Bhandarkar Oriental Research Institute

Pgst-Qradgate and Research Department Series, No, 2$

Fascicule III : Ahnika VI, Part I,

critically edited by V. B. Bhagavat Saroja Bhate


Copies can be had direct from the BHANDARKAR ORIENTAL RESEARCH INSTITUTE fOONA 4}1 004 (INDIA )

Printed at the Bhandarkar Institute Press and Published by R. N. Dandekar, Secretary, Bhandarkar Oriental Research Institute


3 %5R, I ^TK ^

Jin: I

[ ^ ] zm^iziwm\



< \ wi ^mf^^r^fq- [ ^ ],

f is

: i mwz; 3Ti^f eu^- fmfir ^e^i% $fa i s ) I%%: ' (qr. 20

( ? °^ra )


1 ^T


I I% cflt ?

5T T

<TT. ^. R .








1 q;4


I sf^OTjru ff 15


1 f#; crft ?
- 25

( ! °^ ^ )

[ 3. ? ] |?
1^ c[f|

?T, 3f]f





10 :I


2 D

I fa:?

, \S

ficT I





OR4 RIRlf^f7 ffcf | ^

^^r: i


' (m. 6. v. ^ )


^ r [ ? °«r ]



1 ^ g








: [ ftfe: ] 5nqajr3-






[ I T ] ap^qf S^PT



q* c[


, ^0






) sRT 1 3fWlfq I f


f, HfTIT: ^flrfT^im^: fief |

(qT. <:• «. ^ ) #



^.^^ If 3


?• ^ ^ f€ ^ f&ftft 3 ^ 5rwrf^f ^ T ^ f ^ T ^ ^ (qr. vs. ?. c \)


' (qr. ^. ^.


wr. ^ %
l w '
( qi. ^. '?* ^ )


( qr. H. ^ V9? )


^ ^ 5T[ fcf ] q ^ q ^ ||



- [ m$ ] H

# ^


i qfRft- 25

3 # 3 Q;^R


^ »


3, H *r ^ f ^r 6 ^": ? T^J ) ^ N I ft sricf^r i^in^

(qr. ^. ?. ^ r e r : ?rSPTI ' ....^ff^lf^q-....' (qT. ^ ^ . ^ H ), 'cq^lGfRI^....
20 \ o^) f M i *ra ' ^ ^ H ^ T F ^ . . . ' (qr. vs. ^ . Ro ) ffct3

(qr. vs.



1 CT^ ^

< S ^ H ^ I I ^ ^ ( TT. ». v ^ )


, ?I

• \ % \

f § ^ f ' (qr. vs. ?. ^

)^ ^

i mmqf^pf

- 10

( 7T. vs. ?. ^ ? )



=t f

mM^RT ^ mzm ^ mm

SfT- C ) f ^



[ R] * R



vs. t

^^ ) #



^ f R^ 51 aft WTOf?r? I ^ IPT T
I '


f% cfff i

( qr. vs>. ?. % ) #



? 3T?PTT


i ^f ^rerf 10

' (qT.


....? ( m. a. ?. ? )


\ )f
: I

?m I q^T

( ^T. vs. R. { o ^ )
5 3P ft

, ?T:




wr. cc \\


W F F R RRrf


, cflR

I era [^ ]



m m

5 IR^T^;

:5 til



20 ^




3 , tfVS

^ cf^I

# ^ c M


: I]

tm i
<TT. vs # ? .
\3. ^. 15

Hfcft H?T f^T^ I feffiWcf

q ^


• I

^ it^^f: ?
?fi mm






10 m\ mm



ih i -qr^fr fr Btsq H5^ ?T%

( <TT. ^ =1. ^^ )




^ 2 f^ i RXJ I

V 9



« •

% ^%
1 3fiq

?°%? 1 )fcf



V. qt

%qt srrir

mm # , ^ gft

TRl^ R
« •


^ fw


qr. ?. R.

n • ? • ?• ^ 20 ?• star


' (qr. ?. ?. ^^ )
| c|5f

: I 5Tf^T [ ^ ]
arq- %4
IcF 1

m $4 10
i ....



tm i




m m r snfac^: 1

mmmm w
^ vs ^
) | ] % | S^cT i


' ( qT. ? . ?. ^

^ €

' f]% ( q j . {.



( ? Cf| )



• ! i §*er: r




%m i

- 10










srmlfcr i

iff^tf i
( ? sifj^g )

10 ^ (I =5^ ) o q ^ ^ j fp:q% | 5TNT^Rq^: I

15 f^cTs ?

n t ?•
?piqi% ^ t

^^crf | t %

? 3*



iTO F%[f^§T


fcfFT%Tff^JM: ffcf I


im g 1 ei«rT


^f f l
, mi



=f q


( ^ ^. t ^ ^. ?) i

f^T:.... ' ( 7T. V. R. H :' ( TT. ». ?. ? o o ) ?^


cT ' ( ?T. ^T. ^ . ^ . U ) I



m ^ f ITICT ?

?• \9 rriR [fl] i ^ R

i w: qis: I ?i% i ^




:| ts: ^ ^ ' A I Fff^ffi


TWct, m^\ If I %

- 15






) ?i% i


^i%^qRqq r f ^q- sB^qn ' ffar

qr. ^.R. ? R ) fter

, cfcf:


j mm


. 20 (TT.

I cfqi f l -



m. %\.

erar q"|i

H?f% jpr:


em- 10



I cf^TT 20

? '




:' 10 ' e
<TT. v s . ^.

,I ^

i i

f | ^: ^
: ^ B


20 5r ?r si

ft HW




- 10

I cT^RT^ ?T








" ( T h e words) beginning with sarva (are called) sarva( P . 1. 1.27) naman"

"What (kind of) compound is sarvadini!" (This) question is unjustified, because of the absence of doubt. A doubt arises there, where the form ( of the word) is the same and there is no means for under« standing a particular (meaning out of many meanings). For instance, (in the compound) sthulaprsatim it cannot be understood whether this is a tatpurusa or a bahuvrihi, since the accent is not marked. But here It is not so. Now, if (the compound sarvadtni) is regarded ( as tatpurusa meaning) ' sarva which is the first,' the word adi becomes redundant. Moreover, it is not proper to abandon the gender and number of the word adi, since it is the main ( word in the compound ). Further, the inclusion of viiva* etc. in the (sarvadi) list will be purposeless.

Now, if it is regarded that adi means that from which (something) is taken and (thus sarvadi means) that from which sarva etc. are taken, the assumption ( of this meaning of adi) is unjustified, because the word adi is a conventional word.

If (the compound sarvadi is understood as) ' ( p r e s e n t ) before sarva,' this (interpretation ) is also not correct. There is no other ( word) that can be before this ( word sarva in the sarvadi list). Thus this ( word sarvadi) is only a hahuvi'ihi and so the doubt is not justified. Similarly, when it is stated •* ( he ) says that it is a bahuvrihi," (the question ) " what is its constituent analysis? " also should not be asked.

This (question regarding the compound sarvadi) does not ( constitute ) a defect. Here in the formation of ekasesa ' remaining of one \ one who intends to speak gargyau may mean to convey this particular feature (viz) that one (of the two ) has his father alive or (he may mean to express) progeny of Garga in general. And this ( knowledge of one of the two meanings) depends upon the person who intends to speak. For instance,


Ma h ah h a tyja-L Ipi ka , . 1 //• n i ka



( i n the- expression) tadadyacikhyasayam ( i n P. 2 . 4 . 2 1 ) * when it Is intended io speak about its beginning' ( t h e w o r d } acikhyasa Is used when it was possible io say tadadyakhyayam, so that ( t h e r u l e ) should be applicable when the speaker intends ( a n object) to be the first,

And here ( one of the disputants ) asks whether this is a krtuikasesa bahuvrihi or a simple bahuvrihi. sarvadi ( means ) sarva which is adi (the first). Or sarvadini (means) those (words) of which sarva is the first. (The word sarvadini is retained in the ekasesa) by (the rule) napumsakamanapumsakena etc* ( P . 1. 2. 69 ) or by (the F a r . ) svarabhinnanam yasyo-

Uaravidhih (vai\ 24 on P. 1. 2. 64). And in order to obtain (another word with ) the same form (i- e. sarvadi)^ (the compound sarvadi meaning) ' sarva which is the first * is assumed. In this way it is possible to consider (sarvadini) as ekasesa. And without the device of ekasesa it is ( only ) possible to accept that that which is conveyed by the word sarvadi (i.e. visva, etc.) is sarvanaman*

Another (disputant), keeping in his mind the maxim which will be explained (later), states that this ( word sarvadini) is bahuvrihi conveying its primary meaning and not (a word ) which has incorporated in itself another (compound conveying another) meaning: " ( H e ) says * it is bahuvrihi." " What is its constituent analysis ? " From the words obtained through constituent analysis it is possible to understand this fact: whether this is a simpb bahuvriht or krtaikasesa bahuvrihi, just as in the case of sthulaprsatim the special meaning is understood through constituent analysis and thus a compound is understood (from its constituent analysis). Therefore, (the disputant) asks : " what is its constituted analysis? "

Now, why (is it said ) Ss those which are these '* ? (It is) replied that (it is so said,) because without this the meaning of another word is absent. ( The won! citragu) cannot be bahuvrihi by merely (sa\ing) ' his variegated cows,' because (in that case) cows are signified (and not the person who possesses them ). And (it can be) a bahuvrihi there, where the variegated cows qualify their possessor. Thus the relationship between the main and the subordinate (members of compound) is reversed in (compounds such as) gonian, mamjatah and citraguh. And the meaning of another word is understood from the pronoun.

2.24 ]



'* If this is so, to the word sarva (the name sarvandman will not be applicable)." ('U this is s o ' means) if this bahuvrihi conveys primary meaning (i.e. if it is a simple bahuvrihi). " Because bahuvrihi conveys meaning of another word," Because the meaning of another word is the factor which is the cause for the use of bahuvrihi, since the meaning of another word is the main ( meaning) of bahuvrihi. For instance, when (the compound rdjapurusa ' king's servant' in the expression) rajapurusam dnaya < Bring the king's servant,' (is understood as having a ) jahatsvdrthd vrtti 4 an integration with the constituents losing their own meaning/ the meaning of rdjan (namely, king) with no word to convey it becomes a mark which is subordinate to purusa* Since it does not become the main ( member of the compound), the action (of bringing) is not understood as meant for him (i. e. rdjan). In the same manner ( he ) says that the variegated cows also, being known as a mark (of their possessor) are not (understood) as main (meaning of the compound citragu). Then what are they ? With no word to convey them they become subordinate, ( and so) they are not understood in connection with action.

Or, tadguna is that which is accompanied by gunas * qualifiers.* Tadgunasya samvijndnam * knowledge of it accompanied by its qualifiers' means that the meaning of another word along with its qualifiers is understood in connection with action. But how is this (meaning obtained)? Indeed it is said, " Because the former (kind of bahuvrihi) is taught in the sense of matup " {vdr. 5 on P. 2.2.25). Some (suffixes) conveying the sense of matup are restricted to the domain of owner-owned relationship, for instance (in the word) gomdn * possessor of cows.' Some other (suffix is used) in the sphere of association, for instance, (in the word) dandl l (a man ) with a stick.' Some other (suffix is used ) in the sphere of inseparable connection, for instance, (in the word ) visdni * ( bull) with horns.' Verily a hornless bull does not become visdnin * possessed of horns' by merely fastening a horn in his neck. Similarly, ( a person ) does not become yajno* pavitin or nivitin (simply ) because ( he) has a yajnopavita * sacred thread.' In the same manner the bahuvrihi words suklavdsdh ' ( a man ) with white clothes * and lolntakostiisdh * ( men ) with red turbans' convey qualified association as does the word visdvl. This ( word ) citragu conveys only owner-owned relationship. It should, therefore, be understood that this is the nature of the word itself which is diversely present in different meanings.






Moreover, the view, namely, that this, without which that does not exist (is a characteristic of that) is not (accepted ) in grammar. (The relation is to be understood) according to meaning. Just as the ( act of) eating is not without in the same manner (it is not possible without) the pair of clothes Therefore, the pair of clothes aiso is part of (the act of) eating. This is the meaning of the words (suklavasasah bhojyantam). In grammar, the decision about meaning (is made according to the maxim ) of mvanable existence of one depending upon that of another.

Also this (i.e. the view that) the qualifiers and the qualifiers of qualifiers of one who is mentioned elsewhere are not understood (in connection with action) is not accepted here. (For instance, when it is said :) ' Bring him who was seen with ( a pair of) white clothes ' ; here the pair of clothes does not become a mark (of the person when he is brought). When an order .s given ' Give food to those Devadatta etc. who were seen ' then Devadatta partakes in the act of eating although he is mentioned elsewhere ( i . e . in another context). Therefore, the nature of word itself should be accepted { as the cause of the connection or otherwise of qualifiers with the action).

(Objection:) Now, the word sarva may not get the name (sarvanaman) even though (the qualifying constituents are) understood in connection with action. (The words) viim, etc. qualified by it ( i . e. sarva) and accompanied by the word sarva may get the name (sarvanaman). For instance, (when it is said:) 'Give food to (the person) who is wearing a serpent as a sacred thread,' the person eats only when accompanied by a serpent as a sacred thread; not when he is unaccompanied by the serpent. Similarly, (when it is said :) ' Give food to (the man) with ( his) son on his shoulder' (the man ) accompanied by (his) son takes food, ( not when he is unaccompanied by his son ).

(Reply :) This difficulty does not arise. (When it is s a i d ) ' Give food to him along with his s o n ' , owing to the use of the word saha the (action of) eating is shared also by the son even though he is subordinate. For (the statement that) he ate along with his son does not (come true) if the son does not eat. This is a special feature of (the action conveyed by ) bhujitself. ( W e ) do not say that it is (directly) understood from the word




(bhunkte). Just as somebody should eat with two mouths, in the same manner he eats along with his son. Thus the son participating in the action is a qualifier. This is not the nature of the word.

Here also (the sarvadi-vtords ) visva etc. directly ( receive ) the name (sarvanaman), while (the word) sarva (receives it) directly and through inference. In this way (the name sarvanaman) is applied to sarva etc. Words are used in a number of ways. Sometimes only the status of a qualifier and not its close connection (is conveyed by a word), for instance, (the words extra go in the expression ) citragu, * ( a man ) possessed of variegated cows '. Sometimes (the qualifier ) is present near the activity but is not connected with the action, (for instance, suklavasas in the expression ) sukla* vasasam bhojaya, * Give food to (the man ) wearing white clothes,' (or nagayajnopavita in the expression ) nagayajllopavitl bhojyatam, ' Let (the man) wearing the sacred thread of serpent be fed.' Sometimes (it is) a qualifier participating in the activity (for instance, putra in the expression sahaputram bhojaya ' Give food to one along with his son ' ) .

Or the word adi is taken here as meaning ' near' and through nearness it is a qualifier of the group itself, not of the (individual) constituents, visva etc. ( But) since the name (can )not be connected with the group ( i t ) is given to the (individual) constituents. And they are to be inferred. In that case, why does not the word sarva receive (the name sarvanaman) just as the words visva etc., being inferred, are connected ( with it) ?

(The name sarvanaman for the word sarva) is achieved also when the word adi means ' k i n d ' . For instance, (in the expression) kalyanipancama ratrayah ' the nights of which kalyani ( is ) the- fifth ' (the nights ) are not understood as main by (merely saying) * kalyani ( i s ) the fifth.' What then (should be said)? ' O f these' (should be added). Just as, ( when it is said : ) ' Bring Brahmins from the house of Devadatta,' Devadatta also is brought because he is a Brahmin and he stays in the (same ) house, in the same manner the word sarva is also ( one ) of these ( words visva, etc.). So it is included in the meaning of another word ( i . e. other than the constituents).

The objections by assuming that the word adi means * first' and ' near • were raised and refuted. Both these (views) are different. (The


Mahabhasya-Dipika, Ahnika VI


statements, namely,) that 6 the qualifiers are to be understood (in the activity )'. and that ' me collection (is} the meaning of compound' as already stated in the Bhasya are presented { here ).

" Here (in the word) sarvanamani (n replacement by the rule purvapadat.. etc. [P. 8.4.3] is applicable)." (Objection:) Now there is no application ( of terms like ) * earlier constituent • and 4 final constituent' in proper names. Even if it is there, even then it is not (found ) in technical terms. In what then (is it found ) ? In popular names. For this reason also (the objection that there should be n-replacement in the word sarvanaman ) does not stand.

