Mahabhashya Volume 1

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 marginalia 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 this resource, we have 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 from 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 attribution The Google "watermark" you see on each file is essential for informing 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 liability 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/|

/

HARVARD
COLLEGE
LIBRARY

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by }

Google

Digitized by

Google

THE

«* _ .£.

^ASHTADHYAYI^OF PANINL

TRANSLATED INTO ENGLISH

BY

fcJRfSA CHANDRA VASU, B. A,,

Vakil, High Court, M-W. P.

-:-0-0-o-:

JtUahabab:

Indian Press.

i 891.

^

(All rljjhU referred.)

1

Digitized by

\

•y

Google

Digitized by

Google

fEtbntrg

THE BEQUEST OF

Cfl£VUcJ&£i

PROFESSOR OF SANSKRIT
lSSo-t9S<

Digitized by

Google

Ju^vfotyrf/ft*-

<$

/w^***

Digitized by

Google

/.£>

HAfT.'ASD COLLEGE UBRAR?

r^OMTiIICSTATEOF

CHAHLEC f.OC.ttVELL LANMAN

UASCH15.1941

JMlahabai:

PRINTED BY THE INDIAN PRESS.

\

/

Digitized by

\

PREFACE.

+■••»

Since the advent of the British rule and the peace and pros-
perity that has followed in its train, India has witnessed a glorious revival
of her ancient literature, in which is embodied some of the highest
philosophies and religions of the world. Among the various blessings
which our benign Government has conferred upon us, none can be
greater in value or usefulness than this revival of Sanskrit. Our schools
and colleges are annually turning out hundreds, nay thousands of scholars,
who have entered upon the study of Sanskrit literature, and have thus
learnt to appreciate the beauties of this language. Very few of them,
however, have the opportunity of studying the language, with that depth and
fulness, as it was and is mastered by the Pandits of the old school. To pro-
perly understand Sanskrit language, and especially that portion of it in which
is locked up the highest aspirations of the ancient Aryan hearts vis. 9 the
Vedas, the Br&hmanas, the Upanishads &c. it is absolutely necessary to have
a complete knowledge of the Grammar elaborated by Panini. The Grammar is
reckoned as one of the VedArtgas, or the helps to the study of the Vedas; and
it is unquestionably one of the most important of the Vediirtgas. The four
thousand sutras of P&nini contain within themselves almost all that a student
need know to enable him to understand the language of the Vedas.

Not only is this excellent treatise of PAnini necessary for those who
are desirous of learning the ancient Sanskrit literature, but a knowledge of
this is even necessary for understanding the modern Sanskrit, which is
modelled on the rules laid down by that great Grammarian, whose aphorisms
are being constantly quoted in all Vedic commentaries, and classical authors
and law books*

Further, as a master-piece of close reasoning and artistic arrangement,
it ought to be an object of study with every one who wants to cultivate his
intellectual powers. In fact what the Geometry of Euclid has done towards
the logical development of the western intellect, the Ash{idhy4yl of P&nini
has fulfilled the same purpose' in India. No one who has studied this book
can refrain from praising it It has evoked admiration even from the Sans-
krit savants of the west. Professor Max Muller thus gives his opinion about
the merits of this excellent Sanskrit Grammar :— " The Grammatical system
elaborated by native Grammarians, is in itself most perfect, and those who
have tested Panini's work will readily admit, that there is no Grammar in any
language, that could vie with the wonderful mechanism of his eight books of
Grammatical rules."

The style of these sutras is studiedly brief, but then this brevity is its
greatest recommendation. That, which appears to many obscurity and ambi-
guity in the sfltras, vanishes before the clear and exhaustive explanations of

Digitized by VjOOQIC

( >i )

the commentators ; and shows the extreme skill and wonderful ingenuity of
the author of these aphorisms. These sfitras, therefore, which though at first
sight may appear difficult and repulsive, if once mastered, will enable a
student to know and remember more of the Sanskrit Grammar, than he can
ever learn through other methods, with twice that labour. Unfortunately,
however, for our college and school students, and also for that vast majority
of our English reading countrymen, whose number is daily on the increase,
and who depend for their knowledge of what is contained in Sanskrit books,
on English translation of Sanskrit authors, no translation of this important
work exists in English. To supply this want, I have undertaken to translate
P&nini's aphorisms, as explained by the Commentators Jay&ditya and V&mana
in their well-known book, called K&£ik& vritti and issue one chapter every
month. Though, it is not a close translation of the whole of KisikA,
it may be regarded as a free rendering of the most important portions of that
book. I have closely followed on the foot-steps of those authors, translating their
commentary, explaining it where necessary ; and in short, making my work
a help to the student, desirous of studying the KAsik! in the original.

The translation of Sanskrit texts, especially those like K&£iki, is
always beset with great difficulties, even for the masters of Sanskrit learning.
For a beginner, like the present translator, those difficulties were many
and great, and I am fully conscious that here and there, I may have
failed to grasp the full drift of the arguments of the authors of K&£ik&. But
on the whole, I have spared no pains to render this work as free from errors
as lay in my humble power. I shall feel much obliged to those gentlemen, who
will be good enough to point out any errors, or suggest any improvements, so
that I may be benefited by their advice.

I must here acknowledge the great assistance I have derived from the
well-known translation of Laghu Kaumudi by Dr. Ballantyne ; Mr. Iengar's
Guide to P&nini ; Professor Apte's Sanskrit Composition as well as from Dr.
Kielhorn's ParibhdshenduSekhara. I have freely quoted from these authors and
absorbed their rendering into my own, without distinguishing them by marks
of quotation.

When I first undertook the translation of Panini, I had thought that
the work when completed, will not occupy more than 1200 pages. But from
the present sample it will be seen, that that estimate was far below the mark.
The complete translation, together with the Introduction, Glossary and the
Indices, which I intend to add, will take up nearly double as much space *>.,
nearly 2000 pages or more. I have, however, kept the price of the book the
same, namely, Rs. 14 (payable by two instalments), for subscribers, who have
\ already got their names registered, or who will do so within the 31st January
1892, and Rs. 20 for non-subscribers. OCS

14M November, 1891. ^RfSA CHANDRA VASIH

• THE - - . . •;„. r

ASHTADHYAYI OF PANINL ;

Salutation to the Supreme Spirit.

■ o ' *

Now an explanation of words.

The term *pf " now" in the sfttra indicates a commencement, and points
out that a dissertation is to be offered on the science of words, viz. Grammar
and Philology. The term mjsm*M means explanation of any system. This
is an 4J1M4iK sfttra, and introduces the subject.

An aphorism or sutra is of six kinds, qRtf or ' a definition,' qR^WT or the
' key to interpretation/ faffr or 'the statement of a general rule/ RflH or ' a
restrictive rule/ •mnfchK ' or ' a head or governing rule, which exerts a direct-
ing or governing influence over other rules/ and srf^J or 'extended

application by analogy. 9
• ••» ■

The Praty&Mra Sfttras.

The above fourteen aphorisms contain the arrangement of Sanskrit
alphabets fpr grammatical purposes. The anusvira and the visarga, the
jihva-muliya and the upadh mlniya are not contained in the above list.
The final 3r in the consonants f , *r &c, is merely for the sake of articulation.
The final pure consonants in the several aphorisms as «T, ^r, ^&c., are non-
efficient or 73. The JJMJI4H is a grammatical symbol or abbreviation and is
formed by taking any letter which is nota non-efficient letter and joining it with
any non-efficient letter that follows it. This gives a name which stands for the
former ma-efficient letter and for all the other letters intervening between it and
the non-efficient letter. Thus ^p^ means all the vowels, fij means all the con-
sonants,*^ means all soft unaspirate consonants, ^^ means all hard unaspirate
consonants. Though numerous praty&hiras could be formed, practically
however, there are only 42 pratyihlras ; as given below : —

^i^i'^ir^i^ii^i^iH^iifr^i^jii^i^ i«g^? f V{ litter

* The same letter «[ is made use of as" Hjj^q* or an indicatory letter

1 both in the sixth sutra sn^ and in the first sj f ^[ . There arises consequently

i

Introduction.

the doubt, whether the praty&haras «r^ and ^ when they are employed in
P&nini's Grammar are formed with the «£ of the former or with the «[ of the
latter sfttra, and one might, on that account, consider it impossible to ascertain
what P&nini intended to denote by *r«[ and f^ when he employed these
terms. To remove this doubt there is the following paribhAsM :— ■

"The precise meaning of an ambiguous term is ascertained from in-
terpretation! for a rule, even though it contain* an ambiguous term, must never-
the less teach something definite."

Accordingly we learn from interpretation that the term *T% except
in Sutra I. i. 69, is formed by means of the «( of the first s&tra and that,
the term f\ is formed by means of the ^ of the subsequent sfitra.

^

\

\

\

.Google

3h^

BOOK L
Chapter First.

Rf^iHhf ii

1. *n, k and ^r are called v/iddhl

This defines the word vriddhi. The letters qqr, ^ and %fr are vpddhi
letters. The s&tra consists of three words vriddhi , &t and aich. *ni(
means the long ^T, the final ^[ being indicatory only, and is for the sake of
the praty&h&ra aich, and the praty&h&ra^ means the letters £ 'and *ft .

The indicatory ^ in iyn{ serves the purpose of showing that the very
form *jt having two m&tr&s or prosodial measure, is to be taken. This ^ also
joins with the succeeding vowels i£ and sfrbythe rule of OTT(L 1.70) or that
" which precedes or succeeds w," and indicates that these vowels must be taken
as having two m&tr&s only, though they maybe the result of the combination of
vowels whose aggregate m&tr&s may be more than two. A short vowel has one
m&tr&, a long vowel has two, and a consonant has half a m&triL Thus by a
rule of ^f^if or euphonic conjunction of letters IJT+W^WT, as TfT+WUPf: 3 *
Tfnj*H. Here fr has two m&tr&s and not four. So also in ^If^T: the -vowel
$ the resultant of 1JT+^ has only two, not four, m&tr&s.

Thus the initial vowels in the following secondary derivative nouns
are vriddhi forms of their primitive vowels. Hj|<jp: "bodily", from^p; 'body*
HfM* : " optional", from f^[r " option"; *JW»re " the son of Upagu", from the
word ^r^J: meaning " a sage called Upagu." Similarly in ^ifoOq: "belonging to
the house" the in is radical, and is also called vriddhi. The word vriddhi
occurs in sfttras like fffiq^l^K^Mifo (VII. 2. 1.) "Let the final ^ of the
base get vjiddhi substitute before the affix fn^ of the Parasmaipada".

2. *r, s and *fr are called guna.

Each one of the letters *?, 3 and if} whether radical or secondary
is called a guna letter; as the initial vowels in the following:— itf%, "he

4. GunaVriddhi Paribhasha [Bk. I. Ch. I. §. 3

> H •■;. .- V'i \i

moves"; m% " he comes ", <qfrcn% " he goes*" Here the initial vowels of the
roots 3T, f t and ^ have been respectively gunated into m$, ^ and *fr before
the thirdjperson-singular termination ft.

The terth Guna occurs in sutras like ftf^^: (VIl/3. 82). " Let there
be guna substitute for the ik of the root fkf " to melf ." -

*$T 3*135* mi T^rfSril ^5: 13*1-35*, (??%:, g*i:) n

3. Iijl the absence of any ..special rule, when-
ever guna or vriddhi is enjoined about any expression by
using the terms guna or vriddhi, it is to be understood
to come in the room of the ik vowels only (i, u, ri\ and
li long and short,) of that expression.

. This is a paribh&sh& sfitra, and is useful in determining the original
letters, in the place of which the substitute guna and vriddhi letters will
comeu- The present rule will apply where there is the specification of no
other particular rule.

Thus sfttra VII. 3. 84 declares: — " when a s&rvadh&tuka or an irdhadMU
tuka affix follows there is guna of the base/ 9 Here the sth&ni or the original
expression which is to be gunated, is hot specified, and to complete the sense,
the word "ikab" must be read into the sfitra. , The rule then being,
" when a S. or an A. affix follows there is guna of the ik vowels of the base."
The guna of f or f is^r; of gr or sr, is *ft, of ^for^fis «T£,of *y is
*p$; and their Vrddhi is {£, *ft, *tf^ and *jrf£ respectively. Thus sft+Wfrs 8
%+^n% =5 nri% "be leads."

Therefore, wherever in any rule in this Grammar, guna or vriddhr is
ordained by employing the terms guna or vjriddhi, there the word " ika^" in
the genitive case, meaning "in the place of ik", is to be supplied to complete
the sense.

The word 'ik' is thus understood in the following rules, and is there
qualified by the term *j^r or ' base' :— ftr^: (VII. 3. 82) fl^lPfr: (VII. 2.
114) J*FS*TTTO &c » (VII. 3- 86); jfT^yrcfr (VII. 4.1 1) ^Bftftlfc^: (VII-
416) ^rT^T &c, (VI. 4. 156).

In the following sfitras, the term 'ik' is understood and qualifies the
word ingia or base, gfa ^ (VII. 3, 83), ^ijvii^i W^' (VII. 3* 8 4)- x

These stitras will be explained in their proper, place?.
The word ' ik' has been used in the sfttra: to' show that the long *n, ^5 or
the diphthongs andf^ or the consonants, are Hot to take gunaox vpddhtby this

lA.ttf

1/

Bk; l. Ch\- I. §• 4. ] Exception to Guna and Vriddhi. -5.

J; - rule, as in the following examples AH\ "a carriage", TvTRft "He feels aversion,"

4{lH*U " confined."

Here **rpn% is formed by adding the affix f?fy third person singular ter-
** mination, with the class a affix ^JT, to the root t^; as t^+^+ ^=^4.^4.

ft. At this stage, rule VII. 3. 84, would require the guna substitution of the vowel
of the base before the s&rvadh&tuka termination *r; the guna substitute of
^ beings, there would be *^+?rf^,,T*niftr. But this is wrong, the letter
"5, being not included in the praty&h&ra ik, is not affected by rule VII. 3. 84,
and there is no guna substitution in this case. Similarly there is no change in
the long *n of otfj. Thus w + ^ (HI. 3. 1 15) -3TFPJ (VII. 1. i). # Here had
there been guna by Rule VII. 3. 84, the nr of m, would have been changed into
*T, and the form would have been ^T^ which is wrong. But it is not so, as *jr
is not an f^ vowel. So also ^r»jr + fKt - tiP*ML Here, there is no guna
substitution in the place of the consonant ^ ,

The repetition of the words guna and Vfiddhi ir. this sutra,
(though by ^•j^Rf or the supplying of a word from the previous sfitra,
these words followed from the two preceding sutras) is for the
^ „ purpose of indicating that, this is a restrictive rule applicable only there,

, where guna or vfiddhi has been ordained by the employment of the words

guna or vriddhi. Therefore, it is not so in the following cases : — ifh "sky",
q«jr: " way" m " he," fH\ " this." ifh is the nominative singular of the base ^.
It is formed by rule VII. 1. 84, which declares that " in the nominative singular
the letter *ft takes the place of ft^r." Here the letter sit is no doubt a vrid-
dhi letter, but as this substitution is not enjoined by the enunciation of the
term vriddhi, qft does not take the place of the vowel f of f%\, but it replaces
the final*. Thus f*+sfr+^«*ft:.

Similarly rule VII. 1. 85, declares that of the word qf*P![, *tt is the
substitute. The qqr is a vriddhi letter no doubt, but not being enunciated by
the term vriddhi, it takes the place of ^ of <rfat and not of f . Thus we have
q**tt: t the sf being added by VII. 1. 87.

Similarly *r: from tf|, which is formed by VII. 2. 102 "shorty
is the substitute of tyad, &c.," Here *[ takes the place of f . Thus we have
?f, which is changed again into ^ by VII. 2. 106.

4. The Guna and vriddhi substitutions, which
otherwise would have presented themselves, do* not

6. Exception to Guna and Vriddhi. [Bk. I. Ca I. § 4.

take place, when such an ArdhadMtuka (IIL 4. 114)
affix follows, which causes a portion of the root to he
elided.

This is an exception to sfttra 3. By that sfitra iks were to be
'gunated or vriddhied and by s&tra VII. 3. 84, the root is to be gunated
before all Ardhadhituka terminations. All primary affixes, that are applied
directly to roots in Sanskrit, are divided into two broad classes, called sarva-
dhituka and irdhadh&tuka. The affixes marked with an indicatory JJ ( called
ftpi ), and the conjugational affixes, such as fk\ , ft. &c, ( called f%^ ) are
s&rvadh&tuka 5 all other affixes are called irdhadhltuka.

Though by the general rule, VII. 3. 84, the root is to be gunated before ^

Ardhadhituka terminations, yet it is not to be so, before those affixes which I v/

cause the 5Tfa topping off or ellipses' of a portion or a member of the original C

dhitu Thus the termination *?nfir is a first person singular termination of /

the future tense ( off ) and is an irdhadhatuka termination; when this is

added to root f*r 'to excel' the f of fsr becomes gunated and we have %**nPf

"I shall or will excel. 1 ' But in those cases where there is a lopaof the member

of a root form, caused by an affix, there is no gunation. Thus in sffepn "a great

cutter," the original root is Sj/'to cut" which forms the Intensive verb frHo^q.

In forming the noun of agency from the verb oftoUl, an Ardhadhituka suffix

%T\ HI. 1. 134, is added. Thus vOfcgt + *p{. At this stage, by force

of sfttra II. 4. 74. which declares that "before the affix *pzj (IIL 1. 134)

the if of the Intensive verb must be suppressed," the ya is dropped and we

have sffa^+ST. Here by the general rule VII. 3. 84, the final ^ would have

been gunated. But by virtue of the present sutra, do guna takes place, be-^ •

cause here on account of the Irdhadhatuka affix *p(, a portion of the root, \ *

namel y, *, has been elided ; therefore no guna takes place. Thus we have J
feftcgq: "a great cutter." -

The word tfg "root" has been used in the sfitra, in order to indicate
that the exception does not apply where there is an elision of an anubandha
or of an affix. As in sr*+T*IT«5ft+TffT-5TflraT. Here the indicatory letter
3j has been elided, but that does not prevent gunation. Similarly in ^¥
the affix finj is added to the root ft^ "to injure." Thus ft^+fircj (III. 2. 75.)
-ft^+^fl. 3- 2. and 3.)- ftv + * (VI. 1. 67.) Here theaffix ^ is elided, but
nevertheless, the guna substitute must take place in ft^ . Thus we have
ft^+ • -*J (VII. 3. 86 and 1. 1. 62) -^ (VIII. 2. 39).

The word &rdhadhituka has been used to indicate that a sirvadhi-

tuka affix which causes the elision of a portion of a root, does not prevent

guna or vriddhi as in ctoftfa, (Rig Veda IV. 58.3) "he roars much." *+*^

-<fcw(IH. 1. 22). ttTO-rft^-tft+ft (II. 4. 74.). Here the sirvadh&tuka

Bk« I. Ch. I. §• 5. ] Exceptions to Guna and Vriddhi

r

affix finj causes the elision of ^r, a portion of the root The guna substitution
however takes place, and we have tWfr+f+pir (VII. 3. 94)- ffctifit, the
augment f^ being added by sfitra VII. 3. 94.

The word "ika^" of the previous sfitra is understood in this sfitra also.
The ik letters only are not gunated or vriddhied before such ftrdhadh&tuka
affixes; the other vowels may do so. Thus in *pnfir the equation is as
follows:— *J + **3^ + f^[ » *r + *p{ + f ( VI. 4. 33) . Here the d rdhadhituka affix
f^cg has caused a portion of the root ?pn(, to be suppressed, w., the letter ^r
has been elided. But nevertheless the *r of *nj is vfiddhied before f*T£
because ST is not included in the praty&h&ra ik ; and the exception contained
in the present sfitra only applies to the ik letters. So also frssj+ini (III.
3.18) «^+*T (VI. 4. 27.) -*pt: (VII. 2. 116. and VII. 3. 52).

5. And that, which otherwise would have caused
guna or vriddhi, does not do so, when it has an indica-
tory *£ ^ or ^.

Thus the past participle terminations t? and TOf?f are irdhadhA-
tuka affixes, which would, by the general rule VII. 3. 84, have caused guna, but
as their indicatory letter q? is f* $ the real terminations being, n and wm • they
do not cause guna. Therefore, when these terminations are added to a root, the
ik of the root is not gunated. Thus with the root fa "to collect" we have
fofr : or famr% f TI, "to hear," *pr:, ^m% ; 4t ; "to fear" *fftf: t >ftfrnt.

Similarly the terminations TO?^f«pr, U*j[ 3*H^,TOT, f^,&c., are
all fefrj terminations 'the indicatory letters of all being 55*/ the real affixes
being ^r, nf*\*( , ^TCJ &c Sec. &c. Before these, the root is not gunated.
Thus we have from *f "to bear" ^r, from fa^ "to pierce" fift^PT ♦

The terms fr^ &c., of this sfttra are in the locative or seventh case.
The force of this case termination here is that of QfjKf or cause, that is the
guna or vriddhi which would otherwise have been caused, does not take place,
if an affix is ftr^ , ftrjr, fiF£ . Thus all &rdbadh&tuka and s&rvadh&tuka affixes
cause gunation of Hit final ik of the inflective base (VII. 3. 84.) Thus the
S&rvadh&tuka affix ?j\ ( *r ) in the following : —

1+*J?+fay ,3 Ht+*J , + fa-Trf^ "he is." Similarly s&rvadh&tuka
and ftrdbadh&tuka affixes cause the gupa substitution of the short penulti-
mate it of the inflective base (VII. 3. 85.) Thus f^ + ^ + ft^ -*>iRr.

The general force of the locative case in an aphorism is to cause the
operation directed, on the expression immediately preceding it (I. !•
66.) If the force of the locative case in the present s&tra were also that as

8. Exception to, Guna and Vriddhi. [Bk. I. Ch.. L §. 5. * .

explained in I. I. 66, then the interpretation would be this : — An affix having an .
indicatory k f g or ft, would not cause the guna substitution of the vowel ik y
immediately preceding it The case of the penultimate short ik will not be .
covered by it Then though the rule may apply to f% +^RT«ft^: "collected "
it would not apply to ftj+W; there would be nothing to bar the guna subs-
titution of the penultimate f of fiff. This however is prevented by explaining
the locative as one of " occasion" and thus firf + W ■» Pfa": " broken. .

Similarly from the root fSr we have a word ft«0[ : ' victorious/ in which •
the affix *pg has an indicatory i£ Thus fir + *J*5 - fa*J : "victorious" •
(HI- 2. 139), so also **mf : "durable,"

Similarly the affix, *J3^ is a fir^ affix, the real suffix being *r,
the ^ being merely indicatory. Therefore when «n5 & added to a root, there .
is no gunation or vriddhing. Thus we have from firjf " to know', fap

Similarly by I. 2. 4, all sarvadhatuka affixes that are not marked
by an indicatory £, are treated as f|^. Thus the second person singular termi-
nation <r^isft^,and we have faj*r :" they two collect," similarly fanPq,
"they collect," *£ + ?Rr » ye : "cleansed," *p\fct "they cleanse."

The phrase "when indicated bythe term ik" is understood in this sutra. (
The prohibition, therefore, applies to ik\owels only, and not to all vowels )'
in general. Thus by III 1. 30, th^-root gr^" to desire" takes the affix
fJSj^; of this affix the letters^ and 3? are indicatory, the real affix is f.
The force of ^ is to cause vriddhi (VII. 2. 116). The indicatory ^ does not]
prevent such vriddhi, as the letter *? of 3T5J is not included in ik. Thus we!
have 3rrft+ ^~9ff%+snH3ST'ni%" he desires."

According to the wish of the author of the Mahabh&shya, the verb

f^ optionally takes vriddhi, before kit or nit affixes beginning with a vowel.

As MR*j*ifirl or MfiHlwiPd " they rub." Here in this sfttra there is prohibition

of the guna of the short penultimate ik also.

* — - t

The indicatory ^f of the verbal tense affixes *>., 5T^ , f*T^ , g^ , 5J^ t

is an exception to this rule of f§F?£ . This is an inference deduced from y

III. 4. 104 which declares " when the sense is that of benediction, then the

augment *U*J* of f5H£ is as if it were distinguished by an indicatory qf."

There woulcThave been no necessity of making qi$^ a %^, if f*n£ were a

Hfh, but the very fact of making HW\ a f%T7[ indicates by implication or is

mn<h that the final y of feo^ , 5T^ &c. does not make these tense affixes

f|^. Thus in *T^ or Imperfect tense we have 5tf^re^ " he collected" ( *tfaf j

^frf: Q ^Mf^zfWiaj% 5T3Jf^ Hl^i <H& T «nW: ll i Ibigitized by Vj^OQ I

Bk. I. Ch. I. § 6, 7. ] Sanyoga defined. 9

6. Theguna and vriddhi substitutions, which
would otherwise have presented themselves, do not
however come, in the place of the vowels of didhl * to
shine/ and vevi c to go,' and of the augment called jr. if-

The roots $>ft and iffi never take the guf&a or vpddhi substitutions
under ordinary circumstances. Thus the affix 0355 generally causes vriddhi subs-
titution when added to any root ; and so th^aflhfsg^ causes guna. But these
affixes when applied to the verbs didhi and N vest, never cause vriddhi or
guna substitution. As^hft + *5^ - SJTfhft + smrtVII. 1. 1) - «rrfta*r:

(VI. i- 77). So also *n*hft + ^ - *nfcft + «rc (VII. 1. 1) = ^rth^Fi^.

Similarly from %^fr we have Hf|$*>*H1 and Hji^q* :- In the Vedas, however,
these verbs take guna. Thus ^r^[ (Rig Ved. X. 98, 7), and «rfhT3 : (R*g
Ved. V. 40. 5). These two Verbs are confined generally to the Vedic
literature, and the present rule therefore, seems to be an unnecessary
prohibition ; as they take guna in spite of such prohibition.

Of the augment j% the real affix is f . In the general tenses, it is

added to certain terminations beginning with consonants of the *f^ class, *>.,

all consonants except 7 (VII. 2. 35). Thus the future termination, first

person, singular is 441P) ; this when added to the root ^ " to move" requires

/ an intermediate f , and thus we have the form ^ft*srrf% , ' I shall walk.' This f

' is never gunated or vriddhied, though according to the general rule,* coming

■ before an irdhadhituka termination, it ought to have been gunated.

The augment f^ $ could under no circumstances have taken vriddhi,
the prohibition therefore, is in the case of the guna substitute of f^ .

#WntU ? ^ftTC: II 3 II *J^tf*ll f^:, WInTCT:, *S*TO: II
jf%*. 'i f n+ »ii<Ti *h *ip K>m ff <r f a^gteiifrii *sr: *Nt*r ^gr nqfiggypt:

7. Consonants unseparated by a vowel are
called conjunct consonants.

This defines the word sanyoga. The sfitra consists of three words,
5ST:, the plural of the praty&h&ra ft*, denoting all the consonants, HH«^K:
means "without any separation or space" and trvftr: which is the
word defined, and means u conjunct consonants/ 9 So that the s&tra
means, conjunct consonants are those consonants between which there is
no heterogeneous separating vowel and which are pronounced jointly, such as
QT in the word *f^z: 9 " cock." The word u sanyoga" applies to the whole of
the conjunct consonants jointly and not to them separately. The plural
number shows that the conjunction may be of two or more consonants,

2

10 ANUNASIKA DEFINED. [BK. L CH. I. §. 8/

Why do we say " consonant "? If two vowels come together they will
not be called " sanyoga". Thus f frrrggpi " a sieve umbrella," the two vowels *f
and* have come together unseparated by any consonant; yet they are not
called sanyoga, otherwise the final * would have been rejected by VIII. 2.
23 " there is elision of the final of that pada which ends in a sanyoga letter ."
Why do we say " unseparated by a vowel"? If consonants separated by a vowel
were also called sanyoga, then in the sentence <ptf% HHH% , the syllable w^
would be called sanyoga, and by VIII. 2. 29, the initial *r would be
elided.

Trim, m^Hium *

8. That which is pronounced by the nose
along with the mouth is called Anundsika or nasal.

This defines the word anunasika. The sutra consists of four words
ysr " mouth" Hllti^ r " nose" *r^ " utterance or pronunciation" and^npnfiW:
" nasals." Though the words $& + mftw (II. 4. 2) will by the rule of com-
position form ya«ufa<h and not jf<SHlff>* l » yet the latter form might be
held to be an irregularity ; or the phrase jiSHH^H^H may be regarded
as a compound of jj(jHII*l<h + mN*H , the word avachana meaning " partial
utterance." That is a letter partially uttered by the nose and partially by the
mouth would be called anunasika. Therefore the nasals are those letters
which are pronounced from two organs or places, i.e., the mouth and the nose.
The pure nasal is anusvara, while anunasikas are different from this, in as
much as, that in pronouncing these, the breath passes through the nose and
the mouth.

The vowels are generally so nasalised. ' If, instead of emitting the
vowel sound freely through the mouth, we allow the velum pendulum to drop
and the air to vibrate through the cavities which connect the nose with the
pharynx, we hear the nasal vowels (anunasika).' Thus qj f *qr # In the Vedas,
the particle srn£ is anunAsika. As qpst *rf spr (Rig Veda. V. 48. 1); so also

<rft #r ifcft: srf **si (Rig Ved. VIII. 67. 1 1).

The consonants y, sf, nr, ^ and H are also anunasika. Thus y is \ y
pronounced by the throat along with the nose, *f by the palate and nose, UJ by |
the upper palate and nose ; sfby the teeth and nose, and *r by the lip and nose.

The semi-vowels 3f, ^ and 5T are also nasalised and are then called ]/
anunasika. The term anunasika is used in sfitra V. I. 126 &c.

Why have we used the word "by the mouth"? This definition will
not include anusvAra or the pure nasal, which is pronounced wholly through .

Bk. I. Ch. I. § 9 ] Savarna letters. II

the nose. Why have we used the word " by the nose "? In order to exclude
the consonants q?, % *, % ^ &c. &c, which are pronounced wholly and solely
through the mouth.

*r*i% ii

9. Those whose place of utterance and effort
are equal are called savarna or homogeneous letters.

This defines the word savarna or a homogeneous letter. The sutra
consists of four words : — <pir " equal or similar/' q j m " mouth or place of
pronunciation, " inrsr " effort " and toS^ " words of the same class."

The <ujit4 or places or portions of the mouth by contact with which
various sounds are fonpecLare chiefly the following: — I. ^nj* " throat" 2!
*fc ' palate/ 3. w^f ,(head^ 4. f??rr: 'teeth/ 5. *frgV ' lips/ 6, mflfW ' nose. 1 *P

Th^raa^ or quality or effort is of two sorts primarily, MJIKMi: "inter- *
nal" and 5*nQT^"extemal." The first is again sub-divided into five parts:—

1. *WL\ or complete contact of the organs. The twenty-five letters
from q? to T belong to this class. In pronouncing these there is a complete
contact of the root of the tongue with the various places, such as throat,
palate, dome of the palate, teeth and' lip.

2. f^fj^ or slight contact. The letters *, r, 5C, * belong to this
class of contacts. " In pronouncing these semi-vowels the two organs, the
active and passive, which are necessary for the production of all consonantal
noises, are not allowed to touch each other, but only to approach."

3. firf^T or complete opening. The vowels belong to this class.

4. fqr^ f%f«f or slight opening. The letters lj t *r, ^, f belong to
this class. Some however place the vowels and jj &c, into one group and call
them all vivyita.

5* Hf^or contracted. In actual use, the organ in the enunciation
ql the short *T is contracted but it is considered to be open only, as in the case
of the other vowels, when the vowel *r is in the state of taking part in. some
operation of grammar.

The Abhyantara prayatna is the mode of articulation preparatory* to
the utterance of the sound, the b&hya-prayatna is the mode of articulation
at the close of the utterance of the sound.

The division of letters according to &bhyantra prayatna has been
already given. By that we get, 1. the spars a or mute letters, 2. th
a ntasth a or intermediate between sparsas and ushmans, or semi-vowel

the

c\Lc>
1 or

12 SA VARNA LETTERS. [BK. I. Cll. I. § 9.

liquid letters, 3. the syaras or vowels or vivrita letters, 4. the ushmaiw or
sibilants or flatus letters. —

The division of letters according to bihya prayatna gives us first
surds or aghoslia letters, 2. sonants or ghosha letters. The aghosha are also
called svdsa letters, the ghoshas are called nida letters.

The second division of letters according to bAhya prayatna is into ;

1. Aspirated (mahiprina). 2. Unaspirated (alpaprAna).

The vowel *[ has eighteen forms. The acute *T (ucIAtta), grave *T
(anuditta) and circumflexes *f (svarita). Each one of these three may
be nasalised (anun&sika), or not (niranun&sika).

Then lastly everyone of these six may be short, long, or prolated.
Thus:— *T, B^ 4, 4, 4, #, MT, *TT, *d, *ft, *ft, 4K HT*, *f\, 4,*, 4*,

Similarly the letters f, ? and gj have also eighteen forms. The
letter 51 has no long form; it has therefore 12 modifications. The diphthongs
have no short forms, they have therefore, only 12 forms. The antastha or
semi-vowels with the exception of r have two forms each vis., nasalised and
un-nasalised. The semi-vowels have no homogeneous letters corresponding
to them. All letters of a varga or 'class' are homogeneous to each other.

Thus then the homogeneous or savarna letters must satisfy two condi-
tions before they could be called savarna. First, their place of pronunciation
or Mj|*4 must be the same. Secondly their JPT5T or quality must be
equal. If one condition be present and the other be absent, there can be no
savarna-hood. Thus ^ and jj are both palatals, their mm or place of pro-
nunciation being <jr<S ' palate,' but still they are not savarna, because their
TOST is different, the q^ra* of ^ being sprishfa and that of % being vivrita.

Similarly q? and *? though their 733* is the same *>., sprishta, yet their
Mj|*4 being different, one being guttural and the other palatal, are not savarna.
There is exception to this rule in the case of ^ and 3f, which though
having different *TTCT are still called savarna by virtue of the v&rttika of
K&tyiyana ; " the homogeneousness of 3; and of, one with another, should
be stated." Thusffcr+onrrc'-*- ftpCTT: ; here 31 and 5* have coalesced into
long ^r by rules of sandhi, on the assumption that they are savarna letters.

The word savarna occurs in sfitras like*?*?: 97$ tf$: (VI. 1. IOI).
Why do we say "the place of pronunciation"? So that there may
not be homogeneous relationship between mm<i and <| whose prayatna is the
same but whose ftsya is different. What is the harm if we make them savarna?
Then in words like snff and ?rf$ the <r would be elided, if it be held homo-
geneous with % by VIII. 4. 65, "there is optional elision of frt, preceded
by a consonant, when a homogeneous ^ follows. " Digitized by GoCK

Bk. I.Ch. L § io, ii ] Pragrihya defined. 13

Why do we say " effort" ? That there may be no homogeneity between
f and the palatals, and ^ and jj 9 whose organ of pronunciation is the same,
but whose prayatna is different. What harm if it be so? Then in HJ*^«4t-
llftr, the xr would be elided before ** by the rule already referred to, VIII. 4. 65.

TTOSn&T « 90 H Xl^Tfe I1 1 , 3T* ^&T, ( Wfe) »

?i%: f i gWroji^rr*ft ^rspsft ^m m*5*kft t *nr*r. •

10. There is however no homogeneity between
vowels and consonants, though their place and effort be
equal.

This sfttra lays down an exception to the former sutra. There can
be no homogeneity or savarna relationship between vowels and consonants,
though their 3TTCT and TOoT may be the same. Thus f and tfhave the same
SJTW namely m*£ 'palate/ and the samesraa': namely vivjita, but ?till they
are not to be called savarna. Thus in y^cT, j fimfagi &c, the letters *r and f
in the first example, and f and ^r in the second not being homogeneous
letters ; there is no combination or sandhi, which would otherwise have
taken place.

Similarly if XT and f be held homogeneous, then f^THCT + *PJ (IV. 3.
53)«ftTO£+*T^(VI. 4. 148 ; here jj would be elided by VI. 4. 148, but it is not so
and we have ^THJ • " belonging to vipisA or born therein." So also in ijFTJf
"ox-skin/ 9 the f would have been omitted, if held homogenebus with %f.

11. A dual case affix ending in I or *r or s is
called Pragrihya, or excepted vowels which do not ad-
mit of sandhi or conjunction.

As a general rule, Sanskrit allows of no hiatus in a sentence. If a
word ends in a vowel, and the next word begins with a vowel, the two vowels
coalesce, according to certain rules. This is called sandhi. But pragrihyas
are exceptions to this sandhi, "they are certain terminations, the final vowels
of which are not liable to any sandhi rules." (VI. 1. 125). This sutra gives
three of these terminations, viz., the duals of nouns or verbs ending in f
(f^)i^ (3T^), or I? (^). Thus: — qrfr ffc " these two poets/' so also
*T*t *f?T"two winds;" HT^r *T% " these two garlands." V%&fnr"they two
cook" 7%^fi% "you two cook." Here according to the general rule of
sandhi, the final f of 3rft and the initial f of f4t ought to have coalesced into
an I, but it is not so, because ^ft is the nominative dual of WkQhy (

14 Pragrihya. [Bk. I. Ch. I. $ 12, 13.

Why do we say ending in {, H and e ? Because when the dual ends in
any other vowel this rule will not apply. As ^ft + JSpr-fWPTi "these two
trees" (VI. 1. 78). Of course it follows from the definition, that f, *r, or n of
the dual number only are pragrihya; if these vowels terminate any other num-
ber they will not be pragrihya but will follow the general rule of sandhi, as
ipnft ( 1. s. ) + ispf - ^H\^A t " the girl is here" (VI. 1. 77).

Vdrh The prohibition of *pffa &c, should be stated in treating of the
pragrihya nature of f, a? &c. Thus *pft (two gems) + f^ - *nffcr"like two
gems." So also ^FT^hc " like a couple," spanffa " like man and wife," rr^?T
" like heaven and earth/' Some say that this vlrtika is unneccessary. Because
the final member of the above words is notf^, but the particle * only ; which
has the same meaning as iva.

^«H « « Htr^r 11 9?^:, *t<j> ( ^^ mwi )»

12. (The same letters) after the 1 of the pro-
noun *re^ are Pragrihya. .

In the pronoun *ft^ , the dual termination in f, 3T or ij will be pra-
grihya by force of the last sutra, but the present sutra makes this addition,
that in the case of 3flRJ, all terminations ending in f, ^ or ^ and preceded by
*T, of whatever number they may be, will be pragrihya. As a matter of fact,
there are only two such terminations satisfying the above conditions name*
ly ^Pfl" and sp£. The rule may therefore be put in the following words of
Dr. Max Mullen — " The terminations of^pftand^pj^nom. plur. masc. andnom.
dual of the pronoun ^^r are pragrihya." As^pftspsr: "Those horses,"
and *P£WWfr. There is no example of the word adas ending in ^r.

Why do we say " of the word adas" ? The letters f , ^, 5 following
after the ^ of any other wdrd will not be pragrihya. As ^pft + *jpr ■» IF***
"thissami tree."

Why do we say "after the letter it" ? Because in the example m%j%
-Hgj^ = ^TJ^T f ^r, "those here," the ^, preceded not by it but by the q? of ^nr^
(V. 3. 71) is not pragrihya, which it would have been by the influence of
sdtra 1 1, which includes ^ as well as f and 3T, and from which the word dual
is not supplied here, else this rule would be useless.

i* 11 1^ n v^rfk n $ , ( jhi^i* ) 11

13. The affix $r, (the Yedic substitute of the
case-affixes), is a pragrihya.

In the Vedic Sanskrit, the declesion of nouns sometimes differs some-
what from the ordinary Sanskrit. One of the peculiar Vedic terminations, which

Bk. I. Ch. I. § 14. ] Pragrihya. 15

replace the ordinary declensional terminations is called ft (VII. 1. 39.)
Thus tlt'U'ftlJt^'jdt .

Of this substitute ft, the initial ^r is indicatory, the real suffix being ^.
This final 5 is a pragrihya.

In the sacred literature there is only one example of this pragrihya, .
as given above ; while in imitation of this, we have in secular writing phrases
as, stffa, H& ff<t» % Tfa & c -

wn? s i ( wra* ) n

14. A particle 1. 4. 56 consisting of a single
vowel, with the exception of the particle *n?s is a Pra- '
grihya.

This sAtra consists of three words :— f^TRT: " particle," ^FH[ " con-
sisting of one vowel ," ^HT< " with the exception of STT^-"

All such particles technically called nip&ta (I. 4. 56) which con-
sist of a simple vowel without any consonant, are pragrihya and therefore are
not liable to the rules of sandhi. As * %$. "Oh, Indra" gr sf%5" " O arise."

It follows from the above definition that those ni pitas which consist
of two or more letters, vowel and consonant, are not pragrihya, e. g. jj in JJF?r#

^rr^ which is not a pragrihya, has four significations: — As a
diminutive particle when joined to adjectives and nouns, as *rr+ 4uyn
« qfrmj^ "little hot, tepid" (2) as a prefix to verbs, it expresses the senses of
"near" "nearto," and with verbs of "motion," "taking" &c, it shows the reverse
of the action, as TH'togo/srpTT ' to come j'^T + tf?*^, (3) as showing the limit
inceptive ( s *rffifafa) "from," " ever since," as qjM**H ; " ever since his birth, *?
(4) as showing the limit exclusive (*nrr*T), "till" "until" as qrr- HijunHiq »
qj|Ul*HI<t > "until the reading begins." When the particle ^TT has not any one of
the above four senses and is used in expressing regret &c, it is a pragrihya, as
*TT S# f^vlI<A^ "Ah! such there once was," *rr %j >to%. "Oh! do you
think so."

The word JZ^KT* in this aphorism has a peculiar meaning. It means
" consisting of one vowel and nothing else." In other words, the vowels them-.
selves when used as ftmtf, e. g. *jt, 7, f &c. But the term 33THT has a wider
signification than that given to it in this sutra. It means "a word or a portion
of a word consisting of one vowel which can be pronounced at once" — in
other words ^»r^ means a "syllable." In that sense the words it, fir, f% &c^
will also be H«M^, but not so here. Therefore the rule of sandhi will apply

1 6 Pragrihya. [Bk. I. Ch. I. § 15 to 18.

-■■■■-
to them, thus H\**\A JTOftOT. See sfttra VI. 1. 1. where the word 3*fln[ means
a monosyllable.

Why do we say " a nip£ta?" Because if the single vowel is not a par-
ticle, but something else, an affix &c, then it will not be a pragrihya, as the final
*T in n^TT " he did," is the affix of third person perfect tense, and is liable to
sandhi ; as TOftff.

afani v% h nqxfk 11 sftc^, ( uw* ) n

frtfc 11 <j?i<m! % f*nrro *r ^r^pnr ^hnfir 11

15. The final ifr of a particle is a pragrihya.

This sfttra consists of one word *ffr^ the final ^ is a surplusage. The
words "pragrihya" and " nipita" are to be supplied from the preceding sfttras
to complete the sense. The meaning is that those particles which end in 9ft
though consisting of several letters, are pragrihya. As *(H trtft " Halloo, go
away."

^^ajra**ram*Ri3ti^ii rnrfkn ***$, mwwHt

ffrf: llH*ff^[ RAMI *T MjUhK: *T CT3i?*fCTrar4TO ItPT T"IW^at H*fa ,

16. The final sfr of the vocative singular (II. 3.
49) before the word fft according to S&kalya, in secular
or non-vedic literature, is pragrihya.

This sfttra gives the opinion of the /fishi 5%kalya, so that it is an
optional rule. Thus both forms qrartlft (formed according to the general
rule of sandhi) and qpft^Rj (where sjt is pragrihya) are correct.

In arsha literature the sandhi is imperative, not optional, as ^T iff

The mention of the name of any sage, such as that of 5&kalya in this sutra, serves generally
one of the two purposes :— (l) either it makes the rule an optional one, vibhasbartham, or (2) the
mention is merely for the sake of showing respect to the particular sage, the rale would have stood
as well without such mention ; pujartham, as the word Kasyapa in sutra I. 2. 25.

wi: n 93 11 ^^rf* 11 **: , ( g^^ m^n^cnnHi^ ) n

jRf: II **T. ST^RT ^JT *Flfir f^ ^ll^^^l^l^^ H$H I

17. The particle s* before itt\ according to
Sikaly a is a Pragrihya.

This sfttra consists of the single word ^ meaning ? . According
to the same tfishi S&kalya, srsj is a pragrihya, before ffa t in modern Sanskrit.
This is also an optional rule. Thus * tf% or ftft. Di ^

Bk. I. Ch. I. § i8, 19.] Pragrihya. . 17

^ IM6 H vtfk » ^ , ( WWW, «*: , TOF? ) R

18. The particle £ replaces **T in non-vedic
literature, and it is a pragrihya in the opinion of SAkalya.

Before the word jfa the word *sj is replaced sometimes by 3? the
long nasal 4. Thus ^ + f|% - $■ f|% orftft'.

In some authorities the sfitras 17 and 18 form but one sutra, and then it
is divided by the method of yoga-vibhdga, the division of a single sutra into
two separate sfitras having two distinct rules.

ff^r : n f^ry^rq «*f sti ****& *Ah\* *Pi*rk$ *nrf!t 11

19. The final lander of words giving the
sense of the locative case are pragrihya.

This sfttra finds its scope in the Vedic literature generally. There is a
sutra (VII. 1. 39) which declares that "in the Vedic Sanskrit, the case-affixes
sup are often replaced either by 5 (1. s.), or are altogether elided, or they are
changed into a letter homogeneous with the prior letter, or *jt, or *n$ or & or
Wi or ^r, or ssn* or tr% or syr^ takes their place." Let us take the example
where a case-affix is changed into a letter homogeneous with the prior.
Thus the case-affix of the seventh case is Rf (f). In forming the 7. s. of ipj;
"body," we have *FJ+f. The % will be changed into a homogeneous letter with
5T and we have *FJ + 3T « *P£ " in the body," as, *rurc*rf »TPT5ft^ (for ifPRPli Wt).
This is the Vedic form, the secular form being fF^f. Similarly ?frft+f«*ftft
"on the Gaurf," as *?tffr$r# *rfa£w (Rig Veda IX. 12. 3). It is in such
cases, therefore, that the long i and I are used in the sense of locative, and
are pragrihya.

It is only long i and rf which are pragrihya, and not any other vowel,
though used in the sense of locative. Thus by the same aphorism (VII. 1.
39) $l (*) ma y be employed to denote the seventh case-affix. Thus *jffer *+ fr
ss *rf*T+5T=S^ T; ^ , 'in the fire/ as, fipn *jifnrtrVMHI >Wl3 H XWR *J<Ki1*fl 4414.1*1
(Rig Veda V. 37. 5). Here mjhi though in the locative case, its final fs not
pragrihya. Therefore SFSTT+ftlfr - *P#fit \

The f and 3> must be of the 7th case. Therefore *rf?r + *T (3rd s.) «■
1^ + T (letter homogeneous with the prior) » *nfr 'with wisdom 9 . Here t is
in the sense of the instrumental case and is not pragrihya. Thus tpft + fft «■

***** Digitized by G00gle

IS Ghu defined. [Bk. I. Ch. I. § 20, 21.

The word trf is used in the aphorism, in* order to exclude the appli-
cation of this rule to cases like the following. Thus in forming the. compound
of IT^T^ " in the well" and *p* " horse," the case-affix is elided, and we get
CTft + %TS- Here, no doubt, the word 3nft is in the locative case, though the
case-affix is elided ; but we cannot say that the long fbf v&pt has the sense,
df the locative. Here, therefore, we shall have sandhills, *Fro : ** the horse
near the welL" In other words, the present aphorism is not to apply to the (
final for 3? of words standing as the prior member of a compound term \
(pflrva-pada); for words like qrrft &c, in the above example, can never, by them-
selves and alone, denote the locative case ; while words like *ft#, formed in the
way described already, give the sense of the locative, even alone, though they
be placed anywhere in -a sentence.

^rera^iq ii *0 n tr^rfn u ^ , \n , 5 , si^t* ji

fRT: H W*H\ *J*1lO >TPFfr TOt ^T £t, W&W * ARl«ll J^RRT! *nrf^f II

20. The verbs having the form of di 'to give*
and dM c to place, 1 are called ghu.

This defines the word ghu. Four verbs have the form of JT, or assume
it by the application of VI. i. 45, and two, of qr; they are called 5. The ghu
verbs have certain peculiarities of conjugation to be described hereafter.

The tj verbs are the following:— 5^ " to give.," as sjftjwPl ; fTJ
" to give/' as, nfywn ; tf "to cut," as, ijfcroft; *C " to P* lt y " **» 'ffisrc*^ 5>IP|
" to place" as. nftj^rRr, ^t to " fccd " ** ifiinrcfo

In the above examples, because of the verbs being 5, the H of sri% is
changed into ur by VIII. 4. 17.

Similarly ^ "to cut" and |r " to clean" not being called 5, rules MIL
4. 46 and 47 do not apply to them. Thus *r + ^ ~ *rir 'what is cut,* as, frt
Hf%:. So also*rort " washed " as, «n?rt J^r " a bright face." But with tf " to
give," &c. *r + ^f> =*^ +n (VII. 4. 46) «frT: "given ;" so also we have sy+fr
+ W -JPfPJ (VII. 4. 47). The word 5 occurs in sfttra VI. 4. 66, &c.
The word dip includes also daip by the paribh&shi given under III. 4.
19,^ changed into mt by VI. 1. 45.

21. An operation should be performed on a
single letter, aaupon an initial or upon a final.

This is an atide£a\sfttra and consists of four words, qrrft " beginning,"
*|«T "end," ^ "like " Jptfr^L "in one;" the literal translation being ''beginning
is end-like in one." The affix *$ has the force of the locative, and the

Bk. I. Ch. I. § 22, 23. ] Gha defined. 19

meaning is "on the beginning," and " on the end," %** let an operation be
performed on a single letter, as upon an initial or upon a final. Thus by
III. 1. 3. an affix has ud&tta or acute accent on its initial vowel. Thus the
affix tavya has accent on the first if, and we have 3^W^ . But when an
affix consists of a single vowel, it is itself the beginning and the end, and,
as such, will have the accent. As in *fhpi*^ aupagavam the last & baa the
accent, which 4 represents the affix *r<y .

Similarly, by VII. 3. 102, the final of a base ending, in short ^
is lengthened before a case-affix beginning with a consonant of*3| class.
Thus rre+^^TPT' lH l ^l^ . This rule will apply when the word consists
of one letter only, and that letter is a short *T- Thus ir+^W^" 'Jl**! 1 !
"by those two." Here it must not be objected that the solitary if is
initial, and cannot be final.

Why do we say "in one"? If a word consists of more than one
letter, then the beginning is not like the end. Thus in the word Wft the
initial short 3T is not like the final long *JT; otherwise tnft would get the name
of ^5^(1. 1. 73), and the secondary derivative from 4HI4W4*t would be formed
by the affix 5 (IV. 2. 1 14). But this is not so, the affix «T^is employed in-
stead, and the form is *J Him 3*1 : .

22. The affixes tarap and tamap are called *.

This sfitra defines tj affixes. They are two, CRT and ipr, the ^ in
them being f^. The special use of these affixes will be described later on.
These affixes are useful in forming the comparative and the superlative
degrees respectively of adjectives, adverbs, &c. The base undergoes certain
changes before these terminations, which will be treated of in its proper place.

Thus the words apnf^TTT and gjHlftd'U are formed by adding the
affixes ?TT and ?PT to the word ymfl ' a girl/ whose long f is shortened before
these affixes by sfitra 43 of the third chapter of the 6th Book.

1 f^n 11

ftu&1 11

23. The words balm 'many, 1 gana l class, 1 and
the words ending in the affix vatu (V. 2. 39), and cfati
(V. 2. 41) are called numerals (sankhyA). D »

20 1 Sankhya. [Bk. I. Ch. I. $ 23.

This sfttra defines the word sankhyA* or numeral. ^
These are certain words, besides the ordinary numerals, having some %
peculiarity of declension. The words like ?g ' many/ and *n!T 'class 1 , and those
that have the affixes ^ and *rf?r after them, belong to this class. The letters
* and ¥ of ^5 and *(% arefrtfy or indicatory, the actual affixes being ^ and

The word sankhyA means " numerals f and this class contains,
among others, the Sanskrit numerals. The words formed by adding ^ are
generally compound pronouns, e.g.m*$ 4t so much/' *rr*<| " as much," ^TPR[
"so much," similarly words formed by the affix ^rf%, e. £. ^ft " how many," *nfr
"so many," tr* ## as many."

Thus *$+$«**|^ (V. 4. i7)-*3frai? "many times" the affix fTC**^
being added only to sankhyA words in the sense of the repetition . of the
action. So also SJ + ^T(V. 3. 42)«*jqrr " in many ways." Here also the affix
HT is added by virtue of *rj being a sankhyA. Similarly ^f^: (V. 1. 22)
" purchased for a big sum," ^Jtf: (V. 4. 43) " manifold." So also *IUJ$g3 :, iflZPtfi

*rarar:, vrjw* ; so also ciT^g^ir, srn^r , m^gr ; <*ft*w , srfiter, «kRw: ^

When the words qj and ifT& mean "abundance" and "multi- •
tude" they are not sankhya. It is only when they are expressive of
numbers, that they get the designation of sankhyA. The necessity of defining
sankhyA arose in order to exclude such words as, 5jf? " numerous," &c,
which though they express numbers, are not to be treated as " numerals,"
technically so called.

Varti — The word *f^ " half" when standing as a first term in a com- "^
pound number which takes an affix giving the sense of an ordinal, gets the
designation of sankhyA, for the purposes of the application of the rules of
samAsa and of the affix qr^ . The ordinal affixes or p&rana pratyayas are
like ^ (V. 2. 48), *{% (V. 2. 49), U^(V. 2. 51), &c. Thus s^fiW- "eleventh,"
q^if: " fifth," ^3$: " fourth," &c. Similarly we have qjj^**^ '• "purchased for
4} baskets" (a dvigu). Here ^rt+'^fT (an ordinal) =*rt*re5, a bahuvrihi
compound meaning 4J, vis. t that in which half is the fifth term. VJ^'H'l :
^:fflrtn» WJpHH^ I : the taddhita aflfexes snjjandjjRl required by V. 1.19
and 26 being elided by V. 1. 28.

Here, there is compounding (dvigu) by taking qr^q^*PT as a sankhyA,
and then applying sfttra II. 1. 51. The samAsa here is in the sense
of taddhita, and this compounding is only possible by rules (II. x. 50 and
51), when -*j4t«Hq*jl gets the appellation of sankhyA. Having got this name,
it forms dvigu compound in the sense of the taddhita affixes 57^ and snqr
(V. 1.26). The word^ takes both these affixes in the sense of "being

u/

/

Bk. I. Ch. I. §-24, 25.]- Shat. 21.

bought with." Thu$ ^rffap^nd ^Pj both mean "bought with a measure
called $urpa." In the above --compound of ^rtwff^ : the sense is
that of the above-named taddhita; but the affixes have been elided by
V. 1. 28 after dvigu compounds. So also we have Hj^tH<fr : by applying
the affix ^ • For these two purposes, namely samisa and the application
of Qv\, compounds of ardha, like ardha-panchama are treated as sankhyi.

v$&m ^11 m 11 tnjrft 11 *-*r-3ren f tc, (*w) 11

24. The SankhyAs having * or h as their final
are called sha*.

This sfitra defines the term ^ which is a subdivision of the larger
group sankhyi. Those sankhy&s which end in ?r or ^ are called ^. The
word sankhy& is understood in this sutra, because the word dSJF^T is in the
feminine gender, showing that it qualifies sankhyd which is also feminine.

The numerals that end in qr or 5f are six, namely *fa% * five/ ^ 'six,'
TOTJ, • seven/ %TZ% ' eight/ TfJ ' nine/ ^* € ten/

!It is one of the peculiarities of ^ words that they lose their nomina-
tive and accusative plural terminations (VII. I. 22). Thus s^ forfcr,

The word *jp?j in the sfitra shows that the letters *r and H must bo

aupade&ika i.e., must belong to the words as originally enunciated and not

.added afterwards as affix &c. Thus in *Wlfl 'hundreds/ ^ftfrftj 'thousands/

) ^JtlHIH ' of eights ' the H is not radical, but an augment, and therefore they are

,' not ^[ and the nominative and accusative plural terminations are not elided;

^ as sranj + f =» ^Rnfir &c

sfir ^ 11 r\i\ ^t(5i 11 5f?r, n f ( wt, ^) 11

ffrf: 11 ^**wi *tt 4**4i w iv** 1 *&ft 11

25. And the sankhyAs ending with the affix
dati are called sha«.

This sfitra enlarges the scope of the definition of q^ by including
in it, words formed by ^nfr. By siitra 22, words in ^ffir are already san-
khy4 ; by this they are also s^ . Therefore we have gjfir q?fcf, ' how many are
reading/ srrf^T <T*^ ' see how many/ Here «Rfir has lost its plural termina-
tion by VII. 1. 22.

26. The affixes kta and ktavatu are called
Nishdia, Digitized by Google

22 Nishtha. [Bk. I. Of. I. $ 26, 27.

This defines the nishtha affixes, that is to say, the affixes of
the past participle. The actual affixes are ft, ipnt, the 3T and * being indicatory. /
They beiag ft^fatre liable to all the rules applicable to fr<^ affixes, such as rule
I. 1. 5. Thu*jwfe have, gpr:, ^TORT, *«:, g^TJ.

The force of the indicatory * in ktavatu is, by IV. i» 6, to indicate that
in forming the feminine of nouns ending in an affix having an indicatory
vowel of uk pratyahira (7, 3? f Sf), the affix #p^ (f) must be added. Thus
fRTC^ (masc.)i fTCPPft (fern.). The word nisfijha occurs in sutra VIL 2. 14.

WMHUpIl II

27. The words sarva, 'all,' and the rest are
called sarvan&ma or pronouns.

This defines sarvan&ma or pronominals. To know what are the
other words comprised in the phrase ' and the rest* we must refer to Pimm's
Ganapitha where a list of all groups referred to in the sutras, is given. Thus
we know from it, that the following are sarvan&ma : —

*k ' all/ faj? ' all, 1 ^r f two/ sro ' both/ words formed by the affix
qRIT such as qr?nr 'which of two/ words formed by the affix z^H as *nPT 'which
of many/ *p* 'other/ *p*m 'either/ f<nc 'other/ «f£' other,' *r 'other/
Jfr.'half/ wt 'all/ finr 'whole/ nr* 'he, she, it/ ^ 'he, she, it/ ^
4 who/ ipf^'this/ m f it/ %K% 'that/ *rgf 'one/ ft 'two/ 3*** 'you/ *rcr^
'1/^5 'y° u / flf* 'what/ So also ^ ' east,' 'prior/ qr ' subsequent/ ^TT
' west or posterior/ {ftiUJ ' south or right/ *tTC 'north or inferior, subsequent,'
*n^ ' other or inferior/ ^rr 'west or inferior/ are sarvan&mas when they im-
ply a relation in time or place, as shown above, and not when they are
names. So also ^r when it does not mean a ' kinsman or a treasure/ but
means 'own.' So also *F3T when it means 'outej/ or 'an_under^ or lower
garment' isa sarvan&ma.

There are certain peculiarities in the declension of sarvanima words.
Thus ^ though ending in 3T, is not declined like' q(? which also ends in *r,
the plural of ^t being ^, that of 5^ being srt:. Similarly the dative singular
H*PT# *k$t; the abl. sing, srnfc, u$*HI£; the loc. sing. *ft, h\Rh±; the Gen. pi.

The compounds (with certain exceptions) ending in these words are
also sarvan&mas. The word 3*? ' both ' is always used in the dual number" as
^h*, Tyrant* ^ e object of its being inserted in the list of pronominals,
(whilst its declension does not differ from that of qrc), is its taking the augment
*n*^ (V. 3. 71) which it could not have taken, if it were not ^Tbronominal

fV

Bk. I. ch. I. § 28] * S ARYAN AMA. 23

f as &f^(. Another object gained by its being so called, is the application of |

^ II. 3. 27, by which a pronominal may take the case-affix of the 3rd case or I

the 6th case in denoting cause.

There are two Vf in the above list, namely ^gp and ^having the 1

same meaning, and both ending in *q\ But the final if of one is ud&tta, j

and of the other anud&tta. Some give/one **sr ^mly and not the other; 7Vx 6*^
and say that both have anud&tta accent The word H*r Is a sarvanima, when *****
it is synonymous with the word g% * all / and not when it means ' equality/
When it has the latter sense of 4 equality/ it is declined like *ir /. e. like '

ordinary words ending in *r f as PAnini himself indicates in I. 3. io, where he
uses the regular genitive plural BHFTT5T, meaning ' among or of equals/ As
a pronominal the genitive plural would have been tflrei^.

The above words sarva &c. are notto he_treated_ as pronouns when

they are used as appellativ es. For example, if H% be the name of a person

then it will be declined like ordinary nouns, as &fni \(% ' give to Sarva (a

person'). Here the dative singular is regularly formed. Similarly when these

words are so compounded with others^as to lose their original independent

character, namely when they arc(3T€r^ry they are not treated as sarvan&ma,

but follow the regular declension, as mfitmflq $ft ' give to (him who is) above .

/ all/ Here in the compound qjft*i% ' Above-all ' or * Supreme Being,' the word

) ^ loses its independent character and defines and determines the sense of

) another word. Fuller explanation of the word rapasarj ana) will be given later

r on in I. 2.43. This term^TPT is an *F3&R(1 that is a term the appli-

) cation of which accords with its meaning. A,w, 1 '"'' '" '

— , «*

/ t?WWL; (\^^l^ SgsftfT II ^ V^fh » fiim^T,
ifxf: II f%3jt <IHlti) ft<W*U*ft fTyrf^" WTO^ TOft^" ftHTOT ^rafaflft ^5hiH

28. The above words are optionally sarva-
n&ma when they occur in a hahuvrihi compound signi-
fying direction (II. % 26.)

This is an exception to sfltra 29 which follows. As a general rule
(see sfitra 29), in bahuvrihi compounds, these words sarv^ &c. are declined
like ordinary nouns. The present sfitra declares an option to this, in special
cases relating to compounds signifying direction. Thus the bahuvrihi compound
of direction grTrj^T 'north-east' may form Its dative case as a pronominal,
eg. ^rrjjKtir or it may form its dative like ordinary words ending in qjr, i. e.

24 Sarvanama. [Bk.I.Ch.1. § ig.

Why do we say "when signifying a direction" ? The next sfitra will
enjoin that in bahuvrihi compounds! these words are not treated as pronoim-
nals. Therefore, if in the present sutra, we had not used the word ftgsr , it /
would have been impossible to know where there was option and where
prohibition. By employing the word " direction," it is thus determined, that
in a samasa, relating to direction (II. 2. 26), there is allowed option, while
therelsprohibition in all other kinds of bahuvrihi compounds.

