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


This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 

to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 

to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 

are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other maiginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 

publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing tliis resource, we liave taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 
We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain fivm automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attributionTht GoogXt "watermark" you see on each file is essential for in forming people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liabili^ can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at |http: //books .google .com/I 

i, Google 

i,XH--+. 57. J 





.„ Google 

i, Google 

D, Google 

i, Google 

,1,1.0, Google 

i=y Google 








THb Olseb Dialects, of Veda and Bbahmana 
Bi William Dwight Wbithbv. 




itizecy Google 





TS13 WOBX ta copyaiiBT. 




mzec^y Google 

3 1-M^ . 57. 2> 


,1,1.0, Google 



It was ia Jnne, 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-Enropean series projected 
t)y Messra. Breitkopf and Hftiiel. After some consideration, 
and consultation with friends, I accepted the task, and have 
since deroted to it what time conld he spared from regular 
dntiea, after the satisfaction of engagements earlier formed. 
If the delay seems a long one, it was nevertheless onavoid- 
ahle; and I would gladly, in the interest of the work itself, 
hare made it still longer. In every snch 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 
%i& — was urgent enough to recommend a speedy com- 
pletion of the work begun. 

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

1. To make a presentation of the facte 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- 

ioy Google 

Ti Preface. 

cesBors; and a traditional method was thos eBtabliBhed whicfa 
has been perhaps Bomewhat too closely adhered to, at the 
ezpeuse of clearness and of proportion, as well as of scien- 
tific truth. Accordingly, my attention has not been directed 
toward a profonoder stndy of the grammatieal science of the 
Hinda schools: their teachings I hare been contented to take 
as already reported to Western learners in the existing 
Western grammars. 

2. To include also in the presentation the forms and 
constructions of the older language, as exhibited in the Veda 
and the Brahmana. Orassmann's excellent Index- Vocabalary 
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 hare sought 
to complete, as far as the circumstances permitted, from the 
other Vedic texts and from the rarions works of the Brah- 
mana period, both printed and manuscript 

3. 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. 

4. To cast all statements, classifications, and bo on, 
into a form consistent with the teachings of lingoistie science. 

• In doing this, it has been necessary to discard a few of tfa^ 
long-used and familiar divisions and terms of Sanskrit gram- 
mar — for example, the olasBiScation 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 facte of internal and ex- 

* It was published, as vol. XII. of the Journal of the Americao 
Oriental Society, in- 1881. 

itizecy Google 

Pkbface yli 

ternal eaphonie combination, and the like. Bnt care hag been 
taken to facilitate the tTan«tion from the old to the new; 
and the changes, it is beliered, will commend themaelves to 
nnqnalified .acceptance. It has been sought also to help an 
appreciation of the character of the language by putting its 
facta as far as possible into a statistical form. In this respect 
the native grammar is especially deficient and misleading. 

Regard has been constantly had to the practical needs 
of the learner of the langnage, 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 Qsual in European San- 
skrit grammars, is, as a matter of course, retained through- 
out; and, because of the difficulty of setting even a small 
Sanskrit type with anything bnt a large European, it is 
practiced alone in the smaller sizes. 

While the treatment of the facts of the langnage has 
thus been made a historical one, within tiie limits of the 
langnage itself, I have not rentured to make it comparative, 
by brining 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 woik, 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 consistentiy throughout. Explanations 
of the origin of forms have also been ayoided, 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 (com other schol- 
ars. I have had at baud always especially the very schol- 
arly and reliable brief summary of Eielhom, the full and 

itizecy Google 

excellent work of Mooier Williams, tbe smaller grammar of 
fiopp [a wonder of learning and method for the time when 
it wae prepared), and the Tolnmes of Benfey and MtUler. 
As regards the material of tbe Ungnage, no other aid, of 
course, has been at all comparable with the great Peters- 
bnrg lexicon of BObtlingk and Roth, tbe existence of which 
gives by itself a new character to all investigations of the 
Sanskrit langnage. 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, neeessarily liable to correction in detail by the results 
of further researches. For what concerns tbe verb, its forms 
and their classification and uses, I have bad, as every one 
must have, by far the most aid firom DelbrUok, in his Alt- 
indisches Verbum and bis various syntactical contribu- 
tions. Former pupils of my own, Professors Avery and 
£^gren, have also helped me, in connection with this 
subject and with others, in a way and measure tiiat 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 Joum. Am. Or. Soc. (print- 
ed contemporaneously with this work, and used by me 
almost, but not quite, to the end of the subjeot) by my 
former pupil Prof. Lanman; my treatment of'it is founded 
on bis. 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 nn- 
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. von Schrttder is due whatever use I have been 

ioy Google 


able to make (anfortnoately a very imperfect one) of the im- 
portant HaitTayani.-Sanifaita. * 

Of the deficiencies of mj mork I am, I think, not less 
folly aware than any critic of it, even the seTerest, is likely 
to be. Sbonld it be fonnd to answer its intended purpose 
well enoQgh to come to another edition, my endeavor will 
be to improye and complete it; and I shall be gratefnl for 
any corrections or snggefltions which may aid me in mak- 
ing it a more efficient help to the stndy of the Sanskrit 
language and literatore. 

GOTHA, July 1879. 

W. D. W. 


TO THE Sbcohd Edition. 

In preparing a new edition of this grammar, I have 
made nse of the new material gathered by myself during 
the intervening years,** and also of that gathered by others, 
80 far as it was accessible to me and fitted into my plan;*** 
and I have had the benefit of kind suggestions f^om varions 
quarters — for all of which I desire to return a grateful 
acknowledgment. By such help, I have been able not only 
to correct and repair certain errors and omissions of the 
first edition, but also to speak with more definiteness upon 

* Since pabliabed in fall by blm, 1881—6. 
** A pxrt of this new material was pabliehed by myBelf in 1885, 
as a Supplement to tbe grammar, under the title "Roots, Verb-Forms, 
and Primary Derivativee of the Sanskrit Language". 

*** Especially deserving of mention is HoltEmann's oollectioD of 
■naterisl liom the Mafaabbarata, also published (1884) in the form of 
a Snpploment to this work; also Btthtlingk's similar collection from 
the larger half of the Ramaya^a. 

Digitizecy Google 

very many points relating to the material and usages of 
the language. 

In order not to impair the applicability of the referen- 
ces already made to the work by rarioiis authors, its para- 
graphing has been retained nnchanged throughout; for in- 
creased cODTenience of further reference, the sabdirisionB 
of paragraphs have been more thoroughly marked, 'by letteni 
(now and then changing a former lettering); and the par- 
agraph-numbers hare been set at the outer instead of the 
inner edge of the upper margin. 

My remoteneBs from the place of publication has for- 
bidden me the reading of more than one proof; bnt the 
kindness of Professor Lanman in adding his revision {ac- 
companied by other timely suggeetious} to mine, and the 
care of the printers, will be found, I trust, to have aided 
in securing a text disfigured by few errors of the press. 

Circumstances beyond my control hare delayed for a 
year or two the completion of this revision, and hare made 
it in some parts lees complete than I should have desired. 

New-Haven, Sept. 1888. 

W. D. W. 

Digitizecy Google 


Bbiep Account op the Indian Litebatuee. 

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

The name "Sanskrit" [saibskpta, 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 grammaiians, 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 ae 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 sac- 
red language of Buddhism in Ceylon and Farther India, and is 

itizecy Google 

Xli iNTROSDOnOtl. 

still ID service theie as such; and jet latet 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 tact, 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 educatedj 
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 treatmoat of the language by European 

Much in the history of the learned movement is still 
obscuTse, and opinions are at variance even as to points of 
prime consequence. Only the concluding works in the devel- 
opment of the gramatical 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 
hare only our inferences and conjectures to lely 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 
(fSkhSs, lit'ly branchet], phonetic and othei; was eealously 
and effectively followed in the Biahmaaic schools; this is 
attested by our possession of a number of phouetico-gram- 
matical treatises, prSti9SkhyBS (pratt fftkhBm belonging to 
each several text), each having for subject one 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- 

Digitizecy Google 

iMTSODnonoH. ziil 

ceneee between the correct speech of the learned and the 
idtered dialects of the vulgar may have borne in the same 
moTement is not easy to determine; but it is not customary 
that a langu^e 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 giammar 
reached a climax in the grammarian P&pini, whose text-book, 
contaioii^ the facts of the language cast into the highly 
Artful and 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 theit 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 HahSbhSahya greul comment, by Fa- 

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 gtam- 
matical rales; yet the existence of grammatical authority, - 
and especially of a single one, deemed infallible and of pre- 
scriptive value, could not tai\ to exert a strong regulative 
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 grad- 
ual effacement of whatever they might contain 'that was 
unapproved. Thus the whole more modern literature of 
India has been Paninised, so to speak, pressed into the 
mould prepared by him and his school. What are the 
limits of tbo artificiality of this process is not yet known. 

Digitiz^cy Google 

xW Intboddoitom. 

The attention of special students of the Hindu giammai 
[aiid the subject is so intricate and difficult that the number 
is exceedingly small of those who hare masteted it suffi- 
ciently to have a competent opinion on such geneial matters) 
has been hitherto mainly diiected towaid detenuining what 
the Sanskiit according to Pacini leally is, toward explaining 
the language from the grammar. And, 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 
aa 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 
9X6 doubtless earlier than Paitini, it is impossible to tell 
just what parts, or how far they have escaped in their style 
the leveling influence of the grammar. The whole, too, 
may be called so fax an artificial literature as it is written 
in a phonetic form (see grammar, 101 a) which never can 
have been a truly vernacular and living one. Nearly all of 
it is metrical: not poetic works only, biit narratives, histories 
(so far as anything deserving that name can he said to exist), 
and scientific treatises of every variety, are done into verse; 
a prose and a prose literature hardly has an existence (the. 
principal exceptions, aside &om the voluminous commen- 
taries, are a few stories, as the Da9akiunKraoarita and the 
VSsavadattSj. Of linguistic history there is n^t to nothing 

Digitizecy Google 


in it all; but only a history of etyle, and this foi the most 
pait ahoVing a gradual depiaration, an inciease of artificiality 
and an intensification of certain mote undesirable features 
of the language — such as the use of passive constiuotions 
and of participles instead of veibs, and the substitution of 
compounds foi sentences. 

This being the condition of the later literature, it is of 
so much the higher consequence that there is an eailiei 
literature, to which the suspicion of artificiality does not - 
attach, or attaches at least only in a minimal degree, which 
has a truly vetnaculai ohaiacter, 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 theii comparatively simple worship. At what period 
these were made and sung cannot be determined vrith 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 elah* 
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-mateiial, mainly its oldest and most 
genuine part, to the extent of over a thousand hymns and ten 
thousand verses, arranged according to traditional authorship 
and to subject and length and metre of hymn: this collection 
is the Blg-VedB Veda of verses (fc) or of hymns. Other 
collections were made also out of the same general mass 
of traditional material: doubtless later, although the inter- 

Digitizecy Google 


lelatioDS of this period aie as jet too unclear to oJlow of 
out speaking with entire confidence aa to anything twncetn- 
ing them. Hius, the SSma-Veda Veda of chanU (Bftmut), 
containing only about a sixth as much, its veises neaily all 
found in the Rig-Veda also, but appearing hete with nume- 
rous differences of reading : these were passages put together 
for chanting at the soma-saorifioes. Again, collections called 
by the compiehensiTe name of Tojur-Veda Veda of aac- 
rificial formuku (yajua) : 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 oeremomes ; tfaey were strictly liturgical collections. Of 
these, there are in existence several texts, which have their 
mutual differences: the Tajasaneyi-Saiiihita (in two slightly 
discordant versions, MKdhyandina and ES^va), sometimes 
also called the White Yajur-Veda; and the various and 
considerably differing texts of the Black Yajur-Veda, namely 
the T&itUrtya-SaihhltS, the USitrBra^J'-SaihhitS, the Eapiq- 
thala-BaihliitS, and the EBtbaka (the two last not yet pub- 
lished). Finally, another historical collection, like the Rig- 
Veda, but made up mainly of later and less accepted 
material, and called [among other less currtot names) the 
Atharva-Veda Veda of the Athareana (a legendary priestly 
family); it is somewhat more than half as bulky as the Rig* 
Veda, and contains a certain amount of material correspond- 
ing to that of the latter, and also a number of brief prose 
passages. To this last collection 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 brfthma^a), 
and in a language which, though distinctly less antique 
than that of the other, is nevertheless truly Vedio. Two 
versions of it are extant, one of them in only 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 

Diaii.zecy Google- 


the imitative woik of a yet mote modern time), is aoattOTed 
thiough the texta to be later described, the BrShmavas and 
the Satr»B. To aasemble aad sift and compare it is now 
one of the pressing needs of Vedic study. 

The fundamental divisions of the Vedic liteiatme here 
mentioned have all bad theit 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 meotiooed 
above are all that are now known to be in existence; and 
the chance of the discovery of others grows every year 

T^e labor of the schools in the conservation of theii 
saoied 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 texts, 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 
efiaced before it, or lost in spite of it, cannot be told. But 
it ia 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 undistotted one, 
and which goes back a good way behind the classical San- 
skrit. Its differences &om the latter the following treatise 
endeavors to show in detail. 

Along with the vetses and sacrificial formulas and 
phrases in the text of the Black Yajur-Veda are given 
long prose sections, in which the ceremonies are described, 
theii meaning and the reason of the details and the accom- 
panying utterances are discussed and explained, illustrative 
l^ends are reported of fabricated, and various speculations, 
etymological and other, are indulged in. Such matter comes 

ioy Google 

zviit Introduction. 

to be called brfthma^a (appaiently relating to the brahman 
ot u>orship). In the White Yajm-Veda, it is separated into 
a woik by itself, beside the sfubhitft oi text of veises and 
foimulas, and is called the patapatba-Brftbrna^a Brahmana 
of a hundred ways. Other sitnilai collections are found, be- 
longing to vaiious othei schools of Vedic study, and they 
beat the common name of BrSbmaijw, with the name of the 
school, or some other distinctive title, prefixed. Thus, the 
Aitarera and EKuQltokl-Brthma^aa, belonging to the schools 
of the Rig-Veda, the PoSoaTiAfa and Sa^TiAfa-Br&hma^a 
and other minor work's, to the Sama-Veda; the Gopatha- 
Brfthmaoa, to the Atbarva-Yeda ; and a JSiminlya- or Tala- 
TakSra-Brfthmaoa, to the Sama-Veda, has recently (Buinell) 
been discovered in India; the TBittiiiya^BTftbrna^a is a col- 
lection of mingled mantra and brBhmaj^a, bke the saihhitS 
of the same name, but supplementary and later. These 
works are likewise regarded as canonical by the schools, 
and are learned by their sectaries with the same extreme care 
which is devoted to the satbhitSa, and their condition of 
textual preservation is of a kindred excellence. To a cer- 
tain 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 Biohma^ias 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 foimd later ap- 
pendices, of a similar character, called Ara^yakao {foreat- 
sections): as the Aitareya- Ara9.yaka , Tfiittlriya-Arapyaka , 
Bfhad -Aranyaka, and so on. And from some of these, or 
even from the Itrrihmauas, are extracted the earliest Upa- 
nitiads [sittittga, lectures on sacred subjects) — which , 

ioy Google 


howeyer, ate continued and added to down to a oompaia- 
lively modein time. The Upanishads are one of &e lines 
by which the Biahma^a literatuie passes over into the latei 
theolf^cal literatuie. 

Another line of transition is shown in the BtttrM [linea, 
rules). The works thus named are analogous with the 
Brahmaj^as in that they belong to the sohoole of Vedic 
study and aie 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 (9r&utB or kalpa-satros] , they take 
up the great sacrificial ceremonies, with which the Brah- 
manas have to do; in part (grhya-sfltras), they teach the 
minor duties of a pious householder; in some oases [sS- 
mayaoSrika-antraa] 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-fSBtraB), 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 by many, of the Manava 
Vedic school); to which are added that of TBjftavalkTa, and 
many others. 

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 lias 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 r^arded as one of the very 
oldest works of the proper Sanskrit literature, if not the 
oldest (it has been variously assigned, to periods from six - 
centuries before Christ to four after Christ). It is so, again, 
in a still mote striking degree, with the great legendary 

ioy Google 


epio of the MahftbliKrata. The giound-woik of this is . 
doubtless of very early date; but it has served as a text 
iato which matetials of vaiious character and period have 
been inwoven, until it has become a heterogeneous mass, 
a kind of cyclopedia for the warriot-caste, hard to separate 
into its constituent parts. The story of Nala, and the phil- 
osophical poem Bhsgavad-QIta, are two of the most noted 
of its episodes. The BOmSyapa, the other most famous epic, 
is a work of another kind; though also worked over and 
more oi less altered in its transmission to oui time, it is 
the production, in the main, of a single author (Valmikij; 
and it is generally believed to be in part allegorical, le- 
presentiog the introduction of Aryan culture and dominion 
into Southern India. By its Bide stand a number of minor 
epics, of various authorship and period, as the BaghuvaA^a 
(ascribed to the dramatist Kalidasa], the Mfigliak&vya, the 
BhattikSvya (the last, written chiefly with the g:rammatical 
intent of illustrating by use as many as possible of the 
numerous formations which, though taught by the gram- 
marians, find no place in the literature)- 

The Purft^as, a large class of works mostly of immense 
extent, ate best mentioned in connection with the epics. 
They are pseudo-historical and prophetic in character, of 
modem flate, and of inferior 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 Ueghadata and Oltogovinda, 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 Fanis, that of Yama and his sister Yami, 
that of Yasishtfaa and the rivers, that of Agni and the other 
gods — but there are no extant intermediaries between these 

Digitizecy Google 

Introduotioh. xii 

and the Btandaid diama. The begioDuigB of the latter date 
&oni a period when in actual life the higher and educated 
characters used Sanskrit, and the lower and uneducated need 
the popular dialects derived from it, the Prakrits ; and their 
dialogue reflects this condition of things. Then, hotrevei 
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 be made indefinitely on a substrate of Sanskrit; and 
none of the existing dramas need to date fiom the time of 
Temacular use of Prakrit, while most or all of them are 
undoubtedly much later. Among the dramatic authors, 
Kalidisa is incomparably the chief, and his ^akuutalft is 
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 Hroohakatikft of 
^udraka, also of questionable period, but believed to be 
the oldest of the extant dramas. 

A partly dramatic character belongs also to the iable, 
in which animals ate represented as acting and speaking. 
The most noted works in this department are the PaSLoa- 
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 Hitopade9a [salutary imtmction). 

Two of the leading departments of Sanskrit scientific 
literature, the 1^^ 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 Brahmai^aa 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 standi n^ in no 

Digitizecy Google 

xzti ItmoDDonOTi. 

accoidanoe with approved teligious doctrines. All of them 
seek the same end, the emancipation of the soul £rom the 
necessity of continuing its existence in a succession of 
bodies, and its unification with the All-soul; but they 
differ in i^^d 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. Theii medical science, although 
its b^nnings 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. 

ioy Google 


Preface v 

iNTBODncnoB zi 

I. Alphabet 1 — 9 

II. Ststeh of Sounds; Fbohdmoiatiok .... lo — 34 
Tovels,>10{ CoDBonauU, 13 ; Qnindty, 27; Accent, 28. 

IIL Rules of Euphonic Cohbihatiom 34 — 87 

iDtrodactory, 34; Principles, 37; Boles of Vowel Com- 
blnaUou, 42; Permitted Flaali, 49; DeaaplratloD, 53; 
Surd »nd Sonant AislmiUtloQ, 64; Combination* of 
Final B and r, 56 ; ConTenlon ot a to ^, 61 ; Con- 
veraion of n to 9, 54 ; CouTOiilon of Dental Mut«t to 
Llngaals and PaUttb, 66; Combinations of Final n, 
69; Combinations of Final m, 71 ; the Palatal Hntes 
and Sibilant, and h, 72; the Lingual Sibilant, 77; 
Estension and Abbreviation, 78 ; StTenftheoini; and 
Weakening Piocessea, 81 ; OOQit and Vpddlllt 81 ; 
Towel-longthenlng, 84; Vowel-llghtenlng, 86; Natal 
Increment, 86; Eedopllcallon, 87, 

IV. Declbmsion 88 — 110 

Gender, Number, Case, 68; Uses of the Cases, 89; 
Endings of DedensloD, 103; Variation of Stem, 107; 
Accent In Deelenslon, 108. 

V. Nouns and Adjeotiveb HI — 176 

Cla«siflcatlon etc., Ill ; Dedeoslon T., Stems in a, 112 ; 
Declension II., Stem* in 1 and u, 116; Declension 
IIL, Stems in Long Vowels (ft, I, Q): A. Soot-words 
etc., 124; Stems in Diphthongs, 130; B. Derfiatite 
Sterna etc., 131; Deelenslon IT., Stems In f or ar, 
137 ; Declension V., Stama in Consonsnts, 141 ; 
A. Boot'Stema etc., 143; B. DeriTatlTe Stems In aa, 
U, OS, 153; C. DerivatiTB Stems in on, 156: D. 
in In, lei; X. tn ant or at, 163; V. Perfect Far- 
tioiplei in viAa, 169; Q. Oomparatives in yUAaor 
yaa, 172; Comparison, 173. 

itizecy Google 

VI. NuMEEALe 177 185 

Csrdlnils, 177; Oidlndi etc, 183. 

VIL Pronouns 185—199 

Penontl, 186; Demoustnttie, 188; Interrogitlva, 
194; Belttin, 195; other Proiioniu; Empliitic, In- 
definite, 196; Noniu nied pranominally, 197; 
Prauomlntl DedTittTCi, PoBseasivw etc., 197; Ad- 
jective! declined ptonomiaallr, 199. 

Vni. CONTOQATIOK 200 226 

Tolee, Tense, Moda, Nnmtei, Penon, 200; Teilwl 
AdJectiTes and Nouns, ^3; Secouduy Conjogatloiu, 
203; Persontl Endings, 204; Snbjnnettie Hoda, 209; 
OpUtive, 211; Imperative, 213; Uses of the Modes, 
215; PsrtlDlples, 220; Augment, 220; RednpIicltloD, 
222; Accent of tlie Verb, 223. 

IX, The Prebbnt-Btstem 227 278 

OeaenI, 227; Conjugations and Vonlngition Classes, 
228; Boot-Class (second or ad-class), 231; Re- 
dnpllcaUng Claaa (third or ha-claBs), 242; Hasal 
Class (seTsntb or rudh-ctase), 2&0; du and n-ClassCB 
(fifth and eighth, oi Bn- and tan<classes), 254; nS- 
Class (ninth or krl-clasE), 260; a-Clasa (Btst oi 
bhG-class), 2D4; Accented A-Claas (sixth or tud- 
cUss), 269; yft-ClasE (fonrth or dlv-dass), 271; 
Accented y&-ClaE9 or PaislTe ConjagatloQ, 275; 
So-called tenth or cnr-cUsa, 277; Dees of the Pres- 
ent and Imperfect, 278. 

X. The Peepect-System 279 — 296 

Perfect Tense, 279; Perfect Participle, 291 ; Modea 
of the Perfect, 292; Pluperfect, 295; Uses of the 
Perfect, 295. 

XL The Aorist-S^htemb 297 — 330 

ClMsiDcatioD, 297; I. Simple Aorist: 1. Root-Aorlst, 
299; Passive Aorlst 3d sing., 304; 2. the a-Aorist, 
305; n. 3. Reduplicated AoHst, 308; III. Sibilant 
Aorlst, 313; 4. the B-Aoriat, 314; 5. the Ij-Aorlsl, 
320; 6. the Bi(-Aorist, 323; 7. the sa-Aoiist, 32S; 
Piectttive, 320; Uses of the Aorist, 328. 

XII. The Futurb-Systems . ." 330—339 

I. The s-Fntare, 331 ; Preterit of the B-Foture, Con- 
ditional, 334; H. The Periphrastic Fature, 335; 
Usea of the Fatntes aud Condittonat, 337. 

ioy Google 

Xni. Vbrbal adjeotites and Nounb: Partici- 
ples, Ihsthitivbs, Gerunds 340 — 360 

PtulTB Ptfttolpte In t& or a&, 340; Put AcUtb 
Participle In tavant, 344; Fntnie Paetlve Putl- 
ciplei, Geranillvea, 310; loSnitlveB, 347; Uses of 
the Infliiltlres, 351; Qemnds, 365; AdverbUl Oemnd 
Id am. 369. 

XIV. Dbritatite or Secondary Conjuoatioh . . 360 — 391 

I. Pisstie, 361; II. InCenBlve, 362; Present-Syitem, 
866; Perlect, AortM, Fntnre, ate,, 370; III. Desider- 
atiTS, 372; Pieasot- System, 374; Perfect, Aarlst, 
Falnre, etc., 370; IV. Otueitive, 376; PieBent-Syatem, 
380; Perfect, Aeiiit, Futare, etc., 383; V. Denom- 
InatlTe, 386. 

XV- Periphrastic akd Compound Cokjuoation 391 — 403 
The Petiphrutic Perfect, 3fi2; PutlcipUl Perlphiu- 
tic Phruei, 391; Compoiltion wltb Frepoeltlonsi 
PieBiee, 396; Other Verbal Camponndi, 400. 

XVI. Indbcunables 403 — 417 

Adverbe, 403; PrepoiitiDDi, 114; ConJnactioDg, 416; 
iQtaijectloni, 117. 

XVII. Derivation of Deolinable Stems 418 — 480 

A. Primary DeriTaUvea, 420; B. Secondary DeriTa- 
tlrea, 154. 

XVITT. Formation OP Compound Stems 480 — 51S 

Claa^lRcatiou, 180; I. CopalatlTe Conpounda, 435; 

II. Determlnatiie Companndg, 489; A. Dependent 
ComponDda, 480; B. Descitptlve Componrtde, 101 j 

III. Secondary Adjectlre Componnda, 501 ; A. Poa- 
neasite Componnda, 501 ; B. Oompounda with GoTemed 
Final Member, 511; AdJecHye Compounda aa Nonna 
and as Adverbs, 512; Anomalous Componnda 511; 
Stem-llnala altered In Compoaitlon, 514; Looae 
Constmctton with Compounds, 615. 

Appendix 516—520 

A. Bxamplea of Various Sanikrll Tfpe, 516; B. Ex- 
ample of Acoantnated Teit, 516; Synoptia of the 
conjugation of roota bhu and if, 520. 

Sanskrit-Index 521 — 539 

GbnbraIi-Index . 540 55i 

ioy Google 


AA. Aitareya-Aranykka. 

AB. Aitareya-Brihrnana. 
AQS, A(va1IyaDa-^iauU-8iltia. 
AQS. AQvalayana-Qrhya-Sutra. 
Apast. Apastamba-SQtra. 
APr. Atharva-Pratiijakhya. 
AV. Atharva-Veda. 

B. or Br. Brihioanas. 

BAU. Brhad-Aranyaka-Upanisad. 
BbG. Bhagavad-Gita. 
BbP. Bhagavata-PurSns. 
BB. BtlhtliQgk and Roth (Peters- 
burg Lexicon). 

C. Claesical Sanskrit 
Q. 9aknataU. 

9a tr. 9'i^uit>J&ja-U<'hatDijaiii. 
QB. Qatapatha-Brahmana. 
99s. (ankhlyana-Qraata-Satra. 
gaij. Qankhayana-Grbya-SQtra. 
ChU. Chindogya-npantjad. 
9vU. Qvetafvatara-UpaQlsad. 
DKC. Data-Kumara-CariU. 
E. Epos (HBb. aud R.). 
OB. Gopatha-Brabma^a. 
GQS. Gobhinyu-Grhya-Sfitra. 
H. Hilopadeca. 
Har. HarivanQa. 

JUB. Jiimlniya - UpanUad - Brah- 

K. Eathaka. 

K19. EapiBthala-SambitL 
KB. K&a;itaki- (or QtmUtiiyKiia-) 

KBU. EaiiBitakl-Brihmana-Upaai- 

K9S. Katyayana-Qranta-Sutra. 
KS. Ein^ika-Sutra. 
ESS. Eatha-3arit-Sigara. 
KthU. Ka(ha Upanl^. 

EU. EenB-0paniRad. 

LQS. Latyayana-Qranta'Sutra. 

H. Hann. 

HaiU. H^tri-UpaDiaad. 

HBh. Habibharata. 

MdU. Hondaka-Upanifad. 

Hegh. HeghadQta. 

KS. Hsitrayani-Samhiti. 

Nais, Nai|adhiya. 

Xir.' Nimkta. 

PaSc. PaBcatantra. 

PB. PaBcaviiiQa-[or Tiindya-J Br^- 

POS. Piraskara-Grhya-Siitra. 
PD. Pra^na Upaniaad- 
R. Bimsyana. 
Ragh. BsgbiiTad;H. 
RPr. Bigveda-PrattQ&khya, 
RT. RiJa-TaraiigliiT. 
EV. Rig- Veda. 
S. Sutras. 

.SB. ^adTibfa-Brlbmana. 
Spr. lodische SprUcfae (BUhtliQgk). 
3V. Sama-Veda. 
TA. Taittliiya-Aranyaka. 
TB. TUttirlya-Briibmana. 
TPr. Taittiriya-Praticakbya. 
Tribb. Trtbhaayaratna {comm. Ui 

TS. Taittiriya-Sambita. 
U. Upauisads. 
V. Vedaa (RV., AV., SV). 
Vas. Vaaistha. 
VBS. Varaha-Brhat-Sambita. 
Vet. VetalapafScaviiiffati. 
Vikr. VikramorvaQi 
VPr, Vajasaneyi-Pratisakhya. 
V3. Vajaseneyi'Sanibita. 
V3. Kan. do. KS^va-text 
V. TftjiSatalkya. 

Dgil.zecy Google 


1. Tub natives of India write theii ancient and sacied 
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 devanftgarl. 

A, This name ti of donbtTnl oitgin and value. A more compiehenslve 
name l» nfigan (perhaps, of the city'); and deva-n^ari U nagorl of 
iAt godi, or of th* Brahmant. 

S. Hueh that relatai to the hiitoty of the ladlin alphabete ii still 
obKOTO. The earliest written monDineiits of known date In tbe conntrf are 
the inaoriptiona containing tbe edlatg of A9oka or Piyadaal, of about the 
middle of tbe third century B. G. They are In two dllTeient syitems of 
cbanclerfi, of which one ihowa diatinet eigue of dedTition from a Semitic 
tonrce, vhile the other ia alao probably, though mntb leaa evldentlr, of the 
same origin. From the latter, the Lath, or Sonthsm A^oka character (of 
Oiraai), come the later Inilian alphabets, both those of the northern Aryan 
lanfoages and thoie at the sonthem Draridian langoagea. The nftgon, 
devanAeart, Bengali, Quzeriti, and others, are farieties of Its northern 
deriretiTes; and with them are related some of the alphabets of peoples 
outside of India — as In TilMt and Farther India — who have adopted Bindn 
enltare or religion, 

a. There Is reason ia believe that writing was first employed in India 
for pra^lical parpoees — for correspoodeiice aad buslneaa and the like — 
and only hy degrees came to be applied also to literary ase. The Itteratore, 
to a great extent, and the more fully in proportion to its claimed sanctity 
and authority, ignores all written record, and usomes to he kept In eiistence 
by oral tradition alone. 

Whitnsy, OTanmir. 3. «d. 1 

ioy Google 

9. Of the deranfigAri iUelf there are nlnoT varieties, depending on 
iineieatat of locality or of period, as aUo ot indlvidoal hand [sea eiamples 
In Weber'* catalogae of the Beilln Sanekrit HSS., in RajendialiU MKra's 
notlcei of HfiS. In IiidUn HbTaiiei, In the publlbhed fac-^imiles of in- 
«cT[p(loni, and ao on); and these aie In some measure leSected in the type 
prepared for printing, both in India and in Europe, But a itudent who 
maliea himaelf fimliUr with one style of printed chambers will have little 
diniculty nitb the others, and wilt loon learn, by practice, to read the mann- 
scrlpta. A few specimens of types other thin those uaed in thl« work are 
gUvn In Appendix A. 

a. On aci^oitnt of the dlfflculty of combiriing them nlth the smaller sites 
of onr Iloman a[id Italic typo, the devanSgari characters are need below only 
In ronnectliui with the Brst or largest aiie. And, in srcordance with the 
loudibln ntSKa Of teient gramtnari, they are, wherever given, also trans' 
literated, In Clarendon Ictteri ; while the latter alone are D*ed in the other 

1. The Btudcnt may be advised to try to ramiliariie himself from 
the Btart with the deTuiftgan mode of writing. At the same time, 
it U not indiBpensable that he should do bo until, having learned the 
principal paradigms, be comes to begin reading and analysing and 
parsing; and many wilt find the latter the more practical, and in the 
end equally or more effective, way. 

6. The chaiacteis of the devanKgarl alphabet, and the 
European letters which will be used in transliteratiog them, 
ate as follows: 

Hliurt long 

I . q a = HT a 


' ? ' 

' t - 

owela: simple ! labial 

. 3 u 


' B I 

■ ^' r 


• 3 1 

l» ^ 11 

"'"■"■■»»• {r:r :ii 

.. : It 

ra " ^, --. 6 or ifa (Bee 73c,'. 

•Did surd up. 


nn. up. uml 

guttural .' 3R k i. p^ kh 


- q 8h a J a 

palatal i ^ c » ^ oh 

-St J 

n ^ Jh - 31 a 

liugual . r ( > 5 f 

»i <* 

. E ill .1 m » 

dental « tT t « St Oi 

"? a 

«^ dh -q n 

labial «^ p .- T; ph 

.ST b 

->I bt .. H lU 

Diaiiize. ..Google 

Tbeory op THia Mode of Writing. 


»a y 


"T ' 



..g V 


-a ? 


"U » 


-H a 

Aspiration •• ^ h 

a. To thcBe mny be nddcd n. lingual 1 X, which in some of the 
Vedic textB takes the place of 3 4 when occurring between two 
vowels (M). 

8. A few other Boun<le, recogoized by the theories of the Eindu 
frrammaritin?, but either having no acpnrate characters to represent 
them or only very rarely and exceptionally written, will be notlceil 
lielow |T1 b, a, 930). Such are the guttural and labial breatbinga, the 
nasal gemivowols, and others. 

7. The order of aiTangement given above is that in 
which the sounds are catal<^iied 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. 

A, In eomi worka (ae the Petenburg leiicon), i vlaargft whteh Is le- 
gtided u eqalvilent to and «xchang»ibte Tjtb > ilbtUnt (17S) it, though 
written » vlsarga, given tha ilpbabetic pli(e of the sibilant. 

6. The theory of the devanSgivrl, as of the other Indian 
modes of writing, is syllabic and consonantal. That is 
to say, it regards as the written unit, not the simple sound, 
but the syllable (ak^ra); and further, as the substantial 
part of the syllable, the consonant or the consonants which 
precede the vowel — this latter being merely implied, or, 
if written, being written by a subordinate sign attached to 
the consonant. 

9. Hence follow these two principles: 

A. The forms of the vowel-characters given in the 
alphabetical scheme above ate used only when the vowel 


9—] I. Alphabet. 4 

forms a syllable by itself, oi ia not combined with a pieceding 
cODBoiiaat: that is, when it is either initial or preceded by 
another vowel. In combination with a consonant, other modes 
of representation are used. 

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

a. NntWe Sinda usftge, id mftonBcriptH and iQSoriptions, treats 
ihe whole materinl of a sentence alike, not separating its words from 
one another, any more than the aylUblee of the same word: a final 
consonant is combined into one written syllable with the initial vowel 
or consonant or consonants of the following word. It never occnned 
to the Hindus to space their words io any way, even whero the mode 
of writing admitted such treatment; nor to begin a paragraph on a 
new line; nor to write one line of verse under another: everything, 
without exception, is written solid by them, filling the whole page. 

b. Thus, the sentence and verge-line ahaifa rudrebhlr vamibliiQ 
carSmy aham ftdltyftlr uta vigvadev&i^ (Big-Veda X. 115. 1: see 
Appendix B) 1 icander loith the Vtuut, the Rudraa, I with the Aditya* 
and the AU-Oode is thus syllabized; a hadi ru dre bhl rva Bubhi 
90a rS mya ha m& di ty6i rn ta vi Qva de vfti^, each syllable end- 
ing with a vowel (or a vowel modilicd by the nasal-sign amtavftra, 
or having the sign of a final breathing, viearga, added: these being 
the only elements that can follow a vowel in the same syllable); and 
it is (together with the next line) written In the manuscripts after this 

£ach syllable is written separately, and by many scribes the 
successive syllables are parted a little from one another: thus, 

and flo on. 

o. In Wiiatern practice, however, it is almost universally customary 
to divide paragraphs, to make the lines of verae follow one another, 
and also to separate the words so far as this can be done without 
changing the mode of writing them See Appendix B, where the verse 
here given is so treated, 

A. Further, in works prepared fo\ beginners in the language, it 
is not uncommon to maku a more complete separation of words by a 


5 Writimq op Vowels. [—10 

free nse of the Tir&mtt-Biga (11) Qnder finnl conHonuts: thus, for 

or even by Indicating nlao tbe combiDations of initial and final vowela 
(126, 137): for example, 

e. In traneiitentiDg, WeBteni methods of separation of words are 
of course to be followed; to do otbervise wonld be simple pedantry. 

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

a. The short Q a has no written sign at all; the con- 
sonant-sign itself implies a following Q a, unless some other 
vowel-sign is attached to it [or else the Tirftma: 11). Thus, 
the consonant-signs as given above in the alphabetic scheme 
are leally the signs of the syllables ka, kha, etc. etc. (to lia). 

b. The long ^ ft is written by a peipendiculai stroke 
after tbe consonant: thus, ^ kS, Kl dJi&, ^ hft. 

c. Short ^ i and long ^ I are written by a similar stroke, 
which for short i is placed before the consonant aud for 
long I is placed after it, and in either case is connected with 
the consonant by a hook above the upper line : thus, Hi kl, 
^ kl; ^ bhi, >f^ bbl; ^ nl, r) nl. 

Tbe book above, turning to tlie left or to the tight, ie biatorlcally the 
euentiU part of the cbuscter, h>Tliig been otlglDsUy the whole of it; the 
hoolu were only later prolongad, lO u to leuh all the way down beside 
the conBonsnt. Id tbe USS,, Ehey almost never bare the horizontal stroke 
drawn across them ibove, though thij ii added in the prloled characters; 
thus, originally *% ki, ^ M; in the MSS., 1%, ^ ; in print, fe, ^, 

d. The u-Bounds, short and long, are written by hooks 
attached to the lower end of the consonant-sign: thus, ^ 
kn, ^ kll ; ? ^m, 7 4IL On account of the necessities of 
combiBatioQ, du and dll are somewhat disguised: thus, Xi 
g^; and the forms with ^ r and ^ k are still more irregular: 
thus, "^ ru, 5 rfl; ^ hu, "^ kO. 

Digitizecy Google 

to—] I. Alfhabst. 6 

e. The f-voweU, short and long, are written by a aub- 
joined hook, single oi double, openiag toward the right: 
thus, ^ kr, ^ ^i ? dfi ^ <^' In ttie b-fiign, the hooks 
aie usually attached to the middle: thus, ^ hf, ^ hf , 

Ai to the cambliitllDii ot f irltti preceding r, see below, 14 d. 

f. The }-Towel is written with a reduced form of its 
full initial character: thus, ^il; the coTiespondiog 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 ^ o and ^ Su, 
with the K-sign after the consonant: thus, % ke, % k&i; 
^ ko, ^ kSu. 

b. In lome deTanOgan mannsciipts (bb in (lie Bengill tlphibet), iha 
Biiigle Bitoke abov«, or one of the double ones, It repUced bj i sigu like the 
a-(jgn befote Iha conioiunt: thus, (^ fc«, Ri kU; |^ ko, (^ kAo. 

11. A consonant-sign, howevei, is capable of being made 
to signify the consonant-sound alone, without an added vowel, 
by having written beneath it a stroke called the vir&ma 
[rest, atopy, thus, ^ k. ^ d, ^ h. 

0. Since, u wai poiuted out ibove, the Uindiu write the words ot a 
sentence continnonsly like cue woid (Sa, 1>), the vlr&ma ii in geuenl uiled 
foi only when ■ Biul consonant occora before a pause. But It ia alio oc- 
caalonall)' reaoited to by acribea, di Id pilot, in order to BToid an awkward 
or difficult combination of coneonant-eigiis: tbna, 

^Ilft; liijbhl^ %^ lijau, Sl^fq afik?va; 
and It Ib used to make a aeparation of words in texts prepared far begin- 
«eri (Bd). 

12. Under B, it is to be noticed that the consonant 
combinations are for the most part not at all difficult to 
make or to recognise for one who is familiar with the 
simple signs. The characteristic part of a consouant-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 t<^ethet according to convenience, 

ioy Google 

7 Combinations of Consonants. [—-14 

either side by side, or one above the othet; ia a few oom- 
biaations either atrangemect ia 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. 

a. Examples of the side-by-side anangement are: nr gga, 
^ jja, nf pya, nr nma, f^ ttlia, y^ bbys, 7°fi ska, ^ q^a, 
r^ tka. 

b. Examples of the abov&-and-below anangement are: 
^ kka, BT kva, '^ coa, ^ iija, ^ dda, H pta, ^ tna, 
W tva. 

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

a. Of aR k in ^ kta, ^ kla; and in ofTS ki^a etc. 

b. Of ?T t in fT tta; 

o. Of ^ d in ?; dga, T daa, etc.; 

d. Of If m and 7 7, when following other consonants: 
thus, ^ kya, Wf kma, ^ fima, ^ Oya, ^ dma, t^ dya, 
^ hma, ^ hya, ^ ohya, ^ <lii;a. 

e. Of EFT 9, which generally becomes 7T when followed 
by a consonant: thus, ^ Qoa, ff ^na, ? ^va, S7T 9ya. The 
same change is usual when a vowel-sign is added below; 
thus, 5 9U, !>[ 9F, 

f. Other combinations, of not quite obvious value, are 
^ wa, ^ Ua, ^ ddba, ? dbha, '? ^fa, "^ 9tiia; and the 
compounds of ^ b: as ^ h^a, ^ hna. 

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

14. The semivowel ^ r, in making combinations with 
other consonants, is treated in a wholly peculiar manner, 
analogous with that in which the vowels are treated. 

a. If pronounced before another consonant or combination 
of consonants, it is written above the latter, with a hook 


14—] L Alfbabbt. 8 

opening to the n^t (mnch like the ngn of the rowel j, 
as writtea under a eonsoiunt: lOej: thus, ^ rka, ^ rya, 
3 rtva, T^ pmya, rR rt«n^ 

b. Then, if a consonanlrgioup thus containing r as 
first member is followed by a vowel that has its sign, or a 
part of its sign, or its sign of nasality lanusrtca: 70, 71), 
written above the line, the r-sign ia placed furthest to the 
right: thus, ifi rke, $ rka^ ^ rki, ^iki, ^ rko, S^rfelA, 
■^ rkofi. 

o. If r is pronounced after another consonant, whether 
before a vowel or before yet another consonant, it is written 
with a straight stroke below, slanting to the left: thus, 
TT pra, H dhza, H gra, S ara, -f ddhfa, ^T ntra, XZ( grya, 
R srva, ^ ntrya; and, with modifications of a preceding 
cODSouant-sign like those noted above (13], If tra, ^ dra, 
'm ijTtk, ^ bra. 

d. When ^ r is to be combined with a following ^ f, 
it is the vowel which is written in full, with its initial 
character, and the consonant in subordination to it: thus, 

16. Further oombinationa, of three, or four, or even 
five oonsonant-signs, are made according to the same rules. 
Examples are: 

of three consonants, W ^tva, 3T ddhya, &] dvya, ^ 
drya, ^ dhrya, c^ psva, ^ 9oya, ^ 9th;a, fIT hvya; 

of four consonants, ^ ktrya, ^ fik^ya, ^ 9trya, 
rPTT tamya; 

of five consonants, f^ rtsnya. 
a. The mtnuEcriptfl, and the type-fonts u well, differ fiom one knother 
more )n tbelr mintgeiaent of conionint combioillona than In any other respeet, 
often havlnf peruUrittea vhlch one needs > little piacllce to anderatand. It 
1b quite useless to give in a grammar the whole aertes Of potaible combioatlont 
(aome of them ezcesslvely rare) irhlch are pioTlded for In any given type- 
font, or even In all. There ia nothing which due funiliarity with the simple 

ioy Google 

9 Various Siohs. [—18 

rigns tnd wttb tb« ibove rulei of combinttloQ will not enibla the atadent 
leadlly to antlyBB tad eiplsio. 

16. a. A sign called the avagraha [separator] — namely 
>r — is occasionally used in the manuscripts, sometimes in 
the manner of a hyphen, sometimes as a mark of hiatus, 
sometimes to mark the elision of initial ^ a aftei final ^ e 
OT ^ o (136). In piittted texts, especially European, it is 
ordinarily applied to the use last mentioned, and to that 
alone: thus, ^ >I5R^ te 'bruvan, W -I^RtrT ao 'bravlt, for t© 
abravan, so abravlt. 

b. If the elided initial-vowel is nasal, and has the anu- 
av&ra-flign (70, 71) written above, this is usually and mote 
properly transferred to the eliding vowel; but sometimes it 
is written instead over the avagraha-sign : thus, for bo 'b^umKn, 
&om so aJi9umSn, either W 'IJ'IRH or Ht 'ImHH 

0. The sign ° is used in place of something that is 
omitted, and to be understood from the connection: thus, 
cjh^^ft^ o^ d^ vlrasenaantaa -tarn -tena. 

d. Signs of punctuation are I and ii. 

At the eod of a verse, a paragraph, or the like, the Utter of 
them is ordinarily written twice, with the figure of eoiinieratioD 
between: thus, il 1^0 |i. 

17. The numeral figures are 

^ 1. ^ '^ 5 3. 6 <. H 6, ^8, ti 7, US, 5 fl, 0. 

In combination, to express larger numbers, they are 
used in precisely the same way as European digits: thus, 
5_»t 26, ^0 630, bOOO 7000, ^TT^ 1896. 

IB. The Hlndn grammarlana call the different sonndH, and the 
characters representing them, by a k&ra Ifttaktr) added to the sound 
of the letter, if a vowel, or to the letter followed by a, if a consonant. 
Thus, the sonnd or character a is called akSra; k is kak&ra; and 
80 on. But the kfira is also omitted, and a, ka, etc. aro used alone. 
The T, however, is not called rakSra, but only ra, or repba marl: 
the sole example of a speoific name for an alphabetic element of its 
class. The anasvXra and viearga are also known by these names alone. 

Digitizecy Google 

11- System op Sounds. 



I. Vowels. 

19. The a, 1, and u-vowels. The Sanskrit has these 
thiee eailiest and most universal Towele of Indo-Eutopean 
language, in both short and long foim — Q a and 5T 5, 
^ i and ^ I, 3 ti and ^ U. They aie to be pronounced in 
the "Continental" or 'Italian" manner — as in far oi farther, 
pin and pique, pull and rule. 

30. The a is the openeBt vowel, ao utterance from the expanded 
throat, stands in do relation of kindred with anj' of the classes of 
coDHonaotal soniidB, and has no correap ending semivowel. Of the 
close vowels 1 and u, on the other hand, 1 is palatal, and shades 
through its semivowel y into the palalal and gnttnral consonant- 
clusses; u is similarly related, throngh ils semivowel v, to the labial 
class, as involving in its utterance a narrowing and rouoding of 
the lips. 

a. the Panine&n soheme (commehury to Piii)liii'B grsmmu i. 1. 9) 
oluEes a IB guttural, but sppireiitif only Izi older to giv« that series as 
veil BS the rest a Yovrel ; no one of the PTitl9akby)is puts a Into one clafs 
with k etc, AU tbeee anthoritleB dticni In Mlling the 1- and a-vowels 
reipeetlTely palatal and labial. 

SI. The short a is not pronounced in India with the full openness 
of &, as its corresponding short, but usually as the "nentral vowel" 
[English so-called "abort u", of hut, son, blood, etc.). This peculiarity 
appears very early, being acknowledged by Paijini and by two of the 
Pratieakhyas iAPr. i. 36; VPr. 1. 72), which call the utterance Bariivpta, 
covered up, dimmed. It is wont to be ignored by Western BCholars, 
except those who have stadied in India. 

32. The a-vowelsare the prevailing vowel-sounds of the language, 
being about twice as frequent as all the others (includiug diphthongs] 
taken together. The 1-vowela, again, ate about twice as Quoierous 
as the a-vowels. And, in each pair, the short vowel la more than 
livice PVa to 3 times) as common as the long. 

Diail.zecy Google 

1 1 Vowels. [—27 

a. Fat mote preeUe ettlinatea or frsqaeocy, or these ind of Ihe othet 
■Ipbibetic elementB, and Tor the wijr Id which tbey were obtained, teo 
belov, 76, 

28, The T- and l-vowels. To the three simple vowels 
already meotioDed the SaaslcTit adds two otheis, the r-vowel 
and the l-vowel, plainly generated by the abbreviation of 
ayllablee containing respectively ^ ^ ^ °' vl ^ along with 
another vowe): the ^ r coming almost always (see 237, 241-3] 
&om W^ ar or ^ ra, the rT I from lEFT al. 

a. Some ot the Blndit gratnoiKrluis idd to Ihe tipbabet also a long ); 
but this ta only for Ihe aake of an ittlflclal eymmetry, niace the aoond doet 
not occur to a single genuine word In the Imgnage. 

24. The rowel ^ r is simply a smooth or untrilled 
r-sound, assuming a vocalic office in ayllable-making — as, 
by a like abbreviation, it has done also in certain Slavonic 
languages. The vowel ^ I is aa /-sound similarly uttered 
— like the English ^vowel in such words as able, angle, 

a. The modern Hindus prononoce these vowels as n', ri, li [or 
even In), having long lost the iiabit and the facility of giving a vowel 
value to the pure r- and I-sounds. Their example is widely followed 
by European scholars; and hence alao the (distottiDg and altogether 
objectionable) ttansliterationa pi, fl, li. There is no real difficulty in 
tbe way of acquiring and practising the true utterauce. 

b. Some of the gtammirians (see APt. i. 37, note} attempt to deflne mOTe 
nearly the way In which, in these vowels, a real r- oi ^element ta combined 
with something el«e. 

26. Like their corresponding semiTowela, r and I, these vowels 
belong respectively to the geneinl lingual and dental clasaes; the 
euphonic inflaeace of j and f [1B8| shows this clearly. They are 
BO ranked in the Paninean soheme ; but the Pr&ti^akbyas in general 
strangely class them with the Jibvlmallra sounds, out "gutturals" (38). 

29. The short \ is found in evety variety of word and of position, 
and is not rare, being just about as frequent as longfL Long f is very 
much more unusual, occurring only in certain plural cases of nouu- 
steras in f (371b, d. 876). The \ is met with only in some of the 
forma and derivatives of a single not very common verbal root (k)p). 

27. The diphthongs. Of the four diphthongs, two, 
the V e and W o, are in great pait original Indo-European 


87—] II. System of Sounds. 12 

Bounds. In the Sanskrit, they vrear the aspect of being 
products of the increment oi strengthening of ^ i and 7 u 
respectively; and they are called the corresponding gxt^a- 
vowels to the latter (see below, 236 ff.). The other two, ^ U 
and w Su, are held to be of peculiar Sanskrit growth; they 
are also in general results of another and higher increment 
of ^ i and 3 n, to which they are called the corresponding 
TfrddM-TOwels (below, 2S5S.]. But all are likewise some- 
times generated by euphonic combination (127); and ^ o, 
especially, is common as result of the alteration of a final 
qff as (17B). 

28. The ^ e and ^ o are, both in India and in Europe, 
usually pronounced as they are transliterated — that is, as 
long e- (English "long a", or e in tJici/) and o-sounds, without 
diphthongal character. 

a. Such they appsreotly already were to the aathurs of the 
Praticakhjas, which, while tanking them as diphthongs (soifadlijak^ant), 
give rules reBpecting their pronunciation in a manner implying them 
to be TiTtually unitary sounds. But their euphonic treatmeat (131-4) 
clearly shows them to have been still at the period when the euphonic 
laws established themselveB, as they of course . were at their origin, 
real diphthongs, ai [a + ■) and au [a + u). From them, on the same 
evidence, the heavier or vfddbi dlphthongB were dlBtiagnished by the 
length of their n-element, as at {a + ■') and au (a + u). 

b. Tbe ttcosnluble dtstiDCtnsBS of the two elements tn tlk« v^ddbl- 
diphthongB Ib aollced bj the Frat^ikhyu (eea APi. f. 40, note); bnt the 
lelatloD of thote elements U either defined m equal, or the a la made of 
leu qnintltf than the t and u. 

26. The lighter or gu^-diphthonge are much more frequent 
(6 or T timee] than the heavier or v^ddU-diphthongB, and the e and 
ftl than the o and ftu (a half more]. Both pairs arc somewhat more 
than half as common as the simple t- and u-vowels, 

30. Tbe general name given y tbe Hindu grtrnmaiiana to the foweU 
Is avara tone; the simple vowels are called samfttUUc^ara homogentom 
tyliable, and the diphthong! are called saifadliTak^ara comhitiation-syllabh. 
The position of the organs in their nttetsnce Is deSned to be one of openness, 
01 of non-cloBUTe. 

«. As to quantity and accent, see below, 7611., 80 ff. 

ioy Google 

II. Consonants. 

SI. The Hiodu name for 'coDsonant' is vralkjuia numi/ttttr. 
Tbe consonants are divided by the grammariana into Bparqa eontaet 
or mnte, anta^thft, xntermediale or aemivowel, and ^mon spirant. 
Thej will here be taken np and described jo tbia order. 

83. Hntes. The mntea, sparQa, are so called aa involving a 
complete closure or contact (sparfn), and not an approiimatioa only, 
of the month-organs by which thoy are prodnced. They are divided 
into five chases or series (varga), according to the organs and parts 
of organs by which the contact is made; aod each series is composed 
of live members, differing according to the sccompanimenis of the 

33. The five mute-setiea 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 itom point to point, and ending with the fiontmost 

34. In each series there are two eurd members, two 
sonant, and one nasal [which is also sonant) : for example, 
in the labial series, ^ p and m ph, ^ b and ^ bh, and ^ m. 

a. The meruberi are by the Hindu giammturtins called Tespectiiely^^rri, 
leeond, third, fourth, and last oi Jifth. 

b. The surd consonanU aie known aa oghoqa tonelat, and the aonanta 
aa gtao^vant kaving tone; and the dedcrLptlons of the grammarlana ate In 
accordance with these taroiB. All alike recognlee a dKrcrence of tone, and not 
in any manner • dllTaience oT force, vbetber of contact or of eipalsion, aa 
aeparatirig the tiro great classea in question. That the difference depends on 
Tlv&ra opening, or aaifav&ra ch$ur« {of the gloltie), ie alio recognised 
by them. 

86. The first and third members of each series are the 
ordinary corresponding surd and sonant mutes of European 
languages: thus, m k and TJ g, rT t and ^ d, ti p and ? b. 

se. Nor is the character of the nasal any more doubtful. 
What I? m is to TT P, and ^ b, or =1 n to rT t and 5" "^i that 
is also each other na«al to its own series of mutes : a sonant 
expulsion into and through the nose, while the mouth- 
organs are in the mute-contact. 

ioy Google 

ae-] n. SrSTBM ov Sounds. ] 4 

a. The HindD gramiiuriana £!*■ diitlni-tlf this dcSnitiOD. The niul 
(animSaika patting through the note) lotiiidi ue decliied lo be foimed by 
month and nose tofetlieT; dt their niulilj (antmSflil^tt) to be gireo them 

If nndosme of the now. 

37. The secood and fourth of each series are aspirates: 
thus, beside the surd mute ^ k we have the corresponding 
surd aspirate ^ kh, and beside the sonant T\ g, the corres- 
ponding sonant aspirate ^ gh. Of these, the precise char- 
acter is more obscure and difficult to determine. 

a. Thit the upiretes, ill of tbem, ire leil mates or oontict lonndg, md 
not frieiUveg (like Enrop^in th ind ph lod eh, et«.), U beyond qaettion. 

b. It i* *Uo not doubtful Id whit way the said th, fOT eiimple, dilTere 
from the nniBpinted t: laeb aspiiales ire found in mmy Aelitic Imgaages, 
and even in some !E[; they involve the slipping-out of in indible bit 
0^ Jlattu 01 aspintion belneen the bieacb of mate-elosuTe ind the following 
sound, whsteyef it m»y be. They ite anrnritely enough lepresenled by the 
tb etc., vith vhirh, in Imitation of the Latin treitment of the similu indent 
Greek aspirates, we are accustomed to write them. 

o. The sonant aapintea are generally uadentood ind described as made 
In 1 eimilir way, with » peri;eptible A-sound after the breaeli of sonant mnte- 
closnre But there are great theoTeticil difflcuItieB in the wiy of ai^cepting 
thii GiplanitioD ; md some of the best phonetic obierrers deny that the modem 
Hinilu pronnnciation is of inch 1 cbaricler, and dellne the element following 
the mate as a ''glottal bazx", rithei, or u m emphuized ntterince of the 
beginning of tho succeeding sound. The question ii one of great difficulty, 
and upon it the opinions of the highest authorities are much at yirianee. 
Sonant aipititrs are still in use in India, in the pronnnciation of the Terniculir 
as well as of the leimed languages. 

d. By the Priitltalihyis, the afpitttes of both classes are called BOfman: 
which might mean either accompanied hy a rtuh of breath (taking Qfman 
In its more etymological sense), or accompanied hy a tpiranl (b^low, 6B). 
And some native anthorltles define the surd ispirstes is mide by the combi- 
nation of each sard non-aspirate with its own eoneiponding surd spirant; and 
the sonant aspiTites, of each sonant non-isplnte with the sonant spirant, the 
Il-Bound (below, 66). Bat Ibis would make the two claiees of aspjrites of 
qoite diverse character, and would also mike th the same as ts, ^ as ff , oh 
as CQ — which is In any measure pliasible only of the last. Pi^ni baa no 
name for aspirates ; the scheme given in bis comment (to i. 1. 9) attributes 
to them mahapril^ great expiration, and to the non-asplratcs alpaprK^a 
tmali expiration. 

e. It is usual among European scholars to pronounce 
both classes of aspirates as the corresponding non-aspirates 

Digitizecy Google 

15 GuTTrRAL AND Palatal Mutes. [—42 

with a following h: for example, EI th nearly as in English 
boathook, ^ ph as in haphazard, tT dh as in madhouse, ^ bh 
as in abhor, and so on. This is (as we have seen above) 
strictly accurate only as regards the auid aspirates. 

38. The BUDHDt aspirates nre {in the opiaion of most), or at least 
represent, origioal Indo-European sounds, while the surd aspirates 
are a special Indian devolopmeut. The former are more thau twice 
as common as the latter. The unaspirated (non-nasal) mutes are very 
much more frequent [5 times) than the aspirates (for the special fre- 
quency of bb and original gb, see 60 and 66j; and among them the 
surds are more nnmorons (2Vj timesl than the sonants. The nasals 
(chiefly n and m) are nearly as fiequeut as the suid non-aspirates. 

We take up now the several mute-seriee. 

39. Guttural eeties: SR k, J^^kb, Hg, q^gh, l!^ n. 
These are the ordinary European /: and ^-sounds, with theii 
corresponding aspirates and nasal (the last, like English nff 
in singing). 

a. The guttnraii «ra daftned by the PrldcakhyM as nude by contsrt ot 
the buc of the tongae with the hase of the Jaw, and tliey'ars called, from 
the former organ, JlbT&inuliya tongue-root sounds. The Paiiinem ich^uie 
de*cribe9 them limply as made lu the throat (kai^Jhaj. From tho eujjhoiiic 
inllnence n( a k on a rollowin; b (bebw, 180), »>? may perhaps infor that 
In theii ultcranre the tongoe was well drawn back lo the mouth. 

40. The k is by far tho commonest of the gnttnral series occurring 
oonsiderably more often than all the otlier four taken together. The 
nasal, except as standing before ono of the others of the same series, 
is found only as final (after the loss of a foDowiug k: 386, 407) in 
a very small number of words, and as product of the assimilation of 
final k to a following nasal (161). 

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 processeB of alteration 
which began in the Indo-European period, the palatal mutes, the 
palatal sibilant f, and the aspiration b, have come from gntturals. 
See these various sounds helow. 

42. Palatal series; ^c, ^ oh, sT j, iK jb, 31 fl. 

The whole palatal series is derivative, being generated by the 
cormptiou of original guttntals. Tho o comes from an original k — 
as does also, by another degree of alteration, the palatal sibilant f 
(see below, 64), The J, in like m.inuer, couies from a g; but the 


12—] IT. Ststem of Sounds. ig 

Suskrlt j InclndOB in Itself two degrees of altentioD, one eorrespond- 
iBg to the ftlteration of k to c, the other to that of k to q {see below, 
SIB). The o ii eomowhat mote common than the j (aboat as four 
to three). The aspirate ob is yery mach less freqaent (a tenth of oj, 
and comes from the original gronp ek. The sonant aspirate jb is 
exceisivelj rare (occnrring bnt once in RV., not once in AV., and 
hardly balf-a-doaen times in the whole older language); where fouad, 
it is either onomatopoetic or of aoomalons or not Indo-Enropean origin. 
The nasal, fi, never occurs except immediately before — or, in a 
small namber of words, also after (SOlj — one of the others of the 
same series. 

43. Hence, Id the euphonic processes of the language, the 
treatment of the palatals is in many respects peculiar. In some 
situations, the orfginal unaltered guttural shows itself — or, aa 
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 o and j (except artificially, in the algebraic 
rules of the grammarians) do not interchange, as corresponding snrd 
and Bonant. 

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

a. Their deicTiptlon by the old HiDda grunmariana, howCTer, givea tbeui 
t not lee< absolutel; aimple characlei than betonga to the other mules. They 
Me called tilaTya palatal, and declared to be formed againat tbe palate by 
the middle ot the tongue. They «eeni to hsT« been, Ihen, broaght forward in 
the month from the gnttoral point, and made (gainst the hard palate tX i. 
point not far from the lingual one (belov, 4B), but 'nith the upper flat sarftce 
of the tongue instead of ite point Snch sonnds, in all languages, pass eitily 
into the (English) eh- and y-sonnds. Tbe value of the oh u makiug the 
preceding Towel long by position" (SS7), and its frequent orlgluaMon 
from t + f (SOS), lead to tbe suspicion that it, at least, may have had 
this character from the beginning: compare 87 d, above. 

4B. Lingual series: ?1 1, 7 (h, I <}, ^ 4h, HI Q. The 
lingual mutes are by all the native authorities defined 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 
grammaTians mftrdhanya, literally head-sounds, capitals, 
cephalics ; which term is in many European grammars 

ioy Google 

17 LlNflDAL AHD DeNTAL MdTBB. [—47 

lendeied by 'ceiebrals'. In practice, among Euiopean 
SanskritieU, no attempt is made to distinguish them from 
the dentals: 7 t is pionounced like FT t, 3 <} like ^ d, and 
so with the rest. 

46. The lingoale are another non-origin&l series of BonndB, 
coming mainly from the phoaetio alteratioa of the next Berlea, the 
dentals, but also in part occurring in words that have no traceable 
lodo-Enropean coDDection, and are perhaps derived from the ab- 
original languages of India. The tendency to lingnalization Is a 
poBitive one in the history of the language: dentals easily pass into 
liaguals under the Influence of contiguous or neigbbonring lingual 
sounds, but not the contrary i and all the sounds of the class become 
markedly more frequent in the later literature. The conditions of 
their ordinary occurrence are briefly these: 1. q comes from a, much 
more rarely ftom q, J, bf, in euphonic circumstances stated below 
(ISO, 318ff.)i 2. a dental mute following ^ is assimilated to it, 
becoming lingual (t, \b, 9; 187); 3. n la oftea changed to 9 after a 
lingual vowel or semivowel or sibilant in the same word (lB9ff.); 
4. ^, which is of very rare occurrence, comes from assimilation of 
a dental after q {108 a) or h (233]-, 5. \ and 4 come occasionally 
by substitution for some other sound which is not allowed to stand 
as final (14S, 14B-7). When originated in these ways, the lingual 
letters may be regarded as normal; in any other eases of their 
occurrence, they are either products of abnormal corruption, or signs 
of the noa-Indo-European character of the words in which they 

a. In ■ cerUtn nombeT of paMigei anmerlcall)' eximlned (below, 75), 
tbe abnormal occnirencea of lingual mnteB weie Im9 thin half of the whole 
nnmber (71 ant of 169), and moit of them (43) weie of 9: all were found 
more freqaeut in the later panages. In the Eig-Veda, only 15 words have 
an abnormal (; only 6, such a (hi only 1, BDch 1 ^b; about 20 (Encluding 
9 toota, nearly all of whtcb have derlvativee) abow an abnormal 4' besldea 
9 that have ^k^; and 30 (Inoludlng 1 root) show a 1^ 

b. Taken all together, the lingnals are by far the rarest class of 
mutes (about t'/j per cent, of the alphabet) — hardly half as frequent 
even as the palatals. 

47. Dental series: q^t, q^th, ^ d, U dh, ^T^n. These 
are called by the Hindus also dantya dental, and are 
described as formed at the teeth [oi at the roots of the 
teeth], by the tip of the tongue. They ate practically the 
equivalents of our European /, d, n. 

WhllBBJ, Oramiiur. 3. ti. 2 

Diaii.ze.., Google 

47—] II. System or Soitndb. 18 

a. Bat tbe modem Hlndni ue iild to prononoce theit denttli with the 
tip of the l«Dgiie thnut well foiward agiinst the upper teetb, bo th&t theie 
lounda get a alight tinge of the quality belonging to the Engltab ind Hodem 
Greek Ih-ionait. The abeenDe or (htt quality in the Europeui (eipecially 
tbe English) dentili la doubtleas tbe reuon why to tbe ear of ■ Hlndo the 
Utl«r (ppeu more aiulogona with his Ungaals, and he is apt to nae tbe linguale 
in writing European void a, 

48. The denuls are one of the IndD-Enrope*n original rnnte- 
clasHB. In their occDirence in Sanakrit they are just about aa frequent 
ns all the other four ciaBses taken together. 

49. Labial seiiea: 1 p, ^II ph, ? b, H bh, ^ m. 
These sounds are called o^ttiya labial by the Hindu gram- 
marians also. They are, of course, the equivalents of our 
p, b, m. 

so. The nDmerical relations of the labials are a little peculiar. 
Offing to tbe absence (or almost entire absence) of ( In Indo-Earopean, 
the Sanskrit b also is greatly exceeded in frequency by bh, which 
is the most common of all the sonant aapirates, as ph is the least 
common of the surd. The nasal m (notwithstanding its frequent 
euphonic mutstiona when final: SlSff.) occurs Jnst abont as often as 
all the other four members of the series together. 

a. From an early period in the history of tbe language, hot Increasingly 
later, b and t eiohange with one another, or fail to be diEtingniahed In the 
manQSorlpta. Thus, the donble root-forms b^h and vfh, b&db and Tadh, and 
BO on, Jn the Bengal manuscripta, v la widely written inatead of more original b. 

61. Semivowels: 77 y, J r, 3 I, 5 v. 

a. The name giren to this claas of rounds by tbe Hindu grammarians Is 
antahstbA ttanding bttieetn — either from their character as ntterancos 
intermediate between vowel and coneonanl, or (more probably) from the 
clKumBtance of their being placed between tbe mutes sad aplrants in tbe 
arrangement of the consonants. 

b. The semivowels are clearly akin with the several mute aeries 
in their physical character, and they are olaaslfied along with those 
series — thongh not without some discordances of view — by the Hiudn 
grammarians. They are said to be produced with tbe organa slightly 
in contact (i^atapi^t^), or in imperfect contact (duhap^^faj. 

62. 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 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. 

Dijiiizec., Google 

{9 Semivowbls. [^BB 

a. The PtnlneiD leheme leckoni r m ■ llnguil. None of tbe P^ti^ikhyM, 
tao«eT«r, doei lo; dot are ihej enttrelf eoniiitent with one anoUiet In lu 
dwcriptton. For Iha moit p>it, they define it u mida tt "the rtiots of the 
teeth". Thii waatd give It • position like thi>t of tbe Tibnted r; but no 
■nthority hlnti tt & Tib»tlon u beloaging to it. 

b. In point of frequency, r stands yerj high on the list of coo- 
soDanU; it is nearly equal with v, n, m, and y, and only exceeded 
by t. 

68. The ^ 1 ia a sound of dental positioD, and is bo 
defined and classed by all the native authorities. 

a. The peenllai ehanctet of >n ^aoand, u iDTOlviag eipultlon at the 
ilde of the toD^e along with conUct at Ita tip, ia not noticed by any Hinda 

b. The gemlTOweli r and 1 ue very widely interchangeable In Sanakrlc, 
both Id loota and in luffliei, and even In prefixes : tbeie are few toots contaln- 
ing * I which do not ahow also fenna with t; woidi written with the one 
lettei ate found in othet texts, ot in other parts of tbe same text, written 
with tbe other. In the later periods of the langnigs they are mere lepitated, 
nod tbe I becomes decidedly more beqnent, thongh always much ratei than 
the r (only as 1 to 7 or 8 oi 10). 

B4, Some of the Vedic texts have another ^aonnd, written with 
a sUghtly different character (it is given at the end of the alphabet, 
Ca), which is snbetitnted for a lingual 4 (as also the same followed 
by h for a 4^) when occurring between two vowels. It ia, then, 
doubtless a lingnal I, one made by breach (at the side of the tongue) 
of the lingual instead of the dental mute closure. 

a, Ezamplee ate; |^ lie, for ^ i^e, bnt ^^ 1^7*; MIoo^^ 
mOhnfe, for iflBq ml^bn^, bat ^^H mi^Tftn. It Is especially in 
the Rig-Teda and its anxiltary litetatate that this snbstltatlon is naual. 

SB. The n* y in Sanskrit, aa in other languages generally, 
stands in the oloseat relationship with the vowel ^ i (short 
or long); the two exchange with one another in cases ia- 

a. And in the Teda (as the metre shows) an 1 is very often to be read 
where, in oonfoimlty with the roles o( the later Sanskrit eopbony, a y ta 
rnitten. Thus, the final i-TOwel 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 1 Instead of y. Sacb cases will he noticed 
in more detail later. Tbe ooastancy of the phenomenon in certain words and 
classes of word* shows that this was no merely opUonsl interohinge. Very 
probably, the Sanskrit y had everywhere more of an 1-charaotei than belongs 
10 the corresponding European sound. 

itizecy Google 

56—] !I. Ststbm of SotrjsDS. 20 

Be. The 7 is by it* phyaiol character a palatal ntterance; and 
It IB classed as a palatal aemivowel by the Hindu phonetiate. It is 
one of the most common of Sanskrit sounds. 

67. The ^ T is pronounced as English or French v 
(Geiman vj) hy the modem Hindus — except when preceded 
by a consonant in the same syllable, in which case it has 
rather the sound of English w; and European scholars follow 
the same practice (with or without the same exception). 

a. By its whole treatment in the eaphony of the language, however, 
the V stands related to an u-towoI precisely as y to an t-vowel. It 
is, then, a v only according to the original Roman value of that 
letter — that is to say, a ic-sonnd in the Eaglish sense; though (as 
was stated above for the y) it may well bavo been less markedly 
separated from u than English w, or more like French oh in out etc. 
Bnt, as the original w has in most European languages been changed 
to u (English), so also in India, and that from a very early time: the 
Panioean scheme and two of the Fr&titikhyss (VPr. and TPr.) distinctly 
define the sound as made between the upper teeth and the lower 
lip — which, of course, identifies it with ihe ordinary modem v-sonnd. 
As a matter of practice, the nsnal pronnuciation need not be seriously 
objected to; yet the student shoald not fail to note that the rules of 
Sanskrit enphooy and the name of "semivowel" have no implication 
except to a w-sound in the English sense: a c-souud (German u>) is 
no semivowel, bnt a spirant, standing on the same articulate stage 
with the English fA-sounds and the / 

68. I'he V is classed as a labial semivowel by the Hindu phonet- 
leal authorities. It has a somewhat greater frequency than the y. 

a> III the Teda, under the lame clrcam stances as the y (above, B6 ft), 
v ia to be read as a vowel, tL 

b. Ab to the iuCeichange of v and b, see above, 60a. 

69. Spirants. Under the name O^man (literally heat, 
steam, Jiatus), which is usually and well represented by 
spirant, some of the Hindu authoiities include all the remain- 
ing sounds of the alphabet; others apply the term only to 
the three sibilants and the aspiration — to which it will here 
also be restricted. 

a. Tbe term is not found In the Paninean Bcheme ; by dttrerent treatises 
the luttunl and labial breathings, these and the vlaarga, or all these and 
anUBvSra, are aho (In addition to (he sibilants and li) cslled ii^man (see 


21 SlBlLAHTS. [—68 

AFi. i. 31 Dote). The orguia ot Dttennce aie dexulbeil u bslng la the 
poeitlon of the mate-aerlee to which eieb splruit belongs reipedively, but 
uneloBed, oi unclosed in tbe middle. 

60. The TT B. Of the three Bibilants, ot suid spirants, 
this is the one of plainest and least questioned character: 
it is the ordinary European 3 — a hiss expelled between 
the tongue and the roof of the mouth directly behind the 
upper front teeth. 

a. It is, then, dental, u it ie clasBed by all tlie Hinda aDtborities. 
NotwithBtanding the great tosses which it suffers in Saoskrit euphony, 
hj conveTBioQ to the other sibilants, to r, to TiBarga, etc., it is 
■till very high among the coDBonantB in the order of fiequeucy, or 
considerably more common than both the other two sibilants together. 

dl. The ^ 9- 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 
reverted 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 oase of the other Ungual 
sounds: 45} to give it its proper lingual quality. 

a. Its lingaal character ia ehowa by ita whole euphonic Influence, 
and it is described and classed as lingual by all the Hindn anthor- 
iHea (the AFr. adds, i. 33, that the tongue in its utterance is trougb- 
Bhaped). to its audible quality, it ia a »A-sound rather than B«-Boand; 
and, in the considerable variety of sibilant-utterance, even in the 
same community, it may coincide with the jA of some among 
onrselvea. Yet the general and normal «A is palatal (see below, 63] ; 
and threrefore the sign ^, marked in accordance with the other lin- 
gual letters, is the only unexceptionable transliteration for the Hindu 

b. Id madoFD ptonuiidition in India, ; la muoh confounded with kb; 
■nd the mannacilpts aie apt to eicbange the chinetecs. Some Uter giim- 
matlckl tteatlaea, too, take note or the lelitiooablp. 

6fl. This sibilant [as was noticed above, 46, and will be more 
particalarly explained below, 180 ff.] is no original sound, bnt a 
product of the Ungnalization of s under certain euphonic condltiona. 
The exceptions are extremely few (9 oat of 145 noted occnrrencea: 
75), and of a purely eporadic character. The RIg-Veda has [apart 

Digitizecy Google 

•a—] II. System or Soukdb. 22 

from ^ Bah, 182 b) only twelve words whloh show a q under other 
Condi tiooa. 

a. The flDit 9 of a root bu Id lome OMet att&lued ■ more Independent 
TilQe, and doet not levett to B when the enphonic condJtlDna ate teMOVed, 
1>Qt Bhowg aiionlaloDl rorml (286-6). 

63. The ^ 9. This sibilant is by all the native authoiities 
classed and described as palatal, nor is there anything in 
its histoiy or its euphonic tteatment to cast doubt oa its 
character as such. It is, then, made irith the flat of the 
tongve 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 «A. 

a. The two (A-8onnd(, ; and 9, ate made In the »anie patt of the 
mouth (the 9 probably lather tnilliet baclt), but with a different part of 
the tengne; and they are doabtleti not moie nnlike than, for example, the 
two f-aonnds, written ( and t; and it would be not lest proper to pronounce 
them both u one »h Ihan to pronoance the lingnaU and dantali alike. To 
neglect the difference of b and 9 i« much lev to be approved. The very 
□ear relationship of 9 and q Ib attested by their enphonic treatment, which 
1b to a conaiderable extent the same, and by their not infrequent oonfUsion 
by the wTlteie of mannscripti. 

64. As was mentioned above (41), the 9, lilce o, comes from the 
corruption of an original A-sonnd, by lose of mute-contact u well as 
forward shift of the point of production. Id virtue of this derivation, 
it sometimes (though less often than o) "reverts" to k— that is, the 
original k appears instead of it (43),' while, on the other band, aa a 
«A-aoQnd, it Is to a certain extent convertible to 9. In point of frequency, 
it slightly exceeds the latter. 

66. The remaining spirant, ^ h, is ordinarily pronounced 
like the usual European sutd aspiration h. 

a. This ia not, however, its real character. It is deflned by all the native 
anthoiities at not a auid element, but a sonant (or elae an ntterance Inter- 
mediate between the two); and Its whole value In the enphony of the language 
i» that of a »onaut: but what le its precise value Is very bard to say. The 
Pauiuein scheme ranks U as guttural, as It does also a: thia meint nothing. 
The Pratitakbyas bring it Into no lelatlon with the guttural clsss) one of 
them quotes the opinion of some anthorltiea that 'It bas the aame poai^n 
with the beginning of the following vowel" (TPr. 11. 47) — which so Us 
tdentifles it with our h. There is nothing in ita enphonio influence to mark 
It a* retaining any trace of gntturally articulated character. By some of 

ioy Google 

23 VlHARQA. [—89 

tie DatiTe phooetiati it ia identifled vlth tbe upintton of tlie lontnt 
upintei — with the element by which, for eiunple, gh differs from g. 
This view U aopparted bjr tlia derivation of h ftom the aapiiates (ueit 
paiagiaph), bjr that of 1 4- b from ^ (S4), and by the treatment of initial 
h after a Bnal mute (168). 

68. Tbe h, as alreadj noticed, is not an original sonnd, bnt 
cornea in nearly all cases from an older gh (for the few iostanoes of 
its deriTatlon from dh and bh, aee belov, 2S3s). It Is a vastly 
more frequent Boand than the unchanged gh (namely, as 7 to 1): more 
frequent, indeed, than any of the guttural mutee except k. It appears, 
like J !S19), to inclnde in itself two stages of corruption of gb: one 
coTrespondlng with tbat of k to o, the other with that of k to ;; 
see below, 233, for the roots belonging to the two classes respectively. 
Like the other sounds of gnttural derivation, it sonetimos exhibits 
"reversion" (48) to its original. 

67. The : It, or viflarga (visarjRDlya, as it is unifoimly 
called by the Prati^akhyas and by Pacini, probably as belong- 
ing to the end of a syllable), appears to be merely a surd 
breathing, a final A-sound (in the Eutopean sense of h), 
uttered in the articulating position ot the preceding vowel. 

a. One FrBti^kh;a (TPr. U. 48) glvet jnat this lait deicriptlon of it. 
It Is b; Tarfou* authorities classed with h, or with b and ft: all of them 
are alike Bounda In whoae utterance the month-oisans have no definite 
shaping action. 

68. The visorga is not original, but always only a substitute 
for final b or r, neither of which ia allowed to maintain itself nnchauged 
{170ff.). It is a comparatively recent member of tbe alphabetic 
system; the other euphonic changes of final ■ and r have not passed 
through viaarga as an intermediate stage. And the Hindu aathorittes 
are oonsidentbly discordant with one auotber as to bow far 1; Is a 
necessary aubstitnte, and how iax a permitted oim, alternative with 
a sibilant, before a following initial surd. 

60, Before a Burd guttaral or labial, reapectively, aome of the 
native authorities permit, while others require, conversion of final s 
or r into the so-called JlhvSmlUIya and upadbmAniya spirants. It 
may be fairly questioned, perhaps, whether these two sounds are not 
pure grammatical abstractions, devised (like the long {-vowel: 33a) 
in order to ronnd out the alphabet to greater symiuetry. At any 
rate, both manuscripts and printed texts in general make no 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 Uansliterated by x ond 9^. 

ioy Google 

70— J IL Systeh of Sounds. 21 

TO. The ~ anuBv&ra, ft or ifa, is a oasal sound lacking 
that oloBure of the organe which is required to make a 
nasal mute 01 contact-sound (36); in its utterance there is 
nasal resonance along with some degree of openness of the 

71. There li di«Miid»nee of opinion unong both the Hindu phonetlBU 
lad theli modein EniDpean (nicceBaoia reipecting the leal chu&ctei of thU 
eUmenC; henee « Utile detail la neceuuy here with legud to its oecnnence 
sad theii tIcvb of it 

a. Certain dmU» in Ssnikrlt are of eerrile chsitetei, Uvars to be 
uiimlUted to > foUoviDg coasonaDt, of irbateTet cb>T>eter that ma; be. 
Such are Anal m in aentence-combinatiOD (313), the penultimate nasal of 
a root, and a nasal of inclement (266) in general. If one of theae naMl« 
Btandi before a contact-latter 01 mote. It becomei ■ naaal mute Mneepond- 
ing to the latter — that Is, a nual utterance In the aame poiltlon of the 
mouth-organa which gtveg the anccaediog mate. If, on the other hand, the 
following conioaant doee not in^ol^e a contact (being a semirowel or aplnmt), 
the naaal element ia also withoat contact: it ia a naaal utterance with 
unclosed month-organa. The qneatlou ia, now, whether this naeil utteranc« 
becomea merely a naaal Infection of the preceding vowel, turning it into ■ 
nasal vowel (as in French on, en, uit, etc., bj reason of a aimllar loaa of 
a naaal mute); or whether it la an element of more Indlrldnal ebaiacter, 
having place between the Towel and the conaonant; or, once more, whether 
it ti aometimes the one thing and aometlmea the other. The opinions of 
the PAtl(«kh7aB and Knini are briefly aa followa : 

b. The Atbaira-Priti^aUiya holda that the result is everywhere a 
nasalized lowel, except when n or m Is asBlmllated to a following I; in 
that case, the n or m becomes a naeal 1: that ia, the naaal utterance it 
made in the l-posltioD, and haa ■ perceptible I'Charactei. 

c. The other Prltl9akhyaa teach a limilai coUTersion Into a naeal 
counterpart to the semiyowel, or s nasal aemiTowel, before y and 1 and v 
(not before r also). In moat of tiie other obbbi where the Atharfa'FratifakhTa 
acknowledges a nasal Towel — namely, belbre r and the spirants — the others 
teach the intervention after the Towel of a distinct naaal element, called the 
anuBv&ra a/ter-ton». 

d. Of the nature of this nasal afterpiece to the lowel no IntelUglbty 
clear account ia given. It la aald (RPr.) to be either vowel or consonant; 
it is decUred (BFi,, VPr.) to be made with the hobo alone, or (TPr.) to be 
nasal like the naaal mutea; it is held by some (RPr.) to be the louaut tone 
of the naial mutes; in its formation, a* in that of vowel and spirant, there 
ia (BFr.) no conuct As to ita quantity, see fnttber on. 

e. There are, however, certain casea and clasaes of caaea where these 
other authorities atao acknowledge a nasal TOweL So, especially, whereyer 

i, Google 

25 Akubvaba. [—73 

a flual a li UeiUd (308-8) u if it vera hb (its UitoilMlly otdei foim]; 
*Dd also In * Bmall DUiuber of ipeelflad voidi. They al»o mautloii the 
doctrine of Dual vowel lutead of uinsvKrft as held by some (and TPi. 
U nnceTtain and incanilatent Id Iti choice between the one and the olheij. 

f. In Panlnl, Anally, the prevaiUag doctrine 1b that of annsvftra 
eTeiywhere; and it la «ven allDwed in maoy cases vheie the Piatifikfayas 
pieacribe oaly a uual mate. Bat a naaal semivowel la alto allowed inatead 
befoie a semtTowel, and a nsBal vowel is allowed in the caaea (mentioned 
above) where soma of the Fratifakhyas leqolre it by eiaepHon. 

g. It ii evidently a lali qnestlDn whethai this dlseoidance and nnoertalnty 
of the Hindu phanetisls is owini to a real difference of utteraoce in different 
ctasaes of eases and in different localitleB, or whether to a different scholaetlc 
analysis of what la really everywheTe the same utterance. If anuBTftra 
I* a nasal element following the vowel, it cannot well be any thing bnt 
either a proIongatioD ot the same vowel-sonnd with nasality added, or a 
nasalized bit of nenlral-vowel eonnd (in the latter case, however, the altering 
inflaenee of an 1 ot u-vovel on a following a ooght to be prevented, which 
is not the case: Bee 1B3). 

7S. The asBimilated nasal element, whether viewed as nasalized 
Towel, naBal Bemivowel, or independent aniuiTjtTa, has the value of 
eometbing added, in making a heav^ syllable, or length by poeilioD (70). 

a. The Pratlfikhyaa (YPt„ RPr.) give determinatloni of the qaandty 
of the anxuvftra combining irlth a short and with a long vowel respectively 
to make a long syllable. 

73. a. Two different HlgtiB, i and r, are found in the mannscriptB, 
indicatiDg the nasal sound heie treated of. Usually they are written 
above the syllable, and there they seem most naturally to imply a 
D&sal affection of the vowel of the syllable, a nasal (anunaalka) vowel. 
Hence some texts (Sama- and Tajur-Vedas], when they mean a real 
aniuv&ra, bring oao of the signs down into the ordinary consonant- 
place; but the usage is not general. As between the two signs, 
some mannscriptB employ, or tend to employ, the Z where a nasaliied 
(annnaslkal vowel is to be recognized, and elsewhere the -; and this 
distinction Is consistently observed in many European printed texts; 
and the former is called the anunoalka sign: bat the two are doubt- 
less originally and properly equivalent. 

b. It is a very common custom of tbe mannscriptB to write the 
oaTiBvKra-sigD 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 annavSra. Some printed texts follow this slovenly and 
nndesirable babit; bnt most write a nasal mute whenever it Is to bo 
pronounced — excepting where it is an assimilated m [S13). 

itizecy Google 

73—1 H- System op Sochds. 26 

0. It is convenient also In tranBliteraUon to distiDguiBh the as- 
similated m by a special sign, di, from the annav&ra of more inde- 
pendent origin, ft; and this method will be folloved in the present work. 

74. This is the whole system of soands recognized by the written 
character; for certain other transidonai sonnds, more or lees widely 
recognised in the theories of the Hindu phonedsts, see below, S80. 

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

a, s 1 


T. r 



























up. «•'•■ 



Dent. Lab. 

a. The figures set under the characters give the average per- 
centage of freqoenoy of each sound, found by counting the number 
of times which it occurred in an aggregate of 10,000 sonnds of con- 
tinouB text, in ten 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 Brahma^as, and one 
each ftom Hana, Bhagavad-GItti, Qakantals, Hitopadega, and Vaaa- 
vadat» (J.A.O.S., vol. X., p. c1). 

ioy Google 

III. Quantify of sounds and syllables. 

7d. The Hindu giammarians take the pains to define the 
quantity of a consonant (without distinction among consonants 
of diffeient classes] as half that of a short vowel. 

77. They also define the quantity of a long [dlrglia] 
Towel 01 diphthong as twice that of a short (hraava) vowel — 
making no distinction in this respect between the gu]^- 
and the vrddbi-diphthongs. 

7B. Besides these two vowel-qiiantitiee, the Hindus 
acknowledge a third, called pluta (Utetally swimminff), or 
protracted, and having three moras or three times the quantity 
of a short vowel. A protracted vowel is marked by a follow- 
ing figure 3: thus, ^ S3. 

». Tbe protracted vowels are practically of rare occurrence (in 
RV., three oases; in AV., fifteen; in the Brahmana literature, decidedly 
more frequent). They are used in cases of qnestioning, especially of 
a balancing between two alternatives, and also of calliDg to a distance 
or urgeotly. The protraction is of the last syllable in a word, or in 
a whole phrase; and the protracted syllable has nsually the acnte tone, 
in addition to any other accent the word may have; Bometimcfl it 
takes also anuavSra, or is made naeal- 

b. Examples are; adti&l^ avid ftsisd upiri avid a^3t (RV.) tcai 
it, foTMoih, belowT uxu it, forsooth, ahoeef Id&m blitl7a3 idii3m {ti 
(AV.) toying, it thi* more, or is thalf Agni^i p&tnlvos^ admam piba 
(TS.) O Agni! fhou with thy ipouie.' drink the toma. 

c. A dipbtboDg Is piotracted by prolongation of its Biet or a-elemcnt: 
thuB, e to &si, o to jL3a. 

d. The sign ot protraction Is also lometiniei wTltt«n u the lesnlt of 
accentual combination, when ao-called kampa occurt: «ee below, 87 d. 

70. For metrical purposes, syllables (not vowels) are 
distinguished by the grammarians as heavy (guru) or light 
(iBghn). A syllable is heavy if its vowel is long, or short 
and followed by more than one consonant ("long by po- 
sition"). Anusvftra and vlsar^a count as full consonants in 

DijiiLc. ..Google 

79 — ] 11. Ststeu of Sounds. 28 

making a heavy syllable. The last syllable of a pSda (pri- 
mary divuion of a verse) is reckoned as either heavy ot 

B,. Th« dittinctlOD In tanni betveea Ihs difference of long *nd short In 
TOirel-ioiind >nd that of lieiit7 tod Uglit in lyUtblO'ConatnictlaD li laluible, 
and ihoald be obierred. 

IV. Accent 

80. The phenomena of accent are, by the Hindu gram- 
maiians 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 (svaraj or accent-pitches are two: 
a higher (ud&tta raised), or acute; and a lower [anadStta 
not raised), or grave. A third (called svarita: a term of 
doubtful meaning) is always of secondary origin, being (when 
not enclitic: see below, 8S] 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. 

8B. Strictly, therefore, there is but one diatiDClion of tone in the 
Sanskrit accootnal syatem, as described by the nntive grammuiaDB 
and marked in the written texts: the accented syllable is raised in tone 
above the unaccentod; while then further, in certain cases of the 
fusion of an accented and an nnaocented element Into one syllable, 
that syllable retains the compounded tone of both elements. 

88. The svarita or circamflex is only rarely fonnd on a pnre long 
vowel or diphthong, but almost always on a syUable in which a vowel, 
short or long, is preceded by a ; or v repreaeatiog an originally acate 
i- or u-7owel. 

a. In transliteration, in this work, the adStta or acute will be 
marked with tbe ordinary sign of acute, and the svarita or circumflex 
(as being a downward slide of the voice forward) with what is usually 
called the gravo accent: thus, &, acute, yk or vh, cironmflex. 

Digitizecy Google 

29 Accent. [—65 

84. The Piitlfikhyia dUtlaguith and ntme lapuately the cItcdid flexed 
lonei (Tiling by different processes of comblnadon : thoe, Che ciioamflex is 

a. K^&ipTK (quieli), vhen ui tcnte i- oi a-Toirel (short nr long) i* 
eonTeited into y oc v before t, dlislmlln TOirel of grave lone : tbiK, VT^pte 
^m vi-ftpta, apevimt&r from apsu aiit4r. 

b. JStya (native) oi nitya [own), when the ume comblDition Ilea 
farther back, In the make-np of a item oi foim, and bo ii coaatiot, or 
beloDgi to Ibe word In all cireumetancei of iM occurrence; thai, IcTli(from 
kua), nhx (adar), ayhik (niak), bodhnyii (budhnfa), kanyll (kanift), 
Dady&a (nadi-as], tanvll (^tantt-ft). 

o. Tbe worda of both tbe aboie ciutea are in the Veda, in the great 
majority of cues, to be read frith reatoTatioD of the acute vowel ai a lepirate 
ayllable: thna, apsii antb, B&ar, nadlaa, etc. In some texts, part of 
them are written correspondingly: thna, ttivrta, toni^TS, budbniya. 

d. Praflifta, when the acute and graie Towela are of anch character 
that they are fused Into alongTOwel or diphthong (IfiSo): that, dlyi 'va 
(RV. AV. etf.), 'rom dlvf iva; ijldg&t& (TSO, from Bd.adgSMi nSl 'vt 
'fniyKt (QB.), tiom ak evk aqnlyat 

e, Abhlnlbita, when an initial grave a ia absorbed by a final acute 
A or 6 (186 a): thns, ti 'bruvon, from ti abmvan; s6 'bravTt, from 
b6 abraTit. 

85. But further, the Hindu grammariaoB agree in de- 
claring the (natuially grave] syllable following an acute, 
whether in the same oi in another word, to be svarita or 
citoumflex — unless, indeed, it be itself followed by an 
acute 01 circumflex; in which case it retains its grave 
tone. This is called by European scholars the enclitic or 
dependent circumflex. 

a. ThDB, in tdna and t6 oa, the syllable na and word oa are 
regarded and marked as circumflex; but in t^na ti and td oa aviir 
they are grave. 

b. Thla aeema to mean that the TOlee, which la borne up at the higher 
pitch to the end of the acute syllable, does not ordinarily drop to grave 
pilch by an Instantaneous movement, bat descends by a more or lets per- 
ceptible slide In the course of the following ayllable. No Hindu authority 
snggeats the theory of a middle or Interniediate tone for the enclitic, any 
more than for the independent clnnimflei. For the moat part, tbe two are 
Identiasd with one another, in treatment and designation. The enclitic 
eimunltex ts likewise divided into a unmber of snb -varieties, with different 
names : tliey are of too little conteqDonce to be worth reporting. 

D,j,i,...., Google 

80—] II. System of Soukds. 30 

86. The esaential difference of the two kinds ol oircnmflex is 
Bhown cle&rlf enough by these fftcts: 1. the iDdepandent circumflex 
takes the pl&ce of the acute as the proper accent of a word, while 
the enclitio is the mere shadow following an acute, and following it 
in another wotd precisely as in the same word; 2. the independent 
circumflex mUutainB its character in all eituatione, while the encliUe 
before a followiog circumflex oi acute loses its circumflex character, 
and becomes grave; moreover, 3. 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 primary Vedic texts, or suhhitfts, in two 
of the Brahma^as (TMttirlya and ^^tapatha), in the Taittf tlya-Aranyaka, 
in certain passages of the Aitareya-Ara^yaka, and in the Suparnadhyaya. 
There are a number of methods of writing accent, more or less different 
Arom one another: the one found in manuscripts of the Big-Veda, 
wbicb is most widely known, and of which most of the others are 
only slight modificatious, is as follows. ' 

a. The acute syllable is left unmarked; the circumflex, whether 
independent or encliljc, has a short perpendicular stroke above; and 
the grave next preceding an acute or (independent) circumflex has a 
short horizontal stroke below, Thus, 

gfilH agnlm; sji^Ih Juh6tl; FF5|T tanva; Iji kva. 

b. But the Introductory grave stroke below cannot be given if an 
acute syllable ia initial; hence 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 
most all alike have the grave sign. Thus, 

^; Indiah; ^ t«; ^ff'^i^ kari^yiai; r^fcf^lHI tuviJMl 

o. All the grave eyliables, however, which follow a marked cir- 
cumflex are left unmarked, until the occurrence of another accented 
syllable causes the one wbicb precedes it to take the preparatory 
stroke below. Thus, 

H<^nVitt<^ Hudf flOtasaihdf k ; 

but H^^fl'+iH^JJ NlH^audt^lkaBurfidre g&vSm. 

d. If an Independent circumflex be followed by au acute (or by 
another Independent ciroumflex), a figure 1 is act after the former 
ciroumflexed vowel if it be short, or a figure 3 if it be long, and the 
signs of accent are applied as in the following examples: 

^LHItT! apBv aintAl^ (from apad anti^); 

{lUI^H: ifty63 v&ni^ (from ri^6 aT&ni^). 

Digitizecy Google 


81 AOOBNT. [—99 

Th« ntloii»1e of this mod« of deilgnitlon ii not well nnderatood; the 
Frili^kbyu give no aeconnt of it. Jn the ichalaatln attsriDce of the lyllable 
M deilgnftted la made > pecniiu qntvei oi rouhult of the voice, called 
kampa or Tlk&mpanft. 

e. The icceat-mnki a.n written with led ink in the mannBerlptg, being 
*dded (itei iha teit is written, ind pecbape often b; mothet band. 

88 a> Netrljr scoordsnt with thii, the Rlg-Veda method of deeigniting 
Mcent, tre the methods employed in the m>nDHaript« of the Atbu*>-Yeda, 
of the yquineyi-SuhhlU, >nd of the Tiittinya-Sambita, Biihniana, and 
Arinytka. Tbeli diffeieneoB from It are of trifling importance, conitittng 
mainly in peculiar ways of marking the circamflei that precedes an aoate 
(87 d). In lome manuecripta of the Atbarva-Veda, the aueiit-markB aie 
dota inatead of attokee, and that for the dicumflei is made within the 
syllable instead of aboTC it 

b. In most mannseripts of the Haltrayanl-Samhita, the acute syllable 
itself, besides Its snrroundlnga, is marked — namely, by a perpendlcalar 
stroke abOTe the syllable (like that of the ordinary cirenmBei In the BV. 
method}. The independent circumflex has a hook beneath the syllable, and 
the cltcnmllei before an acute (87 d) la denoted simply by b Ognre 3, 
st*adlng before Instead of after the circumfleied syllable. 

a. The (atapatha-Brihma^a nees only a single aceent-algn, the horizontal 
stroke beneath the syllable (like the mark for grave in RV.). Thla ia put 
nnder an acute, or, if two or more acates immediately follow one another, 
only nnder the preceding syllable. To mark in Independent olrcamflei, it 
ia pat under the preceding syllable. The method is an Imperfect one, allow- 
ing many ambigoltles. 

d. The Sama-Teda method la the most intricate of all. It has a dozen 
dlfTeient slgna, consisting of flgacea, or of flgnies and letters combined, all 
placed above the syllables, and varying according both to the aooentaal ohuacter 
of the syllable and to Its snrroDndlngs. Its origin is obscure; if anything 
more it indicated by It than by the other simpler aystems, the fact has not 
been demonstrated. 

8e. Id tbia work, as everything given in tbe devanicart characters 
ia also given in tianal Iteration, 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 really accented syllables, the aonte and the independent 
circnmflex: the latter by the asual evarita-sign, the former by a small 
o (for ad&tta) above the syllable: thus, 

^ bidfo, ^ Agnej fE(^ avkr. ^TuTT nndyia. 

a. These being given, evetythlog else which the Hindu theory recog- 
nlies aa dependent on and aeeompanylng them can readily be nnderatood 
as implied. 

i, Google 

eO— ] II- SrsTEU OF Sounds, 32 

90. The theory of the SiDskrtt ueent, m heie given (& coniiiUnt and 
Intelliglbl* bod; of phenomena), has been eveilaid if the Hindu theoriati, 
eapedallf of the Pridfikhyu, with • nnmbeT of added fealurei, of a maeh 
more qneitlonable chaiactei. Tha»: 

a. The nnmarked grave tyllablei following a oiicnmflei (either at the 
end of a lentence, ot till the neat aproaeh of another acnte) ate deolaied 
to have the aame high lone with the (also nnmarhed) acDte. The; are 
called praoaya ot pradta (atmmulattd: beciDM liable to oceor ia an 
indefloite leries of ancceiiiTe aylUbleO. 

b. The citcnmflei, whether independent or enclitic, It declared to begin 
on a higher pitch thin aeate, and to descend to aoute pitch in ordinary 
cube: the conclndlng Inatant of it being bronght down to grave pitch, 
however, In the cue of an Independent circumflei which ii Immediately 
followed by another Mcent of tbe voice to higher pitch, in acnte or inde- 
pendent circDmBex (a kampa syllable: 87 d). 

0. Kntol gives the ambignona name of dk&^mti-imonotme) to the 
praoita eyllablea, and says nothing of the uplifting ot the dicumtlei to 
a higher plane; he teaches, however, a depresalon below the grave pitch for 
the marked grave syllable before acnte or clrenmflez, calling It sannatara 
(other wise anadSttatara). 

SI. The system ot aecentnation a) marked In (he Tedic texti appears 
to h&ve asBumsd in the traditional recitation of the Brabmanie schools 
a pocnliar and artiflclal form. In which the designated syllablea, grsTe and 
circumflex (equally the enclitic ind the Independent otrcnmflei), have acquired 
s couiplcuous value, wbile the nndeaignited, the acnte, hai sunk into In- 

0S> The SaoBkrit accent taught in the native grammars and 
represented b^ the acceotnated texta is essentially a system of word- 
accent only. No general attempt la made (any more than in the 
Qreek syatem) to define or marb a sentence-accent, the effect of the 
emphasis and modalation of the sentence in modifying the independent 
accent of individual words. The only approach to it is seen (n the 
treatment of vocatives and personal verb-forme. 

a. A vocative is nsnally without accent except at the iKgiDoing 
of a aeotence: for further details, see 314. 

b. A personal verb-form is usnally accentiess in an independent 
clause, except when etaoding at the beginning of the clause: for 
farther details, see &B1 fF. 

98. Certain other words also are, nsnally or always, without 

a. The partlcloi en, tO, u, ama, Iva, dd, avid, ha, and the Tedic 
kam [or k&m), gha, bbala, aamaha. Im, lim, are always without 
accent; also ratbS In BT. (sometimes alio elsewhere) in tbe sense of iva, 
at the end of a pftda or verae-divldon. 

ioy Google 

33 Accent. [—96 

b. The Mine i> true of miUId prauonni ind pronominal itemi : mS, 
me, n&nt nu, tvK. t«, vtm, vm (491 b), enft (600). tva [60S b), 
sama (618 o). 

o. The CUM ot the pnnoinlntl stem a ue Mmetimei ucanted and 
aomstimeB accentleii (602). 

d. Ad accentleas word is not allowed to Btaod at the beginniDg 
of a senteace; also not of a p&da or primarf division of a veTsa; a 
p&da ia, in all matters relating to accenttiation, treated like »u In- 
dependent aenteoce. 

94. Some words have more than a eingle accented ayllable. 
Such are: 

a. OertaiD daal copnlatiTe componndt In the Veda (see 1SB6), aa 
mltriT&TDQfi, dyivftp^ftblTl. Alao, a tew othei Tedic eomponnd* [lee 
1867 d), aa b^haap&tl, t&niin&p&t. 

b. In a fair caaea, the torther oompounda and derivatlTei of anob 
compounda, aa dyitTtpptblvivant, btbaepAtipra^atta. 

o. Inflnltlve dattTea In tavftl (aee 979 a), m ita,vai, ipabbar- 

d< A void natatally bairlone, bvt having ila final syllable protracted 
(aee 78 a). 

e. The particle v&vk (In the Brihmanaa). 

96. On the place of the accented syllable in a Saoskrit 
irord there is no testiictios whatever depending upon either 
the number or the quantity of the pieceding oi following 
syllables. The accent rests where the rules of inflection 
or derivation or composition place it, without regard to any 
thing else. 

a. Tboa, (ndre, agnKvi, (ndreqa, agnfnA, agninam, bKhaoyuta. 
toapaoTuta, parj&nyajlnvita, abhimStitSbi, ^aabhlmlStaTar^a, 
abhi^Bstlalit&na, hlrai^avaqimattama, e&tuQoatv&riA^adakfara. 

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 &r 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 geneial habit with European scholars to pronounce 
Sanskrit words according to the rules of the Latin accent. 

WliitDS)', Onmrnu-. 3. «d. 3 


07—] in. Euphonic CoifBiNATiON. 34 

97. In this woik, the accent of Moh word and fonn will in 
gener&l be marked, so far as there ia aotfaority determining its place 
and cliaracter. Where specific wotda and forms are quoted, they 
will only be BO far acceotnated as they are found with accent in 
accentuated texts. 




86. The words in Sanskrit, as in the other languages related 
with it, are in great part analysable into roots, suffixes of derivation, 
and endings of inflection, these last being added mostly to stems 
containing suffixes, but also sometimes directly to roots. 

a. Thete ue, of conrge, t ceit&in nnmbcr of tinlnflected words — 
indecUntbles, particles ^ &iid iJbo not a few that ire !uc*p<ble of analyels. 

09. The Sanskrit, indeed, possesges m exceptionally analyaable 
character; its formative process^ are more regular and transparent 
than those of any other Indo-European tongue. Hence the prevuling 
method of the Hindu native science of grammar, which sets up a 
certain body of roots, and prescribes the processes by which these 
may be made stems and words, giving the various added elements, 
and laying down the rules by which their combination is effected. And 
the same general method is, for like reason, tollowed also by European 

100. The eaphonic laws, accordingly, which govern the combiuation 
of suffix or of ending with root or stem, possess a high practical im- 
portance, and require to be laid down in preparation for the topics 
of declension and conjugation. 

101. Moreover, the formation of compounds, by joining two or 
more simple stems, is extremely frequent in Sanskrit; and this kind 
of combination has its own peculiar eaphonic rules. And once more, 
in the form of the language as handed down to us by its literature, 
the words composing a sentence or paragraph are adapted to and 
combioed with one another by nearly the same rules which govern 
the making of compounds; so that it is tmposaible to take apart and 
understand a Sanskrit sentence without knowing those rules. Heuce 



BD Increued Aegrw of practical importance beloDging to ths subject 
of eaphonio combination. 

a. Thi* enpbonlc tnterdepsndense of the wUTds at ■ lentenoa Ig nn- 
knavn to an; other Ungnage in anything like the (tme degree ^ and it 
cannot but be snspected of being at leait In part artiflrlal, Implying an 
erecliOD Into neceaauy and Invariable nilet of what in the living language 
were only optional praeticea. Thla is strongly Indioated, indeed, by the 
evidence of the older dialeot of the Vedag and ol tbe derived PrakrIHc 
dialects, In both of wbloh some of tbe m1e« (especially tliat aa to the biatn« ; 
aee 113) are often violated. 

lOS. The roots irblch are anthenticated by tbeir occarreace in 
the literary monnmenta of the language, earlier and later, anmber 
between eight and nine hundred. About half or these Iielong fully 
to tbe language throughout its whole history; some (about a hundred 
and fifty) are limited to the earlier or pre-claaeioal period; some, 
again (over a hundred and twenty}, make their first appearance in 
the later language. 

a. There are In this nnmber roots of very diverae character. Thoae 
occurring only later are, at least In great part, preanmably of secondary 
origin; and a certain nnmher are even doubtleis artifldal, used onoe or 
twice beetose found in the root-Uita of the Ulndn grammarians (103). 
Bnt also of the leit, >ome are plainly lecondary, vhtie otheit are ques- 
tionable; and not a few are variations or dtrferentiated forms of one another. 
Thnt, there are roots ehowliig respectively r and 1, as rabh and labh, 
mmo and mine, kfar snd Icfol; roots with and wltbout a atrengthening 
naial, as vand and vad, maud and mad; roots in & and In a nasal, as 
khft and kbao, g& and gam, j& and Jan; roots made by an added i, 
as tT& from t^, mii& from man, pB& from btaae, yi from 1; roots the 
prodoct of rednplicatloD, as Jakf from gbas, dndb from dtaQ; roota with 
a final sibilant of formative origin, as bhakf and bhlkf from bb^, 
iiak^ from naQ, qruf froin qrn, hfta from hK; root-forms held apart by 
a well-MtabUihed discordance of inneotion and meaning, which yet are 
probably different sides of one root, a« k^f drag and ^j^ plough, vid knoui 
and viA ^nd, vx tndo/e and vf choote; and so on. In many such cases 
it la doubtful whether we ought to acknowledge two roots or only one; and 
no absolute rale ot distinction can be laid down and maintained. 

108. The list of roots given by tbe Hindu grammarians contains 
aboQt two thousand roots, wltbont inctndlng all those which students of 
the language are compelled to recognize. Considerably more than half of 
tbls number, then, are nn authenticated by nsa; and although some of 
those may yet come to tight, or may have existed without finding their 
way into any of the preaerved literary docnments, it is certain that most 
are flctitlODs.- made In part for tbe explanation of words falsely described 
as their derirativea, but in the main fOr unknown and perhaps on discoverable 

ioy Google 

103—] III. Euphonic ComtHA-noN. 36 

a. The looU aniiDthentieited by tnee4ble oee will be mide no tceoant 
of In tbii gramiDir — or. If Doticed, irlll bs apeoifted &s of that cbmetar. 

104. The fonDs of the roots as here nsed will be found to differ 
in certain reepecta from those given by the oatJTe gnmmuiAna and 
tdopted hj some European works. Thoa: 

a. ThoB« root! of *htcb tbe Initiil n and s are regululy cODTeited 
to 9 and f iftet certiin pieflses are by tbe Hinda grunmariani given u 
beglDOing iritb ^ and q; no westeni anthoiity follows this eiample. 

b. The HindQi clsiBify u Blmpls tooM a namber of dariTad »t«m>: 
redoplicated ones, is dldM, J3gr, daridrS; preaent-stamg, u fiipn; and 
dsnomlnatiTG itemt, aa aTAdhlr, kom&r, eabbag, nuuitr, sSatr, arth. 
and the like. Tbeie are In European workt generally reduced to tbeir 

o. A nnmbei of roala ending in an S which ia inegnlarlj treated in 
the precent-Eystem are written in the Blndu list* with diphthaoga — e or 
Si or o; here they will be regarded as &-rootd (see 261]. The o of snch 
loot-foims, eapeeially, i« pnrely arbltruy; no forms oi deiivatives made 
from the loota Jnitlfy it. 

d. Tbe roots ebowlng interchangeably x ^"^ ^ e"^ IX ot va and 
or (&4a) are written by tbe Hindus with f or with f, or with both. Tbe 
f here alee it only formal, intended to mark the roots as liable to <ett*ln 
modlBoations, since it aowbere sbowa itself in any form or deiiTatife. Snob 
roots will in this work be written with 7. 

e. The Toote, on the other hand, ebowing a Tariatlon between x i^d 
ar (rarely ra] as weak and strong forma will be here written with j, as by 
the native grammarlini, although many European anthorities prefer tbe other 
or strong form. So long as wo write the unnrengthened vowel In vld and 
ql, in mud and bho, and tbeir like, consistency seams to require that we 
write it in B^ and kf aleo — in all oaiei alike, withoat refbienee to what 
may have been the more OTiginsl Indo-European form. 

106. In many cuei of roots showing mora than one form, the selection 
of a repreaontativa form la a matter of compantlve indifference. To deal 
with ancb casea according to their historical character ia the part rather of 
an Indo-Enropean comparative grammar than of a Sinakrit grammar. We 
mast be content to accept as root* what elementa aeem to have on the 
whole that value in the existing condition of the language. 

106. Stems as well aa roots hitve their variations of form (311). 
The Hindu grammarians nsnallj give the weaker form aa the nonnal 
one, and derive the other from It by a fitrengthening change; some 
European authorities do the same, while others prefer tbe contrary 
method; the choice is of unessential consequence, and may be deter- 
miDed in any cue by motives of conveaieoce. 

107. We shall accordingly consider first of all, Id the present 
chapter, the euphonic principles and laws which govern the combination 

Dijiiizec ..Google 

37 Introductory. [ — 109 

of tbe elements of words uDd of words as elemenls of thu sODtenoe; 
then will t>e taken up the subject of inflectloii, under the two heads 
of deelenaloii and conjugation; and an aotount of the elasses of 
unimfleeted words will follow. 

ft. The formation of conjngational stems [tense and mode-stems; 
also participles and infinitive) will be tangbt, as is nanal, in connection 
with the processes of co^jngational inflection ; that of nninflected 
words, in connection with tbe variong clsBsas of th<Me words. But 
the general anhject of derivation, or the formation of declinable stems, 
will be Uken op hj itself later (chap. XVII.); and It will be followed 
by an account of the formation of compound stems (chap. XVIll.]. 

108. It is bj no means to be expected of beginners 
in the language that they will attempt to mastei the lules 
of euphonic combination in a body, befoie going on to leain 
the paiadigms of inflection. On the contrary, the leading 
paiadigms of declension may best be learned outright, 
without attention, or with only a minimum of attention, 
to euphonic lule. In taking up conjugation, howevei, it 
is practically, as well as theoretically, better to learn the 
forms as combinations of stem and ending, with attention 
to such laws of combination as apply in the partioalar cases 
oonceined. The rules of external combination, governing 
the make-up of the sentence out of words, should be 
grappled with only when the student is prepared to begin 
the reading or the formation of sentences. 

Principle* of Euphonic Combtnation. 
108. The rules of combination (aadidhi putling together) 
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 eadings to roots and stems; 

b, to the more external putting together of stems to 
make compound stems, and the yet looser and more accidental 
collocation of words in the sentence; 

o. Hence they are usually divided into rules of internal 
combinatioD, and rules of external combination. 

ioy Google 


110. Id both oUwes of oaae^ howtvn, the genenl prinoipl«B 
of combinadon are the same — and likewise, to a great extent, the 
specific rules. The diffeieDcea depend in part on the oconrreiioe or 
Don-oootmence of certain combinatiODB in the one class or the other; 
in part, on the difference of treatment of the same soand as final of 
a root or of an ending, the former being more persistent than the 
Utter; in part, on the occurrence in external combination of certain 
changes which are apparently phonetic but really historical; and, most 
freqoent and conspicnotts of all, on the fact that (1B7) vowels and 
semivowels and nasals exercise a sonantising influence in external 
combination, but not in internal. Hence, to avoid nnnecessair repe- 
tition as w^l as the separation of what really belonga together, the 
rules for both kinds of combination are given below in counectioQ 
with one another. 

111. a. Moreover, before case^endlngs beginning with bb and a 
(namely, bbyftm, bhis, bhyaa, su), the treatment of the finals of stems 
is io general the same as in the combinations of words (pada) with 
one another — whence those endings are sometimes called pada-end- 
ings, and the cases they form are known as pada-cases. 

b. The Importanoe of thi* distinction ii lomewhat eiKggeT4ted by the 
ordlnuf slatement of It. In fact, dli Is the only lontnt mnte initiftl of id 
ending ocoarrlng in conJugatioD, u bh in declenBion; and the diffeience 
of their treatment is in put ovlDg to the one coming Into oolllsion OBtully 
mth tlie flnsl of a root ind the olhei of an ending, and in pait to the fact 
that dh, u a dental, is mote ustmllible to palatals and UsgaaU than bh. 
A more marked and pioblematio diatlnotlon ie made between Btl and the 
verbal endiagi Bl, Bva, etc., eipeclally after palatal Bounds and ^. 

0. Further, before certain of the suffixes of derivation the final 
of a stem is sometimes b«ated in the same manner as that of a word 
in composition. 

d. This is eipeclallr tie ease before leoondu? sufllxea having • 
markedly distinct otBoe, like the poBaeaaive nant and vant, the abitract- 
miking tva, the anffli of material m^a, and ao on; and it la mech 
more frequent in the later langnsge than in the eailler. The examples are 
sporadlo in character, and no role can be given to cover them: for details, 
see the varione infflies, in chap. XVIL In the RV. (as may be mentioned 
here] the only examples are vidy^nmant (beside garutmant, kabiid- 
mant, etc.), p^advant (beside datvinti mar&tvant, et«.), dbj^adTln. 
(beside namaavfn, etc), ^agmii (beside ajmi, idtunii, etc.), mrnmifa 
(betide manasmj^a, etc.), and ahaibyd, kiiiiTU, f adiyd, and afibojA, 
davoy&, Aak^dboyu (beside namaayia, vaoaay^ etc.); and the AV. 
adds only sAtaoran (ItV. sahivaa). 

IIS. The leading *Tales of internal combination (as already stated: 
108) aie those trhich are of moat immediate importance to a beglnnei in 
the langasee, aince his first task ts to muter the principal pandigms of 

ioy Google 

39 Gbkeral Princiflbs. [—117 

inflection ; the rules of Bxteniii comblnttioD may better be left DnUaohed 
until be comet to dealing vitb words iti tenteacM, or to tisniUtlng. Then, 
howeTer, they ue indltpeuMble, since the proper form of the voids that 
compoM the soDtenee it not to be determined irithont them. 

a. The geoenl principles of oombinaUon onderlyiog the euphonic 
rales, and determining their olaeslGoation, may be stated as follows: 

118. Hiatus. In geneial, hiatus is forbidden; every 
syllable except the initial one of a sentence, or of a word 
.or phrase not forming part of a lenteaoe, must begin with 
a consonant (oi with more than one). 

a. For details, and for exoeptioDs, see 12B ff. 

b. In the earlier langnage, howerer, hiatos In evsry poiitlon was 
abondautlT admitted. This appeals ptalnljr from the mantra*, or metrical 
parti of the Teda, where In innnmerable tnstaaess y and v are to be read 
aa i end n, and, less often, a long vowel is to be resolved into two vowels, 
in order to make good the metre: e. g., TWjr&QSm haa to be read as 
vfirl>a^9a-&m, avagvyam as aa-aq-Ti-am, and ao on. In the BAhmsinas, 
also, we find trao, avar, dySUs described as dissyllables, VjrSna and 
satyam ss trisyllables, rl^auya as of four syllaUes, and the like. See 
further ISBe. 

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

116. Assimilation. The great body of euphonic 
changes in Sanskrit, as elsewhere, fells under the general 
head of assimilation — which takes place both between 
sounds which are ao 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 

110. In part, assimilation involves the coDveision 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 sod im- 
portaot occnr in the adaptaUou of sold and Boaant sounds to one 

ioy Google 

117—] III, EuPHOKio Combination. 40 

another; bnt the nanle nod 1 h&ve also in ceiUin cases their i^clat 
aasimilative influence. Thaa: 

ft. Id the tiro cluiei of ddu-dmlI matea sod splrasts, laid uti Bonsnt 
ue wholly isMtiipttlble ; no suid of elQieT dasB can eithei piecede oi follov 
a Booaat of either. 

b. A mute, surd or lonaDt, li wtlmilated by being chtnged U> ft« 
conNpondent of the other kind; of the tpiranta, the snid a U the only one 
having a sonant coneipondent, namely t, to *hleb It Is convertible In ex- 
teraal combloatlen (164 ff.). 

o. Tb« naiala are more freely oombinable : a nasal may either precede 
01 follow a mate of either kind, or the aonuit epirant ti; tt may alio follov 
a aord apiiant (albUant); no naaal, however, ever piecedea aslblUnt in the 
Interior of a word (It ia changed Instead to amUTira); and tn exletnal 
oombinatlon their concaTrenee la nanally aiolded by Inaertion of a inrd mnte. 

d. A. aemlTOWel haa atlll lesi aonantiilng Influence; and a vowel least 
of all: both are freely preceded and followed by sonnda of every other 
olaaa, In the Interior of a word. 

e. Before a sibilant, however, la found, of the semivowels, only r and 
very rarely 1. Moreover, in external oombinatlon, i la often changed to its 
anrd correspondent a. 


f. In composition and ienteDce-eollooatloa, initial vowels and aemi- 
Towels and naiala alio require the preceding Anal to be sonant. And 

g. Before a nasal and 1, the asalmllative procesa la sometimes eanied 
further, by the conieralon of a Snal mute to a nasal or 1 reapectlvely. 

118. Of ooQTersiona iDvolving a change of artionlate position, the 
moet important are those of dental aonnde to lingual, and, leas often, 
to palatal. Thna: 

a. The dental B and n are very frequently converted to ^ and q. by 
the aaaimilating Influence of contiguous or neighbouring llugaal aonnde : the a, 
even by aounds — namely, 1- and a-vowela and k — which have themselves 
no Ungaal character. 

b. A non-nasal dental rante la (with a few exceptions In external 
oombinatlon) made lingual when it oomei into collliion with a Ungual aonnd. 

0. The dental mutea and aibilant are made palatal by ■ contiguoua 

Bnt alio: 

d. A m [not radical) la assimilated to a following consonant, of 
whatever kind. 

Oi For certain anomalous caaea, see 151. 

119. The euphonic combiDatlons of the palatal mutes, the palatal 
ribilant, and the aspiraUoo, aa being sounds derived b^ phonetic 
alteration from more otiginal guttarals (4Sff.), are made peculiar 

ioy Google 

41 General Prikoiples. [—124 

Uid eompticatod by two circiirastftDces: their reverBion to a galtural 
form (or the appesnnce of the unaltered gnttoral iostead of tbom: 
48); and the different treatment of J itnd h according aa they represent 
one or another degree of alteration — the one tending, like o, mora 
to the gattnral rerersion, the other showing, like 9, a more sibilant 
and lingual character. 

ISO. The lingual sibilant 9, also of derivative character (from 
dental a), shows as radical final peculiar and problematic phenomena 
of combination. 

121. Extensioii and abbreviatioa of conao- 
nant-gioups. The native giammarianB allow or require 
ceitain exteneions, by duplication or inseitioa, of groups of 
consonants. And, on the other hand, abbreyiation of cer- 
tain other groups is allowed, and found often practised in 
the manuBctipts. 

122. Permitted Finals. The permitted occurrence 
of consonants at the end of a word is quite narrowly 
restricted. In general, only one consonant is allowed after 
the last TOwel; and that must be neither the aspiration, 
nor a sibilant, nor a semivowel [save rarely FT 1], nor an 
aspirate mute, nor a sonant mute if not nasal, noi a palatal. 

128. 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 different character, not consisting in the mutual adaptations 
of the parts, but in strengthening or weakening changes of 
the parts themselves. 

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

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

2. Bules as to permitted finals (since these underlii: the further 
treatment of final consonants in external combination). 

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

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

ioy Google 


5. Rnles for the conversion of dental Bonnda to lingokl and 

6. RdIos for the oh&ngOB of final nasals, Including those in wbioh 
a former final following the nasal re-appears in oombination. 

7. Rnlee regarding the special changes of the derivative sonnde 
— the palatal mutes and sibilant, the aspiration, and the Ungual 

8. Roles as to extension and abbreviation of consonant groups. 

9. Rules for Btrengthening and weakening processes. 
Everywhere, rales for more sporadic and less elaselfiable 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. 

126. The coDGuitenoe of two von els, or of vowel and 
diphthong, without interveoiDg 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 oonounent sounds into one, by the reduction of 
oae of them to a semivowel, or by development of a semi- 
vowel between them. 

a. For the not InrTeqoent etsat ot compoaitlon uid aentence -combi- 
nation In wfaleh Uie reoent Imi of a s ot y ot v belwMD vowels Isitoi 
> permanent hUtni, tee below, 1S2 ff,, 176-1; for ceitiln fluil towoIb 
which ate malntaiaed uDch&Dged In Benteace-combinatlon before an Initial 
vowel. Me 138. 

b. A Tery few woTdi in tbaii admitted written form show Inteilor 
hiatus; such ate titaii lime [perhaps for titaSQ, BR.), prAiiga vagon- 
pok (foT pr^uga?); and, in RV., BuOti. 

0. The testa of the older dlaleot ato wiltten according to the enpfaonlo 
rules of the later language, althongh In them (aee 118 b) the htatna is 
really of &eitnent occntience. Hence they ate not to be read aa written, 
but with constantly tecnrring reversal of the pioceuea of vowel-combination 
which they have been made artiflciilly to undeigo. See futlhet IBS e. 

d. Also In the later langnage, hiatns between the two p&das oi primary 
dlilsioDB of a metrical line Is tolerably frequent, and it la not nnknown in 
sporadic cases even la the Interior of a pBda. 

e. The mles of vowel combination, as regards both the resulting 
sonnd and its accent, are nearly the same in interna] and in external 

ioy Google 


126. Two sitnilai simple vowels, shott oi long, coaleeoe, 
and form the coiieBpondiug long vowel : thus, two B~ToweIs 
(either oi both of them short oi long) fotm E)T S; two i-vowels, 
^ I; two o-vowels, ^T Q; and, theoietioally, two f-voweU 
foim ^ f, but it is questionable whether the case ever 
practically occurs. Examples are: 

H ?T7lsI: sft oS 'prajatt (ca + aprajat); 

Srrfle< Rtl 'vft (ati + iva) ; 

H^H BQlttam (su-uktam); 

^IslTHifl^riys "sit (rBja + Srtt); 

CtcAlT^: adUfvara^ (adhi-lfvarab] ; 

^^tpCT juhUpabhTt (juliQ — upabh^]. 

a. A« the above examplea iadlute, It will be tho practice oTCTjrHhere 
in tbls work, In trtngUletattoD (not In the devanSgari teitj, to separate 
Independent words; and. if an Initial vowel of a folloning word has caaletoed 
with a final of the pieseding, thia will be iodtoated by an apoatiopbe — 
single If the initial lowel be the shorter, doable if it be the longoi, of the 
two different Initials which In everf case of convblnation yield the same reault. 

X27. An a-vowel combines with a following i-vowel to 
^ e; with an u-vowel, to iHt o; with fj x, to ^ ar; with 
51 I (theotetioally), to 95T al; with ^ e or^ Si, to^ 51; with 
^ o or ^ &u, to €t Sa. Examples axe: 

^IslK rajendra (rRja-indra); 

f^r|l4^5l : liitopade9at^ [hita-apadefalji); 

(f^fN: mabargl^ (iaali&-r9Mk] ; 

^ BSi 'va (sft + evaj; 

( Nut) H TBjfilfvaryam (rBja-ftifvaryam); 

)^^°hH: divSukasat (divS^okasalji) ; 

k( ^ l| ^IT J varSn^adham (jvara-ftu^adham). 

a. In tbe Vedic teste, the vowel f ie ordinarily written unchanged 
after the a-vowel, which, if long, la shortened; that, mahanM^ instead of 
maharfl^. The two vowels, however, are usually pronounced as one syllable. 

b. When saceesalve words like indra S ihi are to be combined, the 
first combination, to IndrB, la made Brat, and tbe resnlt Is indre " 'tat 
(not indrfti " lii, from indra e 'hi). 

ioy Google 

128—] III. Euphonic Coubimatiom. 44 

128. As legarda tbe accent of theaa vowel combiuatioiis, It 1b 
to be noticed that, 1. u a matter of coarBe, the uoiun of acute with 
acute yields acute, and that of grave with grave yields grave; thxt 
of otrcomBex with circumflex cannot occur; 3. a circumflex wldi 
following acute yields Kcnte, the final grave element of tbe former 
being raised to acute pitch; a grave with following acute does the 
same, as no upward slide of the voice ou a syllable is acknowledged 
in the language; but, 3. when the former of the fused elements ib 
acute and the Utter grave, we might expect the resulting syllable 
to be in general circumflex, to represent both the original tones. 
Pacini tn fact allows this accent in every such case; and in « single 
accentuated Brahmana text [^B,], the circumflex is regularly written. 
But the language shows, on the whole, an indispositioo to allow the 
circumflex to rest on either long vowel or diphthong as its sole btaie, 
and the acute element ie suffered to raise the other to its own level 
of pilch, making the whole syllable acute. The only exception to 
this, in most of the texts, is the combination of i and 1, which be- 
comes i: thus, dlvi 'to, from dlvl Ivft; in the T&ittiriya texts alone 
snoh a case follows tbe general rule, while d and a, instead, make 
&: thus, a&dg&tA from ad-odgStfi. 

12&. The i-vowels, the u-TowelB, and W r, before a 
dissimilar vowel 01 a diphthoajf, aie regularly converted each 
into its ovrn ooriespooding semivowel, TT 7 01 ? v 01 ^ r. 
Examples aie: 

\rii\^ Ity ShA (iU + ftbs); 

qfll^ madhv iva (madbu-|-iva); 

i'^'stif dDhitrsrthe (duhltfarthe) ; 

pmr 8try ssra (atrl+asya): 

^ vadhvU (vadha-U). 

a. But in internal combination the 1 and a-vowels are not seldom 
changed instead to I7 and nv — and this especially in monosyllables, 
or after two consonants, where otherwise a group of consonants 
difficult of pronnnciation would be the result. The oases will be 
Dotieed below, in explaining inflected forms. 

b. A radical i-vowel is converted into y even before 1 in perfect 
tense-inflection: so nlnyima (iiinl+ln[>a). 

0. la * few aportdlo cuas, i and a becomo iy and av ena In wocd' 
eompDsition: e. g., triyavl (tii+avi), viyafiga (vi + aOga), euvlta 
(BO + ita): Mimp*ie 1204 b, 0. 

d. Not ver; aeldom, the same word (eapeciall; u found in differant 
texts of the older Unpwge) hu more than one form, ibowiDg vuious treitment 

itizecy Google 


of >n i- or n-va«el : e. g. aror or auvar, tuivi or tonnTS, bodlmya 
ot budhnlyK, T&txyfd or Tatrlyfil. For tbe iDost part, doabtleu, theae 
ITS only two vijs of writing the iftme proDDncUtiaD, an-ftr, bodlmiA, 
and M on; and Gie tlliooTd>nc« bu no otheT importaiKie, hlMoiioal or phonetic. 
There la more or last of tlila difference of trsatment Of an 1- ot u-element 
after > eonioDiiit In all periods of tlie langnige. 

e. In the older langaage, there ia a marked difference, In leipect to 
the freqoencr of to wel- combination for aToldlng hUtiu as compated with 
that of non'Combinitlon and ransaqnent htataa, between the claai of caaei 
vhete two Towel-aonnda, limllar or dlaalmiliT, wonid eoaleaee into one (1S0, 
1S7) and that where an 1- oi U'lowel would be converted Into a aeml- 
vowel. Thai, in word-oomposilion, the ratio of the caies of eoaleaoed vowels 
to thoae of hiatne are in RT. rb Htb to one, In AT. aa nlnateen to one, 
while the cases of seniiTowel'CDnTeTslon are in RV. only one in twelve, in 
AT. only one In Ave; in sentenoe-comblnatlon, the ctsea ot ooaleacence 
are la both RV. and AV. abont se seven to one, while thoae ot semlvowel- 
eonverston ate In RV. only one in llfly, In AT. one in five. 

f. For certain cases of the loss or aasimilation of 1 and a before j and 
T respeetively, see S83 b. 

130. Ab regards the accent — here, as in the preceding cue 
(128], the only combination requiring notice Is that of an acute 1' or 
u- vovrel with a following grave: the result is circnmflex; and such 
cases of circumBex are many times more frequent than any and all 
others. Examples are: 

odfp vyuftl (vi-urti); ^PJJtlfd ftbhyBTcati; 

s^ nadySu (nadi-Su); 

fttiy arista (sii-l^ta); H*<in tanTas (taoa-aB). 

a. Of a aimllsT combination of acnte f with foltowing grave, only a 
alngle ease Itaa been noted in accented tezta: namely, vljaKtr dt4t (1. e. 
▼ijSftt^ et&t: ^B. xiT. 6. 6")', the accentoation is in accordance with the 
rulea for 1 and tL 

ISl. Of a diphthong, the final 1- or u-element is changed 
to its conesponding Bemivowel, IT y ot ^ t, before any vowel 
01 diphthong: thus, ^ e (really ai: 28 a) becomes Wf &7< 
and 3^ o (that is, an: 28 a) becomes Wi &▼; ^ Bi becomes 
Om Bjr, and ^ Sn becomes q)? Bt. 

a. No change of accent, of conrae, occnrs here: each original 
syllable retains its syllabic identity, and hence also ita own tone. 

b. Eiamplea can be given only for Internal eombloation, eince in eitetnal 
combination there are farther ebaagea; see the next psragraph. Thus, 

^ naya (ce-a); 'im n&ya (nU-a); 

q^ bbava (bbo-a); HR bb&va (btiBu-a). 

L,j,i,;ec.y Google 

ISfl— ] ni. EnpHOHic Combination. 46 

182. Id exteinal combinatioD, we have the important 
additional rule that the semiTOwel resulting from the con- 
version of the final element of a diphthong is in general 
dropped; and the resulting hiatus is left without fuithei 

188. That is to say, a final ^ e [the most frequent 
ease) becomes simply ? s before an initial vowel (except 
Q a: see 136, below), and both then remain unchanged; 
and a final ^ Si, in like manner, becomes [everywhere] 
^ B. Thus, 

fTsgrnm: ta SgatK^ (te + agatW>); 

^IT^ ^ uagara ilia (n^are+lha]; 

(TPTT *l(i(ld teams adadSt [tasmU + ada<Ut}; 

fn^l 3Wr^ BtriyS uktam [striySi + uktam). 

a. The liter grimmulana alloiT the 7 to such rombinitlon* to be etthet 
Mtttned 01 dTopped; bat the onl/oim pnctice of tbe msDoecripts, of eTery 
■ge, in secoidance Mth the strict requirement of the Vedic grsmmire 
(PritlfilthTu), ts to omit the BemiTOvel tnd leive the hiatn*. 

b. The percliteace of the hittua csoeed by this omlsBiDn Is ■ plain 
Indication of the compaTatWely recent loie of the iDtervening conaontntal 

o. Insl«ncee, howeTor, of the aioldanee of biatas by combination of the 
Femaiiiing Bnal vowel with the following initial eecording to the nsnai mlea 
■re mat with in eiery period of the langoafe, (rem the BV. down; hot 
they are rare and of sporadic character. Compare the similar I 
the hiatus after a lost Anal a, 170-7. 

d. I 
the MS., 

184. a. The dipfatbong o (except as phonetic alteration of final 
aa: see 176 a) is an nnaaaal final, appearing onlj in the Btem bo 
(861 o), in the voc- sing, of u-Btems (341), in words of which the 
final a U combined with the particle u, aa atho, and in a few inter- 
jections. In the last two olasses it ie nncombiaable (below, ISSo.f^; 
the vocatives BOmeUmes 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 ordinarily lose its final element, bnt 
remains gav or go. A final aa becomes a, with following hiatus, 
before any vowel save a (for which, see the next paragraph}. 

Dijiiizec., Google 

47 ' VowEi, CoKBiHA-noN. [— I8B 

b. The EI T of m^&T from ^ Ku is usually retained: 

rTI=R tSv era (tSu + era) ; 

3»Tri^=^T(]l ubtaBT indr&gnT (ubUu + ii^^Kgi^)' 

O. Id tb« oldei Ungiuge, howe*er, it U In lome texts dropped be- 
fore m n-TOWel; thus, ti abbSA; In otboi texts It la treited like &1, or 
loses ita n-element before every inltUI vonel : thns, ti evi, atihi In- 

1S6. After final ^ e or ^ o, an initial CI a disappears. 