Moreover, since the relation between the name and the named is artificial (to be made ), everywhere it has to be explained how the rule of nreplacement applies. For, it has been already said : * when the relation between word and meaning is ( already ) established (the science of grammar proceeds to explain words)' ( Vai\ 1. Mbh., Paspasa). Further, it has to be stated whether w-replacement is already effected in the name or whether (the name ) without n-replacement effected in it later on undergoes ri-replacement. If (it is said that) ^-replacement is already effected (in the name ) (it may be asked ), ' By which ( rule ) is n-replacement (effected ) before it becomes a name 7 ' If (it is said that) ^-replacement is effected in (the name ) in which ^-replacement is not already effected, even then due to the word undergoing modification there arises the undesirable situation of ( accepting) the impermanence ( of words).

Perhaps this word might be regarded as a single (non-compounded) word since the user is free to choose any form. If (you a s k ) : ' How is that?* (The reply is:) this kind of (constituent) analysis does not exist anywhere from the view point of ultimate reality.

However, in grammatical practice sometimes constituent analysis in accordance with the meaning of the word (in the usage is given). For instance, in the word rajapurusa the meaning of the words accords with that of the sentence (i. e. non-compounded expression). In case of some ( words) a group of words (i. e. a compound) conveying a meaning other

5,3 ]'



(than the meaning of Its own constituents) is grammatically explained by means of the two words conveying a different meaning, for Instance, (words such as) gaurakhara, lohitasah, krsnasarpa, etc. The properties such as fairness and harshness are absolutely not there in the ( meaning understood from) group ( compound). In this way in the grammatical explanation of a group (compound) conveying its own form (i. e. standing for a name ) by means of two words belonging to two different classes (i. e. conveying entirely different meanings) the reference to former and final constituent is not discarded.

And the technical terms are not different from popular names. Whatever is this nature of ( popular names) is also (the nature ) of these technical terms). Therefore, n-replacement must take place here as there.

Nor is the (principle) of permanence contradicted in (assuming) the name-named relationship. Why? Because the restriction of power (of conveying meaning) is man-made. For delimiting the scope of power human effort is (accepted) in both, the one-power and the multi-power theories.

( I n this connection ) the view of others is (as follows ) : The wordss the meanings and mutual relations which are known through direct cognition are permanent, for instance, ganh, aivdh, agnih, udakam etc. But the word, which conveys meaning through inference, namely, the (word) with a convention, is impermanent

Another ( grammarian ) says : Even in that ( newly assigned ) meaning all the words ( being used as) names also are permanent. ( Thus) ri»replacement of that in which n-replacement has already taken place is admitted in grammar. ( When it is said that) n-replacement takes place in a name (it means that) the word in which n-replacement is already effected is correct. (Grammar) does not produce another form. A name without ^-replacement should not be given, because the rule teaching n-replacement (merely) discards (the form without n ) . That is why it is stated (in the Var.) ' Consideration ( of grammatically correct words) while giving name etc. (is proper) due to the existence of grammatically correct ( words ) ' ( Var. 2 on the pratyaharasutra l rlk ' p. 20 ).



AhniJca VI


" The absence of n-re placement in the name sarvanaman is due to its being a nipatana (ready-made form)." (Objection:) Now, on the same ground objection was raised that ^-replacement becomes due. ( R e p l y : ) This fault does not (exist). This is what (the Bhasyakara) states : in popular usage ^-replacement is not effected in this word and in grammar ( also ) it is used by the Teacher without effecting ^-replacement. In this case also it could be an incorrect (word). However, (it is said that) wherever it is used as a name in popular usage, it is without ^-replacement and (the same word ) is used (and thus shown to be correct) by the Teacher here (in grammar) also. It is like this.

Or, this is an argument of the purvapaksa ( as follows): How can this word (sarvanaman), being used for a different purpose and involved in name-named relationship, cancel ^-replacement when it is taught by the injunctive statement purvapadat samjnayam agah ( P . 8. 4.3 ) along with its complementary Items ?

The other (siddhantin replies) that the view of inference is (the same as saying that it is) a smrti. ^-replacement is inferred from that rule ( namely, purvapadat etc.), while absence of n-replacement is inferred from this (nipaiana, namely, sarvanaman). (Thus) n-replacement or absence of n-replacement, both of which exist by their own nature, are merely shown to be correct (by grammar).

*' What is meant by this ( word ) nipaiana ? " Since the prohibition of n-replacement is to be stated, you obtain prohibition by means of nipaiana. This is possible if a nipatana is equivalent to prohibition. How can a nipatana be equivalent? If it cancels an injunction (vidhi) ( a n d ) if the cancellation is without option, then only it can be equivalent to prohibition.

(In this connection) the other (siddhantin thinks that) whatever is the feature of a prohibition is also a feature of a nipatana. A prohibition ( consists in ) the inference of non-application ( of a rule ) by its very nature and not by the law of nail-upon-a-nail. " A nipatana is also of the same kind." ( Thus he ) introduces the question : " What is a prohibition ? "

5# 1 1]



" After having said something m a general way (it is said * n o ' with reference to a specific context)." ( F o r instance, the rule) rnnebhyo nip ( p. 4. 1. 5 ) (is a general statement ), while (the rule ) na satsuasradibhyah ( p . 4. I. 10 ) is a specific ( rule ). " There the intention of the Teacher that this does not take place is clearly understood." The intention of the Teacher, namely, that the Teacher has not intended to speak the word and therefore he has not accepted it is understood.

Why is it that when an injunction and a prohibition are stated ( for instance, in the sentences) ' Food should be given to Brahmins' and * To Kaundinya food should not be given ' injunction is not respected, prohibition is respected ? This is explained ( as follows:)

Tt is not possible to understand either the opposite or the combination. Option also is not possible, because (in that case) the application will be absolutely cancelled. Due to incompatibility (between two opposite actions) combination also is not possible to understand or to accept. Due to the understanding of prohibition option also is not possible. How is the prohibition (itself applied ) first ? It cancels that which is already applicable. If it is argued as such : by all, therefore, first cancellation should take place. And the act of feeding is accomplished in the case of other ( Brahmanas ).

Or words in each sentence are different (although formally they appear to be the same). Because of the same form (the words) are means of understanding the same meaning by (the person ) who hears the identical (words). (The listener) thinks that Kaundinya is (already) mentioned in the word Brahmana which is not different from the specific word (Brahmana).

The view of others (is as follows :) Scales and bones are not intended for eating ( by a person who wants to eat fish ). ( However,) he has to take them since it is unavoidable. Similarly, here also Kaundinya is not intended (by the speaker) but his inclusion (in the word Brahmana) is unavoidable. Therefore, in the injunctive statement Kaundinya is abandoned.


Mahabhaiiya-Dipika, Ahniku



[2.16 ( ! ) ]
Another view (is as follows:) Between direct cognition and inference direct cognition is more powerful. According to this viewpoint ( knowledge of Kaundinya ) is doubtful in the word Brahmana, while it is clear in the (statement of) prohibition. Optional application of the two operations that are (separately) prompted by direct cognition and inference is not correct. (Objection:) But the injunction applies to Kaundinya not because he is Kaundinya but because he is a Brahmin. And direct cognition by the word Brahmana is desirable. ( Reply :) This is not (correct). A general ( mention) does not contain manifestation of difference from a particular ( mention ). Therefore, option should not be applicable. Since Kaundinya is connected with the state of Brahmin, which is a different thing, therefore being understood in accordance with that (Brahmanya) ( h e ) is connected with the action of eating only through inference (in the injunctive statement). However, in the prohibitive (statement he is connected ) through direct reference. For this reason also option is not proper,



The view of others (is as follows) : This is a restriction itseef due to improbability. It is not probable that Kaundinya who is ( already ) fed because of Brahminhood is not fed because he is Kaundinya. Dae to this improbability naturally there will be prohibition.

Others think (in this way ) : It is this act of eating itself which is stated in the form of proceeding and retiring just as vyavasthitavibhasa 4 distributed option.'



The view of still some others (is this :) This sentence (occurs) in a certain context and is subordinate to a certain main sentence. And it is not correct to understand in the main sentence that meaning which is understood from the subordinate sentence, because (it is understood from) different words, just as the words gaura and khara are meaningless (in the compound gaurakhara).

The main injunction is this, ' Give food to Brahmins.' There is no offence when (Brahmins) are being fed. Moreover, the meaning ' cl^ss ' is also fulfilled.


Traits v. m

However, (when it is said) J * A Brahmin should xioi be killed,: \t aiust be ( regarded ) an offence when omitting some Brahmins, one Brahmin is killed. For (in that case) killing is accomplished, while non-killing Is recommended here. In this connection opposite behaviour leads to fault. Thus ( when it said) 'Brahmins should be fed,' (Brahmins) will be fm when some ( Brahmins) are fed. However, when prohibition of (giving) food (is recommended for Kaundinya ) one Kaundinya cannot be fed when some (other named) Kaundinya is fed. With this thought in mind it is said, ' injunctions are not cancelled by merely not carrying them but their prohibition cancels the operation \

It is argued (as follows) : Kaundinya who is understood by the word brahmana retires. However a man proceeds (to do something) by ike words of someone, whereas he does not (do so ) upon somebody else's words. Just as when one says (to him ) : * do ' and another (says): * do not do,' he obeys either, not because the remaining one has not conveyed ( him ) the meaning ( but because he can do only one at a time). The same applies here too. However, it has to be stated why he does not do both ways. And this has been explained in detail. In the same way in popular usage (somebody) after having said : 4 go % again says : * do not go.' He (the listener) understands that he behaves (in this manner) with some intention.

[ 2.20 ]
Or, speakers in this world commit mistakes. However, (Panini wearing ) a ring of darbha on his finger speaks thoughtfully and behaves fully properly. He is treated as father or mother and does not ( allow himself to) contradict sruti and smrti. Therefore non-application (of ?i-replacement) only is to be inferred.

As to (the view) that what is due is prohibited, for instance, (in the statement) * a tame cock should not be eaten'; (we say) there is no in« ference of that which is not due (in the parallel). When the inclination is caused by the object ( such as a tame cock) one does not hesitate (to eat it), knowing full well the visible profit as well as the result of killing (which appears) after death.

The same logic applies to nipatana* Therefore, (the says, " Nipdtana is also of the same kind." Bhasyakara)



JJmilca VI


(Objection:) *' But the absence of ^-replacement because of the nipatana and ^-replacement as per the application (of the rule, both) will be applicable." (The objector) thinks in this manner: A prohibition is of the nature of cancelling; therefore, cancelling (of a vidhi rule by a prohibition rule) is proper. But nipatana is of the form of a word; therefore, its own form is to be understood. In this way ^-replacement taught by another statement is cancelled thus ( by prohibition ). However, so far as this other word (sarvanaman) which is stated as a correct word is concerned, its { use ) is restricted as a technical term.

[ 2.24 ]
" Are there any other rules of this kind ? " [ Yes. ] In the rule iko yan aci { P. 6.1. 77 ) also the relation of original and modiiications does not exist, because it is accepted that ( words are ) permanent. Since the law of ashes from khadira wood does not exist, this (modified form is to be understood ) as a use of another word only. The word dadhy being understood as correct only in close juxtaposition may not cancel the word dadhi.

Here statement is made by genitive." This is the point: When the word sthana is used in connection with that which is not already effected then there is absolute non-occurrence. (For instance,) ik vowels disappear in close juxtaposition. When the mind grasps yan as the word, here also one thought is cancelled by another thought or (one thought) is produced. In both the views (namely, permanence of words and impermanence of words) it is proper (to assume) complete absence (of the original).

Here also (in the rule ptirvapadat) samjnayam (agah) where (the word ) in the genitive, ( namely, nah occurs), there is complete absence of n. However, in this case ( of the nipatana sarvanaman) option is ( applicable ) due to the absence of genitive. Why? Since it ( i . e . absence of nfeplacement) is not directly stated but indicated, option is wrongly applicable.

" Byan etc. will be made substitutes in place of sap:* ( Objection : ) If this is ( d o n e ) the state of being pit will be imposed (on u while




deriving the form ) karoii. ( Reply :) The it ( marker ) n ( attached ) to iyan indicates (that the state of pittva) is not (imposed on n). (Therefore the form) vidhyati (is accented on the initial syllable). The state of being sit however is present in the other ( substitute, e. g. sa, snam, etc.). Therefore the state of iiittva exists there.

(Objection:) And here in (the form) kurvati (the augment) num according to (the rule ) sapsyanor nityam { P. 7 . 1 . 81) is wrongly applicable. (Reply:) This difficulty does not arise, because syan is (separately) mentioned (in the rule sapsyanor nityam ). ( Objection : ) . How does (the rule) ato lopah ( P. 6. 4. 48 apply in the derivation of the form ) dhinutahl ( Reply :) Here also the fact that ( Panini) reads sa in (the rule) tudadihhyah sah ( P. 3. 1. 77 ) is an indication of its not being sarvadhatuka.

[ 2.29 ]
( Objection :) Even then snam which is applicable in place of sap will be wrongly added after that (sap) since it is an exception to replacement. ( Reply :) First of all he should be asked : Why is not it (the ending of the first person sing.) added at the beginning ( o f / ) ? If ( you ) think (that it is not added at the beginning ), because it is implied (as a substitute) by many other (substitutes), here also it is implied by syan etc. (that snam is a substitute).

Perhaps it may be argued : lasya is one genitive. But this is not correct. There are as many genitives as there are substitutes. Here (in the rule) phalipatinamimanijanam guk patinakidhatas ca (Unadi 1.18), since relation in general is understood (by the genitive) it is really the different sentences which convey meaning (separately). Therefore, the genitive having different meanings lays down different operations.

(Objection :) If this is so, snam must take the same place as that of (sap). (Reply:) No. Due to its being marked with m it will be after that (sap) like (the substitute ram taught by the rule) bhrasjo ropadhayo ram anyatarasyam ( P. 6. 4. 47 ). Why it is not so in (the word ) payaihsi has been already discussed (in the Bhasya) under the rule mid aco 'ntyat parah^P. 1.1.47).


Mahabhasya-Lipika, Ahnika VI


' • n u t that {sap) is mentioned in the nominative." (Objection:) But it is nut proper to say that there will be optional application (of sap and iyan) by amrntti (of sap). Since the rule that is (already) applicable cannot be abandoned, the relation between that which is cancelled and that which cancels is admitted even without thn anuvrtti. Therefore, the option < stands) If option is ( understood) also after anuvrtti, there will be contradiction with (the rule) which is already applicable as well as the fault of repetition f of sap) in this situation.

" Her- then (in the case of the rule) avtjayasarvanamnam etc. ( P . 5 3 71) (the option can be observed)." In this case there is no conflict regardin- plac- What then ? Since the meaning is the same (the objector) refers only to option. ( For instance) both uccaiska* and uccakaih will be formed.

•• Akac is introduced when ka is already certainly applicable." It is stated h-re that when a rule of Panini is being understood in its entirety, it neither states nor cancels anything. What then? It being understood implies absolute non-application ( of the other rule). If this is so (with regard t o ) akac there is no specific condition mentioned in the statement of akac. It is understood as taught in a general statement. The relation between general and specific is available only when option or conjunction (of the rules) is possible. It should be understood here that no specific condition is mentioned only general statement is made.

.. Nipatana also is of the same kind." " If nipatanas (also are of the same kind there will be difficulty in the derivation when sam is followed by ,ata\ » Havin" observed this only (a question) is asked regarding (the ule aparasparah) Jcriyasatatye (P. 6.1. 144 ,. Option is found even when a nioatana is (stated, for instance,) satatam and saMatam. Therefore, how to understand that only option and not prohibition (is understood) from nivatanai This being the case, the two-fold derivation is not available. - T n effort has to be done in this connection (to state) that nipatanas I (vidhi rules ) " ( This statement is) in accordance with the argument giv'n earlier. This ( Var.) has to be framed for the sake of the word Ma so that (both the forms) sahitam and samhitam are derived. This (word) should also be included there. Due to the law of avyavika swdatya is not derived. *. * *




" Prohibition with reference to a proper name and a subordinate word (should be s t a t e d ) " The Vakyakara (i.e. Kat.) alludes to difference (between homophonous words), because (according to him proper names and subordinate words) are different (from their homophonous counterparts which are not proper names and subordinate words). However, the Bhasyakdra (thinks) the same (word e.g. sarva) has attained to this state ( of being a proper name or subordinate ). Therefore, ( he ) says, sarhjnopasarjanibhutanani, " of (the words sarva, etc.) which have become proper names or subordinate."