Why do we say " in compound"? So that the option may be in that
bahuvrihi which is purely sam&sa, and injio^ other bahuvrihi. 'There is no/
option allowed in that bahuvrihi where there is a semblance of bahuvrihi \ 1
but not exact bahuvrihi. '

The quasi-bahuvrihi, or analogical bahuvrihi compounds, or to use the
Sanskrit phrase, bahuvrihi-vadbh&va, is taught in sfitra VIII. 1. 9. where words
in repetition are treated like bahuvrihi. Thus 5* + jgk — R^**^ 'one and
one, each one/ Bahuvrihi compounds are possessive compounds, but in the I y
above example, no idea of possession enters. However the compound here \
is treated like a bahuvrihi, for the purposes of the elision of the case-affix of
the first member, and the whole compound gets the designation of pr&tipadika.
In compounds like this, therefore, which are bahuvrfhi-vat, but not exactly
bahuvrihi, the present rule and the succeeding sutra do not apply. Thus
^to"& %ft ' g>ve to each one/ {ftmJtftlUJtil fft II

Why do we say " in the bahuvrihi "? So that there may be no option
in the dvandva compound. Thus {RiUJIiU^fruJIJl ' of the south, north and
east 9 The prohibition in the case of dvandva by rule 31 is absolute and
invariable.

fnT H TOmfT SPTFa wfn 4W«U*Hil*UM H H^TRT II

29. The words sarva &c, are not sarva-
n&ma when occurring in a bahuvrihi compound.

As a general rule (see sfitra 72), when any definition, rule, operation
&c. is made applicable to a particular word, the same would also apply to any
other word which ends in that word. Thus the plural of man being ' men'
in English, the plural of blackman will be blackmen. So the definition of sarva-
n&ma given to sarva &c., separately, will also apply to the words ending
in sarva &c.

This sfitra introduces thus an e xceptio n to sfitra 27. In . bahuvrihi
or possessive compounds, the nature of which compounds will be described

Digitized by LiOOQ 16

Bk. I.Ch. I. § 30.] Sarvanama. 15

later on, the above mentioned words are not pronominals, and must follow
the ordinary rule of declension. Thus the compound fstaft^q meaning
•beloved of all' is a bahuvrthi compound of ft* ' beloved/ and flf^r'all/
(which is a sarvanama). The declension of this compound will follow the
general rule. Thus, in the dative singular, we have ftsrf^^ra, though the
dative singular of firT* by itself is firj*&, like other pronominals. So also
A4W4I4 " to him who is beloved of both." So also t^TOra, »3»3H &c*

Similarly in bahuvrthi compounds, these words not being treated as
pronominals, they do not take also the special affix S^p* f but the general

affix q^. As, qretfteqfr, H^Ptegft (V. 3. 70).

Though the word bahuvrthi was present in this sulra by anuvptti

from the last aphorism, the object of using the word bahuvrthi again in this

r^sutra, is this : — that the prohibition may apply to those expressions also

I which are no longer bahuvrthi, but whose original components were once

I bahuvrthi. Thus the words 4fe4l~<HI and q^HI^KI are bahuvrthi compounds;

but when these two words are further compounded into a dvandva compound,

the present rule still applies. As, q<»ll«*t<H*HMHl : II v

j ifa 11 g<flanma H^h% e3«u*niflifc h *nrf3*r 11

30. In Instrumental Determinative Com-
pounds the words sarva &c. are not sarvanAma.

This is another exception to the definition of sarvanama. There is
a class of compounds in Sanskrit called tatpurusha compounds, in which the
last word governs the preceding one. The word governed raajrjbein^any^one
of the six cases : when it is in the third or instrumental case, the com-
pound is called tritiyS, tatpurusha. When, therefore, there is such a samftsa,
the words mentioned in sutra 26 are not to be declined like pronominals.
Thus the word itch^T is a compound of m& and^~*{|iw^i meaning 'prior by
a month/ where the word misa is in the instrumental case. This compound
will be declined like ordinary words, e.g. its dative will be tliJujjjq, though the
dative of *jjjt, when standing by itself, is <$&; so also uvtH^H l»

The tptiya-samasa of the present sutra has reference to the special
trittya-samasa ordained by the particular sutra II. 1. 31. namely, "that which
ends with the 3rd case affix, is compounded with the words pftrva, sadp£a,
sama, &c." and has not reference to the trittya-samasa in general, such as one
ordained by II. 1. 32, " that which ends with the 3rd case-affix when it denotes
the agent or the instruments compounded diversely with what ends in a

4

26 Sarvanama. [ Bk. I. ch. I. § 31, 32.

kfit affix." Therefore, in phrases (TOOT %n 'done by thee,' TOOT ftf 4 done by
me/ these words are treated as pronominals, though in the instrumental case
followed by a word ending in a krit-affix. In other words, reading this sfitra
with II. 1. 31, we find that purva is the only sarvanftma which is so com-
pounded,, and. to which only the present sutra would seem to apply.

The word wtps may appear to have been mentioned redundantly, as
it has already been mentioned in the previous sutra, the primary object of the
sutra authors being to shorten the rules by the keeping out of every un-
necessary word. But here, the word *i*u« has been repeated, to show that
even in phrases like H\hn itro* though not a compound, but used in construc-
tion with an instrumental case, the word <jj is not a sarvanima, that is to say, the
prohibition extends even to phrases, which bear the sense of the instrumental
compound, though not exhibited in the form of the instrumental compound.

5*5 * 11 ^1 11 q^ifa 11 5*| f % (w^WSf, **wmif3i f

*wi%, 1) 11

fnr« 11 tF^ *t *hi*1 *Hi$ft tj^fiw^iPi h wf^f 11

31. And in Collective Compound, II. 2. 26 the
words sarva &c, are not sarvanAma.

This sutra mentions another exception. In collective compounds the
above words are not sarvan&ma. Dvandva is a compound in which two words
are simply joined together, the compound taking the terminations of the dual
or plural according to the number of compounded words, or the terminations
of the singular, being treated as a collective term. Thus, the compound
^uflNHtR" means, qpfir * the castes/ *|TOT 'the orders' and fiR ' the others'. The
word f<TT in this compound is not to be treated as a sarvan&ma, but an
ordinary word. Therefore, in the genitive plural we have: — iUjfrHtH l UHl
and not *T<f^t9> So also ^m i unq/ of the priors and the posteriors. 9

The word ^ 4 and* in the sftrta indicates that it closes the list of
exceptions or prohibitions that began with sutra. 29.

fom*T sift 11 ^ n tRrrf^ n famm, *rf% f ( qt&(tf£r,

M%*im , 5^5 ^rt%, 1 ) n

ifrf: 11 f^t *nrra *rfa fairer sttltfft «3*nmi<ufl r * *rcfar 11

32. (Tlieir dvandva compounds) are option-
ally sarvan&ma when the nominative plural termina-
tion j as follows.

This is a proviso to sfttra 31, and declares that before the termina-
tion *r*£ ( which is the termination of the nominative pluraj fti £$ b y*re shall find
later on, and which in the pronominals is replaced by jft), the dvandva

Bk- "I. Ch. I. $ 33, 34-] Sarvanama. J7

compounds may be treated either as sarvan&mas or as ordinary words. Thus
the plural of ^R&rOT^T may be either qr^r^^RT: or %^, that is, the word may
take either sr^ or <fr.

This sutra governs the four succeeding suites. The option allowed
in the case of the application of the affix ^J ist pi. to thedvandva compounds
of pronominals, does not extend to the dvandva compounds of pronominals
which take the affix *T%% (V. 3. 71). There the sutra 31 is, absolute.
Thus, gnir:+ *>M*: - <*cK<hri*i*>l; in the plural. This affix ^r (*4«n^)
comes before the last vowel, with what follows it, of the pronominals and
indeclinables.

?far: 11 *m ^rr to spt *r£ srfSrnr ^r t<3ft*fti fairer q$wni<ir

33. And also the words prathama l first 1 ;
charama 6 last,' words ending with the affix taya (V* 2.
42); alpa c few* ardha 'half ' katipaya c some, 1 and nema
4 half' are optionally sarvan&ma, before the nom. pL
termination.

Thus we have stvft or TOTO , ^1% or ^TH , *F^ or *PTC # *F? or
V|r: , 3frf?tt% or 3»faq*ir: , %% or %«TTs , fifr* or fitTOT: II

The words " optionally when the affix *T9 follows " are understood •
in this sutra. The governing force of the word " dvandva " does not extend
to this, and stops at this. By the word <nr in the aphorism, is meant words
ending in the affix *r*T. The rest are prdtipadikas or crude bases. The
word Sfa has already been enumerated among the pronominals, therefore,
the option in its ca p ., is what is called pripta-vibhasA, the result of which
will be that though in the nom. pi. it may be declined like 5f?r, the rest of its
declension is like sarva, while in the case of prathama &c. the nom. pi. may
be like sarva, but the rest of their derlensioiftqust be like that of nara. The
word ubhaya, has been formed by the addition of the affix TO, and it is
already enumerated in the list of pronominals, therefore, by the present
sutra its nom. pi. admits of two forms.

The application of the affix kan or akach must depend upon the
nature of these words, if pronominals, then akach, otherwise kan.

18 Sarvanama. [Bk. I. Ch. I..§ 35:

34. The words pftrva c prior)' para c after, 1
avara f posterior,' dakshina c south,' uttara l north.' apara
c other,' and adhara c inferior,' when they discriminate
relative position, not when they are appellatives, are
optionally sarvanAma before the affix jas.

These seven words have already been mentioned in the list of sarvani-
mas. They are always sarvan&mas, when they have the meaning given to
them in the list, i. e., when they imply a relation in time and space, but when
used in any other sense than the one which has been determined or fixed for
them, they are not sarvanimas, nor are they so when they are used as appel-
latives (H^rf ). Thus, when the word sftru means ' clever/ it is not a sarvan&-
ma, as, f%HJT ?% *l 1*4*7: 4 these clever minstrels,' similarly in ^fTH ^TO
•the northern Kurus/ the word grarr is not a pronominal, for though it
declares a direction, it is a proper noun.

When these seven words are sarvan&ma, they are declined like sarva,
except in the nominative plural, when they may be declined either as sarva-
n&mas, or as ordinary nouns. The option allowed in the case of these words
is a prftptavibhisft.

That there is a 'specification' (niyama), or tacit implication, of a
determinate point (avadhi), with reference to which something is to be describ-
ed by the word itself, is what we mean when we say "that a relation in time
or place (avasthi) is implied." For example, if we wish to describe Benares
as being southern (dakshina), to do this, we may specify some point, say
one of the peaks of the Him&lay& with reference to which Benares may be des-
cribed as a "place to the southward." Again we in the north may thus speak
of the people to the south of the Vindhyi mountains, as being southern, not
with reference to the inhabitants of Ceylon, but with reference (as every one
here understands by tacit implication) to us ourselves, who live to the north
the Vindhya range. Thus, ( *t or Jfh, ^q^r or MH<f : » ^ or *ncn ) *!%*& or
tfrHUi: ; *^ or 4tI<|: ; *FJ^ or *nrn 5 *P^ or 4jtj<i:.

35. The word sva * own,' when it does not
mean a kinsman or property is optionally sarvanima
before the affix jas. \ .••--•

The word ^ when it does not mean a ^mfa ' kinsman ' or >ft 'property'
or ' wealth* is always a sarvan&ma, in every number and case* as it has been

BK. I. CH. I. § 36, 37. ] AVYAYA DEFINED. J9 *

enumerated among the pronominals, except in the nominative plural where it is
optionally so. Thus *% j*tc or w y*T: ' one's own sons/ ^ itXT. or **r TUT:
'one's own ,cows'.

But when it means ' kinsmen ' or ' articles* of property 9 W: alone is
the nom. pi. As f w ^rHnft ' these kinsmen' l^jpr **ft 'much riches'.

Wcri gf|5 fanrean ^T: 11 ^ 11 xi^Tft 11 war^-^ff-
^^•■3y«cMH^r: ( fkflT^r, sift, ^ra^pr, vtinufSr ) n

*1MH I % II

^r^nr* 11 *^T3fh% ^ ^*m*i* 11

<n/n«n^ 11 *f%TT^r »nrrd ^Nw ^r ft^5 ^forprtr<ftg^8nwH^*ii

36. The word aatara being always a sarvanA-
ma, when meaning " outer 11 or "a lower garment," is
optionally so before the affix jas.

The word srwir when it means ^rf**fPT # w*-» ' outer* is always sarvanA-
ma, so also, when it means ^nWOTT or ' a lower garment 9 Thus, <q??ft tor
or HjHff *p3T : ' outer houses, the residence of Chandal, and other low castes/
So spctt or qqpsrcr. JCrr^fr: 'the inner garments, such as petticoats worn
under the upper garments." The word Hj^di when used in the above sense
is always sarvanima, but in the nominative plural, it is optionally so. But,
when it is not used in the above senses, it is never a sarvanima, e. g., *T IH<rt<!<nf
qrsrfit ' he lives between the two villages/ where the regular locative is *JV*T^.
With this sfttra ends the section on sarvan&mas, which began with sfttra 26. :

Vart : — The word M|*<u when qualifying the word jft* 4 a city,' is not to
be treated as a sarvan&ma, and is, therefore, declined like ordinary nouns.
As, m^KU l jfr mft ' he lives out of the city/

Vart : — The words formed by the affix tffa such ns, fi^ffcr and <??fhr,
are optionally declined as sarvan&ma before the case-affixes having an indi-
catory SF (fl^), which are the singular terminations of dative, ablative, geni-
tive and locative, as, ffrfore t or fitftapr, cl<fl4M or ?rf**^.

ffrn 11 ^roffn% *i«4**4ifisr ftmnwiui^Ktaiftf *rcfer 11 • }

37. The words svar •heaven,* &c., and the parti-
cles (1. 4. 56) are called indeclinables.

This defines avyaya or indeclinables. The words **£ &c. must be
found out from the Ganapa^ha ; they are the following: —

**£ ' heaven/ 4j«4|{ ' midst/ qpr^'in the morning/ 3^' again/ Bj?|^
1 in concealment* 3$$ ' high, aloft ' $A% * low, down/ jjS^g * slowly/ ^n^

30 Avyaya ? [Bk. I. Ch. L § 3&

•rightly, separately, aside, singly, particularly, truly,' 5T% ' except, without/
*iTT5 ' at the same time, at once/ ^TO?L 'near, far from, directly/ **T5 * sepfr-
rately, apart/ |0H ' yesterday/ **£ ' to-morrow/ for ' by day/ JX^t 'byjiight
or in the night/ «RT%.' at eve / &&% ' lon & ,on S since » a ,on S time/ ,p ^ ' *
little/ f^t f slightly, a little/ *nc^ ' ever/ sn^ ' gladly/ 3*OT* ' silently/ *&%
4 outside,' qrft g ' below, without, outside. HTOT ' near/ Pi*qi ' near, hard, close
by/ ST*P{ - of one's self/ ?*ir 4 in vain/ srfcir ' at night, by night, 1 spj « negative
particle, not,* %?fr ' for this reason, by reason of/ f^r ' truly, really' $* f
' exclamation/ sp^r ' evidently, truly/ Hfft ' half/ ^*r W * enclitic like, as
Br&hmanavat, priestly/ Wf% *&Fl * perpetually/ ^T^T ' division/ ft^ 'crooked-
ly, awry, over/ MJ*HI , ^-fl^UJ 'except, without' &4l4i, 5T3T 'long/ ^p{ 'expletive
particle,' ^ 'ease/'^RT ' perpetually,' *TfSTT 'suddenly, hastily/ f^jr ' with-
out,' sfpfr 'variously/ ^rfttr 'greeting, peace/ W^TT* exclamation, oblation to
Manes/ *T5T* ' enough/ ^* ' exclamation/ vfrq^, *fcnr (interjection) 'oblation
of butter/ H|v*4<j/ again, moreover, otherwise.' *rfcr ' being present,' Trig* ' in
a low voice, secretly, privately/ ^PTT 'patience, pardon/ RrfRW 'aloft in the
air/ $tar ' at night or in the evening/ $^T, fawrf ' falsely/ yjr 'in vain/ jn'
' formerly/ farot or fa*J5 ' mutually, together/ JTPTfr ' frequently, almost/ g*?j
'again, repeatedly' «l$*^ or 44I$<M 'at the same time,' ^JT^f^^
' violently ' *pft^!T^ ' repeatedly/ HT^PJ or htJ^ ' with' *FVg, 4 reverence/
ff*^ 'without/ Rr^ 'fie I' «ipf 'thus/ iffir^ ' with fatigue/ JRtf^ ' alike/
ipn^ ' widely/ Ht , HI? ' do not'

So, also the words formed by the affixes ktvi tosun, kasun, by the kpit
affixes ending in ^ or v, \, sjt or *ft , and the avyayibhiva compounds, are
indeclinables. So, also the words formed by the affixes beginning with aflte r
(V. 3. 1) and ending with STHJ^ (V 3. 47), by the affixes beginning with ^ (V.
4. 42) and ending with *W|U(<|: (V. 4. 68), by the affixes frenf*, g^, *jp£
and m^by the affixes having the sense of the affix ffe?, or by the affixes %pr
or «n^, ?rf% or qfH , H or ^r^ are also indeclinables.

The words called nipata will be given under sutras, I. 4, 56 to 61.
Thus a word which changes not ( H «$rf<f), remaining alike in the three genders
and in all cases, and in all numbers, is what is termed an HJUI4 or indeclin-
ables.

38. And the words ending in taddhtta or se-
condary affixes (IV. 1. 76.) which are not declined in all
the cases are also indeclinables. Digitiz °8^

Bk. I. Ch. I. § 39, 40.] Avyaya 3*

This expands the definition of avyaya. There are three words in the
sutra taddhital?, cha and asarva-vibhaktib. TaddhiUs or secondary affixes
i are certain affixes, by which nouns are derived from other nouns, such as,

I from H% we have HPTT. , sfcrva-vibhaktis are those words which take all

case-terminations, asarva-vibhaktis are those which do not take all vibhaktit
but some only. Thus the adverb T&: is declined in the singular ablative
only, and does not take the dual and plural terminations.

Those derivative words which do not take all the case-terminations,
but only some of them and which are formed by the addition of Taddhita
affixes are indeclinables. As ?nr: 'thence,' *HT 'there/ Both these words
are formed by taddhita affixes from the pronoun iff ' that/ the one is used
in the ablative case only and the other in the locative. So also TO , Vt 9 TO*

*irr, *er%*r> ^*r &c-

v?it*Rt: 11 ^ u xi^TPf 11 ^, jmj^srj: ( erora* )

39. The words formed by those krit or pri-
mary affixes, (III. 1. 93) which end with 1 or in 5, *fr, $ and

*ft are also indeclinables.

1 All affixes for the formation of nouns are of two kinds: — 1. Those

by which nouns are derived directly from roots : Primary affixes, 2. Those by
which nouns are derived from other nouns : Secondaiy affixes. The former
are called kpt, the latter, taddhita.

Words formed by krit affixes which end in H orn;, ^rf,^, or qqfr (connot-
ed by the pratyahara i£%) are avyayas. Thus the affix jj (technically JJ^)and
*j£ (technically UJJJ^T, III. 3. 10) are affixes which end in ^. The words formed
by the addition of these affixes will be indeclinables. Thus fr«p[ ' to give/ and

,0* VdK% * remembering' are avyayas. So also words like?fta% 'tolive,' frft^r 'to

' i drink/ are also avyayas as they end in n;, fix.

! Thus, in the following examples, the words within quotation are avya-

yas ' jflifoK ,' '^Hnrrf/ 'STTO^rtr/ 2P5#, 'he eats first having made
it relishing' (III. 4. 26). "*$" n*: (V% + % HI. 4- 9"**T>; *T '11%%/

p , Tw^**Tsgf*wtf (* + tr=n^. Rig VedaV. 66. 3). sjSr 3[r* jjm ' *tor% •

vt src% <fUi»*ivM fcr ftrftnj (Rig Veda III. 36. 10). *m : ?irfhr >fa* *****%

TO I W^ ^ *£$ 'ftf,' (Rig Veda I. 25. 21, ^ + %^ HI. 4. 11 ff».

The word *[m is used in this sutra for the same purpose^^asjtwas
employed in sutra 24, showing that these terminations must be(* aupadesika/
and not derivative. Therefore the following words are not avyayas: — *JT*J%

(d. s.), ^n^: (g. s.) ft 404% 1 f^fKr. • ^H^ib w:i *r«i**n^r : 11
s Sn^il^^T: ll 8© 11 Ttfpft 11 gjT, ^Tg^-^^i: , (fosPPT)ll

i . . jfa: f i trt *ftq^ *r%% y*far*rei, qMvw^3<i< i Hqfii 11 D 3ltlzed )y

31 Sarvanamasthana. [ Bk. I. Ch. I. $ 40, 41.

40. The words ending with ktva, tosun (III.
4. 16) and kasun are indeclinables.

As frarr ' having done ' g$*fc # having risen, ' fr^T: # having spread?
Thus in the following examples the words within quotation are indeclin-
ables :— jn^ ' *4)W n$*: 1 ymmmi ' qHn»*u V 1 ycnjrw •f^^E'fir' fl<fr*-3

WFnT^ff*^f%?ll (YajurVeda,I.28)^^f^f^ro 5^ *p|>* **imi' I
(Rig Ved. Villa. 1. 12).

41. (The compound called) AvyayibhAva (II.
1. 5) is also indeclinable.

The Avyayibhava or adverbial or indeclinable compounds are formed
by joining an indeclinable particle with another word. The resulting compound,
in which the indeclinable particle forms generally the first element, is again
indeclinable, and generally ends, like adverbsT^iiT The o rdinaryjermi nations of S
the nominative or accusative neuter. Thus sjrfafft' ' upon Hari/ *IjVI 'after
the form £~*T, accordingly.'

The word ^ in the sutra shows that here ends J.he definition of
avyaya. The avyayas, therefore, are all those words which are comprised
in the above five sutras.

What is the object gained by making the avyayibhava compounds avya-
yas? Three objects are gained, by making these compounds, indeclinables,
viz., we can apply to them ( 1 ) the rule relating to the elision (5J5 ) of the case-
affixes and feminine affixes, *>., sutra II. 4. 82 thus, in the following example, the
words upftgni and pratyagni though qualifying the word Salavah, do not take the
gender and number of the latter, zmf*H JJWF* JJ5THT: HdlSct (2) the rule relating
to^ccent when an avyaya is compounded with the word mukha, f$^T ^T)# *•*•#
sutras VI. 2. 167 & 168: Thus, 53^ yST , here, sutra VI. 2. I67, required the
final vowel to take the udatta accent, but the preceding word being an avyaya,
sutra VI. 2. 168, intervened, and the first word retains its natural accent : (3)
the rule relating to the change of visarga : into ^, this change being technical-
ly called ST^rC. thus **rTO:3rrr: , snfTCWT: , as compared with HJ 441*11:
Here, the compound yppn being treated as avyaya sutra VIII. 3. 46, applies
and prevents the change of : into ^.

The four sutras 38,39,40, and 41 have all been enunciated under the svar-
adigana. see sutra 37. These sutras properly, therefore, belong to the gana-
pajha ; their repetition here in the Ashtadhy&yt indicates that these rules are
*lf%n&or not of universal application, wz., all rules relating to avyayas in gene-
ral, do not apply to them. Thus II. 3. 69. declaring that the genitive case is
not used along with an avyaya is not a rule of universal application, for we see

J

Bk. I. Ch. I. § 42, 43. ] Sarvanamasthana. 38

indeclinables like ?$<ft: governed by the genitive case as, ^fc^fttfc ; or 15CW

fir «^sbn*rcn*re 11 jr 11 vfq$i « fix, ^hnroira* 1

42. The affix ftr (VIL I. 20) is called sarva-
n&masthAna.

This defines the word sarvan&masth&na. ' The affix f^ is the
nominative and accusative plural termination of neuter nouns. It replaces
the ordinary nominative and accusative plural affixes sgj, an< * ^HJ used
in every ocher gender. The 5 of fir is servile, and the actual termination
is f ; the peculiarity of the affix f$r is that it leng thens_Jhejg enultim ate
vowe l (VI. 4. 8) and adds a 5^ (VIL 1. 72). Thus the plural of 9T5T ' fruit' is
**sv\\W ' fruits.' So alscTapfrft fc>^P<1 , and <fr u ^lft trc*. Here, there is no
difference of form in the nominative and the accusative cases. So also
H^pf t T^ftr , sivjpt. The word sarvan&masthina occurs in sfitra VI. 4. 8.

^f%: if gfef%<r^ ^^nHr ^HH^H^^tiH ^rffir srj^raaf^Pi n

43. The first five case affixes comprised in the
praty&Mra suf. (i. e., the three case-terminations of the
nominative, and the singular and dual of the accusative)
are called also sarvanimastMna; except the case-affixes
of the neuter gender.

This defines further the word sarvan&m
tions in Sanskrit are 21 as arranged below : —
~ Singular

The above are the 21 vibhaktis, the sj of srg, the «5[ of ^, the ? of «ft^
&c. are indicatory. All these vibhaktis are included in the pratyahara gj formed
by taking the first vibhakti and the last letter of the last vibhakti. 'Similarly
the first five terminations ($, *ft, sq, «j^ &n & *ft? ) are represented by to
and are called sarvanamasthana in the masculine and feminine genders. In
the declension of nouns, it will be seen, that some nouns have two bases, one
before these sarvanamasthana and the other before the rest D '9' j b v ^"^

5

Nominative

WIT

Accusative

f%a?«iJ

Instrumental

«*<fl*ll

Dative

**3*ft

Ablative

^pft

Genitive

s&

Locative

iaa*ft

>th&na.

The case-termina«

Dual

Plural

19

*9

ft

34 Samprasaranjl [Bk. I. Ch. I. § 44, 45.

Thus the word TRPJ 'king* has the base KRtt% before the 5^ vibhaktis,
while in the remaining cases, the base is tf^and n*T as : —

Sbg. Dual. Plural.

Nominative CPTT K\A\<i \ / IIMM:

Accusative <|J1H*1 Kttrft \*HP

Instrumental CRTF Cni^f *PTfln
But in the neuter nouns there is no such difference.

1%: 11 %fir Jifii^rei' %f<ir f^FT^nfh iRr^f%^^nfff%Hrtfir #?rr H*fif 11

44. "May or may not " is called vibMsM or
option.

This sutra defines the word ft^TRT. We had already had occasion to
use this word in sutras 28 and 32. Where there is a prohibition as well as
an alternative course left open, it is called option or vibh&shi. The commen-
tators mention three kinds of vibh&sha: — pripta-vibhishl., api&ptarvibh&shi
and pr&ptipr&pta-vibh&shft. The first occurs where there is a general rule
already given, and then follows the optional rule ; the second belongs to that
class where there is no such general rule, but there is an • optional rule, the
third is intermediate between the two.

Thus the root fifc ' to swell/ by virtue of the rule fa?rrer *% (V. 1. 30),
forms its perfect tense, either by sampras&rana, or in the ordinary way.
Thus the third person singular is either *JJJT¥ or ftrWT

ipw. ^iw s « »h 11 t^rfon ^, vph 9 ^wwrn^n
^f^r 11 v^t wsr- **uk ^jfi" nnft *r *rro ^^iiuiPI^ni %*cr *f*fa 11

45. The ik vowels which replace the semi-
vowels'yaw are called sampras&rana.

This defines the word sampras&rana. It is the name given to the
change of semi-vowels into vowels. The praty&h&ra ik includes the four simple
vowels f , ^ f ^f and qt , and their corresponding semi-vowels are * , £ f |T f
and 5£ included in the praty&h&ra yan. The word sampras&rana is properly
the name of the vowel which has replaced the semi-vowel. Thus the past
participle of ^\ 'to sleep,' is formed by adding the nistyhl affix 5ff (see
sutra 26). So that we have *^ + ?r. But there is a rule (VI. 1. 15) by which
there is sampras&rana of the ? of ct^ before the ft*3[ affixes, and we have the
form grr: ' slept' Similarly from ^-sfar., ^r-^5L f **X-fo£: , ^-^pjfa^- 1^-
&%. The term sampras&rana is also employed to designate the whole pro-
cess of the change of the semi-vowels into vowels as in VI. i. 13 ; VI. 4. 131.

tf* 11 *nf*%**fir *r*r. f^nfir <rtf Rrffew 11 Digitized by Q 00 gle

Bk. I. Ch. I. § 46, 47. ] Augments £*, frw, and ft^. 35

46. Of whatsoever the augments enunciated
are distinguished by an indicatory 1 or ^ , they precede
or follow it accordingly.

This s&tra explains the special use of two of the indicatory letters
^ and a^. Where the indicatory letter of an augment is ^ that augment is to
be placed before the word in the genitive case with regard to which it is en*
joiued; while a f^r^ augment is to be added after the word exhibited in the
6th case with regard to which it is enjoined. Thus, there is a s&tra (VII. 2.
35) which says " fixdhadhituka affixes beginning with a consonant except*,
have f^". The question may arise where is this f^tobe added, in the beginning
or the end or the middle of the &rdhadh&tuka affix ? This sutra answers the
question. The indicatory % shows, that it is to be placed before the &rdha-
• dhituka affix. Thus the future termination reft, is an &rdhadhituka affix :
when this is added to the root, it takes the augment j^. Thus $+f^+CTfirs»
lAuiPi, ' he will be.' Similarly «Tft?TT ' he will cut*.

Similarly by sfttra VII. 3. 40. the root *ft takes the augment 55? in
forming the causative. This having an indicatory ^ is to be added after
the word *fir, as, *ft +5^+rcr^+*T- 4W^ 'be frightens. 1

ft^T^qic^C: « «to II ^Rftf^f II faf, «!*:, m*cm^ , TO «

47. The augment that has an indicatory^
comes after the last among the vowels, and becomes the
final postion of that which it augments.

This sfttra explains the use of the indicatory *j. The augments hav-
ing an indicatory ^, technically called ft^ augments, are placed immediately
after the final vowel of a word. Thus there is an augment called jij, * n which T
and *( are f^, and the actual augment is ^. When, therefore, it is said "let 3^ be
added to the word," the letter ^ is added after the last vowel. Thus the plural of
q^r?r 'milk' is formed in the following way : — TO^ + Jt+f (VII. 1. 72) « TOrfNr
(VI. 4. 8). Here ^ is added between ^ and ^ t. e. after the %f of *, which is
the final vowel of the word : and before this 5 the preceding short vowel is
lengthened. Similarly J* + tf + fa (III. 1. 77) - yj + JJ + *T+ fir (VI.
»• 59) - 5*31% I Similarly frosn% .

The word %w. is in the genitive case, having the force of nirdhlr*-
na or specification ; it has the force of the plural, though exhibited in the
singular, being taken as representing the class. The meaning is 'among
the vowels.' This sGtra is an exception to sGtras I. 1.49 and III. 1. 2, by
which an affix is added at the end of the word exhibited in the 6th case.

There is, however, an exception, in the case of the root *&%, in

36 . The short of diphthongs. [Bk, I. Ch. I. § 48, 49.

which the augment «pj is added not after the last vowel *j of «r, but between
sr and *£. The result of adding the augment yj between ^and ^ is, that
sfitra VIII. 2. 29. applies, and the 5 is elided. Thus Jf^ + *rr « *rcsr + 3ft +
?IT (VII. 1. 60) - HtJl^ + *rr - *n^r + *rr (VII. 2. 29)=- t^Gt " He will im-
merse". So also H 7 ^: ' immersed/ HH^ f?> H^rfcif.

^ ?*sren^ifr 11 gt 11 ^t(^ 11 ijv, ^fj 5*?-sn^5l n

??%.• 11 ^r swrf ft 3r^rr*far *Ht h^i% tft: .11

48, Of ^ vowels, Ffcis the substitute, when
short is to be substituted.

This sfltra points out the y^f substitutes of ^^. We know there are
short and long vowels in Sanskrit, but properly speaking the 5^ vowels have
no corresponding short vowels. Therefore, when, in any rule, ij^r vowels are
told to be shortened, this stitra declares that the short vowels of If and ^ , ift
jLndjfitJor the purposes of the rule, will be f and ^ respectively. Thus, I. 2.
47 declares: — "the short vowel is the substitute in the neuter of a crude form
provided it end in a vowel." Therefore in compounding sjfif + ^ 1 the ^
must be shortened. Properly speaking ^ has no short vowel corresponding
to it, but by virtue of this stitra, f supplies the place of such a short vowel,
and we have JSffaft ' extravagant 1 so nh *rfifj *' disembarked or landed, 9
*fir 1 OT*I ' near a cow.' All avyayibhiva compounds are neuter (II. 4. 18).

Why do we say ' of qn£ '? Because the short of other vowels will not
be ^5. Thus the short of *rr is *f. As, *n%+ ^r - «rf%*3T|: , *rf*T + HI5TT -

WfiPTTCT:!

Why do we say • when short is to be substituted ? ' Because when
JW or prolated vowels are to be substituted for 3^ , the {*£ will not be the
substitute. As f^r*rT O Devadatta ! fn^T !

to* ^n^nn 11 y% 11 Tn^ M t^ 9 ^n^-^pn 11

W«l*fwfaf*M$% *rfw sir JiRrnfRiT II

49. The force of the genitive case in a sfttra
is that of the phrase ' c in the place of" when no special
rules qualify the sense of the genitive.

This sfltra explains the mode of interpreting words used in the posses-
sive case (sixth case) in the sfttras of P&nini. The genitive case or shAstyhi— ^
denotes many sorts of relations in Sanskrit, such as causa tion, possession, rela-
t ion in p lace, comparison, ne arne ss, proximity, change, collection, component
member, &c. So that when a word is in the genitive case in a sfitra generally,
the doubt may arise in what sense that genitive is to be used. This aphorism

Bk. I. Ch. I. § 50. ] The liIcest is the substitute. 87

P lays down the restrictive rule for the interpretation of such words. It says
L that the force of such genitive is to convey the meaning of ' in the_place at*

Thus in the sutra 3, the word f«£ is in the genitive case ; the* literal
translation being : — " of ik there is guna and vfiddhi." But " of" here means
by virtue of the present rule of interpretation " in the place of." So that the
sfitra means /in the place of f^. ' Similarly in the sutra immediately preceding
this, viz. 48, we have the word jr^ in the genitive case and the sfitra means 'of
*nj.' Here also " of " means " in the place of."

Thus in sutra f?ir£ (VI. 4: 36) ' of hanti, there is if* the word f?%, is the
genitive of *f£?r ; this genitive is used generally and its force is that of " in the
place of." Therefore, *r displaces the verb ^ in*n^ or imperative mood, and
we have srft ' kill thou.'

This sutra is also thus translated by Dr. Kielhorn : — Only that which
is enunciated in a rule, 1. e. that which is read in a rule or rather that which
in the real language is like that which is read in a rule in a genitive case,
or that which forms part of that which is enunciated in a genitive case,
assumes the peculiar relation denoted by the word wpt ' place ' t. e t is that in
the place of which some thing else is substituted, but some thing suggested
by the former does not assume that relation.

The word WR" here is synonymous with TOy or ' occasion. 9 Thus
in the sentence ^rot WR 5TJ VWRw^i the word sth&ne, means 'prasange*
i. e. wherever there is occasion to spread darbha grass, spread there the iara
instead. Similarly in sutras ^r#^: (II. 4- 5 2 )t or 1&i *fa (II. 4. 53), the
words ' asti/ and 'lAru' are in the genitive case, and mean, ' wherever there
arises occasion to use the verbs *r^ or ^ use there the verbs ^ or *^[ instead
respectively/ Thus Hft^T, HN^*, HftPR[, are the future, gerund, and pas-
sive participle respectively of ^5; so slso, q^Gt, ^5^ and ??&ni^ are the
same forms of £.

The word wft^PTT is a bahuvrihi compound of two words, not in

! apposition, (vyadhikarana bahuvrihi, see II. 2. 248 35), and it qualifies the
word shashthL The compound means **n% ^fpfttf^T, - that which assumes
the peculiar relation denoted by the word sth&na.' Consequently the seventh
case-affix is not elided in this compound ; on the analogy of compounds like

^nStorRdV n \o 11 q^ift 11 wt^n«-<UcW 11

?f%: HWI% KlwWHUMW*«KfW WWt *nfa *fJQTTC ||

60. When a common term is obtained as a
substitute, the likest of its significates to that in the place
of which it comes, is the actual substitute. Digitized by (

K

38 The four sorts op proximity. [ Bk. I. ch. I. § 50.

This sutra also lays down another rule of paribh&sha. To explain this
we must take an example. The guna of f , T, 31, and W[ 9 is ST, ?» Mt.
There is nothing to specify what letter is the guna of what It might
be said that *jt is the guna of f because the definitions as laid down up to so
far, do not say anywhere that «TT is not the guna of f. So that when in a
word Hkc(ftrVeare told 'let there be the guna of % ' we do not know what
specific guna letter is to be substituted, whether it is %[, 5 or *fr. To clear
up this doubt, this sutra declares that the likest of its significates is to be subs-
tituted. Now the nearest in place to f is q; both having their place of pro- ^
nunciation, the palate. Similarly *jt is the guna of ? 9 because both are labials.^

There are four sorts of proximity or nearness : — (x) nearness in
^ J place or ctft (i. e. palate, throat &c.) ; (2) nearness in meaning or *nfr (such as,

singular terminations to be used after words used in the sense of singular); (3)
nearness in quantity or n*uuj (such as a short vowel to replace a short vowel,
a long vowel, along one); (4) nearness in quality or ip§ f such as aspirates to re-
place aspirates, and sonants to replace sonants). Of all these approximates
the first vis., the nearness in the organ of utterance has preference, in the
selection of proper substitute.

(1). Thus in applying rule VI. 1. 101, "when a homogeneous vowel
follows ak, the corresponding long vowel is the substitute for both," we
must have recourse to the present sutra for finding the proper voweL Thus— -
V* + TT ™ * a TPT ; here the substitution of the long *jr, for the two short
S^s, is an example of substitution by nearness of place, for both *f and *qx
are gutturals.

(2). A good example of substitution by the sense of the word is
afforded by VI. I. 63 : — " The words qf, ^, *H &c, are the substitutes when
the case-affixes jth (acc-pl.) &c, follow." Here, we know the substitutes,
but we do not know of what they are the substitutes. We apply the maxim
of *pfr: " that only can become a substitute which has the power to express
the meaning of the original, /">., a substitute takes the place of that the meaning
of which it is able to express." Thus we find that *£ is the substitute of
IRi *?r of ^cf,T5 °f Hlf^r, &c., which convey the same meaning and con-
^ tain almost the same letters. Another example of this is afforded by VI.3.34,
by which feminine words in a compound are changed into masculine. Thus
*m w *fl (fern) + *r*f«r = 4l<J u 4il34ft: .

. (3). An example of prosodial substitution (JlHniT^:) is afforded by
rule VIII. 2. 80, "? is the substitute of what comes after the f^of WTO»not
ending in ^and in the room of ^ there is * ." Thus sspt + ^ - *J£ +
*l& - *?$$; ^nr+>^r5 k -^RF+«iT^(VII.3. io2)-sq^+^ni - *p*2^-
Here, short * replaces the short %f t and the long ^ replaces the long *jr.

/

Bk. I. Ch. I. § 51.] The *rj substitutes or ^. 39

(4X An illustration of qualitative substitution is afforded by VII.
3. 52, " a letter of the guttural class is substituted lor ^and ^ when an affix
having an indicatory ijr follows." Applying the rule of *pu*ti we find that
«£ which is an alpaprina and aghosha letter replaces ^having the same
quality; while *£ which is ghosha, and alpaprina is replaced by the letter *r
possessing similar quality. Thus qjj+ t^ (III. 3. i8)>-qpc:; nrir+^—

"inn ; rs-st+tpj^nr 11

Though the anuvfitti of the word sth&na was understood in this sfttra
from thejast. the repetition of this word indicates the existence of the follow-
ing ^aribhAsha

fever there exist several kinds of proximity between that for
which something else shall be substituted, and its possible substitutes, there
the proximity as to the organ of speech is weightier than the rest, 1. e. 9 there
that only is substituted which is nearest as regards the organs of speech
with which both are uttered.

Thus in finding the guna substitutes of \ and 7 out of the three guna
letters *r, ^ and *fr, we find that *j\is a proximate substitute having regard
to prosodial measure, i. e. t %f and f and 7 have all one mitri ; while having
regard to the proximity of the organ of utterance, we get ^ and *ft; the
latter however prevails to the exclusion of the first; as, %gr, ttjfcir.

Why do we use the word " likest " in the superlative degree ? Where
there are many sort of proximities, the likest must be taken. Thus qp£ +
f^rft-SFtrefo Here, by VIII. 4. 63, " after a letter of ^ class, ^ is changed
into a letter homogeneous with the preceding," £ must be changed into a
letter of the class ^. Out of the five letters of this class, «£ and £are both
aspirates, but «r is hard, and (is soft; so ^ and y are both soft, but 5 is
unaspirate, and £ is aspirate, therefore these must be rejected ; the only letter
which has the nearest approach to £, is *£, which is both soft and aspirate.
Similarly in ft«jgw rer> the f has been changed into if.

51. When a letter of *n pratyAhAra comes as a
substitute for m it is always followed by a r.

This sfttra consists of three words vis., *: which is the genitive singular
of 3T and means literally 'of ^ 9 or * in the place of tf; the second word
is *sr$ a praty&h&ra denoting *f, f and * long and short ; the third word
is rTC which qualifies sr^ and means ' having a c after/

This sAtra is useful in fixing the proper substitutes of qf. Thus, there.

40 Substitute replaces the final. [Bk. I. Ch. L § 52, 53,

are only three guna vowels *f 9 if and *ft. Of these what is the proper guna for
3T ? By the application of the rule of ' nearness of place 9 we see that %f is /
the nearest substitute. So that *r is the guna of 57; and further by this/
rule, this %f must have a Rafter it. Thus though technically speaking %f is f /
the guna of 37 , the actual substitute is *jj # Thus fr-Mfr -^frfr; similarly fsfr. /

So also when ^T is replaced by f as by sutra VII. 1. 100, this subs-
titute f must be followed by a £. As ¥? + *qf^-f^ + £+*rf^=farfilr 'he
scatters;' similarly Puft 'he swallows. 9

So also when gr is replaced by 7 9 as by sutra IV. 1. 115 this 7 must
be followed by a £. As ft + *rrar + ST3L " CTHjpr: 'son of two mothers'

This ^ is to be added only when ^f is replaced by *r, f or T~V
(*TJL) vowels , and not -when it is replaced by any other letter. J

Thus by IV. 1 . 97, " the *& of the word spjrar is replaced by sjg^ (tech-
nically MJ«h£ ) when the affix J3£ is added." Thus, 3>ft? + f ^ - *pn£ +
*T3£ + f ■* *ffanrr%: , ' a descendant of Sudhfttri ' Here the substitute of
gj f that is to say, the syllable ^T^is not followed by j.

The * in the text has been taken by some to be a pratydh&ra formed

by the letter ^ and *r of 5TT ; and thus it includes the letters £ and w. In
that case the sutra would mean that 375 substitutes of gj and 5j are always
followed by £ and ^ respectively. Thus guna of p = ar, of ]i «al.

3T%T?cXra H *ft II tr^Tf^T II W3:, WxT^T, (<3TT^9J:) II

52. The substitute takes the place of only the
final letter (of that which is denoted by a term exhibited in
the genitive or sixth case).

The rule of substitution by nearness of place applies to the last
l etter of the word which is exhibited in the genitive case and not to all its
letters. Thus by sutra VII. 2. 102, it is declared ^^ii?HH : "in the place of ?ot
&c # there is %f." It does not mean that the whole word ?ot , is replaced by
%f; though at first sight it may appear to mean so ; but it means, by virtue of
the present explanatory s&tra, that the last letter of nr$, namely w is to be
replaced by *f.

So also in sutra I. 2. 50 yjifr^T: the short % replaces only the final letter
of iftaft and not the whole word, as q^llftj: ' purchased for five gonis.'

wntrar

ffrf: II flTO * ^JIW: %^*Wft W^ *nft Bigltized by GoOQle

1/

Bk. I. Ch. I. § 53, 54. ] ft* Substitutes. 41

53. And the substitute which has an indica-
tory r (even though it consists of more than one letter)
takes the place of the final letter only of the original
expression.

t This sutra is an exception by anticipation to sfltra 55. By that all

/ ^substitutes, consisting of more than one letter, replace the whole word.

Thus in forming the dvandva compound of htt and ffc there is a
rule (VI. 3. 25) by which it is declared:—' In the place of words ending in
3T there is the substitution of ^TPHF in forming dvandvas.' Now the substitute
«TH^ is a ftg_ substitute, and therefore it applies to the last letter of the word
and not to the whole word, vis, the Sff of »nT is only changed into *jr and not
the whole word ; and we have the compound nut i fttu ft, so also fhrriW V.

There is an exception, however, in the case of the substitute wnre (VIL
1. 35) which, though it has an indicatory y , yet replaces the whole expressions
5 and ft and not only their finals. As *fa*nt « may he live, or may you live*;
instead of sffaw, 'may he live,' or srNr ' live thou.'

*n\i xito II «# ii t^rft 11 *n^, xit^t, ( «re: ) ■

54. That which is enjoined to come in the
room of what follows is to he understood as coming in
the room only of the first letter thereof.

When any operation is to be made in a word following another, such
operation is to be made in the initial letter of such subsequent word.
To give an illustration:— there is a rule (VI. 3. 97 ) by which it is declared.'
"Inthe place of *r* used after the words ft and *^ and the particles called
TOP* there is fc" This rule may be stated in other words as:—' In the place
of *n: there is * when ** follows ft or *p^.' Now it is clear that f is not to
replace all the letters of ^ but only one. By sutra 52 that letter would
have been the last letter of *pr, but this sutra makes an exception,
namely, where an operation is directed to be made in a word, simply by reason
of its being placed after another word ; such change is to be made in the
beginning of such second word. Therefore, the f replaces the *r of rnr and
we have ft + ^r -flf (the final SJ is added by V. 4. 74 ) « an island,' *****
•promontory.' So also in the sutra fcm: (VII. 2. 83) ♦ long * is the substi.
tute of *pr when t he latter comes after the root «n^.« Thus ^ + *TT -
WT! + f«r-*mfhr: • Here also the f replaces the initial *r of *pr.

How can we find when an operation is to be performed on a subse-
quent word ? When the previous word is exhibited in the 5th or ablative case
see sutra 67. ( rsr '

Digitized by VjOOQ IC

42 ftr^ substitutes. [ Bk. L Ch, I. § 55, 56.

( sn^w: ) n

55. A substitute consisting of more than one
letter, and a substitute having : an indicatory *T take the
place of the whole of the original expression exhibited in
the sixth case.

Tliis sutra is an exception to sutra 52 by which it was declared that
an &de£a or substitute replaces only the last letter. This sutra on the contrary
enjoins that an &de£a consisting of more than one letter replaces the whole ex-
pression in the genitive case, and not only its last letter. Even where an
&de£a consists of a single letter, but if it has a 5 as its indicatory letter, it re-
places the whole word.

Thus, there is a sutra y^r *fa: (H. 4- 53) which means 'in the place
of *j let there be T^.' Here the fldesa ^ consists of three letters (more than
one), and therefore, it replaces the whole word ^and not only the last letter
3?. That is to say in forming certain tenses the verb !£ is replaced
by *^. Thus the future tense of wis q^GT 'he will speak. 9

So also there is a sfttra (V. 3. 3.) by which it is declared: — "In the
place of f*n there is f^." Here the &de£a f is an &de£a having the *£for its
indicatory letter, and though it consists of only one letter, it replaces the
whole word fy% f and not only its last letter \.

Of course, it must be understood that the phrase ' consisting of more
than one letter* applies to the actual substitute, and not to the substitute with
its indicatory letters, the latter being for the purposes of counting regarded as
mere surplusages and not to be taken into account. To see whether an affix is
Mj^chiq^ or not, the indicatory letters must not be counted; and the affix must
be stripped of all its f£ letters.

^TRrtftr *4Riw m

56. A substitute (&de?a) is like the former

occupant (sthAni) but not in the case of a rule the occa-
sion for the operation of which is furnished by the letters
of the original term.

This is an explanatory sOtra, or this may be considered also as *n
atidcia sutra, in as much as it declares that the operations to be performed

Digitized by VnOvJVlL,

Bk. I. Ch. V. § 56.] "Substitute is like the principal..

43

K*

t

J

I

on or by the original, may be similarly performed on or by the substitute,
but with certain restrictions. The words of this siitra require some detailed
exposition. Sth&ni is the word or the portion of a word which is replaced.
>fde£a is the expression which replaces something. Al-vidhi is a compound
of ' al ' meaning alphabet, or letters in general, and * vidhi ' rule i. e. any
rule applicable to letters.

There is an exception, however, to this rule in the case of rules, the
occasion for the operation of which, is furnished by the letters of the original
term. That is to say for the purposes of orthographic rules, an adeSa is not
like the sth&ni. A rule which is applicable to the letters of the sth&ni, need not
be applicable to the letters of the ade*a.

Thus the gerund of verbs is generally formed by adding the affix
rm e. g. vfy + nrr. Now VII. 2. 35. prescribes the insertion of an intermedi-
ate f^ before ArdhadhAtuka affixes beginning with any consonant exceptor*
This rule is clearly an al-vidhi, because the occasion for its application is con-
ditional on the letters of the affix. The affix ^T begins with ^ and is an irdha-
dhatuka affix and fulfils all the conditions of the rule. Therefore, we have

7£+T + *rr-7#*rr(VI. 1. i6;VH. 2. 37).

Now in compound verbs the affix tqT is replaced by ^r ( tech. ^T^).
Thus iFfe + ^T. This &de£a if which takes the place of *qj has all the func-
tions of *%T viz. it has the power of making gerunds; it makes the word an
indeclinable (see rule 41 ante), as the affix tqT does. But the affix Ctf takes
an intermediate ^ ; should, therefore, the adesa ^r take also the f or not ?
Now the augment f which *qj takes is by virtue of VII.- 2. 35, because it be-
gins with a consonant of ^ pratyfiMra ; while if is the only consonant which
is not included in this pratyfiMra. Therefore, for the purposes of taking f by
rule VII. 2. 35. the adesa * is not like t*r; since that rule is one which has
its scope when there are certain initial letters ; or an al-vidhi. Therefore, we
have^ijfr.

Substitutes replace either a xnq ' root,' or an m^p * base/ or a kpt
affix ' primary affix/ or a taddhita affix ' secondary affix/ or an avyaya ' inde-
clinable/ or a 5 J affix 'case affix/ or a ff(% affix 4 con jugational affix/ or lastly
the full word or pada.

The substitute of a dhatu becomes like a dhatu. Thus sfitras II. 4.
52 and 53 declare : — *^is the substitute of the root *r?J» an< * 1% of the root
W when an ardhadhatuka affix follows.' Here the substitutes ^ and ^. are
treated as dhatu, and as such get the affixes ?jnr &c. by III. 1. 96, &c. Thus

The substitute of an anga becomes like an anga. Thus VII; 2.
fo3 declares:—' q? is the substitute of the base fifr^ when ^c^sej-gLffir follows/

44 Substitute is like the principal. [Bk. I, Ch. I, § 56:

Here, *f gets the designation 'base' and so rules applicable to base, are
applied to m also. Thus in ^r , 4iitqp£ » 4t: &c. f we have fH, the lengthening
of the vowel, and ^3 substituted by VII. 1. 12, and VII. 3. 102, and VII. 1. 9.

The substitute of a krit affix becomes like a krit affix. Thus VII. 1.
37 declares: — 'fsp^ is the substitute of the krit affix tpfr when the verb is a com-
pound, the first member of which is an indeclinable, but not w^ 9 Here, *w%
is also called a krit affix, and as such, sutra VI. 1. 71 applies, and 5^ is added.
As Jjfr+FSTS. - Iff + * + *T - JTf>**; similarly ir^nr &c.

The substitute of a taddhita affix is like a taddhita. Thus VII. 3. 50
declares : — ' ^ is the substitute of the affix * ' as *fir + *^E (IV. 2: 18)
a Vrffci^, here, f^ being treated as a taddhita, sutra I. 2. 56 applies and the
word formed by it, gets the name of pr&tipadika, and is so declined.

The substitute of an avyaya is like an avyaya. As, Jft^nr • if^mv
The affix ?RT makes avyayas (1. 1. 40), its substitute F*r^ will also make the
word an avyaya, and as such sutra II. 4. 82 applies, and the case-affixes are
elided after these words.

The substitute of sup-affixes becomes like sup. Thus VII. L. 13 de-
clares: — c 3T is the substitute of the sup-affix £ after an inflective base ending
in short mJ Here, 3r is treated as a sup-affix, and therefore sutra VII. 3. 102
applies, and there is lengthening of the vowel, as, fSTPT ; 8OTW. .

The substitute of tin becomes like tin. Thus III. 4. 101 declares :—
'm% 9 ^, ?T and sri are substituted for ^5, 1*5, tj, and ftj when tense-affixes
having indicatory ^ follow.' Here the substitutes m% &c, also make the
words ending in them pada by rule I. 4. 14 as, MJ<$t*1^ f HJ<&*4 &c

The substitute of a pada is like a pada. Thus VIII. 1. 21 declares: —
1 ^ is the substitute of jprnj , gm*^ , gmi*^ , and *rf[ of %&tt% , *r**imw
and qrerrapg / Here, *f[ and ^ are treated as pada, and the final ^ is
changed into visarga, as ^t , m .

Why have we used the word ^ in the s&tra? Without it the aphorism 1
would have been ^l^l^Pl ^fcTqfr, and this being a chapter treating of defi
nitions, the meaning would have been, ' an ade£a is called sth&ni.' This
certainly is not intended, for had it been so, then all operations would be per-
formed by or on the substitute, but none by the original* But this is not so.
Thus rule I. 3. 28, says ' the root fg^ preceded by SJT is itmanepadl. 1 *if
the substitute of f^ will be itmanepadt, as *TT?r5?; but f^ also will be so in
its proper sphere, as, *nf*T

The word |de£a is used in the sutra to indicate, that the rule applies
e ven J o partial substitutions : as changing the % of ft into * (III. 4. 86) in the
Imperative e. g. ^f J. \

The word al-vidhi is used in order to indicate that the substitutes of ft*.

tj

Bk. L Ch. L § 57.] Substitute is like the principal. 45

?ft^ , t*f &c, such as the change of ^ into *ft (VII. 1. 84), ^ into *JT (VIL 1.
85), 1 into *j (VIL 2. 102 *, are not to be treated like the original : as «fh# VIP,
*n Had these substitutes been treated like the original, rule VI. 1. 68 would
apply, and the case-affix *g would have been elided.

57. A substitute in the room of a vowel
caused by something that follows, should be regarded
as that whose place it takes "when a rule would else take
effect on what stands anterior to the original vowel.

This sutra consists of three words: — *r*: genitive of the praty&hdra
^T% meaning * of a vowel/ and means an &de$a which takes the place of a
vowel. The words adesa and sthfinivat are understood and are to be supplied
from the preceding sutra. The second word is 'parasmin' loc. sing, of 'para'
meaning 'in the subsequent. 9 The locative has the force of ' on account of or

/ by reason of/ The third word is purva-vidhau loc. sing, of pfirva-vidhi mean*
ing'a rule applicable to a preceding thing. 9 The whole sutra thus comes to
mean that an &de£a which replaces a vowel, becomes sth&nivat (like vowel), pro-

1 vided that the substitution has been occasioned on account of something fol-

\. lowing and when a rule is to be applied to anything preceding such an &de£a.

Thus there is an affix called fy^ the actual affix being f; the other

letters ^ and ^[ being f*. The peculiarity of every affix having an indicatoiy

— ^ is to cause the vpddhi of the penultimate *f ; (sutra V II. 2. 1 16) as
V*+T« a, nf^r. Now in forming the denominative verb from the wordqrj, we have
to add the affix fisj%. Thus qj + f; but before nich, the *T of STJ is substituted vt £
by a 'jopa 9 or blank, and this blank takes the place of the vowel ? ol qj and
becomes sth&nivat by force of the present sutra. The result of becoming
Sth&nivat is that though the ^r of ^ is really a penultimate and ought to be
vriddhied before 'nich 9 it is not so, the lopa-substitute not allowing the *T tobe
regarded as penultimate, and thus we have the form qft the present tense
third person singular of which is q?qfc. The equation being qr? + • + T, the zero
preventing the ^ of q^ from becoming penultimate. So also in the word *fvft*,
the aorist(<j^) of *\r. Thus*r+*>f + fj«H+*X+«+f?r. Here, the *f of
TO is elided by VI. 4. 48. (when an ardhadhatuka affix follows there is elision
of the short Sf of that which ends in short qr;) this elision is thus caused by
something which follows (parasmin.) Now there is a rule (VII. 2. 7.) which de-
clares 'vriddhi is optionally the substitute of a laghu *J preceded by a consonant ^
when r%^(aorist) follows. 9 However in the present case the z^Stftf? 11 *"'**

AM'

'46 Substitute is like the principal. [Bk. I. Ch. I. § 57.

dered as sthinivat, and prevents the application of sfitra VII. 2. 7 ; otherwise
we get the form ^RPftst which is inadmissible.

Similarly *g*3|T +3TX • W$*!|3M Here, the short *r is the substitute of
the long *7T of *rar by VII. 4. 15 ; because of the subsequent (parasmin) affix
*kap\ However, the word H$WL will not be considered as a word ending in a
short vowel, for the purposes of the application of rule VI. 2. 174, (the ud&tta
accent falls on the last vowel but one of a word that ends in a short vowel
and is the last member of a bah&vrihi compound followed by the affix kap).
On the contrary the short %f is regarded as sthinivat to long *jr. Thus

It is needless to say that this rule applies only to vowel &de£a, there- ^j
fore, where an ide£a replaces a consonant, the previous s&tra 56 has its scope \
and prevents its being sthinivat. Thus there is a kfit affix called FTPf, the real J
suffix being ^r, the 9£ and ^ being diacritical letters. Before this 7 the verb
TPITt (to come) loses it ^ 9 and lopa or blank takes the place of W . Here
then lopa or blank is an ide£a and 1 a consonant is sth&ni.

Now there is a rule which says (VI. 1. 71), that before kpt affixes
having an indicatory <g , the letter tr is added after verbs ending in light
vowels. Now in STPPJ , when 5 is omitted, thfe ?r of ?r becomes final ; and
therefore a ^ is required before ^r (which is a faw affix). Thus we have 3TPTO*

But had lopa or blank become sthinivat to *J then the *f of *r would
have remained penultimate as it was when it was *rnp(; and no 1£ would have
been required. But it is not so.

Similarly in forming the word TO:. Thus ij*S£ + ^ (III. 3. 90)
*J*( + f^ (VI. 4. 19). Here, ^ is substituted for fj; this will not be sthinivat
to f£; though this substitute is caused by something that follows (parasmin ?.
Were it to be sthinivat, there would come the augment §g^ ( 5 ) by VL 1. 73.

Similarly *J*rtST5, the aorist of fs^[. Thus *r+5>*r + fli*[+inx-*r+
^^+•4- mi (VIII. 2. 26). Here the 5 is elided on account of what follows,
but still the lopa substitute here is not sthinivat for the purposes of VIII. 2.41.
Had it been sthinivat, ^of f^r would be changed into qr .

Moreover this rule applies where a vowel is changed on account of
something which follows (parasmin). Therefore, where the change has not
been occasioned by anything subsequent, the ide£a (though of a vowel) is
not sthinivat Thus the words j^srjfir: and *g^Nm%: .

The word gi*ui% is a bahuvrihi compound of ^Rr and m^X f mean-
ing ' he who has a young wife.' In forming this compound, the 5JT of W is re-
placed by fir ( tech. fa^) by sfttra V. 4. 134. This fa will not be sthinivat
to *jt, as it is not caused by anything which follows. Had it been regarded
as sthinivat, the 5 could not be elided by VI. 1. 66 ('there is elision of *

\/

Bk. I. Ch. I. § 57. ] Substitute is like the principal. 47

and ^r when any consonant except £ or ^ follows'). Thus, JTOPTT+Rr-f*-
*RC + o + fa (V. 4. 134) - y^rft: (VI.. 1. 66).

Similarly the word ^4IUH€I: , ' a descendant of 14IUHI41' also illus-
trates this rule, wnwf is a bahuvrthi compound ofcqnt and *rf» meaning
' whose foot is like that of a tiger. 9 Now the final *r of 'p&da' is elided in
such compounds by V. 4. 138. This lopa substitution is not caused by any thing
that follows, therefore, this lop&deAa .will not be sthinivat for the purposes
of the application of sOtra VI. 4. 130 (' T^is the substitute bf'tEe~wor3 ^1^
when part of an inflective base ending in the word *Tf and entitled to the
name of Bha. 1. 4. 18'). Therefore, wmrrrj. +*T* (IV. 1. io5)-^nmr:. Had
the lopa been like the original, there could not have been the substitution of
'pad ' for 'p&d.'

Similarly the word * nfr* 5r, 1st per. sing. atm. present tense of the root
#>ft 'to shine.' Thus *nft>ft+ff **sn$r*ft+S(HI. 4. 79). This change off
into £ is not caused by anything that follows, hence ^ is not sthftnivat to f f
for the purposes of the application of rule VII. 4. 53 ('the finals of didhl and
vevl are elided when an affix beginning wither or f follows' J. Had ^ been
sthanivat, the final f would have been elided. .

Similarly an &desa though of a vowel, and even though occasioned *
by a subsequent something, would not be sth&nivat, if there is not the applica-
tion of a rule to something that precedes the &de£a f pfirva-vidhi). If the
rule is to be applied to something subsequent to the &de«a 7 it will be applied
to the exact &de£a without considering what was the letter which the ide^a
had replaced. , It is only when a rule is to be applied to a prior object, that
it becomes important to consider whether the &de£a is sth&nivat or not.
Thus the vocative case of ift is f ifc I The vocative is formed from the
nominative case (II. 3. 49). The nominative of *ft is formed by chang-
ing ^ft into qft and adding the case termination ^ . Here the adesa sft",
replaces 5jft, on account of the subsequent termination fj . Now there
is a rule applicable to vocative cases which declares that the final ^ is drop-
ped after short vowels and ^ and *jt (VI. 1. 69). *

Now is the sft of ift to be regarded as sth&nivat to *ft f which it re-
places; for the purposes of this rule of elision of finals? If so, the conse-
quence would be that the final ^ would have to be dropped in the vocative*
But it is not so. Because the rule VI. 1. 69, which ordains the dropping of f£
after short vowels and ^ and *jt applies to a letter, which is not anterior to
the &de£a *ft, but to one which follows it, and consequently there being no
pfirva-vidhi here, the final fr is retained and we have the vocative *fb I

Similarly the word ^nrffaT:, 'the pupils of B&bhravya.' The word
TOP* .is derived from, *w in this wise. q% + *i£ (IV. \. 105) ■» *pfr

Digitized by VjOOQIC

48 Substitute is like the principal. {Bk* I. Ch. I. $ 58.

+ **£ (VI. 4. 146 and VII. 2. 117) - WTO* + m* (VI. 1. 79) - TO**:.
From this is formed OTrtfa in this wise: — TOFRT + Sf (IV. 2. 114) *■

*n*nr+fr* (VII. u 2) - *ra«5 + f* (VI. 4. 148) - qra* +"t^r(Vi. 4. 150).