»■ The reBnlting accent is as if the » were not dropped, bnt 
rather absorbed into the preceding diphthong, having ite tone dul^ 
represented in the combleation. If, namely, the e or o is gTare or 
dtcnmflex and the » acute, the former becomes acute; if the « or 
o is acute and the a grave, the former traoomes circumflex, as dsu- 
allf in the fusion of an acute and a grave element. If both are 
acQte or both grave, no change, of course, is seen in the result. 
Examples are: 

% '^RR te 'bniTan (t^ abruvanj; 

Ht 'iBlorifT ao 'braTlt [s&it abravlt); 

(^!yH°tll 'dn: hUusitavyo 'gni]^ (hlfiBitavyat^ agni^); 

Ui^"iil A(<^rl y^ indrd 'bravlt (ydd indra]^ ibravlt); 

U^ts)*^ Js<c(1H 7&d vS^&ay& 'bravlt (ydd rBjanyi]^ 

b. As to the use of tbe avagraha sign in the cue of snob *n ellalon, 
see above, 16. Id Ir&nilitention, the reveried apoBtrophe, or roogh bieitb- 
Ing, vlll be used in tbla work to lepreaent it. 

O. Tbls elielon or absoiptioa of InitiU a tftet Srisl e oi o, whlob In 
the Ister Isnguaige ia the innriabU rule, Is in the Veda only sn occaaloTial 
oecnrrence. Thua, In tbe BV., oat ot nearly 4&00 inatancea of snch an 
Initial a, It ia, as the metre ahowa, to be leally omitted only about seventy 
ttmesi In tbe AV., less than 300 timea ont of about 1600. In neitbet 
woili la there any accordsooe in respect to tbe combination in qnesUon 
between tbe written and spoken Form of tbe text: in BV., tbe a Is (u 
written) elided in more than three qnartera ot the cues; In AV., in about 
two thttda; and In both texts it is written in a number of inatancea where 
tbe metre rsqnlres its omission. 

d. In a few cues, an initial A is tbua elided, eapeolally that of 

e. To the rules of vowel combitution, as above stated, there 
are certain exceptions. Some of the more isolated of these will be 



noticed whore they come up ia the processeB of ioflectlon etc; a few 
require mentioo here. 

186, Id internal combination: 

»■ The aagment a makes with the iDltixl vowel of s root the 
combinations &1, ia, &r [vpddhl-TOwelB: S3S), tnstetd of e, o, tat 
lea^ft-TOwels), as reqaired by 137: thus, Uta (a-)-ltA) KubhnU 
(a-f-nbhnU), trdhnot (a+fdlinot). 

b. The flnal o of > item (1S03 a) bsMimel av befoia the snrax ya 
(origiDtlly la: ISlOa). 

O. The flnil TDvel of a (tern ia often dropped irhen & e«eondir)r sofflx 
U added (laOSa). 

d> Foe the Teikenlng and loss of radical Toweb, and for certain InMi- 
tlons, iee below, S49 ff., SET-8. 

137. Id externitl oombination: 

a. The fimil a or & of a prupoBitioD, with initial r of a root, makes 
Sr Instead of ar : Thus, Krchati [&-|-rohatl), avftrohaU (aTa-|-r?liatl), 
np&rqati (^B.: upa-|-r?ati; but AV. upar^antl}. 

b. Ingtanoe* ue occaatonall; met with of a fliiil a or ft being loat 
entirely before Initial e or o: thna, in Terb'formg, av' aqyftmaa AB., 
np' eifatn etc. AV.j in deriiaiiveE, >s apetavya, apetfi in componnda, 
as da^onl, yatbetam, and (permiasibly) oomponnda with o^fha (not rare), 
Ota [not quotable), odaua, aa adbaro^fba or adbarSoftha, tUodana 
or tllftadana; and eien in sentence-combination, ii Iv" etayaa, aqvin' 
ova, yatb' oelqe (ill RV.), tv* eman and tv' odman B.; and always 
with the exolamaiton om or oihkftra. 

0. The form Qh from f^vah sooiBlimeB makes the heavier or vpldlii 
(336) diphthongal combination with s preceding a-iowel : thna, prftu^hl, 
akQjlnhiJ^ (from pra + a^M. etc.)- 

138. Certain final vowels, moreover, aie uncombinable 
(pragrhya), or mamtain themfielves unchanged before any 
following vowel. Thus, 

a. The rowels i, tl and e aa dual endings, both of declen- 
sional and of coDJugational forms. Thus, bandhQ fta&te imjtu; git^ 

b. The pronoun and (nom. pi.: fiOl); and the Vedio pronominal 
forms asmi, yufimi, tvi (493 a]. 

0> A flnil o made by oomblnatlon of a final a-yowel with the particle 
u (llSSb); thus, atho, mo, no. 

d. A float I of a Vedie locatlie ease from an i-stem (336 f). 

e. A protracted iual vowel C78> 

t. The flnal, or only, vowel of an Interjection, as afao, he, ft, i, n. 

g. The older language ihows occasional exceptions to these roles: tbns, 
a daal I combined with a following i, as nfp&ti *va; an a elided after o, 
as itho <«!) s leeitlTe I tamed into a semivowel, »s v6dy asylEm. 

itizecy Google 

Permitted Finals. 

186. The sounds allowed to occur as finals in Sanskrit 
voids standing by themselves (not in euphonic combination 
with something following] aie closely limited, and those 
which would etymologically come to occupy such a position 
are often variously altered, in geneial accordance with theii 
treatment in other circumstances, oi are sometimes omitted 

a. Tha virlety of oontODaDti that would ever coma at the end of eltbar 
ui inflected foim ai ■ derivative (tern In the luigaige 1b veiy imall ; nunel;, 
in tottM, onl; t (or d), n, m, b; in derivatlTe »tema, only t, d, n, r, S 
(>nd, la ■ few lare trardg, J). But simoit all eonaonMite occur as flnali 
of roots; and every root ta ilable to be found, alone or at 1a»t member of 
a compound, iu the chaitctet of a declined stem. 

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

a. But neithei f noT ) ever eettuily occarsi and f la rare (only as 
Renter sing, of a stem in f or ar, or u flnal of each a etem In compoeltlon). 

Thns, {adra, giv&;a, ikKri, nadi, dSta. oamA, Jana;lt^, &gne, 
QlvJrSl, viyo, agnt&. 

141. Of the non-nasal mutes, only the first in each series, 
the non-aspiiate suid, is allowed; the others — sutd aspirate, 
and both sonants ~- whenever they would etymologically 
occur, are converted into this. 

Thus, agnim&t for agnlm&th, auhtt for siihfd, virdt for vlrddh, 
trift^p for trlftubh. 

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

ThoB, dagh becomes dliak, budh beoomes bbat, and so on. 
The toot* exhibiting this change are stated below, 15B. 

b. There wae some qaestian among the EUndn grammarians at to 
whether the final mute i> to be estimated as of surd or of sonant quality; 
but the great weight of authority, and the l&Taiiable practice of the manu- 
tcrlpts, fsTor the surd. 

Wkitn*r, anmnu. 3. si. 4 

ioy Google 


142. The palatals, however, foim hete [om often else- 
wh6ie) an exception to the rules fot the othei mutes. No 
palatal is allowed as final. The ? o teretts (43) to its 
original ^ k: thus, m^ Y&k, ^'^^4 aAhomilk. The W oh 
(only quotable in the lOOt ^(S praohj becomes Z %■ thus, 
m^ prftf. The ^^i either lereits to its original guttuial oi 
becomes 7 ti in acootdance with its treatment in other com- 
binations {218): thus, ftq^ bhi^Ak, fsqi^virij. The g; Jh 
does not occur, but is by the native grammarians declared 
convertible to 7 t- 

143. Of the nasals, the ^ m and ^ n are extremely 
common, especially the former (T{^ m and H s are of all final 
consonants the most frequent); the in 9 is allowed, but is 
quite rare; 1^ fl is found (remaining after the loss of a fol- 
lowing ^ k) in a very small number of words [886 b, 0, 
407 a); ^fi never occurs. 

a. Bat the final m of s root \s changed to a (compare filfi a, 
below) ■ thus, akran from kram, figan, aJagaUf nBanlgan from gam, 
&ii&n from nam, ayBn from yam, pr&9&ii ^om qam ; no other cases 
are qnotable. 

144. Of the semivowels, the ^ 1 alone is an admitted 
final, and it is very rare. The 7 ' '^ C*^^ ^^ nearest surd 
conespondent, n a: 146) changed as final to visarga. Of 
n y and c^ v there is no occurrence. 

146. Of the sibilants, none may stand unaltered at the 
end of a word. The H b (which of all final consonants 
would otherwise be the commonest) is, like ^ r, changed to 
a breathing, the vissrga. The V 9 either reverts (48) to its 
original ^k, or, in some roots, is changed to 7t {^ accor- 
dance with its changes in inflection and derivation: see 
below, 218): thus, f^ dlk, but f^ vi\. The ^v is like- 
wise changed to r t: thus, ^ Mi pr&v^ 

a. Tlie chkDge of f to ( U of r*N ooeniMDCB : lee below, fise d. 

ioy Google 

51 pEaidTTED Finals. [ — 160 

b. Finiil ndictl a i» atii by tliB fnmmuiani to 1>e dhanged to t ; bnt 
no inia axtmple ot tlie MBTfettlon It qootabls: m« 108; and compiTS 

146. The compound ^ k; is prescribed to be treated 
as siinple ^f (not becoming S(i k by IBO, below). But 
the case is a rare one, and its actual treatment in the older 
language inegular. 

a. In the only BY. cusb where the k; hu s qaul-ndlMl chareECei — 
nunely an&k from an&kf, ind imyak from j/mrakf — tbe convanlon 
1« to k. AIw, «/ forait of tlia o-aorlit (lee 690], we hate adb&k. asrSk, 
■rUk, etc (foi adli&kf-t etc.); bot alio aprit, aySf, av«t> auKt ('°' 
apT&kf-t etc.). And BV. ba« twlca ajrls from j/yaj, and AT. tvlce BrSa 
from ynfi (wiDDgly lefened by BR. to yaraAa), both 2d ilag., wheie the 
peiNoal ending haa petbapi oiowded. ont the loot-llaal and tenae-Bign. 

b. The nnmeial QOf lix U perbapa better to be Tegarded aa fakf, vlth 
tta kf treated aa ^ according to tba accepted inle. 

147. The aspiration ^ ta is not allowed to maintain 
itself, but (like ^J and SJ 9) either reverts to its original 
guttural form, appearing aa ^T k, ox is changed to ^ t — 
both in accordance with its treatment in inflection: see 
below, 222. And, also as in inflection, the original sonant 
aspiration of a few roots [given at ISGbj reappears when their 
final thus becomes deaspirated. Where the ^ h is tiom 
original Udh (223 g], it becomes fT t. 

148. The visarga and anusTSra are nowhere etymolog- 
ical finals; the former is only the substitute for an original 
final H B or ^ t; the latter occurs as final only so far as 
it is a substitute for if m (218 h). 

148. Apait iiom the vowels, then, the usual finals, 
nearly in the order of their frequency, are : f^ 1 m, ^n, 
^t, ^ k, ^ p, Tf, those of only sporadic occurrence are 
^ 11, ^1, m 9; and, by substitution, - ili. 

ISO. In geneial, 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 
^ain the last, and so on, till only one remains. 

ug,i,.e.., Google 


a. Thns, tadontB beoomea tudant, and this tndan; ad*&o4 
becomeB ndafik(148), and this adoo; and aah&nUt (a-aor, 3d sing., 
of yctuuid [800 bj) is in like manner redaced to aohSn. 

b. But a nOD-nasal mate, if radical and not anffixal. Is ret^ned 
after r : thue, drk from drj. v&rk from yv^i, avart from Vvft, AmSit 
from ymfi, auh&rt from anb&rd. The case is not a common one. 

C. For rellei ot totmex doable flnsU, preieTved by the 1at«i Unpiage 
under the dUgulfe ot apparent eaphanlc combination*, lee below, fl07 IT. 

IBl. Anomilons conTeraioni of & flnal mule to one of Mother «lus 
are occistonilly met with. Eiimplea are : 

a. Ot final t to k; thna, 1. in a few woida that have aiinmed a 
apecial value ai partlelei, as Jy6k. Uj&k (beside t&iikt), fdhak (beside 
fdhat), p^ak, drllk; and of kindied rharactei is khSdagd&nt (TA.); 
3. In here and there a verbal form, as a&vifak (AT. and TS. Ka^.), 
dambb^ak (Apast), avifTak (Parash.], fihalak (VS. HS.; = Bharatj; 
3. in loet-dnals or the t added to TOOt-items (383 e), as -dh^k foT -dbft 
(Sutras and later) at the end of compounds, soqrAk (TB.), Pfk^U (ST.); 
and 4. we may farther note here the anomalons «flkfva (AB.) forintrra, 
V'ldh) and aTfikaam (AB.), and the feminines in knl from nueealinea 
in ta (1178d). 

b. Of final d or t to a lingual; thus, pad in Tedlo pa^bbia, 
p&t^bM. p^bl^a; upan&fbbyllm (QB.); vy »v&t (US. Ui. 4. 9; 
yva* thine), and perhaps &p([ "rill (MS.; or V^(^j?). 

O. Of k or J to t, in an isolated example or two, as aamT&t, &art, 
■rt^Taaft (TS. K.), and prayitau t^S. Tb.; AV. -k(u). 

d. In Taittinya taiU, of the final of annqt^fah and trlftdbh to a 
gnttoial: as, aatuff^ oa, trlftiigbblB, aiiaftugbIi;aB. 

e. Of a labial to a dental: In kttkdd for and beside kalc6bh; in 
Badiafdbhia (TS.) from v'eiPi >°d in adbbla, sdhhT^a, from ap or 
ftp (393). Excepting the first, those look like cases of disalmilatloni yet 
examples of the combination bbh are not very rare in the older language ; 
that, kaknbbbyftm, triift^bbhla, kakabbhai^^^ anugt^b bhi. 

f. The forms pratidhi^gaa, -ffi (TtUttiriya texts) from pratldi^ are 
isolated anomalies. 

162. FoT all the pTocesses of external combination — 
that ifi to S47, in oompositioa and sentence-collocation — 
a stem-final oi wotd-final is in general to be legaided as 
having, not its etymolc^cal form, but that given it by the 
rules as to permitted finals. From this, however, aie to be 
excepted the b and t: the various transformations of these 
sounds have nothing to do with the viaarga to which as 


53 DBASpntATioN. [—IBB 

finab before a pause they hare — donbtlesa at a com- 
paiativeljr recent period of phonetic history — come to be 
reduced. Words will everywhere in this work be written 
with final B 01 r instead of }/.; and the rules of combination 
will be stated as for the two more original sounds, and not 
for the Tiaarga. 

IBS. An aspirate mute is changed to a non-aspiiate 
before another non-naeal mute or before a sibilant; it stands 
unaltered only before a vowel oi semivowel or nasal. 

a. Sach ■ cue e*n only arlBS In inteinal eombinatiou, ainoe the 
ptoae««ei of eiternkl oombln&tlon praappote the ndacUon D( the uptrate 
Co > non-upii&te inld (IBS). 

b. Ptactic*U]r, »1m, the talea u to changes of aipliatOB coneeni 
almost only the lonant aipliatea, ■ln<»i the lanl, being of Ittec development 
and rarer ooenmnce, are hardly ever found In attnationa that eall for theli 

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

a. Bnt Id the maaneeripts, both Vedlc and later, an aaptiate mule 
It not teldom found written doHble — aipeolally, if It be one of tare occut- 
rente: for example (RV.), akhkholl, jijhjhati 

IBS. In a few roots, when a final sonant aspirate (a 
|(h, 17 dh, H bh; also^ h, as representing an original 3 gh] 
thus loses its aspiration, the initial sonant coosonaat (IT g 
or ^ d or ST b) becomes aspirate. 

a. That Is to lay, th« original initial aaplrato of aueh mota ia reetored, 
when Ita preBenee does not iDterfare with the enphoDia law, of compuatlvely 
recant origin, whleh (in Sanskrit a* in Qreek) forbidi a root to both begin 
and end with an aspirate. 

b. The roots which show this peculiar change are: 
in gb — daghi 

in h (for original gb) — dab, dili, dub, drub, dfftb, ffub ; and 
aUo K>^ah (in the later dealderatlie jlgb(k^); 

In db— bandb. bBdta, bodb; 

lu bb — dabb (bat only in the later deslderatiTe dbipsa tot whloh 
the older language haa dipsa). 

ioy Google 


O. The MOie ohang* ■ppeut when the law u to flnali c&oiei the los* 
of the upliitlon *t th« ead of the loot: Be« ibove, 141. 

d. But from dah, du2i, drota, uid guh ue foaud in the Vedi 
alto fonns without the leetored Initial ispiiate: thai, dakgfftt; adok^at; 
dodnkfa etc.; Jogukya; mitradrAk. 

e. The seme anelog? li followed by dadb, the abhrsTitted snlistitnte 
of tbe pfeaent-itemi dadll&, from )^dhft (667), Id some of the toims of 
coqjDgttion: thus, dhatUitui from dadh+thaa, adhatta from adadh-f- 
ta, adtaaddbvam from adadlL+dbTam, etc. 

f. No CMe U met with of the throwing back of an upiiatlon upoD 
oombiDation with tbe 2d aing. impt. aot. ending Hhl - tbna, dnsdbl, 
daddlil (BT.), but dhogdlivam, dhaddhvam. 

Surd and Sonant Assimilation. 

166. Uadei tluB head, there u especially one very maiked 
and important diffeteaoe between the internal combioations 
of a loot 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 — 

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

b. To thli Tnle there are gome eieepttoD*; thni, »me of the detivatlTet 
noted at 111 d; flnal d of a root befttie the participial lufllx na (dS7d); 
and the fonns noted below, 161 b. 

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

d. It haa been poloted oat above (163) that in the rnlef of external 
combination only edmlttad flnali, along with ■ and r, need be taken 
account ot, all othera being legarded a« lednced to theie befoie combining 
with Initlali. 

168. Final vowels, nasals, and ^ I are nowhere liable 
to change in the processes of surd and sonant assimilation. 

a. The r, however, hu a eorreapondkg surd in a, to which it la 
sometimes changed in exteroal oombiDStioD, under clrcum>tanoes that 
favor a surd utterance (178), 

ioy Google 

55 ASSIMILATIOII. [ — 161 

168. With the exoeptious above atat«d, the oollisios 
of BUid and soaant Bounds is avoided ia oombinationa — 
and, legvlacl}' and usually, by aMimilatiog the final to the 
following initial, ot by legiewive awimil&tion. 

Thus, in interntl oombiiwtlOD : itai, ktO, attMs, atU (yaA + 
Bi etc.} ; qagdlif, ^agdbTiua (Vf ale + dU etc.] j— in esternal combinatioD, 
ibhtkd ay&m, JyAg jiva, fi4 a^It^alh, trlqtdb kpi, dlg-goja, qa^- 
afa&, aroid-dbOmft, brh&d-bhanii, ab-Ji. 

160. If, however, a final wnant aspiiate of a root is 
followed by H t 01 9 th of an ending, the aasimilation is in 
the other direction, or prc^essiTe: the combination is made 
aonant, and the aapuation of the final (lost according to 163, 
above} is transferred to Uie initial of the ending. 

TbnB, gh with t or tb becomes gdli; dh with the same beoomes 
ddh, as boddbi (visndb + ta), roddhia (y^mndh + thaa or taa); 
bb with the same becomes bdh, as labdlii (lOabb+ta), labdb^ 

a. Moreover, h, as representlDg original gh, is treated in the same 
manner: thai, dncdlii, ddgdhum from dah — and compare r&4^ 
and ll^bi from mb and Uh, etc., SSSb. 

bi In thii eombi nation, u the lonuit teplradon li not lost bot tnnafened, 
the raftoratlcin o( the lnlti»l Mplntloir (165) does not take place. 

O. In dadb from ydU (186 «), tbo more normal method is followed; 
the db U made «nid, and the Initial Mptnted: thai, dbattbaa, dhattas. 
And BT. hM ^hntt*™ Initead of dagdbam from f'dagb; and TA. hii 
inttSm iaatead ot inddbsm &am yidb. 

161. Before a nasal in external combination, a final 
mute may be simply made sonant, ot it may be still further 
assimilated, being changed to the nasal of ita own class. 

Thus, either tAd nAmas or tAn nAmaa, vig me or vafi me, bAi} 
mi^hij" or bi-tf mabin, tri^t^b n&nAm or trii^fom a&nAm. 

a. In practice, the converiion into a natal ia almoat iUTariabl; made 
in the mannioript*, ai, indeed, it i) by the Fiaticakbyaa laqniied and not 
permitted meioly. Kren by the general gmnmarlana it ii required in the 
eompoDnd fA^i^vatl, and before mStrft, and the annix mar* (1826) : 
tbn*, TSDmAya, mfninAya. 

b. Etcu in Internal combinatlan, tbo same aisinHlatlon Is made In 
lome of the dedTatiTea noted at llld, and in the na-partleiplea (8ST d). 
And s few ipondle initancei are met with STen in TeiMnllection : thoa, 

ioy Google 


BtUnoU, sUAnajM (UB.; tor stlghn-), mpuilta C^QS.; tux mpdn-). 
jaflmajnuia (KS.; for jBgm-]; thMS, howevei (Uka the doable MpliUet, 
lB4a)i **^ doDbtlsM to be i^ectad u f&lie TOMllage' 

les. Before 1, « finftl t is not merely made BOOitnt, bat fullf 
MsimiUted, becomiog 1: thas, tAl labb&te, Maptun. 

165. Before ^ h [the case occurs only in external com- 
bination), a final mute is made Bon&nt; and then the ^ Ii 
may either remain unehanged or be converted into the 
sonant aspirate corresponding with the former: thus, either 
Hf4<t t^d hi or rri% tid dhi. 

a. In prutlM, tha Uttet method la ilmott liiTaiUbly followed; ind the 
gnmmeTlaiia of the Piiti^khyi pedod ue neuir nnenimosE in requiring It. 
The phonatle difference between the two U tsij slight 

Examples are: v^ «hat&^ f&44bota (^at+hotft), taddhiU 
Ctat+liltk), ann;tab VtA. 

Combinations of final H a and ^ r. 

104. The euphonic changes of H s and ^ r are best 
considered tt^ether, because of the practical relation of 
the two sounds, in composition and sentence-collocation, 
as corresponding surd and sonant: in a host of cases TT a 
becomes ^ r in situations requiring or favoring the occur- 
rence of a sonant; and, much less often, ^ r becomes H b 
where a surd is required. 

a. Id iotemal combination, the two are far less exchaogeable witli 
one another: and this class of cases maf beat be taken up first. 

166. Final r radical or qaasi-radical (that is, not belonging to 
an ending of dedvatioD) remains unchanged before both surd and sonant 
aonnds, and oven before ea In declension: tiios, plparfi, oatorthA, 
oat6r;a, pOr^i^. 

166. Final radical a remains before a surd in general, and usu- 
ally before a, as in qitsal, gftssva, Bsee, a9lna [the last is also 
written fi^fl^^u: 178j: but it is lost In &sl {yaa+ti: 636). Before 
a sonant (that is, bh] in declension, it is treated as in external com- 
binatioo: thus, ft^bhis. Before a sonant (that ts, db) in coiyagatioD, 
it appears to be dropped, at least after long ft : thus, 9&dhl, qag&dU, 
cakftdU (the onfy qooUble cases}; in edhi (v'ae+dbi: 636) the 
root syllable is Irregularly altered; but in 2d perss. pl., made with 
O-hvam, as adbvam, gtdlivani, ar&dbvain (861 a), vadhvam (v^aa 

ioy Google 

57 Fdui, ■ AND r. [—168 

eloHej, it 1b, on ftocooot of the equlvatence sod inteiohange&bility of 

dliv and ddhv (232), impossible to say whether the e i n_ omitt ed ot; 

converted into d. 

S. FtnU tadiol b Is Terjr nre; BY. (twice, both 2d. pars, tiog.) tretts 
iigbaa bom ygluis tn tlie wme muiner u >dt ordinary woid endlog 
Id ap. 

167. In a very few cases, final radical ■ before a is changed to 
t [perhaps by dissfmilation): they are, from y-nn dtctU (also sporad- 
leallf from vas thine, QB., and vaa clothe, Har.), the fature vatsy^ml 
and aorist &Tfttsam; from y'ghas, the deslderative stem J{ghatsa. 

a. For t M apparent ending of tba 3d sing, in a-ieibs, aee S66 a.. 

168. According to tlie gram id art ant, ths &nal B of certain other nioli, 
nied ai noan-itema, becomea t at the end of the word, and before bb and 
tm: thna, dbvaa, dhvadbhla, BradbhyaB, Bratsn. Bat genuine eiimplea 
of inch obange are not qnotabla. - 

a. Sporadic oaaea of a like converaion are found in tbeTeda: namely, 
mSdbbfa and m&dl)liT&8 from tn&: aqidbtala from of^i sv&tttvad- 
bhjas from aT&tavoa; av&vadbhla etc. (not qnotable) bom bv&vos. 
Bnt the actnality of the eontenlan here la open to grave donbt; It rather 
aeama the aubMltntion of a t-sCem for a a-atem. The aama li trae of the 
change of vSAa to Tftt in the decleniloD of perfect participles (468). The 
stem anafvah (404), from anas-vah, is anomalous and leolated. 

b. In the eompoanda duoahuii& (dus-^nft) and p&raoohepa 
(pams-qepa), the final a of the first member Is treated aa If a t (203). 

169. As the final consonant of dertvatiTe stems and of inflected 
forms, both of declension and of oonjngation, a is extremely frequent; 
and its changes form a snbject of first-rate importance in Sanskrit 
eaphony. The r, on the other band, is quite rare. 

a. The r ia fonnd as original final la oertaln case-formi of stems tn 
r 01 ar (368 ff.) ; In root-stemi in Ir and nr from looti in j (383 b) ; 
In a amall nnmber of other sterna, as B-vta, iluur and itdhar (bealde 
ihan and ddban: 480), ivix or dnr, and the Tedic v&dhar, oqar-, 
vaaar-, vaiuur-, frutar-, aapar-, aabar-, athar- (cf. 17do); in a 
tew particles, sa antAr, piftt&r, pdnari and tn the numeral oat&r 
(482 g). 

b> The enpbonlo treatment of a and r yielding preoUely the same 
result after all yowals except a and &, there are certain fonus with regard 
to whleh il Is nneertatn whether they end in S or t, and opinions dilTer 
reapecting them. Such are nr (or m) of the gen.-sbl. sing, ot p-ttems 
[371 o), and ua (oi or) of the 3d plnr. of verbs (B60e). 

.yj, Google 


170. a. The TT ■, ae ahead; notioed [146}, beoomes 
visarga before a pause. 

b. It is letained unchanged only when followed by 
rT t 01 ET th, the fluid mutes o£ its own olass. 

o. Before the palatal and lingual surd mutes — ^o and 
^ oh, ^t and ^fh — it is assimilated, becoming the sibilant 
of either claM respectively, namely KT 9 or ^ q. 

d. Before the guttural and labial surd mutes — ^,k and 
S kh, ^ p and m ph — it is also theoretically aasimilated, 
becoming respectively the JlhTBrnOUya and upadtunSnlya 
spirants (69); but in practice these breathings aie unknown, 
and the conversion is to viaa^ia. 

Examples are : to b. tatas te, oak^ua te ; to o. tataq oa, taoTK^ 
oblyft; pSdaf (nlatl; to d. nelalfy kflimam. purofa^ kbanaU; ya^a^ 
prSpci, v^kfa^ phalavKn. 

171. The first three of these rules are almoat ai^versftl; to the 
last ons there are nnmerons exceptions, the atbilant being retained (or, 
b; 180, converted into f), espeolallj in compounds; but also, in the 
Veda, even in sentence combination- 

a. In the Veda, the reteDtlon of the Blbllftnt la compoandt \a the generU 
mle, th« aieeptloDi to vhleh ate datalled In the Tedlc giunmua. 

b. In the Utoi langnige, the letentton li nuinly detennlsed by the 
Intlmaoj' OT the uitlqiilt; and fiMqoenoy ot the comblnatloo. Thna, the flnal 
afblUnt of « pTeposltion or ■ void fllUng the offloe of a prapodtlan before 
• TSTbal root la wont to be pieaetTed ; am) that of a item before a derivatlia 
Df ykf, before pati, befors fcalpa and kftma, and lo on. Eiamplea ue 
namaakara, vftoaspatl, Syu^Sma, payaakalpa. 

o. The Vcdlc retention of the albllant In aeDtence-coUocatlon ia detailed 
in full In the Piati9il[hyai. The chief claaies of oaaea are: 1. tbe final of 
a ptepoaitloD or Its like before a Terbal form; 2. of a genitive before a 
gOTemlng noon: ai dlv&S pntrA^ if&B pad6; 3. of an ablatlTe before 
p&ri: aa Um&vataa p&ri; 4. of other lesa claaaidable oaaea: aa dyftof 
pit^ trig pQtvlt. yia p&t^i. paridhlq p&tBU, etc 

172. Before an initial sibilant — ^9, ? ti, H a — H a 
ifl either assimilated, becomii^ the same sibilant, or it is 
changed into visorga. 

a. The DatlTB gramnuiluu ate In lome meaanie at variance («•« 
AFr. il. 10, note) aa to vhloh of theie changea (hoold be made, and tn 



p«t their kUov either al pleuni*. The atage of the nuniuulpta U xlio 
dlieoidtnt', the coiiTeitlon to Tiasrga la tbe pTeralent pnotlaB, thoogh the 
■IblUat la alao not Inbeqnently fonnA written, eipeclally In Sonth-lDdlan 
nuuinaetlpti. Eniopeu edltora geQwallj write visarga; bat the later 
dletloiiaiiea and gloturiaa general!; mike the alphalietlc place of a woid tiie 
aams la if the iibilant weie luad initead. 

Examples are: toann]^ avByam or mutoB Bv&yam; indra^ ^Oth^ 
or indrttf qOia^; tMf ^a\ or tfif qB(. 

173. There are one or two escepUooB to these rnleB: 

B. It the initial sibUaat hae a auid mute after ll, the Qnal a may be 
dmpped altogether — and by Boma anthoTlUea li Teqnlred to be *o dropped. 
Thna, T&yftTS etlut or vSyBTK^ Btha; ontaBtanbu or oBtu^tan&m. 
With regard to thii point the osage of the different miDuaorlpti and edltioua 
Is greati; at Taiianoe. 

b. Before ta, tbe B I* allowed to become Ttsarga, instead of being 

174. Before a sonant, either vowel oi consonant (ex- 
cept ^ t: see 179), ? a is changed to the aonant ^ r — 
unless, indeed, it be preceded by St a or ^ S. 

Eiamptes «re: derapatlr iva, fiir iva; manur gaoohati, tanOr 
apaa; sraefr ajanayat; tayor adntabftmal?; BarvUr gtifU^; agner 

a. For a few caaea like d&4^B, dO^iBfa, Bee bedow, 199 d. 

b. The ezdamation blioe (4fi6) loaea Ita e before tokoIb and tonant 
conaonanta; tbai, bho nBlqadha (and tbe a 1* lometlmei fonnd omitted 
also before anrda). 

o. The endings QH as and ?TH Ba (both of which are 
extremely common] follow rules of their own, namely : 

176. a. Final i^ as, before any sonant consonant and 
before ahort SI a, is changed to lat o — and the 9 a after 
it is lost. 

b. The reinlting accentuation, and the feet that the loas of a ta only 
oeeaaioDal In the older language of the Teda, hiTe been pointed ont abOTe, 
136 a. e. 

Examples are: nalo D&ma, bmbma^ro vedaTit; manobhava; 
bantaTjo 'oml; anyonya (anyaa -f anya), yaqortham (ya^aa-f- 

o. Final W\ as before any -other vowel than q a loses 
its H s, becoming simple ^ a; and the hiatus thus occasion- 
ed remains. 

ioy Google 


d. That U lo uy, the o from sa li tt«*ted u ui oilgiaal « U traat«d 
In the lams ritoitloii: Me IS9-3. 

Eiftrnples are: brhadagTR avBoa, tditytt iva, n&mftfibti, 

176. ExcaptioDB to the rnlea as to final oa are: 

a. The uominatiTe mucallne pionoauB B&e and ef&s and (Tedlc) 
ny&s (496 a, 498 a, b) laie thsli b befote any aoiiKiDant: thoa, sa 
dadarqa ha taic, a^a pora^ah Ihit mani bnt ao travlt ha »aid, 
poni^ efa^. 

b. Initancea ate met with, both In the eaiUei and In the Utei lan- 
guage, of elbcement or the hiatni after alleiaUon of oa, by combtnatloD 
of the Temainlng final a with the following Initial Tovel.' thne, tato 
V&oa (tatas+nv&aa), payo^^i (payaB+afql], adHfiaana (adhas + 
Saana): compuo lS3a, I77b. In the Veda, anch a eombinaUoD Ii 
aomettmea abawn by the metre to be leqnlied, though the written text haa 
Iba hiatus. But aa tn BV. la in the gieat majority of caaee eombiued with 
the fallowing Towel; e. g., ai 'd tor si {d, ai 'smU for a& aamKi, 
afta ';a(UiI^ for s& 6fadl^; and almtlar exunplea are fonnd alto In the 
other Tedic texts. 

o. Other ipoiadlc Inegnlarltle* In the treatment of final aa oecot. 
Thna, It la changed to ai instead of O once fn RT. In av&s, once In ST. 
In &vaa (RV. &vo), once In MS, la dambhlfaa; In bhnvaB (aecond of 
the trio of aacred uttarancea bhas, bhavas, Stbt), eieept In It* earlleat 
ocmrrenoea; la a aeilei of words In a Brabmina puaige (TS. B.), vli, 
Jlnvir, ugr&r, bhimiir, trefir. Qmt&r, bbQt&r, and (K. only) pQt&r; 
In Janar and mabar; and aome of the ar-atema noted at 109a are perhapa 
of kindred character. On the other hand, aa la seTsral timet changed to o 
Id BV. befote a anrd conaonantj and e&a twice, and f&B once, retaini Ita 
final sibilant In a like poaiHon. 

d. Ill MS., the final a left before hiatal by alteration of either aa 
(o) or e (133) is made long If itielf nnaecented and If the following initial 
vo*rel U accented: thua, alirS 6ti (from 8lliTaa-|-6tl)i nirapriitft CndrSya 
(from •r&teH-inclOi '"d alio k£r;& 6ka- (from kSr^«, became Tirtnally 
kiHaa); bnt Sdltyi indra^ (fiom &dlt7&8 + Indra^). et&ftara (from 
ati + ftare). 

177. Final 91? Bs before any sonant, whether vowel or 
consonant, loses its CT a, becoming simple ^ B; and a hiatus 
^U8 oocaaioaed remains. 

a. The maintenance of the hiatna la tbeee eaaes, a] in that of O and 
e and fti (above, 133-4), aeemt to Indicate a recent Iobb of the intermediate 
toond. Oplniona ue divided aa to what thla ahonld have been. Some of 
the uatlTB grammariana atalmllate the eaie of Sa to that of U, aunmiiti; 

ioy Google 

61 COMBINATIOKS OF FiNAL r. [ — 180 

the eoDTenlon to ^ Id both «11ke — bnt probtbly odI; m a mtttw of 
formil eoD^eiilenoe In nile-maUng. 

b. Heis, too (u in the *lml|ir etioa of e >nd U and o: 188 o, 
178 b}, there «e eumplM to li* found, both eiTllei mi litei, at effMemsnt 

of the hUtu. 

178. Final ^ r, in general, shows the same form which 
n B would show under the same conditions. 

a. Thus, it becomes viaaisn when final, and a sibilant or Ttsarga 
before an initial surd mate or sibilant (170): thus, mdati pnnal^ 
dvSa tat, BV^ oa, oatuqoatv&riAfat; and (lllo,d) prttastina, 
antastfa, datu^tara. dhOatra; prSta^ karotl, anta^pftta. 

b. But original final r preceded b}' a or & maintaiDs itself nn- 
changed before a sonant: thns, punor eti, prfttarjlt, ifcar Jy6tl^ 
Uiir dimni, TKrdbi. 

o. The T li pre»ar>ed onebuiged btsu bafore & inrd in ■ Dumber ol 
Tedio oompaamlt: thna, abarp&tt; evJUFcanas, BTkroak^aa, ar^rpatl, 
erarf ^ svarf &ti ; dhOrf&d, dhur^ah ; pArpati, vftrkary&, S^trpada, 
punartta; >nd in some of these the r la optlonslly teUlned In the latei 
liogiuge. The RT. alio hu ftvar t&mal^ once In eantencB-comblnktion. 

d. On the other hind, And ar of the TSib-fonn fivar i« ohtnged to 

before a lonint In lOToril cum In RV. And r !■ lost, like s, In one 

01 two cues in the same test: thai, aJc^K {ndu^ itha vrk. 

179. A doable r is nowbece admitted: if snch wonid occur, eitlier 
by retention of an ori^al r or b^ conversion of ■ to r, one r is 
omitted, and tbe preceding vowel, If siiort, is made long by oompen- 

Thus, pnnft rainate, nrpatl r^att, mttA rUiAo, J70&«tlia, 

a. In lome Vedlo text#, bovever, there are InitaDcea of ar changed to 
o before Initial r: tbni, wrb rohftva. 

Conversion of ^a to i3[^9. 

180. The dental sibilant n a is changed to the lingual 
(T J, if immediately preceded by any vowel save 5 a and 
53? 5, or by ^ t or 5" r — unless the H a be final, followed 
by J r. " 

A. The »«iimilat!ng InlliieDee of the preceding llnKoal vowelt and 
■•mlvoirel Is obTions enough ; tbst of k and the other *oveli appear* to 
be dne to • somewhat retracted position of the tongne In the mouth during 

ioy Google 


tbeir atteiuice, uulof It* tip to Teach tti* roof «t th« month men euUr 
»t ■ point fuithST back thaa tbe dental one. 

b< The genenil Hindu gnmmir pieaeribei the tune change tflet a 1 
•leo; but the Pritifikhyai glTS no tnch nile, and phonetlo eonddwitlani, 
the 1 being a dental sonnd, are abiolntely agatnat it Actual caiei ol the 
combiDalion do not occot In the older language, not h*Te anj been pointed 
out In the latei. 

o. Tho TOweU that cause the alteration of a to f may be oalied 
for bieTitf'a sake "alterant" vowels. 

181. Henoe, ia the interior of a Sanekrit word, the dental ■ ii 
not naaaUy found after any vowel save » and &, bat, instead of it 
the llngnal-f. Bnt — 

a. A following r prevents the conversion: thos, lurs, tiarsa, 
tsmlara. And it ia bnt seldom made in the forois and derivativea of 
a root contaioing an r-element (whether r or j), whatever the position 
of that element: thus, aiaartl, alartam, B(u1srP&, tiotlre, pariardt. 
To this mie there are a few exceptions, as vlqt^. vl;tBr&, nifffts, 
vifpO'^li'M' givli}tlilrft, etc. In iOi4>'<ui the final ; of a root is 
preserved even immediatelj before r. 

b. Thli dlaaimilatlng Inflnenee of a foUowing r, ti compared -with 
the invariable Miimilating laSnence of a preceding r, it peculiar and prob- 

0< The leennencB of f In lueceHive eyllablaa Is aometlniea avoided by 
leaving the formers anchanged: thas, slaab^, bnt sifakti; yBoialftliia, 
but yM^TmaJli Similarly, in certain deilderative formaUona: aee below, 
184 «. 

d. Other eatea are sportdlc: RV. haa tbe forme Bisloe and siaiona 
(bnt slfieatofl), and the itemi fblsa, klBti, bfaa, buak, bfBOTft; a 
lingle root pla, with Ita derivative peanka, is fennd onee in (B. ; MS. 
hu mramrfi; m&aala begtna to be found in AT.; and auch casei 
grow more nomerani; for podis and the looti nifis and hlim, aee below, 

18S. On tlie other hand [as was pointed ont above, 68), the 
occurrence of q in Sanskrit words is nearly limited to coses falling 
under this rule: others are rather sporadic anomalies — eioept where 
f is the prodoct of g or kf before a dental, as is draftoni, oaftOt 
tvaftar: see 818, 981. Thus, we find — 

a. Fotir roots, kaf, laf, bht^ bhSq, of vrhlch the lait la commoD 
and ii found aa early ai the Brahmanaa, 

b. Further, in BY., ifo, kav&^a, ca^fla, o&fa, J&Ufa, ptt^yk, 
bafk&ya, v&qaf (for vak^atP), kk^fhS; and, by aDOmaloni alteration 
ot otlginal a, -;ah (torj^^h 'tc.), &4&4^a, apaftdt, and probably apiffhi 
and afthlvint. Snch eaaes gtow more common later. 

o. The nomenl fa;, u already noted (149 b), Is more probably ^akf. 

ioy Google 

63 COHTEXBIOH or ■ TO f. [— 18S 

1SS> The DMatiiKtiOD of tha alterant vowel — or, in other words. 
Ha being followed bj unuvba — does not prevent its altering effect 
upon the afblUnt: thae, haviA^ para&qi. And the alteratloD tiUcea 
place in the Intdal a of an ending after the final a of a Btem, whether 
the latter be regarded u also ciiangod to « or aa oonrerted into 
visargft: thus, bAvi^u or havi^qti, pftrnffn or pam^^a. 

a. Bat llu a of ptufaa (8B4] remalm imebuigsd, appuently on 
tAconnt of the ratiined senie of Ita nlns u pmcB; alio thtt of yhliiB, 
beuoH of iti nine u Una (Unaati Mo.); yoUm (BV. only] is more 


184. The principal oaBea of alteration of ■ In internal combination 
are these: 

&. In endings, inflectional or derivatiTe, beginning with a — thus, 
au; at, a*) era; a of Bibilaot-aorlst, fntnre, and desiderstiTe; saffixes 
ana, ■nn, aya, etc. — after a final alterant vowel or consonant of root 
or stem, or a onion-vowel; thns, Jnhofl, qefe, anUfam, bhavlfy&ml, 
f u^rflfo, defi^ Jlf90f vUcfo, ak&rfam. 

b. The fioal a of a stem before an ending or anf&x: thus: havl^ 
havifaa, etc., from havls; fak^ofmant, qoolfka, nUlnnfa. mannfya, 

o. Bo«t« hKTlng > fln*l alblUiit (eieepl q) sfter in tlterant TOwel ue 
— with the eEteption of fletittooi onei and pia, nUiB, iiliiM — regarded aa 
ending in ;, not a; and eoneemlng the tieatment of thli q In cDmbinattoD, 
see below, S2B-6. 

d. The initial a of a root after a reduplication: thus, aiqyade, 
BOfTSpa, slffisatt, ooqlcayata. aanlfva^at. 

O. Excepted la In geneial an Initial radical a lu a deildeiatlTe item, 
when the deatdentlTe-tlgn becomea f : tbna, aiaipqati bota yof, alaafikfatl 
from l^aafij. And there aro other aeatterlng caaea, aa treaOB (perf. from 
t^traa), etc 

ISB. Bat the aame change occars also, on a considerable scale, 
in external oombiDaUon, especially in oompoBition. Thus: 

a. Both in verbal forms and in derivatives, the final 1 or u of a 
preposition or other like prefix ordinarily lingualizes the Initial a of 
the root to which it Is prefixed; since anch combinations are both of 
great frequencj and of peculiar intimacj, analogous with those of root 
or stem and affix; thus, abhlqio, pratl;tl^ nlfLkta, v{;ita; anu- 
(Tadh&m, anqdka; the cases are nnmberleas. 

b. The principal ezceptlona aro Id acowdance with the priaclplea 
already laid down: namely, when the root contalna an r-alemeut, and when 
■ reenirenee of the libiUnt would take place. Bnt there ate alio othoia, 
of a mors Irregular iharaeter; and the complete acconnt of the treatment 
of Initial radical a after a prefix wonld be a matter of great detail, and not 
worth gtving here. 

ioy Google 

18B— ] III, Euphonic Coubinattoh. 64 

o. Mot Infreqiisntly, the InitUl S, utntUy ftlteied (fter a cettUn 
prefix, Tettins the altered albiUnt evea aflet m interposed a of augment 
or rednplleation: thus, aty aqth&t, abhy a^tblbn, parr a^ftfrajat, vy 
n^aliaiita, ny o^adilma, ulr a^thSpayan, abhy a^ifLoan, vy a^Jabb- 
nlt; vi ta^tba, vi ta^^tiivo. 

A. Harh mom anomilODi is the oecisional alteratioQ ot iiiltlsl rsdiul 
a after in a-element of a prefix. Saeh cues ue ava ^fambb (ag^nst 
ni Btambh snd prati atambh) aad (sccordiog to the grimmsilins) ava 

186. Id other compounds, the final alterant vowel of the first 
member not infrecjueDtly [especially in the Veda) lingualizes the 
initial a of the second: for exnmple, yndhl^thlra, pltn^asf, go^fhi, 
agnl^fomfc, anoqtubh, trl^aibdlil, dlvl^d, parame^t^n, abbifeni, 
pitff&d. puTUftati. 

a. A very few euei dgcdi of the auns alteration after an &-«Iement: 
thus, saftdbh, avaffajnbha, Bavyaf^b^, apa^tba, npaqtut; also 
l/sab, irhen Itj final, by 147, beoomes ( : thas, eatraf^t (but satrft- 

187. The final a of the first member of a componod often be- 
comes 9 after an alterant vowel: thns, the e of a prepositional prefix, 
SB nli^dbvan, duft&ra (for da^ft'ira), ftvlfk^ta; and, regularly, a 
a retained Instead of being converted to viaarfca before a labial or 
guttural mute (171 a), as bavl;pai jyoti^k^; tapan)a. 

186. Once more, ia the Veda, the same alteration, both of an Initial 
and of a final e, it not infreqnent BTSn between the words eomposing a 
sentence. The cases are detailed In the Pratl9akh7a belonging to each text, 
and are of very various oharacter. Thus: 

a. The Initial a, especially at partlolei: as Q f 6, h{ ^ma, k&m n 
fvit; — also of pronouns: as bi ^ify; — of Teih-fomu, eipecially troin 
V^aa: as bi qtb&, divi f tba; — and In other scattering cases: as n Ifttibl, 
vA ;^F&m, trf ^adb&athft, &dM f96^, nUd^ ^U}, y^nl^ ^kaioaAja, 
agai^ ftaye. 

b. A flnal B, oftenest before pronouns (e«peciall; toneless ones): as 
acnif (vA, ni^ {a, lyiif \A, gnolf tv&m, aiidblf fiiva; — bnt also in 
other oases, and wherever a final a Is preserved, instead of being turned 
Into Tiaaiga, hafora a guttural or labial (171): as trff ptttT^ tEjUf 
kiri^otu. vitsto; p&tilp, dy&nq pit^ vlbbif p&tBt. 

Conversion of ^ n to HT v. 
189. The dental nasal =T u, when immediately followed 
by a Towel or by R » or ^m ot CT y or 9 t, is turned in- 
to the lingual HI 9 if preceded in the same word by the 

ioy Google 

65 CONVEBsiOH OP n TO 9. [ — lei 

lingual Bibilant 01 semiTOwel 01 ToweU — that is to say, 
by ^ p, ^ r, or H F or ^ T — : and this, not only if the 
altering letter standi immediately before the naeal, but at 
whatever distance from the latter it may be found: unless, 
indeed, there inteivene (a consonant moving the fiont of 
the tongue: namely) a palatal (except TJj], a lingual, or a 

a. We may thug flgun U DunelTei tbe rtdionale of the pioceu; tn 
the marked procllTlty of the Ungiugc towud Ungual uttetance, •ipedally 
or the Dual, tb* tip of the ton^e, vh«D once teTeitod into tbe looie lin- 
gaiA petition by the Dtlennce of » non-eontaot lingual alement, teoda to 
hang thete and make Ita nest nual contact in that poaltlon ; and doei 10, 
nnleai tbe prooIlTlty )■ latiafled by tbe nttaTance of > lingnal mnta, or tbe 
eigan la throira oat of a4)n*tmeDt bjr the ntttrance of an element which 
cansea it to aaiome ■ different pottote. This la not the oaae with the galtuT- 
ali or labUli, -which do not moTe the ftont part of the tongne (and, at the 
inflnenae of k on following s iboTs, tbe gntcnral podtloo fa*ota the sDccai- 
don of a lingual): and the y is too wetikly palatal to Interfere with the 
alteratlaD (as it* next lelatlTe, the 1-Towel, Itaelf lingnallzea a ■). 

b. This is a rule of constant application ; and (as was pointed 
oDt abOT«, 46) the great mtyorit; of occtirrenoea of 9 in the language 
are the result of it. 

190. Tbe rule baa force eapecialljf — 

a. When anfllsM, of Inflection ot deilTatlon, are added to root* ot 
Mama oonlalnlng one of tbe altering lonnds: tbna, nidrAi^, mdr^&m, 
-rttrlije, vitriol. varlQi. dfitf^l, h&rft^ dvAfSi^, krl^iml, qr^U, 
kqubha^A, t^ir^it kin^ vr^i^ mg^ dr&vl^a, i^kq^, pni^^i, 
rikQJU. o&b^a^a, oQarqamli^, k^amKQa. 

b. When the flnal n of a root or stem eomea to be followed, in Inflectton 
01 derivation, by aach aoonda aa allow it to feel tbe effect of a preceding 
altering catiie: thna, from y'ron, r&^anti, r&^ati, rbaqa, arSijifiia; 
bom briJunazip briUim^n&t Tyr AiiTnfl-ni , bTAlunanit) br^unasiyk, 

c. The form pl^ak (BV. : 2d and Sd sing. Impf.), from Vpl^ li wholly 

lei. This mle (11^^ that for the ohange of a to 9) appUea strictly 
and especially when the nasal and the eaoee of ite alteration both lie 
within the limita of the same integral word; bnt [also like the other) 
it is extended, within certain limita, to compound words — and even, 
io tbe Veda, to contlgaons words in the sentence. 

Whitsfj, Onmaai. 3. ad. 5 

ioy Google 

103—] Ul. Euphonic Cohbihatioh. 66 

18a. Especially, a prepOBition or similar prefix to » root, if it 
contain r or end in euphonic r for s [174}, very often linguatixea the 
n of a root oi of its derived stems &nd forms. Thos: 

a. The Initiil n of • root li nsaally ind regoluly m ftlterod, In all 
fonns und deriTitiiea, iftai park, pari, pra, nlr (for nia), antar, dor 
(for doa): thai, p4r& ^aya, p&ri ^lyate, pik ^udasra; parB^utti, 
pariijAma, praQavi, nii^, duraiiqa. Itoota taBeiiag this chsoee *ce 
written irlth Initial 9 In the n«ttTe root-lliti. The only esceptlona of im- 
pottinra are nj^ naUl, uand, and naq when iti f becomes f (as In 

b. The flnal □ of a loot if liagaallzed in some of the fotmi of an 
and ban: thus, pra '^iti, piSQi, pr& haijyate, prahi^iaiia. 

0. The elaaa-iigne nn and nfi afe altered after the rooti bi and ml: 
thai, pAri hiiyomi, pr4 mi^antl (but the latter not in the Teda). 

d. The Ist sing. tmpT, ending Snl Is sometimea altered: thos, pr& 

e. DerivatlTeB by snlDies containing n Bomedmes have q by InDaence 
of a preposition : thus, praj^a. 

f. The n ot the prepoaltion jol Is sometimes altered, like the initial 
of a root, afier another preposition: thna, pra^ipKta, pra^idhl. 

168, In romponnd words, an alterln| cause in one member lomedmei 
llnguallies a n of the neit following member — either Its Initial or final 
n, ot n In Its Inflectional or deriiative ending. The eterdte of the altering 
Inflnence can be seen to depend in part npon the closeness or fteqaency 
of the compound, or its integration hy being made the base of a deriTstlTe. 
Examples are: grKaaapf, tri^&man, urQ^asA; vq;tralii^aitl etc (bnt 
vftraghni etc: 186a), nrm&ipaa, drugha^A; pravaha^a, nipl^a, 
pfiryaqa. pltryinai Bvargdna, dnrga^i, usr&rftm^e, tryaflgagam. 

184. Finally, in the Teda, a a (usually initial) is occMionUly Ungual- 
ized even by an altering sound la another word. The tonelesa pronouns 
naa and ana- are ofteneat thus affected : tbas, p&ri ^aa, pr&l '^Sn, Indra 
ei^m; but slto the particle ii4 Ukt: thus, viz 9&; and a few other 
cues, as vox ^ama, pnnar i^ayimasi, agn6r tive^a. Hoie snomalona, 
and perhaps to be rejscled as false readings, are anch aa trl^ Imait and 
ak^a^ &va and etihan; ^al^ (t^S.), and vyr^a? ''& (Ap»st). 

195. a. The immediate combination of s n with a preceding guttniiJ 
ot labial ssems in aome cases to hindei the conTereion to 9 : thua, vftraghiUt 
etc., kqnbhnfttii tn»ioti (but in Veda tfp^n), k;epnu, bii^ui]ui&. 

b. The ItV. haa the exceptions &;trinKm and rfiftrin&m. 

Conversion of dentat mutes to Unguals and palatals. 

196. When a dental mute comes in contact ^vith a 
lingual or palatal mute or sibilant, the dental is usually 
assimilated, becoming lingual or palatal respe'stively. 

ioy Google 

67 Bental Hutb9 to Lihguals and Palatals. [—190 

The cases are the following'. 

187. A dental anrd mute or oasal, or the dental sibilant, when 
immediately preceded by a q, Is everywhere converted into the cor- 
responding lingoal. 

B, Undel thl> rale, the Minbiiittlont ff, ffll, and ^ tre very common; 
^ iB ruely ao wiitteD, the viaarga being pnt instetd of the Ibimer sibilant 
(173): thus, jy6tl^qa Initead of jyoti^fU. 

b. Hack leu otUn, dh la changed to ^ after flnal | of a root or 
lenie-atcm, with lo» of the ; or tta conveiaion to 4: eee SSflo. 

C. Those OMei in nhich flnal ^ beconma \ before aa (e. g. dvifsii: 
229 b) do not, o[ MintM, tall under thla lale. 

19B. In the other (comparatively infrequent) cases where a dental 
is preceded by a liogual in internal combination, the dental (except 
of Bu loc. pt.) becomee lingual. Thus: 

a. A n following immediately a q made luch b; the rale ^Ton at 
189, abOTB — or, u it may be eipreBted, a donble aa well sb a itugle a. 
— ia anbject to tbe lingnalization : thua, the paiUciplea ar^^A, kfup^a, 
kfvli^a, chf99&, tf^nA; and, after preflxe* (185 a), alqa^na, pari' 
viv?a, Ti9a9:7a, vifyaij^a. Btit TS. has idhi^fkaasa, and BT. y&Ju^ 

b. Only a very few other instances ooenr: tfte and ilfta from yi4i 
^a44^& (alao qa^dba and^^ha), and qa^^^ (fa^-J-nftm: anomalouG 
gen. pi. of ^af : 488). A amall number of words follow the same rale in 
eitern&l combination: see below, 190. 

O. But tS^i (Vedic: yta^ + dM) ahowa loss of the Anal llngnal 
after asalmilation of the dental, and compensatory lengthening. 

d. Some of the casea of abnormal occnrienee of 4 "^ explained in a 
ttmilai way, as resnlts of a llngnatlzed and afterward omitted sibilant before 
d: thna al^A from nlada, ypi4- '">"' P^"^ V'"'^ 'rom mfsd. For 
words eibibitiDg a like change In composition, see below, 198 o. 

198. In external combiDation — 

a. A flnal t is directed to be astimilsted to an Initial lingaal mate : 
thus, ta(-tikft, ta4 4ayate, tat-thUioI, taf 41>&<i^'>to: bat the case 
never occurs In the older language, and very rarely In the later. For final 
n before a Ungaal, see 20Kb. 

b. An initial dental after a final linguarusually remains oachanged; 
and eu of the loc. pi. follows the same rote: thns, B&ttrlAQat, iua^ 
diT&h, ekar^t tviaa; Qafsu, rlltau. 

o. Exceptiona are: a lew compounds with ^aq tix showing doable 9 
(198 b): namely, ^iiQiiavatl, fanQ&bhl (and one or two others not 
qnotable) ; and JB. baa (an ^iramlmita. 

d. In a few compoanda, moreover, there appears a Hnguali^ed dental, with 
compensatory lengthening, after a lost lingaal slbUant or Ita representative; 



namely, in certain Yedic con pound b v 1th diu: dI14ibha« dd^^Qi ductal, 
du^&qa, dQ^ftfft (compue the anomalous piiro^i9 and -4£qa: part>S + 
)/dftf) ; and, In (lie langaage of eTcry peiiod, cortaln componDds of f a^ 
with change of Its vowel to an alterant qnality (as in vo^am and n mjhiiiw 
SS4b): f64'>9'>> ^o^liA (also (a44ba and fafdhd), ^oij&iit. 

e. Between final t ■°^ Initial b, the iasenlon of a t ie permitted — 
or, accordlug to some anthoiitlee, reqotiedi thns, ^&% aab&sr&l; or tt&^t 

aoo. The oues of usimlUtion of a dental to a contignons p&latal 
occur almOBt only in external combination, and before ftn initial palatal. 
Tbeie ia bnt one case of intenutl combination, namely: 

201 . A ^ n coming to follow a palatal mute in iDteinal 
combinatioQ is itself made palatal. 

Thos, yRofi£ (the ottly ioBtance after a], yajfii, jajfi^, ajiiata, 

aoa. a, A ilnal rT t before an initial palatal mute it 
assimilated to it, becoming ? c befoie %o or W cb, and ^j 
befoie si j (^ jh does not occur). 

Thus, no oarati, atao ohattaram, vldyuj jKyate; jBtwUlana. 
vidyulJUiTa, bphAoohandaB, saooarito. 

b. A final ^n is assimilated before sT j, becoming 3tii. 

0> All the grimmailans, of every period, require this assimilation of 
n to j ; but it Is more often neglected, or only occasionally made. In the 

d. For n before a snrd palatal, see below, SOS. 

208. Befoie the palatal sibilant ^ 9, both fT t and R n 
are assimilated, becoming respectiyely ? a and 3T fi; and 
then the following n 9 may be, and in practice almost 
always is, converted to W ch. 

Thus, vedavlo ohOra:^ (-vlt qn-), tac ohxntvH, lircoliaya (brt4- 
9aya); brbafi ohe^a^ or qe;ab, erapafi diete or qete. 

a. Some anthorltiea regard the converaion of 9 to oh after t or n af 
eTerywheis obligatory, otbcra aa only optional; some except, peremptorily 
or optionally, a q followed by a mute. And some require the same oon* 
verBlon after oTiry mate save m, reading also rfpEJ ohntudrlt ftnaf 
oh^ol, anuffup chSradl, fuk ohucl. The manuscripts generally write 
ch, instead of oob, as resDlt of the combination of t and f . 

b. In the MS., t and 9 are anomalously combined into fi q: e. g. 
t&b qst4m, etftTa&q&a. 

ioy Google 


Combinations of final ^n. 

S04. Final ndiMl n is assimilated in interofJ combination to a 
following sibilant, becoming ajLuSTftra. 

Thus, vAAbI, v&Aava, v&tuuX, ma&ty&to, jigh&Aaati. 

ft. Aecotdlag to the gi&iDiiii,rlau9, It is tre&ted before bh aod au in 
declenalon u in external oombin*tion. Bnt the cues are, at beat, eicea*- 
iTely nre, and RV, bat rifiSQ and viAsa (the only VediB eumplea). 

b. Final n of ■ deriiatiTe tJxtUi it tegulaily and usually dropped before 
a conaonaDt In Infleotion and eompoaitlOD — in oontposltlon, eien before a 
lowel ; and a radical n ooeaaionally followi the game rale : see 481 a, 439, 
1808 o. 687. 

O. For aaslmilatlon of n to a preceding palatal, see SOI. 

Thus remaining cases are those of external combination. 

806. a. l%e assimilation of n in external combination to a follow- 
ing sonsnt palatal and the palatal albiianl 9 have been already treated 
(scab, 808). 

b. The n is also declared to be assimilated (becoming 9) before 
a sonant lingual (^ ^li, 9), but the case rarelj if ever occttrs. 

806. A n is also assimilated to a following initial 1, becoming 
{like m: 313 d) a nasal L 

a. The mannscTipta to a great extent disregard this mle, leaving the 
n nncbanged; bnt also they la part attempt to follow it — and that, either 
by writing the aatlmilated 11 (as the asBimlUtod m, 213f, and jnat as 
leawnably) irith the (umBTSm-slgn , or elee hy doabling the 1 and potting 
a aigo of natality above; the latter, howeTer, Is inexact, and a better way 
vDDld be to separate the two I'e, writing the first vith virSma and a nasal 
algo above. Thnt (^m trln lokftu): 

mannaorlpls afj^+H or 3T1S1!?iR; better ift^ 3l+H. 
The eecoiid of Iheae methode la the one oftenoat followed in printed texts. 

807. Before the lingual and dental sibilants, ; and b, final n 
remains nnobanged; bnt a t may also be inserted between the nasal 
and the sibilant; thus, tan ;&( or tont ^\; mahin a4n or ma- 
liint B&n. 

a. Aocording to most of the gramniarianB of the Frati^Lhyaa (not BFr.), 
the insertion of the t in sneh caees is a neceaeary one. In the mannscrlpts 
it is very frequently made, bat not nnlformly. It Is probably a purely 
phonetic phenomenon, a ttansition-aoand to ease the double change of sonant 
to Burd and nasal to non-nasal ntterance — altbongh the not Infrequent 
cases In which tinal n stands for original nt (as bharan, abharan, 
agolmfin) may have aided to establish it aa a nile. Its analogy with the 
eoBTeraion of n q into fioh (803] is palpable. 

Digitizecy Google 

208—} III. Euphonic Coiibinatiom. 70 

208. Before the surd palatal, lingual, and dental mntea, there 1b 
inserted after final a a eibilant of each of those classes respectively, 
before which the n hecomes anaavira: tbas, dar&fiq oa, bhv&Af 
oUdyftte, kiuu&r&As trin, abhara£a tata^, dadliaAQ (42Bo) oarum. 

a. This rule, iihlch in the cluslcal Ungaage bis eGUbllshed Itself In 
the form here given, ms ■ phonetic rule of unv&Tylng application, really 
iniolrei a bietoric satvivel. Tbe latgs majority ot caaee of final n in the 
langnags (not tai from three quarters) are for original ns; and the retention 
of the sibilant in Eocb casei, when once Its historical gtoniid had been forgotten, 
wag extended by analogy to all otbers. 

b. Practically, tbe mle applies only to n before o and t, since cases 
Involving the otber Initials occur either not at all, or only with extreme 
rarity (the Vedft does not present an example of any of them). In tbe Veda, 
the inseition is not always made, and the different texts bave with regard 
to it different usages, which are fully explained in tboir Prati9aichyas; in 
general, it Is less frequent tn the older texts. When the f does not appear 
between n and a, the n is of coarse assimilsted, becoming fi (203). 

208, Tbe same retention of original final s after a nssal, and 
consequent treatment of [apparent] final ftn, in, ua, fa as if they were 
Aha, IAb, Ofia, f&a [long nHgalized vowel with final a), ebowe itself 
also in other Vedic forms of combination, which, for tbe sake of unity, 
may be briefly stated here together; 

a. Final Sii becomes Sii (nasalUed ft) before a following vowel : that 
is to say, fiAe, with nasal vowel, is treated like fis, vltb pure vowel (177): 
tbne, devad 6 'ha, upabaddhaA th&, mahtA nsi. This is an extremely 
common case, especially in KV. Once or twice, the a appears as h heCore 
p: thus, BT&tavtLnh p&yu^. 

b. In like manner, 8 is Created aRer nasal i, u, ^ as it wonid be after 
those vowels when pure, becoming T before a sonaut saand (174), and 
(much more rarely) fy before a sard (170): thus, ra^miAr Iva, Bunl^Av 
yuvaiiytliAr ut, ufftr abhi; ni&fy patraiu (and iifA§ p*, HS.). 

o. RY. has once -lA before y. MS. usually has aA instead of fiA. 

810. Tbe nasala n, 9, A, occnrring as finals after a short vowel, 
iire doubled before any initial vowel; thus, pratyAAA Ad efi, ndy&an 
{Ullty&^, AB&nii-lfu. 

a. This is also to be regarded as a historical butvIvbI, the second 
nasal being an assimilation of an original consonant following the flist. It 
is always written in the manuscripts, although the Vedic metre seems to 
show that the duplication was sometimes omitted. The RV. has the com- 
pound vT^a^aqva. 

211. Tbe nasals & and 9 before a sibilatit are allowed to in- ^ 
aert reepectively k and ( — as n [307) insBrts t: thus, prftty&Ak 

itizecy Google 

71 COUBINATIONS OF PlMAl, m. [ — 213 

Combjnationa of final ^^m. 
212. Final radical R m, in internal combination, is as- 
similated to a following mute or spiiant — in the latter case, 
becoming anuavKra; in the former, becoming the nasal of 
the same class with the mute. 

a, RetoTe m 01 V (as when final; 14Sa), It la chsoged to n: tbiu, 
from )/gBm come ^anmB, sganmahi, ganv&hi, jagaiiT^fiB (which 
appear to be the oDiy qnotsble cues). Aacording to tbe grimmirlaDE, the 
suns cbaDge is made in the inflection of root-stems before bll and Bo: thus, 
pTsqftnbhla, praQKuBU (from praq&ia: pTa+)^7am), No deriied noaii- 
Btem ends in m. 

b. The (B. and E(S. bafe k&mvant and gimvant, 

218. Final ^T m in external combination is a servile sound, 
bedng assimilated to any following consonant. Thus: 

a. It remaios DDchanged only before a vowel or a labi&l mate. 

b. Bnt also, by an anomaloos exoept[oD, befoie r of the root xij in 
samrij and its deHTatlves aamrl^fU and a&mrUya 

c. Before a mate of any other class than labial, it becomes the 
nasal of that class. 

d. Before the semivowela 7, 1, v it becomes, aocerdii^ to the 
Hindu grammarians, a nasal semivowel, the nasal counterpart of each 
respeclirelf (see 71), 

e. Before r, a slbilaot, or h, It becomes uraevBTa (see 71). 

f. The manascrlpts and tbe editions In general make no attempt to 
distinguish the anal tones produced by the assimilation of m before a follov- 
ing eemlvovel tiom that before a Bpiracl, 

g. But if h be immediately followed by anothei consonant (which can 
only be a naaat or semi-vowel), the m is allowed to be assimilated to that 
following consonant, This Ie because tbe h has no position of the moath- 
orgaiis pecullai to itself, but is uttered in the position of the ueit soand. 
The Piatioakbyas do not take any notice of the case. 

h. Cases are met with in the Teda whete a final m appears to be 
dropped before a vowel, the flnal and initial vowels being then combined 
Into one. Tbe pada-teit then generally gives a wrong interpretation. Thus, 
saiiiT&iiaiu) tiliayaiiikar&ia (RV. tIU. 1. 3; pada-teit: •□ana ubh-i 
ST. -nanam). 
■ 1. It has been pointed oat above (78) that the assimilated m is 
generall; represented in texts by the anusvJbra-siKD, and that in this 
work it is transliterated by di (instead of a nasal mute or A). 



The palatal mutes and aibilant, and ^ b. 

814. These oounds show in some rittDttiona & reverBlon [4S] 
to the original guttunla from which they are derived. The treat' 
ment of J and li, also, is different, according as they repreeeot the 
one or the other of two different degrees of alteration from their 

ai6. The palatals and h are the least stable of alphabetic soonds, 
undergoing, In rirtne of their derivative character, alteration in many 
oases where other similar sounds are retained. 

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

a. Before a of snfax a, final O becomea k in anlci, Qvafiko, arki, 
piki, vSki, fuko. parka, mark&, vfkA, pr&Uka etc., reka, sika, 
moka, rok&, 96ka, tokA, mroki, vraakA ; — final j benamai g in 
tjts&, bhige. bhAg&, yfiga, afiga, bhaagA, saOga, svafiga, pngA. 
tufiga, yoaga, varga, m&TKa, mjsk, varga, aarga, nega, vegA. bhdga, 
yug^ y&sa, loga, r6ga; — flml h becomeB gb in agtiA, maghi, arghii, 
dlrgli& (and draghiyas, drfghiftha], degha, megh4, ogha, d6gha, 
dr6ghn, m6gha; and In dughSua and m^hamftna. In neka (ynij) 
we bare lurther an auomtlom aubatitntlDu of a anrtl for the final couaot of 
the root 

b. In anothei aertei of derl*atlTes with a, tbe altered lonnd appean: 
eiunplea aie i^&, yj^a. ijvuA, qoea, vfaJA, vevlji, ynjat firj^ d6ba. 

O, Before the snfflieB aa and ana, the guttaral only rarely appears: 
namely, in iL&kas, 6kna, r6kaa, f Akaa, bh&rgaB, and In roga^; also 
in Sbbog&ra. 

d. Before an i-vowel, tbe altered aonnd appears (except In ftbhog{, 
Asjyafle, tiglt&, moki, eplilgt): thus, 4j{, tnj£, niol, q&oi, vfvloi, 

e. Before o, the guttnral reappeaii, as a rale (the oasea are few) : thos, 
ftiika, vafUro, rekii, bhfgn, ntiETgnka, raghu (and r&ghlyaAB). 

f. Before n, tbe eiimplei of lererBion are few, except of J (becoming 
g) hefore the participial ending na (9S7o): thua, r^k^aa, vagnu (with 
tbe final also made sonant); and participles bhagn&, rug^, etc.; and 
apparently pfgl^ft from VpfO. 

g. Before m (of ma, man, mant, xnln), tbe guttural generally 
appears: thus, rukm&, tigmi, yngma, fgma (with sonant change); tak- 
miu, v&kman, B&kman, yngm&n; rukmant; rgmin and vSgmfn 
(with sonant change): — bat AJman, ojmitn, bhujm&n. 

h. Before y, tbe altered sound Is used: tbns, paoyo, y^jya, yivjyn, 
ynjya, bhnjyu. Sncb cases as bhogya, yogyn, negya, okya are doubt- 
leii lecondaiy deriiativeB from bhoga etc. 

ioy Google 

73 Combinations of final Palatals. [—SIS 

1, BetoTB r, tfae cisei us few, >iid the naege ippiientl; dMded: tfaua, 
Utkr*, aakrs. vakri, ^vkrk, vlgr&, ugr&, tCigra, mfgra, v&fikii; 
1}at v^ra end p«]ri(!> 

J, Before v (of tbe iafSi«8 va, van, vln, etc., end piitlclpitl vfiAa) 
the gunoril 1* Tegalirly preiened : thus, fkrii pakrA, v&krfti v&kvau, 
fkran, rUcvan, 9akvaii, mrgvan, tdgvan, Tugran; ^kranC, pfk- 
vanti Tfisv{ii, vagToni, vasracu (vltli furtbeT lODaDt cb&age); vivok- 
viAs, ririkT^B, virlkvaAB, mrakv^As, ququkviAs; QUQUkvanA, 
t^aqakr&al: alio bef))ie the anion-TOwel i In ofcivJiAB (RV., once). An 
eiceptloo is yijvan. 

k. The leTenloD of b in deriretlon ia eomptntlTely taie. The floal 
J which Is tnUogoiu with f (S19| ahows much lesi ptocllTlty to reTenion 
then that which coneapoade iiith O. 

1. A like reieraiou ihows ilmlf also to eome extent In conjugttional 
stam-fOimatloii and loBectlon. Thoa, (he initial ladiul becomea guttata I 
after the Tsdapllcatlon In the present ot perfect ot desideratlie oi IntenaiTe 
sterna, ot In dertiitKea, at the toots oi, oit, jl, bl, ban, and InJ&gari (ySf); 
and ban becomea ghn on the ellaion at a (402, 637). The BT. haa 
vlvakmi ^m yv*o and vfivakre from yve&o ; and ST. has saapgrnabe 
(BV, <■!]-). And betare ran etc. of 3d pi. mid. we ba-ve g for radical J 
to BBncran, aargram, aaaargram (>U in RV.). 

217. Final ? o of a root or stem, if followed in internal 
oombiaation by any other sound tbtm a vowel oi semivowel 
01 nasal, teverts (4Sj to ite original guttural value, and shows 
eveiywhere the same form which a ^ k would show in the 
same situation. 

TboB, T&kti, nviktha, v&k^l, vak^&nl, vagdbl; vOgbMs, 
vSk^A; akt4, tiktbi, vakt&r. 

a. And, as final o becomes k (above 148), the same rule applies 
also to o in external combination: thiu, vltk oa, v^&pi, vte me. 

Examples of c remitining unchanged in inflection are; noy&te, 
rirlord, v&d, mumuom&be. 

218. Final Q f) leveits to its original ?i k, in internal 
oombination, only before the H a of a verbal stem or ending 
(whence, by 180, ^ k?) ; before rT t and El tb, it eveiywhere 
becomes zs 9 (whence, by 197, ? 9t and '? Qtb) ; before y dh, 
H bb, and H bu of the loc. pi., as when final (146), it 
r^ularly becomes the lingual mute (TT t or T 1^}. 

ThDS, ivikfata, vek^y^mi; vk^ti, vl^\&, dlde^tu; dldl^^bi, 

ioy Google 

918 — ] 111. Euphonic Combination. 74 

tt. But & few roots exhibit tbe reverBioD of final 9 to k before 
bb and so, and tiao vbeo final (14S}: tbey &re dii;, d|^, appf, and 
optionallj nnqi and TI9 has in V. always vik^u, loc. pi., but vi\, 
vi^bbJa, etc. Eiaioplea are dlkaaihQtta, disbbia, hrdiaptk, n4k 
(or na(). 

Eiamples of 7 remaining anchanged before vowels etc. are: vi9l, 
vivifTfta, avlQran, a^DOml, vnfmi, ngm&sL 

b. A q remains irreguUily unchanged before p In tbe compoand vif p4U. 

219. Final ^ j is in one set of words treated like ? c, 
and in another set like ^ 9. 

TbuB, from ynj: iyukthlla, Ayukta, yunkti, yukti, yoktra, 
yokqyjml, yuk^u; yiifigrtht, AyugdhTam, yugbhia. 

Again, from mpj etc. ; 4mfk^t, arakfynmi ; mir^, mfffA, 
sffti. r&9tr&; mr^^hf, mr4<MiT&m, rS^bbia, r^sii. ri^. 

a. To the former oi yi^-cUu belong (u thown by Cbeii quotable 
rorma) about twenty roots and radical itema: namely, bll«j, aaj, ^1^ (not 
v.), rsj color, Bvaj, majj, nij, tij, vij, 1 and 2 bbaj, yi^, mj, vjj, 
anj. bbaBj, qlfij; t^Ji arij, bhi;^, &arJ; — also, stems formed with 
tbe auttlies aj and U (383. l\), as tj^^iij, va^U; and |i;vij, though 
eontaialug the toot yi^. 

b. To the latter or mFJ-clsEs belong only about one thiid as many: 
namely, y(^, btan^], VTtii, rftj, bhriU, mfj, a^. 

o. A considerable number of J-toota ate not placed in circnmatances 
to exhibit tbe dlitlnction; but such roots are in part stsigaable to one or 
the other clas* on the evidence of the related languages. The disUnetioii 
appears, namely, only when the J occnra as flnal, or it followed, either in 
InDection or in derivation, by a dental mute (t, tli, dh), or, in noun- 
InflerliuD, by bh or su. In derivation [aboTs, 216) we Sad a g some- 
times from the idfJ-cUbs: thus, m&rga, a&rga, etc.; and (316,1) before 
Vedic mid. endlngi, saapginabe, asrgnui, etc. (betide BasTJrlre) — 
while ^m the yuj-cIasG ocnut only yuyujre, ayujran, bubbujrire, 
with J. Aud MS. has vlQvaafk from ysfj. 

220. Final oh fitlla under tbe rules of combination almost only 
in tbe root praoh, in which it is treated as If it were { [pra{ bein^, 
indeed, its more original form): thus, prak^yami, pnt'^ ^i^^ >i1bo the 
derivative praqn4. As final and in noun -inflection (before bh and au), 
it is changed to the lingnal mnte: thus, pr&^vlvftka. 

a. llurt& Is called the participle of mfiroli, and a gerund miirtvi 
Is given to the same root. They (with mltrtl) mntt doubtless come from a 
simpler form of the root 

ioy Google 

75 Combinations of final kf, h. ( — SSS 

SSI. The compouod k; Ib not inft'equent as final of & root (gener- 
ally of demon Btrably seooiKlary origin), or of a tense-stem (ikaorisC: 
see below, 878 ff.]; and, in the not reiy frequent cases of its internal 
combination, it is treated as if a single sound, following the rules 
for 9: thus c&kqe (oakf+se), e&k^Tft; o4|te, &aa^\n, &8rftf(tuD, 
iiei^tft, tv&ftar. As to Us treatment when final, see 146, 

a. Tbna, we are t&nght by the grammarians to make bucIi forcaB as 
gor&t, gor&^bbis, goriffu (from gor&kq); and we actually have ^i.\, 
^a^bhfs, fafsu from ;akf 01 qaif (146 b). For jagdha etc. from (/jak;, 
Bee 333 f. 

b. In the altigld anomalous root vraijo, the compouDd qo i% said to 
loUaw tbe ralee for simple <}. From It aie quotable the future Trakfy&tl, 
the genindB vntv£ (AV.) and vyktvi (RV), and the paiticiple (957 o) 
T^kni. Its o reTcrte to k iti the derivitiTe vraska. 

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

a. In the one class, as dnb, we have a reversion of h [as of cj 
to a guttural form, and its treatment as if it were still its original gb: 
thus, &dhuk;am, dbok^yami; dugdhwrn, dugdb&; &dtiok, dhiik, 
dbugbbie, dbuksu* 

b. In the other class, as rub and aah, we have a guttural re- 
version (as of 9I only before b )n verb-formation and derivation: thus, 
aruk^at, rokfjlbni, Bftkfiyi, sakfinl. As final, in external combi- 
nation, and in nouD-infiection before bh and bo, the b Hike f) becomes 
a lingual mute; thus, tn^^afi pptanSfi^ ayodhy&b, turSs^bble, 
turfifatBu. But before a dental mute (t, th, db] in verb-inflection 
and in derivation, its euphonic effect is peculiarly complicated: 
it turns the dental into a lingual (as would 9); but it also makes 
it sonant and aspirate (as would 411: see 160): and further, it 
disappears itself, and the preceding vowel, if short, is lengthened: 
thns, from rota with ta comes rQ4bil^ from leb with ti comes leijtai, 
from guh with tar comes guijUi&r, from meh with turn comes m^tptum, 
from lib with taB or tbaa comes liijb&B, from lib with dbvam comes 
U^T&m, etc. 

o. This ie ag if we had to asanme as tianaitlon sound a Eonant aspliite 
lingual sibilant sb, with the euphonic effecta of ■ linguU and of a sonant 
wpirtte (100), llselt diiappeaiieg nndet the law of tbe existing laagaage 
Thich admits do eoiisnt Blbilanl. 

SS8. The lootB of the two classeB, as shown by their forms found 
in use, are: 

a. of the first or dnb-class: dab, dib, dub, drub, mub, solb 
(aod the final of u^^Ib is similarly treated) ; 

Digitizecy Google 

SSS— ] in. Euphonic Combination. 76 

b. of the second or nOi-clMB: vah, sab, mlh, rlta or Ith. gnh, 
rah, 6fbiL, tfhh, brh, baAh, Bprti(?]. 

0. Bat mull forms alao (not <n KT.) the participle m&^llA 4nd i^ent- 
noan TaJi^hia, m veil u mngdli&i and mogdfa&r; and dmli sod snih 
»ra allovftd by (he giunmulaiis to do likewise: sach fotmi ts drfljjli* ■"^ 
■nl^lw, bowevec, h&Te not been met with In use. 

d. Fiom roots of the ruh-tlus we find eleo In the Teda the forms 
gart&ruk, nom. sing., end prSl^adb^k and dftdhtk; and hence pomspfk 
(the onl7 oecntrence) doei not oaitalnlf prove yvpjh to be of the dab- 

e. A nambei of otbeT h-root> ara not proved by their occDrrlag fomu 
to belong to either olats; they, too, are with moie or ten conSdenoe aaiigned 
to the one oi the other by oomparlson with the related languages. 

f. lu derivation, before certain snilliei (S18), «e have gh Instead of 
li ftom Terbs of either clasa. 

g. The root niA comes from originil dh Initsad of gh, and Itt reversion 
is iccDTdlngly to a dental mute: thna, natvluul, naddlUi, upSn&dbllla, 
npSn&dyuga, anap&natkft. So alio the root grfth comes tnia (early 
Tedic] grabh, aud sbovs labials in many forms and derivatlTes (though 
h is aislmllited to other h-roots in the desidetatlve item Jighfk^). In 
like manner, tl Is need fsr dll In some or the forms and derlvativaa of 
y'dhS put; and further anilogont facta are the stem kakuh& beside 
kKkabb^, the donble imperative ending dbi and bl, and the dative 
m&byiun beside t&bbyun (4dl). 

SS4. Inegnlaritles of combination are: 

(t. The vowel f le not lengthened arter the loss of the b-element'. thus, 
df^b^ tf^h&, b^^h& (tba only oasej; and in the Veda Iheii Qrst syllable 
has metrical valne as heavy or long). 

b. The roots vftb and sob change their vowel to o Instead of leng- 
thening it: thus, TO^b&m, ToijbKin, To^bif, sd^bnm. Bot from sab 
In the older Imgaage forms vlth S are mors frequent; thus, sA^it, itqS^bit 
(also later), Ba^bMT. The root tfiih changes (he vowel of Its elaas-sign 
nft into « instead of lengthening It: thai, t^p^e^bl, tfl^A^bU, Ktp^et 
(the grammarians teach also tf^ehmi and tpqokfi: bnt no such forms are 
quotable, and, IF ever actuslly in use, they mtut have been made by false 
analogy with the others). 

0. These anomsloas vowel-changes seem to stand in connection with 
the fact that the eases showing them are the only onei vheie other than 
•n alterant vowel (180) comet before the lingnslUed sibilant representative 
of the fa. Compare fb^B^tt etc. 

d. Apparendy by dissimilation, the final of TItb in the anomalous 
oomponnd ana^vftb is changed to d Initead of ^: see 404. 

itizecy Google 

Combinations op finai. f. 

The lingual sibilant ^ 9. 

996. Since the liDgual aibilant, in its nsnal and normal occurren- 
eea, ia (189) the product of liDgnalization of s alter certain alterant 
■ooDds, ve might expect final radical f, when [In rare caaea] it comes 
to Btand where a f cannot maintain itself, to revert to its original, 
and be treated u a a would be treated under the same circumstances. 
That, however, is true only in a very few Instances. 

a. Namely, in the piefli dna (eTldtntly Identiul wltli ydo^); in 
aaiia (adTeiblill; need oia-form from VJof)! in (RV.) vtvAs uid Avlvea, 
from V'viq; in iiyea (RT,), fiom )/1f; and in K^fa, fiom gl^ m lecoud- 
u7 form of v'ffia. All thes«, except the flnt two, sre more oi leis open 
(o qoeition. 

326. In general, final Hngual ^ fi, in intemal combination, 
is treated in the same manner as palatal 1^9. Thus: 

a. Before t and th it remains unchanged, and the latter are as- 
slmllated: e. g. dvlffaa, dvi^fbu, dv^tnm. 

Thli Is a common ind peifeetly n&taril combiiiatlon. 

b. Before dh, bli, and m, as also in external combination (146), 
it becomes a lingual mnte; and dta is made lingual (by 1S8) after it: 
e. g. pl^^ijhl. Tlf^hi. TiTl^^hl, dvi^^hvam, dTi^bhia. dvitsai 

o. So also the dh of dhvam u ending of 2d pi. mid. becomes ^ 
after Anil ^ ef ■ tenn-item, vhstker the | be regarded ts lott or u con- 
Tsrled to 4 hefoie It (the muiasetlpta write simply ^V, not 4^t; but 
this la ambigDODs; see 933). Thus, aftsr ; of O-aorlst stems (681 a), asto- 
flivam. ftV(41iTan3, oyo^hvam (tha only qnotible cues), from oato^-h 
dbvam et«.; bat arBdhvun f^om aria+dhTam. Fortber, after the ; 
of i^-urist Items (801 a), Undhi^TKm, nrtiftaTam, ^anl^vam, 
trepl^Tam (the only quotable oases), from ajani^ + dhvam etc. Tet 
again, in the precatiTe (894), m bllavi^IijItTam, if, as Is probable 
(unfortunately, no example of this person is quotable from any part of the 
Uteratuie), the precatlre'Sign ■ (;) is to be regarded as present In tha fonn.' 
Aecordlng, howoTer, to the Hindu grammarians, the use of 4b or of dh la 
the ig-aot1st and precatlTe depends on whether the 1 of iQ or of I^i Is ot 
is not "preceded by a semlTowel or h" — which both in itself appears 
•enselew and is opposed to the evidence of all the quotable forms. MoreoTCi, 
the same anthoiitles pretcriba the change of db to ^h, under the same 
reMiictlon as to dreumstanoet, In the psrf. mid. ending dhve alto: In this 
ease, too, without any coueeivabie reaeon; and no example of ^ve In tha 
2d pi. pert, has been pointed out in the literature. 

d. The canTorslon of q to ^ (or ^] as final and berote bh snd Bti is 
patallel with the like couvarslon of f , and of j and fa in the mf J and rah 

ioy Google 

326—] 111. Euphonic Cokbination. 78 

cluass of loota, and perhaps with the occi^ioual change of b to t (1674). 
II ii ■ VBi; Inh'eqnent cue, occarrlng (lire as it may be aisnnied in th« 
case of fBf) only onc« in RV. and once In AV. {-dvi% and -pruf], alilioasb 
those texts hav« moie than 40 roots with flual |; in tlie Biahmanaa, 
moteoTer, have been noticed farther ooly -pruf and ■ytf (QB.), and -flit 
(K.). From piil^, RV. bu the anomalona font) pii^ak (2d and 3d alng., 
[or pina^'B and pliia§>t). 

e. Before b id ioternal combiDfttion [except bu of loc. pi.) it be- 
cornea ki thus, dvdkfi, dTefc;yamf, AdTikfam. 

f. This change is of anomalous phonetic cbaiaclar, and dlfflcnlt of 
eiplanation. It ig also practically of very race occorieoce. The only RV. 
examples (apart from ptQak, above) are Tivekfi, from yvl^, and the 
desld. stem rirLk^ from yrif; AV. has only dvik^at and dvlk^ta, 
and the desld. 8t«m qiqlikfR from y'flif. Other eiamplea are quotahle 
^m yy^n and pl^ and vlf (QB. etc.), and qif (^B.); and tbey are by 
the Hindu grammarian* preacrlhed to be formed from abont half-a-dozen 

Extension and Abbreviation. 

827. Ab a general rule, ch ia not allowed by tbe grammariana 
to stand in that form after a vowel, but la to be donbled, beooming 
cch (which the manuscripts sometimes write obcbl. 

a. The various authorities disagree wUh one another in detail as to 
this dnpllcalloD. According to Paniui, ch Is doubled vithin a word after 
either a long or a ehott vowel; and, as initial, necesairlly after a abort and 
after the particles i and rai, and optionally ereryvhere after a long. In 
RV., Initial oh Is donbled after a long vowel of i only, and certain special 
cases after a short vowel are eicepted. For the required ossge in tbe other 
Vedic teits, see theli several Pritl9Bkhya«. The Katbaka vritaa for original 
oh (not ch from cambiuatioii of t oi n with q: SOS) after a vowel 
everywhere ijdh. The manuscrlpta in general write simple oh. 

b. Opinions are atill at variance as to how far this dopUcation has 
an etymological ground, and how far it is only an acknowledgment of the 
fact that oh makes a heavy syllable even after a short vowel (makes 
"position": 78). As tbe duplication ia accepted and followed by most 
European scholars, it will be also adopted in this work in words and sen- 
tences (not in roots and stems). 

aS8. After r, any conBonaDt (save a spiraot before a vowel) is 
bj tbe grammariaoB either allowed or reqaircd to be doubled (an 
aspirate, hy prefixing the corresponding non-aspirate: I54j. 
5R; arka, or 51^ arkka; ^THJ kiirya, or cfiJUl karyya; 
mt &rUia, or qf^ arttha; ^ dirgba, or ^tr^ dirggha. 

Digil.zecy Google 

79 Extension and Abbreviation. [—381 

(u Soma af the aathoiltlet include, ilosg with r, also h oc 1 ai v, or 
mate tbin one of them, Id thl« mle. 

b. A doabled conaDiiant ifEer r Is very common in msnuicripta and 
intciipttona, u alao Id native lexl-edltlona and in the earllei editions pre- 
pared bf Eniopean icbolara — in later onei, the duplication it DUiTenally 

o. On the other h*nd, the roBnaictlpts often write » ilnglo conionant 
aftei r where a donble one Is etymolagically required: thus, kBftlkeya, 
TSrtikft) for kftrttibeyo, vKrttibft. 

229. The flrgt conaonaiit of a gronp ~ whether interior, or Initial 
after a vowel o( ■ preceding word — la by the frammuiing either allowed 
or leqnired to be doubled, 

a, Thia duplication U allowed by Panlnl and required hy the Piatl;ikhyas 
— In both, with mentlOD ot authorltlea who deny It altogether. For certain 
exceptions, aee the Prattt^khyaa ; the meantug of the whole matter Is too 
obaeore to Justify the giving of details here, 

230. Other cues of exteoBion of coDBOnftot-groapa, required by 
some of the grammatical autborities, are the following: 

B. Between a non-nasal and a nasal mule, the insertion of ao-called 
yamas (twitu), or nMal coanterpaTta, la tanght by the Prall^khysi (and 
aasomed in Panini's commentary): see APr. 1, 99, note. 

b. Between h and a (ollowlDg nsaal mute the Priticakbyas teach the 
liiiertlan of a nasal lonnd called tiSdikya: see APr, I. 100, note. 

o. Between r and a following consonant the PrBti9aUiya8 teach the 
Insertion of a svorabliakti or voicel-fragment: see APr. 1. 101-2, note. 

d. Some authorities assume this lusertion only before a spirant; the 
others regard it as twice as long before a spirant as before any other con- 
sonant — namely, a half or a quarter mora before the former, a quarter or 
an eighth before the latter. One [VPr.) admits it after 1 as well u r. It 
is Tarlonsly described as a fragment of the vowel a or of f (or }). 

e. The RPr. pats a STajabhakti also between a sonant conaonant 
and a following mute or spirant; and APr. iutrodnces an eletaent called 
BpllotAlut (distinguuher} between a guttural and a preceding mute of 
mother class. 

f. For ooe or two other cases of yet more doubtful value, see the 

asi. After a nnul, the fonner of two uon-DaBal mutes may 
be dropped, whether homogeueoua only with the naaal, or with both: 
thus, yufidhi for yoQgdhf, yuBdhTAm fur smfigdliv&m, Kfitim for 
AfLktAm, pafltl for petilkti, ohinttun for obinttam, bbintbi for 
bhinttbii, indhd for inddbe. 

a. The abbreviation, allowed by Paninl, is required by APr. (the 
other Pratifiikhyas take no notice of It). It Is the mote usual practice of 
the manuscripts, though the full group Is also often written. 

i, Google 

asa— ] III. EnpHOMio Combination. 80 

233. In geoersl, a double coDeoDant tincluding an aspirate wbieh 
is doubled by the prefixion of a non-aspirate) in combiDation with anj 
other conaonaDt is by the manuscripts written as simple. 

a. Thtkt U to B&y, the ordlnmiy oaage of the mmaectlpts toiikM no 
dlffetenca between those groups In which a phonetic daplication Is allowed 
by the Tulea given sbove (228, 228) end those in which the duplication 
Is etfmologlcal. As erery tv aftei a vowel miy also be properly written 
ttv, so dAttvi and tattv& may be, aud almost invarUbly are, written aa 
datvi and tatT4. As k&rtana is also properly k&rttana, so ^rttika 
(fiom kftti) is written as kSTtlkA. So in inflection, we hsTS always, for 
example, maJSli etc., not majjfi.^ from majj&n. Even in composition 
and aantencc-colloeatlan the same abbiiTlatloDs are made; thus, liT<^ot& 
for h^dyot&; olilniity asya for ottlnAtt? aeijra. Hence It la Impossible 
to detemilne by the eTldeuce of written nsage whether we shonld regard 
ftdbvun or iddhvam (from yia), &dvl^Tam or Advi^^Tam (&om 
/dvlq), as the ttne fonn of > second person plnral. 

238. a. Instances are sometimes met with of apparent loss (perhaps 
after con'erslon to s semiTawel) of i or u before y or v lespectlTely. Thus, 
in the BAbmanss, tu and nn with following Tfi{ etc. often make tr&E, 
nvK£ [alio tTBvA, &n.v&{)i and other eiamples from the older langnage 
are anvart- (onu+ywart); paryan, paryantl, paryfty&t, paryS^a 
(pari+yan, etc.); aUiyarti(abhi+lyarti); antaryM(antar+lTSt)i 
oftrvfio, oSrvftka, ofirvadana (o&ru+vKo, etc.); kyluit for kiyant; 
dvyoga (dvi-f-yoga}; anvft, anvSsana (ana+vS, etc.); probably 
vyiinotl for vi ytuotl (RV.), iitv4qI (nra-vaqi), ^qrari for ^f^-wart 
(RT.); ■vjSja& (yi+y&niai); and the late Bvan^a for aavar^a. More 
anomalous abbreviations are the common tpoa (tri+TOa); and dTTOa 
(dvi+poa: S.), and treql [tri + ei>i: Apast). 

Further, certain oases of the loss of a sibilant require notice. Thns: 

b. According to the Hindu grammarians, the a of B-iorlst items Is. 
lost after a aliDTt tdwbI in the 2d and 3d aing. middle; thus, adithSa 
and adlta (lat sing, adlfl); ak^thSs and ah^ta (let sing, akffi). It 
Is, howBTer, probable that such cases are to be explained in a different 
manner : sea 834 a. 

o. The 8 between two mntes is lost in alt combinatious of the 
. roots ath& and stambh with the prefix ad: thns, dt thns, atthit^ 
ut tliKpaya, dttabdba, etc 

d. The same omission is now and then made In other similar cases: 
thus oit kambhanena (for sk&mbb-: RT.); taam&t tute (for stute) 
■ud puromk tuta (for atata: £.]; the compounds fktb& (fk+atlLft: 
PB.) and atphullflga; the derivative otpfaUa (Vaphal). On the other 
hand, we have vldydt stanAyantt (BT.), utsthala, kakntatha, etc. 

a. So also the tvise-slgn of the s-aoiist Is lost after a final consonant 
of a root before the Initial conionsnt of an ending: thns, aobAntta (and 

ioy Google 

81 Abbreviation of CoNSONANT-GRonPS. [—386 

tor this, by S31, och&Dta) for aoh&ntBta, fftpta for Q&pata, tftptam 
for t&pstam, abhfiktft for &bhfikHta, (un&ubtam fot amftuketam. 
Ihese are llie only quotable casta: compile S83. 

f, A. final s of toot or tBUM-rtem It Is a few Initaucea lost aftor a 
wuant upiiato, and tbe comblnatian of mates 1e tben made at If no ilbilaut 
lk>d BTOi Inteiveoed. Thai, tiom tbe root ghas, irith omiesioii of the 
TOTol and then of the final albtlant, we have the CoTta gdba (foe gbs-ta: 
3d ling, mid.), the ptrtleiple gdha (lu agdhad), sad the derlTitlTe gdhi 
(for gha-tl; in (»&<gdhi); and fulthei, fiom the ledupllrated form of the 
Mme root, ot v'Jakf, we have Jogdho, Jagdhom, Jagdlivfi, Jegdhl (liom 
jaghS'ta etc.); also, In like manner, from bapB, leduplleation of bhaa, the 
fotm babdhSm (foi babbB-tfim). Accotdini to the Hiadn grammaiisnB, 
the «ame uttei loss of the aoilst-slgn a tskes place aftet a final sonant 
upinte of a loot before an ending beginning with t oi th: tbns, from 
I^Qdh, B-aoclst stem arftuta act. and arata mid., oome the active dual 
and plaral penons ar&nddliain and arSuddhftm and ai&addha, and the 
middle slngnlai persons aruddhfia and araddha. None of the active 
form?, however, have been found quotable from the llteratare, ancient ot 
modem; and the middle focmi admit also of ■ different eiplanatlon: see 
S84, S83. 

Sirengthening and Weakening Processes. 

9S4, Under this bead, we take np first the changes that affect 
vowels, and tben those tbttt affect coDSonaDtB— adding for conyenienoe'a 
Bake, in each case, a biief notice of tbe vowel and consonant elements 
that bave come to bear tbe apparent office of ooonectives. 

Qu^a and Vrddhi. 

236. The so-called gu;^ and v^ddlii-ohaDges aie^the most 
i^ulai and frequent of vowel-chaoges, being of constant 
ooounence both in inflection and in derivation. 

a. A gu^a-vowel {gaj^a secondary quality] di£fer^ ftom 
the cotiesponding simple vowel by a prefixed a-element 
which is combined with the other according to the usual 
niles; a vrddhi-vowel (vrddhi growth, increment), by the 
further prefixion of a to the guva-vowel. Thus, of ^ I or 
^ I the oonesponding guijia is (a+i^ ^ e; the correspond- 
ing vrddhi is (a-|-e'=)^ 31. But in all gunating processes 
9 a remains unchanged — or, as it is sometimes expressed. 


S3&— ] in. Euphonic Coubinatiok. 82 

?'» is ita own gu^; ^ B, of courae, temains unchanged 
foi both gu^a and vrddhi. 

286. The Beiieg of coitespondiug degrees is then as 


simple vowel a& il uQ r ) 
giina aft e o ar al 

Ti^dhi a ai Su &r 

a. Thera is nowheie an; ocimnfiiice of f in a litaktloQ to osdeigo 
either gu^a ot Trddhl-chinge; cor doei \ (26) aver inffsr chuige to 
vrddhi. Theontlcally, f iranld hive the same chKUga u f, and the 
vTddbl of } wonld be ftl. 

b. In Beoondu; derivatlvM leqnltEng T^dhl of tho tint vTlIable 
(1804), the o of go (861 o) U itreDgdianed to glu: than, gKomata, 

287. The hlitoriul relatioDB of the memben of «ach Tovel-««ries are 
»t1II malteri of loma dilteieace of opinion. From the »peeial point of view 
of the SaDtkrlt, the simple voweli wear the aipect of being In general the 
original or fundamental onea, and the olhen of being prodaet* of Oieir 
InciemeDt or strengthoDlng, In two aeTBral degreee — so that the ralee of 
fonnatlon direct a, 1, u, f, { to be raleed to gu^a or vTddhl reipectlTely, 
ondet ipeelfled cendidonB, But j has long been so clearly seen to come 
by abbreviation or weakening from an euUer ar (oi ra) that many Euopean 
gnmmuluu hare preferred to treat the gin^a-fonne ae the original and 
the other a* the dertfatlTe. Thai, for example: inttead of (MnmtDg certain 
loola to be bbr and TTdh, and making from them bbaratl and vardhati, 
and bhrta and v^ddlia, by the (une rules vblch from bhd and nl and 
from badb and (dt form bbavati and nayati, bodbatl and oetati, 
bbuta and nlta, buddha and oitta — they assume bbar and vardh to 
be the roots, and give the ralej of formation for tbem in reverse. In this 
work, as already stated (104 b), the f-fonn is preferred. 

S36. The gn^a-increiiieDt is an lodo-European phenomenon, and 
is In many caaea seen to occnr in connection with an accent on the 
incresBed syllable. It Is found — 

a. In root-eyllBblea : either in inflection, as dv^qti from ydvi^, 
d61unl from yduh; or in derivation, as dv^^a, d^bas, dvifttun* 

b. In formative elemente: either conjugational clasa-signs, as 
tan^mi from tanu ; or soffises of derivation. In inflection or In further 
derivatloD, as matAye from matt, bhaniivaB from bhanu, pit&ram 
from pitf (or plt&r), hantavyk from h&nta. 

238. The v^ddhi-increment is BpeclGcally Indian, and its occur- 
rence is lasB freqoeut and regular. It ia foiinil — 

il.zecy Google 

38 <3infA AND Trddhi. [—MS 

a. Id root and snffix-iylUbtes, inetead of paf^: thna, titft6tl 
from yata, sUhSyam from s&ktal, AnBlfam troxa ynl, AkOrfam 
and Ur&Tati and ksrjk from ykf (or kftr), datlbam from dst^ (or 

b. EspMlaUf oft«D, In initial Sfllablee in Becondarj derivation: 
iknt, minaaA fW>m minas, vUdTntA from Tid;4t, bh&am& from 
bhtHinl, pirthiva tiom pfthivi (1204). 

But — 

340. Tbe Ku^a-Increment doea not nanallf take place in a heavj 
syllable ending with a conaonanti that ia tossy, the ml ea prescribing 
giiQa in proceBses of derivation and inflection do not apply to a abort 
vowel which Is "long by poaition", nor to a long vowel Dnless it be 
finitl: thna, oAtatl from >'oit, but nlndati from ynind; oAyatl from 
ynl, bnt jivatl from v'Jiv. 

«. The Trddtai-lnoiemaDt Is not liable to thl« leitriction. 

b. Ezeeptlona to tbe rule are occujoaallf met iritb: tbna, ellA, ebU 
from yfb; be^&yKmi, b6i}(>fl> '*^< ^o™ I'bl^; oofa ete. from yof^; 
ihate etc. tnim y'fih ermnder; uid espeelUly, ftom roots In Iv: dld^n 
derl^yatl, divana, et<^., from ydlT; tlfftheva from y^fblv; Brer&ySml, 
Brdvuka, bom yariv — on lecouDt of vblch It ia, doubtleu, that thsM 
loota ure written with It (div etc.) by the Elndu giammuiani, tlthoagh 
they nowhere «bov a short 1, in either verb-fonna or derivattvei. 

o. A Aw ctMs oeam of prolongation Inatead of incrtment: thai 
dfif^rati from ydnf. gdbatl from y%ah. 

The ehangea of x (more original ar or nt) are so vsrions aa to 
call for farther deacription. 

S41. The inorementa of x arc Bometimea ra and rH, instead of 
ar and Kr: namely, especially, where by ancb reversal a difficult com- 
bination of coDSonants is avoided: thus, from Vd^f, draktrr^mi and 
Adrftk^BLin; but aleo p]lb& and pratli, P|«li and pracb, Iqpa and 

942> Id a number of roots (about a dozen qaotable ones) ending 
in X (for more originsl arl, the x ehangea both with or, and more 
irregnlarly, in a part of the forms, with ir — or also with nr (espe- 
cially after a labial, in p^, mx, vx, sporadically in others]: which ii 
and ur, again, are liable to prolongation into Ir and Dr. Thus, for 
example, &om tr (or tar], wo have tarati, tltarti, tatftra, atfiri^am, 
by regnlar proceaBea; but alao tlrati, tlryati, tlrtvft, -tlrya, OnfM., 
and even (V.) tmySma, tntury&t. tartorft^a. The treatment of anch 
roota haa to be described in speaking of each formation. 

a. For the pnrpoBe of aitlflcially Indicating tbia peculiarity of treatment, 
inch root! are by the Bioda grammarians written wltb long ft "r v'th both 
r and X'- no f actnally appears anywhere tmong tbeir forms. 

D,j,i,...., Google 

848 — ] 111. EnPHOSIC COUBIHATION. 84 

b. The (qootabla) f-ioots »re 3kf ttrew, Igf ting, 2gf hboOow, 
Ijf tctar out, tf, l9r eruth. 

a. Tba (qnoUblo) y and f-rooW sra f, 1 df pierce, 1 pf _;U^ 1 mf iKe, 
2v7 cAoOfK, Btf, bvr- 

d. Fotms iDtlogons vitii these are lomelimet m&de ilio from other 
roots: thas, cln^ olrCrS, oaTOary&, fiopn yoor; HpHrdh&n and 8par> 
dhfcBe from yspfdh. 

S43. In & few cases f cones from tlie coutrictjon of otber syllables 
than ar and ra: tbns, in tpta and t^tlya, from rij Id 9^911, from ru; in 
bhrk&tl, from rQ. 

Vowel -lengthening. 

844. Vowel -lengthen iug conceraB especially 1 arid n, since the 
lengthening of a is in part (except where in evident analogy with 
that of i and u) IndiBtlnguiahable from its increment, and f is made 
long only in certain plnrttl cases of etems In f (or ar: 960 ff.). Length- 
ening is a much more irregnlai and sporadic change than increment, 
and its cases will in general be left to lie pointed oat in connection 
with the processes of inflection and derivation: a few only will be 
mentioned here. 

345. a. Final radit-al 1 and n aie espedall; liable to prolongation 
befora y: as In paasive and gerund and to on. 

b. Final radical ir and ur (from Taiiable f-roots: 843) are liable to 
prolongation before all consonants except those of perEonal endings; namely, 
before jr and tvS and na: and in declension before bh and a (38S). 
Radical is has the same prolongation in declension (898). 

346. Compensitory lengthening, or absorption by a Towel of the time 
of a lost following consonant. Is by no means rommoD. Certain Instances 
of it have been pointed out above (176, 188 o,d, 168 d, 888b). Perhaps 
snch cases as plti for pltarB (371 a) end dhani for dlianins (436) are 

' to be classed here. 

S47. The final vowel of a former member of a compound is often 
made Ion;, especially in the Vedi. Prolongations of flnal a, and before v, 
ate most frequent; but cases ate fnnnd of every variety. Examples are: 
der&vi, Tayun&Tid, prSvff, jthvaaa, {ndrftvant, sadanSs4d, ;auU 
magha. Tigranara, AkSdaga; apijti, parin&h, vlrudh, tnvlmagh&, 
tTifimsLSt, q&ktltrant; Tasted, anuradh, sum&ya, puruT&au.| i 

848. In the Teda, the flnal vowel of a word — genenlly a, mach 
less often i and a — is In a large nnmber of cases prolonged. Uanally 
the prolongation takes place where It is taroied by the metre, but some- 
times even where the metre opposes the change (for details, see the various 

Words of which Ihe flnsls are thna treated are: 

Ugil.zecy Google 

85 Vowel-lengthening. [— SBO 

». Putlclss: namely, Athft, kAhtL, ev£, uta, ghi, b&. Hit, iv&, 
e&, Bma, nt, afigtt, kiia, &trS, yktrSt, t&trft, kutr&, &ayditrS, ubhay- 
Urft, adya, iochS, &p&, pr£; &tT, ni, y&di, nahi, abtd, Tl; u, Hi, 
nd, Bli, mak^d. 

b. C»ae-fonn8- agpeeislly Initr. elng., »■ eni, Unft, y6na, bv6ii&, 
kod othen ; rarely gen. eiiig., sa uyft, hori^&syS. Cue* besides tliese 
US few: BO Bim&i v^fubbft, barlyojanft (toc); tanvi Cl°'-}l '"'I urd 
aad (not rately) pnrli. 

O. Teib-foims ending in a, in great number and variety: thua (nearly 
Id the order of their comparatife ftequ«acy), 2d alng. Impr. aot., aa pibA, 
■yft, gamayft, dhftrAyft; — 2d p). act, in ta and tha, *a BtbS, attS, 
blbhftS, Jayatft, ^rVatS, anadatd, nayathA, jivayathft (and one or 
two In taaa; aviftanK, bantanS); ^ lit pi, art, in ma, u vidmS, 
rlfkmft, Fdtay&mfi, rtihomS, vanuyfimfi, cabr^A, marm^jmS; — 
2d sing. impT. mid. In ava, aa yukq^S, i^i^Tft, dadhl^Tft, Tahaavi; 
— 1st and 3d sing. perf. act., as vedfi, vivegS, jagrabhA; 2d sing, perT. 
act., vatthK; — 3d pi. perf. act., anajfi, oakrA, Of Terb-forms ending 
in i, only the 2d alng. impv. act.: thus, k^rdhl, k^uhl, k^ldhl, qrudhl, 
ffQadm, qrnnhl, didlhl, Jahl. 

d. To these may be added the gerund in ya (883 a), as abhlgdryO, 

Vowel-Ugbtening . 

848. The alteratiOD of short a to an 1- or a-vowel !n the fonnatiTe 
procoHBeB of the Ungnage, except ia ^ or ar rooU [as explained above]; 
IB a sporadic phenomenoc on If. 

S50. But the lighteoing of a long B especially to an l-rowel 
(ae also its loss), is a frequent procesa; no other vowel is so nn- 

a. Of the clasa-iign aS (of the krl-clBss of verbs: 717 ff.), the 
i Is In y-tA forma changed to I, and before Towel-endiaga dropped alto- 
gether. The final & of certain roota is treated In the same manner: thus, 
mft, hi, etc. (6dS-6). And from some roola, fi- and i- or i-forma lo 
Interchange that it Is difficult to claaslfy them or to determine the tme 
character of the root. 

b. Radical fi la weakened to the temblance of the anlon-vowel 1 In 
certain verbal forma: as perfect dadlma from yd& etc (794k); aollst 
adhlthfis from >'dta& etc. (834 a); preient JaMmaa bom yb& etc. (ae6> 

O. Radical fi la shortened to the semblance of stem-a in ■ number 
rednplicated forma, sa tlffba, piba, dada, etc.: see SVl-li also in a 
few aoriets, aa Uivam, &khyam, etc. : see 847, 

d. Badlcal fi sometime* becomea e, especially before y: aa Btlieyfiaaiii> 

ioy Google 

261—] HI. EnpHOMic CouBiNATiOM. 86 

SBl. Oertkin A-rooU, twcftuH of their peealiu BXchingM willt I ind 
i-toTEU, MpeeitUy in ftinnlog tbe pi«*ent (Mm, an gjTan by tlM Hindu 
gnmrnoiiuu u loots ending In a ot U oi o. Thui, from 2 dhft tucit C^Im) 
come the pieeent dh&Tati >nd participle and gamnd <UlIt&, dhitv&i the 
othei loraxt are made from dbft, ae dadbtiB, Kdhftt, dliftByati, dbtttava, 
dUfpi^atl. From 2g& ting (gU) cDine the present gi^tl, tbe parti- 
ciple and geruDd git& and gitva, and paasiie giy&te, and the other fonui 
tlom gS. From 3 d& cut (do) come the present dy&tl and participle dlt& 
or diiiA, and the other forme from lU. The InegQlarltleB of these root* 
nill be treated below, under the varioqe fonnations (see eapeclallf 761 d (T.). 

263. Bj a proeeas at abbreviation eaientlally akin with that of nr or 
ra to f , the va (uiaally initial) of a number ot roota becomes u, and the 
ya of a much smaller number becomes i. In certain verbal forms and derlv' 
atlTes. Thna, from vaa come avaoa, ooyMam, nktvo, ukt4, vkti, 
ukthA, etc. ; from y«J come iyii&, Ijyiiaam, Ul{vA, ifft^ i^^, etc See 
below, under the various formatloDs. 

B. To this change U given by Earopean grammiriant the name of 
aaifapr>»Hnu;a, by adapUtlou of a term used In the native grammar. 

2fiS. A short ft, of root or ending, is not infreqnentl; lost between 
oonaonants in a weakenad syllable: tbna. In verb-forma, gtan&nti, ftpapbun, 
Jagmas, Jftjfins, AJfiata; in noun-forma, Tf^fie, r$jfiL 

264. nalon-voweli. All the simple vowels come to aasnme in 
certain oases the aspect of nnioo'vowels, or insertions between root or stem 
and ending of inflection or of derividoa. 

ft. That character belongi ottenest to 1, which ia very widely osed: 
1. before the of aorist and fatare and detlderative stenu, as in ^Irlf am, 
JiTiqy&ml, jiJIvifSmi; 2. in tenae-lnllection, eapeoiaUy perfect, ai JiJI- 
vim4i occasionally also present, as &iilti, r6dlti; 3. In derivation, as 
JIvit&, ktUoltiun, Janitt. rooi^i^n, etc. etc. 

b. Long I Is used sometimes instead of short; thns, Agrwhiyan, 
grahli^imi; bntviti, T&vadltl; tarltfi savlt^; it is also often intro- 
duced before b and t of the 2d and 3d stng. of verbs: thus, ttiie, itHt. 

0, For details respecting these, and the more irregular and sporadic 
occarrences of n- and a-vowela In the sane character, see below. 

Naaal Inorement. 

265. Both in roots And in endings, & distinetion of Btronger asd 
weaker forms is very often made by the fH^sence or abMnce of a 
naeal element, a nual mate or anasvBra, before a following eon- 
Bonant. In general, the stronger form b doabtleu the more original; 
bnt, in the present coodftion of the language, tbe nasal has cobs in 
great meaaore to seem, and to some extent also to be naed, as an 
actually strengtbening element, introduced luidei certain conditions 
in fonnaUve aad inflective processes. 

Digitizecy Google 

87 Nabai, Ihckbhbnt. f— 900 

». Bxtmple* »«, of roDts: ao >nd afio, srath knd granth, vid 
and vind, da^ tnd d«Ag. eras and sraAs, dpta &nd dfAh: of endlngt, 
bhijrantam uid bbiiratA, m&naai utd m&aaAal. 

SB6. A final n, whathet of item di of loot, it lesa itatile than any 
othei conionant, wh«ie a weaker fonn ig callad for: thna, froin ri^an *e 
luTO r^K and r^abhls, and la composition r^a; frcm jtVin-n fff, dhani 
and dhanibble and dhAnl; hom ybaa we have bathi and baU, etc. 
A final radical m 1« Mmetlmea treated In the same way: thos, from Vgam, 
gahf, gat&m, gati, g&tl. 

SB7. Inserted n. On tbe otbei band, (he nactl n ha* coma to be 
Died nlth freat — and. In the later hiatory of the language, with Increaajng 
— freqneacy i« a niiiiin-eonaonant, Inserted between rowels: tbns, from agui, 
ftg"^"fr and agniD^ ; from m&dba, m&dhnnaa, rr>4^hTiiit, mJidlifiiii ; 
ftom qtvk, qlvtoa. glviiil, ^IvanjUa. 

S58. Inserted y. a. After final g of a root, a y Is often found ss 
apparently a mere nnloD-censonant before another vowel: Qint, in luflectlon, 
&dUrl etc (.844), f i^r&yati eto. (104B), flviyfts ete. (SaS o), gitrati 
etc. (761 e) ; further, in deriistlon, -gSya, -yiyam, djiyaka etc.; 
•BthAyika; pfty&na. •gftyana; dlii^aB, 'hSyae; athlljln etc (many 
ca«ei)i -Utftrin, •taUyin; athSyoka. 

b. other more gporadio cases of inserted y — such as that in the 
pronoon-fonns ayam, lyam, vayam, yuyam, svayam; and In optatire 
inflection before an ending beginning with a vowel (B69) — wUI b« point- 
ed out below in their o 

BedupUo ation . 

868. Reduplication of a root [originating doabtleBs In its com- 
plete repetition] has come to be a method of radioal inoremeDt or 
Btrengthening in varions formatiye proceseee: aamely, 

a. in preBsnt-Btem formation (94S£f,]: as ditdftmi, blbbirmi; 

b. Id perfeot-Btfim formation, almost uniTeTsallylTSSff.): iBtattna. 
dadbftu, oakara, rireoa, lalopa; 

o> in aoTiflt-Btam formation (866 ff.): as ddldbarami, Aonoyavam; 

d. in intenBlve and desideratiTe-Btem formation, thronghont(1000ff., 
lOaeff.): as j&flgbantl, J6taaviti, marmrJy&tai p{pMati, JigbAfiaati; 

e. in the formation of derivative noun-Btems (1143 e): as p&pii, 
o&roara, sAsahi. oibiti^ mallmlnoiL 

1 Bales for the treatment of the reduplication In these several cases 
will be given in the proper coDnection below. 

S60. As, by reason of tbe Btrengthening and weakening dianges 
Indicated above, the eame root or stem not Beldom exhibits, in tbe 
prooessea of inflection and derivation, varietieB of stronger and weaker 
form, the distinction and description of these varieties foime iw Im- 
portant part of tbe subjects hereafter to be treated. 

Digitizecy Google 

IV. Deglehsion. 



Sdl. The general subject of declenBion inelndeB noans, adjectiTos, 
and proaooDB, all of which are inflected in essential])' the same manner. 
Bnt while the corieapondence of nouns and adjectives is so close that 
they eaoDOt well be separated in treatment (oliap. V.), the pronoans, 
which exhibit many pecnlarities, will be best dealt with in a separate 
chapter (Vll.) ; and the words designating number, or numerals, also 
form A class peculiar enough to require to be presented by them- 
selves (chap. VI.)- 

262. Declensional foims show primarily case and num- 
ber; but they also indicate gender — since, though the 
distinctions of gender are made partly in the stem it»elf, 
they also appear, to no inconsiderable extent, in the changes 
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 lavs 
of distribution as, for example, in Greek and Latin. 

a. The only words which show no sign of gendet-dlBtlnetlon >re the 
peraoiul pronouna of the first tmd second person (481), sud the nomeTsts 
iboTe foar (483). 

264. Number. The numbers are three — singular, dual, 
and plural. 

a. A few wotda ue nsed only in the plursl: ss dftr&a wi/s, £patt trotor; 
the nameial dva ivJO, Is dnU only; and, as In other Isngoages, man; words 
are, by the nature of thelt use, toand to occur only in the eingnlai. 

Ses. As to the uses of the numbers, it needs only to be remarked 
that the dual is (with only very rare and sporadic exceptions) nsed 
strictly in all cases where two objects are logically indicated, whether 
directly or by combination of two indiTiduals: thus, fivd te dytU 
Vftpfthlvi ubh^ Btfim mayhtaven and earth both be propitiout to thee! 
dalvaifa oa mSnufaih oa hotfir&u vf^yS haoing choten both the divine 
and the human aacrificert; pathor deTayHnasya pltfyS^aBya oa of 
the two paths leading retpeetively to the gode and to the Fathers. 


89 Cases. - [—368 

a. The dD>l 1« used alone (irittaont AvA Imo) properly irlieD the. 
dnality ot the object* indicated Is -well nndeintood; thas, afvinau Ika finn 
Amint; indraBys bin Indra'g two bagt; bnttaayadvfiT afvBu sttt^ 
A« Aoi two AoTM*. But notr and then the dual Blands alone piegnautly : 
tfans, TOdaih vedftu Tfldftn v& one Veda or fico or more than Ctcn; 
Aaf af(e fate (tea hundred and lixty-ont. 

266. Case. The cases aie (including the vocative) eight: 
nominative, accusative, instiumental, dative, ablative, gen- 
itive, locative, and vocative. 

a. Tb« order In which they ate here raeationed !■ that eslabllBhed for 
them by the Hindu grammsilanB, and accepted from these by WnBtern 
BcholaiB. The Hlndn names of the cases are founded on this order: the 
nominatira is called peatlXAml, Jirat, the accnaattTe dvltlyi second, the 
genitive faffhl eixlh (ec. vlbbaktl dinition, i. e. ccae), etc. The object 
aonght in the arrangement is limply to set next to one another those cases 
which are to a greater or Ihsb extent, in one or another number, identical 
in form; and, patting the nominitWe Brst, as leading case, there is no 
other order by which that object could be attained. The Tocative is not 
considered and named by the native grammarians as a case lilie the rest; 
In this work, it Till be giveii in the singular (where aland it is eier dis- 
tinguished from the nomtnative otherwise than by aec«Dt) at the end of the 
series of Msea. 

A compendions statemeiit of the ueee of the cases is given in 
the following paragraphs: 

867. Uses of the Nominative. The nominative is the case 
of the subject of the sentence, and of any vord qualifying the sub- 
ject, whether attrihatively, in appositioo, or as predicate. 

388. One or two peculiar constructioDS call for notioe: 

a. A predicate nominative, instead of an objective predicate In the 
accasative, Is used wlUi middle verb-forms that signify regarding or calling 
one's self: thns, s6maih manyate papivan (BY.) kt thinii he has been 
drinking eoma; ek maoyeta porft^avft [AV.) he may regard himielf as 
xei»e in ancient things; dur^d va &hartii 'vocathftli (MS.) Aou hast 
claimed to be a savior out of trouble; {ndro brfthma^o bruvS^a^ 
(TS.) Indra pretending to be a BroMnim ; kattbase aatyavftdl (R.) thou 
hoaslsst thyself truth/a!. Similarly with the phrase rflpaih i^: thus, 
kn96 riip&lil k^tva (T9.) taking on a black form (1. e. making shape 
for himself as one that is black). 

b. A word made by Iti (1103) logically predicate to an object is 
ordinarily nomlnatiTe: thus, Kvexgd lok& fti y&ih T&danti (AV.) tchat 
they call the heavenly tcorld; tam agnlffoma ity koakfate (AB.) if 
they style asnlft^ma; vidarbharSjatanaj'&di dam^anti tl viddhl 
mSm (MBh.') know me for the Vidarbha-king's daughter, Damayantl by 


SS8— ] IV. Deolshsion. 90 

atmtt. Both MiMtnietloiii ue eomblaed In ^fiadi bl bUun 1^ Uia^ 
pite 'ty vn tn mmtradam (H.) for to an ignorant man tkag givo Iht 
name of 'chit^, hut that of '/athtr' to ont who mparta the taered ttxtt. 

O. A DomiiittlTe, Instead of ■ Mcood voMtiie, !■ aDmallinei added to 
» TOMtiTe ky oa and: thni, {ndraf ok aimeLih ptbat«di brhaapate 
(RT.) togttker wtth Indra, do j/o two drink the »oma, O Briiatpaii! vifr* 
dev& J^amSnaq ok eidatS (TS.) O ye AU-Oodt, and tht taerijittr, 
tait not*! 

869. Uses of the aCDuiative. The scousatEve [b eBpecialty 
the case of the direct object of a tnnsitlTe verb, uid of kdj word 
qo&lifjdQg that object, u attribate or appoBltive or objeodre pradi- 
c»te. The construction of the verb ia shared, of course, by its par- 
ticiples and iDfinitives; but also, in Sanskrit, by a number of other 
derivatives, having a more or less participial or infiaitivaJ character, 
and even Bometimes by nouns and adjectives. A few prepositions 
are aecompaDied by the accusative. As less direct object, or goat 
of motion or action, the accuwtive is construed especially with verbs 
of approach and address. It is found used more adverbially as ad- 
junct of place or time or manner; and a host of adverbs are accDS- 
ative cases in form. Two accusatives are often found as objects of 
the same verb. 

570. The OEs of tha ■ociuatiTe si direct object of ■ ttaasltiTe verb 
and of its InflnltlTet and paittcipleB hardly needs iUnstratton; an ezampla 
or two ate: agnim I^e I praiie Agni; n&mo bharanta^ bringmg 
homaffe; bhAjro d&um arhasl thou thouldtt give more. Of ptedteate 
irords qoalifyiDg the object, an example is t&m osr&di bf^oml t&lfa 
brahm^iqana (RV.) him I make formidable, him a priett. 

571. Of verbal dBiivatives having bo far ■ participial chacaeter that 
they share the cooitiuctlon of the verb, the variety is coDBlderable : thas — 

a. Derivatl'ea In u from deaideratlve atems (1038) have wholly tha 
character of preaenC paitlclples; thus, damaTantitO abhlpBavab (MBb.) 
detiring to win Damat/anii; didfkfur Jan&k&tmaJSm (B.) detiring to 
lee Jantdide daughter. Barely, alio, the verbal noun io & from such a root: 
thus, Bvargam abhikfifikqayS (R.) with deaire of paradiie. 

b. So-i^alled primary derivatives In In have the eame oharaeter: thoa, 
m^ kftmfnT (AV.) loving me; anam abMbhft^ifl (HBh.) addretting 
him. Even ttae obvianaly aeconduy garbbin lias In QB. the same cod- 
ttruction : tbna, B&rv&^ bhutinl garbby kbbavat he became pregnant 
toiih all beings. 

o. Derivatives la aka, iu the later language; ai, bhavantam abhl- 
vKdaka^l (MBh.) intending to lalute you \ mlthUfim avarodhaksk^ (R.} 
betieging MilAila. 

d. Noam tn tar, very frequently la the older langoage, and as peri- 
phrastic future forma (fl48 IT.] in the later : thua, h&nt& 76 vj^Un 


9l UbBB op the ACCOBATIVB. [—873 

•Anlto tk vtiiaSx dttA mag&inl (RV.) who tlayttk the dragon, lomneth 
beett/, butoteM largemet; Uu hi >duh surradi hArt&r&D (JB.) for 
thty teize on this universa; t7akt&r»h aaibyuge pr&^Sn (MBh.) riaking 
life in hattU. 

e. Tha root iUalf, in the older Uaguige, used with the Talus of a 
pieMDt paiticiple at the end ot a campouud : thns, y&ifa yi^fUih poribhtlr 
itA (BV.) tehat offering thou lurroundeat Iprotectesl); &bim a.p&fy pftrl- 
ffham (KT.) (A« dragon confining tht tcatert. Also a aupeilaiiTa of a 
toot-Btem (468, 471): thus, tv&ifa v4su devayaU T&nl^thati IRV.) thou 
art chief ainner of amUh for the piout; Ut B6maih BomapttamA (RV.) 
they tu>o are the greattet drinkeri of loma. 

t. The deiiiaCive In i from the (especially the redtipUcated) rool, in 
the older langaago: tboB, babhrfr vi^Tadi pap{^ admain dadir %ii}. 
(RV.) bearing the ihattderbolt, drinking the soma, bellowing kiiie; yajfi&m 
fttini^ (RV.) extending the laerijice. 

g, DailTativM in nka, very freqnently In the Bribmana Unguige: 
thus, vatB^Ag oa gblttuko vfkal^ (AV.) imdOie icotf deetroge hie calvee; 
vMoko v&o bharati (TS.) he wine a garment; kamuk& enaih Htrifo 
UuiTaiitl (MS.) the uiomen fall in love with him. 

b. Other casei are mote sporadic: thaa, derlrittves In a, as inAxo 
df^hS eld Sn^&h (RT.) Indra breake up even lehaf ia fatt; nfti V& 
'rha^ p&itrkaifa liktham (H.) bg no mtant entitled to his father'^ 
atale; — in atnu, as vt^d dd Kn^atnubhl^ (RV.) with the breakeri 
of whatever ie etrong; — In atha, as ;^)&th&7a devan (RV,) to make 
offering to the gode; — in ana, as taifa nivSrai^a (HBh.) in reetraimng 
him; BTamftflBam iva, bhojane (R.) at if in eating one'e own fleeh; — 
in ani, aa Bam&tau turir&^l^ pptanydn (RV.) overcoming foee in 
eombatt; — In tl, aa nk t&lh dhOrtfh (RV.^ there ie no injuring him; — 
In van, la ApafoaddaghTt 'rmam bliavatl (H9.) he doet not come 
thort of food; — in snu, as sthlrj oin namayif^vab (IIT.) bowing 
enen Jim thinge. 

S7S. Eianples of an aecnaatiie with an ordinary noim or adjective 
■le only occasional ; snch words as AuuTrata faithful to, prAtirfipa 
corrwtpondiitg to, abbldhr^a daring to cope witk, ptatT&flo oppoeite 
to, may be regarded aa taking an accusative in Tirtna of the preposition they 
contain ; also Annlra, ta ^tik& dsTa v&ni9am (HS.) the gode are inferior 
to Varw^ RT. has tim ant&rvati^ pregnant wUh him; and AV. has 
mtth kltmena through loving me. 

S73. Th« direct eonatractlon of caie* with prepositions is compara- 
tlTely loftrlcted In Saoikiit (llSSff.). With the accuaatlTe are oftenest 
boiid pratl, oppoeite to, in reference to, etc. ; also ann after, in the count 
of; aatar or antarS between; rarely ati aeroee; abM againit, to; and 
others (IISS). CMe-fotms which have assumed a prepositional value are 
alto often Q«ed with the aconaative: as antora^a, attare^a, dakfl^ena, 
avare^a, QrdliTaiii, rte. 

ioy Google 

874—] DECLBNaiON. 92 

S74. Tbe accuBati>e la very often found alM ts object of verbs which 
in the related lingnages ue not tiuiBitiTe. 

a. It aUDilii especiallir a* the goal of motion, with verba of going, 
bilDging, aeuding, and the like ; tbua, vidarbbKn agaman (HBh.) thty 
aent to Vidarbha; dlTaifa yayu^ (UBh.) (A«y went to heaven; vauagnl- 
m&n db&Tanta^ (UBh.) rutming to aooda and buiha; ap6 dlTOm 6d 
TBlianti (AV.) thty carry up toatert to tha tky; dev^ 74« (^^-) ^ 
mtAe offermg to the godt. 

b. With Terba meaning go, this ia an extremely common oonatniction ; 
and the nae of iuch ■ verb with an abstract noon makes pecnliir pbnteB 
of beeoming: thtis, Btunatfim etl he goe$ to equality (i. e. hecomee equal); 
BB gacched badhyatOm mama (HBh.) he ehall become liable to be elain 
by me; Ba paitcatTam agata^ (H,] he ioae reeobied into the Jive ekmente 
{underwent dieeolutien, died), 

o. Veibi ot Bpeaiiing foltow the aime rule: thus, tun abravit he 
taid to him; pr&kro^ad uoo&ir nSi^adham (HBb.) afta eried out loudb/ 
to the Niehadkan; y&S tvo "vltoa CAV.) mho upohe to thee. 

d. The aaSDiDption of an acensatiire object ia exceptionally easy in 
Sanskrit, and anch an object is often taken by a verb or phrue which ia 
atriotly of intranaltlve rharacter: thns, aUiaB& prd 'ay aay^ (.^^^ >'» 
might thou exeelleet (lit. art ahead) others; deTa v&{ br4hma a&m 
avadanta (MS.) the gode tcere ditewaing [lit. wars talking together) 
brabman; ailtii vSi icA yaii&d yantl (MS.) lurely they are cutting 
me off (lit. are going between) from the offering; taih B4iii babbava 
(^B.) he had inlercouree with her. 

27B. Examptea of the cognate aconsitiTe, or accusative of implied 
object, are not infrequent: thus, t&paa tapy&mahe (AV.) we do penance; 
ti h&l 'tarn edbatum edhadi oakrixe ((B.) the;/ proepered loHh that 
prosperity; u^itvA Bukhavaaam. (R.) abiding happily. 

278. The accusative is often naed in more adverbial constraotlons. 

O. Occasionally, to denote meaaure of spice: tliaa, yojanaqataib 
gantnm (HBh.) to go a hundred leaguee; ^a^ aoohrlto yojaninl (HBh.) 
six leagues high. 

b. Hucb more oft«n, to denote meaaare or duration of time: thoa, ak 
BaibTatHarim Ordhrb tlt^that (AT.) he stood a year upright; tlsrA 
ratrir dlkfit4h syftt (TS.) let him be consecrated three nights; gatT& 
trill aborStr&n (HBh.) having traveled three complete days. 

o. Sometimes, to denote the point of epace, or, oftener, of time: thns, 
ylbn aaya di^am d&synh syat (QB.) whatever region hi* enemy may 
be in; t^nftl tliifa ritrlih Bab^ " JagSma (QB.) he arrived that night 
with him; Imfidi r^janiiii vyuffSm (UBh.) this current night. 

d. Very oflen, to denote manner or acoompanying clrcamstance. 
Thns, the neuter acousative of Innnmerable adjeotivea, simple or componnd 


93 Uses of the Accusatitb. [—378 

(1111), U nsed advetbuny, while certain kinds of eompounda ue thus 
DMd to such an extent Ibat the Hindu grammariaiij baTS made of Cbem a 
special adierbial class (1313). 

e. Special cskge are occasionally met vitb: thas, brahmaoAryaiD 
UTtoa t^B.) /la kept a term of »tudenUhip ; phaliifa paoy&nte (MS.) thty 
ripen thair fruit; gaih dtryadlivani (M3., S.) gamble for a cow. 

9,11. Tbe aecusative i«, of course, freely used wiii other cases to limit 
the same Terb, as the sense reqnltes. And wheneveT it is nsable vitb a 
verb in two different const ractions, the verb may take two accasative:, one 
in each constraction: and such combiDatlons aie qaite frequent in Sanskrit. 
Tliiu, with verbs of appealing, asking, having teeonise: ss, ap6 yAoftml 
bhe^iO&m (RV.) / a»k the ualert for medicinsi tvSm abam Batyam 
ioohimi (R.) I detire truth from tht»:, tvfim vayaih qaraaath gatfi^ 
(MBb.) tee have resorted to thee for euccor; — Kith verbs of bringing, 
sending, following, imparting, saying; as, gurutvaiii naradi nayantl (H.) 
they bring a man to re^ectability; utfi oS 'nvetu mfiib vanam (R.) 
and let Sila accompany me lo the forest; Bupeqasaih mi "va Bfjanty 
Astain (KV.) they let me go home weU adorned; t&m idam abravlt (Mllh.) 
this he said to Tier ; — and in olbec less c-ommon cases : at, v^rkf&lfa pakv&lil 
ph&ladi dbonulii (RV.) shake ripe fruit from the tree; tadi vif&m 
tvi 'dbok (AY.) poistm he milked from her; jitTfi rajyaih nalam 
(MBli.} having iBon the kingdom from Ifaia; ixoM^^ltaii paj}im gify (RV.) 
ye robbed the Pani of the kine; draqfum icohSva^ putraiii pa^cimadar- 
^anam (R.) uJe wish to tee our son for the last lime. 

a. A causative form of a transitive verb Tcgnlarly admits tvro acou- 
lative objects; thus, devaA nqat&^ pfiy^B havl^ (RV.) make the eager 
gods drink the oblation ; o^adhir eT4 ph&laib grfthayatl (HS.) he makes 
tht plants bear fruit; vai^ijo dSpayet karfin (M.) he should cause the 
merehanlt to pay taxe*. But such a caoGsttve sometimes takes an instin- 
mental instead of a second aecnsative: see 282 b. 