" Give to Sarva." ( Objection :) Now (the name sarvanaman ) will not (apply to a proper name) owing to the maxim of primary and secondary ( meaning ), for instance, (in the expression) Agnisomau manavakau • the two youngsters, verily the two deities Agni and Soma \ ( Reply :) This is not (correct). This maxim (applies) in (the domain of) operations with reference to finished words. And the name (sarvanaman) applies to nominal stems. There this ( maxim ) does not apply. Therefore, it is said (by Pat.), " Vrddhi and 5-substitution pertain to (the form of) w o r d " (Mbh. on P. 8.3.82).

( Objection :) Even then the proper name samjna ( defined as) " a samjna is that which resides in one object" ( Var. 44 on P. 1. 4.1 ) cancels the earlier ( name, for instance, the names enumerated as) " samasa, krt, taddhita, avyaya, s a r v a n a m a n a n d a s a r v a l i n g a j a t i " ( Var* 4 1 o n P . 1 . 4 . 1 ) . Just as the name samjna cancels the name gunavacana and therefore (the form) daivadattyam is not derived, in the same manner, this ( name, sarvanaman) also would be cancelled. ( R e p l y : ) This (negation) is stated without adopting it (i. e. the name samjna).

" Give to atisarva ('one who has surpassed a l l ' ) . " This (word atisarva) is a pradi compound. In the case of bahuvrihi the negation is already stated (in P. 1. 1.29). How (is this definition of sarvanaman) a statement (applicable) to that which ends in that (i. e. sarvat etc.) ? (Because, it is stated by Kat.), '* the use ( of tadantavidhi can be made) jn the definition of sarvanaman and avyaya.11 Or in the rules teaching





[ 10. 8

operations anga will be qualified by sarvadi. Indeed this must be accepted for accounting (the form p:iramasarva in ) paramasarvasmai dehi. s give to one who is absolutely all.*

[ 3.5 ]
" H o w is that ( p r o h i b i t i o n ) to be m a d e ? " Whether ( i t should be m a d e ) in the ( r u l e defining t h i s ) name, or in the list? Whether in the ganapatka, or in the rule ? Or at both the places ?

" Exclusion from the list (should be stated ) . " If the prohibition is made in (the rule defining) the name, the grammatical operations (stated with reference ) to lists within (the sarvadi list) are not avoided. And if the prohibition is made with reference to the list, it is not applicable to prathama etc. Therefore, here itself ( i . e . in the definition rule) the list is qualified ( by the word denoting prohibition ). " Here only the name (is given) to those which are read (in the sarvadi list)." Sarva etc. which are not proper names and subordinate ( are called sarvanaman). (The others) are not at all included (in the sarvadi list). Those (sarva etc.) which are of this type (i. e. other than proper names and subordinate) are called sarvanaman. Therefore ( a word like sarva which is a proper name), although similar in form, does not receive the name (sarvanaman), because it is not included in the list. For instance, gotra etc. in the sense of contempt or repetition ( are mentioned in P. 8. 1. 27). In other meanings these words are not ( regarded as belonging to) the gotradi (list).

" Is (this prohibition) in a general way ? " Is the prohibition ( applicable to ) that operation whichever (is stated with reference) to all ( i . e. sarva etc. by stating) sarvanamnah ' of sarvanaman ' ? " (The prohibition covers ) also specific (rules )." It means ( both specific) and general (rules are covered by prohibition). The operations characterized by certain groups which are (included) in the list, are (operations meant for ) specific sarvadi words, for instance, dataradibhyah ' after datara e t c ' ( P. 7. 1. 25 ), tyadadinam « in place of iyad, etc/ ( P. 7. 2. 102). As to the ( reference ) yusmadasmadbhyam 'after yusmad and asmad' (P. 7. 1.27), it is in a general way. Although the sequence of these two ( words is found ) in the ganapatha, yet it is not accepted here ( as a group within sarvadi). It is an altogether different group.

U # 9]




[ 3-8 T
" The purpose (is clear) in ad substitution ( with reference) to datara ac/-substitution does not take place after atikatara, because (katara) is a subordinate (member of the compound alikatara). However, k is admitted that here (in the word ) paramakatarat (ad substitution ) should take place.
e tc. f


* A split of rule will be made." Sarvadi etc. are expected (in * the rule) asamjnayam, since it is an unconnected statement. And there (in the rule asamjnayam) the group enumerated in order is expected, not the name (sarvanaman).

" The rule anupasarjanat is rejected." (Objection :) But this rule has to be stated in order ( t o extend ) the tadantavidhi to the main member so that (the words) kaumbhakareya, kaurucareija will be (derived). This (an) cannot be regarded as a mention of krt, because an is also a taddhita (suffix).

( Reply :) This is not the purpose ( of the rule anupasarjanat). It has been already stated " when a suffix is mentioned, except in case of a feminine suffix (it represents the element beginning with that to which the suffix is added )." In the rule stribhyo dhak ( P. 4 . 1 . 120) the statement (of tadadiniijama) is not intended. (Objection:) If this is so, there will be overapplication ( of tadadiniyama also) in the domain of (words where a word ending in feminine is ) subordidate. ( R e p l y : ) with reference to this, a reply has been already given that (tadadiniyama ) does not ( apply) when (the word ending in a feminine suffix is) subordinate.*

(The rule) being thus rejected is understood as a paribhasa % rule of interpretation ' with (the letters) a and at as (its ) marks. The ( word ) anupasarjane is present (in the rule) where a and at are stated. {anupasarjane) means * when the meaning is not subordinate.' (The word anupasarjana is the result of) either the elision of the genitive by the rule supafn suluk etc. ( P. 7. 1. 39 ) or a reference to the nominal base, while the case ending in accordance with the meaning is assumed.


Mahabhasya-Dipika, Ihnika



L 3.13 ]
(Objection :) If it is so, (the forms ) atiyusmat and atyasmat are act derived/ 1 {In the expression ) atiyusmad anaya ( meaning) * Bring from those who have surpassed you f at substitution by (the rule ) pancamya at { ?, 7, 1. 31) does not apply. ( Reply:) '« This is a reference by fusion." anupasarjane (is present in the rule) where a and at are stated after a. (This Is conveyed by ) anupasarjana a a at. According to them ( who accept this meaning of anupasarjanat) at substitution with regard to tyad etc. ( P . 7. 1. 102) remains uncovered (by the Pari. anupasarjanat), because ( a is) not taught after a stem ending in a. Further, bow will (the form priyadva result out of) priyau dvau yasya i one to whom two are dear' ? For here a cannot be (substituted) after a. Moreover, in the form atitad the s-substitution is (applicable) unprohibited. Further, atisarvaya is also not covered,

Or whatever is stated in the angadhikara ( P. 6. 4. 1 ) is ( effected ) with reference to that case-ending which is connected with that which is directly mentioned." In a rule teaching an operation, when a reference is made in the same case-ending or ( a different case-ending), the relation of qualifier and qualified with the words whose forms are mentioned (in the rules) is understood. ( I n this situation ) if (the word ) angasya is treated as the main ( i . e . qualified) and datara etc. as subordinate (qualifying), paramahatarat is derived. However, (^-substitution) wrongly applies to Qtikataram*

If daiara etc. are (treated as) main, then due to the absence of tadantamdhi atikataram is derived. However, paramahatarat is not derived. Similarly atiyusmat (and ) atyasmat ( are not derived ).

If it is understood that angasya is a genitive denoting qualifier ( a n d ) tyadadinam is a genitive denoting place, (the rule tyadadinam ah would mean) * a ( replaces) tyad etc. which are parts of anga when they are immediately followed by a case-ending/ Therefore, paramasah is derived. (However, a-substitution) wrongly applies also here i. e. to

If It is possible to assume tadantavidhi here as belonging to this section - sarvanaman * names for a l l ' is the section and sarvadmi is the

12.13 ]



qualifier - then this meaning results: a sarvanaman is that which ends in sarva etc. And that which ends in a proper name or upasarjana is not a sarvanaman. Therefore (in that case) tadantavidhi is not (possible). So (5??xaj-substitution) will take place in paramasarvasmai not in atisarvaya. However, it is not possible to avoid the operations applicable to datara etc., because they do not belong to the section {sarvanaman). But when tadantavidhi is not (restricted) to the section, (as it is stated with reference to the name sarvanaman in the Var\) ' t h e purpose of tadantavidhi is clear in the definitions of sarvanamanf and avyayaS In both these cases there is application and non-application (of grammatical operations).

Therefore, the Bhasyakara adopts another way : " It is effected with reference to that case-ending which is connected with the directly mentioned (stem ).' grhyamananam means f of those which|are mentioned \ ( grhya* manavibhakti means) * the case-ending which is connected with the words mentioned (in the rules).' The connection characterised by injunction is not adopted ( here ). What then (is adopted ) ? ( The connection) which is a cause of indicating meaning. When a case-ending indicating the meaning which belongs to tyad etc. follows (a replaces the final) of the stem (anga). In this way (the rule add dataradibhydh etc. would mean) c when su and am which indicate the meaning of datara etc. follow (an ahga> they are replaced by add). Similarly, (the rule nasihyoh smatsminau would mean) ' when nasi and hi (indicating the meaning) of a pronoun (ending in a% follow an anga, they are replaced by sniat and smin)S However, (^-substitution) is desired in (the forms) atiyusmad and atyasmad.

•' If this is so, in paramapanca (the Zu/c-operation by P. 7 . 1 . 22 will not be effected )." ( This is ) the idea. The connection of the case-ending is characterised by injunction. Or by assuming that in (the word ) tyada* dinam (the word's) own form (is mentioned, the rule tyadadinam ah is interpreted as meaning when) that case-ending which is connected with the meaning which belongs to the form of tyad etc. (follows, a replaces tyad e t c ' ) . And this case-ending ( namely, jas added after paramapancan ) is not connected with (the meaning of pancan io its) own form.

[ 3.20 ]

" This difficulty does not arise. This compound has sas as the main ( member)." ( Here) the connection is due to indicating the meaning. ( By


Mahahhasya-Dipika% Ahnika


[ 12.14

the word) iyaJadmam the form ( of the word ) is not adopted. A word is just an Indicator ( mark ) of the meaning. When (it is interpreted thus viz.) when Its case-ending which is connected with that which is seen as the meaning of tyad etc. follows, (the stem final is replaced by a ) . Or when the operation is understood in connection with the main (meaning), it can be said that this meaning belongs to that ( i . e . the word paman). For instance, when horses (etc. also ) are going (it is said ) * The king is going/

[ 3.21 ]
However, this objection ( can be raised ) in this connection : (In the expression) pancakah sakunayah ' five birds' the elision by sadbhyo luk { P. 7. 1.22) wrongly applies, because (the suffix) lean is svarthika * conveying (the stem's) own meaning/

" This then (is the form ) priyasakthna ( where anan does not apply ). tf This ( form priyasakthna) will not be derived in the absence of tadantavidhi, because the case-ending is not connected with the meaning of sakthi * shaft of cart.'

Whatever is stated when ( a word in ) the locative is mentioned, is effected when the case-ending is connected with the a?iga" { The substitute) anan h stated ( by mentioning the locative :) iko'ci vibhaktau ' after a stem ending in ik vowel when a case-ending beginning with a vowel follows.' tat prakfiatih'hakkiu is a reference to meaning.

In fact, (there are) three ways (of interpreting the rule asthidadhi... [ P . 7.1. 74] : [ I ] (anan is substituted in place) of asthi etc. which are immediately preceding the case-ending which is connected with the afiga, [ 2 ] or tadantavidhi of X\\&'ahga with asthi etc. (is understood ), [ 3 ] asthi etc. are qualified and ahga is qualifier.

" If this is so, (in the forms) atitad (etc. a-replacement wrongly applies). This ( i . e. a-replacement) is also stated ( by mentioning a word ) In the locative. Therefore, application ( of a-replacement is possible) just as in case of asthi etc. ( application of anan is possible),

13.16 ]




Moreover, it is to be stated.*'

( W h a t is to be stated Is) not


manavibhaktau etc. Then what (is it) ? Saptaminirdiste yad ucyate Qtc.

'* Here the ablative (is mentioned in the rule add) dataradibhyah." ( Objection :) Now, the case-ending ( of afigasya ) becomes identical with (that of atah) by vihhakiiviparinama ' modification of case-ending * in (the rule ) ato hhisa ais (P. 7. 1. 9 ). Or it is understood ( as follows): bhis^ which is connected with anga and which follows short a (is replaced by a i s ) . In this way it is possible to connect (the word anga) with another case in. case of ad-substitution, or (it is possible to interpret that when) the caseending (su or am ) which is connected with anga and follows katara etc. ( it is replaced by ad). How is it said, " (the ablative cannot be connected with anything) other than with the suffix which is stated " ? Moreover, when (the ablative) qualifies (the word ) vihita, the form paramakatarat is not derived. Further this ( pair of forms, namely ) paramasat and atiyusmad is not derived.

( R e p l y : ) With reference to this the intention (of the Bhasyakara can be stated as) this : The modification ( of case-ending) or any other assumption (is to be made ) when the desired ( form ) is not achieved by interpreting (the rule) as it is heard. And this (desired meaning) is summed up here only. Therefore, it is accepted as it is heard. In this wording ( of the rule) there is no possibility of tadantavidhi. This is shown by the ablative (datarddibhyah), * How (to derive) paramakatarat ? ' (the reply is :) It is a doubtful ( form ) as it is not mentioned in the Bhasya. (If you ask :) ' What should be (the form ) ? * (the reply is:) According to this interpretation (the form ) should be paramakataram. How (to derive ) paramasat ? ' ( The reply is :) (the word ) sadbhyah is not an ablative (form). What (is it) then? Dative. For instance*
( i n the r u l e ) ayudhajlvibhyas chah parvate ( P . 4. 3. 9 1 ) ayudhajivibhyah

is a dative form). (Similarly, the rule sadbhyo luk means that) the suffixes jas and- sas which are meant for sas are dropped. * How (to derive) atiyusmad V (The reply i s : ) Just as ato bhisa ais (is interpreted to mean that) bhis which is of anga (is replaced by ais, in the same manner the ablative ending of anga ) which follows yasmad and asmad (is replaced by at), ,Or (the form is. derived ) by ( adopting) tadantavidhi or by ( adopting) partand-whole relationship.


MahabhSaya-Diptka, Ahnika VI

[ 13.17

However, those whose view is that paramakaiarat should be (the correct form ), think in this way : It is stated (in the Bhasya on the) rule ?iyap pratipadikat (P. 4,1.1) that su etc. are stated in the meaning singularity etc. and thus the names ariga, bha, pada etc. will be (given J to the word conveying that (singularity etc.) and therefore, being taught even in the absence of the mention of (the word) pi-aiipadika* they (su etc.) are stated as conveying that { singularity ), In the same manner here also (it can be said that the case-endings ) are being stated in the seose of 4^ara eic » a n ^ hence are added after them. This statement is understood thus as based on meaning,' not on words. Therefore, (the correct form) paramakatarai will be (derived). Because here also (the suffixes) su and am are added in the sense of daiara etc. only.

( Objection :) " If this is so (in the form) atisah a-substitution does not apply." Just as (it does not apply in the form) atitad. ( R e p l y : ) 4 * This compound has tyad etc. as the main (members)." atisah means iobhanah sah * beautiful he.f Moreover, whenever (forms such as) asah * non-he,' asat * non-six' akatarat s not-either * are ram-compounds with the final member as the main member, then also (the relevant grammatical operations) will take place.

* Or rather this is not giving of a name," sarvanamani is not a name of sarva etc. What then (is it) ? '« A qualifier of the list." It is the qualifier of the enumeration. " (sarva etc. are those) which are names for all." Those are mentioned in the list, not others. Therefore, these (words sarva etc, when they are) names and upasarjana conveying a specific ( meaning) will not get (the name sarvanaman). These ( words Barva etc.) are sarvanaman * name for a l l ' in their respective domains just as genus is all-pervading due to its pervading the respective domain. In this way the grammatical operation based on the list, such as ad4 ^aradjM^/ah . . . ( P . 7 . 1 . 25), tyadadinam .. ( P . 7.2.102) applies to those which are names for all. It is not applied here i.e. in (the forms) atikataram aad 'atitad. For here (in these two forms), one who has surpassed is denoted.


[ 3.32 ]

If this is so, (the operation ) based on the name (sarvanaman does not apply, because the name is not taught)." Because the (power of the)

14,23 ]



word sarvanaman is exhausted in qualifying the list. This being the case it can be said that the inclusion of sarva, viiva, ubha, ubhaya etc. (in sarvadi list) is redundant. *• There (the word ) will be understood as significant? that which is a name for all (is a sarvanaman ).<f smai etc. will be added after such a stem.