Here, had the substitute *n£ which replaced the *fjr of m^t 9 been sth&nivat to
VT, rule VI. 4. 150 ('there is elision of the ^r of a taddhita affix coming after
a consonant, when long f" follows'), could not have applied. But this sub-
stitute *r^ is not sth&nivat for the purposes of rule VI. 4. 150, inasmuch as that
rule does not apply to any letter that precedes the substitute %f% , but to the
very letters of the substitute itself.

Similarly the word ^fa 1 :, which is evolved from the root qT, in this

wise fa +*r + ftf (HI. 3- 9 2 )~ft + *T +X (I- 3- 8) = fir+ * +•+ X (VI.4-64)
« ftfa. Let us then form a secondary derivative from the word Reapplying
5^ (IV. 1. 122), the whole of which affix is replaced by m (VII. 1. 2). Thus
fafa + s* " ft* + ** (VI. 4. 148) - ^T (VII. 2. 118). Here, the lopa
substituted for the 3JT of 1ft by VI. 4. 64 is not sth&nivat for the purposes of the
application of IV. 1. 122 by which the affix **r is added. Rule V. 1. 122 says,
'dhak is added to a word ending in f, provided it be a word containing two
vowels/ Now firfa ends in f and contains two vowels ; but if the lopa were
sth&nivat to the elided qjT, then it would be a word of three syllables. That
is however not so ; because the rule enjoining the affix *qr is not applicable
to anything preceding the lop&de£a, but is enjoined with regard to something
after this substitution.

f^IJ, ( qmfa g ^ )

58. Not so, in rules relating to the finals of
words, to the doubling of letters, to the affixing of varach,
to the elision of ya, to accent, to homogeneous lelters, to
anusvira, to the lengthening of vowels, and to the substi-
tution of ja«_ and char letters.

This sGtra lays down an exception to the previous sfitra, by which it *
was ruled that an &de?a which replaced a vowel becomes sthlnivat under cer- /
tain conditions. This sfltra says that a vowel &de£a is not sth&nivat under *
the following circumtances.

i. ipftft: — ' A rule relating to the last letter of a word. 1 Thus the
*T of *nj is elided before the terminations of dual and plural numbers (VI. 4.
in). Thus we have ^m?r 'he is/ w: ' they two are/ *f**r ' they are. 1 Now in
the technical phraseology of Sanskrit grammarians, the *f of W: and mf*Q is

Digitized by VJjVJOV IL

Bk. I. Ch. I. $ 58. ] Substitute is like the principal. : 40-' %

replaced by an idesa callecferx, the peculiarity of which is that the whole of

it vanishes. So that before tKeword W: there is this &de$a existing, though

it is in an invisible form. This invisible &deia will be sthinivat, as it replaces

a vowel, by reason of something that follows, i. e. the affixes ir: and* *rf*l

being f|=*w (I. 2, 4) ; and it will have all the powers of *f by the last sfttra, in the -

absence of any rule to the contrary. .Therefore, in the sentence d£t TO ' whoo^* -* ; -

two are' the final *ft of sfrand the invisible ^ of CT: ought to coalesce by rules

of sandhi into *rnr (VI. 1. 78). But this incon gruity is preve nted by the pre-.

sent stitra, by which the change of *ft into *rnt is a <r?F3 fafa or a rule relating

to the~fihals of a word, therefore, we have q£t w I So also in *fnf^T srfcfj rule VL *
1. 77 is not applied.

2. fiJ^Hfafa : ' a rule relating to the doubling of letters.' In the sandhi

of *fa+spr we have ^^r. Here f is changed into -5. If this ^ were sth&nivat -
to f , then the rule by which ^ could be doubled before 5 would be inappli-
cable. But f[ is not equivalent to f > though its substitutej and we have the doub-
ling of ^and get the form qF5<FT "Therefore, in rules relating to the doubling
of letters the &de§a is not equivalent to the sth&nL

3. . irfafa: 'a rule relating to the affix 5x3/ The affix *?^j[ is a kpt
affix which is added to certain roots to form npunsof^ agency ; thus from f^ *
1 to rule ' we have figp ' ruler/ So from the intensive verb ^HTHT we have
4I4I4<: ' a vagrant7~Tt is formed in thiswise qrara+TCtfll. 2. i76)«^rrarw+%
+^T(VI. 4. 48). Here the *r of if has been dropped and replaced by lopa, on
account of its being followed by the JLrdhadh&tuka_affix jt;. The next
step, is to drop the ^ by rule VI. I. 66 which declares that the letters i£
and ^ are dropped before affixes beginning with any consonant except

?£ t\ e. t beginning with consonants of **[ praty&h&ra. Thus *|PIT + o + *T.
Now there is a rule (VI. 4. 64) which declares that the final long sjt of a base
is dropped before &rdhadh&tuka affixes having an initial vowel. Here there-
fore, if the ade£a 'lopa/ which replaced the vowel ^r, be considered as sth&nivat,
and *r be considered as an affix having a latent initial vowel* then the ^T of
*IP<£ requires to be dropped. But this sGtra prevents this and we have the form

*rarer; as, *fj ^nrrnr: ircfar ft^rnL 11

4* Wnrftfa: ' a rule relating to the dropping of yaJ Thus there is
a verb jh u ^3 'to scratch'. In forming an abstract noun from it, the affix ft
(tech. f^r%) is added. Thus 9Ff* +fir»^ , fl+*+fti the *T of ^T being
dropped by VI. 4. 48. Now there is a rule (VI. 1. 66) already mentioned above
which requires the elision of 3 before consonantal affixes. Now if the lopa-
substitute be considered as stBnivat to *T, then the affix f?r is not an affix
having an initial consonant and the rule, which requires the dropping of jr f
would not apply. The present sfitra, however, provides for this, and we have
the form *«$%:« scratching/ • Digitized by Google

50 Substitute is like the principal. [3k. I* Ch. U § 58.

5^ FTCftfa: 'a rule relating to accent 1 . There is a sutra (VI. 1. 193)
which says that when a word is formed by an affix that has an indicatory
*£, the vowel that immediately precedes such an affix has the ud&tta accent.
Now *$% is an affix that has an indicatory fj, the real affix being jggjr
(VII. iT"i.). It is an affix used in forming nouns of agency. Adding
this affix to the desiderative verb fasftfr € desirous of doing 9 we have: —
fazfift + «W - ft«tft$, + » + *H?» the *f being dropped by VI. 4. 48.
Now if the zero be considered as sth&nivat, then the accent must fall on this
latent <? but it is not so. The accent falls on the f" of 3ft; and we have

^fnNr: , so also fefT4^ a

6. *3N*3ftfa: c a rule relating to the substitution of homogeneous
letters.' To explain this, let us form the second person singular of the verb
ftr% in the imperative mood. The affix ft is the sign of the second person
imperative singular; and the verb ftp^ belongs to the seventh-class of verbs
called RudMdi. This class of verbs take the vikarana gr (H I. 1. 78) the q
and ^ being indicatory the real affix is 5f. This ^ is inserted between the
radical vowel and the final consonant (I. 1. 47). Therefore we have the follow-
ing equation:—^ + rr^+ft-ftf +T +* +ft~ftf+^ + o +^r + ft. The
*T of if is dropped by VI. 4. in, which declares that the ^ of *^[ is dropped
before those s&rvadhitu ka terminations which aregt^orfe*. By I. 2. 4
all sirvadhfituka terminations which have not got an indicatory q are consi-
dered as ftj The affix ft therefore is a ft^ sirvadhfttuka affix'and by rule
IIL 4- 8 7 > ft >s to be regarded as srf^, not having an indicatory <j. The next
stage is :— fa+^+o+^+fa; the ftbeing changed into fir by rule (VI. 4. 101)
which says that 'ft is changed into ft, when' it follows a root -which ends in
a consonant of 55 pratyihara:' here ^ and ^ are changed by general rules
of sandhi into 5 and ^ and we have ftr+^+o+^+ft (VI IL 4. 41).
Then there is a rule (VIII. 3. 24), which says that in the body of a word the
sj is changed into anusvara when followed by a consonant of 5^ praty&hira.'
Therefore we have :-ftr + ♦*^T?^ + ft. This is the form to which the
present rule is applicable. Thus, there is a rule (VIII. 4. 58) which says that an
anusvira followed by a letter of ^ pratyihftra i>., any consonant except f, ^
^ and ^ is-cKanged into a letter homogeneous or savarna to the letter that
follows. Therefore, if we regard zero as sth&nivat to the *r which it replaces,
the anusv£ra cannot be changed into 5 as homogeneous to 5. But IT is' not to
be so regarded, and we have the form ftf + ^ + ^ + ft- frPf or ftrfoj
so also ft?*T.

7- WS^rcftft: 'a rule relating to anusvara.' The above example
will serve the purpose here:— ft + ^ +. • + ^ + ft. This by the rule al-
ready mentioned (VIII. 3. 24), requires the 5 to be changed into anusvara.

Bjc. I. Ch. I. § 58. ] Substitute is like the principal. 51

\ Now if zero be considered as sthinivat, then g^is not followed by a letter of fft^
pratyahira , and cannot be changed into anusv&ra. However, the zero is
■J I not regarded as sthinivat, and we change the 5 into an us vara. Other
; examples areftr + % + • + % + Tfcf - ffr^F* ; so also ffafSv ||

8.- tf^ftnfc f a rule directing the substitution of long vowels for
short vowels/ Let us form the Inst. sing, of the noun *tf%fft^. The ter-
mination of the Inst sing, is Zf or ^n*. Therefore, we have J?f%ft^ +sn«Jri%f^
+ *+gr+*rr. Here the *r has been elided by rule VI. 4. 134, which teaches
that words ending in *r% lose their *[ before terminations technically called
Bha. The termination of Inst sing, is a Bha termination by force of
rule I. 4. 18. The next rule that now comes into operation, is VIII. 2. 77, which
requires the lengthening of the short penultimate vowels, of the {3^ pratyi-
hara, of words that end in ^ or ^ and are followed by a consonant Here
f$\ is a root that ends in \ ; and if the zero be considered as sthinivat, then
this <j[ is not followed by a consonant, but by a latent vowel, and so the short f
of f^ will not be lengthened. This rule, however, provides for such length-
ening, and therefore we have :— Rftrft^ + • + % + *jrr - Ufitffar II

9. *r*3ftft: '& rule requiring the substitution of soft unaspirate
consonants, of pratyihira ?ng, in the place of hard consonants ' In this case
also the idc£a that takes the place of a vowel is not sthinivat

Let us add the affix ftr^ to the root *JJ 'to eat/ compounded with the
word sr or *PfTT. In Vedic literature *r| is replaced by *rg, (II. 4. 39); so we
have, ^+^+f^«sr + tw+fir«^+*{+«+^+f?r (VI. 4. 100)-*+*+%+

•+fir(VHi. 2. 26)- *+^+*+«»+fa(Viii. 2. 40).

It is at this stage, that rule VIII. 4. 53 requires the q^ to be changed
intoir. This is a tst^ rule, and teaches that in the place of letters of 5^
class, the letters of *r*( class are substituted, if followed by letters of fr*J class.
If the zero which was obtained by dropping the si of * t be considered as
sthinivat, then the * is not followed by a consonant of jg^ class, and cannot
be changed into *J. But it is not so. Therefore ^ + • + • +fa "^P**" 1 O^aj. Ved.
XVIII. 9). The various rules, by which first *T is omitted, and then jr is
dropped, have been referred to in brackets, and will be explained in their proper
places.

Similarly the word waft^ , which is the Imperative (lot) 2nd P* r *
dual of the root H^ , and is thus evolved. This root belongs to the Juhutyidi
class, and therefore, there is reduplication. Thus H5+^Rt**^5 + *? + IRt ■
T + H3 + ST5> ¥+H^r + fTT5.-¥ + * + « + ^ + *l% (VI. 4- 100) - * +
*+•+•+ m^(VHI. 2. 26)-* + * + • + ♦ + *n*(Vlll. 2. 40) - w + *

+ •• + .• + m^" 3 *9Jir*£ll Here the change of * into q^ is by the applica-
tion of VIII. 4. 53, which is a ja£ rule, and therefore, the zero is not sthinivat

52 Substitute is like the principal. * [Bk. L Ch. I. § 59-

* 16. ^Tprf^ ' a rule requiring the substitution of the letters of
^pratyih&ra instead of other consonants.' In the application of this rule,
also the substitute that replaces a vowel is not sth&nivat

Let us form the 2nd pers. sing, perfect of the verb m\ 'to eat. 1 *J\

•is replaced by q^r (II. 4. 40) and *rg*r is the termination of the second person

singular of lit or perfect tense. *3 + *rg?[ - *% + -*5 + *f3S ( VI#

1. 8) - * + *H + ST^UVII. 4. 60) - *r + *5 + ^3W(VIII. 4- 54) - * +

5 + % + ^ + ^35 (VI.4. 9 8 ) - ^ + • + *$^-

At this stage comes the ^ rule in operation, which says that the
letters of *R[ praty&hira are replaced by those of ^£ pratyihira, when foHow-
ed by letters of ^ praty&h&ra. Here % is a letter of fl^ class, and if zero
is not sthinivat, it is followed by 5 which is a letter of ^ class, and there-
fore 3 is required to be changed into gr of ^ class. By the present rule, zero
is not sthinivat and thus we have : — ^sra^^^gH-^fffij: (VIII. 3. 60).

Similarly *reT3. the 3rd per. plural aorist of ^ is thus formed:— *j +

i3+i^ + *nt-*r + *g + *FZ ( IL * 8o) " •* + * +%+ S + *ra h IVI.

4.98). Here had the lopa been sth&nivat, rule VIIL 4. 55 could not have
applied. But it being a char rule, the lopa is not sth&nivat- Therefore we

have,*j^T5^

When lopa is the substitute of a vowel, it is not sth&nivat, for the
purposes of the application of the rules of accent, rules relating to the doubl-
ing of letters, and the rules relating to the elision of ya.. In other places,
with the exception of the above three ; the lop.a substitute of a vowel is, and
must be treated as sth&nivat Thus ^NJT^T: * ftifll: » ftr^f: , ^P^fc • In these
cases the lopa-&de£a being sth&nivat, the rules relating to accent, lengthen-
ing and the elision of ya, do not apply.

sn^i: A arc* ) »

59. Before an affix having an initial vowel,
which causes reduplication, the substitute which takes
the place of a vowel is like the original vowel even in
form, only for the purposes of reduplication and no
further, •

This sutra has been explained in different ways by the authors of the
Kasikaand of the Siddh&nta Kaumudi. According to the latter, the sutra
means : — 'when an affix beginning with a vowel follows, that is a cause of re-
duplication, a substitute shall not take the place of a preceding vowel, {whilst
the reduplication is yet to be made : but the reduplication having been made,

J

Bk. I. Cri. L § 59. ] Substitute is like the principal. - 53

the substitution may then take place/ The explanation given by Easika
is more in harmony with the Great Commentary. For all practical purposes,
the two explanations lead to the same conclusion. According to Kasika,
the vowel-substitute is sth&nivat, in the sense, that it is exactly of the same
form as the original, and retains this form only for a fixed time vis. f so long
as the reduplication is being made ; but as soon as the reduplication has been
made, the substitute takes its proper form. This rule has its scope in the
rules relating (a) to the elision of long *JT, (b) to the elision of the penultimate,
(c) to the elision of the affix fij^, (d) to the substitution of semi-vowels for
vowels, and lastly (e) to the substitution of *pr, *p^ *THf and «n? for ^ ^fjr
tj and s?r respectively by the rules of sandhi.

(a). Elision of *JT. Let us form the 2nd person perfect tense of
the root qr ' to drink/ The termination of 2nd person perfect is qTJRf. This
is, therefore, a termination beginning with a vowel, and it causes the redupli-
cation of the root. The conditions of the present sutra all exist in this. *
Thus we have qr + «TJ? - <J + * + *T3? (VL 4, 64). Here the *n is elided
by VI. 4. 64. which teaches that ' the *jt of the root is elided before &rdha-
dh&tuka affixes beginning with a vowel and which are fervor ft^.and before
^ / The affix m^ is such an affix (I. 2. 5, III. 4. 1 15).

Now comes the rule about reduplication which is contained in sutra
VI. 1. 8, and teaches that a root consisting of a single vowel is reduplicated
before the terminations of the perfect. Here, by the elision of sjt, the only
visible root left to us is 5[ which is a consonant without any vowel. The rule
of reduplication, therefore, would not have applied to it, because there is no
vowel in it, but for the present sutra, which solves the difficulty* Here the
vowel-substitute zero must be considered sth&nivat i. e. as if it were the very
*TT itself. Thus we have qr + SL + •+ *T35 ■ THRJ (VII. 4. 59). The long
^TT of the reduplicative syllable has been shortened by VII. 4. 59. So also qj:|

(6). The elision of the penultimate. Let us form the same tense of
S^'to kill/ f^ + WS^ - £ + % + * + WJ^(VI. 4. 98).

. Here also the preliminary conditions are all fulfilled, the affix begins
with a vowel and it causes the reduplication of the root. The penultimate
^ of ^T is replaced by the substitute lopa by VI. 4. 98, which teaches that the
rootsij^'togo'f^'to kill* and a few others lose their penultimate before
affixes beginning with a vowel and which are also f^ or few. We know
^55 to be such an affix. For the reasons already given, we can not redupli-
cate the consonants 5^ as they have no voweL However, this reduplication
is effected by virtue of the present rule^ and we have : — f + 1" + *T3^ -
siifij: (VII.. 4. 62,.VII. 3. 55). The first £ is changed into 9 by VII. 4.62
and the second £ into «£ by VII. 3. 55. ... ^

Digitized by VjOOQLC

54 Substitute is like the principal. [Bk. 1. Ch. L § 5$.

(c). The elision of the affix ftj^ . Let us form the Aorist (lun) of the
causative form of qsnf. The causative is formed by adding the affix ftrg[ to
the root ; and the aorist takes the augment *[. Thus we have s— *r + ^
+RSt^ +^[ s -sn? +f +SC~*rr5+T+^T5;+^. The affix ^ (HI. 1. 48) is added
in forming the aorist of causatives. The letters ^ and 5 are indicatory,
the real affix being sf. Now by VI. 4. 51, the affix fop^ is elided before an
Ardhadh&tuka affix that does not take the intermediate f . Now n^ is such
an affix. Therefore we have 5Sn^ + • + *nj.

At this stage, comes into operation another rule (VI. x. 11), which de-
clares that 'the root is reduplicated before the affix ^C k .' The ^raf is, thus,
an affix which causes reduplication; it is an affix which also begins with a
vowel. Therefore, it satisfies the conditions of this sutra. Therefore, the
zero which was a substitute in place of the vowel f , becomes sth&nivat
to f 9 as if it were the very \ itself. Thus we have the reduplicative syllable
f>, by rule VI. 1. 2, which says that a verb beginning with a vowel reduplicates
its second syllable. But as soon as the reduplication is over, the &de£a takes
up its original form of zero. Thus *±\\\l\ + • + *JT£ = mifidflN

(d). The fourth class of cases, where this sutra has scope, is the
substitution of inj^ for vowels. Thus let us form the perfect 2nd person
singular of fr 'to do' : — fr + *I$f*[ « a^ + STJ^, the semi-vowel £ taking
the place of 3T by the general rule of sandhi VI. 1. 77. Now we have already
learned that *HRl ' s an an ^ x which causes reduplication and it begins with
a vowel. Now £ is a substitute which comes in the place of a vowel (1. e^ 3f),
it will therefore be sth&nivat by this sutra. Had it not been sth&nivat,
the consonants ar could not be reduplicated, as they have no vowel The
substitute £ being taken equal to ^r, we reduplicate fr ; in reduplication the
g* is changed into^by VII. 4. 62, and 3J into si by VII. 4. 66. Thus we
have : — ^T^GJ.' ; similarly ^p$: I

(*). . The fifth case is the substitution of *r^r &c. Let us form
1st person singular of the perfect of sft 'to lead:' 5ft + 1575 " ^ + V ™ *tt% +
^T. The f isvriddhied before the termination si of the perfect (VII. 2. 115),
and then changed into <*JT3T by the general rule of sandhi (VI. 1. 78). Now the
*T of the perfect causes reduplication, and it is an affix which begins with
a vowel It satisfies the conditions of this sutra* Therefore the «n^ which
was a substitute for i£, being a vowel substitute, becomes sthinivat to ij in form.
In reduplicating the word m* 9 we reduplicate as if it were still %. Thus we
have the form f^RPT and not «4«ii4. Similarly Ph*|, 5J5TT, and igstr?. ' N »

Why have we used the word " dvirvachane" in the sutra? Without it,
the aphorism would have stood thus: — " Before an affix having an initial vowel
which causes reduplication, the substitute which takes the place of a vowel

BK. I. CH. I. § 60.] . . LOPA DEFINED. 55

is like the original vowel." Let us form if^ the 3rd per. sing, perfect tense
(fv^r) of the root *& " to be exhausted," in the atmanepada. The affix of 3rd
per. sing. atm. of lit, is ip^(III. 4, 81). Thus tit + l&t - *3TT + ^[ (VI. 1.
45). Now here t* is changed into *rr, on account of the affix ^tj; this affix
causes reduplication, but still SJT is not to be considered as sth&nivat to ^.
For if ^jjt be considered so, then in the next stage »w+ 53^ ■■ *5^ + % + ^
(VI. 4. 64) the Iopa will be regarded as sth&nivat to it and the reduplicative
syllable will be fir and not sr ; the form being fo**r and not 3P5T. But *JT is
not sth&nivat to l£, because for the purposes of reduplication, the change of
*r into *JT is immaterial ; it is the second and further change of *jt into zero
which is directly connected with reduplication, (for if zero be not considered
sth&nivat, there can be no reduplication) ; and therefore, this substitute zero
should be regarded as sth&nivat to *JT #

Why do we say "which causes reduplication?" Without these
words the sfitra would have stood thus : — " Before an affix having an initial
vowel, the substitute which takes the place of a vowel is like the original
vowel, for the purposes of reduplication." In that case we could not get
the form J^f^r, 3rd per. sing, desiderative present tense of ftjr ' to shine, to
play/ This form is thus evolved:— f^+^(III. 1. 7)-ft+3& + ^(VI. 4. 19)
=»|^+3T+^ (VI. 1. 77). Here, the change of \ into ^ is caused by ^ (tech.
3T^), which begins with a vowel ; is this ^r to be regarded as sth&nivat to f ? No,
because 3T does not cause reduplication, it is W£ which causes reduplication.
Therefore the equation is : — *£+ ^ - J*£ET (VI. 1. 9) ^jrapifiir. If w could be
regarded like f , then the form would .have been f$l£n%.

Why do we say "beginning with a vowel"? Otherwise the sfitra
would have run thus : — " Before an affix which causes reduplication, the subs*
titute which takes the place of a vowel, is like the original vowel, for the
purposes of reduplication." The affix if& of the Intensive verbs causes re-
duplication (VI. 1. 9); but as it does not begin with a vowel, the substitute
will not be sthftnivat. Let us form $rffcr% the 3rd per. sing. Intensive of
the root 5TT/to smell.' Thus HT+*y=fft + ^ (VII. 4. 3i)=*rafa*(VIL 4.
82). Here, w is changed into f" on account of ^ ; but f" is not sth&nivat
to *JT; had it been so, the form would have been *ftfta?T. Similarly {ufcft.

^wjf>rtWfr *rer sfrr **fW *%t qrafir u

60. The substitution of a blank (lopa) signi-
fies disappearance.

This defines elision. When a letter or word-form becomes latent,

56 : Pratyaya-lakshana. [Bk. L Ch. L § 61, 62.

is neither heard, nor pronounced, nor written, it becomes lopa or is said to be' -
elided. Lopa is the term for the disappearance of anything previously
apparent. • :

In Sanskrit Grammar, this " lopa" is considered as a substitute or idem, and as such this
grammatical zero has all the rights and liabilities of the thing which it replaces. This blank or lopa
is in several places treated as having a real existence and rules are made applicable to it, in the same
way as to any ordinary substitute that has an apparent form. The Grammarians do not content them-
selves with one sort of blank, but have invented several others ; there are many kinds of them, such
as lopa blank, rlu blank, hip blank, and luk blank, which like different sorts of zeroes of a Mathema-
tician, have different functions.

The word lopa occurs in sutras VI. 1. 66 and VI. 4. 118 &c.
The lopa substitute is a sense substitute, and not a form substitute. Thus
when we say ' let the substitute lopa take the place of such and such a
letter or word/ we do not mean that the letters ft, %jt f <£and *r should be
substituted there, but the sense of the thing, namely ' disappearance. 9

( sr^afcn^) 11

61. The disappearance of an affix when it is
caused by the words luk, ilu or lup are designated by
those terms respectively.

When an affix is elided by using the term luk, that disappearance
gets the appellation of luk, similarly when an affix is elided by using
the terms slu or lup, the disappearance gets the name of Slu or lup.

Thus in *r^+ JT^tf^T + f% - srf^T ' he eats.' Here the vikarana qrji has
been elided by using the word luk (II. 4. 72). Similarly in sjffrft 'he in*,
yokes/ the vikarana jrj is elided by the word Slu (II. 4. 75), so in qr*UT: ' a
city in the vicinity of weeds called varani/ . Here the Taddhita affix indicat-
ing vicinity has been elided by the word lup # (IV. 2. 82),

Why do we use the words " of the affix "? Without these words, the sfttra would hare
run thus: — "The disappearance is called luk, jIu or lup." So that the disappearance of a bos*
would also be caused by the use of these terms. Thus in II. 4. 70, there would have been * luk* of
the bases agastya, and kaum/inya ; which however is not intended. In that sutra, the * luk ' of the
affixes is only intended. So also in sutra IV. 3. 168 there would have been the substitution of 'luk 9
for the bases of the words.

62. When elision of an affix has taken place
(lopa), the affix still exerts its influence, and the opera-
tions dependant upon it, take place as if it were, present.

Bk, I. Ch. I. § 62.] Pratyaya-lakshana. 87

This stitra declares that when an affix disappears by lopa-elision its
potency does not disappear. Thus the term 'pada' is defined to be that
which ends in a case-affix or a con j ligation al-affix. But after certain bases*
these affixes vanish; still those bases become 'pada' in spite of the
vanishing of the whole affix. How can an operation directed by an affix
take place, when the affix itself vanishes, is a doubt which may naturally
arise. This sGtra answers the question. Though an affix may be elided by
lopa, yet an operation of which such an affix was the cause, must take place.
Thus Vjfafil^ is a pada though the case-termination has been elided. Simi-
larly the word *rW3£ ' he milked;' which is the Imperfect (5Hf ) of JR, and is thus
formed, *rtrj+fr*-*re*$+*(III. 4. ioo)«*mjr + % (VI. i. 68)«*ri*?+*
(VIII. 2. 32) (here f is changed into jr by taking the word 'aduh' as a 'pada')
*nft^+« (VIII. 2. 37)=*Nir^ (VIII. 2. 39) "■*!**£:

Why have we used the word smnr in this aphorism again, when its
anuvptti was understood in this sfitra from the last ? This repet? f ion indi-
cates that the present rule applies where the whole of the affix is elided, and
not when a portion of an affix is elided. The present sfttra, therefore, does
not apply to forms like HJltflq and q^*fl«q . These forms are the 1st person
sing. f*T£ Atmanepada of f^andir^. Thus MjlfdL + ^ - *m% + «r(IH. 4.
I06) « *n*l .+ tffy^ + *r (IH. 4. 102) « *wfi +t«[+*r. Here, ^a portion
of the affix *fr^ is elided by VII. 2. 79 ; now there will be no U*34frl\iUI of
this partial elision of an affix. Had it been so, the final nasal 5 of *nt*f
would have been elided by VI. 4. 37 : but that is not so : and we have inrffrr
(VII.3.54and VI. 4.98).

The word CT^TOJ means that by which a thing is recognised. The
word i|U|4U\mJ, therefore, means the effect by which an affix could be recog-
nised. The orthographical changes, like the following, are not however
pratyaya-lakshana. In forming the genitive singular of ^ 'wealth/ we have
^ + «T^ " *TO: f ke rc * the change of I* into %ff^ by the rules of sandhi,
though caused by the affix *rg; is not such an effect, which the affix causes,
by virtue of being an affix, but it is an effect caused by the accidental fact,
that the affix begins with a vowel, and thus gives scope to the euphonic rule
VI. 1. 78. Therefore, when in forming the genitive compound of n*: +
sj*t, we elide the case-affix ^H, we get the form frg^Pfr Here the elided affix
will not cause the peculiar orthographic change mentioned above, which is
dependent upon its letters, and not upon the affix as such.

f $*td!#«« » M 11 *V$* 11 f, §*ran, w^r, ( H5*m-

wH*^) II

68 Pratyaya-lakShana. [Bk. I. Ch. I. § 63.

63, Of the base" (anga), whose, affix has been
elided by the use of any of the three words containing
^ , the operations dependent on it: do not take place-dre* ^
gardrpg^ sjiclr bas f e. * jc &-;&**

This sfitra is a special exception to the previous aphorism which
was too extensive. It declares, that when an affix is elided by using the
words jyqr , ^5 , or jf^ , then the ' sign ' or 5TOTD or the pecu liar virtue of the
affix also vanishes. Thus there is an important difference in the elision of art
affix by force of the words 5?^ &c. ; and by the word 5?nT # In one case the
virtue of the affix subsists inspite of the elision, in the other it does not

The word *j^F in this sfitra requires some explanation. " A root,
followed by a suffix (pratyaya), is raised to the dignity of a base (pritipadika^
and finally becomes a real word (pada) when it is finished by receiving a
case-termination (vibhakti). Every base, with regard to the suffix which is
attached to it, is called Anga, body."

Thus the word *pfc is the plural of the word npfr : . The word «nHr:
is formed by adding the affix zpj to ipi* (IV. 1. 105) ; the force of the indica-
tory sj being to cause the vpddhi of the first vowel of the word (VII. 2. 117),
But in forming the plural of *TF^, rule II. 4. 64 presents itself, which declares
that there is luk of the affix *rsr when the word takes the plural. Here the
affix 3p{ is elided by the word luk, which means not only the elision of the
*T of *ir*A but also the shortening of «qT into W, that is, the affix being elided,
it vanishes altogether, leaving no trace behind, in the shape of the.vfiddhing
of the first vowel, which the affix was competent to cause.

Similarly g?: is the 3rd per. dual present tense of 9*r ' to clean. 9 . It
belongs to Ad&di class in which the vikarana 7J% is elided by ' luk ' (II. 4. 72);
and hence there is no vpddhi substitution, which would have taken place by
VII. 2. 1 14, (the ik of the base 3^ is vpddhied before verbal affixes).

Similarly *J£<r: ' they two sacrifice 9 ; here the vikarana jj«r -has been
elided by using the word ' slu ' (II. 4. 75, let £lu be substituted in the room
of Sap after the verbs hu &c). The effect of this elision is y that there is no
guna substitution of the vowel of the base, which jj^ was competent to cause.

When an affix is elided in any other way than by the enunciation*
of the three terms containing *j, it is competent to cause its proper effect.
Thus ^+fij^+^^+^«g^+it^+%=grr^+^5+&-«CT^. Here the elision
of the affix nich is by ' lopa/ (VI. 4. 51) and hence though the affix vanishes,
its effect remains. Why do we say ' in respect of an Anga or inflective
base?' Observe q*j 'five/ mt 'seven/ <pr: 'milk/ qjH 's&ma veda/
Here in the word paiicha, the plural case-affix is elided by VII. I. 22

Bk. I Ch, I, § 64, .66. ] Tl DEFINED. 59

l> U I I ' U ■

(after the words called sha|, the *r*£ and jj^ are elided by 'luk').
Here the affix is elided by enouncing the word 'luk,' the result is,
that the affix cannot produce any change which it would otherwise, have
caused, in the anga or inflective base; such as lengthening, as in r&jan, 'a
king/ x. p. r&j&najp 'kings.' But though it is inoperative as regards the
anga 9 the elided affix is not altogether inactive. It retains the virtue of rais-
ing the base to the dignity of a pada. Thus: — f fa3 + sn|- f f^2+*, (luk VII.
1. 22). The word pafichan is now *pada % and as such, it drops its final £.

3T^w*?nf^fe 11 qi 11 m^iPc 11 sro, sre*nf^, ft 11

64. The final portion of a word, "beginning with
the last among the vowels in the word, is called ft .

This sfitra defines fi. It is that portion of a word which is included
'between the last letter and the nearest vowel. Thus in the word Mjfnft^l the
portion f^ is f>; so also in writhe portion 7<r. The word f> . occurs in
sfitra III. 4. 79.

The word *pj: in the sfitra is in the genitive case, which has the
force of nirdh&rana or specification ; and though it is in the singular number,
it represents class name, meaning ' among the vowels/

3T% UT C mcMJ i ^preT 11 vi m ^iPi u «ra: # «wm^

^ $ 3TOT II

65. The letter immediately preceding the
last letter of a word is called the penultimate.-

This defines 3TOT or penultimate. Thus in the words q^j, p^ , j>r, f^

the letters ^r , f, ?, and ^7 are 9T^T.

The word ' upadh&' occurs in sfitra VII. 2. 1 16.

1%: 11 trfl*rf%Tfir *m*4*4Pft% ^few 5*nl H^fir •Tituw 11

66. When a term is exhibited in the seventh
case in these s&tras, the operation directed, is to be
understood as affecting the state of what immediately

• precedes that which the term denotes*

This is a declaratory or interpreting sfitra. When in a sfitra, a word
is used in the locative case, the operation ordained by that sfitra is to be per-

60 The force op Ablative case. [ Bk. I. Of. I.. § 67, 68.

formed on. the word preceding it. Thus in the sfitra f«Rfr stoPt (VI. i. 77) the
word *rf% is in the locative case, which sfttra, therefore, means that f; *T, gj tif
followed by a vowel (*r^) are changed into 5 , *£, }T, 5T, (*TT[). The literal
translation of the sfitra being: — c Of ^5^ there is if* in *nr/ The force of
9 in f or the locative case being to induce changes in the letter preceding the
vowel. Thus *fa + 5^7^ - ^5^P9[ ; so Hfcwp% , TF*t&\ U

The word nirdishfa (meaning ' exhibited ') has been used in the
sfttra to indicate immediateness. Thus in q i fafNd , the f of *jfir is not direct-
ly followed by the f of fif£, there being the letter^ intervening! therefore,
rule VI. 1. 77, does not apply here : and we have no sandhi.

sftfc 11 srerrftflr m^mtiPi^i ^rcwr mr% *rrfir t j^w 11

67. An operation caused by the exhibition of
a term in the ablative or fifth case, is to be understood
to enjoin the substitution of something in the room of
that which immediately follows the. word denoted by
the term.

This sfitra explains the force of words in the ablative case when
employed in these aphorisms. As the locative case refers to a preceding
word, so the ablative refers to a succeeding word.

Thus stitra VIII. 4. 61 declares that f after sj, of sthi and stambha,
the substitute is a letter belonging to the class of the prior. 1 Thus 7f +^IHH*
Here the word ??: is exhibited in the original sfitra in the ablative case, and
by the present sfitra it means that a dental letter 5 must be substituted in
the place of sth&na, and by I. 1. 54 ante, this dental takes the place of 5; we
have ^Tf + WTR^^^OT-trpr^. Similarly in sfitra ft*3£ *ft£: (VIII.2. 18, ' a word
ending in a conjugational affix, when following after a word ending in a non-
con jugational affix, becomes anud&tta.') Here the word 'atifiab' is exhibited in
the ablative case, therefore, the operation directed by this aphorism, namely,
the substitution of anudAtta accent, must take place in the word that follows.
Thus ?5f5r <fafH" ; but not so in 4^*4)f«ifVtf) II

^ ^T vc^traatfvi n \c 11 xt^fk 11 ^^, *ww ,

?f%n 11 ot# &h* *ft u*&f infr siui*mi*»4 H^fit i mfitfa v^^l Tfr-
(Vn 11

^rf%^ it * f^tTft^qrot fw«rJ^ * 11

\\

»/'

J

BK. I. CH. I. § 68. ] SVAM-RUPA RULE* 61

qrft*n ii * f^Mfidsiuji^ qwu<i3i * ii

68. In this Grammar, when an operation is
directed with regard to a word, the individual form of the
word possessing meaning is to be understood, except
with regard to a word which is a definition*

This is a rule of interpretation. The word * V which means ' one'^
own' denotes 'the meaning/ and the word CT denotes 'the individual form
of a word.' The sense of the sfttra is that a word denotes both things ex-
pressed by those two words Mr and ct, vis., its meaning, and its form. Thus a
rule applicable to *rfifr will be applicable to the word composed of s%, 5, % , f
and not to the words synonymous with agni, such as, TnfftT:, «q«Fr: &c.

In this science, therefore, the very word-form is to be taken, and
understood and operated upon, and not external sense : with the exception
of words which are definitions, in the latter case the things defined are to be
taken;

This sfitra is begun in order to remove the difficulty which might
otherwise arise from a consideration that a word conn otes a certain meaning,
and as no grammatical operation is possible in a concrete object, we must
take the words having the same meaning. As when a person is called to
bring a 'cow/ he brings the animal cow, and not the letters c, o, w. It may
therefore be said, that as no grammatical operation is possible in a material
thing, we may take all the names synonymous with the word and operate
upon them ; this s&tra removes that doubt. ~

Thus the sfttra *p%£?£(IV. 2. 33) declares :—" that the affix f $hak r
comes after the word 'agniV Thus *jfar + **£ « W%*Rl "belonging to
agni," as in the following sentence H/lH€|Hg|<fc4M fwht • Here the word
%%f*H indicates the individual word-form HJIH and not its synonyms. Simi-
larly sAtra IV. 2. 19 declares : — ' let there be optionally ss£ after Affair / / As
*?lrcP*Wl(VII. 3. 50 and 51) or qjfaftrirqt* The rule applies to the word-form
'uda^vita* meaning 'curd-milk' and not to tmr, HjR?, qrrvHTT, **ST?*f and
trftft* which are synonyms.

There is an important exception to this rule, and that is in favour
of terms which are technical terms of Grammar and have been so defined.

Thus vnfttTO? (!• *• 2 °) an< * ^^^ Hi (I. »• 22). Here the
words * 5 9 and ' *r ' denote the words connoted by those terms and not to
their word-form. There are, however, some s&tras in P&nini which are aj>-

Digitized by VjOOQLC

62 SVAM-RUPA RULE. [ BK. I. CH. 1. § 69. .

parently exceptions to this rule. These exceptions have been arranged
under four heads, technically called f%?r, fir^, finr, and Rf7*r. When there is
a rule about a word, it may mean that very word-form, or words represent-
ing species falling under that word, or words which are synonymous with it.
Thus rule II. 4. 12 declares,: — "j^r 'tree' 3*r 'deer,' giST 'grass/ &c. f
may optionally be compounded as dvandva." Here the word f^T stands
as a generic name for the species falling under it, and does not represent
the word-form, f^r. Thus we have dvandva-compound of jffqr + "Pltar, as
g^T?*pjfrPt or fffTTtpfrrr. I Such sfttras are called f%R(.

Then in s&tra HI. 4. 40, we have an example of fir?£. That s&tra means
"the word yq* 'to feed 1 takes the affix mjj<H (*JTO when compounded with
the word **." Here the word ' w' means both the word-form f^, as well as
the synonyms of W and the species of *r. Thus ^? and *FT 'wealth' are
synonyms; while ^ ' wealth' and if} f cow* are genus and species. Thus

w + yj + *uy* - srrfa^. Similarly rffa*, *mta* , *ftfm, «mre?.

Such sfttras are called f^ .

Then in sfttra II. 4. 23, we have an example of Rft^. " A Tatpurusha
compound ending with the word ^rr # ' court ' is neuter, provided that it is
preceded by the word r&jft, or by a word denoting a non-human being."
Here the synonyms only of the word ' r&jA ' are taken, and neither the word-
form 'r4j4,' nor the special instances falling under it. Thus fr'king' +
^TT-T'WH^ ' king's assembly,' f^T+^RT-f^TOH^ 'king's assembly.' But
not rnr^PT^, but rPTWr. Nor can we take particular kings, such as ^Pjnpr
or y?Pnr. In their case we must have J^lfinreHT, *4»A{IU4I4I«

Then in sfttra IV. 4. 35, we have an example of f^j. The sfttra
means " the words qrf^T ' bird ' Ht&f ' fish ' s*r ' deer,' take the affix s^£ when the
meaning is the killer thereof." Here the words q% &c. denote both the
.word-forms q% &c. as well as the species denoted by those terms, but not
their synonyms. Thus <tffr+saE.= 3 <nf^T3T: 'a bird-killer/ mftwfr :, 'a fish-
.catcher.' JCT^PT+^-^rr^nw. ' a vulture-killer.' But we cannot take the
synonyms of '<rfitT,' &c. except in the case of ' h*rt/ when we may take the
word 4 *fhr.' As *nRW9T: and Sh%^T: 'a fish killer.'

mmfow^fa ^nictra: II \\ n t^rf* II STO , *f^W,

Wta, % 3URTO: ( ^W ) U

69. The letters of the PratyAhAra *rz i. e. the
vowels and semi-vowels, and a term having * for its indi-
catory letter, refer to their own form as well as to their

Bk. I.'Ch. I. § 7o.] Ta-para rule. 68

homogeneous letters, except when they are used as pra-
tydy&s or affixea

The pratylh&ra *T^in thus sfttra includes all the vowels and liquid*.
This is the only instance where BTJ pratyih&ra refers to the second ^ erf
5TT. The letters included in *r^, and the letters having an indicatory*,
refer to, and include their homogeneous letters also. Thus sj refers to all
the five gutturals *r, *r, *T, *, <F, similarly ^, 5, y, and j, as in sfltra 1. 3. 8.
Similarly *r includes short, long and protracted *f. The words 'its own
form ' are understood in this sfttra, being drawn from the previous sfitra:
Thus sfttra *ji^uj: (VI. 1. 87), ' when a vowel comes after *f, guna is the
single substitute for both.' Here though the rule mentions only short *r, we
take the longer also. Thus not only 5TT+fxr:~H*lj:; but **T+fa:»»rifcrt I
So also in sfitra s^T ^ (VII. 4. 32), ' there is long f in the room of *r when
the affix c chvi'(V. 4. 50) follows/ Here long *rr is also included; thus

^nr+^nftRr » *uif|*^fir, and ^}r+^rnf^«fjr^qrrrfir. So also in sfitra

**$% ^T (VI. 4. 148). « When long f follows, there is elision of the f or %f
of the base/ the rule applies to the bases ending in long f and *rr also. >

This rule of course does not apply to affixes. Thus sfitra III. 2. 168;
says, ' after words ending in the affix san, and the words 4£ansa and bhiksha,
there is *.' Here * is an affix and therefore does not include long 3?.
Thus 5nj: ' a beggar.' So also in IV. 3. 9, the short «if only is to be taken
and not the long one.

dM<«c*l»« U too II Ti?r T fSf u H-TTC5, a^-lJTCRT, ( **
^1* ) «

>rcrfir ****<* *ro* 11

70. The letter which has 5 after or before it,
besides referring to its own form, refers to those homo-
geneous letters which have the same prosodial length
or time.

By the previous sfitra it was declared that a vowel standing by it-
self included all the letters of its class. Thus *f includes *rr; and f, f; &c.
This sfitra lays down the rule by which the very form of the letter is taken
and not all the letters of its class. This is done by placing a £ either after
or before.the letter. Thus 3^ means the very letter *j and not all its homo-
geneous letters. Similarly ^ means short 7 only and not long or protracted
7. We had to refer to this sfttra, in our very first sfitra, in explaining the
term *n* Digitized by Q 00 , ^{ e .

$4 Tadanta-rule. [Bk. I. Ch. I. $ 7*> J*.

The sfltra consists of two words IPTC and fTOiTvTCl. Taparafc means
that which has a £ after it or that which is after g. Tat-k&la means having
the same time. According to time the vowels are divided into short, long
and protracted. Short vowels have one measure, long vowels two, and
protracted vowels three. A consonant lasts half the time of a short voweL
Therefore a letter preceded or followed by 5, refers to its own form as
well as to those only of its homogeneous letters which have the same time.
Thus the letter *T*)L W *U include the udatta, anudatta, svarita (both nasalised
and non -nasalised) *? and not the long and protracted forms.

This sutra declares an injunction. The anuvptti of W£ of the previ-
ous sutra does not run into this sutra. Any other letter than «r°[ may be
followed by £ and the rule of this sutra will apply to that also. This rule
qualifies the previous sutra which would therefore mean that an ^rj letter
which is not followed or preceded by a $r includes its homogeneous letters also.

Thus in the sutra VII. 1. 9 *T?nPra 5^. 'After words ending in.
*t% (>• e * short *r), i^r takes the place of P15/ Thus f$r: . But in *$r2T which
ends in long *jt and whose prosodial time is different from 9T, this rule will
not apply, and we have OTjft: *

TffTf ( ^ ^**0 *

*ftr 11 MjiRi^^^<jiH ** jjw*uujw^m « jfii«tH? qrufmt *n?^ >nRr

71. An initial letter, with a final T* letter as
a final) is the name of itself and of the intervening
letters.

This sutra explains the mode of interpreting praty&h&ras. Thus the
praty&h&ra*ra^ means those letters whose beginning is Hand whose final fir
is |f t

Why do we say with the i final indicatory letter?' Because a pra-
tyahara like *j^ would be otherwise ambiguous. It might be doubted whe-
ther it is formed with the final £ of sft^ and refers to the first five case-affixes,
or whether it is formed with the initial \otzf and therefore refers to the first
seven case-affixes. The present sutra clears up the doubt. A pratyih&ra is
formed with the final T3.and not with the initial f^.

FTC* ^ <kHW II , (^ r\r\ci\(>

Digitized by VjiJOv LC

//

Bk. I. Ch. I. § 7a.] Tadanta-rule. 69

srft^rrer ii * ^T^faww^uffl ii *

72. An injunction which is made with regard
to a particular attribute, applies to words having that
attribute at their end as well as to that attribute itself

This sutra consists of three words : — sfa 'by what (attribute)/ f%ft:
' rule/ qq*K3 ' having that (attribute) at its end.'

This is a rule of interpretation. When a rule is made with regard
to a particular attribute or letter, it means also words having those attributes or
letters at their end. Thus there is a rule (HI. I. 97) declaring " let there
be the affix ^ after the vowels." Here the phrase "after the vowels 7 '
means and includes " after the roots ending in vowels, as well as roots con-
sisting of a single vowel." Thus pj + 7^ = *RT*£. Therefore, this sutra
means that when a rule is laid down in this grammar with regard to a partis
cular attribute, that rule, besides being applicable to that particular attribute,
is also applicable to words ending in that particular attribute. Thus a rule
laid down generally with regard to vowels, will apply to words ending id
vowels ; a rule laid down with regard to certain forms, will apply to words
ending in .those forms.

Thus the sfttra ^c^C(IH- 3- 56) 'after \ there is ^5 ' declares that the
affix ach would come not only after X> but after any expression that ends
in f . Thus f + *nt " W$ Pf + *n( - ^P II

Vart : — The present rule, (called tadanta-vidhi) must not be applied
to rules relating to compounds (samisa), and to pratyayas (affixes). Thus
sutra II. 1. 24, declares that a word in the accusative case is compounded with
the words Srita, atita &c. As *6k + pnn « q^ftrcr: f involved in pain. 9
The present sutra must not be applied here, and we cannot say that a word
in the accusative case would be compounded with another that ends in irita ;
for tadanta-vidhi does not apply here. Hence we cannot form a compound
of 3* + TCTfrtr: II

Similarly, in rules of affixes. Thus sutra IV. 1. 99 declares "that
the words to &c, take the affix qrgj ." Thus stt + qr^ - HI^HH: I But we can-
not apply the affix phak to the word ffiRrr , the tadanta-vidhi not being appli-
cable here. The descendant of *npr? will be called <?H* nfo and not as
above.

The above v&rtika is however qualified by the following :—

Vart : — The above v&rtika does not apply to words formed by affixes

Digitized by VjUOvIC

66 Vriddham defined. [Bk. I. Ch. L § 73.

that have an indicatory ?3£ fa 5f, 5f), short or long, or where a rule is pro-
pounded with regard to letters. Thus the affix ktavatu (1. 1. 26) has an indica-
tory w f and we have fTCJTJ. A rule which will apply to kpitavat will
apply also to the word which ends in kptavat Thus the feminine of
kptavat is fT^Ttft (IV. 1.6), the feminine of sukptavat will be sukptavatL
Similarly rule IV. I. 95 *pr 7^ says "after *r there is the affix ^ ." This
is a rule relating to pratyayas, but as it is propounded by mere letters, the
rule of tadanta-vidhi will apply here. Thus J^T + f*J - *if^r ' the son of
Daksha/

Vart : — When a term, which denotes a letter, is exhibited in a rule in
the form of the locative case and qualifies something else which likewise
stands in the locative case, that which is qualified by it must be regarded
as beginning with the letter which is denoted by the term in question, and
not as ending with it.

Thus sutra VI. 4. 77, achi £nu dhitu bhruv&m yvor iyan uvanau,
means " iyan and uvan are the substitutes of what ends with the pratyaya
£nu and what ends in a verbal root in \ or 7, whether long or short, and
of the inflective base bhrft, when an affix beginning with a vowel follows."
Here the word ' achi ' is exhibited in the seventh case, which literally
means 'when a vowel follows;' but by virtue of the present v&rtika it
means, ' when an affix beginning with a vowel follows/ as. Pro, 5^: II

*rf%^ 11 **T Hi«^i4qHw**m4i ^r^ftfir q^m* n
*tf%^ 11 * *r str^tot F5 ^?rr w*&*r. 11 *
flirty 11 * flnpu«hmnRfriiim *3r^ 11 *

73. That word, among the vowels of which
the first is a vriddhi, is called v/iddham.

This defines the word TS% • The sutra consists of five words :—
fftj: ' the vpddhi vowel / *TC*r ■' whose/ 5^^^ 'among the vowels/ *n% f is
first/ wf^ ' that / T&L 4 ' IS called vriddham/

If a word consisting of many vowels, has a vpddhi vowel coming
first in the order of vowels, that word is called vriddham. Thus the words
1THT, JP^t &c-> have the first vowel *jt a vpddhi vowel, therefore, they are
called vriddham. ' itizedby(

Bk. I. Ch. I. § 74, 75. ] Vriddham defined. 67

• The word vriddham has been defined in order to make applicable
to these peculiar words certain taddhita affixes. Thus there is a rule (IV.
2. 1 14) by which the affix gf is added to vriddham. Thus Hn*T + ^ « TTPfl^:!!
Why do we say " the first?" Because a word like 4mj4M4H% which
is not a vfiddha word, will not admit of the affix fj. Its corresponding deri-
vative will be *HWI*H*W : II

Vart: — A proper noun, though not coming under the above defini-
tion, may yet be regarded optionally as a vriddham. Thus ^rf + 5 ■•
Si<*T(Ui, or we may have |**W.

Vart : — A compound ending in a Gotra word, is regarded for the
purposes of the addition of this affix, as if it was not compounded. Thus
rrf% is a Gotra word and also vriddham. Its compound y<t<ifl will be re-
garded as a vriddham too, for the addition of the affix. As ^*n$lt +.5 —
y<U*U?<qr: ' the students of Ghritaraudhi.' So also vJmMlUu*fUl :, f^rwfllT^T:

Vart: — Exception must be made in the case of nr $il^M<3 and ttR<HMM| l
They do not take e£ # Thus Agil^h l d r and g|fid*h|flf : #

c^lflfa ^ II 38 II TJ^lfo II Wf-W^f^, *r n

74. The words tyad &c., are also called
vriddham.

The **r^ &c. have been given in the list of sarvan&mas. See sutra
27. These words are also called vriddham. Thus *rj+3C«<r$te*r, ^nft*l»
*#*£ reft*** sulfas. , * W*n%, ITf nrf%: &c.

^xrrei ^ 11 ^ 11 T^rft u ^, mm\, \§, (tot-

75. A word that has the letters s and *fr as
the first among its vowels, gets also the designation
of vriddham, when it is the name of Eastern coun-
tries.

As if[*f^ is a country of the east, it is a vriddham word and we have
*fa! + SC - *flFTffa: ; similarly ggfrreifa% , Hfrrarfta •

Digitized by LiOOQ 1C

68 . Vriddham defined. . £ Bk. I. Ch. I. § 75.

The word prik in the sutra has been differently interpreted. The
Grammarian Kunin takes the word pr&k as an epithet qualifying the word
frchirya, 'professor 1 understood. In that case the sutra will be thus trans-
lated :-— ' In the opinion of the Eastern (or ancient) professors, a word that has
the letters e and o as the first among its vowels, and is the name of a country,
is called vfiddham.' Under this interpretation, this will be an optional
rule only, giving the opinion of eastern or ancient grammarians.

Others however take the word pr&k as qualifying the word de£a
and the meaning then is as given in the text.

Why do we say ' having the letters If and qjt ' ? If a simple vowel
is first among the vowels, of a word, that word will not be called vriddham :
and as such will not take the affix @C ; as ^jifij^^ :, 3E i«»<|<fr4if : «

Why do we say of ' Eastern country ' ? Because this rule does not
apply to a word like l^tf: which is the name of a town in the Punjab, which
is not an eastern country. Its secondary derivative will be I^TT-i born in
Devadatta.

Why do we say 'country 1 ? The^ rule will not apply to rivers &c«
Thus from ifPTft' we have 4t*raT: f fish born in the river Gomati/

v • .

Digitized by

Google

it %m ^^wml^^ filter* ^j^} ii

BOOK I.
Chapter II.

^rfa^n * wr^: *g?r f|gwreft f<r qrara^ii

1. All affixes after the verb g&ft 'to study 1 and
the verb kufc c to be crooked, 1 and the rest, are as if
they had an indicatory ^ (Ait), except those affixes
which have an indicatory ^ (nit) or w fait).

This is an atide£a sfttra, and makes certain affixes, though not
having an indicatory 5 , to produce the same effect as if they wer£ really ft^f.

The root *TT£ is substituted for the verb ^5^ in certain tenses,
in which the verb f^ is defective (II. 4. 49). The aprft verbs are roots
belonging to the Tudidi class of verbs, and they are forty roots beginning
with 5^ ' to be crooked ' and ending with 5^ ' to groan. 9

After these roots all affixes are treated as if they were marked with
a ^ with the exception of those affixes which are fs^ or ftjw.

Thus from *ni£, we have in the aorist the form ^rwpfhf ; which is thus
evolved. *rfa + *r* + f^ + f|r^ + tf - *W + If + ^t + * (II. 4- 5°) •
Wt + *ft + CT (VI. 4. 66) -»«raift? ' he studied, 1 ^THrifHrni} ' they two ,
studied ;' ^wpffara 'they studied/ Here by VI.. 4. 66, the *rr is changed
into f before the affix ftnj, which is regarded as Rfj.

Similarly, after the verbs *£J &c., all affixes are treated as fir^ f with
the exceptions already mentioned.

Thus ^ + <T3 (III. 1. 133) - ?R + f* + *€% (VII. 2. 35) - 518c,
nominative singular $fd« | ' who acts crookedly.' Here the affix ^^ being
regarded as f|p?£ does not cause the guna substitution of the ? of aj£»
by I. 1. 15.

Similarly with the affixes yyj and ffwrj, we hav£ ^f^fl an *
$f?H«q^ . So on with all the forty roots of kuf&di class : e. g. ^tjfcrf

70 Affixes like ft^. [Bk. 1. Ch. II. $ 2,3*

But the affixes marked with w or ^ will not be so treated. Thus
fitrjr by which causatives are formed, is such an affix ; e. g. ^gf + ftJ5 +
Xr^ + Rrj. - **st^ + f + V + fa - Wfr*!%. Here there is guna of
the -st 'of frar. So also before the affix up* of the perfect tense, wc have
<d*3<k\z, and Wore lhe'affix<*^ we have Vesiftes!?: I

Vdrtika: — The root «^ 'to deceive/ is also treated like
ku{idi verbs, except before the affix 5TH. As fqr^FfTI ' he shall deceive/
fiP^^H ' for the purpose of deceiving ' fipqqwiq &c. Here there is
samprasiraAa before the Ait terminations fir, J^and fl**^, and the* of
«^r^is changed into f by rule VI. i. 16.

But the affix *r^£ is not Ait and consequently there is no sampra?
s&rana before it Thus SF^ren'an imp/

f^r %\\i r n tr^rf^ u fesr: , ^y , (few ) »..

1%: ii ^rf%«ft w^<rH*D:, mwitm f^i^:4<^4) feinrfir II

2. An affix that begins with the augment
IP? (VII. 2. 35) is ftit after the root vij., c to fear, to move.*

The root frogr does not strengthen its vowel before the affixes
that take the intermediate f, because those terminations are treated
like flF^ affixes. The root f%iT belongs to the Tud&di class. Thus ^ftr3|Rir
' he will be agitated/ ^ fi t' fSiH^ ' for the purpose of being agitated/
gftfSryprerS , * ought to be agitated/

But those affixes, which are not preceded by the augment jt, are not
like fi^, and they consequently produce their own proper action, such as
gunation &c. As ^ + f^ + *j? - ^$*H*t » sfpFfor? ' that which ought
to be agitated. 9

f^m^TOT: imixr^rf^r ii fsrwnn f ^n: i (fV^t **) n

*ftr: ii 37^*1 *rr*srr^, *r*m<nr r^rf^^inr^ f^nrr flr^rfir ii

3. After the verb tenia l to cover/ the affix
beginning with the augment ^ is regarded optionally
like Ait.

Thus before the r^rft termination $rir of the future, we may have
either jjr^%*rr or 3 m&f?<n " he will cover."

The equation of the form is : — sy + ^J + f + ITT - tf+ 3T$PJ + *TT
(VI. 4. 77) — sJfjJf^lT. When the affix is not regarded as fir?r , it produces
guna, and we have the second form sfitdhNI II

Of course there is no option allowed before those affixes which do v
not take the augment fj, and before those terminations, consequently, wc have
only one .form : as lfr$+F^~TOiff + srT^« jfafrTf* ' covering/ flufiq*ftqJl
1 which ought to be covered/ ^ ^ r\r\n\o

° Digitized by VjjOOyiC .

Bk. I. Ch. II. § 4, 5. ] Affixes like fa%. 71

!%: ii whrnpfr *refiraf5nrifir m

4. A Sdrvadhdtuka (III. 4. 113) affix not having
an indicatory ^ is like Ait.

The terminations which are added to roots are either S&rvadh&tuk*
or drdhadh&tuka. The Sdrvadhdtuka terminations are all those that are
marked with a ^ and also the tense terminations called fin£, the latter are the
following;—

sing.

dual.

#•

-

3rd Pers.

ft*

«S

ft:

1

2nd Pers.

fa*

^

*T

i Parasmaipada.

1st Pers.

Pr*

*5

*3

i

3rd Pers.

n

Wf^tl*

«r

)

2nd Pers.

*n*r

^i^ii

"**

i Atmanepada.

1st Pers.

T5.

<rft

*rf^

J

Except the terminations marked with a jr such as Rr Rr and flfr
&c. all the other S&rvadh&tuka terminations are fS^. The tf$r<T termina-
tions like ^jTi ^FfT^ &c., are also Sirvadhituka terminations. These affixes
are also flFfT, unless they have a ^ as indicatory, as jp£. Thus the affix
*T^j, being an *rfir^ SArvadhAtuka affix, is a fS^, and before it the vowel is not
gunated: e. g. 3j^r: 'they two do.' The equation is: — fr + 7 + 93 (the ? is
added by Rule HI. i. 79)- «£ + ?.+ 1R[ (VI. 4. no)«^F+*r$ - ^n II
Similarly *|$f£<r ' they do.*

Here the %f of ^ is changed into *r and we have a^r, by virtue of
rule VI. 4. no, which declares that the *[ of the verb *T£is changed into *
before those Sdrvadh&tuka affixes which are feTj or fa%. Here tpj is a
S&rvadh&tuka fi^ affix by the present rule, and hence the change.

These four sfitras describe the fir^ terminations.

Why do we say ' a SArvadh4tuka affix/ ? Because an &rdhadh&tuka
affix though not being marked with an indicatory ^ will not be a fg^. As
q*% 'doer/ *xfa ' to do ' qrfts*PT ' duty/

Why do we say ' not having an indicatory ^'? If a S&rvadh&tuka
affix has an indicatory £ it will not be Ait. As fT + T + ffTX-^Ttf^r 4 he does:'
mtifit * thou doest/ qrrrft ' I do/

fa^, ( ^rftn ) u

Digitized by VjOOQLC

7J Affixes like fti^. [ Bk. I. Ch. II. § 5, 6L

5. The affixes of Lit (Perfect Tense) not com-
ing after a conjunct consonant, are as if they had an in-
dicatory «£ (kit).

With this s&tra begins tbe description of affixes which are regarded
as fe^ though not actually having a ^ to mark them as such. The termina-
tions of f*T^ (Perfect) not being ftfjT terminations, are f^T?r, after verbs
that have no compound consonants. The phrase HjfH^ (not being marked
with <r ). qualifies this s&tra also, being drawn from the previous s&tra by
the rule of Anuvritti.

Thus from firf ' to tear 9 the Reduplicated non-fit second person
perfect is fqrft?<j: * thou hast torn/ faftf : ' they have torn/

. Similarly fqfod^j : 'thou hast divided;' far*3C$: 9 y° u divided;*
fc$: and fsj: »

This rule will not apply if the root ends in a conjunct consonant.
Thus #$ 'to fall down* forms its second person Perfect: — WT3+SET55—
4H?taj : . Here the affix is not treated as kit ; for had it been so treated,
it would have caused the elision of the penultimate nasal (anusv&ra) of
*Nr by rule VI. 4. 24. Similarly from vr^ we have ?*3%.

The pit terminations of the Perfect are not kit ; and therefore, we
have fipiT ' he has divided/ where the root is gunated before the 3rd per-
son singular termination nr*J which is a ft^ termination, because it being
the substitute of Rr^ is regarded like f?^ by I. 1. 56.

6. The lit or Perfect Tense affixes after the
roots indhi l to kindle,' and bhu l to become,' also are kit.

After these two roots the terminations of the Perfect are f%«T.
These roots have been especially mentioned, because the root ff^f
ending in a compound consonant will not be governed by s&tra 5 ; and
the root *£ is mentioned because all the terminations of the Perfect are
fotr after ^, not excepting the ft% terminations. Thus the verbs ^ffa
and t^ are illustrations of Perfect Tense from the root <^as in the following
examples:— *rftt *t****W* f (Rig Veda VI. 16. 15), and yr ft *Plfov
(Rig Veda VI. 16. 14). Here the nasal of the root is dropped by regarding
the affix as fa^ ll Similarly from the root ^, we have ^r f *$?TO II

Varti — This rule should be applied also to the verbs N*£ *to loosen/ h*J
'to fasten, 9 f£ 'to injure 9 and *rg[ 'to embrace/ As ^l?f: 'thou hast

Digitized by VnOOQ 1 6.