878. Daee oT the Instramental. The mstrumenta] is orig- 
inally the ui'fA-case: it denoteB &dja(«ncy, accompaDiment, aisociation 
— passing over into the expreasion of meaofl and iDBtrnment by the 
same transfer of meaning which appears in the English prepositions 
leith and by. 

a. Nearly all the uaea of the ca«e are teadily dedncible from this 
fundamental meaning, and show nothing anomaloai or difflonlt. 

S70. The instrumental is often nsed to signify accompaniment : thus, 
agnir derebhir a gamat (RV.) may Agni eome hither along toith the 
gods; marndbhi rudr&ih bnvema (RT.) we mould call Rudra with the 
Maruls; dTftparena Baht^ena kva j&ayasi (UBh.) whither wilt thou 
go, with Dvapara for eampaniont kathayan nSi^adheita (MBh.) talking 
viiA the Nishadhan. Bnt the relation of simple accompaniment is more 
often helped to plainer expression by prepositions (eaha etc.: 284). 

D,j,i,...., Google 

SSO— ] rv. Declension. 94 

380. The fnatniiaenUl of metng ot Instrnmant or igent U jet more 
fieqaent: thai, blutdr&dl hir^Abhl^ 9r9iV9>i>u (RT.) nuiji ve hear 
teitli our tart what it propitiout; ^astreijift nldhaium (HBb,] deaA fiy 
tht t¥iori; keolt piMlbhyKib hats K^fti^ (HBh.) lome mere tlain by th« 
ekphanU aHh their feet; pithak pt^ibby&ih darbhataro^akUr 
navanltenS "figufthopakajii^tMbSbliy&in akfl^I AJya (AQS.) anoint- 
ing their et/ee with freth butter, by he^ 0/ the 6uncA«i of daxbhn-^aie, 
toiih the thumb and ring-Jinger, ueing the two hande euceeteivefy. And 
this pusea eully over Into tbe expTession ot occulou 01 loason (fai wblch 
the (blstlTe li moia trequent}: thae, k{pay& through piiy; tana satrena 
tn virtue of that truth. 

381. Of ipsdkl applleidoDB, Ibe foUowlDg mi; be DOtlced: 

a. Aorordance, equality, llkeDees, and the like; thas, samiiii Jyotltk 
sArye^a C^V.) a brightrttsi equal with the sun; ye^Sm ahattl Oft 
pSdaraJaafi tulTab (MBh,) to the Aial of tnhoae feet I am not equal. 

b. Price [by which oblalaed): thus, da^&bhi^ kilQ&ti dh»il4bhllii 
(RV.) he huyt with ten kiae; ^vHx ^ataBahaBreoa dlratftih ^baU 
mama (B.) J^' ^abdla be given me for a hundred thoutartd eoae; ea te 
•kiahrdaradi dfttft rajft 'qvahrdayena tU (MBb.) the king uHl give 
thee the eecret eeienee of dice in return for that of horeee. 

o. Hedium, >nd henoe tlta epaee or distance or lOad, traTersed ; thtu, 
ndna nk nivam anaraata (BV.) they brought [him] at it were a A^ 
by water; i 'hi yfttadi path£bhir derar^n&U^ (RV.) come hither by 
god-travekd path*; jagmur vihiVaaa (MBh.) they went off through 

d. Time paaied thioagb, or b; the Upie of whloh aDTtbtng \e bionshl 
aboat: thus, vldarbbin yftttun iooh&my ekShnft (MBh.) 7 wiA to go 
to Vidarhha in the course of one day, te oa k&lena mahatK yftaTaoam 
pratlpedire (R.) and they tn a long time attained adolescence; tatra 
kUena JSyaiite in&navft dirghaJlTina^ (H.) there in time are bom 
men long-lived. This use of the initrameutal boidera upon that of the 
locatiTe and BbUtife. 

e. The part of the body on [or by) which snythlnj Is home Is nenally 
expressed by the inatrnmenial : as, kukkura^ akandheno "bjate (B.) 
a dog it carried on the thoulder; and thli constmctlon is extended to ench 
cases as tulayK krtam (H.) put on [I. e. so as to be carried by) a balance. 

t. Not infrequent are such phrases as bahutia kim pralKpena (R.) 
what is the use of (i. e. it gained by) mucA talking? fco nn me JWltenUl 
Vtbah (MBh.) what obf'eet is life to me? nlrojaa tu kim ftu^adhil^ 
(S.) but what has a well man to do leith medicines? 

g. An Instrumental ot accompaniment is orcaaioniUy need almost or 
quite with tbe Talae of an instrumental abeolnte: thus, na tvayfi tra 
ma7& VaBtbitona ka "pi clnta karyft [Paric.) with me at hand, thou 
needst feel no anxiety whatever on this point. 

Ugil.zecy Google 

95 USEB OF THE Instrdhehtal. [ — 286 

S6S. a. Th« construotlon of t pageive leib (or psrtinlple] irlth an 
Inttnimental of the agent 1» common from the eiillaat peitod, and liecomes 
decidedly moie M l*tet, tbe paaalve participle vlth Instmrnental taking to 
no amal eitent tbe place ot an active verb irlth it> mbject, Tbns, yam^na 
d8tt&^ (RV.) given by Tama; ^Ibblr f^ya^ (BY.) to he praited by 
Mgtt; vTldhena JUuh viMIr^am (H.) by the hunter a net [tea*] tpread; 
tao ohratrft JaradgftTeno *ktam (H,) Jaradgapa, hearing thia, eaid; 
mayft stmtaTyam (B.) / shall go. A. predlcat* to the Inttramental tnbjeet 
of aaeb a coaatniction ii, of conrae, alto in the inatrameiital: thas, adllTinB 
tavft "naoare^a tnar& BarvathB bhaTltavram (E.) hmeeforih I ehaO 
akDoya be &y eon^aaion; avnhltUr bliavltaTTaih bhavadbhiti (Vlkr.) 
you rmiit be attentive. 

b. A caaaatlve vetb eometlmea takea an instrnniaDtal Inatead of an 
accDsatlTe aa second object: tbna, tftdi qvabhih khftd^sd rft}& (M.) 
the king should hace her devoured by doge; t& T&ni9eaK 'grfiliayat 
(HS.) he eauted Varuna to teize them. 

583. Han; iaatramental conitmctiDaa are snch aa call In tianaUtloD 
for other piapoaltiona than with or hy; yet the tree iaatmmental relation i* 
nmallT to be traced, eepeciaUf If the etymological senae of tbe itoidj be 
calefull; coaaideied. 

a. More anooialonatr, however, tbe inatrnmenlal is naed interobangeably 
with tbe ablative with woida algnlf ylng separatloii : tbna, TataKlr vfyutft^ 
(RY.) eeparated J¥om their caloee; mi 'h&m ftbn&ii& vi rftdhl;! (AY.) 
let me not be tevered from the breath of life; Ha tayft vyayujyata 
(MBb.) he teat parted fivm her; pKpm&nftl VAi 'naih v{ pnnantl (HS.) 
they eleanaa him from evil (compare Eagllah parted with). The aame 
meaning may be given to the cue eren when accompanied by Baha teith: 
tbna, bhartrft aalia viyogal^ (MBh.) separation from her huthand. 

584. The prepoflltionB taking the inatmmentai (1 187) are thoae aig- 
nifying v>i& and the like; tbtia, Baha, with the adverbial woids containing 
■a aa an element, aa cAkun, iftrdham, Baratham; — and, in general, 
a word oomponnded wiUi aa, aam, saha takea an iastrumental as its legnlar 
and natural complement. Bat alto the preposition vlnS toiOuiut takei 
Bomettmea the inslrnmental (cf, 888 a). 

8BB. Dsea of the Dative. The dative is the oaae of the 
iodireot object— ~ot that toward or in the direction of or in order 
to or for which anything is or is done [either intraDSitiTOly or to a 
direet object). 

a. In more phyiical connections, the osea of tbe dative approach those 
of tbe accosative (the more proper to-caee), and the two are sometimes 
interchangeable; but tbe general value of tbe dative as the tmoard- or for- 
esee ie almost everywhere diBtinctly to be traced. 

886. ThDs, the dative U aaed with — 

a. Worda signifying give, thare out, atetgn, and the like: thus, yo Q& 
d&dStl B&kbye (RY.) who gives not to a friend; y&oohS 'Bin&l qArma 
(RV.) beetote upon him protection. 

,1,1.0, Google 

286—] IV. Declension. 96 

b> WoidB Bigulfying ahotc, announee, declare, and ibe like: thiiB 
dhanur darfoya rSmBya (It.) lAotc the boa to Sama; Ev{r ebhyo 
abhavat oliryalf (RV.) the tun viat manifattd to them; rtupan^aifa 
bbimf^ pratyavedajraa (MBb.) they announced Eitupania to Shima; 
tebhyah prat^iUlya (HBh.) having promieed to them. 

o. Words •ignifying give atttntton, have a regard oi feeling, atpire. 
Mid the like: thus, nivef&ya mano dadhu^ (HDh.) they aet their nrntdt 
upon eneamping; mat^ 'va putr^bhyo m^^a (AT.) he gracioue at a 
mother to her soni; idm aemibbTaih bf^Iae (RV.} nhy art thou angrg 
at uiT kftmftya epfhayaty Stmii (Spr.) the eoul longi for- lone. 

d. Words Biga\lyiug please, suit, conduce, and the like: thus, yadyad 
rocata viprabbya^ (M.) whatever is pleasing to Srahmane; tad 
ftuantyaya kolpate (KU.) that make* for immortaiity. 

e. Words Bignlfyiug inclination, oheieance, and the like: thus, m^ihyaifa 
uamautfim pradfga^ citaaral^ (I^V.} let the four quarters bow themse'.vee 
to me; devebhyo namaskrtya (UBh.) having paid homage to the gods. 

t. Words Bigiiifsing hurling or casting: aa yina du^^i^s iwyaai (A.Y.) 
aith ichich thoti hurleet at the impious, 

g. Id some of these comtructioiis the genitive and locative are al£o 
used : see below. 

2B7. In its more distinodve sense, as signifying for, for the bene_fit 
of, teith reference to, and the like, the dative is nsed freely, aod in a 
great vaiiely of constructions. And this nee puses over into that of the 
dative of end or purpose, which is extremely commoa. Thas, ifom Imj* 
Tana iaan&ya (AT.) making an arrow for hurling; gphn^ml ta s&n- 
bhagatvaya b&stam (RV.) I take thy hand in order to happiness; rftft'aya 
mUiyaifa badhyat&di sap&tnebbyah parfibhi^ve (AT.) be it bound 
on in order to royalty for me, in order to destruction for my enemies. 

a. Euch a dative is much osed ptadicttlvely (and ofteneet with the 
copula omitted), iu the sense o1 makes for, tends toieard; also is intended ' 
for, and so must ; or is liable to, and so eon. Thus, npade^o mnrklifiQaiii 
prakopSya na q&ataye (H.) good counsel [tends] to the exasperation, 
not the conciliation, of fools; sa oatasy&h aaibtOffiya nfi 'bhavat (H.) 
and lie teas not to her satisfaction; angopa ael nil diibhfiya (BT.) thou 
art a good herdsman, not one for cheating (i. e. not to be cheated). 

b. These uses of the dative are in the older language especially itlos- 
trated by the dative Inflnitivas, for which see 982. 

288. The dative is not nsed with prepositions (1124). 

288. Uses of the Ablative. The ablative is the /rom-case 
in the Tarioug sensea of that preposition ; it ia need to ezprees removal, 
BepaiatioQ, distinctioD, issue, and the like. 

200. The ablative is used where expulsion, removal, distinction, re- 
lease, defense, and other kindred relations are expressed: thus, td Bedbautl 
patho TflEam (AT,) they drive away the wolf from the path; m£ pik 


97 Uses of the Ablative. [—292 

gim& pathil^ (S.V,) maf/ tet not go away from the pa^; ^tl -vt ef& 
y^janmuthit (MS.) Aa va-ily gott moay from V\t fact ofth* taerijiet; 
txb ftsmid astu beti^ (^V.) far from as b» your miuUe; pftt&ib no 
vf kftt (RV.) ime m from Aa tBolf\ istabbnftd dyom aTaardsa^ (RT.) 
he kapt (lit. madt firm) th» Ay from falling. 

SBI> The abUtlTB Is med wbeie ^proeednie ot Ifene from lamethlng 
u from 1 soiuce or itaitlng-poinl Is ligaifled: thas, ^ukra k^i^ad aja- 
niffa (BT.) the bright one hai been bom from the blaek one; lobhfit kro- 
dbo^ pTAhh&vatiiWBh.) paeeion arieee from greed; vittfit te pra^im 
aTldain (AV.) I have won thy lif&-brtatk from the wind; ji pr^oyft di^o 
abhldltsanty aamaii (AV.) who attack ua from the taetem g^tarler; tao 
duntrS sakMganKt (MBh.) having heard Outlfrom the troop of frienda ; 
viynr aatarlk;Sd abhB^ata (MBb.) the wind tpoke ftora the sky. 

a. Hence ilao, procednre m From > cause or oocuJon is Blgnifled by 
the iblKtlve: thla 1b eapedally fraqnent in the later language, and In teoh- 
nical phraseology Is a suiidlng eoDBttnctlon ; It borders on Inatrumental 
cDDtttnctions. Thus, TAJrasya ^u^Bd dadfira (RT.) fi-om (by reason 
of} the fitry of the thunderbolt hs burst aeunder; yuya dK^^abhiyrtt 
Mtrve dharmam anorndbyantl (HBh.) from fear of whose rod all are 
eonetant to duly; ak&ramigrltatritd okanuya (Trlbli.) because e con- 
tains an element of a. 

b. Very rarely, an ablative bas the seme or after: thae, agaOOtiaim 
aliOT&trat Hrtbam (MBb.) they went to the shrine after a whole day; 
takSrBt Bak&re takKre^a (APr.) after \, before s, m inserted t 

S82. One oi two special applications of the ablative constructlan are 
to be noticed: 

a. Tha ablative vlth words implying fear (terrified recoil from); thus, 
t&syS jatiy&l; s&rram abibbet (AV.) eoerything was afraid of her at 
her birth; y&smad r^anta k^t^ya^ (^^0 «' whom mortals tremble; 
yufmU bbiyt (BV.) through fear of you; yaem&n no 'dvljate loka^ 
(BhQ.) of whom (As world is not afraid. 

b. The ablative of compstison (dlailnctlon from) : thaa, prii ririoa 
diT& indxab ppthivyab (RY.) Indra is greater than the heaven and the 
earth. With a eompitatlTe, or otber word used in a kindred way, the abla- 
tive la the regular and almost constant constructlOD: thna, av&d61? svjdl' 
ya^ (RV.) sweeter than the tweet; Idih taamftd dn^khataram (MBb.) 
what it more painful than that? ko mlb^fid anyal^ (H.) who eUe than a 
friend; g& avf^Itba mat (AB.) thou hast chosen the kine rather than me; 
i^fiebhyo granthina^ gre^tbi granthlbhyo dbftri;^ var&t (M.) 
poaseseors of text* are better than ignorant men; rememberers are better 
t/tan possessors; t&d any&tra tvbxi n£ dadbmasl (AV.) we set this 
down eltewhere {away) from thee; pdrvft viflvasmftd bhuvauAt (RV.) 
earlier than all being*. 

O. Occaslonsliy, a probably possessive genitive la used with the com- 
parative; or an iugtrnmental (as in a comparison of equality): tbtis, 
Wkitney, Gnmiui, 9. <d. 7 


892—] IV. DE0I.ENSIOK. 93 

n& *Bti dbanyfttaro muna (R.) thers it no one more forbmaie fhan I 
(j, e. my superior in fortune); patraib mamR prft^&ir gar^asam 
(MBh.) a ton dearer Man my life. 

d. OccuionUly, an ablatiTe fe lued instead of ■ putitiTe gsnitiTe; 
thus, miUmnSd ekadi Jaghfina (R.) he tlew one out of U« pair; 
tftbhya «bam (KSS.) one of them. 

293. Tlie nblitfTe Is naed with & varietf ot prepwittons and wor'ti 
■baring a prepositional chaiactei (1128); but ill these hare lathsr an ad- 
verbial Talue, as stiengthening oi defining the yVom-Tolatton, than any 
propel gotarning foree. We miy notlna here; 

a. In the Teda, Adlii and p&ri are mnch naed u directing and «tce&gth- 
enlng adjuncts with the ablative: as, jSt6 Mm&Tataa piri CAV.] horn 
from the Himalaya ifortK); samndrad &dhl JaJ!Ll;e (AV.) thou art 
bom from the ocean; oirantadi p&rl tastJiufali (RV.) moving forth 
from thai tohiiA etande fatt. 

b. Also para (and par&s), In the senie at forvard from, and hence 
before: aa, pari j&raaa^ (BV.) before old age: and hence also, with 
words of protection and the like, from: as fafamSnA^ para nid&^ 
(RV.) eeeuring from itl^iil. 

o. Also a, in the sense of AitA«r^om, aU the way from: as, ^latbAd 
&au ^uqyatu (AV.) let it dry eompktely up from the roof; t&amftd a 
nadyi nama Stha (AT.) since that time ye are tailed rieert. But usn- 
ally, and eapentally In the later language, the meaeuietaBnt of Interval 
implied Id a la reversed In direction, and the constractlon means aU the 
way to, untii: as jratl gir{bli;a a samtldrat (KV.) going from the 
mountairw to the oceani i 'By& yajfid^jro 'd^a^ (VS.) until the end of 
thi* tacrifke; ft ^^f^^'^ C) '•" ^ eixteenth yeai-; & prad&n&t (Q.) 
untU her marriage. 

284. Uees of the Genitive, a. Tlie proper valne of the 
genitive is adjectivali It belongs to nnd qualifisB a doud, designating 
iometbiog relating to the latter in a manner which the nataie of the 
case, or the connection, defines more nearly. Other genitive con- 
stmctiooB, with adjective or verb or preposition, appear to arise out 
of this, bj a more or leas distinctly traceable connection. 

b. The use of the genilJve has become much extended, eape- 
cialiy in the later language, by attribatlon of a Uoun-cbaracter to the 
adjective, and by pregnant veriial construction, so that it often bears . 
the aspect of being a sobstitnte for other cases — as dative, instru- 
mental, ablative, locative. 

28B. The genitive in its normal adjective construction with a noun 
or pronoun is claBslflahle Into the nsnal varieties: ai, genitive of possession 
or appnrtenance, including the complement of implied relation — this is, 
aa elsewhere, the commonest of all; the so-called partitive genitive; the 
anbjacllve and objective genltivea; and so on. Genitives of apposition or 


99 Uses of the Qbhitive. [—887 

equirtleuee (^eity of Rome), and of thwscteriitlc (man of hcnur), do not 
occar, tad htrJIy that of miteiiil [hoiue o/tcoad). Eiamplai are: {ndrft- 
ayftviO'B^ In^a'* IhundtrboU; pltft ^^ttisf&tct father of sotu; putmb 
pitn> ton oflhafaiher; pltu^ kOmal^ putraeya the father » love of the 
eon; ke nafi whiiA of ui; 9&taib diBlD&m a hundred female liavei. 

a. The eipieaalon ol pMunlao etc. on tba pirt of pronoana Is mats 
almoit entiTel; b; the genitive «ate, and not by a derlvad poMeSilva ad- 
jBBtiVB (618). 

b. Exceptional caaea like nftKarasya m&rga^ the road to the city 
(ft, le ehemin de Parit), ynsyK liaiii d^ta Ipaita^ (HBfa,) tw meteenger 
to tohom I am tcanted, are occaii^nally met with, 

see. The geoltlre U dependent on an adjective: 

A. A so-caUed paitiriie genlil'e vlth a superlative, or anothei ward 
of ElmiUi siibstantlTal valne: thui, Qref(haifa TlrO^iian heti of heroes; 
Vlrudh&ib TliraTati (AY.) of plante the mighty {mighOttt) one. 

b. Tory oftaa, by a tran^ter of the p)tae«slve genillTe ^m noun to 
adjeellve, the adjectiie being treated aa tf it had nonn-Talae: tboa, taaya 
ammai^ or anarQpati or sadrgoh rtaembliitg him (I. e. hit lUte); ttwya 
priyll dear to him (hit dear one) ; taayft Vlditam unk/toion to him (hi* 
unkaovm thing) ; h&vyaq oarf a^Inilai (RY.) to be laerijiced to by mortala 
{their abject of taerifice) : ipalto narauftrS^ftm (HBh.) deiired of men 
and women {their abject of de»ire]\ yasya baBya praofltal^ (B.) of 
tohomioeour bom (hie eon); hantavyo 'omi na ta (MBh.) I am not to 
be tlain of thee; Um arthlngiti TafLoajitavyam aati (R.) why shiuld 
there ht a deceiving of eupplianiif 

O, In part, by a coaatrnotlon ilmilar to that of terbj which take a 
genitive object: thn;, abhijttft rSJalharm&f&m (It.) underttanding the 
dutiee of a king. 

S97. The genitive as object of ■ verb It: 

a. A p»ieiglTe genitive of th» teclplant, by pregnant conatraction, 
with verb] li^lfying give, impart, communicate, and the like: that, Tar&n 
pradftyft 'eya (UBh.) having beilomed gifts upon him (made them hie by 
bettoical); rljflo nivedltam (S.) it wot made knovm to Ike king (made 
hie by knowledge); yad auyaaya pratiJbSya punar aoysaya dlyate 
(M.) that after being promised to one she it given to another. Thlj con- 
nrDC'lon, by which the genitive beomei substitate for a dative or locative, 
ab^anda in the later Ungnage, and U extended eonaetimes to prablemattc 
■nd difficult usee. 

b. A (In niott eaaes, probably) pirtitlve ganlllve, as a Um oomplete 
or lesi abeoluts object than an accnsatlva: thns, with verbs meaning ynr(iiA» 
(eat, drink, et:.), ai p£ba ButAsya (AV.) drink (of) the soma; midhvalj 
pfiyaya (RV.) cause to drink the sieeei draught; — vit)i verbs meinlng 
impart (of Ihe thing Imparted) etr., as d4d&ta no am^tasya (RV.) bestoic 
upon us immortality; — with votbJ meinlng enjoy, be satisfied ot JlUed 


ses— ] 17. Dbclbnbion. 100 

teith: *s, mAtay ftndhosa^ (RV) do fAou M;oy Iht jmee\ tjjaMyfi 
pQn^anU (8,) ftey _^W icitt ftuttw; — with Tsrin meaning perceive, note, 
care for, regard with tMllog of varioni kinda: *i, vAsi^thasya BtnratA 
fndro a^TOt (RT.) Zndm Htlened to FaeieU^a tohc tDiu praiting Atm; 
y&tha mfcma sm&rKt (AV.) <ft«< Ae m«)r think of mr\ taaja onlcopa 
(Mfifa.) he wot tmgry at him. 

O. A genitlre of more doabttul chuicter, with varba meiBlng ruh or 
have authoriig: u, tv&m ifi^ v&eOn&m (RV.) Ihcu art lord of good 
thingt; jiXbA tl&in ef&ih vtr^ftni (AVO that I may rule over &em; 
katlLMh mrtyu^ prabluivati veda^iatraTidSm (H.) Aow Aw dtath 
power over thote who krtote the Veda* and Ireatieeet 

d. A genitlie, IniUtil ol in tblatlTe, li gomeUmet toand oted with a 
verb of teeeltlag of any kiud (beailDg inclnded), and with one of feadng; 
ihuf, yo rajOa^ pratigflu^ti lubdhaaya (H.) whoever aceepU a gift 
from a greedy king; QI^u me [MBb.] learn from me; blbhlmaa tava 
(MBhJ we are afraid of thee. 

B98. A genitive in ita luual pouenlre senie ii often found as pradl- 
eata, and not Beldom wiih the copula amittod: tbna, y&thi 'ao iii4iaa 
k^ali^ CAV.) that thou may*»t be wholly mine; B»rvttfy saihpattlvaa 
tMfya sadltnftuh jaaja mdnesain (H.) all good forfunee are hie who 
hat a contented mind; — n objentt'e predicate, bbsrto^ patnub TiJ&- 
nanti (M.) they reeogniee a ion at the htttbanda. 

[^a8(&.' a. The prepositional conitrnctloiiLOf thejenltWe (1130) are for 
the moat put witb inch pcefoeitions as are realljr noun-casea aod hare the 
gOTeinment of iQcb: thai, agre, arthe, kfte, and tite like; also with 
othei prepositional woids which, in the geneia) looieneat of use of iha 
genitive, havs become a»inillated to theae. A few moie real piepoaitiona 
lake the genitive: either DBnally, like up&ri ahoiie, or o<-eaalonally, like 
adh4a, ant&r, &U. 

b. A genitlTe la occasionally Died in the older langnage with an 
adverb, either of place or of time: thoa, y&tra bvk oa knroklfetrisrft 
(gB.) in whatever part of Karakthelra; y&tra tu btlAmer JityeU (MS.) 
on what tpot of earth he may be bom; iditiulii Uma^ (KV.) at Utit 
time of the day; yimyi ritryft^ pTSUt^ (HS.) on tho mom of what 
night; dvil} smiiTataftntBya (K.) twice a year. Suoli oxpreaaion i« the 
Uit oeonr also later. 

300. a. The genitive It very little nied adveibiilly; a few genitlveg 
of time ocoDt In the older language: aa, aktos by night, vaatoB by day; 
and there are found later inch caeei aa kasya dt kUuya (9) after a 
e^tain time; tatah kftlasya mahata^ pray^ilu (R.) then after a long 
time he went forth. 

b. A genitive, origiaally of poiseiaion, piaging over into one of general 
concernment, cornea in the later langnage (the conittnction it unknown 
earlier] to be med absDlDtely, with an agreeing participle, or quite rarely 


101 Ubw of the Looative. [ — 802 

■a mdjMllve. Form luch cue* u the loUowiog — ps^yftto bakamurkhii- 
■ya nakol&lr btaakfltftl; Bnta^ (H.) of t/te fooiiah heron, while he 
looked on, the young tcert tattn by the ichneumoas, or gato *rdlt(krfttrk^ 
katbSI^ k&thayato mama (K8S.) hoi/ my night mu patted in feUing 
ttoriet, or kftrtavyMya karmo^ah k^pram ahrlyamilpaaya kUal^ 
pibfttl tadrssam (H.] of a work needing to be done but left undone time 
guiekly drink* up iti eatenee ~ come into caneDcy, by InuotBtng Indepeo- 
duiM of the geaitlTB, tnch other cues 4}; diTaib Jag&ma mntllll&tb 
paqyataih tad ft (R. ] he laerd then to heaven, the aicetici looking on ; OTadi 
iJUapataB taaya devadatas t&dft lihyetya vAkyam Uia [UBh.) m he 
thia lamented, a divine meeeenger coming addreaaed him ; iti v&dlna evft 
'l^a dhennr ftvaTrte vanftt (R>eh.) while he that apoke, the cow eamefrom 
the forttt. The genitive ilwayi lDili»t«s ■ living actoi, and the putlclple U 
tifiiall; one of seetag or hearing or nttorlng, espeoUlly the farmer. The con- 
■trnctianUsaldbj' the HindugiimmtrUns to MQiey in Implication of disiegHd 
or detptte; and inch U often to be recogniied in it, thaii(h not preralUngly. 
'801. Uaes of the Locative, a. The locative Is pToperly the 
in-t»M, the caM eipreesing aitaation or location; but ite aphare of 
use haa been somewhat eitended, so as to touch and overlap the 
boondaiiea of other cases, for which it seema to be a Bnbadtote. 

b. Unimportant variatioDS of the sense of m are tboae of amid 
or among, on, and of. Of courae, alao, altnatioa in time aa well as 
place ia indicated by the case; and it is applied to yet lees physical 
relatioDB, to sphere of action and feeling and knowledge, to atate of 
thioga, to accompanying circumatance; and out of this last grows the 
freqnent age of the locative aa the case abaolate. 

o. Moreover, by a pregtiant eonatruction, tbe looative U used 
to denote the place of rest or cessation of action or motion [into or 
on to inataadof moron; Qerman in with aecnsative instead of dative: 
comparo EngHsb there tt>T thither], 

SOS. a. The lootttie of iltnattoD in gptee hardly needs lltnstration. 
An example or two are: y* devi divi athi (A^O vihidi of you goda 
■ are in heaiien\ aa deve^a na yak^eqn t&dfk (HBb.) not among goda 
or Yakehae it tuch a one; it&rvatasya pntb^ (R^O o" '^ "^« of the 
mountain; TldAthe aanta devib (RV.) may the goda he at the asaembly; 
daqaina pade [MBh.') at the tenth aiep. 

b. Tbe locative of time Indicates the point of time at nhlch anything 
takes plaea: thaa, asy^ ti;4bo vyAf^ftu (RV.) at the shining forth of 
thia down; staatnlnn eva kile (HBh.) at juit thai time; dvlda^e -var^e 
(MBh.) t'n the twelfth year. That tbe accusative la occaatonally nsed In 
thU sense, instead of the loeatiTe, was pointed ont above (876 o). 

o. The person with whom, instesd of the place at which, one la or 
remains is pnt in the locative: thn«, ti^hanty aamln paf&ra^ (HS.) 
animaia abide with him; gaiin vaaan (U.) living at a teaeher'i; and, 
piegnantly, tArat trayl bhavifySmi (HBh.) ao long will X eleane to thee. 

ioy Google 

—J IV. Declension. 102 

303. The Iocati*« of epbera or coadltlon i 
freqnent Die : tlii», m&do kbim {ndro JagbSnn (RV.) in fary Indra tUuj 
the dragon; mltr&sya lumatau B7I11UI (RV.) may ure be in the favor 
of Milra; te vaoane rfttam (MBh.) dtlighled in (Ay words. 

a. Thli conBtiuRtian Is, on the one hind, generailied Into in eipiet- 
Flon for wi the matter or cate of, or with referenee to, rtapeetmg, and 
takes Id the later lau^age a totj wide lange, teaching upon i^enltlTe and 
datUe constmctEonB : tbna, b 'm&ih bh^ia grwue ^qre? a g6f u (AV.) be 
generout (o hitn in retainert, in kortes, in cotVt; tijii {t sakhltri imahe 
fRV.) Aim we beg for friendthtp; npftyo "yftifa mayft &Ji^ ftnftyane 
tBV& (MBh.) thi* meotti icae dioised by mt for (with reference to) bringing 
thee hitlier ; B&tltve kSra^iuh Btriyft^ (H.) the eaute of {in the eaee of) 
a iroman'i chaetity; n& qabto 'bbavaa nlvBrai^e (HBh.) he leae not 
capable of preventing. 

( b^^Q the other hand, ^e eipresifon by the loratlTO (f a (ondmon ef 
thing! In whfrh anjtlilng takes place, or ot a conditioning or accompan)ing 
elicoDutuice, pa»ea over Into a well-mtiked absolute conitmctton, which is 
knoirn sven In the eulleat lUge of tbs langnage, hot beeomee more fteqneiit 
later. Traniltlonal eiamples are: hkve tvi Bdnt udite h&ve mti' 
dby&ibdills div&^ (RV.) I eaU to thee altha ariem tun (tohm the tun 
has rieen), I tail at midtime of the day; aparUhe k^tfl 'pi 0& na me 
kopn^ (MBb.) and even in ea»t of an offence committed, there ie no 
anger on my part. 

O. The normal condition of tie absolute conitinetlon i» with a parti- 
clple accom pausing the noun: thni, stinji barhl^l Bamldbftnd ngnftd 
(KV.) ioA«n the barbls it itreum and the fire kindled; kile f Qbha prSpt» 
(MBb.) apropitiom time having arrived; avaatuitiSyBih rfttr&T aat&aalik' 
c&^TBlamblnl oandramasl (H.) the night having drawn to a dose, 
and the moon r.iting on the nanmit of <A< western mountain. 

d. But Ibe noau m«y be wanting, oi may be rrplaced by an adrerbial 
substltnte (aa evun. tRthS, itl): thus, varqati ifAcn it raint; [sOrye] 
aatuuita after mnset; ftdityosya dffyamfUie (3.) whih there is tern 
[tome part] of the aun; Ity ardhokte (<}■) with that words half uttered; 
umabbU? BamannJ&Ue (MBh.) it being fully attenUd to by at; waxa 
ukte kalinK (MBh.) it being thus spoktn by Kaii; tathA *naftl>lt^ (^0 
it being thus aecomplithed. So likewise the participle may be wanting (a 
ropnla aati or the like having to be snpplled] : tbns, dure bkaye the cause 
of fetir being remote; while, on the other hand, the paitlcipla aati etr. ii 
somntimee lednndantly added to the other participle: thus, tath& kfte aati 
it being thus done. 

e. The loeatlieU beqnent'y Died adTerbiilly or prepoaitionally (1116): 
thna, ■artke ot •kfte in the matter of, for the sake of; agre t» ft-ont 
of; rte without; aamlpa near. 

304. The pregnant constructloD by wbloh the locative comes to ex- 
press the goal or object of motion or action or feeling exerrited Is not 

ioy Google 

1 03 Uses op the Locative. [—307 

uncommaD tram the etrliest time. It ii b; no laetnB to be shupl; dlttin- 
gnlihsd from the oidinuT coMtruotioiii the two pus Into one ■nother, trith 
a donbtFnl t«rtll«ry between. It cic«un: 

ft. Eipcclalty vith letba, u of trrlvlng, sending, pltctng, commnnt- 
''4iiug, bestowing, and many othen, in litoationi where in icaasntiTe or 
a datire (or a gauitiTe, S87 b) might be looked for, and extbangeable with 
ihem; thai, ^4 {d dev^Q gaochatl (RV.) Ihat, truly, goet to (to be among) 
the godt; im&ifa no yqlfi&m ajoftefa dhehl [ItV.] »et thi» offering of 
ourt among the immorlab; yi ialfioAntl t&aam dfiultama (^.V.) who 
pour in the juice into the plant* (or, the juiet ihat i* »i the plant*) ; mi 
pi^aoohe "^are dhanam (B.) do not offer wealth to a lord; papftts 
medinySm (HBh.] he feB to (eo a* to he upon) the earth; Bkandhe 
kftTft (a.)pwUing on the shoulder; Badifratya pOrvam aemian (HBh.) 
hoping before promited ui. 

b. Often alio with noniiB and adJectlTe* la iimilai cooalrDctionB (the 
iiutanoei not ilvayi eair to lepMate from tbote of the locative meaning 
with refer«nee to: above, 808 a)-. tbUE, dayA aRrvabhatftfti compataion 
toward all creatures^ anttrac"^ nU^adhe (MBh.) affection for the 
yithadhan; raj& samyae T^tta^ aadS trayi (HBh.) the king haa ataai/* 
befiaved properly iouiard thee. 

306. The piepoiitloiu conatmed with the lowtice (IISS) sUnd to it 
only in the relation of adrerblal elementa strengthening and directing iU 

800. Declensional forms aie made by the addition o£ 
endings to the stem, oi base of inflectioD. 

a. The stem itself, however, in many woids and classes 
of woids, ie liable to variation, especially assuming a stiongei 
form in some oases and a weakei in others. 

b. And between stem and ending aie aometimes inserted 
connecting elements (oi what, in the recorded condition of 
the language, have Uie aspect of being such). 

O, Respecting all these polota, the detaila of treatment, as exhibited 
by ewh claEB of words or by single words, will be given In tbe fallowing 
chaptera. Here, howaTer, It Is desirable also to present a brief general view 
<f them. 

807. Endings: Singular, a. In the nominative, the nana! 
masc. and fem. ending is ■ — which, however, U vantiDK in derivative 
ft and i-Btems; It is also eupfaoQicsllr lost (150) by coDSODant-Btems. 
Henters in general have no ending, but show in this case the tuie 
■tern; a-stemB alone add m (as in the accna. masc.]. Among tbe 
pTODODna, am is a freqnent masc. and fem, nom. ending (and is found 
even In do. and pi.); and nenterg show a form In d. 

ioy Google 

807—] IV. Dbclbbsion. 104 

b. Id the acouBative, m or Bim Is the masc. and fern, ending 
— am being added after a conBonant and r, and after i and S in the 
radical diviaioD, and m elsewhere after vowels. The neuter acoaBative 
is like the nomiiiative. 

o. The instrumental ending for all geuders alike Is &. With 
final 1- and u-vowels, the & is vsrionsly combined, and in the older 
langnage It ia sometimea lost hy contraction with them. Stems in a 
make the case end in ens [aometimeH enX in V.), and those in S make 
it end in oyS; but instaDces occnr, in the early language, of immediata 
addition of & to both a and B. 

d. The dative ending ia in general e^ and with It llkewiae the 
modes of combination of 1 and u final are various (and disappearance 
by contraction not unknown in the oldest langnage). The a-steros 
are quite irregular in this case, making it end in Bya — excepted is 
the pronominal element -ema, which combines (apparently) with e to 
-Bm&i. In the personal pronouns is found bli;ain (or bjam). 

e. A fuller ending &1 (like gen.-abl. fin and loc. &m: see below; 
belongs to feminine stems only. It is taken (with ioterpoaed ;) by 
the great class of those in derivative ft; also by those in derivative I, 
and {aa reckoned in the later language] in derivative n. And later 
it Is allowed to be taken by feminine stems in radical i and n, and 
even by those in 1 and u: these last have it in the earliest language 
in only exceptional iDstances. For the substitution of fti for sbl.-gen. 
fta, see below, h. 

f. The ablative lias a special ending, d (ort), only In a-stems, 
maac. and neut, tbe a being lengthened before it (except in the per- 
sonal pronouns of Ut and 2d person, which have the same ending 
at in the pL, and even, in the old language, in the dual). Everywhere 
else, the ablative is identical with the genitive. 

K* The genitive of a-stems laud of one pronominal u-stem, 
amu) adds sya. Elsewhere, the usnal sbl.-gen, ending is as; but its 
irregularities of treatmeot in combination with a atem-final are con- 
siderable. With 1 and u, it is either directly added (only in the old 
language), added with interposed n, or fused to es and os respect- 
ively. With T (or »r) it yields ur [or us: 169 b). 

b. The fuller fta is taken by feminine etems precisely as fti is 
taken in the dative: see above. But in the langnage of the Brah- 
manas and Sutraa, tlie dative-ending ftt is regularly and commonly used 
instead of fte, both of ablative and of genitive. See 366 d. 

i. The locative ending is 1 in consonant- and p and a-«tems 
(faeing with a to e in the latter). The i- and u-stems (nnlem the 
final vowel is saved by an ioterpoeed n) make the case end in ftu; 
but the Veda has some relics or traces of the older forms (ay-i 1?) 
and av-1) out of which this appears to have sprung. Vedic locatives 

ioy Google 

105 CiSB-EHDlNaS. [—809 

ftom i-it«mB end bIbo in & Bnd L The pFonominal element -bum 
makaa the locfttive -smln. Stems In an in the older lingiuge often 
kwe the 1, and use the bare item as locative- 

j. The ending ftm is the locaUre oorrespoudent to dat. ftl and 
abl.-gen. Ke, and is taken under the same circum stances: see above. 

k. The vocatire (uoleas by accent: 314) is distinguished from 
the Domlnative only in the singular, and not quite alwsja there. In 
a-Btems, it la the unaltered stem, and so also in most consonant-stems; 
bnt nentera in kd and in maj drop the a; and the oldest language 
hae sometimea a vocative in a from etems in nt and fie. Stems in f 
change this to ns. In masc. and fem. 1- and a-stems, the case ends 
respectively in e and o; In neuters, in the same or in 1 and a. Stems 
in ft change K to e; derivative i and Q are shortened; radical stems 
In long vowela use the nominative form. 

308. Dual. a. The dual has — except ao far as the vocative 
is sometimes distinguished from nominative and accusative by a dif- 
ference of accent: 314 — only three case-forms: one for Dom., accus., 
and Toc.; one for instr., dat., and abl.; and one for gen. and loc. 

b. But the pronouns of 1st and 2d person in the older language 
distingnish five dnal cases: see 492b. 

o. The masc. and fem. ending for nom.-accus.-voc. is in the 
later language usually ftu; but instead of this the Veda has pre- 
y^lingly &. Stems In S make the case end in e. Stems In i and u, 
maee. and fem., lengthen those vowels; and derivative I in the Veda 
remains regularly unchanged, though later it adds ftu. The neuter 
ending is only 1; with final a this combines to e. 

d. The universal ending for the Instr.-dat.-abl. is bby&m, 
before which final a Is made long. In the Veda, it is often to be 
read as tvo syllables, Uiifim. 

e. The universal ending of gen.-loc. is os; before this, a and 
ft alike become e (al). 

300. Plural, a. In the nominative, the general masculine 
and feminine ending is as. The old language, however, often makes 
{he case in ftau instead of Se from a-stems, and in a. few examples 
also from ft-stens. From derivative i-stems, ia instead of yaa is the 
regular and nsnal Vedio form. Pronominal a-stams make the masc. 
nom. In e. 

b. The neuter ending (which is accusative also) is in general i; 
and before thia 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 fonns in ftnl. Ini, Qnl are frequentiy 
abbreviated by loss of the ni, and sometimes by further shortening 
of the preceding vowel. 

ioy Google 

800—] IV. Dbclbhsion. 106 

0. The accuaatlTe endiog is also aa in oonaoiiaii totems and 
in the radical division of I- and Q-stems (and in tlio old language 
even elsewhere). Stems in short towsIb lengthen those vowela and 
add in the muonline n (for ns, of which abundant traces remain), 
and in the feminine b- In the neuter, this cue is like the nominative. 

d. In the instiumental, the ease-ending is everywhere bhia 
except In a-stems, where in the later language the case always ecda 
in Sis, but in the earlier either In aia or the more regular ebhia 
(Sbhia in the two personal pronouns; and the pronominal stem a [EOl] 
makes ebUs only). 

•■ The dative and ablative have in the plural the same form, 
with the ending btayas (in Veda often bhlaa), before which only a 
is altered, becoming e. But the two personal pronouns distingaish 
the two cases, having for the ablaldve the singular ending (as above 
pointed out), and for the dative the peculiar bbyam (almost never in 
Veda bhlam), which they ezteod also into the singular. 

f. Of the genitive, the universal ending is Sm; which (except 
optionally after radical I and tt, and Id a few scattering Tedio in- 
stances) takes after final vowels an Inserted consonant, a in the pro- 
nominal declension, n elsewhere; before n, a short vowel is length- 
ened; before a, a becomes e. In the Veda, it is frequently to be 
proDODQoed In two syllables, as a-am 

g. The locative ending is so, without any exceptions, and the 
only change before it is that of a to e. 

b. The vocative, as in the dual, differs from the nominative 
only by its accent 

310. The normal scheme of endings, as lecogniied by 
the native giammarianB (and oonveniently to be assumed as 
the baiBiB of apecial descriptione), is this: 

SinKtkr. Dual. PlmL 

m. t. n. m. r. IL m. f, lu 


a. It Is taken in bulk by the consonantal stems and by the rad- 
ical division of i- and Q-stems; by other vowel-stems, wiUi more or 
less considerable variations and modificattons. The endings which 
have almost or quite unbroken range, through stems of all olaasea, 
are bhy&m and OS of the dual, and bhie, bbyas, Im, and an of the 

ioy Google 

107 Stronq akd Weak Stku. [—3)8 

311. Variation of Stem. a. By fai the most im- 
poitant matter undei this head ia ihe distiaction made in 
la^e claaaes of woids [chiefly those ending in consonants) 
between strong and weak stem-forms — a distinction stand- 
ing in evident connection with the phenomena o£ accent. 
In the nom. and accus. siag. and du. and the nom. pi. 
|the fire cases whose endings are never accented: 316 ft), 
the stem often has a stronger or fuller form than in the 
rest : thus, for example (424), JlsJT^R raj&n-am, ^sTHt rSjan- 
511, (IslHH rtjan-aa, against ^fTslT rajfl-5 and ^fTsjftTT rSja- 
bhia; or (460 b) tl^TrPT mahint-am and (447) g^rPT adfiat- 
am against R^RT mahat-S and Q^ adat-s. 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 cases, and cases of 
middle stem, oi middle cases: the former having endings 
beginning with a vowel (instr., dat., abl.-gen., and loc. sing.; 
gen.-Ioo. du.; ace. and gen. pi.); the latter, with a consonant 
(instr .-dat. -abl. du.; instr., dat.-abl., and loc. pi.). 

b. The class of strong cases, as above defined, belongs 
only to masculine and feminine stems. In neuter inflection, 
the only strong cases are the nom. -ace. pi.; while, in those 
stems that make a distinction of weakest and middle form, 
the nom.-acc. du. belongs to the weakest class, and the nom.- 
acc. sing, to the middle: thus, for example, compare (408) 
ilrOf^ pratydfic-i, nom. -ace. pi. neut., and UriiyH praty- 
dfio-as, nom. pi. maso. ; 51fft^ pratlc-i, nom.-acc. du. neut., 
and HhWIH pratIo-6fl, gen.-loc. du.; ^rll^^ pratydk, nom.- 
acc. sing, neut., and HrUPw pratyig-bhis, instr. pi, 

318. Other Tailitions concern chiefly the Ad*I VQirel of a stem, and 
my be mainly left to be pointed out in detail below. 0( coniequenco 

Diaii.zec.y Google 

312—] IV. Declension. ]08 

enough to mentloD heia la ouly the gu^pa-sbeiigtheDlag of > tiaal 1 ot n, 
vhleh in tba Utet Ungnige ta alwtra made befoie aa of nom. pi. uid a 
of dat. alng. Id inasc. and fern. ; In the Veda, It doea not alvaya take place ; 
nor Is it foiMdden in dat. aing. neul. ilm; and it is leen BometimM In 
loc. ling. Final f has gai^ft-strengthenlng In loc. aing. 

813. Insertions between Btem and Ending. Arter TO«el-items, 
an added n often makes Its appeacanoe before an ending. The appendage 
la of least qneatioiiable origin In nom.-acc. pi. neat, wbeie tbe Intercbange 
in the old langnage of the fotnu of ft- and l-stema viih those of fta- and 
In-steou is pretty complete^ and the a-stenu follow their analogy. Else- 
where, It ia most widely and flnnly eatabllsbed In the gen. pi., wheie in 
the great maas of cases, and tiom the earliest period, the ending is vlrta- 
ally nfim after a tdwoI. In the t- and a-stams of the later language, tiie 
iostr. alng. of maM. and nent. Is aeparat«d by it« presence from the fem., 
and it Is In the other weakest cases made a usual distinction of neuter forms 
from maacnllnei but tbe aipect of the matter in the Teda Is very different: 
there the sppeaiance of tlie n is everywhere sporadic^ the neuter shows no 
special inclination to take it, and it la not eiclnded even from the femi- 
nine. In the ending ena from ft-stems (later Invariable, earllei predomi- 
nating) its presence appears to have worked the most oonalderable trans- 
formation of original ehape. 

a. The place of n before gen. pi. ftm is taken by s In pronominal 
B.' Slid S-stems. 

b. Tbe y aflei tt before the endings ftl, Ss, and &di U moat probably 
ail insertion, such a; Is made elsewhere (SS8). 

Accent in Declension. 

314. ft. Ad » rule without exception, the vocative, if accented 
ikt all, ie nccented on the first Billable. 

b. And in the Veda (the case is a rare one], whenever a syllable written 
as one 1* to be pronounced as two by restoration of a semivowel to vowel 
form, the first element only has the vocative accent, and tbe syllable as 
written is clronmBei (88-4) : ihns, d^Sita (I. e. diftua) when dlisyllabie, 
but dy&uB when monosyllabic; Jy^e when for jt&ke. 

o. But the vocative is aocented only vb«n it etands at the be- 
gionlng of a lentence — or, in verse, at the beginning aUo of a 
metrical dirieion or p&da; elsewheie it is accentless or enclitic: tbua, 
igne y&di y^]&&di parlbhAr Aal (RV.) O Agtii! uihalmer offering 
tkou proUeUtt; but &p« tvft 'gnft ^ 'mosl (RV.) unto OtM, Agni, u>« 

d. A word, or more than one word, qualifying a vocativs — nauaily 
an adjective oi appositlve nonn, but sometimes a dependent nonn in the 
genitive (very rarely In any other case) — coDstltntea, to Car as aocent is 

L,j,l,;ec.y Google 

109 Accent. [—316 

conceiDsd, * DDity with th« tomUv; : thna (ill the exuuples horn BV.), 
it the beginning of a p&da, with first syllsble of the comblnitloD aotented, 
{ndra br&ta^ O brother Indra! r^Jan soma O king Soma! ykvi^\fb& 
data mo»t yonAfui mnimger.' hdtar yaviftba Bukrato nu>»l ymt^ful 
skilUd offerer! 6rJo nap&t aahaeran mig/iiy ton of atrengUi ! — ia the 
inteiioi of & pftda, without acceut, BomOaa ladra girva^a^ the tomtu, 
O long-loving Indra! tSv a^rlnS bhadratuwtS aup&Ql ye, O A^int 
of propiiiout and beautiful hande! & r&Jiluft maha ftaeya K^P^ hither, 
ye two kingly guardian* 0/ greet order! 

e. On the other hind, two oi more independoDt ot ooSrdlnite loostlTea 
>t the beginning of > pSda ue regulitlf sod UBnilly both aooented ; thua, 
p{tar mata^ O father! O mother! 4gna fndra T&ro^a mitra d^a^i 
Agni! Indra! Varuna! Mitra! gode! q&taniute q&takrsto thou of 
a hundred aide! of a hundr/d arte! T&Bi;tha 9£ilcTa didiva^ pttraka 
belt, bright, thining, eleaneing one! ttrjo na]^Kd UiAdya^oott ion of 
elrength, proptiioueh/ bright one! But the texta oftei occmIodiI IrrtgnUt 
eicoptloDi both to this and to tbe proceding mle, 

f. FoT brevity, the voMtlve dnil ud plural will be glTen in the pai- 
edlgma below tloDg irlth the noninattTe, wltboot taking the tronble to 
apedfy in each Inatsnce that, If the latter be aecented elacwhare than on 
(he Br*t gyllable, the aceeut of the voMtlve U different. 

316. As regards the otber 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 pennlt, or any other syllable farther back — as 
is B&rpaQt, vari, bh&gavant, eom&naa, sah&srBTflJa — (he accent 
remains npon that syllable through the whole inflection (except in the 
vocative, as explained in the preceding paragraph). 

a. The only eiceptloni are a few nnmeral sterna: aee 488. 

316. Stents accented on the final (including monoByllables) 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 themselves to take tbe accent. Thus: 

a. The endinga of the nomlnatlTe and aecuaatiie singular and daal 
and of the nomfnatlTe piuial (that ia to Bay, of the strong cases: 811) have 
no tendency to take the accent away from the stem, and aie therefore only 
accented when a flnal vowel of the stem and the vowel of the ending are 
blended together into a single vowel or diphthong. Tbns, from dattA oouie 
dattaii (=datt4 + au] anddattas (=datti + aa); bnt from nadi come 
nadySii CE=nadI+au) and nadyaa (smadf+as)- 

b. All the other endings sometimes take the accent; bnt those begiuning 
with • vowel (i. e. of the weakest cases: 811) do so more readily than 
those beginning with a consonant (I. c. of the middle cases: 311). Thar, 
from 11&68 come n&va and nHubhls; fiom mahAnt, however, come 
mahatlE but mah&dbhls. 

itizecy Google 

317—] IV. Declension. 110 

The general rules of accent, then, maj be thus stated: 

317, In the deolenBion of monosyllabic stems, the accent falls 
upon the ending in all the weak cases (without distinction of middle 
and weakest)- thus, nftv^ nfinbbyam, aftv^, nSufu; T&d, Tfigbhla, 
vftcam, vAkfu. 

a. Rut eoma monosyllible itemi retain the accent thiODghont: Cbns, 
26bhlB, g&Tim, g6fu. For tncb casfg, tee below, 860, 361 o, d. 37fi, 
380, 4S7. And in tlie a^c. pi. the stem Is evm otiensr tccented than 
the ending, «omc woidt also admitting either accentuation. 

Sie. Of polysj'llabic stems ending in consonants, only a few shift 
the accent to the ending, and that in the weakest (not the middle) 
cases. Such are: 

a. Preient paitlclplei In &nt oi it: thug, from tudint, tndati ani 
tudattbs and tudatam; bat tud&dbhj&m and tuditsn. 

b. A few adjeetivea faiving tba (orni et sneh paiticiplea, ai maliat^ 

e. 9t!ina o( which the accented flnal loaei Ita Erllabic chua~ter bf 
tyiicopationorth?Tonel: thue, malJ&it, mQrdlmi, d&mn&a (from nuJJ&n 
et^; 423). 

d. Other aporadlc cases will be noticed nnder the different declensions. 

e. Case-forms used adverbially sometimes ehow a changed accent: 
Gee lUOff. 

319. Of polysyllabio stems etidlng in accented short vowels 
the final of the stem retains the accent if it retains its syllabic 
identity: thus, dattina and datt jya from datt&; agnina and agn&ye 
from agni; and also dattibhyaa, Bgnlbhls, and so on. Otherwise, 
the accent is on the ending: and that, whether the final and the end- 
ing are combined into one, aa in dattfifs, dhenSA, ogniiii, dbends, 
and so on: or whether the fieal is changed into a semivowel before 
the ending: thus, dhenvlt, jfUxt, jftmyos, bBhTds, etc. 

a. But Sm of the gen. pi. from atema In i and 6 anl f may, and In 
the older language alnays doei, take the accent, though e.'parated by n from 
the stem : thus, acQiitJm, dlieailaitm, pi^am. in RV., even derivatlTe 
i-3tcmB Bho'* usually the Fame shift: tbua, bahvln^. Of stems In it, 
only numerals (483 a] follow this rule: thus, saptftnim, dafSnim. 

320. Boot-fords In 1 and il as flnal memberg oT comp'>unds retain the 
ac^nt througboat, not shitting it to any of the endings. And in the older 
language there are polysyllabic words In long flnal vowels which follow In 
this respect aa in others the analogy of the root-dec lenslon (below, 3B5ir.). 
Apart from th^'se, the treatmrn'. ot sl-'ms in derivitive long vowels is, a< 
regards ai-eent, the same ai of those in short vowels — save that the Ion ■ 
is not thrown forward npon tbe ending in gen. plural. 

,1,1.0, Google 



321. &. Tm£ aooordance ia iaflMtion of substantire 
and adjective stems is so complete that the tiro oaanot be 
separated in tieatmeot from one another. 

b. They may be classified, foi convenieaoe of descrip- 
tion, as follows: 

I. Stems in ^ a; 
n. Stems in ^ i and 7 u; 

III. Stems in )9T ft, ^ I, and tJt 11: namely, A. ladical- 
stems(and a few otheis inflected like them); B. denvative stems ; 

IV. Stems in n f (or 51^ ar); 
V; Stems in consonants. 

o. There le nothing ibsolate tn thii clasilBoatian and unngement; 
It U meiely belieTed t« be open to w few objections ta any other. Mo 
general igitement has been reached among Bcbolars as to the number and 
order of Sanskrit declensions. The stems in a are here treated first became 
of the great predominance ot the clui. 

88S. Tbe dlvisioD-liDe between aubstftiitire and adjective, always 
an uncertain one in early Indo-EuropeaD language, la even more 
wavering in Sanskrit than elsewhere There are, however, in all the 
4eclensionB as divided above — unless we except the stems in r oj 
ar — words which are distinctly adjectives; and, in general, they 
are inflected precisely like nonn-Btems of the same final: only, among 
con Bon ant-stems, there are certain sub-claBseB of adjective stems with 
pecnliarities of inflection to which there IB among nouns nothing cor- 
responding. Bnt there are alBo two conBiderable classeB of adjectlve- 
couponndB, requiring special notice: namely — 

3S3. Compound adjectives having as final member a bare verbal 
root, with tbe value or a present participle (383 a ff): thus, sii-dfq weU- 
hoking; pra-badli forekmmoing; »-drdh not hating; veda*vid Veda- 
kaowing; v^tTa-hiui Vitra-slayitig; upastha-s&d titling in tht lap. 
Every root Is liabla to be used in this way, and such compounds are 
not infrequent in all ages of the language: aee chapter on Compoands, 
below (laeS), 

ioy Google 

SaS—] . v. NOCN8 AND ADJECnVEe. 1 J2 

a> Tbli clut U eBientfally only a special cUbb of componnd adjecttvas, 
since In the earlleit Veda tbe timple as well ae the aompoonded root wu 
gometiiiies uBed adJecHvely. But the oompeuDded root vm from the begiDnlng 
much more often so aied, and the later the more eiclualTely, eo that 
prsotically the e1a»» ii > aeparate and impottaut ooe. 

324. Compound adjeotives having a noun as final member, but 
obtaining an ftdjective sense seoondarily, by having the idea of 
poiteuion added, and being infiected as adjectives in the three gen- 
ders (IS9Sff.). Thus, ]^T«4&k&in4 detire of progmy, whence the ad- 
jectire pr^Jaklma, meaning denrotu [i. e. having dtiire) 0/ progtny; 
eabh&rya (sa+bli&ryi) having ont'e wif« along; and so on. 

a, la a few cues, t,Uo, tbe flnal nenn is syDtaeHosIly ot^ect of the 
prec^dmg member (1808-10): thas, RtimKtra imtnodarate (atlmBtrun 
bti/ond mtaiurt); T&va;4ddTef&a driving away anemia. 

326. Hence, under each declension, we have to notice how a 
root or a noun-stem of that declension is Inflected when Goal member 
of an adjective compound. 

a. As to accent, it needs only to be remarked here that a root* 
word ending a oomponnd has the accent, but (380) loses tbe peoa- 
liarity of monosyllabic accentuation, and does not throw tbe tone 
forward upon the ending (except aSo in certain old forms: 410). 

Oecleniion I. 
stems (maBculine and neuter) in q s, 
826. a. This declension contains the majority of all the 
declined stems of the language. 

b. Its endings deviate moie widely than any others 
from the normal. 

887. Endings: Slngntar. a. The nom. msac. baa the normal 
ending a. 

b. The aco. (masc. and neut.) adds m (not am); and this form bae 
the office also of nom. neuter. 

o. The Instr. change* a to ena nniformly in the later Ungual ^ and 
eTBQ In the oldest Tedic this ia the predominant endiug (in RV., eight 
niutba of all casea). Its flnal la in Vedlc leree frequently made long (anft). 
But the normal ending ft — thus, ;ajfL£, DUbivft, mahitvit (foiy^JlLAna 
etr.) — is alao not rue in the Veda. 

d. The dat. has Bya (ae if by adding ays to a), alike in ail age» 
of the language. 

e. The abl. has t (or doubtless d: It is Imposaible from tbe evi- 
dence of tbe Sanakrit to tell oblch It the original form of the ending), 


113 DBOLBHBION I., O-BTEUS. [—819 

bafore vhlob & !■ made loDf: thli ondiDf 1* found fa do oQiot noaii- 
declenston, and elMwhtn only in tbe penooal proaoDns (of til nDmben). 

f. Th« gen. hu mya added to the flnal a; and tlii* endiDg U alto 
limited to K-il«int (with tlie tlngla aioeptlon ot the pranonn amAffft: 
SOI). lU flnal a li In onlj' tbtee oaiw made loag In the Teda; and Iti 
y li TOoallEed (aola) almoit ai laiely. 

g. The iDB. endt In a (*« It by combining the normal •nding 1 with 
tiie final of the item), wlthont eioepllon, 

h. The Toe. la the baie atem. 

828. Dual. a. The dual endlnga in general aie the normal one*. 

b> The Dom., aec, and tdc. maac. end in the later langnage «lira}i In 
fta. In the Teda, hoTBTer, Ae uaael ending la aimple S (in RV., In 
aeren eighth* of the ooonrreneea), The aame caaea in the neut end In a, 
irhioh appeara to be the reanlt of fnaion of the ttem-flnal with tbe normal 
ending L 

0> The initr., dat., and abl. have bhylbii (In only one or two Tedlo 
inttancea reaolved into bhiSm), with the atem-final lengthened to ft befoie It. 

d. The gen. and las. have a y inaerted after the atem-flnal before oa 
(or aa If the a had been changed to •). In one or two (donbtfnl] Tedte 
Inatanoef (as alao In the pronominal forma enoB and yofl), oa ia anbatltnted 
for the final a. 

328. Plaial. a. The nom. maao. hai in the later langnage the 
Doimal ending as combined with the final a to Bs. Bnt In the Teda the 
ending KMS initead ia frequent (one third of the ooonirencei In RT., bat 
only one twenty-flfih In the peculiar paita of AV.). 

b. The ace. maao. enda in Bn (for earlier fins, of which abundant 
traces are left in the Teda, and, under the diagnlae of apparent enphonic 
combination, even In the later language; lee abOTe, 808 B.). 

o. The nom. and «oc nent. have In the later langaage alwaya the 
ending &nl (like the an-atflma; aee 4S1; or elaa with □, aa in the gen. 
pi., before normal 1}, But In the Veda this ending altematea with almple 
& (which tn RT. la to Boi aa three to two, In point of fieqnency; in AT., 
ai three to four}. 

d. The Inatr. end* later alwaya in Sla; but In the Teda la found 
abundantly the mote nonnal fonn abhla (in RV., nearly aa frequent as SIb; 
in AT., only one Bfth aa frequent). 

a. The dat. and abl. have bli;aB aa ending, with e Inatead of the 
final a before it (aa In the Tedle inetr. ebbia, the loo. pi., the gen. loc. 
du. [?], and the Inatr. alng.). The teaoluHon Into abhias U not infrequent 
In the Teda. 

f. The gen. enda lu ftnfim, the final a being lengthened and having 

n inaerted before the normal ending. The S of the ending la not aeldom 

(In leaa than half the inatancea) to be read aa two ayllahlei, aamt oplnlona 

are divided at to whether the reaclntlon It hlttorical ot metrical only. A 

WhltDSf, Orammar. 3. ei. S 

ioy Google 

TOT •null noMbtc (Uf-a-4oHii) of namplM Of timfie Sm u «ading 
liutMd ot laim Munit la BT. 

g. Tha loo. eadi In •fU — tlut ii lo uy, irlth Ui« BCTmtl mdfDg, 
before whiob (lie iten-Onal li Aiaged W • (with OKweqaoDt ihMiga of a 
t« n. ISO). 

h. Of ftccent, In this deolenBion, oothlng raqnim to be Mid; the 
■TlUble Moented in tha Btmn retaini Hi awn aeoant tfarougkont 

880. Examples of declension. Ai ex&mples of the 
infleotion of a-atema may be taken ^iHT kSma m. love; 
^ deT& m. ffod; 51 (TO Bsyi n. mouM. 

M. Einii^ 






A. ^iPFI^ 






1. sm^ 






D. ^mra 






Ab. ^irtn?!^ 






0. WMflT 





L. Jfif^ 






T. ^fiq 







«. A. V. ERT^ 






I.D.Ab. ctiim^lH 





<?■ t. ^iiwt^ 







H.V. ^(UmT 






i, Google 
































EumplM of tin pMuUar V«dlo ftrai v*: I 

& Slag.: Iiutr. nv&theus, yt^Bi (sach genltiT* formi u i^vftolK I 

*Ts fntfAj ap«ndle), I 

b. Dn. : nom. etc. muo. devt; gen.-los. J^aatria (item pastyk). I 

O. PI.: nom.-Too. miM. daTjUasi nest. 7«ai; liwtr. dw^UliS; gm. | 

MrittiCi», davitiwAm. ' 

831. Among oonDB, there ftrd no irregalarities Id thia decleneioa. 
For irregalsi numeral bases in a (or an], see 4S3-4. For the irreg- 
nUrities of prOQomin&l stems in a, which are more or less fully 
sbarod also by a tt\r adjectives of pronominal kindred, aee the chapter 
00 Ptonoana (MSt). 


S3B. Otiginal adjectives in a are an exceedingly large olasa, the 
great majority of all adjectives. There Is, however, no each thing as 
a feminine atom in a; for the feminine, the a Is changed to & — or 
often, tboagh far iMa often, to I; and tta declenaion is then like that 
of aenft or devi (S94). An example of the complete declenaion of an 
a^JMtive »4tem Is tfae three genders will be given below (3W). 

a. Wkether & mwcnent item in a ibill roim iu rsmlnlne in ft or 
in I U > fuestloo to be deterailiied In great put only b; aotokl iiuge,,>nd 
not by gnmmitical idle. Certain Imjtoituit clusM ot vordi, howevsr, etu 
be pointed oat which take the Icti common andtag I tot tbe feminine: Ibni, 
1. tbe (very nnmerom) «econd>ry dertiitlTes In a with v^ddhl ot the Arit 
■ylkble (1304): b. g. &mltr& -trf, nutna^ -;I, ptvamSnil -xd, plor- 
l^amloi -ail % piimarr derlTtllreg In ana with tccent on the radlcil sylUble 
(1160): e. g. oodana -nl, saihgrAfaa^a •ifl, Bnb]i&gBifak4ra9a 'Ul; 
3. primuy deilTitlTM In a, with strengtbeaing of the rtdical tylbble, 
htTlDg » qnail-pirtlciplel meenlng: e. g. dtvikarii -Tl, avalcrSm& -ml. 


Dgil.ze.., Google 


rathRTSh& -hi (but there are nuoy exceptloiu)^ 4. eecoDdo; deiivatlTee 
in mayft (1225) end tana (lS4Se): e. g. ft^asmAya -yl; adyataaa 
-nl; 5. moit oidlnal nnmetela (4B71l}: e. g. pafloam& •mi, navadafi 
-^i, trUqattomA -mi. Not a few words make the femlniae in eithei S 
or I; (. g. k6vala or -U, ugr£ ot -rl, pSpft or -pf, i5m^ oi •mi; bnt 
ordinully only one of theie la accepted aa regntar. 

3S3. There »e no verbal roots ending in a. Bat a ie sometimes 
BubstitDted for the Goal K of a root (and, rarely, for final an), and it 
ie then inflected like an ordinary adjective in a (aee below, 354). 

334. a. A noun ending In a, when occurring as final member of 
an adjective compound, is Inflected like an original adjective in a, 
making its feminine likewise in S or I (367). 

b. For the moat part, an adjectlTe compouad having a noon In a as 
fln&l member makea ita feminine In I. Bnt theie at« numeiom exeeptlon;, 
oeitain nanot tatiog, ueually ot alwayt, i instead. Some of the commonest 
of tbeae are sa (oUowi: ak^ eye (e. g. loblt&kql, dvyakfl, gairftk^i), 
panja leaf (e. g. tilapai^l, aaptapan^l; but ekapar^), mukha face 
(e. g. kr;namaklu, durmukhii but trimukh& etc.), afiga limb, body 
(e. g. anavadyUigi, BarvKfigi; bat oaturaAgB etc.), kega hair (^e. g. 
Btikefl, muktake^I or .qS, etc.), karna ear (e, g. mah3kan;^i; but 
gokar^ft etr.), ndara btUy (e. g. lambodarl), mula root (e. g. pafi- 
oamoli; but oftener qat&mulft etc.). The very great majaiity ot lOch 
nouni (as the examples indicate) signify parts of the body. 

(^ On the other lutud, a feminine noun ending in derivative & 
sbortens its final to a to form a masculine and neuter base: aee 367 o. 

d. In frequent caaep, nouna of conaooant ending are, la Snala of com- 
pounda, transferred to the a-declenaion by an added sufSx a (1206 a) or 
ka ()9S2). 

Declension II. 

stems (of all genders] in ^ i and 7 u. 

333. The stems in ^ i and 7 u are inflected in so close 
accordance with one anothei that they cannot be divided 
into two separate declensions. They are of all the three 
genders, and tolerably numerous — those in ^ 1 mote 
numerous than those in 7 u, especially in the feminine 
[there aie more neuters in 3 u than in ^ ij. 

a. The endings of this declension also difi'er frequently and 
widely from the normal, and the irregularitiea in the older language 
are nnmeroua. 

ioy Google 

]17 DECLBKSIOH II., 1- AND n-STEM8. [ — 836 

330. Endlnca: ainguliT. a. The nom. muc. and fsm. adds to the 
*t«iD the aonnil ending a. The nom. and leo. neat. U the l»re (tern, 
tdthODt ending. In the Tedi, the final a of > fe« neateri 1b lengthened 
(948 b): thai, vai, parA. 

b. The *cc. mue. and fem. adda m to the stem. Vedie fortai In ittm 
and a&m, and, with u, inam and nnam, are eieeMlvely rare, and donbtTal. 

e. The Inttr. fem. Id the Utet Iingnage takes the nonnal ending & 
■imp];, while the maac. and neat Insert n before <t, making infi and vtai. 
But In the Teda, forms in yK and vK (ot IS and nS.) are not ioft^qnent 
In mate, and neat, also; while Inft Is fonad, veiy nrely, m a fem. ending, 
HoHOVer, fem. 7B is often (In two thirds of the ocenrrenoes) oontraeted to 
\; and this is sTen sometlmeB shortened to i. An adverbial Initr. In uyt 
from halr-a-doien stems In n occnrs. 

d. The dst. mase. and fem, gnnitei the final of the stem before the 
ending e, making e^e ancl Kve. The«e are the prevailing endings In the 
T«da likewise; bat the more normal ja and ve (or ne) also occur', and 
the fem. has Id this ease, ai in the instr., sometimes the form I for le. 
Id the later language, the neuter Is cequired In this, at in all the other 
weakest eases, to Insert n before the normal ending: but In the Veda such 
totms are onir sporadic; and the neut. dat. has also the forms Aje, va, 
ftve, like the other genders. 

e. The abl. and gen. masc and (em, have regularly, both earlier and 
liter, the ending a with guuated vowel before it; thns, ee, oa; and In the 
Teda, the neut fDini) the cases in the same way; although unaa, required 
later, is also not infrequent (iaaa does not oeeor). But the normal forms 
yaa (or iaa) and vaa (or tum] ue also freqnent In both maso. and neuter. 
As masc. ending, XUUM ocours twice in RV. Tha anomalons dldy&t (so TS, ; 
In the corresponding passages, vidfAt VS., didyftut K., didiv&a HS.) 
Is of doabtfnl cbarscter. 

t. The toe maic. and fem. has for legnlar ending In the later lan- 
guage &11, replacing both finals, i and a. And this is In the Veda also the 
most freqnent ending; but, beside It, the 1-stems foim (abont halt la often 
in BV.) their loc. In &: thus, agaft; and thli Is fonnd once even in the 
neuter. The RV. has a number of examples of masc. and neat. locatlTes 
in avl (the normal ending and the a gnnated before it) from u-stems; 
and certain doubtful traces e( a corresponding (^ from i-stems. Halr-a- 
doien locatives in I [regarded by the Vedic grammsjians u piagrli;a or 
nncomblnable: 188 d) are made bom i-atems. The later language makes 
the neuter locatives in ini and xad; but the former never occurs In the 
oldest texts, and the litter only very rarely. 

g. The later grammar allows the dat., abl.-gen., and loc fem. to be 
formed at will with the fuller fem. terminations of long-vowel stems, namely 
01, fta (for which, in BAhma^a etc,, Si Is snbstilnted; S07h), &m. Snoh 
forms are quite rare In the oldest language even from l-items (less than 
altogether in RT.; three times as many In AV.); and from 
s they are almost unknown (five in RV. and AT.). 

Digitizecy Google 

8Se— J V. Nouns AMD ADJBOTiTse. 1)8 

b. The TMs. gniutM tk« flnal of the stem, in nwr. and fem., alike 
In the Miller and la th« IMer Uogntfe. In tbe neat., it i« Istai »[lo«ed 
t» be eltbei of the ume foiin oi the uiulteied itea; lud tbia wm fMhaUgr 
the DMge In the older time also; not Cnttaneea enDigb are qaotahls to 
dAtermina the qneatlon. (AV. hu u once, aad VS. o onoe). 

337. Daal. a. The later and eaillei language igtee In making the 
nom.-scd.-TOD. maar, ar.d (em. by lengthening tbe flnd of the stem. The 
■ame cuea In tbe nentei (aeeordlng le tbe nile given above) end latai Id 
Inl and vaiS; bnt tbeee endings ue neail; nnknowD in the Tsda (as, Indeed, 
the oaies Bte of onl; rare oecnnenoe): AT. hai lol twtee (BT. perhaps 
onca)i TS. has unl once; RV, has nl from one n-iten, and I, onee sbort- 
ened to I, ^m one oi two i-steins. 

b. The nnvarylDg ending ot Insti.^dat.-abl., In all genders, 1b bhy&ni 
added to tbe nnobanged Btem. 

tt. The gen.-loo, of all ages add 08 to the stem In masc. and fem.; 
In nent, ttt later Isngiuge Interpoaet, as elsewhere In tbe ireakeet oases, 
a n ; piobahl; In tbe earlier Vedic the form wonld he like that of the other 
gendeu; bat the only occurrence noted Is one onoa In AV. 

3S8. Flaral. a. The nom.-TOC. maso. and fen. adds tbe normal end- 
ing aa to the gnnated stem-llnal, making oytM and avaa. The eieepHona 
in tbe Veda aM veijr few: one word (arl) has las in botb gsndeu, and a 
few feminine* hftve la (like 1-stems); a very few u-atema have uaa. The 
nevt. Dam.4ee. end* later in in! and On! (like tnt from a: 880 e); bnt the 
Veda has I and 1 (abont equally fraqneitt) moch oftener dian mi; and il 
end (moie uanally) a, roore than half as oftea as QnL 

b. The accns. maco. ends la In and fin, for older £aa and ftna, of 
which plain bsoes teaiilB in the Veda tn nearly half the instuee* of oooni- 
rence, and enn not infrequently In tbe later Uagnage, In the gait* of 
phonetie rombinstlon (908 tt.}. Tbe aeons, fern. end> tn I« and b. BhI botb 
mate, and fem. fonnt In las and una are tennd tpsringty in a* Veda. 

O, Tba Insti. of all gender* adds bhia to the slam. 

d. The dat-abl. of all genders add* bhyna (in T., almost Daiar blilaa) 
to the stem. 

0. The gen. ot all gender* la made alike In Inam and Ouftm (of 
which the II Is not seldom, In tbe Veda, to be resolved Into nam). Stem* 
with accented final in tbe later langaage may, and in the euller alway* 
do, throw forward the accent npon tbe ending. 

f. The loc. of aU genders adds an (as ;u: 180) to the atem-llnal. 

g. The accent ia in accordanoe with tbe genera] rules already 
laid iowo, and there are no irregnlaritiei calling for apecial notice. 

369. Examples o£ defilensioQ. As models of i-stem« 
may be taken ^fil agni m. _^re; IT% giti f. gait; ^lij 
vBri n. water. 























gAtaT*. BJfctTBi 







gitM. gitrtB 




JI^, HrHTJ^ 

«< 11(1111 





grf^, ^ 

























N. V. 










^ifTm . 































,1,1,0, Google 




840. In ordti to muk more pUtnty the kbMnce la Tedle Unpuge of 
■oma of the foima vhloh tn oommon later, ■]■ the tomt of Tadio oeourreDce 
ue added below, and In the otdeT of theli tteqnenor. 

B. SiDgnlar. Nom. ognis etc., as a1>0T«. 

b. Ace.: maM. igaixa, jttjitaa, Qnai;i«m(?); fem. and neal. u 

0. IdiU.: mue. ogpfnl, tKyyi and ttrmlt; fem. AoitO, VMt, 
UMtji, onvtktf, dbtefnft; neat wanting. 

d. Dat: maae. aga6j«; fem. tajitja, au, tuirii; neuL 9^o«7"> 

e. Oen.-abl. : mau. agato, kvjam, orUs; fem. &dltea, hetT^i ud 
bbltiniBa; neai bhdiws. 

f. Lac: maac. agnftA, agn£, iU&ri(?); fem. Itgattn, ndltl, dh&- 
iiMKtayl(?), vMl, bhltmrKm; neut ftpratt, BOpUnt^mtu. 

g. Voc; a> aboTe (nent wanting). 

h. Dnal. Nom.-aoa.-TOB. : maic. b&zl; fem. ynvati; neut f6oi| 
mAhl, b&rt^C?). 

i. Insti.-dat.-abl.: ai aboie. 

J. Qen.-loe. : mate, h&rlofl; fem. TUTKlydB and Jlmidsi neat, wanting. 

k, Plata]. NoaL: male, agn&yos; fern, mat&rao, bhAmls; neat. 
9001, bhtirl, bhAitpi. 

1. Aeeoa.: maso, agnin; fem. k^tls, fi^oayaBC^)- 
ID. Initt., dat.-abl., and loc. : aa abOTS. 

n, Oen.; maao. fem. kATinlEiii, ^u^am etc (nent wanting). 

841. As models of u-stems may be taken 917 fitra m. 
enemy, ^ dhend f. cow, ^ m^dhu n. honey. 




0. 5ralTT^ 









dhan&Ta, db lavfti mtdhnne 
dheuda, dhe ivia midbiuiaa 
dhanlu, dhenvltiii miidliniil 
dhino m&dbu, midlio 

ioy Google 



N.A.y. sFrar 






I.D.Ab. 5|5(i-TnR^ 














M. V. 




















D. Ab 





















The rormi of Vedic o 

eeutrence ate given 

liere (or the u-stems 

in the SUDS miDoer u for the i-items ibaTs. 

». Slngnlai. Nom.: masc. and fem. ae aboie; neat. vix&, ord. 

b. Aceaa.: maae. ketfun, &bblniam, Bucetunam[?); fern, dhsndm. 

o, loitr.: maac. k«t&il&, pUjvt and kr&tUS; fern. Adbenuft and 
ponvii, Aguyij neut. m&dliimft, m&dbTft. 

d. Dat.: maafl. keUvs, gf^Ts; fem. farave, £fvfil; neat. piLqTe(v), 
ur&Te, midhim*. 

e. Abl.-gen.: mate, manyos, pitr&s, olimi^Bs; fem. afndhoB, I^vfts; 
nent. mUbvas and mJidlraaB, mddhoB, m&dhunaa. 

f. too.: maac. pOrSi^, sOn&Tii fem. sladhfta, rdJJvSm; neat 
Bltiitu, otnaTi, Bltno, B^uni. 


k. PIniel. Nom.: maac. fbti&raa, mUbnas aod m&dhvas; fem. 
dkea&Tu, qatakratvaa; neat, purd^, purn, pnrfi. 

g. Voo.; ae above. 

h. Dnal. Nom.-acc-TOf. : m 

Bc. and fem 


1. lDitT.-aat.-abl.: aa iboT«. 

J. Gen.-loe, : at above (but voe 

or aoB). 

D,j,i,i« J, Google 

348—] V. NoDHS AHD ADjsonniB. 122 

1. Aflcni. : muo. ftlUn, pafr&a; fem. {fOa, m&dliTaa. I 

m. Initi., d<t-abl., and loc.: u sbOTe; also gea. (but with the kid* I 
lotioD ""flftin in part). I 

348. Irragalar decleiiBloD, There are no irregular a-stema, 
and oqIjt a very few l-8tema. 

«. SAkbl m. friend has for the fi?e strong caioa a peculiarly 
strengthened base (vriddhled), namely a&khSy, iriiich in the oom. 
aing. is reduced to s&Uift (without ending], and in the other oaaeg 
takes the normal endings. The instr. and dat. sing, have the Dormal 
endings simply, without inserted n or eo:^; the aU.-goB. sing, adds 
ub; and the loc. sing, adds an: the rest is like assl Thus: 

Slog. B&ktafi, aikhftyun, s&kbya, o&kbye, s&khyna, aikhrBiir 
s&khe ; Dn. a&UiirSm o&kblbliy&m, B&khyoa; PI. eikhayaB, B&khln, 

b. Ths Vedi hu ninall; B4UlBr& du., >ad oft«a kmItm the 7 to i, 
in a&khlA, s&khloB, etc. Tbe componnds are naiullj dectloftd like the 
simple word, ualeu (1316 b) aaUia be latstltnted. 

0. There Is ■ corretpondlng fern-, Utkhl (declined Uke devi: 864); 
bat the tana* of sakbl ue also sometime* (oniid utad with feminine Tains. 

d. 7&ti m. is declined regularly In composition, and when it has 
the meaning lord, maxUr; when nncompounded and when meaning 
husband, It Is inflected like a&kbl in the instr., dat, abL-gen., and 
loc. sing-, foraing p&tyK, pitya, p&tyus, p&ty&u. There are occasional 
instances of confnsiun of tbe two clasaes of forms. 

a. For patl as final membei of a poeiesslia rompoond is ngalari; 
and nraall; inbstltnted patnl In the fem.; thus, JIvapatid Having a lining 
htuband; dSsapatitl kmviug a barharian for matUr. 

t. J&nl f. wife hai the gen. ling. J^tajua In the Veda. 

B. aA «ag*r, gr»»dy, hotUle his in the Teda sryiB In pi. nom. and 
■ncai., m*M. and fem. Us aeons, line, la Ktim or ai74m. 

h. Vl bird hu in RT. the nom. 761 (betide via). In the plnial It 
accents vibhlB, vfbhyaa, bnt vlnam. 

1. The stems ik^t eyo, &Bthl bone, d&dlil curdi, and B&ktM thigh, 
are defecliTe, their form* eiahanglng with and complementing foims from 
ilems in &a (ak^&n etc.): see the etems la an, below (481). 

J. The stem patbi road U na«d to make ap part of the Inflection of 
p&nthan: see below, 433. 

k. Kr6fta m. jackal lacki the etiojig oases, for which the coneepond- 
ing forms of kroftf ■re snbatltDted. 

344. Original a^eetive stems in 1 are few; those in n are much 
ire nnmetons (many deriTatiTS verb-atems fanning a paitielpial 

ioy Google 

123 DCCLBMSIOK II., 1- ASD ll-STXMB, [—846 

adjeedTe in u). Their isfleotlon is like that of nonni, and hu been 
included Id the rales girea above. In tboie weak casee, however — 
oamelf, the dat, abl.-gen., and loc. sing., and the geo.-Ioc. dual — 
in which neut« nonnt differ from mascniines in the later laognage 
by an inserted n (we have seen above that this differeuoe does not 
exist in the Veda), Uie ne&ter a^jeotive is allowed to taite tither 
form. The stem Is the same for masculine and neater, and generally 
(and allowably always) for femlDine also. 

a. Th«re are a tew InitMoei of a feminine noun In I ittndlng («ome- 
Umei with ohuged uMnt) bwide a mMcullae tn i: tktu, krfml m., feelmi 
r.; e4khl (343 a) m., Mkbi f . ; dondnbhi m., dundabhl r.; dli&ni 
m., dhunl f.; ^ak^nl m., gaknnl or -nl f. In the later langnage, sips- 
Ciallrt tbtr* It a mj baqaeat laterohaage of i ud I m tnah of the lame 
«tem. No adjectlta tn 1 maket a regnUr femlalue In L 

b> With •temi in a the taae ia quite diffeient While the feminine 
may, and la pait ioee, end In a. Kite the ms*oaltne and neuter, a «pt- 
cUI feminlAc-ttem t« oftoa nude b; lengUientng Ike u lo 0, oi alM hj 
adding I; and tot Mine atcmi a feminine Is fonned Into two of theae three 
wayi, or even in all the thne: thoi, kSrO, -dlpad, fundtayA, eaiilfd, 
vaoasylt; •a^vl, nrri, (corn, pQrri (with prolongation of n bafoie r: 
compire S4Bb), bahvi, prabhTl, roghvi, bUJi.t^ ■Tfid'vi; — pfthu 
and pfthTl, vlbbA sad viVtni, mfd6 ard xof&ri, lagbn and laghvl, 
vitra and t4*vI; bablir& and babtkxA, blbtwtau and tnblutelll, bhiru 
and bl^rtl; — tan£i and taaA and tanvi, phalgu and phalgA and 
ptaalffTl, m&dhit and madhtt and midllvL There are alio tome femi- 
nine noQft^tBma in fl altndlng (nanaliir with ehanged accMit) beelde m*i- 
coliaea In u: thu*, igm m., acrti f.; k&dru m., kodrA f.; gueeulu 
m., gngsold f.; Jatu m., Jatii f.; pfd&kn m., pfdftktt f. 

84B. Koots ending In 1 or n (or j: 870 b) regalarly add a t when 
used aa root-wonh or as root-finals of compoonds; and henee there 
ate no adjeqUves of the loot-olass in this decUnaion. 

a. Yet, Id the Veda, a tew woidi ending In a abort ladieal a are 
declined aa if thii were anfflxat: thna, Aamftadhru, BXt^tja; and the AT. 
baa pftantU^ (once), Boota In fl aometimea alao ahoiten Q to u: thna, 
pcabh^ vlbllL&t etc (.304); go (801 e) be<:omea gn in compotitlon; and 
T» perhaps becomes rl (801 e); while looM in K Eometlmea apparently 
weaken i to i (in -dbl f^om yAhiL eir.; 1158). 

340. Componud adjectives having nonna of this declension as 
final member are infleoted in general like original adjectivee of the 
same endings. 

a> Bnt In tneh compoonds a flnal i or n is eometlmea lengthened to 
form a feminine atem: thai, anfro^I, Bwayonl or -ni, -gitraysftl or 
•(1; Tftmora oi -m, dwhiu^Q or ■90, varatand, mltrbandhO; and 
RT. ha* Aql^I from qign. 

ioy Google 

s AND Adjectives. 

Declension III. 

stems in long voveU: )9T S, ^ I| 3i &• 

847. The st«me ending in long rowela fall into two 
well-maiked classes or divisions: A. monosyllabic stems — 
mostly bare toots — and their compounds, with a compar- 
atively small number of otheis inflected like them; B. de- 
rivative feminine stems in i^T S and ^ 1, with a small num- 
ber in ^ D which in the later language have come to be 
inflected like them. The latter division is by fai the larger 
and more important, since most femiiune adjectives, and 
considerable classes of feminine nouns, ending in ^ S or 
^ I, belong to it. 

A. Boot^-words, and those inflected like them. 

848. The inflection of these stems is by the normal 
endings throughout, oi in the manner of consonast-stems 
(with ^ am, not tT m, in the accus. sing.); peculiarities 
like those of the other vowel-declensions are wanting. The 
simple words are, a^ nouns, with few exceptions feminine; 
as adjectives (rarely), and in adjective compounds, they are 
alike ia masculine and feminine forms. They may, for con- 
venience of description, be divided into the following sub- 
classes : 

1. Root-words, or monosyllableB bavlaK the aspsct of such. Those 
in S aie so rare that It is hardly possible to make op a whole scheme 
of forms In actual use; those in I and & are more nomerons, bat still 
very few- 

2. Compounds having such words, or other roots with loog final 
vowels, as last member. 

3. Polysyllabic words, of various origin and character, inolading 
in the Veda many which later are transferred to other declenilons. 

4. As an appandix to this clasa we may most conveniently 
describe the half-dozen stems, mostly of regalar inflection, ending in 

ioy Google 

125 Declension III., i-, i-, and Qi-steiis. [—861 

340. HonoBylUbic atema. Before the endiDga beginDlog with 
Toirela, fioal I Is ohuiged to ly and Q to uv; while final S ia dropped 
altogether, except in the strong oases, and in the aoo. pi., which is 
like the nominative (according to the giAmmarlanB, ft Is lost here also: 
DO instaDoes of ibe occurrence of such a form appear to be quotable). 
Stems in I and Q are in the later langnage allowed to take optionally 
the fuller endlngi U, Sa, Om in the singular (dat., abt.-gen., loc.)i but 
DO such forms are ever met with in the Veda (except blilT&{[F], RV., 
once). Before Km of gen. pi., n may or may Dot be inserted; in the 
Veda It is regnlarly inserted, with a single exception (dhlyiUii, once). 
The vocative is like the nominative in the singular a« well as the 
other nnmbers; but iDstaaceB of ita oocnrrence In nncomponnded iteits 
are not found in the Veds, sod must be extremely rare everywhere. 
The earlier Vedio dnal ending is s instead of &u. 

3B0. To the i- and fl-stems the rules for monosyllabic accent 
apply, the accent is thrown forward upon the endinge in all the weak 
cases except the acous. pi-, which ia like the nom. But the &-stems 
appear (the instances are extremely few} to keep the accent upon the 
stem throughout. 

SBl. EzampleB of deolension. As models of mon- 
osyllabic inflection we may take sIT ji f. progeny; ^ dhi f. 
thought; and ^ bhd f. earth. 

a. The flrat ot tbeie is lathei aibilruil; extended nom the four cues 
which >ctD4lly eccut; ot the loe. ilng. knd gen.-lcc. du., no Vedk examples 
from ft-atemi 4ie found. 















fil&, ^ 



Shisb, dhly&C 

bhuv«, bhuTil 

3^^ JOB 


bhavf, bbuv^ 




Dij.ieo, Google 

V. Nouns itat AsnoriTW. 
































jfaiam, J^ 

dbijr^, dblnim bhuvitm. bhOiUbi 







86S. HonoeylUbio items in composUion. When the nonna 
above described ooenr as final member of a compound, or vben ai^ 
root In ft or I or a fa found in a Hke position, the lofleotion of an 
S-stem is as above. Bnt I- and fi-atems follov a ditided usage : the 
final Towel before a vowel-ending Is either oooverted into a short 
vowel and semivowel (ly or av, as above) or into a semivowel simply 
(y or v). The accent is nowhere thrown forward upon the endings; 
and therafbre, when I and Q become y and v, the resaldag syllable 
is circumflex (8S-4). Thns: 

Mhc. (tad fern. Slngvlu: 

•dhfyam •dliyim 
-dtaiyK -dhylt 

-dhlye -dbyi 

-dhlyaa -dbylM 
-dhiyt -dhri 

-bhuvam -bhvitAi 

•bhi^Tt -bhTt 

■bbdve -bhvi 

-bhiiTaa -bbTJM 

•hhixTl -bhTl 

ioy Google 

27 Dholbhsioh IlL, S-, I-, add a-eT»MS. [— f 


N. A. T. -dtairU -dbrtiik •1>h&Tld -bhrKJi 
I. D. Ab. -dbibhrftm -bbtkbhyim 

0. L. -dhiyoa -dhfAs •bh&voa -bhT6i 


N. A. T. .dblyH -dbylw -bhi^Tu -bbviu 

1. .dhibbiB -bhAbbia 
D. Ab. -dufbhyas •bhAbbjkB 

(-dbiyitai .. .„ r-bhiivam 

L. -dhiqu -bbtifn 

ft. At to lbs ftdmlntblllty of tha tnllei endlnp U, Se, and iia In the 
•tngula (hBlatne), gnMrnatleal iithotltlu tn imaewhtt tt ladMM; bat 
th«; ue nsTBi hand In tht V*da, ind k<*e bw» onltted from A« abote 
■ohmiM u fvobaMy nnial, 

b. It tiro MDHinntB pnceda tkc final I or Q, the diatyUaMc fOnna, 
vldi ty ud Tiv, are legulaily KTitten; after one eanuinant, the mage it 
Tarytng. The gnmmttrian* prescribe ly and ut vhen the monoirllahle »tem 
hai more the character of a noon, and y and v when It la moia pure]; a 
Terbal lOOt with partlolpUl Talue. No raoh ^tlootlon, bowever, It to be leen 
io tbe Te(l4 — where, moreover, the dUfeience of the two foimi la onl; 
graphic, glnoe tbe yS- and vK-forme end the tett ere alwajr* t« be read ai 
dinrllabic; IB or IS and n& ot Oft, and m on. 

c. Ai to neater Btamt foi euch adJeetlTOi, see 867. 

853, A few fluthei Tedte Irtogalarltlea or pemllarttiet maj be brteflT 

a. Of the K-stemi, the forma In fis, ftm, S (da.] are aometlmet to 
be read ti dlaeyllablee, au, aaro, aa. The datiTO of the item aiod at 
InfinltiTe la el (u It £ + e}: thns, prakbyBl, pratim&{, parhUl. 

b. Iiregnlir tniuter ot the accent to the ending in compoondi it seen 
In e ea» or two : thw, avadyabhlyi (EV.), Kdbli (AV.). 

964. Bat oompounds of the olasB above deserlbed are not ia- 
tnqmvtiy transferred to other modeB of InflMtion: the S shortened 
to a for a masculine (and neuter] etem, or declined like a Btem of 
the derlTative s-cUbb (below, 864) as feminine; the I and a sbort- 
ened to 1 and n, and inflected as of the second declension. 

a. Thna, componnd sterna In -gtt, -Ja, -da, -Bthai -bhu, end others, ate 
foand even In tbe Veda, end become freqnent later (being made fTom all, oi 
nearlr all, the lOOta in 4) ; aad sporadic casei from yet others oeoai : for eiampls, 
grtap&a, rayocUilfa and ratnadhibbla, dhommiie [all RT.]; and, 
tfim I and a sompoandt, vefaqriB (TS.), Abrayaa (RV.), gaija^bhlB 
(ST.), kaniuu^s (QB.) and rtanfbbyaa (RV.) and Beninlbbyaa (VS.) 
sad grfimavlbbls (TB.), mp^ft (AT.), qttlbtariTe (TS.). 

b. BtlU more nomeroos are tbe feminine* In ft which hsTO lost their 

itizecy Google 


root-d«eUn(lon: examples m pr^^ (of irhieli the tarthei Mmponndi In 
part bave toot-foima), Bvadhi, graddba, pratlma, and othen. 

o. Then, In the later langaage, a few femlnioea in I ate made fiem 
the ateme 1q a ahaitened ttam t: thna, gopl, goqthi, pumagi, paflki^Ii 
bhnjagl, bbujaihgi, Bur&pl. 

SfiS. Polyay llabio Stems. Stems of tbia division (A) of mora 
than one ajllabte ue very rare indeed in tbe later language, and tiy 
no means common in the earlier. The Rlg-Veda, however, preeeuta 
a not ioooDBtderable body of them; and aa the class nearly dies ont 
later, by the disuse of Its sterna or their transfer to other modes of 
deolenaioD, it may be best described on a Vedic basis. 

a. Of stems in &, maieullnea, hslt-s-dotsn ocdot Id the Teda: p&nthi, 
m&nth&, and rbbukft ue otherwiae Ttsved by the later grammar: sea 
belov, 4SS-4; uqiuB (nom. pr.) has the atiomaloaa nam. elng. Tiq&nl 
(and lee. w irril as dat. uq&oe); mahi grtai is found only In aocui, sing, 
and abnndantly In composition; atfi ^ome hae onlj> atfiau not derivable 
from jta. 

b. Of stems In I, oier seventy sra found In the Veda, nearly all 
feminine*, and all accented on the final. Half of the feniinJne* are farmed 
from maacnlines vlth change of accent: thus, knlyft^I (m. kalyiE^), 
puruft (m. p6rufa]i othars show no change of accent: thni, yaml (m, 
jaxak); others still have no corresponding maacaUnes: thai, nadi, Iftkjmi, 
Btlnni. the masculines are about ten in number: for example, rathl, 
prSvi, atari, ahf, Apatht. 

o. Of stems In 0, the number is amallet: these, too, are nearly all 
femlnlnes, and all accented on tbe Anal. The majority of them are the 
feminine adjectives In tt lo masculines in tt oi U (above, 344 b) : thus, 
oara^yd, carlf^llt, Jlgfaatati, madbd. A fair are ooans in d, with 
change of accent; thus, agtii (igi-a), pfdikA (p^dikn), frogrlt (qvA- 
fora); or vithoat change, as nptd. And a few have no correspanding 
maBcnlinei: thns, tanA, vadhii, oaind. The masculines are only tvo or 
three: namely, pr&gd, k^kadftfti, makftt(?]; and their forms are of the 
QlmoEt rarity. 

960. The mode of declensioD of these words may be iUustrated 
by tbe following examples: ratbl m. eharioUer; nadf f. itream; tanft 

a. No one of the selected examples occurs In all the Torms; forms for 
wbloh no example at all la quotable are put In brackets. No loe. ring, from 
any I-etem occure, to deteraine what the form would be. The stem nadf 
is selected aa example partly in order to emphasize the diSeience between 
the earlier Un|uage and the later In regard to the word* of this division : 
aadi 1* later the model of derivative Inflection. 

ioy Google 

1 DECIJ1S8I0N III., KADIOAL S-, 1-, AND tl-8TBMa, [—868 


















Ab. Q. 











N. A. V. 




L D. Ab. 




G. L. 




M. A. 




















b. ThB cuei 

1 — nadlam, tan6&m, etc. — hb 

written abOTe •ceoid- 

lug to their trn« pbonetlo form, Umoat Invariably belonging to tbem ia 
Qie V«dK; In the wittton tozt, ot coarse, tbe stem-flnal it made a Beal- 
Towel, and th« teiolting tyUable U oltcamdeied; thus, nadykm, tan- 
viaa, etc. ; only, u mail, after trro eciucDanta tb« ruolved form* i7 and 
DV ate mitten ioitead; and alio wbere tbe eombinstlon yv wonld othet- 
iriae remit! thas, oakriyfi, [agrAvSl,] and mltrfiyriTaB. The BV. really 
leadi atarrfam eto. twice, and tanvka eto. four timet; and mob con- 
tractlong ate mots often made In tbe AT. The ending ft of the nom.-acc.-Too, 
da. U tbe equivalent of the later &u. Tbe nom. dng. In a from t-eums 
1* foaad In the older langnige about sixty timeg, ftoia over thirty iteme. 

367. Irregnluitles ot form, ptopetly eo called, ue very few In thU 
division: oamli u loo, aing. [Inatoad of oatllTi) occnre a few time*; and 
there ie another doobtful cue ot two of tbe euoe kind; tbe flnal d ia re- 
garded as prag^bya or uncDnibinable(188); ton'ol ia lengthened to tKOvi 

In a passage ot two; -yuvaa is once or twice abbreviated to -ylSlB. 

368. The process of tranifer to tbe other form of I- and u-deoleniion 
(below, 36S It,), whlcb has nearly esttngniihod this category of words In 
the Latei langnage, hu Its beginolDgs In the Teds; bat In BY. they are 
escesslTely tcenty: namely, datiim, loc sing., once, and gvaqmam, do,, 
once, and dravltllu&, Instr. dng., with two or three othet doabtfal rasei. 
In the AOiaiTan, we dnd tbe aco. sing. kabAm, tonAm, vadHllbn; tbe 
ln«tr. alng. paliUJ and one or two others; the dat. dng. vadlivftl, gva- 
qnuU, agruTSi; the abl.-gen. sln^ pimarbb^Tfis, pfdUcu^. qra^maa; 
and the lee. sing. tonASm (with anomalona accent). Accnsativee ploial in 
la and ua ate nowhere met with. 

WbltnsT, Qismnar. 3. ti. S 

ioy Google 

3B9— ] V. Noons and Adjectives. 130 

360. Adjective componadi from thaie wordi are vaiy tev; thow Thieb 
occui Bie declined like the slmpla atamB; thai, birn^aTftqls tmd aali&a- 
nstaris, AtaptatanOB and B&rrattuiaa, ill nom. sing, mucnllns. 

BtemB ending in diphthongs. 

3S0. There ue certatn raonoByllablo Bteme endiDg in diphthonge, 
which are too few and too diverse in Inflection to Piake a declension 
of, and which may be moat appropriately dlepoaed of here, in con- 
nection with the Btemg In I and u, with which tbey have qiost afBnity. 
They are: 

ft. Btema in &a: naA and Klta; 

b. etems in &1: ttA; 

e. Btema in o: g6 and dyA (or dyu, 6iv]. 

361, a. The stem nftu f. ihip ia entirely regular, taking the 
normal endlnga throughout, and following the loleB for monosyllabic 
accentuation (317J— except that the accas. pi. is gaid (it does not 
appear to occnr in accented texts) to be like the nom. Thus: nfiuB, 
nJvam, afiva, n&7d, nftv&a, nivl; nJvSu,n&ubh7^> n&vAa; n^vaa, 
nivae, n&ubhlB, nAubhy&a, nftvam, n&u^u. The stem glAu m. ball 
is apparently inflected in the same way; but few of its forma have 
been met with in use.. 

b. The Btem ral f (or m.) tcealth might be better desorlbed bb 
rS with a union-consonant j (S68) interposed before vowel endings, 
and Ib regularly inflected as anch, with aonnal endings and loono- 
Byllabic accent Thus: ras, royam, rilya, riji, roy&B, rjl3^; ray&n, 
t&bbTlbm, rfiy^Bi lijaa, rSy&B, r&bhlB, rftbby&a, rSyam, ruii. But 
in the Veda the accus. pi. is either rBty&a or rayas; for accus. sing, 
and pi. are also need the briefer forma ram (RV. once: ilyam does 
not occur in V.) and r^ (SV., once); and the gen.-sing. is Bometimes 
anomalously accented tiyaa. 

c. The stem g& m. or f bull or com is much more irregular. In 
the strong cases, except sccub. sing., it ia strengthened to gKii, form- 
ing (like nftu) gftus. a^vftu, gavas. In accus. alng. and pi. it haa 
(like rsi) the brief forms gam and glua. The abl--gen. sing, is s6b 
(aa if from gu). The reat is regularly made from go, with the normal 
endings, but with accent always remaining irregularly upon the stem: 
thus, g&vft, g&ve, g&vi, g&vos, g&vSm ; g6bhyam, gdbhis, gAbbyas, 
gd^u. In the Veda, another form of the gen. pi. is g6n&m; the nom. 
etc. du. is (as la all other such cases] also givA; and g^m, g6s, and 
g^ are not infreqoently to be pronounced as disaylUibles. As ace. 
pi. is found a few times g&vsa 

d. The stem dy6 f. (but in V. usually m.) sky, day la yet more 
anomalous, having beside it a simpler stem dyu, which beoomes dlT 
before a vowel-ending. The native grammarians treat the two aa 

ioy Google 


IndependeDt irordB, but it is more convenient to pat them together. 
The stem d76 is inflected precisely like k6, as above deacribed. The 
complete deoleuBion ia m follovs (with fonnB not aotnally met with 
in use bracketed): 

Slngolu. Dn>I. FlnnL . 

A. divam dyim K"**'*"' "*^"''" divia. dylin [dyfc] 

I. atvi [dy&v&] i dyiibMB [dy6blilB] \ 

G. divia dy«B 1^., , , . , [divim dyAvaml 

I. divl dyivl If^'"^' ^'^1 dy«,« [dy6,u] ] 

«. The dat. ttag. dy&ve ii not foond In tha early luigntge. Both 
divoa and dlv&a occoi m uont. pL in Y. Ac nom. etc. da., dyarH le, 
M naaal, tha regolaT Tadic form; onu oecnrs dy&vi (iu.), ai If a neuter 
formj and dySl^ is fonnd oqcq used aa ablatiTe. The citee dySOfl, dyBm 
and dynn (once) ue lead In V. sometiinea aa dl<«yllablea; and the first 
at accented ToeaCiTe then becomes dySAa (i. e. dlftua: see 314b). 

f. Adjective oompoandB having a diphthongal item u final membei 
an not nametuiu, and tend to aliorten the diphthong to a vowel. Thus, 
from nau ve haie bhinnum; from go, tevetat words like icu, sapt&gu, 
BDgO, bahngu (f. •gA TB.) ; and, eorreBpandlngly, r&i eegtog to be reduced 
to rl In bfh&draye and rdliAdrayaa (BV.). In dedvatlon, go maintains 
lt« tnll fonu in gotra, ac6t&, -gava (r. -gavi), etc.; a* flnt member of 
a componnd, it la titIoiuI; treated: thne, g&Tftqlr, gkvi^ (but ga&qir, 
gaS^t^ K.), eto.; goa^v4 oi go'qra, gdfjlka, gdopa^a, etc In certain 
eompoandi, also, dyn or dyo i&kea an aaomaloaa form: thus, dySurdA 
(K.), dyKuTlokA (QB.), dySoaadiglta (AV.). In rvritat (nnleu ibit U 
for rayiT&nt) rftl becomea re. BY. has &dlirig&vaa from &dhrlgn (of 
quMtlonible import); and AY. hu ghitastaTaa, apparently accna. pi. of 
ghrtaatu or -eto. 

B. DeriTatiTe stems in S, I, O. 

862. To this divisioii belong all the ft and I-stems which 
have not been specified above as belonging to the other or 
loot-word division; and also, in the later language, most 
of the I and ll-stems of the other division, by transfer to 
a more predominant mode of inflection. Thus: 

1. a. The great mass of derivative feminine fi-steins, subBtantive 
and adjective. 

b. The inflectlDn of theae stems has maintained itself with little change 
tluough the whole hlitory of the language, being almost precisely the same 
in the Tedaa u later. 

ioy Google 

sea—] V. Nouns and Adjeotivbs. 132 

2. o. The great m&as of deriTative faminine i-stema. 

d. This cIms it without exoBptioii In the Ut«i laogiuge, In th» euller, 
it loffera the ezceptloa pointed out iboTe (3E6b): that femlnlset m&de 
with change of accent follow thla mode of declension only wh«n the accent 
U Dot on the i; thn«, titirlql, p&mqql, p&llkiu, T6lii9l> 

•< The I-ilenu of this division in genenl ue leguded u made by 
eoDtnctlon of an eullei ending In yK. Theli InlleettoD hu become In the 
later laDgnage somewhat mixed with that of the otiier division, and so fU 
dlfferont from the Tedlc inDeetton: see below, 363 g. 

f. Toir few dGTlTatlye stems in 1 uo lecognized by the grammaiiana 
IR declined like the lOOt-diTlsion ; the Vedic wotds of that class are, If . 
letalned In nse, tTtnifenod to thia mode of inHectlon. 

g. A very small number of mucnilne I-stema (half-a-doten) are in the 
Teda declined aa of the derivatiTe dlTjalon: they are a few rare proper 
names, install etc.; and r^trl and eixt '(only one case each]. 

3. h. The u-Btems are few tn namber, and are transfers from the 
other dirisioit, assimilated in inflection to the great class of deriyattve 
I-stems (except that they retab the ending b of the nom. sing.]- 

363. EndiDfB. The points of dlstlnctian between this and the other 
dlTUion are a» follows: 

a. In nom, sing, the UEaal s-endlng is wanting: except in the Q-atems 
and ■ very few i-etema — namely, lakfiml, tail, tantH, tandU — which 
have preserved the ending of the other division. 

b> The aoens. sing, and pi. add limply m and b respectively. 

o. The dat., abl.'gen., and loc. alng. take always the fuller endings 
ftl, Ka, Km; and these are separated from the final of the fi-stenis by an 
interposed y. In Biahmana etc., 01 is generally snbatitaled tor Oa (8071i). 

d. Befoie tbe endlnge a of Instt. sing, and oe of gen.-loe. du., the final 
of S-etema Is treated aa If changed to «; bat in the Teda, the instr. end- 
ing & very often (in nearly half the occnrrences) blends with the final to ft. 
The yft of I-etems la in a few Tedio examples contracted to I, and even 
to 1. A toe sing. In i occurs a few times. 

«. In all the weakest esses above mentioned, the accent of an i- at 
n-stem having acVe final Is thrown forward upon the ending. In tlie 
remaining case of the same dua, tbe gen. pL, a n is always Interposed 
between stem and ending, and the accent remaina upon the fanner (in RV., 
however, it is UBuall; thrown forward npon the ending, as tn i and n-stems). 

f. In Yoc. liog., final. & beoomei ej final i and u are shortened. 

g. In nom.-acc.-voo. da. and nam. pi. appeaia In I (and a)-Btems • 
marked difference between the earlier and later langroage, the latter borrow- 
ing the forms of the other dlviilan. The da. ending iu la unknown in 
BV., and very rare in AY.; the Vedic ending la i (a conesponding dual 
of a-at«ma doea not occnr}. The tegnlar later pi. ending as bsi only • 

ioy Google 


DbOLBNSION ni., DESITATIvei S-, I-, AHD &>STEU3. [—864 

doubtful exunple oi tiro in BY., and a Tery small namber to AT.; tlia 
cue tbeN (and It la one at vary bequeot oocarrenccj adda S ilmpl;; and 
thongh yaB-fonna oeoui In tlie Biabmanaa, along witb iB-foima, both aie 
tutd Mthei ludiffeMntl; as nam. and accni. (as, Indeed, they BODietlmet 
Inteichange aba in the eplet). Of ft-Eteme, the dn. nam. etc eadi in «, 
iKth earlier and Itter^ in p]., of eonne, B-forme aie indlstlngnlahable from 
sa-roTniB. The RT. hag a few examples of Bsu for fta. 
b. The lemalniiig ease* call for no remark, 

864. Examples of declension. As models of the 
inflection of denTative stems ending in long vowels, we 
may take WU sinB f. army; ^RIT kanyfi f. yiW; ^sft devi 
f. goddess; ^U vadhli f. i 






































Ab. 0. 
































I.D. Ab. tlHI*-mi^^ etrUIJ-yiiJ^ "^cJl*-«iH^ ^q^^TTt^ 

BinSbhylm kanyJtbbjBm devibliyllm vadhlibliySm 

Bdniffos kany^yoe d«vji>a vadbv^a 

ioy Google 

V. Nouns and Adjectives. 























a. In the Ted> vadhd is a ctem lielonglDg to the other dl<rl(loQ (lik« 
tand, above, 366). 

366. EiamploB of Vedtc forms are; 

a. A-stenu: iDetr. sing. manifK (tM» itmplei form U eep«clilly com- 
moii tTODi Items in t& end it); aom. pi. vaqdlaaa (about twenty eiamplcs]; 
arciu. pi- aradigamaBaa (a cue oc tiro). Half the bhyaa-eases ue to 
ba Tead as bhlaa; the Sm of gen. pi. la a few Hmea to be leioWed into 
aam; and the & and Am of nom. acco*. aing. ue, Teiy nraly, to be 
treated in the same manner. 

b. I-atems: iDati. sing, q&mt, Q^inl; loc. g&tiri; nom. etc. dn. devl; 
nom. pL davlB; gen. pi, bahvinltlll. The llnsi of the atem ia to ba leatl 
ai a vowel (not y) fcequeotly, but not In the majority of inataoeea: thns, 
d«Tli, ABviia, deriam, r6daBio8. 

o. The spoiadlo Inctancea of transfet betweea this dlviilon and the 
preceding have been already GQfficiently notieed. 

d. Of the regular aabstltntlon made la the B^hmaiia language (307 b, 
336 Bi 363 a) of the dat sing, eadiog 111 for the gan.-abl. ending Sa, in 
all elaiBes of vords admitting the latter ending, a few examples may be glren 
heiei abblbhuty&i riipam (AB.) a tign of overpmoering; tTiftnbha9 
oa JagatySi oa (AB.) of the truirtt triftubh andjagad; vSco dSivyftl 
oa m&nufySi oa (AA.) oftpeech, both divint and human; etrly&i paya^ 
(AB.) iconxm'* mvtt; dhenvri v4 atAd rita^ (TB.) that, foriooth, u Me 
ae»d of tht com ; JlnjByU tvaoa^ (KB.) of dead ekin ; JyftyasI y^ySy&l 
(AB.) euperior to the yajyi; asyfii dlvo 'smftd antarlkfftt (^9a.)_/>-om 
thia heaven, from thie atmoephere. The same subilitntton Is made ones In 
the AY.: thus, BT&pantr asyU Jiiat&ya^ let her relative* tleep. 

ioy Google 

135 Declgnsion, III. Desitatite &-, !•, and a-STEue. [—368 

366. The nonn strf f. woman (proliiMy oontraetad ^om Butri gaie- 
ratrii), roUovs * mlxad dadeoBion: thus, atri, Btriyam. or atrim, etiiy^ 
Btriyii, Btriy^ BtriTlbn, stri; strly&a, BtTibbrim, strlyde; atriyas, 
BtrfyaB or tttia, Btrfbh^ atribhy&s, strii^im, sti^o (bat the &caiu- 
atiTu Bb^Qn tnd atria ue not found In the older Uognage, uid the to<i. 
Btrt iB Dot quotable). The accentuation U that of a root-woid; the foim« 
(coDiptenontl; (he nom. liog,) are thoie of the other or derlTati«e dlTttlon. 


867. a. The oocnneDce of original adjectives in long final vowels, 
and of compoands having aa final member a item of the first division, 
has been snfficientlj treated above, ao far aa masonllne and feminine 
forms are oonoemed. To form a neuter stem in composition, th« rule 
of the later langnage is that the final long vowel be shortened; and 
the atem so made is to be infieoted like an adjective in i or u (839, 
341, 344). 

b. Such neater foima are very rare, and In the older langnage almaat 
nnknowa. Of neater« from i-tteme have been noted In tbe Veda onl; 
hariqiijaja, tea. elng. (k muc. form), and auftdhlas, gan. ilng. («ame 
aa maao. and tbni.); from fl'Steme, onl; a few ezamplea, and from atem- 
fonng which might be mate, and fam. aUo : thus, vlbhn, subhu, etc. (nom.- 
Kce. alng.: eompare 364); aapuft and msyobIi4vfi, inatr. alng.; and 
mayobha, ace. pL (compare puru: 34Sk); from a-iteme occnr only half- 
a-dozen ezamplee of a nom. atng. In ta, like the muc. and fern. form. 

o. Componods having nonna of the second division as final 
member are common only from derivatives in &; and these shorten 
the final to a In both masculine and neater: thus, from a not and 
pr^Jft progeny come the masc. and neat, stem apraja, fern. apr^S 
childlen. Snch compounds with nonns in 1 and u are said to be in- 
flected in masc. and fern, like the simple words (onl; with In and fin 
in ace. pi. masc.); bnt the examples given bj the grammarians are 

d. Sterna with ahortened final are occasionally met with: thna, ekn- 
patni, kttalakfmi; and anch adverbe (neut. alng. accus.) u upabhftlmi, 
abliyi^aylai. The etem atrl Is directed to be ahottened to Btri for all 

368. It is conrement to give a complete paiadigm, 
for all gendera, of an adjeotive-stem in ^ a. We take for 
the purpose m^ pSpi evil, of which the femimne is usu- 
ally made in ^ S in the later language, but in ^ I in the 

itizecy Google 


V. Nouns ahi 







pftp&s papim 




















































N. A.V. 

TtqT 'Tlq 



pIpKd pilpe 



I. D. Ab. 















qraif(^ HHlFi 



pap& pSp±nl 




^TTH MIMlfn 




pftpan pSpanl 










D. Ab. 






,1,1.0, Google 


pBp^nSm p&pinBxa pApin&m 

pap^u pKplisa papifn 

Stems in ff T ("^^ Sf^^' «r) . 
860. This deolensioQ ia a compaiatively limited one, 
being almost entirely composed of deiivative notuB foimed 
with the suffix TT tr (oi fT^ tar), which makes masculine 
nomina agentis [used also paitioipially], and a fair nouns of 

a. Bat it inclades also a few noons of relationslilp not made 
with that snfGs: Darnel; darf m., wia^ and ii&aandr f.; aad, besideB 
these, n( m., et^ (in V.) m., not (>d V.) f., sAvyafthf m., and the 
fsmlnine numeralB ttsp and oRtasp (fot which, oee 463 e.t). The 
femininee in tp are only mftt^, dnliit^, and y^tp. 

b. The inflection of these sterna is qnlte close!; analogoas with 
that of ateiDB id 1 and n (second declanaion); its peooliarity, as 
compared with them, conaiats mainlj in the treatmeot of the atem 
itself, which has a doable fonn, fuller in the atrong cues, briefer in 
the weak ones. 

370. Forma of the Stem. In the weak oaaee (excepting the 
loc. sing.) the stem-finnl la r, which in the weakest cases, or before 
a Towel-endlng, is changed regnlarlj to r (139). Bat as regards the 
strong caaes, the sterna of this declension fall Into two classes: in 
one of tbem — which Is very mncb tbe larger, containing ail the 
nomitta agtntit, and also the nouna of relationship niptp and eT&sf, 
and the irregular words atf and aavraft'^T — ^* F ^ vriddhied, or 
becomes fir; In the other, containing most of the nouns of relationship, 
with nf and xie(, the x is gnnated, or changed to ar. Id both classes, 
the loc. sing, has ar as stem-final. 

671. Endings. These are in general the normal, bat with the 
dlowiog exceptions: 

a. Tbe nam. alng. (nue. tnd f«m.) ends ilway* In K (toi orlgloal ars 
OT in). The Toe. sing, endi In ar. 

b. The MCDS. (lug. tddi am to tha (atrengthened) stem; the accaa. 
pL liw (lile 1- and n-stema) n u mage, ending and a u fern, endlag, irlth 
the f lengthened bsfoie them. 

ioy Google 

371—] V, Nouns and Adjectitbb- 138 

0. The Bbl.'goTi. elog. cbangeB p to ur (ot ns: ISOb). 

&. The gen. pi. (u In 1 and a-Btems) Inaerti n befOTB &m, and 
Ungthena the neca-flDal before It. Bat tbe p of n^ may also temdn short 

e. The above ua the rules of the later laogaige. Tbe older prMenta 
certaiD deristloiu from them. Tbni: 

f. The ending In nom.-aw.-Toc da. ie (a« anlTersally In the Teda) 
regularly & Inatead of &a (only ten Sa-forms In RV.). 

g. The 1 of loc. sing, is lengthened to I In a fev voida: thae, barUrl. 
h. In the gen. pi., the RV. has once BV&Br&m, witbout Inserted n ; 

and n&r^ Instead of Uf^^m ll freqaant. 

1. Otber liieguUrltleg of ni are the dng. dst. nAre, gen. n&rftB, and 
loc. nkd. The Veda -writes always nfi^am in gen. pi., but Its p is In a 
majority of oases metrically long. 

J. The stem usf t. datcn has the toc. sing, ofar, the gea. sing, uei&s; 
and the accns. pi. also aar&ii, and loc. sing, asrjtm (which Is metrically 
trtsyllablo: nB^^m), as If in analogy vlth I and &-steros. Once occurs 
usrf in loc. sing., bnt it is to be read as if the regolar tilsylUblc form, 
u;&rl (for tbe exchange of a and ^, see 181 r). 

b. From atf come only tieaa (ippsreDtiy) and Btfbhia. 

L la the gen.-loc. dn., the T Is almost always to be read u a sepa- 
rate syllable, j, before the ending ob; thaa, pltf&B, etc. On the oontrary, 
n&nfcndarl is once to be read n&n&ndrl. 

in- For oeuter forms, see belOM', STB. 

ST2. Accent. Tbe accentuation foUowB closely the rulea for 
1- and a-Btems: if on the final of the stem, it coDtinneB, as aonte, oo 
the correBponding Billable throughout, except In the gen. pi., where 
it may be (and in the Veda always is) thrown forward open the 
ending; where. In tbe weakest oases, r becomes t, the ending has the 
accent. The two monosyllabic stems, at and ett. do not show the 
monosyDabic accent: thas (besides the forms already given above), 
nf blilB, ntqn. 

878. Examples of declension. As models of this 
mode of inflection, we may take (torn the first class (with 
QT^ &r in the strong forms] the stems ^1^ d&tf m. fftcer 
and 7^ BT&ar f. sister\ from the second class (with ^^ ar 
in the strong foims), the stem v^ pit? m. father. 
N. ^THT T^HT ftfTT 

dftti bv&bk piti 

A- ^IfTT^^ Rflipi^ ftrHTI^ 

dfttaram Bv&Bftram pltirtua 

ioy Google 

139 Declension IV., r-STHMa. [—873 















At. 0. 






















N. A. V. 






I. D. Ab, 














N. V. 




















D. Ab. 




















a. The femimne Btem 

qiH mStt, mother, is inflected pre- 

cisely like 

ft(T pilt, excepting that its accusative plural it 


ioy Google 

373 — ] Y. NoDHS AND Adjeotites- 140 

b. Th« pecnliai Vedio tonus hive Iwen •nfflaleutly iDSttnced sbore; I 
the only one* of otbei Aui sporadic oecDtienee being tlie nom. etc. dn. I 
dstar&, ev&aSrS. plt&rfi, and tha gen. pi. of Up, amrim. \ 

C. The nom. pi. forma pitsros and mfitaraa etc. are found used also 
aa accUB. in the epioa. 

874. The stem kro^ff m. jackal (lit'ly Juncler) BubstHutea in tbe 
middle casee the oorreBponding foims of krA^tii (P^^ ^)- 

976. Neater forniB. The gtMomaibaia preacribe a complete 
neuter declenaion also for bues in tp, precisely accordant with that 
of vM or m&dhu [above, 330. 341]. ThoB, for example: 
Slog. Du. FIdi. ' 

K. A. eOAtt dh&t^ dliKt^l 

I. dbStj^ dhst;^bb;am dhftttbhia 

Q. dbStj^ao dhStt^OB dhStr^ini 

y, dbAtTf dbatar dhatTDi dliatrnl. 

a. The weakest cases, bowever (as of 1- and u-stema used ad< 
jeotivelj: 344), are allowed also to be formed like the oorrespondii^ 
masculine cases: thas, dJi&tri etc. 

b. No Bach nentei foiiDa chance to occni In the Veda, bnt they begUi 
to appear In the Biihmanas, ander Inflaeuce of the oommon tendensf 
(compare Oeim. £efter, Stlterin; Fr. metUettr, mtnteute) to glie thli 
nomen agentU a more adJectiTe oharaeter making it coneapood In gender 
with the Donn vblob it (oppositlTely) qnallBea. Thns, we have in 
TB. bbartf and Janayltt. qnaiuviug ant&rlkfami and bliar^i^ and 
JanaTitp^l, qnalUylng nU^trSi^; aa, in H., grahitffi, qualifying 

O. When a feminine nonn is to be qnalifled in IUeb manner, the naoal 
feminine derivatlTe In I ta employed: thns, In TB., bhartryka and bhar- 
tr^&u, Janayltrriis and Jant^tlTSi^ qualifying itpaa and ahorStr^; 
and such inatances are not anoommon. 

d. The BY, ahowa the same tendency very cnrioaely onoa In the aecns. 
pi. mfitrn, instead of mSt^, in appoaitlon with masculine noona [RY. 
I. 36.2). 

e. Other oentei forms In RV. ate Btbfitnr gen. glng., dfamat&rt lae. 
sing.; and tor the nom. aing., Initead of -tf, a few more oi lesa doublfnl 
cases, atbStar, sthftt^r, dliart&ri. 


876, a. There are no original adjectives of this declension: for 
the quHsi-adjectival character of the noanB composing it, see above 
(876bj. The feminine stem is made bf the snffix i: thus, dfiti^ dbltci. 

b. Roots ending to x (like those in 1 and a : 345) add a t to make 
a declinable stem, when ocoorring as final member of a oomponnd: 

ioy Google 


thas, kum&k^ (»^)i ▼(Orabhtt (ybhr), balUift (ylir). Prom some 
r-raota, also, ue made stems in ir and ur; see below, 3B3 a, b. 

o. NoDDB in 7 as finala of adjective componada are inflected in 
the same toaoner as when simple, in the maacullne and feminlnei in 
the neater, they wonld donbtleis have the peculiar aenter endings in 
nom.-acc.-Toc. of all nnmbera. 

d. Bat IS. hw once tr&tpitfinui, uom. pi., having thee for fathtr. 

Declantion V. 

stems ending in Consonants. 

877. All stems ending in consonants may properly be 
classed togethei, as foiming 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. 

878. In this declension, masculines and feminines of 
the same final are inflected alike; and neuters are peculiar 
(as usually in the other declensions) only in the nom.-aco.- 
Too. of all numbers. 

a. The majority of consonantal stems, however, are not 
inflected in the feminine, but form a special feminine deriv- 
ative stem in 1 1 (never in ^ &), by adding that ending to 
the weak form of the masculine. 

b. Exceptions are Id general the items of divisions A and B — 
namely, the radical stems etc., and those in ae and la and na. For 
special eaeea, see below. 

879. Variations, as between stronger and weaker forms, 
are very general among consonantal stems: either of two 
degrees (strong and weak), ot of three (strong, middle, and 
weakest): see above, 811. 

a. The peculiar neuter forms, according to the usual 
rule (811 b), are made in the plural &om the strong stem, in 
singular and dual from the weak — or, when the gradation 
is threefold, in singular from the middle stem, in dual &om 
the weakest. 

ioy Google 

S79 — ] Y' Nouns and Adjectives. 142 

b. Ab in the case of Bt«ms ending in short Yowele [fiB;&ni, 
vartfi, midhOol, dS^^ etc-), a nasal sonetimeB appears in the 
special aenter plaral cases vhicb is fonnd nowhere else in inflectioD. 
Thus, from the stems in as. io, ub, the nom.-acc.-Too. pi. in -SAsl, 
-lA^i, -ilA^ are very ootnmon at every period. According to the 
grammarians, the radical stems etc. (division A] are treated in the 
same way; bnt essmples of snch nenters are of extreme rarity in the 
language ; no Vedic text offers one, and in the Brahmsnas and Sutras 
have been noted only -hunti (AB. vii. 2. 3), -vpia (PB. ivi. 2. 7 et al.), 
-bhaiiji (KB. xxvii. 7), -bhfntl (QB. viii. 1. 3*), and •yuBJi (H!3. ii. 1.8); 
while in the later langnage is found here and there a case, like 
•qruntl (Rsgh.), •piiA;! i^\q.) ; it may be questioned whether they are 
not later analogical formations. 

980. The endings are throughout those given above (310) 
as &e "noimal". 

a. By the general law as to finals (160), the e of &e aom. sing, 
masc. and fern, is always lost; and inegnlarities of treatment of the 
final of the stem in this case are not infrequent. 

b. The gen. and abl. eiog. 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 monosyllabic 
Btems and the participles in itat (accented on the final). For details, 
see below, under divieions A and B. 

a. Bat > tew ot the componndB of the root ailo ox ao ehow an inegnlu 
shift of sGceut in the oldest Iingiuge: aee below, 410. 

882. a. Foi convenience and clearness of presentation, 
it will be well to sepaiate from the general mass of conson- 
antal stems certain special classes which show kindred pe- 
culiaiitiee of inflection, and may be beet described together. 

B. Derivative stems in as, is, ub; 

C. Derivative stems in an (an, man, van); 

D. Derivative stems in in (in, min, vin); 

E. Derivative stems in ant [ant, mact, vant); 

F. Perfect active participles in v&fis; 
Q. Comparatives in ySbs or yas. 

b. There remain, then, to constitute division A, espe- 
cially radical stems, or those identical in form with roots, 

ioy Google 

143 Declembion v., Consonantal Root-stems. [— S8S 

together with a comparatively small numbet of others which 
aie inflected like these. 

The7 will be takeD up in tbe order thiiB indicated. 

A. Boot>Btema, and those Inflected like them. 
388. The stems of this division may be classified a« 
follows : 

I. a. Root-stemB, having in them no demoustrftble elemeut added 
to a root: Uidb, fo verte, gfr tong, p&d /oot, (Uq direction, taiii (V.) 

b. Sueb itami, however, u« Dot alirBys piectsely identtoKl la foim 
with the loot: that, vto from yVao, nii (lom j/sfj, md^ from ynm^, 
vrif from yvTtUj<s(l'), i^ from i^Tas ihirie; — tioin rooU in On&l f come 
■teou in ir and ur: thm, g{r, K-^, atir; J^, ti^, dta4r, pur, mur, 
stdr, sph&r; >Dd pmia from j^psar. 

o. With Ibeae may ba raolced the stems with reduplicated root, as 
oildt, yaviyddb, v&nivan, saay&d. 

d. Words of this division la Qncompounded atg are toisrtbly beqneut 
In the older laDgaage: tbos, in BV. ue found more than ahnndred of them; 
in AT., alioat ality; bnt In the cluslcat Ssnakrlt tbe power of nslng any 
root at will In this way is loit, and the examples are comparatively Taw. 
In all peilodj, bo'wever, the adjective tue as floal of ■ compound is very 
commoD (tee below, 401]. 

e. As to the InflnitiTe nse of varions cases of the root-nonn, see 9TI. 
U. f. Stems made by tbe addition of t to a final sbort vowel of 

a toot 

K. No proper root-stem ends in a short vowel, slthongh there are (364) 
examples of transfer of sach to Hhort-vcirel-decleDeiansi but 1 or n or f 
adds a t to make a declinable form: thni, -Jft, -qrut, -k^. Roots in f, 
hovrever, as has jnst been seen [b], also make itemt in Ir or iir. 

II. As rtgaidl the frequency and nse of these words, the tame is true 
as was stated above respecting root-stems. The Veda offers examples of 
nesily thirty such formations, a few of them (mft, At, stut, brfit, Trft, 
and dyiit If this is taken from d7u] in independent use. Of roots in J, 
t Is added by kf, Hhf, dbTf, bhr, vr, a;, apf, bj, and hvj. The roots 
gK (or gam) and ban also make -g&t and -tUit by addition of the t to 
an abbreviated form in a (thn«, adtavag&t, dTUg&t, dvlgat, navag&t, 
and aaihhit). 

III. 1. HoDOnyllabic (aJBo a few apparently reduplicated) atema 
not certainly connectible with any verb^ root tn tbe language, but 
baviog tbe aspect of root-stenu, as containing no traceable suffix 

ioy Google 

888—] v. Nouns and AwBCnvEa. 144 

thas, tvke tkin, p&th road, hfd )ieart, ftp and vir water, dvix door, 
is mouOt, kaki^bh and kakud ntmmit. 

J. Tbirty 01 forty sneli voiit ue foand in the older ItDgniga, uid 
some of tliem contlnae In Utei oeo, while others btve been tnmferred to 
other modes of declenston or have become extlnet. 

b. Stems more or leas dearly derivative, lint made vitb enf^zes 
of rare or even ieolated occnnence. Thus: 

1. dertTatWei (Y.] tiom prepoeltioni with the Entflz vat: arrsvit, 
&T&t, udv&t. niT&t, parSv&t, praw&t, aadiTit; — 2. derlTftttves (T.) 
In tKt (periupB abbTSTiated ftom tfttl], In a few Ieolated forms: thiu, 
apar&t&t, dev&tat, vr^tSt, aaty&t&t, Barv&tfit; — 3. other derive- 
Utbi In t preceded by vutoag Towele : thug, da^&t, Teh&t, vabit, arav&t, 
aagcit, vfteh&t; n&pat; ta^t, dlvit, yoqft, Tohit, aarft, harlt; 
mar&t; siikft, fUcft; and the numerals for 30, 40, GO, trUif&t etc. 
(475); — ! stem* In ad: thus, dn&d, dlqrf^ bhao&d. van&d, 
far&d, aam&d; — G. items in J preceded by vsrloaa vowels: thus, tf^^ij, 
dhnU, Mo&i, bhi9&j; 119O, va^ij, bhurij, ni^iJC?): Aa^J; — 6. a 
few stems ending in a slblUnt spparenti; formative: thns, Jfiila, •dfia, 
Vbta, voia, bill?; — 7. ■ remnant of nnduslflable oases, each aa vi;t&p, 
vip&q, k&pith. <;nTudh. If Idb, ppkf udta, ragh&t [?), aari^li, viaruta. 

Tiai^lh , kav&fli 

384. en dec. The root-stems are legnlarly feminine as nomen 
aelionu, and maBcnUne as nomen agmtit (which ia probably only a 
snbstsntive use of tlieir adjective value: below, 400). Bnt the femi- 
nine noan, without ohangiDg its gender, is often also nsed concretely: 
e. g., dnili f. [yAraSy be inimkal) means harming, enmity, and also 
harmer, hater, enemy — tbaa bordering on the mascnilne value. And 
some of the feminines have a completely concrete meaning. Throngh 
the whole division, the maacnlines are mnch leas nnmerona than the 
feminines, and the neotere rarest of all. 

a. The independent neater stems are hfd (alao -hard), dim, var, 
B7ar, mts fiuh, ta mouth, bhiia, doa (with which may le meiitioned 
the indecllnablea g&m and 76b); also the apparent derlvatiTea y&fert, 
Q&krt, k&prth, &8|j. 

386. Strong and weak stem-foims. The distinotioii 
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, a^ long or short; less often, 
by other methods. 

386. A nasal app«arB in the strong cnsea of the following words: 

1. Componnda baring aa final member the root ao or a&o: see below, 

407 tr.; and RV, baa once uxuvy&fioam from root vyao; — 2. The 

ioy Google 

145 Declehbioh V., Consonantal Stems. [ — 889 

atem ynj, lometlmei, In the older Unguaga: tbnt, nom. ring. ynH (for 
T^Ak), accaa. 3rufijam,*da. yitLi& (but ilao y^am and TiiJS); — 
3. The stem -drq, u final of a compound in the oldei language; "bat only 
in the nom. ling, maae,, and not alvaya: tbiu, onyftdtS, Idftl, kid^ 
tsd^n, etftdrn, ead^ and pratlssd^: but alio idfk. t&d^k, Bvardtk, 
etc.; — 4. For path tnd pudis, which labBtltnte moie eileDded Btemi, 
tnd foT dant, lee beloir, 394 — 0. 