[ 3.33 ]
And the word sarva denotes totality. It refers to number, kind and a whole formed of parts. (For instance, in the sentence) sarve manusyah agatah ' all people came * ( sarve means ) as many as were written, sixteen, seventeen etc. Therefore, they speak in this way: All foods were eatens without leaving anything over. (The example of the use of sarva) in the icnse of totality of kinds (is as follows :) ( He ) eats all food. {This means he) eats all kinds of food. ( The example'of the use of sarva) in the sense of that which is composed of parts (is as follows :) All rice whatever was served in the dish was eaten.

[ 3.34 ]
" If this is so, (grammatical operations caused by sarvanaman) wrongly apply also to these ( words ) - sakalam krtsnam jagat." (Of these words) to that which ends in short a ( smai etc. are applicable ), elsewhere akac etc. (are applicable). For these (words) also are sarvanamana * names for all' in their respective domains. «' Of these words also " whatever is the word conveying the respective domain, (for instance), sarvah odanah * all rice, • ekah odanah * some rice,' " with reference to that domain grammatical operation caused by sarvanaman wrongly applies.' (In the sentences ) sarvasminn odane * in all the rice • ekasminn odane *in some rice * (the form) odanasmin will be wrongly derived.

C 3.35 ]
This seems to be an impropriety of the Bhasyak5ra who is well-versed in logic. How ? The word sarva conveying a general meaning is used in both the contexts, that of rice and that without rice. In the same manner the word odana also is used in a context of what is all (sarva) and what is not all (asarva)* As for the resultant particular ( meaning ) it is the meaning of the sentence, not of the two words. That is why it is stated " Whatever is the addition ( of meaning understood) here, is the meaning of the sentence.'* ( Var. 11 on P. 2. 3 . 4 6 ) . Therefore, how is it concluded that the word odana has a reference to all ?


Mahabha&ja-DipikS, Ahnilca VI



• •

Though (both the words, sarvah and odanah) refer to the same object, even then a certain word resorts to a certain part of meaning. For instance, when the word nila (is mentioned ) alongwith utpala* it ( resorts to ) the part ' blue, * whereas the word utpala ( resorts to ) the part ' lotus ' ( of the meaning ). The object of these two ( words ) is either composed of both or It is devoid of both. In the same manner the word sarva has resorted to the part * entirety ' while the other word (i. e. odana) has resorted to the part ' rice/ For this reason also (the word odana as) referring to that (i. e. all) cannot be justified.

With regard to this (there is) this view : When something is connected with all objects, the use of the words is in accordance with how one understands it. Therefore the word nila is used ( by the speaker) in accordance with his own understanding as embracing all objects. Thus (the listener) finds the blue quality (associated) with different kinds of objects. Just as the word nila refers to ( any ) substance possessed of blue colour, in the same manner the word utpala also is connected with all objects (i. e. all lotuses with all different colours). Both are connected with that in which ihey reside. Why should one of the two undergo grammatical operations but not the other?

" Alright, then both are achieved by this." Sarva etc. are delimited by the property of sarvanamatva i being name for all.' Only those (sarva etc.) get the name sarvanaman. " But how are both achieved by one effort ? " Words being tied to their own meaning are dependent (upon meaning). Being taken up by one function (of conveying a particular meaning) they are unable to refer to another meaning discarding that ( function ), due to the absence of another function. In that case neither the constituents, nor the collection can be explained as possessed of both the meanings. For instance, the effort for pushing is not capable of sustaining (a thing). Another effort is .indeed made (to sustain it).

" T h i s is a mention by elcasesa." ( O b j e c t i o n : ) But h o w can ( i t ) be ekasem ? F o r , it ( e k a s e s a ) is formed when objects are many a n d

16.5 ]



(consequently there is) an occasion for many (references), for instance* vrksas ca vrksas ca vrksas ca vrksah: ' a tree and a tree and a tree ( are reduced to the ekasesa) trees.' But here neither each of the words sarva, pisva etc. nor the meaning expressed by the word sarvadi are many. The second (i. e. repeated) word sarvadi serves this much purpose : sarvadis which are names for all, those sarvadis are sarvanamans. The others are ( not sarvadis and therefore they are ) not sarvanamans. Moreover, by the ekasesa - words one meaning belonging to one time and leading to one result is referred to, whereas here the meaning concerned refers to different times and leads to different results: sarvadis are those which are names for ail and those sarvadis which are so characterised are called sarvanamans. Therefore it is not possible to maintain that ekasesa is mentioned (here )•

This ( construction) should not be looked upon as an ekasesa in the technical sense. It is, however, called ekasesa, because it is similar to ekasesa. But of what kind is it ? (Tt is) a device called tantra. For instance, a word, being one and the same is understood as having many forms because of many speakers. Or just as one and the same (sound) being produced by the pipe of a wind-instrument is received (in the form of) many echoes due to the peculiarity of the place. In reality, however, it is not multiple. Indeed it is not multiple because when ( a person ) imitates ( a word), they say, * he uses the same word ghata? Just as by means of some device one and the same (word) assumes many forms, in the same manner why are not many ( words) understood as having one form ? Therefore, since the Teacher found the device of a tantra word he mentioned ( one word ) having many forms.

There are only two ways of uttering words: ( uttering) in sequence or (uttering) simultaneously, (If they are uttered) in sequence, their forms are separate, not so in a collective (utterance). Or (the rule) sarvadini sarvanamani is a collective (utterance) mentioned through (the device o f ) ' tantra. If the two parts ( are understood as collective utterances) it is possible to understand a hundred sentences (from such collective utterances).



Or repetition is indeed word's own nature just as (it is a nature ) of actions. (In the sentence) paticakrtvo bhunkte: * ( H e ) eats five times* the number five does not ( refer ) to the act of eating. For, there are not: five acts of eating. What then? The word (root) bhuj in (the f o r m ) .





[ 16,5

bhunkte conveys repeated action. When it is the repetition of that action which is counted (so as to ascertain ) how many times it is repeated, then the word parka conveying the repetition connected with the action being counted generates the suffix ( krtvas). Just as one would read reduplication of that ( root) which conveys repetition ( of action ), in the same manner the word sarvanaman being repeated is adopted by the Teacher.

[ 3.43 ]
Although only one (word) is uttered, the user uses it separately. For instance, among the seventeen samidhems the first and the last rks having the power of repetition are read only once in the Veda. However, the user employs them separately ( as it is said ), " ( He) recites the first one three times and the last one three times/'

[ 3.44 ]
Or we can assume two powers in one and the same (word). For instance, (a person) who desires to see and to burn a wound in darkness prompts the lamp by resorting to two powers. In the same way the word sarvanaman invested with two powers is read by the Teacher. For instance, {in the rk) idam visnur vicakrame (RV L 22.17 ) one and the same word visnu being possessed of many powers conveys self, Narayana and the wooden ring at the top of the sacrificial post respectively in metaphysical, mythological and ritual contexts. In the same manner there is a difference in avagraha in the case of the sentence vrko masakrt (RV. I. 105. 18 ). The word masa used in the sense of moon is separated by avagraha e. g. vrko vidsa s krt. This is the view about repetition.

C 3.45 ]
" ( both samjna and upasarjana) abide in specific (meaning ).*• And the word sarva, when seen in the sentence ( i . e . constituent analysis), does not abide in specific meaning. (Therefore), it is said, niskrantah sarvasmat < one who has gone out from all. 1 In this way the state of being sarvanaman cannot be admitted in the case of these words (namely, idam, tad, etc.) If it is understood that each and every word here (in sarvadi list) is expressive of all, none will be a sarvanaman* Because the word idam refers to that which is being perceived; not to that which is out of sight. Similarly, the word tat refers to that which is out of sight; not to that which is being perceived. Therefore, it should be understood that they are sarvanamans in their respective domains, just as samamjavisesas (particularized generic features ) $re allpervading in their respective domains.

17.12 ]



[ 3.46 ]
(The word sarva) irrespective of the specific ( objects ) odana etc., proceeding by means of the cause of use, namely totality of number, object and kind, abides in generality ( i . e . denotes universality). However, when odana etc. are used, from the knowledge of the specific meaning it is understood that it (the specific meaning ) is the meaning of the sentence. Or being used along with the word odana, without abandoning its own cause of use, it does not express the meaning in the form of odana, because words convey a particular part of meaning ( of the sentence). It expresses ( meaning) in the form of totality. Therefore, the state of being * a name for all • does not cease to be (there ). .

[ 3.47 ]
However, when the word sarva is taken as a name of somebody, then, abandoning its own cause of use, and conveying only (the word's ) own form, it experiences the state of a proper name. This is giving up of its own cause of use. This is abiding in a specific meaning. Therefore it is stated (in the Mbh. on the rule) kasyet ( P. 4. 2. 25) that in an injunctive statement, in both the alternatives (the form) should be kaya. But if the name is caused by the property, (the form) should be kasmai.

In the same manner the upasarjana word also, either abandoning its own meaning or not abandoning its own meaning, becomes a qualifier of (the word) atikranta and conveyes the meaning of (the word) atikranta. And (the word ) atikranta does not stand with its own cause of use. Therefore, abiding in a specific meaning (is the character also of upasarjana). However in the constituent analysis it does not give up its own cause of use. Therefore, the name (sarvanaman) is given.

C 3.49 ]
[ " The name is that than which ( nothing) is shorter." ] The use of words is very brief and very subtle. The name should be made still briefer* This is the way people use ( words). Lengthiness in a proper name (leads} to the inference that in it (name), the cause is intended and it is used because of that cause. For instance, the word Krsva is used in the sense of Vasudeva because of some cause. It is (his) name. Similarly, the name Sahasrabahu is caused by (the presence of) a thousand arms. In the same manner (sarvanaman) is the name of sarva etc. caused by the state of being a name for all.


MahabhSsya-ZHpikS, Ihnika


[ i?. 13

An objection is raised here KEven then the grammatical operation such as sniai etc. does not apply to (sarva etc. ) which have become names or upasarjanas, for instance, sarvaya and atisarvaya. Why ? Because the name sarvanaman is given to that which is a name for all. However, that (operation) which is an operation applicable to an inner group, for instance, ad$ dataradibhyah, tyadadtnam etc. applies generally (to all)." The refutation (of the objection) is stated (as follows:) That may be the case. The rejection in the case of a-substitution and ^-substitution is the same as already ( stated ) earlier. However, the honourable (teacher) says: it is not understood that those which are Dames for all, not others ( are sarvanamans). What then ? An argument is made like this : Why is a lengthy name (sarvanaman) given to these ( words ) viiva, ubha, ubhaya etc. ? Why do they become a cause of ( the use) of a lengthy term ? It is guessed that since they are names for all, therefore, they become a cause of (the use of) a long term. Just as ( Narayana and Arjuna) become causes ( of the use of the names Krsna and Sahasrabahu) when they possess darkness and a thousand arms^espectively ), in the same manner these ( words sarva etc.) also ( become cause of the use of a long term). This being the case those which are possessed of that quality (namely, being a name for all) are included, those which are not possessed of that quality are not at all included. Katara etc. and tyad etc. possessed of (this ) quality are included (in the list). Thus all grammatical operations caused by the name {sarvanaman) as well as by the lists are stated with reference to (sarva etc. which are) not names or upasarjanas.

" Now, what is the purpose of (the name) sarvanaman (being given ) to ubha ? " The meaning of the sentence is completed by (saying) c akac is the purpose.' However, the Bhasyakara does not state the meaning of the sentence in this manner. ( He ) only introduces the purpose of reading (the word ubha) in the (sarvadi) list. " There is an absence of other (operations )9 because (ubha) is subject ( only) to tap and dual endings. This word ubha is subject to dual endings and tap." For, the word ubha is used by resorting to difference as the main meaning, for instance ubhau sammukhau hastau bhavatah 'two hands are facing each other.' Just as the word dvi conveys only difference (in the same manner the word ubha




conveys difference). Therefore, it is subject to dual endings and tap. It is never (a fact that il is not) subject to (dual endings and ) tap. It is subject to tap (in derivations ) other than (those containing) dual endings. (The Bhasyakara ) thinks that (ubha ) is subject to (suffixes } of the kind of tap. What are (the suffixes) of the kind of tap ? Svarthikas: * those which convey ( stem's ) own meaning.*

In this connection other grammarians read (as follows :) « ubhaya ( is substituted in the place) of ubha when ( 1) a suffix other than a dual ending and tap or ( 2 ) the final constituent of a compound follows. ' According to others a dual ending and tap (are added ) after ubhaya and tayap is dropped. In their view by tap tap etc. are understood just as ( by the word ardharca ardharca etc. are understood in the rule) ardharcah punisi ca (P. 2.4.31). The reading (of the statement), according to some other (grammarians), is this: ' A dual ending and tap are not (added after ubhaya)'. Yet some others read (as follows ) : ' ubhaya (is substituted in the place) of ubha except when a dual ending ( follows ) . '

Here in grammar this is proved logically. How? with reference to these (expressions ) ubhayo manih ' a two-fold bead,' ubhaye devamanusyah * both the gods and men,' in the case of the bead as a whole with the difference of parts continued in it (as well as) in the case of the collection devamanusyah containing the ultimate difference of parts, the word ubha cannot be used. In integration also since it is said samkhyaviseso vyaktabhidhanam ( Var. 2. 1. 1. 2, Mbh. I. 362. 13 ) • particular number, clear mention ' ( etc. are understood in a sentence),' the upasarjana has abandoned difference. After the difference has retired, the word ubha also should retire. Moreover, the words eka etc. are not its synonyms. Only that word which is somewhat similar in the phonetic form or (similar ) in some meaning, should be used (in the place of ublia). Therefore, the word ubhaya only is used. In fact, the meaning understood from the phonetic form of these two words ubha and ubhaya is different, but the essence of the meaning of both is identical. (Wtien it is said ) ubhau (anaya) or ubhayam anaya * bring both,' the same' two (objects) are brought.

( Objection :) If this is so, the word dri also has to be (regarded) as incorrectly ( used) in integration since (there) the difference is abandoned.


Mahabham-Dipika, Ahnika VI


v Reply :) Although difference is abandoned, since that which is the cause of the use of the word dvi itself is not abandoned, it is used. Therefore, in the domain of vrtti the relation of identity - * that one ( is) this one ' - exists between the whole and its parts. (The meaning is understood as) 'This group, namely, these two.' Therefore, the word dvaya is not used.

(If you ask,) * Why is (this distinction maintained) in the case of the word ubha1\ (our answer is:) just for this purpose the Teacher mentions the word nityam * always * in the rule ubhad udatto nityam ( P . 5. 2. 44 ). It is settled that this ( word ) always denotes the whole through this form {ubhaya), and not through the form ubha. Here in grammar the word ubha being an upasarjana has abandoned difference. As such it is not able to convey the whole composed of parts, and so the word ubhaya is used.

'• But when this word ubha is subject to dual endings or tap (it is used ).*' Due to the nature of word (itself) or due to the statement (to that effect) or due to logic (it is used in the above mentioned context). " Now, what ( happens) to it elsewhere ?>s " Elsewhere ubhaya (is used)." (So far) it was shown how this word ubha is subject to dual endings and tap* Now (the Bhasyakara) asks a question with reference to the statement "the reading is for the sake of akac " : •* What will happen if akac will not ( be added) here ? " '• Ka will be wrongly applicable " because akac, which is an exception {to ka) is absent. ( The objector) not seeing any difference in form and accent ( between the form with akac and the one with ka) asks : '* Now, what is the difference between ka and akac ? " Avagraha is not respected* because it is stated ( by the Bhasyakara), * the rule (of grammar) need not follow the authors of the Padapatha.'

C 4J ]
" It is already stated that this word ubha is restricted to the domain of dual endings and tap." It has been stated (by the Bhasyakara) while discussing (the Var. no. 8 ) : '« the absence of other ( operations) due to (its being) subject to dual endings and tap," that this word ubha is restricted to the domain of dual endings and tap; therefore the word ubha should be (regarded as) correct when the dual ending is heard (immediately) after (it), not when it is dropped nor when it is intervened. (Thus) this (word ubha) is subjected to the domain characterised by subsequent (ending). When it is stated that if akac will be (added ), this aka, having fallen in the midst of it (i. e. the word ubha ) because of the indication (namely prak




teh)> <ioes not become a cause of a different word, then its being subject to dual eiullng is not contradicted.

" However, when ka is (added) it will not be followed by a dual ending.*' Since the earlier (dual ending) is dropp?d and the other one is intervened, the undesired contingency of having to regard the word iibha as incorrect, or of its non-use arises. Or the use of the word ubhaya wrongly applies, '• There its being followed by a dual ending should be stated." An extended application has to be made. It should be stated that in the domain of ka (it is) treated as subjected to a dual ending or the word ubha is correct before ka followed by a dual ending.