Bk. I. Ch. II. § 7, 8. ] Affixes like fanfc 73

loosened • ^+^T35- , ^+W95- (The % is elided by the rule already
mentioned in the last sutra vis., VI. 4. 24).

«*T*+M*{+ *T3^»$ta3: (VI. 4. 120). Similarly TO^:;**?:, <rfrre*# f

^^^^^^^«<t««5 SRT II b H M^lPl II ^*-^-*f»

lf%: u?r«f ^^ffe^r^^f^^^^nr: tf*pr**p fo**i% ■

7. The affix ktvd is kit after myiij c to be graci-
ous,' mjid * to squeeze, 1 gudh i to wrap up,' kush l to
tear,' kli£ c to suffer,' vad * to speak ' and to vas c dwell*

After these seven verbs, the gerundial ktvd is kit. It might
be objected, what was the necessity of this sfttra ? The affix
ktvd is clearly a kit affix, as it is marked with the diacritical letter 5£, and
it was useless to call it a hit affix after these verbs. The reason of this
apparent repitition is this. Though htvd is a hit affix generally ; but there is
a later sutra in this chapter (I. 2. 18) which declares that the prefix ^r
is not RPT , if it is preceded by the intermediate f. Or in other words,
if tin* is added to those verbs, which are called %? verbs, viz., that
take an intermediate f , it is not then hit. Thus the verb fft € to sleep ' is %^.
Therefore tft + f + *CT~$ + T + «T« *rfo**r.

Now it happens that all the seven verbs mentioned in the sutra are
Sj verbs, and take an intermediate f. The 3FT before these verbs would
have lost its fr/Vness by virtue of sutra 18; therefore arose the necessity of
this sutra, to provide in anticipation an exception to that sutra, in favor of
these verbs, where the affix htvd is necessarily hit. Thus we have ^ftt^F
4 being gracious* ^ff^F ' having squeezed ' ^jf^rT ' having wrapped f <^ftWT
4 having torn/ fefi^^T 'having suffered* ^rf^l 'having spoken' Tftttf
4 having dwelt.' The ^r v of ^ and ^£ are changed into xT by the rule of
Sampras&rana (VI. 1. 15),

Moreover in the case of the verbs *pr f ^[ and ft^, sutra 26 would
have made ^T optionally hit ; the present sutra makes it necessarily so.

^ff-^afir-n^ ^ , ( st fer ) 11

8. The affixes ktvd and san (lit 1. 7) are *iV,
after rud c to weep/ vid i to know,' mush c to steal,' graft
1 to seize,' svap c to sleep,' and prachchh c to ask.' ^ T

Digitized by VjOOQiC

74 Affixes like far^ [ Bic I. Ch. II. § gi

After the above mentioned verbs the ktvd though taking an inter-
mediate f, and the affix ^J are far^. The affix ^ is employed in
forming desiderative verbs. The •£ of ?Erj is ^j, the real affix being fr. The .
root is reduplicated before this fr-T affix. Thus 4ft*lf * having cried/
V*ftqf?T 'he wishes to cry'; f%r?**T 'having known/ fafaft^ * he wishes
to know'; iJ^faT (VI. I. 16). 'having taken, 1 fa^tffa 'he wishes to
take'; jf^r 'having stolen/ gjPrqr** 'he wishes to steal'; gt^r (VI. I.
15) 'having slept/ ffeai% 'he wishes to sleep,' S|f (VI. 1. 16) 'having
asked/ P^fod ^ M ' he wishes to ask \ The verb ^ is changed into 75 , &%
into *p, JJ^££ into 7^5 before ft// affixes by sampras&rana (VI. 1. 16).

The ktvd and san would have been optionally kit after the verb
*T» f%J and yrby virtue of Rule 26 of this chapter, but this sutra makes them
necessarily so, after these three verbs. In the case of the verb if£, these
affixes are kit by virtue of this sfitra only. In the case of the verbs ?E^5T and
3^£ the affix SET is of course Jt/V, because these two verbs are anit, that is,
they do not take the intermediate f; and therefore the *3?T being directly
affixed to them, and rule 18 not applying, the indicatory k would make it
kit. In reference to these verbs, therefore, the present sutra teaches some-
thing new, only with regard to the affix ^, by making that also kit. The
verb JT^er takes the intermediate f when the affix ^ follows, by virtue of
sutra VII. 2. 75.

*%T «t« II % II tl^Tfa II **>: f 9KS^( ^ f f%^) II
ff%: 11 rr^rngnfr : «rn grenf?: ^ fa-iufi n

9. The affix san beginning with a letter of the
Pratydhdra jhalis like &zt 9 after verbs ending in tk vowels.

In this sutra there is the anuvritti of B*£ only and not of stir. The
sutra literally translated means: — "After tk the jAal" The Praty&h&ra fq£
contains the four letters f , 3*, ^ 5f. The Praty&h&ra fT5£ contains all
the consonants, except the nasals and the liquids. After the verbs ending in
f3£ the desiderative suffix ^5 is kit t when it is 4J<rilft, that is to say when the
affix ^r does not take the intermediate f. The affix 9^ is an affix which
begins with ?fj a letter of the ;n? class. But if the intermediate j is added
before it, it is no longer a $T*nf| affix, and in that case it is not f^rt£ . Thus
from f*r 'to conquer* which is a verb ending in f^, we have fi f *1WR t
'he wishes to conquer* 5 'to mix 1 *RJTI%; so also ff 'to do/ ftHflsRi
^r ' to cross' fctfftnt ^ ' to be' fprft .

If however, verbs ending in f or 3* or ^ or $r , take an intermediate f;
they likewise take guna. A$ : — far 'to smile 9 takes guna, and makes its
desiderative firerfiptffr . \

Digitized by VjOOQLC

Bk. I. Ch. II. § 9, 10. ] Affixes like fa\ \ 7$

The fTCTftOTis not kit- after verbs ending in any other letter than
the f^r. As rTTrefi f (from qr ) 'he wishes to drink.' So from WT. we .
have fforafff.

The object of making the affix q% kit affix after verbs ending 1
in ik vowels, is to prevent the gun a of the vowel of the root, as in PrtWSfr .
' wishes to collect ' cfjsrnr ' wishes to praise.' To this it might be objected
" this cannot be : for rule VI. 4. 16 (when ^ beginning with a jhal i. e., not
preceded by the augment fj follows, a long vowel is the substitute of verbs
ending in a vowel, and of the verbs f*r and *r3. ) w f ould prevent gun a by
substituting long vowel." To this it may be replied "if rule VI. 4. i6 f
enjoining long vowel, debars rule VII. 3. 84 requiring gun a, it should a fortiori .
debar rule VI. 4.51 which requires the elision of the affix ftx before Ardha-
dhituka affixes not taking the augment \\. But evidently that is not so;
for the elision of ftr is not debarred by VI. 4. 16. Therefore, the present
sutra is made, so that even by giving scope to the rule of lengthening,
the affix *rj be treated as fifTSf/' Thus in fa^tafcf scope is given to rule
VI. 4. 16, and the vowel is lengthened. But this rule in its turn will be
debarred in those cases where rule VI. 4. 51 will apply: in such cases the
maxim contained in I. 4. 2 is our guide, and a subsequent rule will debar
a prior rule of P&nini when both find their scope in a single case. Thus
in forming the desiderative of the verb g^r ' to inform/ both the rules

VI. 4. 16 and VI. 4. 51 present themselves, but the latter prevails. The verb
^ belongs to the churadi class of verbs, and takes the affix fcJTJ- Thus
fnr + fiiT^r + ^[ + f?r^; here the present sutra comes into force, and H^
being treated as fisrr, the guna of the \ of ^rPr is prevented. Thus sfttra

VII. 3. 84 requiring guna being set aside, sutra VI. 4. 51 comes into play,
and causes the elision of ftj^[ and we have ^rj + fT + 1% - <ffc*U^ (VII. 4. 55
*rr changed to \) ' he wishes to inform.'

10. . And after a root ending in a consonant,
and preceded by a vowel of the pratydhara ik, the affix
san beginning with s.j/tal consonant, is like kit.

Thus from Pt| 'to pierce* we have feH*m3 'he wishes to break/ here
the root Pnf contains the ik vowel \ ; it ends in a consonant f ; the affix 9^
is directly added without taking an intermediate f, and therefore it fulfills
all the conditions of the sfttra ; and the vowel is not gunated.

Digitized by LiOOQ IC

76 Affixes like ftnj. [Bk. I. Ch. II. § n.

■ I I — — — — — ^— — ■ I ———I » I I | ■■!■■■■ || I ■ . ■ | Ill*

If the final consonant of the verb is not preceded by a vowel of
the it class, the affix *l«£ is hot kit after such a verb. As ^jr + m +
% - RlUMfr 4 wishes to sacrifice. 1 Had the affix 9^ been here a kit affix,
there would have been sampras&rana by rule VI. 1. 16 which declares that
there is samprasirana of the verbs ??, *TT, and irw, &c, before JkiV affixes.

If the san takes the intermediate f and is then no longer a f!5Trft
q%, it is not tit ; and causes gunation etcetera, As f^ + f + ^ « ftiftlH^
4 he desires to be.' Here there is guna of fi.

Roots like $»?, #£ &c, are governed by this rule, though they end
in conjunct consonants, for the word f *J of the sfttra means jdti or class; %
thus *ȣ + ^ + fax - 1% + *rfa (VI. 4. 24 the 1* being elided by treating
^asfirj) - tftarftorfacsfa(VII. 4- 5^).

11. After roots ending in a consonant, that
adjoins a vowel of *£ praty&h&ra, the substitutes of lift
(III. 3. 161) and sich (III. 1. 44) when they begin with a
consonant of jhal pratyAhAra, are kit y when the &tmane-
pada affixes follow.

The sfttra consists of two words fag-Rl'tft and Hm*HH<*J 1 literally
meaning " the affixes lin and sich in the Atmahepada." The other phrases,
given in the above text must be supplied from the previous s&tras, to com-
plete the sense. The anuvrittiol the word *%, that began with sfitra 8.
does not extend further. As : — fir^te ' may he break ' and $*tft€ ' may he
know/ wPHT ' he broke ' ^rf^ ' he knew/

Thus the root f*r$ ends with a simple consonant, is preceded by f
a letter of tt pratyah&ra; after this root the jhaladi affix tfhr of Benedictive
in the Atmanepada is treated like fa<£:; therefore there is no guna subs-
titution of the vowel of the root before this affix, fo% + *fir - Pfttffcr.
Similarly in the Aorist we have *n^T - *r + fa* + fe^ + % the sr of
Aorist being elided by VIII. 2. 26. So also f>^ + tfhr - JpJ + sftr (VIII. 2. 37)

If the final consonant is not preceded by an tic vowel, this rule will
not apply. Thus Ben. inn + tfhr - ^ + ^hr (VIII. 2. 36) - ir^ + *fl* v
(VIII. 2. 41) - ^fhr (VIII. 3. 58). Similarly its Aorist is *TW . Had the V
affixes tfte and n been tit 9 they would have caused sampras&rana of * by

Digitized by VjVjVjy IC

Bk. I. Ch. II. § 11-13.] Affixes like ft^ # 77

This rule will not apply, if the Benedictive and Aorist affixes are
Par&smaipadi. Thus the Aorist of irsr 'to let off/ is *C + *p* + 4tiQ • *T
+ *T + *T + ^ + iStH ( VI - *• 5 8 ) " *TOl + *5tVI. 1. 77) » anireftv
(VII. 2. 1). Similarly from f^we have Aorist Parasmaipadi ^TO^ftj. Had the
Aorist been ftf£, rule VI. 1. 58 (sq^ is the augment of the verbs g^and f^
when an affix beginning with SLjhal and not having an indicatory 3£ follows),
would not have applied.

This rule will also not apply, if the verb ends in a vowel. Thus of
the verb f^, we have Ben. ^nfte, Aor. *q^ g. Had the affixes been fa%^ f there
would have been no guna substitution of i into e.

The rule will not apply if these affixes are not jhalddi but take the
augment f%. Thus Ben. y^ + ^Jte-^fWte, so also Aorist *r*f§*. Had they
been f^r?ti there would have been no guna substitution. -

This rule, of course, is confined to the Benedictive and the Aorist.
Other tenses will not be affected by this rule. Thus the 1st and 2nd Future
of fitSt * s fST and ^q|V

&J II 9^ n q^Tfo II ^: f ^ f ( fe#R&T Wh3<4$3

12. And after verbs ending in ^r, the substi-
tutes of liii and the affix sic/i, are kit, when they begin
with j/tal, and the Atmanepada affixes follow.

As from fr ' to do ' the benedictive is frfte ' may he do.' So also
fa*, as Vb* ' he did/ So also fpfte and *mi.

Of course this rule would not apply if these terminations take the
intermediate f ; as f + ftfte , -«rfHte' 'may he choose/ (Benedictive) ; *T+T +
ff-^TTftT or^r^hr 'he chose or* covered/ the f being lengthened by
j VII. 2. 38.

* 13. Optionally after the verb gam % the liii

and sich beginning with jhal consonants, in the Atmane-
pada, are kit.

The verb n£ 'to go 9 is Atmanepadi after the preposition WH by
Rule L 3. 29.

The result of its being kit is that the nasal is dropped in the Atmane-
pada, in forming the Aorist with fcr, and the Benedictive. Thus we have

2 Digitized by G00gle

78 ' Affixes like fi^ m [ Bk. I. Ch. II. § ti<

the forms :— ^pNRH' or *Hi<J)g in the Benedict!?* ; and OTPHT or CTPTCT in the
sick Aorist. These forms are thus evolved :—

eX+ir^+3hr-*r , J+*r+*fte' (VI. 4- 37) - tfrefte 'may he combine/
Here the nasal \ of n^ is elided by treating the affix tfhrasftp^by Rule
VI. 4. 37. In the alternative we have &HhT.

So in theaorist^+n^+ft^+V-^+^^+^+^(VI.4*37)-^nnr*
The f%^ being elided by VIII. 2. 27 (after a short inflective base, there is
elision of fin;, if ajA*/ follows). So when q[of *p( is elided by VI. 4. 37,
there remains only *r which is a base ending in short %?, hence, the elision of
ftn^. But in the alternative we have OTfctr*

14. The sick Atmanepada affixes are kit after
the root han l to kilL*

As ^nfW, ^nyETmp; , ^rr^RT are the aorist of fsj. When a fy% or second
aorist is kit the verb loses its nasal, by Rule VI. 4. 37. After the verb f*j the
f%^ is only kit and not the Benedictive ftqj terminations, as was the case in the
previous sutra; hence the repetition of the word sick in this sfitra. The sfitras 12,
and 13 were governed both by f^r^and ftrjp fromand after this sfitra, the govern-
ing power of f*fi£ ceases. The governing power of the word " Atmanepada"
in sfitra 1 1 does not however cease, and its anuvritti runs through the succeeding
sfitras ; though as regards this sfitra, its anuvritti is redundant, inasmuch
as before the Parasmaipada terminations, the root f<£ is always replaced by
^ in the aorist, there is no necessity of making the affix sick a ktt. For the
object served in making sick, kit, after the verb f* , is to elide the nasal, while
in the case of q>r , no such purpose can be served.

Tfa: 11 *£reHpt*i% ^hhwc first ***w. l%^,H^Rr ^l^H^ qtw: II

15. The affix sick before Atmanepada affixes,
is kit after the verhyam, when meaning c to divulge.*

The word gandhana means 'to divulge/ 'to make known that
which had been concealed/ The verb *pj has various meanings.

The result of kit is that the nasal is elided. The verb ^ becomes
Atmanepada when preceded by the preposition ur^ by rule I. 3. 28. Thus
5J3[+^rr+^+V 5 -^nnf -' he devulged/ <m<qm<iH , ?fra*nT, the sick is elided
by VIII. 2. 27 and 5 by VI. 4. 37. When the verb *£ has any other meaning '
than 'to divulge/ then the fa^in the Atmanepada is not ibV; as^p^T qif^
•he raised the foot/ So also T^Rrer yny^C 'he raised water from the

Digitized by GoOgle

Bk. I. Ch. II. § i6, 18. ] Affixes like ftnj. 7f

well 9 Though here the vc&yam, is in the transitive form, and takes aa ob-
ject after it, yet it is Atmanepadi by Rule I. 3. 75 which declares " after the
verb^aa* to strive, preceded by ^r*, ^ and *rn£ when it does not refer to a
book, the Atmanepada is employed."

tRt^^trh^^i fk^) 11

vnrc 11

16. The sick before atmanepada affixes is kit
optionally, whenjKtfw means * to espouse. 1

Thus^rnw <h^P^ or OTPT^r ^+n\\ 'he espoused the girl,' 3fPTO
or grrra^l HPftn ' he espoused his wife. 1

The word f Upayamana' in the sfitra means ' to accept in marriage,*
'to espouse. 9 The verb ^ preceded by the preposition T* is Atmanepadi
when it means 'to espouse* by Rule L 3. 56.

*ar^if^ it 13 n mrf* « ^^t:, ^ * <f%*RW~

17. The affix sick before Atmanepada affixes,
is kit after sthd " to stand n and ghu (1. 1. 20) verbs, and
these verbs change their *rr into r before these termi-
nations.

The roots frir'to stand/ jr ' to give/ ^r 'to place* ^ 'to pity' %
1 to feed' ^ 'to cut' change their vowels into f before the terminations
of the Aorist of the Atmanepada ; and those terminations are also kit* As
4<Jlfw«, 'he worshipped* gHlfrnMHI^ , OTTf^rre ; «rf|* 'he gave* *rf*T*
4 he fed. 1

1 ejt %^n %t n q^r u % en, %^, { fro^) «

18- The affix £tvd when it takes the augment
«^ («. e. when it is set) is not kit.

We had had occasion to refer to this sfitra in explaining sfitra 7.
The affix lctvd {ihe actual affix being tvd 9 the k being merely it or indicatory
letter) is on the face of it a hit affix. This sfitra however limits its kitness to
cases where it is not sit. If it is sH i.e., has before it the intermediate f it
will not be treated like a kit affix. Thus in ff^r, %*if, the affix e*r is added
directly to the root ff ' to do ' and f ' to sacrifice/ without an intermediate f
and therefore the root has not been gunated, the affix here being kit. But in

Digitized by VjOOQL'6

80 Affixes like fen^. [Bk. I. Ch. II. § 19, 2d.

*
*^ft**T, \ffal, *f%**T, from ft*|, ftj; and m ' to sweat/ ' to shine/ 'to be/ the
affix is not it*?, because there is an intermediate f before tCT and the root has
been consequently gunated.

Of course this s&tra relates to ^pr ktv d only. Other affixes though
having an intermediate \ will be treated as hit, if they have an indicatory
«£. Thus the affix frfT^tHI. 3. 94) by which feminine nouns are formed
from roots is a f%*£ affix. This affix when %*, will remain RfTjr, asf^+ifs +^
fif^Pr+TJ+^+fir (VI. 1. 16) -Rpjfiftr: (VII. 2. 37). Here there is sampra-
s&rana by treating ktin as kit. So also in y rf^rfgfi r and fH*l>lP f: there is no
guna. After ir^&c, the affix Retakes the augment j\ by a Vartika under
sfitra VII. 2. 9 which see.

(hum sf^fef^rf^flcif^^r: II %% H M^lfa II fa*3l,

19. The s<*t Nishfhft affixes are not kit after
the verbs 61 c to lie down/ svid ' to sweat/ mid ' to melt/
kshvid c to be unctuous/ and dhyish c to offend.*

The Nishthis are kta and ktavatu (see s&tra 26, Chapter I). They
are obviously kit affixes. But in the case of the above five roots they are
not kit, and the vowels take guna Thus jjffcnr. slept, *lfa<W*, ifttft*:, S^ftff^

sk$R<i: ; nwRtwii , tf>R<nm , n^rfenrr^ , snrffo, mrffo*^ &c. In all the

above roots the Nisktha affixes are also sit.

The roots ferf (ftfarar ), ftf ( ftFur ), Rr|( fafircr) and ^ ( ft^r)
have all long qqr as indicatory, as well as the syllable ftr • These roots
therefore by Rule VII. 2. 16 are generally anie, and consequently beyond the
scope of the operation of the present sfitra. As ft^r: , ferarer% . But when
the Nishthi affixes denote 'a state or condition, bhdva/ or 'beginning of
action/ then the above four roots optionally become set. (VI I. 2. 17) It is in
that case that the present stitra is applicable to them.

qq fcrtfd*aiqi*M I V II t^lPi 11 ^: f fdfcmqwj $VSl

20. The set MshtfiS, is not kit after the verb
mpsh when meaning • to forbear/

Thus»nf*fcn 'forborne' ^r fq^rar^ * forbore. 1 x

Why do we say 'when meaning to forbear?* When it has not this
sense, the se| nish(h& will be ft^ . As srTgft^ ^TTWHH ' he spoke unintelli-
gible language.' ■**»** GoOgfe

Bk. I. Ch. II. § 21, 22. ] Affixes like ftnj. 81

21. The sit NishtM is optionally not iiV, after
the verbs with a penultimate * if used impersonally
or denoting the beginning of action.

According to some the sit Nishthd suffixes are not kit after the verbs
having a penultimate ? if those verbs are used impersonally (bh&va)or denote
the beginning of action (&di-karman). Thus *fgr/to shine* qyffrOT^ or qftfifrPFN
' it has been lighted/ Similarly IJ4J7M4: or licjfcta: *he first illumined.' ' The
word penultimate or uj>adh& has been defined in sutra 65 ante.

If the verb has any other vowel than ? as penultimate, then the
rule does not apply. As ftjfotW^*! ' written by him. 9 Here f is penulti-
mate ; and therefore no option is allowed.

This rule has no application where the Nish{h& does not express
either HTT ' condition f or 'Impersonal action 9 ; or VJlft4i$ 'beginning of
action 9 but is used as a verbal adjective. As ?ft* gfrfard ffrfa'he
gives a shining coin (called k&rsh&pana). 9 Here the word *fr* • shining *
is used as adjective.

The rule applies to the Nishfh& which takes the intermediate \ and
not to others as jtj^k *fnp£ 'the rice that has been eaten. 9 Here no
option is allowed.

This optional form of past participle is restricted in its scope. Those
verbs which take the vikarana JJ\ i. e. the verbs of the Bhu class &c, admit of
this option. Thus *n^ f to wrap up 9 belongs to Div&di class, and therefore,
its participle has one form only as *jf>rt.

fT% u <jjFi <rft f%OT unpr: ^r n %^r fortf* 11

22. The sQ EishtM and the s£t tivd are not hit
alter the verb puii 'to purify.*

The word 'optionally 9 of the previous sutra 21 does not govern this
sutra, because the word ' optionally 9 has been again used in the next sutra
23; of which there would have been no necessity, had the word 'optionally*
governing force in this sfttra, for then this would have also governed the next
sutra. The anuvfitti of * $? however is to be read in this sfctnu 1

Digitized by UOOv LC

8* . Affixes uke .ftr^ # { Bk. I. Ch.II.§ 23,24.

By rule VIL 2. 51 after the veib ^ the affixes Nishthl and ^r
ktvd take the intermediate \. So that yr ktvd being sit (by virtue of Rule
VII. 2- 51) would have been non-kit by Rule 18; and it might be aske<f what
is the necessity of repeating it in this sfitra: when sfitra 18 comprised the
case. The reason of this repetition is that by mentioning iff ktvd in this
sfttra, it has governing force over the succeeding sfitras. As qfiFtt 1 <ii^*Hil •

23. The s€% ktvd is optionally kit, after
the verbs having a penultimate % and ending in %
and ^.

The sit ktvd may or may not be kit after the verbs ending in 3
x>r 3: preceded by a nasal. The result of being kit will be that nasal will
be dropped. In other words the dropping of the nasal is optional in such
^rerbs. Thus from inj we have *rftr**T or irf^rerr ' having twisted/ *jfat*T
or $fars*T 'having tied round 9 *lfteqT or Nfct*4T •

The word " NishthA" does not govern this sfttra, and comes to an end

with the last sfttra*

This rule does not apply to verbs that have not a penultimate ^
though they may end in t^ or q^. As^f%*3T 'having reviled ' ifrftt^l 'hav-
ing tied/ Here by rule 18, the m tvd being non-kit, the vowels of the root
fc% f and jpg have been gunated.

A verb may have a penultimate ^ 9 but if it has not final q* orq? t
the rule does not apply. As tfn + ^TT«^f%^T 'having dropped 9 i^toct
* having killed. 1 Had it been kit, the nasal would have been dropped.

%^1 f*S\) *

24. The &6$ ktvd is optionally kit, after the
verbs va iich % i to cheat * lunch i to pluck * and jit i to dare
or abhor. 1

After the above three verbs, the ktvd affix being s{t % may or may not
be kit. As *f^STT or *T%^r # having cheated/ So also *rft*CT or tjfttfT
4 having plucked 9 ^ffiMC or *rf5*CT 'having dared/

When the verb gj^ rfi takes the affix fn^ it is not governed by
this sfttra. See sfttra III. 1. eg. \

Digitized by LiOOQ 16 *

Bk. I. Ch. II, § 25, 26] Affixes like far^. 83

When the ktvd is not sff, but anif, then there is no option allowed.
As T^+ W*T"-**Rrr

fnr: 11 *fa sFr ^rRr *3np*r. qr: ^ nero ^ «snpwCTr*rfoq «rt *r *

25. The s£$ £/*><* is optionally kit, according to
the opinion of Rishi KfiSyapa, after the verbs tjish c to T>o
thirsty ,' mjish i to sprinkle/ and kyiA c to become lean.*

As ?rfs^r or rifq^r ' having thirsted/ sfa**r or *rffar ' having sprin-
kled/ ^rT^TT or *$fk*U ' having become lean. 1 By sfltra 18, the sil ktvd is
never kit. This sfttra adds another exception to that general rule.

The special mention of Kisyapa is for the sake of showing respect.
The *T is already understood in this sfltra by Anuvptti from sfttra 23.

^T^J^T^T^ *g II ^ II ^ifir II VSl 9 *- *-*TOI^,

^-3TT^: # *^ f * ( *T%^f%^) ||

^ftrs 11 ^rrtW>rifV*iOwwi*i*u*rr i'«i«<ii<4«tft: rc *Sv ^ni %*lr it faw
w: 11

26. After a verb which begins with a conso-
nant, and ends in a letter of Ral pratydhdra^ and has as its
penultimate \ orf,* or *r, the ktvd and san affixes are
optionally hit.

The wording of this sfitra requires a litofc explanation. . It consists
of the following words:-—

1. ?sr: ralah, ablative of r^ (after the consonants included in the •
Prat&yh&ra rati. e. f all the consonants except ny and %v): — Verbs having
their final letter any single consonant except ^[ y or ^ v.

*• *QHm*l means "after penultimate *ft" It is a compound of
4t + *TOT . The *fr itself is a compound of w + X - fa, in the dual 4t*
Therefore the phrase means w after (verbs) having for their penultimate either
f tor * u long or short/ 1

3* ?5tt?: means "beginning with a consonant."

All these three words qualify the word " verb" understood.

4. 3fv means " and san/ 9

The sfttra means, those verbs which have a *5 letter in the end, a
f^in the beginning, and an {, f or ?, ?? in their penultimate, may optionally
treat the affixes ktvd and san as ft^. T^

Digitized by VjOOQLC

84 Short, long" and prolated vowels* Bk. I. Ch. II. § 27,38.

Thus *g^ 'to shine/ makes **ff?T*TT or iftffcCT 'having shone/ ftnj
f to write' makes fofcOTT or $1%?*T.

Similarly in mji affix we have f*rf«rftraf<r or fc*rfcraft ' he wishes to
write' fo*/fcfS« or ffeftffoft ' he wishes to shine.'

The rule does not apply if the verb ends in a letter other than rsj.
As f$\ + ^p » $f%**r; its desiderative will be fttftift. No option is allowed
here. • _

Similarly there is no option when the penultimate vowel is not f or
^ . As j^ + ^gpT • *ffor. Des. ftifdMtt .

The rule also does not apply if the verb does not begin with a con-
sonant As irfas^T and ^pvfiraft. There is no option allowed here.

The rule does not apply if the ^T and *n{ are a ntf. As *^T, J**nt
. In these 26 sfitras the description of affixes which are treated like (%^
and fo^ is comprised.

5f^r: nyyy^ ftfrf qrrerM^ rare* **i<Rfrg4 x*%* *fcft >nfa 11

27. A vowel whose time is that of short u$
long & and the prolated w, is called respectively krasva
short, dirgha long, and pluta prolatecL

The x[ u having one m&tri or measure is short, having two measures
is long, having three measures is pluta. This sfltra defines the three kinds of
vowels according to their mdtrd. The letter 3T il has been taken to illustrate
the rule. The phrase Jgt4<ft3Md: is a Dvandva compound, in the Singular,
and is exhibited in the masculine gender, contrary to the general rule II. 4. 17.

The word kdla means measure of time required for pronouncing a
vowel.

As *fa 'curd ' nq 'honey' here x and * are short, gpnfl , *ftft, here
f 1 is long. ct?xTT^ *P*W Devadatta ! Here *jt o, is pluta or prolated.

These words are used in the following sutras I. 2. 47, VII. 4. 25.
VIII. 2. 82.

3TOJ II ^6 || mpfa || <**', % ( 3*S ^ *«: ) II

?f^tt 11 vw 0§ spir: w&^wRhwiuji mj^hhwm Vf*i**jn 11

28. The shorty long, and prolated, when
enunciated as such, by using these terms, are to be
understood to come in the place of vowels only.

This is a paribMshi sfltra, and declares the object to which the terms
* short/ ' long' and f prolated ' are to be applied. When in this book • short 9

Digitized by LiOOQ 16.

Bk. I. Ch. II. § 29. ] Udatta- defined.. • • &

' long' or ' prolated' arc taught by using the words hrasva, dirgha pr pluta,
they must be understood to apply to vowels. As will be found in the sfltra .
I. 2. 47. l "there is the substitution oiharsva or short in the neuter gender
of the crude-form." The word ' vowel' must be inserted in this sutra' to com-
plete the sense. That is : — the hrasva is substituted for the final vowel of
the crude-form.. As t" srfeHr, Hi ^TT%5, *ff W$ ,

The substitution must take the place of vowel and not consonant.
As g*H£ «l$JUJ<JyH*l •

Similarly sCitra VII. 4. 25 • teaches "There is the substitution of
dirgha or ' long vowel ' before all affixes beginning with the letter ^, except krit
and S&rvadh&tuka affixes." Here also the word * vowel ' must be read into the
sutra,' which will. then be ".of an inflective base ending in a vowel, the long
vowel shall be the substitute &c." As fH + *H£ + «* - ^fN% f it is collected';
^ + *T + W - **J?T% ' it is heard'. Why do we say " in the place of vowels
only"? The substitution must take place in the room of a vowel and
not of a consonant. As fa| + ^r + n - f*teT% 'it is divided/ fifrqfl 'it is
pierced'.

So also in the case of plu fa substitution. Thus rules VIII. 2. 82 and
84, 'declare that in calling from a distance, pluta is substituted for the ft por-
tion of the word in the vocative. (The word fi is defined in I. 1. 64). Here
also the phrase ' of the vowel' must be read into the sutra to complete the
sense. The sutra will then read : — " Pluta is substituted in the place of the
vowel of the f> &c." Thus &*^\, O Devadatta ! q m ^| O Yajfiadatta !

Why do we say " in the place of vowels only"? The substitution must
be in the place of the vowel of the fi and not of the consonant. As
mfnP rVfr *tlH*J V| . Here the syllables f^ and ^ are f>, and the vowels f
and ? have been prolated and not the consonant t(.

The substitution must be enunciated as such by using the terms,
hrasva, dirgha, and pluta. Therefore it is not so here ; «ft :, *F*n* :, m, *pm%
*JI% II See I. 1. 4.

^fj^nn 11 ^ 11 tr^rf^r n *f : , g^iw , ( on* ) *
ff^r: is ir*Ui«f»«i«ii«fi 4i4i^ sr 431*4 <j*i) >ref?ri

29. The vowel that is perceived as having a
high tone is called Uddtta or acutely accented.

With this sutra begins the description of the three sorts of vowel ac-
cents. The accent placed on a vowel by raising the voice is called uditta or acute
accent. As SjPn: agnfh, here 1 has the acute accent. The uditta is not marked

(*) *etir 5ry=p£ inf?r<rf53ve* 11 (•) «nrwr}^f!i^i^nJ: 11 ^^fwi:

$3 3^:11 ^TCSttall

3 *

86 -- ANUDATTA DEFINED. [ Bl{. I. O*. II. f 36,3!,

in writing. The word ' vowel' of the previous s&tra is understood in t this
also. These accents are the qualities of letters, and are well known in the sacred
and profane literature. The vowel which is perceived as having a high
tone is called ud&tta. This tone depends upon the organ from which the
vowel is pronounced. In the various organs, such as the throat, palate &c. f
if a vowel is pronounced from a higher part of that organ, it is called udatta*
The vowel in the pronouncing of which there is the restraining and checking
of the bodily parts, and in which there is dryness and . toughness of tone,
as well as a contraction of the larynx/ is said to have the acute accent.
The word is used in sfitra III. 1.3. * "the first vowel of an affix has
the uditta accent/' As fr + ?|5^=3rtfs^kartavyarn.

ift ^qiw 11 \o it tr^rf^r u ift^: f argons:, ( s^r ) n

. 30. The vowel that is perceived as having a
low tone is called Anuddtta or gravely accented.

The word" vowel " of s. 28 is understood in this stitra. As *!&*{ : agnflj,
here a| has the grave accent The vowel accent known as grave of
anudatta is pronounced by lowering the voice. In writing, the anudatta is
marked by a line underneath the vowel. The vowel which is uttered from
the lower portions of its special place of pronunciation gets the grave
accent.

This word is used in sfitra III. 1. 4. • "The case terminations called
?=pj and those affixes that have an indicatory J are anudatta " &c. As TO:, 3t>?Tft.

In pronouncing an anudatta vowel, there is slackening of the
organs, and mildness, as well as softness and sweetness of tone : while there
is expansion and widening of the larynx. .

ffift II ^iTilj^M^i^Hl^lO *ftj^ t *ll ?*HI <ft H^f» II

31. The vowel that has the combination of
Uddtta and Anuddtta tones is said to be. svarita or circum-
flexly accented.

The word " vowel " of s. 28 is understood here also. The svarita
or circumflexed accent is pronounced by the combined raising and falling
of the voice. It is marked by a perpendicular line on the top of the letter.

The word is used in sfitra VI. 1. 185. * "The affixes having an indl- v
catory ^ / have svarita accent." As m^ir kanyft firr^siky&m , ^ kv&.

0) *jnpnreii (•) aij^nftgfSmhi (*) ft?^rftgnu ~ ~~ ■ 7""

\ *

Digitized by LiOOQ 16 .

Bk. I. CH. II. § 34, 33.) The Svarita defined. 87

m 1 ■ ■ ■ 1 ■ 1 li 1 1 ■ ■ . ■ ■ 1 1— — ^ 1 1 ■ ■ n

This sutra is not to be understood to mean that a svarita is the resdi-
tant of the combination of two vowels, an uditta vowel with an anud&tta voweL
It means the accent midway between the two well-known qualities of letters^
known as acute and grave accents.

32. Of it (svarita) the first portion is uddtta^ to
the extent of a half measure, or pi % osodial length.

It has been defined that svarita accent is a combination of uditta
and anudatta. It remained doubtful what portion was accute and what
grave. The present sutra clears up the difficulty, by declaring that in svarita,
the first half is accute and the other grave.

The phrase ardhahrasva is used to indicate half the measure, of
a short vowel. Thus if a short vowel is svarita, its mitri being one, half
will be uddtta and the other half anudatta. If a long vowel, whose mitris
are two, be svarita, then J will be uddtta, i\ will be anuddtta. If a plutd
(protracted) vowel be svarita, then first half measure will be uddtta, and the
remaining 2 J measures will be anudatta. In short, the uditta portion of a
svarita must not occupy more time than is taken in pronouncing a half short
vowel. Thus in fW^l £ikyam the a is svarita, half being acute, the other
half, grave or monotony; in § r »4l hanyd, the long & is svarita, its first half
measure is acute, the remaining 14 is grave; in m«H^I ^ manavak^
here a prolated is svarita, its first \ is accute, the balance 2\ measure
is grave. _

^nr: 11 %$?*{% tot *nfH ^15^*^^ ^ft h

33. In addressing a person from a distance,
the tone is called EkaSruti or monotony.

Monotony or Ekagruti is that tone which is perceived when a person .
is addressed ; in it there is an absence of all the three tones mentioned above;
and there is no definite pitch in it. It is, therefore, the ordinary recitative
tone.

The word * Sambuddhi 9 means here addressing a person from a dis-
tance ; and has not its technical meaning of the singular number of the voca-
tive case. As ^hi^ *ft mm* **l*n\ « O boy Devadatta ! come/ There
is vanishing of all the accents in the above case ; and the final short vowel of
Devadatta is changed into plu ta by the rule already referred to in s. 28 ante
(VIII. 2. 84 ttt^t**) II

Digitized by VjOOQ IC

88 Ekasruti defined. * [ Bk. I. Ch. II. 4 34, 35;

■ '. •

If the person is not addressed from a distance, then the word

will get its proper accent. As Hjm*^ nt 1UU44i' ^PPT 'come, O boy.

Devadatta.'

^^^^^^^^m g 11 ^ 11 m^iPi 11 *re-^fa f sn*-

34. In sacrificial works, there is Monotony,
except vol jap a (silent repetition of a formula), Nyftitkha
vowels (sixteen sorts of om) and the Sdma vedas.

In " sacrificial works" or on occasions of sacrifice, the mantras of
the Veda are recited in Ekasruti or monotony. But on occasions of ordinary
reading, the mantras are to be recited with their proper three-fold accents.

" Japa" is the repetition of mantras, and their recitation in a low
voice or whisper. Nyfinkha is the name of certain hymns of the veda, and
the names of 16 sorts of " Om." Some of these are pronounced with uditta
and others with anudAtta accent SAmas are songs, or the musical cadence
in which some vedic hymns are to be uttered. As: — SfpH^llJl. *$^f?f: afa><lf
^r* 1 *rrc3.rmra fit^uft \ 5.11 (Rig Veda VIII. 14. 16).

When a mantra is recited as a japa, then it must be pronounced with
accents : — As *j*jp% ^T figfoi*j (Rig Veda X. 128. 1).

When not employed on occasions of sacrifice, but are ordinarily read,-
the mantras must have their proper accent, and there will be no Ekasruti.

In the s&ma chanting, the mantras should have their accent, there can
be no Ekasruti there. As ^\ f^npr *WTTO *?T* • Here ti^re is no Ekasruti.

^fhcTtf -m qMd4>l< : II ^H II V \\\k II ^ffcc^T^ , *t 9

*wzww , ( *ra »Afo i, ^5%: ) n

jfc 11 v&xnftj <m^K: «lt*uf qr *r*n* h^nPiIi i f

35. The pronunciation of the word vashat
may optionally he hy raising the voice (accutely accent-
ed), or it may he pronounced with monotony.

The phrase "yajna-karma" is understood here. Even in yajfia-karmas
or sacrifices the word *ftaj; may optionally be pronounced in % a raised
tone. The word VT^ in the sutra signifies *fta^. But why was not
the word 4Wl. used in the text,, it could have been used with .as mucji
ease as*q*? To this natural query, the only answer is, ftf^FIT: ft *Jjre*
.frf^: <nf%J^: strange and wonderful is the structure of PaniniV aphorism*.
Thus *fl*WU«3 <ft#* *ht HI or ^r^ *ft*H tft* S%. : . ^ ^ ed b v ^ 9S - '

BK. I. CH. II. § 36, 3/.] EKASRUTI DEFINED. *- 89

fswnu^fa n \\\\ v^ifaw f^nrar, ^fa, (ij«*Sn)
^frT: 11 SF^ft* fir^rar fiprrer ^3jfti>fai% ^^f^ *«4*i* *nf%ii

36. The Monotony is optional in the recita-
tion of the Vedas, or they may be recited with accents.

In the Chhandas or the Vedas there is option either to use the Ekairuti
tone or the three tones. Even on the occasion of ordinary reading, the
Chhandas might be uttered either with the three accents or monotonously-
Some say this is a limited option (vyavasthita-vibhAshA).

The option allowed by this sfitra is to be adjusted in this way. In
reading the Mantra portion of the Veda, every word must be pronounced
with its proper accent : but in the Brahraana portion of the Veda there might
be Ekasruti ; while some say there must be Ekasruti necessarily and not
optionally in the recitation of the Brahmanas.

. Thus :— *rffctfr^ .3*lT$f or simply srffcp&?C jnffc*. " I praise Agni
the purohita."

The word *r u optionally " of the last sfitra could easily have
been read into this sfttra by the rule of anuvritti; why then use the word
f%HRr "optional" again? To this we reply, this apparently redundant repe-
tition is for the sake of indicating that the anuvritti of the word jd^flif,
which was understood in the last stitra, does not extend to this. Had we
taken the anuvritti of qr, the other word also would have been attracted ;
hence the separate use of the word PoTTST

37. There should he no Monotony in the re-
citation of the Subrahmanya hymns and in those hymns,
the vowels, that would otherwise have taken the svarita
accent, take the uddtta accent instead.

The subrahmanya hymns are portions of Rig Veda mentioned ia
£atapatha Br&hmana. •'-.,'-

This sutra prohibits Eka£rutt in the case of certain prayers called
subrahmanya. By rule 34 read with 36 ante, prayers might be optionally
uttered with Ek as rati accent This ordains an exception to that rule.
In subrahmanya prayers there is no ekagruti ; and in these hymns, a
vowel which otherwise by any rule of grammar would have taken a
svarita accent, takes an udatta accent instead.

As ^m^t y ftTgPTCi *ft* ^mr^ . Here the word giw* is formed
by the addition of the affix ^to the.wordtpigr3 , and this * will get svarita

Digitized by VjDOQ

00 SUBRAHMANYA HYMNS. { Bk - ~ I; CH.il/ $ 3$,3^

accent by VI. I. 185, ( ftresfttp;) as it has an indicatory ?£; by ; the present
sutra, this nascent svarita is changed into ud&tta. In the phrase f?3C ^rnr'Sf
the word Indra being in the vocative case, \ is uddtta, the *r of Indra is
anud&tta VI. 1. 198 * The anuddtta preceded by an ud&tta is changed
into svarita (VIII. 4.66) \

Thus the *f of f** must become svarita, but by the present sutra this
nascent svarita is changed into an ud&tta* Thus in f^r, both vowels be-
come uditta. In the word -iJHl^ , the *jt is uddtta ; the next letter which was
anud&tta becomes svarita, and from svarita it is changed to uddtta by the
present rule. Thus in the sentence x*X IPT^SE the first four syllables
are all acutely accented, the fifth syllable is only anuddtta. So also
in ffcr <*JMI*dU for the reasons given above, the letters 7 and «3 are anuddtta,.
the rest are all acutely accented*

; ^rsrsr^ng^nr: w\i\\ xj^f^r n ^»-w^5ii: > stitch, n

38. The word devd and Brahm&ija in those
hymns have anud&tta accent,

By sutra 37, it was declared, that in subrahmanya hymns, svarita
accent is replaced by uddtta accent. This sutra makes an excep-
tion in favor of the words far and srcrrcr occurring in those hymns.
These words have anud&tta accent. As far TOTO" W | T 5 39 f come ye Dev&s
and Brahm&nas.' Here .the word far gets ud&tta accent .. on the first
syllable by rule VI. 1. 198. l (in the vocative the accent is on the
beginning): q? vd has originally an an uditta accent which • by rule
VIII. 4. 66 ■ (an anuddtta following an uddtta is changed into svarita)
would have been changed into svarita. This svarita, by the previous sutra
required to be changed into uddtta ; but by this rule, it is replaced by anu-
d&tta. In other words, the original anuddtta remains unchanged.

ffir. ii #fg*rraT^ffafr ^iT^wCNiq^iMHft^ft jqfll 11

39. The Monotony takes the place of the
anuddtta vowels which follow the svarita vowels, in close
proximity {sanhitd).

SanhitA is the joining of two or more words in a sentence, for the
purposes of reading or reciting. When words are thus glued together, then

e ) VWf***k n H f ) ^TWT g^^W «lfff : H digitized by GuO^It:

BK. I. CH. H. § 40, 41. ] SANtfAtAitA DEFINED.* 9V

th£ anuddtta accents become Eka^ruti if they are preceded by svarita
vowels ; and are pronounced monotonously. As f*T tf. T§F *3^r WCWftr (R'g
Veda X. 75. 5) O Ganga, Yamuna! Sarasvati ! this mine.

Here the word y4has uddtta on the last syallable: the word % is
originally anuddtta, but by rule VIII. 4. 66 * following an uddtta % it is
changed into svarita ; after this svarita all anuddtta like *i^, &c., arc
replaced by ekairuti. All the vowels of the words i^$r *rg% &o, had anuddtta
accent by rule VHI. 1. 19 • (all vocatives get anuddtta if standing in the
middle of a sentence and not beginning a stanza.)

The word " sanhita" has been used in the sutra to show .that when
there is a hiatus between the words then there is no change of anuddtta into.
eka£ruti. The word sanhitd will be defined in sutra I. 4. 109 *•

^TO^fePKV! ****** n 80 H H$ \ (k H ^TO*S*ft?r-

to* , wz*K*. , ( srg^irem ) 11

jfa: 11 ^m«kw ^famro ^uth* qq<umiun *nfir n

40. The accent called Sannatara is substituted
in the room of an anuddtta vowel, which has an uddtta or
svarita vowel following it.

In the previous sutra it was said that an anuddtta preceded by a
svarita becomes EkaSruti. If however such an anuddtta is followed by an
uddtta or a svarita, it does not become Eka£ruti but becomes sannatara i^*
lower than anuddtta.

The sannatara is therefore that accent which was originally auu-
datta, and which is preceded by a svarita and is followed by an uddtta or a
svarita.

This is one explanation of the sutra. There is another explanation
which does not take the anuvptti of Eka£ruti in this sutra. The anuddtta ie
replaced by sannatara when such anuddtta immediately precedes an ud4tta
or a svarita. The sannatara is also called anuddtta. As^n *fgw. uffiHMifo : II

Here the word mice, is anuddtta. The word *JH*: has uddtta on the
last syllable by VI. 1. 171. 4 In the phrase ifam: (*rnTT: + *pr:>The syllable ft
is anuddtta, because anuddtta + anuddtta - anuddtta. This anuddtta *jf,
preceding the uddtta T- 9 is changed into sannatara.

STlrR ^BTSTHcW II 81 N H^rf* II WJ^R: f ^-STCT f

rare: ,11

C) **r*m*rTKmfK*:\i C) <OTrfNm**fii (•) *c. crfrft&ofttirii

(*) ^f^HilSJFJ^tJWBT: II

Digitized by VjOOQLC

OS .- Karmadharaya defined. " [ Bk. I. Ch. II. § 4*:

# 41. An affix consisting of a single letter is

called an Aprikta.

This defines Aprikta affixes. They are those affixes which consist
of only one letter of the alphabet. Thus the affix f£ / in *r4frft<rts a one
lettered suffix and is an Aprikta. Similarly the affix fej kvin is an Aprikta,
because the letters gr, f, and g^are merely indicatory, the real affix is ^r.
This % v being a single letter, and therefore an Aprikta^ when added to a
base, is elided by rule VI. I. 67 * (the Aprikta \y is elided). Thustpre«p£ +
fa% - \^^i\ + ^ (III. 2. 58) • « tpRS^ll *faPT?t + fa*« stfjpn^ 'co-sharer.'
(III. 2. 62). • An affix which consists of more than one letter is not an
Aprikta. As ?tff : ' a ladle '' formed by the Unidi affix ft* (^f^f ftr^ ).

%*tmiRw>c&k , **>4^iw , n

42. A Tat-purusha compound, the case of each
member of which is the same, is called karmadhdraya ox
Appositional Determinative compQund.

£* " Karmadhiraya or Appositional Determinative compounds form a
sub-division of the determinative compounds (Tat-purusha). In them the first
portion stands as the predicate of the second portion, such as in black-beetle,
Os . sky-blue." J ([Max Muller).

'^^^.tr/' The compounds like «fl<rU<M»l 'the blue lotus' *rc*n*TT 'the supreme
\ J spirit ' &c. are examples of karmadhiraya compounds. " The component words,

;' -. if dissolved, would stand in the same case, (saminidhikarana), whereas
in other Tat-purushas, the preceding word is governed by the last."

As, m>*R|W|*jL, is a karmadhiraya compound meaning " the best
Government." Had it not been a karmadh&raya compound, the uddtta
accent would have fallen on the syllable n by rule VI. 2. 130, 4 (the word
<l*41 has uditta on the first syllable in Tat-purusha compounds, except in
karmadhiraya).

Similarly «nft^+rerit^° Hiq*3^lf<4il I (VI. 3. 42.* the feminine
adjective is changed into masculine in the karmadhiraya).

Each member of a compound may have the same case, but if it is
not a Tat-purusha compound, it won't be karmadhiraya. As *ftPCT+HF5r-
1lP<MiHl4: • ' He who has a wife for a cook. 9 This is an example of a
i Bahuvrihi compound, and therefore the word <j|fc«M retains its gender sign.

(') ^ci^kwii (•) *Tfffr3**ffc*» (*) ?nitf**:ii ( 4 ) *r**m*m*wn

\ Digitized by VjO

it

i

Bk. I. Ch. II. § 43, 44.] Upasarjana defined. 93

A compound may be a *f at-purusha, but if its component parts do not
refer to the same thing and are not in the same case, it won't be karmadWU
raya: — srregg /isj^ Brihmana r&jyam ' a Brahmana kingdom. 9 Here *T5^gets
uditta on the first syllable by the rule already mentioned (VI. 2. 130J.

sre*nfr f^H wt*3*H3*!*w 11 *$ 11 ^tf^r 11 mmr-f^ff^j

ffrT: 11 sn^nTr fire^n* ^Rrflf^ *mhhuh!1 sjroibreRi H*fil 11

43. In (this book, in the sdtras relating to)
compound, the word that is exhibited in the first (Nomi-
native) case, is called upasarjana or the secondary word.

This defines or rather describes the Upasarjana. In. the sutras refer-
ring to compounds, the words that have the case termination of the nomina-
tive, are called Upasarjanam.

Thus sutra II. 1. 24 (frcffar fimT^ sffiflJlflKq^M I MH*! is a samisa
sutra enjoining composition. In this sutra the word factor is exhibited in
the nominative case, and the remaining words are in the Instrumental case.
The sutra means, " An accusative (fSt^tarr) is compounded with the words
P*tf, 'snfhr, <rf<nr, *nr, STTOT, TOT, and ?rrnr and forms Tat-purusha." There-
fore, in forming accusative compounds, the word having the accusative case
will be called Upasarjana. Thus ^re + f^f: « «K*P>W. Here the word
zr£ being Upasarjana, stands first in the compound, by Rule II. 2. 30.
STO^rrf^, (in a compound the Upasarjana should be placed first)* Simi-
larly in rules relating to the formation of Instrumental, dative, ablative, geni-
tive and locative Tat-purusha, the words standing in those cases will be
Upasarjana. Thus fPJSTTCr*?: , ^HFTF ; T^>ra^, rPTJTO, HJ^gUll"* : .

Upasarjana is, therefore, that word which either by composition or
derivation loses its original independent character, while it also determines
the sense of another word called JTVTPT or principal.

44. A word which has one fixed case, (while
the word compounded with it may vaiy its case) is also
upasarjana^ except for the purposes of the rule which ^
requires the upasarjana to stand first in a compounds / p5^

4 Digitized by VjOOQIC

94 ' Pratipadika defined. ; [ Bk. I. Ch. II. § 45;

A word which in the Analytical statement of the sense of a compound
has one fixed case, whilst the word with which it is compounded may vary
its case is also called upasarjana, but does not necessarily stand first.

That which has one fixed case-termination is called eka-vibhakti,

t\e., a word whose case termination has been determined by a rule of samasa

i.e., while the second word in composition with it may take many case-

^\Jt affixes, it takes only one case affix; apurva-nipite, such a word will be an

i3»*n* v upasarjana for all purposes, except for the purpose of the rule (II. 2. 30)

nun-o/i which requires an upasarjana to stand first in a compound.

Thus the prepositions fit: &c, when meaning across &c., are com-
pounded with another word in the ablative case. Though the first word may
have any one of the several case terminations, the second member must
always be in the fifth (ablative) case, when analytically stated. Ex fti4*l~4:

+ * h U » * « m r - fWhcrrffci: . fa**rn*r + rfku*««if - fWNrrf*^ , ftuhi*<N +

Similarly the word PhN j * UI % : II

If the word, when analytically stated, does not retain the one and the
same case, it will not be called upasarjana, as the word ^H\(\ in the compound
iM^Hld, ' the king's daughter,' because n*n + aprrtf - <l*t$Hlfl ; n*n +

*5*rpft^ - im^hiO ; n^: + y rn ft ~ rmyrnft , *r*n + *5*nif - <M<fr*u«f .

Had it been an upasarjana, the long f* of ft would have been shorten-
ed into X as *n the previous example.

This upasarjana, however, contrary to the genera! rule (II. 2. 30) l
can not stand first in the compound, because there is no such form as, 3?)!UIH?l3:

45. A significant form of a word, not being
a verbal root (dhdtu), or an affix (pratyaya) is called a
prdtipadika or crude-form.

This defines the word " praiipadika" or crude-form. The pratipadika
is the intermediate stage in the development of a full word (pada), from the
undifferentiated ore called the dh&tu or root It is that part of a word which is
capable of receiving the case-terminations.

The word ^rf" means having signification. With the exception of
verbal roots and affixes, a word-form having a significance or meaning is called

* * ^ Digitized by V^OOgle

BK. 1. CH. II. § 46. ] PRATIPAD1KA DEFINEd. . 95

a pratipadika, as &&Z: ' a wooden elephant/ qrffcf: ' the wocH-apple tree,' apT*
'■ a bowl/ <ft*£ * a seat*.

Why do we say "significant"? Of course it must have some meaning,
thus of the word ^PT, we cannot say that the portion *^ is pr&tipadika ; as it
has no meaning, and had it been a pr&tipadika, the final ^ would have been
dropped by Rule VIII. 2. 7 l (the final ^ of a pratipadika is dropped).

The word must not be a verb, as the word *rf^ (the third person
sing, number, Imperfect tense of f^); he killed. Had it been a pr&tipadika,
the ^ would have been dropped.

The word must not be an affix or ending in an affix : — as 35*% * in the
bowl. 9 Here the word ends in the affix X of the locative case, and is conse-
quently not a * pr&tipadika ' or crude form, but a complete word. Had it been a
pratipadika, the 1£ would have been shortened into f by rule 47 succeeding.

A nip&ta or particle though not a significant form of a word, yet gets
the designation of pr&tipadika in the opinion of the author of the Mah&-
bhishya. As sy uiHl^ l fr, H<d**i<l .

46. The forms ending in Kjit affixes, or Tad-
dhita affixes, or compound are also called Prdtipadika.

This further explains the use of the word pr&tipadika. " Suffixes for the
formation of nouns are of two kinds. Those by which nouns are derived direct
from roots, primary suffixes. Those by which nouns are derived from other
nouns ; secondary suffixes. The former are called Kfit, (III. 1. 93) • the latter
Taddhita (IV. 1. 76). 8 Thus^PT jana, man, is derived from the root *f^ jam
by the krit suffix ST; but spffrf janfna, appropriate for man, is derived from
3FT jana by the Taddhita affix fn tna. The name pr&tipadika would apply
both to 9CT jana and spfbr /a«i«a, as nominal bases ready to receive the termi-
nations of declension" (Max Muller).

Of the compounds, of which there are six varieties, we had occasion to
refer to five already viz., Tat-purusha, Karmadh&raya, Dvandva, Bahuvrihi and
AvyaytbhAva. They will be fully dealt with in their proper places. J>'

In the last sutra, the words " not ending in an affix " were too exten- \ M ^*
sive; this sutra makes an important provision in favor of words ending in
Kfit or Taddhita affixes. So also compounds are also called pr&tipadika.
Being restricted to compounds only, it, of course, does not apply to sentences.

(*) srefrr: 3 iftMfl<m«w« 11 V) f*ft£ii (•) nf^iir.ii

Digitized by VnOOQ IC

96 Shortening or the Neuter bases. [ Bk. 1. Ch. II. § 47;

The words formed by krit affixes are pr&tipadika a$, f? (to do)-*-*^-
fT + J^^KTT+^pK'-^iTT^K'. The affix vjsj is a krit affix, and it is added
to the root to form agent by Rule III. 1. 133. l The letters^ and *£ are indi-
catory, the force of ^ being to cause the vriddhi of the vowel of the root.
The real affix J is changed into 5H? by (VII. 1. 1 • the affixes ^ and j are
replaced by *Ff and *[zr respectively}, thus the form *H* is a pr&tipadika,
and is capable of declension by receiving case terminations, as, *ttK'<t,: ||
Similarly fT+gr^(HI. 1. 133) * - gf$; nom. case ?fSt.

The word formed by Taddhita affixes are also pr&tipadika ; 3TJ+
^^=^r T Pfr+^r aB ^TPTT 'the son of the Rishi Upagu': nom. case ^ftlJjqr,
similarly TKTTZT- •

In fact in the chapters treating of these affixes, the derivative subs-
tantives formed by them are always shown by the Indian Grammarians in the
nominative case singular number. The examples of nouns given in this
book will be always declined nouns and not prAtipadikas, except in peculiar
cases. So also in the chapter of sam&sa or compounds.

The words ending in compound are pr&tipadika : as n^r:+5^r: = CRT-
$$H, nom. case rPTJ**: .

ff%: 11 Tj^*riw^43f gcmfl s fW «r*l ?rc*r **f)r >rcf?r ^fcu1triM*iHU , i: 11
47. The short vowel is substituted in the
neuter, for the final vowel, of a pr&tipadika.

In the neuter gender, the pr&tipadika that would have ended in a long
vowel, changes it into short. Thus *fft + ^«vrftf)r 'extravagant* as snfaft*-
*$?*$ ' an extravagant family' ; *n^ 5ft « *rf%5J ' very hairy* *rf^<y ^*T f -a very
hairy family/

There is no substitution of short vowel, if the noun is not in the
neuter gender, as *rPT + 'ft =» TTTOft, nom. case inntft: ' the leader of a village
or head-borough/ similarly d^rnfr-'the leader of an army/

A word may be in the neuter gender, but if it is not a pr&tipadika,
the vowel is not shortened, as «hl u 4 f^S^r: ' two cantos remain/ Here the
^ of *l u 5 is formed by the coalescence of ST (of the pr&tipadika *6T*sg) and
^ (the dual case termination) ; thus Sfrnrs+f-grr*^. Now it might be urged,
that here one letter ^, replaces two letters *f and f, therefore this substitute
is equivalent to each of those letters separately, and consequently n is the
vowel of the pratipadika and ought to be shortened. But this is not

(*) ^C^hl ( f ) jtamftii — — — —

Digitized by V^iOOgle

BK. I. CH. II. $ 48, 49. ] A RULE OF LUK-ELISIOtf. 97

to be so, because the word pr&tipadika has been especially mentioned
in the text, and therefore this ekid£sa is not treated as the final of the
form which precedes it

( mfef^rf^B^i ) n

48. A short vowel is the substitute ofpr&ti-
padika which ends with the word *rlr c a cow,' and of that
which ends with what has as its termination a feminine
affix ; when regarded as upasarjana.

The word upasarjana has been defined in sutra 43. In compounds
the word *ft becomes shortened when it stands last and is an upasarjana.

Thus f^Fr + *Tt~P^np ' possessed of a brindled cow\ This is an ex-
ample of a Bahuvrihi or possessive compound. So also SU'TrPf ' a spotted
cow 9 . Similarly in those pr&tipadika upasarjanas which end in feminine
suffixes having a final long vowel, there is a shortening. Thus {%:+<& 2111*41
sf^c^rgrrK?: . Here the final long f is shortened. So also Pi^KlUjfa: .

The word pratipadika of the previous sutra is understood in this.
The word *ff in the sutra means the form *ft and not other words synonym
I with it The word ^ft means " a word ending in a feminine affix," because

j it has the svarita accent The word upasarjana qualifies both these words,

t. e. when ifr is an upasarjana and when a word having a feminine termina-
tion is an upasarjana.

Other examples are: — ^rnr + ^rjf-^rRn3j: ' without a bed stead f
qjfimw * surpassing a necklace in beauty 9 .

If such a word is not an upasarjana, there is no shortening, as,
*M^HlO ' the king's daughter. 9 Here the word TPTJ is upasarjana, and not
the word aprrffr, which latter is therefore not shortened. The word must be
a derivative word formed by the addition of an affix % which makes it femi-
nine. Thus the words 5T^fr, ^ft &c. are feminine originally, and are not de-
rived from the masculine nouns, by the addition of any feminine affix. There-
fore we have *fft?4^4|: 'surpassing Lakshmi in beauty 9 *rftMfc II

Vart : — In Bahuvrihi compounds, exception must be made of femi-
nine nouns ending in the affix t*r^h as sr^M^f, ftfJHH^mJ l* |

^tra^TOi ) n

?f^r: 11 ^f^r^f5i *f*r sfbro^OT ssrcnfo^ 5**nfa ||

Digitized by LiOOQ LC

98 A RULE OF LUK-ELISIOX. [ BK. I, CH. II. $ 49.

49. When a Taddhita affix is elided by the
word luk there takes place the elision by luk of the
feminine affix of the upasarjana.

The word jg^ defined in sutra 61 Chapter I, is twice used in this
sfttra. Where a Taddhita affix is elided by the word JJ3£ ; there the feminine
suffix of the pr&tipadika is also dropped, or becomes luk. Thus there are
certain Taddhita suffixes by which patronymics are formed (one of them
being ^73) from other nouns. Before this sr% the initial vowel of .the word is
vriddhied and the final is gunated. Thus flnjf 'ocean'; forms Rfj + ^T^"
%^r+sr=%^H' ' born in ocean or oceanic.' But there are certain words e. g.
MpTST, y^ m ift &c. which are all feminine gender (see sfitra IV. 3. 34) *
after which the patronymic suffix is totally elided, without leaving any trace
of its action, by using the word luk. Thus *n%BT + (*T;raps) = WI^S": il Here
then the present sutra finds its scope. The Taddhita suffix being luked; the
original feminine termination *jt also becomes luk. Therefore the patrony-
mic derivative of %rf^T5T (is not a longer word, as might have been expected)
but actually a shorter word, namely ^rf^f: ; similarly from <ti<^Jj«ft we have
^p*pT: &c.