887. The vovel a it lengthened in strong ouea as follows: 
1. or the loots vao, eao, aap, nabh, gaa, la a few Inataneea (V,), 
St the end of componnds; — 2, Of the loota vKb uid Bah, bat IrregoUtly; 
■ee below, 408 — B; — 3. Of ap water (tee 893); also In Ita oompound 
rityiip; — 4. Ot pAi foot: In the eomponnde of thla word. In the Utet 
langDige, the aame lengtbenlDg li made tn the middle caiea alto; and in 
RV. and AV. the nom. elng. neat. le both -pat and •pfit, while RY. baa 
once -pftde, and pftdbhia and p&tan oocnr In the BrahrnKQas; — 5. Of 
nas note (? nitaft nam. da. fem., RT., once); — 6. Sporadic esaei (V.) 
are: yJU (!], toc. aing. ; pSth&s and -rSpao, bccdb. pi.; T&nlT&naa, 
nom. pi. The atrengthened fornii t>h^ and rfij are conatant, through all 
clasgee of oasee, 

388. Other modes of differentiation, by elision of a or contraction 
of the syllftble containing it, appear in a few stems: 

t. In •ban: eee below, 402; — 2. In kqam (v.], along with pro* 
longatlon of a: thua, kfllmS du., k;&naapl.; k;ain£ tnatr. sing., kf&mi 
loe. slug., kfin&B abL sing.; — 3. In dvar, contncted (Y.) to dur in weak 
caaee (but with some canfoslon of the two clacaee); — 4. In mkr, which 
become!, in KY., sQr In weak cases; latei it la indedioable. 

889. The endings aie as stated above (380). 

a. Bespeoting their combination with the final of the stem, as 
well as the treatment of the latter when it oooutb at the end of the 
word, the rolea of enphonic combination (chap. IIL) are to be oon- 
anlted; they require mnch more constant and various application here 
than anywhere else in declension, 

b. Attention may be called to a tew exceptional eases of combinalloD 
(Y.): mSdbbia and m&dbhy&s from vaia numih; the wholly anomalous 
pa^bhia (RY. and VS.J AY. has always padbh{B) from p&d; and aar&t 
and ear&4'>'VB oonespondlng to a nom. pi. 8ar6ghaB (instead of Bttr&bas: 
SaS). D&n Is appatenU; for d&m, b; 148a. 

O. According to the grammarlana, nenter stems, nnleaa they end In a 
nasal oi a aemlvowel, take In nom.-acc.-ioc. pi. a atieagtheolng naaal before 
the final consonant Bat no such cases from nsntei nenn-etema appeal eiet 
to haTe been met with In nse; and as regards adjective stems ending in a 
root, see abOTC, 379 b. 

WhilaeTi Onmrnai. 3. ed. 10 

ioy Google 

890—] V, IfoDNS AND Adjectives. 146 

890. MonoByllabic atenis have the regular accent of inch, Haoir- 
iDg the tone forward npon the endings in t^e weak casee. 

a. Bnt the acoasatiTe plnral has its normal acoenttution as & 
weak case, npon the ending, in only a minority (hardly more than a 
third) of the sterna; namely in dat&a, path&s, pad&a, nld&s, ap&a, 
of&B, jaSs&B, pn^B&a, mas&s, mahiuii and sometimes In v&oAs, 
snio&s, hrut&n, srldh&s, k^pia, vip&s, dur&a, i^ka, dvlf &a, dnih&a 
[beside v^as etc.). 

b. ExcsptloDal InBbmou, In vhiob a weak caw bat the tons on the 
Btem, oceni as tallows: s&dS, n&dbhyaB, t&nit (also tan£) and t&ne, 
b^dha (infln.), ri^e and r&fisu, viimn, arinl, vfpoa, kf&ml, adrt 
and sAraa (^bnt m5x6), kbhaa, and T&nas and b^has (in vinasp&tl, 
bfhasp&tl]. On the otbei hand, a strong case Is aecented on the ending 
in mali&s, nam. pi., and kfts&m (AV.: perhaps a falee reading). And 
pre;^ Inati. aing., Is accented as if pr^ weie a simple stem, Instead of 
pro-ff . VimrdbA^ Is of donbttnl character. For the lometlmeB anomalons 
a'scentnitlon of stems In ao or a&o, eee 410. 

301. Examples «£ inflsxioo. As aa example of 
QOimal monosyllabio inflection, we may take the stem 
^R ■^o f. voice (liom y^ vao, with constant prolongation] : 
of inflection with strong and weak stem, ^p&d m.foot. 
of polysyllabio inflection, R^ mamit m. xoind tit vsind'-god; 
of a monosyllabio root-stem in composition, Bl^a trivft 
three-fold, in the neuter. Thus: 


















































Dij.ieo, Google 

Declension T,, Consonantal Steus. 















marudbh;&m triTfdbbySm 




















vSoAa, vaoas 





















H t>ril^^ 












B; wAy of illustration of the leading methods of treatment of 
ft Btem-final, at the end of the word and in oombiDation with case- 
eodingg, characteristic Dase-forms of a few more stems are here added. 

a. Stenu in j: ynJ-clasB (210a, 142], bhl^^ pht/sieian: bhiq&k, 
blilq&Jaia, blii^cbhls, bMi!4k9u; — mrJ-claBs |2ieb, 143), samr^ 
univerttd ruhr: Bamraf, eamrl^am, Bamra^bhiBi samratsu. 

b. Stems in Hh: -Tfdb mereating: -vft, -TfdluuD, -ytdbhis. 
-trftBa; -b^dh (16B) waking: 'bhdt, 'bi^dham, -bhtidbMB, -bhiitBU. 

o. Stems in bh: -atubh pramng: -atup, -stubham, -st^bbhiB, 

d. St«ms in q: dfq (216 a, 14B| direction: dik, dlgam, dlgbhfa, 
dik^A; — vlq (818, 146) Me peoph: vif, vi<}taa, vl^bhle, vifeu (V. 
vlb^u: 218 a). 

e. Stems In f (226 b, 14S): &vi^ mtmy. Avi%, dvifam, dTl^bhls, 

t. Stems in h: doh-class (a32-3a, I&6b, 147). -d^ rnOking, 

D,j,i,...., Google 

391—] V. Nouns and Adjectives. 148 

yielding: -dhi^ -ddbam, -dbugbhia, -dbukfn; — roh-olaM (823b, 
147), -lili liekmg: -Ut, -liham, -U^bhla, .li(au. 

g. Stems in m (143 a, S12a: only pra^iia, aom. «ing., Qaot&ble): 
-q&m quieting: -{ibi, •q^mam, -fltabhlB, -qibiBn. 

393. Tbe root-Btema Id Ir and nr (388 b] lenfrthen their vovel 
when the final t is followed by another oonsonant (246 b], and also 
in tbe nom. sing, {where the cue-SDding a is lost). 

a. ThDB, from g{r f. tong cone gtr [gih], gfram, gtrt etc.; 
gfrSo, girbhyam, K(r6B; giras, girbbfe, sirbhy&B, Eivitm, eSrfa 
(166); and, in like manner, from pur f. stronghold come pdr (plih), 
p&ram, por^ etc. ; pur&n, pilrbhyim, pur6B ; pAraa, pQf bhla, pOr- 
bby&B, puram, pQr^ix. 

b. There ace no looM In is (except the eiceBBWal; nre pla) ot In 
OB; bat from the root qUa with Ita & weakened to 1 (S&O) comeg the 
oonn S4jiB t. bleeiing, which is Inflected like g£r: thne, Bqla (Mtflb^, 
B^am, ifi^ etc; ftflfSo, ftqirbliySin, iqi^oa; ft^tfAB, fi^irbhia, 
aqirbhyaa, S^Cf&m, Rfi^fO. And sajda together \t appaieutl; a steieo- 
typed DomlnittTa of like formation from the root juf. The form affapruf 
(TS.), from the root-etem pruf, is Isaltted aod inomalons. 

O. These stems in Ir, ur, la show a like prolongation of Towel alao 
in composition and deiiTation: thna, glrrftifa, pQrbllfd, dhargsta, 
dbOatTO, fiflrda, (Iffrvaiit, eto. (bnt also gfrran, glnra^aa). 

d. The native grammar lets op a clasa of qnasl-ndical stems ' like 
JlgamiB detirirtg to go, made trom the deslderatlTe conjngatiOD-stem (1027), 
and pceseiibes for it a declension like that of iijia: thna, Jigamla, Jigo* 
miq&r Jlgamirbhis, Jlgaml^^n, etc. Snch a class appears to be a mere 
flgmsnt of the grammarians, since no example of it has heen found quotable 
from the literature, either euUet or liter, and since there Is, In fact, no 
more a desideratife alem Jigamle than a cansatlTe slem gama;. 

3B3. The stem kp t. uater ia inflected only in the plural, and 
with disBimilation of ita final before bh to d (161 e): thus, Apae, 
ap&a, adbhfs, adbhy&a, apam, a,paii. 

a. Bnt KV. has the sing, luslr. apli and gen. ap&s. In the earlier 
langQBge (especially AT.], and eren in the epics, the nom. and accns. pi. 
forms are occaslonilly confosed in use, Spaa being employed as accns., 

b. Betides the stem ap, case-forms of thla word are sometimei used 
in composition and derivation : thnE, for example, abja, ftpodevata, 
ftpom&yo, apBtunaot. 

394. The stem puiha m. man is very irregular, snbstitating 
pianidiB in the strong cases, and losing its b (necesearily) before 
initial bh of a e&Be'etiding, and likewise [by analogy with this, or 
by an abbreviatioD akin with that noticed at 231) in the loc. plural. 
The vocative is (In accordance with that of the somewhat similarly 

ioy Google 

149 Dbclbssioh T., Cossonantal Stemh. [—898 

inflected perfect partloiples: Bee4es&) pi^nuui in the later lui^age, 
bat puni&a id the earlier. Thua: picim&n, pdmsAsEun, pmhao, 
pTufasd, pu^u&a, puAsf, pAmau; pumftAeSu, pumbhyun, poifasdB; 
piiun&fiaaa, poifaeie, pumbbb, pvunbhy&s, paihsaiii, puihBu. 

a. Tha accentsitioD ol the -weak fomii. It vrlll be uottced, is that of 
ft tnie moDDfyllibic item. The loitOB vith bb-endlnga nawhete oceoi lu Ibe 
older language, not do the; appeal to have been cited from the later. 
iDEtaaces ot the confnslon of itrong and weak farms are occasionally met 
irltb. As to the retentton of a anUDgaillzed in the weakest cases (whence 
necessarily toUowi th»t In the loe. pi.). BBe 183 a. 

b. This stem appears undei • conaidaiablB Tarlety of forms in com- 
position and detiTatlon; thus, as patbs in podiqoalf, podietva, pudia- 
vant, -paiheka, etc.; u pmn Id puihTataa, pnibrQpa, podiTat, 
pnmairtha, etc.; as podisa in piufaeaTant; — at the end of a coispoDnd, 
either irilh Its full Inflection, ta In strlpAifaB etc. ; or aa purhsa, In 
stripmhaa, mabSptufaBa; or u puma in atnpoma (TS. TA.). 

S0E. The stem patb m. road is defeotive in declension, forming 
only the weakest cases, while the strong aie made from p&utbs or 
pinthan, and the middle from patbi: see nnder an-atems, below, 433. 

396. The stem d&nt m. tooth is perhaps of participial origin, and 
has, like a participle, the forms diiiit and dkt, strong and weak: 
thus (V.), dka, d&ntam, dat£, etc.; dat&« ace. pi. etc. But in tbe 
middle cases it has the monosyllabic and oot the participial accent: 
thns, dadbbJis, dadbhy&a. la aom. pi. occurs also -datae instead 
of -dantae. By the grammarians, the strong cases of this word are 
required to be made from d&nta. 

897. A number of other words of this divisioti are defective, 
making part of their inflection &om stems of a different form. 

a. Thna, b^ heart, m^Aa or mas n. meat, mas m. mtmlh, n&s 
f. note, ni^ t. night (not fonnd in the oldei language), pft f. army, are 
Mid br the grammarians to lack the nom. of all numbers and tha accns. 
sing, and du. (the nentera, of conise, the ace. pi. also), making them 
lespectiTBlr from hfdaya, milfia&, mitsa, naslks, nlqS, pftanfi. But 
the usage in the older langnage U not entirely in accordance with thla 
requirement: thna, we find m^ _fieah aocng. alng. ; mia month nom. sing.; 
and uitsft nottrih dn. From p^ occnta only the loc. pi. p^u and (RY., 
once] the same case with doable ending, pftaufu. 

398. On the other hand, certain stems of this division, allowed 
by the grammarians a fnli inflectioD, are used to fill up the defioien- 
cles of those of another form. 

a. Thos, Asfj D. hlood, g&krt n. ordure, yUcft n. liver, d6s n. 
(also m.) fore-artn, have beside them defective stems in &ii: see below, 
488. Of none of them, howoTer, la anything but the nam.-acc. sing, found 
in tha older langnage, and other case* later are but very scantily represented. 

ioy Google 

89S— ] V. Noons asd Adjetites, 150 

b. Of Ss n. mouth, iDil i^d uialer, anl; s ewe 01 two ue fannd, !□ 
thn older langntee, bealde fis&n and aajr^ »d ud&u and ndakft (48S). 

306. Some ot the illeniatiTe Btem* menttoned tbo^e are IntUnces of 
trtngllion from ^^^e eoDeoDant to a Toirel deolenilon: Ihne, dAnta, mjsa. 
A Dumtier of other similar cues oconr, epoiadlMllr In the ald«i laofnage, 
mote commonly in tbe Uter. Sach are -pi>d», -ni&da, -dfifo, blir^&, 
vi^t^pa, dvfira aad dura, pura, dbnra, -df^a, iuia&, nidK, kflpft, 
k^pll, Kf a, and perhaps a few others. 

a. A tev Inegalat atemi will find a more proper place imder the head 
of Adjectives. 


400. Original adjectivea having the root-form are comparatively 

rare even in the oldest langaage. 

a. Ahont a doaen are quotable from the RV., for the most pari ouly 
Id a few scattering cases. Btit mah great is common in RT., though It 
dies ont rapidly later. It makes a derivative feminine item, malli, vUch 
contlnaee in use, t» meaning ettrfh etc. 

401. But oomponnd adjectivee, hsring a root as final member, 
with the value of a present participle, are abnndaDt fn every period 
of the language. 

a. Posaeaaife adjective compoanda, also, of the same form, are 
Dot very rare: ezamplea are yat&aruo uiith offered hotel; adryatrao 
gun-tkiimed; o&toqpad foar-foated-, subard kind-hearted, friendly; 
ntyap (i. e. ritf-ap] having etreaminff lealert; Bah&sradv&r ^/tfrnisA^iJ 
with a ihotuand doore. 

b. The inflection of sach compounds is like that of the simple roet- 
gtems, mascnllne and feminine being throughont the same, and the nenter 
varjlng only in the nom.-aoc.-voc. of all uombers. Bat special neater formt 
are of rare occarrenee, and maac.-fem. are sometimes nsed instead. 

O. Only rarelji is a dertvatlTe feminine stem la I formed: in the older 
language, onl; from tbe compounds vith ao or ajle (407 ff.), those -with 
ban (402), those vith pad, as 4kapadl, dvipAdl, and with dant, as 
vt^adaU, and mahl. kniMSH (AY.), upaaadi (?9B). 

IrregnlaritieB of inflection appear in tho following: 

402. The root ban ilay, aa final of a compound, is infleotei 
somewhat like a derivative aonn in aa (below, 480 ff.|, becoming bft 
in the nom. sing., and losing its n in the middle casea and its a la 
the weakest caaea bnt only optionally Id the loc. sing.). Farther, when 
the vowel ia loat, b in contact with following n reverie to its orig- 
inal gh, Thoa: 

ioy Google 







vrtrahibhyilm 1 ^^ahAbhyM 

« oeaal, the Btem-fonn ■hoirn 

N. vftraliM 

A. Tftrali&9am 

I. vftraghna 

D. TTtraiglmA 

Q ■ VvTtraghnia 

V, TJlraliKD 

a. Ab to the ehtngB of n to 9, aee 1' 

b. A famlDine U made ij adding I ti 
In the westeit Mies: thus, Trtrastmi. 

0. An ucni. pi. -biinaB (Uke the uom.) kIio occqib. V}^nih&bhia 
(RT., ono«) it the only middle esBe-tocm quotable from the older langnage. 
TranBltloae to the a-deeUnaion begin already in the Teda: thua, to -h& 
(EV. AV.), •glini (RV.), -hftna. 

403. The root vah carry at the end of a componod is said by 
the grammuians to be lengtbened to v&h in both the Btrong and 
middle cases, and contracted in the we&kest caeea to ah, which with 
a preceding a-TOwel becomea Su(lS7o]: thus, from havyavili laeri- 
JUfiaaring [epithet of Agni), havTftTBt> baTyavabom, ba-vyft^S, 
etc.; bavyavibftu, haTyava^bbySm, bavySuhos; bavyavabas, 
bavySohaB, Jtavydvi^bblA, etc. And qvetav&b [not quotable) is 
said to be farther irregular in mailing the nom. eing. in t&o and the 
vocative in vaa or tBs. 

a. In the earlier langaage, only Btiong fonua of eompounda with vab 
have been fonnd to occur: namely, -v^f, -vabam. -Tabfia or -vabB, and 
-V&as. But feminines In 1, from the weafceat atem — aa tury&ubi, 
dity&uhi, paf^fiobi — are met with in the Brihmiijas. TS. haa the 
inegakr nom. iiag. paffbavat. 

404. Of very irregnlar faraiatioD and infleotioD is one common 
compound of irab, namely ana^v&b (anaa-l-vah burden-bearing or 
eart-drawing, i. e. ox). Its stem-form in the strong cases is ana^^ah, 
in the weakest ana^uli, and in the middle aua^ild (perhaps by dis- 
similation from ana^u^'- Moreover, its nom. and voc. sing, are made 
in T&n and van (as if from a Taut-stem). Thus: 



Dual. Plural. 


}•»•*'*- nis." 


. ana^uhe 

)»•*'-" z^ 



toaijvahftu ina^vaiiaa 

i, Google 

4o4— ] V. Noons asd Awectivbs. 152 

a. Ana^udbbyM (A.T., once) It the only mtddle cue-fotm qaotable 
from the older language. But compatiQdi ahoiriDg the middle item — u 
aoo^i'^B'i'^^ ana^ndarhft — ^re met with In B^manu etc. 

b. The GOiteapondtng feminine stem (of very infreqnent occanenee) 
U eitbei noa^uhi ({B.) or ana^vUU (K. HS.). 

406. The root sah overcome haa In the Veda » double liregalaritr: 
its B is changeable to |f even after au a-Towel — u alao in ita tingle oc- 
currence as an Independent adjective (RV., tv&di ^f) — while it toma- 
times remaina nnehanged after an i oru-TOwel; and its a ia either prolonged 
or remains unchanged, in both strong and weak cMes. The quotable formt 
are : -^i\, -f abam or -Baham or -s&bain, -aUtS, -situ or -oAhs, -^aliaa 
or -9AI1IUS 01 -B&haB; -s&lia (du.); -;ahas or -sUuw. 

406. The compound avay^ (v'yJ mafce offtring) a certain priat or 
(BR.) a certain tacrifice is said to fonn the nom. and toc slug, avayaa, 
and to make its middle casei froui avay&a. 

a. Itt only quotable form is avayaB, f. (RV. aad AV., each once). 
If the stem is a deriTative from ava + i/yaj coneQiaU, avayas U very 
probably from ava-HF'y^ which haa the aama meaning. But aadbamas 
(BV., once) and pttrodee (RV. twice) thow a simiLai apparent anbatitution 
in nom. aing. of the caae-ending a after long & for a final root-oonaonant 
(d and f teepectiiely). Compare also the alleged qvetavfis (above, 40S). 

407. Compounds with alLc or ao. The root ao or alio 
makes, in combination with prepositiona and other words, s consid- 
erable class of familiarly nsed adJectlTes, of quite irregular formation 
and inflection, in some of which it almott loses its obanicter of root, 
and becomes an ending of derivation. 

a. A part of these adjectives hare only two stem-forms: a strong 
in afic (yielding aa, from afiks, in nom. sing, maao.), and a weak in 
ao; others diatinguiah from the middle in ao a weakest stem in o, 
before which the a is contracted with a preceding i or a into 1 or tL 

b. The feminine is made by adding 1 to the stem-form used in 
the weakest cases, and ia accented like them. 

408. Aa esamples of inflection we may take prftilo fonnard, «a*t, 
praty&&o oppottte, vest, vi^vafic going apari. 

Singnlar : 

N. V. pr^ pr&k praty^ pra^&k vl^aa vffvak 

A. priUioam prak pratyificam praty&k vi^aftoam vf^vak 

I. prac& pratioi vifucfi 

D. praoe pratioi vifuoe 

Ab. a. praoas pratlo&s vf^ilcBA 

L. pr^oi pratioi v{faoi 


N. A. V. pi^cfiu praoi ptaty&ilaftii pTatiof v{;Tafioftu vi^uol 
I. D. Ab. pragbliy&in praty^bhy&m v{|vagbhyltm 

O. L. pracos pratioAs vi;uoos 

Digitizecy Google 

153 DECI.EKSION v., C0N8OSANTAL Steus. [ — 4t3 


N, v. pritfioaa prafioi pra^Altcaa praty&fici vi^vahoaa vi^va&ei 
A. prioaa pra&ol pratio&a prftty&fiol vffaa&B vl^vofiai 

L pr^bhiB praty&gbhia viifvaqbbia 

D. Ab. pritsbhyas pratr^bhyas vi^TOgbhyae 

Q. prjioftin pratleam vi;uo&m 

L. pritkqu pratyik^u vl^Tabqn 

a. The femiaiae stems are pi^i, pratioi, vi^i<a, raspectiTel}'. 

b. No eiBjnple of the middle tormi excepting the nom. etc. siog. 
DiDt. (and tMi genenllj ojed >■ adverb) le found eitfaei In RV. 01 AV. 
In the tuae textt ii lacking the nom.' etc. pi. nant. in fiol; bnt o( this a 
number of examples occnt in the Brahmanu: thus, pra&ol, praty&Siol, 
KTvB&el, Bamy&fiol, nacUuTa&oi, anva&oi. 

400. a. Like pc^o are inflected &pfifio, kvSho, p^ftiio, arvafic, 
adliarfific, and othors of rare occnrrence. 

b. Like proty&flo are inflected nykfio (i. e. nlaiio), samyi&o 
{aam + afio, with irregularly inserted li, and udafio (weakest stem 
^dio: ud+a2o, with i inserted in weakest cases onlj), with a few 
other rare etema. 

o. Like vi^tt&c is inflected anv&fio, also three or foar others of 
which only isolated forms occur. 

d. Still more irregular is tiry&iLo, of which the weakest stem is 
tir&90 (tlr&s+tM: the other stems are made from tlr + afio or ao, 
with the InseiCed 1). 

410. The acnentnaUon of thete words is Inegolar, u regard* both 
the stemi themseWes and their inflected foims. Sometimes the one element 
has the tone and tometimea the other, wlthoat an; apparent reason for the 
difference. If the oompoand i« accented on the flnil efllaMe, the accent 
Is shitted in BV, to the ending in the weiliest cases provided their stem 
•bows the contraction to I or il: thoa, priEoft, arvaoft, adhorioas, bat 
pratloli. auao&B, samlci. Bat AT. and later texts nsnillr keep the 
accent apon the stem: thne, pratloT, aamloi, anAcn (BV. hag pratiolm 
once). The shift of accent to the endings, and even in polysyllabic stems, 
li against ail nenal analogy. 

B. DerivatiTe sterna in as, 1b, us. 

411. The stems of this division are pievailingly neuter; 
but there ate also a few masculines, and one or two 

412. The stems in ^(^ as aie quite numerous, and 
mostly made with the suffix ^ as (a small number also 

ioy Google 

412—] Nouns and Adjectites. 154 

with fITT tas and RTTnaB, and some aie obscuie); the others 
ate few, and almost all made with the suffixes ^ is aod 
3?r OS. 

418. Their inflection is almost entirely regular. But 
masculine and feminine stems in 3^ as lengthen the vowel 
of the aiding in nom. sing. ; and the nom.-aco.-TO0. pi. neut 
make the same prolongation (of 9 a or ^ i or 3 u) before 
the inserted nasal (anasTftraj. 

414. Examples of declension. As examples we 
may take WJ^ m^as n. mind; ^T^J^&AgiiaB m. Angiras; 
*f^ bans n. oblation. 











































mfaiaal AfigiroBila bavfyl 

m&uobhyfim &&glrobliyfini bavfrbhyKm 
m&aaaoB AfiglrasoB haviqoa 

ioy Google 

Declension Y., Stems ik aa, is, i 

M. A. V 














D. Ab. 





















In like 

manner, g^H 

odlcfua n. eye 

forms g^TT odhjuflB 

^Tm^odkeurbhySm, 'g^f^ odk^ttjigl, 

and 80 on. 

415. Tedlc etc. IiisgnUiltiei. a. In the oldei langn&ge, tbe 
ending* -aaam C^or. Ring.) and -aaaa (genenlly nom.-ice. pi. ; ODce at 
twice gen.'abl. ting.) of stems la as ire not Inftequentl; conCiacted to -Sm, 
-&B — e. g. ft^am, TOdbam; eur&dbfts, ^nOgfts — >nd out of each forma 
grow, both aiiUer >nd Uter, tabetltDte-itemB In fi, as &<}&, jar£, medbi. 
So from otbai forms grow etems in a tad In asa, -which exchange more or 
leas with thoae in aa thiongh the whole history of the Unguage. 

b. MoTo scatt«ting tnegnlaritlea may be mentioned, aa follows: 1. Tha 
luaal maac. and Cam. da. ending In & lnst«td of ftu; — 2. af&a f. daten 
often prolonga ita a In tha other atrong euee, as In the Bom. aing.: tbna, 
Tifaaain, afltafi, nf^Laaa (and once In a weak case, ufdsaa); and In Ita 
biatr. pi. ocean once (RV.) Uf&dbbla Instead of u^bhlB^ — 3. from 
to^&a la onoe (BV.) foand a elmllar dost, Uxftei; — 4. from avdvaa 
aitd Bvitavas occnr in RV. a nom. ring. maac. in T&n, aa If from a stem 
In rant; and In the Brahmanaa ia ,foaud the dat.-abl. pi. of like foimatloD 

C. The sttms In ia and ub also ehow transitions to alemi la 1 and 
M, and In Ifa and Ofa. From Jan^ la once (KT.) made the nom. ting, 
Jaada, after the manner of an as-stem (of. al«o Jan&rvasaa (B.). 

416. The grammariana regard ug&naa m. aa regular stem-form of the 
proper name aotlced abore (3B6 a), bat give it the iripgnlar nom, U9&nft 
and the toc ufanoa or uqanA or ufanan. Forms from the as-atem, 
even nom., are aometlmeB met vith In the later literature. 

a. As to forms from aa-atemi to &bail or Uiar and ddban or tidhar, 
see below, 480. 

ioy Google 


4t7. a. A few neuter nonnB io ae witb aoceot on the udioal 
Billable have corresponding adjectiveii or appellatives In &a, with 
accent on the eadiog: thus, for example, -&paa uorA, ap&s active; 
t&raa quichntu, tar&B quick; jiqaa glory, yK9&a giorimu. A few 
other similar adjectives — as tav&s mighty, vedb&s pioaa ~ are without 
corresponding nonoa. 

b. Odginal adjectives in is do not occur (as to alleged desider- 
atire adjectives in ia, see 398 d). Bat In na are fonod as man; ad- 
jectives as nouns (about ten of each clasa); and in several instances 
adjective and nonn stand side by side, without difference of accent 
such as appears in the stems in an: e. g. tipus heat and hot; T&poa 
wonder and leondtrful 

418. Adjective compounds having nouns of this division as final 
member are very common: thoa, BunUnafl favorably minded; dlrgli- 
iyuB long-lived; (iikr&foolB having brtlliaui brightnest. The etem- 
form is the same for all genders, and each gender is inflected in the 
usual manner, the stems in (m making their nom. sing. maso. and 
fem. in fia (like iSigLroB, above). Thus, from Bum&caa, the nom. 
and accus. are as follows: 

Slogulu. Doil. Pland. 

rn-Ln. m-f. ii.m,t n. 

N. Butn&nSa 

^^^ > Bom&naBSu -naal anm&naaas -nftAsl 

and the other cases (save the vocative] are alike is all genders. 

a. In Veda and Biilhmana, the neat. nom. sing, ii in a cDDSlderabte 
namber ot instuicGs made in Ss, like the otfaei genders. 

b. From dirgbayua, in like muinei: 

N. dlrfcbiytiB \ _, . 

A. dirftbSjnigam -j^f ^Bl'«^9«" -r^9> dftgliayn,aa -y&4^ 

L dlrgbayu^ft dirgbaynrbhySm dlrgb^Torblila 

416. The stem aneb&s unrivalled (defined ib meaning time in the 
later langnage) fonna the nom. a'lng. muc. and fem. aueluL 

C. BeriTfttdTO stems in an. 

420. The stemjB of this divieioa are those made by the 

three suffixes ^ an, IH man, and ^ van, together with 

a tew of more questionable etymology which are inflected 

like them. They are almost exclusively masculine and 

421, The stem has a triple form. In the strong oases 

itizecy Google 

157 Dbclbnsioh v., Stems m an. [—484 

of the maBCuline, the vowel of the ending is prolonged to 
3^ ft; in the weakest oases it is in general struck out 
altogether; in the middle cases, oi before a case-eading 
beginoing with a consonant, the final ^ n is dropped. The 
^n is also lost in the nom. sing, of both genders (leaving 
^ ft as final in the masculine, q a in the neuter}. 

a. The peculiar cases of the neuter follow the usual 
analc^ {811b): the nom.-aco.-voo. pi. have the leiigthening 
to ^ S, as strong oases; the noni.-aoc.-Toc. du., as weakest 
cases, have the loss of i? a — but this only optioiially, not 

b. Id the loc. sing., also, the a may be either rejected or retained 
(compare the correspondiDg usage with patems: 373}. And after the 
m or T of man or van, when these ate preceded by another con- 
BOuant, the a 1b alwaj^ retained, to avoid a too great accnmalation 
of coDBonantB. 

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. Ab to accent, it needs only to be remarked that when, in 
the weakest caseB, an acute & of the suffix is lost, the tone ib thrown 
forward upon the ending. 

424. Examples of declension. As such may be , 
tak^Q jlsR r&jan m. kinp; qir^H fttmin m. soul, self;') 
sfPFJ^nfanan n. name. Thus: 

SlDgnUi ; 

raj a Stmi nama 

rJUJbuun Stmanam n&na 

I. fnn yirHHi =iraT 

rijSa atmAnft uanmK 

rijILa atm&ne namne 

ioy Google 

434 — ] V. Nouns add Adjegtives- 

Ab. O. jnW cJTffpTFr ^TBIT 

i^&as fttm&naa nimnafl 

r^jfil, r^ani fttm&nl 

^an atman 

N. A. V. ^IsfT^ yirHI*Jt 'TTHt. ^Wtpft 

raJSn&u fitmioSn namnt, 

I. D. Ab. TTsFnnr^ *iirH*-u'!H^ HiH'-ym^ 

r^abbyim StmAbhyain namablirlia 

r^fiOB fttmiinoB nimnos 


N. ^urns tfirHHH^ hwiIh 

r^J&nas fttm^as nlbnCul 

A. TntTi^ laifH-fH^ ^mi^ 

T^fLaa Ktm&uaa Dlbi&nl 

I. TTsT^H^ WlrHPlH^ =?THft^ 

TE^abUs fitmibhia nimabhla 

D. Ab. ^iyMJ4^ Mlr4^H^ ^H^UH, 

r^abhyas fttm&bbyaB ' n^abbyaB 

TaJiiAlll Stm&nftia n rfmngm 


r^asu fitm&aa namasa 

a. The weakest cases of murdb&n m. head, would be accented 
mardbna, murdhnfi, mQTdlui6B, mordhn&B (ace. pi.), m&rdhnaia, 
etc.; and bo Id all aimilar cases {loc. sing., miirdlmf or mOnUx&nl). 

436. VediD IrregaliTltles. a. Here, u elsewheie, the endljig of 
the nom.-Mc.-TOo. da. miac. li QsniUy K Initetd ol &u. 

b. The briefer form (with ejected a) of the loc. sing,, uid of the nent. 
nom.-acc.-Too. dn., is qaite unnsDal In the older language. RV. writes 
onoe 9atad£vni, bnt it la to be lead ^atadavani; and slmilai cases occur 
In AY. (bat aho'seversl times -mni]. In the Brihma^as, too, aaeb forms 
as db&mani and a&mani are very much more eommoo than saoh as aluti 
and lonml. 

ioy Google 

159 Declenbion V-, Srsua m m. [ — las 

a. But IbTODghont both Tsds uid Biahmins, uk abbreviated torm ot 
Iha loo. alng., with the euiltng i omitted, ot IdentloU with the 110111, ij of 
eoniidenbly moia frequent ooenneQce thm the legulu foitn : thua, mOr- 
dhin, Umuuit fctUtTOD, beside mardh&ni eto. The u h>« nU the 
viiul Mmbliwtloii* of ■ 6ml a: e. g. murdliaim aa;*, mOrdbo&t ea, 
mordhafis tvB- 

d. In the nom.-aoc. pi. neat., ■Iso, ui sbbiertatod foim Is eommoD, 
ending In a or (twice as often) a, instead of Kni: thus, br&bma ind 
briluna, beside tarAhroHiji : compare the ilmilai wrles of endings from 
a-stema, 3S9o. 

e. Fiom a few items in man is made an abbreviated Instr. ling., vtllh 
lou of m. u well as of a: thus, mahina, prathini, vari^a, dSni, 
pre^ bbOni, for mahlmni etc. And dxfighma and rafma (BV., 
each ODoe] are perhaps for drAgbm&fft, ra^taixxi. 

t, Other of the weakest oasei than the loc. sing, ate eometlmes found 
with the a of the snfflx retained: tbns, for example, bhdnuulS, damans, 
yamauaa, "'^;^?"* (scooa, pL), etc. In the laflniUve datives (970 d) 
— triuna^B, TidmAne, dlv&ne, etc. — the a alwajs remains. Abont as 
numerous are the instances In which the a, omitted in the written form 
of the text, is, as the metre shows, to bo restored In reading. 

g. The voc sing, in vaa, wMoh is the asual Vedic form bom stems 
n vant (below, 464 b] Is found also from a few In Tan, perhaps b; a 
transfer to the vant-decIenBlon: thas, ptSvaa, evaySvaa, l£bidTaa(?), 
prUaritraB, mStari^oa, vibhSvaa. 

b. For voids of which the a Is not msde long In the strong oases, 
i the nest paragraph. 

' 480. A fev stems do not make the regular lengthening of a in 

the strong cases (except the nom. sing.}. Tbns: 

a. The names of dlTtnlties, p&fAn, aryam&n: thus, piifa, pilfA- 
9am, puffi, etc 

b. In the Veda, ok^fal, bull (but also akfi^ain); y^qan maidtu; 
yfqan virile, bull (but vf^fti^am and v^&^aa are also met with); tm&n, 
abbreviation ot Ktm&n; and two or three otiier scattering fcrms: anarvA- 
9am, Jimanft. And in a nombet of additional instances, the Vedic metre 
seems to demand a where & is written. 

427. The stems qvAn m. dog and ravan young have in the 
weakest caaes the contracted form ifim and Tilin [with retention of 
the accent); in the strong and middle cases they are regular. Thos, 
^i, Qv^iuim, qdnA, ^{me, ate,, qv&bbySm, qvibbis, etc.; yiivK, 
y^vsaam, j6ah, Ti^vabbla, etc. 

a. In dnal, BV. hss once ytina for yuvEnft. 

498. The stem magh&van generous [later, almost ciclasiTely a 
name of Indra) la contracted in the weakest cases to magbdn: thus, 
matCtaATS, magh&T&aam, magbinfi* magh6ae, etc. 

ioy Google 

428—] V. NouMS AND Adjectives. IgO 

a. The RV. bu odcb the ve*k form niBghdnaB Id nom. pL 

b. PuaUel wltli tbla ii found the stem magli&Taiit (dMiton E); 
and from the Utter ilone In the older Isuguage ue mads the middle cum: 
thDR, nutglukTadbMB, magbaTBtBu, eto. (not ntoghavabliis etc.). 

4Se> a. Steiot in a, ma, va, pn>llel irlth those In an, man, van, 
KDd doD^tlags in muiy uses deriTed from tham thiongh traneltionil fonoi, 
ue frequent in botb the uiUbi and the Utei luig;uige, puticDluly w And 
memben of oonpotiDdg. 

b. A Dumber of an-atems ara more or less defective, making a 
part of their forms from other stems. Thas : 

430. a. The stem &lian n. day is in the later language used 
only in the strong and weakest cases, the middle (with the nom. 
siDg., which nsnallf follows their aoalogy) coming from dbarorUiOB: 
namely, ibar aom.-aoc. sing., ibobhySm, ibobhle, eto. (PB. ha* 
abarbbia); bat Uuift etc., A-hni or Abanl (or Aban), UmS or &baiii, 
&hfinl [ud, io V., &ha]. 

b. In the oldest Uogaiige, the middle uses Ababfals, ibftbhTAB, 
ihaan also occnr. 

o. In composition, onl]r abar ot abas la oaed aa preceding member; 
as final membei, abar, ahaa, alian, or the derivatlvea aha, ahna. 

d. The stem Adtaan a. udder eichangea In like manner, in the old 
tangnage, with ddbar and l^dhaa, bnt has become later an as-stem only 
(except in the fem. udbnl of adjective componnda): thna, tidbar oi ddbae, 
tidhnaa, lidban or ddbani, ddbabhls, ^dha^sa. Aa derivatiTes from 
It are made both Qdhany^ and udbasya. 

431, The neuter stems akf&n eye, asth&n bone, dadb&n curdt, 
aaktta&n thigh, form in the later langQage only the weakest caaes, 
(tfcf(i^ aatbn6, dadlm^, aaktlml or saktb&nl, and so on; the rest 
of the inflection is made from stems in 1, &kfl etc.: see above, 

a. In the older language, other cases l^om the an-stems occur : thna, 
ak^i^, akf&bbie, and ak^sa; asth^, astbAbbls, and astbibhraei 

433. The nenter sterna aain blood, yakiin licer, ^ak&n ordure, 
fis&n mouth, ad&n water, doq&a /ore-arm, ynf&n broth, are required 
to make their nom.-acc.-voo. in all Dunbers from the parallel stems 
^J, ji^tft, Q&hit, Asya, udaka (in older langaage ndaki;, d6B, 
yCtfi, which are fully inflected. 

a. Earlier occnra also the daat dof&ni. 

433. The stem p&atban m. road is reckoned in the later langnage 
as making the complete set of strong cases, with the irregularity that 
the Dom.-voc. aing. adds a a. The corresponding middle cases are 
made from patbf, and the weakest from path. Thus: 

ioy Google 

161 Dkclbksion v., Dbritative Stems is an. [—438 

from p&ntbsD — p&nthSa, pintliSnam ; p&ntliftiifiu; p&nthSuaa; 

from pathi — pathlbby&m ; patlifbhlB, patlifbtayu, pathlqu; 

from path — patb^ path6, path^, patbl; patbos; path&s or 
pithsa [accna.), pathibn. 

a. In tlie oIde«t Ungntge (R'V.)i bovevei, tbe etrong Bl«m Is onlr 
pinthA: thai, pJinthBa, nom. alng.; p&athKm, ue. elng.; p&nthu, 
nam. pi. ; and evan In A.V., piiiithliiaia tnd pADtbSnaa ue rue eoni- 
pand MU Uie othere. From patbf ocoui also the nom. pl. pathiyaa mi 
gen. pi. pathln^m. ST. h>i onoe pftth&B, >oc. pi., with long S. 

434. The stems mAnthan m. ftirring-$tick, tnd ^-bbakf&n m., mi 
epithet of India, ue given by the gtunmBrims the time Infleotion vltli 
p&ntban; but only s (ev cues haTS been fonnd In nse. In V. oecDi Trom 
fhe fonnei the vx. sing. m&ntbSm, and gen. pl. mathliilEm (like the 
eerregpondlDg cases from p&nthan); from the lattei, the nam. alng. ^bba- 
fcfas and voe. pL ^bbiik^SB, like the oorreipondlng Vedlc fonns of p&nthan ; 
bat also the aoc. sing, rbhakfifam and nom. pi. rbbnkf^pas, which 
aie anei qnite anotliei model. 


436. Original a^'ective stems in an are almoBt ezolusively tboee 
made with the suffix 7an, aa jr^van aaeriJUing, Butvon preumg tht 
aoma, jltran eonqutrtng. The stem is masc. and nent onl; (bnt 
sporadic o«b«s of its nae as feu. occur in RV); the correBpoading 
fem. stem is made in varf: thus, rfijvari, jltrari. 

486. Adjective compounds having a noon in an as final mem- 
ber are inflected after the model of noun-stemsi and the masculine 
forma are sometimes nsed also as feminine; bnt UBnallf a special 
feminine is made by adding i to the weakest form of the maaonline 
stem: thus, B6maTajfii, h^llUodlmi, dkamOrdbni, dnn^amnl. 

487. But (as was pointed out above: 420a) noune In an oconnlng 
as final members of eempoonde often snbstitatg a stem in a for that in 
an: thus, -rSJa, -Janma, -adbva, -aba; tbelr feminine is in s. Occa- 
sional eiohangea of stems In van and In vant alto oocnr: thus, Tiv&avan 
and vlT&svant. 

a. The remaining dlvialons of the consonantal declenaion are 
made up of adjective Btems only. 

D. DeriTatlve Btema (adieotlve) in in. 
488. The stems of this division are those fonned with 
the suffixes ^^to, f^min, and 1^_^vin. They are mas- 

ioy Google 


y. Nouns ahd Adjectives. 


culine and neuter only; the conesponding feminine is made 
by adding ^ T. 

a. The atema in In &ie very nuEaeroiie, since almost any iionn 
Id a in the langaage may form a poasasBive derivative adjective with 
this Biiffix ; thuB, b&la ttrtnglh, balfn m. n. balfni f. potteuing ttrengih, 
strong. Stems in vin (IfiSS), boweTer, are very few, and those io 
mln (1831) atill fewer. 

439. Theii inflection is quite legulai, except that they 
lose theii final ^ n in the middle cases (be£oie an initial 
consonant o£ the ending], and also in the nom. sing., wheie 
the masculine lengthens the ^ i by way of compensation. 
The voc. aing. is in the masculine the bare stem; in the 
neuter, either this oi like the nominative. 

a. Id all theae reapecta, it wili be uotiqed, the In-declension 
ttgreea with the an-deolenBion ; but it differs from the latter in never 
losing the vowel of the ending. 

440. Example of inflection. As such may be taken 
olidH balin strong. Thus: 

Singnlu. Dn&l. Plural. 


ballnam ball 


bal{nSu balinl 


baliuas balini 


(HIhhIH^ ballnfim 

Slfllft,/'; ' f ballnos"" ^%5 

ballnl ;: , j baU^n 

aif^l^aiRl'l.^% ^fe^ Slfepft ^f^I^ M^fH 

b&lbi bUin. b&ll bUln&u b&lini bUinaa bUini 

ioy Google 

163 Declehsioh V., Derivative Stems in In. [—444 

a. The darived femiDltie stem in Jul is inflected, of coorae, like 
saj' other feminine la derivstire I (3d4t. 

441. Ik There are no irregolaiides in the inflection of in-Btetus, 
io either the eulier language or the later — ei cept the ngual Vedio 
dml ending in ft instead of au. ' 

t>. Sterna in In eicbange with Btems lu i throughoat the whole hlt- 
tory of the lingotge, those of the one cUsa being developed oDt of thoae 
of the othei often throagh truuitlonal forms. In a much amaller niimbeT 
of casea, atema In in are eipiiided to ateiiig lu ina: e. g. QftkinA (RV.), 

E. DerivatlTe stems {adjective) in ant (or at). 
442. These stems fall into two sub-di visions: I. those 
made by the suffix s^rl ant (or WJ at), being, with a veiy 
few exceptions, active participles, present and future; 
2. those made by the possessive suffixes R^T mant and 
^rT vant [ot RH mat and c(rT vat). They are masculine and 
neuter only; the corresponding feminine is made by ad- 
ding ^ I. 

1. Fairtioiples in ant or at. 
44S. The stem has in general a double form, a stronger 
and a weaker, ending respectively in ^rT ant and ^frf at. 
The former is taken in the strong caacs of the masculine, 
with, as usual, the nom.-acc.-voc. pi. neuter; the latter is 
taken by all the remaining cases. 

a. But, in sccoidauoe with the cala foi the formation of the (emlnine 
Item (balow, 448), Che future partlciplee, and the present partlciplea of 
Terba of flie tud-dma or accented ^-claaa (762), and of verba of the ad- 
elaaa oi loot-cUsa ending In fi, ue by the gtammariaiia allowed to make 
the Qom.-ace.-vDC. da. neat from either the stronger or the weaker atom ; 
■nd the present paitlciples from all other present-atems eudlng In a are 
required to make the aame from the Eirong alem. 

444, Those verbs, however, which in the 3d pi. pres. 
active lose \n of the usual ending f^T nti (ESOb), lose it 
also in the present participle, and have no distinction of 
strong and weak stem. 

Digitizecy Google 


V. Notms AND Adjectives. 


a, Sneli are tbe Toibi fon&tng tbeii preaent-Btem by rednpllcition 
without Kdded &: oimelr, tbote of the ledapltMtin; oi ba-clwi (BBS) and 
the intenslTei (1012): thns, from yha, pTeieDt-«tem Jnho, ptttlcipla- 
alem Juhvat; tutenrfve-Btem johu, IntaoelTe putlclple-item j6llTtit. 
Farther, the puticiplea of loots appneutly contktnlng > contncted lednpU- 
cation: nimelr, e&kfat, da^at, daeat, qasat, a&^oat; the aoiiit putl- 
elple dli&kf at, and vBgh&t C'). VavfdliAiit (RY., once], which haa the n 
notwithstauding Ita ledupllaation, oomes, like the dettdentiTe partieiples 
(1083), from a stem in a: compace TSvfdli&ilta, vSv^pdlliBva. 

b. Eieu theie vaiba are allowed by the grammarlaDe to make tbe 
nom.-acc.-toc. pL nent. in antl. 

446. The inflection of these stems is quit« regular. Hie 
Dom. sing. masc. comes to end in ^ an by the regular 
(160) loss of the two final consonants fiom the etymological 
form iQt^iuitB. The vocative of each gender is like tbe 

449. Sterna accented on the final ejllable throw the acceot 
forward upon the case-ending in the weakest cases (not in the middle 

a In tbe feminine item) from enoh particlplea, 
la iBtained, ati if it la loat 

447. Examples of declension. As such may serve 
i^elft bbivant betiiff, 51^ ad&nt eating, g^JfT jfihvat socn- 


bh&van bb&vat 

bh&vantam bb&vat 






ad&ntam ad&t 





JAbvat J^vat 
JlUivatam J&bvat 





ioy Google 

Dbolbmbion v., Derivative Steiu in tut [ — 448 


^^ H^ 

^^ «^ 


bhivftn bh&vat 

idan iOat 




5^ a^ 
ad&nUa adaliY 

sj^Hl ^ 

bh&vnntftu bhivantl 

JubvatKu JubvatI 


. ^j^mi^ 












Plutil : 

N. T. 

H^?ff[^ iRiH 

5^tT^ 5^ 

s^^HH, i^^^ 

Jobvatas jdhvati 


'HrlTl^ i^srfW 

^rJ^^ ^^ 

5^fT^ f^ 

btaiv&tM bhivantl 

odatis ad&nti 

J^vataa Jiibvati 








D. Ab. 




















, The fatnre participle btaavi^yint ma; fonn in nom. etc. dnil 


bhavi«7«n; tndint, either tad&ati or 

tadati; siat (KrKj, either yintS or jrftti. And J^vat, i 

pinral neuter, may make aiso Jubvantl (beside jlUivatl, aa given in 

the paradigm alwve). 

b. But Utege atrong forma (u irell u bb&vanU, dn., >nd lt» Iik« 
tiom pTSUDt-itemB fn anaccented a) ua quite contniT to genetil inklosy, 
and of (omevrhat dnnbtfal chancEsT. No example of them li. qaoUble, 
eltbei fiom the older oi from the Ittei Isngntge. Tbe ouei cDnceined, 
indeed, would be everrwheie of iice occnnencs. 

448. The Tedlc derlTtUona tnta the model m eboTe glTsn are Ivw. 
The dual eediog In It only one aUth ta common sb K. AaomiloiiB accent 
ia aeen la a c«e ot two: aood&t«, rathirBy&t&m, aod Tfigb&dbhle (if 
thU la a participle). Tbt only Initanee in T. of nom. etc. pi. neal. la 
■intt, with iengtheued ft (compaie tlia fonu in &Qtl, below, 4B 1 a, 454 o) ; 
one ot two examplea In antl aie quotable from B. 

ioy Google 

449—] V. K0UF8 AND Adjeotives. 166 

449. The feminine paiticiple-stem, as already stated, 
is made bj adding ^ 1 to either the stiong or the weak 
stem-form of ihe ma^c.-neut. The rules as to irhich of the 
two forms shall be taken are the same with those given 
above respecting the nom. etc. dual neuter; namely: 

a. Participles from tenBe-Btems ending in unaccented a add T to 
the strong stem-form, or make their feminine in anti. 

b. Sn<ili ftre the bhQ oi nnaeeeiited a-elasR ind the iUt oi ya-ciui of 
pieient-Btemi (ehtp. IX.), tud tbe desidsT4tlve< and c&oaktitM (chap. XIT.): 
thus, from ybbu (etem bh&va), bbi4vaiitl; from ydlv (stem divya), 
ditryanti ; frcm bnbhi^B ind bh&v&ya (de«ld. ind cans, of ytib^), 
bdbbiifand and bhfiv&yants. 

o. EiMpUoni to thli role ue now and then mat with, sTen from the 
earliest period. Thna, RT. bae J&raU, andAT. tbe dettderatiTe slfBaaU; 
In B. oeeoT vadaU, gocati, tppyatl, and in S. fnrthei tlffliatl, and the 
caaaadve namayaQj whtle In the epics and later anch cues (Inclndtng 
deiideratiTei and cinaatlTea) aie moie namerom (about Ilfty are qnotable), 
though itlll only iporadic. 

d. Participles from tense-stems in accented k maj add the femin- 
ine-sign either to the strong or to the weak stem-form, or msj make 
their feminines in &nU or in atf (vith accent as here noted). 

e. Snch are the preaent-stems ot the tod or aceeuled &-cl&aB (751 ff.), 
the B-fnlnraa (033 ft.), and the denominaUvea (lOSSS.): tbna, from ^tud 
(stem tud&), tnd&nti or tndaUi horn btaaTi^yi (fnt. of yVtOi), bha- 
vi^yinti or bbavlfyad; from derayA (denom. of devi), deray&ntl 
01 devayati. 

f. The forma In 4ntl from thia claaa are the preTalllnf oneg. No 
future fern, participle In atl li quotable from the older language. From 
prei-'atems in & are found there piUatl and sULoatf (RV.), tndatl and 
pinvati (AY.). From denemlnatlTaa, devayati (KV.), dnraayati and 
^atruyatS (AT.). In BhF. occnre dbakqyaU. 

g. Verb* of the ad or loot-elags (61 1 ff.] ending Id & are giren 
b; the grammarlana the eime option aa leguda the feminine of the present 
participle; thua, ^m yyi, yantJ or ySti. The older language aSordi no 
example of the former, ao fat ai noted. 

h. From other tense-Btems than those already specified — that 
is to say, firom the remaining ciasaee of present-stems and irom the 
IntensiveB — the feminine is formed in ati (or, if the stem t>e other- 
wise aoeeoted than on the final, in atl] only. 

L Thus, adati from ^Ad; JfihvatI from ybu; ynfijatj from y'ynj; 
sunvati from ym; knrwati from ykj; kri^tf from ykn; dAdlqati 
from d^dig (Intena. of ydlg). 

ioy Google 

167 Declension V., Debitative Steug in &nt. [— 4BS 

J. Feminine iCeiDa of this olus &ic ocoiioniJIy (bnt the cise is mach 
lesi fieqaent tb4n Ita opposite: kboie, o) fonnd with the nual: tbae, 
T&ntl (AT., once), Tiiid4iitl ((^B.; bnt prabtblr from the (ecoDdary &-alem), 
grh^antl (S.), and. In tbe epliu tni lil«r, ench fonns la bruvnntl, 
mdnntl, ctnTuitl, barrantl. JftnantI, mof^antL 

450. A few words are partidpUl in form and Inflection, tbongh 
not in meaning. Thus: 

a. brbint (often written vrh&nt) great; it is inflected like ^ 
participle (with bphatl and bfh&ntl in du- and pi. neat.). 

b. mab&nt ffrtat; inflected like a participle, bnt with the Irreg- 
olaritf tbat the a of the ending is lengthened in the strong formB: 
thus, mali&i, mataiatam; mata&it&u (neut. mabati); mahlintaa, 
mab^tt: instr. tnahati etc. 

C. pffant tfeekled, and (in Veda onl;) rd^a&t thming. 

A. Jigat mtnxAh, lively (in tbe later lanfoage, w neuter noan, world), 
» redaplioated fDmatloD from ygam go; lt« nom. ele. nent. pi. Is allowed 
b; the gMmmBriana ta be only j&gantl. 

e. rh&nt *miiS (only once, la RT., fhati). 

f. All these form their feminine in atl only: thus, bfhati, 
mahati, pffati and rdfati (contrarj to the rule for participlCB), 

g. For Halt tooth, which ii perhaps of partlciplU ori^n, see abore, 


451. The pronominal adjectWee lyant and kiyant are inflected 
like adjectives in mant and vant, having (46S) {y&n and klySn as 
nom. masc. sing,, (yati and klyatl as nom- etc. du. nent and as 
feminine stems, and {yantl and klrantl as nom. etc. plnr. neut 

a. Bnt the neut. pi. {ySnti and the loc. sing.(f) kiyKtl are fonnd 
In RV. 

2. FoBBesBlves in mant and vant. 
462, The adjectives formed by these two suffixes aie 
inflected precisely alike, and very nearly like the participles 
in ^f\ ant. From the latter they differ only by leogthenicg 
the Q a in the nom. sing. masc. 

a. The too. sing, is in an, like that of the participle (in the 
later language, namely: for tbat of the oldest, see below, 454 b). 
The neut. nom. etc. are in the dnal only aU (or &tl), and In the plnral 
anti (or inti). 

b. The feminine is always made from the weak stem: thas mafl, 
Tata (or m&tl. vka). One or two oases of nl instead of I are met 
with: thus, antirratDl (B. and later], patlvatni [C.]. 

D,j,i,zec.y Google 

4B2— ] V. NODHB AHD AsjECTtTZS. 168 

0. The accent, however, is never thrown forirud (u In the 
partEciple] upon the eue-BndIng or the feminine ending. 

468. To illustrate the iofleotion ot snoh atems, it will 
be sufficient to give a part of the forms of ISRH pa^umiat 
possemng cattle, and HiraH bhigavont fortunate, blessed. 


pstfuman paqom&t bhigav&n bli&gavat 

A. tpiHrtu^ MajMf(_^ HJleWH^ m^[^ 

pwjumf^tam pa^nmit bhigavantam bh&gaTat 

paqtun&ta bh&gavat& 

piqumaa piqomat bb&gavan bbigavat 

N.A.V. q^Rrit tigqal >PT^ i^lRrft 

paqnm&ntiii pagumiiU bbigavantsu bh&gavatl 

etc eto. 


pagumiutaa pafonduitt bb&gavantBS bh&gavantl 

A. cj^TrTTl^ ntjuffl HIHrTfT H^wfrl 

pagnm&taa pa^am&ntl bbieavatas bh&gavantl 


pafum&dbUB bhigavadbhla 

etc. etc 

4B4. V«dio liTeguUTltiei. a. In doil muc nom. eU., ft (for 
&n) ii the grefttlj pier&Ulag ending. 

b> In Toc. »lDg. muc., tbe ending In the oldest Unciuge (RT.) la 
ilmoit tlwfty* in aa Instead ot an (m In tbe perfect puttcipU: ttdoii, 
462 a) : thai, adrlvas. haiivas. bUrLumaa, bavt^maa. Sach voMtlvea 
in RV, oooar moie than \ hnndred tlinea, while not ■ single onqneitlenible 
Initinoe of one to an la to be found. In Ihe other Yedia texts, TOMtivee 
In as ue Gxtiemelr lue (bat bbagavas uid ll« contrutioa bhagos are 
met with, ETen in the Uter langnige); uid in theti piodaotton of RT. 

itizecy Google J 

169 Dboliuisiok V., Derivative Stbus in nnt. [—456 

pMsagai the as li luatlly ebKng«d t« xa. It nu pointed out aboie (4SBk) 
that the BY. mtkei the toc. in (m ■I(d appuontlr (rem ■ te-u ail-stems. 

O. Id BT., the nom, etc. pi. neut., In tho only two inBttnceB that 
occar, ends In ftntl Initead of AUtl : thae, ghrtivUnti, pAfOmintl. 
No sach fonnB bare bgen noted elsewhere in the older Iuignag«^ the 6V. 
tt$,i* ftntl In it! verilon of the rorreiponding paslagel, and a feir ciam- 
ple( of the lame ending ire qnolable from the Brahmanu: tbna, tftvanti, 
otivantl, yjtvsatii shrt&vaiiti, pravuiti, rtomonti, yugmantL Com- 
pare 448, 461. 

<L In a few (eight oi ten) moie oi leas doabtftil caiei, a contoiion 
of strong and veak fonna of stem is made; they tie too purely iporadto to 
leqnlco tepoitlDg. The aame la tine of a case 0( two trbeie a mascnllDe 
fonn appeara to be oied with a feminine noun. 

4ftS. The stem irvant nmninjr, tUtd, has the Dom. eing. arvA, 
from irvaax; Mid in the older tanguage also the voo. arvan and acoue. 

466. Besides the participle bhAvont, there Is another stem bli&- 
vftnt, frequently ased in lespeetfnl address as subsljtnte for the 
prononn of tlie second person (but construed, of oonrse, with a verb 
in the third person), vhloh is formed vith the suffix vant, and so 
declined, having In the nom. sing, bh&vtn; and the contracted form 
bhoa of its old-style vocative bbavaa is a common exclamation of 
address: you, «iV.' Itfl origin has been variously explained; but i^ Is 
doubtlesB a contraction of bhigavant. 

457. The pionomlnal adjectlfei ttvant, etitvtut, T&vaot, and the 
Tedic ivant, mivant, tvivant, et«., are Inflected like ordinary deiivstlTea 

F. Ferfeot Fartioiplea in vUia. 

468. The active pactioiples o£ the perfect tense-eystem 
aie quite peculiar as regards the modifioations o( their stem. 
In the strong oases, including the nom.-aoo.-voc. pi. neut., 
the form of their suffix is ^fH vS^, which becomes, by 
regular process (160), vKn in the nom. siug., and which is 
shortened to ^^vsn in the voc. sing. In the weakest 
cases, the suffix is contracted into 3IT uf. In the middle 
cases, including the nom.-a43c.-voc. neut. sing,, it is changed 
to ^tT Tat. 

a. A nnion-vovel i, if present in the etrong and middle cases, 
disappears in the weakest, before of. 

ioy Google 

V. Nouns j 

1 Adjectives. 


469. 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^ ti^f. 

460. The accent is always opoQ the snffix, whatever be its form. 

461. Examples of inflection. To show the inflection 
of these participles, we may take the stems {cj^lll vidvAfiB 
knowinff [which has irregular loss of the usual reduplication 
and of the perfect meaning] from y^^ vid, and afp!I^f?T 
tasthivBAs having stood from y^^ sthB. 

BlngnUr ; 


iMAH_^ &SrT^ 

flit*lqii fTWRrT 

Tidvlia TidvAt 

tastblv^ tastbiTit 


^trn^ ^^ 

r)fwi=?l?m^ rTfera?^ 

vidvaAeam vldv&t 

taathivafisam taathivM 











Al). 0. 










MaH Mift 

FrfeR^ rlft^cffi 

T^dvan vfdvat 

t&atbivan t&atblvat 


H. A. V. 

l^StHt ftpft 

rlffBlctlfil rlfa^l 

vidvlbbBftu vldii;! 

1. D. Ab. 



G. L. 





i, Google 

1 7 1 Declbnsion v., Participles in vftAa. [ — 162 


vldvi^BaB vidvaAsi tuthiv^aaB tuthiv^el 
A. HrS^H icJAlitl FTRI'TO FrfeRtftl 

Tldu;as vidvifisl taathdqas taatblvaiiBi 

vldT&dbblB tutbiT&dbhlB 

TidvidbhTU twthlv&dbhriM 

Ab. Q. i^i"E|W ffWllll 

vid^&m tasth^ffim 

L. KlSrH rlffEWrH 

Tidvitaii tftsthlT&tsa 

a. The feminine stems of these two participles aie feTjyl 
Tidd^ and H^g^l tastbd^. 

b. Odier emmples of the diffeient stems are: 
from ykx — oak^^^a, cakrr&t, oakra^ cakru;!; 
froin ym — ninlvAAs, nlnlT&t, 1111176^ ninTU^I; 

from yljhQ — babua'ribiB, babhtlv&t, babhuTiif, babhuvu^i 
from Vtan — tenlT^s, tenlv&t, t«nnf , t«nu^. 

488. a. In tbe oldest luisuige (RT.), the TOcitive alng. mate, (lika I 
that of vant and muit-atema: aboTS, 454 b) haa the ending iraa inatead 
of van: tho«, oUdtras (changed to -van in a parallel pauage of AT.), 
titirraa, dldivas, ml^bvoa. 

b. Fonni tiom the middle stem, In vat, aie eitremel; rate earllei : 
ODlf three (tatanv&t and vavrtvit, nent. elnf., and JBgfr&dbhta, initr. 
pL], ara fonsd In RT., and not one in AY. And In the Teda the weakest 
Item (not, aa later, the middle one) U made the baale of comparlion and 
derivation: thai, vidi^fara, 4dSfa;tara, ml^oftama, mltJliA^ant. 

o. An extmple at two of the nee of the weth etem-fonn for eaiea 
regnlarlf made from the strong are found in RT,: the; are oakri!lfam, 
acfl. (ing., and iblbliTDfas, nom. pL; emu|&m, by ite accent (nnleu an 
enor), is taAer from a derlvatlTe stem emu^; and QB. haa pro^fam. 
Similar inttcneea, eipeclally from TtdviAB, are nov and then met with 
later (see BR., under vidvltAs), 

d. The AT. has onee bhaktlv^Aaaa, as If a paiticlplal form from a 
noon; hot S. and TB. gi*e in the ooriupondlng pasaage bhaktiv^afl-, 
eakhT^aam (RT., once) la of donblful character; oUvlAaK (ST., once) 
■how* a reiertion to guttnial form of the Ilnal of yno, elsewhere nnknown. 

Diaii.ze.., Google 

G. Comparatives in ySAs or ysB. 

463. The comparatiTe adjectives of primary fonnatioii 
[below, 467) have a double form of stem for masouUne and 
neutei: a atioogei, ending in QtH rS&s [usually f^jtff TySAs), 
in the strong cases, and a weakei, in m yas [or ^ITH lyae), 
in the weak oases (there being no distinction of middle and 
weakest). The voo. sing. masc. ends in TR yan (but for 
the older language see below, 466 a). 

a. The feminine is made by adding ^ I to the weak 
maec.-neut. stem. 

464. As models of inflection, it will be sufficient to 
give a part of the forms of WI^ 9r^yaB belter, and of 
i^lUH g^Lrlyaa heavier. Thus: 



nftaR^ n^teni^ 



Kfaiyto giriyaa 


iTftatH\ ufhra^ 


K&rtyjUHsam g&rlyM 












giriyan gfcriyoa 



nfhitqt iTTtarft 



S&nyiAsftu g&rlyaBi 

9r6yBABBe qriyafiBl giriyiAsaa g&riyOABi 

friyoeu Qr^yftAsi K&riyasae g&rly&ftal 

gr^obhia g&rlyobhiB 

ioy Google 


a. The feminine stems of these adjeotires are vIuhI 
frdyasl and )|(1uh1 girlyasl. 

466. a. The Yedio too. mue. (u In tbe two pracedlng diiialoDB: 
464 b, 461 a) ii in yas lutetd of yan : thiu, ojlyaa, Jy&raa (RV. : no 
ei tin pie* eltewhere have be«n noted). 

b. No eiunple of » middle cms ocodti in RV. oi AV. 

0> In the later Ungnsge ue found t, veiy few ippireiil eiimplee of 
■tiong ouei mide from the vedcer ■tem-form: thus, kanlfaaam and 
yaviyaBam Me. muo., kanlyasKu do., yavlyaaaa nom. pi. 


466. Deiivative adjective stems having a oomparstive 
and superlative meaning — or often also (and more origin- 
ally] a merely intennve value — ate made either directly 
&om toots (by primary derivation], or from other derivative 
or compound stems (by seoondaty derivation}. 

a. The mbject of eempulson belong! more pToperl7 to the ebspter of 
deiiTttlon; bnt It Mandi In snch neu lelttion to tnllectloo that It is. In 
Mceidtnee with the antl onitom In griinmut, conTenteDtly and suitably 
enoDfb treated briefly here. 

467. The suffixes of primary derivation are ^T]TT Iras 
(or ^ITtTT lySbs] for the comparative and ^ i^^ for the 
superlative. The root before them is accented, and usually 
strengthened by gunating, if capable of it — oi, in some 
oases, by nasalization or prolongation. They are much more 
frequently and freely used in the oldest language than 
later; in the classical Sanskrit, only a limited number of 
such, oompaiatives 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 
theit corresponding positives; but in part also they ate 
aztifioially connected with other words, unrelated with them 
in derivation. 

a. Tbns, from ykflp hurl come kf^Iyas nnd fcfipi^ha, which 
belong in meaning to k^prt fuiek; from yvx tneompatt oome v&rl- 
yaa and viriffha, which belong to urn broad; while, for example, 

itizecy Google 

467—1 V. NOCHS AND Adjbotivbs. 174 

kdjilru and k&niftba are attached b; Che grammarians to yuvan 
young, or &lpa mnall; and v&rqiyu and T&rqi^tba to v^ddbi old. 

468. From Veda and Brahmana together, considerably more than 
A hundred inntanceB of this primary formation in lyas and Iffha (in 
many cases only one of the pair actnally occurring) are to be quoted. 

a. About half of thus (<n RV., the deolded mKJorlt;) belons, In 
meaning u in form, to the 'bars root In its sdjeetlve idae, u DBed etpe- 
clatly Bt the end of oomponDds, bat Bametlme* also ludependantly ; thus, 
from ytap bum comes t&pi|f^a excenivelt/ burning; from I'yBj offer oome 
r^iyas uid yftjlf(lia bvlttr mi bttt (01 vay icalfjiacrifieing; from j/yndh 
Jight aomeB yAdhlyaa fitting befttr; — In a few Instances, the simple 
root is also found used ss conespoadlng positive; thus, Jill hotly, rt^id 
with J&Tiyaa and j&viqttut. 

b. In a little class of instinoes (eight), the loot bai s ptepesiUou 
prellxed, which then tskes the accent: thus, igaml^tba eipteialiy commg 
hither; vfcayiffha beat clearing moaj); — In a conple of oases (Aqrami- 
(t^a, ipar&vapiqtha. ^theyaa), the negative paitlole is piefiied; — 
In a single word (Q&mbhaTlqtlia), an element of anotbu kind. 

o. The words of this foimatlon somatimea take an Mcnsatire object 
(see 271 e^ 

d. But even in the oldest language appears not infrequently the 
same attachment in meaning to a deriTative a^ective vhioh (as point- 
ed out above] is usual in the later speech. 

a. Besides the eiau^les that occur alto later, others are met with like 
T&rlftlia choicat (v&ra choice), b&rb.l;tlia greatait (b^li&iit great), 
6;lt;(ha quiditit (btftaa quidcfy), and so <m. Fiobsbly by aDaJog; irltb 
these, like formatlona ue In a tew cues made from the appuently ndioal 
syllsbles of words which have no otherwise traceable root la the language: 
thus, ktadhiyaa and kradhlft'ta (K.) from k;dll&. Btb&ViyBB aud 
Bthiviq^bK from Bthar&, f&flyaa ifiy.) from q&^ant, iafiyaa (AV.) 
aud A^itlia (TS.) from ai^i^-, and so on. Aud yet again, In a few eieep' 
tional cases, the suffixes lyaa and iffha are applied to stems which ue 
themselyes palpably derivati'e: thus, ^l|(lia from &qu (RV.: only ease), 
tikQ^iyas (AV.) from tikf^i, br&hmiyaB and br&hmi^tha (TS. etc.) 
from brUimtui, dh&nniftba (TA.) frooi db&rman, dr&^hifta (TA.: 
Instead of d&rhiqlha) from dp^bi, r&gMyas (TS.) from raghn. These 
are beginnings, not followed up later, of the extension of the formadou to 
unlimited use. 

f. In n^viyaa or o&vyaa and nAvi^Iia, bom a&va neu>, end in 
atoyaa from eitaSk old (all RV.), we have also formations nnconnected 
with verbal roots. 

466. The stems in i^tha are inflected like ordinary at^eotives 
in a, and make theii feminineg in S; those in iyaa have a peculiar 
declension which has been described above (468S.J. 

ioy Google 


47a Of peoolarities and irregalaritieB of formatioo, the follow- 
ing may be Dotioed: 

a. Thg satflx lyaa hu In s tew instances the briefer form TAB, gener- 
tlly aa alternative iritii the other: thne, t&viTaB and t4vyBS, n&viyaa 
and n&vjOB, v&^aa and v&syaB, p&ulyaa and p&nyae; and sa from 
rabh and aab; a&nraa dcoiub aloae. From bliu coma bhdyas and 
bhdyiftha, beatd« irbich RV. has aUo bMviyas. 

b. Of roota In &, the final blends with the InitUl of tho BuffliL to b: 
thuB, 8tti6yaa, dhi^tlia, r^ffliai but snoh forma are in the V«da gener- 
ally to be reaolved, aa dliil^tha, yiiSq^a. The root JyS ronns Jyeftha, 
but Jyayaa (lik* bbtiyaa). 

o. The tvo ioot« in I, pri and 911, form prdyas and prd^flia and 
^r^yaa and Qr^fha. 

d. From the loot of fjA come, wlthont strengthening, ^Jlyaa and 
^Jiftha; but In the oldu language also, moie regularly, r^iyaa and 

471. The suffixes of secondary derivation aie FT^ tara 
and fT^ tama. They aie of almost uuestricted application, 
being added to adjectives of every form, simple and com- 
pound, ending in vowels oi in consonants — and this from 
the earliest period of the language until the latest. The 
aooent 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). 

a. Examples (of older as wel) as later oooarrencel are: from 
Towel-stema, prlyfttara, v&hnltama, raOiitara and rathltama (BV), 
oAmtara, potftama, aadiTaktatara ; — from coDsonant-Btema, Q&ih- 
tama, q&Qvattama, mr^arittama, tav&atara and tav&otamai tuvif- 
(ama, v&pa^tara, tapaavltara, yaqaBvitama, bbiigavattara, bira- 
QyavSfiiaattama ; — from compouDda, ratnadhatama, abhlbbiitara, 
sukfttara, pOxbhittama, btaOylfthabhaktama, bhiindaTattara, 
fuolvratatama, strtkfimatama. 

b. Bnt in the Yeda tbe final n of a stem ii regularly retained; thns, 
mBdlntara and madintama, vff&ntaina; and a few sterna oTon add a 
nual: thuB, anrabhlntara, rayCntama, madb^ntama. In a case or 
two, the strong item of a present participle is taken: thna, TTadbanttama, 
B&banttama; and, of a peifeot participle, the weakeat atem: thaa, vldinQ- 
(ara, ml^buft&ma. A feminine final I la shortened: thus, davltamS 
(BV.), tajaavlnit&mft (K.). 

ioy Google 

471—] V. Nouns AND Adjbotivbs. 176 

0> In UiB older luigDage, tlie iroTda of this fonnatloTi ut not maeli 
more freqnent than those of the other: thas, In RT. tho itemB In tarn 
and tama are to those In Itm and Iqtha u three to two; In AV., ooly 
as six to flye: bnt liter the former win a great ptepondetanee. 

47S. These compaTatives and BoperlatiTfls are inflected libe 
ordinary adjectivefi in a, forming their feminine in S. 

473. a. That (espeolallr in the Veda) some stems which are 
nouns rather than adjectives form derivatives of comparison is natural 
enough, considering the uncertain nature of the divieioD-line between 
substantive and adjective value. Thus, we have vir&tora, vir&tama, 
T&hultama, m&t^ama, nftamo, maruttoma, and bo on. 

b. The suffixes tara and tama also make forms of comparison 
from some of the pronominal roots, as ko, ya, 1 [see below, 6S0); 
and from certain of the prepositions, aa ud; and the adverbially used 
aconeatlve (older, neuter, -taram; later, feminine, •tar&m] of a com- 
parative in tara from a preposition is employed to make a corres- 
ponding comparative to the preposition Itself (below, 1119]; while 
•tax&m and -tamiLm make degrees of comparison from a few ad- 
verbs: thus, natarim, natamim, katbaditarSm, kutaatarKm, 
addh&tam^m, nleaistarlim, etc. 

o. By a wholly haTbaiona_ combination, finding no warrant In tlie 
earlier and more geuuloe DSiges of the laogaage, the snfllXBS of compariMD 
la their adverbial feminine form, -tarftm and tamBm, are later allowed 
to be added to poTBonal forma of verbs: thna, aidatetarim (R. : the only 
ease noted in the eplos) i> more detpondmt, vyatbayatltarSm ditturbs 
more, alabbBtataiftm oitaintd in a higher degree, hasifyatitarKm will 
laugh more. No examplol of this nse of -tam&m are quotable. 

d. The gufflxea of secondary compariwn aie not iDfceqnently added 
to those of primary, (ormtng double comparatlTee and snperlatfTes : thus, 
lariyaatara, Qreftbatara and ^^(hatama, pKpiyaatara, pftpi^flia- 
tara and -tama, bhuyaataram, eto. 

e. The use of tama as ordinal suffix Is noted below (4B7f J ; with 
this value, it is accented on the final, and makes its feminine in I: 
thus, qatatami m. n., qatatami {., hurtdredth. 

474. From a few words, mostly prepositions, degrees of com- 
parison are made by the briefer suffixes ra and ma: thus, &dbaro 
and adbomi, &para and apamA, Avara and avami, 6para and 
upamli, &ntara, &utama, param4, madhyami, oaram&, antltna. 
Sdima, paqoima. And ma is also used to make ordinals (below, 487). 

ioy Google 



476. The simple oaidinal numeials for tbe first ten 
numbers [which are the foundation of the whole class), 
with theii derivatives, the tens, and with some of the higher 
members of the decimal series, are as follows: 







qat4 C*t- ■ 









30 fsfUR^ 














60 yyiwr^^ 







80 q% 

10,000,000 ^Ifi; 





10 Rnf?! 







80 M!*TlfH 




















Tha aoMnl 

: sapti iDd mk li thtt 

beloQgiDS tt 

> theie woT4a In all 

Mti texta; 

re B&pta and i^\t, 

In tbe later laDgnige. aee below, 483. 

b. The series of decimal numbers maj be carried still further; 
bnt tbere ue great differences smoiig the different authorities with 

Wkiti*;, anrnmiir. U. id. 12 





legard to their names; and there is more or leas of diBCordknce even 
from ayi^ta on. 

c. Thna, In tb« TS. and MS. we find ftyuta, oiydta, prayuto, 
&rbuda, ny^buda, lamudrd, m&dhya, Anta, parftrdh&; E. iBTeno 
the ordei oF alyiita and prayuta, uid Imerts badva >ftei nyarbuda 
(reiding njrarbudha): these ue piobsbly the aldett lecoided aeriea. 

d. In modem time, the only numbers In pnetlcil nas abate thoutand 
■IB lakfa (lac or lath) and koti (crore); and an Indian sum ia wont to 
be pointed thus: 123,45,67,890, to signify 123 erorts, dS laJAi, 67 Uou- 
taiid, eight hundred and ninety, 

0. Aa to tbe alleged ilem-forma patLoan etc., aee below, 484. A* 
to tbe roim ^akQ tuatead of ;af , aee aliove, 14S b. The alem dva appeara 
in compoiition and derivation alao u dva and dvi; oatur in compositian 
la accented o&tur. The older (oim of aqta Is a^i&: see below, 488. 
Forms in -qat and -gatl /or the tens are occialonally interchanged: e. g. 
TlAQat (MBh. R.), trlAQatl (AB.), pafioB^ati (itT.). 

f. The other numbers are expreaaed by the varioua compoBition 
mid ayntacticftl combination of those given above. Thoe: 

470. The odd Dnmbera between the even tens are made by 
prefixing the [accented) unit to the ten to which ita value la to be 
added: bnt with variona inegalarlties. ThtiB: 

a. eka in fl becomea ekft, bnt la elsewhere unchanged; 

b. dva becomes everywhere dv&; bnt in 42-73 and In 92 it ta 
Interchangeable with dvi, and in S2 dvi alone ia uaed; 

0. for trl U snbstltnted Its nom. pi. maac. tr&yaB; but tri ilaelf la 
alao allowed In 43-73 and in 93, and in as tri alone Is need; 

d. eaf becomes ;o in IS, and makes the inlUal d of daqa llngval 
(199 d); elsewhere Its final undergoes the regular conversion (336 b, 198 b) 
to t 0' 4 or ?; >nd fn 96 the a of navfttl It assimilated to it [198 o); 

e. aq(a becomes Bf^K (488) In 

succeeding combinations. 

1 ia-38, and has < 

either form In the 

f. Thus; 

u ekadaqa 
11 dvada<fa 

31 ^katriiqat 
31 dvatpiiijat 

«i Akafa^ti 

31 ikSfitl 
Bl dvy^ti 

13 trdyodaqa 

33 tr^aatrU^at 


B3 try4qitl 

14 o&turdaqa 

15 p&fioadaqa 

IK ^bi^B. 

IT BBptddaqa 

31 B&tuBtTi&Qat 
3 a p&fioatrlAQat 

36 ^ttrlAqat 

37 Bapt&trlJiqat 

«i o&tnbfaftl 
es p^oaqaffi 
U 9&t9a9tl 

«7 Mpt&Wti 

81 oituraqltl 
30 p&aoftoitd 

37 aapt^ti 

IS ft?5^aqa 

38 aft^triA^at 


S8 aftS^iti 

le navada.;a 

3» oivatrii<!«t 

BB ii&Ta?a^l 

Bii n&vSflti 

Dij.ieo, Google 

179 Odd Huubeks. [—478 

g. Tbe nmnbacs SI-29 ue mide llks thole foi 31-39; the nanbets 
4l-d9, il-59, 71-79, and 91-99 ara mide like thoee for 81-69. 

h. The fanns mtde with dirS ind trayas up more usual than ihose 
with dvl and trl, which are hardlr to be quoted From the older literature 
(T. and Br.). The forms made with aqtd (initead of a^td) are almoit ei- 
clntlvely used In the older Uteritote (48S), and are not Infrequent In the 

477. The above are the DortanI ezpresaioDa for the odd niim- 
bera. But equivalent BabBtitatea for tbem are also variously made. 

a. Br oae ot the adjectives Qua deficient and adhika redundant. In 
composition with lesser uomberB whlrh are to be subtracted or added, and 
eiUier Independently qnalifrlng or [more usually) in composition with larger 
onmbers which are to he iiicreaaed or dimlnUhed by the others: ihue, 
tTTuaafaiftl^ lixty deficient by three (1. e. S7); a^t&dlilkaiiaTatlh 
nwty increaeed by eight (i. o. 98); ek&dUkaih fatam a hundred tn- 
ereaeed by one (i. e. lOl); paficonaifa ^ataia 100 let* 6 (1. e. 9S). For 
the nines, eapeclally, inch subitltutes aa ekoiiavlAQati^ SO hee I, or 19, 
are not unoommon ; and later the eba 1 U left oiT, and QnavlAQatl etc. 
have the same value. 

b. A caae-ronn of a smaller number, generally Aa one ii connected 
by ok not with a larger number from which It Is to be deducted: tll^^l, 
iikaji ni trliLf&t ((B. PB. KB.) not thirty by one (30); dvitbhyfiih 
a& 'qliim (QB.) not eighty by two (7S) ; pafie&bMr ii& oatvari fat^ 
(^B.) not four hundred by Jive (396) ; ehasmfia n4 paSoa^&t (in ordinal) 
i9 (TS.); HkMayU [abl. fern.: 307 b) oipaacil^t ^» (TS.); moat often, 
akia (1. e. i'ktt, irregular abl. for dkaamAt) nil vldfatfl; 19i 4k&ii ii& 
Qat&in 99. Thii last torm ie admitted also In the later language; tbe 
others are found In the Brahmanas. 

o. Instances of multiplication by s prefixed number are occasionally 
met with: thus, trl^ptA thrice seven; trinavA thrict nine; tridai;& 
thrice ten. 

d. Ot coarse, the numbsts to be added together may be expressed by 
Independent words, with connecting and: thus, n&va oa navati? oa, or 
n&Ta navatiq oa ninety and nine; dvtu ca viA^atf^ oa tioo and 
taenty. Bat the connective Is also (at least, In the older language) not 
seldom omitted: tbna, navatfr niiva 99; triA^&taifa trih 33; aqitfr 
a^tftn 88. 

478. The same methods are also variously used for forming tbe 
odd numbers above 100. Thus: 

a. The added number is prefixed to the other, and takes the accent: 
lor example, ^ki^atam 101; a:^ttia.ttaa 108; tri&f&cobatani ISO; 
aft&vififatlfatam 1^8; o&tu^aabaaram (RV. : unless the accent is 
wrong) loot; o^tleahasraan 1080. 


Dgil.zecy Google 

478—] VL NVHBKAi^. 1$0 

b. Or, ihe aoiuber to be added i» compoanded with adhUu rtdundatU, tha compauiid ii elthac made to qualify tbe other niunbai or la fnithet 
compounded vlth it: thus, pa&o&dMkaifa ^atarn oi pafioSdhikaqaUun 
lOi. Of coune, una deficient [ai alao other woida eqiUTalant to ana oi 
adblka) maj be used lu tbo »udo way: tbni, pafioonadi qatam 95, 
qa^ti^ paiiaavarjitfi 5S; qatam abhyadtalkaifa f&ftltal^ 190. 

O. Syntactical cambiiiatlons ate made at cou-venience: for example d&fa 
gat&ih oa no-, gat&in dkaib oa 101. 

478. Anotber OBiial method (begioning in the BrAhmaoaa) ef 
forming the odd numbera above lUO is to qnalify the lar^r nniober 
by an adjective derived from tbe smaller, and identical with the 
briefer ordinal (below, 488): thus, dvOdaq&ih qat&m. Hi (lit'ly a 
hundrtd of a IS-iorl, or eharaeUritadbj/ 12); oatuqoatv&ri&f iib f atdm 
idd; fat^aqt^di ^at^m IBS. 

480. To multiply one number by another, among the higher or 
tbe lower denomiuationB, the simpleat and least ambiguous method 
ia to make of tbe multiplied number a dual or plural, qualified by 
the other as any ordinary noon would be; and th I a method b a com- 
moD one in all agee of the language. For example: pAfioa palLoK- 
ijAtoB Jive fi/liei {S50}; niva aaTatAyaa nma nituliti (810); aqltibUs 
tlB^bhia with Ihree tigMie* '\:H0); p&iioa ffttitni ^ve hundredi; trlfl 
aah&ar&i^ three thoutaadt; ^a^tlifa. aah&arfi^ 6I),000; daqa oa aahaa* 
rai^ af^u oa qat&ni 10,800: and, combined with addition, tri^i 
9at£ni tr&yaatrl&qataib oa 333; aahaare dve pafkoooaiii qatam ova 
oa 3095. 

a. In an excepUonal caie 01 two, the ordinal form appean to take 
the place of the cardinal aa mnltlplicand in a like combinaUou ; thna, ^\- 
trtiiflaAq oa oatdra^ {KV.) 3ex4 (lit. /our 0/ the thirU/tix kind); 
ttiAt ek&daqan (KV.) or traya ekadaqjbM^ C^. Till. t\. i) iJxS. 

b. lly a peculiar and wholly Illogical conetmetlon, iuch a combination 
as tri^i qai}tiQatanl, which ooght tu aignify jau (3x100 + 60], la repeat- 
edly us<!d In tbe Bnhmanaa to mean 360 (3x100 + 60); >o alao dvA 
oatuatriAq^ qaXb S3t (not 768); dvi^mt^hai trlnl qat&ni 3eS; and 
other like cuei. And even K. has trayah qataqat&rdli&h 360. 

461. But the two factors, multiplier and multiplied, are also, 
and in later uaage more generally, combined into a compound (accented 
on tbo final] ; and this is then treated as an adjective, qualifying the 
numbered noun; or else its neuter or ftminlne (in I) singular is used 
aubstautively: thus, daqaqataa looo; qafqatftih padfttlbbil^ (UBh.) 
with 600 foot-aoldier»; tiaraatriiiqat trlqata^ qa^ahaar£^ (AV.) 6333; 
dvlqaUiia oc dvlqati 200; a^t&daqagatl tsoo. 

Oi Id the usual absence of acnentnatlon , there ulsea (ometimet a 
queatlon ai to how ■ compound nam bet shall be nuderitood: whether aft*- 
qatam, foi example, is afft&qatam lOa 01 a^ta^atAm WO, and the like. 

Digitizecy Google 


181 Intlbction. [ — 46S 

4Sa. Inflection. The inflection of the cardinal numeTals 
is in many respects irregular. Gender is distinguished only 
by the first fooi. 

a. Bka one is declined after the manner of a pronominat adjec- 
tive [like B&rva, below, 6S4); its plural is used in the eenae of tome, 
certain ontt. ItB dual does not oocnr. 

b. Oi'.cMlonal formi of the ordinirj declensicn tf« metirfth; thai, 4ke 
{loc. ilng.), 6k(lt C477 b). 

a. In the Ute literatnrp, eka is nsed in the sense of a certain 
or even Bometimes almoat of a, as an indefinite article. Thus, eko 
vyBghra^ (H.) a certain tiger; ekaamln dine on a certain day; haste 
doq^am ekajn idftya 'H.) taking a itick in hit hand. 

d. I>Ta two is dual onljr, and is entirely re^lar: thus, N, A. V. 
dvin (dv&, Veda) to., Avi f. n.; I. D. Ab. dT^hySm; G. L. Av&yot. 

e. Trl three is in mnsc. and nent. nearly regular, like an ordinary 
stem in t; but the genilJve is as if from tr^& (only in the later 
language: Ihe regular ti^&n occurs once in RV.). For the feminine 
it has the peculiar stem tis^, which ie Inflected In general like an 
f-etem; bat the nom. and accns. are alike, and show no strength eniog 
of Che r; and the j is not prolonged in the gen. (excepting in the 
VedaJ. Thus: 

m. n. f. 

N. tr&T-aa tri^l tiarAa 

A. tnn tn^i tlar&a 

I. trlbtals tiatbhis 

D.Ab. trlbfa7&8 tia^bhyaa 

0. tray&^&u ttap^im 

L. trlt^d tU^ 

f The Vada h>B the Rbbtevlated neat, nom, and >cen». trl. The 
acrentDatlon tispbhie, tiarbhy&a, tls^nSm, and tlsni^ ■• aaid to be 
alao alloved In the lat«r laegiiage. The item tlBf occnra in comi>ogitioD 
In tlardhaav& (B.) a bou? with three arroxoe 

g. Oatdr four baa oatvar |[be more original form) in the strong 
csees; in the fem. it aubatitotes the stem o&tasr, apparently akin 
with tlaf, and inflected like it (but with anomatoug change of accent, 
llke^that in the higher numbers: aee below, 483). Thus: 
m. n. f. 

N. catvtras oatriri o&tasraa 

A. osturaa oatv^ o&tasras 

I. oatuTbhis oatas^bhls 

D. Ab. oatdf bhyas oataatbhraa 

O. oatar^itm oatasf^ttm 

L. oat^TQU oatasftn. 

Digitizecy Google 

48S— ] VI. Numerals. 1S2 

h. The uae ot a berors Km ot Hie gen. taaac. and neat. 4fl«i > Bnil 
conBODATit of the stem is (w in ^oq: below, 463) a itiiklng IrregoUiltf. 
The maiB regalir gen. tern. oatasf^Sm Ubo sometlmea oecnre. Id tbe 
later language, the accentuation o( the final eyllable Inalead of the penult 
l9 eaid to be allowed In initr., dat-abl., and loc. 

463. The nnmbera from S to 19 have no distinetion of gender, 
Dor any generic character. They are inflected, Bomewhat irregularly, 
as plurals, save In the nom.-acc., where they have no proper plural 
form, but show the bare stem instead- Of f&q (as of catdr), nam 
is the gen. ending, with mutual assimilation (108 b) of etem-final and 
initial of the termination. A;(& (as accented in the older language) 
has an alternative fuller form, aqt^ which is almost exclusively used 
in the older literature (V. and B.), both in inflecUon and in compo- 
sition (but some compounds with aff^ are found as early as the AT); 
its nom.-acc. is a^^k (usual later: found in £V. once, and in AT.), 
or ofta (BY), or oft&u (most usual in RV.; also in AT., B., and 

a. The Bcrent is in many respeota peeollai. In all the accented teits, 
tbe stiesi of voice Ilea on the penult before the endings bhls, bbjas, and 
BU, ftom the st«nia in a, whatever be the accent of the stem ; thna, pafL- 
o&bhla from p&fioa, nair&bhyaa ftam n&va, daq&su from d&qa, nava- 
daq&bhlB from n&vadafa, ekfida^bhraa from ^ksdafa. dTftda^&sa 
from dvida^a (according to the grammailaiu, either the penalt or the 
llnal li accented In these forma in the later language). In the gen. pi,, 
the accaut is on the ending (as in that of 1-, u-, and f-Btems): thus, pafi- 
oadaqftnam, Baptadaf&r^n. The caaea of ^a^ and tiioae made from 
tbe Btem-furm aft^, have the accent thioaghoat npon the ending. 

b. Examples of the ioRection of these words are as follows: 
N. A. 



a^tsi aft4 



a^tabbis a^bhia 



a,t»bbyiB a^tibhyaa 





aftaafi aft&sTi. 

0. SaptA (Id the later language s&pta, at dL^^a for af^) and n&TR 
and d&Qa, with the oomponnda of d&qa (11-19), are declined like pAAoa, 
and with the Bame shift of accent (or with alternetlTe shift to the endinga, 
aa. pointed ont above). 

484. The Hindu grammarlana give to the Btems for 5 and 7-19 a 
final n: thus, pafioon, aaptan, o^tan, navan, daqan. and akada^an 
etc. Thii, however, haa nothing to do with the demonBtrabl; oilglnil llnal 
nual of 7, 9, and 10 (compare stptem, novem, decern; ' eeven, nwie, 
ten); It Is only owing to the fact that, starting from suf^h a stem-form, 
theli InQectlon la made to asBume a more regqlac aspect, the nom.-icc. 
having the form of ■ neut. sing, in an, and the Inatr,, dat.-abl., and loe. 
that of a neot. or maae. pL In an: compare nluna, n&aabhlfl^ xiiaui- 

ioy Google 

183 Imflbotion. [—487 

bhyaa, tUtmun — the gen. alone being, nllieT, like tbit of >d ft-item: 
eompue '<'m;*'"^"' *itl> {ndrfti^ftiD tnd namnSm oi Um&n&m. No trace 
whktsTBi of > Bnal n is found uiywhaie In the Uofnage, In Inflection oi 
derlTMIon ot comporition, ^m any of theie voidi (tboogli 9B. hM twice 
dAftufadK^, (oi the U8ua] doqodiM;!!!), 

4S6. a. The tens, viAqftti hdiI tiifiq&t etc., with their compuunde, 
are declined regularly, as feminine atema of the same endings, and in 
all numbers. 

b. 9at4 and sahiiara are declined regularly, as neuter (or, rarely, 
in the later language, as masouline] stems of the same final, in all 

o. The like Is true of the higher nnmbers — which bare, Indeed, 
DO proper nnmeral character, but are ordinary nouns. 

486. Conatrnotlon. As regards their construction with the 
noons enumerated by them — 

& The words for i to /A are in the main used adjecdvely,- 
agreeing in case, and, if they distinguish gender, in gender also, with 
(he nouns: thus, daq&bMr Tirftif ""'^ ''" h^oei; ji devtt diyy 
ik&dafa Bth& (AV.) what eleven godt of you are in heaven; pafio&su 
j&nequ among the five trihea; oatasfbMr glrbbll^ tci'M four eongi. 
Rarely occur such combinations as diqa kal&qSnftm ^KV.) ten pitcher*, 
rtenfiih vat ;E.) six seatong, 

b. The numerals above IB are construed usually as nouns, either 
taking the numbered noun as a dependent geuItiTo, or standing in 
the singular in apposition with it: thus, ^atadi dfts^i or qatadi 
dlaiuKm a hundred tlavet or a hundred of tlaves; vi&qaty^ hirlbM^ 
teith twenty bagi; qaffyaih far&tBu iii 60 autumne; ifatiaa, p^&l^ 
teiOt a hundred fellera; qat&ih aab&sram KTutaih ny^bnd&ib ja< 
gh&ia qakrd d&aynn&m (AV.) the mighty [Indra] tlexv a hundred, a 
thoutand, a myriad, a hundred million, of demone. Occasionally they 
are put in the plural, as if used more adjectivelyi thus, pafic&^od- 
bhir b&QiU^ uiith Jifty arrows. 

o. lu the older language, the numerals for 5 and upward are 
sometime used in the nom.-acc. form (oi as if indeclinabty) with 
other cases also: thus, pafioa Iq^tf^u among the Jive races; Bapt4 
rfiqSifa of seven hi^ds; aah&aram ^Ibhl^ with a thousand barde; 
$at&ib pOrbbl^ with a hundred itrongholde. Sporadic instancea of a 
like kind are also met with later. 

487. Ordinal s. Oi the clasBes of derivative words 
coming &om the original oi cardinal aumerala, the ordin- 
als are by far the moat important; and the mode of their 
fonnation may best be explained here. 

ioy Google 

487—] VI. NUMBRALB. 184 

Some of tho first ordloAlfi are irregularly owde; thus, 

a< 6ka l forms no ordinal; instead is used praQuuaii (i. e. pra- 
tama foremo$t); ftdya (from Odl beprnting) appears fint in the Sutras, 
and &dim& much later; 

b. from dvk 2, and trl 3, come dTttlra and tftij* (seooudarily, 
through dvlta and abbreviated trlta); 

o. oatdr 4, f&f 6, and aapti 7, take the ending tha: thus, 
oaturtb&i fsfthd, aaptitha; but for feuriA are used also tnriya and 
torya, and aapt&tha belongs to tbe older language only; pa&oatha, 
tor /iflh, is excessively rare; 

d. tbe numerals for 5 and 7 usually, and for 8, 9, 10, add ma, 
forming pafioami, aBptai]i&, af(am&, navami, dafami; 

e. for nth to I9t/i, the forms are ekAda^A, dvSdaqi, and so 
on (the same with the cardinals, except change of aooent) ; but akK- 
da^ama etc. oocasiooally occar also; 

f. for the tens and intervening odd unmbers fh>m 50 onward, 
the ordinal has a double form — one made by adding tbe full [super- 
lative) ending tain& to the cardinal: thus, viAqatttami, trlA^attami, 
aqltltamii, etc.; tbe other, shorter, in a, witb abbreviation of the 
cardinal: thus, viA^A 90th; tTlikf& 30th; oatTftrlA9& iOth; paiLoftfi 
6olh; faft& eoth; aaptati ?Oth; aQitA 80th; navati 90fh; and so 
likewise ekaviA^^ 2iit; oatiutriAg4 34th; a)f\lMKtvirUii}k 4Sth; 
dTftpafioftq& B2d; eka^a^fd etsl; and ek&nnaviAfA and OnavlAqi 
and ekonaTifiqA /9M; — and so on. Of these two forms, the latter 
and briefer is by far tbe more cossmoD, tbe other being not quotable 
from the Veda, and extremely rarely from the Br&hmanas. From iOth 
on, the briefer form is allowed by tbe grammarians only to the odd 
numbers, made up of tens and units; but it is sometimes met witb, 
even in the later language, from the simple ten. 

g. Of the higher numbers, 9at& and aahAara form gatatamA aod 
aabaaratamii; but tbeir compounds have also the simpler form: thus, 
ekaqatd or ekagatatama lOUt. 

k. Of tbe ordinals, pratbami (snd Ura), dvltiyat tptiya, and 
turiya (with t^rya) form their feminine in i; alt the rest make it 

488< Tbe aidlntls, u in other langnagea, have other than erdlnsl 
offlcee to flit; and in Sanskrit especiillr they are general adjectives to the 
cirdlnale, with a conilderable Tartety of meanlnfs, as fraetlonala, a« lipilf 
fying compottd of to many porta or so-mam/-/old, or eetOatntng to many, 
or (b8 wm seen «l>0Te, 478) having to many added. 

a. In B fractional seme, tbe grammariaiu direct that their aeceat be 
shifted to the first sf liable: tbne, dviOynhalf; tftSytk thirdpart; O&tur- 
tba quarttr; and bo on. But in accented teita ouly t^lya Mini, and 
c&turtha ((!B.) and tArIra guart«r, are found lo treated; for half ooouis 

ioy Google 


only ardhi; and o«tnrth& (M8. et«.}, pttltoam&, and lo on, ue woented 
u in theii oidlnil ue. 

488. There are other nnmeral dsrivatiTeH: thus — 

a. moltiplicative itdverbs, ae Avia tviict, tria fhriet, oat6a four 

b. adverbs with the snffixeB dhft (1104) and 9H (1106): for 
example, ek»dhA »i one may, g&tadha in a hundred tcoya; ekaqaa 
one by one, i;ataq&s by hnndredi; 

o. colleotives, us dvftaya or dva^i a pair, d&fataya or dag&t 
a decade; 

d. adjectives liico dvlka eompoied of two, paiLoaka comitting of 
fiee or fives ; 

and BO on; but tlieir treatment belongs rather to the dictionary, or 
to the chapter on derivation. 



480. The pronouDs differ from the great mass of doudb 
and adjectives chiefly in that they come by derivatioa from 
another and a very limited set of roots, the so-called pio- 
□ominal or demonstrative roots. But they have also many 
and marked peculiarities of inflection — some of vrhich, 
however, find analogies in a few adjectives; and such ad- 
jectives will accordingly be described at the end of this 

Personal Pronouns. 

481. The pronouns of the frrst and second persons are 
the most irregulai and peculiar of all, being made up of 
fragments coining from various roots and combinations of 
roots. They have no distinction of gender. 

ioy Google 

Vn. Pronouns. 

a. Their inflectioQ in the 

later language is a follows: 

3d p 


m^, mft 

loUiyam, mft 
Ab. t^ 

m&ma, me 
L- Rfil 

N. A. V. qiEim^ 

r.D.Ati. WTT^UFl^ 

a.L. srraatH^ 

uidA.D.O. ^ 

A. WfHR TO^ 

asm&i, nas 

I. mwtPiti 


tr^, tvft 



t^bhytfnit te 


t4Ta, te 




yugnUtn, tm 

Digitizecy Google 

Fbbsonal Fronodns. 



asm&am, nas yuqmikam, vas 

asmisn yuftnasu 

b. The briefer second forms for accus., dat., and gen., in all 
numbeis, ure accentieaB; and hence the^ are not allowed to stand at 
the beginning of a sentence, or elsewhere where anj* emphasis is laid. 

O, But th«y may be quallQed by icceated sdjaneta, u adjecUTeB: e. g. 
te JAyata^ .0/ Ihte wh«n a conqueror, vo TTtabhyalb for you thai tcere 
confined, naa tribbyi^ to ut three (all RV.). 

d. The abUUie mat is tccentleiB In ons or two AV. pMiiges. 

482. Forms of the older language. All the forms given 
above are found also in the older language; which, however, has alao 
others that afterward disappear from use. 

a. Thus, vre 'Bud a, few timee the luBtr. alog. tva (only RV.; like 
manifa for manif&yfi); fnrther, tbe loc. or dat. ling. md (only VS.) 
ind tv6, and the dtit. or loc. pi. aam6 (which ia by fii the commonest 
of these e-formsj and yi^md: their Biial e la uncombinable (or pra- 
gl^lTai 198b). The TS. makas tirice the ace pi. fern, yl^m^ (as if 
yu^niMi were too diitinctiTely a masculine form). The datives Id bhyam 
are In ■ number ot cases written, and la yet others to be read as if written, 
with bhya, irlth Iota of the final naial; and in a rare Instance or two wo 
have In like manner aomika and yOfinaka In tbe gen. plural. The asnal 
retolotions of semlvawet to vowel are made, and are especially frequent in 
the forms of (he second person (tii&m tor trim etc.). 

b> But the duals, above all, wear a very different aspect euUer. In 
Veda and Brahmana and Sutia the nominatives are (with occasional 
exceptious) &vfcm and yaT&m, and only the accusatives Bivtia and ynTlun 
(but In RV. the dual forou ot Ist peis. chance not to occur, uulesi In 
viin[f], once, for Sv&m); the luetr. in SV. is eltber ynv&bhj&m (occurs 
also once In AfS.) or yaTtbhTftm; an abl. ynv&t appears once in RV., 
■nd a,vkt twice In TS.; the gen.-loc. Is In RV. (only) yuv6s Instead of 
ynT&yoB. Thus we have hero a distinction (elaewbere unknown) of five 
different dual cases, by endings In part accordant with those of the olfaer 
two nnmliers. 

488. Peculiar endings. The ending am, appearing in the nam. 
sing, and pi. (and Vedlo da.) of these pronouns, will be tonnd often, 
though only In ting., among the othsr proDDUns. The bhyam (or bTam) 
ot dat sing, and pi. ts met with only here; Its relationship with the 
bhrKm, bhyas, bhls of the ordinary declension Is palpable. The t (or 


4B3— ] TTI. Pronouns. 188 

d) of th« sb1., though hen pTseaded by > ihort lowel, I* donbtlei) the 
«une irilh that of the ft-decleoiioD of nonm md idJeollTca. Thit the nom., 
dat., and abl. ondinga ihonld be the same Id slag, and pi. (and Id part 
In the earllei da. alio), only the item to which they are added being dif- 
ferent, Is DDparalleled elaewheie la the language. The elenient nnti appear- 
ing tn the ploral forniB will be fonnd frequent In the inflection of the 
etngnlai In other ptonamlnal words: In bet, the compooDd stem uma 
which underllea the plnral ot ahAm Memg to be the game that fntnUhes 
part of the elngnlu forms of ayam (EOl), and ita valae of ice to be a 
apeclalliatloii ot the meaning thete ptrtoiu. The genltWes aingnlar, m&nut 
and tAvA, have no analogies elsewhere; the detlTatloD from them of the 
ndJeeHTes m&makn and tjlvaka (below, KlSb) ioggeete the poaelblUty 
of their betn; themseW^s stereotyped stems. The gen. pi., asmitkuil and 
yTiynftcBin, are certainly of tbls character: namely, nenter ring, cueforma 
of the adfect<Te Btema aamSka and ynfmfika, other cases of which are 
found In the Yeda. 

40^ Stem-farmB. To the Htndu gcammsriftiiB, the Btems of 
the personal pronoons kre road and a«mad, and trod and mfnuul, 
because these are forms nsed to a certain extent, and allowed to be 
indefinitely nsed. In derivation and compoaicion (like tad, kad, etc: 
see below, under the other pronouns). Words are thus formed from 
them even in the oldpr language — namely, loitkfta and "'^^■■*''* 
and aamitsakhl (RV.), tvkdsotd and matt&a (A7.), tvAtpltp and 
tv&dvtvftoana (TS.j, tviitpraautii and tvaAd»VKtyk and yuvad- 
devatyk and fnfmaddevatrh (QB). aamaddevatya (PB.j; but much 
more numerous are those that ahow the proper stem in a, or with 
the a lengthened to fi: thus, mivant; aamatr^ aamadr&h, etc.; 
tr&yata, trttvant, tritdatta, trSnid, tvitvasa, tr&ata, etc.; jvuf- 
m£datt«, Tu^miflta, etc.; ynvltTaiit, yuvjLka, TtiT&dlilta, yavt- 
datta, ynviulta, etc. An-I the later language also has a few words 
made in the same way, as m&drt;. 

a. The Tedaa haTe certain more Irregular romblnatlons, with complete 
forme: thua, tvitdilE&ma, trftm&buti, mftdipaqyA, mamaaat7&, asm^ 
hlti, abadipfkrT&, aluimiittar&, ahaiiiTii, abaihsana. 

b. From the stems of the grammarians come also the derivative 
adjectires madl^a, tradiya, asmadira jiifma^djra, having a pos- 
seseive value: see below, 6iea. 

e. For Bva and narfcin, aeo below, 61S. 

Demonstrative Pronouns. 
496. The simplest demoD strati ve, rT ta, whioh answers 
also the purpose of a personal pronoun of the third person, 
maf be taken as model of a mode of declension usual in 

ioy Google 

189 Deuonstrative Pbonoums. [ — 466 

•o manf pionouos and piODominal adjectives that it is 
iaitly to be called the genera) pronominal declension. 

a. Bnt this root hu also the special iiregulurity tbst in the 
nom. siog, luaac. and fem. it bu a^ (for whose peculiar euphonic 
treatment see 176a,b) and ai, instead of t&e and ta (compare Gr. 
o, is, JO, mnd Goth. «a, «o, tkaia). Thus: 

m. a. t. 

H. WI flfl^ HI 


a^ fiH 


t4m t&t 




























N. A. V. 

* s 


ttn tk 


I. D. Ab. 





0. L. 





■ Plnw 


ff fT#I 


M Utnl 



fiH mft 


tin ttni 







D,„i,i.o, Google 






b. The T«du show no othei Irreinluttlea of inflection thin theae 
nhlch belong to aJI Btoma in a &Dd S: nunely, Ud& iDmedmai; DsnaUy 
tt ht t&a, dn.; often tA lor titni, pi. nent.; ntnally Ubhis for tsfs, 
InstT. pi.; and the ordinary lesolutlona. The RV, haa one more cAge-roim 
from tlie toot aa, iiame1>- B&Bmin (occniring nearly half as oft«n as t^ 
mltl)i and OhU. has once BBBm&t. 

466> Tbe peculiarities of the general pronominal deolenaion, it 
wilt be noticed, are tliese ; 

A. In the singalar, the use of t (properly d) aa ending of nom.-ace. 
neutj the combination of anether element ama with tbe root In maac. and 
neat, dat., abl., and loc, snd of By In fern, dal, abl.-gen., and loo.j and 
tbe mass, and neat. loc. ending In, vhlch is rastiicted to thla declension 
(eiecpt in the anomilouB y&dtfmin, RV., once). The inbatilution in B. 
of U for fis as fern, ending (307h) was illuatiated at SSSd. 

b. The dual Is piecisely that of noun-atema in a and S. 

o. Ill the ploral, the irregnlsritieg are limited to tb tar tits in uom. 
masc, and the iDsertioa of s instead of n before Km of the gen., the stem- 
final being treated hefore it In the same manner as before an of the loc. 

497. The Htem of this pronoun is by the grammariaQS given 
na tad; and from that form come, in fact, the derivative adjective 
tadiya, with tattvfr, tadvat, tanmaya; nnd numerous conipoandfl, 
such as taoobtla, tajjfia, tatkara, tadonantara, tanm&tra, etc. 
These compounds are not rare even in the Veda: bo t&danna, tadWd. 
tadvai;&, etc. But derivativen from the true root ta are also many: 
Hspecially adverbs, as t&taa, t&tra, t&thft, tad^; tbe atyeotiveB tJE- 
vant and t4tl; and the compound tsdf^ etc. 

498. Though the demonstrative root ta ia prevailingly of the 
third person, it is also tVeelj need, both in the earlier language and 
in the later, as qualifying the pronouns of the 6rst and seoond person, 
giving emphasis to them: tbna, ad 'b&m, thit I, or I h«re; bA or aa 
tv&m. thou there; te vayam, tee A«r«; taayamama o/me Aere, taamlfia 
tvayi tn thte there, and so on. 

490. Two other demonstrative stems appear to contain ta aa 
an clement; and both, like the simple ta, subBtitute aa in the nom. 
sing. masc. and fem. 

ioy Google 

191 Dkhohbtrativi: Pronouns. [ — BOl 

m. The one, tya, is tolerably common (although only a third 
of Its possible fonna occnr) In RV-, but r&re Id AV., and almOBt 
noknown later, its Dom. ring., In the three genders, is ajiAs, ayd, 
tyit, and it nutkes the accna&tlves tyAm, tjtja, ty&t, and goes on 
throngh the remainlDg crhob in the same manner aa ta. It has in 
BV. the Inetr. fem. tyi (for ty&yji), loBtezd of ayft as Dom. sing, 
fern, is aleo fonnd ty&. 

b. The other is the osual demonstrative of nearer position, tht> 
here, and la In freqnent nse throngh all perlode of the laognage. 
It prefixes e to the siiople root, forming the numinatives eqis, e^a, 
«t4t — and BO on thiongh the whole in6ecdon. 

c. The stem tya haa ni'Uher couiponnda oor derivativea. But 
from eta are formed both, in the same manner aa from the simple 
ta, only much less nnmerone: thus, etaddj (QB.), etadarttM, etc., 
from the so-called stem etad; and et&d^ and eUEvuit from eta. 
And ofm, like aa (408), is used to qualify pronouns of the lat aud 
2d persons: e. g. e^K "ham, ete vayam. 

600, There is a defective pronominal atem, eoa, which is accent- 
tess, and hence used only in eituationa where do emphasis falls upon 
it It does not occur elsewhere than in the accusative of all numbers, 
the Instr. sing., and the gen--loc. dual: thna. 

Slog. A. enam enat enftm 

I. eaena etia;& 

Dd. a. enln ene e&e 

Q. L. enayoa enayoa 

PI. A. enSn en&ni en&a 

a, Th« BV. bu enoa iasteul of enayoa, and in ons or two InsUnces 
Mcsnta K form; tlin*, en^m* enJbC?). AB. aaee enat ilao as nom. neut. 

b. As ena ia alwi>ys osed Bubst&ntively, it hu more nettl; than ta 
the value of a tbtrd penontl pronoun, unemphatlc. Apparent eiamples 
of its adJeetlTal nae here &nd there met with are donbtlesa the reeult of 
confusion with eta (409 b). 

o. This Btom forms neither derivatives not componiids. 

BOl. The declension of two other demonstiatives is so 
inegulatly made up that they hare to be given in full. The 
one, mm Biyim etc., is used as a more indefinite demon- 
strative, this or that; the other, SI^ asSii etc., signifies 
especially the remoter relation, yon oi yonder. 

a. They are as follows: 

ioy Google 

VII. Pronouns. 

^^ V^ ^^ ^ ^^ ^^ 

i^m Idiun H&m tuOA ad&a aai&. 

Wl Wl^ Wi. ^fl. ^^ ^i^ 

IfnAfn id&in <TnjtT¥i 

i^mi^tT) ad&a nmrtTn 















































W. A. 

^ ^ 



im&fi im6 



I. D. Ab, 









5^ ^Rlft 


a« a^ 


Imi tmitnl 


ami amibil 



^^ 5*nft 


5^^ «(^ 


Im^ Imtni 


amAn amftnl 











D. Ab. 




























193 Dbhonstbatives. [—603 

b. The same foiau ue nied In the older Unguage, vlthont variation, 
except that (as nraal) ima ocean roi im&ii and ''nff#'><, and Ktai for 
amdiil; amnyft vhen iiaed adverbially 1b accented on the final, omurd; 
as&u (with aeceat, of oourge, on the fliat, 4b&u, oi vlthout accent, aafiu: 
314) la need alto as vooatlTe; ftml, too, oci-nra u vocitlve, 

602. a. The fotmei of these two pionoDoa, ay&m etc., plainly 9how« 
Itself to be pieced together from a iiDmber of defective etems. Tba majority 
of forms come from the mot a, vlth which, aa In the ordinary pronomlDil 
declension, bhul (f. By) Is combined in the slngnlar. All these forms fiOQi 
a have the pecnlUrlty that In tbelr substantive use they ate either accented, 
as In the paradigm, oi accentiess (like ena and the second forms from 
ah&m and tv&m). The remaining forms are always accented. From nn& 
come, with entire regularity, an^na. au&yS, an&yoB.^ The ationg oases 
In daat and plnral, and in part in singular, come not leas tegnlarly from ■ 
stem imit. And ay&va, ly&m, Id&m are evidently to be referred to ■ 
simple root 1 (iiUm being apparently a double form: id, like tad etc., 
with ending am). 

b. The Teda ha* frem the root a also the instrumentals ena and aylt 
(nsed In general adverbially), and the gen. loc. dn. ay6B; from Ima, 
im^Bya ocsars once In RV., Imasm&l in AA., and imSla and ime^u 
later. The RT. has In a email number of instances the Irtegnlar aocen- 
toation iamal, Aeya, abtiis. 

0. In analogy with the other pronooaa, Id&m la by the gram- 
mariaoB regarded aa repreBentative stem of this pronominal declen- 
sion; itnd it is actoallf fonnd ao treated in a very small number of 
compounds (Idamm&ya and Id&dirupa are of Brabmana age). Aa 
regarda tbe actual atema, ana furnishes nothing further; from Ima 
comes only the adverb im&tbft [RV., once); but a and i fumisb a 
number of derivatlveB, mostly adverbial; thus, for example, Atas, 
Atra. &tha, ad-dba (?|; IUb, id [Vedic particle], IdB, il^, ftara. im 
(Vedio particle], tdf9, perhaps ev& and ev&m, and others. 

603. Tbe other pronoun, aafi4 etc., hts amu for itj leading stem, 
which In the singular takes In combination, like the a-stema, the element 
ama (f. ay), and which shifCe to ami In part of the maso. and neat, 
plnial. In part, too, like an adjactivs n-stem. It lengthens its final in the 
feminine. The gen. stng. amu^ya is the only example In the Unguage 
of tbe ending aya added to any other than an a-stem. The nom. pi. ami 
is unique la form; its i is (like (hat of a dual) pragfhya, or exempt 
from conblaatlon with a foUowlug vowel (188 b). Aeftu and adAa are 
also without analogies aa regards their endings. 

a. The grammarians, as usual, treat ad&a. as representative stem 
of tbe declenaiOD, and it ie found in this character in an extremely 
small number of worda, as adomnla; adom&ya is of Brahmana age. 
The QB. bas also aaftunaman. But moat of the derivativea, as of 

Wtaitn.y On.n.i.u». S. »d. 13 

DijiiLc. ..Google 

the cues, come fromamu: thus, am^tas, amutra, amdthfi, amudl, 
amurbi, amuv&t, amuka. 

b. Ill the oilier Ungaage ocean tbe root tvB (icceiitleta), meaning 
one, many a one; It U oftene«t fonnd tepetted, u one and aaotier. It 
fallows tbe ordiDsry pionomin*! decleasian. Fiom II Is uisde the (alBO 
iceentlese) adiatb tvadftidm (MS.). 

o. Fri|menta of mother ilemanstratlTe root or two are met with: thus, 
dmos he occurs in ■ tormuU in AV. and in Bishmsnu etc; avdfl ss 
gen, -Ion. dual is tOnnd in EV.; the peiticle a poiuti to a root u. 

Interrogative Pronoun. 

604. The characteristic part of the interrogative pro- 
uomiaal root is ^k; it has the three forms ^ ka, 1% ki, 
sR ku; but the whole declensioDa] inflection is from ^ ka, 
excepting the nom.-acc. sing, neut., which is &om 1% kf, 
and has the anomalous form 1^71 kJm (not elsewhere known 
in the language from a neuter i-stem]. The nom. and 
accus. sing., then, are as follows: 












ision is 

precisely like that o£ FT 

ta (above, 498). 

a. The Veda bu its n«ual varlstlona, ki end k^bhis tor kiai and 
k&Is. It ntsa has, aloug with kfin, (he pronominal I y tegular oenter k&d; 
and k&m (or kam) l9 a frequent particle. The masc. foim kia, oorreB- 
ponding to klm, occurs as a atereotypud faae in the comhtnatlens n&kis 
aud mltklB. 

BOB. The grammarianB treat kim aa repreBentative stem of tbe 
interrogative pronoun; and it is in fact so used in a not Urge nnmber 
of words, of which a few — klmm&ra, kbhkara, klifakHmya, kiifa- 
devata, klihQlli, and ttie peculiar kimyu — go bacl( even to tbe 
Veda and Btahmana. In closer analoj^y with the other prononna, the 
form kad, a couple of times in the VeUa (katpayi, k&dartha), and 
not Infrequently later, is found as first member of compouTtde. Then, 
from the real roots ka, ki, ku are made mnnj derivatives; and 
from kl and ku, eBpccially the latter, many compouDds; thus, kfcti. 


Ids ' Relatives. [— B09 

luith^ hath&m, kadJt, katar&, lutaiii&, k&rhl ; kiyant, kid^ ; kUtas, 
kntra, kutaa, krk, knoarii, knkarman, knmuitria, etc. 

BM. VarlouB forms of this pronoun, u kad, kirn, and ka land 
rarely, ko), at the beginning of componndi, have passed from an 
interrogative meaning, through an escliuiiatory, to the value of pre- 
fixes signifyiDg an unuBual quality — eitlier something admirable, or, 
oftener, something contemptible. This use begios in the Veda, but 
becomes much more common in later time. 

507. The interrogative prononn, as in other languagus, turns 
readily in ita independent use also to an exclamatory meaning. 
Moreover, it is by various added particles converted to an indefinite 
meaning: thus, by oa, oanik, old, 4pi, vi, either alone or with the 
relative ya (below, Bill prefixed: thus, k&g obii& any one; nk k6 
'pi not am/ dim; yitnl klini oit tckaUotvgr\ yatamiit katam&a oa 
whatever one. Occaaiooally, the luterrogBlive by itaeif acquires a 
similar value. 

Relative Proneun. 

'60S. The loot of the relative pronoun is TJ ya, 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. 

609. It is inflected with entire regularity according to 
the usual proootninal declension: thus, 

Sinsular. Du»l. Plunl. 

m. n. f. m. n. f. m. d. t. 

N. jjB|^ TifT^ UT 1 & mft Tira^ 

y&s y&t 7± I nt !T Q T^ y^i y^ 

A. xp\^ iXH^ oitT j y*"^ y^ y^ hr^ trrf^ nra 

jkta ykt y^ I yan y&ni yits 

ytaa yAya nT^TTm ^^ y^bh^ 

D. OT^ n^ ya y m ^.jjq^ UP^jq^ 

yimi&l y&sy&l y^bhyaa yabhyas 

a. The Ved* ihovB ita ueu&l latiationa n( these forjiis: yS to 
•ml for yini, and ydbhls Tot y&{B; y6a lor yayoa also occnra 
yteB, wilb prolonged flnal, U In KV. twir^ as common ne ytaa. 

D,j,i,...., Google 

609—] VII. PKOHOnuB. 196 

latloQB occur In y^biu, *ad yifosm and yisMun. The eoninnotioD 
yat Is in ftbUtiTe foim ucotdlng to the ordinate declanilOQ. 

610. The use of y&t as representative stem begins very eaily: 
we have y&tkSma in tbe Veda, and yatkBrln, yaddevatya in the 
Biahmans; later it grows more general From the proper root come 
also a considerable series of derivatives: y4t(w, y&tl, y&tra, y&th&, 
y&dfi, ykdi, y&rhi, jivant, yatar4, yatamii; and tbe componnd 

611. Tbe combination of ya with ka to make an indefinite 
pronoDD baa been notioed above (fi07). Its own repeUtion — aa 
y&d>yat — gives it sometimes a like meaning, won through the dis- 

Sia, One or two marked pecnliarltes in the Sanskrit ose of the 
relative may be here briefly noticed: 

a. A veiy decided pieference for putting the relative clansa before 
that to which ll relates: thus, y&h BOnvatA!^ S&kbft t&Bmft fndrfiya 
gSyata (RV.) who M the friend of the loma-pruaer, to that Indra ting ye; 
y&di yajli&di paribbfir ^i ■& Id dev^^n gaoetaaU (RV.) tchat Buffering 
thou proieeteet, that in truth goeth to the godt\ yi trlfapt^ parly&Dti 
b&l& t4f&di dadhStu me (AV.) tohat thrice eevan go oiouf, their ttrength 
may he attign to me; as&d y6 adhariid grhia t&tra aantr arSyyal^ 
(AV.) what houte is yonder in the depth, there let the wttekee he; sata& 
y&D me istl t^na [TB.] along with that toAtcA it mint; baAsiaSdi 
vaoanaiii yat tu tan madi dahatl (MBh.) hut what the wordt of the 
swana were, fiat buvTu me; aarvaBya looanadi QOatraifa yaaya nK "aty 
aadha eva aa^ (H.) who doe* notposieet learning, the eye of everything, 
blind indeed ii he. The other amngement, thoagh, frequent enough, la 
notably leas usual. 

b. A frequent conveiaion of the tubject or objaot of a verb b; an 
added relative into a snbttantlve clauie: thui, mi "milh pri '^at p&4- 
nifeyo Tadli6 y4h (AV.) may there not reach him a human deadly 
weapon (lit'ly, what ie such a weapon); pfiri no pahl y&d dhAnam 
(AT.) protrct of ui what wealth {Utere is]; ap&iiiSrB6 'pa ruftr^t^ 
k^etriyiifa fap&thaq oa y&fy. (AV.) may the cleansing plant cleanse 
ateay the diteate and the curse; pu^are^a Iqtaih r^yadi yao oK 
*ayad vaau kiihoana (UBh.) by l^uhkara was taken away the kingdom 
and whatever other property [tAara wa(i\. 

Other Pronouns: Emphatic, Indefinite. 

618. a. The isolated aud uniafleoted pionominal woid 
HQH avayarii (from the root ava) signifies self, own aiif. 
By its form it appears to be a nom. sing., and it is often- 

ii.zecy Google 

197 Phonominai, Dbrivatitbb, [—-816 

est used as QominatiTe, but along with wotds of atl peisone 
and numbeis; and not seldom it lepiesents other cases also. 

b. Svaynm ie also need as a stem in oompositioD: thus, sra- 
yathji, sra^ambbli. Bat sva itself (osually adjective: below, Bide) 
has the same value in composition; and even its inflected forms are 
(In the older language very rarelf) used as reflezive pronoun. 

o. In RT, alone are found a few examples of two indefinite 
prononns, sama (accentless) any, every, and slmA every, all. 

Nouns used pronominally. 

514> a. The noun KtmAn loul is widely employed, in the siU' 
gular (extremely rarely in other numbers), as reflexive pronoun of all 
three persons. 

b. Tbe noDQ tanA bo^ li employed In the ume manner (bat In ell 
numben) la tbe Vsds. 

0. The a4jeative bhavant, f, btaavati. Is used (as already pointed 
out: 4IM) In respectful addreas as snbBtltute for the pronoun of 
the second person, TtB couHtruction with tbe verb is in accordance 
with Its true character, as a word of the Ihiid person. 

Pronominar Derivatives. 

616. From pionominal loots and stems, as well as from 
the la^ei class of loots and from noun-stems, are foimed 
by the ordinary suffixes of adjeotive derivation certain words 
and classes of words, which have thus the character of pro- 
minal adjectives. 

Some of the more important of these may be briefly noticed here. 

614. Poasessives. a. From the representative etema mad etc. 
are formed tbe adjectives madlya, asmadlra, tvodlra, yaqmaiUTa, 
tadlyn, and etadiya, which are used in a possessive sense: relating 
to me, mine, and so on. 

b. Other posse ssives i 
t&Taki, from tbe genitives 

O. An 4iialogoas deilittlTe from the genltlTe am^^ya Is ftmn^yj^ 
ya^i (AT. etc.) de*e»n4<tnt of eueh and meh a one. 

d. It W4B pointed oat ebOTe (498) that tbe "genltlTCB" asmakam 
end jofinjtkam ire really itereetyped caies of posieMive adJectlTet. 

ioy Google 

6ie— ] VII. Prohodkb. 198 

e. Correspond lag to BTay&m (618) is the poasesBivo bt&, meaning 
own, aa relating to all persons And numbers. The RV. baa once the 
correspondiDg simple poasesBivc of the second person, tri thg. 

f. For the uie of Bvft as lefleilTe pioDOnn, aee aboTe, BlSb.^ 

g. All these words fonn tbeli feminlneB Id &. 

b. Other deiliBtivea of a like value hare no claim to be mentlonBd 
here. But (eieeptlng BTCi) the poasesslTes are so rarely used as to make 
but a sDisU figure In (he language, vhloh piefera generally to Indicate the 
poBSMsive relation by the geniUte case ot the prODOnn Itself. 

S17. By the suffix vaat aro formed from the pronominal roots, 
with prolongation of their final vowels, the adjectives movant, trit* 
vant, yu^mitTant, yuvavaiit, tarant, etitvant, yavant, meaning of 
my tort, like me, etc. Of these, however, only the last three are in 
use in the Ittter language, in the sense of lantue and qaantut. They 
nre infiected like other adjective stems in rant, making their feml- 
ninea in vatl (462). 

a. Words of similar meaning from the roots 1 and ki are {yant 
and klyant, inflected In the aame manner; aee above, 461. 

618. The pronominal roots show a like prolongation of vowel 
in combination with the root d^ »te, iook, and its derivatives -dr^a 
and (quite rarely) dfkfa; thus, mad^, -dfqa; tvadpq, -dffa; yu;- 
m&dfq, -dr9a; t&d^, -dfi}a, -d^k^a; et&dtq, -dfqa, -d^kfa; ylUlff, 
-d^a; id^, -d^a, -d^kfa; kidfg, -d^a, -d^rk^a. They mean «/ my 
^ort, like or reuembling me, and the like, and tftd^Q and the following 
are not uncommon, with the sense of lali$ and qualit. The forma in 
drf are unvaried for gender; thoao in dfqa (and drk^a?] have fe- 

618. From ta, ka, ya come t&ti lo many, k&tl koto numyt ji$i 
as many. They have a quasi-nnmeral character, and are Infiected 
(like the numerals piiflca etc.: above, 4BS) only in the plural, and 
with the bare stem as nom. and accns.: thus, N.A. tAti; 1. etc. t&tl- 
bhia, t&tlbbyaa, t&tiDbn, t&tiqu. 

830, From ya [in V. and B] and ka come the compsrativee and 
superlatives yatard and yat&m&, and katar& and katami; and from 
i, the comparative {tara. For their inflection, see below, 538. 

631. Derivativee with the auffix ka, sometimes conveying a 
diminutive or a coDteraptuous mciining, are roade from certain of the 
pronominal roota and sterna (and may, according to the gr^immarians, 
Im made from them all): thus, from ta, takim, takiit, tak^s; from 
aa, ntHsi,; from ya, yak&a, yak^ yak4; from aasA, aaak&u; from 

amu, ftwrt iilra.. 

8. For the numeroaB and freqnanlly used adverbs formed from pio- 
nomiual roots, soe Adverbs (belo?, 1007 ff.). 

ioy Google 


Adjectives declined pronominally. 

622. A number of adjectives — some of them coming 
from pronominal roots, others mote ot less analogous with 
pronouns in use — are inflected, in part or wholly, accord- 
ing to the pronominal declension (like FT ta, 496], with 
feminine stems in S. Thus: 

623. The comparatives and BDperUtives from proDOminal roots 
— namelj, katar& and kntami. yabuA and yatomi, and ftara; 
Also ajiyi olh^r, and Ub comparative anyatarft — »re declined like 
ta througliout. 

a. Bnt even boia these voHb fonns mide ici'aidlng to tbe sdjective 
declciuion are (paiulically met with (e. g. itarfty&in K,]. 

b. Anya ttkes ooculonally thu foini anyat Id compoiltloD : thug, 
anyfttkBma, anyatsth&na. 

624. Otber words are so inflected except in the nom.-aco.-voc. 
sing, neut, where they fa»ve the ordinary adjective form am, instead 
of the pronominal at (ad). Snch are B&rva atl, vfgva aU, every, 
4ka on«. 

a. These, also, its not wlthoDt exception, >t letit in the etiller 
lanfuge [e. g. vi^Kya, v{fvat, vjgre KV.; 6ka lot. sing., AV.). 

626. Tet other words follow tbe same model usually, or in some 
of their ai^i&catioDB, or optionally; but in other senses, or without 
known rule, lapse into tbe adjective inflection. 

a. Snch ne the corapuidTes and inperlatlieB from piepositlonal stema: 
Adliaia uid adtiam&, &ntara and 4iitama, 4para and apami, &vara 
and avam&, uttara and uttami, upara and upam&. Of theae, pro- 
nominal forms are dscldedly more nameiDue tiom the compatatlves than 
tiom tbe BupeTlatiyes. 

b. Further, the «aperUtlTes (fllthont conesponding comparativea) 
paramA, oaramA. madbyamA; and also anyatama (whoae poaldve and 
eomparatlye belong to the clasa first mentioned : 528). 

O. Further, tbe word* p&ra dittant, other; ptilva, prior, east; dikqiija 
right, touth; paqoima behind, tcettern; ubhAya (f. ubhiiyi or ubhayi) 
of both kinds or parties; nAma the one, half; and the poaaesalva svA. 

626. OecailoDal forms of the pronominal declenslen are met with from 
numeral adjective!!; e. g. pTatbam&syAs, trUyasyftm; and from other 
wordi having au Indefinite numeral character: thna, &lpa feie\ ardta4 half; 
kAvalaotf; dvftaya o/ <A« two kinde; hiixsa, outeide — and others. RV. 
baa once saminismSt. 

ioy Google 




627. The subject of ooDJugation oi verbal infleotion 
involves, as in the othei languages of the family, the dis- 
tinctions of voice, tense, mode, numbei, and person. 

A. Further, besides the simplei oi ordinary conjugation 
of a verbal root, there are certain more oi less fully de- 
veloped secondary or derivative conjugations. 

828. Voice. There are (aa ia Greek) two voices, active 
and middle, distinguished by a difference in the personal 
endings. This distinction is a perrading one: there is no 
active personal form which does not have its corresponding * 
middle, and vice versa; and it is extended also in pait to 
the participles (but not to the infinitive). 

628. An active form is called by the Hlndn gntmm&rianB 
parasm&l padam a icord for another, and a middle form is called 
fttmane padam a word for one'a self: the terms might be beet para- 
phrased by trantitwe and reflexwe. And the diatiDOtii>D thuH expressed 
is doubtlesB the original foundation of the difference of active and 
middle fonnsi in the recorded condition of the language, however, 
the antithesis of transitiTe and reflexive meaning is in no small 
measure blurred, or even altogether effaced. 

a. In the epics Iheie Is mncb etbcement of the dliHnctloQ between 
■ctlTB and middle, the choice of voice being veiy often dUeimined b]i 
metrled consideration b alone. 

680. Some verbs ate 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 oi in both; of 
a veib 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. 

'. oyGoogle 

20t Tenbb ahd Hodb. f— 633 

6S1. The middle £oim8 outside the present-syetem (for 
which there is a special passive iaflectioD: see below, 768 ff.), 
and sometimes also withia that system, aie liable to be 
used likewise in a passive sense. 

682. 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 : ■. simple; b. reduplicated; o. sigmatic 
or sibilant; 6. a future, with 7. a oondidonal, an augment- 
tense, standii^ to it in the relation of an imperfect to a 
present; and 8. a second, a periphrastic, future (not found 
in the Veda). 

a. The tenses here distingiiiflbed (io aecordance with pTeTsiling 
naagej as imperfect, perfect, pluperfect, and aorist receive those 
names fh>m theii corraspondenoe in mode of formation with tenses 
so called in other Ungnsges of the &mily, eBpecittllj' 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 — Dor of perfect time, except in the older language, 
where the "aorist" has this value i later, imperfect, perfect, and aorist 
are BO many nndlBortminated past tenses or preterite: see belov, 
under the different tenses. 

833. Mode. In lespeot to mode, the difference between 
the classical Sanskrit and the older language of the Veda 
— and, in a less degree, of the Bt&hmaniae — is especially 

a. 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 lees frequent occurrence, as belong- 
ing to the perfect; and they are made also from the aorlsts, being 
of especial frequency from the simple aorist The future has no modes 
Ian occasional case or two are purely exceptional). 

b. In the classical Sanskrit, the present adds to its in- 
dicative an optative and an imperative — of which last, 

ioy Google 

583—] VIII. Conjugation. 202 

moteoTer, the first peisons aie a lemnant of the old sub- 
junctive. And the aorist has also an optative, of somewhat 
peculiat inflefltios, usually called the piecative (or bene- 

584. The piesent, 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. 

636. Tense-systems. The tenses, then, with their 
accompanying modes and participles, h.\l into oettain well- 
marked groups or systems; 

I. The present-system, composed of the present 
tense with its modes, its participle, and its preterit which 
we have called the imperfect. 

II. The perfect-system, composed of the perfect 
tense (with, in the Veda, its modes and its preterit, the 
so-called pluperfect) and its participle. 

III. The aorist-system, or systems, simple, re- 
duplicated, and sibilant, composed of the aorist tense 
along with, in the later language, its "precative" opta- 
tive (but, in the Veda, with its various modes and it« 

IV. I'he future-systems: 1. the old or sibilant 
future, with its accompanying preterit, the conditional, 
and its participle; and 2. the new periphrastic future- 
636. 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 ate supplied from the subjunctive. 

ioy Google 

203 Vbrbal Adjeotivb8 and Nouns. [—1^0 

6S7. Verbal adjeotives and douds: Participles. 
The paitioipleB beloogiiig to the tenBe-eystems have been 
already §poken of above (684). Theie is besides, coining 
directly &om the lOOt of the veib, a participle, prevailingly 
of past and passive (oi sometimes neuteij meaning. Future 
passive participles, or gerundives, of several different for- 
mations, are also made. 