" Then just as it is not followed by a dual ending when ka is ( added), in the same manner also when ap is (added, it is not followed by a dual ending)." This is not reasonable, because ap is mentioned. This is how it is stated : •* Due to its being subject to dual ending and tap" When (its) correctness, when (it is) subject to ap, is thus stated, it is improper to say, u Also when ap (is added), it will not be followed by a dual ending.*9

With regard to this ( some grammarians) describe (the situation as follows): In the statement ubhasyobhayo^dvivacane there is no mention of tap, therefore this objection is raised in this view. The honourable ones however say, * You (i. e. the Varttikaka ra) have shown that the word ubha is correct when a dual ending is heard immediately following it, (and ) tap also is mentioned. However, you have not achieved your purpose by this much. It has also to be stated that this tap is followed by dual ending, not by singular or plural ending. ( Thus) the Bhasyakara, saying, " Just as (it is not followed by dual ending) when ka is (added ) so also (it will not be followed by dual ending) when ap is (added ),'• has staled the same point, namely, that it should be stated that also when ap is added (ubha) is followed by a dual ending.

" ( Even) without a statement (it is) understood that it is followed (by a dual ending) when ap (is added y diivacanaparatal being followed by a dual ending' should not be understood like this : that is subject to dvivacana after which dvivacana is added. How then (should it be understood)?


Mahabhasya-Dipika, Ahnika VI


Being followed by a dual ending is characterised by meaning. (The word ubha is) incorrect where (n ) conveys another meaning, since there the difference retires. ( On the other hand ).when ap (is added ) and the relation between the main and the subordinate is reversed because of the svarthika nature of tap etc., (the word ubha) with the difference not abandoned is correct even when followed by ha,

( Objection :) If this is so, how (is the word ubha regarded as correct) here in (the forms) ubhabhyam and ubhaydhl (Reply:) These two also do not convey an outside meaning, therefore (the use of) the word ubha has to be (regarded) as correct. (Objection:) Then the (wrong forms) *ubhatah and *iibhatra are derived. If you think that tral conveys outside power just as (in the word ) adhistri the compound (is possible) only when the adhikarawti * substratum' is outside stri * woman,* (this is not correct, because) in that case correctness (of ubha) cannot be (justified) also here in (the forms) ubhabhyam and ubhayoh. ( Reply :) Therefore, this is the reply : Because tasii etc. are indeclinab?es. ( An avyaya) is similar (in form ) in the three genders and before all the case-endings. When tasil etc. which convey only the meaning of case-ending ( follow ), all difference is abandoned. Therefore in this case (the use of ubha ) is incorrect. The word ubhaya is used according to (the Var. no. 9 ) ubhayo 'nyatra * ubhaya elsewhere.'

( Objection :) If this is so, (the mention of ubha) has to be ( regarded as) incorrect in (the expression) ubhayo mariibi because (the word ubha ) conveying another meaning, namely, the whole (avaijavin), abandons difference. (Reply:) This fault does not arise. This itself is the first original form of this (word ubhaya). Having obtained a form here this (word ubhaya) is used elsewhere even when the word ubha retires, if the word ubha is engaged in conveying another meaning.

Option with reference to (words listed in) dvidaridyadi (gana) is stated by means of nipatana, (for instance,) ubhanjali (and) ubhayanjali, ubhapanih ( and ) ubhayapanih etc. All this correctness and incorrectness in case of the word ubha was stated by means of ka and akac. However, it should not be read (in the sarvadi list) fov.akac but for some other ( purpose). ( For instance) this ( form ) ubhabhyam hetubhyam * due to both the reasons," as well as (the form) ubhayoh hetvoh * for both the reasons' are obtained ( according to the rule) sarvanamnas trtiya ca ( P. 2. 3. 27).


Translation [4.15] • ..;


Now ( what is the purpose ) of ( calling) bhavat ( a sarvanaman) ? •• Why is this (word) also included (in the sarvadi list)? The question (is asked ), because smai etc. are not possible after ( a stem ) ending in a consoaant. •' ( The purpose ) of (the inclusion of) bhavat (in the sarvadi list is ) akac, sesa and alvar §esa is not characterised by sarvanaman. What then ? ( It is) dependent upon the list tyxdadlni. 4< This is just an example — " "Instrumental also is desired." The Bhasyakara also (gives) just an example. There are also other grammatical operations dependent upon the list. (For instance, bhavat is called sarvanaman) so that (the incorrect form fbhavatya should not be ( derived ).


vibhasa diksamase bahuvrlhau / P. 1 . 1 . 28
" In a bahuvrihi compound containing words expressing directions, (sarva etc. are) optionally (called sarvanaman)."

" What is the purpose of the mention of (the word) dis ? " Here the intention (of the Bhasyakara) is to say whatever has been already stated in the vrttisutra in connection with the mention of (the word) bahuvrihi, (The Bhasyakara) introduces the whole ( discussion in ) the book vrttisutra io such a way that the impression of ( a homogeneous) work is not blurred. Moreover, the intention (of the Bhasyakara also) seems to be to ask the question. ( H e ) thinks in this way: The arrangement about application and prohibition (of sarvanamasamjna) should be made and option (should be applicable) to other cases. Otherwise, the prohibition will be redundant, since the optional statement has already provided (for i t ) . Contradiction also is not desired, because (in the science ) whatever is beneficial to human being is stated. " (Panini) will state the prohibition na bahuvrlhau ( P. 1. 1. 29 ) (sarva etc. are not [ called sarvanaman ] in bahuvrihi)." This ignorance ( about the definite domain of application ) will not be removed unless the specific domain ( of application ) is ascertained. (The Bhasyakara) means to say by this sentence that in this connection no blame should be put (on the grammarian) since the grammar itself explains (all the problems).


Mahabhamja-Dipika, Ahnika VI


" Now, what is the purpose of the mention of (the word) sam a s a ) ? " ( Panini) thinks in this way : Just as by the mention of ( bahuvrthi) samasa the elision of case-endings and pnfnvadbhava ' treatment as masculine * is obtained, in the same manner why is not this option ( regardIng sarvanamasamjna ) also obtained ? " The bahuvrihi which is a compound only (is optionally liable for sarvanamasamjina ) " Just as (it is said,) * when a kartr which is kartr only (is to be conveyed, parasmaipada endings are added)', in the same manner " ( here it is said that optional sarvanama* samjna) may apply to that {bahuvrihi) compound which is a compound only"; (it) may not apply to that which is also a non-compound. Or (the word) samasa is pronounced with svarita accent (so that the rule applies to) that bahuvrihi which is introduced by (the adhikara) samasah.

(Objection :) Now, what kind of an appropriate expression is this i that which is bahuvrihi by the extention of the status of bahuvrihi 1* For bahuvrihi by means of extended application does not exist. Only the grammatical operation ( due to bahuvrihi) applies. ( Reply :) This difficulty does not arise. Just as (in the sentence) * He is Brahmadatta' extended application (of Brahmadatta status) is understood without (the suffix) vati, in the same manner (the extended application bahuvrihivat) is understood as a designation {bahuvrihi). Or with regard to extended application, the view of extended application of designation also exists without any (special) reason. Also for this reason this expression can be said to be proper. * # #

" Now what is the purpose of the mention of (the word) bahuvrihi ? " That which will be stated later (as a reply) is in the mind (of the Bhasyakara). At present he thinks only this way: Those (words) which are constituents in a dvamdva (compound) are read in the sarvadi (list). After the pumvadbhava (which is) applicable to any integration in general is effected in their case, the same constituent expresses... Therefore there is absence of sut.

" In dvamdva (the rule ) should not be (applicable), (for instance in the compound) daksinottarapurvanam" There are different (ways) of

22.1 ]



effecting pufnvadbhava in connection with a pronoun in any integration in general. Some (grammarians) effect pufnvadbhava in connection with a!! (pronouns). Others ( effect It) with the earlier member oi a coriipouad when the final member follows. Such being the case, according to those who effect (pufnvadbhava) in connection with the earlier constituent before the final constituent,... there will be absolutely no pufnvadbhava in (the expres« sion) aparadaksinottaranam on account of the popular understanding.

(Objection:) If pufnvadbhava is effected in connection with all constituents, the statement of (a compound having gender) according to that of the final member etc. cannot be justified. ( Reply :) This difficulty does not arise. The gender of the final member is noticed at the stage of constituent analysis, that is extended to the compound. Or by pufnvadbhava only words are made to withdraw. Since the meaning is cot made to withdraw the gender of the final member ( remains and ) is extended. In that case, either after the earlier ( feminine suffix ) tap is dropped tap is added after the group, or one more tap is added after (the group ) which ends in tap.' (The compound can be described as) ending in tap by (the dictum of) treating (one single substitute) as the final ( of the preceding) and initial (of the following element). Moreover, tadantavidhi also is (introduced by the Var.) prayojanam sarvanamavyayasafjijnayam ( Far. 8 on P. 1.1.72). ( Objection :) How can the stem ( be described as ) ending in tap, when pufnvadbhava is not effected in the case of (the final member) ending In tap and when a second tap is not added ? Because (the paribhasa says) that when a suffix is mentioned, that after which it is added (including the suffix is understood ). ( Reply:) It is already stated in this connection that (the mention of a suffix stands for the element to which it is added including the suffix) except in the case of feminine suffixes.

" When ( a rule) is already applicable (another rule that is stated cancels it ).'* ( Objection :) This can be said to be incorrect, because (the rules operate in ) different domains. Indeed there is no prohibition by (the rule) na bahuvrihau in the domain of dvamdva, nor (is it possible ) by (the rule) dvafndve ca in (the domain of) bahuvrihu

(Reply:) An answer is given : (The rule) yena naprapta... etc. is presented in a different way. How? This option is applicable to that the





[ 22. i

domain o( which is positively reached, i.e. positively pervaded by prohibition. In that case the prohibition is to be cancelled by this option. Then how is it (that the option is cancelled by the prohibition) ? This option being already applicable in (the domain of) hahuvrihi and thus having served its purpose is cancelled by another prohibition since it is later, due to the (rule) lipratisedhe paratva * in case of mutual conflict the later supersedes the earlier/

" Or the preceding exceptions (cancel the immediately following general rules, not the further ones)." ( Such exceptions are) cancelled by other rules. And the grammatical rules are expected (to apply) separately. Therefore, a statement having fulfilled its purpose with respect to the nearer one competes with other rules.

'• Of that which is implied and that which is directly mentioned (that which is directly mentioned undergoes grammatical operation)." (Objection :) The compound is (stated in the rule) dinnamanyantarale. This (parihhasa) is stated with reference to that in which a word in its accurately defined form is stated, for instance, (the rule ) srenyadayah krtadibhih. An answer is ( now ) given. Here direct mention through designation is adopted. What deserves receiving the designation diksamasa ? Is it not that ( compound ) about which the word dik is mentioned? ( I t is) proper (to call a compound diksamasa) when the word dik is mentioned (in connection with it).

(Objection :) It has already been stated that every bahuvrihi is pratipjdokta, ' a direct statement' because of the very statement regarding each one, for instance, (the compound ) mahabahuh. ( Reply :) That was stated in comparison with other (kinds of) compounds. (The distinction between) the implied mention and direct mention, however, exists within its own domain, for instance, (it exists in the case of the rule) sadhakatamam karanam ( P . 1.4.42).

This refutation (in connection with the problem of bahuvrihi) is justified when (the rule) dinnamany antarale is being stated. However, when it is rejected by the Var. diksimasasahayogayos cantaralapradhanabhi-




dhanat ( Var* 1 on P. 2.2.28) s because the compound of dik words and a compound with saha convey space and main item respectively,' then this refutation has no scope.

" This then is the purpose." The mention of the word bahuvrihi is thus settled to be meant for the subsequent ( rule ). And a statement is (assumed to be) purposeful. The option ( of sarvanamasamjna) before jas which is bahiraiiga is cancelled by the prohibition which is antaranga. Otherwise the continuation of the second word bahuvrihi would be simply redundant.



How is (the rule ) vibhasa jasi applicable here 7 For the two words, namely, vastrantara and vasananlara of which the dvanidva is constituted, are not read in the sarvadi list. (If you argue ) that by tadanlavidhi (they will be called sarvanaman), it (tadantavidhi) cannot be (accepted), because the word antara is upasarjana (in this compound ) and ( also because the term sarvanaman) is a significant name. Names and upasarjanas are not at all (included in ) sarvadi (list) for instance (in the word ) atisarvaya (sarva is not sarvanaman as it is upasarjana). The purpose of the mention of (the word) bahuvrihi is not found.

na bahuvrihau / P. 1. 1. 29
* (sarva etc. are) not (called sarvanaman) in a bahuvrihi (compo* und ) " .

" What is the example ? " The intention (behind this question is this): (A bahuvrihi) does not end in that (i. e. sarvanaman), because sarvanaman is placed first. And that ( word )5 in which the ( word) other (than sarvanaman) is placed first, for instance, priyavisvaya, will not be used. Why ? Because of the existence of another word having the same meaning, namely, viivapriyaya, that will be used.

'• This then (is the example ) : dvyanyaya." A numeral being always dependent upon that which is counted becomes (its ) qualifier. Therefore, it must be placed first. Hence tadantavidhi is (possible). (Objection:) 11 Now here also a sarvanaman only ( must be placed first). " (The word


Mahabhasya-Dipika, Ahnika VI


sarvanaman ) is additionally enumerated also here (in the VarttiJca) so that in a conjunctive occurrence of sarvanaman and samkhya (numeral), purva* nipata * first placement' will be effected for that (constituent) to which is it not already applicable. For which is it not applicable ? For sarvanaman,

( Reply :) Sl This difficulty does not (arise ). ( Kat.) will state this." This is meant by this statement: The sentence does not mean that what you have understood. " This also is the example." ( The Bhasyakara) points out this: If you think that doe to the absence of usage (of priyavisvaya) the other one {visvapriyaya ) will be used, (this is not correct, because ) this is not a matter of the intention of speaking of (just) one speaker. What then ? It Is based on popular practice. And this word priyaviivaya is found to be used in popular practice. Moreover, the mention of va (in the Var 2. 2.2. 35 ) va piiyasya is an indication of this observation.

" What is the purpose (of rejecting sarvanamasafnjna in that bahuvrihi which does not end in sarva etc.) ? " The intention (behind asking this question is this ) : 5-substitution, instrumental, smas-substitution etc. are the operations due to sarvanaman* And this is not possible in the case of the bahuvrihi ( which does not end in a sarvanaman). For example, the first one ( i . e . 5-substitution) is stated before drk, drsa and vatu. The instrumental is not heard, so also (the substitutes) smai etc. Therefore, (the Bhasyakara ) introduces another grammatical operation, '• akac should not be ( added)." (The objector) not seeing the difference between ka and akac in the case of ( words) ending in a vowel says, " What now, is the difference between ka and akac ? " " The difference (is found) in the case of ( words) ending in a consonant." This ( statement) is meant to show the special feature ( of the difference ). The difference is (found) also in the case of ( words) not ending in a consonant, for instance, dvikaputra and dvaki' puira.

C 1.5 ]
" How (is it possible ), in spite of ( your) wish (to cancel antaranga by means of bahirahga) ?" Even before the anyapadarlha, which is the cause of bahuvrihi, comes into existence (the sarvanamasafnjna and akac will take place). The cause of the prohibition is bahiranga, while the cause of akac is the name sarvanaman which is dependent upon one word (and therefore antaranga). Therefore, akac should be effected in connection with (the sarvanaman) which is the substratum of bahuvrihi*

24.6 ]



" For instance (in the derivation of) gomatpriya (a bahiranga operation cancels even an antaranga one)." The indication praiyayottarapadayoi ca (P. 7.2.98 ) which is found in grammar is justified in this way : This is understood as an indication in general ( so that) the bahiranga compound cancels the antaranga ( operation ) or the operations caused by compound cancel the antaranga (operation). "(Special) effort is made in connection with that." (This is) the implication: The indication is concerned only with luk, not with all ( operations) so that the antaranga paribhasa should not be cancelled.

'* There is (another purpose ) of the statement of this ( prohibition ).'* . (It is effected) where both the rule and its prohibition have a common basis (for instance, the compound ) " priyavisvaya" " This ( prohibition) is achieved also by the (statement of) prohibition (of sarvanamasamjna) in the case of an upasarjana." Whatever alternatives are stated in (the discussion on) the rule sarvadlni.^etc., such as giving a significant name, are to be understood here.

Now (the Bhasyakara) demonstrates that this (namely, the constituents) is the meaning of the word itself. " (The use of) the same word (to convey another meaning) is ( possible), because (that another meaning is) meant for the (literal) meaning ( of the word )."