Thus there is a feminine suffix called sft«[ (actually f- ) and certain
words like f*x &c, (IV. 1.49) f take an additional suffix called *JTJ^ (actually
*TPr). Thus the feminine off** will be f?ff + ^TJ^+^hC. " T*X + *n% . +
t". *» f^ftruft. Now in forming Dvigu compound (Numeral Determinative
compound), the Taddhita suffixes are dropped. Thus Hi^^M: * an offering
(purodash) made in a dish with five compartments/ In this compound. the
Taddhita suffix *r<^ is dropped. Otherwise the form would have been <*Hl<ri: f|
Similarly when the word ysgpgft forms a Dvigu compound, the Taddhita suffixes
are dropped. But not only that, even the original radical affixes are also
dropped by force of the present sutra. Thus <pbn which means the cake
sacred to the five goddesses (HUUft- Here the full form was ^hf + f^TOft +
^r . (The suffix *r^ has the force of " sacred to God." Thus f*X + *TJ ■»
£np£ sacred to God Indra). In the above expression, first the Taddhita-suffix
W? is lukedby rule contained in sutra IV. 1. 88 • relating to Dvigu compounds :
we have then remaining sfa + f^ c mfl • In this stage, the present sutra
comes into force and declares that where a Taddhita suffix has been luked,
the final feminine suffixes of the upasarjana pr&tipadika are also luked.
Therefore, the feminine suffixes 3n33£ and ^9 are also dropped, and we
have vfazf: which means " cake sacred to the five Indr&nis (goddesses)."

(*) ^ftHI4i^^H<IMK*(filf?K*! J^gft<lftW*<lgl4f??re5I3£ll (') f^^^UJ-

v Digitized by VjOOQ 16

BK. I. CH. II. § 50,51.] A RULE OF LUK-ELISION. "99

The words ^ft and upasarjana of the last sfitra are understood in this.
By the last sfitra, a word ending in a feminine affix was told to be shortened
when it was an upasarjana ; this sfitra teaches the total suppression of such
an affix in special cases. To take another example: Thus Rule IV. 3. 163
declares q?5Trj3r " luk takes the place of an affix signifying modification or
product when fruit is meant." Thus *njis an affix which means modification.
Thus ift + *n{ =■ *r*3PT * cow's milk or the product of cow/ . This affix is sup-
pressed when applied to a tree, in the sense of the fruit of the tree. Thus
the fruit of the tree qrragraft will be called %jw*t3i i.e., first the Taddhita affix
is luked, and then the feminine affix is also dropped. Similarly fruits of
Vfft or qr^ft trees, are ^t or *TF*t.

This rule only applies when a Taddhita affix is elided ; and not so
when any other affix is luked. Thus * \ \*M\ : + ^fr — nnff*^ ' the family
of G&rgt/ Here in forming the sam&sa, .the case affix of the word g&rgt is
luked by II. 4 71 1 . But this suppression of the genitive termination does
not necessitate the elision of the feminine affix $ of gArgl.

This rule would also not apply if there is no elision by luk of a Tad-
dhita affix, as in *TPffepg *

Nor will the rule apply when the word is not an upasarjana. As

^f *frpn: 11 \9 11 v^rtk 11 ^h, Stawr: , ( qr^dgfo ) ■

ffrf: 11 *n u 4it<i 1^431% ^r% f^nrf^ft *nt% II

50. The short is substituted in the place of
goi # ii, when the Taddhita affix is elided by luk.

By the former sfitra when a Taddhita is elided, the long f of the femi-
nine also required elision. But this rule makes an exception in favor of the
word iftaft| in which, under similar circumstances the short f replaces the long
ff, as in the compound T^PTTftn ' purchased for five gonis. 9 Here the Taddhita
suffix which conveys the meaning of " purchased for/' has been rejected
after ifrtft by V. 1.^28, • therefore by previous sfitra , the feminine suffix f
also required rejection. But this sfitra intervenes and changes the long ^
into short f.

This sfitra is divided into two separate rules by the process called
yoga-vibhiga. Thus *?^is made one sfitra by itself, and then it means, when
there is a suppression of Taddhita affix by 5^, then the long | is replaced by
short f . Thus purchased with five needles -tprr%: $*ftPr. *ft*n -T^Pr, TO^fa: •
This is restricted to such examples of purchase only.-

( l ) ^tjrjmf^fi^:!! (*) ^*^ftifr^«rai^ii

Digitized by VjOOQLC

* IjAjl^L** a /(fk' o>^*J ^J^u<~J- urns. >v^(t d± t^cL ZxfatJ,-
100 A RULE OF LUP-ELISION. [ BfC. I. Ch. II. § 51.

51. When a Taddhita affix is elided by using
the word sjst , then the gender and number (of the deri-
vative word) agree with those, of the original word.

This sutra consists of four words : gflr, * where there is a rejection
by using g^ 2 3prey?rJ appropriate to the same/ 3 wrf^K, ' gender/ 4 qn^T ' num-
ber/ The form vyakti vachane is dual.

When a Taddhita affix is luped, then the derivative word retains the
gender and number of the primitive word. Note the difference between the
rejection of a Taddhita by saying * let there be a 5J^' and " let there be a SJj" II
The direct action of both words is the rejection of the suffix ; but a derivative
formed by luk rejection may not retain the number and gender of the original
word ; while a derivative formed by lup — rejection does retain the number and
gender of the original word.

Thus in the secondary derivative word ftrcfar: meaning * a village not
far away from fqrffa trees'; the masculine gender and plural number of the
original word has been retained, though made applicable to a village, which
is in the singular number and neuter gender; similarly the word v^fRTP is
masculine plural in form though applicable to a single province.

This sutra is not an approved rule ; as will be mentioned by P&nini
in sutra 53 following and is consequently of little practical importance.

Thus <farrerr: are a class of Kshatriyas, as such it is masculine gender
and plural number. The country in which Panch&lis live is also called PafL-
chAJAs; similarly 3ft*: , *n*pjr: &c. There is a rule 5pn%5J. (IV. 2. 81, the
lup takes the place of an affix when signifying the proper name of a country
or kingdom). Thus *f ^THT: + ST* (this is one of the affixes forming country)

This rule does not apply when the affix is suppressed by the word
. jyar &c. and not g^. Thus gPTOT : *Jjn ' salted soap* STTORraT*: 'salted gruel/
SMUJSTT* 'salted vegetable.' Here the affix ^ has been elided by the word
/uk f by IV. 4. 24, MIUJI^^ (after the word lavana; the affix is luked).

These derivatives retain only the gender and number of the original,
but no other quality. Thus as we have already seen that fttfNr. £irishAlj f
is the name of a village and is formed by the Ixip of the affix. Now com*
pounding it with ** we have fin!fmq^ 'the forest of the village SirishA/ Here

Digitized by VjOOQLC

BK. I. CH. II. § 52. ] RULES OP LUP-ELISION. 101

the word ftiflm : means village, and not trees. Had it meant trees called
girisha, then the H of *5f would have become t!j optionally by Rule VIII.
4- 6. ( ftajmqfa q^Tfiw II the H of the word ^r is optionally changed into n |
when preceded by words signifying annual herbs and trees). *

Vart : — In the case of the words Haritaki &c. the gender only is
retained and not the number. As ^flraw^ q»gllPr ' the fruits of Haritaki.' Here
<K*T is neuter in form, while ^fltuft is feminine.

Vart : — The words i^ftcfc &c. retain their number only but not gen-
der, as <^fcjch%HIH 'mountain forests i. e. forests in the vicinity of mountains. 9
Here *35TnfcT is singular and <Mlft plural, but the gender of both is the same
viz. neuter.

ffrn 11 9<j<U44ift faftmuiRi tfaprf? «rf%^^ >rrefr *nf% T^firar 11

mfin^ 11 »T5^^ Jrt%wT T^wn 11

52. And of adjectives, (the gender and num-
ber are the same as of the word formed by lup elision of
the Taddhita and which they qualify) so far as the jdti
(or kind or species) is concerned ; (or when not expressing
jdti).

When a Taddhita is luped, the adjectives which qualify such secondary
derivatives agree with them in number and gender. This is a rule of syntax. Thus
■Hwr: ililitai: *gTO, H^fU qw: ^Sm^^f : ' the pleasant Panc hjtLAs having ^J
plenty of food, plenty of milk and butter.' Here the adjective rroffax* pleasant ^\
&c, agrees in number and gender with the Panch&l&h; being also in masculine **** t '
plural. But when such a derivative is used as a jlti word then the Vv*']
adjective does not agree in number and gender, as 4^11*41: imTi *it$t Wit.
Here 3PH1 though qualifying Pafichala does not agree with it in number
and gender, as it is a noun of class. All attributes of such words formed
by the lup of Taddhita agree with their substantives in number and gender:
provided that such an attribute is not a common noun used as an ad-
jective. Similarly " adjectives which do not directly qualify such Taddhita-
formed nouns, but do so through the medium of a noun of class, do not
agree in gender and number with the Taddhita-word. Thus tfaraT:
*\HHi\ r*nrfHt ^f^ ' th e country of Pafich&las is pleasant and fertile/ Here tlie
words TVifa: and ^JiT: qualify directly the word ' SRqj: 1 and indirectly
the word PanchtU&s, and therefore do not agree with the latter.

5

Digitized by

Google

102 v LUP-ELISION NOT. APPROVED,. [ Bk.; I. CH. II; $• 53v

The. explanation of this sutra as given by Patanjali.in the. MahA*
bhAshya is somewhat different from that of the KAsikA. The phrase m*rnfc : *
is analysed by. hira as n and -*JMI^ :, the sense being *!and as. far, as the.
jati goes." The word jAti is thus explained : — m

A word expressing whatever is distinguishable by. its form or figure,
or.which being the name of an individual is easily found applicable to. others,
without any further teaching, (a common name) provided it is not. a word,
used in all genders, (common noun 34*4*{ift genus); and (only, for the
purpose of grammar as coming under the rules relating to the above mention*
ed words), a word expressing descendants by their parentage as well as a word
expressing a person by the branch of the Vedas ( sirar ) which the family to,
-which the person belongs is competent to read. W<g(iui<g (sAkhAdhyetfi)
OliuMlft ) - ^ifiHl^*!: (jativAchakAfc) sjTgjft ' y^nfl'fcifStiiwI .^T whl^l

The following examples will illustrate the meaning of the above de-
finition:— (i) HZ is ajAti word because it denotes what is distinguishable
by its form, i. e. by its being contiguous to a place covered with water, (2)
TOT is such a word, because when given as the name of an individual, is
applicable to other sudras (common name), (3) ^jtarcr is such a word, and
(4)^7 a person belonging to the family competent to read the branch of
the Vedas called katha is also such a word. But $ 5 3*3 is not such a word,
because it is a proper name applicable to only one individual and trf? is
likewise, not such word, because it, though a common name, is used in
all genders.

Varti — Prohibition must be stated in the case of nouns formed by.
the W{ elision of an affix relating to man. Thus the Taddhita affix 3T3j[ (gff)
is employed in denoting representation by imitation. Thus ^J^^r means .
a figure like a horse (in wood or clay or paper). But when the imitation is
that of a man, this affix is elided. Thus ^[f means a " doll of straw." But
when we want to say a doll in the figure of a man, we will have ^fr+^rTOT .
(hhtto} V. 3. 98) « ^jr. Here this word is formed by the elision of an affix
relating to man. Adjectives qualifying this word will not agree in gender and
number. Thus «^r 3jfW<T: ' the delightful straw figure of man.' ^fVfarr: fy
{3Fffc: * a beautiful representation of an eunuch* 9 ' S

M*U*W*THj U

Digitized by

Google

BK. I. CH. It § 53. 54- ] V "LUP.ELISION NOT APPROVED. 10S

53. This (concord of gender and number, ot
primitive and derivative nouns, and of attributes and
substantives* need not be taught (or approved), because it
has the authority of sariijM (or conventional term or
idiom).

The agreement between the number and gender of the nouns and
attributes ; of primary and secondary words formed by suffix rejection, is
not an approved syntax. The gender and number of particular deri-
vatives or adjectives are not blindly to be fixed by the primary words
or the governed substantives ; but it is a matter more or less of usage
or idiom: and no hard and fast rule can be laid down for it. Thus the word
CTCT: 'wife' has the form of masculine plural, but is always applied to females,
similarly ^jr?: 'water' is always feminine plural in form, so also *JfT 7^h
faehdr t- This sutra, therefore, modifies the former sutras, and leaves the
whole question of syntactic agreement to be decided by usage and idiom.

The words like PanchAtes, Kurava &c, should not be considered as
derivative words formed by the lup elision of Taddhita ; but they must be taken
as proper nouns and appellatives of certain countries ; consequently the
gender and number of such words are fixed by the nature of such words and
not by any rule of grammar.

^msrr ?fir 11

54. The s(itras declaring lnp elision need not
also be taught, because of the non-currency of the etymo-
logical meaning of the words supposed to be formed by
lup elision.

In explaining certain words, it is a round-about way of going to work
to say that there was a suffix after them, which has been rejected, without pro-
ducing any effect. It is easier to say that the form is what we find it to be, and
that its peculiar meaning depends upon usage. Moreover, in many cases,
the would-be etymological meaning is not in fact the current meaning which
the word has in ordinary language. It is therefore a useless circumlocution td
presuppose the existence of a suffix, to add it to a word, and, then after all,
to make it vanish. ^ f'" " ~

Digitized by VjOOQlC

104 PjRADHANA AND PRATYAYA. £ BK. 1. CH. II. § 55,56*

Thus *TOT: Varan&b is the name of a city." To say that it is derived
from *TOT: a kind of tree, because the city is situated near such trees, and
the trees gave the name to the city, is a cumbersome process. Because there
might be no trees <|<UJI at all near the town Varani. Let us therefore call
words like q r qjf , TVRTr, &c, as simply proper nouns. Therefore there is no
occasion for the application of the Taddhita affixes given in IV. 2. 69 and 70
(WT f^rrcr:, STgrW^), relating to residence and vicinity; much less of rules
relating to Mup' elision contained in IV. 2 81 and 82 ( SFntsyj, ^T^lTf^TW).

55. And if the etymological meaning be held
authoritative, then when such meaning is absent, the
word also should vanish.

This sfttra strengthens the former sfttra. If it be said that a word
always retains its etymological meaning in current language ; then when there
is an absence of such etymological meaning, then the word itself ought to
vanish. But we know, as a matter of fact, that, there are words which have
altogether lost their etymological sense, and connote a thing different from
their radical meaning. Therefore to explain such words by hip elision of affixes
is improper, because they do not retain the meaning of their primitive words.

It has already been said that words like <raT5RP &c. are not deriva- •
tive words, &c. but are appellatives. If these words be taken as proving
their etymological meaning, that is, as expressing or dependent upon, the
sense of the original word, then when there is a loss of the original word
signification, the secondary word must also cease to be employed. Thus if Pan-
chilis be the name of the country in which the Kshatryas called Pafichilis
live, then when they cease to live in this country, it should cease to be called
Panchalis. But we know, as a matter of fact, these words are now used
without any reference to the people who once inhabited the countries." These
words are consequently not derivative, but their sense is determined by
usage. They are what may be called ^f3r words.

Bk. I. Ch. II. § 56, 57. ] Kala and Upasarjana. 105

56. (SSor need be taught) the rule relating to
the dependence of the meaning of a word on the princi-
pal (pradhana, in a cojnponnd) or on the affix (pratyaya),
because the authority of the meaning (of a word, com-
pound or derivative) consists in something else.

This sutra may be divided into two parts. The first part stating the
proposition and ending with the word vachanam. The second part giving
the reason. The first part consists of four words; Pradh&na: 'Principal
in a compound. 9 As in the compound ll*iq<*H: ' king's man/ the word raja
is Upasarjana ; and <T5TC is pradhAna." 2. Pratyaya : ' affix. 9 3. qpi' f mean-
* n g-' 4- 4^*i4 ' word or sentence/

In the time of PAnini, there were some grammarians, who held the
opinion, that the meaning of a word is determined, if a compound word, then /'* J
by its principal word ; and if a derivative word, then by the suffix. PAnini * ^
controverts this opinion by declaring that in as much as the meaning ofa^ r
word is fixed by usage and idiom, therefore the proofs for the meaning of a ' „V»
word are not to be searched either in the affixes which go to form it ; or by its '
principal word, if it is a compound. Because, says PAnini, the proof of the
meaning of a word is in something else, that is to say in the usage of a people
and not in the suffixes &c. A person who has never heard the name of
grammar, understands full well when told to bring a Raja-purusha. He brings
an official and not a king, nor even any person in general* Similarly when
told to bring an 5rPPH he brings a cowherd's child, not a cowherd, nor a child
in general, nor both. When, therefore, the sense of a word is determined by
usage, there is no use to strive to find out the sense by grammatical rules.

57. And a rule fixing the meaning of Tense
(kala) and Upasarjana iseqnence) is equally (unnecessary,
and need not be taught).

Here PAnini controverts the opinion of those grammarians who
would even define time and sequence. Thus there were authors who de-
fined pluperfect as a tense relating to time hundred years past, others said
a thousand years past &c. PAnini declares that these are redundant
definitions and reprehensible, for the same reason as given in the previous.

Digitized by VjOOQiC

"106 Singular denotes THfi plural. [ Bk. I. Ch. II. § 58.

To give another example. Thus *nwt. 3RT5T: or the period of
the current day, has been thus defined by some Acharyas*: — "The period of
time beginning from the point ^when one rises (or ought to rise) from his bed,
according to the law, and ending with the point when one goes to bed,
according to rule, is called adyatana." Others say "from midnight to midnight
is the period of the current day." Similarly others have explained the
1 upasarjana' to m^an '"that which is not the principal word in a com-
pound is called upasarjana." The sage PAnini has not thought it worth
his while to define such words as upasarjana, in as much as their sense is well
understood by all as a matter of usage. Thus common people, who have
never heard of grammar, also know the correct use of tenses, they say.
" This is to be done by me to-day, this will be done by us to-morfbw,
"•this "was done by us yesterday."

Similarly common people also know the use of the word upasarjana,
for they say : — " *m*m tft *tt% 3T 4<{4i4*11 ' we are secondary or unimportant
"persons in this house or village/ meaning that they are not the principal
persons. Therefore that which is valid according to the usage of the people
is needless to be taught.

Why was not this sutra included in the last ? What is the necessity of
making a separate rule of this ? The answer is that it is an illustrative sutra,
and does not exhaust the subject There are other rules made by former
sages which are also not taught by PAnini. Thus the following rules of
old Rishis are not taught herein, they being supposed to be well-known.
"A Bahuvrihi compound connotes ownership." " In Dvanda compounds the
'sense of both the members of a compound are principal" " In AvyayibhAva
compounds the sense of the prior word is the principal " &c.

5ngnwmfr^T^gg^ wM^d<^i*i N ii H* u *mfa »

58. In (a common) name (expressive of) class,
in denoting the singular, the plural is optionally em-
ployed.

The words of this sfltra require some detailed explanation. The
name of a j4ti or species, always denotes one, since it is a collective noun,
and will naturally take the singular number ; by the present sutra such words
may take the plural number also. In such nouns of class, the plural number has
also the force of the singular. The anuvfitti of a£ishyam does not go further.

Bk. I. Ch. II. $ 58-60, ] Dual may denote plural. 107-

It follows from this that adjectives which are. not themselves common
noun, such as *rqrsr: &c. f agree with the noun and will also get the. plural
number. Thus writifV. or wmt *nrn ; " Full grown barley " <cfa*T KTtmn

j^^Jhr: or ^T^Hr m&tur. **?$m: 11

W hy have we used the word jati in this sutra? For this aphorism.
does % not apply to Proper Nouns. Thus 444-4: Devadatta, 4{J{<h: Yajfiadatta.
Why have we used the word Hjimuwi ? A word may be a class noun or jati, .
but it may not be the appellation of a class. Thus the statue of ka£yapa may.
also be called ka£yapa ; this is a class noun, but as it is not the appellation
f*j|444f or name of a class, such words cannot take the plural form and still-
denote a singular sense.

Why do. we say 'ekasmin'? When two or more class names are
compounded as fffrferil ' the nee an< * barley:' the rule does not apply.

Vart : — An exception must be made when a class noun is qualified
by a humeral adjective. As 5[5RT tffrft: *nf: **Pl^T ^oftf •

Trffo*^ II 3**tf! iKM^iMI*JL II

.59. The plural of the pronoun asmad, l If is
used optionally, though the sense requires a singular or
dual number.

Thus "I speak "or "we. speak" (*n? frftftor Tqrinr.), may be
spoken by one person, similarly two persons may either use the dual case or
the plural case, thus *rnrt ^ « we to speak* or **f ipn f we speak. 9

Vart :— There is prohibition when the pronoun of the 1st person is-
qualified by an attribute :— as *Tf ^frft IHfH% . ' I Devadatta am speaking.':
*li TF'ff frtW ' I GArgya am saying.* Here we cannot use the plural.

Vart: — The word g^ ' master' may have plural form though referring
to one person, when following the word *[«If ; as *t 5f *j^: or ^jf % ipcr^ l Thou
art my master', or 'you are my master. 9

60. And the dual of Phalgunt and Proshtha-
pad&, when* signifying asterisms, (also connotes option-,
ally plural).

Digitized by LiOOQ 16

•?

108 SOMB RULES OP NUMBER. [ BK. I. CH. II. § 61-63.

The word ** in this aphorism draws in the anuvptti of the word
fpf: from the last. Of the stars phalgunf and prostyhapadA, the forms may be
either dual or plural As *f$t <jjf qpp*r?4lr (dual) ; or sRprr: «pfc SRMf«^: (plural.)
So also ^5f ifrOTfc or ^jffr iJTSW: II There are two pairs of stars of the name of
phalgunt and prostyhapada, both being of feminine gender, and their logical
number being ' dual. 9 The present sfitra ordains plural optionally. If phalgunt
and proshjhapada are not names of asterisms they must have their proper
number : as <sqpft »mufNj II

61. In the Vedas, the two stars, Punarvasft,
may op|)tioiially be singular, (and connote a dual).

In the Vedas, the star pxinarvasd which is always dual in form, may
be in the singular form and connote a dual meaning. As 'grfg^RcR or
^^•UHlftfl^dl II The option is only allowed in the Vedas and not in
profane literature. In the latter it must be in the dual, that is, JT#^ f l Similarly
when it is not the appellation of an asterism but of a man, there is no option.
As tprt^HTOTOCT •

sron* ) n

62. In the Vedas, the two stars Vi6akh& may
optionally be in the singular number.

The word ftftjnsr is in the dual number as a rule. In the Vedas,
it is found sometime to have the singular form, denoting duality. Thus
f%JF3T *mn\ or flnjr^ sftMftHM 1 *!) $^nT II

?fH: 11 f^ryr^TR?Rf%^^t <n^«iw^ Rr^r ft*^f *renr 11 .
63. In the D vandva cc mpounds of the stars Tishya
and PunarvasCl, the dual constantly comes in the place
ofthepluraL

When there is a collective compounding of the words Tishya and
Punarvasu (meaning two groups of stars), the final compound is always in the
dual. Thus sftcft AM^H^t^^ 'the Tishya and Punarvasu having risen are

Seen (dual). 9 Digitized by C OOgk

Bk. I. Ch. II. $ 64. ] EKA5ESHA. 1W

There is one star called Tishya ; and there are two stars called j
Punarvasu. As the Dvanda compound of these stars refers to more than
two stars, it ought to have been in the plural number. This rule teaches *
that there should be a dual number, where otherwise there would have been |
a plural number. Why do we say 'Tishya* and 'Punarvasu'? Observe
f^jTOT3TT>IT: ' the stars Vis&khd and AnurAdhA/ The compound is in plural
number.

This rule would not apply if these words do not connote stars but
are used in any other sense. As fa*zr3pfrH%! *UUjq4ip » ' Boys called Tishya
and Punarvasus.'

Why is the word ' star 9 repeated in the sutra, when the previous

sutras and the context show that stars were being treated of. The repetition

is for the sake of including the synonyms of Tishya and Punarvasu. As

h^Sh^t; 3*^5^%^, R^X^U*

The rule only applies when these words form a Dvandva compound,
and not when they are compounded in any other way. Thus ftiq<j«i5444: is
an example of Bahuvrihi compound meaning " persons who confound the
Tishya with the Punarvasu."

The rule only applies to the plural Dvandva. If the Dvandva takes
the singular termination, this rule has no application. As Rr^rynf ^f *Trfl%.

This sutra indicates by implication (jn&paka) that ''every Dvandva
compound may optionally take the terminations of the singular 19 for otherwise
the employment of the term q$qHHtq in the sutra would be superfluous.

The word ' always ' has been used in the text to show that the
governing power of " option " stops here with this sutra and does not extend
to it or any further.

G4. Of the words having the same form, and
all in the same one case-teimination, the last one. is only
retained.

This is a very important sutra, and deals with what is technically
called Eka£esha (or retention of one). When there are two or more words
of the same form, and the same case termination, standing together, one is
retained and the rest are dropped. Thus f^T: + f^P « f^ft (trees, in dual)

Every individual thing requires an individual word to express it ; one
word therefore could not express many things; to prevent the repetition of

6

110 Ekasesha.' [ Bk. L Ch. II. § 65."

* — — ■ . , a t

the same word to express many objects of similar form, is the purpose of this
sutra, by which one word is only retained.

This applies only to words having the same form and not if they
have different forms ; as JRT ?^nfhn: the plaksha and nyagrodha trees.

The word " form " is used in the text to show that even if two words
have different meanings, but the same form, the rule of Eka£esha will apply,
as *r?tt (die) + SRT: (eye) + *rer: (axle) =*nTT: Similaiy *(F[f: and HfTC .

The present rule only applies to the retention of one, and prevents
the retention of two or more.

The word ^r (retention) is used to show that there is not a
substitution (Adesa) of one for many, but the retention of one out of many.

The rule will not apply if the two -words are in different case-termi-
nations, though they may have the same form. As <np(i. s.) q% (2. s.)
sron% ; ftrerop^rf (3. d.) "^frsrsiTCTOP^rf (4- <*•) 1 ^% •

^c^-^Pl:, 3^ lj*, faito II

G5. The vjiddha (or patronymic Gotra word
becomes ekaSesha, and is retained, when conrponnded)
with a patronymic word called yuvan, provided that the
specific difference in form between them be in their
signs (affix) only.

The word sesha of the previous sutra is understood here. The word
" yfinA" in the text is in the instrumental case. The yuvan word is dropped
when read along with a Vpiddha word. The word Vpddha means gotra: the
old Acharyas (grammarians) used the word Vfiddha for gotra. The vpddha
(or in other words, the gotra) word designates a person between whom and
original head of the family a son (or a word formed by an apatya affix) inter-
venes. In using together two words, one formed by a gotra affix and the
other by a yuvan affix, the Vpddha is retained, where the difference between
those words is in their affixes only. These two words gotra and yuvan .are
defined in sutras IV. 1. 162 and 163. farof <JhpPjfir *TJr*PjH sftafa 5 **% 3*0 "
Roughly speaking a gotra word is formed from the original noun which is the
name of the head of the family when a son of such person is living; and a
vpddha word represents thus a grandson, or lower than that, but not a son ;
while a yuvan word is derived from such last mentioned gotra word by
another affix and thus represents persons lower than grandsons, g[g

Bk. I. Ch. II. § 66. ] Ekasesha. Ill

Tbe word 'tad' in the sfitra refers to the gotra and yuvan affixes.
The word lakshana 'sign' in the sfitra is synonymous with 'cause' or
'occasion'; meaning the specific difference is caused or occasioned by the
gotra and yuvan affixes. The word %^ in the sfitra means 'if? The word
S* ' only* is used to exclude others. The word f%^r. in the sfitra means the
'dissimilarity of form.'

Thus * i mJ: + irrrqfcffTr : « irrnjf. Here the first word denotes a clan
called G&rgya (the offspring or grandsons of Garga formed by the affix ^T (IV.
I. 105) l the second word denotes a younger branch of the same, and is formed
by the Taddhita affix qrar denoting a secondary derivative. - In this then, the
first or " Vriddha " remains ; the second or the yuvan is dropped ; and the form
" GArgyau " means both the old and the new clan. Similary *fTC*n + qKWKM?

This EkaSesha is to be done only in those cases, where the two words
have the same radical form, the only difference between them being in their
derivative formation.

The words must differ only by their affixes, but their radical element

must be the same. The rule therefore does not apply to : — irp^: (a Vriddha
word(- MicHU€H ; (a yuvan word)ti4||'2|1l4t4l4H) ll

This rule of ekasesha does not apply if one of the words be not a
Vriddha word. Thus Tpj-: + JIH^HUJ: - *T ?HII*JHuft II

Similarly vice versa *TT^: + *P ! h a »'Tr f 3ppff I Here the one word is
Vriddha, but the other is not yuvan.

The word " only " is used in the sfitra in an exclusive sense,* # tbe
force being, that two words may stand in the relation of Vriddha and yuvan,
their radical element, 1. e. form may be the same ; but if the secondary word
has another sense over and above the sense of a yuvan affix, this rule is not
applicable. Thus Rule IV. 1. 148 8 says a " Vriddhi word may optionally take
the affix 33J to' form a yuvan word and to indicate contempt and born in the
country of Sauvira." Thus H Wfa fiT: + Wlftft*: - HHTft fiWMlfliq*V
Here the second word, though a yuvan word, connotes the additional sense
of contempt and degradation. The affix sa£ is not merely a yuvan affix but
conveys the other sense also. The pure yuvan word being HPTftw^PT .

N. B. — The system of domestic economy among the ancient Aryas wns considerably differ-
ent from the prevalent one. The surnames or the gotras represented the patriarchal system. Three
forms of surnames are constantly met with in ancient literature. The first was most important : It
ws the patronymic by which the lord of the united family the patriarch was known. For example,
Garga or Gnrgacharya was the recognised head of all the Gargas who may be a hundred. His eldest
son and heir wns called Gargi. This is the first form of the patronymic (apatya). The grandsons or
sons of Girgi were named Gargyas (Vriddham or Gotra). On the decease of the old patriarch,
Gargi began to be called Garga ; and the eldest of the grandsons was called Gftrgi ; and those
great grand-sons who belonged to the family, and were designated Gargiyanas (yuvan) were now
ci lied Gargyas."

(') *Pirf|*?t *pj 11 (•) *rf?T5?rv n (') ^*Tj*^hWf wjwj n

Digitized by VjOOQLC

112 _ Ekasesha. [Bk. I. Ch.IL §.67.

66. And so also a feminine word, ending
with a vjiddha affix, when similarly spoken along with
the same word but which ends with a yuvan affix, is
only retained ; and it is treated like a masculine. .

When a word denoting gotra and in the feminine gender is com-
pounded with another word denoting a yuvan, the first is retained and the last
is dropped; and the feminine noun so remaining is then treated like the
masculine. This aphorism makes an addition to the previous sutra and is
governed by all the conditions mentioned in that sutra. Thus *Tntf + 4||«$l4Uj:
«=*TF4fr. Here note that the resultant form is the same as in the former
sfitra namely the masculine.

Similarly ^ft+^WHniTI^SRTT (masculine dual of ?rftr).

Vart — The word strixn the sutra means VfiddhA stri, that is the eldest
daughter of a grandson, or a further descendant considered as the female
head of the family. The words ' tal lakshanas chedeva viSeshafo ' of the
previous sutra govern this sutra also t. e. the distinction between the two
words must be in their formative element and not in any thing else.

g^T^f^nn n 39 11 ti^TPi » ^x% fercr ( 5fa: , <rer-

G7. A word in the masculine gender, similarly
spoken along with the same word, but ending with the
feminine affix, becomes ekasesha, and the latter is
dropped.

When two words of different genders differ only in their derivative
elements but not in their radical elements, that is to say, one is masculine
and the other feminine because of certain affixes, the masculine' only is
retained and the feminine is dropped.

The governing force of the words ' vpddha ' and 'yuvan,' of sfitra
65, does not extend to this sutra, but the remaining portion of sutra 65 is to
be read into this sutra. When a masculine word is read along with a femi-
nine word, the feminine word is dropped and the masculine is retained,
when the difference between the two words is caused only by the feminine
and masculine affixes: as JTTSTO: + W^ft = OT^nift 'tlie Brihmana and the^
BrAhmani'; cfr$j»d : + aj**£| - $ gyfr • ' the cock and hciBOgitized by G( \ „

\

Bk. I. Ch. II. § 68, 69. ] Ekasesha. 113

But not so in the following : — ^^. + 'T^3'= a ^fra^nj4f 'the cock and
the peahen/ Similarly 4|iy*: + ***&& " *IUWHU»4l * Here the feminine
affix 5^» by which the word ganaki is formed from ganaka; in addition to its
feminine signification, has the further signification given to it by Rule IV, 1. 48,
j qmisHs^mi-M • meaning the " wife of " a ganaka and not a feminine ganaka ;
so also ^srv T^CTUft ^-Vftl u *li II Similarly ^n^: + ^rnJr-^rr4J but SErr*: +
-*il4iuft — ^rnfr^P 7 ^. Because the feminine affix in aryani conveys more
than its feminine signification. See Rule IV. 1. 49.

This rule does not, of course, apply when it is not a masculine noun
which is compounded, but it is a noun without any gender or neuter noun
which is compounded with a feminine noun. Therefore JTP£+'Jr'fl , "*n^55T*4t
The word irr*? is an indeclinable and is without any gender* -

fpcf : II WTO" tiKM-HWill^jIt^*) ^Tf^^T STOjf? JS P*R% W

G3. The words bliratvi, i brother,' and putra,'
c son,' when spoken of along with svaspi, c sister ' and
dnhitri ' daughter' respectively are only retained, and
the latter are dropped.

With this sutra ceases the governing force of the remaining condi-
tions contained in sutra 65. Even where the radical elements of two words
are different, there may be an ekasesha under special circumstances. Thus
OTT + STO^nRT (brother and sister or brothers), similarly JT+jf^r-yfc,
(son and daughter or sons).

N. B. — The practical application of this sutra is very neatly illustrated
by the following example. There is a verse in YAjfiyavalkya smriri
<^fi%<H^TflmirMM<W*4r which declares that on the death of a sonless per-
son his property passes to his wife, daughters, parents and bhratara^.

Here the word bhratarah, if it be taken as simply the plural of bhratri,
it means " brothers " and excludes sisters. But if it be interpreted according
to this sutra, it means " brothers and sisters" ; and thus gives sisters a right
to inherit property of their brothers. .

^^* M HMq$3*qqKq i« Md<^ m II $« II Tf^rik N

*i^raw f «raq«Qui f ^^ f ^ 9 are? f 4t*Md«*jm , ( Sfar f

Digitized by VjOOQLC

114 _. Ekasesha. [ Bk. I. Ch. § 70, 71.

— ,, " —— — — — ■ — — — ■ ■ 11 1 ■ *

69. A neuter noun, which has the same form,
only differing in affix ; is optionally retained, and the
other is dropped, and it is like a singular number. '

A neuter prevails over a masculine or feminine noun when spoken of
along with them ; provided that the words differ in gender signs only ; but
radically they be the same. Thus ^rrereft , %*&, fa*T <MlM Pujfl f Idleness,
lust, and sleep, indulged in, increase.' Here the adjective ' sevyamAnam ' is
in the neuter gender, though it qualifies the words * Alasya ' which is in the
masculine ; ' maithunam ' which is in the neuter, and f nidrA ' which is in
the feminine gender. The rule, therefore, may be stated in these words, that
where one adjective qualifies several nouns, in different genders, but one
of which must be neuter, the adjective agrees with the neuter.

The above illustration also shows that the neuter ekaSesha may be in
the singular number.

If the neuter is compounded with another neuter, this rule does not

apply. Thus HW+g#+5lsr=5WrPf the word is in the plural; there is no
ekavadbh&va.

The words dM%lUj3ft4 f%$TO "of sutra 65 governs this sutra also:
thus fjs^i **«<H:, ^T^^R^ir.^lf ^qret "The white (masc.) blanket,
the white (fern.) wrapper, and the white (jteut.) garment," may all be spoken
of collectively as *pjj (neut.).

ftCTTOIT II 30 II T^TfH II fqcTT, m^l (^3TOcTTOP*)ll

70. The word pitji, c father' is optionally only
retained when spoken of along with m&tri, c mother.*

Thus firarr + nnrr = fadifr , f father and mother, or parent ' or nnuffiqft.
The verse quoted under sutra 68 illustrates this sutra also. There the word

* pitarau ' has been explained by all commentators as " father and Jmother."
The word " ekavad " of the last sutra does not govern this sutra ; though the
anuvptti of ' optionally * is to be read into this.

WW^) II

71. The word fiva&ura c father-in-law, 1 is op-
tionally only retained, when spoken of along with &va&rft 9

* mother-in-la w.> I ^

Bit I. Ch. II. § 71.73. ] Ekasesha. 115

Thus **gt:+S^jmay be either **qjTT ' father-in-law and mother-
in-law,* or x^M^npcrl

fw* (5fa: ) H

72. The pronouns c tj T ad &c.' when spoken of
along with any other noun, (pronoun other than c tyad
&cJ) are always retained as ekaSesha, (to the exclusion,
of others).

The list of ' tyadAdi ' pronouns has been given under sfitra L I. 27.
They are 12 words beginning with ' tyad' and ending with ' kim/ when these
tyad&di pronouns are compounded with any other noun, (or a pronoun, which
does not belong to their class), they prevail, the rest are dropped. The
word ' sarva ' is used in the aphorism to indicate ' universality ' ; r. e. what-
ever may be^the word in composition with them, the 'tyad &c f are' retained
to the exclusion of others. The word 'nitya' is used to indicate that this
is an obligatory rule, and not optional, as were the last rules.

Thus ^: (he) + \\\^x (Devadatta) - fft (they two).

Vart : — When nrcrft words are compounded among themselves, the
word which is read last in the order given in the Ganap£{ha remains, the
rest are dropped. Thus *n+^:«*rr; ^: +^1*^11 Thus ^| is read after
*T£ (see I. 1. 27), so*r$ will be retained and not fif ; similarly f%*^ is read
after ^f, and therefore fc\ is retained to the exclusion of irj[.

^1^3 , ^ , ( fr?: ) «

73. The feminine noun is only retained, when
denoting a collection of domestic animals, not being
young.

As a general rule, when there is a compounding of masculine and
feminine nouns, the masculine noun is retained (see 67). The present
sfitra is an exception to that rule. It ordains that when there is a com.
pounding of words denoting group of domestic animals, of different genders ;

116 ^ Ekasesha. " [Bk. I. Ch. II. § 73:

the feminine noun is preferred ; provided that the words should not relate to
the young of such animals. The word 'ataruna' in the aphorism qualifies
the word ' pasu * with which it is in construction, and not the word ' sangha.'
As nr* r*T: ' these cows (and bulls).' *pir $Hn ' these goats' (both males and
females).

The rule only applies to domestic (gr&mya) animals and not to
toild animals ; and only to beasts (pasu) and not men. As ct? f%, ' these
wild deers' (male and female), $^rr f^r, ' these antelopes ' (male and female),
ICTSmrr:, ' the Br&hmanas ' (male and female) Vlfaqf :, ' the Kshatriyas.' In all
these, the masculine noun is retained by Rule 67.

Vart* — This rule must be confined to domestic animals having
divided hoof. Therefore vj^il f%, ' these horses ' (male and female).
Here the masculine noun is retained, because horse has no divided hoof.

Why have we used the word (sangha) " collection " ? Observe ^^
*n4V ^C<n, ' these two cows are grazing.' Here the word ' g&vau ' does not
connote ' bull ' also.

Why do we say " not being young." ? *c*?r & ' these, calves ' ;
^*tr f% ' these kids. 1 Here the masculine noun is retained, by the applica-
tion of Rule 67.

c$'^

*H1

BOOK I.
Chapter III.

^T^^t tow: it 1 II tj^ifSr ll ^l-an^H: f TOW: n

1. The words beginning with bh& c to become/
and denoting action, are called dhatu or verbal roots.

This defines the word dhAtu or root The list of Sanskrit verbs
begins with the root bhfl ' to be. 1 Sanskrit grammarians have divided all verbs
into ten classes, according to certain modifications which their roots undergo
before certain terminations. The ten classes are as follows :—

i. Bhft class. 2. Ad. 3. Hu. 4. Div. 5. Su. 6. Tud. 7. Rudh.
8. Tati. 9. Krt. 10. Chur.

There are nearly two thousand verbs, and they are conjugated in two
forms, some taking the Parasmaipada terminations only, the others the
Atmantpada terminations and some both. The present chapter deals mostly
with rules determining the nature of the verbal root, whether it is Parasmaipadi
or Atmanepadi or both. A verb expresses action. The word dhAtu is a word
coined by older grammarians than PApini and they employed the term in
denoting a word expressing action. Here also therefore, the word dhAtu
means a word which expresses action.

The qt in the sfitra is for the sake of auspiciousness ; for the regular
sandhi of l+*nft is *TTft and not *prf|; while Dr. Ballantyne considers that J
*T is a separate root: he translates the sfitra thus : — " Let the verbal roots bhA
1 be,' vA ' blow ' and the like be called dhAtu."

^gffii3T 3 ? Hfa *E m II ^ II TT^lft || *T^$t 9 3I^ f Snj-

^W II OTfaf^TJ^fifr 3*rt*T:, 5llfel4l<f3lfif H^TOT: R&M4IAV *m ^fW^j-iifii-

2. The nasalized vowels are * n in Upade&a, or
original enunciation.

118 _ Indicatory letters. [ Bk. I. Ch. III. § 3, 4.

Anunasika or nasalized vowels are f?£ that is indicatory, when occur-
ring in technical words. Thus in fa*, f%^, f%^ the anunasika f is fsfc
The original sfitra contains the word ' upadesa ' which we have translated as
technical term. 'Upadesa' literally means 'instruction' or the first mentioning
of a thing either in a s&trap&tha, or ganapAfha or dhatupi^ha &c

From this sfitra up to sfitra 9, there is a description of servile or indi-
catory letters called f* " In PAnini's Grammar there is no visible sign of the
nasality of a vowel — hence we can know a vowel to be nasalized only from %
Panini's explicitly asserting that it is so, or from our finding that he treats it
in such a way that we must conclude he regarded it as nasal." An upadeSa
is defined as signifying an original enunciation, that is to say, an affix
(pratyaya), or an augment (agama), or a verbal root (dhatu), or in short, any
form of expression which occurs only in technical treatises of grammar, and
which is not a word ready for use, but one of the supposed original elements
of a word. Thu$ in the root sttj- ' to increase,' the final qf is indicatory, the
real root is 3^. So also the final sjof *?s$ is ^r II j

If the word is not an upadesa, then the nasal vowel Is not fir II As
: ?ra *ft W: II A word may be an upadega and nasal, but if it is not a vowel,
it need not be jw as the affix *rf%* in sfitra III. 2. 74. ( srrtft yft%gft*r qft<*V ) -
It is only the nasal vowel of an upadeSa that becomes ^ , and not all the
vowels. The word f?r occurs is sfitras y. 2. 16. ( mfotV &c).

f«*w* n ^ 11 ^t£i 11 ^w^ ( g^ifr %qj it
\P*z 11 \n*ri$: *T5*nre*r ^w* *% irfcatf *rcfa 11

3. In upaclega, the final consonant of roots &c.
is *w.

All the final consonants of roots, affixes &c. are indicatory. As in the
pratyahara sfitras ^ T ^ ^; lierec^ is *wil ^ 5* sr; here a* is *w Hit a|fcr;
herey is ^5 11 $ VfVS; her? ^ is 5^11 It is only in upadesa, that a final
consonant is X* . Not therefore, in <*rftft^ or *ta§^, which are complete
words.

* fvmst H^n: II 8 II ^tOi II H fafRR*, g-«-*T: N

%f*: 1 1 <nfa smmiTft^>rRrf f%H^ v^sirit) *wfr writ *nKrrptf qftft?
smrcii

4. The final dental consonants, and the final
e , and n, are not ^, in affixes called vibhakti or inflective
affixes.

Digitized by VjOOQtC

Bk. I. Ch. III. §5.3 Indicatory letters. 119

This is an exception to the previous sfttra which declared all final
consonants to be ?^. In case terminations, the final w, *^ f ^, ^ f ^, ^ f and ir f
are not f^ m Thus the case termination of nominative plural is 5f^. Here the
final H , though a consonant is not an f^; and is not rejected, as in MIIUUI: II
The word 3 in the sfitra means the letters of the class *, by virtue of sfitra 69
Chapter I. (see ante). The word vibhakti will be defined later on in sfitra
I. 4. 104. These are terminations applied to nouns in declining them through
various cases ; and to verbs in conjugating them. Thus in sfitra VII. I. f 2,
'4tef*l*4llf'Hl*b4i:) the vibhaktis f*[, *inj and *^r replace 2T, vfa and 3^. Here
the final f£ of the affix W*[ is not indicatory and hence not rejected, as f^TIW .
Similarly verbal terminations ^ and *j^ as <p*PT: f T^T II So also the final
H as trnj and ?PJ in *ttoipt and ^TT^T^ II

This exception applies only to vibhaktis and not to affixes in generaL
Thus the finals of ^ (III. 1. 97 *p^T*;) 3« (V. 2. 123 3TCjfar 3^) and mr
(III. 1. 78 F^TTf^^Wl) are ^ and are rejected. This exception, however
does not apply to the vibhakti «pj (in V. 3. 12 ftftJTJ^ or III. 4. 106, ^4^)- I«*
those two sfitras the final ^of the vibhakti is fw. The reason of this is, that
the present sfitra is an anitya sfitra, i. e. a rule not of general application. la
sfitra V. 3. 24, (f^rwj:) we find, c the word fj* takes the vibhakti «rj in the
sense of mode or manner.' As fi^+trj-^w^, 'in this manner. 9 Now tfj
is a vibhakti by V, 3. 1. ( sTP^foft firefor) the ? is ^, the real vibhakti is tj^.
Now had the present sfitra been of universal application, then there would
have been no necessity of adding this ? as indicatory letter, which has been
evidently added to save the ^ of ^ from becoming an fif. This sfitra is con*
sequently an anitya sfitra.

5. The initial ni, tu, and du are ^11

The syllables fir, J and J, standing at the beginning of a root are indi-
catory; as firf^r ' to have affectionate pleasure.' The real root isft^, the fir
is servile. So j*rft ' to rejoice, be glad ' the real root being 7^. So vpn^'to
give ' the real root being jr.

The fif shows tha't the past participle ^ has the power of the present
III. 2. 187 (Tfhf: **:). As pflxrfl — qp^r: 'blown •; filflfar— ft*p ' fat'; fiT|*r— ^€:
' bold/ fsrf^rer-f^rir: « soft/ fsj^^— f^: « kindled/

The J subjoined to a root serves to show that it admits the affix
*jtf to form a noun indicative of the act (III. 3. 89 f|tftj^*); as 5%^— **!p
1 shaking ' ; J^frft^- TTERf: ' swelling.*

The 5 indicates that the root to which it is subjoined admits of the
affix fir* (III. 3. 88, f%w. fa:) to form a noun or adjective atributive of the act

120 Indicatory letters. [ Bk. I. Ch. III. § 6, 7.

"by which the substantive thing has been produced, as £«kii— **ft«^ ' artificial 9

3*^V"*flPfl[ ' "P* ' » m— ^J^R^ ' sown.'

If these three syllables do not stand in the beginning of a verb,
they are not indicatory ; es «rrjrf^, gfp^rft &c.

m vphf® 11 % 11 r^jfk w *: f u<nnw, ( snfi[ *<ij u

G. The initial h, of an affix is indicatory.

The initial ir of an affix gets the name of f^ and is rejected. Thus
by Rule HI. 1. 145 (fafarf* cjsr) the affix sj^ is added to a verb to indicate
proficient in the art expressed by the verb. The ^ and *[ are j£ and 977
takes the place off (VIL 1. 1, g^prerafr )- Thus ^+«jsr*«r§r+*^««rfai:
• a dancer' (by profession) *fem. qfagt (IV. 1. 41, foftnf^lV ' words formed
by affixes having an indicator)* «[ take the affix ^[ in the feminine.') : so
also i"^«h: fem. oraft.

The initial «£ of an affix only is f% and not every initial *. Thus not
in qfrT" ' six/ or qft«K: II The ^ must be initial, therefore it is not JJ in the
affix fzqr^ in *rf%T*nfff<r9, (Un I. 45) as srfiTO ' an ocean ' *f^r: * a buffalo/

55 11 s 11 vyfk n 3-5 ( uwmu snf^: f ^) u

7. The initial palatals and Unguals of an affix
are indicatory.

The consonants ^, f^ f sr f 5^, 3j, ^, j, j, 3:, and «r are always
*c£ and are rejected when initial in affixes. The word 'affix ' of the previous
sfitra governs this also.

As the affix *qn£ chphafi in sfitra (IV. 1. 98 nt% gf^rf^rVKTj). Here
^ is indicator)-, as, e^granp II The initial ^ chh of an affix is always replaced
by f*r (VII. 1. 2 muStfMlfi m: <W^r«5^f g «<w{Nlf* ). The initial 7 of an affix
is*f£. As. *f$ (Nom. PI. term.) srnjnrn:: || The initial ^of an affix is always re-
placed by wr, as *l+*r$+fir - *T*f«r II (VII. 1. 3 tfTJ^T:). The initial^ of
an affix is ^ as in t^ which comes after the words *rf*T3* &c. As 4llf*>444:
(IV. 3. 92 \lfi^^»R»4i^ :)» The indicator)' *r causes vriddhi by rules VII. 2. 115,
116 and 117: (^^^PhPr, *I*imnn :, *l^«M , *l | lftO«

The initial ^t of an affix is f% as in III. 2. 16 fafe: the root ^£ takes
the affix z when compounded with a noun in the locative). As <£**H: 'who
goes among the Kurus.' The feminine of words formed by this affix take
#*X(IV. 1. 15 ftrgpr^jp?*n3t &c.) As *wQ, wmfc The initial 5 of an affix
is always replaced by f^j (VII. 3. 50 *TO3n.) The initial ^ of an Sflke is t/„ ^j)
as in III. 2. 97 (OTTOt *T^T: the root "sr^ takes the affix 7 when in composition

iW-

Bk. I. Ch. III. § 8.] Indicatory letters. 121

with a word in the locative, as 3TODT: , H*|i*f : . The V indicates that the
final of the word called fi should be elided when an affix having indicatory
d is added. The initial 5 of an affix is replaced by ^r (VII. 1. 2)-

The letter 5 is always an ft£ when initial. As in the aphorism IV, 4
85 fa^TCf:) the affix ur comes after the word wr ' food.' As inr + ^-Uraf: .
The p causes VriddhL

This sfitra could very well have been included in the last aphorism.
Why has it then been made a separate aphorism ? . The reply is that the
present sutra declares an anitya rule — that is to say, a rule which is not uni-
versally true. The initial palatal or lingual . of every affix is not indicatory.
Thus in the affixes ^^[ and ^r^the initial ^ is not indicatory but forms
part of the real affix. As fi 44133 : f%*Trr*: ' famous for one's learning/ Se£
S. V, 2. 26 >r ftrr^p^^nfh Similarly the initial ? of the affix Hz%^ (V.
2. 31 migffill'W , srif?l fa<*i i qr : *U»nf ?fe*r IHir S?^), is not it. As «TT*te£
'flat nosed/

Of course it is the initial palatal and lingual which are servile ; if
these letters occur in the middle of an affix, they will not be indicatory. As
? of the affix *|?^ is not indicatory in sfitra V. 2. 35 favrffa **lrJ**r) ^fa:
1 dexterous/

*"% 1*0 ■

\fa 11 qfltmfijfW i q<m*H3ift<D i *4mm w^rr *rcrt «*irf %*&& to£? 11

8. The initial 1 and £, and the gutturals of all
affixes, except taddhita, are indicatory.

The initial g, ^j, g£, ?qr, i£, tj, 5 of affixes are indicatory, except in
Taddhita affixes. Thus the initial fj in ra^ is m(III. 3 1 15 *33£J l) Ft+*J£«"
< *HR^(VH. 1 1) 'collecting/ The initial 5 in jj^ is f^(W. 1 68 qftrfirjrj)
^+^rSL+fi r X s= ^ , +M+f?r= s Wfit 'he is/ The initial ^isf^as in *E and ^3
(I. 1. 26), as 5^: ' eating* ^qi-jK The initial *3 is {£ as in CT^dH. 2. 38
ffcmf W- *&% the word q% takes the affix khach when compounded with f*m
and qsr) ft?rc?: ' speaking kindly ' ^TO ' submissive. 9 The initial *j is f^as
in the affix r^ (HI. 2. 139 «<HlOwi*| ^5), **rre j: 'languid f f3i*j: 'victorious,'
^mj: ' unmoveable/ The initial ij[ is f^ as in ^c^ (III.* 2. 161 H^pirefift
5T^r) H^H- 5^ « H^fTj 'brittle/ The initial ^ is f£, as in yftr (termina-
tion of the ablative sing.) yj? ditto of genitive sing. (IV. 1 2), as *pFTH+
3 r 5=g , re5+*r3 as 3 , Fre:'of a well minded/ In Taddhita affixes however,
these initials 5*^ and gutturals are not fj; and are therefore not elided.
Thus *rjr+i*^«*prrar: (V. 2. 96 wfiUwNHft ^ni^mrwr^) 'crested'; ?frr+*

122 A RULE OF ASSIGNMENT. [ Bk. I. CH. III. § 9, IO.

— ^W.(V. 2. 100 <Hl*HRqtHlfifl*tdfo *r: ^§15) 'hairy;' f^r + qr^» T?TO
IV. 3.-8 1 -86 *nftTOK **% ) ' a small tree/

?tct sto 11 ^ 11 tT^rfSf n «^r f sita:, ( *qj n

9. Of this, (namely of that which has heen
called *h), there is elision.

This sutra declares the function of fijj namely it is a mere indicatory
letter, and must be rejected. It is useful only as a mnemonic and is not a
part of the t$rm.

The word tasya in the siitra indicates that the substitution of lopa
must be in the place of the whole term called fj, and not only in the place
of the final letter, which last is the general rule (see. I. 1 52). Therefore, the
whole of fsy, 5 and J is rejected and not only their finals.

*j*n q^qjfe n wnw ii 10 r h^tto ii w-^ f

^frT II HTffif *M*I^Hl 4W4f<4f£dHI*f^fiH IHH^f^Rf ^7IT 9iH^Rl(^ij$R|«|«

10. "When a rule involves the case of equal
numbers of substitutes and of things for which these are
to he substituted, their mutual correspondence or assign-
ment of each to each, is according to the order of enum-
eration.

Thus siitra III. 1 134 (Hfaqfojq i ffo ft »ip%?ro) declares :— f the
affixes Fg, f*n% and *f\ are applied to srf£f, ^ and t^ class of words.' It
means the affix fj is applied to the words of jrfSj class, the affix fafir to words
of irf| class, and %[% to words of T^ class. The application must be respec-
tively according to order, and not hap-hazard. Similarly in VI. 177 j^fr
*HJjf^T (it a vowel follows) in the place off, ^, 3?,^ there is ^r«r 1. *., %, ^, ^,
* II So also see S. IV. 3. 94 (^ l^^lrji^^^^Kl^ g^Bg r^^rer:) where the
four affixes dhak, chhan dhafk, and yak are applied respectively to the words
tfidl, £al&tura, varmatt and kuchav&ra ; u e. 9 the first affix in the order of
enumeration to the first word, the second affix to the second word &c.
As *H|*r:, UfoUgOq: , fJ^fcj: and 5ifoCTc£: l f

Why do we say ' of equal members ' ? This rule will not
apply if the number of substitutes and of things for which these are
to be substituted are unequal. As in sfitra I. 4. 90 ^rdfq^n^TPPimfNrr
9ft q&PT for here the words lakshana &c in the first part <

Bk. I. Ch. III. § II, 12.] Atmanepada verbs.

1S3

the sfitra are four in number, while the words prati &c. are only three. That
sfitra therefore must be read as thus, " The words qfir, qft and *r«f are karma-
pravachaniya, whenever they indicate either lakshana (a mark), itthambhut4-
khy&na (a statement of mere circumstance) ; bhAga (division); or vipsi
(desire)."

^iRu^Tf^^n:: II 11 II V\rfk H ^ft^T, ^f^^i<: H

11. In these aphorisms, when a word is
marked with a svarita accent, by that an adhik&ra or a
governing rule is to be understood.

When in this collection of grammatical sfitras, there is any sutra
that has a circumflex accent, it denotes that it is either the beginning of a
subject and the subsequent sfitras are governed by it, or that it ends a sub*
ject and separates the previous sfitras from the following.

As a rule, the sfitras are not marked with accents ; it is therefore
from commentaries that one has to learn whether a sfitra has a svarita or
not. Thus the following sfitras, must evidently have svarita as they are go-
verning sfitras; qnra: (III. I I.) ^rat: (III. I. 91), Mlfr ' W (VI. 4. 1), TO*
(VI. 4. 129).

^llcHrlM^*! II

*jfc 11 w^nHTr 5 >n*rfr fiF*ra w^rwW *nf?r m*$**r. »

12. After a root which has an indicatory anu-
datta vowel (anudattet) or an indicatory ft (Ait), the
affixes are those of the Atmanepada.

The word ' atmanepada ' here governs all the subsequent sfitras up to
sfi. 77 inclusive, and is understood in them all ; and will not therefore be re-
peated. Thus of the verb *?ra the final «f is anud4tta and is f*. It is there-
fore an Atmaccpadi verb. As mR?T. So «££— «^ f tfa£ jtft

Mnrs3*ft: II 1^ II * mfa H WTO-S^ft:, ( sUcH^ ^T) h

^frT- II HT# «h«ifiSj , *Ht*i^ , 4$fJL*nrf% 11

13. An atmanepada affix is the substitute of
the affix <* (III. 4. GO) when it~denotes the action of the
verb or the object of the verb.

Sfitra III. 4. 69 declares the " letter *f C*nO fa* &c.,) is placed after
transitive verbs in denoting the object also as well as the agent ; and after
intransitives in marking the condition (£ e. f the action itself which the verb
imports) also as well as the agent." The Paras, and Atman. affixes would

124 ATMANEPADA VERBS. £ Bk. I. CH. II). § 13, 14.

have come, therefore, generally after all these verbs. The present sfitra res-
tricts Atman. affixes to bhava and karma verbs.

This is clear. Impersonal verbs and passive verbs as a rule are conju-
gated in the atmanepada and take the terminations of that pada. These ori-
ginally were verbs expressive of states rather than of actions. As of *ir* we
have *tSnre WIT 'you dislike' (lit. 'it is disliked by you'} tpr<T WfT 'you
sleep.' Similarly passive verbs as ftr*T% 5*?:'the mat is made ' ; ft"*T^r HTC:
1 the load is carried.'

All verbs in Sanskrit may have three voices vis. active, passive, and
middle or impersonal ; active voice is generally parasmaipada. The reflexive
verbs known as SP^Sft- are also atmanepadi. Thus f&fil <*?$K: STO^* ' the
wood cuts of itself.' See sfitra 78.

fl^ « ( snw**^) «

14. In denoting the agent, when reciprocity
of action is to he expressed, the affixes of the Atmane-
pada are employed.

The active or ^^r^r verbs when denoting reciprocal action are
atmanepadi. As, mffrwft , ' they cook for each other,' c^rfitpnt 'he perform^
cutting of wood which was the appropriate office of another.' ''J'-*^

The words " reciprocity or interchange of action " of tliis sfitra
govern and are understood in the two following sfitras also.

The word karma in the sutra means action and not the technical 'karma '
meaning ' object ' and the word vyatihara means reciprocity or interchange.
When an action which was appropriate to one person is performed by another,
whose duty it was not, that is called • interchange of action.' When such in-
terchange is not meant Paras, is used. As «J*fe*T ' they cut' The word
kartari here is used for the sake of the subsequent aphorism 78, which see.

* ifa fgwfar: II V\ U x^Tfa II * f ^-ffOT-Vlfai: ||

( ^RH^^^^^fgrfir ) ii

^fa; 11 «rniuwit fs*rriifajv ir***: *& sqfatf ^^ * *wf% n

^ifSehflL 11 *<«KiifWta 11

15. After verhs having the sense of c motion, 1
or * injury,' when expressing interchange of action the
Atmanepada affixes are not used.

BK. I.--CH. III. $ 16, 17.] THE ATMANEPADA VERBS. 125

This aphorism prohibits Atmanepada affixes in certain cases where by
force of the last sfitra Atmanepada affixes would have been obtained. As
**lfn*7ufN ' they go against each other/ sqfdnfcfc r ' they injure each other/
iq fo 'd fo f * they fight together/

The word c not ' of this sfitra is understood in the subsequent sfitra.

Vart. — This prohibition extends to the verbs $$ • to laugh/ and the
rest As g^f^rf^r, mft^^Pfl, ^Rl4*Pa II

Vart. — Prohibition must be made of the verbs f*f?f ' to injure,' as
4WC«?Kl*iM:il

^RT: II ^T^dW^T ^frlipw^iql ^vforf^T? «|l*|3?{ H *Plfa II

16. And after the verbs which take the words
itaretara 'each other/ and anyonya c one another/ as
upapada (or dependant qualifying words), the affixes of
Atmanepada are not used, though reciprocity of action
be denoted.

Thus jd?*UHI **tf<* sprf^r / they cut each other ' ^F^inr utfd«f«lP4
' they cut one another.'

Vart. — This rule must also be applied when the word <u'fH is in com*
position with the verb, as an upapada. As «H*3U*J wif^npfl ,

Stfhjt II 3b II q^Tfvr II $:, fo^: f ( micH^H^q ) «

17. After the verb vi& c to enter,' when pre-
ceded by the preposition ni, the Atmanepada affixes are
employed.

As firerfir ' he enters,' but ftfiicft ' he enters in/

Even when the augment ^j» of the Imperfect, Aorist and Conditional
tenses intervenes between the root and the preposition, terminations are or*
the Atmanepada. As ?trf%TOT ' he entered in/

The f% of the sfitra must be an upasarga, because that has a sense,
and not any f%. On this there is this paribhishi : — wi^TOl) s f PnNi^ " a
combination of letters capable of expressing a meaning denotes, whenever
it is employed in grammar that combination of letters in so far as it possesses
that meaning, but it does not denote the same combination of letters void of
a meaning." Thus in irjft farf£? *TO: f the beetles enter the honey flowers.'

2

126 The Atmanepada verbs. [Bk./J. Ch.JH. § i8, 19.

-Here the f% is a part of the wcrd ijgfir and has no sense by itself, and there*
fore the verb is in the Parasmaipada and not in the Atmanepada.

{ vno ) 11

^ftr 11 Hftmfr * qrerergi gftarlft itH^N 4 *&fa 11

18. After the verb krl c to purchase/ when
preceded by pari, vi or ava, the Atmanepada affix is em-
ployed^ even when the fruit of the action does not
accrue to the agent.

The verb f^lffaff ' to buy or barter/ has an indicatory ^ and therefore,
by sutra 72 it will be Atmanepadi when the fruit of the action accrues to the
agent The present sfitra, therefore, refers to cases where the fruit of the
action does not accrue to the agent.