638. 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-in&nitive sense; 
most often in the dative case, but sometimes also in the 
accusative, in the genitive and ablative, and (very raiely) 
in the locative. In the classical Sanskrit, there remains a 
single infinitive, of accusative case-form, having nothing to 
do with the tense-systems. 

680. Gerunds. A so-oalled gerund (or absohitive) — 
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 frequent 
in the later language, where 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. Another gemnd, an adverbially used accasative in form, ie 
found, bat only rarely, both earlier and later. 

640. Secondary conjugations. The secondary or 
derivative conjugations are as follows: 1. the passive; 2. the 
intensive; 3. the desiderative; 4. the causative. In these, 
a conjugation-stem, instead of the simple root, underlies 
the whole system of inflection. Yet there ie clearly to be 
seen in them the character of a present^ystem, expanded 
into a more or less complete conjugation; and the passive is 


640—] Vin. CoNjnaATioN. 204 

BO puiely a preBent-system that it will be desoiibfld in the 
chapter devoted to that part of the iafleotion of the veib. 

a. tinder the same general head belong! 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. 

641. 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, rather than to the pure 
root, the personal endings are appended. 

a. In this chapter will be given & general account of the per- 
Bonal endings, and also of the formation of mode'Stems from tense- 
stemB, and of {bose elements in the formation of tense-stemB — the 
aagment aod the reduplication — which are found in more thfui one 
tense-aystem. Then, in the following chapters, each tense'System 
will be taken up by itself, and the methods of formation of its Btems, 
both tenBe-atems and mode-atema, aud their combination with the 
endinga, will be deecribed and illuBtrated in detail. And the com- 
plete conjugation of a fen model verba will be eshlbited in syate- 
matic arrangement in Appendix O. 

Personal Endings. 

643. The endinge of verbal inflection are, as was pointed oat 
above, different throughout in the active and middle voices. They 
are also, as In Greek, naually of two aomewhat varying forms for 
the same peraon in the aamo voice: one fuller, called primary; the 
other briefer, called aecondary. There are also less pervading differ- 
eneea, depending upon other condltiona. 

a. In the epica, exchanges of prlmuy »nd wcondirf tctlTo endlnp, 
(eipeddlly the BubetUutloii of ma, va, ta, far mas, vas, tha) ire not 

i, Google 

205 Pebsonal Endihob. [— M5 

b> A condeDBSd atttement of til the vutletlM o( endtog toi ewb pei- 
son and nnmbet here folloirs. 

643. Singular; First peiaon. a. The prim&ry ending in 
the active is ml The Bubjanctive, liowever (later imperative), liaB 
ni instead; and in tiie oldest Veda this ni is Bometimes wanting, 
and the person ends in K (as if the ni of ftni were dropped}. Tbo 
secondary ending is properly m; but to this m an a has come to 
be so peraisteutly prefixed, appearing regnlaily where the teDse-stem 
does not itself end in a [Tam for Tarm or vanun in RV., once, and 
abhiim HS., avadlilm T3. etc., sanem TB., are tare anomaliesX that 
it is convenient to reckon am as ending, rather than m. Bnt the per- 
fect tense lias neither ml nor m; its ending is simply a (sometimes 
ft; S48o]-, or, from ft-roots, fto. 

b. The primary middle ending, according to the analogy of the 
other persons, wonld bo regularly me. Bat 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 corresponds 1 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 ftl for e. 

644, Second person, a. In the active, the primary ending 
is Bi, which is shortened to e as secondary; as to the loss of this 
B after a final radical consonant, see below, &&B. Bat the perfect 
and the imperative desert here entirely the analogy of the other 
forms. The perfect ending is invariably tha (or thft; 248o). 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 ma- 
jority of verbs (including all a-stems, at every period of the language) 
no ending is present, but the bare stem stands as personal form. 
Id a very small class of verbs (722-3), ftna is the ending. There is 
also an alternative ending tfit; and this is even used sporadically in 
other persons of the imperative (see below, fi70-l). 

b. In the middle voice, the primary ending, both present and 
perfect^ is ae. The secondary stands in no apparent relation to this, 
being thSa; and in the imperative is fonnd only sva (or avS; 248 o), 
which in the Veda is not seldom to be read as sua. In the older 
language, ae is sometimes strengthened to a&l in the subjunctive. 

546. Third person, a. The aetive primary ending is ti; the 
secondary, t; as to the lose of the latter after a final radical con- 
sonant, see below, 66S. Bnt in the imperative appears instead the 
peculiar ending ta; and in the perfect no characteristic consonant is 
present, and the third person has the same ending as the first. 

b. The primary middle ending is te, with ta as corresponding 
secondary. In the older language, te is often strengthened to tftl in 


646—] Vm. COKJBOATION. 206 

the subjunctive. In the perfect, the middle third person hu, like the 
active, the enme ending with the Gret, namely e simplj; and In the 
older langu&gc, tbc third peruon present also often loses the distinctive 
part of its tcrmioation, ami comes to coincide in form with the first 
land HS. has adnha for adngdha). To this e perhaps corresponds, 
as secondary, the i of the aorlst 3d pers. passive (848 ff.). The Im- 
petatlTe has t&m (or, in the Veda, rarely Am) for its ending. 

640. Dual: First person. Both in active and in middle, the 
dual first person Is in all its varieties precisely like the correspond- 
ing plural, only with substitntion of v for the m of the latt«r: thus, 
vaa fno vaei has been found to occur), va, vahe, vabi, vahSl. The 
pprsoQ is, of coarse, of comparatively rare nse, and from the Veda 
no form in voa, even, is quotable. 

547. Second and Third persons, s. In theactive, the primary 
endlog of the second person is thaa, and that of the third is tas; 
and this relation of th to t appears also in the perfect, and runs 
t!irou(;b the whole series of middle endings. The perfect endings are 
priniHTy, but have u instead of a as vowel; and an a has became so 
persistently pre6xed that their forms have to be reckoned as athna 
and atus. The secondary endings exhibit no definable relation to 
tho primary in these two persons; they are tarn and t&m; and they 
arc used in the imperative as well. 

b. Id the middle, a long a — which, however, with the final a 
of a-stems becomes e — has become prefixed to all dnal endings 
of the second and third persons, so as to form an Inseparable part 
of them (lUdhltliSm AV., and jlbltham pB., are isolated anomalies). 
11)0 primary endings, present and perfect, are fttlie and Ste; the 
secondary (and imperative) are Btbjbn and SUbn (or, with stem-final 
a, etli« etc.). 

o. Tbe Rig-V«(U hu a very few forms In jUthe uti Sit«, appuently 
from athe and ete with iDbjanctive atrangtheDiDg (ttey are all detailed 
below: Bee 616, 701, 737, 762, 880, 1008. 1043). 

648. Plural: First person, a. The eariiest form of the 
active ending is maai, which in the oldest language is more frequent 
than the briefer maa (in RV., as five to one; in AV., however, only 
us three to four). In the classical Sanskrit, mas is the ezolnsive 
primary ending; but the secondary abbreviated ma belongs also to 
the perfect and the subjunctive (imperative). In the Veda, ma often 
becomes mft '248 o\ especially in the perfect. 

b. Tho primary middle ending is mabe. This is lightened In 
the secondary form to matal; and, on the other band, it is regularly 
liu the Veda, not invariably) strengthened to mah&l in the Bubjnnctive 

549. Second person, a. The active primary ending is tba. 
The secondary, also impciativo, ending is ta {in the Veda, tK only 

Digil.zecy Google 

207 Perbonal Endikos. [— SSO 

once Id impv.). But In the perfect an; characteristic canBonaut is 
wanting, and the ending is simply a. In the Veda, the syllable na, 
of problematic origin, is not infrequently added to both fonjis of the 
ending, making thaua (rarely thanft) and tana. The forma in which 
this oconrawill be detailed below, under the different formaUons; the 
addition la very rarely made excepting to persons of the first general 

b. The middle primary ending is dbve, which belongs to the 
perfect as well as to the present. In the subjunctive of the older laQ' 
gnage It is aometimeB strengthened to dhvU. The sncondary (and 
imperalavej ending is dhTam (in RV,, once dhva); and dhv&t is 
once met witb in the Imperative (S71 d). In the Veda, the t of all 
these endings is aometimee to be resolved into u, and the ending 
becomes disayllabic. As to the change of dh of these endings to 4h, 
see above, SS6o. 

660. Third person, a. The full primary ending is onti in 
the active, witli ante as corresponding middle. The middle second- 
ary 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 (307); the ending is an. In the imperative, 
(uita and «nt&m take the place of anti and aote. The initial a of 
all these endings is like that of am in the Ist sing., disappearing 
after the final a of a tense-stem. 

b. Moreover, anti, antu, ante, ant&m, 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 non-a-stems (and 
after a few roots which are treated as if reduplicated: eSBS.i; in the 
middle, it occurs after all tense-stems save those ending in a. 

0. Further, for the secondary active ending «a there is a sub- 
stitute us (or ur; 16&b; the evidence of the Avostan favors the 
latter form), which is used in the same reduplicating verbs that 
change antl to ati etc., and which accordingly appears as a weaker 
correlative of an. The same as is abo used universally in the per- 
fect, 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 root- 
stems ending in ft, and a few others (621). 

d. The perfect middle has in all periods of the language the 
peculiar ending re, and the optative has the allied ran, iu this per- 
son. In the Veda, a variety of other endings containing a r as dis- 
tinctive consonant are met with: namely, re (and Ire) and rate in 
the present; rata in the optative both of present and of aorietj; 
rite in the perfect; ranta, ran, and nun in aoristB (and in an im- 
perfect or two); r&m and ratfim in the imperative; ra in the imper- 
fect of dub (MS-)- The three rate, ratftm, and rata are found even 
in the later language in one or two verbs !6&d). 

Diaii.zecy Google 

6B1— ] VI. Conjugation. 208 

651. Below are given, for convenience, to UbnUr form, the 
Bohemes of endings ae accepted in the clafiBical or later language', 
namely, a. the regular primary endlnga, used in the present indicative 
and the future (and the subjunctive in part) ; and b. the regolar 
secondary endiags, used in the imperfect, the conditiooat, the aorist, 
the optative (and the subjunctive in part); and farther, of special 
schemes, o. the perfect eudings (chiefly primary, especially in the - 
middle); and d. the imperative eudings (chiefly secondary). To the 
so-called imperative endings of the first person is prefixed the B which 
is practically a part of them, though really containing the mode-sign 
of the subjunctive from which they are derived. 

662. 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 unaccented 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 onaccented endings (this being a characteristic 
of the subjunctive formation which they represent). 

S63. The schemes of normal endings, then, are as follows: 
«. Primary findings. 


m6a « vJJie 



tbk ai ithe 



&iiti, ku u, iu> 

date, Ate 

b. Secondary Endings. 


mi. I. k v41il 


t4ta '■ 

' "• t& thoB atbSin 



'-, << ^ Os t& atfim 
c. Perfect Endings. 

intft, &ta, r&o 


m& e Tibe 

tn&fae , 


& Bi i,the 

dhv6 X 


UB e' ite 


d. Imperative Endings. 

1 &ni SiVa ama ill Svotiai ftmaliSi ^ 

2 'Stti, bi. — t&m t& av4 ath&m dhv&m 

3 tu t^ Antu, &ta tiim atftm &iit&m, &t&m . 

E64, Id general, the rule ia followed that an accented ending, it dls- 
syllalilo, iB accented on Its Ant ejrllable ~ and the oonstant nnlon-vowels 
aie regarded. In this respect, aa integral parts of the endlnga. But the 

Diail.zecy Google 

209 Febsonui EKDraoH. [—667 

3d pL ending ate ot the piea. Indio. middle liu in BY. the acMDt tA6 in 
■ nombM of Tub* (lee 618, 686, 660, 718); and m oeoulonU InsUnee 
U met with in othei eDdiugi: that, m&h^ (tee 718, 73B). 

C66. The leeondaiT endings of the seoond uid thlid pettODi dnfolat, 
u eonalittng of en tdded conioaant wlthoat vowel, shoald legnUdr (IBO) 
be lost wheoerer the loot oi item to irhich they eie to be added Itself ende 
in a contontnt. And thli rule ia In geneial followed; yet not without es- 
eeptioni. Thna: 

a. A Tool ending in » dentU tnnte aomeUniM diopa tUa Snal mnte 
Instead of the added fl in the second person; and, on the other band, a root 
OT stem ending In B sometinieB drops this B instead of the added t in the 
tUid penon — in either ease, establisblng the ordinary relation of ■ and t 
in these persons, instead of b and b, and t and t. The examples noted are : 
2d sing, ftvae (to 8d sing, avet), Yvid, AB.; 3d ling. akat, ykr, ^B.; 
agbat, Vghao, JB. AfS. ; acakSt, yoakUi, RT.; aqfit, VqtM, AB. HBh. 
R.; aarat, ywraa, VS.; ahinat, i^hiAB, (B. TB. OB. Compare also the 
B-aorlst forms ayfiB and btSb (146 a), in which the same Inflaence is to 
be seen; sod fnrthei, t^Utetc. (888 a), and preoatlTe r&t fotyBB (637). 
A similar loss of any other Bnal conMnant is escesslvely rare; AV. hae 
onee ablianaB, for -nak, f/hhalti. There are also a few eases where a 
In sing. U Irregularly modeled after ■ Sd sing. : thu, atr^am (to atf^at), 
ytfi, S.V., aoohlnam (to acohlnat), yohid, UBh. ; compare farther 
the 1st sing. In m Instead of am, 643 a. 

b. Again, a union'TOWel is sometimes Introdnced before tbe ending, 
either a oi 1 or I: see below, 6B1 b, 631, 818, 880, 1004 a, 1068 a. 

a. Id a few isolated cases in the older langaage, this I ie changed to 
U: tee below, 804 b. 836, 1068 a. 

666. The ohangeB of form which roots and stems undergo in 
their combinatioDS with these eodingB will be pointed out iu detail 
below, under the varioua formations. Here may be simply mentioned 
in advance, as by far the most important among &em, a disUnction 
of stronger and weaker form of stem in large classes of verba, stand- 
ing in relation with the accent — the stem being of stronger form 
when tbe accent falls upon it, or before an accoutiesa ending, and of 
weaker form when the accent is on the ending. 

a. Of the endings marked as accented in the scheme, the ta of 2d pi. 
la not infreqnently In the Veda treated as nnaoeented, the tone reating on 
the stem, wbleh la strengthened. Unch less often, the tam of Id da. is 
treated in the same way; other endings, only sporadically. Details are given 
nnder the Tariona formations below. 

Subjunctive Mode. 

6B7. Of the subjnnctive mode (as was pointed out above) only 1 
fragments are left in the later or olasdcal luignage: namely, in the i 

WliltDST, Omamar. 3. ei. 14 I 

ioy Google 

B67— ] 



Bo-called fint pereonB imper&tiTe, and in the nee (679) of Iha imper- 
fect and aorist perBons without augment after mi probibitive. Id 
tbe oldest period, however, it was a very freqoent formation, being 
three or four times as common aa the optative in the Big- Veda, and 
nearly the same in the Atharrao; but already in tbe Br&bmanaa it 
becomes comparatively rare. Its varieties of form are conaiderable, 
and sometimes perplexing. 

658. In its normal and regnlar formation, a special mode-stem 
is made for tiie subjunctive by adding to tbe tense-stem an a — which 
combines with a final a of the tense-stem to 3. The accent rests 
npon the tense-Btem, which accordingly has tbe strong form. Thus, 
from the strong preeeot-stem doh (|/duli} is made the snbjnnctive- 
Btem d6hai from Jah6 tyh-a), Juli&va; from yim^ iVToi), yimija; 
from 8un6 (v'BU), BuuiiTa; from bh&Ta (i^hfl), bhivS; from tudA 
(>^tadj, tndi; from uoy& (pass., v'Tae), noya; and so on. 

S68. The stem thns formed is Inflected in general as an a-stem 
would be inflected in the indicative, with constant accent, and S for 
a before the endings of the first person (73S 1) — but with the follow- 
ing peculiarities as to ending etc.: 

660. a. Id the totive, tbe lat dog. has nl la eudlog: thus, dohltiu, 
yun&jBui, bh&T£aii Bat in the Rig-Teda somettmes S slmpl;: thns, 
kyA. br&Tft. 

b. Id lat do., let pi,, and 2d pi., the endings aie the seconduy : thus, 
dohKva, d6ltfima, ddhau; bbivSTa, bh&vftnia, bh&T&n. 

O. Id 3d and 3d du. and 3il pi,, the endings are piimaiy: thna, 
dohatbaa, d6Itatas, d6liattia; bbiivftthaB, bh&TStas, bb&vfitba. 

d. In 2d and 3d sing., the endings are either primary or aecoDdaiy: 
thns, dohaal or d6ItaB, d6hati or d61kat; bh&TOsl or bh&T&a, bh&T&tl 
m bh&vKt. 

e. OcoasioDally, forms nlth double mode-algn ft (by aasImilatioD to 
the more numeroQB eubJuaetlTea from tenie-stemg (□ a) ue met nlth from 
nan-a-stenu; thus, iafttha from as; Ayfts, 4y&t, iyfiu from e (yi). 

661. la tbe middle, forms with seoondaiy instead of primary encl- 
Ings are very rare, being found only iu the 3d pt. (where they are more 
freqnent tban the primary), and in a case or tvo of the 3d alDg, (aDd AB. 
has once aayftthfiB). 

a. The striking peculiarity of sabJunctiTe middle iDHectlon is the i^ 
qnent strengthening of e to fil in the endings. This is less general in the 
very earliest language than later. In lateing., Ki alone Is fotud as ending, 
even In RV, ; and in 1st du. also (of rare occniience), only &v&h&l la met 
with. In 1st pi,, ajnabftl preyalls In BV. and AV. (ftmahe li found a 
few times), and is alone known later. In 2d eing., sfti for ae does 
not oecnr in BV., bat Is the only form in AV. and the Biahntanas. In 
3d slug., tU for te occors once lu BV., and ia the predominant form 

Digitizecy Google 

211 Sdbjunotitb Mode. [— 68S 

In AV., aad the oa\j one liter. In Sdpl., dllTii foT dbTe It found In 
one woid In RT., uid t few times In the BAhnunu. In 8d pi., nt&l 
for iit« is the Brihrnua form (of fu fiam Sequent occnnence); It occurs 
neither In RT, nor AT. No inch dail endings ss th&i and t&i, for tbe 
•nd te, sre uiywliere fonnd; bnt RV. bis in i tew words (nine: abore, 
647 o) Utbe ind ftlt«, wUoh ippeii to be a Ube lubjnDctiTe itreDgtfaeulng 
at etbe and ete (slthongli found In one IndlcitWe form, ki^T&ite). Be* 
fore the Sl-endlngs, the vovel it regularly long &; bat antSi Instead of 
ftntU ii two or three time! met with, and once or twice (TS. A.B.) atfil 
for it&i. 

663. The subjunctive endingB, then, Id combination with the 
BubjonctiTe modc-nign, sie as followB: 

l&TAhU /AmahU 
IKvahe iKmahe 

/aal foae fadhve 


attaan attaa { Uthe 

(ati . lata ... fante, anta 

a. And in farther oombinatlon with final a of a tODBe-Btem, the 
initial a of all thoBO endings becomes &: thus, for example, in 2d pers,, 
aai or So, athas, fitha, Sse, &dhve. 

068. Besides this proper subjunctive, with mode-elgn, in Its triple 
form — with prlmuy, with strengthaned primary, and with leeondary end- 
ings — the name of subjanctlva, in the forms ^'imperfect sobjuncttve" and 
"improper tnbJancUTa", his been also given to the Indieatlve forms of Impar- 
feot and aorist when used, with the augment omitted. In ■ modal sense 
(below, 687): rach nse being quite common in RT., but rapidly dying out, 
so that in the Brahmana language and later It la birdly met with except 
after m5 prohibitive. 

a. As to the general uses of the lubJunotlTe, see below, 674 IT, 

Optative Mode. 

664, a. As baa been already pointed out, tbe optative is of com- 
paratively rare Dccnrrence in the language of the Yedae; but it gains 
rapidly in frequency, and abeady In the Brahma^aa greatly out- 
numbers the sabjuDctlve, and still later comes almost entirely to take 
its place. 

b. Its mode of formation is the same in all periods of the 

666. A. The optfttive mode-sign is in the active voice a dif- 
ferent one, according as it is added to a tense-Btem ending in a, or 


Dgil.ze.., Google 

666—] VUL CoNraaATioM. 212 

to one ending in lome other finkl. In the Utter eue, it is ylt, Moented; 
tUB yR is ftppendBd to the weaker form of the tenao-item, and taket 
the regular series of secondary endings, with, in 3d plnr., as in- 
stead of an, and loss of the a l>efore it. After an arBtem, It is i, 
nnacoented; this 1 blends with tiis 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. oa), by means of an Interposed 
enphonlc y. 

b. In the middle voloe, the mode-sign is I thronghont, and Uikm 
the secondary endings, with a in Ist ning., and ran in 3d pi. After 
an a-stem, Uie rules as to its combination to e, the accent of the 
latter, and its retention before- a vowel-ending with interposition of 
a 7, are the same as in the active. After any other final, the weaker 
form of stem is taken, and the accent is on the ending (except in 
one class of verbs, where it falls npon the tense-stem: see 646); and 
the I (as when combined to a) takes an Inserted y before the vowel- 
en dinga [a, AtUbn, Stam). 

O. It Is, o[ eoane, ImpoMlble to tell from the form whethei i or i 1* 
comlslned with th« flntl of in a-item to e; bat no good reuon •ppeua to 
eiiat for uiaming i, r&thar thui the I whiBh showi Itaelf In the other clags 
of Btetni In the middle Tolce. 

660. The combined mode-sign and endings of the optative, then, 
are ai follows, in their double form, for a-Btems and for otfaere: 
a. f«r Bon-a-ateBB. 

—tr. T- ■■ <"■ 



yItTR ytmn fci IvAhl 
yHtam yitta ifhiU ^thSm 
yitam y6B. it4 Mttam 







eva ema eya erntii 


etam eta ethfis ^tham 



•tain Ua eta ^jbtam 


C, The yfi is la the Tedii not leldom leiolTed Into iS. 

d. The contMotod aanem, foi eaDeyam, it fonnd In TB. uid Apast. 
CeittiD Vedio 3d pt middle formi In rata will be mendaaed below, under 
the vailoai foimatlont. 

667. Precatlve. Precative forms are Bnoh as have a sibi- 
lant inserted between the optatlve-Bign and the ending. They are 
made almost only &om the aorist stems, and, though allowed by the 
grammarians to be formed from every root— -the active preoative 
from the simple aorist, the middle from the sibilant aoiiit — are 

ioy Google 

213 Optativb Modz. [—670 

prustioally of nre occDrrenoe at every period of the language, and 
especiall; later- 

B. The Inierted B todb In tbe 4CtlTe through the vholo Boitea of per- 
>oi»; Id the middle. It li Ulowed only In tbe 2d and 3d penoDS elng, md 
dn. and the 2d pL, and Is qnotible onlr for the 2d and 3d Bing. In the 
2d Jing. act., the preutlTe form, by leuou of the neMBiuy loss of the added 
B, It not diftlngnitbtble from the almple optatlTe; in the 3d ting, act., the 
aame la the cue la the later langoage, irhlcb (oompua 666 a) aavei the 
penonal ending t Initead of the procatlTe-Blgn a; bat the RT. aiually, and 
the other Vedlo text* to some extent, haie tbe propei eodiog yfts [for 
7iat). Ab to 4I1 In the 2d pi. mid., see 336 o. 

b. Tbe accent 1b bb In the ilmple optative. 

MS. The precative endtogB, then, accepted In the later lanpwge 
;inclading, In brackets, those which are identical with the simple 
optative), an as follows: 

actiTe, middle. 

B. d. p. 8. d. p. 

1 yaaua yiava yiUma [iy&] [Iv&hl] [ImAhi] 

1 [yan] yintam yiata Iqthia ly&tbSm l^vAm 

1 [yif] ylbt&m yiboo i^ iy&t&m [ir&n] 

a. SeipecttDg the precative, Bee farther 891 IT. 

b. As to the general ueee of the optative, see below, 573 IT. 

Imperative Mode. 

669. The imperative has no mode-sign', it ia made by 
adding its own endings directly to the teoBe-stem, just as 
the oAei endings aie added to fona the indicative tenses. 

a. Hence, in 2d and 3d dn. and 2d pi., ita Ibmu are Indlstlngnlsbable 
I with its angmeDt 

b. The iDlee u to the nee of the different endings — espeetallv In 
2d elng., where the variety U coUBlderable — will be givan below. In connee- 
tioD with the variens lenBe-ayatenis. The ending tfit, however, has ao moch 
that 1« pecDiiar to Ita ose tlut it cbUb for a Ultle eipUnaHon here. 

670. The Imperative in t&t. An imperative form, usually 
having the value of a 3d pera. elng., but sometimes also of other per- 
sona and numbers, is made by adding t&t to a present tense-stem — 
In its weak form, if tt have a distinction of strong and weak form. 

a. Eiamplea are: brtltfit, bat&t, wlttltt; plp^tftt, Jahitst, 
dliattat; kr^aMt, knnitfit; gr^M^tat, J&nitit; &vat&t, r&kfat&t, 
vaaat&t; vl4;at&t, s^jatftt; asyatSt, naqyatit, ohyatSt; krlyatftt; 

Digitizecy Google 


gftmayatfit, oy&vaytttBt, v«ra;sUt; IpMt&t; JigftOt. Ho esimplw 
h>Te been found fiom & naul-cluB yeah (690), dot tny other thui thoM 
bete given from ■ peitlve, intenjiTe, oi deaidan^Te. The few aeoented 
caiet indiMte that the foniiatlon folloiri the fenetal rule foT one made with 
ail accented ending (662). 

b. The impeiatlTe in t&t is not a Ter; »Te rormatlOD in the older 
language, being made (Id T., B., and 8.) tram about flft; loota, and in 
toTard a hundred and fifty occnrreuces. Later, it ii very unufnal: thu, 
only a elngle siimple haa been noted in HBh., and one in R. ; and corree- 
pondlngly few in yet more modern texta. 

S71. As regardB its mesniog, this form appears to have pre- 
TailiDglf Id the Br&hmarLu, and tracesbly bnt mach less diatinctlf in 
the Vedio texts, a specific tense-value added to ks mode-yalue — as 
signifying, namely, an injunction to be carried oat at a later time than 
the present: it is (like the Latin forms in to and tots) a posterior 
or future imperative. 

a. Eiamplee are: ihti 'vi ma tlq^bantam abhyehi ti briilii 
tadi tu na igatSifa pratipr&br&tftt (9B.) *ay to her "coma to meat I 
stand jiul here," and [^aftertcard] announce her to ue a* having come; y&d 
urdhT&B tfqtbi drivi^e 1i& dhattat (RV.) when thou thalt atand up- 
right, \(hen\ bettow rtchei here (and limilarly in many cases); utkfilain 
udvalk6 btaavo 'duhya pr&tl dtaftvatftt (AT.) 6s a carrier up the aeeent; 
after having carried up, run hack again; vimaap&tiT &dhl tvfi Sth&syRtl 
t&B7a vittSt (TS.) the tree will ateend thee, [then] take noU of it. 

b. Examples of its nse as otlsr than 2d slug, are ae foUovrs: islaing., 
avyuf&ifa JKgft&d ah&m (AV, ; only cue) let me tcatch till day-break; 
as 3d sing., punoT m^ "vifatSd rayl^ (TS.) ^^ viealth come again to 
me, a7&m ty&aya rl^ift mOrdhlUaih vi pBtoyatfit (^B.) the king here 
ekall make hit head fly off; u 2d da., naaatya? abruvan deva]^ 
punar tt vahatSd itl (RV.) the gode said to the two A^ini "bring tliem 
back again^; as 2d pi., ipal^ . . , d«v^|a naib euk^ bratat (TS.) ye 
Koters, announce us fo &e gods as well-doert. In the later laugaage, the 
prevailing value appear* to be that of a 3d sing.: thus, bbav&n praaftdadi 
koTutftt (HBh.) may your toorMp do the faoor, anaih btiavta 
abtalrak^atfit (DEC.) let your exeeUency protect him. 

a. AocordlDg to the natlTe grammarians, the Imperative in tftt la to be 
used with a benedlotlve Implication. No Initance of such nse appears to 
be quotable. 

d. In a certain passage repeated severa]. times in different Brahmanas 
and Sutras, and containing a number of forms in t&t naed as 2d pi., 
TSrayadhvKt Is read instead of T&rayntfit in some of the texts (K. AB. 
A^S. <;<;S.). No other occnrrence of the ending dhvSt has been anyvhete 

itizecy Google 

Uses of thb Modes. 

Uses of the Modes. 
672. Of ihe thiee 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 oi injunction — an attempt at the exercise 
of the speaker's will upon some one or something outside 
of himself. 

a. This, howevei (in Sanskrit as in other languages), is by no 
means slvays of the same force; the command Bhades off Into a 
demand, an exhortation, an entreatf, an expresBion of earnest desire. 
The ImpeTBtiye also BOmetimes eignifies an assumption or conceBaion ; 
and occasionally, hy pregnaat constractioD, it becomes the ezpTcssIon 
of something cooditionBl or contingent; but it does not acquire any 
regular uae in depend ent-oUnae-maki ng- 

b. The ImpflratiTe Is nov and than used in an InUrrogatlte seDteDCe: 
thua, bravihi ko 'dySi Va mayft vlTi^yatfim. (R.) *peak! who thall 
rtoto be ttparaltd by mef katbam ete gu^vantah krlyantam (H.) 
how are they to be made viiiuouet kasm&i pi^^Ui^ pradlyatAm (Vet) 
to whom shall the offering be gioent 

573. The optative appears to have as its primary office 
the expTession of wish or desire; in the oldest language, 
its prevailing use in independent clauses is that to which 
the name "optative" pioperly belongs. 

a. Bat the expression of deBire, on the one hand, pasBcs naturaily 
over into that of requMt or entreaty, bo that the optative becomes 
a softened Imperative; and, on the other hand, it cornea 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 Itkely or usual, and ho 
becomes at last a softened statement of what is. 

b. Further, tbe optaUve in dependent clanses, with relative 
pronouns and co^jnnotlons, becomes a regular means of expreasion 
of the condiHonal and contingent, In a wide and increasing variety 
of use*. 

O. The Bo-MllBd pTeeittie foimi (667J ue oidintTil; med in the 
proper opIatiTe leuie. But In the Uter language they are occailonall; met 
with in the othei Die* of the optative : thus, na hi prapa^yftmi mamft 
^anudyftd yao ohokaum (BbG.) /or I do not perceive what ekould diipel 
my grief; jnA bhuytonr vibhataya^ (BhP.) that there ehculd be 
changea. Also raiel; with mS: lee 678 b. 

ioy Google 

674—] Vin. COKJUaATIOK. 216 

674. The subjunctive, as has been pointed out, becomes 
nearly extinct at an early period in the history of the 
language ; thete are le£t 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 
unaugmented forms (670), with the negative particle ITT mi, 
in a prohibitive or negative imperative sense. 

a. And the geDeml value of the subjnuctive from the beglniiing 
was what those relics would seem to indicate: its fundamental mean- 
ing is perhaps that of reqaiBition, less peremptory than the imperative, 
more so than the optative. But this meaning is liable to the same 
modiflcationB and tranBitions with that of the optative; and sub- 
junctive and optative run cloeely parallel with one another in the 
oldest language in their use in independent clanaes, and are hardly 
distingnlshable in dependent. And instead of their being (as In Greek) 
both maintained in use, and endowed with nicer and more distiDctive 
values, the Bubjnnctive gradually disappears, and the optative assumes 
alone the offices formerly shared by both. 

676. 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 leas exchangeable with one another, and com- 
binable in coordinate clauses. 

a> Thus, In AV., we have in impv.: qat&ib. Jiva ^ar&dal^ do 
thou live a hundred autumna; ubli&ii tl& Jlvat&ih jar&daf^ let &«m 
both live to attain old age; — in Bubj., ady& jiTfinl let m» liv4 Oiia 
day; qat&ih Jivfttl qar&da^ lie ehalt live ahtoidred autumns; — in opt, 
jivema far&d&di qatani may we live Aundreda of aulumiu; airvam 
iyxtx JIvyiEaam (prec.) I teould fain Uve out my whole term of life. 
Here the modes would be interchangeable with a hardly perceptible 
change of meaning. 

b. Examples, again, of different modes in cotirdlnats construction 
are: ly&m ague nitn pitiih vldeffa . . . navfiua pntrin m&hi^ 
bhavStl gatvi p&tidi sabhigft v{ ri^atu [AY.) may thi* wmnaa, 
'0 Agni! _find a ipotue: giving birth to tont the thali become a chief' 
fatnttt; having attained a tpouee let her rule in AopjWMw; KOpftyi 
nah BTast4;« prabi^dhe na^ punar dada^ (T3.) watch over ut for 

ioy Google 

217 Uses or the Modes. [ — B78 

our a^^fm■»; grant unto ut io wakt again; myta na3f Bfinali . • . Bn te 
uumaXii bhatv aami (BV.) may Hurt be io ut a ton; Ut that favor 
of tkine be our*. It iB not veiy Beldam the cue that versloDs of 
the ume pBHSge in different tezta ahov different modee as varloaa 

0. There is, in &ct, notliing In the earliest employment of tbeee 
model to prove that they might not all be specialized luea of forms 
originally equivalent — having, for instance, a general future moaoiog. 

&7fl. As examples of the less charaoterlatic use of subjunctive 
and optatJve in the older language, in Independent clauses, may be 
quoted the following: it shS ti gaooh&n nttarK yug^ (BV.) Ihote 
tattr agte teitl doubtlett eomt; y&d . . . ok marft {tl m&nyaae (RV.) 
if thou lhinke»t "I thatt net die" ; n4 til nagantl n& dabhStl t&akara^ 
(BV.) thtg do not become lost; no thief eon harm them; k&smfil derliya 
liav{^vidheiaa(RV.] to what god thall wa offer eblatioat asndlS rayfm 
a^navat . . . div6-dive [BV.) by Agni one may gain taealth efery day; ' 

utftl "nfiifa bralim&9e dadyKt t&thA Byonlt qivS ayKt [AV.] one 
thould give her, however, to a Srahman; in that eale the will be propitious 
and favorabh; ^bar-ahar dadj^t (^.) <m» should give every day. 

677. The obos 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 (MBb.) 
will be enough to Illustrate them: uoclii;taiii nftl Va bhaSJiy&di na 
kaiTilh pidadbtTanam I wiU not eat of the remnant of the tacrifiee, 
I wiH net perform the foet-lavation; jfifttm vrajet let her go to her 
relative! ; nfil Vaiii sK barhloit knryllt the thould not act thut at any 
time; kathadi Tidy&di nalBith nppam how can I know king Nalaf 
ntaai^e sadigaya^ sy&t tu Tindetft 'pi tokbadi kvaoit but in case 
of her abandonment there may be a chance; she may aleo find happiness 
somewhere; katbaih vfiso -rikarteyaifa na oa budhyeta me prly& 
how eon I cut off the garment and my beloved not wakef 

678. 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, divySva let u* two 
play; klih karavK^l te what shall I do for theet 

S79. The imperative negative, or prohibitive, Is from the earliest 
period of the language regularly and nsaally expressed by the particle 
mt with an aagmentlees past form, prevailingly aorist 

R. Thus, pr& pata mi tik raifaatlift^ (AV.) Jty away, do not stay 
here; dvlf&A^ oa m&byaih radhyatu mi o& li&ih dvi;at6 radluua 
( AV.) both let my foe be subject to me, and let me not be eubjeet to my foe ; 
niv hfyfim Abbayadi jyotlr indra mi no dlrghi abb{ naqan 
tamlsrt);! (RV.) I would win broad fearlets light, O Indra; let not the 
long darknetee* come upon ut; mi na i^nh pr& mofl^ (BV-) do not 

Digil.zecy Google 

679—1 Vm. COHJlliAl 


tteal away our Ufa; sanil^vveihi mi ^aoab (MBfa.J bt tiomforUd: do 
not grieve ; m& bhftifl]^ ot bhfti^ [MBb. R.) dot not be afttad; mft bhut 
kUaaya paryayal^ (R.) let not a change of time take place. EsampleB with 
the imperfect are: ma bibber ni. marlfyasi (RV.) do not fear; thou ailt 
not dig; n^ Bm&l 'tint fli^hin kurutbftti (AY.) do not make friendt 
of them ; mfi putram tmatftpy&tbih (HBh.) do not sorroui for My ion. 
The reUtiou of the imperfect to the aoriet conBtrnotiOD, in point of 
freqaenc;, is in BV. about as one to five, Id AY. atill less, or abont 
one to six; sod though instaaces of the imperfect are quotable from 
all the older texts, they are exceptional and infteqaent; while in the 
epicB and iater they become extremely rare. 

b. A BlogU opt&tlve, bhujema, Is need prohibitively with ma in 
RV. ; tha older Uognige preaenti no other example, uid the coDBtmctlan 
ie very lue aleo later. la en eiemple or tvo, ileo, the preeative {bhOT&t, 
R. PeSc.) follows ma. 

o. Tht RV. ha« once appiirently ma with an Imperative; bat tha 
passage is piobably corrupt. Mo other Buch cue ia met with In the older 
langnege (unlets sppa, TA. L li; doabttess a bad reading for afpaa); but 
in the epics and later the eonatmction begtni to appaar, and becomee an 
ordinary torm of prohibition: thai, mfi prayaoohe "9vare dhanam (H.) 
do not beitow trealth on a lord; aokhl mSi 'vaih vada (Vet.) friend, 
do not ipeak that. 

d. The (B. (li. 5. 1'] appesie to olTer a single example ot a Crne 
subjunctive with mi, n{ padyftaii; there Is perhaps something wrong 
iboot the reading. 

«, In the epics and later, an aorlst form not deprived of augment ii 
ocoaaionally met with after mfi: thus, mfi tvftdk k&lo tyagfit (HBh.) 
let not the timepaei thee; ma v&llpatliam anv agfi^ (B.) do nolfoUoui 
V^lCe road. But the same anomaly occurs also two or three times in the 
cider langnage: thus, vyiipaptat (9^.), agfia (TA,), anRi^at (K8.). 

B80. Bat the use also of the optative with ni not in a prohibitive 
sense appears in the Veda, and iMComes later a familiar construction; 
thus, n& ri;yema kad£ oan& [BY.) may ica suffer no harm at any 
time ; ak oK 'tisf J^n n& Juhnyftt (AY.) and if he do not grant permierion, 
let him not taciiffce; t&d u t&thil nk kuryfit (^B.) but he muei not 
do that 10; na dlvfi gayita (^S.) let him not sleep by day; na trfidi 
vidyuT Jaafil^ (HBh.) let not people knoie thee. This in the later 
language is the correlative of the preBCriptive optative, and both are 
eitreuelf common; so tliat in a text of prescriptive character the 
optative forma may come to outnumber the Indicative and imperative 
together (as is the case, for example, in Hann). 

S81. In all dependent constructions, it is still harder even Id 
the oldest language to establish a definite distinction between sub- 
junctive and opUtive; a method of nae of either is scarcely to be 
found to which the other doea not furnish a practical eqalvaleot-— 

ioy Google 

219 Uses op the Uodes. [—681 

and thoD, In the later language, anch wna sre repreBOnted by the 
opUtaye slone. A fev ex&mplee will be eafficleDt to illuBtrate this: 

a> After relative proDOQDs and ooDJQDctlooe in general: ji 
vyfl^ar yaq oa nUn&ih vyaedhia. (RV.) tchieh have shone /orth ihith- 
erio}, and tehieh thai! hereafter ihin» forth; j6 'to Siy&tB. aamjkadi 
B& 6ko 'aat (TS,) whoeeer *hall be bom of her, let him be one of ut; 
yo Tfti tin vidyit piaty&k|aifa b4 brahm^ v^ditK ayfit (AV.) 
tehoever thail know them fac* to fact, he may poet for a knowing priest; 
pUtr^Slh . . . jataBBdi JanAyftf oa yan ( AV.) of tone bom and tchom 
Mow mayett bear; yisya . . . ititbix grhia Sg&oohet (AV.) to whote- 
loever houee he may come ae guett; yatamAtbA kSm&yeta t&tha kury&t 
(Qfi.) in whatever way he may choose, eo may ha do it; y4rlii h6tS yija- 
mftnaoya n&oa gr^U^^t^ t&rhl brOy&t [TS.) uAeh the laerijicing 
priest ihail name the name of the offtrtr, then he may speak ; BTorupaiii 
yada dra^tum ioobetbJU? (HBh.) when thou shall desire to see tJiine 

b. In more distinctly coodiUonal cone true tione : y&j&ma devan 
y&di 9akniTfiins (BV.) we will offer to the gods if we ihaU be able; y&d 
ague eylba aUiifa tv&ifa tvUb vfi gha syn ab&di syn? t« satjit 
ihi "({^a^ (BV.) if I were thou, Agni, or if thou wert I, thy withes 
should be realised on the spot; yo dyam atlB&rpftt par&atSn ait b& 
muoyStai T&nujaaya r^fialji (AT.) though one steal far away beyond 
the sky, he shall not eteape king Varuna; y&d &iifif vftn upav&Mt k^- 
^ih-niTMh ayftd y&d aqmyad rudrd 'eya pa^liii abhi manyeta (TS.) 
if he should continue without eating, he would starve; if he should eat, 
Sudra would attack his cattle ; piirtbayed yadi mSih keloid dajj-dya^i 
■a me pumftn bbavet [HBh.] if any man soever should desire me, he 
should suffer punishment. These and the like conatmetioDB, with the 
optatiTe, are rery common in the Brahmanae and later. 

c In final clanees: y&th& 'b&ih ^atrnho "sSnl (AV.) that I may 
he a slayer of my enemies; gr^&ni y&th& pibfttbo Ai^ilh^h (RV.) that 
being praised wiOi song ye may drink the draught; urftu y&thft tAva 
f&rmam m&dema (BV.) in order that we ry'oice in thy wide protection; 
&pa JSnlta y&tbe 'y&m pdnar Sg&oobet (^B.) contrive that she come 
back again; kfpfiib kuryftd yatha mftyl (UBb.) so that he may take pity 
on me. This Is in the Veda one of the most freqaent ueea of the 
subjunctive; and in its correlative negative form, with n6d in order 
that not or lett (always followed by an accented verb), it cootinnes 
not rare in the Br&hmanaa. 

d. The IndlOkUie ia also very commonl]' used tn RntX cliniet after 
yathfi: thai, ;itb&] *7kii pum^ 'DtArikfam anuo&ratl ((B.) in order 
that this man may traverse the atmosphere ; yathS na vigbnab kriyate 
(R.) so that no hindrance mag arise; yatba 'yaili naqyatt tathft vidhe- 
yam (H.) ■( must be so managed that he perish. 

Digitizecy Google 

881—] VnL COMJUOATION. 220 

e. Vitli tbe eonditloQ&I me of inlijaDcUTe and opUtiTe Ls fnithei ttt 
be cempaied that of the M-eallod cODdittanal tenie; Bee below, 860. 

f. As li indicated b; many of the examples given above, It is naoal 
in a conditional eentenoe, oontalnlng piotaiii and ipodosls, to emplo; always 
the same mode, whether subjauotiie or optative (oi oondltional}, in each 
of the two oUueas. For the older language, this Is a lole well-nigh or 
qnlte without exception. 

682. No distinction of meaning baa been ettftblished between 
the modes of the preseDt-stem and thoae (in the older langnage) of 
tbe perfect and aoriat-aystems. 


688. PaiticipleB, active and middle, aie made &om all 
the tenee-Btems — ezoept the periphiastic future, and, in 
the latet language, the aoiist [and aorist participles aie lare 
&om the beginning). 

a. The psTtidples aaconnected with the tense-ayetems aie treated In 
obap. XUI. (863 ff.). 

684. The general participial endings ate QtT ant [weak 
foim 9(7 at; fem, Qrft anQ oi Qrf) Ml: see above, 448) fot 
the active, and 9H Bna (fem. ?1HT 5nB) foi the middle. But — 

a. After a tense-stem ending in a, the active paiticipia] aafGx 
is virtually nt, one of tbe two a's being lost in the combination of 
Btem-final and sufSx. 

b. After a tenge-atem ending in a, the middle participial suffix 
IB mSna instead of &na. But there are occasional ezoeptlone to the 
mle as to the use of m&na and &iui respectively, which will be 
pointed out in connection with the variona formations below. Such 
exceptions are especialiy frequent in the cansativer see 1043 f. 

o. The perfect has in the active the peculiar suffix vfifia (weakeat 
form n^, middle form vat; fem. u;l: see, for the inflection of this 
participle, above, 4S8ff.). 

d. For details, as to form of stem etc., and for apecial exceptlona, 
see the following chapters. 


686. The augment is a shoit Q a, prefixed to a tense- 
stem — and, if the latter begin with a vowel, combining with 
that vowel irregularly into the heavier or vfddhi diphthong 

ioy Google 

221 AudKHNT. [ — B87 

(136 a). It is always (without any exception) the accented 
element in the verbal form of which it makes a part. 

a. In the Tedk, Uie angmaQt la In & few toxmt long f. thai, ixiA%, 
avar> Kt);^, avr^ak, Kvldtayat, Syunak, iyulcta, iyok^fitSiD, 
tri^ak. IbUk, (ud y&a ta tvldliat, ST. 11. 1. 7, 9?]. 

586. The angmeDt la a rign of put time. And an aagment- 
preterit is made &om each of the tense-Btems from which the system 
of coiyi^B^o^ <s derived: namely, the imperfect, from the present- 
stem; the plnperfect (in the Veda only), from the perfect-stem; the 
conditional, from the f\itnre'item ; while In the aorist snob a preterit 
stands withont any correspondiDg present Indicative. 

K87. In the early langoage, especially in the BV-, the occarrence 
of forms identica] with those of sngment-tenses save for the lack of 
an augment is quite frequent Such forms lose in general, along with 
the augment, the specific character of the tenses to which they belong; 
and they are then employed in part Don-modally, with either a pres- 
ent or a past sense; and in part modaily, with either a sabjunotive 
or an optative sense — especially often and regolarly after mfi pro- 
hibitive (676) ; and this last mentioned ose comes down also into the 
later language. 

a. In RV., the tngmentleis forms are more th*n hUt m common u 
the togmeoted (abmit 2000 and 3300), ind are made (lom the present, 
perfect, and aorltt-syitems, bat eonsldenblr over balf from the aoriat 
Tbelt non-modal and modtl use* ue of nearl)' eqtul freqaenoy. The tense 
valae of the non-modally uasd foims is mora often psst than present. Of 
the modall; used foima, nearly a third are constrned with mt piohibltlTe; 
the reit have tvloe at often an optative as a proper inhjunctlTa value. 

b. In AY., the numerical relations ate Tery different. The augment- 
less forms are leas tban a third at many as the angmented (about 476 to 
14(iO), and are pieialllDgly (more than four flfUks) aoristlc. The non-modal 
uses aie only a tenth of the modal. Of the modaily oied forms, about 
foor fifths ate construed with mS ptohiblHTe; the test are cbieD; optative 
tn Tslne, Then, la the langaage of the Biihmanas (not Inclndlng the 
mantea-miteiial which they contain), tbe loaa of augment la, save in 
oceaslonal sporadic cases, restricted to the prohfbitiTe eonitrUBtion irith mfi; 
and the same continues to be the case later. 

o. The acoentnatlon of the angmsntless forms is thioDghont in accord- 
ance with that of unaDgmented tenses of similar formation. Examples will 
be given below, under the Taiions tenses. 

d. Besides the angmentless sorlat-forms with ma prahibttive, there 
are also found occistonally in the later language augmeutless imperfeet-forms 
(totj rarely sorist-forms), whieh have tbe same Talna as if they were aug- 
mented, and are for the most part examples of metrical license. They ue 
especially fteqnent In the epies (whence some seores of them are quotable). 

Digitizecy Google 

688—] Vni. Conjugation. 222 


688. The deiivatioQ of coajuga,tional and declensional 
stems from roots by i^duplication, either alone ot alon^ 
with other fotmative elements, has been already spoken of 
[260], and the foTmations 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 desideiative secondary conjugations contain in 
their stems the same element. 

688. 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. 

690. The consonant of the reduplicating syllable is in 
general the first consonant of the root : thus, ^US papraoh 
from V^f^ praoh; WTJf 9i9ri from yfST 9ri; a^U bnbudh 
from yWJ- But — 

a. A non-aspirate is substituted in reduplication for an 
aspirate: thus, ^ dadhs from yfJl; f^ blbhy from y^bh^. 

b. A palatal is substituted for a guttural or for ^ h : 

ioy Google 

223 Rbduplication. [—698 

thus, ^^ cftky from ym kj; Rffaj oikhid from I'fej kMd; 
giM Jsgrabh from yzpt grabh; s!^ jahr from y^ hj . 

O. The occuioQil retenion, on the othei hand, of a pil&lsl In the 
iadic*l lyllible to guUuial form bu been natieed above (S16, 1). 

d. Of two initial consonaQtB, the Becond, if it be a 
noQ-naaal mute preceded by a sibilant, is repeated instead 
of the fiist; thuB, rTRT tastr from vTfT sty; HFIT taatlUi from 
yreif sthS; ^Ri-f oftskand from yHi-^ akand; ^f^^T 
oasktaaJ from yV^vi akhal; ^^j^ oufout from yWl 9aut; 
qpW paapfdli from yPJ^ Bpfdh; qftfij poaphuf from y^^JiZ 
Bphut: — but Tt^ saanS from y^ anS; Tt^ eaamr from 
y^ amp; ;w aoaru from ^^ arn; fulfill M fi^U? from yftn' 

Accent of the Vert>. 

681. The statements which have been made above, and those 
which will be made below, as to the accent of verbal forma, apply 
to those cases in which the verb is actual 1 7 accented. 

a. Bat, according to the grammarians, and according to the in- 
variable practice in accentnated texts, the verb is in the majority of 
its occnrrences unaccented or tonelesa. 

b. That ia to uy, ot couiee, the verb In itf proper rarma, ita personal 
or lo-called finite formg. The verbal uonua and adjecllvea, or the infinitives 
and paitlclplea, are BDbject to precUet; tbe aime laira of accent as other 
nouns and adjectives. 

692. The general rale, coveting most of the cases, is this: Tbe 
verb iii an independent claoae is nnaccented, unless it stand at the 
beginning of the clause — 01 also, in metrical text, at the beginning 
of a pida. 

a. For the accent of the verb, as well as for that of the vocative 
case (above, 314 c), the beglaniug of a pftda counts a* that of a lentencs, 
whatever be the logical connection of the pSda with what precedes It. 

b. Eiamplea of the nnaccented verb ars: agnlm I^ pur6hitam 
Agni I praise, the houte-prittt; sa id dev^n gaochati that, Hubf, goet 
to the godt; Ague aupayaa6 bhava O Agni, be easy of aeeett; Idiim 
indra ff^tilil somapa thU, O Indra, loma-drmker, hear; o&maa te 
rodra kp^ma^ homage to thee, Rudra, tee offer; y^amftnasya paqlhi 
pUii the taerifieer'i eattU proteei thou. 

o. Hence, there are two principal situations in which the verb 
retains its accent: 

ioy Google 

B88— ] VIII. Conjugation. 224 

69S. First, the verb ia accented when it stands at the beginning 
of a clause — or, io verse, of a pfida. 

a. Extmplet ot tbe vaib accented at the bead of the aeuteiice me, Id 
proBs, q&ndlutdhTtufa dsivyS^& k&rma^e he pure for Iha tUeitu 
ceremony; &pn6ti 'mUi loUm ht iciiu (hie tcortd; — in verte, *heig 
the lieul of tbe sentence Is also ttiat of the p&da, ayamd 'd Indrasra 
fjixtaaifi may u>» be in Indra's protection; darf&ya mft yStndhanftn 
thou) me the toreerere; g&road Tl^ebbir t n& nah mat/ he come milh good 
thingi to tt>; — In verae, vheie the head of th« cUaie is within the p&da, 
t^fitb pftbl Qrudhi hivam drink of them, hear our call; B&stn mftti 
B&Btn pita B&stn (va B&atu vlfpAtih hi the mother »le»p, let the/other 
ale^, let the dog «iup, tat the matter eteep ; vifrakarman nAmaa t« 
pKhy aamlln VicBokarman, homage to thee; protect us.' yav^. ■ .r^jtia 
Qoe duhlti pf aohi v&ih nar& the king's daughter taid to you "I pray 
you, ye men"; vay&ib te T&ya Indra vlddlil ^u ^al^ pr& bharftmatae 
tee offer thee, Indra, ttrengthening ; take nott of ut. 

b. Eitmplea of tbe verb accented at the head ot the pftda when this 
is not the head of the aentenee are: &thB te &ntamftnSil) Tldyaiaa 
somatln^ >o may we enjoy thy motl intimate favori; dhtti 'aya 
agruvU p&tiifa d&dhKtu pratikamykm Dhatar beeUno upon thia girl 
a huehand aecording to her leiah; yfitndhdnaaya Bomapa Jahi pndim 
tlay, Soma-drinker, Gie progeny of the toreerer. 

&94. Certain special cues under this hesd aia as followe; 

a. As a vocative forms no sfottcdcal put of the sentence to vlilch 
It ia attaohed, but Is only an externil appendage to it, s veib following 
an initial vocative, oi more than one, l9 accented, aa if it were lUelf initial 
In the clause or p&da: thus, ttqmtkar^a ^rudhf hAvam tium of 
liatehing eart, hear our caU! Bite T&nd&malie tvil Sita, we rtvermca 
thee ; viqve devS visaTO rAkf ate 'm&m all ye gode, ye Vatue, protect 
Ihia man; ati "go^ oakr^adi derS dfrvA JivdyathS piinah likeuiiie 
him, O gods, leho hat committed crime, ye gode, ye make to live again. 

b. It more than one verb follow a word or words sjnttetlcally con- 
nected with them all, only the Arat ioaea its accent, the others being treated 
as if they were Initial verba In separata clauses, with the same adjuncts 
understood: thus, tor&qir IJ Jayatl k^dtl pu^yatl aucceiaful he eonquera, 
rule), thrivei; amittin . . . p&rSoa indra pr& mf^ Jalil oa our foea, 
Indra, drive far away and alay; aam&bhyaifa Jeql y6tBl oa for us 
mnquar and fight; &Kiiifoinfi havif^a^ pr'4sthltaBya vit&lfa b&ryataiii 
vna^ft Juf^thftm O Agni and Soma, of the oblation aet forth partake, 
etyoy, ye mighty onti, take pleaaure. 

C. In like manner (but much less often), an sdjnnct, as subject or object, 
standing between two verbs snd logically belonging to both, la teokoned to the 
tint alone, and the second has the Inllisl accent : thus, Jahl praj&il D&yaava 
ea slay the progeny, and bring [itj hither; fp^tn n&l^ anbhigK bddbatn 
tm&Q& may the bleiaed one hear ue, [and may she] kindly regard [iw]. 

ioy Google 




d. It hu even come to be > (onnal mie that a verb immediately 
followlDK aootker verb li accented: thoa, b& 74 ttkm vv&m npltste 
ptUy&te pr^&y& pa^iibhil^ (9^0 whoever worMpt km (Aim it filkd 
toith offipring and eatlU. 

696. Second, tbe verb ii accented, whatever its position, in » 
dependent clanfe. 

a. The dependency of a clauie ia Id the very great majoilty of caaei 
conditioned by the relative pronoun ya, oi one of its derivittvea oi com- 
poands. Thus: y&ifa ya|&4iii parlbhiir 4si tchaf offering thou proteeUtt; 
6 U yontl jh aparl;u p&qySn they are coming who shall behold her 
hereafter; sabA y&n me Aatl t^na along ictth that tchieh i» mine; y&tra 
no^ ptb-ra plt&rah pareyulb whither our fatheri of old departed; 
adyi moriya y&di y&tudhiao &aad let me die on the epot, if I am 
a toreerer; y&thit "hADy anupUrv&ih bb&7aatl a> daye foUow one 
another in order; yivad id&ih bh'^vonaih tI^tojii 4stl how great thi» 
iohole creation ie; y&tkftmfta te Juhum&B t&n no astn what desiring 
we taerijiee to thee, let that become oura; yatamAa tftfpaSt whii^ever 
one deeire* to enjoy. 

b. Tbe pretence of a relative word In the aentence does not, of conne, 
accent the verb, nnle«a tbia la really the predicate of a dependent dauae; 
thus, &pa t;6 tfty&TO yathft yanti they make off like Ihievee {at thieve* 
do); y&t atha j4gao oa rejate whatever [is] immovable and movable 
tremilee; yatbfikamadi n{ padyate he liee down at hi* pleasure. 

e. The particle oa »ben It meana if, and obd (oa + id] if, give an 
accent to tbe verb: thne, braluna o6d db^Btarn Agrablt if a Brahman 
has grasped her hand; tv&ih oa soma no v&qo Jlvatudi ok marSmabe 
if thou. Soma, willeet u* to live, toe shall not die ; i oa g&oobftn mitrim 
e&B dadbSma ■/ he will come here, we will make friends with him. 

d. There are a very few paeaagea In whlrh the logical dependence of a 
danae containing no tnboidinatlng word appeara to give the verb Itg acoant: 
thaa, nbxa i>Qvapar9&9 o&raatl no n&ro 'amakam indra ratbino 
Jayantn when our men, horie-winged, coins into conflict, let the chariot- 
fighters of our side, O Indra, win the victory. Barely, too, an imperative 
EO followlna another Imperative that Jtg action may seem a consequence of W* ^ ^** cteX*. 
the lattec' a Is aocenled: thne, ttiyam A gaU kio^vefu b4 a&oK p(ba A -»*ywK>^rM 
come hi^er quickly/ drink along with the Xanvas (l. e. in order to drink), art yyt^^ U,^ 

e. A few other particles give the verb an accent, in viitae of a alight ^ - 
anhordiaating force belonging to them: thne, especially hi (with ita negation 
nabi), which In Its fnllest value means ^ttr, but ahadea off from that '"to^^ 
a mere aseeveratlve aense; tbe verb or verba connected with It ue alwayevr'' 
accented: thua, -vi t^ mnScantSib vimaoo hi 8&nti let them release 
him, for they are releaeere ; jke old dbi . . . anK^astii ira Bm&ai if 
we, foreooth, are as it were unreitowned; — also n6d (ni + id), meaning 
leet, that not: thai, nSt tvft t&pSti Bfiro orol^a that the tun may not 
bum thee with his beam; vir^aib oAA Tlocbln&dfini tl saying to himself, 

Wbltiaj, Onamar. 3. *i. IS 

J ..__., Lioogic 

896—] Vin. CoNJUOATios. 226 

"Imt I cut off the viraj" (inch cuee us fteqnent in the Brihmanai); — 
kod the inttnogative kavid toltethirf thus, uktbibtal^ kavfd Sgiunat 
leiB he eome hiiier for our praiiett 

B86. Bat further, the verb of k prior cUdbs is not iafreqnentlf 
accented in antithetical ooDstnioUon. 

a. Sometimei, the reUtion ot the two oUnies ti iskdil? etpsble of 
bslDg TsgtTiled >9 that of pcotMi* and apodoais ; but often, tXso, lach a 
TelatioQ Is tbit indistinct; and the cues of antltheata shade off Into ^ote 
of OTdtnsiy coSidination, the line between them appeadng to be nthet 
arbltralilr dravD. 

b> In many cs«eB, the antitbeals is made distinctei by the picaence In 
the two clauses of conelatlve woids, eapeciUly onya— anya, eka — eka, 
Tft— vft, ca— «a: thna, pp&-prft 'ayi jkati p&ry aayi fisate tome go 
on and on, oOttra tit about (u IF it where u>hile eoma go etc.); nd v& 
BiiiofcdhTam &pa vS pr^adhvam either pour out, or fiU up; a&ib oa 
'dhy&avft 'gne prii oa vardhaya 'm&m hoth do thou thyself become 
kindled, Agni, and do thou increase thie person. But it U also made with- 
oot suoh helpi thas, pra 'JatAh pr^a janiyati p&rl prAjStt gflu^tl 
fAe unborn progeny he generates, the bom he embraeet ; &pa ynfmAd &kra- 
mln ni 'amitii upavartate [<Aouf;A] she has gone aicoy from you, she 
does not eome to uf; na 'ndli6 'dhvaryur bh&vatl Jt& yajtiaih rakfSAai 
gboantl the priest does not become blind, the demons do not destroy the 
taerifice; kina e6mft grhy&nte k^na huyante by whom [on the one hand] 
are the somas dvpped outf by whom {on &e other hand] are they offeredt 

597> Wbeie the Terb would be the same In the two antithetical cUaies, 
it ta not infreqaently omitted In the second: thns, beside complete expres- 
glons like urvi att "bI t&siti eft ^1 hoth thou art broad and ^u art good, 
occai, much oftenei, incomplete ones like agofr amnpnlA lokA ielA 
yam& 'amfn Agni was in yonder world, Yama [u>iu] in this; aBtbna 
"n^^ praj^ pratlti^thaiiti mSAB^nft 'iiy£^ by bone some creatures 
stand ^rm, by Jtesh others; dvipao ca s&rvaih no rik^a o&tuqp&d 
y4o oa na^ BV&m both protect everything of ours that is biped, and 
also whatever that is quadruped belong* to us. 

a. AccentnittDU of the *etb in the formet of two antithetical clansee 
la a TQls moie itiictty followed In the Brabmanaa than in the Teda, and 
least strictly In the RT.: ihaa, in EV., abbi dyidl melllnll bhuvam 
(not bbi^vam) abht 'm£ih pf tblviih mahTm J am superior to the sky 
in greatness, also to this great €arth; and e^en {ndro vidur ifiginuw; 
oa gkoT^^ Indra knows, and the terrible Angirates. 

6eS. There are certain more ot less donbtfal cases in which a 
Terlhfonn ia perhaps accented for emphasis. 

a. Tbns, sporadically before oan& in any wise, and in connectton 
vlth aaseTetatUe pBrHclea, as klla, K&gk, evi, Md (In (B., repilarly) 
h&Dta: thus, b4iite liUbfa pfthlvlih vlbhU&maliU come on! kt us 
share up thie earth. 

ioy Google 

IX. Pbssbkt-sybtbm. 


509. Thb present-system, oi system of forms coming 
iiom the piesent-stem, is composed (as was pointed out 
above] of a present indicative tense, tt^ethei with a sub- 
junctive (mostly lost in the classical language), aa optative, 
sn 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. 

a. Tbeae tormi often go Id Sanikrlt gTMnmug br the name ot 
"ipeclil tenses", while the other tenM-systems ace styled "geneTkl tsnaes" 
— u If tlie farmei vere made ^om i special tense stem oi modifled root, 
while the Utter came, all alike, rrom the root Itself. Thets Is no reason 
why end) * dltUnetion and notoencUtnra should bo letained; ilnce, on the 
one hand, the "Bpedal tenses" come in one set of verb! directly from the 
root, and, on the othei hand, the other tense-sy stems ate mostly made from 
stems — aDd, In the case o( the aoiist, from stems baling a variety of fonn 
comparable with that of present-steias. 

600. Practically, the present-system is the most prom- 
inent 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 

A, Thus, In the Veda, the occurrences ot personal forms of this syetem 
are to those of all others about aa three to one; in the Aitareya Brahmana, 
a* five to one; In the Hitopadega, as ali to one; in the Qakuntala, aa 
eight to one; In Hann, aa thirty to one, 

601. And, as there is also great variety in the manner 
in which difTerent roots form their present stem, this, as 
being theii most conspicuous difference, is made Ibe basis 
of their piiaoipal classification; and a verb is said to he of 
this 01 of that conjugation, or class, according to the way 
in which it8> present-stem is made and inflected. 

itizecy Google 

eos— ] IX. Preshmt-btbtbk. 228 

602, In a Bmall minority of verbs, the piesent^tem ia 
identical with the root. Then there are besides (excluding 
the passive and causative) seven more or less different meth- 
ods of forming a present-stem from the root, each method 
being followed by a larger or smaller number of verba. 
These are the "classes'* or "conjugation-claases", as laid 
down by the native Hindu grammarians. They are ar- 
ranged by the latter in a certain wholly artificial and un- 
systematic order (the ground of which has never been dis- 
covered); and they are wont to be designated in European 
works according to this order, or else, after Hindu example, 
by the root standing at the head of each class in the Hindu 
lists. A different 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 &om one another by 
wider differences than those which separate the special 

608. The classes of the First or HON-a-CoNjoGATiON 
re 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, af ad eat; ^ i go; f^IH Ks sit; UT 
y5 go ; fi:^ dvif hate ; 3^ dnh mHi. 

II. The reduplicating class [third or hu-class); 
the root is reduplicated to form the present-stem: thus, 
^^ juhu from y^ hu sacrifice; ^^ da6& from y^ dft 
ffive; f^ bibhr &om yij bhr bear. 

III. The nasal class (seventh or radli-class) ; a 
nasal, extended to the syllable 'I na in strong forma, is 
inserted before the final consonant of the root: thus, 
"^^ rundli (or "^HR ruQadh) &om y^ radh ohstrucl; 
g^ yufij (or n^ ynnaj) &om vTOf yuj Join. 

uiaii.zecy Google 


IV. a. The nu-olass (fifth oi bu-oIius]; the syllable 
f nu is added to the root: thus, TR budu from yQ su 
press out; ^ffS Spnu &om yWI &p obtain. 

b. A ret; small numbei (only half-a-dozen) of roots 
ending already in ^ n, and also one very common and 
quite iiregulaily inflected root not so ending ^ kf make), 
add 3 u alone to foim the piesent-stem. This is the 
eighth or tan-olass of the Hindu grammarians; it may 
he best ranked by us as a sub-olass, the u-olass: thna, 
ag tanu &om yrJ^ tan stretch. 

y. The nft-olass (ninth or krl-class); the syllable 
^ nB (or, in weak forms, Rt nl) is added to the root ; 
thus, ^fttnr krli^B (or ^hft krlvl) from yuft tel Jwy; 
WJl BtabhnB (or frlHt stabhni) &om yTt^ Btabb estab- 

eOi. These classes have in common, as their most found- 
amental charaoteristic, a shift of accent: the tone being 
now upon the ending, and now upon the root ot the class- 
sign. Along with this goes a variation in the stem itself, 
which has a stronger or fuller form when the acoent rests 
upon it, and a weaker or briefer form when the accent is 
on the ending: these forms ate 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, theii 3d pi. middle, 
and their middle participle, in a difl'erent manner &om 
the others. 

605. In the classes of the Sbcobd or a-CoirjuaATioi*, 
the present-stem ends in a, 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., the 3d pi. middle, and the middle 

ioy Google 

606—] IX. Pbesbnt-svsteil 230 

paitioiple, are (as just stated] unlike those of the oUiez 


606. The classei of this oonjugation are as follows: 

TI. The a-olass, oi unaccented a-class (first oi 
bhn-class); the added claas-sign is a simply; and the 
root, which has the accent, is (if capable of it] strength- 
ened by gu;^ throughout: thus, H^ bhiva from yHf bhtl 
he; •RT n&ys, &om y^ nl lead; ^f^ b6dha from y^ 
badh wai«; ^ vftda &om y^ Tad tpeak. 

yn. The d-olass, or accented a-olass (sixth ot 
tud-olass) ; the added claas-sigu is a, as in the preceding 
class; but it has the accent, and the unaccented root 
remains unstrengthened : thus, ^ tudi from y^^ tnd 
thrust; ^ s^ji firom yTn a^j let loose; ^ suvi &om 
y^ sfl give birth. 

VIII. The ya-class (fourth or div-class); y» is 
added to the root, which has the accent: thus, ^t°lX 
dirya iiom y1^ div (more properly ^cf dlv: see 766) 
plat/; R^ n&hya from y^ nah bind; "^^ knldhya 
from y^i^ kmdh 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 &om the remaining classes in having a specific 
meaning, and in being formable in the middle voice 
from all transitive verbs. Its inflection may therefore 
best be treated next to that of the ya-class, with which 
it is most nearly connected, difiisring from it as the 
i-class from the a-class. It forms its stem, namely, by 
adding an accented yd to the root: thus, ^PS adyi from 
yVp^ ad eat; "^E? mdhTi from y^ rudh ohstrvet; 
^^JS bndhyd &om 1'^^ budh ieake ; H^ tadyi from 
yH^ tad thrust. 

,t,zec.y Google 

231 GOMJUQATIOlI-OLAaaES. [—81 1 

607. The Hindu grftUDurians reckon t, tenth clau ot onr-clua, 
having » clus^ign &r« added to a etren^hened root (tbos, oor&ya 
firom yonr), and na inflection like that of the other n-stema. Since, 
however, this Stem is not limited to the preaent-atem, bnt extenda 
alao into the rest of the ooqjngfttion — while it also has to a great 
extent a causative value, and may be formed in that valae firom a 
large inmber of roots -- it will be beat treated in full along with 
the derivative oonjngatloiis (obap. XIV., 1041 ff.)- 

608. A small number of roots add in the present-system a oh, 
or Bubatitnte a oh for their final consonant, and form a stem ending 
In oha or ebk, which is then inflected like any a-etem. This Is 
historioalty, doubtless, a true olaas^ign, analogooa with the rest; bnt 
the verbs showing it are so few, and in formation so irregular, that 
they are not well to be pat together into a class, bnt may best be 
treated as special oases falling nnder the other classes. 

a. Roott addiDg oh ua f ind yn, whlnb mske tlie st«mB rooh& and 

b. Roota inliititiitlng oh for thetr floil uo If, Tiq (ot van thint), 
gam, yam, vhloh make the atemi iochi, nooh&, g&ooha, y&ooha. 

O. Of the (o-ullfld root* endiog In oh, seTenl ue more or leii 
desdy atami, whoie nta bw been extended from tlia preaent te othei ayatenu 
of tenaea, 

608. Boota are not vholly limited, even In the liter Ungnage, to 
one mode of (ormitlon ot their pieaent-item, but ub aometlmes reckoned 
at bolonglng to two oi more dlffnent conjngation-etiaaea. And aueh yariet; 
of fonnatlon is etpecially fMqaent In the Teda, being exhibited by a 
ooDtiderable proportion of the roots there oecQTring; already In the Biihmanaa, 
howeiBi, a conditlan Is leached nearly agreeing In this respect with the 
olsMlcal langnage. The different present- formadDns sometlmee have diOei- 
ence) of meaning; yet not more Important ones than »re often found belong- 
ing to the aame foimatloD, nor of a kind to (how olearly a diffeienos of 
vftlne as otigtoilly belonging to the sepante elatsea of preeents. If anyflilog 
of this kind ia to be eittblUhed, it must be from the derivative eoujngatlooi, 
whieh ue sepanted by no Aied line horn the present-iyftems. 

610. We take up now the different classes, in the order in which 
tbey have been arranged above, to describe more la detail, and with 
llluetrstion, the formatioo of their present-stems, and to notice the 
inegutarities belonging under each class. 

I. Root-oUss (second, ad-olau). 

611. In this claw there is no olaas-aign; the root itself 
is aln> jaesent-vtem, oad to it ate added dizeotly the pei- 

ioy Google 

flll— ] IX. PSE18BNT-eS8TBH. 232 

eonal endings — but combined in subjunctive and optative 
with the respective mode-aigns; and in the impeiCect. the 
augment is prefixed to the loot. 

R. The accented eadlo^ (fiSS) regnlulj take the Kocent — except 
in the imperfect, where it falta on the augment — and before tbem 
the root remaina unchanged; before the nnacceated endings, the root 
takes the gufa-Btrengthening. 

b. It ia only In the fltst three oluBei thiit the endings eome Imme- 
diately In contact with a llnal conBOnant of the root, and tliaC the mlea for 
conaonant combination have to be noted and applied. In these ctiaaBB, then, 
additional patadigmi will be gl<en, to lllustiate the modei of combination. 

1. Preaent IndioatiTe. 
612. The endings aie the piimaiy (with ^g^ ^te in 3d 
pi. mid.), added to the baie lOot. The root takes the accent, 
and has gui^, if capable of it, in the thiee persons sing. act. 
Examples of inflection: a. active, root ^ i yo; 
strong form of rootr-stem, ^ 4; weak form, ^ i; middle, root 
Rs tit, stem Sb (irregularly accented throughout: 628). 
aetlTe. middle. 

1 ^ ^=IH ^^^ 51^ ?ITtJf% ^1FT% 

Ami iv&s im&B ose asvahe asnuthe 

69I lth&8 itliA asae aaUhe addhve 

3 ^ ^fTlT^ 11% 5n^ MIHIr? SIHFT 

6tl It^ yfcnti Ute i^te Jaate 

b> root dvi; hate: strong atem-foru), dv6q; veak, dvlq. For 
rules of combination for the final ;, see 226. 

1 dv^fmi dvl^v&B dvlfinds dvi;6 dvi^viihe dvl^m&he 
3 dv^kql dvi;tIi&B dvlfthi dvlkfg dvlq^Uia dvi^^vd 
s dvd^ti dviqt^a dvifinti dvi^t^ dvi^^te dvif&te 

o. root duh milk: strong stem-form, ddh; weak, dah. For rotes 
of combination for the final h, and for tbe conversion of the initial 
to dh, see 2B2a. 1B6. 160. 

i d6hiiil duhT&B duhm&B dubd duh.v&he dutamUie 

3 dh6)c9i dugdta&B dugdhi dlra^^ dnl^the dhiigdliv6 
3 d6sdhi dogdb&B duh&nti dngdlii dnh^ dnh&ta 

ioy Google 

233 HOOT-OLASS (HBOOND, ftd-OLASS). [—016 

d. root lib Uek; strong Btem, lih; we&k, lib. For roles of 
combinatioD of the final b, see SBS b. 

1 l Ahmi liiiv&B litun&e; Hh A llbT&be llbmUie 

1 Itefi U^b&a 114b& ' Ukf6 lib^tbe liijbTg 
a U^bl U^b&fl llb&ati U^hi llhite lib&te 

61S> Exunples of the 3d (tog. mid, colnddeDt in form vith tbe Ist 
■Ing. are not tn« In the oldsi langaiige (both T. and B.) ; the m09t fteqaent 
eztmplei ue fffl, dub^, vldi, Q&ye; moTe Bpoialie ue dU, brave, bav6. 
To tha of the 2d pi. Is added na In stta&na, p£thin&, yBtb4na. 
The iriegnlu aceent of the 3d pi. mid. Is found in RV. In rlbatd, dnbat^. 
Eianplei of the tame peiton In re and rate also oocnr: thus (beeidea 
those mentioned beloir, 628^0, 63S), vldr^, and, ultti auxiliary fovel, 
Arbire (nnleit theas tie to be lanlied, rather, at peifeot focma vithont 
ndnplleatlon: 780 b). 

2. Freeent SubjnnotlTe. 

614, Snbjanctire fonns of this clua are not untwmmon in the 
older language, and ne&rif all tbose which the formation anywhere 
admits are quotable, from Veda or from Brihmana. A complete 
paradigm, accordingly, is f^ven below, with the few forms not 
actnally quotable for this elaes enclosed In brackets. We may take 
as models (as above), for the active the root 1 go, and for the middle 
the root Sa ait, from both of which numerous forms are met with 
(although neither for these nor foi any others can the whole aeries 
be fonnd in actual use)- 

a. The mode-stems are 4ya (6 + a) and £aa (ia-f-a) respectively. 
setlTe. middle. 

.g:f 4^tb« *^tba { J- [^.tbei f^:2 

\i^tit »'***■ "i"^ l&jfttai I aatotai 

615> The KT. has no middle foime in Si except those of the flnt 
penon. The Ist sing. act. in ft oecon only in KY., in ayS, bravK, 
otAvK. The 2d and 3d sing. act. with primary endings are Teiy unnsail 
In the BAbmanas. Forms Irregalaily made with long ft, like those bom 
pTMODt^stems in a, are not nn In AT. andB. : thns, arfts, e^it, iyftn; 
&«ftt, br4vBt; bravfttbaa; asfttba, ay&tbA. brav&tba, baastha; 
&dftn; dobftn. Of middle forms with secondary endings are found h&nailta, 
3d pi., and iQftta, 3d slog, (attei mi prohibltlTs], which is an Isolated 
example. The only dnal person in tlte is br&vfilte. 

ii.zecy Google 

IX. Prbsbkt-btbtbm. 

3. FrsBent Optative. 

616. The perBooal endinga combined with the mode- 
signs of this mode (7T 7ft in act., ^ I in mid.) have been 
given in full above (866). The stem-fonn is the unaccented 
and unstrengthened toot. 

utlvB. mIddlB. 

1 ^UTR ^J\^ ^UIR MIHly MIHleji^' WH^tf^ 

iyom lyava Ijr^a istya iaiTabl ^matal 

s ^DTT^ 5mrT\ ^HTfT MrahJTC|^ MwluiBim yiHluH^ 

ijt& iyittaa iyittt italthAs asIrfttliKm aaidbvam 

lyjt Irit&m ly^ taitA ieHyitSia ieiinui 

a. Id the same manner, from ydvif, dvi^jim and dviqiyi; from 
ydxih, dubylCm and diilil7&; from yUh, lili;£m and lUuy4. Th« 
infleotian is so regular that the example above given is enougb, with 
the addition of dvlf^i, to show the nonnal accentuation in the 
middle: thus, sing. dviqlyA, dviqlth^. dvli}It&; da. dTiflv&hl, 
dvlQlytoiSin, dvlqlyitsm; pi. dvlfim&hl, dvl^dliT&m, dvi^ErJtn. 

b. The RT. hu ODce tana in 3d pi. act. (In syBtMia). 

4. Present Imperative. 

617. The impeiative adds, in second and third peisons, 
its own endings [with yHH atftm in 3d pi. mid.) directly 
to the loot-stem. The stem is accented and strengthened 
in 3d sing, act.; elsewhere, the accent is on the ending 
and the loot remains unchanged. The first persons, so called, 
of the later language are &om the old subjunctive, and 
have its strengthened stem and accent; they are repeated 
here from where they were given above (614 a). In the 2d 
sing, act., the ending is regularly (as in the two following 
classes) iv dM if the root end with a consonant, and i^ bl 
if it end with a vowel. As examples we take the toots 
already used for the purpose. 

ioy Google 




3& Boot-class (second, ad-oLASs}. [—619 

a. Thus, from the roots ^ i and ^STH &b: 

letiTt. middle. 

*(UfiH ?niR ?raT7 ^ tllHIsi^ MIHTn% 

Ar&ni &T&Ta ArKma is&l is&v&bSl ivKmahU 

lt4 isBva £astlism ^dlivun 

yjmtu itatftm itafttfim UsaULm 

b. From the roots dvlf sod dnh llh: 

I dvifft^ dT^f Rta dv^Bmft dv6f Si dvdf (iTahU dTdf&mabU 

1 dvl^^ dTlftim dTi4t& dvlkfT& dvl^ltthSm dTl44bvim 

3 dviftn dTi;(jEm dvlfintu dvlttliim dvifitBm dTifitSm 

1 d6hKnl d6hSva d6h&ma dbhU d6hSvahU d6hSmfthfil 

1 diigdhl dagdh&in dogdliA dhiihfv& duIiithBm dtmgdhT&m 
S d6(dlin dugdliim doliinta dngdliAm doliltuuii da2i4t(Un. 

i lUbMni lih&v& l^ibBma UhU ItiUvahU libSnutb&l 

2 n^iti It<pi4m U^hk Ilk9v4 lltiithSia lii^vim 

3 Id^liu H^liium UMnta ll^ikim UlLat&m libit Km 

618. The 2d ling. >ct ending tKt li t«niid in tbe oldei lungnige tn 
> few T«ibB ot thia oliu: namelr, vltttt, rttftt, brfltit, hatSt, yUBt, 
■tntfttt Id 3d alag. mid., two ot three ^erbs baie In the older linguege 
the sadtug Km: thni, dnlLim (only BV. cue), Tldfim, qt^Sm; and In 
9d pi. mid. AV. hu dnlurjni and dohmtAm. Tlie uie of tana tai ta 
In 2d pi. aet. 1b qnita frequent In tlie Veda; thni, Itana, y&t4na, attana, 
etc. And in Btota, ita 6tana, bravltana, QSst&na, hantana, ve haTe 
exunplM In th« aamc pgtiod of a ttrong (and accented) stem. 

6. Freaeiit Fartioiple, 

619. a. The active paitioiple has the ending ^1? &nt 
(veak stem-fonn CIrT at) added to the uaattengthened lOot. 
Mechanically, it may be foimed &om the 3d pi. by dropping 
die final ^ i. Thus, foi the veibe inflected above, the active 
participles are n? y&nt, ^^^T duh&nt, i^NH dvlt^t, i^I^tT 
Uhint, The feminine stem ends usually in 5fft ati: thus, 
Hfft yati, i«(.d1 duhati, \i,nS dvi^ti, fe^fft llhati: but. 
from loots in &, in En# BntI oi SHrft Bti (449 gl 

Dgil.zecy Google 

eie— ] IX. PRHSENT-SYflTEH. 236 

b. The middle paitioiple has the ending ?H Kn^, added 
to the imstrengtheiied root: thus, ^ITRly^ni, T^H duhSni, 
fiMHn dvij^d, fe?^H lihSni. 

o. The root aa forms the anomalooi and isolated ittina (in BY. 
alBO laSn&l. 

d. But a number of these participles in the older lanpnage have 
a double accent, either on the ending or on the radical syllable: 
thus, l9fiiiA and iqRna, obSnA and 6hULa, dnli&iiii and duliSna (also 
dnghiaa), rUift^A and tih&^a, vldSni and vidina, sa-ritni and 
BuvSna, BtuT&n& and staTftnii and at&VBna — the last baring in part 
also a strong form of the root. 

e. Imperfect. 

620. This tense adds the secondary endings to the root 
as iQCteased by ptefixion of the augment. The root has the 
gui^s-stiengthening (if capable of it) in the three persons of 
the sittgulai active, although the accent is always upon the 
augment. Examples of inflection are: 

a. From the roots ^ 1 and ^n^ Kb : 

aetlTe. middle. 

1 STTTH^ "^ "^ arfti cnt^f^ Miwf^ 

Ityam &{va sfma itsl favabl itsmabl 

tie ftftam &{U isQiBfl as&Ui&m addbvam 

nit slt&m ayan iatat asKtam asata 

b. From the roots dvi^ and duh and llh; 

1 idva^am Advi^va AdTl^ma &dvl;l &dTi;vaIil &dvlqmahl 
1 &dvet 4dvl;tam dd-visfa AdvifthSs &dvi;fitham advl^^vam 
s 4dv»t &dvl9(&m iidvifan Advi^fa Advlf&t&m idvl^ata 
1 Adobam Uutava &dubma Adohi dduhvahi Adnhmahl 

1 &dbok idugdbam idugdlia &dugdhfta AdubfithSm AdhofcdliTam 
3 &dh<dc &dngdhKm &duhaii &dusdba Unhttfim Adnlista 
1 Uetiam Uibva Uihma Uihi AlihTahl AiihmnW 

1 aiet Ui^liam Ul^a ill^bfia aUh&tbBm ili^hvam 

3 Uei au^ftm Uihan ail^ba ililiKtSm ilihata 

621. a. Roots ending in a may in the later langoage optionally 
take UB instead of an in 3d pi. act (the s being lost before it); and 

Digil.zecy Google 

237 Boot-class (second, ad-cLAss). [— flSfi 

in the older they alwaja do so: thus, kyan from yji, &pUB from 
yp& protect, (ibhuB from ybh&. The ssme ending is also allowed 
and met with In the case of a few roots ending in coDsonsnts: namelf 
vid know, oakf, dvif, duh, inrJ- RV. has atvifiiB. 

b. The ending tana, 2d pi. scL, 1b found in the Teda In iyBtana, 
isastaiiB, filtans, ^bravitoiui. A rtrong ilem H leeo in the lit pi. 
homa, and tbe 2d pi. abravlta ind Aliravltana. 

o. To Mve tbe ehsTMteristia endings In 2d and 3d sing, act,, the root 
ad inaerti a: thus, idaa, iAati the root as Inserts i: thtu, jisia, ttSX 
(see below, SSd); compare also 031-4. 

628. The use of Iha peieona of thii tenae, withonl ingment, in the 
older liDgaege. ^^u been noticed tboye (687). Aagmentle«B imperfect* of 
thii clui are rather ancommoD In the Veda; thus, b&O, t^ 2d aing.; 
ban, vet, stftut, d&n (?), 3d sing.; bruvaa, dtUtuB, cakfus, 3d pi.; 
vasta, Bilta, 3d sing. mid. 

ea3. The first or root-form af aoriet is idenlieal in its formation with 
this imperfect: see below, 8S9fr. 

624. In the Teda (hot hardly outside of the RV.) are found certain 
2d ling, farms, having an imperative tsIdb, made by adding the ending si 
to the (accented and strengthened) root. In part, they aie the only lool-forms 
belonging to theiaota from which they come: (hue, J68i (for J6;9i,^m yju^), 
dh&kfl, p&r^l (ypr pas'), praai, bhak^l, ratal, o&tal, ho^; but the 
majority of them hue foiroa (one or more) of a loot-preseat, or sometimes 
of a rool-aarlst, beside tbem: thus, k;6;i (ykql rule), J6fi, d&T;l, nakfl 
(ynaq attain), ak^i, m&tsl, mftsi (y'mfi measure), y&k;l, y&disl, ySSi, 
76tsi, raai, v&kqi (/vah), t^;!, froqi, aakffl. Their formal oharicter 
la eomewbac disputed ; bnt they ire probably indicative peraona of tbe root- 
class, used impetatiiely. 

6SS. Forms of this class are made from nearly 150 roots, either 
io the earlier language, or in tbe later, or in both: namely, from 
about 50 through the whole life of the language, from 60 in the older 
period (of Veda, Brafamana, and Sutra) alone, and from a few (about IIJ) 
in the later period (epic and classical] only*. Not a few of these 
roots, however, show only sporadic root-forma, beside a more usual 
coDJugatiOD of some other oIhss; not is it in all cases possible to 
separate clearly root-present from root-aorist forms. 

a. Many roots of this class, as of the other classes of the first 
conjugation, show transfers to tbe second or a-conjngation, forming 
a conjugation-stem by adding a to their strong or weak stem, or 

* Soch statements of nnmbete, with regard to Ibe varions parti of the 
system of conjugation, are in all cases taken from Ihe author's Supplement 
to tiiia grammar, entitled "Kaota, VoTb-Formi, and Primary DerlvatlTet of 
the Sanikrit Language", where Kata of roots, and details m to forms etc., 
are alio given. 

i, Google 

636—] IX. PaBsanT-srsTBH. 238 

even to both: thus, from |^m|j, both m&ijft (027) and mfjft. Snob 
tansfers are met with otod Id the oldest Ungoage; but they iiBOally 
become more frequent later, often esUblishing a new mode of present 
infleotioo by the aide of, or in aubstitation for, the earlier mode. 

b. A nnmbei of roots offer irregnlaritiea of inflection; these are, 
in the main, pointed out In the following paragrapha 

Irregularities of the Boot-ola&a. 

630. Tbe roots of the ctasa ending in a have in their strong 
forms tbe v^ddhl instead of the gima-streagthening before an ending 
beginning with a consonant; thus, from yata, atfiiimi, &atsnt, and 
the like: but Aatavtun, st&Tftnl, etc. 

a. Root* found to exbibit thie pecQlUrlt; in •ctaal nae uc kq^U, yn 
unite, BU (or aft) impel, eku, atn, ana (these in the Birllet Ungoage), 
na, ra, and hnn. RV, hu ones Btoql and anSvan. Comp&re alio 638. 

627. Tbe root mji also baa the TTddhl-Towel in its atrong 
forms: thus, mlt^fml, iim&rjain, 4m&rt (160b); and the same streng- 
thening is said to be allowed in weak forms before endings beginning 
with a vowel: thus, mSiJantn, amKrJao; but the only quotable case 
is m&^ta (LCS)- Forms from a-stems begin to appear already 
in AV. 

a. In the otb«r Unse-Bystems, also, and in derlTation, tufj shows oftan 
the TTddbi Inatead of the gnna-itrengtheniug. 

628. A number of roots accent the radical syllable throughout, 
both in strong and in weak forms: thus, all those beginning with a 
long vowel, fia, i^ Ir, li^; and also oak;, tak;, tr&, niiiB, tob cloth*, 
qljlji 9I tie, and afi. All these, except takf and trS (and trll also in 
the 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, m&tsva, y&k^vtt, B&kova, sik^va, fdbat. 
Middle participles so accented have been noticed above (616 d). 

629. Of the roota mentioned in the laat paragraph, ^ lie baa 
the gu^a-strengthening throughout: thus, 9&7«i f ^9e< f^v^n, g&y&na, 
and so on. Other irregularities in its inflection (in part already noticed) 
are the 3d pi. persons q^rate (AV. etc. have also qAto), Qdratim, 
itforata (RV. has also Aqersn), the 3d sing, prea q&;e (R.) and impT. 
^AySxa. The isolated active form A^ayat is common in the older 
language; other a-forms, active and middle, occur later. 

0SO. or the same roots, i^ and 19 Instrt ■ nnion-TOvel i before 
certala endings: thus, 19!^, iQldhve, i^^va (these three being the only 
forms noted in the older langnage); but RV, has ikfe beside ifiqe; the 
9tU. has once I^lte for i^\e. The 3d pi. Iqlre (on account of Its aooent) 
ti also apparently present rather than perfect Tbe MS. btt ones ths 3d ilnf. 
Impf. &i9a (like adutaa: 6Sft). 

ioy Google 

239 EOOT-OLABS (SKOOND, Bd-«LAfiS). [—680 

681. The roots rud weep, Bvap akep, an breathe, and qvaa hlour 
insert a onion-TOwel i before all the endingB begioDing with a con- 
■onsnt, except the b and t of 2d and 3d sing, impf, where they Insert 
Instead either a or I: thas, Bv&piml, ^Aeiql, Juiltl, and anat or 
inlt. And in the other fornB, tbe last three are allowed to aecoDt 
either root or ending: thus, sT&pantu and ^vAeantu (AV-)t or 
Brapiutu etc. The AV. huB aviptu instead of BV&pltn. 

a. In ths older Ungnage, yvaia inikea the same Inscrtlona: thug, 
vamiti, avanut; tiid othei cues occasionaily <hcqt: thns, Jfcui^Ta, vasifva 
(yvM cloiht'), 9natliUii, aUnihi (»11 RV.], yamiti (JB.), goolnii (MBh.). 
On the other hand, y'an early makes forma from an a-item: tbns, &aatl 
(AV.); pple inant (^B.); opt. anet (AB.)- 

932. The root bru tpeak, eaii (of very frequent use) takes the 
union-Towel I after the root when strengthened, before the initial 
consonant of an ending: thns, br&vimi, brkVi^, br^viti, ibrayis, 
&bravlt; bnt brfim&a, briiy£m, &bravam, &bmvaa, etc. Special 
occasional Irregularities are bruml, bravlhl, abruTam, abrQvan, 
brajr&t, and sporadic forms irom an a-stem. The eubj. dual br&v&ite 
has been noticed above (61B); also tbe strong forms abravita, 
ftbravitana {631 a). 

633. Some of the roots in u are allowed to be inaecCed lite bru; 
namely, ka, tu, ru, and atu; and an occasional instance Is met wfth of 
a form so made (in the older language, only tavitl noted; in the Jatsi, 
only Btavimi, once). 

634. The root am (hardly found in the later Ungnage) takes I ai 
nniou-vowel; thus, amlfl (RV,), amltl and Smlt and amifva (TS.). From 
y^am occur ^audgva (VS.; TS. qami^va) and ^amldhTam (TB, etc.). 

63&. The irregularities of i/^Ub in the older language have been 
already in part noted: tbe 3d pi, Indlc. mid. dnbatd, dotard, and dutarita; 
3d sing. Impr. duh^m, pi. duhram and duhrat&m; impf. act. 3d sing, 
idotaat (which is found also in the later langoage), 3d pi. aduhran 
(bealde iidutasui and duhiis]; the mid. pple dugli&ika; and (quite un- 
exampled elaewheie] the opt. furms dubly&t and diibiy&ii (RV. only). 
The MS, has adntaa 3d sing, and adutara 3d pi. impf, mid., apparently 
farmed to correapond to the pree. dntae (613) and dnhre as adugdtaa and 
adnliata correspond to dogdtae and duhate: compare fttga (630), related 
in like manuer to the 3d sing. l<je. 

Some of the roots of this class are abbreviated or otherwise 
weakened in their weak forms: thus — 

686, The loot 9H as be loses its vowel in weak forme 
(except where protected by combination with the augment). 
Its 2d sing, indie, is i^i^ isi [instead of aaai]; its 2d sing, 
impr. is 1^ edhi [iiiegulaily &om aadhi}. The inseition of 

Diaii.zec.y Google 


IX. Pbesbmt-ststeh. 


^ I in 2d and 3d sin^. impf. has been noticed already 

a. The forms of this extiemely common veib, 
as follows: 

:e, then, 







































iBia *" 








Participle Hrl s^t (fern. T^ satl). 

b. Besides the forms of the preBent-syBtem, there ie made from 
this root only a perfect, Jaa etc. (800), of wholly regular inflection. 

c. The Vedle tnbjnnctlie form* are the ainsl onea, made npon the 
>teiD 4fl&. The; are In IVeqaent use, and appear (asat especially) eTen 
in late teita where the BDbJanctiTe li almost loat. The resotntlon aUtm 
etc. (apt.) Is common la Vedic Terae. Aa 2d and 3d Bing. impt. li a few 
time* met with the more normal Ss (for IIb-b, as-t). Stb&na, 2d pi., wu 
noted above (618). 

d. Middle fotoiB from yas ate aho given 1); the grimmsrians aa allow- 
ed -with certain preposltlong (vl + atl), but the; are not quotable; smohe 
and ByKmahe (0 occar In the epics, bat are merely initancsa of the ordl- 
narr epic coQfnslon of voicea (BS8 a). ConfuBlana of primar; and Becondaiy 
endinga — namely, Bva and sma (not rare), and, on the other hand, sfftvaa 
and ByBmaa — ace alao epic. A middle present indicatlre U aaid to be 
componnded (in lit and 2d persona) with the ttomen egentii In tp (tar) 
to form a pertphraatlo fntore In the middle voice (bat lee below, 047). 
The 1st ting, indlc Is he; the rett Is In the nsual relation of middle to 
active forms (in 2d pera., oe, dhve, bta, dbvam, with total loia of the 
root itaelt). 

ioy Google 

241 BOOT-OLASS (eBCOND, ad-OLABfi]. [ — 940 

687. The root has tmtt«, <&iy is treated aomewbat after the 
manner of DOUD-stemB in an in deolenaioD (421): in weak forms, it 
loBes its n Iwfore an initial coDBOuaQt (except m and v) of a peraonal 
ending (not in the optative), and Its a before an initial vowel — and 
in the latter ease its b, in contact with the n, is changed to gh (oom- 
paro 40S). Thai, for example: 

PreaODt Indiettlie. Imperfect 

1 h&iunl hanv4B hanm&a A>i't"a i" iJianva Ahannut 

1 bUal batiiiA tiath& &Iuui ihatam Uiata 

3 hAntl hat&a gbntoti ihan &hatlm i«hiuui 

a. Its participle is ghn&nt (fern, gtmatf). Its 2d aing. Impv. is 
Jah£ (by anomaions dissimilatiOD, on the model of reduplicating 

b. Middle form* from thli not ue frequent Id the Biifamanu, iui. 
thoae tlut ooctti ire foimed In general iceoiding to the ume lalea: thus, 
lutt*, haruoabe, glinate; ahata, aghnfttSm, agtmata (in AB., alio 
ahata); ghnlta (but Ueo hanlta). Foimi rrom tranifei-stams, hana and 
ghua, are met vlth ftom an eaily period. 

638, The root va^ be eager Is in the weak forms regularly and 
nsually contracted to uq [as Id the perfect: 704 b): thus, ugm&al 
(V.: ODce apparently abbreviated in RV. to fmaai), uf&ntl; pple 
nqant, n^ftnA. Middle fonoe {except the pple) do not occur; nor do 
tbe weak forms of the imperfect, which are given as Au^va, Kn^tam, etc. 

a. KT. has in llbe manner the participle u^Ki^ rtom the root vaa elciht. 

830. The root 9&a order shows some of tbe pecnliarities of a 
rednpHcated verb, laoklog (046) the n before t in all 3d persons pi. 
and in the active participle. A part of Its active forms — namely, 
the weak forms having endings beginning with consonants (iocluding 
the optative) — are said to come from a stem with weakened vowel, 
^ [as do the aoriet, 864, and some of the derivatives); bnt, except- 
ing the optative ($l^&m etc, U- 8- and later), no such forms are 

a. The 3d alng. impf. Is aqftt (665 a), and the Mme form i* said 
to ba allowed alio aa 2d alng. The 3d sing. Impv. la (Sdhi (with tetal 
loaa of the a); utd RT. Iiaa the atrong 2d pi. gfiat&na (with auomilona 
accent]; and a-farma, from ttem ffiaa, oecaalonall; oceni. 

b. Tbe middle inflection Is regnlar, and the accent (apparently) 
always upon the radical syllable (Qiste, qltaate, giaftna). 

a. The root dSq tcorihip has in like manner (RT.) the pple diEfat 
(not d&qant). 

640. The doolie lO-callad root Jakf m(, laugh la an evident ledn- 
plioatlon of gbaa and has rMpecHvely. It baa the abnnee of U In act. 
WkitniT, Onanu, 3. »i. 10 

ioy Google 

640—] IX. PEBSHNT-flYBTBlI. 242 

3d penoDi pi. *nd pple, and flie uicent on the root bafora lowel-eDdlngi, 
which belong to radnpllutad Terba; ind It also takes the onfou'vovel 1 
In the nunoei of rod etc. (abore, 631). For Its fonuB and derivatlTe* 
made with nlt«i Iom of'the final dbtlant, ■•« SS3C 

041. Certain other obTioaaljr rednpltcated verbs ue tre»ted b^ 
the native grsmmariuis as ff simple, >nd referred to this conjngstioB : 
raeb are the intenatvely rednplic&ted Jigr 11090 a)> daridrfi (1024«), 
and Ten (1094 a}, dldl^ etc. (676). and oaUs (677). 

II. Reduplleating Class (third, ha-elaM). 

642. This class forms its pieseat-Btem by prefixing a 
reduplicatioa to the root. 

648. a. As legards the oonsonant of the reduplication, 
the geaeral rules which have already been given above (890) 
are followed. 

' Xi. A long vowel is shortened in the reduplicating syl- 
able: thus, ^ dadS from v^ dS; f^Ht bibhl firom y^ bhi; 
^ jubll from y^ htt. The vowel K r never appears in the 
redapUcation, but is replaced by ^ t: thus, ^^ blbhr from 
VH bby; ftq^ pipTo from jTR pto, 

O. For verbi In which a and ft also are inegnlaily repreaantad In the 
rednplicatlon b; 1, lee below, 660, The root vpt (V. B.) make* vavarttt 
Btc; oakr&nt (RV.) is Tory donblful. 

d. The only root of this class with initial vowel is j (or ar); 
)t takes as redaplicatloD 1, which ii held apart irom the root by aa 
interpoaed 7: thos, lyar and iyr (the ]att«r has not been found in 
actual nse). 

044. The present-stem of this class [as of the other 
classes beloogii^ to the first or non-a-conjugation] has a 
double form: a stronger form, with gunated lOot-vowel; 
and a weaker form, without gnva: thus, from y^ liu, the 
two forma are ^f juho and a gT Juliu; from y>ft bhI, they 
are ^^ bibhe and f^^ bibbl. 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 (SS3], 
and the weak stem before ihe accented. 

ioy Google 



040. According to all the wtalogieH of the first general ooi^a- 
gatioD, we Bhonld expect to find the accent upon the root-syllable 
when thU is stiengthened. That is actnallj the case, however, only 
in a small minority of the roots composing the class : namely, in hu, 
bbl [00 test-fonns Id the older language], hri (no test-forms found in 
the older langnage), mad (veiy rare), Jan (no forms of this olass 
found to occur), ol notica (in Y.), yu ttparaU (in older laogaage only), 
and in bhf in the later language (in V. it goes with the minority: 
but RV- has blbbArti once, and AV, twice; and this, the later 
accentuation, is found also in the Brithmanas); and RV. haa once 
IrArfl. In all the rest ~ aiqtareDtly, by a recent transfer — it rests 
npon the reduplicating instead of npon the radical syllable. And in 
both classes alike, the accent is anomalously tbrowD back upon the 
rednplioation In those we>^ forms of which the ending b^ns with 
a TOwel; while in the other weak forms it is upon the ending [but 
compare 666 a). 

a, Appaisntly (tlhe ouea iritb written ucent tie too few to detarmlne 
the point Mtlifactoiil;) the middle optttiTO endinga, tya ato. (B66), ire 
iMkoned thiODghoat u endings with Inttltl vowel, ind throw back the 
■cc«nt npon the lednplleition. 

646. The veibs of this olass lose the ^n in the 3d 
pi. endings in active aa well as middle, and in the imper- 
fect have 3H OS instead of ^Pf an — and before this a final 
radical vowel has gOQa. 

1. Present Indicative. 
647. The combination of stem and endings is as in 
the preceding ola««. 

Examples of infleotioa: a. y^ ha sacri^e: strong 
stem-form, g^ jah6; weak form, ^gr jubu (or jdba). 
■edTo. middle. 

^H^ gs^^iH^ |^w*i^ 

Jobiioi JnliavAs JulminfaT 

ivhb^i iubuth&ajubuthi 

^•^id sJ^HH^ 5^fii 

Jidi6tl ialiat4a J^vaU 

J&hve Juhuv&be Jnbnjniba 
JubofA J^vfttba Jubudbvi 
Jnbntd J&hvftte J^vate 

Digil.zec.y Google 

647—] IX. pREeEKT-BTSTEIf. 244 

b. Boot H bbr l>^'"' (given with Vedic accentuation): 
stion^ stem-form, fsp^" bibhar; weak, fspT blbhj (or bibbr). 

> Isprfrf f^^HH^ ^^'?^ ^ ^JR^ f*WH^ 
bibharmi blbtiTiri« blbhrm&a blbtare bibhrv&li» bibbrm&he 

bfbharql bibbptbAa blbbftbi blbb|^« btbbrftths biUiT<UiT« 

3 isprfif Imhhm^ ^^ fii>^ f^aj^ fea^ 

bfbbartl bibhrtis bfbbratl blbhrti bfbbrtto bfbhnta 
O. Tha n or hu (like that at tlie elus-ifgnt nu and a: we below, 
607 a) ll mli to be omlulble before v &nil m of the endings of lit dn. 
ftnd pi.: An*, Juhv&B, Jobvibe, etc.; but no Rtiah toima ue qnoUbls. 

2. Preaent SubjunotlTe. 

648. It Is not poeaible M ptoient to dn« & dittlnot line betweon 
those Babjqiicdve fonui of the older Isngnage which ihoald be reckoned w 
belonging to the pTeaent-ayitem tnd those which should be uilgned to the 
peiteet — or even, In some ct«es, to the rednpllMted Mrltt uxi iotsnalTe. 
Here will be noticed only those which most cleuty belong to this clsi'^i; 
the mora donbtfnl rases will be treated nndei the perfact-iystem. Except 
Id Dnt penons (whioh contlnne In □■« is 'HmperetlTeB'' down to the later 
lingnafe), sabJuncUves from ioot» htvlDg onmlstikebly ■ redaplieated 
present-systsm ate of far from frequent occairence. 

648. The subjunctive mode-Btem ia formed in the saoal manner, 
with the mode-sign a and cnna of the root-vowel, if this is capable 
of snoh strengthening. The evidence of the few accented forma met 
with indicates that the accent is laid in accordance with that of the 
strong indicative forms: thoB fWim yhu, the stem wonid be Juh&va; 
from ybbf. It would be bibtaara (but blbb&ra later). Before the 
mode-sign, final radical & would be, In accordance with analogies 
elsewhere, dropped : thus, d&da trom Vda, d&dba f^m i^dht (all the 
forma aotnallf occurring would be derivable from the aeoondary roots 
dad and dadbj. 

eso. Inatead of giving a theoretically complete scheme of 
InflectiOD, it will be better to note all the esamplea quotable from 
the older language (accented when fonnd bo occurring]. 

a. Thus, of i(t persons, we h&ve in the actlTO JnbivSnl, blbbarii^, 
dadini, dadbKni, JabUd; Johavftma, d&dbbna, J&bSmai — in the 
middle, dadbftl, mlmU; dadbfivab&l; JnbavamabU, dadSmabo. 
dadbuablli, Hajihgir^Ahgi . 

b> Of other persons, wa hsTe with primary endings In the active 
blbbavisi (with double mode-sign: 660«), d&dbatbaa, Jubavatba (do.) 

ioy Google 

245 BBDUPUOA-nno Class (TBntD, hn-OLASB). 

and JuluTBthai in the middle, d&dhase; dAdhnte, r&rate, 6 
dftdUAl;— with MconduT ending*, d&dhfla, vivafOB, JohaTat, blbharat, 
yny&rat, d&dliat, dodh&nat, bsbluuMt; dadliau, yusravan, Juhavon. 

S. Freaent Optative. 
661. To fonn thU mode, the optative endings given 
above (566a), as made up of mode-sign and peisoaal endings, 
are added to the unstrengthened stem. The accent is as 
already stated [646 a). The inflection is so regular that it is 
unneoessaiy to give here more than the fiist persons of a 
single verb: thus, 

Mtlre. middle. 

1 sJ^UIH^ ^^UH d^UW ^<^1U sJ^IhPj 5^ftR% 
Juhnr^ Jnhnriva JabnylEma ji^vlya JAhvIvahl J^vtmahl 

4. Present Imperative. 
662. The endings, and the mode of tbeii combination 
frith the root, have been alieady given. Id 2d sing, act., 
the ending is f% M after a vowel, but f^T dU after a con- 
sonant: ^ hn, however, forms d^flT Jolindbi (apparently, 
in order to avoid the recurrence of ^ h in two successive 
syllables): and other examples of fv dhi after a vowel are 
found in the Veda. 

668. a. Example of inflection: 

kotlTe. middle. 

■■ ^ d. p. •■ ^-^ P-^ 

sJ«tNliH g^^T^ 3^PT ^^ sJiJeHel^ sj-^ojllj^ 

JoIi&vBnl JnhAvKva Jub&v&ma Jnli&vKl Jub&vSvahU Jaliiiv&maliU 

Jnliudbl Joliut&m JobutA jQliiifv& Ji^vJtthftni Jnhudbv&m 

jnbAtu Jnliutiin Jiibvatu JnliutAm JfihvfitBia j6hvatSm 

b. The verbs of tbe other diThion differ here, as fa the indicative, 

in tbe aoceotnation of tbeir strong forms only: namely, In all the 

ioy Google 




first persona [borrowed sntaJnaotiTes), and in the 3d alng. ut: thus, 
(In the oldsr Ungiuige) bibharOqi etc., bibhftrtu, bibharU etc. 

664. Vedlo liragolultles of inflection an: I. tke oocwioiud vm oC 
itrong foiMi In 2d penoDB : thus, yuyodbf , flQKdhi (beilde fi^Ihf) ; 
ynyotam (bMlde yuratim); {yorta, dAdftts uid dHdfttuu, dj&dhftta 
and d&dbStana (ne below, eSB], piputona, Juli6ta ^nd jnli6t(ULB, 
mrota Mid ymrotUUi MUftBira (066); S. the nge of dlii iiut«*d of 
bi *ttu > Towel (only In the two iBittncei Jnit qnoted)) 3. Uie endlag 
tana In 2d pi. act', nunelf, baddes those Joit glran, In Jigitana, 
dbattaan, mamAttana, vivaktana, dldlgtaoa, blbhltana, Jojnffuut. 
Johatana, vav^-ltana: the caM* aM proportienall; mDch more naauMV* 
Id tbla than in anjother olaw;- 1. the ending tfit In 2d (lug. act. In 
datCAt, dhatttt, plpptst, JabltU. 

8. Freaent Fartloiple. 

686. As elsewhere, the active particdple-atem may be 
maie mechuiioally £tom the 3d pi. iadio. by dtopping ^ i: 
thus, sr^fT jdhvat, firSrT bfbhrat. In inflection, it has no 
distinction of strong and weak forms (444). The feminine 
stem ends in ^(^ atl. The middle paitioiples are r^olarly 
made : thus, ^tJH JahrSna, (m^IUI bibhrSi^. 

a. RV. ihom u Inefnlat aooent in plpftoi (VpS drtnA). 

6. Imperfeot. 

686. As already pointed oat, the 3d pi. act. of this 
olass takes the ending 3^ns, and a final radical vowel has 
gniM» before it. The strong fonns ate, as in present indio- 
ative, the three singular active persons. 

687. Examples of iafleotion: 

a«ttre. middle. 

UnbaTam ijuhnva djohuma 
AJnlLOB ^ohutam Uahuta 
Uubot i^J&lintBni &}mhaTUS 

UubTl ^obavabl AJabnmabl 
^ubuthKa AJnbvatham J^utandtaTaai 
ijuhuta AJubvAtam ijohvata 

ioy Google 

247 BEDDPLioATUia (tbibd, lin-oLASS). [— ees 

a. Fiom vH bbr, the 2d and 3d aing. act. are of^HT 
ibibbar (foi ftbibbar-6 and abibhar-t) — and so in all other 
casea vheie the strong atem ends in a consonant. The 3d 
pi. aoi. is QHH'^H ibibbarus ; and other like oases are 
ibibhayna, aoikayiu, amifaTiiB. 

b. In H8., ODce, ablbhmfl li donbtlM* i filie iMdln;. 

688. The DMul Yadto InegnlultlaB tn 2d pi. tct. — stiong fonui, 
■nd tbe fendiiig tana — ocom In tUi WiuetlH: thoj, UadBta, idadUUa; 
Uattana, J^ahatana. The BY. hu alio odm aplprata foT apip|ta 
In 3d ting, mid., and aUblunui fbi ablbharoa in 3d pi. Mt BumplM 
of uignantleu formi tie flQla, tIt^, jIsAt; jOdtM) 9lfna, Jitaata; 
ud, with InejDlu •iiengtheiUac, mroma (AV.), TojotbMt, rnrota. 

669. The roots that form their present-Btem b; redapUcatioQ are 
a very imall olass, especiall; in the modem language; ther an oalj 
50, all told, aDd of these only a third (16) are met with Uter. It is, 
however, Tei; difficult to detenniDa the precise limits of the claas, 
becaUBa of the impoesibility (referred to above, under sabjanctiTe: 648} 
of alwaya distingaishing its forms from those of other redoplioating 
eonjngations and parts of ooiuugationB. 

a. BMldM the Impilulttei In teiue-lnO*etlon (liMd; pointed ont, 
othecs m>; b« notl«ad as follovs. 

Irregnlailtles of the BednpUoatlng Class. 

990. Besides the roots in j or ar — nameljr, r, gbf (asnallr 
written i^iar}, tp, pp, bbr, sf, bif, pfo — the following roots having 
a or B as radical vowel take i instead of a in the rednpHoating 
sylbble: gA go, ma meantn, mB btJtow, fB, liB remove (mid.), vac, 
aao; vaq has both t and a; rS has 1 once in RT-; for stbB, pB drwk, 
ghrB, ban, hi, see below (670-4), 

661. Several roots of this olaas in final fi change tbe R in weak 
forms to I (ooeaslonally even to i), and then drop It altogether before 
endings beginning with a vowel. 

a. Thli it In oloBB Ukalog; with the tieatment of the vowel of the 
elus-ilgii of the nB-oltu: below, 717. 

ThMe looti ue: 

661. qB nWpon, ut. mud mid.; tbtu, flqatl, ^fOnasi. ftfltal (aJw 
^fSdbi: abOTe, 654), flqBto, a^iQBt, qlflte, ^{flta. 

663. mB beUoto, aot., and mB mtanurt, nld. (taialj alao aoL): thnt, 
BibaBtl« voimlyBt; minmia, ubuate* iunlndtat minHilt niniBtu. 
KV. hai oooe mimanti 3d pL (for mlmatl). 

ioy Google 

604—] IX. PaBBBHT-sraTBif. 24S 

604. h» remove, mld.i thm, j{liJta, Jlbldhve, J&iate; JUtlqva, 
Jlluttfim; UihIU. ^Jlhat*. QB. bu JUutta&m (foi JlUthAm). 

60fi. hS ^i(, Mt. (oitglTully identical with the formei), may farther 
shorten the i to 1: thai, Jahftti, jaMtft. JftbitAt (AT.); jahlmas (AT.), 
Jaliltu (TB.), jt»tait«m (TA.), ajahitam (T3. AB.). Id the opUttTe, 
the Ttdietl Towel 1« loit altogethei; thai, jRh7Siii, jRhTiu C^T.). The 
2d ling. ImpT., Mcoidlng to the stammariani, 1« JahUil or JaliUll oi 
JaliSbl ; only the lint appean qnotable. 

a. Farmi bom an a'Stem, Jali&, are mads foi this lOOt, and eten 
deilvatlTes from a qnasi-root Jab- 
SOS. r& gtot, mid.: thai, mldhTtua, ruithia [tmpf. without 
angment) ; aod, with 1 In tedaplloaUon, rii^hL But AY. hat ntrWra. 

a. In thoH TMbs, the aeeent la generally conitant on the redapIlEaUng 

607. The two roota dS and dh& (tbe commonest of the dUm) 
loM their Tftdioal vowel altogether in the weak forms, being shortened 
to dad and dadh. In 2d sing. impv. act., they form respeotiTely 
dahf and dbehL In combination with a following t or tb, the fin^ 
dh of dadh does not follow the special rule of combination of a 
final Eonaot aspirate [becoming ddh with the t or Ol: 100], bnt — 
as also before a and dliv — the more general rules of aapirata and 
of enrd and sonant combination; and its lost aspiration ia thrown 
back upon the Initial of the root (IBB). 

608. Tbe lofleotioD of ydhs is, then, as follows: 

Present Indicative. 

actlTe. middle. 

s. d. p. a. d. p. 

1 tJAHhftnii dadliv4B <<iii^hm^i> dadhd d&dltvahe diUlunahs 

1 didliSal dliatth^ dhatthA dhata^ dadh&tlia dhaddhve 

3 dAdhfttl dhattiia d&dhati dliatti dadlitte d&dhate 

Present Optatire. 
1 dadlir&a dadtartva dadtaritma didbiya didhivahi d A dM m ahl 

•te. otc etc. eto. Mc. et«. 

Present iMf entire. 
1 didhtnl didhftva d&dh&ma d4dh&l d&dhBvabU d&dhamaJiM 
S dhehl dbatt&m dliatt& dhatara dadhAtta&m dhaddbvam 
s ditdhAta dliattim d&dhata dhattOm dadhfttftm dadlutsm 

1 &dadliKm Adadhva ibdadlwia Madhl Adadlivahi Adadbmahl 
3 Adadbfta &dtaattam 4dhatta Adbattliaa &dadbathBm Adhaddlivniii 
3 idadh&t idhattba &dadliiu &dtiatta fcdadbStSm Adadhata 

itizecy Google 

249 REDUFLI0AT1N& Clasb (THIBD, bU-OI-ASS). [—476 

Putioiples: kcl. dAdhat; i 

a. In th» middle (except Impf.), only those fonoi are here iceeoted 
foT irMcb tbete li ■ntbority In the wcentnated tezta, u tliBie 1> dUeoiduice 
between the aotDU tcoent and that which the tnalogleB of the cIub would 
leid oe to expect. BT. bw ones dbitse : dadbi and dmdhitte might he 
peitecta, bo fai ai the feno !■ coDoeined. BT. Mcente dndhltA once 
(d4dliltft thrice] J aeTeial other teit* hsTC didblta, d&dUran, d&dlte. 

b. The root dS is Inflected In precisely the ume vay, with 
change everywhere of (radical) db to d. 

009. The oldei language hat inegnlarltle* aa follow*: 1. the nanal 
ntoDg foima in 2d pi., didhfita and AdadtaSta, d&data and idad&ta; 
2. the nsiial taoa endings in the urns peraon, dhattana, d&ddtaua, ete. 
(6B4, 668); S. the 3d dog. Indlo. act. dadhi (like lit ilog.); *■ the 2d 
ling. impT. act. daddli£ (foi hoth dehi and dhehl). And R hai dadmi. 

670. A number of roote have been traneferred from this to the 
a- or bhit-claSB (below, 74B), their reduplicated root becoming a 
Bteieotyped stem inflected after the manner of a-etems. Theae roota 
are aa follows: 

671. In all periods of the language, from the roots etlUi stand, 
pB drink, and ghrS tmtU, ire made the presents t£ffPiSml, plbftml 
(with irregular sonantlzing of the second p), and JigbrSml — which 
then are inflected not like «"<"'«■»'. bnt like bbAvfiml, as If from 
the preseDt-flteme tfqthn, pfba. jCgbra, 

678. In the Tedi (eepeelally; alio later), the ledapUcatad toota dB 
and dh& are lometlmai tBraed into the a-itema d&da and d&dba, or 
Inflected at It roots dad and dsdb of the a-dua; and single foms of the 
same character are made from other roots: thns, mlmantl (ymA btthui), 
rirate (^rft givt: 8d sing. mid.]. 

673. In the Teda, also, ■ like secondary stesa, Jlghna, is made from 
yban (with ornisslan of the radical Towel, and eonTenlon, usual in this 
root, of h to gta when in oontaet with d: 687); and aone of the forms 
of aaqo, from J^aae, show the same conversion to an a-stem, sa^oa. 

674. In AB. (vlU. 28), a similar seoondu? form, Jisbya, U glTen to 
(/hi or hft: thoB, Jighyatl. Jlghyatn. 

676. A few lo-oalled roota of the first oi root-olus are the prodaots 
of redapllcatloa, more or leas obilons: thus, Jak; (040), and probably 
qla (bom Vfaa) and oak^ (from yk&q or a lost root kas lee). In the 
Teda fs fonnd also eaqo, from ^Boe. 

676. The grammarians reckon (as already notioed, 641) Bereral roots 
of the most eTidently reduplicate character as simple, and belonging to the 
root-class. Some of theae (JSgf, daridrft, Vtvt) are regular IntenslTO 
•terns, and will be described below nnder lotensiTes (lOflOn, I024a)i 
^dbl iUrs, together with Tedic didl Mat and pipl tweU, are sometimes 
also clasMd aa intensiTes; bat they haie not the proper rednplieatlon of 

itizecy Google 

070—] IX. Prmbut-ststbii. 260 

•Doh, uid in*y paihapE 1w 'bett noticed hen, m ndopllcUed pMMmt-ttems 
witli inesQlulf long ledupllcktlag voweL 

a. Of piea. Indtc aecnn In the otdsi Iingnt^e onlr ^dyatl, S4 pi., 
wlUi the pptoB didyat ind didliyat, ind mid. dlil7«i dldli7«, dldli- 
yKtbSm, vltli tba pplei dfdy&ao, didhrKoat ptprKna. Xbe nibj. lUnis 
are did&ya, dldhi^ra, pip&ya, and tram thsm us made fOrma wltb both 
primary (bom did&ya] and aeeondary eadlngi (and the Irrafnlatly aeoented 
didyat and didijat and didliayaa). Ko opt. oocori. In impv. ve haT* 
iUdih£ (and dldlh£) and pipihl, and ptpyatatn, pipyaUba, plpyata. 
In impf., adidea and plpeoi &dldet and &didhat and aplpct (with 
angmentlMi fonni), aplp«ma (uttli ttiODg form of root), and adldliariu 
and (Inegalu) aplpyan. 

b. A lew fonni from all the three show tranifei to an a-lnSeedon: 
thni, dldliaya and plpaya (Impi.), Aplpa^at, etc 

0. BlmtUt forms from yml btUoto ate amlmat and mlmvat. 

077. The item oakBs ehme (sometimBi oiJcfif) it also regarded by 
the grammailaQB as a root, and inpplled u ineh with teneea oatdde the 
preeent-syttem — whlob, however, hardly occur In genaine n«e. It ie not 
known In the older langnige. 

678. The root bliaa chea loMi Iti radical >owel In weak forma, 
taking the form bape: thus, b&bhastl, bat b&psaU (3d pi.), b&psat 
(ppla). Fat babdliam, *ee 833 f. 

070. The root bU fmr U allowed by the gtammaiUni to ehorton 
ii« vowel in weak forms: thus, blbt^maa or bibhlmaat btbUrim ot 
bibhiTftmt; and blbhlySt etc. are met with in the later language. 

06O. Forme of this cdaea (Tom )/Jan givt hiriA, with added i — thai, 
JaJtU^e, J^J&ldliva — ue giisn by the grammailana, bnt have neToi been 
foaad In Die. 

081. The tooti ei and Oit have In the Veda rerenion of o to k in 
the toot-eyllable after the rednpllcatton: thai, ollt^l, olkithe (anomalooa, 
for eUETttha), oikltim. aoikat, ofkyat (pple); oiUddlii. 

681. The root vyaa hat 1 in the rednpllcation (from the y), and !■ 
coDbaoted to rlo Id weak forme: thna, vlvlktAa, dtvlvlkt&iii. So the 
root tavar (If it> forme are to be reckoned here) hu a in rednpUoatlon, 
and contraeti to bar: thai, JnHflrthfia. 

II). Natal Class (seventh, rodb-class). 
683. The loots of thu olans al] end in oonsonants. And 
theit class-sign is a naaal preceding the final consonant: in 
the weak forms, a obmI simply, adapted in chaiacter to the 
consonant ; but in the strong foims expanded to the syllable 
1 nd, which has the accent. 

ioy Google 

2!»1 Kas&i. Class (bevhnth, radb-cLAiis). [— aS6 

a. In B bw of the veibt of thi clut, the qh*1 aitendi tlw Into 
other ttnie ttamt: they ue afij, bhftlU, hlAa: hc belov, 004. 

1. Freeent Indioatlve. 
684. Examples of inflection: a. the loot ^ 7uJ 
join: strong stem-fotm, n^yunij; weak, Q^ ynAj- 

Foi the nlM of oomblnition of Bn»l J, lee S19. 

MdTB. middle, 

i. d. p. I. d. r- 

> gniiM 3^^ 3^^ 3^ u^$ ^^ 

jmiijmi jrafijv&s ynfiJm&B mllji yaajvAhe ynltjinUu 
ytm&kfl ynfiktli&e yuakthi yuOkfA ynfUtthe yuAgdlivi 

3 gnfrti g^^ gsfn gi^ g^ar^ ^^ 

Ttmiiktl ynaktis mlUbiti ynflkM TnfiJJk* yuiU&t* 

b. the root "^U mdh obstruct; bases pnu roi^dta and 
"^ftT rondh. 

For the rales of combliutloii of final dh, see IBS, 160. 
i ■^tlrfBT "pH^.^ ^-Mffj^ "^ t^y^ (i-W)^ 

ni9&dlimi randhv&o randhmJui randlid mmULvilie mndlim&lw 

Tu^ittl raxkddbim moddbi runtsA mndhiJthw ronddbT^ 

3 "^mfe "5*?^ "^=3% '^'^ ^)*yiS t»-y{) 

mj^Jtddhl mnddliAa nuiiUt&ntl randdhi nutdliiEta rtuidli&t« 
o. Instead of rnBkthaa, ynficdhvo, and the like (bera and in 

the tmpT- and Impf.), It is allowed and more nanal (SSI) to write 

rnAthas, Tiifldhve, etc.; and, in like manner, rondhns, rondbe, for 

mnddhast mnddlwi aad so in other like cases. 

606. Vedlo Inegsluidet of InBeotloD *re: 1. the ordlnuy nte of ■ 

3d tins. mid. like the lit ling., u TT&]e; 2. the eccent on ti of 3d pi, 

mid. In alUati, indhati. bhiifijat& 

a. Tnnoai^l, In BhP., It doabtloM & felte re»dtng. 

2, Present SubJnnotlTa. 
686. The stem Is made, as nsnal, by adding a to the strong 
present'^tem: Uiiis, jrmija, ra^^idba. Below aie given as if made 

ioy Google 

eSfl — ] IX. PKB8MIIT-8T&TUL 252 

froDi KT^f ftU the forau for which esimplM hara been itot«d u 
actiwllj occuriog In the older Unpiage. 

■ctiTe. middle. 

1 ynniijaiil ToniyBTa ynn^Sma jnnajfil yuni^jbiuliSl 

1 ynniJttB ynnajKdlivli 

3 Ttuijat ynnijatu TOnAltui ynnijata 

687. The KT. hw once -*)»»«, irUeh 1* Knomaloaa u being made 
from the weak teow-rtem. Forma with doable mode-aign ue met -with: 
thai, tniAhtn (AY.), ridhaiTit &nd ywuO&n KB.]; and the only 
qoeuhla example ot 3d dn. *et. (betidaa oBJatia) Is hlnABltan ((B.). 
(IB. has alio hltUtoSvaa u lit dn. act: an elaewhsie nneiampled Rmn. 

8. Present OptatlTe. 

688. The optative is made, as elsewheie, by adding the 
compounded mode-endings to the weak form of piesent- 
stema. Thus : 

aettra. tniddle. 

I «sm*i^ gfuH g^iF? ^^lu (jaTc(f^ t^lni^ 

TnAJyim yxiSijivii ynQjjltms rnfijlvi rufijlv&bl yofijim&hl 
et«, etc. eto. etc. et«. etc. 

a. AB. hat ones the anomalont 1st aiag. act. v^f^Iralii. And fomiB 
like bl»ilUi7Si& -rSt, Tofijirftt, are here and there met with in the 
eplci (bboSJIr&tSm once In OOS.). MBh., too, has onoe bhaOJttAm. 

4. Proaent Imperative. 
688. In this class (as the roots all end in consonants) 
the ending of the 2d sing. act. is always fu dhL 
active . middle. 

OHsnPi ^HdM gHstiH g^ g^isn^ y^m^ 

ymi^Jfinl Tnn^&va ynnj^ftma ynn^Kl yua^Avahftt ynn^iAmatiU 

WW 3^ 3^ W\ a*iwiH, H^jf^ 

yniiAktii yuakt&a ynmintu roAkUtin jn&iit&ia TnOi&tSia 

i, Google 

253 NiBAL Clabb (sbvbnth, radli-OLA£B). [—094 

690. Theie Is no oecunenM, so hi u noted, ot the endlDg t&t in 
Terba of tbi* oIs«i. Thi Teda bu, u usniJ, gometiiDeB stiong foima, uid 
■omedmet the eoding tana, In the 2d pi. &ot. ; thni, un&ttit, ynnikta, 
anaktana, plna^tana. 

6. Present Fartlaiple. 

691. The paiticiples aie made in this olass as in the 
pieceding ones: thus, act. irgH ynfijdnt [fem. H^r\\ yufijatf); 
mid. iJyn ynfijSnA (but RV. has indhOna). 

092. The example of the legular inflection of this tense 
needs no introduction: 

5g=rair «a53 sigsn ^15% ?igf^i% 

i^naaitaa irufijTa JtTofiJma iiynfiji dyniijvahl Ayui^mahi 
Ayrmak &ynfiktam iyuakta &ytiakth&s&7^ftth&m&7Ufi«dtiTai& 

iyuiiak Ayuilkt&m iiyufijan ftyufikta iyuiij&tftm AynfUata 

a. The endings a and t are necesaarily lost in the nasal olaas 
throngfaont in 2d and 3d sing, act., nnless saved at the ezpense of the 
final radical consonant: which is a case of very rare occnrreDoe (the 
00I7 quotable examples were given at BEtBa). 

693. The Tedi ihows no irregnl4rltlea In Ihla tenae. Occarrannee at 
anfmentleaa forma ate round, eapeelttllT In 2d Kud 3d sing, aot., ihoiring 
«n aoBent Ilka thu of the preasnt: foi ezunple, bhlnit, p)^94k, vp^&k, 
pijjAk, riij&k. 

a. The lat gin|. act. atp^am and aeohlnam (foT atr^^adam and 
aocMnadam) were noted abOTe, at BBS a. 

694. The roots of this class number abont thirl;, more than 
half of them being fonnd only in the earlier language ; no new ones 
make their first appearance later. Three of them, afij and bbafij and 
talAa, carry their nasal also into other tense-systems than the present. 
Two, fdli audnbh, make pre Bent-syBtemB also of other classes having 
a nasal in the class-sign: thas, r<llmotl (nu-olaas) and ubhnSti 

ioy Google 

a. Hut of tba roota Make fonuB [tok MoandM; k-fMniB: thu, bom 
ftfija, nndB, ambUi, cblnda, tpfihi, piA^ PffioA, bhaSJa, randha. 

IrregulATities of the Nasal Claw. 

OOB. The loot tfh Mmblnet tf^^ab with tl, tn, etc. Into tT^ofliii 
ty^^^ui and, MCDidlng to the grimmulMu, hu alio nich fauna u 
tr^ebml: Me above, 824 b. 

686. The root blAs (by oHgln *ppiTentl]r a desldeTUlve from ybui) 
Mcentt irregaUily the root-syllable Id the weak forms: (liiu, bibMmtl, 
biAata, biAB&na (but blnAaot etc. and blAayltt (B.). 

IV. iTn- anil a-elasies (fifth and eighth, an- and tan-elatSM). 

6B7, A. The present-atem of the nu-olaaa is made by 
adding to the loot the syllable ^ nn, which then in the 
strong foims leceirea the accent, and is Btiengthened to ^ n6. 

B. The few toots of the u-olass (about half-a-^ozenj 
end in ^n, with the exception of the later inegulax ^ kp 
(or kar) — for which, see below, 714. Hie two clawes, 
then, are olosely coirespondent in form; and they are wholly 
accordant in infiection. 

a. The n of either cluB-sign ta allowed to be dropped before 
V and m of the lat du. aod let pi. endingB, except when tiie root 
(nn-claaa) ends in a consonant; and the u before a Towal-ending 
beoomea v or uv, aooording aa tt is preceded by one or by two 
oonaonants (ISO a). 

1. Present Indicative. 
BBS. Examples of inflection: A. nu-clasa; root 
n Bii press out; strong form of stem, ^t Ban6; weak form, 
^ sunn. 

aottTO. middle. 

1 g^ ^m^ ^qR ^ ^it ^!% 

aaii6iiii Bannvia snnnm&s Bxmvi anDUT&h« mnmaihe 
Bui6(l sanaOi&a atmntbi nmafft mmvitbe BBswIbrt 

ioy Google 

255 Nn- AMD n- (futh and biohth, bu- and tan-ci,A8BE8). [—700 

3 HHlin HHH^j^ h-^Ih g^ g^^i^ g=^ 

Baii6tl Baniit&B "^ ilWSftH Bnnut^ aoavit* Btmv&te 

a. The fonui simv&a, amun&Sf Bonvilie, aiuuii&h« are alter- 
native viA tboBB given here for let da. and pi., and In practice are 
more common. From yVp, however [for example], onl; the forms 
with n can occur: thus, Spnavia, kpnnmiha; and also only ftpuu' 
v&ntl, ftpiiuT^, SpnaT&te. 

B. u-class; root WJ tan stretch: stiong form of stem, 
fH? tanA; weak, fTT tann, 

i fRlft ffMH^ rf^H^ ?T'^ rp^ ^7"?% 

taii6iiil taiiT&a tatim^ tanvd tanviiha taum&ba 

eU. etc etc. etc. etc. etc. 

b. The inflection is so precisely like that given above that it 
Ib not worth writinfr oat Is fnll. The abbreviated forma in lat du. 
and pi. are presented here, instead of the fuller, which rarely occur 
(as no double consonant ever pieoedes). 

608. a. In the older Unguage, no itiong 2d penons da, oi pL, 
and no thana-endlng, obtnee to ocboi (bat thay ue nnmerona In tbe 
tmpT. iDd impt.: lea below). The BT. hu leveial cubs ot the Irregiilii 
secant In 3d pi. mid.: thn*, kp^vat^i tanvat^, manvat^, vf^vatd, 

b> la RT. oooni ilio leveiil 3d pi. mtd. In ire ftom pcesent-alenu 
Of this olMi: that, Invlre, p^vire, piiLvlre, iff^viri, aimvlri, hlnTlri. 
Of those, pluvlre, utd hinvlrd might be peifeoti without lednpUctdiin 
from the samnduy roota plnv *nd hlnv (below, 716). Tbe 2d sing. mid. 
(with puslve valne) fffvlqi (R^O is cf inomslotu and qnestlonable 

2. Freeent SnbJunotlTe. 

700. The lubjunctive mode-atom is made in the nsual maoaer, 
hj adding a to the gnnated and accented elass-Bign: thus, aaniva, 
tani-va. In the following scheme are given all the forms of which 
examples have been met with in actual nee In the older language 
from either division of the claas; some of them are qnite numerottslf 
represented there. 

active. middle. 

1 vtmkrKai mm&vSva anniTSma waakvtl Bnn&vftvahSi BwUiTfiiaaliU 
) annAvaa aonAvatha BunAvase aunA-vUtb* 

fannAvate aunivanta 


9 BOnAvat Biin4van 

itizecy Google 

701—] IX. Frbsukt-stbtbu. 256 

701. Of the briefer 1st ling. set., BY. bM krfftVt and hlnttTi. 
Fonnt vltb double mode-tlfn oeenr (oot Id BT.): thiu, kpqiv&t uid 
kuavKt (AT.); aqnavBllia (E.), knWTBtb* (VS.; bat •Tfttba in 
Ki^Ti-text), karftvfiiUia KB.). On the otber huid, a^navatfil Ib found 
once (Id T3.]. Fdidib like SpnUT&nt. udhni^Tttt, aqnavat, met with 
now and then in the older t«xt«, ue donbtleaa to be reguded u filie 
r«&dinga. BT. hu In i ■ingle putage kfi^TSlta (inite(d of kf^AvUta); 
the only form in iithe ii »{n&vUthe. 