( Objection :) If this is so, the prohibition of akac is wrongly applicable to (the construction) ahakam pita, because it is also meant for that (the compound). (Reply:) This difficulty does not arise. Since the word is characterised by difference, this construction cannot be (said to be) meant for (the compound ). Only the alauUka (grammatical constituents ), which are not worthy of being used, are meant for that (i. e. compound ).

" However, Gonardlya says, (that tvakatpitrka etc. are correct examples ).'* What is the view of this Teacher ? Some (grammarians) describe it (as follows): The mention of (the word ) bahuvrihi should be made for the sake of the example vastrantaravasanantarah. The purpose thus being (available) it is impossible to cancel the antaranga (akac) by


Mahabha&ya-Dipika, Ihnika


[ 24. 6

, :.•;• haldrakga) bahrmhi or to understand tadarthya. This is not(correct). just now it was stated that this (prohibition ) can be obtained also by the prohibition with regard to upasarjana. The word antara (in the above compound) is also upasarjana. Therefore, the Bhasyakara rejects this rule.

" Give to one who was formerly rich.'* The additional statement is made, because the prohibition is not applicable due to its not being a bahw •vrihi. It is said that when relative position (is conveyed, purva etc. are called sarvanaman ). *' And here relative position is not understood.'' This means (that relative position is not) mainly ( understood ). Because in the integration the word purva in (the word) bhutapurva which is construed with adhyatva * richness,' is subordinate to bhutapurva* Since it is upa* sarjana due to its being subordinate,^ the word purva ) is not (included ) in sarvadi list, just as (the word sarva in the compound ) atisarvaya.

[ 2.2 ]
Other ( grammarians ) explain this (as follows :) The relative position is not understood here (i.e. in the word bhutapurva). Just as the skill (conveyed by the word daksina in the construction) daksina gathakah is perishable, in the same manner the richness (conveyed by the word adhya) in (the compound) adhyapurva is perishable. This ( explanation ) is not possible. Why? Although richness is perishable, the state of being purva * previous • is fixed. Moreover, (the word) vyavastha is an adjective of purva etc., not of adhyafva*

[ 2.3 ]
Some other grammarians describe that the meaning vyavastha is not understood in the jahatsvartha integration (that in which constituents give up their own meaning ). # * * . .

trfiyasamase / P. 1.1. 30
" (sarva etc. are not called sarvanaman when they are used ) In an instrumental compound. "

" When (the word ) samase is being continued ( where is the necessity of the word samasa in this rule) ? " ( The Bhasyakara ) thinks in this way :

25.7 ]



when (the word ) samase is not used and the rule h ,i:^..; _.s if:l 7 / "'i/ir sentence is generated by means of what is heard and what is Implied (I. c« continued) : In the instrumental compound sarva ete* are not called sarvanaman. In (the discussion on ) the previous rule (it is stated that) akac is not heard in the sentence when (the principle of) the use of the same word by reason of being meant for that (is accepted.) And here the grammatical operations due to sarva tic. are already prohibited in a sentence. Tadarthya is resorted to because of the fact that due to the cooccurrence of certain similar feature it is possible to derive a sentence of this kind, However, if a sentence is completely dissimilar, like e. g. (the sentence ) Devadattena purvasmin kale * by Devadatta during an earlier time ', there is no prohibition,

" Or, rather the repeated mention of samase when {the word ) sarnSse is continued ( is meant for the purpose that it should be a part of the rule)." (Objection:) But this is the mention of (the word) asamasa. This is the reply : Although it is so, still, after the homogeneous a is substituted, (the word ) samase and not the other word (asamase) is being heard* Therefore, (the Bhasyakara says) ' samase is mentioned \ Arid after the rule is split, according to the maxim of naniwyukta, the prohibition is made of that which is not a compound but which is similar to a compound. (It is) not (applied in the sentence ) Devadattena purvasmin kale :« by Devadatta during an earlier time \

vibhasa jasi / P. 1 . 1 . 32
" (In a dvandva compound sarva etc. are) optionally ( called sarva* naman) before jas ( i . e. nominative plural ending ) . '

" The optional (sarvanamasafnjna ) is with reference to the operation effected to jas. For, akac is not effected." Here both, the operation effected to jas and akac, are dependent upon sarvanaman. In this connection when akac is prohibited and ka is added the replacement (of jas) by it does not take place due to the intervention (of ka, for instance), katara* katamakah. Some (grammarians) say that this should be added to the statement. Some other ( grammarians understand ) jasi as jasa ih * i (substitute in place) of jas\ However, according to still some other (gramma-






rians), since samjna is taught in each rule, in the rule jasah ii (the sarva7iamasamjna) is stated, therefore the option is with regard to the operation relating to jas alone, while elsewhere the prohibition (of sarvanamasafnjnft is understood).

purvaparavaradaksinottaraparadharani vyavasthayam asamjnayam / P. 1.1 • 34

" T h e mention of avara etc. again in the sutrapat\ redundant, because they are (already) read in the ganapatha ". From the examples (given) by the Vrttikara (on this rule) option (of sarvanainasamjna of purva etc.) before jas is known (to the Varttikakara ). This also is known (to him ) that purva tic. characterised by specific meanings are read among the sarvadis so that (they) will have sarvanaman as a permanent name. Therefore, here the Vakyakara is not questioning the formulation of the rule. By means of the ( above ) statement ( he ) introduces (the question ) : Why is the mention of each one of all these (words) in their own form made in the course of three rules ?

Avaradinam % of avara etc.' means ' of those similar to avara9 whose own respective form is meant (by the rule ). The Vakyakara has rejected the reading of each of the ( words) avara ete. in the three rules with the thought that they will be recognised when they will be mentioned in the ganapatha. However, in order to explain the intention of the Sutrakara the BhasyakSra justifies (by asking) " But how is it known that that is a repeated text while this is an earlier t e x t ? ' '

What kind of priority and posteriority is here when (the words are) listed in (twc ) separate texts ? By deciding the succession (the fault of) gaurava * prolixity ' based on reading can be removed. With regard to this other commentators explain the intention of the Bhasyakara in this way : This word adi is found to be used in connection with some ganas the enumelation in which is known as fixed and already arranged, for instance, in the lists, irenyadi etc. There itself (it is) described as conveying vyawstha. However, sometimes popular lists containing (words) similar to the usage of the learned accompanied by some partial similarity ( with other words in




the usage) are read with the word adi conveying kind. (These lists) are put among the examples in the vrtti and in the groups which are not prior to the Sutra, for instance the krtadi list. Therefore, (the Bhasyakara ) will say * srenyadi-s are being read, krtadi is an open list.' In the case of those ( rules ) which contain the word adi as conveying vyavastha, the ganas with the enumeration already known to be fixed and arranged (exist earlier and) the rule is taught (later ). However, in the case of those rules where the word adi means kind the rule containing only (some words) similar to the usage of the learned is taught earlier, the (other) words in the form of reference in ganas are read later just as the vrtti and examples (are added later). The question about length and brevity is not to be asked with reference to such Bhasyasiitras, Therefore (the Bhasya -. kara says) " And surely the Teachers having once formulated the rules do not take (them) back " {Mbh*). For the Bhasyasutras are meant to be illustrative and explanatory by means of injunction and elucidation. For instance, the cases are mentioned in rules in specific forms. Among them (some) are changed (in some rules) while some remain unchanged. In the same manner in other grammatical systems there are specific lists of ganasutras the traditional sequence in which is unbroken. Among them some (ganasiitras) are modified while some others are read in the same manner. Moreover, references to the technical terms of ancient teachers or to the names (given by) ancient teachers as well as profuse references to the lists of stems are found, for instance, krnmakarasamdhyaksarantavyayibhavai ca 1 a krdanta ending in m or a diphthong and an avyayibhava compound ( are called avyaya),' etc.

There, if the word adi is understood as conveying vyavastha', that •"•( ganapdtha) is (to be regarded as) the earlier text, read (taught) earlier by the Teacher. But the sutrapatha is later. Therefore the repeated reading of each of the words avara etc. in the three rules was rejected ( by the VarttikakSra suggesting thereby) that shorter rule should be formed. However, if the word adi denotes kind (it should be understood that) the teacher has picked up some ( words in the sutrapatha) since afterwards he read them in the ganas just as the grahanakat or they are (to be regarded as) constructed "by skilled composers for the sake of memorization. Therefore there is a disregard for the length or brevity with regard to them. (Sometimes) that which is already read is read again. Additional statements are inserted. For instance, upad dvyajajinam agauradayah; nirajinodajinopajinah; iunah


Mahabhasya-Dipika', Ahnika VI

[ 26. 17

prasamnain dirghatvam; dvitiya 'uupasarge; pravrddhadisu khaitarudhah; nadyajadyudattatve brhanrnahator upasanikhyariam. Among the gauradis brhatyah, mahatyah, saudharmyah (are read ), The gopavanadi (list) is read in the bidadi (list) as a samhitapatha just as {it appears) in the second chapter, kanva etc. are read in the gargadi (list), they are again read as kanvadibhyah. But (under the rule) haritadtbhyo 'nah harita etc. are not again read. ( The words lohita etc.) are also ( not read again under the rule) sarvaira lohitadi-katantebhydh. ( I t is stated that) iakala is intended to be earlier than kanva and later than kata. (The words appear) different in separated (text) and different in samhitapatha* Mindas read in a different way. Aharyas read in a still different way. ( For instance), sthuiabahu sthilla bahu, triveni and trivaya, paian lunaviijate, paiau lunavihine, pasau lunavighate. kapigaviyudhis are read in the (context of) derivation from pra (?). jarabhara, svapaca are mentioned in their ready-made forms, vipratisedha is explained at udanapace ubhayam. In the nyankvddi list kutva is stated with reference to ivapaka. All this is, as it is remembered, somehow summed up (here) as an evidence. In this way the word adi conveying kind is used in (some ) rules so that the subsequent mention of (the words) in the ganapatha in a comprehensive manner, after the meaning is ascertained, does not lead to fault (of gaurava). This (subsequent mention should be understood as representing) an alternative view.

It is also stated, how it is known that this word adi means vyavastha and not kind. How does the question regarding kind (as the meaning of adi arise) ? Because the grammar makes use of both (the meanings of adi,) so ( both ) the uses are found. For instance, indragnl paramasyam prthivyam madhyamasyam avamasyarn uta sthah / samanatra tisthan juhoti /

'* Those purva etc. These avara etc.*' Haying decided the sequence from the indication (in the rule) purvadibhyo navabhyah ( he) makes this statement. The other (disputant), however, thinks that he (the opponent) has introduced the recitation in connection with the sequence of the two readings. (Therefore, he asks :) " How to know that this is the earlier ; text ? " The succession, that is sequential enumeration in the text, is stated i here. That is decided. For instance, when a question (is asked ), ' How is it known that this is the oldest one and this is the youngest o n e ? ' the decision is stated on the basis of ( o n e ) being born earlier and (the other) being born later. In the same manner he (the Bhasyakara) has adopted the utterance referring to prior period and posterior period for the sake of

27.28 ]



decision, •* ILubu are purva eic*'* By the mention of pun-i is understood ' the word purva included in the ganapatha. '* These also ( are) avara etc. *' ( Here also) the word purva only included in the sutrapatha is referred to.

However, some ( grammarians ) think that the word puna is referred to by mentioning (the word ) avara. With reference to this (they argue ) that because of its being earlier, the being itself is referred to as purva (bhava) with reference to all activities. Birth is the name of one of the modifications of being. Being which is the beginning of it, is described as avarabhava with reference to all activities, because destruction is referred to as ania due to the disappearance of the specific being. Similarly here also, since the two readings whose sequence is fixed are separate, the word purva itself is referred to as avara as well as purva. By the word adi beginning of vyavastha is understood.

" Because those are purva etc." The word purva which has already obtained a specific position in the arrangement is at their head. The commencement of others is indicated by it. * These are avara etc.Jf It * means that avara is the iesa i. e. beginning ( of the list), which is placed first (in the sutrapatha) and which is understood ( as indicating) the posteriority of the rules in the sarvanaman section. Or by the word purva a. cross refe» rence is made to these rules beginning with avara etc.

One who thinks that the decision regarding sequence cannot be based on an indication like this says, " These are also purva etc.'* They cannot be explained by merely making a statement. Since the grammar makes use of both (the meanings of adi) and since ( both the meanings are) found in usage, it is possible to understand this word adi as conveying kind.

Therefore, the knowledge of indication already introduced before is made manifest as the cause of the decision on sequence : " (The fact) that he mentions (the word ) nava (in the rule ) purvadibhyo navabhyah (indicates that the gav^patha is earlier)." The mention of nava is purposeful in the ganapatha since it is meant for excluding tyad etc. (in the ganapatha from the operation of P. 7.1.16). If the ganapatha is well-arranged, with the sequence fixed, and with each word delimited to its own form, then ( only) it is possible to construe it with the word nava, as for instance in iamam astanam


Mahabhasya-Dipika, Ihniha



( P. 7.3.74 ), phanafn ca saptanam ( P. 6 4.125 ), kiras ca pancabhyah ( P. 7.2. 75) etc. In this way, since only nine ( words), purva etc. are listed in the sutrast and t hecomprehension ( of all the nine words) was possible by a briefer way, the mention of thQ word nava is an indication of the fact that there exists another list outside the sutrapatha which is earlier enunciated by the Teacher ( himself or, being borrowed ) from some other grammar which is continued through unbroken tradition and (that it is) not merely imagined. Sometimes (the Sutrakara ) mentions ( words merely ) by (referring to ) the list, for instance, cadisu. From an indication like this, therefore, this word adi (Is to be understood as) conveying vyavastha. Moreover, availability in the usage cannot be an indication of (the fact that adi) conveys kind* because (such usage) is restricted to Vedic literature.

Another (grammarian) says, * These words puna etc. are all read in the (sarvadi) list. There the close association of these words (all) connected with a special meaning is understood. Recognition even from a reference to a part of rules is well-known. For instance, (in the rule) Jctena ca pujayam (P. 2.2.12) the rule matibuddhipuja ( P . 3.2.188) etc. is reminded { by kta ). In the same manner, those who thought, having seen the two words purva and avara read in the gana, that it is possible to indicate (these words) through association, read one part of avaradis in the rules. (In the sentence ) " Also these purva etc. " the ganapatha is intended ( by the word purvadini). Therefore, as an indicator, one or the other word is mentioned. From that the whole group (is understood ). From the order which is obtained from that a reference to the sequence (is understood).

If so, then the reading of each word (purva etc.) distributed over three rules is again adopted so that (the words purva etc. conveying) the same meaning should be understood without exception when optional (sarvanamasamjua) is being taught in the sutras with reference to purva etc. characterized by specific meanings in the ganapatha. For, when certain characteristic marks are first mentioned in some context, the function caused by their association is restricted to the operations relating to that context. For instance, when it is said ' Bring the man with white clothes, (one) with matted locks, (one) with hair tied on his head * (the qualifiers such as the white clothes and matted hair are connected only with the action of bringing). However, when a certain (object) already characterized by some mark and intended for another purpose is mentioned again in a different context, the expression expressive of the marks is not (understood as)

29.6 ]}



associated with (the object) through imagination. Therefore when the: meanings (of piirva etc.) are mentioned along with an optional statement, the repeated mention ( of the same meanings ) is meant to indicate the identity of words. And this will be achieved by him (i. e. the Bhilsyakara ) at the end.

According to the other (disputant), just as (in the rule) purvadihhyo navahhyah (the words) purva etc; conveying relative position etc. are understood, in the same manner here also (the words purva etc. conveying relative position ) will be ( understood ). And the meanings are distinctive. For instance, (the expressions) tailamatra and ghrtamatra are not included (in the domain of the suffix nip) although (the word matra in them) is similar, ( because it is) not included (in the pratyahara matrac). There, just as the item marked by the code letter (c) is understood as a cause of. distinction of ( one) word (from another) in spite of (the same) verbal form, in the same manner, relative position also becomes a cause for distinguishing (a word from another) on the basis of specific meaning. Considering this (he ) says " These (words, purva etc.) qualified this way only are read in the ganapatha. "

" There should be no exclusion ( of purva etc.) by the exclusion of dpi e t c . " The repetition (is) meant for this purpose. The followers of Apisali (read first) tyadadi-% beginning with kim and ending with asmat, then the three rules purvaparavara etc. and then s-varadis. According to this arrangement of the text (the exclusion) of purva etc. (through) the exclusion of dvi etc. wrongly applies. For instance, the name karmapravacaniya (taught by the rule) anur laksane ( P . 1.4. 84) is (first looked upon as superfluous), because the general statement has (already) given ( i t ) (for instance, in the expression ) vidyud ami vanam * lightning near the forest \ Owing to the repeated statement ( of the karmapravacaniyasamjna) the accusative caused by karmapravacaniya (samjna ) applicable by another rule ( viz. P. 2. 3. 8 ) cancels the instrumental which has already cancelled it (i. e. the accusative ), in the same manner the sarvanamasamjna being taught again by means of reading each word (i. e. purva etc.) leading to the application of taml etc. by means of another statement, cancels the prohibition (of applicalion of tasil etc.) which is an exception. With reference to this the (sarvanama) samjna is permanently given to purva etc. by the three rules. And owing to the attachment of svarita mark a separate sentence (is understood) by means of (which) optional (sarvanamasamjna) before jas is understood in every rule. This (statement about) option is not possible



Ahnilca VI


If (the rule) is one sentence. The reading of each word is meant for con» necting (the statement) with option through a separate sentence. la the absence of expectancy, however, It is proper ( t o understand the rule a s ) one sentence,

According to the other (disputant), for those who assume the sTdrapdtha to be earlier, the arrangement in the sequence does not give an indication for the fixed arrangement. For instance, the statement of anu* datta in connection with deva brahmanah ( P. 1. 2. 38 ) in the subrahmavti* indicactes that in addition to the existence of another reading, a different kind of arrangement exist?. Moreover, separate mention of kirn, caused by the exclusion of dvi etc. indicates that in this section suirapdtha is not assumed to be earlier.