The root ifl" ' to buy ' as a general rule takes the terminations of both
padas, but when it is preceded by pari, vi, or ava, it is restricted to Atmane-
pada terminations. As JflfafffefHr ' he buys ' ; fliftfMKi ( he sells ' ; srqrafhrfl% ' he
buys/

The word pari, vi, and ava must be upasargas, therefore, in the follow*
ing example, there is no Atmanepada termination, because the vi there, is not
a preposition but a noun, as, gff%5?Mrf5 i ^pt^.

fiwronRit §: 11 *% 11 ti^Tft 11 f%tr^i^qt\, 5t: (we ) 11

19. After the verbji c to conquer,' preceded
by vi or par&, the Atmanepada affix is employed.

This sfitra debars S. 78 by which the root flf is generally Paras-
maipadi.

As f%*PT^ ' he conquers ' <HT*J4<) ' he conquers/ The words vi and
pari must be upasargas (prepositions), for the application of this rule. In
the following examples^ they are used as substantives «Kfq*nfll 3W%,

«mft ^tonw $*%*$ 11 ro 11 TRjTfSp 11 are:, ^:, spit.

^I-faf^, ( wo ) 11

qiRf«n^ f i *m*iftt$r i 'KWHftraT3ft *?f?t$>St <Mit*i: 11

*lf^ll^*tf^ft**R«ppi Digitized by GQOgk

Bk. I. Ch. III. § 20, 21.] The Atmanepada verbs. 127

2(X After the verb d& € to give, 1 preceded by
Aft, and when not meaning € to open the inouth,' the At-
manepada affix is used, even when the fruit of the action
does not accrue to the agent.

The root ?r is generally both Parasmaipadi and Atmanepadi (see S. 72).
But when preceded by the preposition sjt, it is restricted to the itamanepada.
As fawn*^ 'he acquires knowledge.' But when it means "to widen the
mouth 19 whether actually or metaphorically, it is parasmaipadi. As VJTOT
^KIlH r* 'He expands his own mouth.'

Vart: — The prohibition also applies when the action is similar to
the act of opening the mouth. As ftmR<hf wrr^lftr ' he opens the tumour/
a£J* g*r H<jlfi *rft ' the river breaks the bank.'

Vart 1 — When the action does not affect the agents' own body the

verb is itmanepadi, as ^W^ft-ilf^C <Jttat< 5?3d 1 ' the ants °P en the
mouth of a locust.'

* > ( W^: , W?q% *o )

ffrf: II sftf ft?T? H<HHIJ ^H^H^R Jtift ^^^T^^^m^T* ^ft It

qrrfSsir^ 11 ^tflw**i^ ffa q^m*^ 11

^rf^^r* 11 ftr^rftr^rraRrr^ 11
TiftNr^ 11 ^nftrPnrnr: 11
qif?!*^ f i $T%*nrcfT^B£Mfr 11

wrfWj 11 v* ^TT5T»pr ?f?r ^raro^ it

21. After the verb kricj l to play' preceded by
anu, sain or pari, as well as &ii, the Atmanepada affix
is used.

The word "W^" is to be read into the sfitra by virtue of the conjunc-
tion ^ in the text. As srnfln^ S*fte% MJj*Cu %, or qfiufhsw 4 he plays/

P*>*'.— When the verb *ft^ compounded with ^ means ' to make a
rattling or creaking noise/ it does not take Atmanepada terminations. As
^wr^cT^RCTfir'the carts rattle or creak.' The word anu, pari &c, being
taught along with sam, indicates that the upasargas anu, pari &c., are to be
taken, and not the karmapravachaniya anu pari &c. Therefore, when these
prefixes are used as karmapravachaniya, they do not cause the verb *ff^ ta

Digitized by VjOOQLC

128 The Atmanepada verbs. [Bk. I. Ch. III. § 31-22.

take the Atmanepada terminations as munqiHjttfUffr I * he plays in imitation
of the boy/ For the definition and action of karmapravachanlya, see* I. 4. 83
and II. 3. 8.

Vart. — The verb hjit^ takes the Atmanepada terminations when mean-
ing 'to wait for\ ' to over look' 'to have patience' as, sjnpnrw *K* *m^ I
' Have patience with the boy.'

Vart. — The verb ftrcr takes Atmanepada terminations, when meaning
* to enquire ': — as, fimg fw% ' he investigates sciences.

Vart. — The verb s?t^ when meaning 'to bless/ takes Atmanepada
terminations, as, ^rfr^ HWR 'he blesses with clarified butter.' iiyfr ?rm% I Why
do we say 'when meaning to bless?' Observe, *iUU4«hHHHU4ft f 'he begs of
Manavaka/

Vart. — The verb f^fir takes Atmanepada terminations when meaning
' to take after the nature of the parents.' As feq f rara r *^<^1 ' the horses
resemble their father.' m<t<£ TTPfr *H^<«-3 'the cows always imitate their
mother.' But when not having this meaning, we have HM^iftj ' he resembles
his mother.'

Vart. — The verb fchifa takes Atmanepada terminations when
meaning ' to scratch out or scatter with joy, for abode or food.' As »rof*3ft%
fTOlrSET: 'the happy bull scratches with joy.' Hjqfehid ajf^fr HW*ft 'the
cock scratches in search of food,' fr^fc^id **I\MM3TOT ' the dog scratches to
make his abode.' When not having these senses we have gpfqK^fH ^HP^C/ he
scatters about the flowers.' The dental *r in apaskirate is added by sfitra
VI. i. 142. •Wwami-H4$flMll3<rt »

Vart. — The verbs 5 ' to cry ' and spsg[ ' to ask ' take the Atmanepada
terminations when preceded by the preposition SJT^ , as m i jft Jjnn*: ' the
jackal howls.' iniHjfl *ns% ' he questions the Guru.'

Vart.— The verb *rx when meaning 'to touch the body by the word'
takes Atmanepada affixes. As ^FfrTRT *TT% ' he promises by oath to'Devadatta/
Otherwise we have wrfH 'he curses.'

22. After the verb sth& < to stand/ preceded
by sam, ava, pra, vi, the Atmanepada affix is used.

As «f?te% * he stays with,' q y qffra^ ' he waits patiently/ irf^% ' he sets
forth/ f%fifar% ' he stands apart/

Digitized by G00gle ■

Bk. I. Ch. JIL § 23-25. ] The Atmanepada verbs. 129

Vart. — The verb w when preceded by *JT£ and meaning ' solemn
declaration ' takes Atmanepada terminations ; as atflsr 4H4tKm-H4lfd8<) « Or to
take another example, sjc* f%tf *T SPC ttilUUKIWIt^ * for thee I shall surely have
recourse to water or poison/

* , ( ^IT , W&&0 ) II

23. After the verb sthil when meaning c to,
indicate pile's intentions to another/ or c to mak e - aft

c k&&biP&& an arbitrator/ the Atmanepada affix is em-
ployed.

The word prakAsana means disclosing one's intentions. The word
sthey&khya is a compound of stheya * arbitrator' and Akhya ' name. 9
A u As frrar^r ^rromi? ' the wife expresses her wish to the husband^ t*fr»

frtg% , ' he refers to thee for settlement/ *ftro Mi^uT^ firrt ^:, 'who, when
he is in doubt, has recourse to Karnaas his judge or umpire.'

ffrT: 11 ^$ 1??^ fi*a<l **&$^*$fa *£hhw*h*N$ ^nft ii

24. After the verb sthst, preceded by ut, when
not meaning 'to get up or rise/ as from a seat; the At-
manepada affix is employed.

The word ' karma' in this aphorism means ' action, 9 and does not mean
the grammatical karma or ' object. 9

As #? 3ftnr<r ' he strives for U13 house' so also «5J*% ^Otc% II But
^liHI^friyftl ' he rises up from the seat.'

Vart. — The force of the preposition ut must be to express f^r, 'effort,
exertion, wish or desire, to surprise or excell.' If this be not the force of ut,
the terminations are those of the Parasmaipada. As qrenX^TPTT^ tm^frtgft

* a hundred is yielded by this village. 9 The word ffT qualifies the word
1 anurdhakarmani,' and does not debar the latter.

*mw3i*<3l 11 % n vztfk 11 ^n^ t to-^j , ( w,
en* xto ) n

Digitized by LiOOQ LC

130 The Atmanepada verbs. [ Bk. I. Ch. HI. § 25-27.

25. After the verb sth&, preceded by upa,
when meaning * to adore,' the Atmanepada affix is used.

As ifejUHllJMmgqfrtgfr ' he approaches with prayers or worships the
Girhapatya fire with aindra hymns/

If it does not mean ' praising with hymns' the terminations which
the verb takes, are those of the Parasmaipada, as WdfiJJlfiyft ^hRH ' she ap-
proaches the husband through youth/

Vart. — The verb *sjr after the preposition 3T takes the terminations
of the Atmanepada when meaning^ worshipping a deity/ ' to approach for in-
tercourse or uniting or joining/ ' to form friendship with/ and ' to lead to as a
way/ Thus: — 1st Deva pujA Hjlll^^kiy^l 'he worships the Aditya/ 2.
Sangati karana tfd^ftud HlO 'the wife approaches the husband/ mi^ljHI^a^
' forms union with charioteers/ 3. Mitrl karana ^raryrf^OT *ST£i ' the good
man approaches the saints to make friends with/ What is the difference bet-
ween Sangati-karana and Mitri-karana ? Sangati-karana means drawing near
and approaching together in space, as ^HimjJHI^lrlaJlr 'the Ganges joins the
Jumna/ While the friendly relations may be established without coming in
physical contact* 4. Patha : — ^q 7«ir: 5** J**?^"^ ' this roa( ^ leads to Srughna.

Vart. — It must be stated that the Atmanepada is optional when the
sense is ' desire of getting/ As faspRf ^T^TCT^TyTf?!^ or OTT%gf3 ' a beggar
waits at the palace of a Brahmana with the desire of getting something/

26. After the verb sth&, preceded by upa,
when used intransitively, the Atmanepada affix is nsed.

As H)T3R3tT5) VMf^^f) ' he stands ready at the time of dinner/ 41443^3*
qri%3% 'he is present whenever it is dinner time/ 4H4WH3<jf<ta<t 'he is present
whenever there is food, that is he comes at the time of dinner/ The word
bhukta is formed by adding the affix kta to the root, and has the force of
* condition ' here*

The phrase ' when used in the Intransitive y governs the three suc-
ceeding sfitras also*

If it is transitive ; the verb upasthi takes the parasmaipada termina-*
tions. As iMH^Mfrty^ • he approaches the king/

gf^Rit cw 11 ^a 11 n^ifk 11 g^-favcn^, ^cn f ( aro^-
to^stw xio ) n

*iNN*n 11 wi^*i «*i*4«w*ifJi q-rhmn 11

Digitized by LiOOQ LC

Bk. I. Ch. III. § 27, 28. ] The Atmanepada verbs. .131

27. After the verb tapa c to shine/ when used
intransitively, and preceded by ut or vi the Atmanepada
affix is employed.

As ZrHrt or ftmn fkvrm ftsmff ' The scholar shines with knowledge/
But in transitive verbs, it is Parasmaipadi. As ^xnf% 4}4u7^4lJ«hli:> ' the gold-
smith heats the gold/ R<wft gy ttf^dl * the sun heats the back.'

Vart. — It must be stated that the terminations are of the Atmanepada,
though the verb ut-tapa or vi tapa be transitive, when the object is some limb
of one's own body. As ^rTT^ITfaj or fir^falftj^ or %£ ' he heats his own hand
or back ' The word wy means one's own body and not the pdribhdshika or
the techinical ^y meaning " a thing which not being liquid or gaseous and
being capable of being perceived by the senses, and not being one produced
by a change from the natural state, exists in a living being and though found
elsewhere actually or at any particular time, had previously been known as
existing in only a living being, or is found to have actually (not figuratively)
the same relation to the being it is in, as a similar thing has to a living
being." (See IV. 1. 54 Hl^lClMqJHK^JihMlO Therefore not so in
the following ^-cft *|<Him ssj^nft • Devadatta heats the back of Yajfia-
datta. -When the preposition is other than ut or vi, paras maipada affix is
employed ; as, fi^rft.

3TT#taHTf?|: II ^C || TRJTft II 3n^t, *nr-fr:, ( ^+A*Ic^
9TT0 TW ) II

«ri >wfir 11

28. After the verb yam i to stop,' and lian * to
injure * when used intransitively and preceded by &&,
the Atmanepada affix is employed.

The word intransitive of s. 26 is understood here also. Both these
verbs are generally parasmaipadi. When they take the affix *jr they become
Atmanepadi. Yam belongs to Bhv&di class; and han to Ad&di class. Thus
W+^r + *TT+^"W+*r*3 + ST^ (VII. 3. 77 *S*lffcritf 5: ' chh is the substi-
tute of the finals of ish, gam and yam, when an affix having an indicatory
*T follows) = srr*r*5^ 'it spreads.' **nr«gpr and ^rrar^u?^ ; so also W + *% sfa*^
+ ^-W+f +% (I. 2. 4 and VI. 4. 37 gr3^ mAli^fddHl^i4>HI*IJHlRl4if?W>
fTf^fi^j-BinjlT, 'he strikes.' Brr-^+^r^-W + ^+HT^ (I. 2. 4 and VI. 4.

gS^MiH^HR^ im<4*M<gMti<ii ott: a^Sr wrffcJ-nnnMVH. 3. 54), (*ir?^r-
fe«5)Pl.«nft. I

Not so when these verbs are used in the transitive, as MJI4«t»f<14JtMnc- | **
^npj € he draws up the rope from the well/ fttt*f£* j^r* «T!$t ' they kill the 1

sinner with the foot.' j

1$2 The Atmanepada verbs. [ Bk. I. Ch. III. $ ig, 30.

Vart. — Wheto the object is some member of the agent's own body
these verbs take the Atmanepada terminations though transitive. As a?nre*it
qrf^ ' he puts forth his own hand.' *ff?% Rnf f he hurts his own head.' When
the object is some limb of an others body, the terminations are of the Paras-
maipada, as mnPd far; qr^fhr ' they hurt others head.'

q%T^{^g f^a<c^f%faR * 4 S H *<! II U^lPl II ^t: f
*n**I%<rt *Hfif II

29. After the verbs gam € to go,' jichchh * to
become hard,' prachchh l to ask,' svar c to find ihult,' ji
4 to go,* Srfr c to hear,' and vid c to know,' when used in-
transitively and preceded by sam, the Atmanepada affix
ia employed. -

The above seven verbs when preceded by the preposition *3PJ and
used in the intransitive, take the terminations of the Atmanepada. As oq^s^r
(VII. 3. 77) ' he joins/ *Ui*u^ ' he becomes hard or goes.' «s*€% ' he asks/
g^q?^ he * blames/ tm^i ' they are attained. 9 In the case of this last verb
which belongs both to Bhvidi and Juhfltyftdi class it is Atmanepadi only in the

Aorist As ^j+^r+an^+tr (HI. 1.' 56 *rfTrcrr$nrf%vq , v) « srq*r, as >if^pr
(VI. 4. 75 q^rig[^^pn ^flfl jft) ^ + *T + »^ + w*r - mvxi (VII. 4. 16
^l^Mf^-^JUJ: )- This occurs generally in the Vedas. The verb faj must have
the meaning of 'to know/ and not that of 'to acquire'' cfff3^ 'he hears/
^f%?Hr ' he knows.'

Vart. — The root <[*r (to see) after the preposition ^ when used in*
transitively takes the terminations of the Atmanepada as tw^qft ' he sees/
But when transitive, it takes Parasmaipada terminations as mn*( ilMUqft . ' he
sees the town.'

(«no no) n

30. After the verbs live l to call,' preceded by
ni, sam, upa, and vi, the Atmanepada is used, even, when
the fruit of the action does not accrue to the agent.

Digitized by LiOOQ LC
\

Bk. I.Ch. III. § 31, 32.] The Atmanepada verbs. 133

The verb \: 'to call' whether transitive or intransitive is Atmanepadi
after the above prepositions. The condition of being Intransitive does not
apply to this sfitra, and from this sfitra forward general rules of Atmanepada
are treated of. As f%£^r%, *3|*WT» OT|*J% II

The verb^ is marked in the Dhatupa(ha with a ^ and therefore by
sfitra 72 of this Chapter, it will take the terminations of the Atmanepada when
the direct fruit of the action accrues to the agent. But even when the direct
fruit of the action does not accrue to the agent, the verb hve takes the ter-
minations of the Atmanepada when preceded by the above prefixes.

Vart. — The verbs sj^r 'to throw* and 3T^ 'to note* take optionally
Atmanepada affixes when compounded with upasargas. As ftUw i tl — »fr ' he
casts out/ ^jipft — •% ' he collects.'

31. After the verb hve, when meaning c to
challenge ' and preceded by fifi, the Atmanepada is used,
even when the fruit of the action does not accrue to the
agent.

As i{F$t qFSTPTTf ^[ir ' an athlete challenges another athlete (in order
to conquer him)' fe M^ nm^ 'one student emulates with another student*

This sfitra is also for the purpose of showing that the root hve takes
the Atmanepada terminations even when the direct fruit of the action does not
accrue to the agent.

The word spardhi means to emulate, to vie, to desire to conquer
another. When the verb hve has not the above signification, it takes the
terminations of the parasmaipada. As nrarf^ffT ifrrreT: 'the cowherd calls the
cows/

^n^^^qiir^yi^^f^q^f ^q^H^^^i!!^^ ^|: II ^ II

(WO TW) ||

iTw: 11 «i?*?rrft«Rr^ ^4»tri^ *0?Uich3hj *rrfHr 11

32. After the verh kji when meaning i to di-
vulge,' 'to revile/ c to serve/ 'to use violence/ 'to
cause change/ < to recite/ and « to do an act tending to
effect a desired purpose/ the Atmanepada is used, even
when the fruit of the action does not acci^e. t to d g@ agent.

3

Chx^i s'\

184 The Atmanepada verbs. [ Bk. I. Ch. III. § 33.

The verb fps[ by sfitra J2 would take the affixes of the Atmanepada

when the direct fruit of the action accrues to the agent* This sfltra has been

begun to show that this verb takes the Atmanepada terminations under certain

circumstances even when the direct fruit of the action does not accrue t'o the

agent

The word gandhana comes from the root gandh f to injure/ of chur&di

class and means to inform against another maliciously with the object of injuring
such person! Avakshepana means ' to revile' 'to over-come/ sevana means 'to
obey and serve/ sihasikya means ' an act of violence/ pratiyatna means ' im-
parting a new quality or virtue/ prakathana means ' to narrate fully/ upayoga
means ' the disposal of a thing for the object of attaining merit &c/

Thus 3*$?%, xT4l^»?l means he informs against (2) ^SHt *f3<Mii<tf**<l

QajlaXl ^ 'the hawk overcomes or reviles a^ail* (3) HUj<hlj4^»*> ' he serves thej>rosti-
tutes/ »rgrqnrpr^T% ' he serves the mahamatra/ (4). qr^rn^ *fajF% * he outrages
/ /''another's wife.' (5). ip ilr^i^iPT^y^ ' the fuel gives a new quality (boils) to the

/t ^ ' water (or he prepares the wood and water for a sacrifice)/ The object of the
verb kfi takes the affix of the sixth case 1. e. genitive, only when the verb
means pratiyatna, see S. II. 3. 53 (frsf: Wflqfr ) I Therefore udakasya, though
an object, has genitive form. The verb takes the augment «£ only when it is
preceded by upa and signifies pratiyatna see VI. 1. 139. (^TT^ q ffi*U T ^^f
qi44ltqw?f) II Therefore, there is the insertion of sibilant in upask unite.
(5)« 1HIT: U*ihit 'he recites stories/ 3remT$r% M^hh 'he recites slander/
(6). CTT 73}$% 'he devotes a hundred ' pieces of money, for the sake of merit.
<i4<M Hij*S he devotes a thousand.

Why do we say in these senses ? Witness sj# *itfa ' he makes a
mat/ In this case the Atmanepada affix is not employed.

The verb kp is understood in the three succeeding sfitras.

?f%T: II %jfa <rj??[qRtlr*T: *W*3 ^r{HHii\9H^Hi ^rft II

33. After the verb kji preceded by adhi,
•when the sense is that of l overcoming or defeat/ the
Atmanepada is used, even when the fruit of the action
does not accrue to the agent.

This sfitra is commenced to show that kfi may take the Atmanepada
affix even when the direct fruit of the action does not accrue to the agent.
The word prasahana means 'to over come, to be not defeated/ As fnyf^Ni 'he
overcame him or he was not defeated by him/

Why do we say, ' in the sense of to overcome?' Witness *fl faptarftffr ' he
learns the meaning ' in which example the Atmanepada affix is notr^mploycd.

\

Bk. I. Ch. IIL § 34-36. ] The Atmanepada verbs. 135

The necessity of making a. separate sutra of this rule, instead of in-
cluding it in the last apporism, arose with the object of giving a distinctive
meaning to the prefix adhi.

mo tio) n

^ftr: 11 fk ^5. *hCl«li*^ftqr8 fi^^ii^i^ v*^c^*4m *iw^N< *nflr 11

34. After the verb kvi preceded by vi, even
when the fruit of the action does not accrue to the agent,
and when the sense is that of c making sound/ (liter-
ally, having { sound ' for its object) the Atmanepada is
employed.

The word karma in the sutra indicates objective case or kAraka; and
does not express 'action/ as in some previous sutras such as 24 ante.

As nfrST f%3f$% WI41 * the birds are making noise/ t*Nft ft<£*£

The word fk governs the succeeding sutra.

Why do we say " when governing a word expressive of sound in the
objective case"? Witness f^r R^uft 3TPT: ' love affects the mind.'

«rc*M*5n| n \\ n *mfa 11 erevfan^* % 9 few:

SITO VO §:) H

35. After the verb kji preceded by vi, when.
used intransitively, the Atmanepada is used.

As fofr «* T»tl &*j<ii « the horse^move gracefully/ afFFTCT yjfasPn (%aH%
4 the students being full of food are acting as they will or are aimlessly wander-
ing about. 9

ffa: 11 cot^, qnd t<9<tw^i^>^ftmitfti<»^^w»i< *nrfo, virniftj
firafaiSy **5 11 *

36. After the verb nX ' to lead,' when used in
the sense of c to guide so as to render the person guided
worthy,' c to lift up,' * to make one a spiritual guide, 1
1 to determine the true sense,' c to employ on wages, 1

Digitized by LiOOQ LC

136 The Atmanepada verbs. [ Bk. I. Ch. HI. § 37.

c to pay as debt/ and 'to give as in charity/ even when
the fruit of the action does not accrue to the agent, the
Atmanepada is used.

This sfitra is begun in order to" show that the v6rb sfr may take the
terminations of the Atmanepada, even when the direct fruit of the action
does not accrue to the agent, in the following cases, vis., when it means
to ' respect ' &c.

SammAnana means to respect ; as spqftf *5pft <HicfaTq% 'the ChArvi gives
instruction in the Lokayata £astra.' The word charvi primarily means intelli-
gence, and by secondary use it has been extended to the preceptor also, such
a preceptor gives instruction in Lokayata Sastra, that te to say, having estab-
lished the truth of the doctrines of that philosophy by argumentation, imparts
them to the pupils. Those doctrines being thus established by reason, be-
come respected (Sanmanita) and honored.

Utsanjana, 'to throw up, or lift up/ as *TFnpfr g ?Pra^ 'he lifts up
Manavaka.'

Acharya-karana 'acting as a teacher/ that is to say, to bring a boy
near oneself in such a way, that being so brought near (upa-neta); he may
himself become an achArya. As Hmi^^indfr 'he initiates Manavaka (/. e.,
making himself the preceptor he brings the boy near himself/) .

Jnana means ' knowledge, a demonstrated verity.' As spnt t ^rff *tfa*T-
^% ' the Ch&rvt investigates the truth of the Lokayata doctrine.'

Bhfiti means 'wages.' As qvfcgrc r jT?^ 'he employs the servants on
hire or wages.'

Viganana means the paying off as a tribute, debts &c As «OT *R(
fipRFT, l the Madras pay the tax due to the king, that is, they discharge the debt.'

Vyaya means ' allotment of money on works of merit &c/ As *ffi Rw^tffc
' he expends a hundred pieces on religious acts.' qr&t fiH^3> 4 he devotes a
thousand.'

Why do we say, in these senses ? Witness ^prf TOfct MW5. ' he carries
the goat to the village' Here there is no Atmanepada affix, as the verb has
not any one of the above significations. —

*>*Rfa, (fro: zno no) n

?Rf: II TO%: «E MU* 4 *lfi<$*H3l*4l : *&i$ cfewSu^qft? Sift SRWiKH^*

*nifaii

37- After the verb ni, when it governs an

incorporeal object existing in the agent, as its object, the

Atmanepada is used. Digitized by Google

\

Bk. I. Ch. 111. $38.] The Atmanepada verbs. 137

The verb ft is Atmanepadi when governing an object which has no
material body, but is an abstract noun, that is an object which has abstract
but no concrete existence ; and when such an object has its seat in a
portion of the agent of the verb. As *£W f%TO*t ' he subdues his own anger/
T?J f^rerJfr ' he suppresses anger/

The word sarira means the body of living beings, any portion of such
body is also called iarira.

The object must reside in the agent, (kartristha), otherwise the verb
will be parasmaipadi. As *43ift <J*i4*(t4 **t>t ft^niRr ' Devadatha removes
Yajnadatta's anger/

The object moreover must be immaterial, otherwise the affix will be
of parasmaipada. As irf fenqfo ' he removes his own wort.' qraf ftroftr
' he bows his heck/

Why do we say ' in the case of an object ? ' Witness f^jrf^Prfa ' he is
submissive through knowledge/ sncRTT fi«i^fl| . Here the verb fi?nri% has not
taken an accusative case, but is in construction with an instrumental case,
and therefore the present rule does not apply.

«rRf: || *T4lRt4*?^ JhttafJHlfH 1 *! *Wft II

38. After the verb kram c to move' when used,
in the senses of l continuity/ c energy ' and development,'
the Atmanepada is employed.

The word * kram 9 governs the succeeding sutras up to 43 and is
understood in them all.

The verb kram by sfitra I. 3. 78 (after the rest, let the parasmaipada
affixes be employed in marking the agent), would have taken the affixes, 'of
the parasmaipada ; the present aphorism enjoins Atmanepada affixes in cer-
tain cases when the verb means to have a taste for &c.

Vritti ' continuity ' means unobstructed, or want of interruption (t. e. f
a taste for, or facility in, anything), 'sarga energy' means application
resolution and determination. T&yana ' development ' means increase and
growth. ,

As (1) 44^4*4 ffPT% jfqp — ' his reason proceeds unobstructed through
the Rig. scriptures u e. % he can easily comprehend the Rig/ (2). wrraTSJWf-
*R|7T *FT% ' the pupil shows energy or exerts to study the grammar/ (3). ^lft%^
^rr^rrfi£T *W% ' the shastras are developed in him/

Why in these senses only? Witness 9TTOPrft 'he runs away/

Digitized by LiOOQ LC

133 The Atmanepada verbs. [ Bk. I. Ch. III. § 39-41*

39. After the verb kram, preceded by upa
and para, when used in the senses of continuity, energy
and development, the Atmanepada is employed

The words vfitti &c., of the last sfitra are understood here also. Why
has this been made a separate sfitra and not included in the last ? The reason
is to make a restrictive rule in the cases of upasargas ; that is to say, only ip
the cases of the upasargas upa and pari, there is Atmanepada affix j but not
so when any other upasarga precedes the verb kram, though the sense may
be of continuity &c, Thus OT%reft 'he commences to advance/ «w*hHft 'he
inarches to attack/

Why do we say after the upasargas ' upa and part. ' ? Because, after
any other preposition the affix will be of the parsmaipadaj as OTTRft ' he
makes progress/ If the sense is not that of " continuity " &c- t parasm^ipad^
will be employed, a? 3*TOPrf*t, TOSTPTftr.

«ii«44+u) 118011 ^ft n sn^:, ^pra^ fan*:, sna tjo) u

Ttffcsqin »«if*i»ym3 jht tt&^n

40. After the verb kram, preceded by 46, the
Atmanepada is employed, when used in the sense of the
rising of a luminary.

As stlliMid ^?: 'the sun rises/ qjrCTri) ^HPtf: 'the. moon rises ' HjUMjsfr
*4>qi'ft 'the stars rise/

Vart. — This ascending must refer to heavenly bodies. Therefore in
WfrlHfft Ujft^JfltM^ , 'the smoke rises from the surface of th$ terrace/ the verb
iCT is parasmaipadi. Why do we say in the sense of " to ascend" ? Witness
4ij l HiHpi qrara*: <$W*{ # the boy assaUs4h£^»/ M*]4 tK, <td Ct *~t - £a iv mj

^: qi4&{tifc H 89 II Wfjfk II *:, tn^-f^S*, ( *&t
«IW W ) II

Tfo 11 fa^rfcsprer: mifi^iul^ yiprnrnrf^i^w H*ftr 11

41. After the verb kram, preceded by vi, the
Atmanepada is employed, when used in the sense of
* placing of foot-steps. 1

As qnfr ft<tiHtl ' the horse is pacing/ The term yikramana is applied to
the special movements of horse &c.

Digitized by VjOOQlC

Bk. I. Ch. III. § 4*, 43-] The Atmanepada verbs. 139

Though in the DhAtupA^ha (catalogue of verbal roots), the veib nrf%
and*rj have the meanings of pada-vihaxana and pAda-vikshepa, that is
11 throwing or placing of foot," and so it might be objected that the present
sfitra is a useless repetition, yet as verbs have various other meanings than
what is assigned to them in the dhatupa{ha, the present sfitra is not un-
necessary.

Why do we say "in the sense of throwing of foot"? In any other sense,
the terminations will be those of the parasmaipada. As f%3tprf3f *t0w4rf^i|:
thefetdof the antelope skin is -rupture}* 9

lifamit «q$r«n* u « ii ^T^f it n-^THRn^, *n*rt-

42. After the verb kram, the Atmanepada is
employed when it is preceded by pra and upa, both con-
veying the same sense; viz. that of "beginning an
action."

As topi% Ht-rj»*t , 3WT^ *¥iT3^ ' he commences to eat 9 &c.

The prefixes pra and upa are synonyms, when they denote the com-
mencement of an action.

Why do we say, " when they are synonyms." Witness the following
tjjfaj: W^lHp l *nfrgFnKWft 'he goes during the first part of the day and he
comes back during the latter part of the day. 9 Here in one case the sense
is " to go," in another it means " to return."

Why in the latter case of ' aparedyur upakramati ' there is not At-
manepada by virtue of sfitra 39 ante, as there is the upasarga upa here? The
answer is that Rule 39 is not applicable, as that rule is limited by the condi-
tion of the sense of ' continuity ' &c while in the present sfitra, that limita-
tion is not applicable.

43. After the verb kram, the Atmanepada is
optionally employed, when it is. not preceded by any
preposition. *>*•'•

The root ' kram # may always be conjugated fc as Atmanepadi when
not having any upasarga. The upasarga will be defined in I. 4. 59. The
option allowed by this sfitra is an example of what is technically known as
apr&pta vibhAsA viz., an option which is not an alternative limitation to a
general rule already found or known.

•Digitized by

Google

140 The Atmanepada verbs. [JBk. L Ch. HI. § 44H6.

Examples: ICT^ or %mfo 'he goes over/ Why do we say ' when it is
without any preposition V For no option is allowed when it takes preposition.
As 3**PTft.

44. After the verb jna, when used in the
sense of * denying,' the Atmanepada is emlpoyed.

The root jfia which ordinarily means 'to know' and is'parasmaipadi
by rule 78, becomes Atmanepadi when through the virtue of any prefix it
means * toltective: ' as <U4*u«ftfr ' he Reives/ fEPPTniPfl^ ' he denies the debt
of a hundred rupees/ 4J{UH<JJiH'ft 'denies a thousand/

Why do we say when meaning 'to deny?" observe: — ?f^f3ffovft
TTRTft 'thou knowest not anytb : ng/

qre & grer » 8H » itvft • sre^n?*, * (v, sno tro) 11

lf%: II g f Hl^i*^ »3 *A * faqi*^HKUH^4 *nft tl -

45. And when used intransitively, after the
verb jfia, the Atmanepada is employed, even when the
fruit of the action does not accrue to the agent.

. As citfjt *iHl'?f«^ft^rrft% 'he engages in sacrifice by means of clari-
fied butter or honey/ (See II. 3. 51).

This Atmanepadi form of jna is employed when the fruit of the ac-
tion does not accrue to the agent From sfitra 76 it will be seen that
when the fruit of the action accrues to the agent, the verb is Atmanepadi.

How in the above example the verb jna is intransitive ? Because the
word sarpil? is not the object of knowledge, but it is an instrument of sacri-
fice, into which one engages from complete knowledge (jn&na) ; and it (sar-
pishah) is in the genitive case by virtue of II. 3. 51 ( *frjf^for *MUJ ) by
which the instrumental IcAraka of the jfia is put in the genitive case.

Why do we say of." Intransitive"? Because in the transitive it is
parasmaipadi, as re?«i jpfr sfFftfr 'he knows (recognizes) the son by his voice/

arrant , ( v snow ) n

46. After the verb jna, preceded by sam, and
prati, the Atmanepada is employed, when not used in
the sense of "remembering with regret."

Bk. I. Ch. III. § 46-48. ] Thb Atmanepada verbs. 141 •

The word jfia is understood in this sfitra: and the aphorism applies to
that jfia which is used in the transitive. As %pt ^JTPrtit # he l o oks f or a hundred/
so *m sf?f3Pftf « he promises a thousand/ ^^ * C^Yf^V

Why do we say when 'not meaning remembering with regret?'
Because it is not Atmanepadi, when it conveys that meaning. Thus IT*:
tfjTRTfit ' he remembers with regret his mother 9 fay: ftaHlft II

ff^r: 11 >»^HiRa f%$r*ndf **9 *f*ta*rfrrt *wflr »

47. After the verb vad, the Atmanepada is
employed, when used in the senses of " showing brilli-
ance, or proficiency in," "pacifying," "knowledge/*
cc effort," " difference of opinion," and u flattering."

By sfitra 78, vad generally takes the terminations of the Parasmai-
pada, by the present sfitra, it takes also the affixes of the Atmanepada when

having certain meanings.

Bh&sana, (shining) means to illumine as wjifr *wif tfta*f*?t The chlrvi
illumines the Lokayat shastras, by his discourses and clear exposition.

Upasambhdshd (persuasions) means to appease, to conciliate as qnl-
STnjTOT^ he conciliates or cajoles the servants.

Jildna (knowledge) means to know completely as fyyfr *m{ ffoKryl the

chArvi knows completely to discourse upon Lokayat shastra.

Vatna (endeavour) means energy as vft wfW. ' He toils in the field/
«tfr **% 'he toils in the house. 9

Vitnati (disagreement) means dissension, want of unanimous opinion.
As tf^ft*^. 'They disagree over the field' 1. <., holding different opinions
they talk diversely.

Upamantrana (enticing) means to coax in secret as, ffi?THrfr3<mpi 'he
entices the wife of a respectable family (1. <., seduces her in secret)' *HjKljM-
**% 'he flatters another's wife. 1

Why in the above senses only ? See TTfr fafanfa *hc says something'.

srerarotsg^T^ 11 a* 11 u^rffi 11 wnwiw N , **-9^t-

^ f ( *5 : f WO X(0 ) 11

ff^r: 11 ***STr«tf «rg^mrf ^fhrmd *Hrefrrrenrc%n*ifrrt H*ffr 11

48. After the verb vad, the Atmanepada is
employed when used in the sense of "speaking articu-
lately in a similar manner/ 1

4

142 The Atmanepada verbs. [ Bk. I. Ch. III. § 49, 50.

As #TO3?9 Al^nnr: ' the Brihmanas are speaking/ But in *fqq?fcr
pT: 'the cocks are crowing' it is properly Parasmaipadi.

The sense of the sfltra is that when men, who are only capable of
articulate speech, speak all in one and the same time, then the verb vad
takes the affix of the Atmanepada. When lower animals make a chorus of
noise, the verb does not take the Atmanepada.

There must be samuchchirana for the application of this rule, when
there is no samuchch&rana or speaking in a chorus, this rule does not apply,
as OTSTOt Ufa ' the Br&hraan 'speaks.'

wo n» ) g Hingreri 11

49. After the verb vad, preceded by anu
when it is intransitively used, the Atmanepada is em-
ployed, when the sense is that of" speaking articulately
in a similar manner."

As 4djj<f4<) gjT: cfartutq kafha is echoing or imitates kalapa. He
reutters exactly what the kalipa-reader or the teacher says. The word anu
here means similarly.

Why do we say • in the Intransitive '? Because when it is used in
the transitive it takes the affixes of the Parasmaipada. As <rf$* H*^flfl1j34fll
he repeats the yajurveda which he had learned before.

The words ' uttering of articulate speech' are understood here also*
Otherwise ar^q^fi 4t*\f the lute resounds, here it is Parasmaipada. *

50. After the verb vad, the Atmanepada is
employed optionally, when the sense is that of " con-
tradicting each other.*

As ftOTsF^ or#?r tor: (the doctors are at variance). The words
vyaktav&ch&om (articulate utterance) and samuchchAraoa (speaking to-
gether) are understood in this sfltra also.

No option is allowed when the sense is not that of contradiction or
wrangling as dUi^A flngTO „ The Brdhmanas are speaking together.

The phrase ' articulate utterance ' is necessary in this also. Because
as ftrorfN tr^TO: the kites are quarrelling, the veib is in the Parasmaipada.

Bk. I. Ch. III. § 51, 52.] The Atmanepada verbs. m l«

The wrangling must be by talking at one and the same time other-
wise this rule will not apply. As wfru #OT $^r *T* Rw^ the doctors con-
tradict by turn another doctor.

st^t^i: 11 \% n q ^ifn a-eraif 9 v: t ( «no w )

ffrT: II ^W^ll J fc f*H«UlwJl4 Wfiti II

51. After the verb gri c to swallow/ when pre-
ceded by ava, the Atmanepada is used.

As ^ K ft ii P f (he swallows). The root grt to swallow preceded by the
preposition ava is Atmanepadi. The verb gri taken in this sutra means to
swallow and belongs to the Tudidi class, it is not the gft ' to make sound '
which belongs to the kryidi class. Because there is no word formed with the
latter verb by affixing the preposition ava to it. By sutra 78 this verb gfi
'to swallow 1 would have been Parasmaipadi, the present sutra debars
that.

When it is not preceded by ava, it is Parasmaipadi, as ftfcft be
swallows.

ff^T: 11 <u\jfis^ftK& 5rf^rr% ^HPnrr^R^t *r*f% 1

52. After the verb gri preceded by sam, the
Atmanepada is used, when employed in the sense of
1 promising.'

As *ptf bPik9 (he promises to pay a hundred rupees) If it does not
mean to promise or acknowledge, it takes paras maipada terminations. As
*iftf*frt *iro*rhe swallows the mouthful.

( «na va ) a -

ifip II ^^ITOr%: q*^*flh^l**Hljl**l>J1 nrfir i f

53. After the verb char * to walk * preceded by
ut, when used transitively the Atmanepada is employed.

As $*<jmd he strays away from home ; «iM^Hjq<<> he transgresses
the commands of his preceptor.

Why do we say "when used in the transitive"? Observe qimgqifl
the vapour is rising. Here it is parasmaipadi.

q^ciliiigitii^ a H« » *RFfa a qm 9 ^ftin-g^r^ ,
(w WO tro) 11

?fif: II ^I^K^M^ f 3 -rMlltHHMj i snfif H n

Digitized by VjG>OQLC'

144 • The Atmanepada verbs. [ Bit I. Ch. ML § S4# 55-

54. After the verb char, preceded by sam, and
connected with a noun in the Instrumental case, the
Atmanepada is employed.

As «n[%7 3*nrd he rides on the horse-back. ■

The rule does not apply when it is not in composition with a word in
the instrumental case. As &tt stigfr <JMlfa %$ ^T$ ^ %^T ! O Devala ! thou
wanderest through both regions, this and that Here though the sense of the
Instrumental case is implied, yet as it is not expressly stated, we use the
parasmaipada terminations.

*rftw*£ 11 v^jt^icft *nforr ^igwn} H**ftftr *v*n^ n

55. And after the verb d& to give, preceded
by sam, and connected with a noun in the Instrumental
case, the Atmanepada is employed, provided this Instru-
mental case have the sense of the Dative case.

When is the 3rd case used with the force of the 4th case? That is
answered by the following.

Vart: — The Instrumental has the force of the Dative when immoral
conduct is implied. As {TOH or f^zrr *tore»% (he gives with the object of
enticing to the female slave or prostitute).

The verb di to give, is generally parasmaipadi, it becomes Itmane-
padi under the above conditions.

Why do we say " when it has the force of the dative case" ? Because
when it has not the sense of the dative, parasmaipada terminations will be
used, as qrf*RT shW^sfa, ' he gives with his hand/

It might be asked how the Atmanepada terminations are employed
when the preposition sam is not directly applied to the verb, but another pre-
position pra, intervenes between sam and the verb* The reply is that the
word samah in sfitra 54 is not in the ablative case (which would have required
its being placed immediately before the verb), but it is in the genitive case,
and is used as an attribute.

**!8|*r* *5TOC3* H 1$ II V^tft II *VTH, *W , W^ ,
( 9TW X(0 ) «

fftf: II V*Vftf5W : ^ 1 ^ 1 ^ ^HHHIWPff H*f*l II (^ r\r\n\o

^ Digitized by Vji (J KJWVC

Bk. I. Ch. III. § 56, 57. ] The Atmanepada verbs. 14*

56. After the verb yam, to give, preceded by
I upa, when used in the sense of " espousing,' * the Atmane-

; pada is employed.

J This is clear. As ^1^^^ he espouses or knows his wife. But

not so when another's wife is meant, there it is parasmaipadi.
j By rule 78, the verb yam would have been parasmaipadi, but the

I present sfitra makes an exception, when the root takes the preposition upa

j and means " to marry, to espouse, or to accept The meaning of sva-karana

is confined to accepting in general. This is according to Kisika; according
to MahAbhAshya, sva-karana means to make one's own what was not
previously his own.
I Why do we say in the sense of " marrying"? When it has any other

sense parasmaipada affixes will be'employed as S^r^ft <UU*m lnfjfWfcW^
Devadatta has illicit intercourse with the wife of Yajnadatta.

( sno tro ) *

i 57. After the Desideratives formed by the

affix san, of the verbs jM to know, sru, to hear, smyi to
remember, and djiS to see, the Atmanepada is employed.

The above four roots are Atmanepadi when in the Desiderative
form, the suffix" san" being the pratyaya by which desideratives are formed.

The verb jna takes the terminations of the Atmanepada when used in
the senses indicated by the three sfitras 44, 45 and 46 already explained.
The desiderative form of jfiA would ex necessitate take Atmanepada affixes
when used in the above senses : see Rule 62. The present sfitra however
enlarges the scope of Atmanepada by declaring all desideratives of jnA to be
Atmanepada, in whatever sense they may be used*

Similarly by sfitra 29 ante and the vArtika under it, the roots £ru 9
and dpi, take the terminations of the Atmanepada, when preceded by sam,
the desideratives of those would of course have taken Atmanepada affixes by
62, but the present sfitra makes it general.

I ^ The present sfitra however enunciates a new rule in the case of the root

smp. 1 ■

As y&if f*i<!Nj3 he wishes to know (i>. enquires after) religion *[$
spjjnt he serves the teachers, *n* 5^% he wishes to remember the lost yt
ftf^T% he wishes to see the king.

Why do we say, when taking the iffix san ? Because the primitive
verbs will take the parasmaipada terminations as *n*fft he knows; tpflftrhe
hears, mtfH he remembers Tpqf% he sett.

Digitized by LiOOQ LC

146 The Atmanepada verbs. [ Bk. I. Ch. IIL § 58-60.

*It^; H \c 11 q?[Tf*r II *T , ST^T* , W , ( VT: WO^o) ll

?fa: 11 m«fljil*i»iMifr ; ^n^m^T^rt * wflr 11

58. After the desiderative of jM when pre-
ceded by anu the Atmanepada is not employed.

This is ah exception to the last sfitra. The desiderative of jfid with
the prefix anu is not dtmanepadi.

This sfitra read with sfitra 45 limits the scope of parasmaipada to the
transitive, desiderative verb anu-jiiA. As J^prjfinrfraft he enquires after
the son.

Why do we say when preceded by anu ? Because otherwise it will
take the Atmanepada terminations. As >**&{ ft'HTOQ' he inquires after religion.

( **T: WO HO ) h

yfa: 11 *fa «rr^*Mr ^gft^gf^-: q*renrreft*N * H^rfir ii

59. After the desideratives of 6ru when pre-
ceded by prati and &fi the Atmanepada is not used.

The verb 6ru when taking the prefixes prati and An, is not Atmane-
padi, though taking the desiderative "san.*' This is also an exception to the
. rule of sfitra 57. Thus sn%%ycffi and miyyiifr l»

The word prati and in must be upasargas; if they are used as karma-
pravachaniyas (see I. 4. — 83) then the rule will not apply As ?^HT iflRr *pg*% II

sor^: fifca: II ^0 ll xr^Tfa II V%t , fsrcr: , ( WO HO ) II

60. After the verb 6ad to decay, when' it has
one of the affixes with an indicatory & (6it) the Atmane-
pada is used.

The root 'sad ' when taking any affix which is marked with an indica-
tory jr, is conjugated in the Atmanepadi. In connection with this, must be
read sfitra VII. 3. 78. TniB^^liiKIU^t^^^^^^^li^^^^^^^^^^^^^^^^"
^fNttft^T: by which the root qr^ is replaced by tffar before affixes, having an
indicatory *r . Thus *Tf + *r + % « *ffa + b^« ^ffaHfr he decays or withers,
*fHft they two decay *ffai*$T they decay.

Why do we say before affixes having an indicatory *r? Before
other affixes, it is not Atmanepadi. As HjAI<^^ if he decayed, TC**rf*r he will
decay, firaroft .

The well-known vikaranas like ^ , *r &c, the affixes like to &o,
are fiflj affixes. In other words the root Sad is Atmanepadi in all conjuga-
tional tenses. C^r\r\a\o

Digitized by V^OVjy IC

Bk. I. Ch. III. § 6i, 62. ] The Atmanepada verbs. 147

3TT0 T|0 ) M

61. After the verb rnji to die, when it has one
of the affixes having an indicatory *r, as well as when it
takes the affixes lufL (aorist III 2. 110) and lift (Benedic-
tive III. 3. 159) the Atmanepada is used.

This is a restrictive sutra, the root ^ (to die), is marked with a ^
as an indicatory letter, so by sutra 12 it would be always Atmanepadi. But
the present aphorism restricts it to W3F lun (Aorist) and ffr& liA (Benedictive)
tenses, as well as to those tenses which are fil<£. It will be found hereafter
that out of the ten tenses, those that take (^ affixes are the special tenses
i.e., the present, the Imperfect, the Potential and the Imperative.

It is only before these three affixes, namely, lift, luA and sit affixes,
that the root mp takes the terminations of the dtmanepada. Thus the aorist
yQ1% ampta he died ; Benedictive s*frr mrishishja ' may he die/ Similarly
before sit affixes thus 1%^ ' he dies/ - 5 + v + % (VII. 4. 28 ft?£ mftH**l )

rtfa + ^r + tf «■ Rre% (VI. 4. 77. »rf%r ^j^Tj^ft ^^Irft^^w) fii^r , fini^

The root $ belongs to the sixth class of verbs called Tud&di which take the
vikarana %t in the conjugational tenses (III. 1. 77).

In other tenses, viz., the two Futures, the Perfect and the conditional,
this verb is parasmaipadL As: — nfi^Pl he will die. UjnRtq^ .

4 ^r**^*: 11 ^ u *5t0c n *nNr^ f *** 9 ( srw no ) n

62. The verb which is Atmanepadi in its
primitive form before the taking of the affix san, will
also be Atmanepadi when it ends in the affix san. In
other words ; after a desiderative verb, Atmanepada is
employed, if it would have been used after the primitive
verb.

If the primary verb is parasmaipadi, its desiderative will be also
parasmSiipadi ; if the primary verb is Atmanepadi, ifs desiderative will be
Atmanepadi. This is the general rule. Some exceptions to it have already
been mentioned in sutras 57, 58, and 59. A root which was Atmanepadi,
before taking the Desiderative affix ^san, will be Atmanepadi even when it
takes the affix *sr san. In other words, that by reason of which the Atmane-
pada affixes were ordained in the primary verb, will cause the same termina-

Digitized by LiOOQ 16

*148 . The Atmanepada verbs. [ Bk. L Ch. lit § 62.

tions to be applied when the verb ends in *p^san. Thus it was said in sAtra

12, that roots having an anuditta accented vowel as indicatory or a^ A as

their indicatory letter take the terminations of the Atmanepada. Thus vm%

he sits down and *ftr ' he sleeps/ The verb Bfl^ (to sit down) and *ft^ (to

lie down) will remain Atmanepadi, even when they are used as Desideratives.

Thus *rRrfifa?r ' he wishes to lie down/ Similarly sfitra 17 declared that the

verb fifty, is Atmanepadi as ftfrcrit » this will be Atmanepadi also in the

Desiderative form, as f%f%i%WT% ' he wishes to enter/ So also by sfltra 40,

•TnCTft is Atmanepadi, the Desiderative «flfttto% will also be AtmanepadL

*f ^ ^%* uCL --^But though by 60 and 61. *ffo% and ft^ are Atmanepadi, yet

^ } K ftrcTOI% ' he wicheo to lie dow n * 3^<fa% ' he wishes to die/ are parasmaipadi.

: :l Because the Atmanepada affixes were ordained after the roots qjf and

', *-^ 55 only under excep-tional conditions and not generically, and as those

V conditions do not exist in the desiderative, the latter does not take Atmanepada

/ / terminations.

'^ Of course that which in its primary state would not have taken the

terminations of the Atmanepada, there being a prohibition to that effect
will not allow them in its Desiderative form. As HjHftmflflft ' he wishes
to imitate,' Milftchtfft. Here the root f?sj by rule 79 has been especially de-
clared to be parasmaipadi, to the exclusion of Atmanepada affixes which would
otherwise have come by Rule 32 and 72, and therefore its Desiderative
is also parasmaipadi. The force of ^ causing Atmanepada is counteracted
by 79.

Now it might be asked: — True, this rule provides for those cases where
a root is conjugated in two forms, one a Primitive and another a Desider-
ative conjugation. But what provision do you make for those partial verbs
which take the affix 3^ even in their primitive form ; and in whose case we
have no prior form to look upon as a guide in the application of Atmanepada
affixes. And there are at least 7 such quasi -roots which take ^ in this way,
called also the self-descriptive ^ . Those quasi-roots are ^r, f*n^, f%?<r, 1FJ,
*>f , $P( and wj. What are we to do with these quasi-roots, which are always
conjugated with the affix ^r J and have no simpler conjugation ?" To this we
reply ; " In the case of these partial verbs which take the self-descriptive q%
the pada will be regulated by the indicatory letters which these quasi-roots
have. For though the full roots are ^ppa fifa^T, fofam , «ftaF€r , <Nw ,
*Ihl*U and $<IHI 1 yet by the maxim a^R# fftf f*T^p ^Jfnn^ fMfa* HT filr ,
"a sign made in a portion of a thing, qualifies the whole thing/ 9 the sign
made in the expression *J£ &c, will qualify the whole verb SRp^r &c" Thus
we have *np«% he despises, ftfc?*r% he cures ; *fhrfcr% he investigate^nfcc.

Bk. I. Ch. III. § 63. ] The Atmanepada verbs. 149

if^T: II MI«J€iHW^fRlfc ^pfrJi^^lWW^Hl *T*Rf II

63. like the verb that takes the affix &m f
if the verb be conjugated with the Atmanepada termina-
tions, so of the verb kri when subjoined thereto as an
auxiliary, the terminations are of the Atmanepada,
oven when the fruit of the action does not accrue to
the agent.

This sfitra applies where the fruit of the action does not accrue to the
agent. There are in Sanskrit two forms of the perfect tense (f«*^):
one formed in the ordinary way by the reduplication of the root and adding
the terminations; the other called the periphrastic perfect, is formed by
adding 3^ to the root, and then affixing to it the perfect tense of the
roots f? (to do), ^ ( to be ) or Hf[ (to be). These latter verbs are
called anuprayoga or auxiliary verbs, as it is with their help that the perfect
tense is formed. The question then arises, what conjugation, parasmai or
Atmanepada, do these auxiliaries take in forming the periphrastic perfect ?
This sfitra supplies the answer: — the auxiliary y follows the conjugation of
principals, the other auxiliaries do not. Thus the root ^>f (to increase,
prosper) is Atmanepadi, therefore in forming the periphrastic perfect, the verb
f? will be also in the Atmanepada. Thus it vjH* ' (he prospered). Similarly
f^(to appear with great splendour) is parasmaipadi and the auxiliary $r after
it will be parasmaipada, as <«il4j4hlt (he appeared with great splendour).

But the other auxiliaries «^and H^ retain their own peculiar conjuga-
tion and are not influenced by the conjugation of their principal. Thus ipir-

The word Am-pratyaya of the sutra means 'that after which the affix
Am (III. I. 35 and 36. ^ gmrw»W g% f*& ) comes/ Of the verb krifi, when
subjoined to another as an auxiliary, the termination is that of the Atmanepada
like the verb that takes the affix Am. If this sutra enjoined a rule of injunction
(vidhi), then there would be the termination of the Atmanepada even in
examples like Sjfcirfcresit and gjfrqpyKTC where the fruit of the action ;
accrues to the agent. This is no valid objection^ to the present sfitra.
It is to be interpreted in both ways, that is, both as a vidhi (a general injunc-
tion) and a niyama (a restrictive injunction). How is that to be done? Be-
cause the word purvavat of the last sutra is to be read in the present aphorism
also. The second explanation in that case will be for the sake of establishing
a niyama or restrictive rule.

5

150 . The Atmanepada verbs. [ Bk. I. Ch. Ill § 64.

It has already been said above that the auxiliary verbs *R,\ and
ira are used in forming the periphrastic perfect This is done by force of
the aphorism III. I. 40 fr^M^ft fe£> meaning, the verb kr/ifi is
annexed in the perfect tense, to verbs that take Am. In this sfltra (III. I. 40)
the word kr/in is a pratyAhAra, implying the verbs 55^ , ^ands*^. ** mi ght
be asked how is this pratyAhAra formed? The word kji is taken from
the fourth word of sfitra V. 4. 50, *J*JJW.*I? ^w*R<l4W Sec-, and the
letter n is taken from sfitra V. 4. 58 spfrffsffcf &c. This is the way
in which the commentators explain how not only the verb krin is used as
an auxiliary in forming the periphrastic perfect, but also the verbs ^
and tra . These three verbs denoting absolute action unmodified by any
especial condition have in all languages been fixed upon as the best auxi-
liaries.

It must however be noted that the word kjifi in this sfltra is not used
as aprat3'AhAra and consequently does not include the verbs ^ and sj^.

jftqwt g5ft *re rere g 11 v* n n^ifH n h-4<iihihi, git* ,
swiHn% , ( ano xto ) n

qffffr^»i M<K4^iM<*aiftui ygfs^^n

64. After the verb yuj to join, tlie Atmane-
pada is used, when it is preceded by pra and upa except
with reference to sacrificial vessels even when the fruit
of the action does not accrue to the agent.

The verb *ifSl^ 'to join* is svaritet, and consequently by sfitra
72 it is always Atmanepadi, when the fruit of the action accrues to the
agent. The present sfitra declares when this verb may take the termi-
nations of Atmanepada, even when the fruit of the action does not accrue
to the agent After the verb *r^ preceded by the prepositions ir and 57
the terminations are of the Atmanepada, when not employed in reference to
sacrificial vessels. As R^^tfe 'he joins or employs' ; ^TJ^% 'he fits or uses/
Why do we say "when not employed in reference to sacrificial
vessels ?" Because there the terminations will be of the parasmaipada. As

{Rf «rf% <rnrrT% *&*&* -

VArtika: — It shall be rather stated that the root takes the terminations
of the Atmanepada when preceded by any preposition beginning or ending
with a vowel. This is an important modification of the above rule. Thus
^*|£^ and faqw^. In fact, all upasargas, with the exception of ^, firr , and
3^, either begin or end with a vowel, and therefore the VArtika ^mounts
to the inclusion of all prepositions with the above excep^on^ CoOQl

Bk. I. Ch. HI. § 65-67. ] The Atmanepada verbs. lot

With ^n, f%^ and jr , however, the root will take parasmaipada termi-
nations ; as tig^frtr.

^T: ?^: II Vl II *^i£r II **?: , *$V , ( 3H0 90 ) B

65. After the verb kshnu, to sharpen, pre-?
ceded by sam, the Atmanepada is employed.

The verb ^3 'to sharpen, whet, or grind/ is generally parasmaipadi,
but it is 4tmanepadi when it is preceded by the prefix g^; as ^^jtttrc^Pr
'he whets the weapon' 3^[41?r 'they two whet' *J^<gq% 'they all whet'.

It might be objected, why has a separate aphorism been made of the;
verb kshnu, when it could well have been included in the sfitra 29 ante, which
also speaks of the force of the prefix sam, when used with the verbs gam
richchha &c. To this the answer is that, that sutra treats of intransitive verbs,
the word akarmaka being understood therein, while the present sfitra,
as is evident from the example we have given above, treats of transitive
verbs. > •

MiftR^ 11 \%\\ 9^rf*r 11 wt: f <mvsfl , ( mow ) h

Tf^rftii :

66. After the verb bhuj, the Atmanepada is*
used, except in the sense of protecting.

The root bhuj when it does not mean to protect, is Atmanepadu
This root belongs to the rudhadi class and has several meanings, as, to feed,
to cherish, to preserve, to eat, and to enjoy. As jpft ' he eats or enjoys ' also
bhufijAte, bhunjate &c. But *m\ yfa ftor 'the father cherishes the sons'*
^JHTrH^HpHilR^n .

The root bhuj belongs also to the tudadi class, but there it has the
sense of being curved or crooked. The bhuj belonging to the tudadi class
is not to be taken in this sfitra, because that bhuj has never the sense of
protecting. Therefore f%*prf3)r Tlf^r ' he bends the hand/ is in the paras-
maipada.

Slv^i 93 Wrft 3ft *frf^H «itll*livqi^ It ^9 II q^lfa II 5fr:^

*&b$ ^t ^Af ^» * : f *5^tf ,3RT*qj^ # ( W0 90 ) II

67. After u verb ending in the affix jji (causal)
the Atmanepada is employed, provided that when the-"
object in the non-iji or non-causal sense becomes thfr

o

152 __.: The Atmanepada verbs. [ Bk. I. Ca. HI. § 67.

agent in the causal ; and when it does not mean c to re-
member with regret ;' even when the fruit of the action
does not accrue to the agent.

The causatives (f*P^) are conjugated in the Atmanepadi, when
they are used in the passive voice, that is to say when a word which was an
objective case in the ordinary verb becomes nominative case in the causa-'
tive ; and the nominative case of the ordinary verbs becomes objective case
in the causative. In short when causatives are used in the passive voice,
they are Atmanepadi. This round-about phraseology has been employed,
because the base of the active and the passive causatives do ^ot differ in
form; as*ji0te*j9 ff^pranp^fpfr. ^

The phrase ' *fc (after the causatives) ' of this sfitra governs the four
succeeding sutras and is understood in them. The sfitra consists of the follow-
ing words : — dfr: ' after the causative f *f°tt ' in the non-causative ;* ^ • what;'
qp? 'object;' «& # in the causative;' «tqr 'if;' ^: 'that;' sraf 'nominative ;' *Hreitf
'except to remember.'

In general, by sfitra 74, the causative verbs take the terminations
of the Atmanepada, when the fruit of the action accrues to the agent The
present sfitra applies to the case where the fruit of the action does not accrue
to the agent.

After the verbs that take the affix f«r^, the terminations are those of
the Atmanepada. How ? When that which was the object when the primi-
tive verb was used non-causatively, becomes also the object when the deriva-
tive verb is used causatively ; and even that object becomes also the agent
as well. There is exception however, in the case when the verb means to
• remember with regret.' As HjlO«Ptf *R*PT tftdl«h|: ' the elephant-keepers
mount the elephant.' Let us paraphrase this sentence by transforming th^
primitive non-causative verb <4jl<]T<Ptr into a causative form. The sentence then
will be, <qi0^3 f **fr trcfo 'the elephant makes itself to be mounted'. Similarly
^Mf%qPd {fcw 3fci14ti: 'the elephant-keepers sprinkle the elephant'; and
OT&^fr 1V$t **I*I*N ' the elephant makes itself to be sprinkled ;' <Tprf^I *J**ir
THTR^ ' the attendants see the king 9 ; and $afcr% n*TT ^5R%T ' the king makes
himself to be seen/

Why do we say 'after the causatives?" 1 Because the rule of
this sfitra will not apply, if the verb though conveying the sense of a cau-
sal verb, is, however, not formed by the affix (5j^; as atffttfN ?ft*Pf ffttfT^T:
'the elephant-keepers mount the elephant;' and Ml ft gem mil ffvft WWiX f lh
'the elephant that is being mounted, mounts gracefully 9 . Here the termina-
tions are of the parasmaipada in the second case also. y

Bit I. Ch. III. § 67. ] The Atmanepada verbs. 153

Why do we say ' in the non-causative non-ni ' ? For this role will
not apply, when the primitive verb itself is formed by the affix ftj^ ;
such are the roots of the tenth class or chur&di in which the nich is
added in the self-descriptive sense, and not for a causative purpose.
Thus the root *mj belongs to chur&di class ; whose causative and primitive
forms are the same. As *rcreft *ntf *Ti4M«h: 'the cowherd counts the herd; 9
and the causative form is *JUj^ft TT: *TO*ta 'the herd makes itself to be count-
ed/ The terminations in both cases are those of the parasmaipada.

Why do we say ' when the object (karma) becomes the agent.' Be-
cause the rule will not apply, when any other k&raka or case, than the karma
or object, of the primitive verb becomes the agent in the causative. Thus
spnf?* siw ' he cuts with the scythe/ «snrcft *PT ^ra^f 'the scythe is made to
cut of itself/ Here the word d4tra was the instrument of the primitive verb
which became agent in the causative, and therefore in the causative the termi-
nations are of the parasmaipada only.

Why do we say' if in the causative 1 in the sutra? It is for the purpose
of indicating that the verb must be one and the same, both in the primary
sentence and the causative sentence. Therefore the rule does not apply
here where the causative verb ttaqflt in the second sentence was not
used in the 6rst sentence ; HJHlsPq ffert tfct<44ii: • the elephant-keepers
mount the elephant;' and HjlflqqHlUJ) **fl4l<U! % 1 «R1% iJrtT 'the elephant
that is being caused to mount is made to sprinkle with urine the frightened
men*.