8. Present Optative. 
702. The combined endings (688) aie added, as usual, 
to the weak tense-Btem: thus, 

ictive. middle. 

1 IJj^UIH^ ^^^ HHIIR y*=ftU tJ-=)icTf^ ij-c|ii((^ 

snnnyim snnuytva annu^bia avnviyk eunvlv&hl sunvlmilii 

a. Froo) yhp, the middle optative would be Kpnaviyi — and so 
1 other like oases. 

4. Freaent ZmperatiTe. 
703. The inflection of the impeiative is in general like 
that in the preceding classes. As i^aids the 2d sing, act., 
the rule of the latei language is that the ending ^ hi is 
taken wheneret the root itself ends in a consonant; othet- 
wise, the tense- (or mode-) stem stands by itself as 2d pet- 
son [foi the earlier usage, see below, 701). An example of 
inflection is: 

active. mlddlo. 

5*RlfH ^RR g-wlH W^ g5RR% (j>MIH% 

sunivani aonAvSTa sim&vSma soojitU Bon&vftTata&i aoniTtmaUl 

sand Bunat&m Biumt& BniiiifT& sanTi^hSin stiniidhTAiii 

aan6ta sunubtm sunv^tu sounbtia sunvitfiin BunvitBrn 

ioy Google 

257 Nn- AND U- (FIFTH AND EIOHTH, BU- AMD tan-) CLASSES. [—706 

a. From ytp, the 2d sing, aot- wonld be ftpnuhl; from ya^, 
afiralif; from fdhr;, dhrf^ulii; and 80 od. From y&p, too, would 
be made Spnuv4ntn, ftpnuvaUiftni, ftpnuvlitsm, fipnuT&tani. 

704, Iq tlie ettllut linguoge, the idIb ib to the omlailOD of bl 
tftei ■ toot with Anal Towel does not hold good: in BV., luoh formi u 
Inohi, krvolif, olnnlil, dhOnnhl, ^^9°^ app^nhl, hlnuhl, and 
taDuhi, BftTiiihl. ate nearly thrice aa fieqnent in Die as Infi, QP^Q, 
Bund, tanu, and thelt lite; in AV., howoTer, the; ate odIt one siitli 
as flteqnent; and in the Brahma^as they appear only ipotadkaUy; eiea 
979Udb{ (with dill] ocears eeveral timet in BV. RV. haa the Ist eing. 
act. hlnavi. The ending tSt Is found In kp^utSt and hinutftt, and 
kurut&t. The rtrong atem-fotm ia found In 2d dn. aet. In hlnotam and 
kpnotam; and in 2d pi. act. in kf^dta and kpi^dtaila, g^ota and 
^r^otaca, Biui6ta and sundtana, hta6ta and hlnotana, and tanota, 
kardta. The ending tana occdib only la the fotma Just qnoted. 

6. Present Parttoiple. 
706. The endiDgs QtT &nt and )3H Smi aie added to the 
weak foim of tense stem: thus, from v^ sn come act. UHrl 
suLVint (fern. HMrfl BunTsti), mid. T^RIR simvSnil; from yrm 
tan, fFSfrt tanv&nt (fem. H-^fffl tanvati), cl^R t&nvbiB. From 
y^Sfm 5p, they are CIT^TT Kpnuvint and Om^TH SpnuvSni. 

e. Imperfect. 
706. The combination of augmented stem and endings 
is according to the rules already stated: thus, 
active. middle. 

5g=i^ q^p g^H sihP^ ^33^^ ^^^ 

iBtmavam &annaTa iennuma JMunvi Aaunuvabl to"iTiniahi 

5g^q^ ^^^ *wjtT yjjjsnt^ Mg-^ierni^ ^35^ 

AsunoB &8iinutam Asimuta AannathSs iBunvftthSm ieunudliTam 

Asimot &Bim\it&ia Aannvan &Bunuta &aimTat&m Aflnnvata 
a. Here, as elsewhere, the briefer forms isunva, Asmima, Aaun- 
vahi, &Biuunahl are allowed, and more neoal, except from roots 
with final consonant, as dhn= which makes, for example, alwajB 
Adbf^qoma eto., and also Adb^^^uvan, &dfan9UTi, idbr^ijuvKthlm, 
Adhr^uvfttitm, Adbn^nvata. 

WbititT, Onmrnar. 3. cd, 17 

ioy Google 

707—] IX. PRKfiBNT^rsTEx. 258 

707. Strong Btem-roniu and tana-ending are foond only Id RV., In 
akf^Ota, akfi^Uuia. Aagmentleae tormt wlib acceot sk tainvka, 

708. About fifty roots make, slthei exclnaiTely or is paxt, their 
preeent-formB after the manDer of the ud-oUbb'. half of them do so 
only in the older language; three or four, onlj in the later. 

ft. A« to truiifan to the »-conjng*tlDn, tee b«loT, 710. 

709. The roota of the other diviaion, or of the n-olaaa, are 
extremely few, not exceeding eigrht, even iodnding tf on acoonnt 
of tarutA BV., and ban on acoonnt of the oocanenoe of hanoml 
once in a S&tra (POS. i. 3. 27). BR. refer the etem iau to In of the 
u-clasB instead of 1 of the nu-elass. 

IrregularitleB of the nu and a-olaBBea, 

710. The mot tfp be phated Is uld by the grammuriini to rstiln 
the n of Iti clMi-atgn nallQpi&ltzed in the later luignage — whera, however, 
fainis of conjugation of tbti oIub are vet; nre; while in the Ted* the 
regnlai change la made: thue, t^P^tl. 

711. The root ^ruhmr is oontraoted to 9r before the claoa-aign, 
forming <}pt6 and ilf^n as stem. Its forms qpfviqA and ^f^-viri 
have been noted above (698 b). 

712. The root dhS ahak« in the later language (and rarely in 
B. and S.j shortenB its vowel, making the stem-forms dboiid and 
dtanna (earlier dhiiii6, dhnnn]. 

713. The *o-ealled root ilri^ii, tceited by tbe nttlTe grammariani aa 
diuyllabie and belonging to the Toot-claia (L), ia propeHy » present-Meni 
Of this claaa, with anomalooi con&actlon, from the root v^ (or var). In 
the Teda, It bai no formi wUch are not tegululy made aocordlng to the 
nD-claiBj bnt in the Bribmana language are found aometlmea snch forma 
as Or^autl, as If from an u-iool of tbe root olasi (686); and the gram- 
marians inita n>T It a perfect, aoilst, futate, etc. Ita 3d alng. impr. art. 
Is firjfn or Ori^nlil; iti impr., sdrpos, Sur^ot; Its opt. mid., aii;iavita 
[E.) or ur^TitA (TS.). 

7L4. The extremely commoa loot ^ kr (or kar] maie 
ia in the later language inflected in the present^etetn ex- 
olusirely according to the u-olass (being the only root o£ 
that olau not ending in ^ n) . It has the insularity that 
in the strong form of stem it (as well ae the olass-eign) has 
the gu^a-stiengthening, and that in the weak form it is 

Digitizecy Google 

259 Na- AND u- iPIPTH and eiqhth, bu- abd tan-j CLAaaES. [—714 

changed to km:, so that the two foims of stem aie ^f!JJ kRT6 
and ^^ karu. The olass-sign 3 u is always dropped be- 
fore cF T and R m of the 1st du. and pi., and also before 
JT y of the opt. act. Thus: 

1. Present IndioatiTe. 
active. middle. 

a. d. p. B, d. p. 

EfJ^tPr ^^ 5*'^ ^ =W^ ^H^ 

kardml karv&a knrmiM kurri karr&he karm&b* 

*jtfR) ^r^TO^ j^ 5^ ^n^ ^^ 

kar6ql kaTatli4a karath& fcnm;6 kurr^the komdlivd 

^ijlifl 4|))rm jqfti jfirt ^i-^irl ^^ 

k»T6ti kornt&B fcanr&nti kuruM knrrite kiirv&t« 

z. Fresont Optative. -' 

JIIF^ ^dw JOTR 5=(Tu ^i4ftf^ 5i^ft% 

kur;^ kuTT^vR ktur^a kurri7& knrriv&hi kurvlmihl 
etc. <tc. etc. etc. etc etc. 

3. Preae^t Imperative. 

E?rf^lfm *t=(H ^it^TR ^il^ i^I^% <*^m? 
kar&Tft^ kar&vSva kariTftma karivftl kar&T&vahSI kar&vBmah&i 

kuni korat&m kurat& kam9T& karrSthftm knmdhv&m 

kar6tu kurutltin knrv&ntu i korutam knrTatllin kvirT&t&m 

4, Freaent Participle. 

^c^rT kurvAnt (fern. ^3rft knrvati) ^4lUI kurvSni 
6. Imperfect. 
M*(e(H 3^3 3^1^ 5(5R^ U<4)c]f^ Q?i4f% 

Akaravam Akurra Akurma &kurvi iktuvahi &kunnahl 
55iftl^ *J*t>HH^ «*JirI M^^W MiJi^lEIIM U4i^MM 

ikaroB UEoratam &koruta &kurathSa ikortrsthSm UmrudhTam 

ikarot iLknraULm Akurran iktiruta iknrr&t&m &karvata 

itizecy Google 

Tl6— ] IX. PaBBEKT-BTTTEM. 260 

71&I In BV., (kit root U legululy inflecMd in th« pr«wnt-«y(Um 
■ecotding to tbe na-clut, mtkiDg the atem-foinu k)n6 and ^ija; the 
only siMptiong ue kormaa onee and kom twice (>U in the tenth book); 
In AT,, the Da-tormB in Mil more than tix timet u fteqoant u tlie 
a-form* (nttrly Iwif of which, moreoTet, ue in pioM p4si4ge(); bnt in 
the BribmuM UDgn«ee ud Itter, the u-fonns ue ued (o the escliMtoii 
of the others. 

a. As lat (ing. piea. art. is foand konai In the epos. 

"b, Whit inesolu totmt boot kf w a verb ot the na-cltis occdi in 
the older langnage have been already noticed abore. 

O. The iiotated form tamtA, from ytf, shows aa apparent analog 
with these a-formi from k^. 

7 Id. A. few Tflrba belonging origin&llj to Utese oUsses have been 
shifted, in put oi altogether, to the a-olasa, their proper clua-tign 
having been etereotypod ae a part of the root. 

a. Thai, tn BY. we Hud formB both from the stem Inu {yi or in), 
and also from inTft, representing a deiivaHve qnaai-rool inv (and these 
litter alone occnr In AT.). 60 likewise fonns fioa a stem p^va l»eslde 
those from ^a (Yf); and from hinva beside those from hina (v^t). 
The M-oaUed coota Jinv and pinv are donblleu of the ume origin, ilthoDgh 
no forms from the stem pina are met with at an; period — aolesa plavlre 
(ibOTe, 6SSbJ be so tegatded; and AT. baa the participle piUT&nt, f. 
pinvatl. The giammarius set up a root dhlnv, bat oolr forma tZont 
dhi (stem dhinu) appear (0 occot In (he present- BTslem (the aotist 
adhln-nt ii fonnd in PB.]. 

b. Oocalional a-forms are met with also from other roOU: thus, 
cinvata eto., diuiTBsva. 

V. ns-clast (ninth or kri-dass). 

717. The olasB-sigD o£ this class is in the strong forms 
the syllable ^ nft, afibented, which is added to the toot; 
in the weak foima, oi where the accent falls upon the end- 
ing, it is ^ nl; but before the initial vowel of an ending 
the ^ I of ^ nl disappears altogether. 

1. Present Indicative. 

718. Example of inflection: root sfil krl buy: strong 
form of stem, Wt^ krl^S; weak form, c^toit krO^ (before 
a vowel, gftal krii).). 

ioy Google 


1 t+TlUlliM 5lpl!li"=(H^ stillliiHH^ ^fffit 4ti1llT|c(^ tti1inH$ 
kri^lbnl ktij^lvie krli^Im&a : lui:^6 kri^v&lie krDjlm&he 

i ^R^unftr ehiuniiiH crWIh s^OuflN sfrluiia stTiuilQ 

kri^lttl ki^Qlt&B kri^iintl kr^j^t^ krl^Jte krli^te 
718. In the Veda, Ibe 3d linf. mid. hsi the game foim wltb tbs lit 
In K^s; the pecaliai tCMDl of 3d pi. mid. Is seen Id puofttd *Dd rl^at^; 
And TTVlmahd (bBBlde Vfulm&he) ocean once In RV. 

Z. Freeent Sabjimotive, 
7S0. The SDbjnnctive fonns whtch hare been fonnd exemplified 
in Veda and Brahmans aie given below. The snbjunctive mode-stem 
is, of conrae, Indiatin^ishable in fonD from the strong tenee-stem. 
And tho 2d and 3d stng. act (with secondary endings) are indistin- 
gDishable from angmentless imperfecta. 

■ctlie. middla. 

1 b^Snl ki^£m& krli^&l kri^fivalifii kri^fimaliili 

2 kiii^ae kriQfttha krlf&BAl 

3 krl^tt kri^an ki^StSi kru^ntai 

3. Present OptatiTe. 

721. This mode is foimed and inflected with entiie 

tegulaiity; owing to the fusion of tense-sign and mode-sign 

in the middle, some of its persons ate indistinguishable fiom 

augmentless imperfects. Its first persons are as follows: 

Active. middle. 

1 ^luftuiii^ ^oihTR stiluiluiH ^fhrftu ^Tuiferf^ ^fhrfNi% 

kriiji?^ krtpiyaTa kn^Iylma krmir& ki^vihl kri^^m^hl 

1. Present ImperatiTe. 
722. The ending in 2d sing, act., aa being always pre- 
ceded by a vowel, is l^hi [never fu dhi}; and there are no 
examples of an ofyssion of it. But this peteon is forbidden 

Digitizecy Google 

728—] IX. PBESBHT-SySTBM. 262 

to be formed in the classical language fiom loou ending in 
a consonant; foi both class-sign and ending is substituted 
the peculiar ending ^ETH Sni. 

sctlTS. middle. 

I fehlii?Pi ^luiH thhinn ^Wt ^tnnsi^ cfOinH^ 

Innwitnt TtpTw a qu kll^RIIia kTl^Sl Irr i-^it irnhltl kril^JtmallU 

TryTnThf h-Tinit-ATn krl^tA krinlfri kn^thfim kri^dtiv&m 

» ^fhnR yiluilHiH^BtniiTj siftoflrn^ ^fhni?n^ ^fhncnR^ 

k^^itn knnltim kri^&ntu krl^ldun kriQatSm kH^&t&m 
a. Examples of the ending &a& in 2d sing, act are a9ftii«, 
gfh&9&, badbftnji, atabUni. 

7S8. Tha ending ftna U knowa lUo to th« eullMt luisotc*; of tbe 
exunpiea Jnat elven, nil are found Id AV,, uid the Itiit two iu BV.; other* 
ue ifS^a, moffi^, skabhftna. Bnt AY. hu also g^bb^ibl (alsoAB.), 
and even gfh^&hi, with Etroag stem; BhP. has badlmlhi. Strong (teioa 
sie fnrthei fouod in KT^Uli and stj^hl (TS.), PI'i^Uil (TB.), and 
QTjjp&hi (Apatt), and, with anomalong lecent, punUii and ffi^Shl (ST.) ; 
and, In 2d pi. act., Id pimita (RV.). The tndlng t&t of 2d sing. act. 
occnn in gfbi^t&t, J&nitat, pumtftt. The ending tana is fonnd in 
punlt&na, pj^itana, 9ri9itana. 

5. Present Participle. 
724. The participles are tegulaily foimed: thus, foi 
example, act. sliimtl krT^&nt [fem. ^luirfl kitfarf); ^^• 
sblmw krl^ni. 

e. Imperfect. 
726. There is nothing special to be noted as to the 
inflection of this tense: an example is — 

•eUTe. middle. 

^' ^ '> ^ '^ ^'^ ^ ^ ^ ^ V «*^ 

Utri^Km ikriijlva ^Ttri^jima Url^l Akri^lvahi AkrlQlmahl 

Ala^ftB &kri];kltam Url^Ita ^kri^ltliBB Akri^thftm ikrlqidhvam 
iikru^t 4ki^t&m Ucra^an &krl^ta Ak^fttKm Umpata 

ioy Google 

263 Tlil-CLABB (NINTH, kri-CLASBj. [— 78« 

726. It hu bean pointed oot >bo«* tbst aagniaDtlen penoiiB of thli 
tenie tie In part IndiitinguiBhable Id fonn from snbjanctiTO >nd opUttve 
penont. Sacb u cartatnly belong heie uo (Id V.) kfi^im; a^ui, 
rl9&n; gfblu^tn, vf^ata. The AY. bu once minlt inatead of miuiL 
MBh. hu ftfnlfl aftei xai. 

a. AB. haa the false fonn aiioBnas, and in AA. ocean avf^ita ai 
3d plual. 

797. Tbe roots which form thoii piesent-fijstemB, wholly or in 
put, After the mkniter of this cUu, are over fifty in QDmber'. bat, for 
about three fifths of them, the forms are qnotable only from tbe older 
language, and for half-a-dozen they make their first appearance later; 
for less than twenty are they in nse through tbe whole life of the 
language, from tbe Veda down. 

a. Aa to aeeondar; o-atems, eee 731< 

IrregulEuitlds of the nB-olaaa. 

7SB. a. Tbe roots ending in a shorten that vowel before the 
class-sign: thus, from yp% pnuiti and punlt6; in like manner also 
JO, dhfl, la. 

b. The root vll (B.3.) foims either vlmfi or vlinS. 

7S0. The root grabb or grata (the former Vedlcj is weakened 
to grbta or sr^. 

a. As the perfect alao in weak forms baa gfbta or gp'^ '* '^ v* 
easy to sea why tbe grammarians abonld not faaTe written p Inatead of ra 
in tbe root. 

730. a. A few of the roots have a more or less persistent nasal 
in forms outside the present-system; such are without nasal before 
the class-sign: thus, gratta or graatli, badh or bandta, math or 
mantta, akabh or skambh, stabh or Btambta. 

^bJThe root J&S also loses its nasal before the class-atgn; thus, 

731. Not rarely, forms showing a Iranefer to the a-oODJngation 
are met with : thus, even in BV., minaCi, mlnat, amlnanta. from 
ytDi; in AV., fiT^a from Vft; later, grb^a, Jftna, prii^ mattana, 
etc. And from roots pj and vaj are formed the stems pf^ and 
m{^, which are inflected after the manner of the i-cUss, aa if from 
roots pf9 and 11179. 

782. Id the Veda, an apparently denominative inflection of a 
stem Id &y& is not infrequent beside the conjugation of roots of this 
class: thos, grbbaji, mattaST&tl, a^rath^aa, akabhSy&ta, aatabb- 
Kyat, pmfar&nte, mufSyAt, and so on. See below, 106eb. 

ioy Google 

7S3— ] lIX. Pkbskkt-«tbtem. 264 

Seetnd or ft-CtnjugatiAn. 

783. We come now to the clawes which compose the 
Seooad ot ft-Coajugation. lliese aie more markedly 
nmilar in theii mode of inflection than the preceding classes; 
their common characteristics, already stated, may be here 
repeated in summary. They are: 1. A final a in the present- 
stem ; 2. a constant accent, not changing between stem and 
ending; 3. a briefer form of the optative mode-sign in the 
active, namely I instead of 7ft (combining in both voices 
alike with a to e); 4. the absence of any ending (except 
when tst is tued] in 2d sing. impv. act.; 5. the conversion 
of initial B of the 2d and 3d du. mid. endiogs with final a 
of the stem to e; 6. the use of the full endings ante* ante, 
ant&m in 3d pi. mid. forms; 7. (he invariable use of an 
(not na] in 3d pi. impf. act.; S. and the use of mBna instead 
of &na as ending of the mid. pple. Moreover, 9. the stem- 
final a becomes S before m and t of Ist personal endings — 
but not before am of Ist sing, impf.: here, as before the 
3d pL endings, the stem-final is lost, and the short a of the 
ending remains (or the contrary): thus, bhdvanti (bhiva-|- 
anti), bb&vonte (bhiva -(- ante , Abbavam (ibhava-(-am). 

«. All these chaiacteristicB belong not to the indecdon of the 
a-present-ayBUm alone, bnt &lso to that of the a-, rednpl lasted, sad 
sa-aorista, the s-fdtnie, ftnd the desiderative, caosstive, and demon- 
inatlve preaent-stems. That ia to aa^, wherever in conjugation an 
a-atem ia found, it is Inflected in the same manner. 

VI. A-c)ass (first, bbft-ciass). 

784. The preaent-stem of this class is made by adding 
9 a to the root, which has the accent, and, when that is 
possible (2SB, 240), is strengthened to gu^ia. Thus, ^ 
bh&VB from y^ bbD; sHT jAya from v^ ji; ^^ bodha from 
V^q budh; Wf s&rpa from yw\ syp; — but ^ vida from 
V^ vad; ?Rte kri^A &<>" V^^ krl^. 

ioy Google 


1. Present Indiofttive. 

736. The eodingB nod the rulea for theii combinatioQ 
with the stem have beea alieady fully given, fot this and 
the other parts of the pieseDt-system; and it only lemains 
to illostiate them by examples. 

ft. Example of inflection: lOot T bhil be; stem HS) 
bhivft (bho+a: 181). 

MtlTe. middle. 

>. d. p. 1. d. p. 

1 ^^\^ H^Rg^ H^Fin^ >R H^ra% H^n% 

btk&vBml bh&T&TOS bh&vfimM bh&T» bh&TftTaha bh&vfimah« 

bh&vMl bh&vathoa bb^vatba bh&TaaebbiTetha bhdvadhve 

bhiivati bh&TRtas bb&vantl bh&vate bMvete bh&vante 
b. The Y. hu but a Bingle eiiniple of tlie Uiana-Bsdlng, nunely 
T&datbana (tod no othei in lay cIub of tbls eonJagtUoD). The l«t pi. 
mid. manSmahA (RT., oncej U piobtbly id error. RV. }iu f6bho once 

I 3d alngnlar. 

2. Present SubjunotiTe. 

786. The mode-stem Is bh&v& (bb£va+a). Snbjonctire fonDB 
of this coDJngation are very numerous in the older language; the 
following soheme Instaaces all that have been found to occur, 
■dive, middle. 

bh4vanl bb&vSva 
' Ibh&vftB 

bbftvatbas bfa&vStba 

bh&vfttas bh&T&n 

bh&vama bhivai bhAvfiTiibU bb&v&mali&i 


ibhivatal »i^v«to IbbAvtotal 


737. The 2d da. mid. (bh&Tftithe) doee not cbHnce to occm in tbig 
clua ; uid y&tUte Is the only example of the 3d penon. No aach pi. 
mid. forms u bhivfidbve, bh&Tftnte ate made ttom any ctua with item- 
flnil a; inch » bh&vanta (which are very common) are, of eonrse, prop- 
erly aogmentleu imperfects. The Biiibmanas (especially (B.) protei the 
2d sing. act. in ftBi and the 3d in at AB. ha« tbe 3d iing. mid. baraUU; 
and a 3d pL In antfti (Tartantfii KB,} ha« been noted once. BY. has 
examples, areft and madS, of the briefer 1st tlnf. act. 

itizecy Google 

7S8— ] IX. Prebbht-«y8tbu. 266 

8. PreMBt Optative. 
788. He scheme of optative endings a« combined with 
the final of an a-stem woe given in full above (See). 
■etlTe. middle. 

l^UTT >fe ^R >1^ H+jf^ *WIlf^ 

bh&veyam bh&veva bb&vema bb&veya bh&vevahl bbivemahl 

bh&v*a bhivetam bhiveta bbivethfta bh&vey&thKm bh&vedb.vain. 

bh&v«t bhivet&m bh&veyna bb&veta bh&veyat&m bh4ver&n 

a. Tbe RT. hu once the 8d pi. mid. bhorerata (for one oUiei 
•iMuple, ice 763 b). AV. h» udeyam tiom y'vad. 

b. A. faw Initanoes are met vith ot middle 3d penoiu from a-stemi 
In Ita >nd (very tardy) Iran, Insteiid ot «ta and eran. Foi conveiilence, 
they ma; be pal logethei here (eiceptiag the moie nauetoua cantatiTe 
tarmi, for which tee 1048c); they are (to fu as noted) theee: oajita S. 
and later, faAaita S., 9ra^ta S. ; dliaylta S., dtayayita U., hvajita 

., dhmSylta V. An active fonn qafialyftt C. is 

4. Present Imperative. 
788. An example of the imperative inflection is: 

acttTO. middle. 

H^nft JTsn^ JHPT ^ H<4N^" H=(IH^ 

bh&vsnl bb&vava bh&v&ma bh&vU bh&v&vahKi bh&vSmahU i 

bh&va bhiivatam bhiiTata bb&vaava bh&veth&m bb^Tadhvain 
>1^ H^rTFI^ J|^ HNHIH H^HIH^ M^lllM^ 

bb&vatn bh&vat&m bhivanta bh&vatftm bh&vetftm bbAvantCm 

740. The ending tana la 2d pi. act. Is u lue Id this nhole conjuga- 
tloo u la thana in tbe pteaent: tbe V. afforda only bhaJatana Id tbe 
a-oliH (and nahyatana la the ya-clsss : 760 o). The ending tfit o( 2d 
alng. act., on the other band, la not tito; the RV. tau avatftt, Ofatftt, 
dahat&t, bhavat&t, yaecbatSt, ySoatat, r&k^atat, vahat&t ; to which 
AV. adds Jlnvatftt, dllfivatSt; and the Brahmanaa bring other eiuBplec. 
MS. has twice Bvadfttu (parallel teita both tliae« svadfitl): compare 
almllar cases In the &-clasa: 7KSa. 

ioy Google 

267 A-0I.A88 (FIBST, bbO-CLASB). [—744 

S. Freaent Participle. 

741. The endings ^ ant and ITH mSna ate added lo 
the pteBent-stem, with Iosb, before the foimer, of the final 
Stem-Towel: thiu, act. H^fT bliivont (fern. Hcflfl bbdvana); 
mid. H^IIH bhivamSna. 

(k A small nnmbei of middle pirtlclplea tppetr to be made from 
sterna of tliU clatB (m of other a-elaiaea: aee 763 e, 104S f^ b; the 
snffli l&« inatead of mftna : thna, namlna, pao&nn, f Ikf &Qa, STaJ&na, 
hTa;ftua (all epk), mi^ftna and ka^ll^a (later); and theie «e Vedlc 
examples (aa oy&Tftna, prath&nfc, y&t&na oi yatftni, (i&mbh&iia, all 
RV.) of which the character, vhethei piesent or aorlat, ia doubtful: compare 
B40, 863. 

6. Imperfect. 

742. An example of the imperfect inflection is: 

aetlve. middle. 

HOWl ^^^ W^(\^ 5PR gH=IRf^ SIHMIHf^ 

AbhATam AbhaTava &bhavfima dbbave ibhavaTatii Abhavsmabi 
^iism^ ^>1^?T^ ^^l^rT SpigSira^ 5H^I^ ^il^ia'J^ 
^bbavas &bhavatam dtbbavata AbhaTathAs AbbaTAtb&m AbbBTadhTam 

Abbavat Abbavatim Abbavan Abbavata dtbbavst&ni Abhavanta 

743. No forms In tana are made In tbia tense tiom any a-c1ala. 
Examples of angmentleBa forma (which are not ancommonj are: oyfcvam, 
ivas, d&baa, b6dbat, bh&rat. o&ran, niiqan; b&dhatbKe, v&rdhata, 
g6oanta. The aabJonetiTely u«ed forms of 2d and 3d sing. act. are more 
frequeot than those of either of the proper subjunctive persons. 

744. A far larger Dumber of roots form tbeir preBent-syetem 
accordiog to the a-olasB than according to any of the other claasee : 
in the RV., they are about two bandred and forty (nearly two fiftfaa 
of the whole body of roots); in the AV., abont two hnndred (nearly 
the same proportion); for the whole language, tbe proportion is still 
larger, or nearly one half the whole number of present-steins : namely, 
over two hnndred in both earlier and later language, one hnndred 
and seyenty-five in the older alone, nearly a hundred and fifty in the 
later alone. Among these are not a few transfers from the classes 
of tbe first conjogatioD: see those olassce above. There are no roote 
ending in long & — except a few which make an a-atem in some 
anomalons way: below, 749 a. 

ioy Google 

■746 — ] IX. PRE8ENT~STSTEH. 268 

Irregolaritlea of the a-olaes. 

745. A few Terbe have irregQlar Towel-cbaoges in fonaing the 
present-stem: Urns, 

n, Qh comider )lu Eu^-ittengthenlns (•gtlngt 240): thiu, 6tiate. 

b. kfp [or krap) lament, on the contiuy, remslna nnchingad: thus, 

o> KUb hide hsi prolangatlon loetead of KU^a: that, gdbatL 

d. kram stride lesnluly leDgthena iu Tovel in the BctlT«, bat not 
Id the middle: thus, knbnati, kr&mate; but the TOwBl-qnaotltlfls are 
gem«wh&t mixed ap, evan fioin the oldest liDgnage down; — Uam. lire li 
■aid to Fonu klSmati etc., bnt la not qnotable; — cam with ths piepo- 
■itioD B rtnia the mouth fotmg aoKmati. 

•. Id (he Utei language are found oceasloQal fonns of thli elasa torn 
mfj u^e ; and thef show the lame vfddM (Instead of ffu^a) which belongs 
ID the root Id its more proper inllectlaii (627): thna, mBljaava. 

f. The giammarlaas giTe ■ nombei at tools in urr, which the; declare 
to lengthen the n in the pieseat'Btem. Only three are foand in (quite 
limited) use, and they show no forms anywhere with »hort n. All appear 
to be of Bccondary formation from roots In f or ar. The root maroli or 
m&rotl coagulate ha« likewise only a Id quotable foimi. 

g. Tfas onomatopoetic root f}\iuv epew is wdtten by the grammarians 
as (thlT, and declared to lengthea It) Tovel Id the present-ajstem : com- 
pare 240 b. 

746. The roots daAf bite, rofu color, nafij hang, svaiU embrace, 
of which the Dsaa] is in other parts of the conjugation not constant, 
lose it in the present-ay stem: thus, d&qatl etc.; eafij forms both 
aajati and aajjati Iprobably for sajyati, or for aasjati from aasa- 
Jati]; math or manth has mathatl later. Id general, as the present 
of this class is a strengthening formation, a root that hu snob a nasal 
anywhere has it here also. 

747. The rooU gam go and yam reach make the prosent-stems 
g&ocha and y&ooha: thns, gfcaahftml etc.: see 008. 

748. The root sad lit forms aJda (conjectured to be contracted 
from Biada for aisada): thus, sid&mi etc. 

748. Tranefeis to this class from other classes are not rare, as 
has been already pointed ont above, both throughout the present- 
system and in occasional forms. The most important oases ate the 
following : 

a. The roots in I, eth& eland, p& drink, and ghifi tmell, form 
the present-Btems t{ftba (tifth&ml etc.), ptba (plbfimi etc.], and 
jfghra (j{ghraml etc.): for these and other similar cases, see 671-4. 

b. Secondary root-forms like Inv, Jlnv. pinT, from simpler roots 

ioy Google 

269 Accented &-ctASB (siith, tud-CLAss). [—763 

of tbe hu-cIbbs, are either foand nloDgude their originals, or have 
crowded these out of use: see 716. 

7B0. Oq the other hand, tbe root dham or dhm& blow farmB 
its present-Btem ^om the more original form of the root: thue, 
dh&mRtl etc. 

VII. Accented i-class (sixth, tud-class). 

761. The piesent-stem of this class has the accent on 
the class-Bign ^ i, and the root remains unstrengthened. In 
its whole inflection, is follows so closely the model of the 
preceding class that to give the paradigm in full will be 
nnnecessaiy (only foi the subjunctive, all the foims found 
to occur will be instanced). 

762. Ilzample of inflection: root ^^ VI9 enter; stem 

I. Present Indioative. 

■etiie. middle. 

i ieJWlfi f=15II=(H fejailH+i ^15f fil5FR% ^RITR^ 
viQ^mi vl^ltvaa vlgimas viq6 vlf^vahe vlqamahe 

etc. etc. etc. etc. el''. etc. 

2. Present Subjunctive. 
1 vl^ltnl vlfjva vigitina vi^ii vlqavabU vlq&nahKi 

■ C5? "'*'» -'" KsSi "'•"• "'*•*" 

a, A single example of tbe biJefei Ux elng. ict. is mifk^a. The only 
forms in ftitlie and Bite ue p^^filtbe and yuvlite. 

3. Present Optative. 

vlq^am viq^va vi^dma viqeya vlq^ahi viqtoiahi 
etc. etc. etc. cfc, et«. etc. 

b. The BV. bu tbe ending tana once In Hretana 2d pi. act, and 
rata in jofeiata 3d pi. mid. 

ioy Google 

76S— ] IX. Present-system. 270 

4. Present Imperative. 
The fiiBt persona having been given above as subjunc- 
tives, the seoond aie added heie: 

vigil vlq&tam vlqiita vlq&ava vlgitliSm vlQ&dlivain 

flic. etc. etc. etc. «ta. etc. 

0. Tbe ending tat Is found In RV. and AV. In mr^atftt, vrhat&t, 

suvat&t; otber eumplei aie not Intreqnent in tbe BnbmiiDa languaige : 

thus, klildAtftt, ebyat&t, proohatAt, viqatftt, B^jatsti and Uter, npi^a- 

t&t. The 3d slug. act. audfttu and muiloStu oocqj In 8utrai (ef. 740). 

G. Present Partioiple. 
The active participle is fc)HIH vl9dnt; the middle is 

d. The feminine ol the actlTe participle it asnallT made from the 
strong Etem-fotm: thns, vig&ati; but aometimea l^om the weak: thus, 
Blfio&nU and alfioati (RV, and AV.), tad&nti and tudati (AT.): tee 
■boie, 449 d.e. 

e. Middle participles in &na instead of m&na aie dbuvftni, dhrfA^&i 
Uqana, gy&na. In the older language; kpqKna, muilo&Da, sp^^&na in 
the later (cf. 741a). 

e. Imperfect. 

i qiysiH ci^Rira sri^'sniT ^i^ qfaroR^ tii^simi^ 

&vlqam &vlQfiva Avlfftma &viQ« ivlfSvahl &viQSmahi 
etc. etc. etc. etc. etc. etc. 

f. Examples of angmentlesB form« accented ue ttjiko, apjit, tirinta. 

g. Tbe a-aorlit (846 ff.) ia In general the eqniyalent, aa regards Ita 
forma, of an Imperfect of this class. 

763. Stems of the A-cIsbb are made from nearly a hundred and 
fift}* roots: for aboat a third of these, in both tbe earlier and the 
later language; for a half, In the earlier only; for the remainder, 
nearly twenty, only In the later language. Among them are a number 
of tranafera from the otABsea of the non-a- conjugation. 

a. In some of these ttanafen, aa pff and ni|^ (731), there takes 
place almost a setttng-np of Independsnt roots. 

b. The atoms looliii, uoob&, and jocHkit m reckoned aa belonging 
reipecttiely to the roote 1; Jttire, vas ehine, and j go. 

0. The roots vrllten b7 the Hindu grunmariana witb final o — 
namely, tiho, do, qo, and ao — and forming the preeant-itemi 0b7i> 

ioy Google 

271 AOCBNTBD A-CLAS8 (SIXTH, tud-CLABS}. [—'69 

<l7&> <}yk, B7&, ara more propeily [u hiilDg ui ucraUd it In the «tem) 
to b« tMkoned to thii cUu than to the jK-claia, where (he natlTe elaail- 
flMtlon pati them (tee 761 g). They appear to be analogoiu with the 
ilemi k^ya, SVB, hvm, noted beloir (766). 

764. The loota from which &-(temi ue mad* hare certain noticeable 
pecularitlea of form. Hatdlf any of them hiTe long Towala, and none ha*e 
long Interior Towela; very few have final tdwcIi; and none [aaTo two ot 
three tranafers, and ylaU ba athamtd, which doea not occur In any aoean- 
tnated text, and la perhapt to be referred rather to the ft-clais] hare « n 
radical rowel, except aa thli forma a combination with r, which li then 
ledneed with It to f or aomo of the nanal labttltatea of j. 

IrregularitieB of the d-olaes. 

7&fi. The roots in 1 and a toA n change those vowels into I7 
and OT before the oluB-sigo: thus, kfiT'ii jrawi, mvA; saT&, etc.; 
and BVB, hvk oocar, instead of auva and htiva, in the older language, 
vhile TS. has the participle kqy&nt. K. has dbfiva from |<'dhQ. 

766. The three roots Id x form the present-stems kirft, eiri 
(also glla), tlT&, and are sometimes written as Ur eto-i and eat. Jar, 
tur are really oolf varieties of wCi Jf> Ti and blinr and sphur are 
evidently related with other ar or r root-forms. 

a. The common root praob atk makes the stem prcohi. 

767. Aa to the atema -driy& and •prlya, and mrlyi and dhriyi, 
(omeltees reckoned a* belonging to thU elaa«, lee below, 773. 

768. Although the preeent-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 bjr a pennldmate nasal. Thus, 
the stem mnfLo& is made from ymuo reUtat; aiiloi from yslo tprirMa; 
viadi Irom yviA find; kputi from yt^ cut; PIA9& from ypiq 
adorn; tfinp& from v'tTP "V'oy; lnmp& from ylup break; llmpA from 
yllp *maar\ and occasional forms of the same kind are met with from 
a few others, as tnnda from /tud thnut; hjthk from ybfb »trength«n\ 
A^fiiik (beside dfAha) from y'dfh make firm; <fam.bhit (beside qumbha] 
from yqnbh ihint ; TS. has fpithatl from i^qratb (instead of f rathn&tl) ; 
nfioha, vindbi, aombba, are of donbtful character. 

&. Nasalized &-alemi are *l)o In isTeral Inttancea made by tianafer 
from the nasal clati: thni, unda, nmbha, jiik, piAf&, ynflja, mndlia, 

VIM. Ta-orass (fourth, divdass). 

769. The piesent-stem of this class adds TS ya to the 
accented but luutiengthened root. Its inflection is also pte- 

ioy Google 

769—] IX. PRMEKT-BYSTBlf. 272 

cisely like that of the ftHslass, and may be pieaented io the 
same abbieviated fonn as that of the i-class. 

760. Example of infleotion: loot ^ n*lt bind; 
stem ^^ n&hya. 

1. Freaeut IndloatiYe. 
■ctiie. middle. 

i H<^iiH R?iraT!^ -RariH^ r^ -wjm^ 4^m^ 

n&bTtmi nUiy&vaa DUirftmaa natara nihrtTaha nibriinaha 
2. Present BubjonotlTe. 

nUiyBsU n&hyftdlivCf 

nUyfitas n&hy&a nUi^atfii n&hrintU 

I n&hr&ni n&hyftma nJihyfii nUirkvahU aUiyi 


a. A 3d pi. mid. in antftl (JXyant&i) oocare once Id TS. 

3. Present Optative. 
1 R#OT^ =1^ R^ R^IT H-S-Mf^ H'SmI^ 

nAtayaTsm nihreva nihyema nibyera nitayevahl nibysiaalil 
etc. eto. ato. etc. etc. etc. 

b. Foi two or tliie« 3d ting. mid. Itnmt !□ ita (foi eta), see 788b. 

4, Present Imperative. 

nihra niiliyataia n&hyata n&hyasva n&hyetliSm n&hTadlivaia 

etc etc. etc, et«. etc. etc. 

0. Of the ending tana, RV. hu ooe ei^mple, nahyataoa; tb« end- 
ing tat ii found In aayatSt, khySyatftt, na^yatAt. 

6. Freaeat Farttolple. 

The active participle ia q^itl n&liyant (fern. R^Jrfl nih-- 
yantl); the middle is R^iTTH aihyamSna. 

e. Imperfect. 
1 fcH-WH^ MH-SR SR^iT SR^ tW^Rl% HH-AIIMf^ 
&nahjram ^natayftva inahyBma iuahya &nahy&vabi Anahylmabl 

ioy Google 

273 ■ya-CLlM (FOURTH, dlv-CLASS): [—761 

d. EitnipleB of angmentlBas foims ghowtng the Moent belonging to the 
pTwent-iytteiii ue gt^Rt, p&^rftt, p&q7ui, J^atbSa. 

701. The 7ft-claBB Btema are more than a hundred and thirty id 
nnmb«r, and nearly half of them have forms in ose in all periods of 
the langnage, about forty occurring: only in the earlier, And abont 
thirty only In the modem period, 

a. Of tbe toot) making yft-Btemi, i yery ooniideitble part (OTei flfty} 
tignlfy a Bt4te of feeling, or t oondition of mind ot body: that, kap b» 
<mgTy, klam b« leeary, k^ndli be hungry, mnh b» confuttJ, lubh be 
hatful, qof be dry, etc. etc. 

b. A fnithei numbeT bMe ■ more or leu dliUnctly pustre mdh, 
dnd are In part evident and In pait pielumable transfeca train the paasire 
or y&-clatB, with change of accent, and sometimeg alto with aaBOmptioa of 
teOy* endings. It it not poailble to diaw precisely the Umila of the divi- 
■lon; but there are in the older langnage a nambei of clear oaaea, In which 
the accent vaTers and changes, and the othen ace to be Judged by analogy 
with them. Thna, ymne forms m&oyata once or twice, beilde the nsnal 
mQcy&ta, in BY. and AY.; and in the Biihmanas the former is the 
regular accent. Similar changes are found alio in ya-foims from other 
roots: thus, bom k(i dettroy, JI or JyS Hirwre, t^ hwt, d^ makt firm, 
ptto cook, px fi^' »^ damage, rio have, lup break, hft leave. AetiTa 
fonos are early made tcoia aome of these, and they grow more common 
later. It Is worthy of special mention that, {torn the Yeda down, Jtfyate 
M bom et«. Is fonnd as altered piuire or original ya-Cormatlon by the tide 
of i^aa gioe birih. 

0. A conElderable body of loola (about forty) differ from the above in 
having an apparently original transltlTe oi neater meaning: example* ire 
as throtB, nail bind, paq tee, pad go, glif claep. 

d. A nnnber of roots, of rattans meaning, and of somewhat doabtfnl 
character and relationa, having preaent-atems ending In.ya, ate by the native 
grammaiians written with flaai diphthongs, U or e or o. Thos: 

e. Boots reckoned as ending ia ti md belonging to the a- (or bhil-) 
class, aa ffU >mg (gityati etc.). As these show abandamly, and tor the 
most part exclnsively, O-forms ontalde the ptesent-syitem, there seems to 
be no good reason why they sbonid not rather be regarded as A-roeta of 
the ya-elasB. They are kf^K bwm, gS »ing, gift be aeary, tr& itae, dbyB 
thiuk, ■py&JiUup, mis relax, r& bark, t& be blown, 97ft coaguMe, fr& 
boil, ntyS ttifflui. Some of them ate evident eitensiona of simpler roots 
by the addition ot II. The secondary roots tKy rireieh (beside tan), and 
dy obiertte (beside ei) appear to be ot dmilar character. 

t. Roots rsckoaed as ending in e and belonging to the a- (or bhtlO 
class, u dhe tuc/t [dh&yati etc.). These, teo, have i-forms, and some- 
times I'forms, outside the present system, and are best regarded as ft-roota, 
either with ft weakened to a before the class-sign ot this class, or with ft 

Whitasy, Onmiiiar. >. ti. IB 

D,j,i,...., Google 

761— J IX. PRBBBNT-BreTBM. 274 

weikeiiAd to i or 1 and InllecMd UMndlng to tha a-clui. The; ue dhi 
tuck, mS exchange, V& weaoe, vji envelop, hvft call (leixiiidary, tram 
htk). Ai of kladred form mi; be mentiooed day ehare tnd vyoy sxpantf 
(pTobtlilT denoiniaUiT« of TTft;a). 

g. A tew tooU ■nlfldilly written vlth flnil o uid reckoned to the 
jtb-diM, with ndictl vowel lost before tbe clus-elgn: thm, do cut, bind, 
pres, dy&ti etc Theie, u haTlDg in icceated A In the lign, h»e 
pUiDly no right to be put tn thia cImi ; &nd they ire better referred to the 
JkeUiB (lee >bOTe, 768 o). Outalde the present-Byitem they ihow ft- and 
i-formt; (Dd In th*t system the ya la oRen Teaolved Into la in the olde*t 

7€8> The ya-olua la the only one tbas far described whioh ahowa 
any tendency toward a reitrictloD to a certain variety of meaning. In this 
tendency, le well aa In the form of Ita (Ign, It appsan related with the 
claia of distinctly dsflned meaning which is next to he taken Dp — the 
paaslTe, with y&-algn. Thongh very far bom being as widely as»d as the 
Utter beside other present-iystecDs, It la In some cases an IntranaltlTe 
conjugation by the side of a tranaltlve of aome other class. 

Irregularities of the ya-olasB. 

763. Tbe roota of this olass ending in am lengthen their rowel 
in forming the present-steiii: they are klam, tam, dam, bhram, qam 
be quiet, Qnun: for example, Utmyati, gramyati. From tcqam, how- 
ever, only k;amyate occurs; and fam Mor makes ^amyati (B.). 

764. The root mad has the same lengthening: thne, mitdyatl. 
76B. The roots in iv — namely, dlv, aW, anv or ^riv, and 

ijtliiT [from which no forms of this class are qnotable) — are written 
by the grammarians with Iv, and a similar lengthening in the present- 
Bjstem Is prescribed for them. 

a. They appear to be properly dlQ etc., alnoa their vocaliied flasl 
in other forms le always a-, dlv Is by this proved to have nothing to do 
with the aaanmed root dlT ehine, which changes to dyn (S61 d): compare 
840 b. 

766. pTom the roots J; and tf (also written as Jnr and tlr or tnr) 
come tha stems jirya and tirya, and Jirya and tiarTa (the last two only 
in RV.); from pf comes plUrya. 

767. The root wyadh la abhrevlated to vldb; thus, vidhyatl. And 
any root which In other forms has a penolttmate nasal loses It here: thus, 
df hya from d^Ah or d^h ; bhrafya from bhraAf or bbraf ; rskiy% Iron 
ralU or raj. 

oy Google 

275 ACCHHTBD yd-CLA8S (PASSIVE). [—771 

IX. Accented ra-class: Passive conjugation. 
768. A certain foim of preseat-stetn, inflected with middle 
endings, is used only in a passive sense, &nd is formed 
from all loots fot which there is occasion to make a. passive 
conjugation. Its sign is an accented U yi added to the 
root: thus, ^^ hanyd from y^ han slay, W^ 5pyi 
from v5iq Sp obtain, IRET ex^^ fto™ V^ gTh {or grab) 
seize: and so on, without any reference to the class accord- 
ing to which the active and middle forms are made. 

760. The form of the root to which the pasBive-sigD Is added 
is (BiDCe the accent is on the sign) the weak one: thus, a pennltim&te 
nasal [e dropped, and any abbreviation which ie made in the weak 
forms of the perfect (794], in the aoriat optative (929 b), or before 
ta of the pusive participle (964), is made also in the psBsive present- 
system: thns, b^j6, from yaftj, badliy& from )<l)andli, aajk from 
Vvao, ijyA from VyaJ. 

770. On the other hand, a final vowel of a root is in general 
liable to the same chaoges as in other parts of the verbal system 
where it is followed by y: thus — 

a. Final i and u are lengthened^ thuB, miy& from yml; a&yi 
from yen; 

b. Final S is usually changed to i: thus, diy& from ydl; hiy& 
from v^ft: but jiUyi from j/ifiKi and so Uiy&y&, kIiKy&, nmSyi, etc.; 

o. Final r i" i" KCoeral changed to ri: thns, kriyi from ykp; 
bat if preceded by two consonants (and also. It is claimed, in the root 
f), It has instead the gu^a-strengthening: thus, smaryi from ytmf 
(the only qnotable case); — and in those roots which show a change 
of r to It and ur (so-called ^verbs: see S42), that change is made 
here also, and the vowel is lengthened: thus, qlry& from ytj^; pOryi 
from ypf. 

771. The inflection of the passive-stem is precisely like 
that of the other a-stems ; it differs only iu accent from that 
of the olasB last given. It may be here presented, therefore, 
in the same abbreviated form: 

a. Example of inflection: root ^ Ilj make; passive- 
stem f^m kriyA: 

ioy Google 

771 — ] IX. PBEeEHT-STSTEH. 276 

1. Fresent IndioatiYe. 

». i. ^ p: 

kriri ki4;itvalie krly^Klie 
etc. etc. etc. 

a. pTSBent SubjunotlTe. 

b. The forms noticed u (MonrriDg ia the older lasgokge ue alone 
here inatanced: 

d. p, 

1 krlTll kriy^ahU 

3 kriyidhvU 

' CSi '^''-"^ 

0. The 3d pi. ending antfit l» foand ouoe (uoyant&l K.). 
3. Present Optative. 

kriyAya kriy^Tahl kriyfenutlil 
ete. etc. ate. 

d. No toimt of the puaiTe opt&tive chince to occur In BV. or AT.; 
th*T ate toead, howoToi, In the BnhmkDU. GhD. hu once dhmi^ta. 

4. Freeeiit ImpertttiTe. 
kriyftava kriy^th&m kriy&dtavam 

B. Freaent Farttoipld. 

e. This is made with the suffix ifR mftna: thus, f^UMim 

f. Id Dse, thli putlcipla Is wall dtstlngaJBhed ftom the otlieT putlve 
putlclple by iu dUttuottTetr pieaeni meaning: Qtm, kft& done, bntkrljA- 
mft^a in proee** of doing, oi being don«. 

e. Imperfflot. 

&kriye AkrlyKvalil UrlyBmahi 

g. The pitiiie-Blgn ie ne^ac leaolTed into la in the Tedi. 

77S. The roots tan and khan oBoally form their paBiivea from 
parallel roots in &; thus, t&y&ta, khSy&te (but alio tanyate. kbaa- 

ioy Google 

277 So-called Tbnth or aut^Lxae. [ — 77B 

Tftte) ; and dluun, in like muiDer, makes either dlutmyRte or dtamKrite. 
The correapontUng fonn to VJwi, namely J Jtyat* (above, 761 b}, ii 
Appareotlf a trsnefer to tbe preceding elaaa. 

773. By their form, miijkte diet, and dbriy&te matntaint itseff, 
it ittad/oft, are pMsives from the toots mj die mi dtaf hold; although 
neither is used id a proper passive sense, and mj is not transitive 
except in the derivative form jtvp^ (above, 731). With them are to 
be compared the stems ft-(Iri7& heed and &-prly& be buty, which are 
perhaps peculiar adaptations of meaning of pasBivee from the roots 
d^ pitree and pf JiU. 

77i, Eitmplea of the tnniteT of stems bam the yi,- m pwiiM 
clwa t« the ya- di iatnnsltl*e class were gtven sltoTe (761 b); and it iru 
also pointed mit that acUva Instead of middle endlnfB are occaitoBall;, evui 
in the eadlei laofnafe, asanmed by forau pi^erly pasalTe; examtilas ate 
i dtamftyatl and vy hprnqyat (QB.), bbOyatl (UIID.). In ths eplca, 
however (as a paK of thalt genaral coufniioQ Of aetlTe and middle fonna: 
BSSft), active endlnga ai« by no m««nt infteqnently taken by the psi«iv«: 
thus, Qakyati. QrOyanti, bhrlyantn, IJyimt-, ate 

The BO- called Tenth or our-OlMs. 

77S. As was noticed above (607), tlie Hindu grammarians — and, 
after tbeir example, most European also — recognize yet another 
coqjngation-clasB, coordinate with those already described; its stems 
show the eUss-sign iya, added to a generally strengthesed root (for 
details as to tba strengtheDlng, tee 104S). Though this is no proper 
class, but a seoondary or derivative conjugation [Its etene are partly 
of causative formation, partly denominative with altered accent) an 
abbreviated example of its forms may, for tlie sake of accordance 
with other grammars, be added bere. 

a. Kxample: root oint think, meditate; stem clnt&jra: 

setiTs. mlddlp. 

Pies. Indlo. otat&ySiai elnUya 

Subj. otnU^&nl oint&yU 

Opt. clat&yeyam olnt&yeya 

Pple. olnt&yant ointiyam&na 

Impf. Aolntayam Aointaye 

b. The ^nfleeilon, of eouise, la the tame vith tbit of other tormi from 

o. The middle paiHciple, in the later language, ia more often made 
with Sua IniCead of m&na; thtie, clntayOna: a«e 1043 f. 

ioy Google 

Tie—] IX. PBESEKT-STflTEK. 278 

Utts Of Ihe Present and Imperiect. 

776. The naw of the mode-foraiB of the preBent-STfltem have 
been already briefly treated In tbe preceding chapter (672 ff.)- The 
tense-uHH of the twQ indicative tenaes, present and imperfect, call 
here for only a word or two of eiptanation. 

777. The present has, besides ita strictly present nse, the same 
subsidiary nsea which belong in general to the tenae: namely, the 
expression of habitnal action, of fnture action, and of past action in 
lively narration. 

a> Examples of fataie msftnlng are: itnilh oAd vA imi dnv&te 
t4U eri no 'bhlbhavanti (^R.) verify if t&Me build thit up. Dun Otg 
uiili itraightteay gei the better of m»; agnir Stmabhavadi pr&d&d yatra 
Tft&ohatl nftifadha^ (UBh.) Agni gone hi* ovm pretence lehereeer the 
Xitftadhan should desire ; Bvigatadi te 'Stn Ulh karomi tava (R.) vd- 
eome to thee; tehat thall I do for theet 

b. EiRiDpleB of paetmMntng sre: 4ttar& sAr idhara^ putri tud 
dbili^ faye aab&Tatail a4 dheno^ (KT.) the mother woe over, the eon 
under; there Dinu lie*, lUce a eoto toith her eatf; prahaaanti oa t&ib 
keold abbyaanyanti oK 'pare afcurrata daySdi kecit (MBh.) *oine 
ridicule her, aome revile her, tome pitied her; tato yasya vaoanU tatrft 
Valambltaa tadi earve tlraakurvanti (H.) thereupon they all /all to 
reproa<Aing him by uthoee advice they had lUighied there, 

778. Id connection with certain particles, tbe present has rather 
more definitely the value of a past tense. Thus; 

a. With ptu:^ formerb/: thus, Baptarfin u ba ama vftl pari 
rkf ft ity acakfate ((IB.) the *ecen tagee, namely, are of old called the 
bear*; tanmstram api oen mahyaih Da dad&ti porS bbavKn (Hfih.) 
if you have never before given me even an atom. 

b. Wltb the uieTentlve puttde ama: tbna, qr&meqa ba ama v&i 
t&d devi jayantl jkd BtfBib jiyyam £a4 r^ayaq ca (^B.) in tmlh, 
both god* and eages were icont to win by penance what wa* to be won ; 
ftvlft*^ kallBS dyQte Jlyate ama ualM tadft (HBh.) then Nolo, being 
poieeteed hg Kali, woe beaten in play. 

0. No enmpla of thts last conUiuction la found In either BV. or AT., 
OT slsevheie in the metiiol p*rt* of ths Vedi. In tbe Bnhmuiu, ODljr 
habitual action la expiHted by It. At all period! of tbe Ungnige, (be use 
of ama *Uh > verb ae pure aiseveratlve particle, with do effect on tbs 
teaae-mesnlng. Is very common; and tbe examples later are hardly to be 
diitlnguished ^m the present of liielf narration — of vblch tbe vfaole 
coDBtrnction la donbtleas a form. 

778. 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. 

a. Compare what is said later (end of chap. X, and obap, XI.) as to 
tbe taluB of the other past tensea, the perfect and aoriat. 

ii.zecy Google 

Characteristics of the Pehfeot. 



780. Thb petfect-system in the later language, as baa 
been seen above (B3B), consists only of an indicative tense 
and a participle — both of them in the two voices, active 
and middle. 

a. In the oldest language, tbe perfect has also its modei and 
its augment -preterit, or pluperfect, or is not less full in its apparatus 
of forms than is the preBeDt-Bystem (see SOS CF.). 

781. The foimation of the perfect is essentially alike 
in all veibs, differences among them being of only subord- 
inate consequence, or having tbe character of irregularities. 
Tbe chaiacteristics of the formation are these: 

1. a stem made by reduplication of the root; 

2. a distinction between stronger and weaker forms of 
stem, the former being used (as in presents of the First 
or non-o-conjugationj in tbe singvl&i active, tbe latter in 
all other persons; 

3. eadiogs in some lespecta peculiar, unlike those of 
the present; 

4. tbe frequent use, especially in the later language, of 
a union-vowel ^ i between stem and endings. 

7S2. 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 (see 648] — but 
with this exception, that radical Q a and ^ ft and ^ r (or 
5^ ftr) have only sj a, and never ^ i, as vowel of the re- 
duplicating syllable: thus, from )/q PT_fill comes the preaent- 
.stem f^q pipr> but tbe perfect-stem tpn papr; from yij\ mS 

Diaii.zecy Google 

78a—] Z. Perfeot-ststeh. 28tf 

mstaure comes the piesent-Btem fifljf mimS, but the perfect- 
stem ifiTT mamK; aad so on. 

a. IcieguUrltles of TooMTKhlnitlsl coniiHiaDti will b«gl<enbe)o«, 784. 

7S8. For roots b^Qniog mth a vowel, the tules of 
reduplication ate these: 

a. A root with initial 9 a before a single final consonant 
repeats the ^ a, which then fuses with the radical vowel to ^&, 
(throughout the whole inflection): thus, iffl^' ftd horn v^ »d 
eai; and in like manner CfH ij, ETR ftn, tTTH Sb, sn^ Bb. The 
root ?? T forms likewise throughout STF^ Br (as if from ^ ar), 

b. A root with ^ i or 7 a before a single final conso- 
nant follows the same analogy, except in the strong forms 
(sing, act.); here the vowel of the radical syllable has gu^, 
becoming ^ e oi ^ o; and before this, the leduplicatii^ 
vowel maintains its independent form, and is separated bam 
the radical syllable by its own semivowel : thus, from i^ 
if comes ^ If in weak forms, but ^(^ iyep in strong; Irom 
yZ^ uo, in like manner, come ^Si^Qo and 3?)? uvoe. The 
root ^ i, a single vowel, also falls under this rule, aad forms 
^ ty (y added before a vowel] and ^ iya. 

o. Roots which begin with vowels long by nature or by 
position do not in general make a peifect-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, ehap. XV.: 1070 ff.). 

d. To tbtg nde, howeveT, y&p oilain (pratubly crlgiiuJlr ap: 1067 f) 
eonBtitute* tm eicepUoii, maUog the eonaUnt peifect-sMm ftp (u if from 
ap; Kbove, a). Alio are met irith 1^6 (BT.) and I<^re from yl^ and 
Irip* (V.) f"" y^- 

e. For the peealiu ledoplioatlon ftn, helonglng to eetttin wots widi 
Initial Tovela, aee btlow, 788. 

784. A nnmber of roots beginniag with va and ending with a 
single coQsoiunt, wlitch in various of their verbal forms »nd deriv- 
stivea abbreviate the va to u, do ft also in the perfect, and an 
treated like roots with iaitial u (above, 788 b), except that they retain 

ioy Google 

281 Bedupuoation. [—788 

the full form of root in tbe strong peraonB of the Biognlu active. 
ThiiB, from y'rao ipeak come uo and avao; from yv»a Aoeil cone 
flf and avas; and bo on. 

»• The roots showing this abbreviation are tao, ts^ vad, ro^i 
vaa, Tabi and tB Keavt Is said to follow tbe same role. 

b. A single root beginning with ya, namelf j^ ojf^, has tbe 
aame oootraetioD, fonning tbe stems lyaj and U- 

O, OccMioiuI ex<ieptlani ire met wltb ; m, vaTftea and Tavakfi 
(RV.)i vavfipa ind Tav&ha and vaTSbatua (E. and i»ut); yeji (T.). 

785. A namber of roots having ya after a first initial coDBonaut 
take 1 (from the y) instead of a in the reduplicating syllable: thus, 
from v'vyao comes vlvyao; from WJt comes plpyK. 

a. Tbeae tooti are vyao, vyath, vyadh, vyil, jy&, pyft. syand; 
and, In the Teda, alao ty^, with oyn and dyut, irhich hare the lOM- 
Tovel n. Other iporadlc cues occur. 

b. A single root with va is treated in the same way: namely 
Bvap, which forma sofrap. 

o. Tbne rooti arc for the moat pan abbrerfated in (be weak forma; 
Me below, 794, 

7B6. A considerable number of roots have Id the Veda a long 
vowel in their reduplication. 

a. Thnt, of roots redapllMtlng with ft : kan, Up, gf^b, tfp, tpf, 
drb, dbf, dbn< nam, mab, mfj, mi^, ran, radb, rabh, vafio. Tan, 
TAf, vaa elotht, V&9, v|j, vft, vrdb, vj^, <)»Aprtoaii, aah, skambh. 
Some of tbeae occur ouly in taolaled cases; man} bave alio (onna wilb 
ahoit vowbI. Host ue Vedic only; but dSdbiira fi common alio Id the 
Brghmana Ungaage, and la eTen foand later. Aa to jfigf, see 1030 a. 

b. or roots reduplicating witb i : the so-called loots [676} didbl and 
dldl, which mske the perfect from the same stem «lth the present: thus, 
dlditba, didtya; didbima, tfidbyns (alio dldbljrua, dldiyua). Bat 
pipl has plpye, plpyns, etc., tdtb ehort 1. In AY. occur* once JDu^a, 
and in AB. (^and AA.) bibblya. 

O. Of roots lednpllcating with tl: to, JO, and ^ (or QVS). 

767. A few TOota beginning with the (detlTative: 4S) palatal mote* 
and uplrttlon show a reTersion to the more original guttural in the radical 
ijllable after the rednplUation : tbaa, )/ol forma «ikl; yoit forms oikit; 
yji form* jigi; ybt forms Jighl; yiitca forme Jagban [and the same 
reTeialoaa appear in other rednpllested forms of these roota; 816,1). A 
root dft protect la said by the grunmarlane to form digl ; but neither root 
nor perfect Is quotable. 

788. A small number of roots witb initial a or r (Bx) show the 
anomalons rednplicatlon Sn iu the psrfect. 

a. Thus (the foimi occarrlng mainly In the older Ungnage only): 

ioy Google 

788—] X. Pbrpeot-btsteii. 282 

t^afij 01 aj, which foiniB the piei. an&ktl, bas the peifnct Snaflja 
&nd finnjd etc. (with (uu^& ind an^r&t}; 

yof ottotn (from *hlch coiaee odc« in RV, aa&^Kinahftl), hu the 
veik loratn Aiaqtna etn. [with opt. Snafyfiin), Saa^d etc. (and LgS. 
hu finaf adhve), and (he aliong fome ftn&Aqa and finSf a — along irith 
the regular B/jO, etc.; 

Y^^ ((torn which cornea once p^idliat) haa ftnfdliaB and ftnrdhe; 

y^fo 01 arc hw Knro&s and &aja6, and Utec fiiiaroa and Saarooa; 

yarb haa On TS.) Snrh&s; 

RnBIia (RT., odcb) has been teferaed to * root all, eliewhere ouknown, 
and explained aa of thia faimation; bat with altofetbei doabtfal piopriety. 

b. The latei grammar, then, acta ap the rnle that roots begnnning 
with a and ending with more than one oonaonint have &n aa their regular 
rednpHcation ; and aueh perfect! are taoght from roots like akf, Uj, and 
aiic or aO; bat tha only other quotable forms appear to be Suarohat 
(HBh.) and Onarfat (TA.) ; which are arcoidlngi; nckonsd aa "plaperfeots". 

788. One or two indiTidnal cases of irregalarity »re the followingi 

a. The extremely common root bliu be has the anomalODS redn- 
plication ba, forming the stem babhfi; and, in the Veda, f'an forma 
in like manner eaaQ. 

b. The root bhf bear haa In the Veda the anomalons lednplieatlon Ja 
(aa also In Intenilve: lOOS); bat RV. haa once aleo the legnlar babhre, and 
pple babhrft^. 

e. The root ((hiv ipmo forma either tllfthlv (^B. et tl.) or (ifthlv 
(not quotable). 

d. Tlvabv^ (KV., once) 1« doabtleas participle of /rao, with 
irregntar lednplicatlon (aa In the preaent, 990). 

780. Absence of reduplication is met with in some oases. Thus: 

a. The root vid hum has, from the earliest period to the latest, 
a perfect without reduplication, but otherwise regutarlj made and 
inflected: tbna, vida, v6ttha, etc., pple vldviAa. It has the mean- 
ing of a present. The root vid ^nd forms the regular vlvAda. 

b. A few other appnentl; perfect forms lacking a tedapUealion ate 
found in RV. : they are tak^athua and takqna, ram&tua, akambb&tbna 
and akambbuB, nlndlma (roc nlnldlma P), dbi^e and dhlre (P ) dh&), 
and vidti and artalxe (P see 618). And AV. ST. hiTc oetatoB. The 
participial words dUqvaAs, miipivaAs, SftbT^S are common In the oldest 
Ungoage; and RV. has once Jftniifaa (v'Jfift), and kbldvas (toc), perhapa 
for olkbidvas. 

o. A few aporadlc cues alio ate qnotabte from the later langnaga, 
especially from the epics: thus, kar^atue, oe^f ""^ oeffatuB, bhrS- 
jatus, earpa, qaAaua and fafiaire, dtav&Aalra, Bra&slre, jalpire, 
edhire; also the pplee faABlTftfts and dar^lvKfis, the latter being not 

ioy Google 

283 Stronh akd Weak Stbm-fokms. [—798 

791. Fat to uiomtlong cue oi vo of Tedaplkitsd piepoBittoD, eee 
beloir, 1067 f. 

792. Stioug and weak stem-foTtna. In the three 
peisons of the singulai active, the lootrsyllable is accented, 
and exhibits usuallj a stronger form than in the rest of the 
tense-inflection. The diffeienoe is effected paitly by strength- 
ening the toot in the thiee peiaons leferted to, paitly by 
weakening it in the others, paitly by doing both. 

793. As regards the strengthening: 

a. A final vowel takes either the gn^s oi vrddM change 
in Ist sing, act., gu^a in 2d, and vrddhi in 3d: thus, fiom 
1^ bbl, Ist ^ hitihi 01 f^ bibldi; 2d ^ hibh6; 3d 
fsp^ bibbRi; from v^ ky, Ut tJc^,^ oak&r or *l*l^ oakfir, 
2d Tjtfi^' oak&r, 3d -f\4\\\ oakSr. 

b. But the u of fbha remains uDchaaged, and adda v before a 
vowel- an ding: thas, babhttva etc. 

o. Medial ^ a before a single final consonant follows 
the analogy of a final vowel, and is lengthened or viiddhied 
in the 3d sing., and optionally in the first: thus, from yrJ'J 
tap, Ist rra^I tat6p or fTrH'^^tatSp, 2d FTrP? tatdp, 3d ?THTO^ 

d* In the suiter langaige, howeTer, ibe vesker of the two forma 
kllowed hj tbese mloa in the But person U almost escluaivetT in use: thus, 
lat only blbhAya, tatipa; 3d bibbitya, tat^po. Exception* tro oakftra 
*nd Jagr^a (donbtfnl resdlog) In AT., oak&ra in A^S. tnd BAL'. fgB. 
eakars), jlgaya In A(S., u drst persons. 

e. A medial short vowel has in all three persons alike 
the go^a-stiengthening (where this is possible: 240): thus, 
from y^ drub comes ^<i^ dudT6h ; from yf^ vl^ comes 
f^c^l viT^9 ; from y^iH krt comes ^^ oakart. 

f. An iDitial abort vowel before a single final conaonant Is to be 
treated like a medial, bnt the qnotable examples are very few: Danely, 
lyafa from yi^ letk, uvooitha and uvooa from yao, uvofa from 
yof. Aa to roota 1 and p, wboae vowels are both initial and final, 
see above, 783 a, b. 

g. Theie ralea ite said by the griDuntiitin* to apply lo the 2d king. 
4lir*;s when it hu simple tha u ending; if It lia£ itha [below, 707 d). 

L,j,i,.ec.y Google 

199—] X. PBRPB0T-8T8TEIi. 284 

the accsnt la ftllowed to fall on any one of tli« ByllsUw of the word, uii 
the loDt'Bylltble If ansc«ented hu eomstkines the weak rona (nuaety, Id 
contisRted etemB with a (oi medi&l a: below, 794 a-, and Id ceitala otber 
leibs, Bs viidjitha]. The esTliBi Ungntge, howeiCT, affoide no example 
of a ^ sing., what«T«r Iti ending, aocented on any othai than the ladinl 
■Tllabla, oi raiUog to conform (o tbe (olei of itrengtheniDg as glvaD abofs 
(I. ^ o, .). 

b, Ocuilona] Inatancea of BtFengtheoing in other th&n the slngnlu 
pereona are met irlth ; thoa, yuyoplma and -vive^us (RV), paspkTfoB 
(ReU.), and, In the epici, ookartus and otOurtlJta, Mtkarf&tua, Jn^iu^ 
bire, nan&mira, blbbedtu, TftTUiatOB, viva^atuB, v-avarfiui. The 
loots df, pr, and 7r, and optionally Jf, are aald by the grammariana to 
ba<e tbe atrong atem In weak fbrna ; but no examplet appear to be quotable. 
AV., howeieT, hu once Jabarua [probabl; a falae reading); and in the 
later langaag* occur oaakara (f^ scatfer) and taatare. 

i. The root mfj hae (ae In tbe proient-syatem : 637) vrddbl tnateul 
of ipuja In strong forma: thna, mam^Ja; and v'gob (aim u In pteaant: 
746 o) haa Q Inatead ot o (but also Jagnbe E.). 

794. Ab regards tbe weakening in weak forme; 

a. It has been seen above (788 b) OuA roots beginning with i or 
n foM rednplicating and radical syllable together to I or q in the 
weak forme; and (784) that roots contracting va and ya to a or 1 
)n the reduplication do it also in the root in weak forms, the two 
elements here also coalescing to Q or I. 

b. A few roota hsTing ya and va after a first initial conaonant, and 
reduplicating from the aemivowel (IBS'), contract the ya and va to 1 and 
a: thoa, vlvio from j/vyac, vividb from yvyadh (but vivyadbus 
UBh.), sn^np from i^arap. The ettnnded roota Jy&t pyS, vyft, qvi, 
hT& abow a aimilai apparent contraction, making their weak forms tiom 
the almpUr roota JI, pi, VI, fil, bQ, while bvS maet and fvA may get 
their atrong forma slaa from the aame (and oaly jUyK& 1° qa»table from 
the othert). 

C. The root grabh or grata (If it be written thoa: eee 7B8a) con- 
tracta to Krl^i making the three forma of «tem JagrUi (1st and 2d slag. 
act.), Jagrib (3d), and JajTh; but praob (U It be so written : see 768 a) 
remains unchanged thioogboni. 

d. Soma roota omit In weak forms of this tense, «r la songe of tham, 
a naaal which le found ia It^ atrong forms; thus, we hare 0&kra<i6 etc. 
(RV.) from vtrand; tatasr^ (RV.) from ytaia; dada^viAa (RV.) from 
ydaA;; bedh^, bedb6, etc. (AV.) from )/baiidb; aejae (^B.) Crom 
VsafUi oaakabb&ni (AV.) from i/ekambb; tastabhus etc (T.), 
tastabh&n& (V.B.), from y'stambb. Compare alio 786 a. 

e. A number of roots having medial a between single oonaonants 
drop that Towel. These are, in the later laogaage, gtun, kban, Jan. 

ioy Google 

285 Strohq akd Wbak Steu-fobhs. [—796 

hKD, ghM; the^ form tbe weak stems Jagm, cMiUm, Ji^O. Jaghn 
(comure 8S?), }alai (eoBpan 640]: but RV. bu onoe J^anua. 

CtJJn tbe old ItDgnage ue foand in like mumei manm&tlia >nd 
ma%zkSt« fiom y ^non; ya-ra* ttom >'t(ui; tatne, tatniqe, tatniro 
rtom ytcm (boBlde tatane, and tate, u If ttam ytS); paptlma ind 
pt^Ub iDd pkptiTlEAB (ram Vpat (bciida pet-fonoB ; below, g) ; papn6 
from ifan; Baqolma uid bb^oub, aaqoe sod safoirA, tiom Vsao. 

(g.) Roota In gODeral having medial a before a aingle final con- 
sonant, and beginning also witti a single consonant that is repeated 
onobanged in the ledaplicaiion — that is, not an aspirate, a gattuial 
mate, or h — contr act their roo t >nd_rfldnpli-C*t'9.''. together into one 
ay liable , harfng e'^fi'lts vo wel: thus, yaaA forms the weak stem eed, 
ypao Torms peo, yyttm forms yvm; and so on. 

h. Gertiln looU not baving the form here deflned ue decUted by the 
giammBiliD* to andetga the eune contrtatlon — mait of them optlonallyj 
•nd examples of them are tn general of very rare occnrteQce. They are as 
follows: r^ (K.C.) and rSdh (radhf), notvlihitanding thelt long Towel; 
put?, pbal Obelire 0.), bhtO (occurs ftam RV. down), thongb their Ini- 
tial is changed In lednpllcatlon; trap, tniB (timuE.O.). QTatb, Byam, 
Bvan, thongh thej begin with more than one eonaonant; dambb (debb&e, 
BV., from the weaker dabh), thoogh it ends with more than one; and 
bhram (btaramoB etc KSS.). btarft], Krantli, btsIU, in iplte of more 
reasons than one to the oontnry. And fB. has aeJOB from VBBiU, and 
EB. bas qremiis fitom y^ram. On the other hand, RT. haa onca rarabb- 
mk, and R. has papatua, for petos, tioja /pat. 

L This contraction Ig allowed also In 3d ting, act. when the ending 
1b Itha: thtia, tenltha beside tatantha (bnt no examples are quotable 
fnini the older langnage). 

J. The roots qa; and dad (from dft: 872) are said to lejeot the 
contraction; bat no perfect forms of either appMC to have been met with 
In nse. 

k. From ytf (or tar) occurs teras [R.)i and Jeros from yjf Is 
aathorizcd by the grammarlaaB — both against the general analogy of roots in f. 

1. Boots ending in S lose their ft before all endings beginning 
with a vowel, including those endings that assame the nnton-TOwel 1 
(788) — nnlesB in the latter case it be preferred to regard the i aa a 
weakened fonn of the ft. 

795. Endings, and theii union with the stem. 
The general scheme of endings of the perfect indicative has 
been alread)^ given (55S c]; and it has also been pointed out 
(643 a] that toots eading in ^ A have ^ Su in Ist and 3d 
sing, active. 

ioy Google 

796— J X. PBKFBCT-B7BTEN. 286 

a. Tbe ending maB Inste&d of ma Is fonod Id ^nqrumaa (E.G.). 
Foithe alleged (MscmreDce of ^ve Insteul otdbv* In 3d pi. mid., eeeSSe c 

796. Those of the endings which b^n with a con- 
sonant — namely E[ tha, ^ va, IT ma in active ; H se, sj^ 
Tahe, R% mahe, 3 dhve, ^ re in middle — are veiy often, 
and in the later language usually, joined to the base with 
the help of an interposed union-vowel ^ i. 

a. The nalon-ioirel 1 is found wtdel; otei also in otbet ptkrU of the 
general leibsl syitem : namely, in the aiUluit aorlEt, the futnies, and the 
lerbal noani and adJectiTee (as also In other clasaea of derliatlTe ttcms). 
Ill the later langnage, a certain degree of ooReapondence i< seen amoug the 
different part^ of the same veib, ae regarda their oae or non-nae of the 
connective: but thla corteapoD deuce is not ao close that general rules res- 
pecting it «an be glien with advantage; and it will be beat to treat each 
fonnation by itieU. 

b. The perfect is the tense in which the use of 1 bu establisbed 
iteelf most widely &Dd firmlj in the later language. 

797. The most important rules as to the use of ^ i in 
the later language are as follows: 

a. The "^ re of 3d pi. mid. has it always. 

b. The other consonant-endings, except ^ tha of 2d 
sing, act., take it in nearly all verbs. 

o. But it is rejected thronghoat by eight verbs — namely Icp make, 
bhf bear, sf go, vf dtooee, dm run, 9ra hear, stu praite, srajlou)] 
and it ie allowably (not nsnallyj rejeet«d by some others, in general 
accordance with their nsage in other formations. 

d In 2d sing, act., it is rejected not only, by the eight 
verbs just given, but also by many others, ending in vowels 
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, including 
those in StT fi (of which the ^ S is lost when the ending 
is ^ itba), and most of those in ^ i, | I, and 3' u. 

e. The lulea of the grammartaua, eapecially aa reguds the nae oF tha 
or itba, lun out Into infinite detail, and are not wholly oonHatent with 
one anothei; and, as the forms are very Infrequent, it la not possible to 
criticise the statements made, and to tell how fai they are founded on the 
facts of uaage. 

i, Google 

287 Endinob. [— SOO 

f. With tbia i, a final radical i or i is not combined, bnt chang- 
ed into y or iy. The Q of ybbH becomes Qt throughout before a 

798. Id the older language, the usage is in part quite other- 
wise. Thus : 

a. la the RV., the anion-^owel i ia taken by moti ending In cod- 
■onantt proTided ths last gyllablc of the Etem it t heiTf one, bat not other- 
wise; thua, iaitha, uTooitha, viv^ditba, bat tat&nttut and vlvy&ktha; 
ncim&. paptiml^ aedima. ynyopimi, but Jaganma, Jagrblun&, ynrui- 
ma; aoif^, Ji^fii^d, sastthlqe, bnt vivitae and dad^k^i ; babboJmUie 
and ija^admahe etc. (no example* of Ivabe oi Imalie ebance to oeeni, 
not any of either idhT« or dtave); ijiti, Jaj&lr^, yetird, tatakqlri, 
but c&k}pr6, vividr^, duduhrd, pasprdhr^, tatasrd (and ao on: 
tneety-twa forou). The only exception in BT. ia v^ttba riom j/vid, 
withoot i (In Bt., also Sttha from yali: below, &01 a). Th« otbei Vedic 
teita present nothing tncoDBistent with tbia tale, but In the Bcabmanas 3d 
pi. forma in ire aie made after light ayllablea also: thua, sasfjire, bubadh- 
ire, rnyujlre, mradhira. 

b. In roota ending wltb a vowel, the early Daage la more.neuly like 
the later. Tbnai for roota in & the rale la the aune (except that no 2d 
alng. in itha ia met with), aa dadlilm&, dadlLi;^, dadhldhvi, dadliird 
(the only peraona with 1 quotable from RV. and AV.; and RY. baa dadhre 
twice); — roota in ^ *ppe" also to follow the later rale; aa oakffi, 
papFfe, vftTTf^, vavpoUie, bat dadtari^e and Jabhrl^, and in 3d 
pi, mid. botb oabTli4 and dadbrire-, — yhbA hu both babhritba 
(uanally) and babhtivitha, but only babbaTlm& (AV.). Bat there are 
found, againat the later mlea, su^uma, olcytiqe, Jutauri, and JataOrA, 
without i: the tnataneea are too few te fonnd a role upon. 

79B. The ending riti of 3d pi. mid. la found In RV. in aix forma: 
namely, olMtrlta, jafcbhrlr^, dadrire, bubhujrlrd, vlvldrlre, eaaiij- 
rlre; to which SV. adda dudultrire, and TB. dadrfrlre. 

800. Examples of inflection. By way of illustra- 
tion of the rulea ^ven 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 STI bndh &notc: its strong 
form of perfect-stem is <^4l^ bub6dh; weak form, ^^ 

babddha bubadhlvi •dhim& bnbadbe -dhlT&he -dtaim&tae 

ioy Google 

800—] X. PBSFB<7r4r8TKH. 288 

bal>6dbitlia -dh&tbnB -dbk bnbadhlf^ -dhJUia •dhidtavi 

babbdha -dbAtiu -dbAa babndb6 -dbit« -diitri 

b. Tbe UMrted i&iiet]' ol posilble aecsDt to 2d ling. »t. (iboTe 788 g) 
needs to be noted both Ln tfali Uld li) the lemiinlng ptndlgmi, 

0. As example of the noimal inflection of a root with 
final i or u-vowel, we may take the root ^ nl lead: its forms 
of stem are i^FRT nlndr or ft^ni nii^y, and ^^ atid. 

nin&ra, nlnitTa ninrlvi ninyimi ninyi nlnytT&lM uinrimihs 
nln^ths, nln^ltha nlnrAthns nlnyA nlnyl;^ nlnylttba ninyldlivi 

ninitya ninyitiui niity6s ninr^ ninyito nlnjiri 

d. Tlie root krl would mnke (129 aj in weak fonns oUcrl;lvi, 
olkrlr&toB. olkriyns, etc.; and yhbXt is inflected u follows in the 
active (middle forms not quotable): 

1 babhdva babbuvivi babhtlvlm& 

2 babhiitha, babhAvltha babhQTithoB babhaT4 

3 babhdva babhQ vitas babhilT^ 

Other roots Id a or n change this to nv before the initial vowel of 
an ending. 

e. As example of the inflection of a root ending in ^ 8, 
we may take ^ dS ffive: its forms of stem are ^ dadS and 
^ dad [or ^ dadi: see above, 784, 1]. 

^ ^ 



^^ ^it'% 

dadS& dadtri 



^^■^ ^^^ 



^ ^ 

daditha. dadlthi dad&thus 



d«diUu dsUcUTi 

^ ^^^ 



ws <© 

dadftvi daditus 



d»dtt« d«dir« 

f. The BY. ku on» papri t«i pq)Tt& (and •)«&* fttr JaUiP). 

ioy Google 

289 fixAJO'LBS OF Inflection. [ — 800 

1"^ As example of a loot with medial Q a showing fusioD 
of loot and leduplicatioD, resulting in medial ^ e, in the 
weak foims (794 g), we maj take FH tan stretch: it£ forme 
of stem aifl ^rR tatto oi RrTH tatKn, and ^^ten. 
> FTfPT. fTfTH ^ftej rri^tr ^ ^ft^^" ?rf-I*1% 

tat&na, tatiuia tenivk tenlm& ten6 tenlv&he tenlm&he 

tat&ntha, tenitU tan&tliiu ten& tenif^ tenitbe tenidlivi 

tatina ten&tna tenua ten6 tenSt« tenir6 

h. The root Jan, with the others which espel medial a id weak 
fonas (7640), makes Ja)&ntha or ji^filtha, Jajfiivi, JaJfLdBi Jajfli, 
JaJfUmihs, JajfUrd; and ho od. 

i. As example of a root with initial cf va contracted 
to 3 u in the reduplication, and contracted with the ledu- 
plication to ^ Q in weak forms (784J, we may take ^ vao 
speak: its forms of stem are 3^^uv&c or Jc(|tl uviU), and 
3^ tic, 

nv&ca, uT&oa noiTii Qolmi Oai aolT&he aolmUia 

uv&ktha, nv&oitha Qo&tlinB flo& aolqd uo^tha noidhT^ 

3 3STra ■3i^^^ '^^-, "^ >A^irt -A (^7 

UTROR ncatas uoub ao4 Qcite Qoird 

j. In like maDDer, yjai forms iy&ja oi lyaJa, iy^tha or iyUitha; 

I}Ai Ql^i and BO on; ^^nc has avoca and UT6cltIiR in the strong 

forms, and all the rest like vno. 

(t) Of the four roots in H r mentioned at 797 o, the 

inflection is as follows: 

oak&ra, o&kara oakfvA oak)rm4 eaikxi oak^&he oakrmAhe 

s g^ ^shy+1 ^isTT ^^ ^rara ^I^ 

oak&Ttha 0&kr4thuB oakrit oakf^ 6 cakiathe oakrdhv^ 

oak£ra oalcr&tuB oakrus caiire oakrate cakrir6 

Whitney, Gnmiur. 3. ed. ' ' IS 

D,j,i,...., Google 

800— J X. Pbkfbct-ststbm. 290 

1. 0£ the lootB ID n r in general, the first persons are 
made as foUows: 

dodliAra. dAdbAra <Udlinv4 dmdhHmA dwUirA dadhrlT&h* dadlirlmibe 
m> We miT fartheT add here, Badly, the iietiTe laflection (the middle 
i« not In nie) ot the peirect of as bi, irhleh (like babhltva and oaMbi^, 
glTsn above) is freqaently employed u m ■nxfUuy. 

1 ixA ftBiv4 Katmi 

2 ialthn U4tbuB ia& 

3 asa Bs&tos Uus 

801. A few miscellaneous iiiegulaiitiee calr still (or 

a. The root ah »peaA occare onlj in the perfect iDdicstive, and 
only in the 3d persone of all Dumbers and in the 2d alug. and dn., 
in active (and in 2d eing. the h is irregularly changed to t before 
the ending): thus, Kttha, jEha; Ahathue, fihatua; £h&B (in V., only 
£ha and ShAe are met with), 

b, t'tom yv& weave, the 3d pi. set. fivua OEonra in BV., and no 
other perfect farm appesra to have been met with In nse. It li allowed 
by Che gramiDariauB to be Inflected regularly as vS; and also kb way (the 
preient-Btem U w&ya : 761 f), with Gontracllon ofwa to n in weali torms ; 
and further. In the weak fome, as simple u. 

0. The root vyft envelop has in RV. the peifMt-tOTou vlwyatltils and 
tItit^, and no otberg have been met with In uae; the pammariins leqalre 
the strong fonna to be made from wyay, and the weak from tI 

d. The root i go forma In RV. and AV. the 2d alng. aet. iy4tha 
beiide the regalar ly6tlia; and beilde iiird from ^^, RT. baa ureral 
times erlrd. 

e. RV. hat an anomalous accent In d^d^qe and didpfre (beaide 
dadfkqd) and the pple d&dTf&na- And ciketa (once, beaide dkita) la 
paihapa a kindred anomaly. 

f. Peraana of the perfect from the ir-foima ot rootB in changeable f 
(S48) are tltlrua and tistire (both RV.); and tbey have cortoiponding 

g. The bastard root iir^u (Tt3) is said by the giammariana to make 
the perfeet-atem vLn^unn; the roots m^JJ and nog are said to Insert i 
nasal In the 2d sing, active, when the ending Is simple tha: tbna, ma- 
maaktha, nanaAftha (also mamajjitha and neqltlia). 

h. Farther may be noted Basajjatue (MBh.: yeafij, which has In 
passive the aeoondary form saJJ), rurundhatus (R), and dudubaa (BhP). 

1. The anomalans ajagrabbfii^aiii (AB. vl. 3b) seema a farmatian on 
the perfset-atem (but perhaps for ajlgrabbif^ail, deeld. ?). 

ioy Google 

Ferfeot Participle. 

B02. The ending of the active participle is ^fH tSAb 
(that is to say, in the ationg foims: it is contracted to 3? 
lif in the weakest, and replaced bj ef?J y&t in the middle 
foims: see above, 468 ff.). 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, 
roeohanically, the weakest participle-stem is identical vrith 
the 3d pi. active. Thus, ^tiUlH bubudhvSAs, fMHl^lfl 
ninlvi&B, g^jcftn cakrvR£kB. 

903. If the weak form of the perfect stem is monosyl- 
labic, the ending takes the union-vowel ^ i [which, however, 
disappears in the weakest cases): thus, FTf'i^n tenivSA,a, 
■JTRinIh noivSAs, nf^Rt? JajfiiT5]^, Qlf^eilH ftdivK^ (from 
y^ id: 788a), and so on; ^1^1^ dadiviA^ and its like, 
from roots in ^ &, are to be reckoned in the one class or 
the other according as we view the ^ i as weakened root- 
vowel or as union-vowel (794, 1). 

a. Bat piTtieiples of which the perfect-aUm Is mouoarlltbli; hy 4b- 
sence of the redapllcsdan do not tike the union-towel; thns, vidvaAs, 
Slid in v.. dft^lt&B (SV. (UqivaAs), nu^v^B, aShv^B, khld- 
v&Ab(?); uid R. bt» (Im dadv^fia (AV. dadivaAs and oace dad&v^AB) 
horn ydft (or dad: 872); sn &ii-a9vSAB Cv'aq eal) occnn in TS. and 
TB. B«t AV. baa viqlv^As and varjivll^ (la negative fem. ivarju^i). 

804. Other Vedic iTragularitiea calling for notice aie few. The long 
vowel of the radopIUatlan (786) appeirg In the pirtleiple ii In the iiidlcatiTe: 
thug, vavfdhTttAB, B&aatavlbbs, JQJuvlbbB. RV. and AT. have BasavifiB 
from i/san ot bS. RV. makes tbe ptiTtlclpUI foims of ytf or tar from 
different modlflcitions of the root: thns, tttirvAAa, but tatar&^BB. Re- 
specting tlie occasional exchange! of strong and weak stem in Inflection, 
see above, 482 o. 

SOB. a. From roots gam ftud ban the Veda makeg the strong stems 
jngaavaAs (aa to then, aeeSiSa) and jacbaavltAs ; the later language 
allows either these or the more regalar JagmlvKAB and JaghuivaAs (the 
weakest atem-fonns being ererywheie Jafmuf and jagbnuq). RV. baa 
also tatanvi&a. 


Digii.zec.y Google 

806—] X. Pbrfbot-btsteu. 292 

b. Fiom thiee looU, vid ,^nd, viq, ind dfq, the lalei Unguige tUoni 
ittODg puticipl«-iteia« to be mide vitli tbe nnlon-vowel, u veil u in the 
TcgBln manner wlthoat tt; thua, TiviglvSAB or viviQTSAs; dadrQlTtbha 
occara In KtbU. PB. bu once oloohidiTfiAB. 

80e. The ending of the middle participle is Kna. It 
is added to the weak form of perfect-stem, as this appears 
in the middle inflection: thus, ^3UH bubudhftnd, (+^JH 
ninyftnd, ^^ dadftnA, rHR tenSni, sT^H jajfiKni, ^'^H 

a. In the Veda, the long rednplioating vowel li shown by many middle 
putfoiplei : thus, TaT7db&n&, vftTaBftnA, dKdfh&fii, tiltnjftn&i etr. 
HT. baa 9aqa;&n& tiom yi;t (nlth irregular ffu^a, aa in the preient- 
Byatem: SSSJi UatiriL^ from yatr; and oace, with mSna, aaB|iDft9& 
from Ybj, a tew participles witb long Tedupl. vowel ba<e it trregululy 
accented (as !f lather Intenaive: 1013): thua, tlituJSiia (alio tatnJAnil, 
b^ibadb&na, q^adfina, QlSQi^&na, giifuv&ua. 

807. In the later language, the perfect participles have nearly gone 
out of use; eren the active appears but rarely, and is made from 
very few verbs, and of the middle hardly any examples are quotable, 
save such as the proper name ynyudhSna, the adjective anuoKna 
Uamed in icripture, etc. 

Uodea of the Perfect. 

808. Modes of the perfect belong only to the Vedic langasge, 
and even are seldom found outside of the Big- Veda 

a. To diaw the line sviely and distinctly between tbaae and the 
mode-forms from other lednplicated tense-stems — the present-ttem of the 
rednpllcating clus, the rednplloaled aorlst, and the intensive — is not pos- 
sible, since no criterion of form exists which does not in gome cases faii, and 
sluce the geueial eqnlralence of modal forma &om ail stems (582), and the 
nommon use of tbf> perfect u a present In the Veda (823), deprive na of 
a criterion of meaning. There can be no reasonable doubt, however, that 
n conaiderable body of forms are to be reckoned here; optatives like &aa- 
QyBm and babhOyOs and babhilyat, Imperatives lilie babhQta, sQbJanc- 
tives like JabhArat, show such dUtinctlie 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: &om a reduplicated perfect-stem, as (for example) 
mnmuo, an imperative would be made by simply appending, as 
usual, the imperative endings; the derived enbjunctive mode-stem 
would be mum6oa (accented after the analogy of the strong forms 

ioy Google 

293 Modes, [—818 

of the perfeot lodioative), uid would take either primary or secondary 
endiugB; and the optative mode-atems would be mTunuoya in the 
active, and miunu<d (accent on personal endiogs) in the middle- 

And the great majority of the forms io question (about three 
qnartera) are made in these ways. Thus: 

810. Examples of the regular subjunctive formation are: 

a, with secondary endings, actUe: 2d ling., papT&thas, oftk&nas, 
m&inUuui, pipr&yae, bnbodhas, rftr&^BS; 3d ilDg.,oftk&nHt, jnbh&rat, 
r&r&^at, B&s&hat, paep&ifat, pipr&yat; lit pi., cftk&nSma, tat&QSma, 
9ng&viina; 3d pi., tat4naii, papr&thnii (other penouB do not occdt). 
This ii the Urgett clus o[ cue», 

b. witb primary endinga, acdie: here aeem to belong only dftdhAr- 
fatl and vav&rtatl: coiopaie the formation with different accent below, 
81) a. 

o. of middle forms obcdi only the 8d sing. taUtpate, faQ^otate, 
Tuyojate. Juj6^te (SV. ; BV. haa JiUo^ate); and the 3ii pi. cftUtnanUt 
tat&nanta (and psihapg two or three othera: below, 811 b, end). 

811> Bot not a few anbjnnctiies of other formation occur; tbaa: 

a. With Btrengthened root-syllable, as abOTe, bnt with accent on the 
leduplieation (as In tbe majority of pTMenl-rorms of the reduplicating clua: 
above, 646]. Here the forms with primary endings, andTS, preponderate, 
and are not very rare: for example, jujofasl, Jojofatl, J^ofathas, 
Jnjo^atha (other persona do not occur). With secondary endings, J^oqas, 
jAjo^at, and JuJOf an are the form* that belong moat dlstinrtly bnie (since 
dAdJ^as and afif&das etc. are perhapa rather torlsta]. And there Is no 
middle form but Jnjo^te (RV.: see above, 810 o). 

b. With unstrengtbened root-syllable occur a small body ef forma, 
which are apparently also accented on the redaplicatton (accented examples 
are found only in 3d pi. mid.); thua, active, for example, mnmuoaBi 
wavptat, vlvidat, ^uqavat; the only middle forms are dadhf^ate, 
viTpdhata, 3d sing.; and c&kramants, didh^^anta, ruruoanta (with 
dadabhanta, paprathanta, mftmahanta, Juburanta, which might also 
belong elsewhere; BlOe). 

0. Accented on the ending are vKvpdb&nta and cakpp&nta (whloh 
are rather to be called aDgmenllets plupetfecta). 

d. As to forms with double mode-sign, or transfers to an a- conjugation, 
aee below, 81ft. 

819. Examples of the regular optative formation are: 

a. In active: lat sing., ftnaqTam, Jagamy&m, pappcyftm, rirlO' 
ySm; 2d slag., vavptySe, viviqyaa, fUfrOy^, babhOyfta; 3d slug,, 
Jogamyftt, vavptyftt, tatiiJyAt, babboyit; 2d du., jafpnyltam, qn^rQ- 
^tam; lat pi., eBsBhy&ma, waTrtyftout, faqay&mai 3d pi., tatanyua. 
vaTpjyufl, vavptyuB. The forms are qi 

ioy Google 

81S — ] X. PsRFEcrr-STSTEH. 294 

b. In middle, the loima ue few: namtlr, lit sing., Tayy tl yOi 2d 
•Inf., Tftvrdblth^, oakfunithSs; 3d aJng., Jogrufitft, TavTtit&. 
m&mrJita, dtidhuTita, qngnoltai l(t pi., TavfQnwtai. And sSsa- 
h'fthB* uid rirlqifta ippear to fainiih exunplM of pT«ektiTe optattia 

o. There 1> no IrregaUi mode of fotmaCloii of perfect optidTec. Indl- 
Tldual irregnluitieB ice aho^n by eertilu tbrme: thui, oalo^lyis, p&plyftt, 
qnfruyaa »nd fu^ruy^taia, witli tteitmeDt of the flnd u before the 
paeBire-ilgn 74 (770) ^ onaJySt with Ehort initiil; fl^rlti from yifri; 
JftkqiySt 1« tnomalous: rirl^B ia the o&ly form that ebowt a tinlon-Towel 
a (nnlesa also sifst, from y»B,). 

818. or regnlir InpentlTe forms, only > very im).!! number are to 
be quoted: namely, acCive, cfikandhl, rArandhi, oiklddhl, titigdlii, 
mumogdlil, gaquKdbl, and plpnhi ; o&kanto, rfirautu, momoktu. 
and babbatu; mnintiktam and vBTfktajii; jnjuftaaa and vBTTttana 
(nnleis we are to add mamaddbi, mamatta, mam&ttaiia) ; — midrtle, 
vaTTteva and TavrddhTom. AT. has onca dadrfrtm. 

814. At Irregular ImperatlTes may be reohooed eeveial which ahow 
a union-TOwel a, or have been trantferted to an a-oonjagitton. Such are. 
In the active, mumAoatam and Ji^o^tam (2d dn.), and miiia6oata 
(3d pi.); In the middle, pipr&yaBva (only one foand with accent), and 
■aftmobaavH, vSTrd^asTa, Tfivrfasira (2d alng.), uid mftmabantftm 
(Sd pi.: probably to be accented -Asva and -4at&m). 

816. Snob imperatlTei as these, taken In oonneRtlOD with some of 
the snbJnnctlTes given aboTe (and a few of the "pluperfect" forms: below, 
8tt}), SDggeat as planaible the sisumptlon of a double preaeot-aten, with 
redDpUeatlon and added a (with which the deaideiatlva steme wenld ba 
comparable: below, 1020 ff.): for example, JvOoQB ^om KJu;, from which 
wonld come J^o^aal etc. and Jiijofate [811a) as indlaative, J^o^a* 
etc. as subjuactiTely used angmentleas Impeifect, and Ji^o^tant as im- 
peiatiie. Host of the forma given above as anbjanotivea with primary 
ending lack a marked and constant snbjanatlve eharactet, and woald pass 
falrl; well aa Indicatives. And it appears tolerably certain thst from one 
root at least, vplb, such a doable stem la to be recognized; from vAvrdliA 
eonie reidll)' T&TTdhate, v&vrdh&ata, and from it alone can come regu- 
larly v&vrdhaeva, vftTrdhdte and vfivrdli&ti (once, RT.J — and, y^t 
more, the participle TaTrdbint (RV. j AT. vSvfdb&nt: an Isolated case): 
yet even heia we have also vSTrdbith^B, not vUvrdbMhfia. To utsme 
donble piesent-stems, however, in all the cases would be highly implau- 
sible; It la better to reeognlie the formation as one begun, but not car- 
ried ont. 

a. Only one other aabjnnctlve with donble mode-sign — namtdy, 
PftPToSai — la found to set baslde vtwrdbftti. 

816. Forms of dlBerent model are not very seldom made from the 
same toot: for example, from ymno, the subjnnotivea mum6oH) i 

Uiail.zecy Google 


oati, >Dd mumuoaSi fcom ydhn^ dadli&rfati and dadhnate; fnm 
ypii, tbe tmpcutlTei piprDif *Dd plpr&TUva. 


817. Of KD AogmeDt-preterit from the perfeot-Btem, to whioh the 
name of pluperfect is giren on the ground of its fotmatioD (thongb 
not of Its meaoing), tbe Veda preaents a few examples; and one or 
two fonuB of the later language (mentioned above, 768 b) have alao 
been referred to it 

a. Then U mach of tbe nme dlfDculty In dlitingalihlng the ploptifecl 
M the perfect model fioca kindred redupUosted foimatlona. Betveen It and 
the urbl, howerer, a ditference of meaning helpi to make > seputilon. 

818. The normd plupeTfdct ihonld ihow a itroni item in the sincniar 
■ctlTe, and a weak one eliewheie — thui, mnmoo and mumao — with 
angment preflied and Becondary endings added (us in 3d pi. act,, ata in 
3d pi. mid.]- 

a* Of forms jnado kccordlog to thla model, «e have, In the aotlvs: 
iat ling., ^agrabbam and aoaoakftun (vtbicb, b; ita form, might be 
■oilit: 880): 3d ling. ^agan; 3d ling., Rjagau and aolkst-, li do.. 
».mi-niTiir»jm ■ 2d pi. i^agftiita, and i^jagantana and tOabltaitana (■ 
atrong fonn, ai often in thii penou: Efi6a); 3d pL (perhapf), ama> 
mandiia uid amamadno. To tbeie may be added tbe augmeDtleaa cSk&n 
and rSr&Q, oibitam and oakaxam. in the middle, the 3d pi. aoakrlrau 
and ^lagmlran |_wlth Iran Inatead of ata), and the aagmentlesi 2d sing. 
JngOrthAa and Bti;upUila, are the moit legnlu fonni to be fonnd. 

818. Sevenl foimi from rooti ending In oouMUiDti me the endings 
In 2d and 3d ling. act. by laierting an i (566 b): Uids, 4bxibbOj&l. 
avlTa^lB; artreoit, Ajagrabhlt (avftvarit and aTava9ltftm aie rather 
iotentlTM); and the angmentleai jibl&flia (accent?) aud dadliar^t belong 
with them. 

&SO. A few forms show a item ending in a: they a-e, la the active: 
3d ling., asasvaJat, aolUtat, aoakrat; In the middle: 3d sing., Aplp- 
lata; 3d do., ApaaprdhettlKlD; 3d pi., atltvlffaata (which by its form 
■i^t be aeiiit), Adadrhaot^v; and oatoadat, oak^p^ta, TtTTdh&nta, 
jnlmraiita, wadM perhapi be beet elaiilfled here as augmentleia forms 
(eompaiB 811, abors). 

UtM of tfie Perfoct 

881. Perfects are quotable as made from more than half tbe 
toota of tbe language, and tbey abound in ase at every period and 
in almost all bianehes of the literature, thongb not always with the 
same valne. 

a. Accoiding to the Hlnda gnmmailan*, the perfect li naed In the 

ioy Google 

Stl— j X. PnracT-ttTVrKM. 2% 

••fnii'iB •( latu mM wixmtmti tj the aamtx; tat Ikac u no eiUtnra 
•( iti Will! <^tb^ esdviinly «r ditttoctnclr *• cafUfed it lay pnlod. 

b. la the later bnfnafe. it ia h^It a preterit or pest tenw, 
•q«iT«lent with the imperfeet, aad freel; teterekSMgtAble or coJh- 
dinated with iL It U od the whole Icm tommom thaa the inperfeet. 
altbongh the prefereatea of differeat aatbora are diTerse, aad it soBe- 
tlaea exceedj the iBperfeet in (reqacaej compare 9S7 . 

e. The perfects rada and iha are everjrwhere tued with preseot 
value. In the Brahmanar. also otbere, eapeeially dAdhira, also dtdlya, 
Mbhftja, etc. 

82S. Id ibe Brikmkiiu. the dlitioctioD of teate-Tsloe between per- 
fect and Imperfect ii almoft tttog^thcr loft. u ia the litei languce. Bat 
In molt of the text* the impeifeel U the ardinuy teaie of nuntion, the 
perfect bdn; only eieepHomny nud. Tkoi Id PB., the imperfects ue to 
tbe perr*rts u more than i handred lo onai in the BAhmani puts of TS. 
■ud TB., u oTci thlrtj-font to oae: and in thoM et MS. In aboot the 
lame proportion ; in AB., ■• more than foai to «n«, the perfect appearing 
moitlr in certain pawagea, where it take* the place of imperfect It ia 
only In (^B. that the perfaet It mneh more eommonlj nsed, and eTon, to 
■ coniili-riblc extent. In roordlnatlon with the imperfect. Tbiongbout tbe 
BAbmanii, bowcTet, the perfect pirtldplea Iuto in general the true *^r- 
fed" Talae, iodieatlDe a completed or proximate puL 

823. In the Teda, the eaae la very different. Tbe perfect I* Died 
as paat tenae in narration, bat oDly rarely; sometlmfa alM it hu a tme 
"perfect" lenae, or atgnlflea a completed oi proximate paat (like the aoilet 
of the older language; 938); but ofteneat it baa a Talne hardl; or not 
at all dlaUDgniahable in point of time from tbe present. It ia thna the 
equivalent of Imperfect, aorist, and present; and it oceai* coordinated with 
them all. 

a. Examples are : of perfect with present, n& ^rSmyantl nil vf 
mufioanty it* T&yo tJt paptnh (RV.) iha/ uieary nol nor tlop, theyfiif 
like bird*; ni 'd n rtEjS k^aratl car^ai^inMO arin ak nami^ p&ri 
tit babhfiva (ItV.) ht in truth rule* king of men: he tmbraet* Uitm all, 
a* the wheel the epokee; — of perfect with aariat, dpo runioe yuvatlr 
I1& jb^ . . . ibnad agii£l;i aajnidhe mllniiq&^ftm &kar ]y6tlr bidh- 
amKnB t&mHAsi (BV.) *he u come beaming like a young mtudtn; Agni 
hath appeared for the kindling of mortal* ; *he hath made light, driving aieay 
Ike darkne**; — of perfect with imperfect, Uumn Uilm kav ttpka tatarda 
(BV.) he *le%c the dragon, he penetrated to the u>ater*. Such a coordination 
u this lut Is of conalant ooonrrence in the later langnige: e, g. mnmude 
'pfUayao oU 'iibn(B.) he tea* glad, and paid honor to her; waatrftnte 
Jagraha akaadhadefft 'arJat toajra srajam (MBh.] the took hold of 
the end of hi* garment, and dropped a garland on 6it ihoulder*. 

ioy Google 

Varieties of Aorist. 



624. Under the name of aoiiat aie iacluded (as was 
pointed out above, 532) three quite distinct fotmations, eaoh 
of whioh has its sub-Tarieties: namely — 

I. A SIMPLE AORIST (equivalent to the Greek "seoond 
aorist'^, analogous in all respects as to foim and inflection 
with the impeifeot. It has two varieties: 1. the loot-aotist, 
with a tense-stem identical with the root (coiiesponding 
to an imperfect of the loot-class); 2. the a-aorist, with a 
tense-stem ending in Q &, oi with union-vowel Q a before 
the endings (coiresponding to an imperfect of the ^lass), 

n. 3. A REDuFLicATiNo AOBI9T, perhaps in origin iden- 
tical with an imperfect of the reduplicating class, but having 
come to be separated from it by marked peculiarities of form. 
It usually has a union-vowel Q a before the endings, or is 
inflected like an imperfect of one of the a-claases; but a 
few forms occur in the Veda without such vowel. 

III. A 8IOUATK' or SIBILANT AORIST (corresponding to the 
Greek "first norist"), having for its tense-sign a n b added 
to the root, either directly or with a preceding auxiliary 
^ i; its endings are usually added immediately to the tense- 
sign, but in a small number of roots with a union-vowel 
Q a; a very few roots also are increased by H b for its 
formation; and according to these difietenoes it falls into 
four varieties: namely, A. without union-vowel 9 a before 
endings: 4. a-aorist, with T\ s alone added to the root; 
5. i«-aoiist, the same with interposed ^ i; 5. sif-aotist, 
the same as the preceding with ^ s added at the end of 
the root; B. with union-vowel ^ a, 7. sa-aorist. 

ioy Google 

SaS— ] XL A0R18T-8YSTBM8, 298 

826. All these vaiieties aie bound together ajid made 
into a single complex system by ceitain coriespondences of 
form and meaning. Thus, in regard to form, they are all 
alike, in the indicatire, augment-preterite to which there does 
not exist any corresponding present; in regard to meaoiag, 
although in the later or classical language they are simply 
preterits, exchangeable with imperfects and perfects, tfaey 
all alike have in the older language the general value of 
a completed past or "perfect", translatable by /lave done and 
the like. 

8S6. The w)ri«t-i;8Mm la b fOTmaiion of iufieqaent occotTenas in 
ntDcb of the dwalul S&iukiit (iu forma mo foand, for eiunple, only 
twenty-one tlmeB in the Nils, eight in the Hllopadofk, geren in Uanu, tU 
Mch Id the Bfaagkvtd'Oita and ^ikantsla, and iizty-eli times, from tour' 
laen roots, in the Ant book, of about 2600 linn, of the Samaya^a: com- 
pare 037 b), and It poaaeises do psiticiple, nor any model (^excepting ia 
the probibidTe nae of ite aiiKmentleaa forma: aee S79; and the Bo-c>11«i 
piecatiie: see 921 ff.); in the older language, on the other hand, tt I* 
quite common, and has the «boIe variety of modes belonging to the present, 
and sometlmea partioiplei. Ila desoription, tcooTdingly, must be glTen 
mainly as that of a pari of the oMet language, with dae notice of Its res- 
triction in later nae. 

827. a. In the BV., neuty half the looia occurTing ebov aorUt form*, 
of one or aneiber claasj In the AV., rather lesa than one third; and In the 
other texts of the older lingnage comparatively few aorlsle ocenr irhlch an 
not found in these two. 

b. More than llfty roots, in BV. aud AV. lagether, make aorlM forns 
of more than one class (not taking into account the reduplicated or "canaa- 
tive" aorlst); bat no law appears to underlie this variety; of any relation 
■neh as la taagbt by the granmarlsns, between active ef one daaa and 
middle of another as correlative, there la no trace discoverable. 

o. Examples are: of claetes 1 and 4, adh&m and dhBans from 
|/clta&, aynjl and ayukfata from yyMi; — of 1 and 6, BBrabham and 
■grabUfma from ygrabh, mnU^AB and oiBrQifthSfl from v'mn! — 
of 1 and 2, ftrta and irat from y'f; — of 3 and i, ntrldam and avltsi 
from y-rid find, anljam and an&ikflt from yolj ; — of 2 and &, Ba&6- 
ma and aaAnl^am from ysau; — of % md T, arubam and amkfat 
bom yrtili; — Of 4 and 6, amataus and amsdi^oa from t^mad; — 
of 1 and 6, bftsmahi and bfisl^us from f^bK; — of 1 and % and 1, 
atnata and atanat and at&n from y'tan; — of 1 and 4 and fi, abudb- 
nm and Abbatai and b6dhi^t from j/bndb, &atar and s^^lTa and 

itizecy Google 

290 1' ROOT-AOKIST. [—881 

aotkrls bom yvtf. Otteo the second, ot teeaiid and thltd, ilasa ii rep- 
Tuentad b; onl; >n IsoUted ronn ot two. 

1. Simple Aorist. 

8S8. This is, of the three principal diTisions of aorist, the one 
leut removed fiom the analog of forms already explalQed; It U 
like an imperfect, of the root-olaM or of the A-claas, without a coireB- 
poudiog preBOnt indicative, but with (more or leas fragmentaiH;) all 
the other parte which go to make up a complete present-ayBtem. 

1, Root-aorlst. 

829. a. ThiB foimation is in the Idter language liuited 
to a few loots in ^ & and the loot ^ bbO, and is allowed 
to be made in the active only, the middle using instead I 
t^e s-aorist (4), or the if-aoiist (5). - — \ 

b. The roots in QT ft take 3TT na aa 3d pi. ending, and, 
as usual, lose their W S before it; ^ bhQ (as in the perfect: 
788 a) retains its vowel unchanged throughout, inserting 
? Y after it before the endings W\ am and ^ an of 1st 
sing, and 3d pi. Thus: 
















For the classical Sanskrit, this is the whole story. 

880. Is the Veda, these same roots are decidedly the most fre- 
quent and oonspicDOQB representatives of the formation: especiaUf 
the roots k^ 6a, 6iilL, pS drink, sths, bbQ; while sporadic forms 
are made from J&t. prt, sS, bft. As to their middle forms, see 
below, 884 a. 

a. inatetd ot abbuvam, KV. h»9 twice abbnvam. BhP. has agan, 
3d pi., Initcad of affua. 

881. But aorlsts of the same class are also made from a nnm- 
ber of roots in r, and a few In i- and n-vowels (short or Ion;) — 

itizecy Google 

881—] XI. AoRiBT-srsTEMS. 300 

with, M required bj the ADftlogj of the tense with an Imperfect of 
the root-clftBS, gu^a-BtrengtheDiDg in the three persona of the Biognlar. 
a. Thus (In the nctiTe), bom Vqra, &fr«Taiii >Dd &<}Toti from y^ri, 
Agree >nd &gret; from y^ make, iikaram led &kar (for akara and 
akart]; fiaai vj enelote, avar (686a); and so fcatac, aepar. D nil and 
plural foima are mach leas i^qDent than ainfulir; but for the moat part 
they aleo abov an irregalar atrengthenlng or the root-Tovel : thus (inclndlne 
aagmeaUeiB farmi), &karma and karma and ikarta, T&rtBm, HparCami, 
&hama and Utatana, bhemo, agravan; legnlar are only avran, Ucran, 
ah7aii> and igrlyan. 

832. Further, from a few roots with medial [or initial) vowel 
capable of icn^a-strengthening and having in general that strengthen- 
ing only in the Blngular. 

a. Tbni, &bhedam and abbet from f/bhld; Amok from y/muo; 
yojam from yyaj ; rok (VS.) tnm ymj ; arodham and aradtanut from 
yrudb; avart from yvft; v&rk fmm VvfJ (AY. hai once aVTk); adar- 
gam from ydp}; irdhma from y^dh; and adrfan, aTTJan, agvibuk. 
Itut chedma, with gui^a, from yohld, and adarqma (T3.) from ydjq. 

833. Again, from a larger number of roots with a as radical 

a. Of ikwe, gam (with n form when flnal or followed by m: 14Sai, 
SIS a) ii of decidedly most frequent occurrenee, and ehowa the greatAd 
variety of forms: thug, Agamam, &gaii (2d and 3d alng.), Aganma, 
aganta (itrong form), frgman. The otber caaea are akrau from ykram; 
itan from ytKa\ abhr&t from )/bbT&)i askan from >'BkaiLd; asrat 
l^om i/arafia (P VS.); dhak and daghma from ydagh; £iia( (G8Ba) 
and ana^fSm from ynat} ; iighaa or aghat, Aghaetftm, aghasta, and 
&k;aii (tor aghean, like agmaii) from VKkas; and the 3d pi. in as, 
Akramae, ayamue, dabhua, nftna (pf.?). mandua. 

634. So far only active forms have been oouaiderad. In the 
middle, a considerable part of the forms are such as are held by the 
grammariane (B81) to belong to the a-aoriat, with omiaalon of the a: 
thej donbtleaa belong, however, mostly or altogether, here. Thna: 

a. From loott ending in vowela, we have adbitbfis, adhlta (aleo 
ahita), and adhlmahi; adlttuta, adita, and adlioahl (and adimaU 
from yd& cuf) ; &qtta(7); sim4hi; iathithSB and Aathita and AathiraD, 
form) of K-Toota; — of ^-loots, akri, Ultbfia, Akpta, akrStfim, ikrata 
(aTid the anomalous kr&nta); avri, av^thSs, avpta; Arta, firata; mrth&a, 
amrta; dhrtbas; adfthU; aetata; ah^thaa; g&rta;~of i and u roota, 
the only exaoiples are ahvl (? AY,, onc^-O, 4bumahi, and 4aldbvam. 
The ibaeaoe of any laalogies whatever for the omiiBlOQ of a S In aueb 
forms, and Ihe oFcuirence of avri and akri and 4krata, shoir that their 
reference to the s-Bariat U probably without sufflciont reiBon. 

b. As regards lOOts ending iu conaonante, the case is more qnestian- 
able, since Iota of a after a flnil canaonant before thSa and ta (and, of 

ioy Google 

301 '- HOOT-AOBIST. [ — 836 

conne, dhvam) would bs in msriF euet required by euphonic rale (9S3 o fT,). 
We And, howeTer, inch nnmlBtikable middle ioDectlon of tbe Toot-ioriat ui 
ayaJU iynkthfts, itynkta, aTnJmalil, &yugdhvam, Ayuiraii; i^ 
uid ^ata; niAfl; apadi (1st sing.) and apadmahl and apadran; 
Amamnahl; g&nvahi and &ganmalii and igmatn; atnata; AJani 
(111 ling.) and ^fiata (3d pi.) ; from ygasa are mads agathBs and agata, 
(torn >'taii, atath&e and 4tata, and from yman, amata, with ttestinent 
of the Bnal like that ot ban in present inllection (637). The ending ran 
It eapeolally fteqneni In 3d pi., being taken b; a nnnibei ot vecbe vbinb 
hare no other middle person of this aorlst: thus, agrbhran, kefgrBXi, 
adr^raD, abudtaran, &v|tran, t^fu^ran, ak^ran, asprdhxan, avas- 
ran, &Ti9raii; and ram ii found beside ran in ^dp^ram, ibudbram, 

o. FiDm roote of which the Unal vould corablne with a to k;, It 
seems more probable that aorist-Forms abawlng k (Instead at ^) before the 
ending beloog to the root-aorist: such aie amnkthfia (and itmugdhTam), 
BpTkthSs and aprkta, ibhakta, avrkta, aaakttaOa and asakta, rlk- 
tbfia, vikthOa and vikta, arnkta; apraffa, ayaffa, Aapa^tBi asntl><^ 
and &Kn%m, and mpfthfts would be the same In either case. 

d. There remain, as cases of more doabtfnl belonging, and probably 
to be ranked in pan with the one formation and In part with the other, 
■ceordlng to their period and to tbe occurrence of ether perEons: OhlttfaAs, 
sattluU and inntta and inuddhvam, patth&s, bhltth&a, amatta, 
atapthSa, alipta, aa^pta; and dually, &mbdba, alabdba, aruddba, 
abnddba, ayuddba, and drogdh&e (MBh. : read drngdhia): see 88S. 

Hodea of the Boot^aorist. 

83B. Subjunctive. In «ubjanctlTe use. forms Identioal with the 
augmentleea indlcallYe of this aorist are much more treqnent tlian the more 
propel BubjanctlTei, Those to which no corresponding form witb augment 
occurs hsTe been given above; the others II is unnecessary to report in 

836. a. or Crae snhjunotivea the forms with primary endings are 
quite few. In the acllvp, k&rft^i, gSni, gambii (for bbuT&ni, eee be- 
low, o); k4raai; athfttl, dati and dbati (wblch are almost indicative In 
Talne), karatl, Jofatl, padftti, bbMati, rftdliati, varjati; ath^tbae, 
karathas and karatae, darqathas, qravathas and ^r&vatas; and 
(apparently) karasti, gAmaatl. In the middle, Jo^aoe; idbat^ (f), 
kArate, bhojate, yojate, T&ijate; dbdtbe and dh&ithe; kirfimabe, 
db&mabe. g&m&mabfti. 

b. Forms with secondary endings are, in tbe active, d^rqam, bho- 
jatn, yojam; k&ras, t&rdas, piiraaB, yamas, rfidbfts, viraa; k&rat, 
gamat, garat, jo^at, daghat, padst, y&mat, yodhat, rftdhat, varat, 
virtat, qrdvat, B&ghat, ap&rat; kitr&ma.gamSma.r&dhmnai g&maD, 

D,j,i,...., Google 

886—] XI. Aoiu»T-8TSTBH8. 302 

gafftn, iktijaSL, 7Uaan< No lulddle foimi u« clMilfltbla with confldence 

c. The series bbuvam, bhi^Tas, bhi^vat, bbfrvaji, &nd bbartni 
(comptte ftbhoTam; 830 a), >nd the isolated cravat, an at donbtml 
belonglnfi; nlth a different icceot, they vould >eem to be of the nan 
elat*; hoTe, a gtU^-itrengthening woold be more legular (but note the 
abaence of gu^a io the toilit IndlcatiTe and the perfect of j/bbn). 

837. Optative. The apUtiTs active of this aorirt conetttutea, with 
a i Interposed betweeo mode-ilgn and pereoDa) endtDga (567), the pnea- 
tive active of the Hindu ^ammsrlana, and la allowed b; them to be made 
from evefy vatb, they Tecognlilng no eonneetlon between It and the aotUt. 
Rut In the 2d sing, the Interposed b Is not dlitingnlshable from the person^ 
ending) and, after the earlleBt period [see 838), the endlog crowde out the 
sibilant In the 3d sing., which thns comes to end la y&t Instead of yta 
(compare 665 a). 

a. In the older tangnage, howerer, pare optative forms, wlthont the a, 
are made ftDm this tense. From mots In ft oeeur (with change of ft to e 
before the y^ 860 d) deyftm, dbeyftm and dbeynB, and Btbeyftma; 
111 u-vavele, bhQyima; In f, kriyftma; In consonants, aqyim and 
aqy&na aiid a$yua, TTJyftm, qakyim, ynJySva and yi^jyitftm, aftbyt^ 
ma, and tpdyuB. 

b. The optative middle of the rooi-aarljt Is not recognized by the 
Hindu grammarians as making a part of the precatlve formation. The RV. 
has, however, two preoailve forms of 11, namely padlffi and mncl^. 
Hnch mote comiDon In the older langaage are pare optative forms: namely, 
eifljk and aflni&bl (this optative is especially common), indbiya, gmlya, 
intiTiya, ruolya; arlta, ubita, vurlta; idtalmahl, naqimabi, nad- 
mabl, proimahl, mudlmabi, yamlmahi; and probably, from a-root«. 
suniihl and dMmahi (which might also be angmentless Indlcstivs, slDce 
adblmabl and adhitftm also oeonr). All these forms eicept the thiee 
In 3d ling, might be preoative according to the general undentandlog of 
thai mode, as being of persona which even by the native authorities are not 
olalmed ever to eihlblt the inserted nibilant. 

8S8. Prerattve active forms of this aorlst are made from the eariiest 
period of the language, in ST., they do not occur from any root whleh 
has not also other aorist forms of the same class to show. The RV, forma 
are^ lat alng., bhOyfiaam; 1i sing., avyfia, Jfieyfts, btaflyaa, mfdbyia, 
aabyfts; 3d ting. (In -yfts, for-yftet; RV. has no 3d slug. In jlX, whieti 
if later the unlveraal ending), avy&B, agyAs, f dbyfts, gamyia, daghyfts, 
peyfia, bhOy&s, yamySa, yiiyiks, vrJyBa. ^riiy&B, sabyfts; 1st pi., 
kriyftama [beside krlyfima: 837 a), AV, has six 1st persons slag. Id 
-ylsam, one 2d in -y£B, one 3d In .yftt (and one in .yfts. In a RV. 
pastago), throe 1st pi. in -j&natt (beside one in yftma, In ■ KV. passage), 
and the 2d bhayftstba (doublteas a falae reading; TB. has •Bta in the 
corresponding passage). From thia time on, the pnr* optative A>nBi uenlr 

ioy Google 

303 1' RoOT-AOEUBT. [—640 

dlwppew (the eieaptlDDB are ilren iu 687 *). But the pieoiitlT* farms 4rB 
nowhare eommou, ezecpdns *a mtde from >'1)liQ; ind rrom no other root 
li »ajth<ng Itka & complete leriea of peiioDB quottble (only bhilyftava 
aad bhOrSstftm 1>etng wiDtlog; *ni) theee two panons have no repreaeiit- 
itWe from any mot). All together, actiTO optatlTe oi pTecatfTO formi ^re 
mkdo in the older Ungntgo from over fifty roota; and the epic and claealcal 
tKita add them tiom hardly a dosen more: aee fnrlher 9S6. 

889. Imperative. Imperative forms of the root-aorist are not rare 
in the eaily langaage. In the middle, Indeed, almost only the 2d sing. 
oeeuTs: It U accented either regularly, on the ending, ae k^vA, dhlfri, 
jmkfvi, or on the root, as mUsvn, y&kfva, vifisra. r&va, B&kfva; 
dlfva and mKoVB are not found with accent; the 2d pi. is repraaenled 
by k^dhTKiDi vo^vun. In the acllTe, all the peraona [ii and 3d) are 
ftiiind in use; eiamplea are: 2d sing., k^dbf, T^^dM, qagdhl, frudhl, 
gftdhl, yadidlif, galil, mfthl, B&hi, mogdhi; 3d sing., gaditu, dStu, 
a^fn, 9r6tu, sAta; Id du., d&tam, Jitam, qoktain, ^rut&m, bhOtAm, 
apftkm, gat&m, rikt&m, vo^bam, sitam, sut&m; 3d dn., only gaih- 
t&ni, d&tftm, vo^bim; 2d pi., K&t&, bbat&, Qruta, kfta, gata, dsta, 
dbStana; 3d pi., only dbftnttt, gruvantu. These are the most legulic 
forma; but irregnlaritieg aa to both accent and strengthening are not infre- 
quent. Tbni, strong forma In 2d du. and pi. are yadit&in, varktam, 
vaitam ; kirta, g&iiita [once gadit&), y&ibta, vartta, beta, grdta, B6ta ; 
and, with tamti k&rtaQa, giafatana, yaditana. ootana, and the irregular 
dbetana Cv'dbK); In Sd do,, gKifatftm. Much more iirepilar are y6dbi 
(Instead of ynddbl) from f'jrudb, and bodbi from both )/budb and |/l>bil 
(inatead of buddl^ and bbndbf). A single form (3d stng.) in Ut is 
fonnd, namely gaat&t. We find krdbl also later (MRh. BhF.). 

a. Aa to 2d penons aingulat in ai from the simple root used in an 
imperative sense, aee above, BS4. 

PartiolplflB of the Boot-aortot. 

640. Id the oldeet language, of the RV., are found a number of 
paitioiplea wbleb moat be reckoned ae belonging to tbiB fonnadon. 

a. In the active, tbay are eittemely few: namely, kr&nt, dt&nt (f), 
Ciu4nt, Bthint, bbidtuit, v^dbint, dyutant- (only in composition), 
and probably ^dbint. And BhP. has nLffant (but probably by error, for 

b. In the middle, they are in RV. much more numerous. The accent 
is uinally on the final of the at«m: thus, Brll9&, tdbSni, krSni, jtifb^ 
tnt^ii, aldand, pigftnA, p)-<!&°^> pratbSa&, budbSni, bbiy&ii&, 
DumBni, mandftni, ynJUni, ruosni, Tip&ni, vtBjfk, urS^, fubb- 
ftnik, Baa&ii&, Buv&n4 or BTSnii, B^jBuA, BprdbilD&, blyBnii; — but 
sometimes on the root-syllable; thus, oftSna, oyavftaa, rub&na, ilhSna 
(prea.?), T&sina, gumbhina; — Tibile a few show both aocentuatioua 

Dgilize. ..Google 

840—1 ^I' AORIST-eTBTBHS. 304 

[compu« 018 d)r thna, (lT9aii& and dffftua, d7Tit&ii& ■nd dr^tfinai, 
yatftnA >iid yitftna ; and oetbia ind hrnyS^a ocenr only In eomposltloD. 
A jety few of tliei« >re found once oi twice in othei tatU, nimely oltiiM, 
dyntftna, mhS^A, Taa&aa, snTftna; and -kup&na oceius once tn 
Aput. (liv. 18. 4). 

841. All toother, the roots exhibiting in the older langnage 
forms which ue with fail probability to be reckoned to Hie root- 
iMiia [-system are abont a hundred and thirty; over eighty of then 
make such forms in the RV. 

Fafisive Aorist third peraon singular. 

843. A middle third pereon singular, of pecnliar formation and 
prevailingly passive meaning, Is made from many verbs in the older 
Uoguage, and has become a regular part of the passive conjagatioo, 
being, according to the grammarians, to be substituted always for the 
proper third person of any aorist middle that is need in a passive 

848. This person is formed by adding ^ i to the toot, 
which takes also the augment, and is usually strengthened. 

a. Th« ending 1 belongs eliewheie only to tba Qnt person; and thla 
third petBon appiienily standi in tba game reUtion to ■ Brat In i u do, 
In tba middle toIcg, the regular 3d sinii. perfect, ind also the fteqneot 
Vedlc 3d alng. pieienl of th« root-claia (613), which are identical In fonn 
with their respective Orst pereona. That a fuller ending has been lost off 
it extremely Improbable; and henoe, aa an aotlst formation from the aimple 
root, this is most properly treated here, in connection with the ordinary 

844. Before the ending ^ i, a final vowel, and usually 
also a medial Q a before a single consonant, have the vrddhi- 
strengthening ; other medial vowels have the guna-stiengtheD- 
ii^ if capable of it (Z40]; after final ^ S is added JJ y. 

a. Eiunpleg (all of them quotable from the older langoage) u»: 
liom Tooti ending in S, ^fiBfi, iidh&yi, Ap&yl; in other vowels, fifrfiyi, 
iistfiTi, 4hftTl, jikftrl, Aetftri; — from loote with medial 1, n, j, aoetl, 
&oohedi, aqe^, &bodbi, Itmooi, ^oji, &darql, asarjl, vaibi; from 
roots with medial a atrengtheued, agKmi, ApSdl, dyjlml, avftoi, TSpl, 
isBdl (these are all the earlier caaea); with a nncbanged, only fcjani (and 
RV. has once Jjuli), and, in hexy syllables, imyakfi, vandl, Qafiai, 
ayandl; with medial K, AbhrOjl, &r&dlil ; — from moU with Initial 
Towel, firdbl (only ease). 

b. According to tbo grammarians, certain roots in am, and >''Tadh, 
retain the a unchanged : qootable are i^jani (or lOIni), agaml (or agjtml). 

ioy Google 

^05 SiupLi AoftiBT: 3. a-AOKisx. [ — 840 

Mtranli avodhit >Uo araol; uid there ue noted beddei, trom root* 
■omstiine* ihowtne & Baul, KdaAgi) antmbhl, araiidlii, i^jambhi, 
«bh&aji or abhajl, alunbhi (alwayi, irith ptepcwItioM) or ftUbbi, 
Htftmbbli Qh. hM OM&JL 

a. AagmeDtleu toima, m in itl othei like euM, >re met vlth, with 
eltlkat IndlettlTe oi mbJunetlTe iklae; eiunplM (beiidet the two oi thiee 
klready flTen) aie : dh^yi, qtIItI, bhtri, real. v6di, rool, J&nl, pftdi, 
sidi, ftrdbl. The ■<»«», when pieient, it »lw>ya on the root-eyllible 
(8V. dlUtfi >■ doubtlwi * t*Ua leading). 

S4ft. Thate forms ue made in HT. from forty roots, and all the other 
«atller tnti oomblned add only about twenty to the number; from the 
later languaE* are qnot^l* thirty or forty more; In the eploi they ara 
nearly unknown. When they come ftom roots of nentet meinlDg, aa kbul, 
pad. Bad, bhrBj, rftdta, ruo, ■aflj, they have (Uke the ■o-called paatiie 
paiticipls in ta; 659) a lalne eqnlTalent to that of other middle forms; 
in a e»»e or two (BT. Til. 73. 3 [T]; VS. xivili. Ift; TB. it. 6, 10*) they 
appeal eren to be used ttaniitiTely. 

2, The a>aoriat. 
846. a. This aoriet is in the later language allowed to 
be made bom 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 a-olass 
(878 ff.) or the if-claes (898 ff.). 

b. Its closest anal<^y is with the imperfect o£ the &-^lass 
{761 fr.]; its inflection is the same with that in all partieulars; 
and it takes in general a weak form of root — save the roots in 
^ r (three or four only), which have the gu^a-strengthening. 

o. As example of inflection may be taken the root 
ftra aio pour. Thus: 

actiie. middle. 

i sft^ srirpira wfw^rH ^ft% aftqrsrf^ afentr^ 

Asloam AsloftTa fwlcSma ftsioe &BloAvatai Uoftmohl 

1 aftra^ ^ftrarTT^ sftPifT wRiyEim^ yftwmiH,^ ^ftraan^ 

AalOMS ielcatam &aicata iaioatbia 4aloatbftm ^loadbvam 
Aaioat AsioatSm toioan AsioBta Asloetflm Asioanta 

ioy Google 

847—] XI. AORIST-BYSTBKB. 309 

847. Th« n-toilit makes tu the BT. i imall flgnte bMlde tii< root- 
■orlit, being repreaeuted by lais tban half the lattoi'i nnmber of root*. It 
become!, howeTer, more common latei (It ii the only tonn of aoritt wUah 
la made from more verba in AT. than in BT.]; ■"•^ In Teda and Brihmana 
together about eighty root* eiblblt tbe tonoatlon more or lata folly. Of 
then a laige nombei' (fully half) are of the type of tbe loot* vhich nake 
theli pieaent-ayatem according to the &-ela*i, haTlng a vowel capable of 
gn^a-strangtheDlng before a final conaonant (754): thns, with 1, obid, 
bbid^ nlj, no, Tl;, lip, Tld, Ifif (fia), 2i}l^ qri^, i;llf, bIo, uidh; 
— with a, krudh, kfadli, gab, duf. dyat, drnk. puf, badh, bbuj, 
muc, mmo, jmj, rao, rud, mdb, muli, mh, quo; — with f, rdb, 
kft, grdta, gfh, tn>, tff, tph, dn>, drq, dtaff. ur*^ mrdh, mn. Tft, 
TTdh, vn, Kfp, kff . A anall nombec end In Towela ; thna, f , kf , sf 
(which have the gui^a-itrengtheaing thioDghaut), hi (? ahyat once la 
AT.), and several in ft, apparent tranafer* freu the loot-claaa by the weak- 
ening of theii S to a; the*, khyS, b-vft, V7&, ^B, and dA and dh&j 
and fidthmt, regaided by the giammariana aa aoilat to ^as thrmo, la donbt- 
lea* a like formation from f/sthB. A few have a penultimate naial in the 
preacnt and elaewhera, which in thla aorlat li lost: thas, bhrafl^ taAa, 
dhvftAa, srftAB, krand, randb. Of lest claaalflable cbaraclei an ftf, 
kr&m, gam, Rbas, tain, gun, from, tan. Ban, nad, &p, das, jnM, 
9ak, dach. The roots pat, nag, tso form the tenae-etems papta, ii«f a, 
TOOa, of which the flrat la palpably and the other two are probably the 
resolt of lednpli cation ) but the language has loal the aense of thelt being 
inch, and makes other redopllcated aoriats from the same roots (lee be- 
low, 8U> 

a. Many of these aorlits are almply tiansfeis of the root-aorist to an 
a-lndectioQ. Oonaplcuoos eiataplai are akarat etc. and agantat etc. (In 
the earliest period only abar and agan). 