In order to remove (the possibility) of the indication (ptlrvatra in the rule) purvatrasiddham ( P . 8. 2. 1) being restricted to its own domain, (the Bhasyakara ) adduces another indication referring to another word (sayIng) (viz uttaratra occurring in the Var.) "jasbhdvad iti ced uttaratra.'*

( I ) will teach optional (sarvanamasamjna) before j a s " for those (words ) pilrva etc. alone, for which it is already established. (With this intention) the repeated mention of each word by means of the three rules (is made by Panini). And these meanings separated in three rules cannot be brought together in one rule without confusion. Therefore, although another reading, ( namely) purvddayah sapta, is imagined, (the making of three) separate rules does not retire (i.e. cannot be avoided). Moreover, just as there is no inconsistency in the matter of brevity or lengthiness when the same three mentions (are assumed), in the same manner, there is no difference in the matter of brevity or lengthiness when the same three rules ( are there).

But for which reason is praptavibhdsd not possible if there is no mention of meanings by referring to each word (purva etc.)? In the case of prdptavibhasa the use ( of words) is based on repetition. Owing to the difference of use, and to the repetiiion the established use of words is obstructed. And the fault of gaurava results when (word is ) used in a different way. Moreover, it is said ( by Pat.) : " With a lesser effort greater.

30.12 ]



and greater purpose (should be achieved )". ( Also) : "( When a doubt arises whether a rule) should be ( understood ) as a new statement or a restriction [it) will be ( understood as) a new statement, not a restriction."

For which reason should not the non-difference of the words seen in the ganapatha be understood when they are included in the sutrasl "(Reply :) In view of contradiction with the general practice there is no possibility of function caused by the association of a mark which has already served its purpose ( of distinguishing ) elsewhere. Therefore the maTk is not (regarded as being) near (the object mentioned at another place) due to the (possibility of) contradiction with the general practice. For instance, (the mark is not present in the context of another action in the case of the following statements :) st Quickly bring that goat which is tied ''; " Give food to him who is practising pranayama in water*'; " H e who has controlled his speech, who is firm and motionless should be employed as a witness *\ However, when there is no contradiction (the distinguishing mark is present also in another action even) without the use of a word ( t o express i t ) . ( For instance,) '• Give food to him who was seen wearing white clothes.' 1 In this way the distinguishing mark is inconsistently present or absent ( i n another action ). Therefore, the use ( of the words purva etc.) in a different sense is not understood, because the topic (sarvanamani) is not marked with svarita.

( Objection : ) Now here also, since purvadi-s are sarvadi-s the topic of saruanamasafnjna, which is already known, is marked with sxarita. Or the word sarvanaman is understood as an adjective of the samjnin (i. e. purva etc.) just as the word samkhya (is understood as a samjnin) in the definition of sat. ( Reply : ) It is not possible to continue sarvadi-s here. If the continuation ( i s accepted), praptavibhasa will have to be understood in connection with the mention of taija ( i n P. 1.1.33). Nor is it possible to understand the word sarvanaman as a qualifier of samjnin, since it conveys its own form and is used as a samjna also in the subsequent rules requiring (its presence ) for (teaching) optional samjna. ( I n the rule) purvadibhyo navabhyah etc. ( P. 7.1.16) the word sarvanaman, ( continued from the rule ) sarvanamnah smai ( P . 7.1.14 where its semantic) connection is already established and conveys a samjnin, is understood as a cause of distinction of purva etc. However, when distinguishing marks have served their purpose already in antarganas, an effort should be made for their presence or otherwise in such a way that there is no contradiction.






svam ajncitidhanakhayain / P. 1 . 1 . 35
* (The word ) $m~ (Is optionally called sarminantan before the nominative plural ending jas) provided it does not denote a relative or property. 1

This word sva is a synonym of self, ( one's) own, kinsman and wealth. And in the domain of (the suffix ) aminic ( i t ) conveys master or mastery •
For instance, svakrta irine paraclr nidadhati (MS* 3.2,4), "(The

Adhvaryu ) puts those (three Nairrti bricks ) so as to follow one another in a hollow made by himself,0 na hi svah svam hinasti ( TS. 5. 1. 7. 1 ), u One does not kill oneself." ( l a these two citations sva means the self). Similarly in the sense of ( one's ) own (it is found in expressions such as) svam grham 1 own house,' svah punisah * own servant.' ( I t is) used in the sense of wealth For instance, yavanta evasya svas tan asma anukan avivadinah karoti (MS. 4.3.7),. " ( H e ) makes as many kinsmen belonging to his family free of quarrel as belong to him."

" What is the purpose of reading akhya ? " (This is) the idea ( behind asking the question ). The meaning of a sentence (is that which is understood from the mutual) connection ( of words). For instance, ( when it is said * one should slaughter) the animal, the goat' the word pasu * animal * conveys pasutva * property of being an animal * or the object characterized by it ( by that property ). In the same manner, when the word chaga ' goat' is used, the pasutva is restricted to sacrificial oblation, nothing more or less (than that) is possible. As a result the donkeys, mules etc. arc excluded. It is not possible that the pasutva is connected with only one ( animal, namely,) goat and ( at the same lime it conveys something) less or more or connected with only one ( animal namely,) donkey. Similarly, the word sva also conveys (the meaning ) one's own. The additional (meaning) is the meaning of the sentence. Thus, (the follower of) samsarga view rejects the mention of akhya (in the rule ).

" (The word akhya is mentioned ) so that the word sva which is a synonym of kinsman or wealth should be understood." The remaining part of the above sentence is " in the prohibition." The reading according to some other grammarians is ajnatidhanakhyayam. They interpret (the sentence by supplying the word) vidhau " in the injunction.*' (The view of) this (disputant), however, is that the specific ( meaning) is the meaning




a sentence. (To explain:) There are as many words (of the phonetic sequence ) sva as the meanings con\eyed. Owing to the sam^n,.^ of die vsi'ixd form the specific meanings are not manifest. Through (connection with) another word (they ) are stated to be conveying specific meanings. Therefore, (in the expression ) sre putrah * ( one's) own sons * the word sva Itself conveys the meaning ' kinsman.' The prohibition is applicable in this view, Xhe word akhya serves this purpose: The prohibition applies there where the knowledge of (the meaning ) * kinsman ' or * wealth ' proceeds (directly) from the form of the word (sva) itself, without (its connection with) another word.

Or the word akhya is mentioned so that the prohibition should not apply when the meaning f kinsman • or ' wealth * is understood either from the meaning ( of the sentence) or from context.

antaram bahiiyogopasathvyanayoh / P. 1 . 1 . 38
' (The word) antara (is optionally called sarvanaman before the nominative plural ending jas) when it conveys connection with outside or a clothing.*

With reference to this (rule) there are different views. Some (grammarians ) think that (bahiryoga means) * connection with outside/ According to them by antare grhah or antara grhah are expressed the houses of out-caste and low-caste people which are outside rampart. However, for those according to whom the meaning of bahiryoga is * connection with that which lies outside', by antare grhah or antara grhah are expressed the interior (parts) of houses. Just as there is a difference of opinion in this regard, in the same manner it is there also with regard to (the meaning of the word ) upasamvyana, because (it can be regarded as a word conveying ) either karman ' object' or karana ' instrument': (upasamvyana is that which is covered, or) upasamvyana is that by which ( something) is coveredt

" The mention of upasamvyana is superfluous. •' " Then it should be stated ( as ) meant for a pair of garments." ( This is the) meaning : when (it is assumed that) bahiryoga (includes the meaning of upasamvyana ), it is difficult to understand the shape etc. of the garments of equal size which arc


Mahabhmya-Dipika, Ahnika VI


not worn. ( Clothes like ) a cloth with painted flowers or a loin-cloth can be recognised as outer-garments or inner garments (respectively) even when they are not worn. This is not possible in the case of a pair of garment (of equal size). Therefore, (the sarvanamasathjna) is not accomplished ( i n such a case ) by ( merely saying ) bahiryoga,

" In this situation also the man who (thinks before he acts, i. e. an intelligent man, can recognise which is the inner-garment and which is the outer-garment) ". This is the meaning : ( T h e objector) thinks that ( t h e sarvanamasamjna ) is not accomplished in the case of a pair of garments ( of equal size ) even if (the word ) upasamvyana is read. Some commentators state (as a reply to this) that in view of (their being used as) upasamvyana in future the name (upasamvyana) will be used ( with reference to the pair of garments cf equal size). However, the other (grammarians) think as follows : (The word ) adi is mentioned in the Varttika ( viz. iatahayugadya* rtham). It is meant for (the inclusion of) three garments or four garments. In that case when three garments are worn (sarvanamasamjna) is not applicable to the first (innermost) garment if (the word bahiryoga) in the rule (is understood as meaning) connection with ( something ) lying outside. If (the word bahiryoga is understood as meaning) connection with outside, even then neither does the first nor does the middle (garment) has it. And it is desired { for the name sarvanaman) for both (so that the expressions) antare satakah and antarah satakah ( can be used to refer to them ). Therefore, (the mention of the word upasamvyana may be) rejected for a pair (of garments of equal size), whereas ( i t ) has to be stated for the sake of this (i. e. three or four garments).

" It should be stated : * not when pur is referred to.' " (The sarvanamasathjna) is applicable ( t o antara) by (the paribhasa) pratipadikagrahane lingavisistasya (grahanam) : * When a nominal base is mentioned (it should be understood as a reference) also to (the nominal base ) with a specific gender,' Hence the statement of rejection. And in some country a town is built within a rampart, while in some other country (it is built) outside a rampart. Therefore bahiryoga in connection with pur * a town ' is of both the kinds.

P. 1.1. 27 : ^RWTR FT The sixth Ahnika of the Mbh. deals with the siltras from P. 1.1. 27 up to P. L 1.44. Actually, however, out of the 18 rules, 15 rules only are treated by Kat. Pat. and Bhar. The present Ahnika deals with the following four samjna-s : ^T^ and fiRFSI. The first ten stdras ( P. 1. 1.27-36) deal with The present fascicle of the sixth Ahnika deals with this first group of rules teaching the srfirow^rT. This rule is quite extensively discussed by all the three commentators. The discussion on this rule Is divided into five srf^f^rris traditionally known as follows : ( I I ) ^TI^qWl'ic^TftlW3T^ I ( I I I ) i (IV ) ^vm^rsft": H^f^TW sratennfaqRaiis; I ( V )

Summary : A snpftfir compound always conveys s ^ q ^ s r ' a meaning other than the meaning of constituents' or ' the meaning of another word.' The meaning of the constituents is gor 'subordinate' to or 'qualifier' of spqq^isr. The «r|sftf| is of two kinds : c r ^ g ^ ^ j j R and ara^goitffl^ra.' In the former the meaning of the constituents is understood in the action with which 8T?gq^rsJ is connected. For instance, in the sentence g^i^rgr?i?PT ' Bring the person with white clothes ' the goj, namely the white clothes *s understood in the action of bringing the person who is the In other words, white clothes also are a part of the object of bringing. On the other hand, in the sra^gi^niRTR type of 3|[£ft%, the meaning of the qualifiers is not understood in action with which the sr^q^rst is connected. For instance, in the sentence ferg??Hq « Bring the person who is possessed of variegated cows' the variegated cows ( f ^ n *\T) do not become part of the object of bringing. In other words, the variegated cows are not brought when the person who is the SF^q^i^, is brought. is a <sn|sft% compound. Assuming that it is an p kind of «?|cft% the objector raises a difficulty that the word H # cannot get


Mahabhasua-Dipika* Ahnika



the designation ^T^fam^. ^ ^ T % means ^ # : «nf3[W m^ *! ^ o s e ( words) the ftthi ,:mong which is (the word) sj# \ The word z$ being the rpi ( qualifier ) Is iiul understood la the meaning of the ^?^q?pj i» £• the list of words «g# etc. The reply to this objection is that the compound ?jc?ff|fif should be understood as a ^"gowf^TR kind of crpffff so that the word ^f# also will be understood in the meaning of S ^ T I T P I and will get the designation <yej*iW^. This point regarding the compound s r ^ t f a is briefly discussed in the Mbh., whereas it has been elaborately treated in the Dip. The author discusses, inter alia, the following points : 1. Three possibilities of dissolving the compound and 2. Two possibilities of dissolving the s?|fsft% compound simple «r^tf| and f§r^%q ^g^tff. 3. 4. Reference to arp^q^T^1 by means of the pronoun ^ ant feature of the constituent analysis of 3n|^ff. : an import-

Various kinds of relation understood from the possessive suffixes such as 3q?|^, etc. as well as from sigcftlf which conveys the same sense as that of ^ \ according to Kat.

5. Other possible replies to the main objection.

1. Passages 1.1, 2 and 3 discuss different possibilities of dissolving the compound *?sri^H%. Two among them being discarded owing to the difficulties involved in accepting them, the remaining one, namely, « i | ^ f | is accepted. In the beginning the author tries to establish that there is a scope for doubt about the meaning of the compound $ ^ f t % . 2. fe%^Tf£mj?ftfTp3r «nfel etc. Tradition speaks of fourteen different means which help ascertain the meaning of a homonymous expression (see VP, 2. 315-6). The word f?*$m means the specific meaning intended by the speaker. f^^fiOTwfjrfax* is, therefore, any one of these fourteen means. A doubt regarding the meaning of a word exists in the absence of such a means. The author perhaps implies that there is no scope for doubt in the case of ^ t f t f t since one or the other ficfr^f^Rf¥im exists. Although the author has not explicitly stated it, it follows from the discussion in the subsequent passages that it is the gender and number of the word H^Htfo which is a to ascertain it as




3. F<j^l^#^ etc. The Mbh. {Paspiia pp, 1 and 2) cites a Vedic injunction : f^^#mi^r^#?7?TfgTftJ?T^cr {(^ne should sacrifice a cow with big spots dedicated to srfsi-^Tpf\ This injunction, as it is, is not so far traced to any srhs text.1 Pat. cites this statement to explain, the doubt and its removal by means of some fw^lf^TOf^fJrrr. The compound trf^t is ambiguous. It can be understood as q ^ ^ n ^ : ^SJSJT ^mj 1 (the cow) which is fat and spotted \ Or it can be regarded as $rqm$i ^ f o r 2?^?f m 4 {the cow ) which has big spots ( on her body ) J , The sentence ^35\H<fta«r|T|fVn etc. cannot be correctly understood unless the meaning of ^sj^j^cft is ascertained by means of some fg%^TBm*?ftrftrTr» According to Pat. the accent of the compound is the f^%qifv^iiiif?ffi?Tf that resolves the doubt about its meaning. Since the accent is found on the first member of the compound, this word is understood as srpftff according to p. 6 . 2 . 1 . 4. a#fflrmr 3Tl%j. The meaning of the compound H^Ml^i c a n ^ e understood as ' *?# which is the first'. In that case it will be a 5HNTR3. However, this cannot be accepted for three reasons pointed oot by the author as follows : ( 1 ) The word srrf^ becjmes redundant. Since in this meaning the name ?g^«T!*n| becomes applicable only to HW, the word sicftr has no function to perform. ( 2) The word srif^ which is the gifirq^ of the cE^nro compound frcfftfo is the main member. The gender and number of a cfrfhrRq ( which is a variety of $?*gm) a gree with those of the 3W?^ ( P. 2. 4. 46), Accordingly, the proper form of the ^ ^ r i ^ i sreif^ must be OTT%, oot SR^Vfi*. ( 3 ) Since only the word &$ gets the designation ^r^iTO^ in this interpretation, the listing of f^«j, etc. in the H3tf?7Tur is redundant.