The word 9: 'that 9 is employed in the aphorism to show that the
objects in both sentences must be the same and not different* Therefore in
the following example where the objects are different, the terminations of
parasmaipada are only used. HjiO^fcl f^PT * ft<nrarn ' the elephant-keepers
mount the elephant'; and ejrctfWT) **ft tqwjli)* <lft *T3«*P£ ' the elephant,
that is being caused to mount, makes men mount on the land/

Why do we use the word karti (agent) in the sfitra? If the former
object is not the agent in the second case, the 'terminations of the parasmai-
pada only would be used. As WCirff^ vfcpt cftOTOR ' the elephant-keepers
mount the elephant ;• and UH l fls^ft «J?pnw: ' the elephant-driver makes it to
mount' *

Why do we say f except when meaning to remember with tender-
ness?' Observe, CTtfir <H^*^U3 t£lfo<ri: 'the dove remembers, with regret, the
forest tree ;' and W3<54 **npl: W3^* 'the forest tree is made to be remem*
bered of itself. 9

Digitized by VjOOQLC

13* Thb Atmanepada verbs. [ Bk. 1/ Ch. III. § 68-69. *

sfo*^f*nft n \* 11 xt^ifk n *a-**ft: ; %^-fl^ , ( «t ,

WO X(0 ) tt

ihr: nf»>fo: grafts q««*nq«H3«i< »nfolyw »i

G8. After the causatives of the verbs bhl to
fear, and smi to wonder, even the fruit of the action
accrues not to the agent the Atmanepada is employed, -
when the fear is produced directly by the causative
agent.

The phrase 1 of the causative ending in ftr' is understood in this
sfitra and is to be supplied from the previous aphorism. This sfitra is also
restricted to cases where the fruit of the action does not accrue to the agent.
The word |tj in the aphorism has been defined later on in sfitra I. 4. 55
( Wf^T ^r^fi" ); it is the agent which is the mover of another's agent When
a fear is caused by a $5 it is called £<TTO- The word H*T 'fear* in the
aphorism is illustrative, and includes by implication fire*RJ 'astonishment 9
also. As *\(l<8t *ft**r% ' the jatila, the cock -headed frightens/ *r«*?fr Hfrra#
' the munda, the shave-headed frightens '; *ffe3t ft^TTnr^ ' the jatila astonishes/
5**fr fiwiinil 'the munda astonishes 9 i.e. the very fact of matted-hair or
shaven-head frightens' &c. .

Why do we say tyre? For if the fear or astonishment is not
the direct result caused by the agent, but arises from something else, then
the Atmanepada will not be used. As ^f^T^f qnrefc 'he frightens him with
the kunchika/ ^TD fagTPreffr ' he astonishes with his form*. Here kunchika and
rupa are the instruments, which cause fear or astonishment, and are not the
hetu thereof.

***% , ( «*: STW *0 ) tt

69. After the causatives of the verhs_gyidh to
•covet, and vafich to go, the Atmanepada is employed,
when used in the sense of deceiving, even though the
fruit of action does not accrue to the agent.

The phrase 4 of the causatives ending in fiSj* is to be supplied here from
sfitra 67. This sfitra is also restricted to the cases where the fruit of the
action does not accrue to the agent. The word iMH«f of the' sfitra:
means deceiving. AsiPPRi trf^'be deceives the boy/ «fpf*3i *^T*# ****
cheats the boy/ izedbydo gle

\

.Bk. I. Ch. III. § 70-71. ] The Atmanepada verbs. 155

Why do we say 'when it means deceiving.' For these verbs have
not this meaning, their causatives take parasmaipada. As J^pt ir&qft ' he
causes the dog to bark/ *rf8? q^^fa ' he avoids the serpent'

9*n^-^l<^«ft-^*$T: , ^ , ( ^: H^H% WTO TW ) N

70. After the causative of the verb 11 to melt
or stick, the Atmanepada is employed, when used in the
sense of showing respect, subduing and deceiving, even,
though the fruit of action does not accrue to the agent.

The phrase ' of the causative ending in ftj' is to be supplied here from
sfitra 67. This sfitra has its scope only when the fruit of the action does not
accrue to the agent. In the Dhdtupatha, there arc two roots 5ft , one is
technically called ?ft? and means to stick, and belongs to divadi class. The
other +tt meaning to ' melt' belongs to kryidi class. As there is no specifi-
cation in the sfitra what & is to be taken, both are therefore taken.

The force of the word ^r in the sutra is to include the word SffiT»Fr
*to delude' of the last sutra into the present. The word *wi*m means to show
respect The word ui^4)«h<^ means to subdue. As ^.ipKMI^S ' he gets
jespect or causes respect to be shown to him through or on account of his
matted hair.' ?pi% qfa*l3feU4*M 'the hawk subdues the partridge', *K^rrgSFni%
• who deceives theeA •

The sfitra ftwir tffoit: VI. I. 51 declares that the | of 5ft is optionally
changed into *JT before certain terminations. But there is no option allowed
when the root 5ft has any of the above three senses: in these cases the substitu-
tion of *rr is necessary and not optional. For the option allowed by sfitra
VI. I. 51 is a c*nfeptf% irqT and not a general ftqrar applicable every-
where.

Why do we say • when it has the meaning of, to show respect &c. ' ?
Because otherwise there is parasmaipada. As *nTOi5?t*ITOfir«

fi^m^i <^ M T4^^T% 11 toi 11 n^ i ft 11 f^T«^OTnjra x ,
f*: ; 3l«n% , ( ^t: WO xw ) ft *

71. After the causative of the verb kpi, the
Atmanepada is used, when it has the word mithya,
incorrect, as an upapada or dependent word, and is.

156 ___ The Atmanepada verbs- [ Bk. L Ch. IIL § 71-72.

employed in the sense of < repeated wrong utterance*
even when the fruit of the action does not accrue to
the agent.

The phrase ' of the causative ending in fa f is to be supplied from

sfitra 67. The aphorism is restricted to the cases where the fruit of the action

^ does not accrue to the agent. The word abhyAsa means doing again, or

~?f repetition ; as <rt fiwjr ***Hqfa means ' he repeatedly pronounces the word

incorrectly that is with wrong accent &c, not once but constantly*.

Why do we say 'when the word mithyA is used as an upapada?'
The causative of kp will take parasmaipada when it has any other upapada.
Thus <rt ygfc *hHq f% ' he pronounces the word correctly.'

Why do we say ' of the verb frsj? Because the causative of any other
verb used along with the word mithyA will not have Atmanepada; as <rf
ftw ^T^^lf^r ' he repeatedly utters the word wrongly/ Here T^ takes parasmai-
pada.

Why do we say 'repeatedly'? For if the incorrect utterance is not
habitual, then parasmaipada will be used; as <rt ftiqi4iH4fd * he pronounces

wrongly, not always but once/ "

72. After the verb marked with a svarita,
(svaritet) or which has an indicatory fi rfiit), the termi-
nations of the Atmanepada are employed, when the
fruit of the action accrues to the agent.

The anuvptti of the phrase i^: does not go further. The word
ftronrtt is a compound, meaning * fruit of the action.' When the principal
object for the sake of which the action is begun, is meant for the agent indi-
cated by the verb, there the Atmanepada is used after verbs having an
indicatory sy or a svarita accent. As *nr% ' he sacrifices for himself ' «p*% ' he
cooks for himself/ Here the verb tpr and *r\ are marked with svarita accent
in the DhAtupAtha, and therefore they take the Atmanepada terminations.

Similarly gj^ 'he presses the so ma- juice/ ^^ 4 he does/ Here the
verbs gsj and ^rsj have an indicatory 3j.

In all the above cases, the principal object of the action such as getting
heaven by performance of sacrifice, eating of food &c, is meant for the agent
That is to say, he sacrifices in order that he himself may attain heaven, he
cooks in order that he himself may eat, &c. • ed by l

Bk. I. Ch. III. § 73, 74. ] The Atmanepada verbs. 157

Why do we say ' when the fruit of the action accrues to the agent 9 ?
Otherwise it will have parasraaipada. As *nrf£*I *U*i4(l: the priests sacrifice,
(not for themselves, but for their clients)' q^rf^ct «rr^CT: 'the cooks cook (for
their masters)' apifcf chJ^MC ' the menials work (for their masters). 9 Here
though the fees and the wages are the fruits which the agent gets, yet as that
fruit is not the principal object for which the action was begun — the principal
object of the sacrifice was not that the priest should get his fee, but that the
sacrificer may go to heaven — the verb takes the terminations of the Parasraai-
pada. Here the principal fruit of the action does not accrue to the agent ;
but to a third party.

*«% ( sno tw ) it

fPrf: 11 wtiliw: ^rlpttra fe^nr^ *TT*H<I* >rcfa II

73. After the verb vad to tell, preceded by
aj>a, when the fruit of the action accrues to the agent,
the terminations are of the Atmanepada*

The phrase " when the fruit of the action accrues to the agent," is
understood here. The sutra is clear. As qpraiPft -m^q-H^ 'the wealth-seeker,
forsakes justice, that is to say, he wishes to acquire wealth at the sacrifice of
justice. But when the fruit of the action does not accrue to the agent, the
parasmaipada is employed. As W*tf%.

ftmg 11 ba a <i^TfSi 11 ftre: # 1 , ( ^Hifwni* feqrR%
wo xw ) n

f Rr- 11 ftnig <mi^u< >nrH *r%Pisr^ ftnmra 11

74, After a verb ending in affix i.ii (causal)
when the fruit of the action accrues to the agent, the
Atmanepada is employed.

The phrase "when the fruit of the action accrues to the agent" is to
be supplied from sfttra 72. The verbs that take the affix fiir* nich are gene*
rally causatives. As 3vi <KK4% ' he causes the mat to be made for himself sfo«f
<n^Rl*T ' he causes the food to be cooked for himself. When the fruit of the ac-
tion does not accrue to the agent, the parasmaipada is used. As qrcr 5nmft
TCttT he causes anothers ' mat to be made. *►

V*v » ^t^l^rr^^5rt ijrNr^: ^%Pmr% f*r**n?£ ^n^m* H*f?r *rof%.

158 - The Atmanepada verbs. [ Bk. I. Ch. III. § 75, 76.

75. After the verb yam to strive, preceded by
sam, ut and to, when it does not refer to a book, the
Atmanepada is employed, when the fruit of the action
accrues to the agent.

The phrase ' when the fruit of the action accrues to the agent ' is
read into this aphorism by anuvritti from sutra 72. As ffrsfcr ^RJ^3% ' he
gathers rice ' HPCyT^S^i 'he lifts up the load' qfemi4*&?t 'he draws out the cloth/
The root yam preceded by the preposition ^rr^f takes the terminations of
the Atmanepada by virtue of aphorism 28 ante ; but in that aphorism the verb
was intransitive, here it is transitive, and this explains the necessity of making
two different sutras for one compound verb SJRP£ .

Why do we say ' when it does not refer to a book.' For otherwise
the verb will take the terminations of the parasmaipada. As OTTtgft fef^Rqf
§3?: 'the physician studies diligently the medicine.

When however the fruit of the action does not accrue to the agent
the above compound verbs are parasmaipadi. As dq^ ffr, W4<k»ft, *7nre$fo

wgq^nhra: u b% 11 vFrfk 11 ^ly ra^, «: ( ^fam^

<3TTd qo ) II

ff^r. 11 vijwiTviiMri: 9&f>nira ferarcr^ *n*i%<T* >Rfir 11

76. After the verb jM when not preceded by
any upasarga, the terminations are of the Atmane-
pada, when the fruit of the action accrues to the agent.

The phrase ' when the fruit of the action accrues to the agent ' is
understood in this sutra also. Thus *rf siPfttl * he recognises the cow as his
own* ; ^TC^ *THt% ' he recognises the horse as his own.'

Why do we say when not preceded by any upasarga ? For when
compounded with prepositions, it may take the Atmanepada terminations ; as
^A <S|p3£ T *J*fHlPt *JT. ' the fool does not know the heavenly regions.'

When however, the fruit of the action does not accrue "to the agent,
the simple root <rr takes parasmaipada terminations. $???rer «tf STPTTft f he
recognises Devadatta's cow/

^PfT: II *rfn WWTOJ 90PtlecHJ|<|<|{ *PT R*fN*T% ^%Pm% ftf^RT^ fiPfHT*-
wfc*>Pinill Pnoalp

Digitized by VjVjijy IC

Bk. I. Ch. III. § 77, 78. ] The Parasmaipada verbs. 150

77. The Atmanepada is optionally nsedj when
the fact of the fruit of the action accruing to the agent
is indicted by an upapada i.e., by a word used along with
the verb.

By the preceding five sfitras 72 to 76, Atmanepada terminations were
ordained to come after verbs which implied the accrual of the fruit of the
action to the agent, that is to say, where the verb by itself denoted this. But
when the same idea, instead of being inherent in the verb, is expressed by
an upapada, that is by a word governed by the verb, the necessity of using
the Atmanepada affixes, to denote the same idea, is obviously removed, and
in such a case it is optional whether we use the Atmanepada or the Paras-
maipada terminations. As etf K«t *T5rf?t or *nnt ' he sacrifices for his own
yajna' H grfsrctfl* or 3j*% ' he makes his own cot/ ** J^nmffa or Bm^f,
FPfr? 5 * <rrec% or qre*rf% &c. So on with all the above five sfitras.

fnr: 11 i[miH*3Rq<*3M4 H*f?r flNi^H^w r ^j^fed hji^^jj^ ii

78. After the rest i.e., after all those verbs not
falling under any one of the previous provisions, the
terminations of the Parasmaipada are employed, in
marking the agent (i.e., in the active voice).

The rules of Atmanepada have been declared in the preceding 66
sfitras, 12 to 77. The terminations of the Parasmaipada, which are the
general verbal terminations, will come everywhere else that is to say where
its operation is not debarred by any one of the preceding aphorisms. The
present sfitra declares this universal rule. The word sesha or the rest, means
that which is the residue after the application of all the previous restrictive
rules. Thus it was declared by sfitra 12 ante that a root having an AnudAtta
vowel or a sf as it, will take Atmanepada terminations. As *mit, *r# . The
converse of this will take parasmaipada terminations, namely all verbs which
do not have an anudAtta accent or a r as it. Thus irrft ' he goes, 9 *rf% ' it
blows/ It has been declared by sfitra 17 that the root fttr when preceded by

f% takes Atmanepada termination, asftf%g3. Wljen not preceded byftbut
any other preposition it will take the parasmaipada termination. As 4jlfi*lft,

Why do we say "when marking the agent"? For when used in the
passive voice, the root will take the Atmanepada termination. As <TO(it # it
• is cooked ' ip^ 'it is gone/ y

Digitized by VnOOQ 16

160 The Parasmaipada verbs. [Bk. I. Ch. III. § 79, 80.

Why is not parasmaipada used when the verb is employed reflexively?
As in the following example : — q^r^ ^tRFK vmti-i ' the food cooks of itself.'
Because in the present sfitra, the word 3T<ifc of sfitra 14 ante is to be read in
by anuvritti, so that, in fact there are two 5t5ft in this aphorism which thus
means " when the agent of the verb is an agent pure and simple then paras-
maipada is employed." While when a verb is used reflexively ( ch*4ch<ift ) the
agent of the verb is also, in one aspect, the object of the verb; and it is there-
fore, that in reflexive verbs the Atmanepada terminations are employed.

79, After the verb kji to make, preceded by
ami and para, parasmaipada is employed, even when the
fimit of the action goes to the agent, and when the sense
is that of " divulging" &c.

By aphorism 32 ante, the root fr took the terminations of the Atmane-
pada when the sense denoted was that of " divulging, reviling" &c., and it
also took Atmanepada termination when the fruit of the action accrued to the
agent by virtue of sfitra 72, because the root fr»J has an indicatory sr. The
present sfitra makes an exception to those rules, and ordains parasmaipada.
Thus mjchflfi he imitates, qilchfift he does well.

fan: II ( w ) n

ffrp 11 ^iPnrf% ^ri^^re 'pt^Rrr: mwi *roftr 11

80. After the verb kship, to throw, coming
after abhi, prati and ati, parasmaipada is used, even
though the fruit of the action goes to the agent.

The root fin to throw, is svariteta therefore by sfitra 72 ante it would
have taken Atmanepada termination, when the fruit of the action accrued to
the agent ; this aphorism ordains parasmaipada instead. As sfftftmft ' he
throws on' itfcrftrrft 'he turns away or rejects,' aiftftrrft 'he throws beyond/

Why do we say "when coming after abhi, prati, and ati"? Because
when compounded with any other preposition, it will not take parasmai-
pada, but will be governed by sutra 72. As grrf^TTt * he throws down/

The second q^ of sfitra 14 is also understood here, so that when the
verb is used reflexively, the agent not being purely an agent, the present
sfitra will not apply. As qr ftfiim3 ^j^r ' it is thrown on of itself/ Ogl

\

Bk. I. Ch. III. § 81*83. ] The Parasmaipada verbs. 161

81. After the verb vah to bear, coming after
pra, parasmaipada is used, even though the fruit of the
action accrues to the agent.

The root q? to carry is svaritet, and by sfitra 72 it would have taken
the Atmanepada termination when the fruit of the action accrued to the agent.
But the present sfitra ordains parasmaipada instead. As JFTffir 'it flows.'

Why do we say " when coming after pra"? Because after any other
preposition it will not take parasmaipada as 9ITO9# 'he brings. 1

jft^K n ^ n ^fa II *ft* I Wl ( TO W ) B

ffrf: 11 ^rfr^x^^Tc^rt *nfif 11

82. After the verb mjish to bear, preceded
by pari, parasmaipada is usedj even when the fruit of
the action accrues to the agent.

The root *n* fi to suffer" is svaritet, and by sfitra 72 it would have
taken Atmanepada termination when the fruit of the action accrued to the
agent. This ordains parasmaipada instead. As qpnrorflt he ' endures or he
becomes angry. 1 When not preceded by this preposition, it takes Atmanepada
termination. As W|**pl .

According to some authors, the root.*? of the last aphorism is said to
be understood in this, so that q? preceded by <rft will also take parasmaipada
terminations ; as qftq^ft .

( QV qo ) H

sfa: 11 Ptmi<sRfriU 'C"T r,r ^ : ^^* *Frf* 11

83. After the verb ram to sport, preceded by
vi and &ii, parasmaipada is used.

The verb r* means 4 to sport.' It is anud&ttet and therefore by sfitra
12 it would have taken Atmanepada terminations ;'the present sfitra ordains
parasmaipada instead, as f%^7if% f he takes rest '; aiirofct ' he delights in,' qrftTRflt
1 he sports. 1

With other prepositions than these, it will take Atmanepada termina-
tion. As hjPKw) •

Digitized by VjOOQLC

162 - The Parasmaipada verbs- [ Bk. I. Ch. III. § 84-86.

*m^ 11 t% n t^Tftr 11 *nm, i ^ i ( w wfo<z?\) n

84. And also after upa, the verb ram takes
the affixes of the parasmaipada.

As foj-dyuqffr 'he causes Devadatta to refrain.' The sense of the
.verb uparamati is causative here, as if it was <fUH4ft • This is an example
of a verb involving in it the force of the causative affix f*pf .

It might be asked why was not the preposition 3T read along with the
other preposition in the last aphorism, instead of making this a separate sfitra.

The answer is that the necessity of making a separate sfitra with the preposi-
tion upa arose from the exigencies of sfitra making. It is desired, that the
next sfitra 85 should apply only to the root upa-ram, and not to other com-
pounds of ram ; and this could only be done by making these two distinct
sfitras.

fwmw*4«i^ ii i\ ii v^fk ii fkwm 1 s^forc* ( w
to w ) n

iPr 11 qHHflIt<H3w T <Mft'H si m£M *ref*r 11

85. After the verb ram preceded by upa,
parasmaipada is optionally used, when employed intran-
sitively.

The last aphorism ordained parasmaipada absolutely, this declares
an option under certain circumstances. As qii^-'h^'UHft or 4<U<|3 he
desisted from or was quiet while it was eaten.

ffrf: 11 J>T5>T^EnPf ^p 3 5 ? ***f^*lT ,y 'l^ v ^ <hA<i4 *f*i% 11

86. After the verbs budh to know, yudh to
fight, naSto destroy, jana to be born, ifi to go, pru to move,
dru to run, and sru to flow, ending in the affix iji (i.e.,
when used in the causative), parasmaipada is employed,
even when the fruit of the action goes to the agent.

By sfitra 74 ante, causatives took the &tmanepada termination when
the fruit of the action accrued to the agent This makes an exception to
that and ordains parasmaipada. As *ftrqfct he expands 4ta*rf3T he causes to
fight ; OTraft he causes destruction, srraft he begets; qtuOTOft he teaches ;
snroft he causes to obtain ; grroft he causes to melt ; traraft he causes, to

trickle; Digitized by Li O

Bk. L Ch. III. § 87. ] The Parasmaipada verbs. 163

Of the above eight roots, those which are intransitive would have
taken the parasmaipada terminations under the conditions mentioned in sfitra
88 sub, that is to say, when the agent is a being endowed with a reason. The
present aphorism in the case of such intransitive verbs makes this additional
statement, that those intransitive verbs will take parasmaipada, even when the
agent is not a being endowed with a reason. As ^nrafa TO*J he makes the
lotus to expand, ifrroflr 3TOTfir he makes the woods to strike each other, sire-
*rf?r J .^T he destroys the sorrow, *H*jfa *p!P£ he produces pleasure.

Again of the above eight roots, those that have ordinarily the sense of
"moving" will get parasmaipada by sutra 87. The present sutra, in their case'
makes this additional statement, that they will take parasmaipada termina-
tions even when the sense is not that of moving. Thus *, means both 'to move/
and 'to obtain/ * means • to run' as well as ' to melt,' and w, means ' to flow'
as well as to ' to trickle/ As sret he obtains; mJiftqfft the iron melts ; Kp^r
*T?rf*fr the water-vessels drip. The examples in the first paragraph have there- .
fore been thus translated

The root $y is always compounded with sjfa •

fSni^I^WI^ig U 4$ I) tJ^lfSl U fSprc^l^-3|W: I
^ I ( *t: q<* fr q qq ) H

*rfW* II *1%: gnt»ftTOg* : II

87. And after the causatives of verbs which
have the sense of the ( eating or swallowing* and c shak-
ing or moving' parasmaipada is employed, even when
the fruit of the action goes to the agent.

The phrase ' when used in the causative (%:) of the last sfitra is to
be read into this also. The parasmaipada is ordained as an exception to
Sutra 74 by which Atmanepada was ordained when the fruit of the action
accrued to the agent The word ftmilf means ' eating,' and ^ratf ' means'
• moving/ ' shaking.' As fijimft he causes to swallow ; ?mroft he causes to
eat H)r*nrf?r he feasts. ^5^rf?r he moves ; <$TTOft , 3R*nrfq he shakens. This
aphorism applies to transitive verbs, and to verbs whose agents in non-causa-
tive state are inanimate objects *>., not possessed with reason.

Vart : — The prohibition of the root %n to eat, must be mentioned.
The causative of *ff to eat, takes Atmanepada. Thus a*r% fc^T: Devadatta
eats; 3tf?*i3 I^^R he is made to eat by Devadatta.

wm^^^^f^i^^^Tc^n 44 u xnpfa N ^^ , g^
^t-t^i mi^i «^v.^( ^: *w qo ) n

164 The Parasmaipada verbs. [ Bk. I. Ch. III. § 88.

88, The affixes of the parasmaipada are em-
ployed after the causal of that verb, which in its non-
causal state was intransitive and had a being endowed
with reason for its agent ; even when the fruit of the
action accrues to the agent.

The phrase d": is understood here also. fThe Atmanepada was
ordained by sutra 74 when the fruit of the action accrued to the agent
This aphorism is an exception to that, and ordains parasmaipada. As *rrc$
S-TStT: Devadatta sits: ijra^ fa? ^T he makes Devadatta to sit *}$ t*TV
Dcvadatta sleeps, *rnr*rf% fafW he causes Devadatta to lie down.

Why do we say ' in its non-causal state? For if the causative' root is
from another causative root, and not from a simple non-causative rOot, then
the parasmaipada will not be employed. Thus ' if one person such as Deva-
datta, causes another person Yajnadatta, to cause a third person Ramadatta,
to do an act denoted by an intransitive root, as ' mounting * for instance,
though Ramadatta mounts for his own benefit, the verb to be used in such a
case is qtf<tl*Ft in the Atmanepada, from the root 3]TC? to mount; and not
mifl^fo ' (Iengar's Guide to Panini).

Why do we say ' which was intransitive"? This rule will not apply
if the verb in its non-causative state was transitive. Thus from qoft 1 the
causative from ^r, if the person caused to do the act denoted by the verb does

it for his own use, comes only thH4$ though the person has a will ; for the
otiginal root fT though a non-causative root, is not an intransitive root 9

(Ibid).

Why do we say 'having a being endowed with reason for its agent?
For if the agent is a non-sentient object, the verb will be Atmanepadi. Thus
from sfrfa the causative from lyr 'to dry' if that which dries is a thing not
possessed of a will as jft^r: ' paddy/ for instance, though the fruit of the
action i.e. t the drying, affects only the paddy itself, comes ;fr<nid in the
Atmanepadi, though 'to dry* is an intransitive root ; e. g ., piFR# gflfolK W :
the sun-shine causes the paddy to dry/ (Ibid).

Bk. I. Ch. II. § 89-91. ] The Parasmaipada verbs. 165

89. Bub the affixes of the parasmaipada are
not used, after the causals of the verbs p& to drink, dam
to tame, &yam to extend, ayas to exert oneself^ parimiih.
to be bewildered, ruch to shine, nyit to dance, vad to
speak ; and vas to dwelL

The last two aphorisms had ordained parasmaipada instead of Atmane-
pada, even when the fruit of the action accrued to the agent, thus debarring
the latter's action which would otherwise have taken place by virtue of sutra
74. This sutra prohibits the last two s&tras, and re-instates Atmanepada of
sutra 74. Thus root qr to drink, has the sense of nigarana or swallowing,
the roots 5* &c, have sentient beings as their agent; the root tr^ to dance has
the senre of «5RTO or moving, but still these verbs have Atmanepada affixes,
and do not take parasmaipada in the causative. As qnr^ he causes to drink,
$Hnd he causes to be tame, ^iqm^d he lengthens, Mjiqmq3 he troubles,
<T KH)gq3 he entices, n^trt he makes agreeable, «r5*rt he causes to dance,
<iimd he makes to speak, qrcro^ he causes to dwell.

Vart : — The root *iz should be enumerated along with qr &c Thus
>TTr$$ fii^ji ^?Nfc The doe suckles a young infant. .

srere: II %0 II vc^fik II ^T I *mz I ( wfl q ^ ) ■ " "

90. The affixes of the parasmaipada are used
optionally after the denominative verbs ending in the
affix kyash.

The affix z&m is ordained by sutra III. 1. 12 *prrft^ WT^[ t^nTV $5f:
after the words <f)^r &c. These root take optionally parasmaipada. As
<*fR«iqf*r or 3 he reddens. 444414ft or % he makes pa( paj.

^«ftg% 11 <si n q^rfSr u ^rp 1 ^fe i ( *n *ns?fc-

ff^r: 11 wrrf^^ g% *r 4<**>4* H*ft 11

91. Alter the verbs dyut to shine &c. the
terminations of the Parasniaijmda are optionally em-
ployed, when the affixes of lull (aorist) follow.

The Dyutadi verbs are 22 in number, to be found in DhatfipAtha in
the Bhuadi class. By the use of the word «£?«?: in the plural in the sutra,
the force is that of &c. These verbs are anudatta and so by sutra 12 they
would have been invariably atmanepadi, this aphorism makes them optionally

166 The Parasmaipada verbs. [ Bk. I. Ch. III. § 92,93-.

. * J ,■ ■ n 1, 1 ■ . 1 1 1 ■ ■■ ii 1 ... 1 ■■!.■■ ■

so in the aorist Thus ^rfyr^ or «f*4TM8 he shone. In other tenses than the
aorist ( W ) these verbs are invariably Atmanepadi. As «fta% he shines.
For a list of Dyutddi verbs see Dhatupa^ha.

tn^fcr* ) »

92. After the Verbs vyit to exist &c, Parasmai-
pada is optionally employed when the affixes sya (Fu-
ture and conditional) and san (Desiderative) follow.

The TcTlffr yerbs are five in number and are included in the *J<nft sub-
class. They are ?sr to be, f% to grow, *re to fart oi break wind; ^^J to ooze,
and ^J to be able. As 1st Future ?*t?rf<t or qfV^tff, it will be, conditional
^T*^^ or *nfa«^; Desiderative ftf^pf or fir^RW &c.

In other tenses than, the above, they are always Atmanepadi As
*3%itis.

§f^$*r: 11 ^ ii r^xfk 11 §fe i ^ i ^t: » ( ^re%r *n

93. After the verb kjip to be fit, Parasmai-
pada is optionally employed, when lu$ (1st Future) is
affixed, as well as when sya and san are affixed.

The verb fpv is one of the five verbs of the sub-class, ^mf$ of the last
aphorism. Therefore it will take both parasmaipada and Atmanepada, when
the affixes ^or^T follow. The present sutra makes the additional declar-
ation in the case of rst Future or 5?. Thus in Lu|wehave: — 3TFtn% or
*h<-Mlfl thou wilt be; in 1st Future we have : — 3fifimqd or tfiwqfi f he will be;
in the Desiderative we have: — f%3|-*tf^ or Pf*rffc<re% ; in the conditional we
have : — H1***W^ or tn^TO*. —

Digitized by VjOOQiC

BOOK L
Chapter IV.

s*ts3sto^5t *>st ii 3 ii rnpfa n sn «h«IW » ^T-wr n

tnr ii**i4i<<H^<flf5i $Rci»^ ii

1. From this sdtra up to the aphorism Kada-
r£h Karmadh&raye (II. 2. 38) only one name of each
thing named is to be understood.

What is that name then ? That which comes last, where the claims
are otherwise equal (I. 4. 2.) and that which were its claim disallowed, would
have no other opportunity of conducing to any result, would be the recognised
name. As a short vowel is called ' light ' by 1. 4. 1 o f and it is also called 4 heavy '
when it precedes a conjunct consonant (I. 4. 11.) Thus a short vowel has two
names * light' and * heavy/ But it will not be called ' light ' when it precedes
a conjunct consonant, but will have only one name, i*. 9 * heavy.' Thus in
f%% to divide fe? to split, the ^ is ' light ' while the same letter is 'heavy 9 in
ftrw teaching ftw begging.

Thus in the root ^ the m is 4 heavy ' and therefore in forming its
aorist we have the form 3prera ?T . Similarly MJU*i^ . The rule VII. 4. 93 not
applying here as that rule is applicable to laghu vowels only.

2. IVhen rules of equal force prohibit each
other, then the last in the order herein given is to take
effect.

^ yhe word ftitftro means ' opposition of rules of equal force.' When

two topics having different objects in view find scope of action simultane-
ously in ojne particular case, that opposition of equal force* is called

16$ - Nadi defined. f Bk- I. Ch. IV. § 3.

vipratishedha. A general rule (utsarga) and its exception (upavAda), or an in-
variable (nitya) and an optional (anityo) rule, or an antaranga and abahiranga
rule, are not rules of equal force. There the stronger prevails against the weaker.
As an example of rules of equal force, see VII. 3. 102. and VII. 3. 103. The
first rule declares, 'when a case-affix beginning with a letter of yan pratyAhAra
follows, the long vowel is substituted for the final of an inflective base ending
in a short Hf. f As Vpksha + bhyAm « VpkshAbhyam. The next rule de-
clares : — ' When a plural case-affix beginning with a letter of jhal pratyAhAra
follows, if is the substitute for the final short «T of an inflective base. 9 As
Vpksha + su=Vriksheshu. But when the plural case-affix bhyafc follows, what
rule are we to apply ? For the letter bha belongs both to the pratyAhAras yafl
and jhal. Are we to lengthen the shorter, or substitute^? The present
sutra gives the reply, ^ is to be substituted because VII. 3. 103 ordaining q
follows next to VII. 3. 102. Thus Vpksha + bhyarj - Vpkshebhyafc.

fl W H Wft **$ 11 ^ M T*Vft " t^^-^HRT, ^ II

*jfif: 11 |*ui?d*i«hKiwl ^ ***iwwi w%&<? i^f *W ^nrfir 11

3. "Word-forms ending in long I and ft being
names of females are called Nadl.

The word sms compound of f +.^T. The word stryAkhya means that
which by itself denotes the name of a female. These words must be always
feminine, having no masculine of the same form, as the word grAmani has.
As the words <*>H\(t KumArl a virgin, *r*pr yavAgfifo rice gruel. The declen-
sion of .nouns of nadl class is somewhat peculiar which will be treated of
later. As see Rule VII. 3. 1 12 qgr^r is the augment of the case-affixes having
an indicatory 9 s when they come after a word ending with a Nadl.

Why do we say ending in f and 3T ? Because feminine nouns not
ending in these vowels will not be declined like Nadi words. Thus while the
dative of ^$ttO will be mm^t , the dative of «f|c will bejf*% .

Why do we say ' which are feminine ? Because if they are names of
males, they will not be called Nadl. As STHUfr: leader of a village ; ttanffc
leader of an army; *jp**r: a sweeper; their dative being *TPF$ , ^^TF^ tgvict .

Why have we used the word AkhyA ' name ' in the text / Because
feminine gender must be denoted by the word itself and not by any other
epithet used along with the word. Thus if the head-borough or the sweeper
should be of the female sex, the Dative Singular would still be iffT^ fori,
and 3rvF%f€r4*

% qvq^ranr *sft n 8 it vqfa 11 *r, : .f^**«^sOTit 9
«R*£, (3*$) 11 v

Bk. I. CH, IV. § 4*6 J ; NADI DBFfNED. 169

4. Feminine words ending in I and ft which
admit the substitute (v^) iy«A and ( **^) uvaft (VI. 4. 77)
are not called Nadi ; except the word stri, (wliich is called
nadi notwithstanding its substituting iyaft).

The definition of Nadl given in the last sfttra was rather too wide,
this limits the scope. Thus *fr, happiness, ^ brow, admit the substitutes
iyaft and uvaft respectively, and are not nadi. Their vocative singular is
t *ffc , $ *£ &c. f while the vocative singular of stri is $ f%T

tuwfa n h it v^rfk ii *t, -onf*, ($*^re-$-?n^) *

5. Feminine woixls ending in 1 and ft, though
admitting iyaft and uvaft substitutes, are optionally
termed Nadi, when the affix &m (Gen PI.) follows, hut
not so the word stri, which is always Nadi.

*fr + 8ip£ - *fr*^ + *n^« firo^;^* *n* - *gn^; or %fr + van
=*& + 5\ + ^^(VII. i. 54.) - *fiajP£; ^ + %n\ - *gOT*. But stri i* always
nadi, and we have V»AUJIH strtn&ra.

To the absolute prohibition enjoined by the last sutra, this allows an
option in the case of Genitive Plural

f|pfH|^Ri h % ii wpfii 11 fefk fe;ffe) |w f *,

^f%: 11 fefa <rctfr s**v cfft *p*&ft v. €«ro^ pnpnreirfr n w *r
h^Wiit wr: It

6. When a case-affix having an indicatory
it (nit) follows, then feminine words ending in short 1
and ft are optionally termed Nadi, as well as feminine
jiouns in long i and ft which admit of iyaft and uvaft ; but
not so the word stri, wbich is always Nad^

Feminine words in long t and ft have been defined as nadl, words in
short vowels can never be termed nadt, while even some words in long vowels
have also been excluded from the scope of the definition if they take iyaft and /
uvaft. The present sfttra declares an option i nthe case of all ^he above words, H&*
when a case-affix having an indicatory «r follows. ~~ ££ *W **** ^^tijX^

, The case-affixes having an indicatory A are the Dative, Ablative/* V u

Genitive and Locative singulars. Thus we have 1^ Digitized by GoOgk

170 _ ' Ghi defined. [ Bk. I. Ch. IV. $ 7-£

-------- ' * ■ - ' * "* p

Dative *&& or mil S?rt or $?! faH or fa£ ; Ablative Genitive iflfc
or «rarr: Srr or d^n fa*: or finff: ; Locative *nft or ^nr^^fi' or $*H prfir QT
fiptf . So also ^ or gt &c. But stri is always fe& &c.

3&t t^ref^ ii 3 ii ti^TtSr ii Sta: , fa, sref?s, (&&) a

7. The rest of the words that end in short * and
* are called ghi with the exception of the word sakhi.

This defines the word ghi. The peculiarity of the declension of ghi
words will be treated later on* The word hrasva is understood in this sutra.
The word £esha or f the rest* implies ' the words ending in short f or 7
which are not the names of females, or if they are feminine names, they are not
nadi words. 9

As atffcr fire. Before case-affixes having an indicatory ft, the ghi
nouns gunate their vowel VII. 3. 11 1. As stffcf + # « sp# + 3~9P7$ to the
fire. So also ^nj x £- ^nr? to the wind. But the declension of sakhi in
these cases is : — wsm ^^ ^*£ and tjiHu .

*ft: WRm q* II * II M^lPf II t|fer: , ^fT%, ^E| f (fa) 9

.8. The word pati is called ghi only when it
is in a compound.

The word pati would have been ghi by the laist sfitra ; the present
sutra is therefore a niyama rule, restricting the use of ghi in the case of pati to
its occurring in composition. The word eva ' only ' is used in a restrictive
sense.

As q?qr by the lord but *MmpH» by the lord of creatures. So also
q?*t and S |«u*i<i3 ; q?*p and mn*pfc ,' <*?*& and inirmV &c. When qfSt is ghi,-
there is guna of the vowel before the four f&^ affixes by VII. 3. 1 1 1.

TOig^Eara^fa m 11 % 11 t^rfa ii TOt-gw, ^fa,
*T, (qft: fa) It

9. The word pati when used in connection*
with a noun ending in the sixth or genitive case, is ghi,
optionally, in the Chhandas (veda).

The woid pati is understood in this sutra. By the last sfitra, pati
"would have not been ghi when not in composition. This sutra makes an
exception to thai when this word occurs in the Vaidic literature* y

Bk.I. Ch. IV. $ 10-12. ] Laghu and GurIa 17*

tfrc^Hf **& or q?§ H*r. salutation to the lord of the Kulunchas.
Why do we say 'when used in connection with a noun m the gen if'
tive case '? Observe *RT <TW Miif^jqm : .- j-

Why do we say * in the Chhandas '? Observe «rre^T*t . .

^fff: ii j^rorc g y *a *wf% ii

10- A shoi-t vowel is called c light ' (laghu).

This defines the word Laghu. A hrasva vowel which has already
been defined (I. 2. 27), is under certain circumstances called laghu or * light 9 ;
thus the i of ft? to break, is laghu; and by being laghu it is gunated before
the affix fir + *JT , as >tor he will break, by virtue of the rule VII. 3. 86, which
declares that a laghu penultimate vowel is gunated before a Sarvadhatuka or
an drdhadhatuka affix. So also 3*TT , aHtaft^ and wflfT^ • '

^TSfcj* 11 n it ^rf^r ii ^nt, 3s, (ipew) »
<jfa. 11 *Uf)r<f<rc<Tr jrwrstf «r^Efif *wft 11

il. When a conjunct consonant follows, 9
short vowel is termed i heavy * (Guru).

A hrasva vowel however is not to be called f light' when it is followed
by a conjunct consonant. As the i of fin?TT 'learning' is a heavy vowel. Thus
Rrtt is derived from ftw + *? (HI. 3. 103, let the affix a? come after that verb
which has a heavy vowel and ends in a consonant when the word to be
formed is feminine}. So also ^**T and fiwT •

^5^r « ir 11 v^Tft 11 ^Hh, * f (gs) n

^fin 11 $ttaiiku T {|44]f mfa 11

12. And a long vowel is also termed heavyv
(Gnru).

This is clear the anuvritti of the word conjunct is not understood in this
sfitra. As 1 of f ffa* he endeavoured, fcrfcnfc he saw. Here the letter f is
guru, and because of its being called guru, the rule III. 1. 36 is applied in
forming the perfect tense by the addition of the augment W\. Thus all long
vowels, and short vowels followed by conjunct consonants are guru or heavy,-
AI1 other short vowels are laghu. *

q^nc^Xlc^fafMW^lf^ *c*nfc J^MII^H ^ffa H TON* II

lf%: ti wti^imret frfhnl qrmtat nrffiprffacrirs^ m^^i n&A iKwlypinlt'
*rcfii 11 . _ *Q$\£ ,

"IT* _ ..-Anga defined: . [ Bk. I. Ch. IV. § 13.

*^ m ^*** m ' " * * " ' »^— ■ ii ■ ■ i ■ i .I ■ I . .I. . . < ,

13. After whatsoever there is an affix enjoin-
ed, whether verbal root or crude-form, that which begins
therewith in the form in which it appears when the
affix follows it, is called an Inflective base (anga).

The words of this sfitra require some explanation. Yasmat after

whatsoever ; pratyaya-vidhih compound of pratyaya meaning an affix and

vidhih, a precept, i.e. a rule enjoying an affix ; tadidi, i. S. that which begins

therewith ; pratyaye 7. SL=»in a pratyaya i.e, when a pratyaya follows (I. 1.) is

. called a base.

After whatsoever there is an affix enjoined, whether it be a verbal
root (Dhatu) or a nominal base (pratipadika), the word-form having that as
its beginning, is called an anga, with regard to the affix that follows. The
word yasmat is used in the aphorism to point out the thing named, as anga ;
because the word tad&di follows it. . This defines the word anga. The
word ' base ' iS thus a relative term with regard to its affix. Thus ff + nr «
qrs? he will do *?tf he will lose, 3*ft«*rfir , *ft**fH Here because the root
^p and J gets the name ^fr they are gunated by (VII. 3. 84) similarly because
37J &c. get the name Anga, their vowel is vriddhied in a?fai]T: &c.

The. words kfi hri are anga with regard to the affix -tt&c. 3*TJ +
»l^ — ^nif: ; ^rrnr: • Here upagu and fcapafu are anga with regard to an.
Similarly ff + &[ + *: - 4ifttq + *: = 4iR«4l4: . Here the whole word*
form Karishya is regarded as anga, and as such the short a is lengthened by
^ VII. 3. 101 ; because though the affix vafc is enjoined after the word kp, the form
-which begins with kfi i.e. karishya will also be calfed anga when the affix is
to be added. The word tadddi, therefore, has been used in the sfitra, to
make the definition of Anga applicable to the forms which a word may as-
sume after taking the. intermediate affixes like sya &c», or ^ before the
final affixes. Thus ap*t + H% + f ~ ap**^ + T - * u *iift (VII. 1. 72 and
VI. 4. 8) kunda + num + 1 — kundan + t = kundini, bowls. Here the
whole form kundan is called anga, and as such it lengthens its vowel before
the case-affix t by VI. 4. 8.

Why have we used the word pratyaya? Without it the rule would

have ran thus : — ' After whatever there is anything enjoined &c., is called

1 1 Anga.* Then in sfr + finft « ^Tinft - Here sandhi of vowels is enjoined

between f + f — f . If &ft was here an Anga, then its last vowel would

have been replaced by j*T3£ (VI. '4. 77), the form being farfSrsfa ./'U ^ ^

Why have we used the word f%ftr? Had we omitted it the rule would
have run thus :— " After whatsoever there is an affix, whether root or prati-
dadika is Anga.' 1 Thus in *fa W^j though the affix adfaunA is placed after

BK. I. CH IV. f 14, IS- £ PADA DEFINE*.. Wtf

the word dadhi, it is not enjoined by any rule ; and the word qtfSfr is not- called -
anga. Had it been so called, then the f of dadhi should have been elided
by rule VI. 4. 148.

The word pratyaya has been . repeated twice in this sutra in order to-
show that when an affix is elided, the terra anga will not apply to what stood
before it. Thus in the compound word fe$ -I- *& ■» && for the sake of
the woman. Here in forming the Tat purusha compound the dative case-affix .
after the word wit has been elided, for as a general rule case-affixes are<
elided in forming compounds* If the word jpft after its affix had been elided,,
still retained its old designation of Anga, the ^pfr + qjtT would have been;
ftsnrtf , the augment iyart being added by the rule already referred to above.*.

sfiwei t^*r 11 98 11 *KpfSf n ^-{erer^rei^, *V^«

ff^r: if g^F*f fire^f ** *r*T5tf *%*$ fifii 11

14. That which ends in sup ( case-affix )*
IV. 1. 2; or in tift III. 4. 78 (tense-affix), is called a pada
or inflected word. • #

The sup or case-affixes are those by which nouns are declined ; and
tin are tense-affixes by which verbs are conjugated. They have already*
been given before. Thus f^^5ro + T^-l^^5TO^ , : the Br&hmanas qref&r they cook.

It might be atked by a caviller why the word *t??r has been used in

the aphorism, for by the rule of tadanta given in'Sfitra I. 1. 72 a rule relating

to sup will mean and include also that which ended with a sup-affix. To this

we reply, that, the very fact that the word anta is used in this sfitra, indicates

by implication (jnipaka) that the Tadanta rule of Sfitra 72 Chapter I doei

not apply to rules of sanjnA (definition) made with regard to affixes. Thus

tarap and tamap affixes are called gha by Sutra I. 1. 22. The tadanta-vidhi

will not apply here ; words ending with these affixes will not be called gha:

Thus Hitufujflrj will not be called gha, for had it been so called, the long %

will be shortened in irrft fTRgfitffiTT • In short, ' an affix when'employed in a

rule which teaches the meaning of a technical term (sanjna) does not
denote a word-form ending with the affix/

H: ^ II ? II tr^Tfa H *P • *^> (*VQ *

15.. The word-form ending in n, is called
pada, when kya follows (Le., the affixes kyach, kyaft,
andkyash).

These are affixes by which* denominative verbs • are formed from
nouns. See. III. 1. 8, 1 1, and 13. A word ending in «r is called pada, when
these affixes follow. Thus *npr % + **C\ - *nft* 3rd J^dSy cnfhtf* he

-174 ~ Pada defined. { Bk. U Ca IV. $ 16, 17.

behaves like a king. So also CHP^ + «R^«tt*Fnr9 (VII. 4. 25.) n5«^ + «R^«
Wihrif or ^ifarfa . The result of its. being called pada is that the *r of tnr^ ,
"^^&c, is elided by SAtra VIII. 2. 7. (there is elision of sf final in a pada
which is entitled to the designation of pr&tipadika). Thus CTSP^ + W\ •

*rjr + it - cnfhr (VII. 4. 33).

These three affixes ^l^\ &c, come after case-inflected words (i>. t
words ending in sup), and though, before these affixes, the case terminations
are elided, still by Sfitra I. i. 62 such words would have retained the name of
pada which they got by Rule 140! this Chapter. The present sfitra however
makes a restriction (niyama). It declares that only words ending in *j retain
the name of pada ; while all other case-inflected words before these affixes
do not retain that designation. Thus the words qm speech, g^ a ladle are
not treated as pada and we have *pnnl and ymd . Had they been pada,
the ^ would have been changed into a? by VIII, 2. 30.

ftfar * 11 94 n ^if^f 11 *-^, *r f (^j)

3%: 11 ftrf?r *&& wr: <# tt^t *rcfir 11

1G. When an affix having an indicatory o

follows then that which precedes it is called pada.

The sfitra 18 of this chapter teaches that before certain affixes, the
preceding word is called bha. This sfitra declares an exception to that by
anticipation. Thus Rule IV. 2. 115 declares: — "Affixes 8^ and ej^ come
•after the vriddha (I. 1. 74) wordspn^." Here the affix 13^ has an indicatory
^ , therefore the word *nnj, standing before it, will be called pada. Thus
*l^ + ^ m »ra?fcf: VII, 1. 2. belonging to you. The result of being pada is
that n is changed into f (VIII. 2. 39). Similarly ^is ^ in *ra[ (Alter the
word firni there is yus V. 2. 123). Thus ^mfcj :; so also q^V. 1. 106, has
jE^as indicatory. Thus ^jftfTj: There is no guna because of its being pada.

^f%: .11 ^nf%5 '^rSj <ron «$HN*qMif*JdM <nf irtfst *r*f% it

17. When the affixes beginning with g
(IV. 1. 2) and ending in ^ (V. 4. 151.) follow, not being
Sarvan&masthana (1. 1. 43) then that which precedes is
.called pada.

The affixes beginning with su and ending with leap ana meant by the
pbove sfitra. Thus the case-affix iqr>((Ins Dual) is an affix included in the
iibove. Thus xm% + «IP£ - CfSf*^, KXsrfH'. , CHPf # KHtol, CHTOn, KtWJH: •
The.Jf is elided by being pada* Digitized by v

Bk. I. Ch. IV. § 18, 19. ] Bha defined. 175

f Why do we say ' when not a SarvanAmstliAna ? Observe <m^-+ WV

• Cnrnft" two kings, nw*: kings. The ^ is not elided. - - ,

ift WT II %C || TT^lfa II *T-3?fa f <OT f fanf^TOO)

<iifn«&^ II THt4f^-ilHH^f qi^HtaitWtll
*lf3<*H II ^HUH^ifr II

18. And when an affix, with an initial y or an
initial vowel, being one of the affixes, beginning with
sn and ending in k, follows, not being Sarvan&masth&na,
then what precedes, is called Bha.

This debars the application of pada. Thus ipfr + *P^(IV. 1. 105)
m*& a grandson of Garga. So also TOOT: . The affix yan begins with a
ya and the word garga being treated as Bha its final *r is elided before the
affix by VI. 4. 148.

So also qn=T + r*J - *n%: (IV. 1. 95) grandson of Daksha Rlftf: .
Here the affix %j^ begins with a vowel the word standing before it being
Bha, it causes elision of the final a? of daksha.

The word irf^f is in the 7th case meaning when ^ or *pj follow; and
by the last Vart of Sutra I. I. 72, it means c when an affix beginning with
ya or ach follow. 9

Vart 1 — The words 5T*I*J , Hjf^-i^and qyj should be treated as Bha
when the affix w^ follows. Thus ipflef^ like the sky. HjQ{*<*4<l like the
Angiras. q^srw like the man. By being Bha, tfie ^is not changed intor ,
which it would have been, had it been a pada (VIII. 2. 66).

Vart : — The words ^\ is treated as Bha in the vedas when the
words **J and *r* follow. Thus ^V**£ ; ^MUI^W 5# . Here had the word
vpshan been treated as pada, the 7 would not have been changed into IJF
(See Rule VIII. 4. 37) ; and this h would have been dropped before the affix
vasu by VIII. 2. 7.

*§l *Tc*ni U <*% II H^rfk II ?l-%T f T^-^, («\) H •
*jfa: II ff<4*HI«4 4J<fcKI?rf JJ*W* H&& H*«*«J Sfrtt >Whf H*f*f M

19. The word-form ending in t or in s is
called Bha when an affix with the force of matup ('whose
is it,' c or in whom it is ' V. 2. 94) follows.

The word Bha is understood in this sutra. Thus is the word
4lft<4« having butter milk Norn. Sing. *qf*3W^ tfta: the herdsman having
butter milk, ftqf?PT TOOT*: the cloud full of thunder. So also TOPr)* famous <

"• ^ Digitized by vJfV^VJV F

2

If 6 BHA CASES. (Bk. I. Ctt. IV. § 20>2I.

TOSft full of milk. Thus 3*fa*c + **n - S*f*?*q;, mj*L + f%fa (V. 2. 121)
— *Hjft*^ . By making these Bha, the operation of pada rule is
debarred in their case. That is to say the final tf of uda£vit &c., and the *£
of yaias &c., are not changed into f and r respectively, which had they been
pada words would have been the case by the action of Sfitra VIII. 2.' 39,
and VIH. 2. 66. via., gjpMt « and *n{jtf%*r which are incorrect.

20. Words like ayasmaya &c, are valid forms
in the chhandas (veda).

These words being taught here in the topic relating to pada and
Bha, show that they have been properly formed in the chhandas by the ap-
plication of the rules of Bha and pada. Thus *nr*£ + TO =* ^n^HT made of
iron. Here the word aro^is treated as Bha and hence the H is not changed
Into r . Thus 9FEKTO *# iron-made coat of mail. WPSTOrft <rPTffij iron
vessels. The present form of this word is *nft*Ri . In some places both
these pada and Bha apply simultaneously. Thus in the word ij^ formed
by ^C\ + **[* ^ e % * s first changed into wby treating the word rich as a
pada. Then the word gj^ is treated as bha, and therefore the g^is not
changed into *£ before qrtf • For had it been pada, the form would have been
sjpr^ Rigvat. These irregularly formed words occur only in the chhandas
or Vedic literature. Thus ^r *|^!ra 3fto<tl *i*fa •

*§g snprnm n ^i 11 *^ift 11 335, *g-**R\ii

21. In expressing multeity, a Plural case
affix is employed.

When it is intended to denote multeity those affixes should be
employed after nouns and verbs, which denote plural number. Thus KTOTO
H*P< t The Brahmins read.

This rule applies to words which are capable of expressing numbers.
Indeclinables (Avyaya) do not admit of numbers, and consequently they are
always in singular number which is the general form*

£\*&\ fs^hrcr*^ 11 rh n ^ftr 11 fs-Tjwfr: , fjpwi •

Bk. t,Cii. IV. § 22-24.] Numbers Karaka. 177

22. The dual and singular case-affixes are
employed severally in the sense of duality and unity.

This is also clear* When duality is to be expressed, a dual case*
affix should be employed, and in the case of unit, the singular case-affix.
Thus MiWUjV TOT: the two Brahmins cook. ur&ro V^ftr the Brahmin cooks.

^rc$ n ^ 11 tr^Tft *nt$ h

23. The phrase c k&raka' (meaning c in the
special relation to a word expressing an action ') is to be
understood in the following aphorisms.

Any thing that helps towards the accomplishment of an action is a

kAraka. Thus in 'cooking/ the fire, the furnace, the vessel, the cook are mil

helpers in the accomplishment of the acton. Thus ' Rama cooks food in a

vessel, by the (ire from the furnace for his master. 9 Here Rama is agent

kAraka ; 'food 9 is object kAraka; 'vessel' is a locative kAraka, fire is an ins-

trumental kAraka, 'furnace 9 is the ablative kAraka and 'master 9 is the dative

kAraka.

The word 'kiraka 9 thus is synonymous with the word cause (hetu)

and occasion (nimitta). All the various causes and occasions that are re-
quired to complete an action will be kArakas. Out of the seven cases in
which a Sanskrit noun is declined, six represent such relation with an action.
The Genitive or the sixth case can never be directly related with an action,
and thus can never stand in the relation of a kAraka to a verb.

wnronfc jm^Hd it sqj it q^ift n wr, <*im5fc, snmjT*

«[ftr. 11 ig*f sryira^K *mx «i*5 ^rc^rfipg* q*ftH**wuH*J<f >rcf% 11

24. A noun whose relation to an action is
that of a fixed point from which departure takes place
is called ap£dana or ablation.

This defines the Ablation or ApadAna kAraka. Thus UNiqintofl he
comes from the village. <r$4f{4tfcft he descends from the mountain. *q|tll<4)*r:
lost his object. T*n^qfcifT: fallen from chariot. The ApadAna takes the 5th
case -affix (II. 3. 28) and the above examples show this. When therefore
this relation is to be expressed, ' the fixed point (like grAma, parvata &c., in
the above) which is the limit denoted by a word dependent on a verb, is
called ablation. 9 Digitized by Google.

178 ._ The Ablation. _ [ Bk. I. Ch. IV. § 25-27..

Vari: — The objects of verbs denoting ' aversion/ 'cessation 1 and
* negligence* are also called ablation. Thus: — q renfrssnrar ft he dislikes in-
justice; MJWii&iJjft he ceases from injustice. vmfwmftl he neglects justice.

^T: II f4+ta|tji*lt IRt^wf^ff ^ HT*j\ *riH*t ^W?g^ Wf*U*H4HHiM-
*Wf% II

25. In case of words implying c fear * and
'protection from danger' that from which the danger
or fear procedes is called Ap&d&na k&raka.

The verbs signifying 'fear' or ' protection* govern the object feared
of, or the danger to be protected from, in the ablative case. Thus nt^ft
f^j% he is afraid of thieves, , *to*r sitr^ ^e is agitated because of the thieves.
^ft frqHIMfl or TOfif . He protects or saves from the thieves.

Why do we say 'the cause of fear (bhaya-hetu) is put in the ablative
case? Observe *rc°3f fk^fa or <iraff he fears or protects in the forest.

*jRfc II <RH*}k* *Hl3: 34)4)441) 4t *&& .^TJ •tiU<t*|ft *!^H*H4KM^*f>Wf<* Jl

* • 26. In the case of the verb par&ji, 'to he
tired or weary of,' that which becomes unbearable, is
called Ap&dana k&raka.

When the verb fir to conquer, with the preposition pari has the
sense of 'becoming tired or unbearable ' it governs ablative case of the thing
become unbearable. As M^nRHJ muitlft he finds study unbearable.

Why do we say 'that which becomes unbearable. 9 Observe jpnr
TCnraS he defeats the enemies.

$faci: (qno vrar^ro) 11

27. In case of verbs having the sense of pre-
venting, the desired object from which one is prevented
or warded off is called Ablation or Ap&dana k&raka. .

The obstruction to one's n.atural inclination is called v&rana or pre-
vention. As *rewft irf ^nfrfflr or fr*forf?r he wards off or withholds the cow
from the barley. jOC

Bk. I. Ch. IV. § 28-30. ] The Ablation. 179

Why do we say 'the object desired' is put in the ablative case'?
Observe irf ^H7ri% %% he wards off the cow in the field.

tjfa: » W^rf^f^fW 3HK^qiW**fsftfl^K«W4l<H*Ui *|*f% Q

28. When concealment is indicated, the per-
son whose sight one wishes to avoid, is called Ap&d&na
karaka.

Thus OTT^rnrr^f^ ^ or ftoftqJ I he conceals or hides from the teacher;
so that the teacher may not find him out or see him.

Why do we say 'when concealment is indicated/? Observe ^Jkp*
? fi£ VI ^ he does not wish to see the thieves. Here the term chauran is in
the accusative case.

Why has the word ichchhati been used in the text ? The ap&dAna
k&raka is to be used there only, where one desires that he should not be
seen, yet he shows himself.

< MK^kim%Hl n ^ 11 ^ifSi 11 sn-^rnn, ^ra-^i?*,

(mo 3TTT*) II

29. The noun denoting the teacher is called
Ap&d&na or ablation^ in illation to the action signifying
formal teaching.

The word 4ji4$4idl means teacher, and greftl means acquiring know-
ledge in the regular way. Thus TOumrrqpftt or Mnpruft he learns from the
preceptor.

Why do we say 'when meaning to learn'? Observe *GRf *JtDtft he
hears the player.

?lfw<ftTi: Hsbfa: II $p 11 t|^ip| 11 ^rRr-^S:, J*?rft: 9

(«tp wno) 11 .

^r. 11 m~4<U4j<»jm: ^r *rra*r?: i^^r nfrffc *ur*«f ?g: af^H^^m*-

30. The prime cause of the agent of the verb
jan to be born, is called Ap&d&na.

That which is the Agent ( qm ) of the verb jan, is called nfiqprf .
That which is the prime-cause ( UfrfJr ) of the agent (or product) of the root jan

180 The Dative. [ Bk. I. Ch. IV. § 31, 32.

is put in the ablative case. As MJfrcrft *TOfr , the arrow is produced from
horn iTtaqr^^fcqft *rr*Ht the scorpion is produced from cowdung.

£*: WW: II ^1 U M^lPt it gw ; mre: , («19 «pn») H

^frf: II *J?E^: wft *f WWH**WI<jmM Tift II

31. The source of the agent of the verb bhu,
to become, is called Ap&d&na.

The phrase ' of the agent (kartuh,) is understood here. The word
iMW means the source or that from which anything arises. As Q*i4<ft ifar
ipiqft the Ganges has its source in the Himalayas. *ttivft?**ft ft^WIT JPfafir.
The Hydaspes has its source in Cashmere. .

^ftf: II W&GJt, *RVZP*jfcr 9T*lf ^Rnft *Jf*K*li ffrf^H^f *faf% II

qiRkh^ 11 fa^moj «ift wrfcn^ 11

vf*^ 11 **irc «M.ujil«n fm*m <Jq*H<^ n *tfaar 11

32. The person whom one wishes to connect
with the object of giving, is called Samprad&na or re-
cepient.

Though the word qpfen in the sutra is indefinitely used, meaning
• with the object' yet it is not every object of any verb. The object must be
of the verb ' to give.*

Ai'iWWW *tf W&t ^ e gives the cow to the teacher. «Uiy44il4 ftw^
OTftt he gives alms to the boy. Here the words * cow 9 and ' alms' are the
object of the verb 'give'; the persons connected with this object are the
teacher and the boy respectively. These latter are in the Dative case and
take the 4th case-affix,

Varti — The person whom one wishes to connect with the action

should also be called recipient. As VT5TO f%*I$it he censures for the sake

of Sraddha. J^m 9TOJ3 he prepares for battle. <nd $ft she sleeps for her

husband. _

Varti — 2. After some verbs (especially to sacrifice), the object

(karma) gets the name of Instrument (karana) and the recipient (Sara-
prad&na) is called object (karma).

Thus qjpr *t frifr or *£ tot* Wf^ he sacrifices with an animal to
Rudra, which is equivalent to, ' he gives an animal to Rudra.'

Bk. L Ch. IV. § 33-35 ] The Dative. 181

33. In case of verbs having the signification,
of the root ruch 'to like,* the person or thing that is
pleased or satisfied, is called Samprad&na or recepient.

The verbs having similar meaning (artha) as the word ruchi or
1 liking* are »*ml . A desire or longing caused by something else, is called
ruchi. As foj-diq *ta3 Ht^S: the sweet meat pleases Devadatta. ^np*TO
*^3 4TO Yajnadatta likes Apupa. Here modaka is the agent that draws
out the longing which is latent in Devadatta.

Why do we say 'the person pleased? Observe $44*114 *ta*fr qtfV
qfir . Devadatta likes modak in the way. The word pathi being in the
7U1 case.

^k ii 5^nr ' ! 5 3 <^t jj* v53<rf ^ftM-sHi^i *SM*J w^ *h\x<£ *lw-nW *fln% it

34. In case of verbs 61stgh to praise, hnu to
take away, sth& to stand, and 6ap to curse, the person
whom it is intended to inform of or persuade by, these
actions, is called Samprad4na.

The word <fttQCTPT means whom it is desired to make known or in-
form. As <N4tU3 VrilM^ he praises Devadatta, m., while praising Devadatta,
he wishes that he should know this praise. So also with other verbs. As
?^4*tH %9 he hides from (wishing that Devadatta should know of it) Deva-
datta. 3q$Tuq fr\u3 she offers herself to Devadatta for (sexual embrace,
wishing that he should know of it). ?44tU4 jjq9 he reviles Devadatta.

Why do we say 'the person whom it is intended to inform'? Observe
93TTOT Vrimft *f*i • Here pathi is in the Locative case.

^Rn ii wspfc jrftff sp*pp5i *Jt J5r fw*iirf ^twrtfif *wfa u

35. In the case of the verb dh&ri c to owe/
the creditor is called Saniprad&na.

The word swrf is compounded of two words ^nr best and sjtjj debt
meaning (whose debt is best) creditor, as opposed to «POT& debtor. As
V^T^nr ^pt WT^rf^ he owes hundred to Devadatta.