848. The Inflection of this aorlst U la general so regnlar that It will 
be suffldent to give only examples of Its Yedlc forma. We may take aa 
model avidam, from jiMd^^nd, of which the varlona persons and model 
are more frequent and In fuller variety than those of any other verb. Only 
the form* actually ijnotable are instanced; those of which the example* 
found are from other verba than vid are bracketed. Thua: 
aotive. middle. 

1 &vidam &Tldftva &vld&ma &Tide [&vidftvabl] ividftmahl 

1 &vidaa [&Tldata] [Avidathfta] 

» &Tldat iivldaii [avidata] [avidet&m] Avldanta) 

a. The middle fonos are rue in the earlier language, as in the later: 
we have Uive etc., &kh;e etc., Avlde (7) and aridanta, avooatbAa 
and aTOOSvahi (and avidSmabe QB. and aelofimahe KB. are doubtless 
to be amended to •mabl). 

ioy Google 

307 Simple Aoubt: 3- h-aorist. [— 8BS 

b. Angmentleu farmg, with Inillc&tlTe ot subjunottie Talse, an not 
tnfreqnent. Eismplw, ihoirlng aecGot on the tenie-slgD, iMOrdlag to the 
general maloglet or the formttlon, us: mb&m, Brpaa, bbnj&t, Tid&t, 
KnttSm, TOoatA, qaikAn; vldata mA vy&ta (3d iln|.), arSiaalii, 
flfamabi, vldinta, bndli^ta, mntuite ((or eiceptiona u tegudi 
ucent, aee below, 8B3}. 

Uod«B of ihe a-aorist. 

848. The tubjuncttve foTiDi of tUi loHit are fev; thoie nhtoh occoi 
m ln>t4DMd below, in the method which w»s followed foe the IndlcatlTO: 
1 [vldava] vidsma [vld&mahe] 

1 |lj^ vidsthlta vldaUift 
s TldAt [TidUai,P] 

a. The endlpg tliana is fonnd once, In riffithana. Of middle forma 
ocoor oul<r gifttU (AY.: bnt doobtleee mtgrBidlng for flqr&tfti) and 
fiQKmahe (AT., for RT. flfftmtthl). The form B&dathas aeema an Indic- 
ative, made from ■ leeondair prasent-atem. 

BSO. The opUtiTei are few In the oldest Ungnage, bnt become more 
frequent, and in the Brahmanaa are not rate, Eiamplei are: In artive, 
bhidayam, Tid^yam, eandyam (TR. odco sanem); vid6fl, camea; 
garnet, Tooet; gametam; gam^nia, fahdma, san^ma; vareta; in 
middle, (onlr) vidsya; gamemahl, vaaomahi: robethfta etc. In the 
•plea muBt be ilewed rather as preient forma of the &-olaM. 

a. A ilogle middle preoatlre fbrm oceuri, namely vldef^a (AT., 
once); it t* lo laoUted that how mnch ma; be iDterred from it it lery 

8K1> A complete leriee of active ImperatiTe forma are made froin 
y^ad (Including sadatana, 2d pi.), and the middle sadantSm. Othet 
impecaUves ate Terjp rare: namely, B&na, B&ra, raba, vldA; rub&tam, 
TldAtam; kby&ta. 13. has once vrdbfita (compare 740). 

Fartiolples of the a-aorist. 

86S. a. The active parti cipl ea ■ t^pint, H^ant or riqant, vfdhtot, 
flf&nt, 9ao&at, B4daiit, and (in participial oompoanda, 1309) k^taat-, 
gobaot-, Tidant- (all RV.), are to be aaalgned with planaibliltr to this 

b. LIkewiae the middle partlclplea gab&mfina, dbff4mS9a, d&sa- 
m&tia (?), nftiunftna, gao&mSaa, and perhaps vfdhBJi&, erldhftii&. 

IrregularltieB of the a-aorlst. 

8B3. A few inegnlaritiea and pecnliaritles may be noticed here. 
The roots In f, which (847) show a atrengiheuing like that of the 

D,j,i,...., Google 

8B»— ] XL A0KI8T4TSTEMS. 30g 

present of the niue«ent«d ft^lui, hiT* liknrlte the seetTit apon tba 
radiral (yUibls, llk« tiiat eUw: thm, ft«m Yf, iranfai (ansmentlwa 3d 
pi.). sArftt ami aArk, Tke root skd fiillowi tka ume ial«: tboi, a&d*- 
tam; nd from >/s«ii ue foDnd a4lUM lod vinat uid BineaM mui 
■4n&, bulde •antrun md Mninut. It !■ qnutionablc v1i^«r tkew 
ue not tms antlofne* of tfa« bhQ-clu« (anMcaiitcd a-clut) pietent-irftan. 
On (he other hand, rdhat (bMlde mhAm, rali«ra, rohAtam), fl^t 
■nd ^{fttU (T), and rffutt or rtfutt are more iMiited cues. In Tiev 
of inch u these, the foimi from the item bbdva and f ruT& (886 o) 
ue perhapa to be lefaned hither. From y-vma, the optatire ii Mc«nted 
vooiytuu, Toote, Toeima, voo^jiu; elsewbera the accent U od the nwt- 
lylUble: thus, v&0«, T6eHt, vAoati, viatatm. 

8B4. ». The etem Toe hai In Yedic me ireU-nifh aeaaned the 
Tilae of a root; it* forma aie tbt; Tartona and of freqnent Die, in BT. 
aipoeially far ODtaoinberfng In occnirencai ^I other farm* from y^raa. 
Bealdei thoao alraad; given, we And tooA (let alng- impi.) and vooSti, 
mo&TahU; voce*, Toa«7a, vooemahl; vooatU (2d aiug.), vooAto, 
Tocatam, vocata. 

b. Of the stem De^a only na^at occnta. 

0. The root qfis (w In soma of ita preient forma: 688) i* ireakened 
to Qif, and make* a^ifam. 

S68. Isolated forma which haTfl more or lew completely the 
upeet of iDdieative presenta are made in the oldest laognage from 
tome roots beside the aorist-sjstems of the Grat two classee. It must 
be left for matorer research to determine how far they may be relics 
of original presenta, and how far recent prodoodona, made In the way 
of conversioD of the Aorist-stem to a root in valne. 

a. Such forms ue the following: from ytcf makt, Urfl, k^tbao, 
kftba, ^n^i from Vgam, gat]i&; from yei ffathtr, aetii from y'dS 
fine, d^tl, dStn; from ^dbS put, dhfttl; from yp& drini, pSth&S, 
pftntl; from yhbx, bhartli from ymao, mue&ntl; from ymdh, radb- 
maa Q); from v'vrt, Tftrttl. 

II. (3) Reduplicated AorlsL 
SB6. The reduplicated aorist is different from the other 
fotma of aorist in that it has come to be attached in almost 
all oajei to the detiTatiTB (causative etc.) conjugation in 
WI ijBi, as the aorist of that conjugation, and is therefore 
liable to be made ftom all roots which hare such a conju- 
gation, beside the aorist or aorists which belong to theii 
primary conjugation. Since, howerer, the connection of 

ioy Google 

309 3' Reduplicated Aokist. [ — 868 

the two 18 not a formal one (the aorist being made directly 
from the root, and not from the oauaative stem), but rather 
a matter of established association, ovring to kinship of 
meaning, the formation and inflection of this kind of aorist 
is best treated here, along with the others. 

8B7. Its chaiaflteristio is a reduplication of the radical 
syllable, by which it is assimilated, on the one hand, to 
the imperfect of the reduplicating class (6BS ff.], and, on the 
other hand, to the so-called pluperfect [817 ff.). But the 
aoiist reduplication has taken on a quite peculiar character, 
with few traces left even in the Veda of a different con- 
dition which may have preceded this. 

868. ». As r^ards, indeed, the consonant of the re- 
duplication, it follows the general rules already given (680). 
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 x (or arj are usually (for exceptions, 
see below, 860] repeated by an i-vowel — as they are, to a 
considerable extent, in the reduplicated present also (680), 

b. But in regard to quantity, this aorist aims always at 
^tablishing a diversity between the reduplicating and radical 
syllables, making the one heavy and the other light. And 
the preference is very markedly for a heavy reduplication 
and a light root-syllable — which relation is brought about 
wherever the conditions allow. Thus: 

S60. If the root is a light syllable (having a short 
vowel followed by a single consonant), the reduplication is 
made heavy. 

a. And this, Qaaally bj lengtheDing the reduplicating vowel, with 
I for radical a or p or } (in the single root containing tliat vowel): 
tbna, arirlfam, adudufuo, ^ijanam, avivrdliam, a(iIk|pRm. The 
great m^ority of rednplicated aorlBte are of this form, 

b> If, however, the root begins with two coDBODants, so that (he 
rednplicating sfllable will be heavy whatever the qnanlltr of its vowel, 



SftB— ] XI. AOKIBT-SrSTEHB. 310 

the Towel renutiDB short: tbtu, aolkf Ipuo, ftookradtuun, aUtrasun, 

860. If the root is a heary syllable (tkaving a long 
vowel, or a short before two ooosonante), the vowel of the 
redupJioation is short: and in this case Q a or ^ &, and 
ff p (if it occurs), are Teduplicated by Q a. 

a. Thus, adidikfam, abubbOfam (not quotable), adadakfam, 
adadhfivam. atataAsam. And, in the caseg in which a root sttotild 
both begin and end with two oonsonants, both ayllables wonld be 
necessarilj beary, notwithstatidliig the short vowel in the former; 
thne, apapraoobam, aoaskaadam (bat no suob forma are found in nse). 

b. A medial j li allawsd by the gnmrnarluu to retiln the itrenKtben- 
Ing of the eanntlTe ttem, togethei with, ot mutib, lednpUeitlon by ar thw, 
aoakarfat, avavartat (beside atdk^fat, avivrtat); bat no rach form* 
hire been m«t with in lue. 

o. These torlMt u« not dlattngnlibable In fntm from the so-ulled 
pluparfecu (817 ft.). 

861. a. In order, howeveT, to bring al>ont the &vored relation 
of heBT7 TeduplicAtion and light radical syllable, a heavy root is 
sometimes made light: either by shortening Its vowel, as in ariradham 
from yrKdli, avivaqam from Vv&q, a^adbam from yMdh, ^JyiTam 
from yjiv, adidlpam (E. and later: RV. has didlpas) from ydip, 
abibMfam from ybhi^ aailBuoani from ysilo; or by dropping a 
pODUttimate nasal, as In aoLkradam from f'kraiid, asifyadam from 

b. In those oases in which [1047) an aorist is formed directly 
from a causal stem in Sp, the & is abbreviated to 1: thus, ati^ftalp- 
am etc., ajlJiUpat (but ESS. ajljfiapat], jitaipas, ajijlpata (but VS. 
^yapata); but from frap comes afl^rapftma (QB.)- 

86S. Examples of this serist from roets with initial vowel aia very 
raiej Ibo older language has only Kmamat (or amamat) from yam, 
ftplpan (^B. : BAD. Bptpipat) from yip, and arplpam (aogmentleM) 
fTom the Musattve stem arp of yf — In which lattsr the root U aieess- 
iTely abbievlMed. The fiammaTiana give other similar formatlona, u Srai- 
oam ^m yaro, SnbjlJani from yubj, Brjlbam from yarb. ftlolkfam 
from V^ti ■fdidham from ypdb. Compare the similar redaplication In 
desldeiatlie Btema: 10S9b. 

803. or special irregularities may be mentioned: 

a. From ydyut la made (T.B.) the stem dldyuta, taking Its redu- 
pUcaUng Towel from the radical eemlTOweL From yftap, Instead of JOcn- 
pa (B.S.), JB. hu Jugupa, and aome tens (BS.) have Jognpa; ud 
Jlbvara (B.) li met with beside the legatai Jlhvara (Y.B.). 

ioy Google 

311 3. Bbduplicathd Aorist. [ — 867 

dft (Nli-)i "*^ *^^ "*"" "' ^'** •lonbtfol paipr&tlu uid fOfraoA tod 
Baav^ft (BV.) we have a lusMul of i in the rediipUcatlon. 

b. In gnpport of thelt hlw Tlew of this loriat u made from the 
MOMtlTe Item ln«te»d of directly from the root, the natWe giamnuriuia 
taach thit loola ending in in a-TOwel mey rednpllute with 1, >■ repieunt- 
Ing the i of the atrentthened stem; thne, blblMTa tana bh&T-aya, ai 
well u bflbhava from bhIL No example of (ueh a formation, howoTer, 
ii met with except AplplaTam iQB., once); igalnat It we And dndtuva, 
bfibhuva, rdraTa, ^afraTO, and other*. 

o. Aa to ftpaptUD, avooam, and ane^aiii, aee aboTC, 847. 

884. The inflection of the reduplioated aoriat is like 
that of an imperfect of the second general conjugation: that 
is to say, it has ? a as final stem-vowel, with all the pe- 
culiarities which the presence of that vowel conditions (7SS a). 
Thus, from I'SR jan ffiee birtA (stem jljaua): 

aetiie. middle. 

ijljaoam ^^antva ^^anbna AJ^ane Ajijan&vahi iijyaa&mtthl 

AjyanoB jj^anatam iKJ^anata AJIJajtathJU JJXJanath&m ^yanadbvam 

ijyanat AJiJanat&m ikj^anaa lijijaiiata 4]UanetSin AJlJananU 
8a&. Tbe middle forms are rare in the older langnage (the 8d 
pi. IB decidedly the most common of thein, being made from eleven 
roots; tiie 3d s. from Beven); bnt all, both active and middle, are 
quotable except lat and 2d du. middle and Ut dn. active. 

a. Atitape ippeara to be onco naed (ftT.J aa 3d stng., with paulTO 

866. A final f has the cu^a-etrengtbeDiDg before the endings: 
thus, ai^karat, aplporam, atitarai, didaras, adidbarat, amimarat, 
avivaran, JihTaras. Of similar strengthened forms from I and u-roots 
are found apiprayau [TS.j, ablbbayauta (BV.}, aplplavam IQB.j, 
aoQoyavat (K.), aQu^ravat (MS.), atuftavam (RV.). Not many roots 
ending in other vowels than j make this aoiist: see belov, 868. 

867. Forma at the Inflection ullhoat nnlon-iowel are occaalanally 
met with: Damely, from roots ending In conionanl», af^ap (2d aing., 
angmentleai) from ysvap, and aqiqnat from |/qnath; from roots In f 
or ar, didhar (?d aing.), and ajigar (3d and 3d sing.); for rootg In t- 
and tt-Towela, see 868. Of 9d pl. In na are fonnd almost only ■ form 

ioy Google 

887—] Xr. A0KIBT-»78T«1IB. Zit 

ot tiro ttom 1- Mid a-TiMt», with gnifm before the ending: thiu, A^qrayiu, 
AoDOTaviu, afngrainu, uo^Toe; but tUo ablbh^nB ((B.), uid 
Dinaf oa (HBb.). 

808. In the later Isn^Etge, ft few roots are uid by the gram- 
marians to make this aorist as a part of their primary coDJagatioa: 
they are gri and qvi, dm and arn, kam, and dlift *utk (ijvi and dbs 

a> In ths oldn langnage ire found nrom Vqrl aqifret >nd a^^rarns 
(noticed in the preceding paragiipb) uid a^l^rlyat (^B.); from y&to, 
adndrot uid adudravat (TB.: not nscd u loriit)-, fiom v'aru, asnarot 
■nd (lugmeDtleii) BOflToa uid aaaroti fiom yktaa, aolkamstsiii uid 
-manta (B.S.)' Of forms anUogoiu with theie occai > nambec from rooti 
in u or Q: that, anOnot uid nOnot from yaa; TQyot from yyn 
teparate; dfldliot ftou ydtaa; apapot from v'pQi tOtoa and tfltot from 
yta-f aaufot from f^fl; — uid one or two from roots in 1 or I: (hut, 
alfflt from yei (or aft) hind; amtmet from yiai beUoie ; apiprea (wiih 
i^iprayan, noticed above} from yplA (and the Hmperfects" from didhi 
etc., 670, are of corresponding form). And from yaya are made, with 
nnlon-TOwel I, aanoyavlt and aauoyaTltana. Few of these forms pcuei* 
a necessully cansstlTo or a decidedly aoriitle value, and tt la ver; donbtfo] 
whether they should not be atslgned to tbe perfect-ajiitem. 

b. From the later language are i^notable only aqifrlyat etc. (3d iL, 
-yan or -yua) and adudiuvat. 

UodeB of tbe BedupUoated Aoiist. 

869. a. As in other preterit formationa, the augmentleea indicative 
peteenB of tbia aorist are used labjanctively, and they are very much 
more frequent than true anbjanctlyeB. 

b. Of tbe latter are found only rixadhft (1st slug.); tjtaploi; 
oiklpRtl and fllfadhfttl, and piap^fatt (aa if oonespondlng to an indlC' 
ative aplapf'k, like afi^nat); and paihaps the lit slag. oild. gafvaosf. 

o. The augmentless indleatlTe forms are accented In genaral on the 
rednplloatlon : thus, didliaraa, nina^as; J^anat, piparat; jijanan; 
also aifwapj bnt, on the other band, we bsTe also pip&rat, ^fr&thas 
and ql^n&that, and dadrivat and tnqt&vat (which may peihepa belong 
to Uie perfect: compare 810). According to the native grammarlana, the 
accent rests either on the radical gyUable or on tbe one that fallows it 

870> Optative forms are even rarer. The least qaestlonable oase 1* 
the middle "precadve" ilriqifta (riri^|ta hat been ranked above with 
ateahigta, as ■ perfect; 819 b). OucyuTimahl and onoyavlrata be- 
long either here or to the perfect- ay stem. 

871. Of impeiatlves, we have the Indubitable formi pQparanta and 
^gratbantu. And JlgftAm and jlgilA, and didhftam and didtaTti, 

ioy Google 

313 Sibilant Aorist. [—876 

and JajuUm (lU RT. onlr}, ind peihapi BtUfttiUta (AV.), aie to be 
TBtened hither, as coTTespondtng to the IndiutlTOi [wlthoDt nnlon-Towel] 
i^lxgar andadidliAr; their ihart ledaplieatlng vowel and tb^li accent 
•HlmlUio them closely to the rednplleated Impetfeett (9M It.), with which 
wo aie probably to tegaid tbU aoitit ai Dltlmataly related. 

879, Ko puticiple is found belonging to the redaplioated aorist. 

87S. The nnmbet of roote from whicb tblB ftoriet Is met with 
in the earlier langnftge ia about a hundred and twenty. In the later 
Sanskrit it is unusual ; in the eeiiee of later texts mentioned above 
(8S6) it oecQTs onlj' twice; and it has been found quotable from hardly 
fift; roots in the whole epic and classical literatnre. 

III. Sigmatic or Sibilant Aorist. 

874. a. The commoQ tense-siga of all the varieties of 
this aorist is a H b (convertible to tf 9: 180) which is added to 
the root in forming the tense-Btem. 

b. Tliia aibilant baa no analoguea among the olaaa-slgD* otthe pteaent- 
■yatemj bnt it Ib to be campared with that which appears (and tlkewUe 
with or withODt the tune anlon'iowel 1) Id the item* of the lature tenae- 
tyBtam (888 11.) and of the desjdetaiiie coojogatlon (1087 fl.}. 

o. To the root thus increased the augment is prefixed 
and the secondary endings are added. 

876. In the case of a few roote, the sibilant tense-stem 
(always ending in ?! k?) is further increased by an CT a, 
and the inflection is nearly like that of an imperfect of the 
second or a-conjugation. 

876. a. In the vast majority of cases, the sibilant is 
the final of the tense-atem, and the inflection is like that 
of an imperfect of the first or non-a-conj ligation. 

b. And these, again, fall into two nearly equal and 
strongly marked classes, according as the sibilant is added 
immediately to the final of the root, or with an auxiliary 
vowel 5 i, making the tense-sign ^ i^. Finally, before this 
^ if the root ia in a very small number of cases increased 
by a H B, making the whole addition TTT^ sig. 

ioy Google 

877—] XI. A0BIflT-8TBTEM8. 314 

877. We have, then, the following olasnfication (oi the 
varieties o£ Bibilaot-aoiist: 

A. With endings added directly to the sibilant: 

4. with Hs simply after t^e root: a-aorist; 
h. with ^ i before the ? s: i^-aorist; 

5. the same, with TT a at end of root: siv-aoiiat. 

B. With S a added to the sibilaot before the endings : 

7. with sibilant and 9 a: sa-aorist. 
a. Aa legarda the distinction ^etveea the fouilb uid flfth fonni, It 
may be sild in a genenl wtr that thoae roots Incline to t^e tbe (nxllUty 
1 in ttie lorbt which take It also In other forrnktlonei bat it la impoadble 
U) lar down any atricl inlaa m to this accordance. Compare 803. 

4. The B-aorist. 

878. The tense-stem of this aorist is made by adding 
IT 8 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 If? t) has the vrddbi-ohange 
in tbe active, and (excepting ^ x] iu^ in the middle: thus, 
from y^ lead, active stem 5^^ anSi^, middle stem ^l^^ane^; 
from y^ 9ru hear, USflH a^r&u^ and MVll^ a^rop; from 
y^ kr make, *mi\fi ak&rf and SRi^ akff. 

b. A medial vowel has the vfddhl-change in the active, 
and remains unaltered in the middle: thus, from ya-^ ohand 
leem, active stem yc*lr+1 aooh&nts, middle stem M<%^ 
aoohants; from y^ rio leave, ^^ arUk^ and ^tj^ arik^; 
from v"^ rudh obstruct, Q^fH arRute and W^r^ amts; 
from vTlsT Sri pour out, UHM aBrSk? and ER151 asfkf. 

880. a. The endings are the usual secondary ones, with 
3TI us (not ^{ an) in 3d pi. act., and ^ ata (not^ antaj 
in 3d pi. mid. 

ioy Google 

315 Sibilant Aobibt: 4. b-aori8t. [— S8S 

b. But before R b and FT t of 2d and 3d sing. act. is in 
the latei language always inserted an ^ I, making the end- 
ings ^ Is and ^It, 

O. This inaerUon li nnknoim In th« ttfliest linfntge (of the RT.) : 
•M below, 688. 

861. R. Before endings beglnoiDg with t or th, the tenBe-sign ■ 
is (898 a-«) omitted After the fiDXI coDsonsnt of a root — unlesa this 
be r, or n or m (converted to anaavSTft). 

b. The same omiialan li at eoune made before dbvun after a con- 
Mnant; aod after a totoI the libiUnt Is either omitted ot siBimilatad (tho 
equlTalenee ot dhv and ddhv io the tiieorlet of the grammiiUae and the 
ptaetJM of tbo maDiiacripts makea It Impoeilble to aay which: SSB); and 
then the endlDg bceomea ^hvam, prorlded the sibilant, if retained, would 
have been ^ (886 0): thus, uto^vam and Avj^vKm (beside aatof- 
ata and avTfata); dr<^T«m (>'dx regard: (B., onee), whlah 1* to 
dfthfta (2d slog.) as KV^^VKca and av^ata to avTi and aTTthU; and 
kr^vam (H.). 

O. According to the- grammsiians, the omlsilon of a before t and th 
takes place also after a short vowel (the case can occur only in the 2d and 
3d ilng. CDld.); bnt we bare seen above (884 a) that this Is to be ilewel 
rather aa a substltutlDn In those persons of the forms of the root-sotlst 
Neither in the earlier nor In the later language, howeier, does any example 
occur of an sorist-form with b retained after a abort vowel before these 

d. After the final sonant aspirate of a root, the sibilant before the 
same endings is said by the Hindu grammarians to disappear altogether, the 
oontblnation of the aspirate with the th or t of the ending being then 
made iccotdlng to the ordinary rule for such eases (160): thus, from the 
•tern arHnts, for ar&ndh-s, is made arftuddha, as If from nr&udh •(- ta 
directly. No example of such a form Is qaotable from the literature; but 
the combination Is established by the occurrence ot other similar cases 
(883 f). In the middle, in like manner, aruts + ta becomes amddha, 
u if from arudh + ta; bnt all sach forms admit also of being understood 
as of the root'aorial. Those that have been fonnd to occur were given 
above (884 d); probably they belong at lea«l in part to this aorlst. 

e. From the three nasal roots gam, tan, man are made the 2d and 
S4 ling. mid. persons ngathia and acata, atathSa and atata, and amata 
(amathBs not quotable), reckoned by the native grammarians as B-aorlst 
forms, made, after loss ot their flnal root-nasal, with loss also of the sibilant 
after a short vowel. They are doubtless better referred to the root-aorlst. Bnt 
JB. haa a oorreiponding 1st sing, ataal bom I'tan. 

882. As examples of the inflection of this variety of 

ioy Google 




sibilant aorirt we may take the toota Hi nl lead, and f^^ 
ohid cut of. Thus: 

mctlTB. middle. 

1 #fN^ ^^ ^^ ^ft M^ai^ 

inftlf&in inftlfva inU^ma &n«fi AnOfTRhl Aa«fm«fai 

^Sl^ inUftain Anftlfta AnSfthOs Ane^Kthlm. 4iie^Tain 
3 fcflHtft^ tR?;i'^ *filj*i^ ig^ WHHIHTH^ tf^^H 

Aufil^it &n8lQt^ &naifus &ne^ tee^tftm ineqata 

^ «• ^1- ^ P- 

1 *f4^rHM^ «l^rH M^rW 

&col)&itBain &aohSitBva ioohUtama 

ioohtitelB iooliUttam 4ooliiitU 

I y^'rHlH M^tIIH yc^'rWfl 

AoohftltBlt &ccIi&itUm &ooh&itsua 

iootaltsl icohltBTahl icohitamahl 

iioohlttha* ioobitaathixa &ooblcldliTtua 

&oobltta &oohitsfttam ieohltuta 
a. From yr\idh. obtbtiet, the 2d and 3d du. and id p). set sod 
the 2d and 3d eing. mid. would be Ar&addbaiii, ^taddliaiii, 
irladdha, &raddIiBe, iroddba; from ytfi pour out, iuSftam. 
iarB^tAm, aarftqtai aant^^ aar^; from ydfq im, idrftftam etc. 
(aa from bi-J). But from yk^ do the same pereoDB in tlie active are 
ikSr^^am, ikarfffim, Akftr^ta; from Vtau ttreteh tbey are AUUiatam, 
iULOetftm, &tUsta. 

689. The amliiian of a Id the ictlve penons (&oollUttain, iooUlt- 
tarn, 4oobaitta) 1b ■ cue of nrj Tire ooouRence; til the qnottble eius- 
plea vere giveo above (938 e). Aa to the like omiulan In middle parHna, 
■ee 881. The OhU. ha» twice AvKstam for avita-tun (yvM dtBtOi: 
thia may be Tteved aa anothar case of total disappearance of (he elbllant, 
and eonaoquent leatoratlon ot the final radleal to tta origliial form. 

ioy Google 

317 Sibilant Aorist: 4. s-aokibt. [—888 

8S4. Cortain roots io & weaken thA & in middle Inflection to i 
(u also in the rooc-aorist: above, 834 a): these are said to be atbB, 
41, and dbB; in the older language have been noted Adlql and adifata 
fVom ydt giv» (and adlfl perhaps once from yH bindj, adM^ and 
Adbifata (with the optkUve dhi^iyai] from ydh& put, and aathi^ta; 
>lBO ftcl9(hfia and «e3t»tti from ygi go (with adlil). 

R. The middle inflection of the aorist of ydi would be, then, 
-Moording to the grammarians: Ul^i, 4dithflB, &dlta; A&i^vSd, 
idifftthAtn, idlffttftm; Adl^mahi. fcdi^Tua, idl^ata. 

88B, Root* ending in ehanguble f (so-oalled roou in f: 242) are 
Mid br the giftmnurtsDi to eonisrt thl* vowel l« Ir In middle farmi: thns, 
-astlr^ ast&-qthBa etc. (bom Vatf); of anoh forms, honevar, hu been 
ionnd in the older Ungoige only amr^ata, PB. 

686. The a-aorist is made in the older language from abont a 
hnndred and forty roots (in RY., from about seventy; in AV., from 
about fifty, of which fifteen are additional to those in RV.J; and the 
epic and classical literature adds but a very small number. It has in 
the Veila certain peculiarities of stem- formation and inflection, and 
.also the fall series of modes — of which the optative middle is re- 
tained also later as a part of the "precative" (but see 026 b). 

887. Inegularities of stem-fomiaUon are as follows: 

K. The itrengthentng of the rooc-eylisble U now md then iiregalBTlf 
made or omitted: thus, ayok^t (AB.), obetaia (B.3.; Uao oecnra In 
HBh., ohlEb hM tuiUieT rotAs), rottia(KU.)i amatooB (KY.^ arSifaBi 
■nd arSntol (AB.), ulkqi etc. (T.B. : Keah), mUsta (AT.) ind mSABtAm 
(TA.); lopMra (v.); and HBb. has drofldbSB. From yani Is made 
Blfik;It (V. etc.), >nd from ymtiii, amgflt^t (not qnoUble). The form 
aTaflkqinabi (BhP.J is donbtleei i falae reading. 

h. A ndicil flnil nuil ia lott In agaamaU (RV.) and gaa&UiKm 
(TA.) from Vgam, and In the opUtiTee mat^a and vaalmahi (BT.) 
from yjuan and van. 

O. The rooU bu, dba, and iiQ haTS d Instead of o in the middle: 
thai, abufata, adbo^ata, emOfi and anilfStSm and anQf^ta; f'dbur 
(or dbuTv) makes adbOrfata. 

d. (JB. has once atr&Bat&m for atrAatfim (yttft). 

8SB. Tbe principal peculiarity of the older language in regard 
to inflection is the frequent absence of I in the endings of 2d and 
.3d sing, act., and the consequent loss of tbe consonant-ending, and 
sometimes of root-finals (150). The forms without I are tbe only ones 
foDud in EV. and E,, and they outnumber the others In AV. and 
TS.; Id the Brahma^as they grow rarer (only one, adrSk, occtirs in 
QB.; one, arS^ in KB.; and two, adxfik and &yi\. In QB.; PB. has 

ioy Google 

S6e— ] XI. Aobibt-btbteus. 3i8 

88B> If the loot ands In ■ vowel, only the eousoiiiiDt of tba ending 
ii neeeiBirll; Itut: thus, aprfis (for both aprSa-S ftnd sprlft-t) bom ypAi 
»Dd in Ilka mtnner abb bom yliftj — aJUa (fOi aJUf-t) from yii; tad 
in like mannei ac&is from yol, >nd nftio (oijiaBntlMB) fMm >^m; — and 
7&UB (for ayHtif-t) from v'yn. 

a. Bnt (aa In other like oaaee : BBS ■) the ending la aometimec preacr- 
Ted at the eipenae of the tenaa-ilgnj and we haie In Sd ilng. ^Slt (Ira- 
aide oj&ia and i^U^t) from yH; and in like manner kcBit, afrSit, 
ah&lt, nSit (no examples have been noted except fiom loota In i «nd I): 
compare Kris and BTto, 2d sing., 890 a. 

800. a. If the root (in either ita simple oi itiengthened form] esda 
in a conaonuit, the tenae-algn is loet with the ending. Thna, ablulr (for 
abhftrf-t: beside abhSr^am, abhSr([{fttIl) from ybhj; other like cases 
ate obfir, and (bom roola in ar) akfir, ata&r, aevir, bvKr. Farther, 
trUk (B8Ba; for ar&ikf-t) ftom yiio; Uke case* are a^Ut bom 
ytfvit, and (bom root* with madid a) adfftut bom ^dynt, ariat bom 
y^dh, and mfiiUk bom v^nc. Further, bom roots ending in the pUt- 
tall and h, aprftk from yp^o, aorftk from yn^, abhftk bom ytihai, 
adrfik from y^Ji}' n^^^f^ f""" l^dab; bnt, with ■ different change of 
the Una], ayS( bom v'yaj, aprftf from t'PT'^i avfif ftom vVab, and 
aarfit bom ysfj; and (above, 146 a) btKb appeara to stand twice in AT. 
fbr BTfif-8 from ytfi; RT. has also twice aySa bom yj»i. Further, 
from reota ending in a nasal, atSo bom ytan, khftn from yUian, aySn 
and an&a bom yyjfm and nam ^143 a). 

b. If, again, the roots end lu a double conaonant, the latter of the 
two is lost along with tense-sign and ending : thna, aooban (for aoohftnta-t; 
bealde aooh&Otta and aocbKntsun] bom ych&Ddj and other like oaaea 
are akrOn, aak&n, and nay an. 

891, A relic of this peenliaiity of the older inflection hat baen 
preaerved to the later language in the 2d sing. bhSlB, from ybhl. 

UodeB of the s-Aorlst. 

88S. The indicatlTe forma without anient are ased in a sub- 
junctive seoae, especially after mi prohibitive, and are not QncommoD. 
Examples with accent, bowever, nre extremely rare; there has been 
noted only tr&Aai, middle ; jadging from this, the tone woald be found 
on the radical syllable. According to the Hindu grammarians, it may 
be laid on either root or ending. 

893. Proper subjunctive forms are not rare in RT., bnt are 
markedly less common in the later Tedio texts, and very seldom met 
with in the Brahmanas. They are regularly made with gu^a-atrength- 
ening of the radical vowel, in both active and middle, and with accent 
on the rooti 

ioy Google 

319 Sibilant Aorist: 4. b-aorist. [—698 

a. The forms with prim»iy sndlngtarei la aotlve, stog^il; dar^aal; 
ii«(atl, parf ati, pftanti, m&taati, yo^ti, Tak^tl, aak^ati; dAiattau, 
dhftBatbu, pir^athaBi vakfathaa, vosTfathM; pfiaataa, yuhsataa, 
yakfatas, vafc^ataa; dhaaattia, nOQatha, p&r^atba, m&taatha; — 
Id middle, naifasftl, m^Aaftl; m&ABaae; kradiBatc, trtaate, dar^ts, 
m&Asate, yakfate, r&sata, vaAaatA, s&kfate, Usate; trasSthe (not 
ttAaUthe, u we Bhould rsthsi expert); u&ihaaiite, m^LAsaute: and, 
with the fnller ending In 3d sine-, miallt&i. 

b. Tbe foims wlih BMonduy endings ire (Ktlie only) : J^M, vikfas; 
lUrf at, ii6fat, p&kf at, p&n^at, pri^t, y&k;at, y6^t, viiAaat, rUq^at, 
v6fat, 84taat, ohantBat, etc. (some twenty othere); yakfatftiBi viJi- 
BSma, eti(9&iiia, stOfSma; part[aii, yadiBan, yo^an, rasan, vakfan, 
fdfan, qToqaa. Of theia, yakfat and vak^t ire found not itiely in 
tbe Briliniinia; iny others, hirdl; more thin spaiidiolly. 

894. Of inegntiTltiei lie to be noted the rollowlng: 

a. The fotmi dfkqaaa ind pfkfafs (3d sing, mid.) lick the gu^a- 

b. Je^am, ato;am, ind yofam (AT. ynfam, with Q for o i« in 
an&fata etc.) ippeir to be flret penong formed under government of the 
inilogy of the lecond end third — nnlell they ire lellcg of i itite of 
tbingi interior to the v^ddlii-strenethenlng: in whti^h cue Je;ma is to 
be compired vith them (we Ebonid expect IKi^ma or Je^ftma). 

C. From root* in K ere mide ■ few formi of pioblemitic chiricter: 
nimely, yefam (only cue In RT.), khye;am, Jfiet[ain, ge^am lod 
ge^ma, detfma, sefam md aet, Bthefiam and sthequa. Their vilne 
le optatiTO. The inalogy of Je^am ind Jeqma suggeite the poasibllity of 
their derltatioa from 1-form* of the a-roots; or the sibilant might be of 
a preeative cbiiacter (thaa, ya-l-B-am). That Ihey really belong to ths 
i^iorlat ippeiT* highly Improbable. 

d. The RV. bat a (ew difflcalt Brat peraoni middle in ae, which in 
perbipa best noted here. They are; 1. fhim the simple rrot, kffe, hl^ 
(and olii^eP), Bta^i; % from preaent-atems, aroase, fnjaae, y^jaae, 
gftyife, gF^^ and ptmi^d. They hive the Talua of indicative present. 
Compare below, 897 b. 

895. Optative rorms of this loriet ire mide in the middle only, and 
they have in Id ind 3d sing, always the preeative b before the endings. 
Those fonnd to occur in the older language are: dlflya, dhlflya, bhak- 
^A, ntaslya (for maAsiya), mukflya, rialya, lop^ya, e&kf^a, 
stfflya; moA^t^fis; darfma, bhakfifitA. maAeifta, mpk^ffa; 
bhftkqjtmahi, dhakflm&hl, maABim&hi, va&Bim&hi, vaslmatat, 
sakflm&hi; maAairata, PB. bu bbuk^iqiya, which (honld belong to 
a slf-iortBt. The RV. form tr£iBlth&m (for trfislyath&m or trOsftthftm) 
is an isolated anomaly. 

a. This opiitlve makes ■ part of the accepted "ptecadve" of the IMer" 
liDgvige: see below, 983, 926 b. 

ioy Google 

886—] XI. AORtBT-STBTEHa. 320 

896. ImpuitiTe perrons bom thli urUt uc extcemelT ma: wt flnd 
th« 3d aiDg. a«t. n«fa >ud par^A uid the 2d pi. ytuhBat& C^m a-itama, 
and ihowing lathei, thetefare, > tieitment of the aoriit-item u > root], 
»nd the 3d ilng. mid. rBBat&m kod pi. rSMuitftm (of which the wioe 
may be Mid). 

Participles of the s-aorist. 

867. tu AetWe particlplat 4r« dikf«t oi dbikfst, ud ■4kfat 
(both EV.). 

b. It flUaM (above, 8M d) It to be teokoned m xn B-aoTiat fom, 
ffljM&il4 la an B-aoriat participle; and of a kindred oharact«T, appuendy, 
arA ArQctB&uA, 61iasbm, jwty mjw 4^ Hh4y fiaftn^j wmfcwrt^ inftTiftj yaniA- 
■ftii&, rnbhuAu^ vrdhaoSni, Mtliaa&iii, qavu&nA, all in RV.; with 
namoaSni, bhl74sbia. In AY. In RV. ocean aUo onoe »ihi^»m«Titt 
apparently an a-foixn of an B-aoriat of ydhi. 

S. The If-aoiiat. 

898. The tense-stem of this aoiist adds the, general 
tense-sigD H s by help of a piefized auzitiaiy voirel ^ i, 
making ^ if , to the toot, which is usually sttengthened, 
and which has the augment. 

899. The rules as to the strengthening of the root are 
as follows: 

a. A final vowel has rrddhi in the active, and ga^A in 
the middle: thus, <9m^? apftvif and Ui^JMN apavi^ from 
y^ pO cleanse; UrTlfj^ at&rlQ, act., firom v?T tr ptus; Q^lf^H 
a^aylf, mid., from y^ 9I lie. 

b. A medial vowel has gu^a, if capable of it, in both 
voices: thus, i^f^HT alefif, act. and mid., from y%ts U9 
tear; ^ji^^ arooiQ from V^ ruo ahine; Uc^f^tl avarfif 
from CoW vTf '■<»i ; but flsUNH ajlvi* from ysf^^jlv live. 

o. Medial Q a is sometimes lengthened in the active; 
but it more usually remains unchanged in both voices. 

d. The loota in the oldei langaage wblch show the lengthening are 
kao, tan, ran, stan, svan, ban, vraj, sad, mad, oar, tsar, svar, 
jval, das, tras. From ran, san, kram, vad, rak^ and aah oecnr form* 
of both kinds. From ]/tnath or nantb aie made the two atema mattdq 
and iwnTiflii^ 

i, Google 

321 Sibilant Aorist: 5. iq-AOiusT. [—808 

900> «• Of eiceptloDi ma; be noted: Yiaji hu (■■ elMwh«ia: 6S7) 
vrddbl iDsCesd of gn^n: thai, unftrjlfom; i^str Iub ostarla, sod y^ 
hsi ftQailt C>1bo «;arUt Id AT.)i "^f' KolfA Id KcHve. 

b. The root grabh or grab hu (u In fataic «tn., below, 9369, 066) 
long I instwd of i befaTo the •Ibllant; thiu, aftrabblqiDS, AgrRbiqta, 
Bgrabhi^nta. The loota In ohaagaKble f (ao-called roots in f: 84S), and 
yin[ ue iBtd by the griminirlins to do tbe ume optionally; bnt no foims 
with long I fiom meh roots have beea foand quot»blo. A Sutr* (PG3.] 
baa once uwyl^fa from i<'iii (doablleai a false loading). 

eoi. The endiags are as in the preceding Coimation 
(3^ UB and W{ ata id 3d pi.]. But in 2d and 3d aing., 
the combinations if-s and if-t aie &om the earliest period 
of the language contracted into ^tflB and ^ It. 

a. The 2d pi. mid. shonld end alwai^ in l^IiTaiii (or 1441>Tam, 
from i^^dbvam: SSS); and this is in Tact the form Id the only exam- 
ples quotftble, n&melf ajanl^bvain, arti^liTam, WtiJhijhTam, ve- 
pl^bvam; SB to the mlea of the native grammsrianB reepeDting the 
matter, see 236 a. 

902. As examples of the inflection of the i^-aorist may 
be taken the loots ^ pll cleanse, and ^ budb wake. Thus: 
actlTe. middle. 

i grni^ftm s^ri^ yqiJ^iH tiTMpjf'ii srri^9i% *44i=(^H[^ 

&pSvlqam &p&Tlfva Apftvl^ma ^pavlfi Apavlfvalil ApaTt^mabi 

a SHRtH^ MMlftfet^i^ ^fTliW ^mfirWTH^ ^^^^T^nsITH^ qnf^bH^ 

&pKTt8 ^pSviitam ApSvlfta tpavlf p*^ Apavl^fttbKm ApavlfbTam 

ApBvIt ip&Tift&m Apftvlfus Apavifta &paTlf&t&m Apavl^ata 

1 g^tfcraii^ sratftra' sRtRiKq a^fuft Msi^fu«i% yih^iHf^ 

Abodbl^am ibodbifra &bodhifma &bodbifi &bodbif vahl &bodbif mohi 
eto. etc. etc. etc. ets. etc. 

908. Tbe nnmber of roots from which forms of this aorist have 
been noted in the older language is Dearly a haodred and fifty (io 
BT., abont eighty; in AV., more than thirty, of which a dozea are 
additional to those in RV.^ tbe later texts add lees than tweoty. 
Amung tbeae are no rooCe in K; but otherwise they are of every 
variety of form (mreat la final 1 and I). Active and middle persons 
are freely maile, but sparingly from tbe same root; only about fifteen 

WbltiitT, Siammar. 3. ad. jt 

Dij.ieo, Google 


roola hAve both Mtlve mud middle formB In the older Imngiuge, and 
of these « p«rt onlj exceptionftllj In the one voice or the other. 

a. No tale appears to goTem the choice of nuge between the 
If- and the ■•aorlst; and in no small number of cases the same root 
flbows forms of both classes. 

904< TrrefQlultiei an to be notlud u follovc 

a. The Mmtruted farms {tkram&n, ograbhlm, xnd a-vadbtm (vltli 
ingm«ntle» v&dUtn) ue roqnd Id lit alng. act. 

b> For fcqarlt octms Id AY. i^arUt; &1io (in » part of the mtDiiicripti) 
<}arKla for ^ails; agrahUfain Is ronnd Id AB. (alio the monrtroDi form 
t^JagrabbJUfam : tee 801 1). AJaylt, with short i in the eodlng, occnn 

in xa. 

o. AT. hu ODce nadi^^ta, *ithoot gta}A. 

d. The formi atSrima (RV.), &vSdlran (AY.), ind bKdblthia 
(TA.), thongh they lack the aibiUnt, ara pechapi to be refeiied to thl* 
aorist: compare avlti, 008. A few elmllar caies occur Id the epics, and 
are ot like dODbtfal character: thai, jSuitliSs, mSditUo, vartithBa, 
qa&kitbia, aod (the eaiuatlTe: 1048) aghStayithiB. Agrbitbn and 
gfhlth&S and (Cpmta, If not falie readlnji far gflml-, are probabl; 
Irte^lar present-formatloDB. 

Modea of the if-aorUt. 

806. Ai DBQal, anfmentleu Indicative formi of thle aortut are more 
common than proper labJnnetiTei, Example!, of all the peraoni fonnd la 
occnr (and indadlnf all the accented woidi), are. In the actiie: qUslfam, 
v&dtalm; mitUs, tr&dhls, yavls, s^tIb; Avlt, jdrvit, m&tlilt, vUb- 
It, veqlt; mardhi^tani, doft^tam, hiAsiftam; aviif^Siii, J&niftam, 
b&dhlft&mi framtfiiia, T&dlqma; vadblffa and vadhi^tana, math- 
Iftana, blAelfta; bTSrifiiB, grahl^ns; — in tlie middle: rKdhlfii 
J4nlftli&8, marflfthfts, Tyatblftbfta; kr&mlffa, J&niffa, paTl^ftai 
pritbiffa, m&iidi;ta; Tyathifniahi, The aeoejit la on the root-iyilable 
(tSrl;6B, AT. oDce, ii donbtlesi an errot). 

806. a. Of iDbJnnctlTe forms with primary endlagi aecnr onlr the 
Ist ling. act. davlf&^i, and the lit pi. mid. (with nnsttengthened e) 
y&dflmabe and Bani^lmabe. 

b. Formi with seeondarr endlagi are almost limited to 2d and 3d 
■ing. act. There are fonnd : avl^aa, kinl^as, t&ri^«B, rak^^aa, T&db- 
Iqaa, Titdl^ae, viqifaa, gaAelqaa; karl^at, Jambtd^at, Jd^l^at, 
takflqat, t&rifat, D{odifat, parliat, b^dhlqat, mirdbl^at, y&oifat, 
yodhlqat, rakfi^t, vaniqat, yyathi^at, ^afialqat, sanifat, sKvlqat. 
They are made, it will be notkeil, with entlie regalarlty, by adding a to the 
tenie-etem lu if before the endings. The only other peraoni fonnd to oocnt 
ai« the 3d pL ac:t. aanifan and mid. B&nifanta (and T8. hai vanlfanta. 

ioy Google 

323 Sibilant Aorist: 5. I^aorist. [—©II 

foi the pToblematio Taanfaata of RT.), irblcb are alio TegvUr. BhRvlfSt 
(AB. onae) !■ m iciliUry siampU of a taim with doable mode-ilgD; o&ni^ 
fbat (KT. ; BY. iottaad J&nif (bat) leen* hopeleealy coinipt. Tbe ladlul 
■yllable tlwayi liu the Accent, and ita rowel aeoally accord* with that of 
the iQdlrittTa; bat ve bare man- in the labJonctlTB agatnit aB&nlfam 
(as to oay- and ran-, lee below, BOS). 

907. The middle optative of this aoriat alio fonna a pait of the ac- 
cepted "preeattTa" of tbe later Ungnage (OSS, 926 b). It la veiy rare at 
all periodi, being made In RT. flKim only Ave lOota, and In AV. from two 
of the «ame and tiom three additional one* (ili of the eight have other 
Iif-forma); and the remaining leila add, to far ae noticed, only fonr other 
roots. All tbe forms found to occnr are as foUowa: Jftnl^ya, iDdhiflya, 
sdhifi7&, ruoli^R and rooIflTa, gmi^lyKi modl^thBs; Janiql^tft; 
Tuni^ts; Bahiflvahl; Idhifunahj, edbl^Im&hJ, Jani^Imahi, tSri^I- 
mahJ, mandl;lmahl, Tandl^lm&hi, vardhl;Im4hl, aabi^Imatal and 
afihiflm&hl. The aoeent la on tbe ending, and thia would lead na to ex- 
pect a weah form of root throagboot; bnt the nsage in this respect appears 
to be TSTioiu, and the oases are too few to allow of aetttng up any rule. 
The forms Juiifeyua and -70, ffom a secondary ft-stem, occnr in E. 

908. Of imperative forms, vb have from y^v a series: namely, 
avl^^hl, avif^, avlft&ill, avitA (if this, as seems probable, staads 
anomalonaly for aTlf{&) and avl^f^bia; two of theae ate of unmistakably 
ImperatiTe form. Other forma occnr only In 2d dn. and 2d pi., and are 
accordingly snch u might also be aabJnnctiTea used imperatlTely (which 
is further made probable for two of (hem by their accentnatiDn on the 
loot-Byllable) : they are kramiqtAia, gaml^fam, oanlq^m, oayi^t"')' 
(agtinat acftTii^ain), titriftam, rodhl;tam, vadblq^am, 9nathl jfHP ; 
riL^iffana (againgt arS^ifUB), qnathiftitna. 

909. Ko words baviDg a participial eDdiog after if are found 
auTwbere to occur. 

910. This is the only aorist of whicli formB are made io the 
teoondaiy and denomlDatlve conjugations: see below, 103B, 1048, 

6. The Blf-oorlBt. 
911. According to the giammaiiaDa, this aorist is made 
from loots in ^ S [including j^ mi^, ^mi [01 ml] damage 
and ^ IT cling, which substitute Eorms in B), and from 
•IIT nam iow, Tra7am reach, and ^l? ram he content, and is 
used only in the active; the oonesponding middle being of 
tbe a-foim (878 ff.) . Its inflection is precisely like that of 
the i^-aorist; it is unaecessary, then, to give more than 


ioy Google 

911—] XI. AORIST-STSTEHS. 324 

ita first persons, which we may form from the roots TJ\ yS 
ffo and sm nam bote. Thus: 

■- i. p. •. a. p. 

AyKslfom ArOeifva &7fiai;ma AuufaBi^Am inaiheifTa AnadkHtftna 
etc. etc. etc. etc. et«. etc. 

91fl. The sif-MTlit U propeilj only * sub-foim of the if-wiUt, 
bavtDg th« tenae-iiso uid ending* or the Utter «dded to a fonn of loot 
increued b; vi added s. It la of extreme tully In the older Ungutge, 
being made in RY. only from the roots ga ting and jft go, and In A.V. 
only from hft Itava, *nd doabtleM alio from pyi/Ul up »nd van wm 
(«ee below, 914b)i the remaining older teiu *dd JfiS inov (B.), jjft ocer^ 
pote»r, dbji think (^B. once; the edition reads -dhft-), and ram be con- 
ttnt (ST.; a bad Tarlant for BV. riaiya); other Biahmana forms wbloh 
migbt be also of the s-aorist are adrftalt. avfi^t, and ahvaait; and bhnlc 
fllflTa (FB. S.} must bo regarded as an anomaloni tonaation from yhhVij, 
nnless we prefer to admit a secondary root bbuk;, like bliakf from bhiO- 
In th« later langnsge have been fonnd quotsble from other roots only glSale, 
adtamSait, anaifaelt, apsalt, mUsIs, and amnKslf us. 

a. Tbe participle hihaiiiKiia and caDsatlTe hftsayanti (KY.) show 
that li&S had assumed, even at a lery early period, tbe valne of a secon- 
dary root beside bA for other forms than tbe aotlst. 

013. The whole lerles of older IndlestiTe forma (omitting, as donbt- 
fnl, the 2d and dd alog.) Is as follows: agialfam, aj&&8tfam, aria- 
tiam, adhTialfam; ^rBaiftUn, ayaai^tom; aja&siqma; aJfiSalfta, 
Ayfiaifta; agfisl^iu, ayBalfOB (Kkfifaa is from y^ak; attain). 

a. Forms wlthoat aagment ere Uiese; JfiSalfam, radtsi^am, hBsl- 
;am; hOai^m; hisl^timi baalffa; hftaiqua, gfisl^ns, JUal^oa. 
The accent would doubtless be npon the root-syllable. 

914. a. Of proper subJonctlTes are fonnd two, gdstfst and yftaii}at 
(both KV.). 

b. OptatlTOS are not less rare : namely, rlsiBl^thaa and py&8i|Fimalil 
(tor which the AY. mamucrlpts read pya^lfTmalil, sltered in the edition 
to pyfiylf-); and doubtless vaAqlflya (AY., twice) is to be corrected to 
vaAfli^a, and belongs hare. As to bhnkql^Iya, see aboTe, 912. 

O, The accent of yaaJqt&m (like aTi^t^m, 908) ikow* it to be a 
ttne ImperatiTO formt and yfial^fa (BY., once) is doubtless the same, with 
anomalous t for L 

915. Middle form* of this sorist, it will be noticed, oocnr from the 
optative only; but, considering the great rarity of the whole formation, wa 
are hardly Juitilled in concluding that In the ancient language the middle 
persons in -sl^ -BlqtbfiSi etc., were not allowable, like those in -ifl, 
-IfthftB, and the othor* of the Iq-sotltt 

ioy Google 

325 SiBILAMT AORIST: 1. BVA0RI8T. [—919 

7. The sa-aoriflt. 

916. In the latei language, the toots allowed to foim 
this soiist end in 31 9, ^ 9, or ^ h — all of them sounda 
which in combination with the tense-sign make vf kp; and 
they have ^ i, 3 u, ot ^ ^ aa radical vowel. 

a> They &ie- ta followB: dig, rif, ilq, vlq, kUq, kruf, ruf, mff, 
spfq; tvlf, dvi^, glif, t1^ krf ; dih, mlh, llh, gUh, duh, nlh, tph, 
v^h. Btfb; fiom (bout faiir al them aa-foinK, aiilier or lalar, ub quotibla. 
Some of them may, or with ceitiln memingi most, tike loiitt* of olhar fonnf. 
And B few are allowed to drop both tenae-sign and nnlon-vowel a la cei- 
talu persons of ths middle; that Is, they may make Instead forma of the 

917. As the tease-atem ends in Q a, the inflection is 
in the main like that of an imperfect of the second general 
conjugation. But (according to the grammaiiaDs: the fotms 
unfortunately have not been found quotable} the Ist sing. 
mid. ends in ^ i instead of 1^ e, and the 2d and 3d du. 
mid. in ^g(M SthBm and ^TrTR St&m, as in imperfects of 
the othei conjugation. Both active and middle inflection 
is admitted. The root is throughout unstiengthened. 

816. As example of inflection we may take the root 
1^ dl9 point. Thus : 

actlTe. middle. 

idik^ua &dikqftva idlk^Kma fcdlkql &dikfSvahl Adlk^SmaM 

wi^irtH^ 5l^irtf<*i^ flf^iHH Mi^eriaif)^ Mf^miR^ Mi^rfUH^ 
MUtfas &dlkfatam 4dlk(ats fcdlk^thSa &dlkfBtliKm Adik^adhvam 

idikqat &dlkfat&m' &dlk;an Adlk^ata Adlk^UBm idik^anta 
819. In the euUai laogaage, the forma of the Mt-aorlat are hardly 
nne than apondle. They are made in BV. froia Mien roota; iu AT., 
from two of these >nd from two otfaen; *nd the ramalning tnta add ten 
more, making nlnetceo in all (the Ikler laognsge makea do addHlons to 
this ntunbei). A* later, all have i or U 01 p u rMt-nwel, lud a final 
ant which oombiDeB with s to ks; bnt tbwe ate in the lilt alio two 


91S — ] XI. AOEIST-SYBTEHS. 326' 

ending Id J, nimely mf] uid t^J. All the ezamplw noted tie ginn 

A. So fiT u the middle ronni ue conoeined, thlt loilit would be fnlljr 
eipUia»d 4» * tnntfei of ceit«in ■•wtlets to an a-inllection. The miiked 
difference In the itiangth ol ledlcel TOwel Id t)>e letlTe, howeier, ittndi 
in the way of the incceHfuI «ppUc*tioii of ineli an tiplanation to Ilia aoll«a 

820. A. In the IndleatiTe, we find, In the actlre : av^k^am; mimfcqiiiti 
adhuk^aa, anilqkaB, aknik^aa, aeprk^aa (and HBh. adda amrkfM); 
odlkfat, amlk^at, Kllk^at, avik^t, &krnk^t, aghuk^t, adukfat 
and Adhuk^at, druk^t, av^kfat, akfk;at, 4mfkfat, 4apfk^t; 
asbukqatbn; arnkflma, amrkftma, avrkfftma; ftdhak^an, aplk- 
^ait [v'pl;)i nrukfao, aappk^an; — in the middle, only ak^k^atliU 
(ykf^), Adhukqata, atid amfkfaiita (and UBh. adda om^kfataP). 

b. FDrma without anient (no tcne «ntijunetiie« ocenr] ate, in the 
aGtiTe: drlcfam, mpk^am; dnk^as, ruk^aa, mrkgaa; dvlkfat; 
m^k^ta; dhtikf&n and duk^ia; — In the middle, dvik^ata, dakfata 
and dhlUcfata, dbukq&iita. 

o. Tbeie are no optative forma. 

d. ImpsratiTe are: in the acttte, m^k^tam; in the middle, dbob* 


e. The few aceented lonna witbont angmeot wUeh oeonr haie the 
tone on the tenae-eigD B^ In analoey with the a-aorlat (2) and the Imper- 
fect of the &-claH: a alngle axceptloD la dbdkfata, which probabiy needa 
emendatioD to dhnkf&ta. 

f. The aapiratlon of Initial d and g, after loat of the aspirated qnalitf 
of the root-flnal (16B), la aeen tii forma from the roola duh and gub, bnt 
not bom drub (eal; a ilagle case, AB.); BV., howerec, hu alto adokfttt 
and dnkfas, dnkf&n, dukqata. 


981. Ae the Bo-oalled precative ia allowed b^ the grammarians 
to be made in the later lan^oge from eveiy root, and in an inde- 
peDdeDt way, without reference to the mode of formation of the 
aorist from the same root, it is desirable to pnt together hare a brief 
statement of the rules glTen for it. 

022. The precative aotire is ma^le b^ adding tbe active 
pieoative eodingo (above, 668) diieotly to the root. But: 

a. Of final rool-7owela (aa before the paMlvB-algn yi: 770), I and 
U are lengthened; j la nanally changed to ri, but to Sr and Gr in thoao 
loota whieh elsewhere ihow ir- and ur- forms (so-oalled f-ioota : 242}, and 
to ar in T ""d smf, R 1» changed to « In tha roota di, dbK, atb*, p» 
drink, gft ting, and a few others, in part optlonaUy. 

ioy Google 

327 PBBCATiYa. [— M4 

b. Tbe loot In geneikl utnmei Iti weiAeit form : > pennltliDKU d*m1 
i< loit, u fa badbyBaun from ybKadb; the loou Khieh ue kbbieTUted 
In the whIe penoDi of tho perfect (TM) hare the Hme ibbiaviitioD here, 
u lo neytanin, JJrBsBLin, vidhySMum, tnirylniin. grbyftwun; V9KB 
foimi i^yiawn (compue 68S, 8(Mo); uid to dd, 

O. It ku been pointed out tbOTe (887) that the aetlTe preoadTO U *ii 
Dptatlre of the n>ot-ioriit, irlth > problemttic Intertloii of > ■Iblltnt betmen 
mode-iign and endlos. 

928. a. llie pieoative middle is made by adding the 
middle precatire endingB (abore, 668) to the loot inoTeaaed 
by n s or ^ if — that is, to the tense-etem of an B-aorist 
01 of an if-aoiiat (but without augment]. 

b. The root is 8tteI^;thened accoidiug to the rules that 
apply in forming the middle-stem of the b and of the if- 
aorists reepeotirely : in general, namely, a final vowel is 
gunated in both formations; but a medial vowel, only be- 
fore ^4- 

o. A* wu pointed out above (667) the middle preoatife ii letUy the 
t^tatlTo of oeitalu aotiiti, with the inutiion ot a itbllaol between mode- 
tiga and endlnf onl; (*0 far »« anthenticaled by use) in the 2d and 3d 
■tngnlar. Id tlie older laD^iugo, lueh fotmi are alteneit mado Aom the 
«-aorlat (886) and the Iq-aoilat (807); but alio from the lool-aorlat (837 b), 
the R-aoriEt (860 a), the redopUoatad aorlit (870), and the slf-aoilat 
(014b); and even Itom the peiToet (SlSb). 

924, As example of inflection, we may take the root 
H bbtt he, which is said (no middle aorist or piecative from 
it is quotable] to form its middle on the ff-stem. Thus: 

bhOrlUjuD bhOyisva bbOy&ima 
bbfiyfa bbOy&atam bbflyista 

bhilTAt bhOyJetam bbttf^UB 

ioy Google 


bbnvl4ly& bhavl^vUil bhKTi^Im&hi 

bhATl^ftb'^ bhavlflyltstbKta blxavifl^liv&m 

bhavl;If(& bhsvlfiylUtam bhavlfirfcn 

ft. Tbe foimi given by the gmnmailBDi u 2i &nd 3d dual are of 

very qnestioDable value, aa rsfardi tbe plaee ueigned to the ilbllant 

Tboia pMionB, and the 2d pi., have uevei been met with Id nae. Far the 

qautlOD teapactlQg the ending of the 2d pi., as dbvom oi i^v'am, we 


826. ft. The piacatlve active is a foim of varjr rare oeciut«iee to (he 
elatilcal laagaage. In e»«h of the teiti already more than once Teterred to 
(Hanu, Nala, Bhagavad-QTta, (akontali, Hltopade9a) It oecort once and no 
moie, and not half-a-doten torm* have been foood qnetable traoi the epia. 
Ai to Ita value, see 673 o. 

b. The pieeative middle is ylrtnally nnkoown in tbe whale litei 
lit«t>tare, not a single oecmreDce of It having been bioaght to lleht. Tbe 
BhP. his onoe rirlfl^fft, whloh is also a BV. form, belonglDg probably to 
the rednpllcated loiist: see 870. 

Uses Of the Aorist. 

9Se. The uses of tba aoTiBt mode-forms [fts bu been ftlreftdj 
pointed out: 68S) appear to accord with those of tbe mode-formB of 
the preBeDt-BfBteiD. The predilection of the earlier language, cod' 
tinaed sparingly in the later, for the augmentlees forme in prohibitive 
expression afier rai was aufficiently stated and illiutrated above 

ft. The tense-value o( the aorlil indicatiTe haa alio been more than 
once referred to, and oall* only for somewhat more of detail and for illut- 
tratlon here. 

927. The aoiist of the later languag^e is simply a pret- 
erit, equivalent to the imperfect and perfect, asd frequently 
codrdinated Tvith them. 

a. Tbui, tato^ aa gardabbaib lagu^eua tA^arSin&Bft; ten& 
'bBu pafioatvftm agamftt (H.) tk^M^on Ae beat tlu donkey with a ttid ; 
and hereof the latter died; tata^ aS vidarbhUa agamat puna^; tSih 
tn baiidhiijaiia^ aamapl^ayat (UBb,] thereupon the toent back le 
Vidariha; and her kindred paidher reoerenee; prltlm&n abbQl, uvAOft 

ioy Google 

329 UsBa OF the Aorist. [—889 

oU *Dun (HBh.) he watJilUd ui'tA affection, and laid to him; tarn, ftda- 
bat kftfthiil^ so 'bliad dlvyavapas tftdft [B.) As bunted him with 
aood, and he became then a heaeeniy form. 

928. The aorist of the older Unguftge hu the valae of a proper 
"perfect": that Is, it elgnifiee somethiDg past which is viewed aa 
completed with reference to the preBent; and it lequirea accordlDgly 
to be rendered hy our tense made with the anxiliary have. In general, 
it iadicateB what has just taken place; and oftenest Bomelhing which 
the speaker has experienced. 

a. Eiunples bom the Tsd> are: p&ri 'mk gtia anefata p&ry 
agnim ab^f ata, deT^qv akrata qr&va^ k& ImiA i dadharfatl (KV.) 
iheee here haoi ted lAout a eoic, they have carried around the Jire, they 
have done honor tc the godi — who thall venture anything againtt ihmtf 
y&m siob&ma iiiiiiiaa& ad 'y&m it 'g&t (Etv.) he ichom we (formerb/, 
Impf) taught teiih our mind hat {now, aoi.) eome; y^iia "ndro havifft 
kptvy Abhavad dTOiniiy ilttaiii&^ Id&ifa t&d akii derS aaapatni]^ 
klli'bbavain(RT.) that libation by which Indra, making it, became {iaft.) 
of highett glory, I have now made, yegode; I hone become fi-eefiom ettemie*. 

b. Exunplei ttOBi the Briihnii^a luigaage ire: oa hft 'amlfl J76K 
UT&ea... t&to ha gandbarrt^ a&m ddira: Jydg va iy&m urv&^i 
manii^dfV avfttalt (^B.) the lived with him a long time. Then the 
Oandharvtu taid to one another, "thie Vrvaci, foreooth, hat dwelt a long 
timt among mortah"; taaya ba dantft^ pedlra: taifa hoVBca; apat- 
aata vft aaya daut&^ 1^3.) hie teeth fell out. Be taid to him: "hit teeth 
truly have fallen out"; Indraaya TTtr&lfa Joghn^a indriyiuh viry&ili 
pfthlvfm inu vy krebat t&d d^adbayo virudho >bbavan ti. 
prajApatlm ap& 'dbftvad v^tiiih me jaghnoqa indrly&ih viryhih 
pfthlTim &11U vy irat tkA &;adhayo virddbo 'bbuvann fti (TS.) 
of Indra, when he had ilain Vritra, the force and might went away into the 
earth, and became the herhi and plante; he ran to Prajapati, toying: "my 
force and might, after tlaying Vritra, have gone aaay into the earth, and 
have become the herbt and plante"; BTay&m enam abbynd^tya brilyftd 
vrd^a kvi 'vS.t^ify (A.V., in proao passage) going up to him in person, 
let himeay: "yratya, wh«reha*t thou ahod^f yk& idjtnrdi dTftuvlv4da- 
m&n&T eyitSm ab&m adar9ain ab&m aqr&u^am ftl jk evb, brOylid 
ab4m adarqam Itl t&smft avA fraddadbySma ((B.) if now two thould 
come dieputing with one another, [tt« one'] toying "I have teen", [the other\ 
"I have heard", we thould believe the one who taid "I haee tetn". 

929, a. Thli dlslinctlon of the aoiiet bom tbe Impeifect and perfect 
M tenses ot nsnatlon is Tsry common in the Brahmana langasge (iucladlng 
tlie older Upaniahads and the Siitns), and is closely obterved; violation at 
it is Ter; rare, and is to be regarded a* either due to corruption ot text or 
IndlcatlTe of a late origin. 

b. Id the Yedlc hymns, the same distinction is prsTalent, hot it both 
less clear and lets strictly maintslned) msny pttsagea would admit an 

Dij.ieo, Google 

9S0— ] XII. FuTUBE-aTSTEiia. 330 

iDtarpreUtlDD Implying elth«t teuo; tind erldAUt Miiit-fonu ue •omMlmte 
Died nutUiiely, vhlle Imperftiat-fonni ue also aceuion4llr employsd Id 
tbe wriit tense. 

880. The baaodaTy betireen wh&t hu Jnit been >nd what li Ifl ut 
BTineMtot ooe, and U ■omvUmei oieiitapp«d, m that an aortic ippaan 
when a piaieiit migbt lUnd, ot was eTsn ratbar to be eipected. Thm: 
■rKsoBtho bhavfttam indave it& Itl eomo vil rlUa 'ndo^ aomO^ 
yU 'vU 'no etsd rlJBa iaada -olklpKt (AB. I. 18. 7] "£« ye comfof ■ 
tahU teati for oar Indv", ht tm/i ; Inda it kmff Soma ! bg this nuotw ha 
hat mada thtm (iiiitaad of nuiAwi thfm) laitabhfor king Some to sit upon; 
Tftm^ii ipo y&d adbhfr abhi^c&ti vkra^Mn ovBi 'nun akar 
(HB. It. 3. 10) the watert art Varuna'i; in thai ha bepourt him with teaUn, 
he hat mada him Vanma; pa&o&bhir vyigbftrayati p^fikto yi^lL6 
yliTkii «T& yajfiia Um ilabdfai 'tho yivftm *vk yajfiia tAamU 
rikfUay &pahanti (MS. Ul. 2. 6) A« (nuor* v>Hh Jive i _fivefoU it Ih* ogtr- 
itig; at great at it ^t offering, of it ht hat {ihtrehjf\ taken holdi then, a* 
gnat at it iht offeriiig, from it ha tmitet awm/ the demont. Tbi* Idiom ta 
met vltb In all tbe Bnbnufai) bat it ii eapeolalty fre^oeat Id th« HS. 


981. The verb has two futures, of veiy diffeieot age 
and character. The one has for tease-sign a sibilant folloired 
by Tl y*, and is an inheritance from the time ot Indo- 
Euiopean unity. The other is a petiphrastio formation, made 
by appending an auxiliary verb to a derivative noun ot 
agency, and it is a recent addition to the verb-system; i(a 
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. 

ioy Google 

331 The a-FUTORE. [—984 

I. The B-future. 
982. The tense-sigQ of this future is the syllable ^ aji, 
added to the^joot^either_diiBcUy or by an auxiliary vowel 
^ i (in the latter case becoming ^iqr iVJi]- The root has 
the gu^a^tiengthening. Thus, fiom y^ dft give is formed 
the futuie tense-stem ^THT 6t»jit; horn y^ 1 go, the stem 
JffHI e^yi; fiom kJ^ du^ mtlt, the stem ^t^ dhokn^; 
from yil bbn be, the stem Hf^UJ bh&Tifyi ; from y'^ rdh 
MrtDe, the stem crf^TCZr ardhif^i; and so on. 

a. But fiom yflv live the stetu is JIvlfyii, from y'okq *prinkl« it 
is nk^loTi, and so oo (940). 

b. There *re budly toy Tedlc cuea o( raMlutlon o[ the tengg-slgn 
B;a into sia) BV. bu k^e^Uoitas onoe. 

883. This tense-Stem is then inflected precisely like a 
piesent-stem ending in !| a (second general conjugation: 
788 a). We may take as models of infleotiou the future of 
y^ dft give, and that of ym Iq^ make. Thus: 
*cdf«. mlddla, 

dsayiml dfisy^vas diayiimaa dAsyi dftayttvabe daay&naho 
dlay&ai dAayitbas daay&tha dOsyiaa dlajAtlie dBsy&dbve 
dOsyiU dssyitas dieyintl dBay&ta dBay^te dftay&nte 

1 ^ifi^n#T ^lliuiwH^ *!("yjinH^ ■*i|w) ^f^KtiH^ *!^aoiM^ 

barifjiml karif yivas katiqyiunas karl;yA karlisriTabo karlf ytmalie 
•to. etc. etc. eto. etc. etc. 

a. Id tbe tplci >re found occuional cieei of Ist du. and pi. In va and 
ma: e. g. radiByftTa (R.), bhak^yftTa tunutite: HBh.); sqyama 
(UBb.), Tatayama (R.]. 

984. With regard to the use or non-nse of the aoailiarf vowal 
i t>efore the Bibilant, there is a degree of general aooordanee betweeo 
this tense and the other future and the desideratire ; bnt it is by oo 
means absolute, nor are any definite roles to be laid down with re- 
gard to it (and so much the less, becanse of the iafrequeDcy of the 
two latter fonnations in actual use): between this and the aorist 

itizecy Google 

884—] XII. FurnRE-BYSTBus. 332 

(■••oriit on tfae one Bide, or if-aoTist on the other), toy coTretpondeDce 
is Bttll less traceable. Practically, it ii neceeBary to lean, u a nut- 
ter of UBage, how any given root makea theae varioas parte of its 
coDJogational Bjetem. 

08K. Belov ia added & lUtomeDt of the auga, u legudg the anzUiarr 
Towel, of &I1 the loots foaod qnotible — foi the moit put, Id the form of 
■ ipeulflMtion of thoie which utd the tense-ilxn dlreotly to the loot; io 
bnrketB >te rurther menttoned the other tooEb which tec«rd)ng to the gnm- 
muluu ilso lanice the •uxlllary Towel. 

tt. Of TODtB ending In rowelt, tha great mtjotitij (excepting those fa 
j) ttke no 1. Thne, all in & (nDmeroae, and DOnaceuary to sped^: bnt 
oompare o betow); — thoae In 1, as kfl poit4t, ol g^h*r, ol note, ml, b1 
or B& fti'fuf (alfya), hi; from 1, kfi dtttroy, and Ji occur forms of both 
elMsei; qrl [and <jvi.] has 1; — those In I, as krf, bhli ml, t1i; bat fi 
tit and nl have both forms [and ^I takea 1]; — those in n, aa oyn, dm, 
pin, 9ra, ha; bnt au prui out and ata hare both forma [and kfU, 
kf^O. no. yn, ro. anu Uka i]i~of those in a, dhil and bhu take i; 
bQ hu both forme. Bat all In p (namerona, and nnneeesaar; to apeclfy) 
take 1 [those In ohsngeable f, oi ao-called f-roots (24S), are aald br the 
grammartaea to take either 1 or I; no I-torma, however, are qnotable]. 

b. or roots ending in mntes, about half add the tenae-algn dttectlr. 
Thus, of roots ending in gutturals, q(tk; — In palaUls; in O, pao, mac, 
rlo, vao, vie, vraqo, aio (but y&o takes 1); in eh, praeh; in J, bhofi}, 
mrJ (m&rkfya and mrakfya), jtti, bhuj, yuj, ttJ, erJ [alao bhr^, 
rafij, aaiy, arattj, nij, ruj], white (yaj, bh^, and ma,U (mailkfTa and 
m^lJifya) hare both forma, and vij (viji^ya and vajifya) and vr^ 
lake 1; — in dentals: in t, kpt eat and vrt [»'*" o^t and nft] make 
both forma ; In d, ad, pad, qad fall, akand, syand, ohid, bhld, wld 
^nd, nad [also had, khld, avid, k^ad, tod]; white aftd (aataya and 
■idlfja) and vid knotc make both forma [also ohfd and tfd], and vttd 
has i) in dh, wyadh (wetaya), rfidh, aldb tuccttd, budh, yadh, radh, 
TTdh Talao aftdh, kradh, k^adh, fodh], and bandh and eidb rtpil 
have both foimi; In n, ton, while man and ban have both forma; — in 
UMala: In p, Sp, k;ip, gnp, tpp, afp (arapaya and aarpsya) {alto 
9ap, lip, lap], while tap, vap, avap, dfp, and k)p ba*e both forma; 
In bh, yabh and rabh, labh having both forme ; In m, ram, while kram, 
fcfam, nam, and yam make both forma. 

O. Of the roots reckoned by tha grammarians as ending in teml*ewel* 
(761 d-g) all Uk« 1. And wS oi vl wtme, vy& or Ti enveiop, and liTi 
or bQ cal4 take a y-form, aa In theli pteaent-syatem, to which then i Is added: 
thus, wayl^ya, vyayi^ya, hvayi;ya (but also hvaaya). 

d. Of root! ending in spirants, the minority (about a third) are with- 
out the auxiliary »owel. They are: roota in q, iiq, wlQ, dji) (drakfya], 
api^ (aprakfya) [also daAf, rig, lif, kruq, mrq], while na^ ba hit 
has both forms (na&kqiya and naqifya); — tn 4, pi;, TI9, ^if [alio 


333 The s-FnrDRB. [— 9S8 

trlf. dTi(, QUf, tU9> Axuh pn;, qof], wbile k^ bu both formi (kmk- 
fya aod karftfya); — In ■, vas (Airw, vas c^[><A« [lUa ghats], while vaa 
daeU hM both lorms; — tn h, mih, doll, drnh [alio oaJi, dlh, lib], 
irbile dab, vah, aah uti rah b>Ta both fonns. 

•. In the older language, > mijotlty (aboat Ave nintha) of simple nuts 
add the aya wltboat auitliur i; of tbe fntarea occnTrin; In the later 
laognafe only, nearly thiee quarteii have the 1, tbia being fenetally taken 
by any loot of late origin and derivailTe charietu — aa It li alao nnlfonnly 
taken in (ecoodaty conjngatlon [1019, 1086, 1060, 1068). 

886. Aa ti>« root is Etrengtbened to form the item of thle future, ao, 
of a root that baa a itroDger and a weaker foim, the stronger form la naed : 
thus, frem vlwndh or badh bind, bhantSTa or bandbl^a. 

a. By an Irregnlst atrengthening, naUfya (beside na^lfya) is made 
from ^Da9 he loit, and mafikfya (berida majjifya) from ymajj sink. 

b. Bat a few roots make fatare-atems In the later language wltboat 
atrengthening: thus, UUiifya, mUioya {also TS.)> vljl^a (also vejlfya), 
aifya (KsS or b1), Mfya (988 b), aphnt^Vra; and f^adb make* 
VAtaya from the weaker form vldh. 

o. The (B. hae once tbe monstToni form afnavlfytinabe, made 
apon the preaani-atem af ml (607) of y'a^} attain. And the later language 
makes i^dlfya and Jabifya from the preaent^tema of yaaA and yb&. 
Compare farther hTayiffa etc, B36 e. Alto kfayfiyi^ya from ykhyfl 
(beside kbyOsya) appears to be of similar obsracter. 

d. A number of roots with medial ^ strengthen it to ra (S41): tbaa, 
krak^jra, trapeya, drapsya, drnk^a, mrakqya (beside mKrkfya), 
sprakfya, arak^a, arapaya (beside aarpsya), and mradlfya (beside 
mardlfya); and y'k^p forms klapaya (beside kalpifya). 

ft. The root grab (also Its doublet glah) takes I InatMd of 1, as It 
doea also in the aori*t and elsewhere. 

987, This fntare Is eomparaUiely rare Id the oldoat langnage — in 
part, apparently, becaase the uses of a fntaie sre to a large extent answered 
by sobJunctiTe forms — bnt becomes more and more common later. Thns, 
the RV. haa only seyenteen ocenrrencea of personal forms, bom nine diffneni 
roots (with participles from six additional roots); the AV. haa flfty oconrreneea, 
from twenty-Hie roots (with participles from soTen more); bnt the TS. has 
occurrences (peraoaal forms and partiolplea together) from orer sixty roots; 
and forms from more than a hundred and Qfty roots are quotable from the 

Hodee of the s-fatore. 

B38. Hode-forms of the tutore are of the utmost rarity. The only 
example In the older language is karl^ia, 2d dug. snbj. act., oecnirtng 
once (or t«lr«) in RT. (AB. has once notayftTShtl, and GB. has efyft- 
puthii, taAsybnabU, sthfiayKmahSl, bnt they ue ionbaen fdae 

Digitizecy Google 

938 — ] XII. FrrnKE-arBTEMs. 334 

TMdlags for -he. Tvo or three opiaihe feioii >r« found In tbe epics: thni, 
dhsk^yet tai maAsyeran (MBb.), and drskqjata (K)j ileo >n Imper- 
ative patB7antV (Bai.). And aeTeial 2d pi. mid. In dhTnni ne quotable 
from t1i« eplca: thus, vetayadhTun, savl^yadliTam, uid (the caiuatlTe) 
kUs^qyadhvam (PB.) and Jivaylqyndhvam [M6h.: and ona text has 
mokgyadhvftm at 1. 133. 13, wheie the othei leidi mokfafftdhvnm], 
and bhRTigyadbvam [HBb. B.): it li a matter of qneatlon whether these 
are to be accounted ■ real tmperatWe formation, or an epic aabatitntion ot 
Mcondary for primarr endfngt (compare B4S a). 

FartioiplflB of the s-fatare. 

939. Faiticiples are made from the futuie-atem precisely 
as from a piesent-stem in ^ a: namely, by adding in the 
active the ending fT nt, in the middle the ending ITH mKoa; 
the accent remains upon the stem. Thus, from the yeibs 
instanced above, 3.inJM dfisy&nt and <^|'HIHR d&syAmftna, 
^if^mn kari^yAnt and ^i^HIQI karify&ml^a, 

a. According lo the grammarlani, the feminine of the active participle 
li nade either In &ntl 01 la vtS; bat enty tbe former bai been noted ai 
OCCDtrlng In the older langnagp, and the latter ia avenrwhere extremely 
rare : aee above, 449 e, f . 

b. In RT. oreura once sflfyanti, ftom yoU, with inomalon* accent- 

Preterit of the B-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 present- 
stem in Q a.- This preterit is called the conditional. 

a. It atanda related to the fotnie, in form and meaning, u the French 
conditional auraii to the future aural, or as tbe English would havt to 
Kill have — nearly as the Oeiman Kitrd» haben lo teerdt Aoftm. 

b. Thus, from the roots already instanced: 

active. middle. 

i. d. p. s. d. p. 

idSayam idftayavti idSsyftma Adfiaya AdasyftTahi &<UayKmaM 

ioy Google 

335 Tub Comditiohal. [—848 

3 S<IVUV1^ Si^lHJHM^ tl<ll-Ua «<IfUe)IH^ 5RTF&EIFT^ WiITWIIi^ 
kdB»j»B iAaajAtaxa fcdbyftta Mbyatbfts AdAsyetliftni idftsymdhvatn 

idioyat idMyatfim &diayan 4dftayata &dAByetKm ^dftoyanta 

) fl*f(mH w+ft^uM «4i[|iwjm a^f^ M+fi'wjwr^ q^if^wjwl^ 

ikaxi^jaxa Utarlfylva Akarl^fina iklrl^ye AkarlfyftTahl Ucari^yftinRhi 
etc. etc. etc. etc. etc. eto. 

841, Tbs conditional i* tbe nreit of ill the ronni of the Stnikrit 
Teib. The BT. hat bat > tingle eiusple, ibltari^yat was going to carry 
off, and none of the Tedio texte rnrniihei another. In the Brahmanaa it 
la haidly more common — eioept In ^., where It U met with more than 
BRy tlmei. Not does It, like the fntore, become more freqacut Utei: not 
•n example oeetira In Nala, BhagaTad-Oiti, or Hitopade9>; only one In 
Mann; and two In (akontala. In the whole HBh. (Boltimann) It ta tovnd 
aboat twenty-llTe time*, from thirteen T00t«. The middle forms ate ei- 
liemely few. 

II. The Periphrastic Future, 

Ma. ft. This formation contaioB 0DI7 a single indicative 
active tenae (or also middle: see 647), without modes, or 
paitioiple, ot preterit. 

b. It consists in a derivative nomen ageniu), 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. 

948. The noun is formed by the suffix r{ tf (or rT^ 
tar); and this (as in its other than verbal uses: see 1182 
is added to the root either directly or with a preceding 
auxiliary vowel ^ i, the root itself being strengthened by 
gnna, but the accent resting 00 the suffix: thus, ^TrT d&t^ 
from v^ d§ give ; ^ kartf &om v^ ky make ; Hf^ bhavitj 
from y>T bha be. 

a. Ai regarda the pteaenee or absence ot <he vowel 1, the naage la 
said by the gtammiilaQs to be generally the same as in the a-futnre from 
the game root (above, 98B). The moat ImporUnt ei<-eplion la that the 
root* in j take no 1: thae, kart^ (against kariqya); toota ban and gam. 
■how the lame diSetenee; while vft, VTdb, and ayand have 1 here, thnngh 

itizecy Google 

943—] XII. FtrrOKB-BTBTEHB. 336 

not in tbe B-fotare. Tlie fsw fono* irhlch oecDt Itl the oldei li]kg«M« 
agree with thue italementi. 

944. In the third persons, the nom. masc. of the noun, 
in the three numbers respectively (S73j, is used without 
auxiliary: thus, HfllfTT bhavitS he or she or it mil be; 
Hferrft bliftTltarftu both will he; >rikfnpi bbBvitiraa they 
will be. In the other persons, the first and seoond persoos 
present of v^9H m he (ese) are used as auxiliary; and they 
are combined, in all numbeia, with ihe sin^lar nom. maao. 
of the noun. 

a. Thus, from V^ dft ffive: 











b. Occaiionally, In the epl«B 4nd liter (almost nsTei in the old«t 
lingaege), the norm of the tanse aa giTen abova Is in Tuioni letpectt de- 
parted from: that, by we of the anzUlary In th« 3d p«r«on alto; by iti 
omiulon in the lit or 2d penon; by iDTertlon of the order of uodd and 
anillilTTj by intarpositlaa of other wordi between them; by nie of a dual 
or plnral nom. with the anxlliaTy ; and by naa of a feminine form of the 
Donn. Examptei are: vaktA 'stl (MBh.) he u>iU ipeak; nllumtl (BIBh.) 
I thaU 01 thtm wilt itrike dovm, yoddlift 'ham (H.) I ihall Jight, ohadi 
draft* (HBh.) I shall tet, kartS "haih to (BhP.) I miU do far thtt, 
Crtuh bhavltft (HBh. Hegb.) thou will b»; aami KantS (HBh.) / thaB 
go; pratlgrahits tSm aaml (HBh.) I uiill rtctivt her, hetntft tram aal 
(HBbO tJ""' viilt tlay; kart&rSa ava^ [MBh.) tct Iteo thall do; draftT 
asml (HBh.) I (f.) thatt tee, adbhavitrl (Nal;,] the wiB inereate, 
(Cantrl (T.) *he vjtU go. AB. ha* onca sota ai 2d ting., thou wit pre**; 
JOB. makei the combination gma^&a&nl bbaTltSraa the etmeteriet 

O. An DptatiTe of tbe auxiliary appeara to be once uaed, In joddlift 
arSm I tOBuld Jtght (R. 1. 22. 26 P«ter*on; bnt the Bombay edition readi 
yoddtmifa yftayftml). 

04B. Tb6 accent in these combinationa, as la all tbe ordinary 
,oa>eB of oollooation of a verb with a preoeding predicate nonn or 

ioy Google 

337 Pbhiphrastig Future. [—948 

ft4Jee^^o(B82J. ia on the noan itMlf; itnd, unlike all the tnie verbal 
forms, the combination retains its acceot everynhere even in an in- 
dependent olaose: thus, t&rhi vt atlnB^trA bbftvltasml (QB.) thm I 
thall be out of danger (vhere bliaTifyftml, if need, would be acoent- 
lees). Whether in a dependent clause tbe auxiliary verb woald take 
an accent (B86), and whether, if so, at the expense of tbe accent of 
the nonn (as in tbe caee of a preposition ooiepoiiBded with a rerb- 
form: 1088 b), we are without the means of determiniDg. 

946. In the Veda, the nomina agenti$ in ^ or tor, like tuIodi other 
derlTkttve nouns (S71), but with eipecial heqnency, are naed fn partlelplil 
conattoetion, govarolng the accusative If the; come from roots whose verbal 
Carma do so [1 18S}. Often, alio, tkey are naed predieatUelr, with or without 
accompanTing copula; yet wllhoat any Implication or ttme; they are not the 
begtnulDgs, but only the foierannera, at a new tenst'tonnaUDn. Oenerally, 
when they hare > participial Tilue, the root-tyllable (or a prefix preceding 
it) haa the accent. The tenie-nae beglne, bat rather apariDgly, Id the 
Biahmana* (from which about thirty tornu us quotable); and It grows more 
common later, though the perlphraatlo future is nowhere nearly so treqaent 
as the B-future (it ii quotable later from abont thirty additional roots). 

847. a. A Few iaolated attempts are made in tbe BrihnaDaa to form 
by analogy middle persons to this fntnre, with endtnge CDrreaponding after 
tbe niual faihlon to those of tbe actlTB persona. Thna, TS. bas once pra- 
7okt&e I toill apply (standing related to prayobtftaml aa, for eiample, 
Qftse to fioml); C^B. has Qayltioe thou thait lie (similarly related to 
qafitOai) ; and TB. has yaf tiamahe v>e witt make offering. Bnt in TA. 
is fonnd (1. 11) yaffle as lit sing., showing a pbonetlo oorreapondonoe of 
a problemaHo character, not elsewhere met with In the language. 

b> On the baais of ancb teutatlTe formations as these, the naUte 
grammarians set up a complete middle Inflectloa for the periphrastic future, 
aa followe: 

s. d. p. 

I d&t&e datiawabe dst&mahe 
1 d&taee dsttta&the dttadbve 
3 dfiUi d&tirau dat&TM 

O. Only a single example of such a middle baa been bronght to light 
in the later language, namely (tbe cansattTo) daT9(^tUie C^alf.). 

Uses of the Futures and Conditional. 

948. Aa the a-futore is the commoner, so also it is the ooe 
more Indefinitely used. It expresaea in general what is going to take 
place at some time to come — but often, as in other languages, add- 
ing on the one band an implication of will or intention, or on the 
other hand that of promise or threatening. 

Whitmy, Qrammir. i, ti. 32 

ioy Google 


a. A few extmplei are: varflfyAty M;Amah p&ij&nyo vffflmAn 
bhaTl^TRtl (9B.) it it going to rain .- Parjanya it going to be rich in ratn 
thiiytar; ykm tka nJi vdds klm jai ktwlfyati (RV.) tohomer doet not 
knoui that, iohat will he do with vtriet t vti vayjun agni dtiBaySmkh^ 
tha yuy4ih k(ih karln^tha (58.) we art going to build the Imojirat; 
then tchat laill you da f t&m Indro "bhyltdadrBvA lutalfy&n (QB.) Aim 
Indra ran at, intending to elm/; y4dy era kar^y&tha Bftk&lh dar&ir 
ya^tiyiiBo bhRvify&tha C^^) >/ J" "'^ ^ *^'< V ■Anti bt viorthj/ 0/ 
the eacrifiee along with tht godi; dintfla to qataymntl (AT.) thg tetth tciU 
fall out; 11& mari^Ml mjt bibhe|;i (^T) thou thail not die; be not 
afraid; brKhl kva ySsyasl (MBh.) tell ut; tehere are you going to got 
yadl mftih pratySkhyftsyacl viqam asthftsye (HBh.) if you ihallr^'ecl 
me, I icill retort lo poiton. \» in other UognigeB, the (enie la tlao lome- 
timei uaed for the expreaslou ot a conjectare or presumptton : thns: ko 

^'yaifa devo gaiidlutrTO tS bhavifyati (HBh.) toho it thit t he is doubtleu 
a god, or a Oandharva; adyaavapsyaiLtl(MBh.)(A«yniu9f be tieeping now. 

b. Tb« «phere« of fatare uid de«tdei«tiTe border npon one uiother, 
uid the one ia lometlmel met with where the o^er might be expected. 
EiimpW of tbe flitiire taken In > qnwi-dealderatlTe aeme ne w followt; 
y&d dft^&fe bhadriifa kwrlfyial t&vi t t4t saty&m (RV.) ■"'^ 
favor thou wHlett to bettotc on thy worshiper, that of thee becomelh actual 
[ie surely brought about}; yAthft *ny&d Tadlfy4nt b6 'aykd v&det 
(9RO as if, intending to eay one thing, one were to say another, 

949. The periphrutio future ts defined by tfae gnmmarians as 
eipresalag something to be done at a definite time to come. And 
this, thoQgh but faintly traceable tn later use, is a distinct aharaater- 
iBtlo of the formation in the language where It first makes Its ap- 
pearance. It iB especially often uaed along with fria tomorrow. 

a. A few examples are: ady& Tar^i^yatl ... qvo TTaf^(HS.} ii it 
going to rain today 1 it aill rain tomorrow; yatarftn vft ime qva^l kaml- 
tftraa te JetSrae (K.) whicJiever of tmo parties these thaU ehooss tomorrow, 
they will conquer; prftt&r yaft^bmalie (TB.) we shall taeri/tee tomorrow 
morning; ltyah6 vah paktiUml ^B.) on such and such a day I wilt 
cook for you; t&n ma ^kOih ratrlm into 9ayitlUe JfttA a to "yidi 
t&rhi pntr6 bhavitlE CV-) '^"i y^" '^"^ '■' '"^^ "" ""^ "■j'^'i "i^ o' 
that time this son of yours will be bom. lu other casee, thl» deflnitenes* 
of time 1* wanting, bat an emphtsia, aa of apecial eerlatnty, leemi perhapa 
to belong to the form; thna, blbhFbl m& pSrayit^yami tvi ti: k&smftn 
mft pSrayi^yftsi ty fttigbA lm£h aArrafai pn^lt nlrro^blt, titas tvK 
p&rayltfami 'tl d^.) support me and I will save you, said it. From 
what will you save met said he. A flood ie going to carry off all these 
creatures; from that I will taee youi said U; parldewayBih oakrlre 
mabao ohokabbayaib prSpt&ama^ (OB] thoy set up a bmmtation : "we 
art going to meet with great pain anddrea^; y^e "yakfl yaf^ha oa 
(TA.) I sacrifice, I hoes mcrifieed, and I shall saerifiee. In yet other oaies, 

ioy Google 

339 U8B8 or THB FtmiKBS ahd Conditional. [—860 

In tlie Dldai lingnige even, and yet more in the Uter, thli (ntare ippean 
to be oqniTilent to the othei; Ihoi, pr^JlTam anadi T^JtUttSsmo JtM 
vldvSii tB Jnhoty ftvldTftii vl (AB.) in.Au chtUrtn tee thaU kmne him, 
lehether ha u on» that tacrifieta loith knaaUdge orieifhout hnoicledgt; vak- 
tluno vft Idoih dsvebhya^ (A.B.) w thaO itU thit to the god*; yadi 
•Ttrtho mAmft "pi bhaviti tota svodL BVKrtluidi karlqySml (M8h.) 
i/ later my own affair ahall coma up, Uian I teiU altand to my oum affair; 
katbaih m bliavltaBr eka, lU trKdi n^pa qooiml (HBh.) but Aote miU 
you get along alonet that, O king, it the eaute e/ my grief about you. 

960. The cooditional would aeem to be most originally and 
properly used to si^lfj tfa&t something wm goirtg to be done. And 
thiB valne it baa in Ita only Vedio oconnence, and ocMsionally elae- 
vhere. Bnt Dsnally it has the Bcnae ordinarily called "conditional"; 
and in the great mfyority of its occarreocea It tg found [tike the sab- 
JDnctire and the optative, when used with the same valne) In both 
clausea of a conditional eeotence. 

a, ThiiB, y6 Tftrtra abiam Atri "bharl^yat pr& t4ili J4nltrl 
vlddqa nv&oa (RT.) him, whc tfoi going here to carry off Vritra't weallh; 
hit mother proclaimed to the knmoing one; Qatftyiuh gfim akarl;yam 
(AB.) / tea* going to make (ehould have made) the eotc live a hundred yeari 
(in other i«rfjan« of the lams itory U mdded the athei clanae, In vhlch the 
conditional hu a value more romoved tram Ita original: thna, In OR., if 
you, tillain, had not ttopped [prKgrablfyah] my moulhY, t&ta aril *aya 
bhay&ih vl VSya k&am&d dhy Abha^yad dvltiytd vti bhayAdi 
bbavati (9^-) thereupon hi* fear departed; for of whom tea» he to be 
afraidf otcation of fear arieee from a eeeond pereon; ntpapftta oir&di 
t&n mene y&d visa^ paryAdbftayata ((B.) he leaped up: ha thought 
it long that he ehould put on a garment; fl4 tAd avA ni Vlndat 
pr^tpatir j&trt "taofyat (H3.) Pr<^apati, verily, did not then _find 
where he wa* to {ehoiUd) tacrifiee; evadi cen nft Vak^o milrdbB t« 
vyapatlfyat (OB.) if you ehould not epeak thue, your head would fiy 
off; bA yAd dbai tivad o-vt "bbavlfyad yivatyo hU 't< 'gre praj£^ 
■nt^ titvatyo hU 'vi. 'bhavl^yan oA pHt 'Janlqycmta (QB.) if he 
had been only eo muth, there would hajie been (mly eo many living creaturei 
a* viere created at Jiret; they would have had no progeny; Ulfa v& 
'bbavl^yad Bm^aa tamaaBjh vibhatui taifa oet aahaarakira^o 
dhorl n& "karl^yat ((.) would the Dawn, foreooth, be (he tcatttrer of 
the darkneti, if the thouiand-rayed one did not aet her on the front of 
hit chariot f 

Digil.zec.y Google 




961. a. Thosb verbal adjectives, or partioiplee, which are made 
from teQae-BteiDB, and so coDBtitnte a psit of tbe various tense- 
syatema, have been already treated. It remaina to deacribe certain 
others, which, being made directly from the root itself, belong to the 
verbal system as a whole, and not to any particular part of it 

b. The infinitive (with a few sporadic exceptions in the older 
language) also comes la all cases from the root directly, and not from 
any of the derived tense-BtemB. 

o. Tbe same is trne of the so-called gerunds, or indeclinable 

Passive Participle in t& or n&. 

962. hj the accented sufEx rT t^ — or, in a compar- 
atively small number of verbB, =1 ni — is formed a verbal 
adjective which, when comii^; from transitive verbs, quali- 
fies anything aa having endured the action expressed by 
the verb: thus, ^ datti given; T^ ukti spoken. Hence 
it is usually called the passive participle; or, to distinguish 
it from the participle belonging to the passive present- 
system (771), the past passive participle. 

a. When made from aa intransitive oi neuter verb, the 
same participle, as in other languages, has no passive but 
only an indefinite past sense: thus, TH gatd gone; HcT bhati 
been; ilfdR patiti fallen. 

968. In general, this participle is made by adding rT 
td to the bare verbal root, with observation of the ordinary 
rules of euphonic combination. 

a. Some roots, however, require the prefixion of the auxiliary 
vowel 1 to die But^x. For these, and for the verbs that add n& 
instead of t4, see below, 966, 967. 

ioy Google 

J41 Passive Participle in ta or na. [— 9BB 

b. Aa to the accent when the root is preceded hj a prepoBition, 
see lOSBa. 

964. The root before FT t& 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 fs not seldom dropped: examples are 
akt& (Va&j;, baddhi (vttaadh', frabdba (v'qrambb], daffi (v'daAq), 
arosta (i/BraAsJ, bft^ba (>^afib}. 

b. Roots which are abbreviated in the weak forms of the per- 
fect (784) suffer the same abbreviation here: examples are uktA. 
iyvtiO], aft& (yVaa thme), npt& (|/vap: also vapta), Q^b^ (v'TSh), 
«npt& (i/srap), IftA iYs^], vlddtaA (yvyadh! ; — and, hj a similar 
procedure, y^r^tih (or praq) makes Prft4i Vbhrafif makes bh^ft^ 
(beside the regular bbra^fA), and y<jTi boil makes 9^ [beside frat&). 

{^^Flnal 4 li weakened to i in gitA (/gft >uif), dhlt& ((^dhK *uc(J, 
Plt4 (l/pft drink), Spbita; and JIti, vlt^ qlt& are made from the looti 
Jy4> tt4> Qy4, (01 JI ete.); — and farther to i in ellit& (beside oh4t&], 
dlta (ydA divide and di bind), drita (P ydii tUtp) , bltk (y'dhft jittO- 
with b^ for db ; bat dblU aUo ouara In V.), mlt4 i^niA meature), qlU 
l^aiso Qftta), Blt&, StbiUt. 

d. A final m Is lost after a in fati. nat&, yatA, r«t& (hom ygnxa 
etc.); and a final n In kqata, tat&, mat&, hat&. Aa to the other roots 
in am and an taking ta, lae 9BK a, b. 

a. Hare isolated cases are -flta (RV. : yar), nt& or tlta (yvi wtaoe), 
giqti (al«o qSata: KQ^s), mfatk (refeTied to ymflrob). As to -gdha 
and jagdb&, «ee 233 f. 

f. Oil tbe olhet hand, ^avad makaa Bvitt4. 

ess. Of more irregular character are tbe following: 

a. A number of roots ending in am retain the nasal, and leogtben 
the radical vowel (as also in some others of tlielr verbal forms: thus, 
kSibtA, krbbt&, USibtA, k^&tbta, oidita, t&iht&, dKiiit&, bhrSihta, 
vSiiitA, QBjht& {; qam be quiet], grftifat& {from ylcam etc-); and one 
in an, dbvan lound, makes dhTfintA. 

b. A few loota In an nuke their partlefpla from another loot-funn 
Id ft: tbna, kbftti, J&t&, -vfita, sAt^i dbam hai both dbamlti and 

O. Certain rooti in Iv Uke their yQ-foim (76Ba): thua, dyfit4 (ydiv 
flag), ftbrDta, BjrftU; but ynuv makea -mata. 

d. From roots in changeable x (geneiallr taking na: 9B7b) are made 
also panfc (_yvi fiS- l>edde Pfta), qlrta and ifOxtA, (y^ cnuk); and 
9lrta Is tnitber made from V9ii mix. 

i, Google 

066—] XIII. Vbrbix Adjrotiveb and Nouns. 342 

•. Doubl« foimt (re mngdhi and ma^lUki Stfl^ "'^ ao^lM, dhOrtft 
■>nd dhruta, hv^ta tod hrati. 

f. Tlie loot dB give makes datt4 (ftom the secondarr lOot-focm dftd; 
but data alw Id T.), Bat tbt anomaloail; cantraet«d form -tta (a* If 
tor dftttt, witb th« radical yowal loit) ta also fieqaent in eompoaitlon, ei- 
peeUltf wltbprepoaltLona: thui, iua, AnuttA, pirltt», Fr&tta,priiUtta; 
rarelT with other elemantt, as dev&tta, punutta, mftrdtta(T). And the 
aame abbieTiated fbim eomea from yH dmidt \a Avattk. 

g. The root! maUng parUotplM In both ta aod Ita, or ta and na, or 
In all thne, will be noted In the next two puagnphi. 

966. The suKx with ^ i, ot in the foim ^ Iti, ib 
used espeoially with roots having finals that are only with 
difficulty, if at all, combinable with (T t acooiding to the 
1UU&1 analogies of the language, and often with roots of a 
secondaiy, derirative, or late chaiactei; but also not seldom 
with oiiginal roots. 

a. Thai, of root* piewDting diffleultlea of combination: — 1. all that 
end In two consonanla (eaie those of whloh one eonionanl la loat by 4 weat- 
enlng piocess: 964 a, b); e. g. gaflk. Talc, vKfiob, 1«J]. nbj, oaft, 
Khtt>9, katth, oind, Jalp, oumb, nmbh, kball, plnv, ;aAt (aiw 
9BBt&), rakf, ItlAa, satb (In all, over flit;); but takf makea tofta; — 
3. all that end In tingoals (Inolndlng 9 after a 01 K): e. g. af. truf, pafh, 
luth, I4> Tm^i bha^, kaf, bbSf; — 3. all that end In nud ^iMott: 
e. g. llkh, ((ral}i, tiKtb. kntb. ripbi gi^pb ; — i. all that end in 1 : e. g, 
oal, gill mill lull kbel; — 6. all that end In othet paniaient aemlTOwela: 
namely, oarv (alio oOr^), jiv (for the oth«T loota In It, see 9S6 o), 
db&v run, urn, day, vjay, pllr; — 6. u]h. — Thli daw iDoladet mote 
than halt of the whole nombei tb*t take only Ita. 

b. Of other tooti ending In ooDaonantt ; — 1. in guttarala, oak, ^ftuk 
({ak haa both ta and ita); fUgh; — 2. In palaUla, ac (al«a akui), 
ne, kno, kbao, rfto, mo; a\V, kOJ, vraj, also ty^j and mpj In l>t« 
text! (naually tyBkt4 and d>ct{&); — 3. In dentaia, at, pat, qoat, alio 
yat io epos (eliewbeie only yattji); krad, khftd, gad. end, nad, mad, 
mfd, rad, rod, vad, vid hnofn, brad; alio nod in epoa (elaewhen 
nutti and nunna); mad haa both matt& and madlt4 (the mtjorlty 
of loou in d take na : 857 d) ; adb, k;udb, gadb, dndb, nSdh, 
bftdb, Bpardb; an. In, kvan, dbvan, pan, ran ring, van, atan, 
avan, and dbwan (aUo dbvftuti); — 4. in labials, cap, ynp, mp, 
aud aiaally kup (knpta late) and lap (lapta epic), oecaaionally kflp, 
gup, tap, dfp. Tap, qap, while Jap has both ta and ita; grabh 
(g^bbltA), 9ubb, Bkabh, and occaalonally lubh, while kfubh and 
■tabb have both tarmB; tim, dbam. {am ItAor, atlm, and kfam in 
epo* (alto kfiiiita);— d. In apiianti, af «ot, Iq, ktq, kfq. Ttq, qaq. 

ioy Google 

343 Passive Participlr in ta or na. [— 9B7 

irblle piq hu both foimi, and mqr9 Ukei Ita only Iste; If rand, I;, kn;, 
tn, tvlf , pmq, mlf , ra;, httf, bre;, aUo moq eirept Ute, irhila dhn, 
roq, and hff ahow both tormt; ta, bhM, bhfta, ratt, lag, vas cfoMa, 
has, also aa throw occasionally, «hile kaa, graB, yas, vaa ihme, vas 
dwell, qOs (witti i;i;{4 and qllta), gvaa, and hras make both forms; 
lb, frail (gi4dt4). Jab (iccondair form of hi), mah, rah, and occsslonalljr 
ah rtmotie, wbile (Ctb has both forms. 

o. Of roots eiidlug In vowels, only qi lie, which makas qa^ta (vlth 
gui^ of root, as elMwhore: 629). 

d. Id geD«ril, a root malntalna ita fall form before Ita; but thaie 
are a few exceptions: thus, g^bblti and gfUti (the loot being reckoned 
as grabh sad grab: see 729), udlt& (also vadita in the later language). 
Of ita [yvae ihine; beside a;t^), n^ta [yVas dwell : also iporsdloally 
TUita and aqfa), uk^ti [y'vak; inereate), qfihitk (f/qratb). From 
ymfi ate made botb miJlta and m&rjlta (with strengthenlag a» in present 
and elsewhere: 8S7), beside mnt^. 

e. Instead of 1, long I Is talien lo glftAdtk and gfblti. 

967. The sufjix ^ n4 (always without auxiliary ^ i) ie 
taken instead of FT t& by a Dumber of roots (about seveDty). 

a. Certain roots In &: Ibne, k^ft, gift, drjl run, drft tltep, (also 
drltaP), mis (also ml&tA), Tft blow (also T&ta), qyft [also qini), styft, 
hft Uave (also hln& and hftta), hi go forth; and dft divide mshei UnA 
(also dlta and -tta). Farthar, cerUln roots in 1- and u-voweli: thus, kfl 
dftiroy (kfiija; also kfltA), ^ pi, II ding, vU, ql or qyft coagulate 
(beildo ijyilna and qlta), hri (beside hrlta)i dtl bum (also duta), IQ, 
qQ; and diT lamtnt makes dyOna (compare 766). 

b> Koote in f, whirh before the suffli befomes Ir ot vkr: the forms 
are, ar^a (late; bealde {t&), kln^ Cv^ scatter), (fieifh (v'KT «toiifioic), 
Jln^ and JiLn^ (yij watte atcay), XSn^ and titr^ (also tuTt&), dlr^A 
iV^ pi*rt»- >>■<> <lrt<^), pfLr^i (.VW^- >1bo pSrti and pfta], mOr^i 
(ymr erutK), qir^i iVfi crush: aleo qlrta and qart4P], sOrj^ (atao 
■tfta). Of like chaisctei with these are Ir^A from y\i, olrqa (beside 
oarlta) ftom f'aar, g&n^ (beside gurt4) fram i^gur, a secondary form 
of gf, and oQn^a (beside carvita) from >'oarT', which is also plainly a 
■econdary root 

O. A few tools ending In j (which becomes g before the bdUi againtt 
the Dsnal tale of Internal combination: 218 f): thus, bhagna (^hafij). 
bhngiia (v1>baj bend), magni (ymi^), rng^A, vlgna (betide vlkta). 
Fortiier, two oi thtee ending in o (similarly tcealed): thus, akll& ()^ao 
OE atio: alto aolta and a&elta), vrkfi (f'vraqo), and apparently -pntpo 
(BV., once: with doubly irregnlar change of root-Hnal, from V'PFO). And 
one root in g, lagna. 

ioy Google 

QS7— ] XIII. Verbal Adjzotitbb amd Houms. 344 

d. A eoQatdeisUe number, (ome of them lery cemmon ones, of loott 
in d (wtileli, *g>lnit erdlnirr rule, becomes a befoie the inffli: lB7b). 
The forme are: oniia (alM ntte), KHftfAf, kllima, iryiwni^, kqrli^K. 
wii<iiii» ^ cbumot ohlBnii, cbfi^ijA, tiumi, tp^i^A. ntuuia (>lio nattii 
■nd nudlta), p«iiii&, bhtnnA, vlnna (v'vld ^nd: t,\so Tltt&), fonna 
[/^ad/aU), Buinfc (ilio Batt&), skannfc (y'skond), sTumi (ywyxaA), 
BTI1111&, htuma. And kanA /'ood. In epite of lie different vxeni, >ppe*n 
to be B Ilko foimalion (lom y'ad «al. 

9BS. The DAtlTe gTAmiiuttianB leckoD u participlea of this for- 
mation a few miBCellaDeoua derivativfl adjectirM, coming from roots 
which do not make a regular participle: Buch are kqlaui burnt, k^Qi 
emaeiaied, pakT4 r^e, pholU eicpanded, QJifka dry. 

Fast Aotive P&rtloiple in tavant [or navant). 

9B9. Fiom the past paasiTe participle, of wfaatevet 
foimation, is made, by adding the possessive suffix 31? 
Tsnt, a secondary denTatiTe having the meaning and coo- 
sttuotion of a perfect aotive paitioiple: foi example, ^rT 
eTiH°rH tdt krt&Tbi having done that; taih niglrQavftn having 
awallotced him down. Its inflection is like that of other 
deiivativea made with this suffix (4S2 ff.) ; its feminine ends 
in cTrf^ vatlj its accent remains on the participle. 

860. DeiintiTe woide of Ihla formadon t» found In RT., bat withont 
uijtfalng like * pirtlclpl&l valaa. The AT. be* a elngle OEtunple, with pu- 
tlclpitl meuilng: a^lttvaty &tit)ita on«'i gtitti having taten (loo. abe.). 
In the BAhmaQai alio It le hudlr met Mth. In the later langaage, howoTot, 
It eomea to be quite oommon. And there it li ehiedy naed predloattTely, 
and ofl«nect witbovt copnU eiproMed, or irlth the Tain* of a persona] TOib- 
fOrm in a paet teoee: primarily, and not leldom, elfnifring immediate p*«t, 
or lu<rlns a tra* "perfect" Talne; bat also (like the old perfect and the old 
aorlft In later nee) oomtng to be freely need for lodeflnlte time, or Mth the 
Tatae of the imperfect (779). For example; mfiih na bagoid dfffavfin 
no on* hot teen (or 1010) dm; sa nakuladi vyKpAdltavSn ha de»tro^»d 
iht iehatumon; or, with oopnla, mahat k^eolirBih prSptaraty osl Mou 
hatt fali«n i^on great mietry, Althongh originally and piopeily made 
only from traneltlve verbB (with an object, to which the participle in ta 
Btandi in the relation of an obJectiTO piedleaiive), It fi flnally foond alto 
from intranaitlTes: thne, odtena aath^tavaH (9,) Am bacome united 
wtih Ma mango-tret ; gatavati (lb.) the hat gone. 

a. The rame participle ia alio made in the secondary oonjugatlone: 
e. g. daiqitavant having ^unon, prabodhltavant having atoakaned. 

ioy Google 

345 Gerundives. [—863 

b. PoMMllTBi klio Id In nude from pulive ptrticipteB tie eams' 
limes found Q)ed In to kntlognos iDiQiieT, Dearly ■■ partect tctlTe partlcl- 
plM: e. f. Iftin having ittcriJUtd, vljltbio mui7ambi&^ (AB.) thinking 
th«mt«het to have conquered. 

Future pMBlTe Participles: Gerundives. 

961. Certain derivative adjectives (foi the most put 
more oi less clearly secondaiy deiivativee] have acquired in 
the language a value as qualifying something which is to, 
ot which ought to, auffei the action exptesaed by the root 
&om which they come; and they aie allowed to be made 
fiom eveiy verb. Hence they are, like mote pTOper pai- 
tioiples, sometimes treated as a part of the general verbal 
system, and called future passive participles, or gerundives 
(like the Latin forms in ndua, to which they correspond in 

662. The suffixes by which such gerundives are regu- 
larly and ordinarily made are three: namely Uya, rTc?7 tavya, 
and fl4l(J MUlya. 

a. DerlTUlveB in ya biting tbii VKlue >r« made tu all pBriods of the 
luigntige, ftom the eultest down; the othei tno are of moie modem origin, 
being entliely wanting in the oldest Yedi (RV.), and hardly known In the 
later. Other dariTatiTei ot * similar character, which afterward disappear 
from nae, ate foand In the Veda (960). 

96S. Tbe suffix 7a in Its gemndive use bas notbiog to dia- 
tiD^ieh it from the Skme saffix as employed to insks adjectives and 
nouns of other oharaoter (Bee below, 1313). And it exbibits tXm the 
Mine variety in tbe treatment of the root. 

a. The original valne of the snffli Is la, and as such it hag to be read 
Id the Tery great mijorlty of Its Vedlc occuiiences. Hence the conTersion 
of • and o to ay and av before it (see below). 

b. Thus: 1. Final & becomes e before tbe suffis: Ahj% dliyeya, 
khy^TS, mdya (perhaps dlt-ia et«., wtth eaphonic ; interposed]; but 
RV. baa onoe -Jftftya. — 2. The other vowels either remain unchang- 
ed, or have the gu^ or the vrddhi strengthening; and e usually 
and o always are treated before the ya as they wonld be before a 
vowel: thoB, -kfayya, J&yya, bhiyya, l&yyai n&vya, bMvya, bAvya, 
blL&vy&; virya: and, in the later language, nlya, j«ya, dhtiya (such 
cases are wanting earlier). In a few Instances, a short vowel adds t 


969—] Xm. Tbrbal ApnonvES and Nouhb. 34Q 

before tb«BnfBz: thus, lty&, mityA, fr&tTft, Bt6t7a, kf^ra (the oaly 
Vedic ezuDplea). — 3. Hedi&I n rem&inB unchkoged or is lengtiiened: 
thns, lUbhya, vAndya. s&dya; mtdya, v^laya. — 4. UedUl 1-, «•> 
and f-Toweli are unchaaged or have the cn^-Btrengtlieniiig: tbu, 
i^ja, guhja, dhf^a; dvifya, 76dli7a, miijya. 

e> Tbe BY. hM »boDt forty eiimplet of thta gamadWe, uid the AV. 
Kddi litlf u muT mors. Kscept in bhSrlA (one«), the koe«nt in RV. 
U ■.IwiT* on the not; AT. h&a uTenl cum of leeent on the 1 of th« 
rani (beoM written idjk, tqyl^ -vyidhyli, -dbarfyk). According (0 
the gnmmtniuii, the lueiit Is on the loot ot eUe the ending Is oirenm- 
fleied: tlviji the temei, if Iha ya fbliow t Towel. 

9d4. a. The aufSz tavya is a secondar; adjectlTe derirativa 
from the iDfiDitival dodh in tu (below, 868], made by adding the 
nffiz ya (properly ^ wheooe the aooent yi), before which the final n, 
aa maal (ISOSa), has eo^a-Btrengthening, and is reBOlved into av. 

b. Hence, as regards both the form taken by the root and the 
nu or omisalon of an aQxitlary vowel i before the tavya, the rules 
are the eame as for the formation of the infinitiTe (below, 968). 

O. Ko nunple of thU fonnttion it found in HT., ind In AY. oocdt 
only tvo, Janitavyh and hlAsltavyh. In the Bnhmsnt Isngatge It be- 
gim to be not rue, and is mtde both from the simple root snd fnm the 
d*Tl*ed conjDgationsI stems (next obspter); In the duiles] Ungnsge It is 
still more frequent. Acooiding to the grimmariaiis, the scc«nt of the word 
Is either otrcnmaei on the flnil or scute on the pennit: thus, kartavyk 
or karUvya; Id the scoenlasted leiti. It is tlwsTs the former (the Mcent 
t&Tya glren to certain gemndlvei Id the Petenbuig Uiicona is sn error, 
growing out of the •mhlgaons arcentnitlon of QB.: 880]. 

eeb. a. The snf&x anlya ia In like manner, the prodact of 
secondary detiTation, made by adding the adjective snffiz lya (ISIB) 
to a nonun aetionit formed by the common snffiz ana. 

b. It follows, then, se regards its mode of formation, the mlea 
for the suffix ana (below, 11 BO). 

O. This deriTslWe also la unknown In RV., and in AY. la found only 
in npt^lTaniya snd AmantaraQlya (in both ot which, moreoTer, It* dis- 
tinct gerundlTe value admits of qnestlon). lu the Br«hmanas (where less 
than a dozen examples of it have been noted), and in the later language, 
it is lets oommon than the gerundive in tavja. lu accent, a« in all the 
derivatives with the sntflx lya, ft on the penatt: thui, kara^iya. 

866. Other formatioBs of kindred value are found in the Yoda at 

a> Gerundive* in tna 01 tva, tpparently made from the Influitlval 
noon In tn with the added autflx a (1208). They are kirtfla (In two 
<t k&Ttva), •gaihtraf.Jfuitiu^ Jitua, n&ditaa, v&kttut, a6taa. 

ioy Google 

347 INFINITITHB. [—968 

■nttofti bintna. hAtuft, botva; uid, with auxUiaiy i (or I), J&nitva, 

b. QernadlTst In enlB ot enya (compue 1817): they are Oi^Qitt, 
I^^nlA, oari^la, dfqinU, -dvlqa^la, bbflfi^TB, Tudh^ola, tj^tk^Ik 
(and bbftlBliyB BhP.) ; with one aiample from an apparent aoiist-stem, 
TUfaBtoyo, and thiM oi ftmi from aecondaty TetVit«ms [>ea betaw, 1016, 
1088, 1068 a). 

O. GerondlTBs In l^la (onoe Ayya: compua 1218): tliey are dak- 
ftylai panJtTla, vidtyla, <}TKvtjimi hnnv&yla; with a few from aeeon- 
daiy oonJngadan-atema (below, 1018, 1088, lOKl, lOeSa); and ItOfiylH 
la of elote Undted with them. 

d. A few atDeotttei In ellma, as paoelima, bhldelima (only tb«sB 
qootable), aK lookoned ae gOTDndiTea b; the grammariana. 

867. The dlriBion-line between participial and ordinary a^ec- 
tivM is lesB itriotlj drawn in Sanskrit than in the other Indo-Euro- 
pean langnagea. Tbua, adjeotives in u, as will be Men later [1178), 
from Becondaij oonjngational Btems, have participial ralue; and In 
the Br&hmanai (with an example or two in AV.) is fonnd widely and 
commonly naed a participial adjective formed with the suffix oka 


868. The later laoguage has only a single infinitive, 
which is the aooueative case of a Terbal noun formed by the 
suffix ^ tn, added to the root usually directly, but often also 
with aid of the preoediog auxiliary vovrel ^ i. The form of 
the infinitive ending, therefore, is ^tl turn or ^tT itmn. The 
root has the guva-strengthening, and is accented. Thus, for 
example, ^^T ^tom &om f^ i; tf^ kdrtum &om ym kp; 
tjftHH oirltum from y^ oar ; HleTrR bhivitum from y^ bhO. 

a. Ab regards the nae or omisBion of 1, the infinitive (aa aleo 
the gemnd in tvK; 991) followB in general the analogy of the pasaive 
participle (666). Ezamplea are (with the gerand added) as follows: 
rtagdh^ digdbtun, dagdbvjt from ydahi bbinni, bbittum, bbittrit 
from )/bbid; mat&, m&ntoin, matri from v'man; a^bi, vA^nm, 
&4bTit from I'vab; patltA, pititnm, patitva from v'pat; yftolt&, 
ytoitnm, yftoltvl from yyio; Qarlt4, 9&7ltum, ^ayltva from v'fi. 
Bat certain exeeptionB and speoial cases require notice. Thns: 

b> Of loota haTing no quotable paitlolpl*, inflnltiTe stems In tu ate 
made ttosa ad, saghi in Itn from nfiobi ilb eontider, kfap, In^tb, 
lok, Bvar; and in both from yabb. 

ioy Google 

868 — ] SIII. Verbal Adjectives and Nouns. 348 

O. Ot iMt( miking piitlelpUi of both forma, an inOnltlve Btem in 
tn only is quotable for k^p, kfubh, tap, tya]. loffi lubh, vas Mute, 
Qkk, aUbh; only in itu for gKb, oarr, Jap, mad, yat, van, 9afiB, 
qvaa; In both fOi as throw, Oil remove, gup, oar, mpj (mkr^fa, mJUr> 
jltu), lap, vaa dwtU, ^ap, f&a. 

d. Alio In 1 nQmbeT of othei cues {berideB thoie ilietdy natic«d) ui 
iiiAnltlve alem la made both vlth ind irlthont i. Thai, In addition t« th« 
more legnlu tmm, > »Km In itn I* occadonally met vltb from n»U a^ 
attain, 1; eeek, baadh, bbaj, yi^ (^itnm), mdh oMrua, mh, vn, 
sad (■Idltum), Bah, ban, bf ; and one In tn from roots fta, bbSf, vid 
httow. Both forms oocur alio from certain am-rool«, namely nam, yanii 
ram, and, with ft before tn a* in the pple, kram and bhram (kfam 
haa only k^adttn, against the analogy of k^Sihta); farther, from certain 
roots in larlable ^, namely tf (tartu, taritu], vf cover (T&rtn, varitu), 
and atr (Btartn, ataritu, Bt&rltn) (but from Qf ctimA occur only ij&rltn, 
farita, and from -vT dutote only vwitn ; while gp luiallow and pf Jilt 
make their inflaltlve ^m other root-forma, namely glrltnm, pfiritnm); 
farther, from a few Towel-roota, namely nl, cyn, BH (Bllttn); and flually 
from kn, npt, fno. 

fl. Against the analogy of the participle, Inflnttiie-atems In Itu after 
a final conaonant are made from the rooU av, kqan, kban and Jan (the 
pplea coming from ktaS and Jft), gnb, Jabb, tarn, diT play and div 
tontent (both devitu], m(OJ> vjt, vjAh, Bfp; and after a Anal vovel, 
from roots In Q, namely pil, bbfi, bQ (alao sQtu), and from 9ri and fri; 
aa to roots In variable f, see Just above, d. 

f. Aa the Inflnltive Is made from the (aocented and] etrengtheaed 
root, so It naturally haa, as a rule, the stronger or foller root-form where 
a weaker or contracted form ts taken by the participle (and gerund in 
tvi): e. g. vUitu against alct4 (and aktv^, y&tfu against l^fa (and 
i;tvA), banddbom agalnet baddtaii (and baddhva}, and so on. Deserv- 
ing special notice are gStn (V'B^ '"'g) against gtt^ and db&u ()/dhft 
tuck) against dlllt&; and eo from dft give and bft leave are made only 
ditu and h&tu; but dbft put, mB,meaeure, and Btb& add to the regotar 
dbfttu, mfttn, athfttu the late forms -dhitn, -mitu, •Sthitu; and sS 
ot ai has s&tu, aitni and -eitn; vft weave (pple ut&) haa both vitta 
and 6tu; hfk or hvft baa havitu, hvJ^tu, and hvstn. The root vyadh 
makea Its only qnotable infloltive, veddbum, from Its Tidb-form; from 
■alU or Baj occur both Bafiktn and gakta. The anomalous epic forms 
lyitnm (>'yaj) and sidittum (Vead), were mentioned above. The root 
grata makes gr&hitum. 

g. In the later language, the Influltlve-steDi forma poasesslTe com- 
pounds with kftma and manaa (especially the former): e. g, araptu* 
kftma having the leitk to tleep, yaf(ukftma deiiroui of lacriJUing, 
vaktomanaa minded to speak. 

b. In very rare inetancee, dalive infinitives In tave or tavOl are 


349 Infinitives. [—970 

mid« bom the inflnttire stem la the Utec Unfnige (ii ibundtntlr In the 
esiUer: e70b): tbiu, pratUiartava (BhP.). And JItkm (97S a) U 
OD«a rooDd In UBh. [1, 3. 67 =: 732), in a quail-Vedlc bymii to the A9t1ni. 

900. Id the Veda and Brabmana, however, a DQiuber of verbal 
DouQB, nomina actionit, in varioQB of their cases, are used in con- 
structions which assimilate them to tbe infinitive of other laDgrnages 
— although, were it not for these other later and more developed 
and pronounced infinitiyes, tbe constractloiiB in question might pass 
as ordinary case- constructions of a somewhat peculiar kind. 

970. The coans thas need infinitively are the following: 

a. The root-DOun, without derivative suffix, is so used In its 
accusative in am, its dative in e oi (from i-roots] U, its genitive 
and ablative is aa, and its locative in 1, 

b. The verbal noun in tu is so used in its accusative in tiua, 
its dative in tave or tavftf, and its ablative and genitive in toa. 

Of othei nouns only single cues, genermlly datlTei, lie reckoned ss 
nfed with Inflnitire vilne; tbui: 

a. From the verbal noun In as, the dative in aoe; and also, in 
an extremely small Dumber of instances, a dative in se (or ^}, from 
a noun formed with s simply. 

d. Prom Doans in man and van, datives in mane and vane. 

e. From nouns in ti, datives in ts^e, or (from one or two verbs) 
in tyU. 

f. From nouns in 1, dttives in dye. 

K< From nouns in dhl and fi, datives in dby&i and fySl. 

h. A few Infinitives in ^ai^ are perhaps locatives from nouns in 
an added to a root increased by a. 

i. From a single root, dhr, are made InSnitively used forms in 
t&rl, of which tbe grammatical character is qnestionable. 

J. Among in these, the forms -wblnb hiTe best right to apsctsl treat- 
ment M Inflnltiyet, on sceoant of being of pecnIUr formation, or from 
snfQzes not found fn other nses, or for both reaaone, sre thote In ^, ^tu^ 
tarl, dhyfii, and tav&i. 

k. Except the various cases of the derivative In to, and of the root- 
DODO, these inBnitlYeB are almost wholly unknown ontside the Rlg'Veda. 

L Other safflzes and forma than those notioed above might be added; 
for It Is imponlble to draw an; Bied line between the nies classed as 
lullnltlie and the ordinary esse-nsee: thus, prt^ipatlih praqiL&in UtSm 
(T80 tA«^ went to oik Prq/'l^aii; rffvaifa Jlv&ih pruuv&nU oariyU 
(ItV.) quKkming every living being ta motton; apA^ a&rmftya ood&ytm 
(RV.) impelling the watm'a tojloui; qaknuy^ sr&ha^Sya (Instead of the 
oinal gr&liltiua: ^B.) mta/ be abk to apprehend; S tamanSt (Instead ot 
tbe usual tamito^: 8.) luidV exhaustion. And the so-called inflnltlTes 


970— -1 Xni. TiRBAL Adjectives and NouNe. 350 

are fannd coSrdtDtted In tlis iiima HoUnce witli common aomu, and eren 
with componnd nonni: e. $. uirltsTe . . . Sbboc4ya ]|tA7B rlyi O^O 
to go abroad, to anjoy, to $tek teeaUh; Srtfttr&i^l7« na pralurtnm 
anicui (^) /or &« ruea» of iht dittreutd, not for hurling at M« 

Hon spedil ralei u to the vuion* foimttloni are u tollowi; 

671. Tli« loot-Donn nsed u infinitive hw the urns form (eioeptthat 
It doei not take an added t; 388 f), and the eame aeeent, Iioth when tlnple 
and vhen eomblned with piapoiltloni, at In Itt otlier naei. In the Tery 
great majoiit; of initanoea, It 1» made fiom loota ondinB in a ooniODant; 
but alio l^om ■ reir In & Qchyft, di, ShK, pftP, mft, rft), from two or 
thiee In 1- and u-TOWel* (ttl, ml, bbfl), and from one or two In changeable 
f, which takea the ir-fDin (Ur, Stir). 

ft. The roots in ft [oim the aecna. In fim, the dat. In U, the abl. In 
U (undertlanding avaai beloie A u toi &▼«»& and not ktmAI in RT. 
111. 53.20), and the localtTO In • (only two ezamplei, of which one ii per- 
hapi better nndantood *■ datUe). 

67S. Tbe inflnltlTe nonn In ta )i made freely from looti of aTeir 
form. The toot takei the KU^a-etiengthening, If capable of It, and often 
adde the aQiiliary towoI i befoie the infflx (ucoTding to the ralea already 
■tated, 968). The mot ia accented, anloM the noan be combined with a 
prepoiltlon. In which caae the later haa the accent Initead: thus, Urtnia, 
dtave, h4ntoB; but nikartum, nlrAtave, nlrhantoe. 

a. The datlie in tavU ig In two respects anonuloai : In having the 
heavy feminine ending U along with a ittengthonad a; and in taking a 
donble accent, one on the root or on the pieflxed prepoiltlon, and the other 
on the ending It: thaa, iXtevii., hintATSf, AtyeUrit, ipabhsrtaTif. 

978. a. The Inflnltlve In u« ia made In RT. from abont Iwenty- 
flve root*; In AT. and later there hare been noted no other eiamplei of 
It. In nearly thiee qaartere of the eaaea, the accent la on the infBi: e. g. 
(lljiae. Jlvise, bhijr&a«, CdJ&m; the eiceptloni are oAkqane; dta^Tacs 
(with y inserted before the anfflx: fl58); and i^tMn, bh&roee, Bpirase, 
hiroae (with ga^a-atrengthonlng of the root). Strengthening of the root 
it alto shown by Jbv&m, doh&so, bhoj&m, Qobhiao. In pofy&se It 
■een, apparently, the piestDt^tem Instead of the root. 

b. The ending ae It extremely rate, being fonnd only in Jifi and 
perhaps Btn^, and one or two still more doubtful ease?. 

974. Inflnittvea in mana are made from only flre roots: thni, tc&- 
mo^e, dimaoa, d&Tmaqe, bbJoma^a, and (with diffetenl accent) vld- 
m&na. From ydft comes dftT&iie ; tturril^e may oome directly from ytf, 
or through the secondary tool tnrv; dbdrra^a la talhei from ydhltrv 
than from )^dhT^. 

97B. a. The loflnltivet In tay are Ifffcys {yi^), plt4ya (|/pfi 
drink), TlU^e, sit&ye, and perhaps Qt^a (flt4ra n^ to hdf hit m*n: 

Digitizecy Google 

351 Infihitites. [—981 

S7.). In t3«i, the only eumplM noMd no Urii (RT.) ind sd^brU 
(M8. AB.). 

b. With h;« tie roTmed If&ye, ti^Aye, df^&ye, mahiya, yndhky, 
■aniya; end oiUja (T9.)i STbaya (K.). 

976. The ending dbyU 1«, more than tn; other, liregalu uid rari- 
aus In Iti treatment. It ha> alvaya an a befbre it; and in tb» majority 
of cuea It 1* aceented apon thia a, and added to a weak form of root: 
thni, Qao&dtayOi, pp^AdbrlU, dhly&dhyftt. huv&dhySl. But the form 
of root ii the itTOng one In a fow caaet: name!)', Qay&dliyU, BtaT&dh> 
;ai, tarfcdliyU, Jar&dbyU. mandidhrU, Tand&dhyU. In half-&- 
donen fonna, again, tha root bas tbe accent: namely, kqiiradhyU, g&madl)- 
yU, yJOadbySl (bat once oi twioe alio yi^idhyU), v&hadhyU, 
Bibadhyll, bhiMdhyU. In a single Inatance, plbadhyU, the anfflz 
la added dlatlnetly to a present-atem ; and In ons, vXv|41i&dliy&i, to a 
perfect atem. Finally, In a number of Inalaneea (ten), thti InSnitlve ti 
made from a oaaaatWe atem In ay: thna, m&dayUhy&i, ri^ay&dliyll, etc 

a, Thla mflnitiie ia by no meani rare In RV., balng mado In thlrty- 
nT« different form* (with terentr-two oocnrreneei). Bot It ia hudly known 
ODtalde of tbe RV. ; tbe AV. bai It bnt once (in a pauage fonnd alao In 
RT.}; and eltewhere b*lf-*-dozen examples have been noticed, in mantra* 
pataagea (one ot them T3. faldely reads gimadhyo); in the Bnhmana 
language proper II appeira to be entirely wanting. 

977. An example or two ara mat with of an InftaiUvB in ;yU: thus, 
t6U^U(TS.), aT7atht9yU(K. Kap.; US, avyittaife; TS. Tyathifat], 
and perhapi -dbSayU (PGS.). 

978. Tbs inflnltlva 1° 9&9I *'«■ if&^i C) f""^ /W f"*^. -bhOfiifi 
from ybtaQ; qOfi^^i trota j/fa 01 fvS; naf&i^l from yni; aakfi^ 
from t^Bah; par9&9l ^m yfj, ttaeif&I}! from ytf; and gp^Ilffagd and 
-stp^f&fl from V'v'ET *"d "tp — the last containing eiident preaent tenae- 
stgna (compare the lat alng. gf^^fi, 884 d). 

979. Tbe only InflnlliTe In taxi la dhart&rl [with Itt oomponnd 
TldhaiiAri), from ydhp. 

Uses of the InflnitiTefl. 