Why do we say ' the creditor is called Recepient? Observe ^^TPf
lltfmiqftl VT% he owes hundred to Devadatta in the village. Here village is
in the Locative case. . Digitized by (

182 - The Dative. [ Bk. L Ch. IV. § 36-38.

_ • . £

(mo qn^r*) n

36. In the case of the verb sprih to desire,
the thing desired is called Sarnprad&na karaka.

The verb wp to desire, belongs to the churAdi class. .The word
governed by this verb takes the Dative case* As j*<Mr: ^nprft' he desires
flowers, «K^*: *jpifit he desires fruits.

Why do we say ■ the thing desired/ Observe yfcft *% *^lfa he

desires flowers in the forest Here rf is in the Locative case.

37. In the case of the verbs having the sense
of krudh to be angry* druh to injure, Irshya to envy,
asfty& to detract; the person against whom the feeling of
anger &c, is directed is called Samprad&na.

Krudh means non-tolerance, droha means doing wrong or hurt,
irshya jealousy ; and asfiya means to find out the faults of another. The
word kopa includes all the above four sorts of actions. As qq<j-dm ^i^ft*°
55Tf?T-i^5f% or «r*n?ft he is angry upon, (i>. with) or bears malice to, or is
jealous of, or finds out the faults of Devadatta. Here Devadatta is in
the Dative case.

Why do we say 4 against whom the feeling of anger is directed.'
Because if there be no feeling of anger, then these verbs will not govern the
dative case, but the accusative. Thus Hnrfiftcjftt he is jealous of his wife,
*>., does not wish her to be seen by others. Here the word bhirvA is in the
accusative case. f

H*i§ \t ; mi (mo *i nfe^nj:) U

^f: II Ig^TJjt ^TO 5 ?^ twfcfrMiql 5 llft^HflWiHM! ^PT €Tlf H*f% II

38. But in the case of the verbs krudh and
druh, when preceded by prepositions, the person against
whom the feeling of anger &c.y is directed is called karma
karaka or object, N . - Dgtze

Bk. I. Ch. IV. § 39, 40. ] Sampradana Karaka. 183

This aphorism is a proviso to the last, and enjoins accusative case,
where by the last, there would have been dative* The word upasrishta
means having upasarga or preposition. As frqqvmftyqfa qjfo^uft . When
used simply, those verbs of course govern the Dative case by the last
aphorism. As $44^4 flv«lPlr .

fkv&: , far© ^w^to) n

^r 5f^i >r*rf% *re*i spirept 5^m 11

39. In the case of the verbs r&dh, to propi-
tiate^ and iksh to look to, the person about whose good
or bad fortune questions are asked is called Samprad&na.

The word vipraSna means literally asking various questions; and
denotes questioning about the good or bad luck of another. Thus foiTm
nwrf^-t^T^ *T*fc Garga is favorable to or looks to Devadatta, the sense is that
being casually asked by Devadatta, he reflects upon the good or bad fortune
of Devadatta.

Another explanation of this sutra is, the agent that puts various ques-
tions is put in the dative case. As ftronr TOftftr or f ^T$ qT *[F: meaning, the
pupil asks the teacher various questions to resolve his doubts, and the teacher
answers those questions to the pupil.

Hcmspari sw treh* <Eftf 11 so 11 trcTft 11 nfa-snsronw,

^ Nft ^^ ■ \ ^

^f%: 11 *rfcr»n^juN *JJR*t , JioHfc 4ift<* ^unrest *rcfir <&i\i *jfr*t ^K^rf 11

40. In the case of the verb Sru preceded by
the prepositions prati and &ii ; and meaning ( to promise/
the person to whom promise is made (lit : the person w)io
was the agent of the former verb) is called Samprad&na.

The compound verb Jtffaf and sjpi^means to promise. A promise is
made on the motion or at the instance of another. The person so proposing
who was the agent of the former action becomes the recipient of the promise
in the latter case. As fcf^TRr irf *tfkq«irf$— Hrajajrffir he promises a cow to
Devadatta. ....

3?3wf g q*Tq 11 n n v^ti* n srg-Hfa-vi*, *, (*w
**rc[w tHfo vsri) g

H * ft " Digitized by G00gle

.184 Instrument Karaka. [ Bk. I. Ch. IV. § 4*-43-

41. In the case of the verb gri> preceded by
ami and prati, and meaning c to encourage by repeating/
the person who was the agent of the piior action, which
is repeated, is called Samprad&na.

The phrase pfirvasya karta of the last is to be read into this aphorism.
Asflifr jy uu iffr . They encourage the Hotri, i.e. 9 the Hotri priest invokes
first, the others then follow him in invocation and by so doing encourage him.
The word tqiff: and srftflTC: mean encouraging the invoker.

qi v i<feg H 5>r y w ii *k ii tr^Rf ii ^w^w^, *<<n*^
(«n%) ii '

42. That which is especially auxiliary in the
accomplishment of the action is called the Instrument
or karaka k&raka. ^ _^

As fr%UT fprffcThe cuts ^vith the sickle, qf^rf f§PfRf he divides by the
axe. The instrument 4fcU<fc takes the third-case affix.

Why do we say € especially? Because in the case of the other
karakas, the non-mention of the word 'especially 9 makes it possible for u&
to use those cases, in not their strict sense. Thus though the locative case
should be used where strictly a thing is located in another, yet we may say
i|<fl.|*lf qfr: the herdsman in the Ganges, not strictly in, but on the banks of,
the Ganges. gr^*j*J*c5Jjthe family of frogs in the well, t>. on the sides of
the well

^Rf: II ffa: ^PJ^T?r4 «*tkhl<* tfeh'&U? H^fif *l*hUI^ 4fc<UJ4l<f ^T II

43. That which is especially auxiliary in the
accomplishment of the action, of the verb div to play, is
called karma object, as well as karaka, Instrument.

The present sfitra ordains accusative case, where by the operation
of the last aphorism there ought to have been Instrumental case. The force
of ^r in the sfitra is to indicate that karana is also to be read into this. As
WT5£ ft&rfH or Sflff tfo^ft he plays the dice or with the dice.

Bk. I. Ch IV. § 44-47. ] Adhikarana Karaka. 185

44. In the case of hiring on wages, that
which is especially auxiliary in the accomplishment of
the action of the verb paiikri, c employing on stipulated
wages, 1 is optionally called Samprad&na or recipient.

This ordains dative case, where there would otherwise have been
Instrumental case. The word qRthq fr means to engage for a limited period
on payment of wages, and not absolute purchase for all time. As ^rara

Adhikarana.
STTqiftrf ^^W II 8H II ^rf^T H 3*T*H: f Slfvi-
WW, (too) It

«[f%: II ^i^f^Tf: f^lWT'gl^tafoj flfarf lrf% T HJWH*<l«*lt<* l if?l<**<UJ

45. That which is related to the action as the
site where the action is performed by reason of the agent
or the object being in that place is called Adhikarana
or the Location.

That in which the action is supported or located is called 4dhAra~
As 9$ we9 he is seated on the mat ^ Hi he is sleeping on the mat. ^qr^irt
TOfrt he cooks in the pot The Adhikarana takes the 7th case-affix.

srrom , *&* , faro arret*:)

46. That which is the site of the verbs si to
lie down, stha to stand, &s to sit, when preceded by the
preposition adhi, is however called karma karaka or
object

This ordains accusative case, where otherwise by the last sfitra there
would have been the Locative case. As iTWTftfW *jfaf?tefa or *!U[rc# he
lies down, occupies or lies in the village.

*, faw stpstc: ^*) 11

^tt II *tf*tf%*J?J^I ftHJdilWHlT *H:<i**K«h ^Wf *Wfa II

47. Thatwliich is the site of the verb abhi-
nivii to enter, is also called karina-k&raka. Di

186 * _ Karma Karaka. [ Bk. I. Ch. IV. § 48, 49.

As «TPFrPtf^f%7Jtf he resorts to the village. This is an optional rule,
as the word " option " of sutra 44 should be read into it. Thus we have the
following forms also <n^jf*rf%t$p resorting to sin. *K^r5jPif%#sn resorting to
good. The rule here is that of vyavasthita vibhAsa.

(VBIO S?P*T*: **) II

48. Tliat which is the site of the verb vas to
dwell, when preceded by npa, anu, adhi, and ail, is called
karma-karalca.

As iiiMJiq^ft On? the army dwells in the village q^gryrTO^r -M^rofr
Hjftquft or MlTOft.

Vart : — Prohibition must be stated when the verb ^ means " fast-
ing/' or does not denote lying in a locality. As gr$ yr^r ft he fasts in
the village. Here the verb upavasati governs the locative case and not
the accusative.

Karma.

S^H^^RW 3>3 II 8<5 H M^lPl II ^tt 9 ^ftoiRPPT,
*&* («W) II

49. That which it is intended should be most
affected by the act of the agent is called the object
or karma.

That which is especially desired by the agent to be accomplished
by the action is called karma. As sr£ <M)ft He makes the mat. qtf *r*&fa
he goes to the village. Why do we say " desired by the agent " ? Observe
*ff$*? xi TOTft he ties the horse in the gram field. Here gram is no doubt
most desired by the horse, but as horse is not the agent of the verb, the
word imj takes the locative case. Why do we use the word " most " ?
Observe <Rj*fosf *J3^ ne eats tne f° 0( * a ' on g with the milk. Here milk is no
doubt desired by the agent, but not being the principal object desired, takes
the Instrumental case.

Though the word stf was understood in this sutra by anuvritti from
the last sfitra, the repetition of this word here is to indicate that the anu-
vritti of the word AdhAra does not extend to this sfitra, because as we do
not take the anuvritti of the word karma into this sfitra, we do not take the

Bk. I. Ch. IV. § 50, 51. ] Karma Karaka. 187

anuvritti of any word of the previous sutra into this. Had we taken the
anavritti of the word karma from the last sfitra, then we could use the accu-
sative case in those limited instances, where the word is capable of taking
the Locative case also, but not in other cases* Thus while we could very
well say *it ^ fi ^ ifr T he enters the house, we could not say *Hrpf T*fH he cooks
the food, ^rj fttfftf he drinks saktu. By repeating the word karma in this
sutra, such examples become valid everywhere. The karma-karak, takes
the second case-affix.

sw grH toHSto* 11 \o 11 mrfk n *m 9 ipsa, * f

W«nfal*W f (^5T0 S>*) II

*f%r: 11 3* zznfcj vRrf tffotnrf fk^m q*«i3 Hh* ^ *Hii?uj ^{•frfam
3% mfH ^tot cr^tkt ft^Jhid u

50. If that which is not intended to be most
affected by the act becomes however similiarly con-
nected with the action it also is called karma.

That which is not desired by the agent is anipsita or object of
aversion. Thus f%tf w^raft he eats poison, ijhr^ **VXf* He sees the thieves.
aTH^ip^^r 1 ^ «jTO*rfcr going to the village, he plucks the roots of the
trees.

sreftm * 11 \% » ^^fc " w^rf^rcPL, *, (*» *5) »

«[f%: 11 *rafZfcf ^ ii^kk«ii H«%Wjff *wflir II

51. And that karaka which is not spoken

of as coming under any of the special relations of abla- ^
tion &c, is also called karma. ^J^£ju^

There are some verbs in Sanskrit which take what .is called an ^*"
akathita object, in addition to their usual direct one. As its name indicates, ^ /;,
it is that object which is not otherwise kathita or mentioned by way of any
of the other case relations, such as ftfTOFT HjfacM"! &c. and is, therefore
optional. If the noun capable of taking this akathita object be not intended
for any other case, it is put in the Accusative case with such verbs ; as, $£
$tfnr q?r: he milks the cow (her milk; 9PPTOCT& *tf ' he confines the cow to
the fold. 9 Here §5 and jpt are akathita or optional objects. If the speaker
does not intend to have this object, the words will be put in their natural
cases ; as, &?f: (ablative) T*fl$rfnr, ir# (locative) aPTt&ft «rf.

The roots that are capable of governing two accusatives are mentioned
in the following kirikA :— JSJPJ <r* **^*fa *f*5 fa ^ JTHJ fif *IW <*&{ II Sffap^

188 - Karma Karaka. [Bk. I. Ch. IV. § 52.

In the case of the roots j? N ' to milk ' qr\ f to beg/ T^ ' to cook/
^SH[ 'to punish/ to' to obstruct or confine 9 H*^ ' to ask/ fa 'to collect*
Wto tell jjt^ 'to instruct* f5r 'to win' (as a prize of wager) f^
9 to churn ' 5^ f to steal/ and also in the case of «ft, 5, ^, and W N all mean-
ing ' to take or carry ' and others having the same signification, that noun
which, besides the direct object, is affected by the verb, is put in the Accusa-
tive case ; as, irf $tfi* W (S. K.) ' he milks the cow ' ; qfii w*& *f*tf .

' He begs the earth of Bali ' ; similarly tf^gjpftyf W<t, *Plfc^ *Rt TO**lfa,

*r are examples of the other roots in order. «1IUI4^ *T$ HPfft sfa? *r, qrfS-
*^*t ftsnt, flt art dWWUt f TOlft firar*3lr are instances of this kind of
object, because *n*T or qt\ and f^ or ^ have the same meaning as f^ and
iTPT, the roots given in the kasiki.

Obs.— The roots Pr, J^, T^ «f^, ^^., fsT, ^ 5 and everq^ are of
every rare occurrence as governing two accusatives, in classical literature,
though given in the above list

The roots mentioned above and others having the same sense,
take two objects. One of them is principal, and the other, secondary. In
the case of the first twelve roots from 5^ to y{, the nouns <nr ' t «Rptf 9 *h*\lfk,
*nrf, &c. are principal objects, and irf, sfift, ^f, vfhftfSf &c, are secondary
objects, for they can, according to the speaker's volition, be put in other
cases. And in the case of the last four roots SPTT, is the principal object
and qnt the secondary. Thus that which is necessarily put in the accusa-
tive case in order to complete the idea of the verb, is the principal object,
and that which may be put in the Accusative case, depending upon the
speaker's will, is called the secondary object.

rppfk II nm-^-IlWWTO-^-S^^ 9*fWrrf, *:,

JjfrT: II l^whrf $*W*lW RHR^Hnjprf n ^TRglt WTOS? ^chim H«h4<4ilUfl

3?9 VRnrf *n '^raf h ^ii^dHl **ta*iY lift M

^Tf^^ II ^mnrj ft **it: uforo* ^rarar: 11
*ifi!*HJi t?t ft^i*^ «n<£ cut-in ^m***^ it
qif&h«i II *jrff mtit: Jrfswt Tifi^f: 11

^nf^«R^ 11 *^Jf fifanfc^r iiftWlr q^vni: 11

52. Of the verbs having the sense of ^mo-
tion ' ( knowledge or information * and l eating,' and of

Bx. I. Ch. IV. § 52. ] . Karma Karaka. 189

verbs that have some literary work for their object, and
of intransitive verbs, that which was the agent of the
verb in its primitive (non-^i or non-causal state), is
called the object (karma) in its causative state (when the
verb takes the affix (13a).

In the case of roots that imply ' motion/ ' knowledge ' or 'informa-
tion ' or some kind of f eating/ and other roots having a similar sense ; also
of roots that have some literary work for their object, and of intransitive
roots, that which is the subject of the verb in its primitive sense is put in the
Accusative case in the causal, the object remaining unchanged ; e. g.

Primitive. " Causal.

suft*if*ui ifrer ft#f ^f$& *m*mn^

But in inraft UHt *Jtf*Ft (Rama makes Govind go) if some body else
(f%*3 Pft) prompts R&ma to do this, we shall have to say fforfoft trfnf
*ltf*** UH^fft ' Vishnu mitra prompts R£ma to cause Govind to go/ Here
1 Rima' is not put in the Accusative case, because it is the subject of the
verb, not in its primitive, but causal, sense.

Patanjali, in his Mah&bhishya, adds this explanation on the meaning
of the word 1&&& in the sutra qfapt &c. *re*ri may be either ipsft ^f
firar or u*$t *fafr *rf*

When we take the former interpretation, the roots f?n% ( £ ) iiftft
(ifa) and a pr^rerfl (denom. of $rs* ) have to be excluded from the rule ; as, K*f%
w^f: iXQvfif fa^fcr t* itfff<ir-$n^nc§-for*fs ; iK^ifir-^^nnn%-5if^f . And the

roots *| , ot with fir and w^ with ot must be included in the rule; as, UHjffif
f% <iHlft -g*req3%^: WWft-ft«ll , wRl-4 , l*l | *^ft-?*f^n| ( . When we adopt the
second interpretation, the roots sif^ , HPJ with sjf and 5W with ft , must be
included in the rule; 3f^7fe-ft<»«filM*JWIMfl-?*TTO 1 M*S^ft-fi^W*lfi-^l*im^Rl-

There are several exceptions and counter-exceptions to the preceding
rule, which are important

Vart : — The causals of *ft ' to lead' and *f % f to carry/ do not govern
the Accusative, but the Instrumental ; e. g. *jnifr *nt infa tfjft *r A servant
carries a load. **fcr *mf ITOift TOtrflr *r (S. K.) (He) causes a servant to
carry a load.

190 _ Karma Karaka. [ Bk. I. Ch, IV. § 53.

Varti — But ** H $ when it has for its subject in the causal a word
signifying a ' driver/ obeys the general rule; as, qur K*i «T?f^f . Horses
draw the chariot. qtff% Vffl\ TOtatf: •

Varti — (b). The causals of the roots ^ and igf\ f 'to eat/ govern

\ the Instrumental case; e.g. H^tUtRt ^HffitqT . The boy eats his food.

C*\ mH\+m\mTh.r*ttTQfa *t . (He) causes the boy to eat his food.

Hj m m Jc Varti — (c). H*^,when it has not the sense of fifar ' injury to a

sentient thing/ governs the Instrumental ; as, *rcnrf% W^f fop*:, W*rf% fihff

*^frr ; but «Kraft sren^^ft^f : » wrf<lr 'raft ^f *j irf^ .

By ' intransitive 9 roots mentioned above is meant such roots as are
not by their nature capable of governing an object other than that of
' time,' ' place 9 &c., and not those roots which, though transitive, may some-
times be used*intransitively according to the speaker's volition, or when their
meaning is, quite, evident ; as, fc&h<: q^r: ¥51% . Here q^f?r , though transi-
tive, is used without an object, because it can be easily understood ; hence
fifaftqr qnraft and not fifcac ; but n i iwmfa frcyif .

In forming the passive construction of casual verbs, the princi-
pal object in the causal, which is the subject (agent) of the verb in its primi-
tive sense, is put in the Nominative case, and the other object remains
unchanged ; e. g. ~

Primitive. Causal Active. Causal Passive.

Kama goes to a village. (He) causes Rama to Rama is caused to go &c.

go to a village.

*p*n *rf ttitlft . ^wta *|**f ^t tt *iwRlr. tot mi 4tii£d .

The servant prepares a mat. (He) causes the servant The servant is made to

to prepare a mat. prepare &c.

Govind sits for one month. (He makes Govind sit &c. Govind is mode to sit &c.

(a). But in the case of roots that imply f knowledge 9 * eating, 9 and
those that have a literary work for their object, the principal object is put
in the Nominative case, and the secondary in the Accusative, or vice versa
e. g. 5m* V^^nrf^ 'he makes Mananaka know his duty 9 ; *mu*3iti|4
ifcztf or *mim» \pft qfc?3 ' M. is made known his duty 9 or ' duty is made
known to M.' ; ^H>3«i ^nrf?t ' he makes the boy eat food: 9 7|ft?? Ht*Rl5 or
^JHt^t Hts^f (S. K.) .

With regards to roots that govern two accusatives, the rules
mentioned above hold good in their case also ; ,' i.e., those roots that
imply motion <£c, govern the Accusative of the subject of the primitive
Case, and others, the Instrumental case, sometimes; as, qpritqfvi Tfprt-

Bk. I. Ch. IV. § 53, 54. ] Agent defined. • " 191

nrtfl: ( fn?*t ) im** *fwt ^rf^mft 4 (God) makes Vamana ask Bali for
Earth/ ntfTjsrf *ptc *rfr; ( *wft ) «tW iftfcr *nrt *mf (ITOft . '(The master)
makes the cowherd take the sheep to the town.'

% Q<**4<H ^n^ n *$ 11 W^ n ^ste f ^TOUM,
(5>w eif^r *><<tf * *ft ?&) h .

^frf: 11 jx& ^rfWw^^nirt: spa? ^r °ii«iiilf twmwil irftfsfr *wfa 11

53. The agent of the verb in its non-^i (pri-
mitive) form in the case of hji to lose and kji to make,
is optionally called karma or object when these verbs
take the affix.

nf (Causal)..
Primitive. Causal.

As ftf% nit tostto: tiroft *?rr <mmtf or uncroftr .

The boy takes the load. He causes the: servant to take the load.

Devadatta makes the mat. He causes Devadatta to make the mat.

Vart: — The subject of the primitive, verbs *rf3ft^ and ro s when used
in the Atmanepada, is either put in the Accusative or Instrumental casein the
causal: as,

Primitive. Causal.

Devadatta bows down to the Guru. He makes Devadatta bow down to the

Guru. • y. £jL^->*~tJ^'r

The servants see the king. He makes the servants see the king.

KARTRI.

w?f*: srat II t$ II q^rfSf *ro?*: *rcrt I (*w) N -

^[Rf: 11 firarr sr^np w«i**ihu fro^qd *m*i<*' Hi&itf *rofa 11

54. Whatever the speaker chooses as the in-
dependent, principal and absolute source of action is
called kartft or agent.

The agent is absolute and unconditioned, as, |*f*T: q^lftr • Devadatta.'
cooks, **mfr T^rfa the pot cooks. t

Digitized by LiOOQ LC

f

19J - NlPATA DEFINED. [ BK. I. CH. IV. § 55-57.

55. That which is the mover thereof, L e., of
the independent source of action, is called Hetu or
cause, as well as kartA or agent.

Thus SKTCqfij he causes to be made, ?nroft he cause to be taken.

When hetu is employed as agent the verb is put in the causative form (III. 1. 26).

The force of the word ^r is to give both names to the mover of an

agent, v/*., Hetu and kartA: otherwise by sutra 1 of this Chapter only one name

would have been given.

Nipita.
The Particles.

56. From this point forward upto the apho-
rism Adhirt-6vare (1.4.97), all that we shall say is to be
understood to have the name of Mp&ta or Particles.

The word sjraj of this sutra serves the same purpose as the word cha
of the last ; namely, it makes the words Gati, upasarga and karmapravachntya
take two names, t\ e., their one name as well as the name Nip&ta* The letter
C in mifNiiqJ s to remove doubt, i. e. t aphorism I. 4. 97 should be taken as
limit, and not sutra III. 4. 13 which has the word iswari also.

*T^t to?3 11 h> 11 mpfa 11 *-sn^j: sre?3, (Etotot) n
^f^i: 11 *n^ir f^nrcUir >wfwi t ^^r^ q&9 11

57. The word cha c and,' &c, are called Nip&ta
or Particles, when they do not signify substances.

The following are particles (nipita) ^f'and' m 'or' f 'an expletive'
fjf 'vocative particle' <j* 'only * 'exactly.' <r*n' so, thus,' snp£ 'certainly/
tf**^ •continually/ vnr^ 'at once/ *jjn^ 'repeatedly/ ^rr^ ^q^ 'excel-
lently/ vf%n ' abundantly/ 4^^'if ^j < if/ [the n is indicatory], ^r
'where/ fpc 'there/ *&*** 'what if?/ «* 'no/fe* « ah/! nrfan [*rt»n)
«rftpr ' do not' M4l«*tl^ ' indeed !' m^ ' do not,' ^ ' not/ *r*^ ' as much as,'
*T*^'so much/ tf ?* I 'perhaps/ * (disrespectful interjection) m^4W

Bk. I. Ch. IV. § 58. ] Prepositions defined. 198

*3T?r (intcrjO ' oblation to the gods/ fT^T ' oblation to the manes/ forefathers
q^' oblation to the gods/ *fr^ (mystical ejaculation typical of the three great
deities of the Hindu mythology), inj € thouing/ tnnft * thus/ introducing an ex-
position, (?5P* • certainly/ fa*? ' indeed/ *m f now ' auspicious inceptive), *[$
' excellent/ w (attached to the present tense gives it a past signification),
U& 'fie' trf??? 5? <C 3 *ft *fK The vowels a, i, i, 1, u, ft, e, ai, o, au, when, as
interjections, they indicate various emotions, differ from the ordinary vowels.

§t> (ft), 'n*. c^t ^rr|t, srarft, sfr, «rft, ift; Off), *rit, 115, «Fd, ftvsr, «A, fj^# 5,
2» tf^* TO *t» *fl& ^*» Wi [ ^Pfc to*Ji ftrar^, f)«ii<( J ^5^» s«nr» (tft* ) •

*f*i ^pR^» fff^t ^RH^ ^5^» Tft^fc STTOfc ?iPP[f *F3T, I3Ti nWty *^l» ^t»

srnj, 5k*tjt, apr:, 9^, ^k, s^, fr, ?, ( $ ), «rrftRf^» JCPfc ^*. ^. fS^sf* ^. *?.
^f » ( 15^)1 vii^N^, **3Fi qs*» snr^. ( vw ), *r&, *^, *i^» (^i|)» ^i ^Pfc ^»
•**, ?^, f-fhj, fti^, ft:, ^ .

To the list of indeclinables belong also what have, without the reality,
the appearance of an upasarga (No. 59), of a word with one of the termina-
tions of case or persons and of the vowels. In the example MT*M*£ avdattam
' given away/ the *{% is not really an upasarga, for if it were, the word (by VII.
4. 47) would be »mPJ avattam. In the example ahanyuh *rtj: € egotistic/
the ahan is not identical with the aham • 1/ terminating in a case affix — be-
cause a pronoun, really regarded as being in the nominative case, could not
be the first member, in such a compound. In the example, atfercfrrr asti-
kshiri, a cow or the like • in which there is milk/ the asti*jft* must be regard-
ed as differing from the word asti, • is,* which ends with the affix of the
third person singular, otherwise it could not have appeared as the first
member in a compound.

xn^q: 11 \c 11 n^ffn n n-sn^r*, (famdi ere?*) *

58. The words pra &c., are called JSTip&ta when
not signifying substances.

The following is the list of *npn or * prepositions:' * , TO , Hf , *%, *IJ.

«w 1 ft» » 5^ t ft . *ros£ » fir , *rftr , nfir , «*flr , 9 » ** • *rf* » ***» *»**»**"

The making of this a distinct aphorism, when it could well have been
included in the last, is for the sake of giving the Pra &c* words two names,
namely, those of NipAtas, as well as upasargas under certain circumstances,
t. e. t when they are in composition with verbs.* Not so however the cbi Ac,
words. They never get the designation of upasargas.

When these words signifying substances they are not Nipitas. As TO
*F(f?r fhir the excellent army conquers. Here the word TO is not a Nipita.

Digitized by VjOOQiC

104 Upasarga. and Gati. [ Bk. I. Ch. IV. § 59, 60.

^f*c 11 sir^r: ftrsr ^A 3*nh1 3ht mfas 11

*lf*?<h*(JI f$*&^W *ft*f*I*43U*f ^Mfcu^ll

59. The words pra &c, get the designation
of upasarga or prepositions, when in composition with
a verb.

. As q + «Rlft « JflSrafir • Here the dental *r is changed into cerebral
qj because of the JT getting the designation of upasarga (See sfttra VIII. 4. 14).
Similarly qfrtOTft, jtotrFR:, qftuffWH .

Why do we say when in composition with a verb? When in compo-
sition with a noun they are not called upasargas. As ipj^t inTOMOT^Stfni
— • *Hmehl%U : a country destitute of a leader. Here *t is not changed into «r.
Thus while sroqcff: means ' a leader 'the term TOPraft means destitute of a
leader; though both have the same radical elements: — pra in one is an upa-
sarga, in the other a NipAta pure and simple.

Vartz — The word ij^^ should be included in the list of upasargas.
As *r*fie$^r: - *$*£ given by Marut. *TF^ + *T + ^ « KS\ + f£ + n
(VII. 4. 47). Here Marut being treated as an upasarga, though it does not
end with a vowel, the $r ls replaced by n by rule VII. 4. 47 which declares
1 n the substitute of $r which is called ghu, when it is preceded by an upa-
sarga that ends in a vowel, and is followed by an affix beginning with n
which has an indicatory k» 9

It might be objected, that as *$!( does not end with a vowel, sutra
VII. 4. 47 does not apply. To this we say that otherwise the giving the
designation of upasarga to H^ is superfluous; and in order that this should not
be so 9 the fact of its not ending with a vowel is overlooked.

Varti — The particle m^ should be included in the list of upasargas.
Thus m*J + W + *J3£ - **5T (HI. 3. 106). Here because v^is treated as an
upasarga, that sfitra HI. 3. 106 is made applicable.

*tfhn 11 \o n o^TfSr w nfat, * f (ni^r. fom en t ) it

60. The words pra &c, are called also Gati,
when in composition with a verb. Google

Bk. I. Ch. IV. § 61.] Gati. 195

As JT|TO» Now the gerund of the simple root f> would have been
^FTT, but as it is compounded with a gati word (II. 2. 18J the C4T is
replaced by *** see S. II. 2. 18 and VII. 1. 37. H + y + f^ - *&**' O
(VI. 1. 71). Here pra being called gati we have compounding by rule
II. 2. 18; and then we apply S. VII. 1. 37. Similarly it^ra^.. Here pra
being a gati word retains its own accent by rule VI. 2. 49. ' a gati retains in
a compound its own accent when it immediately precedes a second member
that ends in kta 9 provided the latter denotes the object of the action which
is expressed by the root of which kta is added.'

Similarly sjcfctlft , here pra being treated as gati gets annuditta accent
by rule VIII. a^f. /. 7/ C*~A *''*47)

The yoga-vibh&ga or the separation of one aphorism into two, is for
the sake of the subsequent aphorisms. 'The annuvptti of gati only runs
through the latter sfitras and not of upasarga. So that while pra &c., have
two names upasarga and gati ; uri &c., have only one name, namely, gati.

Thus in JT>ffapj and *rf*rfikK^ , by treating the words * and Mfa as
upasargas we change the 7 and 9 into QJ and * by rule VIII. 4. 14 and
VIII. 3. 87 : and again treating them as gati we regulate the accent

Varti— The words SRTftqfn should be included in the list of Gati. As
(1) qhlRchl^m ( 2 ) frlR^lfr^ fc) H^lR*r"^rtf?T . In the first by taking it
as gati we have samasa (II. 2. 18) and f^ (VII. 1. 37.) The other two cases
illustrate accent.

Vart: — The words 5^ and *&Q are treated as Gati in the Vedas. As
yreqqqnfarq fr 5*Pl . Here the word «p^ being gati, causes &nt to take annu-
dAtta accent (VIII. 1. 70) ^itft?!: . Here also the accent is regulated by
VIII. 2. 71.

^T: II 3T*lfof: JJ*ftt *J**J*3I >II*W1IW ft>4(4)i) ffirtHF *CTf5*T II

61. The words uri, assent, &c, and those
that end with chvi (V. 4. 50), and those that end with -
cj&ch (V. 4. 57), (when in composition with the verh bh&
kji or as) are called Gati.

The affixes chvi and ^4ch are ordained when the verbs in composi-
tion is either 7 f ^or qq^ (V. 4. 50 and 57) firi &c., being read along
with chvi and d&ch ; shows that the verb in composition with them must also
be any one of the above three verbs, and none else ; in order to entitle M>tt
to the name of gati.

Digitized by LiOOQ LC

196 Gati. [ Bk. I. Ch. IV. § 6a, 63.

The words ??ft and qx(t mean to assent and spread As irfl^W
(II. 2. 18 and VII. 1. 37) *fl$«H> (VI. 2. 49), gffififtft (VIII. j. 71). So
with the word wft . The following is a list of these words: — \

snfc

^*-

*fe[lPK5ff.

ITifr.

*n*Ih

XFIiW.

^rer*-

«JJ3R*5I.

sir:

^Nifel

ufararcjir.

^.

The words ending in chvi are also gati. As ^tfmj having made
white what was not white.

So also words ending in *V\ as <FTT?ffRq having made the sound
pat.

62. A word imitative of sounds is also called
Gati, when it is not followed by the word itL

The phrase <ijfaftm^ is a Bahuvrfhi compound ; H^-fft-qr^ *• ' that
which has not the word ffa after it 9

As <U^£kH| having made the sound kh4t. iai^<|\ (VI- 3. 49),

^rr^tfir (Viii. 2. 71).

Why do we say ' when it has not the word ffa after it?* Observe

^j^-^to^ 9 ( i*m^i5t ufa* ) n

^fa; 11 *njtT hh<41 &ii**i ^wkj*w *iftuw *fw 11

63. The words sat and asat when in com-
position with a verb are called gati, when used in
the sense of c respect or love,' and * disrespect or in-
difference. 1

As ««<£<* having honored (II. 2. 18 and VII. 1. 37) HCTJ^r?* , QtQRI*
or ^TOUj^CVI. 2. 49) *^*IW sftftft or ^^lETtf* (VIII. 2. 71).

\ Digitized by LiOOQ LC

Bk. I. Ch. IV. § 64-66. ] Gati. 197

Why do wc say when meaning ' respect or disrespect ' ? Observe

(feo ufar:) n

64. The word alam when in composition
with a verb is called gati, when used in the sense of
1 ornament.*

The word m^ is an indeclinable and has four meanings : — 'prohibi-
tion, competent, enough and ornament/ The term »T5^ is gati when it means
ornament As * H<i^m (II. 2. 18, VII. 1. 37); tfti^rt (VI. 2. 49) qqt^flfa (VIII.
2. 71) when it does not mean ornament we have t|c$*qvE*T JPfftr he goes hav-
ing eaten enough.

<H**K*ifTO% n ^ 11 <4<\i(h 11 snEP^sprft**^ f
(few) ift: 11

*if&***t u m*^: jjs*sm f^fkfa ^t^TOife^inr^iwn 11

65. The word antar is called gati, when used
in the sense oi * non-accepting/ in composition with a
verb.

The word ?fof means ' taking 9 or 'accepting/ iftfrq? means there-
fore the opposite of this namely ' rejecting/ abandoning. As «r?qt wr tpjipr
he went away having abandoned home ; i renfrwyp^ falsehood being aban-
doned. 3ftR| if*G •

Why do we say 'when meaning to abandon 9 ? Observe qjsafaf
ijfTOtf jSnUltf: the hawk went away taking hold of the mouse.

Vart : — The word «J^rr is treated as an upasarga for the purposes oi
the application of the following rules ; HI. 3. 106 by which WT is added ; rule
III. 3. 92 by which f§K is added; and rule VIII. 4. 14 by which *r is changed
into ot. As WSTgf f a^if$: and Hj«4i<$J|ft #

*51m^^ sjjnrcihn?! ii %$ 11 v^ifk 11 «*Mnrol f sijt-
inft-tn^, (nfir:) u

*fa: 11 «$jj*fr wj j umi 'ror irfhm* iflttak wr: 11 '

66. The words ka^e and manas are gati when
in composition with a verb and used in the sense of 'reac-
tion by satiation.*

Digitized by LiOOQ LC

108 ~ Gati. [Bk. I. Ch. IV. § 67-69.

The word M^Mftftara means ' satisfaction of desire/ As qrfftqr
W. ffafir he drinks milk to his heart's content or till he is satisfied? So
also T4t?nT TO ft^ft . That is to say, he drinks so much till his desire ( WET )
is satisfied by revulsion of feeling. When not having this sense we have q*4
or q^terqnpr : he went away he satisfied.

trtt wmi^ n ^ « *^Tf3ir 11 1**: , stow* , ( ifife ) a

67* The word purah in front of, when indeclin-
able, and in composition with a verb, is called gati.*

The pronoun ^jf 'front' with the affix «rflr forms ^rr^, and by S. I.
1. 38 it becomes an Avyaya. (See S. V. 3. 39 pur being substituted . for
pfirva). The object by making it gati is threefold (1) compounding by II. 2.
18 (2) accent by VI. 2. 49 (3) to change the: fc into ^ by VIII. 3. 40. As <v?*$ft* 9
H<*fr*iH and 11^ y\*±ht\fb, when not an indeclinable we have ^: , j?ft , JT: ff?*r

srer * i' \c 11 n^r 11 s^si* f *, ( *rf?r: swmnr) n

68. And the indeclinahle word astam c at
home/ is called gati, when in composition with a verb.

The word H/uiu is an indeclinable as it ends in^(I. 1.39); and
means ' not visible.' As tpsforo ^fiffiT yr**f% the sun having set, rises again
4J4d<|tMft >*mPt riches that have vanished. Tppnt Jj«t»ft, when not an indeclin-
able it is not gati as Httt CEfQCT the arrow has been thrown.

era ict^^j 11 \% n ^rf^r u srra 9 *rfif-3i^*^i| 9
( nfai siaws^ ) *

*wfa II

69. The indeclinable word achchha, meaning

* before in the presence of,' is called gati, when used in
composition with verbs denoting c motion* or with the
verb ,< vad ' to speaik.

The word *i«3 is an indeclinable and has the force of the word lift. :
As Ws&t?*, H*5WJ , and qq?mi*bft . . So also *!*$m «TK5tftn^ , and H1*Q-
qjfa when not an Avyaya, we have 3f3iTOi *I*6fa .

•^ JjF^J it bo 11 wrfk 11 w?« , •ig'WiX'^O w .

Bk. I, Cii. IV. § 70.73. ] Gati. - 109

TO. The word adas c that * is called gati when
in composition with a verb and not implying a direction,
to another.

The word upadeSa means a direction to another* When a person
cogitates within himself and does not address another, that is anupadesa.
That is when it is not- a demonstrative pronoun. As *jf: $>*?• W f^l
and w^(: 37ctf% when used as a demonstrative pronoun we have *TT

f?Jtt J*dffi II to? II ^?[Tf^ II fH^: , e^lfl, (ifei:) H

71. The word tiras when used in the sense of
' disappearance,' is called gati when in composition witlx
a verb.

As f3ttt*gr» fifl*t*l^ and ^ fftfftiqft why do we say when meaning
disappearance ? Observe fiftt Tttfflqw: i. e. t standing apart.

fawror^fe 11 \s* u mrfk H fsmnw, ?fa 9 (*rf&: «F?it*) »

72. The word tiras meaning c disappearance *
is optionally called gati, when the verb kpl follows.

This is an example of sftff-ftTOTT. As fsr. fff* or ft<*$U| (VII L $*
42): fift: fwr or fiii^^l (VIII. 3. 42). The change of visarga into sr Is
optional VIII. 3. 42. When not meaning disappearance we have ftfc
fJW 3ft# f<tefir he stands having laid aside the stick.

^m%J*QfT% 11 s^ n y^iHt ii ^rm%-^FmSt 9 (fenrai
^fsi vfoi) 11

^frT: 11 3*n^j?*r^s(j5ffc ftn^K *ifiwran fan™ j^w otto?dii# wSfa m

frfs? f%Wrf *lf*W<" Wf: II

73. The words up&je and anv&je both mean-'
ing c supporting or assisting the weak/ are optionally
called gati when used along with the verb kjri.

As «n^ frar or gqr£ ffffT having given support Hj^ili $nr or

*ir*f ^ fror* *•

^Tsmwj^fa* 11 t»g 11 *^rf* 11 *TOi^ f n^ftf* f *,

«^T: II ^TOTOPpftft JJ^VJlfr yft? ftqrar *rf*RfHlft TOfal II

5

200 - Gati. [ Bk. !• C*C IV. § 74-77-

74. The words s&ksh&t, c in the. presence of*
Ac, are optionally called gati, when used along with the
verb kri.

Varti — In the words **hrU^ &c. the force of the affix f%^r is under-
stood, namely making a thing what it was not before. As *U4fin ^HRf or ffcfr
fkiziff%m ffaztt^Wf making evident what was concealed before.

iRrefin fnn it

75. The word urasi, - 4 in the breast,' a^d
rnanasi ' in the mind ' are optionally gati when the verb
kpi follows, provided that they are not used in the sense
of c placing/

As xiifin$< ? or 4K(ht>*tt ; <Mft$*3 or *HiV^f . ' When it has the
sense of placing wehave ^ifa^ fr qifa ^ he lies down having clasped the
hand on the breast, *Hfa^>* 3T *^f mmfil he ponders having . placed 'the word
in his mind.

1^ ^ fiwfl * ll b$ li H^ifa II *virtj^ f fore%,

^fxT: ii *Mfrqi Pta^^nN} 3cr*^r, HT^rnn^ f%Hm «fSr *rfcrcr*n *m/5<i h

76. And the words madhye c in the middley
pade * in the foot ' and nivachane c speechless * are option-
ally gati, when kpi follows, the sense not being of
c placing.*

As H&i ^pnj or ^stqr. qf f??ar or ^j&r ; firare? fiHT or ^Rqr.
But when it has the meaning of placing we have fffenr: «l% $R*T fijjT:
d# he lies .down having put his head under the foot of the elephant.

i*fi?j ij^t *ii*nicjmw^ n 33 ii xj^Tf^r ii f*i?i* , ips!-*n*fl

<4<4J4+& (nfg: ^ftl) II

*fir: u ^ <rn& x**$t jf$I ^ftr fro* *rf^ft wr OT3H3 ii

77. The word haste c in the hand,' paijiau 'in
the hand f are always and necessarily called gati when
used with the verb kji in the sense of c marriage;* •

As qnnr ^pi^r or f^ ^j?^ having married. . But f^d fftCT'^irfcrf *ni:
he went out having taken in his hand a.k&rshipana (a coin). r^

Bk. -I. Ch. IV. § 78-81.3 Gati. 201

*ifnffwf^r«f: arf^i f%?4 *rfo«J^ *i*fif 11

78. The indeclinable word priidhvam followed
by the verb kj*i, is always called gati when used in the
sense of binding/

The word *n^ ends in ^ and means ' favourably suitably. 9 When
however it means ' bound f it is a gati : as, snW^TOT having bound. But
when not meaning to bind, we have : — muf fr^t VRii W. having made the
carriage agreeable, he is gone.

efta*2t, (nf?r: stf* ) 11

^Rt: II «flf%cRT >WpiMR*A<tf IFw **tv£t fi^KI SjfsT ijfa&tft TOT: II

79. The words Jivik& and upanishad followed
by the verb kvi are called gati when used in the sense of
* likeness or resemblance.'

As ?ftf%3\T3yKr having made it as if it was a means of living ; OTftTO-
ipm — having made it like an upanishad. But 4tfa<*»l fffyr ipf: having made
his livelihood he is gone.

f* xn^m^: 11 co 11 q^ifir n ft, fra^.innt:., («rftr:

^t: 11 $ ireyrepfcrerarn qrfh'mi xi^K^ir: 11

80. The particles called gati and upasarga
are to be employed before the verbal root : (that is to say,
they are prefixes).

The word # has been employed to include the term 3<rerfalso. The
preceding examples all illustrate the application of this rule.

m**\ mo) n

^ffc 11 E*?flr fire* m^ro* ^f«r; <tf Jft <j*ft mJ^E^ir: ti

81. In the chhandas (veda) these gati and
upasarga are employed indifferently after the verbal
root, as well as before it.

As.— ^ruft^rv *?** ^m i <wft* y*nj n if&m fa*? «ixr (Rig. I. 2. 6).

** Digitized by VjOOQLC

202 Karmapravachaniya. [ Bk. I. Ch. IV. § 82-84.

In this we have WTOHPgq instead of tiqnnPj/ vayu and thou Indra, ye
heroes, come ye both quickly to the soma of the worshipper by this sincere
prayer/

arafgmq « t* 11 mpfa 11 sroni<n: f *; (no ««

82. I11 the chhandas (veda) these gati and
upasarga are also seen separated from the verb by inter-
vening words.

As:— *** *ra ^ *nrr 3* Jnftfacnm^ii f^€lr *r gqjfcaft (Rig. I. 2. 4).
" Indra and vayu, here are soma — libations for you. Approach, ye with
pleasures for us. For the libations are desiring you." Here ot is separated
from the word 3n*T9 hy the intervening word sritfir. .

Karma pravachaniya.

83. From this point as far as aphorism I. 4.
97 the particles treated of, are to be understood as hav-
ing the name of karcna-pravachaniya.

These particles govern a substantive and qualify a verb and
differ from other particles termed upasarga and gati. As snsjtf ?&m well
praised by you. *| fori? ?mr you have sprinkled. Here *f is a karma pravach-
aniya and qualifies the sense of the verb, /. e. it is an adverb. It is not an
upasarga, for had it been so it, would have changed the *ET into * (VIII. 3- 6 5)-
The term karma pravachaniya is a big term compared with other technical
terms such as, f>, fir , 5 &c, which generally do not exceed more than two
syllables. The word karmapravachaniya is not however merely a technical
term ; it contains within itself a definition of itself. It means that which
qualifies or speaks about (qjffiCTHTj an action ( q& ) is so called.

«3««3| II ch 11 ti^Tfif 11 3*«|: «stf* f ( mj« qyftqr:) II

84. The word anu when it denotes a sign, is
called karma-pravachaniya.

The word lakshana means a sign; an attendant circumstance an
invariable concomitant or accident in its logical sense and not merely an
occasional concurrence. The force of sj^ in this case is that of* • after/ 4 in
consequence -of/, because of, or being indicated by.' As JiTO^-TOfcl it rained

Bk. I. Ch. IV, § 85, 86 ] Karmapravachaniya. 203

after or in consequence of the muttering of prayers. To put it in other word*,
anu is a karmapravachniya when it governs the word which indicates the
cause or the attendant circumstance of an action.

So also grrarereq ^tftnnr5i?Rr^-?Tra^r g*j*tf tfft*rprrftjp*r«qr: mro;.

The God (clouds) rained having listened to the well performed reading, by
£&kalya. The force of a karmapravachaniya, therefore, when fully analysed
will be found to be that of an upasarga whose verb is not expressed in the
sentence but is understood; and because of its being so understood, the
particle itself seems to govern the noun which in fact was governed by the
unexpressed verb. " Thus in the above example ff|%<rt is in the accusative case,
apparently governed by the preposition ^5 , but really governed by the
verb wf%3[F8f understood.

So also ^pr?^ auwqfiit-q^ . The Taurus sprinkled water after the
sacrifice, <*j4j*U4W4ftta<^ ifsiT: the people began sprinkling water after the rising
of the canopus.

Why make unnecessarily this sfitra, when *& would have been called
karma parvachanlya even by force of sfitra 90 following, where also the word
lakshana occurs ? This is for the sake of indicating that the karmapravachanfyas
should always govern the accusative case (II. 3. 8) even in expressing $g
(II. 3. 23). Otherwise Rule 23rd of the third chapter of Book II would have
set aside Rule 8 of the same by the maxim of to? I. 3. 2: and would have
caused a karama pravachantya to govern an Instrumental case where the
sense was that of hetu.

85. The word anu is karma-pravachantya
when it has the force of the third case.

The meaning of *?•[ in this case will be that of ' with f or along with.
As .fftqj ?Hrfo<TT sfaT the army lying along side the river. t}<PP[ 9nfo*r *hrf
lying along the slopes of the mountain.

^% II 6$ 11 n^rfa 11 ^, (313: qnfno) II

^Rt: 11 ^ cfoWfisni^: 3v$ Jnr^fhr sWt *raf?lr 11

86. The word anu is kanna pravachanlya
when it is used in the sense of inferior or subordinate to.

The word ftn means c inferior/ and being a comparative term re-
quires the other term with which comparison is made and which is superior ;
to complete the sense. In other words *f?r governs the person to which others
are inferior, in the accusative * case. As HJH^l^4l44$4l<h a lUJI: ' all gram man-
ans are inferior to&katayana/ " '^^GoOgle

204 Karmapravachaniya. [ Bk. I. Ch.. IV. § 87-90. -

«iftfa$ * h 63 11 gt?. srfafc, * , (^ ^ no) » -

- 87. The word upa when it means c superior '
or c inferior to ' is kaiTaa-pravachaniya.

That is when the sense is that of 'over, above/ or 'inferior* as OT
*3Tr£f Kto: a Drona is above a khiri. 3<r fir«^ wqfTCT. a karshapana is more than
a Nishka. In this sense 37 governs the Locative of the thing which is inferior
(II. 3.9). So also 37 mcteu rf $*U<hlU»: 'all grammarians are inferior to
S&kat&yana. 9 In this sense 37 governs the accusative case.

ere **S ^% u cc 11 sre-xi(* f «pfc,.(«5no) n

*jf^f: II 5JT <rff ^T*^f ^f$^R$ «hjfc4^«iNl4i*H *TO II

88. The words apa and pari are karma-prava-
chaniya when meaning c exclusion. 1

The force of *TT and qft" is in this case that of ' with the exception of/
As wfaipfr'iT *jst fo: it rained outside of or with the exception ofTrigarta.
So also TftPnRwJlr ^flfcp • In this sense they govern the noun excluded in
the ablative case (II. 3. 10). When not having these senses, they are not .
karma-pravachaniya. As sjfctf Mftfa^fa . He waters the rice. Here it is an
upasarga, and hence changes the 9 into * .

sn^^rohp w& 11 1% 11 ^^r n«rra^ f ^n^ET-^ra^,
(tow) n j

89. The word £& as far as, is karma-prava-
chaniya when it expresses limit (e. g. when it means c as
far as inclusive of or, c as far as exclusive of.)

v The word ^7 in the text shows that both sorts of limits are here

meant. As *jrm£l*W4U^t[?r $7: it rained as far as (but excluding) Pitaliputra
*»m*Rw4 ^?T $7: it rained as far as (including) Pitaliputra mi*h*U<
*nj: TTpl^: ^luf^hl^m , WT HUimi : when it means 'little' or is a verbal prefix
it does not get this name.

sa^C^^til^qRWI^cqigsift ^WJ II & 11 u^lPt n

^ n srer* i*tf *pron* ht* ^T^rnrf ^ fir*rcpijir*t *f% 7ft «pj p«*

Digitized by VjOVJvLC •

Bk.. I. Ch. IV, § 91; 92. ] • Karmapravachaniya. . 205

90. The words prati, pari and anu are karma
pravaxshaniya when used in the. sense of * sign* (in the
direction of) * mere statement of circumstance , ( c as re-
gards *) * division * c share of and c pervasion 9 severally.

As ( 1 ) ^f itf^-TR 1 or ^[ f%s?t*nfr fk*£\ f the lightening is flashing in the
direction of the tree.' (2) m^*V$t *TOt xrf?f — qrft* - H«r ' Devadatta is a good
man as regards his mother.' (3; fr Jffa %*U4*\ 'the poison fell to the share of
Hara.' mAHl *rf% **jnj that it may fall to my share. . So with qft and *?•[. (4)
^pf ^irsftr fa^sfir he sprinkles one tree after another. So with ^rft" and Hj.

stfH*m?t u \i n ^i h n «?ft: , aw** (ssn* *s*r«

*f*r. 11 g round* wnntemfl r; *Hm<*fU*<i4ft wfir h

91. The word abhi is kai-ma-pravachanlya, in
the above senses of c in the direction of,' c as . regards/
and c each severally \ but not when it means division,
1 share of? ^

The illustrations given under the last sfitra muttra mutandi apply
here also, with the exception of those given under head (4^/ As ^TOft
f%dta*t; HlWlft &c. But it is not karraa-pravachantya when* tot is meant. TOT
means the share which is allotted to one as his own. ' As 4{MHHlft c^RT ^
$ter?rpj give that which falls here to my share. Here abhi is an upasarga, and
therefore fr of ^r is changed into <r

92. The word prati is karma-pravachantya
when used in the sense of representative ( c representative
of 1 ) or exchange ( c in exchange for.^

That which is like to the principal is called srftftftr • Giving in
return for what is obtained, is srf?r?Fr . As Hjftn^i ^«i*f: *rf*T Abhi-manyu is
the representative of Arjuna. *ilNH^ flf^f: ^rf^^uft he exchanges M&sh&s
for these sesamum. In these senses prati governs the Ablative (II. 3. 11).

srf^m* 3to$%i 11 ^ n *5t6i h .vifr-vft-, s^r5% f

^jfiT: II Ufa itf tT«$T MTfeRi MfqfalTTTpRT **?iW^*ftvidi HWtfl^ ' » •

Digitized by VjOOQLC

206 KarmapravachaniVa. [ Bk. I. Ch. IV. § 93-9&

93. The words adhi and pari are karma-pra-
vachaniya when used as mere expletives.

As qfrfo frqm^kffr whence has he come ? or gj*f: <ntf *PS5f%. These
words though mere surplusages, have been called karmapravachaniya so that
they may not get the designation of upasarga or gati and thus produce
results peculiar to those words.

94. The word su is karma-pravachaniya in
the sense of respect, (when it means l excellently*).

As §Rtt1? ^f^rr , H *§tf H^ir excellently sprinkled by your honor. Well
praised by your honor. By not being an upasarga, it does not change the
* into v (VIII. 3. 65).

Why do we say ' when respect is meant 7 Observe gftrff ftf fmw
has this your place been well sprinkled to-day.

95. The word ati in the sense of super-abun-
dance {* excessively ') and ' excellently ' is karma-prava-
chaniya.

The word ujftftiiuj means to do more than what is necessary for the
accomplishment of kn object. The force of ^T in the aphorism is to draw in
the word <j5rranj from the last. As atf^Rffq? *mr. It has abundantly been
sprinkled by your honor, a rfirepfftq «WfT so also stffcmf TO3T excellently
praised by your honor, *rf?t ft^f *rro .

sift: , ^^-««nnf-w?^^^-ii^-^gi^3 ( *w ) ii

^f: 11 T^wf tfror^ 4^N*iiI *i?htf ^rj^d ^ h&hw. sift: ^Rtiwtitef&St
>*rflrii

96. The word api is karma-pravachaniya,

when it implies, the sense of word understood ( c some-
what ') or possibility (e. g., c even * in the sense of such a
great person), or permission to do as one likes, (' if you
like '), or censure ('even' in the sense of what v is disgrace-
flil) ; or collection (* and*); >> •

Digitized by V^iOOgLe

BlC I. Of. IV. $ 97, 98. J KARMAPRAVACHANIYA* •• - fe(>7

The word «renf means the sense qf another word which is not
expressed but has to be supplied. As *afhH\jft ^n^ there may be perhaps a
drop of ghee ^jfttffi *Rin^. Here some word like vindu 'a drop/ stoka 'a
little' mitri - a measure &c, has to be understood. The word B*TTOT means
possibility, a supposition. As, ^ftr fiNfc^ «J$FraTC*# ^f*l%£*[ possibly may
sprinkle a thousand trees in one moment; *rfr *d*U$ N <l*iH he' may praise
possibly, a king. In the surfive cases it is used with the potential mood*
&4«<wl means indifference ort the part of the speaker where he permits
another to do as he likes. It is used with the Imperative mood. Astjft f*ta
you may sprinkle if you like. *ifir*5[ftyou may praise, if you like.. The
word ij^f means censure, contempt or reproof. As fat[*tteqr 5*^wft firt^
q»ii"^ . The word ^ijsnr means cumulative. As sjfa sgf| — Mfir f^N praise
as well as sprinkle.

In all the above instances, the word not being an upasarga, does not
change the 9 into tr.

mfi<k^ 11 csa n tt^tto u srfa:, t^ • ( *^w* ) » *

97^ The word adhi is karmapravaohaniya
when used in the sense of " lord " (" being as a lord " or
" having as a lord").

• »

The word f^qpc means ' master/ and it therefore requires another
correlative word denoting 'property 9 of which one is master. The word adhi
governs a Locative case. Sometime locative of the person possessing,
sometime locative of the property possessed: as *fffr TOT*^ tf^rar: °r*lft
<ftr^r JTOT^ • Brahmadatta rules over Panch&las.

f^riRWi ?>fa n %t> n • m^Pi 11 finnan 9 s>fif .

^f%: 11 sjfin <tiClni fipiiNi 3rfcnwffarcf<ft H*flt 11

98. The word adhi is optionally karma pra-
vachaniya when the verh kji follows. .

As *f$*r *Plfi)F <h(lmfa . Here the word *rfa may be treated either as
a irfH or a ^pfopnftq . When it is a iifiir the accent will be regulated by
S. VIII. 1. 71 ; otherwise not. - Dieted by

6

208 Parasmaipada and Atmanepada. [ Bk. I. Ch. IV. § 99, iod.

99. The substitutes of 5? are called paras-
maipada. ' . ,.

The word w is in the genitive case, and means ' of m. 9 The word
•nfSJ 'substitutes' must be supplied to complete the sense. The term's?
is a generic word for verb in general, viz., «^, f*^ ^? ^, 'S^'sSt^',
«*£ f5n^ sp£ OFF} . The substitutes of m are the well known personal
terminations by which the verbs are conjugated in those tenses. Namely the
following. :

Parasmaipada.

Sing.

Dual.

Plural,

1st Pers. R^

^

^

2nd Pers. f%«^

'*V

*T

3rd Pers. ft^

. **. ,

m

And the affixes 3? and qr$ ( III. 2. 107, and 124, and
HI. 4. 78).

*j$wi«i<w3<4Vi « see u ^Tf^ n ^;3n^ mi^q^ ,

100. The nine affixes comprised under the
Praty&hftra tafi. and the two ending in &na (S&nach and
K&nach), which are substitutes of m are called Atma-
nepada. - *

The following are the Atmanepada affixes :—

Sing.

Dual.

PluraL

1st f^

*ft

*lft^

2nd m%

Wtt\

**l

3«i *

unti^

11

And the affixes Jjm^and «pr^(III. 2. 106). 30gle

Bk. I. Ch. IV. § 101-103. ] Person defined. * 209

*rcf5*rn

101. The three triads in both the sets Paras-
maipada and Atmanepada, of conjugational affixes (com-
prised under the general name tift, a pratyah&ra formed
of the first and last of them, viz., tip and mahifi) are
called, in order, Lowest (3rd person of European Gram-
mar, the middle (2nd person), and the highest (1st person).

Of the 19 conjugational affixes, above given ; 9 are Parasmaipadi and
9 are Atmanepadi. Each of these two classes is subdivided into three classes,
according to person, as shown in the above list.

^f%: 11 *»*qcM^H ffrrw w^r tfsrrfa *rcfar sw- ^Nf *f«j n

102. These three triads of conjugational
affixes, which have received the name of Lowest &c,
are called (as regard the three expressions in €ach triad)
severally " the expression for one v (singular), " the ex-
pression for two " (dual), and " the expression for many n
(plural).

Of the six triads thus formed, each is divided according to number
into three classes, viz., singular, dual, and plural.

§*: II 30^ II trcpft II gq: f (*tf*I * ^WHR f^nff ^J-
^Rlf^T ^*p ) II

^t: 11 «jto tfftj *ftj q*n% mxjj ir*>w ft^npr *^w tfsifir *w&* n

103. Of sup (which is a Pratyfihftra formed of
su the first of the case affixes and the final p of the last
of them) the three expressions in each successive set of
the three, are also severally called singular, dual and

Plural. Digitized by G00gle

210 VlBHAKTI DEFINED.; [ BlC. I. CH, IV, § JO4-I06.

The sup or the case-affixes by. which nouns are' declined are 21 in
number corresponding to the seven cases, see sfitra IV. 1.2*. They also
have three numbers singular, dual and plurah ;

^f%: II *ftf*l ^ftftj fiPffrS" SflTCV Wlfrl ^fttfi-V II .

104. The triads of conjugational affixes and
case affixes are also called vibhakti or Inflective affixes. .

The word fiprf^ir means a complete triad. Thus wpfr fa*F$fi( means
the three affixes of the seventh case, t. e. f the locative singular, dual,
and plural. So TOTT firHf^K" means the third person, singular, dual, and
plural. '

^f%: 11 *ror*J«nf$ *rf% unRft ^i*^qf^8 ^% WMifa^M.^ *t*iMrPH$ 3^-

105. "When the pronoun yushmad, "thou"
understood, and also when the same expressed, is the
attendant word in agreement with the verb, then there
is the verbal termination called the middle (2nd person).

This defines the 2nd person of conjugational affixes. As *f TOftl
thou cookest or T«rf%; ^t <TW or <rw: you two are cooking ; zri q^qr or vr*m
you cook.

mjro ^ tuifiqq^ ^r^^viH q^sng 11 %c\ 11 ti^if^r n

U^re f * f ^q-^qq^ f *T?q?t: , ^W: f Tjqj^, * J ( *T«TO: ) II

***fa 11

106; When joke is implied with reference to
an action, the verb denoting it is used in the 2nd person ;
provided that the word many a * to think * is the atten-
dant word (upapada) of such verb, and of the verb many*
itself, the affix must be of the 1st person and singula?
niimter. \ . -«Xoo^e • .

BK. h CH. IV. $ 167-109. ] AVASANA DEFINED^ 211

The word jjsw means joke, sportt Asqt *&$ Bfftif $t&& fft; «rf>
HT^rt , ^fst: ^Nmfaf**: • Thus thou thinkcst ' I shall cat ri ce/. thou sh alt not
eat, that has been eaten by a guest 5ft *&t kOh uraifir 1$ ^fWfir , TO^thf
$ f^rr • I think thou wilt go by the chariot, no, thou wilt not, thy father has
gone before thee on it. When joke is not intended, the proper persons
should be used: as, lift iftmft iftvt Ht*£ • Thus thou thinkest 'I shall
eat rice/ • "

sreT|jsw n %#& n vimfk u QRPRf^, ^pt*, (4i<i$ vb-

^nfvi^5t ^[fwrqfq ) H

^fa: 11 grewynfr OTTTTPir^ h^iwi* « w^wr* jft vwwtit ^rt% u

107. When the pronoun asmad " I," under-
stood and also when expressed, is the attendant word
in agreement with the verb, then there is the verbal
termination called the Highest or the 1st person.

This is clear. As Mtf TOlft I cook, or merely V*lft-*Jf*t ^TOHT: or
merely TTOT .

«^ »*!*: 11 *o6 11 rvrfH 11 5ft mni: n

108. In the other cases, namely where,
" thou " or " I " are not the attendant words in agreement
with the verb, there is the verbal termination called the
Lowest (or 3rd person).

As q^Rr he cooks, toj: they two cook. qrrfSfl they cook.

109. The closest ^proximity of letters, there
being the intervention of half a m&tr& or prosodial length
between them, is called contact or sanhitA.

When words arc in sanhita, they are glued together by the rules of
sandhi. As ^a?nr - ffa + *nc .

Sit AVASANA DEFINED. [ BK. I. CH. IV. § IIO.

110. The cessation or the absence of succeed-
ing, letters is called pause or avas&na.

The word avas&na occurs in s&tras VIII. 3. 15 &c.

Digitized by

Google

^

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

* *'- •*

Digitized by

Google

12 «2225 OWER W,LL 8E CHARGED
** OVERDUE FEE IF THIS BOOK IS
NOT RETURNED TO THE LIBRARY ON

BELOW. NON-RECEIPT OF OVERDUE

«oo£L S ° OES N0T "eSpT ?HE
BORROWER FROM OVERDUE FEES

Harvard College Widener Library
Cambridge, MA 021 38 (617)495-2413

L

D]igitized by

Google