980. The uses uf the so-called Infinitirea are for the most part 
closely accordant with those of the correBpondiiig cases ftom other 
abstract soniia. Thus: 

981. Tbe acoasalive, which is made only from the root-nonn and 
the nonn In to, is used as object of a verb. 

a. Gapecially, of forms ftom the root fak bt dbU, and arh b* icorthy, 
Aofa the right or the power. Thas, ^akdma tvh aamldham (RV.) may 
to* (Keompliih thy hmdiing; xai ^akaii pratldta^ {qam (AV.) may they 
not ba able to Jit the arrow to the ttring; mAno vt imidi aadyi^ P&ry- 

u,j,i,._.., Lioogle 

881—] XIII. VutBAii Adjbotives and Nouhs. 352 

iptum ftrbati m4na^ p&ribll&Tltum (TS.) Me mind, fortooth, eon at 
oaet attain and turpait her; k6 by iti»jA 'rhatl gdliyatfa nlboa cr&- 
bltum (QB.) for who u worthy to takt hit iteret name T la the Y«da, Iha 
conttroctiOD with ihe«e Terbg 1b only od« tmoDg Mhen; Id the Bnkmtoa, 
it bacomei the greatly pretslent one (thiee qasrten oi mon of all the cue*). 
b. Further, of lerbs of motion (next moit rieqnent cue); thaa, 
dik^ftni hittua stl (T9.) h» got* to taeriJUe thing! pertaining to 
tacrijicial gtfti; indraifa pratlram omy jfu^ (RV.) / go to Indra for 
(I, e. hiMtch of him) tht Ungthening out of life; ^of ydhf peritit m, 
undertake: as, ak id&ih Jftt&ti s&rvam ovk digdhtuh dtMlbre ((B.) he, 
at toon at bom, began to bum thit untoerte; — of vecba meaning datire, 
hope, notice, know, and tbe like: u, p&fSn vlo^aiij vettlui B&rr&ii 
(AV.) thou knowett how to looten all bonde ; t&amfid agnidi ai " drijeta 
p&ritaantuin (VO therefore one thould not be careful to rmother tkt 
_^r<; — and of others. 

96fi. Of the iafinitiTfl datives, the fandamental and uaoal sense 
is that expressed by for, in order to, for the purpote of 

Eiamplea are: viijvaA Jiv&ih oar&sa bodbiyuitl (RV,) awakening 
every living creature to motion; titn Apa ySta plbadhyKl (RV.) come 
to drink them; nsi t&ifa te Aevi adadur fcttavs (AT.) the gode did 
not give her to thee for eating; pr&f "d yndb&ye d&syuili {ndra^ 
{yi.\.) Indra went forward to fight the demon; C&kfor no dbehl Tttbyftl 
(RV.) give ui tight for looking abroad. 

Some pecnliar eaiiatmetloDi, howeTsr, grow ont of thU me of the In- 
floltlTe datlTe. Thui: 

a. The nonn which i( iogic&Uy the aabject or itbe ot^ect of the action 
•ipretsed by the InflnitiTe \t frequently put beside It in the d»tlTe (by a 
conatmctien wbl«b i« In part a perfeotly simple one, bnt vhloh is stretdiad 
beyond It* natnral boundwies by a kind of attraction): thus, oakira 
atiryftya p&ntbfim AnvetaTit n (RV.) he made a track for the tun to 
foUow (made for the eun a track for hit foUtrwing'); fiqit* qt^^C^ 
r&Itqobhyo vlnfkfe (RV.) he wheit hit home to pierce the demont; 
mdraya dta&nur t tanoml brahmadvlfe q&rava taintariE a (RV.) 
/ ttretch the bow for Mudra, that with hit arrow he mai/ ilay the hrahma- 
hater; BSm&bhyath d^qiiye sdrySya punar dStftm &amn (RV.) mojf 
Iheg grant life again, that we mag tee &e ran. 

bi Ad inllaite with y'kp make U uied nearly in the aence of ft 
cauiatlTB Terb: thna, prit "ndb&ih gro^&ih oAkqasa 6taTe bftba^ (RV.) 
jfe make the blind and lame to tee and go; agnlth aamidbe oaUrtba 
(RV.) thou haet made the fire to be kindled. Of almiUr cbaraeter la an 
occasional constmcUon with anothei verb: as, y&d im u^m&ai Urtava 
Uurat t&t (RV.) what we with to be done, mag ?ie do Uof; kavtbr 
iecbfiml aaifadf f « C^V.) I deeire to tee the lagee. 

a, A dative ioflnltlTe 1« not soldem nsed as a piedlaate, aometline* 

ioy Google 

853 Uses of the Inpihitives. [—984 

wltli, bnt mora luutllT wltbont, > copuls eipnuedi thus, ogidr IvK n& 
pratldta^ bhaVHti (TS.) likt Jire, he m not to be resitUd; mabimA to 
auytaft n& fluhnA^e (YS.) thy graOnttt it not to ht attaintd by anothtr; 
nAUm {ndro nlkartsTO ait qaktUf p4rlf&ktaT« (RV-) Indra i» not 
to b* put doum, &t mighty ont it not to he overptneertd. 

d. aometlmM ui inflnitlva aa naed wlthoat ■ copula hu quite neaily 
the itloe of an Imperative: thua, tyi me yofdeft . . , KaqlJ6 liuv&dh;&l 
[astl] (RV.) thttt glorious one* thaU the son of Ucy invoke for ms; 
sOktAbbir va^ . . . {ndrS nv ogni &vaaa bnTAdhyll [Bta^] (RV.) 
with your hymni thall ye call now on Indra and Agnifor aid; TUadidhy & 
■gniiii ii4iiiobM^ [BBml] (KT.) Ut me griet Agni with homage ; asmakO- 
•■9 Oft BBjriyo vit^vi a^fis tariffci^l (RV.) and let our laerificert erou 
aU regioni; tin nii VAih k&rtBT&I (HS.) that mutt not be done to; 
brnbinadvlfa^ ^rave h&ntava u (RV.) let the arroui ilay the brahma- 
haiera. The liiflnitifes in dhySl and ^a^l (which latter is io all Its asei 
aecoidant with dttlvei) ara thoie iu which the ImperatlTe lalae Is most 
diitlDotly to bs reoognited. 

a. In the Btshma^ai and Sutras (especially In ^B.) the dative in tavU 
is not iBldom used with s vetb signifying y>eak (brQ, vao, ab), to sipreis 
the orderinf of anything to be dooe : thns, tAamU Afadbln&m wk m^llLny 
Aoobettav&I brttyfit ((B.) therefore let him direct the roott oftheplantt 
to be eut up {ipeak in order to l?ieir cutting up : ct. y^ ywjtjt &dftiijlya 
T&daaU who ditmade from giving the cow : AV.). 

888. The ablative Infinitive — which, like the accusative, ia made 
only from the root-Douo and that in tn — it fonnd especially with 
the preposltlonB i until and port b^ora. 

a. Thna, t timlto^ (TS. etc.) until eoAouttion; puri Tftoitk pr&- 
Tftdlto^ (TS.) before utterance of the voice. In the Biahmans langaage, 
tUs ii the wetl-ni0 exclnsive conslnictian of the sblatiTe (it ooonrs also 
with pTSkt-arrBlc, etc); in the Yeda, the lattec is nsad also after ^ 
without, and after several verbs, as trft and pft protect, yn teparate, bbl, etc. 

b. In a f^ir instancei, by an attraction similar to that illnstisted 
above for the dative (982 a), a noun dependent on this inflnltive is pnt in 
tha ablative healie It; thns, purS vBgbhyah aampravadito^ (PB.) 
before the utterance together of the voieet; trodbvaih kart^ avapida^ 
(BT.) taee ue from falling down into the pit; purK dokfi^bhyo aeto^ 
(Apaat.) before the gifti are taJeen away. 

884. The genitive infinitive (having the eame form as Che ab- 
lative) is- in common use in the Brahma^a language as dependent on 
I^rarA lord, matter, employed adjectlvely In the sense of capable or 
likefy or expoted to. 

a. Examples are; td [devUV?] Igvari enoiii pradihah (TS.) 
they are likely to bum him up; itba ba vt IqvarA 'gniifa tAtvi kiih- 
old d&tirlt&m jtpattor vf vi hvUito^ (QB.) «o m truth ht it Uable, 

Wkitmj, Onmmar. 3. ed. 33 


884—] XIII. Verbal Adjbctivbb amd Nouns. 354 

after piling the Jiri, to mtet miffi tome miAap or otA<r, or to ttaggtr; 
IfTKT&ih tU rattaanUiram ndgfttoQ cakfu^ prunfttUtob (PB.) tht 
rathaniara it Hahle to Anoeit out tht tya of the chanter. 

b. Tha dtttia la nied In gB. iDpt«*d of tlie genitive In & iliifla 
plitue (I^arK& J^ATltaT&Ol *^^> li ^^ lUar l&ngnaga, lometlmM tke 
ucQiattn in torn. In 1 cue or two the ntao. ring, nam. I^Tftrah i* 
Qied, wlthoat regiid to tlie gander or nnmbei of the iraid vhleli it qoallflei: 
thai, tAcye "^ari^ prajA ptpiyASl bh&vito^ ((B.) hit progvng ii 
UtAh to deteriorate. And in ■ very few initances tha word I^ram li 
omitted, and the genftlTe bai the aame lalne withoat It: thai, dTB nudhy- 
^ThiUnfctw abhi pnttyato^ (AB.) two mag b* added to the noon i^ation ; 
ttto dlkfiU^ pfimanA bhAvlto^ ((B.) then the comterated m liabU 
to get the iteh. 

0> This oonatmctlon with Igvara, which is the only one for the geni- 
tive inflnltiTe In tha BtEhmana, ia unknown In the Yedi, where the genl- 
tlTB is fennd in a vsiy amill anmber of eiamplea with madhyi, and with 
tiie root Iq: thug, madliyi Urtol^ (RV.] in the midef of aetion; jfe 
rSr& ^tol^ (If^-) ^ M maater of the giving oftotatth; Iqa y6to^ Ot^O 
M (Me to keep away. 

686. UnleM (he inHntttvea In (ai^ and tari are locaUTe In form 
(their naei are those of datlTea), tlie locative tnflnitive ii *o tare, and haa 
10 little that ia pecnliat in its use, that it la hardly worth making any 
account of. An aiunple U Of&so bndhl (RV.) at the awakening of the 

8S6. In the Teda, tbe dative infiDttlve forms are very macli 
moie ntunerona than the aecaBative (In RV., their ooonrreDces an 
twelve timea as many; in AV., more than three timet); and the ac- 
ouMtlve in torn Is rare (only fonr formB in RV., only eight in AV.). 
Id the Brahmanas, the aoeiuative bat risen to comparatively much 
greater freqnenoy (Ita forms are nearly twice as many as those of the 
dative]) bnt the ablative-genitive, which is rare in the Veda, has 
lUso come to fnll eqaaiil? with it. The disappearance in the olassical 
language of all excepting the acousative in turn (bnt see 868 h) is a 
matter for no small anrpriae. 

887. The later infinitive in torn is ofteneet used in oonstmctlong 
oorrMponding to those of the earlier acoosative: thns, na vifpam 
aqakat so^tun. he could not rutrain hie feari; tuii drnftom arhoai 
thou oughteet to tee him; pr&ptma iochanti they de»ir» to obtain; saiii- 
khy&tuia ftrabdham having begun to count. But also, not infrequently, 
in those of the other cases. So, especially, of the dadve: tbns, 
avaathfttuih sthKnftntaradi ointoya deeita another plate to ttay in; 
tvftm anve^tuni ibS "gata^ he hai come hither to eeek for thee; — 
but likewise of the genitive: thus, samartbo gantiun eapabU of 
going ; aadidhBtaiD I^vara^ able to mend. Even a construction as 
nominative Is not unknown: thus, yuktaih tasya mayK aamaqvK- 


355 GBRmnDS. [—WO 

Mtyittah bUUTltn (HBh.) it it proptr for me to comfort hit wife; 
na tuptlraifa avayaih nyATraifa 9aptiim eT«m (R.) it it not luitable 
thut to cur*« one't eten grmdton; t»d vaktiuh na pftTTate (Qatr.) H 
it not poisibie to say tiuf. 

ess. In the Utei liDguag*, u la the etrlter, the InflnltiTe Id eei- 
Uio counimtlaDB hu whit we look npon >i e puitTe vilae. Thai, kmrtnm 
irftbdhtt^ beffutt to he made; qrotudi na ynjyftte it it not ^ to ht 
heard (for hearing). Thla it eipedillT frequent ilong vlih the putiTa 
fomi ot y<}ttk: thni, tjaktmh na qakyato )f eatutat be abanAmtd; 
faky&T ihS "netnm M<y two can be brought hither; na oa vlbbfltaya)^ 
qakyam avftptum orjita^ nor are mighUf lueeeteei a thing capable of 
being attained. 


888. The so-called gerund is a steieotyped case (doubt- 
leas instnimeDtal] of a reibal douh, used geoeially as ad- 
junct to the logical subject of a clause, de&otiug an accom- 
panying oi (more often) a pieoediog action to that signified 
by the verb of the clause. It has thus the virtual value of 
an indeclinable participle, piesent or past, qualifying the 
aotoi whose action it describes. 

a. TfanB, for example: <}TutvU 'va oft *bmvaii and hearing (or 
hatting heardi they ipoke; tebhya^ pratlJ&ByK thU tin parlpa- 
praoAa haoing given them hit promite, A« then quettioned them. 

890. The gerund is made in the later language by one 
of the two suffixes ^ tvS and U ;a, the former being used 
with a simple root, the latter with one that is compounded 
with a prepositional prefix — or, rarely, with an element 
of another kind, as adverb oi noun. 

a. To thie dtstiibntioii of uaei betweea the two suniiei tbeie ire 
oeeulonal exception!. Thvi, gerund« In ya from •imple root* m not 
Ter7 rate In the epic languags (e. g. grhya. X^Jt^ IV^** dwell], ar^a, 
Ikfya, dn^a, tynjya, lah^ya; sI»o from Mui*tlT«s and deaomlnttiTet, 
u Vftoya, yojya, pUVTB)> and "^ not anknown elievheie (e. g. aroya 
uti ikfya H., prothya AQS., sthftpya QvD.). And gamndt ia tvS. 
from compounded looti eie met vltli In conelderable numben from AT. 
(oalj pratyarpayltvJl) down: e. g. samlrayltTit HS,. virooayltvA 
TA., utkftptvK U., pratyiiktvs S., pra^aaltvS S., prahaMtTft 
HBh,, saihdarfayltvft HBh., vltnuktrft R., nivadayltvK R., prsktrS 
Psiic., anupltvl TBS. : the great mijoilty of them ue made from tlie 

Dgii.ze.., Google 

eoo— ] Xlll. Verbal ADJEO-rivsa and Nouks. 356 

b. Tbe pretliton of the negatire psrtide, a oi Kn, doet not oaute 
the gerund to take the form In ;■: thus, lAftvft, &mrayiWft (bat R. 
has aeln^a)- Of eompouiidg with other than veib>I prefliei, BV. hu 
punardirt^ karpag^byBi ptdag^hya, haatag^bya, araifakftya. 
akkhalik^yB, mithasptdliya; AV. bia further aamaak^jra. 

881. The suffix ^ tvB has the accent. It » ustullf 
added diiectly to the root, but often alao with iutetpositioo 
of the auxiliary vowel ^ i — with regaid to which, as well 
as to the form of the loot befoie it, the fotmatioD nearly 
agrees with that of the participle in ?T ta {8BZ if.]. 

a. Examples of tbe general accordance of pauive participle, in- 
finitive, and gerund in regard to tbe nse of i were giveo above, 
SflSa; farther Bpecifloations are called for, aa follows: 

b. The quotible roots In virfibla p (S4S) change it to ir: tho*, 
tartvi. aUrtri (alio atrtvtji and oar makea alio oirtrS (like cSn^); 
— loota Id S ihow in general tJie lame weakening aa in the partlelple; bst 
from db&put ii quotable only dhltrlt (MtrS), from m&meaturt mltwi 
and ml tv a, tiomda^io« only dattv^ from ohfi obl^tTfti — of roots in 
am, kratn. and bhram and yam make foinu both with and without i 
(u in the laflnltiie), bnt ram haB ratvit and raihtrS, and dam and Tain 
make damltrft and TamltT&. 

C. The auxiliary vowel la taken by loota gras, muf, fap, and (Ba 
(qSaltTB) (vhoie participle* have h«th forma); alto by oSy, nft [nar- 
titT&), lag, and Brai (agalnet analogy of pplc); and foo makei fcxritrk. 
On the other hand, from ruj (rug^) and vraqo (vrk^a) come raktri 
and Vfft^^ And both forms are made (as alao in Inflnltlve or partioiple) 
ftoni oar, vaa duell (u^^vK, o^tri), ni (nltvA, naritvft), and m^J 
Cmnfv^ mSTjitrft). 

d. While tbe formation ia In general one requiring, like the ptiaiTe 
participle (e. g. Qptrft, like upt4; odltvi, like udlti), » weak or weakened 
toot, there are lome caaea lo which it ia made from a atrong or stiengtli- 
ened root-form. I^hna (beeidea the Instancea already given; ohX^Cvf, 
raditvA, gfisitrfi, c&yltTft, fooltTfi, naTitrft, mBrJltrt), w« And 
oharditTfi (Apaat.), daAftrft, and epharitvft, and, from a nnmber of 
Toats, a second strong form beside the more ragnlar weak one: namely, 
a&ktTft, bhai^tra, bhufiktrS, syanttva (beilde aktvA etc.) ; oayltrS. 
BmayitvB, BmaritTft [beside oitrlE etc.); rodltrft (beside mdltvaj, 
and BiAoltvft (beaide siktTS). Tbe last shows the Inflnence of the 
preaent-stem ; as do also mBijltrji (above) and JlgtirltrS (yghrt). The 
form ffhutra (Apast.) U doubtless a false reading, for fthyiltTa. 

892. The suffix ? ya is added direotly to the lOOt, 
which is accented, but has its weak form. A root ending 

ioy Google 

3fi7 Gerqhd in ja. [—993 

in a short vowel takes flT tya instead of IT ya: thus, %f7I 
-jitya, H^rU -stiitya, ^ItTI -ktty*- 

a. Roota in Tulable f (S4S) change that Towel to Ir or Or: thua, 
klrra, girya. ttrya (and tArTa), dizya, pOrya, qlrya, Btiiya (also 
Btftya); — roots In ft have for the mo*t pin •ftya; but dbft >uol makes 
dl^a, and doable fbrmi ue foDnd tiom g& awi^ (gfiya, glya), p& rfrin£ 
Qpltya, piya), dft giee (dtja^ d&dya), dft dioide (d^a, dltya), ma 
maamrt, txchaitgt (nUtjra, mitya), b& bind (elEra, aya); li c/iVijr baa 
l^a or Uya, m It an ft-vetb; and kban and dbam maha khSya and 
dlunltya, trom their R-forma; — the roots In an and am making their 
participle in ata (954 d) make the gerund in a^a, bnt also later in anya, 
amya (e. g. gitya, gamya; h&tya, hanya; but tan makes as aeeond 
form t&ya, and ^m ram only ramya Is qnotable); — the roots In it 
add ya to their Iv-form; that, tfthlTya, aivya; — a few loola In i and 
a add ya to the lengthened vowel bealdea adding tya: thne, 1 go (lya, 
I^a; alao ayya), ol giOhtr (dya, oi^a). and plu, yu unite, bh, atu 
(pl^ya, plQ^a, etc.); while b^l dtttroy baa only k^iya. 

b. Thia gerund, though accented on the loat-ayllible, Is generally a 
weakening formation: thus are made, without a strengthening nasal fonnd 
In some other forms, Aoya, Uya, idbya, udya, ubhya, sratbya, t4oya, 
da^ya, bidhya, bhajyo, lipya, liipya, vl&gya, ^rabbya. aajya, 
sk&bbya, stibliya, syodya, Bvajya; with weakening of other ktndi, 
K^hya and g^bhya, proobya, lioya, udya, upya, ^ya (tab dirdl), 
6hya, vidtaya, viya, v^qoya, Bp^dbya, hfiya; — but fram a nomber 
of roots are made both a itroDger and a weaker form: thus, maathya and 
m&thya, mErJya and m^Jya, mndbya and rddbya, qaAsya and q&a- 
ya, fSaya and qlfya, sk&ndya and skidya, ST&Asya and araaya; — 
and only atrong fonue are tound from raote aro, av, ofty, 91 (qayya), u 
well u from certain roots with a constant nasal : e. g. uflob, kamp, 
nand, lamb, qaftk; isolated cases are oijya (y^nq burn), prothya (alao 

0> Other gpeclal eaaw ate nllya and tlbya (V'uh remove), guTya and 
gtbya, gnhya and giUiya, rubya and rOhya, bhramya and bhrSmya, 
&yya (beside {tya, lya), gbrSya and Jighrya; and ilr^utya (beside 

ess. The older langnsge haa the same two gerund formations, 
having the same distinotion, and used in the same way. 

a. In RV., however, the Bnal of ya la in the great m^oilty of In- 
atanses (fully two thlrda) long (at if the Inatrumental ending of a derlv' 
ative noun in i or tl). In AY., loug ft appears only once in a RV. 

b. Instead of tvK (ilone, the Veda has three fbrms of the anfili, namely 
tv^ tv^a, and tvl. Of these three, ttri 1> decidedly the eouiinonast in 
BV. (tiilTty-flTe oocnrreneei, against twenty-one of tvS); but it la unknown 

Digitizecy Google 

raS— ] ZIII. Verbal Adjbotivbs and Notnts. 358 

In AT., ud *eTT rute elnvhere In the oldei luigatgB; tvtya li fosnd nlns 
timet Id RT. (enlr onc« onUtdt the Centh Book), twice In AT., >nd but hair->- 
dozen times elaevhete (In QB., onee from » eiQMtlie tiem; Bpft^ayltT^rs). 
The hiitoriul reUtlon of the thiea formi Im obscnie. 

O. T>o other gerund infSzei, trftiuun and tvlnam, ue ueDHonod 
b7 the gnmmirluia m of Tedic nae, bnt they hiTe noirhera been bmnd 

894. The nse of this gerund, tboofrb not ch&ngliig !a Its chtr- 
Kot«r, becomeB much mora freqaent, and even exowaiTe, In the liter 

». Thni, In the N&ta *nd BhagsTtd'Qiti, vhlch hiTe only one tentii 
u many Tetb-fonns u KV., theie >ia more than three timea m many es- 
amplea of the geiand u in the latter. 

b> In genera), the gemnd ts an adjunct to the anbjeet of a aentenee, 
and expreaiea an act or condition belonging to the anbjeet: thai, T&Jre^a 
hatrlt nfr ap&^ eaearja (BT.) smiting teith hia t\undtrboll, Ae pourad 
forth the wattn; pltvf sAmura vtTrdhe (RT.) having drvnk of the 
soma, hs waxed strong; th ji^fi&aya r&eaifa dtaltra TldoliTa y^tUjfa 
Tllpina yopayltTit tir& 'bhavan (9^0 ^^ing sucked out the sap of the 
offering, hoeing miiked the offering dry, having blocked li teith the saerifieitd 
post, they disappeared; fratrll 'va Oft 'brnvan (HBh.) and having heard, 
they said; taib oa dOre drftvft gardabhl 'yam Iti iftatvS dbSvita^ 
(H.) and having seen him in the distance, thinking 'it it a the-ats', he ran. 

a. Bnt if the logical lableet, the real agent, 1> pat by the conatmotion 
of the Mntenofl In a dependent caae, It la atill i^ntllfied by the gernnd: 
thai, Btrlyaih drffv^ya kitav&lil tatXpa (RT.) it distresses the gambler 
(i. e, the gambler is distressed) at seeing hie ujife; tkia li&i ^ladl dntri^ 
bhir Tlveda (^B.) fear came upon Kim (I. e. he was afhiid) when he 
Sato him; TldlLfiya pTOQlte vfttlm [U.] when he stay* away after provid- 
ing for her nyipori; kiih na ms a;ad idaifa k^ft (HBh.) vihal, I 
iconder, teould happen to me if I did this; — and oipecially, when a paaiive 
form la given to the aeotence, the gerund qnallOea the agent in thelnatmmental 
caie (ass a): thna, tat«t> Qftbd&d abhijtkftya ea vrBahre^a hata^ (H.) 
iherei^on he was slain by the Itger, who recognised him by hit votee; 
trayK sa rWi yJrontaHUfa paraakfty^ vakta-rra^ (^.) presenting 
(^akuntala, thou mutt say to the king; ha&atoftifa vaoanaih fratvS 
yatbt me (gen. for initc.) nSlqadbo vrtal^ (HBh.) at the Nithadhan 
was chosen by me on Hearing the words of the swans: thla eonetractlon 
Is extremely eommeo In mneh of the later SanikilL 

d. Oeoaiionally, the gerund quallllea an agant, eepecially an Indellnlte 
one, that ia Qnezpretied : thus, tada trU *v« paktvt kmUUtawTal; 
(H.) then A« shall bs eaten [by im] cooking him on the spot ; yad anyacya 
pratiJiUya punar anyaaya dlyate (H.) that, after being promised (lit 
when one has promised her) to one, the it given again to another ; anotntTa 
00 'ktaib aavloilya yat kftam (H.) what one says after mature thought, 

ioy Google 

359 Obe& of thh Gbrumd. [— OdB 

and Ami afttrjiill MA*ration. Henca, itiXt mora slUptloUr, ifter alam: 
Ihns, aluil TlaBrrft (^.) tnough of httilation; tad altufa te vanub 
- Bmtvft (R.) to hoot dong tmM jrom; to tkt fortti. 

•. Otkot 1«M isgoUT BoiutiDctiDai HB mat Tltb, eipeeUllf in tha 
■Idac Uapuca: thnt, in tha maoiiat of a partlelple witli man *nd t)i« lllie 
(see a), u Udi lilAaltTA 'va men* ((B.) h» Oouf/U kt had luirl him; 
M adbbir abhlfloya aijItsrU 'vR 'maajrata (AB.) hoeing tprinkUd 
then with icattr, Aa btlitvad himttlf to have gxhauittd them; — In Iha 
Dioaat of a putleiple faimlag a ovotinnoai tanie «lth yi (1075 a), aa 
Indram erU tftlr KrabbTa yaatl (AB.) by meam of them they kt4p 
taktitg hold of Indra; — aa qnalifjlng a saboidtnaU mamber of tha ientsnoa, 
aa poro^ltqam wb. kOrmAtb bbtltrlt B&rpaatam (^B.) to th« taori- 
fieial eake creeping lAotU, having become a tortoiti; ^odtaj&m . . . 
saphsnKih BBaTantih bltQtva Jalormlm Iva (B.) into Ayodkya, Uke a 
eurge that had been foamy and roaring; — even abaolatelr, at atithytaa 
vti dovi Iftvi t^t Bam&d avlndat (f B.) when the gode had laeri- 
Jiced with the gttett-offering, ttrife befel titem. 

t. At In the two examplti btforajthe lait, a ptsdloate word with 
bhatvA is pat In the lame oaae with the Bubjaot: thna, faitfaat, t4d If&m 
OTti t&d bbQtvi yt^atl ((B.) to having thvt bteomt thit earth he 
maltM offering; yena vSmaaenS *pl bhatvft (Vet) by whom, even toAen 
At had become a dwarf. The coultnictioa li a laie one. 

g, A numbeT of gemtida haie theii meaning attenoated ionietlmei to 
the temblanee of a piepoaltlon or adTub: inch art adhikftya making a 
tu^««t of, 1. «. retpeeting, of; Uiya, apBgfhya taking, 1. e. leiih; ud- 
dlfya pointing toward, 1. e. at; SaSdja, arriving at, 1. a. along, by; 
Irabhra beginning, 1. a. from; BatnbhOya being with, i. e. with; sadihatya 
ttriking together, L t. in uniton; prasohya tuing force, 1. e. violently; 
tyaktrK, parltyajya, mnktrK, vlbBya, uddh|^at TarJaTltvt leaving 
out etc., i. e. excepting, withotU; and othtia. Eiamplea are: faknntaUm 
adhlkf tya bravlml (y.) I am epeaking of fakuntala ; tam nddif ya 
kqiptalagufal^ (H.) having thrown the cudgel at him; nlmittaih klifaold 
Ssftdya (B.) for tome reaton or other. 

b. The geniDcl la Id the later langaage aometlmea fonnd In compo- 
aitloD, aa It a nonn-stem: a, g. praaabyaliara^ taking with violence; 
pretyabhSva exiifence after daath; vlbhaJyapStha eeparate enunciation; 
Kam'tiixviyhgiaaKa.ti going together. It la alio often iiipeated (ISdO). tn a 
dlatribntlTt teote: e. %. b4 vk{ ■ammfJya-BanuiiiJya prat4pya-pra- 
tapya pr& yaoobati ((B.) in eaeh caie, after wiping and warming them, 
he handt them over; gfhltv&-gfMtv& (K^iS.) at aocA taking; TUUUUnyo- 
*iuuuiiya (PaKo.) mery time that Ihey arite. 

Adverbial Oenmd in am. 

WB. The aoonMtive of a derivative nomen actionit in a, naed 
adverbially, BMomea aouetiiues a value and conatrvotion ao aeeord- 

ioy Google 

998—] XIV. BscoNDAxr CoNjnaATioH, 360 

Knt wltb tbftt of tbe ninal gerund that it oannot well be called by 
a different Dame. 

a. No ei4iDple of 4 paoDltai geiondltl eomtiactlon wllh (uch > foix 
oecnn either in RV. or AV., altbongh i doien 4dvetbl>l (coaiatiTea ue U 
be clu«ed u repteienting the formitlon: thni, abhy akrittn am , pr(Lt&&- 
kam, pra96dsm, nUltr<U>l> abMak&ildam, etc. Tbis genind la fonnl 
etpecimr io tbe Bribmau** end SQtru, wbeie It ii not nre; in tfae epia 
It ia eitr«melr Infreqaeat; Utai, eUo, It ocsnn -very ipulnglr, 

b. A floU TOwel h» TTddtal-attengthening before the anttx: tliot, 
n&Tam. qrivam, kSrtun; fln*! ft edda j: thai, kbTayun, rBT^m; a 
medial Towel hu gnvti (if oipable or it: 940): thoa, kfepain, kroqun, 
TUtam (bat Ikfuu, pOrain)! > medial a before a aiogle eontonaiit i« 
lengthened: thna, kramam, oKraiD, gHUutm. BVldam (but Entutliaia, 
lambham). The accent la on tbe radleal lyllable. No nnoompoanded ei- 
emplea are found in the older langaage, and eibemeif few in the later. 

O. Ezamplea aie: k^aih vi tmioj Ugftni vjfttT^e&ifa fete 
((B.) ht lit changing tAa poiition of that limbg at pleaaurt; ^ttarfim- 
Qttar&b fAkhSiii aamMAmbhaifa r6het ((B.) be aould climb, talcing 
hold of a higher and ever a higher litrib; apari^n mahfaiftgfan ivft 
■bhlBaduiraiti didfkfitba^ (9^.) hareaftor, rvnning together at it tuere 
about a great snake, th«g toiB with to tee htm; nltmSn; fta&m ettoi 
ngmagrfamm (9^0 filh teparate naming of (AeM their namet; 76 
▼ipaijAaam &Tag6hati (^B.) tahoevtr buriee it t^tide doton; hilhfltV^e 
padi krandittuh praTrttA (^.) the proceeded to cry, throteing tip her 
armt {wiOt arm-totting') ; navaofltapaUATftnl darf adi-dar^aiii madha- 
karSQiih kvai^ltftni {rSvadi-frRTaih parlbabhrftma (DEC.) he 
wandered about, constantly teeing the young thoott of Q^e mango, and hear- 
ing the humming of the beet. Bepeated fonna, like thoae in the tut «s- 
uatile, are appiofed In the later langnage; they do not ocent earlier (bat 
inatead of thetn the repeated ordinaiy gennd; 994 h). 


996. Sbcomdart oonj ligations are those in which a 
whole system of forms, tike that abeady described as made 
from the simple root, is made, with greater or less com- 
pleteness, from a derivatiTe oonjugation-etem ; and is also 

ioy Google 

361 Piiwm. [— OM 

usually ooonected with a oertaia definite modification of 
the original radical ienae. 

ft. We luTe Hen, Indeed, th«t Ui« tente-aratetnt ue *lio tot the moat 
part made tnm dsrtiatlTs-steiDt; eed e*en that, In lome otiei, aoch stenu 
■Mnme tke appeuuice and -ralae of looti, and aie made the baifa of a 
completB ooojagatlonal gyitam. Nor ia there any dlatlnet dlvUfon-llna to 
be dravn between tenta-cratenil end derlTBtlTs oonjngationa; tbe Utter tie 
praent-aytteDU which have heen expanded into eonJngatloDi by the addition 
of other tonne, and of parUolplea, InflnltlTei, and lo on. In the eaiUeat 
language, tbelr forma ontalde ef the preaenl-ayatem ace still qnite raio, 
hardly more than apotadlo; and eien Utei they are — with the exception 
of one or two formatlona wUoh attain a comparaHie freqnenoy — maehleae 
common then the correapondlng forma of primary eonjugatlon. 

907. The secondary conjugations are: I. PaMive', 
II. Intenme; III. Deaideiatire; IV. Causative; V. Denom- 

a. The paatlie la olaited here ti a aeeondaiy oooJogatlon became of 
Ita analogy with the othen In reapsot to tpeciflo valae, and tteedom of 
formation, altbongb It doea not, like them, make ita foima outaide the 
present eyatem ftoro Ita preaent-atem. 

I. Passive. 

SOB. The paBsiTe conjugation has been abeady in the 
main described. Thus, we have seen that — 

a. It has a special ptesent-syetem, the stem of whidi 
is present only, and not made the basis of any of the le- 
mainJDg forms: this stem is formed with the accented class- 
sign IX r&,- and it lakes (with exceptions: 774) the middle 
endings. This present-system is treated with the others, 
abore, 768 ff. 

b. There is a special passive 3d sing, of the aorist, 
ending in ^ i: it is treated above, 842 ff. 

o. In the remaining tenses, the middle forme are used 
also in a passive sense. 

d> Bnt the ptailve nae of middle forma la not common; It la oftoDeit 
met with In the perfect The participle lo a great extent takea the place 
of a paat paMlTo lenae, and the geiandlie that of a fatnie. On the other 

D,j,i,...., Google 

998—] XIV. Secondary Cohjuqation. 362 

hand, In the oldeat Unguige (RV-)i middle fonna «f other pi«Miit-«;tMiBE 
ue In a «OD«lder>ble nnmbei of caaei emplored with ptssiTo meuifDg. 

«. AocordlDg to the gnmmariMkB, there majr be formed from Bome 
verba, for ptMive nao, a apeoUl item for the urlit (nd the two foMue 
syatema, colDctding in form with the peonlUr 3d aing. u)riat. 

f. ThuB, from yM (loi. 3d alng. adiyl), beside AcUtBi. dia^, 
dfttb». alao idlTlfl, dRyijr^i dfiyltJEhe. The permiwion to make this 
double farautlan extend* to ell loot^ ending tn towoIs, and to Ersb, d^ 
and lum. No inch paaalTo form* ooonr in the older langnage, and not half- 
a>doien are qnotablo ftom the later (we Hod adhjlyl^i and astblylfi Id 
DKC, and anJIyltMa In KnT*].). 

g. Aa to the alleged paialve inlleciloii of the pertphnede perfect, «ee 
below, 1079. 

h. Besides the participle Irom the piesent tense-stem 
(771. 5j, the passive has a past participle in rT ta (062), or 
^ DB (9B7), and future participles, oi gerimdiTes, of various 
foimation (961 ff.), made directly from the root. 

S99. As already pointed ont (28S aj, the laogu&ge, especinllr 
later has a decided predilectioD for the pusive form of the eentence- 
ThJB IB given In part by the use of finite passive forms, but oftener 
by that of the passive participle and of the gerandive: the participle 
being taken in part in a present sense, but more nsaally in a past 
(whether indefinite or proximate past), and eometimeB with a copula 
expressed, but mnch oftener without it; and the gemndive represent- 
ing either a pure fatura or one with the sense of neoesaitf or duty 
' added. A further example is: tatrU "ko ynvB brUuna^o dpyt*4* 
taib dntvS k&mena pl^itA aaibjfttft: Bakliy& agre kadtitam: uklii 
pnru^ "yaih g^hitvK mama mfttii^ aamlpam inetavra^ (Vet) 
thtrt ihe tatc a young Brahman; at lighi of him $lu ftlt Me pang* of 
Umt! tht laid to her frimdi "friend, gou rmut take and bring this man 
to my mothtr^. In some styles of later Sanskrit, the prgvailfaig ex- 
pression of past time is by means of the passive partioiple (thns, in 
Vet., an extreme case, more than nine tenths). 

a. Ai ill otber langntgee, a 3d aIng. puaive la treel; made from 
iDtraDBltive aa well as tranaitive verba: thua, ih& "gamyat&m eome AsfiUr; 
tvaya tatrftl 'va sthlraUtm do you ttondjtut there; oarvflir JUau 
jUUyo '4^atftni (H.) Ut aU fiy up wifh the net. 

II. Intensive. 

1000. The intensive {sometimes also called frequent- 
ative) is that one of the secondary conju^^ations whiefa is 
least removed from the analogy of formations already 

ioy Google 

363 Ihtbnsite. [— lOOS 

described. It is, like the piesent-system of the secocd 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 lew extended beyond the limits 
of a present-system than any other of the derivative con- 

a. The intensive conjugation signifies the lepetition or 
the intensification of the action expressed by the primary 
coajngation of a root. 

1001, According to the grammarians, the intensive 
conjugation 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, lOSe], 
and in general those beginning with a vowel. 

a. In fact, howBTaT, ioteDBiTss to the lat«i Uagutge are very ttn, 
■0 tUB tti*t 11 li burd to tell pi«at)«lr what tbIhb ia to Im glveii t« the 
iale( of tlia natlTC grammii relpectiDg them. Nor are the; at all commoa 
aailier, eioapt foompiratlTely) In the RT., whtch containa about aix leTtntha 
of tbe vbole nuinbei [lathei over a handled] ijDotable from Veda aod Btah- 
mana and Sutra-teitti AV. haa leaa than half ai man; aa RT., and many 
of them In RT. paaeagea ; from the later language aie qaotable about twenty 
of theae, abont forty mote, bat toT the moat part oulf Id i 

b. Hence, In the deicilption to be gl*eii beloir, the aotnal aapeet of 
the formation, aa exhibited in the older langaage, will be had primaiUy and 
eipeciilly In view; and tbe eiamplea will be of forma fouDd there in nee. 

1002. The strong intensive reduplication is made in 
three different ways: 

I. a> The redapIiOBtiiig syllable is, as elsewhere, composed of a 
single ooDsoDsnt with rollowing vowel, and, ao far ae the consonant 
is concerned, follows the miss for present and perfect redaplfcaljoD 
(B90); bat tbe vowel is a heavy one, radical a and r (<>r ar) being 
radaplicated with t, an i-vowel by e, and an a-vowel b; o. 

Eiamplea are: vkvad, babadta, qa^vas, rSrandli; dftdf, dJblbr; 
e«kit, tetlj, neni, vevU; qofno, popratb, oofkn, joha. 

II. b. The reduplioatiog syllable has a final coDgonant, taken 
^m the end of the toot With an exception oc two, tbiB coDsonant 
is either t (or its substitute 1] or a nsaaL 

ioy Google 


Extkmplei are: ooToar, oaloal, suvp, rnarmrJ. Jnrhnt oaOkrun, 
Jft&ghui, t»ABtui« dftadaq (ydaAq or daf), Jftl^abli (vJambli ar Jatth). 
tamtu ()/(«&■ or taa), niMiiiiun (_ynam), yubyun (yyam). The (mmI 
li >MiiDlUi«d to the inlttkl ooDionini, 

d. IrregnlaT fDimatlons of thli slua aie: irith > final othei than r 
01 n Id the itdnpIlcattoD, badbndb; with a Anal naaal in the redopll- 
eatton which ti not found in the loet, Jafigab (RV.), jafijap (QB.; and 
jaflgOyat PB, ii pethapa frem yga; the latei langaage haa fnittei 
dandah); with an aDamaloiu initial contooant In redaplt cation, jarbhni 
from i^bbor (compare the Tedlo peifeot Jabllira fiom ClibTt 789 b), 
galgal from ygtd ; with lariou tteatment of an ^ or KT'element, dardar 
and dardir, oarkar and oarklr, tartar aod tartar, oaroar and ear> 
our, Jargnr and Jalgol 

a. The roota 1 and f are the oDljr onai with Towel initial ronning an 
inteoiiTe item: i makes ijB^ (t PU., ODoa); f m»kea the iTTegalar alar 
or alf. A« to the ttem ^a, see below, 1031 b. 

III. f. The reduplication is dissyllabic, an i-TOwel being added 
after a float oouBonant of the ledaplicating syllable. This 1-TOwel ii 
in the older langnago short before a doable consOQant, and long be- 
fore a single. 

Examples are: ganigam (bnt g&nlgmatam), Tarivft, TanJvUi, 
oanifkad, aanifTau; navlna, daridTut (aDd tbe ptitidplea dividhvat 
bat ti-vltnat). A single exception as to the qnantltr of the 1 la d«Tl> 

g. This method of lednpll cation Is followed in the older langaafs 
by abont thirty roots. Thus, of roots haTing final or penultimate n (once 
m), and n in the rednpllcating syllable, pan, pban, san, avan, haa; 
gam; krand. qoand, akand, syand; ot roots having final or medial f, 
and r In the tednpUcating syllable, kf tnuita, tf, bh^i VT> mfj, mni 
V|j, Tft, Bfp; alio mine (malimlno); — fnrlher, of looti scsnming ia 
the reduplioation a n not foond In the root, only vah (QB.: tke gram- 
marians allow also kaa, pat, pad; and panlpad Is qnotable later; and A{$. 
has oanlkbadat, for whlek TE reads kJmikbunat) ; floaUy, of now 
haling u or Q aa radical Towel, with kv before the l-vovel, to, dhfii 
au. dTUt. 

h. In this class, the general rales as to the farm of the redapUeatipg 
eoniontnt (B90) are Tiolated In the case of ghanlghan and bbarSbbf. 
and ot ganigam, karlkf (bnt the regular oarlkr also occnra), kaai- 
krand, and kaniipcaad (bnt also oanlqkand occurs) j tlsc In tntfiiHinn- 

i. The reierslon to more original guttural form after the rednplieallm 
In oekit, and Ja&ghan and ghanlghan, Is In accordance with what takea 
plue eliewbere (31 e, 1). 

ioy Google 

365 Intensive. [—1004 

1008. The Bame root Ib allowed to form ita inteiiaiTfl stem Id 
more than one way. 

Thns, in the oldei lan^oaga, iidf aod dardf; d&dlir «nd dardbr; 
oMal and oarottr (and ottrour); tartar (and tartar) ud torltr; 
JafigNii and gtmlgam; Ja^hns and ghut^han; pamphan >nd 
panlphan; marmrJ and marlmiij; nuurmrg and maHmrgi Tarvrt 
■Dd TarlTTti Jarbhf and bbarfl>hT; dodhQ and davldhn; nonn >nd 
navlnu; b&badh and badbadh. 

1004. The model of normal inteneire infleotion ia the 
piesent-system of the leduplioating ooajugation-class (642 If.} ; 
and this ie indeed to a consideiable extent followed, in 
lespect to endings, stiengtheniag of stem, and accent. But 
deviations from the model are not taie ; and the foims aie 
in general of too infrequent occurrence to allow of satis- 
factory claasifioation and explanation. 

a. The most marked irregularity is the frequent insertfon of an 
I between the stem and ending. According to the grammarians, this 
IB allowed in all the strong forma l>efore an ending beginning with 
a consonant; and before the I a final vowel has ffu^-strengthetiiDg, 
but a medial one remans unchanged. 


1005. We will take np the parts of the preseat-syatem in their 
order, giving first what is recognized as regnlar in the later language, 
and then showing bow the formation appears In the earlier t«xts. As 
most grammarians do not allow a middle inflection, and middle foims 
are few even In the Veda, no attempt will be made to set np a par- 
adigm for the middle voice. 

1006. As example of inflection may be taken the root 
icl^ vid know, of which the intensive stem is ^f^i vevid, 
or, in strong forms, ^1^ vived. 

a. Neither from thit nor ^m any oltier root are more than a few scat- 
tering rormi Bctaally quotable. 

1. Present IndioatiTe. 

T&vedmi, T^vl^ml vevldv&a vevidm&i 

ioy Google 

1006—] XIV.- Sbcondarv Cohjuqation. 366 

T^Tetai. T^Tidlfl veiritth^ vevlttU 

i ^cnrT, ciH<;iif? eirclflH^ =lT*l<ia 

vbvatti, r^Tidfti vevlttia v^vldati 

b. From )^ha, the singulai foims with auxili&iy Towel 
would be sfi'^efifM i6hftTlml, Mt<^'^fl| jShnvl^i, sn^cj)^ 

1007. a. The fatmt found to the alder Unguiga *giee iiu-generil 
witb tbe pukdigm. Kxanplai in: l>t tiag., tsKXkarmi, Teraqmli 2d 
■lug., ftlw^. d&rdwfi; 3d ling., Uartl, dfidhartl, vevvO, ii«nekti, 
jBfighantl, Unlkraatti, ganlgaihtl ; 3d da., jKrbhft&a; let pi., nonu- 
maM; 2d pi., SsgiaOM; 3d pi., dStUmU. n&nadatl, bharibhraU. 
v&rvrtatl, d^Tldyutatl, n^nijati, aad, IrregnlBily, v«vlfanti; uid, vfib 
the kuiilUry yowbI, johavlini, oSka^Imli oika^IU, nonavitl, darda- 
Iltl, jarbhnritl. No item wttli dlMjUabio raduplleitloa UkM th« >uU- 
Ut; 1 In mj of ite foimi. 

b. A alnfle dD*l fonn with I tnd atrong •tem oeean: nunelr, ta^ 

o. The middle forniB found to oocor >Te: Itt dug., J&fnrre, nen^e; 
8d ling., nenlkM, aararto; uid, with ImgnUr Mceat, MtUcta, AMi^; 
with InegnUr loei of flnal ndiol nuil, n&imatei with ending e liutMd 
of ta, ciklt«, JiAgalie, j6ffiivo, yoyaT«, btbadhe, and (with iiregaUt 
acMut) badbadbd; 3d da., ■arsr&tei 3d pi., d6dlqata. 

2. Fresent SubjunotlTe. 

1006. a. Subjoncttve focms with primarj andinga am extremely me: 
there hsTe been noticed oiil; Ja&gh&iiSnl, JftgarBsl (AV.); and. Id tbe 
middle, tantaa&Ite (3d da.). 

b. ForcriB with aecoDdary eodingi are more fKqneot; that, 3d (ing., 
Ja&gbanas, Jalgolaa; 3d ting., jfigarat, oikltat, bobhavat, oArknat, 
J&flghauat, b&rbrhat, minnrJat, mkrmj^at, parpbarat, dardirat, 
oani^adat, davidyutat, Banl^va^t; lat dn., JafigbaniiTa; lat pL, 
oarkirama, vavldftma; 3d pi., pipatan, 96gaoan, oarklraa; and, 
wllh double mode^ign, oakaqSn (AY.). Of the middle are lonnd only 
3d penona plural : tfaua, JiSghaoanta, Jarbrfanta, marmrJanta, aona- 
vanta, f o^oanta. 

3. Present Optative. 
lOOS. This mode would show the unstrengthened Btem, 
with the usual eudiags (B66), accented. Thus: 

ioy Google 

I iM^'dW clicf^M ^ftyiH 

TOVldyim vevidySva Tsvldyama 

a. The opUdva U rapiesenled b; odI; an ezampla oi t«o tn Ihe older 
lufDif": tlini, actlTs, vavify&t (AV.), Jftgryaa (KB.), Jiffrly&t (AB.), 
JigfTiiM (VS. HS.; botJSerlyimaTS.); RV. huanly oftkanyftt (pft.?); 
middle, nsnljlta (K.). 

4. Present Imperative. 
1010. The regular forms of the impeiative, including 
the usual subjunctive first persons, would be as follows: 






T^Tsttn, v^Tidltu vevitt^ 


Oldei imperitlTe 

(oTma are les« 

ire than opb 

1011. a. Oldei imperatlTe (oTma are les« rare than opttUve. The 
DtbC peiGonB hare been given above (JaAghiui&Di, the only aecented ex- 
ample, daei not conMpond with the model, but la Id eonroimlty with the 
BnbjFincdie of the reduplicating p[e«ent)i the propel Imperativea are: 2d 
■Ing., dldrbi, dardi-lii, oarkrdlii, J&sr'i*' nenigdhi, rAranddtal ; the 
ending Ut li fonnd in oarkTtftt and jKRrtflt; and the Utt«r [sb waa 
pointed oat abore, 671b) ia nied In AV. u Ant person aiag.; barbrlli 
ahowa tn ebewhere nnpuallded lou of h before the ending M; 3d alng., 
d&dliartu, TBTe^tn, dordttrtn, murmaxttn; 2d do., JfigftSiin; 3d do., 
Jitgrtftm; 2d pi., JigflA; oafiknunata (RT., once) hia an anomaloua 
QDion'VOwel. In the middle Toice la found only nenlk^va ((B.). 

b. Of imparitlve tormi with aaxiliary i, KV. hu none; AV. baa 
vftvaditu and Jofaavitn, and aoch are eometimea foond In theBrihmanaa; 
AV. haa alio, againat rule, taAatanltai and Jafigbamht; TS. haa oftka^Ibi. 

5. Present Fartloiple. 

1012. The intensive partioiples, 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. 

uiaii.zecy Google 

lOlS— -] Xiy> Secondary Cohjooatiom. 368 

EiuDplcBin: icUve, oakftqftt, ninadat, oikitst, m6in;»t, fd^n- 
cftt, rAmvat. dardrat, in&rmfjKt, J&aghanat, n&iuiamat. p4nl- 
phBnat, Utolkradat, d&vtdrutat; — middle, bj^adta&na, mteorijia. 
odkltftna, 76yav(Uia, tdmoSuA, Jirbhnrt^m, s&rsrfiijA, Ja&JabUiu. 
n&nnam&na, djindaflliia. No middle puttdple ihowi tbe dlJtrUsUi 

1018. a. On ■ceonnt of theli accent, rSrahK^A, rMnk^M^k, ud 
jahff Ki^^ (betide JirhrfA^*) *re probtblf to be regttded u poifect puti- 
ciplea, ■Uhcmgb no othei perfect tonni with heaiy ledapllcaUoD from A* 
B«me rooti ocoot. Tbe iQfeienM Is, bo«e*er, rondeied nBOeittln by tbe 
unmistikibljr intensiTe badbadhlnA and menufibik (beilde minnpjina). 
Ai to qdqaoSiia etr,, aee 806 a. 

b. Tbe RT. bu onoe JifiolmataB, gen. sing., with Toot-Towel cut 
oat; k&nlkrat ippeui to be med oare tat ktnikradati l( oSk&t ii to 
be Nterted to f^S (Oiuamann), )t la the only example of an iDtentiw 
(nm a root Id ft, and Ita tcoent it aDomalooi. JCarmrfantas (AB.) Ii 
peibipa a tain reading; but fonna with the nual Inegnlsrly retained are 
found repeatedly In tbe eplca aad later: tbtu, lalUian, dadlpyanOm 
(HBh.). jiJvaUnt (HBh. R,), Barlarp^ntiBu (BhP.), riratanU (R-) 

6. Imperfect. 
1014. The imperfect is legulaily inflected as follows: 

I. d. p. 

Arevidam ivavldTa ivevldma 

ivevet, iravldls &Tevittaiii ivavltta 

3 S^^rJ. 5Rf^^^^ *fif=(HIH^ «l^i"H^ 

itTevet, ^TSTldlt &vevlttfim &Tevldiu 

101B> The Imperfect fonna foond In tbe earlier texti ue aot nnmec 
ons. They are, looludlng thoae from which tbe angment la omitted, a* 
follows: In actWe, lat alng., aoSkaqam, dadlgam; 2d atng., ^Ig^, 
adardar, d&rdar; 3d slog., adardar, adardbar, avarlTar, dardar, 
k^lqkan, diividjrotin&vlaot; "id dn., odardfiam; lat pi., marmpjmt; 
3d pL, anannamne, adardinu, a«arknu>> ^ohavna, anonawaa; 
■nd, wttb •nxlllirr I, In 3d ilag., avftV»Mt, ivftTO^Itt irSTSlit, 
iyoyavlt, iroravlt, i^ohaTlt; and, frreguluty, Id 3d dn., avftwafitftm. 
The middle forma are eitremely few: namely, 3d ling., idedlffa, inaa- 
nata (with loia of the final radlotl in ■ weak form of root)) 3d pL 
marmpjata, and avftTB^aiLta (Khkh, If It belonga bore, ibowi a trantfti 
to ail a-atem). 

ioy Google 

369 IMTENSIVE. [—1017 

loie. Derivative Middle Inflection. Ftom every 
intensive stem, as above described, may be formed in the 
piesent-system a further derivative conjugation ivfaich is 
formally identical with a passive, being made by the accented 
sign U y&, along with middle endings only. It has not, 
however, a passive value, but is in meaning and use in- 
distinguishable £iom the simpler conjugation. 

a. A final vowel before this ya is treated u before the paasive- 
sign ya (770). 

b. The inflectlOD is precisely like that of any other stem ending 
in a in the middle voice: thus, from yxaji, intenefve stem marmfj. 
Is m»de tbe present iDdicative mannrJy^, iiiarm|'jy4Be, marmrJy&te, 
etc-; optative marmriy^ya, mann^jyittaia, marmijy^ta, etc.; im- 
perative marmfjy&Bva, marmfjy&tam, etc.; participle maTm{'Jy&- 
m&nai imperfect &marmfjye> imarmTJyathOB, Amarmfjyata, etc. 
BobjuDctive formB do not occur. 

o. Id a very raw aportdlc cues, thete y&-foTniB are glien a pusWe 
v*lne: thua, jafighanyamfea In MdD.; bambhramyate, dadtunfi- 
yam&na, peplyamftna in tie later linguBi;e. And active participles 
(CSS a) ats not anknown : thae, detUpyanUm (UBh.), dodhOyant 
(MBh. BhP,}. 

1017. This kind of intensive inflection is more common 
than the other in the later language; in the earlier, it is 
comparatively rare. 

a. In RT., y^foims are made from right roota, Bve of which have 
alao fatms of the glmplei conjugation; the AT. addi oae mote; the other 
earlier texts (ao fai aa obaerved) aboat twenty more, and halt of them have 
likewise form* of the simpler conjugaiion. Thaa: from ym^U marmij- 
jr&te etc, and marlm^Jyeta; from ytj[, tartnryante; from yvtx, 
oaroury&mil^a; from yu, neniy^ran, etc.; from yvi, veviyate; ftom 
yrih, rerihy^te etc.; ^m vii, Tevijy4te; fiom v'aku, co^kOyise eic.j 
from ydiif, dedigyate; from ykit}, cSka^y^ts etc.; from yvad, 
. vfivady&mSaa ; from ^nom, nannamyadhvam; from f'vah, vaniv&h' 
y^ta etc. (wlih lengthened loot-Towel, elsewhere anknown); frum flrand. 
kanlkradyimSna; from yvjt, varlvarty&maaa (9^'' should he 
▼arivrty-); fiom yiaxq, amarlmr9yanta (9B. ? the text reads amarimrt- 
■yanta); from i^ynp, yoyupy&nte etc.) from y^nd, anonadyauta; 
ftom yvli, avevllyaata; from y'Jabh, JaiUabliy&ta etc.; ftom yjap, 
Jaiyapy4mllna ; and so on. 

WhitBST, Onnmar. 3. >d. 34 

ioy Google 

1018 — ] XIV. Secondakt Cokjuqation. 370 


1018. The giannDariana aie at vanance as to whether 
a peifect may be fonned directly &0111 the intensive stem, 
01 whethei only a peiiphtasUo perfect (belov, 1070 ff.) ii 
to be admitted. 

K. Ho eximple of ui lotensive periphruUc peif«et hu larwhcni com* 
to light (except h'DDi Jigf : 1020 a)- A few anmiitaksbla perfect tonnt us 
mide from the inlensiTely ledupliMtAd root in RV. : namely, daTldbin 
and ndnKTa, 3d »iag,, and nonaToa, Sd pi.; and thve oecni futiicr 
dodrftva (IS.), yoyKva and leUya (HS.), and lel&ya (? ^.), ta wed 
In the aenie of pieaents. To them ma; be added JKgara l«t ling, and 
Jfigira 3d sing.: bat as to tbeie, see below, lOSOa. 

Aoriat, Future, eto. 

1019. Ab to the remaining parts of a full verbal con- 
jugation, also, the grammarians are not agreed [oocuirencee 
of such forma, 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. 

a. ThuB, from v'vid, iDtensive stem wvid, would be made the 
aoriat avevidiqam with precative v»vidyieam, the fntnres vavld- 
Ifyiml and vevldltiami. the participles vevidita, vevidiUvya, etc, 
the ioGnitive vevldltuin, and the gerunds vevlditvA and •vevldya. 
And, where the intensive conjugation is the derivative middle one, 
the aorist and futares would take the corresponding middle form. 

b. Of all thU, in the iDcient Uagnaga, theie Is baldly a trace, Ibe 
RV. has o^k^e, 3d sing, mid., of a fotmatlon like hl^ and StUfi 
(8Q4 d), and Ibe geiundlTes vltantasayjra, and marmrJeDya and vKv^- 
dhdnya; and (B. has th* participle vanlvahit^ and the Inflnitlie d^dlyl- 
tav&{, As to j5gari;y&nt and J&garitA, see the next paiagispb. 

1020. There are systems of inflection of certain roots, the in- 
tenaive character of which is questioned or questionable- Thtis: 

a. The root gf (or gar) woAe has from the fltit no present-system 
gsTo one with tntenslve rednpUoatiou ; and its intensive stem, jigf, begiss 
esrty to aisnme the Talue of a toot, and totm a completer conjugation; 
while by the grammadans this stem is reckoned as If simple and belong- 
ing to the loot-olasB, and ie inflected throughont aocordlngly. Those of 
Its forms which occur In the older language have bees given along with 


371 Intensive. [—1024 

tbe othec tntendTW *bo*e. Thej m, for the preisnt-tTStAin, the suae 
with those uknoifledged a> lefuIaT Utei. The older perfect is like the 
othei intenelTe perfects found in RT.: nemely, Jigara eto., vltb the 
partldple jBg^vita ; and a ftatuie Jfigari97&-, » piulre putlclple JI[gaTlt4, 
•nd > gemndiTo Jig&ritavyk, ne met with is the Biibmiisai. The old aorlst 
(RT.) iB the luakl tedaplkBted OTBO-called ceiuetlTe uiist: thnn, ^igar. The 
pammulMi* gfTs it la ihe UteT Ungnege a peifect viih additional redapU- 
calioD, JEOftSifft etc., in i§-aorlst, ^j&garlfam, with pieeiUveJfigaryftsani, 
and everything etge that ie Deeded to make Dp a complete conjugation. 
The perf. JaJBgSra ia quotable from the eploe and litet, ag aUo the peti- 
phraatlc Jagar&m iaa. And MBh. hai the mntilated Jfigpnl, and alio 
a-torma, as JSgaratl and JakgramSj^. 

lOSl. a. The stem lr^)ya (active only) rtgulaft, from which a 
nnmher of forma are made in RV., has been Tisvred aa an Intenaive from 
yr^ 0^ TJ' It lacks, however, any analogy with the intenaive tormatloa. 
Tbe same is true of iradli propitiate (only iradhanta and ir&(Ui;&i, 
■ppuently for iradhadhy&l). 

b. Tbe middle item iya, not Intreqaent In tbe oldest language, 1< 
often called an Intenaive of yi go, but wlthoat any propriety, as it has no 
analogy of form whatever with an intensive. Tbe isolated lat pi. Imahe, 
common in RY., Is of queatlonable character, 

1029. The root U totter, ifitb constant intensive reduplication, Uli, 
Is qulto irregnlai in inlleotion and accent: thua, piea., lelayati and lelll- 
yate, pples lelfty&ntl and lel&yatas (gen. aiag.) and lelSyamSiia, Impf. 
alelftyat and alelet and aleUyata, perf. Ul&ya and lel&ya (?). 

1029. Tbe RV. anomalous form dart (or dard), 2d and 3d aing. 
from |/df or dar, Is doobtfally referred to the intenaive, aa if ahbrevlated 
from dardar. KV. baa once avariVUB (or -Tur) where the aenae reqaire* 
a fonn from V'v|t, aa aTariTftaa. The form rarft^&tft (RV., once] seema 

10B4. A marked intensive or frequentative meaning is not always 
easily to be traoed in the forms classed as intensive ; and in some 
of them it is qnite effaced- Thns, the roots oit, nij, vif use their 
intensive preBent-aystem as If it were an ordinary conjugation-class; 
nor is it otherwiBe nitli g^ J&gi'). Tbe grammarians reckon the 
inflection of nlj and vlf as belonging to the reduplicating preaent- 
system, with irregularly strengthened reduplication; and they treat in 
the same way vio and tU; JSgr, as we have seen, they accoant a 
simple root. 

a. Also darldrS, intenaive or [/dri run, ia made by the grammarians 
• simple root, and furnished with a complete set of conjngatlonal forma: 
aa dadarldrftu; adaridrfislt, etc. etc. It does not occut in tbe older 
langasge (duIbss diridrat TS., for which VS. HS. read dAridra). The 
so-called tool vevi flutter it a pure tnteniive. 




lOSB. It it tillowed hj the gnmmtriuii to make from tbe futeDdTs 
Item also ». pitalve, deaideraltve, cauMIive, and ao on : tbiu, from Tttrld, 
ft»a. TBTidyd; deild. vdvidlQ&mi; ciui. vevid&yilmi; desid. of caoaa- 
tin, T^vid&yiffimi. Bnt lucb foimatlong ars ezceaaiTelr ran; qnotable 
are rulTarj&yuitl AT., Jfigar&ya&t TB.. •!«.; d&dliftrayftti JB., 
dandafayitrft DKC. 

III. Desiderative. 

1026. By the desiderative ooDJugation is signified a de- 
siie for the action oi condition denoted by the simple toot: 
thus, (MsHfn pibftmi / clrm&, desid. fl|M(HliH pip&a&ml / wish 
to drink; sft^Tftr jiTSml / live, desid. ftsftftinftl jijivi^Bini 
I desire to live. Such a conjugation is allowed to be formed 
fiom any simple root in the language, and also fiom any 
causative stem. 

a. Tbe desiderative conjagation, although Its forms outside the 
pre Bent-system are extremely rare in tbe oldest language, is earlier 
and more fully expanded into a whole verbal ByBt«ni thau the ioteo- 
sive. Its formB are also of increasing frequency: much fewer than 
the inteoBiveB in RV., more numerooB in the Brabmanae and later; 
not one third of the whole number of roots (about a hundred) noted 
as having a deaidetaclve conjugation in Veda and Brahmapa have 
such in RV. 

1027. The desiderative stem is formed from the simple 
loot by the addition of two characteristics: 1. a reduplication, 
which always has the accent ; 2. an appended Tl b& — which, 
however (like the tense-signs of aorist and future], sometimes 
takes before it the auxiliary vowel S i, becoming ^ i^s. 

a. A few inaUnces In tbe concluding part of f B. in which the acc«n( 
i( othetHise laid — thns, tifttiaset, yiyfis^tam, vividi^&atl, ips4at«a 
— must p 10 b ably be regarded as eiToii. 

102S. The root in general remains unchanged ; but with 
the following exceptions: 

a. A final 1 or a is lengthened before aa: thus, oik^fa, ciklfa, 
Jigifa; fuqra^a, JnliDfa, cuk^ufa. 

b, A final X becomes ir or ur before sa: thus, olklr^a, titSrqa 
(also irregularly tQtOr^a RV.), didliirfa, sisirfa, tiBtirfa (also tn- 
BtOrqaj, JUiIr^a; bnbhfir^a, mumQr^ (the only examples qaotable}. 

itizecy Google 

373 Dbbidbrative. [—1020 

0. Before Ifft, ft final i- or u- or r-vowel necessarily, and a 
penultimate i or u or f optionally, have the gnna-strengthenlng:; no 
«xampleB are qnotable from the older texts; later occur gl9ayifa, 
flfarifia; elkarti^a, nliiartlfa, mimardifa, vivar^i^R, fuqobhl^; 
but Turudif a. 

More special exceptions are: 

d> A few roots In & weaken tbig vowel to i or eien 1 : thus, Jigifa 
trom yei go; pipifa (beside plpOaa) from ^pS drink, jUiifa (AY.) 
from yhft remove QU^ite: 604); dldbi^a (bet[de dhitsa) from yAhR. 

«. A few tooM In an or am lengthen the vowel : thus, jig&Asa (beside 
Jigamlfa) from j/gam; JlgbUsa from yhaxii mun&bBa from f/man; 
and tltabaa trom ytva. 

r. Reversion to gnttuial foin of in Initial aftec the reduplication is 
Men in oikl^a from yd, oikitaa fiom ^cit, jlgi^a from yji, Jlgh&flsa 
item ylXKO; and J^lli is said to make Jigblfa (no Dccnnence). 

g. The lOots van and aan make tIvSbb and ai^&aa, from the root- 
fotms Y& and Bi. 

h. The root jiT forms Jajyu^a (9B.: JtJiTi^a, VS.); »nd Hie other 
roots in Iv (760) ire leqalred to make the same change before ga, and to 
■lave gu^a before l|a: thus, sosyafa oi BiBevi^ from >^bIt. Bvap 
forma Bu;aps&. DhUrr forms dudhOr^a. 

1, Initial B is nsnallj left unchanged to if after the redapliution 
when the desiderative sign has ^ (184e): thus, aiBflLiikqa (gB.: ^'BaSJ), 
and BUBTuqa and alBOniQa, acoording to the grammarians; but tU^tOqa 
1* met with. 

J. Further may be meDtloned as prescribed by the grammarlanB: 
nina&kfa (oi nlnafifa) from ynBiq be loit; mlma&k^a from V'maJJ 
(occurs in mimaikkfn) ; mim&rjtqa (or mimj-kf a) from v'mf], 

1028. Tbe consonant of the reduplication follows the 
general lules (880] ; the vowel is ^ i if the root has an a- 
vowel, or H X, or an i-vowel; it ia 3 u if the loot ha^ an 
u-vowel. But: 

a. A few roots have a long TOwel in the redapllcating syllable: thna, 
blbhataafrom ytiadhorbftdb; mim&fisa from )^man; andtatur;a(RV.) 
from |/taT; dadblfu (AV.) and dadaflkf u (C.) are probably false forme. 

b. From yaq is made (^B.) aglqtfa, and from )/edb (J3.) 
edldhi^a (with a mode of reduplication like that followed aometimei in 
the reduplicating aoiitt: 862). In the older language, these aro the only 
roots with Initial vowet which form a desideratl>e stem, eieept ftp and 
fdb, which have abbreviated stems : see the next paragraph. In the later 
language occur foitbor efiqlqa (|^lf ««tA) and loikglqa (^Ik;); aud the 
giammarians add others, as arjlhl^a (V'arh), undidi^a (Viind), ardi- 
dhlfa (Vrdb). 

ioy Google 

1039—] XIV. Secondaky Cohjugation. 374 

0. BV. has the items fnak^a and fyak^a, legtided as dMicloittiTu 
rrom v'v'aa; attain snd y^, with matll&ted ledupUuUon. 

1030. A numbet of TOOU, including some of very com- 
mon use, form an abbreviated stem apparently by a con- 
traction of leduplication and root together into one syllable: 
thus, ^ Ipsa from yWi 5p ; ^rH ditsa from y^ dB. 

a. Snch abbieTlated stems are rooad in the oldet Unguige ss fnlloTi : 
dbitaa (besido dldtiifa) trom ydbS,; ditaa (beside did&sa) from f'di; 
dipsa (dldpBB JB.) from i/dabh; gib^a fiom y<;ah; uk^a from yaaii: 
these lie foand la RV.j in AY. ire added ipsa from y&p (RV. baa apaa 
oDce), and irtsa from yjdh; the othti texts rnmlsh lipaa (^B.) or 
llpsa (TB.) from yi&bh, rlpsa (QB.) from yrabb, pitaa ((^B.) from 
ypad, and dhlfe^a (^B.] fcam f'dah (not ydlh, sin«a no toots with 1 as 
medlU Tovel show the contracted form). In the later laDguage are farther 
fonnd pltsa from f/pat also, Jilipsa from the cansatlTe qnasl-root jfiap 
(below, 1042 J), aod the anomalom mltsa from ymK mtaiure (allowed 
also from toot* ml and ml) | and the grammBrians give ritsa from yvBdb. 
Also mokfa la (lety qneationably) sieved as a desideratiTe stem from 

108X. 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 grammajians in many stems 
which have not been found in actual use. 

a. It is declared to follow in general, thoogh not without ex- 
ceptions, neoeBsary or optional, the analogy of tbe fntnres (934, 
943 a|. 

b. No eiample of tbe a»e of i Is found In RV., and onlj one each in 
AV, (pipatli^a), TS. (jljivi^), and TS. (Jlgamlfa). The other ezamplea 
noted in the earl; texts are a^ifli^. oltramlyt, Jlgratalqa (with i fot i, 
ai elsewhere in this root), oloarl^, edidhi^, jUanl|a, didlkqi^n, 
bibBdbl^ia, rumoifa, vivadlqa, vivldifa, qigOBifa, tl^figlilqa, jllil&- 
alfa: most of them are foaiid only In ^R. Sterna alio withoot the aniil- 
lary vowel are made from roots gam, grab, car, J!v, pat, bSdh, vld. 

1082. Inflection: Present-System. The desider- 
ative stem is conjugated in the present-system with per- 
fect regularity, like other o-stems (733 aj, in both voices, in 
all the modes (including, in the older language, the sub- 
junctive), and with participles and imperfect. It will be 
sufficient to give here the first persons only. We may take 

ioy Google 

375 Debidbratite. [— 10S8 

u active model ^ ipsa seek to obtain, from yWJ Sp ohtain\ 
as middle, frlldH titikfa endure, from yfrTsI tij be sharp (see 
below, 1040). 

1. Preaent Indicative. 

mHts. mlddte. 

i. d. p. a. d. d. 

tpflSmi ipsSTfts ipBlmaa titLk^e tItlk|ftTaIie t{tlk;&mahe 
etc etc. etc. etc. etc. etc. 

3. Present Subjunotlve. 

1 ^:f\t^ \mi4 ^t-HIH fHid^l iniHidlsi^ fniddlM^ 
fpsBni IpB&va Jpaftma tftlkq&l tltlk^STahai tftik^amaliai 
etc. etc, etc. etc- ete. etc. 

3. Present Optative. 

ipseyun ipaeva ipsema titibfeya tftLk^etrahi titikijemahi 
•te. etc. etc. etc. etc. etc. 

4. Present Imperative. 

a |cH ^CflfTT ^cnfT frjlaiHW iflfrl^lH^ frT^<^a^ 
ipsa ipeatam ipsata tftlk^aava tltikqethim t(tlk;adhTam 

ete. etc. etc. etc. etc. etc. 

5. Present Participle. 

^^Hrl ipsant (f. ^(44Tn JpsanH) Idfrli^HIUl titlkfsma^a 

6. Imperfect. 

'"^"^^ "^^R \'MW 5%f?l% MfniVldHi^ qfrrfHcinrf^ 
&[psun ilpsftva K[pBKma atlUk^e Atltik^ftTahl Atttik^ftmaht 

a. There ue almott no iiregnkiltlea of inflection to be lepoited tiom 
the older langiuge. No lit pi. In masi, or 2d pi. In thana or tana, i* 
met with; ot the ImpT. in tftt, only Ipaatftt. The quotable rabjuaetlio 
feimi ue Ihoie in sjtnl, sSt and sat, a&n, and santa. KBU. haa JtJtLfialta 
(ef. 7S8b). Bat the fem. pple sifBaatf (Initead of SlifBoantT] eccDn 
once or twice in the elder tetta; and RT. hia didhlfS^a. 

b. In the eplcj and later are found ipotadlo formi of the non-a- 

ioy Google 

loss— J xrv. Sboondaby Conjcoation. 376 

coDjagttlira; thus, gisyk^mas (BhP,), tltikfinalis ind bnbhfifate 3d 
pi. (HBh.); *ad tbe fern, putlnlplea lipaatj uid oil^ni^tj [MBh. : tgUnat 
440 b). The anomklans jlghftAaifit ocean kIbo in MBh. tnd Vm. 

1038. a. Desideratire fomiB ontaide the preaent-ByBtem &ie 
extremely rare in the oldest language. The BY. has only perfect 
forme from a stem mioilkf — thui, mlmik^tbos, mlmikf&tas, 
mlTnllcjfia; mimlkfe, mimlkfire — along vich the preBent fomu 
mlmlk'^fttl, mlmlkf a etc., mlmihyint (ppte] : they show tbat wiirw<fc^ 
or mik; has taken on the character of an independent root la AV. 
are foond two aorist forms, irtaia and oolbltalB, and a participle or two 
^m mlmfiAaa (see below, 1037a, lOSBa} — all of them from atena 
which have lost their distinct desiderative meaning, and come to bear 
an independent valae. The forms noted from the other earlier teiU 
will be given in full below. 

b. Id the later language, a complete Bystem of verbal 
foims is allowed to be made Id the deeideiative conjugation, 
the desiderative stem, less its final vowel, being treated aa 
a root Thus: 

1084. Perfect. The desiderative perfect is the peri- 
phiastic (1070 ff.}. 

a. Thus, ipalib oakSra etc.; titik^ftm oakre etc. Sach foraiB 
are made in ^B. from v'V'kram, dhurv, b&db, mb; and in ChU. 
firom maa. 

b. Apparant perfect formg of tbe oidlnuy kind made from mtmiln 
in BV. haie been noticed In the preceding paragraph. And A6. (vtU. 21. 10) 
haa once didiaitba thou haii detired to gtee. 

10S5. Aorist. Tbe aoiist is of the i^-form : thiu 
^iuiqq^ftipBi^Bm, l^iratdfsr Atitikfipi. 

a. The AV. has aoikltsis, and IrtalB (logmeDtlesa, with in& pro- 
hibitive: 678]. TB. hu Upsltj ^B. airtait, aolkiTfis and aJIghSAaie, 
and amIm&fisiqitbiB; KB. Jijfi&slfi; JUB. Upalfma; and AA. adblt- 
■Ifam. No eiamplaa have been found in the later langcage. 

b. A preoatlTe is alio aliased — thot, IpeyasaiD, titik^lftya; bot it 

1036. Futures. The futures ate made with the auxil- 
iary vowel 5 I: thus, ^iUiVllfH IpsifySmi and ^fUlHltVfl 
IpBltiami; (dfrfldsS tMl^i^ji> and raidRlHI^ titik^it&he. 

a. The QB. hae titikqi^yate and dldrkfltiraa. Sach foims as 
JiJfiioy&inaB (MBh.), didbak^yKmi (B.), and tnlmlAByant (OOS.) are 
donbtleii preaentB, with -ays- blanderingly for -an-. 

ioy Google 

10S7. Veiba.1 Nouna and Adjectives. These too 
ate made with the auxiliary vowel ^ i, ia all cases wheie 
that rowel is evei takeo. 

a. Id the older Ungnage b*ie be>D noted: p&rtinlple In ta, mltnSfl- 
Bit& (AT., OB.), Jljyaflta (AB.), qnijriiflU uid cUukfit4 (9B.)i — 
gernndiTe in tavya, UpBitavya (AB.), didhyfiBitavy^ (fB,); in ya, 
Jtjfilayk (9B.)i — geiand in tra, mimAAaitrB (K.). 

1088. Of other decli'itble atema dariied from the desideTitive stem, 
by tai the most comiiion are the adjeottve In u — e. g. titUc^U, dipslk, 
bibttatsA, alffian (RV. once did^kfu) — and the abatiact nonn In ft — 
e. g. ipsft, bibhatai, mlmSAs^, ^ufrfi^S — both of which are made 
«ltb incieaslng tieedom ^om an eaily epoch of the langnige: especially the 
former, which hat the value and coiiBtrnetlon (S71 a) of a ptesant pSrtl- 
elple. A few adjeetlvea la anya (having a gaiaodlve character: 96Sb) 
occur in the eailtet language: thna, dldrkfj^^a (RT.), ^uqrfiqd^ya (T9.), 
nlDlfia^ya (P6,), jijilaaeiiya (AB.), and, wlib Iriegutai ledapUcatlon 
(apparently) papr^^^nya (BV.), dadlil^e^ya (JB.); and dld^k^a (RV.) 
fs ■ almllai formation. RV. has also Bifftfl&nl and mrak|&^, and sifB- 
a&ttl(P). In the later language, besldea aome of the formatloni already 
iastanred (ihDae in u and ft. and In sya and sltavya), are found a tew 
derivatives in aka, as oikltsaka, bubbufaka; in uka, aa Jljfiiaana, 
dldbyftaana; and, very rarely. In anIya(oikitBanIya)and tr(i;agrufltT); 
further, secondary derlvailvea (doubtleta) in in from the noun In S, as 
ipsln, Jislfbl (one or two of these occur In the older language). And of 
an adjective tn a we have an example In bibbatii4 (B.S., and later), and 
perhaps in avallpaa (AVP.) ; auch words aa aju^psa, da9oikitea, are 
rather to be undentood as poiseulve eompoands with the noun in S. A« 
to nonn-atems in is, see 392 d. 

10S9. Oeiivative or Tertiary Conjugations. A 
passive is allowed to be made, by adding the passive-sign 
IT yi to the desiderative loot (oi stem without final aj: thus, 
i^UrtSH Ipsyite it is desired to be obtained; — and a causautive, 
by adding in like manner the causative-sign WJ &ya (1041): 
thus, ImUliM IpaiySmi / cattse to desire obtainment. 

a. Of these formstione iu the older langnage are found mimUayi- 
nUtna (donbtleaB to be read for -a&mftna, AV.), llpByimfina (^B.), and 
raratayainSna (R,). Half-a-doien suoh passives are qootable later, and 
one or two causatlies: e. g. oikltsyate, vlvak^yate, Jijfiaayata; clklr- 
yayant, olkitaayi^yati. 

b. Pot the desiderative conjagation formed on causative steins, 
which is found ai early as the Brsbinanafl, see belov, 106Sb. 

Digitizecy Google 

1040—] XIV. Secomdabt Conjcgation. 378 

1040. Some stenu which kre desiderative in form have lost the 
peonliaritj of dOBiderative meaniog, and aBBnmed the valne of inde- 
peDdeot roots: eiamplea are dklts eurt, jngope etupue, tltikq mdwrr, 
blbhata abhor, mlmiAB pondtr, ijxitpf^ obt^. Donbtlete Bome of tie 
apparent roots in the language with sibilant final are akin with the 
deBideratives in origin: e. g. glkf, desiderative of gak. 

a. On (ccoanl of the nan ralation of dmldantlfe >nd fatnn (cf. 
848 b), the fomeT li occulonalty fonnd wfaece the litier wu uthei (« W 
•xpected: thus, r^Snaih prayiybuuitaiii (^B.) a king about to i^mrt 
prfii^a nooikramlfan (ChU.) the breath on the point of expiriitg; imn> 
mtlrqur iva 'bhavat (H.) he urm fain to dif. 

IV. Causative. 

1041. a. In the latei language is allowed to be made 
from most roots a complete causative conjugation. The 
basis of this is a causative stem, formed by appending the 
oausative-sign SRT Aya to the, usually strengthened, root. 

b. But by no means all conjugation-stems formed by 
the sign ^(U &ya are of causative value; and the grammarians 
regard a part of them as constituting a conjugation-class, 
the tenth or our-class, according to which roots may be 
inflected as according to the other classes, and either alone 
or along with others (776j. 

o. In RV., the ptoponioo -witboot oaasatlve Talne i« folly one tbitd. 
The formUiOD U » more obilonsly denomioitlTe one thui tiiy of the other 
conJngttioD-cluaaB, ui iDteimediate between then) ind the proper deaom- 
Inatliei. h. cauiatlve meaning hia eitibUshtd itaelf In eonnedioD witk 
the fonnatioD, and become p red oml riant, though not eielnilve. A number 
of TDots of late appearance and probably dirlTatiTe cbaiacter are Incloded 
tn the ctati, and tome pUpable danomlDatives, which lack only the anal 
donomtDiliTo accent (belon', 10B6). 

d. The cauMtive fomiKlOD ia of much mora frequent nae, and nore 
decidedly etpanded Into a fall conjogation, than either the iatentiTe oi the 
dedderatlTe. It la made from more than three hnndred looti In the early lan- 
guage (in BV., from abont one hnndred and Oftj) ; but in the oldeat. Ha 
formi onteide the preaent-iyatem are [apart from the attached rednpllnted 
aortet: 1046) exceedingly few. 

1042. The treatment of the root before the catisative- 
sign WX aya is as follows: 

ioy Google 

379 Causative. [— 104S 

ft. Hedial or ioitial i, u, r, 1 have the Ku^a-streagtheiiiiig (if 
capable of it: 240); thus, vedaya from v'vid, oodaya from ymxd, 
tarpaya from ybfp; and balpa^a from ykfp (only example): but 
olntara, gulpbaya, dfAbaya. 

b. Bat X few loota lick the etrenglhening: Ihete uc. In ths older 
Ungusge, oit (oitaya 'od cetaja), U), 11, rlf (rifaja and rsQaya), 
vlp (vipaya uid vepaya), toj, tur, tuf (tufaya >nd to^aya), dyut 
(dyntaya ind dyotaya), rue (ruoaya and rocaya), fuo (^uoaya uid 
Qooaya). gubb (Qubbaya and qobhaya), krP< "'V^' 'Prb i ">■' S^abh 
makes in RT. Kfbhaya. Zhi^ and glib leugthea the lowel initead. UrJ 
aometiin«i hu vrddhl, as in alher faima: thns, mftrjaya (beilde mar- 
Jaya). On the other hand, gwfa ippaare Irregularly (240 b) In aravaja 
(beside frivaya), be^aya, mek^aya. Similar irregntarltles in the later 
Isnguage are giraya. tulaya (ako tolara], ohnraya (also oboraya), 
mufaya, apburaya. No Toims made without atrengthening bave a cansstive 
Tslee In the older Isngoige. 

0. A final vowel has the vrddbi-Btrenglheiiing; thus, oByaya, 
fSyaya, oySvaya, bbKvaya, dbftraya, aSraya. 

d. But no root in i or i baa vrddbi in the Veda (aulesB pOyaya 
[k, below] cornea from pi rather than p&] — as, Indeed, tegnlu caoaa- 
tifti from Bseh roota are hardly qaotable: only RV. hu k^ayaya (besldo 
kfopaya) from ]/kfl pogtem; for i few alternatlTely permitted forms, aee 
below, 1. In B, and 3., however, occur (ftyaya and aayaya {yi or aS); 
and later -Oyaya, aayaya, amftyaya, AAy^et, nByaya. 

e. A few loota have a form lUo with gu^a-atrengtbening : thna, oyu. 
dru, plu, yu separaU, ^ru, pil, sta, aru; Jr watte aieay, df pierct, ay, 
amr, bf; v^ ehoote makes varaya later (it U not found in V.; epic alto 

f. A medial or initial a iu a light syllable iB Bometimee teogth- 
ened, and sometimea remaiDs anchanged; tbos, bbfijaya, sripaya, 
ftdaya; Janaya, qratbaya, auaya (but mandaya, valgaya, bbak^ayaV 

g. The roots in the older language which keep their abort a are Jan, 
pan, avan, dban, ran, atan, gam (gftmaya once in RT.), tam, dam, 
raj (usnally rafijaya), pratb, fratb, gnatb, vyatb, avad, obad pletue 
(alto obandaya}, nad, dbvaa (also dbvaAsaya), rab, mab (alio 
maAbaya), nabb (also uambbaya), trar, avar, bvaL In the later 
langDtge, further, kva^, Jvar, trap, day, pKf, ra<r, ran ring, vadb, 
val, vaf, flatb, akbal^ ntbag. Both forms tre made (either in the 
eyllat or In the later language, or in both taken together) by ad, kal, 
kram, kfam, khan, gha(, oam, oal, Jval, tvar, dal, dhvan, nad, 
nam, pat, bbram, matb, mad, yaxo, ram, lag, lal, vam, vyadb, 
qam be qaitl, ^ram, fvea, avap. The roots which lengthen tba vowel 
are decidedly the more nnmeroui. 

h. If a uaaal Is taken in any of the strong forms of a root, it nanally 
■ppaara in the caiuMiTe item: e. g. dambbaya, daAqaya, Imdlwya, 

ioy Google 

104S— ] XIV. Secondary Conjuqation. 38o 

llmpayo, rondbarOt ^undhaya, kpntaya, df&baya. From m Dombet 
af Toota, tUiDM batb with lod vitbont the nsBal ara madai thne (besidet 
tho«e mentioned ■boie, g), kufioaya and kocaya, sranUtaya and gratb- 
aja, bp&haya and barhaya, bhraAfaya and btirSfaya, Qundbara 
lud 9odbaya, safijaya and sajjaya, sificaya and aeoaya. In a few of 
thege 1« aeen the infloence of present- it? ma. 

1. Host roots in fioul &, and the root j, add p before the con- 
jugation-sign: thoe, dBpaya, dhSpaya, ethSpaya; arpaya. 

J. Such stem« are made in the older langoage from the looti kfft, 
hhyfi, ES (in^ (alao g&yaya), glfi, gbrS, Jfi&, dS giee, dft divide, dr& 
run, dhA j)uf and dbft sucA, mS mtature, mlft, yft, tS 6bui, athfi, anS, 
bft remove; the later language adds kqmK, dbmS, and b& leave. From 
JB& and enft are found in AV, and laler the ghoitened forma Jfiapaya 
and anapaya, and from qrS only Qrapaya (not in RT.). Alio, la the 
later language, glS forms glapaya, and ml& forms mlapaya. 

k> Stems from fi-roote ahowlng no p are, earlier, gSy&ya (also gKpa- 
ya) from f e& ting, obSyt^a, p&yaya from y^pfi drink (or pX), pySy- 
aya from ypyB. or py&y; efiyaya from yall (or al); also, later, bvRy- 
aya from yhv& (or bu); — and further, from roots vS weave, vyK, and 
93 (or qi), according to the grammarians. 

1. The same p is taken also by a few i- and i-toot», with other 
accompanying irregalarities : thai, in the older language, k^paya (R'V-, 
beside ksayaya) from yk^ posaeei; JSpaya (VS. and later) from yH; 
lApaya (TB. and later; later also l&yaya) from yS cling; frftpaya (V3., 
once] from f'frl; adbyfipaya (S. and later] from adbl + yli — in the 
later, kfapas^ (beside ki^ayaya) from yk^i deitroy; m&paya from 
yml; amKpaya (beside smSyaya] from ynni; brapaya from ybxi; 
— and the grammarians make further krBpaya from y'krl ; o&paya (beside 
oftyaya) from yoL gather; bbSpaya (beside bbByaya and bbiqaya) 
from ybbi; repaya from ]/rI, and vlepaya from yvli. Horeofer, f'rab 
makes ropaya (B. and later) beside robaya (V. and later), and yiam 
makes knopaya (late). 

m. More aoomalODs easee in which the eo-ealled caDeative li palpably 
the denominatiTe of a derived noun, are: p&laya from yp& protect ; prifaya 
from ypxi; llnaya (aecording to grammarians] from yii; dhOnaya (not 
canaatiTe in sense] bom ydbu; bhif aya from ytttu ; gb&taya from )^an ; 
spbftvaya from yspba or spbfty. 

n. In the Prakrit, the causative stem is made from all roots by the 
addition of (the agnWalent of) apaya; and a nnmbei (aboot a doren) of 
like foimatians are quotable from Sanskrit texts, mostly of the latest period : 
bat three, kri^Spaya, Jiv&paya, and dik[|&paya, occnr to the epica; 
and two, a^Spaya and k;&Upaya, even in the Suuas. 

1018. Inflection: PreseDt-System. The causative 
stem is inflected in the pieseot-system precisely like other 

ioy Google 

381 Caosativb. [—1043 

steins in 9 a (733 a,) : it will be sufficient to give beie in 
geneia.1 the fiist persons of the different foimationg, taking 
as model the stem m^ dhSriya, fiom y^ dbr- Thus: 
1. Present IndioatlTe. 

'• ^ J- p. 

1 m^uiSr yi(uii^H^ ui^RH^ 

dh&r&yfiml db&r&y&vas dh&r&yimMi 

1 ^JfJTI m(lIN<f UT^HlH-t 

dliAr&ye dh&r&y&TObe dbiriyftnuhe 

etc. etc. ttc. 

a. Tha 1st pL act. in mosl gieatly oatnumben (ii tm to ons) that 
in mu in both RV. and AT. No eiampJo ocean ol 2d pi. act. In tbuia, 
DOT at 3d sing. mid. In e foi at*. 

2. Present SubjunotlTe. . 
For the subjunctive maj be iDstaoced &1I the forms noted as 
ocourriof; in the older language: 


1 dtiSr&yS^l dh&r^y&va dh&r&ySma 

) I *y ')h=>^''&thgts dliftriyatlia 

dhftr&y&taa dta&r&yftn 


i dh&T&yU dhar&yftTab&l 

j.^. , , fdh&r&ySdhve 

. dliiriyaee idhariySdhvW 

(dhSriyite ^^ 
IdhSriyfttii *^'*™J'"™ 
b. Only one dual mid. rorm In Kite occqib; mOd&yUte (RV.). The 
onlT RT. mid. foim In U, except In Ist da., U m&dayfidhT&l. The 
primary endings in 2d and 3d elng. act. ice maie common than iba secondaiy. 

S. Present Optatdve. 
dbftriyeyam dbftr&yeTa dh&r&yama 

ioy Google 

1048— J XIV. Secosdary Cokjiqation. 382 


etc. etc etc 

c. Optitiie toimt ue very rare in the oldeit lansnags (fani In RT., 
two Id at.) ; they become more e«mmon In the Btibmuu. A 3d elng. 
mid. in ita tait«wl of eta (ct. 738 b) ocean once in B. (ULnuiTittt AB.), 
it not Tery tub in B. (i (core oi two of eiimplea ue qttetible), uid 
ii alto found in MBh. uti liter. Of ■ corresponJing Sd pL In Inn only 
one or two instennea on be potnted «nt (kAmsTlran A(S., kAlpai;iTan 

4. Freseat ImperatlTe. 


dbBLTiya dhftri^tun dli&riyata 

etc. etc etc 


I fJHUH mh13e)ih^ eihov^ 

dh&r&y^BVA dta&riyethSni dh&r&yadhvttm 

d. Imperitiie persons with (he endicg t&t occiu: dbftToyKtat (AT.) 
and oySTByatit (^B.) ire 2d sing.; pUayatat (1,'B.] la 3dBiag.; gam»> 
yatAt ind cySrayattt (K. etc.), ind vftrayatat (TB.) ue used u 2d pL 
VSrayadhTKt (K. etc.) It 2d pi., *nd the only knowii eiimple of Bnch 
>n ending (sei above, 648 b). 

5. Freaeat Participle. 
Mlfyri^dhariyant UI^UMIUI dbfaiyam&^a. 

e. The feminine of the active pirttclpls is regolirly ind usnilly inkde 
in antt (440 0). Bat i Tery few eiimplea in ati are met with (one In 
the oldsr langaige: namayati Apiat,). 

t. The middle participle iu mana Is mide through the whole hlitery 
of the language, from BV. (only y&t&yam&na) down, and is the only 
one met with In the eitlier Unguige (for Iray&nas [aid], MS. ii. 7. 12, 
is eridently a faUe reading, perhaps for Irasi tUtS). Bat decidedly more 
common in the epics and liter i* one farmed with ftna: e. g. kSmay&ns, 
ointayfina, pftlay&na, -vffday&aa. It is quotable from a luger nombei 
of roots than is the more regular participle in mSaa. Ai it occurs In no 
accentuated text, its locent cinnat be given. 

i, Google 

1 ^TttjH^ ^I^UM ^mTJTTW 

Adhftrftyam &6hEraja.vti kihBxayamA 

1 am^ sMuuwf^ ?wr(7iTn% 

&dh&r^a idUray&Tahi AdhSrayftinKlii 

1044. At iru ibDve pointed out, the tormttloDB ttom the c4iiutlTe 
st«m Bya oataidA the praseut-iyitem »re in the oldest luigaige very 
limited. la ET. are fooud two rorms of the fatuie In syfiml, one puiive 
psrdciple (aodlt&), end teo inllDltlTeg In dbylU; >1bo one oi two derlv- 
Mwe noDns tn tp (bodboylt^, oodayltrf), five In i^i^u, ssTen In Itnu, 
and ■ few In a (atipBrayi, nldh&ray&i vftoamlfikhftri, vlfTuneJtiya), 
>nd In a (dlifirayi^ bhfiToy^ ntandayn). In A.T., also two s-fotnia 
fotiai and (oar genmde In tvi; and a few deriiatiTe noan-a(emB, trom 
one of which is mads a petiphtattle perfect (gamaTibh cakSi»). In the 
BrihmainM, verbal derliatlTe forms beoome more numerone and lariou, m 
will be noted In detail below. 

1046, Perfect. The accepted causative perfect is the 
periphiaatic (1071 a) ; a derivative noun in S is made from 
the causative stem, and to il« accuBative, in Sm, is added 
the auxiliary: thus, 

?J|^!l1 ^tRI^ dhftrayaih cakftra (or Baa; 1070b) 
MTJUl ^^ dbftrayidi oakra 

a. Of thti perfect no example occart la RT. or SV, or VS., only one 
— CamayiEih cakftra — in AV., and but half-a-dozen In all the varloni 
testa of tbe Black Y^ur-Veda, and these not In the numtra-parta of the 
text. The; are alao by no meana treqaeat In the Brahmanl*, except in 
QB. (where they abound: chiefly, peihapa, tor the reason that this work 
ntes Id considerable part the perfect inatead of the imperfect aa lie nanatiTe 

104B. Aorist. The aoiist of the causative conjugation 
is the reduplicated, vrhich in general has nothing to do 
with the causative stem, but is made directly from the root 

a. It haa beeo alieady fally deactlhed (above, 858 S.). 

b. Its asBOciatioQ with the oanutive is probably foiuded on an 
original Intensive character belonging to it as a redaplicated form, 
and is a matter of gradaal growth ; in the Veda, it is made from a 

ioy Google 

1(M6— ] XIV- Secomdabt Cohjuqation. 3gj 

considerable number of roots (id KV., more than a third of its in- 

staDces; in AV., about a Srtb) irhich have no cansative stem in aya. 

o. The cauBative aoiist of yrj dlir, then, is as follows: 

1 M^iyfH^ Q^NiTT^ y4lu|m 

&dldliaram ididharfiva Adidharftma 

Adldbare AdidharSTahi adidbar&malil 

Ad example was inflected Id fall at 664. 

1047. In a few cases, where the root haa asenmed a pecaUai 
fonn before the causative sign — as by the additioo of a p or f 
(above, 104Siff.]~the reduplicated aorist is made from thie form 
instead of from the simple root: thus, atiffhlpam from ath&p (stem 
Btb&parai for yotht. Aorist-stems of this character from quasi-roots 
Id 8p are arpipa (Krl. JUapa or jTjlpa, jijflapa or jJjaipa, (Ifrapa. 
ti^tbipa, JDilpa; the only other example from the older language is 
bibhi^ from bhif for ybbl. 

1048. BdI b few iporidio toixot of >n l^torlat from canutiTs con- 
JngaUon-EteiDB »e met with; thai, dhvasayit (ftV.; TS. bM initead the 
wholly enoiOBlons dtavanaTit), vrathayia »nd Ulaylt (AV.), py^Biylf- 
(hfiB ind av&daylqtbSs (KBU.), tn the oia«t Ungnige (RV. hu alio 
Qnayia from s denominitive stem); in the Ute^ ahlftdarlfata (DKC), 
■nd probably aghatayitlifia [MBb.; for •Ifth&a: cf. B04 d). The ftasWa 
3d Blng. aropi, from the oauEstiTe ropaya, his > late oGDnrreDee (9>tr-}. 

1040. A preotlve is of conree allowed by tha gmnmariani to be 
made for the cinaattre conjugation: In the middle, from tha oanaillTe atem 
with the ftoxilisry 1 substituted for its ilQil a; In the sctiTe, from the 
form of the root as BtTengtheiied in the cansitive stem, bat without the 
OUiMiTe sign: thus, 

^lUIMH dhiryasam ate. ^I(uJNIU dhSrayifiya etc 

This formation is to be regarded as pnrely Bcttttoni. 
1060. Futuies. Both futures, with the coaditioaa), 
aie made from the causative stem, with the auxiliary ^ i, 
which takes the place of its final 9 a. Thus: 

yi{fymi(M dhtraylfyimi et«. EJI^fll^ dh&rayi^yfi etc. 
CJTjfijsnTt dhSrayiflyAnt Ml^ni^HIUI dMrayifpyimiifla 

ioy Google 


?iyH (U^U H idhteiqri^yam etc. MMI(R|i^ AdtaKrAjrl^r* etc. 

PoriphraBtio Potore. 
En^lhfnfFT db&raTittnml etc. 
a. It bM been neatianed sbovA tbat BT. 4Dd AT. «0Dlaln only two 
eumpIeB euh of th« ■-fatore, and none df tke peilphiutio. The lOrmer 
begin to appeu in iha Btihtiu^ts more nameionsly, bnt itill tpuingly, 
with paitfciplGB, ind conditional (onlj adhArsyi^yKt (B.; ftlipajii^jra- 
tbSa ChU.); of the Isltec, ()B. affoidB two tn«tiaces (pSrarltiUial and 
janayit^). Eiunplei of both fonm-dODa are qnotable bom the later 
Ungaage (iuclading the middle form darqayltUio: 047 a). 

1061. Veibal Nouns and Adjectives. These are 
made in two diffeient ways: either 1. &om the fuU causa- 
tive stem (in the same maonei as the futures, just des- 
cribed}; or 2. from the causatively stieogthened loot-foim 
(with loss of the oausative-sign) . 

a. To the latter claas belong tbe paaaive participle, u dh&rita; 
tbe gerandive and gerund in ya, hh dhSira. -dh&rra; and the gerund 
in am, as dbSrom; also, tn the older language, the root-infinitive, 
as -dhbam etc, (870 a). To tbe former class belong the infinitive 
and the gerand in tvS, as dh&rarltnm, dtaara7ltT&, and the gerundive 
In tavya, ae dbftrayitavya (also, In the older language, the infioitivea 
in tavU and dliy&l. as J&najrltavif. iray&dby&l, etc.)- Tbe auxiliary 
1 in taken in every formation which ever admits that vowel. 

b. Eiamplea of the pualTe parUdpU ate: teUk, vBalta, frtvltA. 
Bnt tiom the qnMl-root JE^p (104SJ) la m»de Jfiapta, without nniOD- 

o. Eiamples of the inflnltive and geinnd in tvl aie: Ji^ayltiim, 
dharayltain; kalpajitvlt, arpayitvS. Bnt in the eplca, and even later, 
inflnitlvefl are oeraaianally made *llh loae of tbe canaatlTe'Bign : e. g. 
qe^lttun, bbSvitom, dh&rltnm, mooitum. 

d. Examples of the gemnda tn ya and am are: -bb^;a< -Kblvyai 
•pftdyn, -vBaya, nSyya, -athftpya; -bh^am, athftpam. Bnt atems 
ahowlng in the toot-gyllable no diffeienca from the root retain ay of the 
cansadve-Blgn In the gemnd, to diatlngolah It from that beloagiiig to 'be 
primary conjugation: e. g. -kram&yya, -gamiyya. -Jan&yya, -Jval&yrs. 
-balayya, -qamayya, -racayya, -ftpayya. 

e. Eiamplei of the gerandive In tavya are; tarpayitavyli, gaxa.- 
arltavya, hvftyayltavya ; of that In ya, sth^pya, tiirya, yl^ya; of 
that in aniya, sth&paniya, bb&vanlya. 

Wbltaar, Orammar. 1. ed. 35 

Diaii.zecy Google 

1061—] XIV. Secondart Conjuoation. 386 

f, Eximplea ot otli«i foimatiODi DcnrriDg in tha older )ui|aife an 
M follovs: lOOt-in&nitlTe, -BtlUlpain, -v&saa; — inflnttlTe In tu, otbtr 
ctses thiQ accnutlTe, •JamtfitaTS ; j&nAyitaTS£, piyajitKvSi, -foot- 
oyltavAl; f&mayitos; — infinitive Id dbyU, l^ay&dhy&l, ira}-4dhyU. 
tafisay&dhyU, nftqay&dhr&i. manday&dhyU, i&&da;4dhy&l. ri^- 
ay&dhyU, vartayAdhyftl, viinjkdiijtii, Byanday&dbySl (ill RT.); 
— gemndiTs in Syya, pana^yra, Bpfbay^yya, trayayayya (? ytii). 

g. Other Doan-derivatiTat ftom lh« Mnutiva stem in not iDrreqiient, 
being deddedl; moie numerous and tuIoub thui from *ny oAei of tke 
seceudarir eoi^ngatlon-neiiu. Eiunplei [of other klndt thin those InituiMd 
la 1044] ue; fcrpa^s, dSpuiA, prl^AtiM, bldfaqft; JiUlp«ka, ropnlu; 
patRyaiA. BprlurUu; J&DOyfttl, Jfiaptl. 

h. All the oIuMt ot deiivitlTei, It vill be uotieed, follow In tcfud 
to accent the tnalogy of limilsr fornikUoni from the limple not, and >hoir 
no InDnence of the ipeolal aecoDt of the cavsatlvo-stem. 

lOSa. Derivative oi Teitiaiy Co aj ligations. 
From the causative stem are made a passive and a de- 
sideiative conjugation. Thus: 

a. The paasive-stem is foimed by adding the usual 
pasaive-aign 17 y^ to the causatively strengthened toot, the 
causative-sign being diopped : thus, umFT dbftry&te. 

b. Such puclTea ue hardly fonnd in the Tedi [only bbftjyi- AT.), 
but Bome tbiitjr Instanoea are met with in the Brahmanae and Suttu: ei- 
amplee are Jftapyi- (TS.), sBdya- (K.), pUya- (AB.), vKdya- (TB.), 
Btllftpya- (OB.); and they beoome qnlte common Uiai. 

o. The desiderative stem is made by leduplication and 
addition of the sign ^ i^a, of which the initial vowel replaces 
the final of the causative stem : thus, I^UI^nj^fH didb&rayifati. 

d. These, loo, are found here and there Id the Brahmaoaa and later 
[about fort; Btema are quotable): example* are piplyayifS (E.), bibbftv 
aylfB and olkalpayl^s and lulobbayi^B (AB.), didrfipayi^ and rirfidb- 
ayl^ (^B.), and bo on, 

e. A> to DsusatlTes made from the Intenalve and deslderatlTe atema, 
«ee aboTB, 1036, 1038. 

V. Denominative. 

1058. A denominative conji^ation is one that has fot 
its basis a noun-stem. 

a. It ii a view now preiailingly held that moat of the preaent- 
ayslenu of the Satukrit Terb, along with Mhat fonn»tlona «n«logoai with a 

ioy Google 

pteMot-syitem, >k in thelt alllmtte origin denomtnttlTe ; Mid that tahuj 
appuent loota are of the same charactei. The denominatiTea which aie go 
called differ fram theie only tn that their origin ia recent and nndlggnlsed. 

10&4. The gTammarians teach that any noun-etem in 
the language may be converted, without othei addition than 
that of an Q a (as uniou-vowel enabling it to be inflected 
according to the second geneial conjugation] into a present- 
stem, and conjugated as such. 

ft. But aach fonnaliona are rare in actual uae. The BV. hu a few 
Isolated and doablfol eiamplea, the clearest of which U bhif4kti he healt, 
from hbi^ phytician; It ia made like a form of the loot-claae; abhi^yak 
eeeml to be Ita imperfect according to the Da»l olaiB) and p&tyate he 
rule* appeals to be a deDominatlie of p&tt matter; other possible caiee 
are i^a^as etc., k^pAi^anta, tarofema etc., vanufMita, bhun^anta, 
vinanvati. From the other older teite are quotable bavy&at (TS.), 
4>9lonat (JS.), uamfilati (^B), Bvadhfimabe (((3). And a consider- 
able DDmber of instanoea, mostly isolated, are found in the later langDige: 
e. g. kolahant (UBb.), arghanU (PaFic), abJaU (^atr.), gardabliati 
(SD.), utka^fhate (SD,), Jagannetratl (Pru.), kell9TetaBahaBTa- 
pattratl (Pras.). 

lOSB. In general, the base of denominative conjugation 
is made from the noun-stem by means of the conjugation- 
sign tr ;&, which has the accent. 

a. The identity of this ya with the ya of the so-called eausatlve 
eoDjDgaUOD, as making with the final a of a nonn-etem the caasallTe-slgii 
aya, la hardly to be qneitioned. What relation It SQitalns to the ya of 
the ya-cliBS (759), of the pasaire (766), and of the dorivativa iDteasive 
■tern (1016), is much more doubtful. 

IOB0. Istermediate between the denominative and causative 
coDJngations stands a class of verbs, plainly denominative in origin, 
bnt having the caaBativ« accent. Examples, beginning to appear at the 
earliest period of the language, are mantr&yate ^eake, take* counsel, 
[ftom mantra, yman + tra), kirt&yatl eomm»maratea {from kirti, 
)/kf praiei), arth&yati or -te mak»* an object of, leekt [from &rtlia goal, 
o^ect), var^ayati depict* (from var^a color), katbayati or -te gives 
the how of anything, relates (from katham howf), and SO on. 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 cansative meaning, are reckoned by the grammarians as a 
separate oonjagation-olaeB, the our-class (above, 607, 775). 


D,j,i,.c.., Google 

1067—] XIV. Secondabt Cokjuoatioh. 388 

1067. Denominativea aie formed at evety period in the 
history of the language, boxa the earliest down. 

a. Tbey are frequent in RV-, which contains over a hundred, 
of all varieties; AV. has only half as many (and personal forms from 
hardly a third as many; from the rest, present participles, or deriv- 
ative DOuns]; AB., leas than Inentf ; ^B., hardlf more than a doceQ; 
and Eo OD. In the later language they are quotable by hundreds, 
but from the vast majority of stems occur only an example or two; 
the only ones that have won any currency are those that have assumed 
the character of "eur-olass" verbs. 

1068. The denominative meaning is, as in other lan- 
guages, of the greatest variety; some of the most frequent 
forms of it are: be like, act as, play the part of; regard 
ot treat as; cause to be, make into; me, make application 
of; desire, wish for, crave — that which is signified by the 

a. The modes of treatment of the stem-final are also various; 
and the grammarians make a certain more or less definite assignment 
of the varieties of meanicg to the varieties of form; but this allot- 
ment finds only a dubious support in the usages of the words as met 
with even in the later language, and still less in the earUer. Hence 
the formal dassificatioo, according to the final of the noun-stem 
and the way in which this is treated before the denominative sign yi, 
will be the best one to follow. 

1069. From stems in a. a. The final a of a nonn-etem 
oftenest remains unchanged: thus, amitray&ti plays tkt «nm^, u 
hottile; devay4tt cuttinatM the godi, it piout. 

b. But final a is also often lengthened : thus, agh&y&ti plaav 
miiehief; prlfSyite holdt dear; a^Hy&ti leakt for horttt; a^anKyUl 
dtiirea food. 

o. While in the Veda the larioaa modes of denominatiTe fonnatloa 
■le veil diitrlbaled, no one allowing > maikid pTepondennee, In the later 
lingnage tbe vaat majoritr of deDoninativei (fully aaven eightha) are of 
the two kind* Jaat noticed: namely, made fiom a-atema, and of tbe form 
aya or Sya, the former predominating. And there li seen a decided ten- 
dency to give tbe denominativea In ajra an active form and tianaltiTe mean- 
ing, and those in &ya a middle form and intranaittie or Tefletive meaning. 
In >not a few caaM, parallel foimationa from tba aama stem illuatrita dila 
diaUnction: e. g, Icalnqayati makt» turbid, kolofSyate m or frawnwt 
turbid; tani^ayati ref'u nmo^M, tami^Sy ate u rtfuvenattd; qittallayati 
lootetu, fitbilftyate gr<no» iooie. No dtaUnct traces of this diatinctoD ua 

itizecy Google 

389 Dbnohinative. [—1064 

MMgDluble in the Toda, illhoagli there alio coneipondlug [onns itith abort 
a and irith long R aomettmea atuid aide hy aide. 

d. Final a U sometimes ch&nged to 1 (ver^ rarely i); thus, adhTttriT&ti 
performi the tacriJUt ; tavlfly&tt ■« mighty ; patrl;&tl or patrly&ti deiiret 
a ton; mkhsiyiM erana JluK; Bi^iyats it rtady; oandrakKnOyati it 
moonstorulikt. Not Bftr itema ot thia form ate quotable. 

e. It is oecaaionally dropped (after n or r): thna, taraQy&tl >> rapid; 
adfaTUT&tl performs (Ae laerifice. 

t. Other modea of treatment are sporadie: thaa, the addition of s, ■■ 
Id etanaByatl leehi tht breatt ; the change of a to e, as In varey&tl 
play a tht uiootr. 

lOeO. From Items in ft. Final A ninallT remains, as in gopfty- 
&ti pbtyt tlt4 htri*man, proUelt; vV^KaijaXA. fighU\ bat it it somettmea 
treated la the other methods of an a-stem - thus, pftanTatl Jighti; tllottft- 
mlTatl actt Tilotiama. 

1061. From atems in 1, I, and a, d. Suoh items are (especially 
those in u, u) very rue. They show regnlarly i and Q before 7a: thni, 
arStlr&tl (also -tl7-) ploU irjjury; JaniyUl (also -niy-) tttk* a tei/e; 
gakbiy&tl dttirtt fritndthip; nftiijats turtu urotnan; — qatrflf&ti ael* 
the foe; rJOy&ti it ttraight; vasily&tl dtairtt tctaith; asEljriti frumiira, 
i» diiconUnted ; with short n, gttuy&tl lett in motion, 

a. More rarely, 1 01 a li treated as a (or else Is gnnated, with loss 
of a y 01 v): thus, dimttay&tl eamtt tnorting; laghayatl makat eatier. 
Sometimes, aa to a (abo«e, lOBBf), a sibilant is added; thni, avl^tl 
M vehement; tirafy&ti tavet. From dhi, BV. makes dhlySyUe. 

1063. From other io«el-stem«. a. Find r i» changed to ri: 
thus, mitrlTiti Irtatt at a mother (only qnotable example). 

b. The diphthongs, in the few cases Ihat oacar, have their Anal ele- 
ment changed to a semivovel : thus, gavy&tl letkt cattle, goet a-raiding. 

1068. From oonsonant-slema. A final conionant osually remalna 
before ya: thus, bhlqajy&tl jiloya t&e phyeician, cures; iik^ai^y&ti actt 
like a bull; apasy&ti itacliee; n&mtiay kU payt reverence; BumanasyAte 
u favorably ditpoeed; tarnfy&ti Jighti. 

a> But a final n Is sometime* dropped, and the pteoedlng vowel treated 
as a final; thns, r^fiy&ta or r^tyiU m kingly, ftom T&jan; -korma- 
yaU from •kanuan; evftmlyati treiUt at master, frocD avftmin: v^f^ 
y4ta bon vrfan is the only example quotable from the older langnage. 
Sporadic cases oocnr of other final consonants similarly treated : thna, oJ&- 
y&ta ttma ojas, -manilyate from -monaa; — while, on the other hand, 
an a-TDwel is occasionally added to such a con aon ant before yai thus, ifa- 
y&tl ftom I9, Batvanftyati from satran. 

1004. The largest claia of conaonantal atems are those showing a a 
before the ya; and, as has been seen abote, a albllanc ia sometimes, by 
analogy, added to a Bnal vowel, making the denomltlie-sign virtnally sya 

Digitizecy Google 

1064—] XIV. Secondary Conjuqation. 390 

— 01 eves, with ■ tlao Bddtd aflei ui i- or n-To«e1, uyRj and thli come* 
to be Tecogaized b; the gfnmmiriuiB u tu indepsadanC atgn, formlag deaom- 
initiv«( that expiegi de»lre: thni, Bum&khuy&te m merry; jIvanasyB- 
(In •ayt /ace of life); vnasyati ifafirM <A« tTwie (the only qooUble eum- 
ples); madhofTatl 01 madhvMyati loagt for honey; k^irasyatl cram 

1065. The grammiilinB leckoD ai » special clan of denominaiiTe* 
in kllmya what are teslly only ordinary ones made (iDm a eompoond noun- 
■tem having kSmK ai its Qnal member: thiu, rathakSlOTati long* for 
the chariot (K, : only example foand iu the older lanpiage); artbaUm- 
yatl deeiret wetdth; pntmk&myKti loithee a eon (the only quotable eiam- 
plei); coming from the paiaeBiiTe compounda rathakSma etc And &rtli&- 
pfiyati treate at property i» a (tele qootable) eiample of a atem kanng 
the Prakrttle cauiitlve fonn (104Sii), 

a. Sterna of anomalone formation are drliglutya from dirgha, dra^ll' 
aya from dr^bft, and perhapa mradaya from mpdu- 

1060. a. A anmber of denomlniti«e atema occur iu the Veda foi 
which no correspoitding noan-Btema are found, allhoogh for all or nearly 
all of them related words appear: thus, a&kuy&, Stabhuyi, ifudhya; 
dtaiqa^yA, ri%B^4, ruvai^ya, hnvaDya, iqa^yi; ratharyi, gratharyi. 
Baparya; iyasya ((B.), iraey&, da^aayi, makhasyi, panasyi, aa- 
casy&. Those in aoya, especially, look like the beginninga of a ne« 
con] Dgatton- class. 

b. Having atill more that aspect, howeT«i, are a Vedic group of etemi 
in Kya, which In general have allied themseWes to preaent->y*t«ma of the 
nS-elaEs (7S2), and are fonnd alongside the forms of that clasa: thu«, 
grbhfiy&tl beside g^bh^tl. Of sucb, HV. has gpblU^&, matfaiyA, 
pru;fiy&, mu;By&. grathSya, skabbfiyi, stabhSyA. A taw others 
have no nil-class RompaniDus: thus, dam&yA, gamfiy&, tudSy& (AV.); 
and pon&ya, na^Sya, vr^Bya (KvrS rain), vas&y& (yvaa tiothe), and 
perhaps a^Sya (Vaq attain). 

o. Here may be mentioned also quasl-denomiuitlves made from ono- 
DiatDpoetic combinations of sounds, geneially with repetition: e. g. kitaki- 
^ya, thatatbatarAya, mi^ami^fiya, 9ara9arBya. 

1067. The denocDinatlYe sterna in RV. and AT. with cauaatlve accent- 
uation are: RV. a&kli&ya, artb&yo, i^a (also Ifayd), iirj&ya, r^^ya, 
kfpiya, mantr&yo, mrg&ya, vavr&ya, vSJiya (also vfijayi), Tfl&ys, 
BUf v&ya [also BUf vayA) ; AV. adds klrt&ya, dbup&ya, p&Utya, vlr&ya, 

a. The. accent of &iuiiya and li&staya (BV.) Is wholly anomalona. 

1068. Inflection. The denominative stems aie in- 
flected with regularity like the other ateme ending in Qa 
(78Sa) throughout the present-system. Forms outside of 

ioy Google 

391 Denouinativb. [—1069 

that syBtem — except horn the stems which axe leckoaed 
to the causative oi our-olass, and which follow in all le- 
speots the rules foi that class — are of the utmost larity. 

ft. In RT. oecuM no form not belonging to the preaent-gyttem, aieept 
ena^e (with mi piohibldTe), » if-totitt 2delng. (at. 1048]. Further 
eiample* of thii loriBt ue KaOyit (9B.), ptpftyi^^a (TS,: pi., vlth mi 
prohibitive), »nd Btqrf^fatH (VS. etc.). The form &AaparT&it (AV- 
Tlv. 3. 20), with U for I (eSBo), might be *ari«t; bat, as the metre 
*bow(, is probably ft corropt re»d<Dgi amanasySit, certainlf imperfect, 
sppsars to occur in TB. (11. 3. &). Other forma begin to appear in the 
Biahmtnu: e. g. the tntnres gopKyifyati ((B.), mesb&ylqyiuit, Ita^- 
^nyiqy^t, qlkSrifrJait (TS.), the pirtioiplea bhi9aj;lt& (? JB. -jita) 
and lyaaitii ((fi.), ka];k4ii7lt&, qlklti, >nd meshiU (TS.), the gerund 
Hjfa^l&kq^a (^B.), and M on. In the later Uugmage, alao, forms out- 
side the preBent-tfttem (except the participle in ta) ire only aportdie; and 
of tertiary oonjngatlon forme there are hardly any: eiamplea are the caiisa- 
tivea dliamaraya and aatlyaya (MBh.), and the desideratlTe abhi^e^a- 

y^a (915.). 

b. Honn-dertvatiiea ^m denominatlTe etemB totlow the andogy of 
those from causative atema (lOSlg). In the older language, thoae in n 
and & (especially the formtr) are much the most Dnmeroas; later, that in 
ftna preTaiis over all others. 



1060. One peiiphrastic formation, the peiiphiastic 
futuie, has been already described (042 ff.J, since it has 
beoome in the later lanpiage a lecognized part of every 
verbal conjugation, and since, though still remainig 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 
olasB is — 

ioy Google 


The Periphrastic Perfect 

1070. This (though almost uaknown in the Veda, and 
ooming only gradually into use in the Brihma^as) ia s 
tense widely made and frequently used in the olaaeical 

a. It is made by piefixing the aousative of a deriva- 
tive noun-stem in EIT S (accented) to the peifeot tense of an 
auxiliary verb : namely, of ym kr maAe, moie often of 
yC(Q as be, and very rarely of VH bha be. 

b. In tbe oldei Ungnige ((m below, lOTSd), kf !■ tlmoat llie onir 
aatiliu; oMd In making thii tanse, aa oeonnlng Tery few tImM, tnd bfaO 
navBE. Later, also, bbQ ia qnite nre (it i» fonnd nine time* in HBh., 
tix timet In Bgh., and a fev timea elMwhete), but as gaini Tery gr«afl]> 
in coiiancy, having bsMine tlie naual aniillary, irbila kf li only exeepUonal. 

o. Somewtiat ilmilat formatloni irjtb yet other auxUUiiu ue not 
absolntaly noknoitn in the later language: thni vaxajMx praoakramtta 
(MBh.}. pflraySm (ete.) vyadhtu (VTncailtra), mrgs7«m avfttaTt (ib.). 

1071. The periphrastic perfect ooouis aa follows: 

a. It is the accepted peifeot of the deiivatire conjuga- 
tions: intensive, desideiative, causative, and denominative; 
Uie noun in SIT i being made from the present-stem which is 
the general basis of each conjugation: thus, &om y^fj hndh, 
intensive M HM I H bobudhSm, desiderative ^HrHIM bubhutsim, 
causative Mt^J U IH bodha^bn ; denominative H ^ U (4 mantxay- 

b. The farmitlon from caaeative atema (Ineladlog those deDomlnativea 
which have asiumed the aapeot of oaniatiTea: lOfifl) ia by far the moat 
Tiequant Only a few dealderativea are quotable (1034a), and of intMi' 
liTca onl; JSgarKm. ftaa (1020a; beside JsOBg&ra). 

o. 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, <b|IMIH Ssftm from y^[ff 
Ss sit; ^5IPl Dtfltai from j^_^Ik* see; 3«hT^ujjli&m &om 
|/3(F Mih forsake; ^tfm edhBm from y^edh thrive (the 
only examples quotable]. 

ioy Google 

393 PEaiPHRABTic Perfect. [—1078 

d. EsMptod >ie die loola Ap ind ilkotl, tnd tbote beslnniDg wllh ft 
Itefote two conwointi (ind Ukiof Bn u redupUeMlon : 786). 

•. The HWtt (thit It, «tem» reckoned bf the gniDiniriuis m iooU) of 
mora thui one irUiblo have their perfeot of this formation; thnt, oftkSsltin. 
But Hxipi (713] 1* add to Coim Or^onava ouly; vhile Jigf (lOSO) 
nukei a perfeot of either fomation, and dftrldrft (1084ft) la add to do 
the eame. 

t. A ttw Dthar root) make the periphraatlo In addition to the lunal 
reduplicated perfect. Thai, in the older language only are toond the atema 
oB7lin, tSfftm, ailaytja, vBe&m (^yvoB dwell), vidim (f'vid knovi), 
TJtts^aa, and the reduplicated ttema bibh&rtia and JuhftTim; the later 
language adds ftyim, Jftyim, dfty&m, nftyKm, smajBrn, bvayAm, and 
the reddpllMted biblinTlin; and the grammarlana teaob like formation* 
from Of, kb, and the reduplicating bri. The alem la made in eyery mm 
tzma the pteaent-ttem with go^a of a final TOwel. 

1072, The periphrastio peifect of the middle voice u 
made with the middle inflection of i^ kf. Foi passiTe 
use, the auxiliaiiea ERT m and H bha aie said to be allowed 
to take a middle inflection. 

ft. One or two lata example* of bhfi with middle inflection have been 
pointed out, but none of aa. 

b. It is mmacesutry to give a pandigni of this fonnatiOD, aa 
the inflection of Uie ansiliairiee ia the same as in their iadependeot 
use: forthatofykr, see 800k; of y^bu, see SOOd^of ^as, see 800m. 

o. The oonnectloQ of the noun and auiiliary ia not lO doae that other 
word* are not occaalonailf allowed to come between them : thua, mlmfiA- 
Bltm «v4 oakrA ((B.) Aa tntrely ipaculattd; vidftdk vK Idftm ayadi 
oakfiTft (JB.) A« i>«rt7y hnew> ihii; prftbhr&fifoySib yo naghn^aih oa- 
kSra who made Naghu»ha faU headlong (Rgh.)- 

107S. The aboTO is an aocoant of the peripliraatio formatiOD 
with a derivative noun in Km as it appears espeoially lo the later 
language; earlier, its aspect is rather that of a more general, bnt 
quite infrequent, combination of such a noun with various forms of 
the root Iqr. Thosi 

a. Of the periphrastic perfect occora only a atngle example In the 
whole body ot Vedic texts (metrical): namely, gamayiih OftUia (AT.). 
In the Brihmapat examples from cauaatlTe atoms begin lo appear more 
freely, bnt are eTerjwhere tew in number except in ^fi. (which haa them 
fk-om twenty-foDf roots, and a few of theae in leTeral oocunenoea). From 
deaidetatlve atemt they are yet rarer (only aeven occnrtenoea, flve of them 
in (jS.: lee 1034a); and from intenilTes they are nnknown. The peri- 
phrastic perfects ot primary conjagatlou were noted above (1071 f; in (B., 

ioy Google 


eight Btemi ind ibout eighty occuirancea, chleOy fmn ikf, bid, ind -rid; 
tbst ttom vld ii found in tb« gieiteat number of text*). 

b. Fornu with the u>rist of the •niilluy aie in the oldest Biihmanu 
M nnmerout u tbotevith tbe perfect. Tha*, with Bkar oecnt ranuvwn 
(K.J, Jaiuqribii and sfida^tm >nd avadt^bu end ath&payiia (H^.); 
end Kith akran, vid&n (TS. TB. MS.). With tbe eoii*t optetlTe or pre. 
fetiie bu been found only pSvayalfa krlySt (HS.). 

o. Like combiuatlonB with other temet ute not entirely unknown; 
tboa, JuhavSlfa karoti (^^S.). So alao in the liter Imngnaee, where heT« 
been found quotable half-a-dozen inch caaea ai vldftdi karoti (Pifir.), 
vldsib karotu and kurranta (Paric. etc,]. 

d. Only two or three caaei of the nae of ae Intteid of kf as aniil- 
iary are met with in the older language: they are matitrftTKin taa (AB. 
OB.), JanaykD Sbs (?<U.), and Ikfam Baa (^(S.). 

e. A BlDgle example of an accented auxHiary ia met with in the aCMRl- 
nated lexU; namely, atiraoayaifa oakrua ((B.). At was to be expected, 
from the nature of the combination, the noun alM letaina Its accent (com- 
pare 646). 

Participial Periplirastlc Phrases. 

1074. The frequent use, especially in the latei language, 
of a past or a future passiTe participle with the copula (ot 
also without it) to make participial phrases having a value 
analf^ous to that of veib-tenses, has been already noticed 
(899). But othei similar combinations are not unknown in 
any period of the language, as made with other auzibaries, 
or with othei participles. 

a. They ocenr evan in the Veda, but ire far more common and 
oonapicuotiB in the Btihminas, and become again of xnlQor iccount in the 
later ImgDige. 

107B. Examples of the various formations are aa follows: 

a. A (uBuatly present) paiticlpis with the teneei of the Terb 1 go. 
Thia ia the combinatioD, on the whole, of widest and moat ileqaent occur- 
lence. Tbu»: iyajvano vibh&jann 6ti vMo^ (RV.) he eeer gieet mcay 
the wealth of the tum-offarvr; yatliK B&cyA v&sa^ saihdadliad ly&d 
evam erli tfibhir ynjflasya ohidraih sadidadhad eti (AB.) j%ut ai 
one toould mend IkabituaUy] a garment tctth a needle, to utith thete one 
mende any defect of the taerifiee \ agnir v& idaib T&tfvKnaro daliann 
tit (PB.) Agni Vaicoanara kept burning thia creation; ti 'aorfth p&rfi* 
jlt& y&cto dyivSpptliivi dpS^rayan (TB.) thoae Asarae, getting beaten, 
took refuge toith heaven and earth; ti 'syag^hah pa9&va upamnryi- 
mi^S iyu^ (^B.) the animale, his family, would be eontinuaUj/ destroyed, 

ioy Google 

39f> Participial Phrases. [— 107B 

b. The ucDa with the retb oar go (continualfy or ItdbitvaUy) ilgnirying 
■tUl moie distinctly thtn the pioeedlng a continued or habltnal sctton. Thai: 
agn&v agnif cnuratt pr&vlffa^ (^^0 Agni it eonttantly prt$tnl in the 
Jire; adBi;4yuh da^i^Da ghnanta; oarantl (PB.) tkey makt a practice 
of beating taith a rod ^e/uit it undaerving of pvnithmtnt. 

c. The aims with the veihs Ss tit aod Bthft (toni^, with a like nieen- 
lag. Thni, JnliTata Bsate (K.) they continue tacrijicing \ te ■pakr&mya 
prativlvodato 'ti^lh&n (AB.) they, having gone off, kept vehmtmtly 
refuting. In the later language, Btb& la the verb ortenMt uced, wllb piedt- 
ntea ot t«t1oui kind, to make a verbal phrase of contJaaauce. 

d. A present or rnlnie or perfect participle with as and bli& be. 
The parUdple la aftenest a fature one; em oalj I* need In the optatlie, 
bhd oinally In other forma. Thus: yab pttrram anljfiiia^ ByU (AB.) 
vhoever mag not have made laerijice before; samfivad evayi^&e kur- 
vft^ ftsan (OB.) they did the tame thing at the lacri^ce; parlkrl^anta 
Baan (HS.J they were playing about; y&tra euptvA pjinar lA Vadrfi- 
uyka bh&VBtl (VB-) when, after tleeping, he ii not going to fall atleep 
again; havysrii hi vak^yan bbavatl (AB.) far he it intending to carry 
the lacrifice; dOsyant syKt (K.) may be going to give; y^na vahanena 
■yantajriiit syjit (VB.) uith ujhat vehicle he may be about to drive. True 
ezpreastona for perfect and plupeifecl and future perfect time aie capable 
ot being made by su«h meana, and now and tlien are made, but in no 
regalu and contlnawfaeMoo, 

Composition with Prepositional Prefixes. 
1076. AH the foims, peiBonal and other, of verbal con- 
jugation — of both primary and secondary conjugation, 
and even to some extent of denominative (ao far aa the 
denominative stems have become assimilated in value to 
simple roots) — ocoui 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. 

a. Practically, In the later language, It la aa If a compounded root 
were formed, out of root and prefix, from whioh then the whole conjngatlon 
(with deilvattvea: below, chap, XYll.) la made, Just aa from the aimple 
root Yet, eren there (and gtlU more In the older language: 1081 a-o), the 
combination ia so loose, and the members retain ao much of their independent 
Taliie, that in most dictlonarioi (that of Sir Honlei Wlllluni la an eieeptlon) 
the eonjagatlon of each root wltb proflzet is treated under the simple root, 
and not In the iJpbabetie order of the pteHi. Deriiatiie words, however, 

i, Google 


MB by unlventl tgTeeiiient giien In theli iDdspeudent tlpbtbetic pUca, like 
alaipl« -woidi. 

1077. Thoae verbal prefixes whioli have value as such 
throughout the whole history of the language are given 
below in alphabetic order with their fundamental meanings: 

sin 4ti acroM, b eyond , past, over, to excess; ■ n 'i-~ 

isf^ Adhi ahone, o ver, on, on to; I . 

^p{ ktxa. after, alonff, tovxtrd; '■ ■'■ '-^ -.^'-^^ 

At! 51 *ot*J^ between, among, witMn; yii/^"^' 

qq 4pa away, forth, off; ' 

^wk Api unto, close upon t>r on; 

QTH abhi to, unto, against (often with implied violence); 

515 4va dovm, off; 

m t to, unto, at; 
■ - 35" «d up, up forth or out; 

:5^ lipa to, unto, toward; 

H ni dovm; in, into; 

f%?^^nia out, forth ; , . 

qjT pirS to a distance, atoay, forth; ,. L '- <■ •" 

qf^ pAri round about, around', 

51 pt4 forward, onteard, forth, fore ; y^/ - 

V^ prdtl in reversed direction, back to or against, 
in return; .'•' 

R vi apart, asunder, away, out; Hi^ 

W\fAm along, with, together. 

a. Some ot tbase, of ooatte, ue naed much mote widely «nd fteqaently 
than othen. In order of frequansy In the oldei languige (u eitlniated br 
the nomber of loots with which they aie lOand Died In EV. and AT.), they 
■uad aa followi : pra, A, vi, sam, abhl, nl, ud, part, ana, upa. pratl, 
ava, nia, atl, apa, para, adbi, apl, antar. Apl la of very limited 
uae aa pTcDx in the latu lingoage, having beoome a conjanotloa, too, alto, 

b. Xke meanlngi given above ate only the leading onee. In oombinatlong 
of root and piefli they nndetgo much modification, both llleial and flgutative 
— yet aeldoai in euch • irty that Ihe atepa of tianaltion hom the fnnd- 
■nental aenie aie not ewy to trace. SometlDiei, indeed, the Talue of a 

ioy Google 

397 Vekbal Pbefizeb. [—1081 

root U hwdl} perceptibly modtded bj the eddltlon of tlie prellx. An Iri' 
teadTi force is not infreqneatly glTen by purl, tI, and aam. 

1078. Prefixes eaeentially akin with tbe above, but more diB- 
tJDctly adverbial, and of more reBtricted nse, are these; 

Aotui (or &eb(l) to, unto: tolerably frequent tn RV. {used with 
over twenty roots), but already nnnsnal in AV. (only two roots), quite 
restricted in B., and entirely lost In the later language; 

Kvia forth to tight, in view, used only with tbe roots bhd, as, 
and h^; 

tlr&a through, erouujayt; out of gighf: hardly nsed except with 
br> dbS, bha (in RV., with three or four others]; 

pnr&s HI front, forward: nsed with only half-a-dozen roots, 
especially hf, dbftf 1; 

prftdus forth to viete: only with bhQ, as, kf. 

a. A few otben, >b bahla outtide, vinft teithitut, Alam (with bbQ 
and kr) luffieienily, proptrfy, B&k;&t in vimo, tre itlU leu Temoied from 
ordinary adrwbs, 

1070. Of yet more limited nse, and of noun- rather than adverb- 
Talae, are: 

TTSd (or 9»thP), only vlth db& (in RT., once alio with kf): 
^addbS believe, trtdit; 

bin, only with kf (and absolcte in the cluilcil language): blfikf 
make tlie tound hlng, low, murmur. 

a. And beilde theie itand yet more fortnttoas combination a: lee 

belDw, loei. 

1080. More than one prefix ma; be set before the same 
root. CombinationB of two are quite uBual; of three, much 
lesa oommon; of more than three, rare. Their order is in 
general detenuined only by the requirements of the meanii^, 
each added prefix bringing a further modification to the 
combination before which it is set. But EIT S is almost 
never allowed, either earlier or later, to be put in front 
of any of the others. 

a. Tbe very rare caaes of apparent piefliion of K to anotfaet prefix 
(» Svihanti HBh., ftvitaavftaftb BhP.) aie pethapa beat explained as 
having tlie a used Independently, ai an adverb. 

1081. In classical Sanskrit, the prefix stands immediately 
before the verbal form. 

a. Id tbe earlier langnage, however (especially In the Teda; In 
the BrUiinana lesa oft«n and more restrietedly), its poslUon is qnite 

ioy Google 

1081—] XV. Pbeiphrastic asd Cohpoithd Comjdoatioh. 398 

free: it may be sepanted from Uie verb by another word or wotAm. 
and may even come after the form to which it belonga; it may also 
Btand alone, qoalifying a verb that is nnderatood, or conjointly with 
another prefix one that ia expressed. ^^^- -Ij- 

b. Tbna, ai devofi i 1i& vakqyati ^V.) he »luill bring Hi* god* 
hither; pr& 9a jtyQA^ ttri^t (A-T.) may ie lengthen out our licee; t&v 
i T&tam upa dravit (RV.) do y« two comt Ai'tAar quickig; g4mAd 
v^ebbir i bA na^ (R^O •'U'y A« eome teitk gift* hither to tu; piri 
malb p&ri me prt^iib p&ri ^a^ pfihl y4d db&nam (AT.) protect me, 
my progeny, and lohat weaUh we oien ; ^tal^ Bad;& t, oa p&rfi oa y&nti 
(AV.) from tehence every day they advance and retire ; vj ib&ih BJ