1. ^S^TT^T^T?^ f%. Another way of understanding the compound srarf^ is as follows : sn^te?^1 fC?TI% I ' snf^ means that which is taken (mentioned ) ' . The word snf^ is derived from m + ^J with the krt suffix % {hi) according to P. 3.3.92 in the sense of object. The word «nf^[ is further compounded with ^r#» S?3rf^*> * that which is taken from ^ \ This is an ablative ^ p q " . The compound H^"T^ understood in this way can convey all the words which follow ^ onwards. However, this interpretation also is discarded, because here the word anf^ does not convey its conventional meaning, namely ' beginning • or ' first \ The conventional meaning of a word is always more powerful than its etymological meaning :


Cf. however, 3?[f?fqr^ff^|TftJTr^cT Rath. Sam. 13.6.


Mahabhasya'Dlpika, Ahnika VI


The Ms reads, ^ ^ s ^ l % 3 1 l % I *NN*TT3T^fa ?f?r I These two statements however cannot be reconciled, because they lead to two different meanings of wrf%» Although both the statements imply the derivation of ST?% from the same root, namely, wi4-^i with the same suffix f, in the former the suffix conveys srqFJF?, whereas in the latter it conveys $ $ 3 . Accordingly, the word m\f% derived by the former sentence would qualify the word 7^ itself while the 3?r% derived in the tatter statement would qualify the list, fn order to bring accord between the two statements the wording of the former has to be emended into 3TT<fte^ ^ 1 % and of the latter into: *$S53T° etc. as it is presumed in the above discussion. In that case both the statements derive the word 3*1% from 3 T + ^I in the sense of «pj"3[. The C statement ^ S ^ I T I ^ 3CI£R;% has to be regarded as a semantic paraphrase rather than the constituent analysis of the compound SJ3tfi[. Understood this way the word ^c^f§[ would rnran ' that (group) which is taken from ^ i. e. which follows sarva in the list. * 2 sHN*?T3[ s n ^ f t t Strictly Ht^TFi; would be an incorrect form here, because the word H # h e r e denotes its own form and so, not having the name s^TOT^ it cannot be inflected as a H$«iN4«^ deserving the operations such as ^ n ^ substitution ( P. 7.1.15). Correction ! In the light of the emendation suggested above, read which for from which (something) in the Translation 1.2, line 1. similarly, read all (i. e. visva etc.) for sarva etc. in line 2.

1. H * ! W % etc. The third alternative for the interpretation of tho compound 5^1% is to understand it as a f%ft w ? S ^ meaning { previous to Spf / This alternative also fails, because it signifies nothing. There is no item previous to ^ # . Thus, finally, the word *reifi[ has to be accepted as a 3fsftf|" as stated in the Jtfbh.: 2. f^flfq1 ^ S T E ^ ^ etc. Just as the disputant has turned down all the proposals about the various ways of compounding 3 $ and anf^ except the ooe leading to si|^t%, he similarly turns down the possibility of dissolving the gipfti*i n&ftf% in different ways. No arguments are given by him in explaintion of his view.

1. After it is finally accepted that the word e^fr^% is a compound the author points out two possibilities of interpreting this j ( 1) simple *tfcft% and ( 2 ) ^35%*? «rgsft%. Unless one knows the intention of the speaker, one canno^ decide which of the two alternatives is correct.




This fact is illustrated here by means of an example of the word ?ifi^f which is ambiguous since It can be interpreted in two ways. 2. ^5rafaa%Trorat. KW^n s retaining of one ( out of many ) ' is one of the five kinds of ^ % ' integration ' in Paoini's system (See P. L2,64-73 )« ft consists in the abbreviation of two or more words having the same form into one, e.g. fajr is an q^^hr of two words S[S>T: and f§r:. The dual and plural forms of any stem are examples of q;s|>$Ff. Pamni has further enumerated some specific cases where w^%^ takes place even when the words differ in their form or meaning. P. 1.2.65 : f.ft ^ n ^^TcnSr^? ft§r^: teaches, for instance, that when a word conveying TfteTqc^ is accompanied by a word conveying g^iirsj, the former is retained for the both if its difference from the latter is only due to its being fTTorPTc^ 'Tt^T^r^ is any descendant from grandson onwards ( P. 4.1.162 ). A T\T% is called s p ^ if an elderly member of the family such as father is alive ( P . 4.1.163, 164). These two meanings are conveyed by different suffixes. The word Ttfrqf can be understood as an %^m of two words TTR^: * n^TTR of ^ ' a n d * T r* O * g^TTr^ of i p i * T «f T T The word m&t is different from the word JTTT^^OT only due to the fact that it is derived by adding the suffix q ^ conveying ^rV^nqc^ ( P. 4.1.105), whereas the word Tnnfpm is derived by adding the suffix <Ef» conveying sj^rqc** ( P. 4. 1.101 ). The difference of TTR§ from ij\*zfam is caused by the suffix ^ conveying jffa. Therefore <£*?i«* of the two, namely ni*ff, is allowed by P. 1.2.65. However, the same word mi^\ can also be regarded as a nominative or accusative dual of Tnrq" and, therefore, an ordinary qq^rq meaning two offsprings of Garga, The word TUI?TI is thus ambiguous. Its correct meaning cannot be understood unless one understands the intention of the speaker. If the speaker intends to convey the sense of offspring in general the word Tirrcff is an ordinary ^ ^ t q . However, if he.intends to convey the existence of the father of one ^Ttenqri as its special feature, ?n*^f becomes an qq&fre according to P. 1.2.65. The intention of the speaker can be understood only through its constituent analysis. 2. cr^r^n^nf^f[ffff§r ^TESS^. P. 2.4.21 : referred to here in order to emphasize the importance of the intention of the speaker. The rule teaches that a compound ending in the word ^q^rf or ^T^JTJ is used in neuter if it is intended to express a starting point. Apparently, the word sttf^^THT * intention to speak ' is redundant. The meaning of the rule would have been the same even when cf^liur^T^T^ instead of 3^i3nf^n;rnqi^ were read in the rule. The redundant mention thus indicates that Panini wants to convey importance of the intention of the speaker through its use. The author suggests that in the present case it is necessary to understand





[ 1.16

speaker's ( Punini's) intention while interpreting the ambiguous compound

Like the word Tffi^f, the compound H ^ ^ H also is ambiguous. It may be regarded as ( 1 ) a simple «r|sft%, ^ anf^fqi ^nfif * those of which *r# Is the first ' o r ( 2 ) fnh%q 3§sft% snpfrff which is the result of The latter kind is explained in this passage. L p f W f t s ^ The word ^ f r : means ' KJ# which is sn ,V It is a srafarc?! ( P. 2.1.57). The word H^T#fr means « those of which ^ is the first ( m: wrf^TT mft)•' It is a ^|*ft% ( P. 2.2.24 ). The word sret^Tfir in the Paninian rule is an rr^m of two w o r d s - ^ f ^ : and ST^Yft, the ^^n?«? and the s?pftf|\ The former is in the masculine, while the latter is In neuter. The latter (i.e. the sr?rsftf|) is retained for the both by P. 1.2.69 : 5 ^ s p j ^ ^ % | ^ ^ ^ | ^ | H ^ ^ q | ^ . The word *T3fr£?fir in the rule has thus a double character: it is basically a snpft% but it is igsfft* one retained for both srgf% and *rafa[fcfir. Therefore it is described here as § i w ¥ , f^f: m&w: q&l * that which is made an trspr^.' [ Correction ! In the Text, 1.5 line 3, read 69 for 68. ] The word H^TCT^ a ^ an tg&m of % can be justified also by referring to the accent of both the compounds. According to P. 6.1.223 the accent of sreif^ is on the last syllable while the word H^pftfir is accented on the first syllable according to P. 6 2.1. According to the Var. 24 on P. 1.2.64 : s^firarRT ^^HRftftr', when q ^ q is to be made of words having different accents, that word, the accent of which Is taught by a later rule, is retained. Therefore, s R ^ f i r , the accent of which is decided by 6.2.1 which is later than P. 6.1.223, is retained. The f^^r?3j;pftf| is absent in the Mbh. The author of the Dtp. explains it for the first time. Later grammarians like Kaiyata also have discussed it. According to the author of the Dip this complicated way of interpreting the word H ^ H H ^ is necessary if we accept the view 3|sftcr «ra^» 3<TOfejR^ { See Note 1.), because in that case the word ^ cannot get the designation H*NT*T^. The %*&?$ in the present case is thus a device as described by the author himself in this passage. 2. ^ ^ f ^ q ^ f g r q ^ ^ . The basic condition for q^^f is that all words must have the same form (cpR^T). The sfiTbrr^ H^rff?: is specially designed to obtain a form similar to the sr|pftf| ^ref^fir so that both coukj be reduced to the $ # f

2,8 1



Virtually, there is no difference of meaning between HW and since both denote the same item, namely, ^ # . Therefore, the question arises : why is the lengthy word *?^rf%, instead of the shorter word *i#, used to make the designation H ^ T ^ applicable to &$ ? The reply to this question is given in the sentence g^^qsriffqW-I etc. If fHf instead of w^l^k is accepted there would be no ST^IF*. The compound of ;g# and H^t^fif would be s?W*Ffa is possible only if the similar form nwif^ is accepted. 3C??r\cji %c?%q3?^\ The line reads in the MS as follows : 3T?<T^JT qRT^m ^ % r ^ i r ^ f 8**T$fs ' Without the device of ^^m it is possible to accept that the name ^r#?rT£F^ l s applicable to that which is denoted by SRjf^r (i. e. fe^j, etc. ) \ However this is obviously not intended by the author of the Dip. He wants to emphasize the fact that without the device of vj$3w it is not possible to apply the name sHfcrra^ to e t . After the »p»%^r is accepted the word *?# becomes H3t%STS^r^? * expressed by the word H3lf^' and therefore gets the designation ^ n f a r ^ . A more proper reading would be : ^ ^ 3.

etc. The Ms reads srsjrsrsrn- However, if that reading is accepted it is difficult to construe it, since there is no other word in nominative plural which can be qualified by 5re?T*?5TT:. We have, therefore, changed it to STcqiH^n^: and construed it with %rfr*X. Similarly fir$p*n!fafir has to be understood as

1. Since sr?qq^pJr is always conveyed by ^jpft%, it must be indicated by some word in the constituent analysis (fejjf). This important feature of the constituent analysis of Sfcftff is illustrated in this passage with the help of an example. 2. q- % fen *TRfrs^f=r etc. The constituent analysis of the compound fecg is given as f ^ n *TT3: 3^T ST ' one who is possessed of variegated cows.' The 3T?3Tq^p} is conveyed by the form ^:. If the pronominal form conveying gT?^r^l^ is missing in the constituent analysis, it will denote only the meaning of the constituents. For instance, if the constituent analysis of f^rg is given as f%^j TTI^ts^T ' his variegated cows,' only variegated cows, and not their possessor, will be denoted. 3. f^q^TT etc. The constituents of ggsftff are always subordinate to or qualifiers of sfrqq^p}. In the absence of a word for sTF^q^pJ they become principal.


Mahdbh&sya-Dipika, Ahmka


{ 2. 10

1. It has been already stated that the word ^rrfHw is understood as a simple gipftfir ( 1.6). Now Fat, introduces an objection, namely, that the wrrd *r# does not get the designation ^#?iwi if the word ??iif^ denotes &p:i— : , This objection presupposes the view ^Ipftfr ^ ^ g ^ f w ^ R ^ (see the summary). The objection is discussed in this passage with two illustrations: ?T3T<pqf and ferjj. 2. s m p s m F R etc. The word *rs?<pq is a r^m meaning SIST; p q : king's servant.' In the r^^p*? compound the latter constituent is always the main member ( ^ ^ ^ ! T O ? 7 ! W 5 ^ : ). In the present compound *p*f is the main member while ^nr^ is the subordinate member. If the word ^foTtpq is accepted as a case of srf ? ^ ? ^ ff% ' integration in which the constituents give up their own meaning,' the meaning of XV3F{ is not understood In connection with the action with which the compound *i3^p*f is connected. For instance, when it is said ^fjp^inqR ' Bring king's servant • the servant is brought, not the king also.

sr^c^P-rf I% : » an integration in which the meaning of the constituents is not understood. In some formations the meaning of the constituents is present in the meaning of the formation whereas in some others it is not. Mbh. on Var. 2 on P. 2.1.1 : qi% STfr^rm^ f%: STsrg^C5?R^fcg% *pTO?^TR*R V R % . Kaiyata on this Bhasya passage explains the term as follows: sr^% q^ifn ^rpl ^m\ *n ^ ^ r ^ T ^ . In fact, however, in case of fT^pqf the meaning of the constituent is not totally abandoned; it qualifies and therefore restricts the meaning of g^r. 3. 4. sreTS^. The constituents in the compound have no separate existence. A compound once formed is treated as a whole. Therefore, after the word fjsr. is compounded with tpq: there is no word to convey the meaning * king.' The meaning ' king ' is thus s r ^ r s ^ * without a word to convey it.' Correction in the Text! *r should be added before fw^T^%. Translated accordingly.

1. The reply to the objection, namely, that the word ^ does not get the designation jgsh?*?^, is given by Pat. as follows: %w $\q: \ Hlpftlt cREFTCJf^iraft : * This difficulty does not arise, because in | compound the ( meaning of the) qualifier is understood (in connection with the action ).' This reply is elaborately treated in this and the following two passages by the author of the Dip. He points out that like the suffix conveys various kinds of relation between two objects, one,

2. 22 ]



denoted by the qoalifier( s) and the other, understood from the whole formation. The w?W?T# or the object understood from the whole formation is always connected with the action. However, sometimes even the object denoted by the qualifiers) participates in the action. This participation of the subordinate into the action is dependent upon the kind of relation conveyed by the formation. The author further observes that it Is the very nature of words themselves that sometimes the subordinate is connected with the action and sometimes not. This is discussed by means of various illustrations. Although the author here discusses the notion of ^[goi^fir^R and treats it as a reply to the earlier objection, he does not cite the qgfa from the Mbh. in this part of the Dtp. Moreover, this passage introduces the reply with the word sm ni ' or rather f which suggests the possibility of some other reply preceding this. This leads us to a doubt whether some lines containing reply to the objection ( between 1.8 and 1.9) are missing. 2. 3Tr**f W etc. Tar. 5 on P. 2.2.25 : wm 3T <#??? ft*RT^ states T that the «r|5ftft taught by the first rule ( P . 2.2.24: a i ^ q w n ^ ) conveys the same sense as that of the taddhita suffix *Hfq^ P. 5.2.94 : cr^^qrc&fclfcft agH teaches that the suffix * g n is added after a nominal stem in the sense «that belongs to this> or < that is in this.' In other words, *jg<^ is a possessive suffix. However, the possessive relation conveyed by * y ^ is of various kinds. In the word *rfrn^ * owner of cows' owner-owned relationship is conveyed by * g ^ , while in the word <tf*Z\ ' a man with a staff' the suffix ^k merely conveys relation in general. On the other hand, the same suffix %fk conveys inseparable relation in the word f^nfii^ * one having horns ( i . e . a bull).' What is true of the possessive suffixes is also true of

3. q ^ q ^ M f t . The word ^j^MVfe^ literally means «a person possessed of a sacred thread.' However, the word is used only in connection with a person who wears a sacred thread on his left shoulder hanging down under his right shoulder. Any person just having a sacred thread either in his hand or on his neck is not described as q^tq^%^. Similarly, f?rfrf^ is a person who wears a sacred thread on his neck (like a garland ). If the sacred thread hangsin any other position the respective terms canaot be used,

1. t& ftHI ir * ?$h. This incomplete line is to be completed as T I M 7ft <$m fasr. mma ^ * r g ? m f ¥ ^ ^ . The posses-


Mahabhasya-Dipika, Ahnika VI

[2*25 wftm or

slve suffixes do not necessarily convey permanent relation, Impermanent relation Is also conveyed.

T... This line is badly written. It is very difficult to make meaning out of it. Some words after smgiif seem to be missing. Again the next words %^f ^ gf% etc. represent perhaps the last part of some sentence the beginning of which is missing. The scribe seems to have dropped a complete line of the Ms from which he was copying. We propose tht following reading after reconstruction : mn q?n&rer <O tRT s?rp?4: I TT TOT glwr %3T * ! % % f ^ 3rcrtat (ikmfq) 1 ' The meaning of n ^ is in accordance with the essential meaning (of the derivation ). Just as the act of eating is not possible without mouth, it is also (not possible) without the pair of clothes.' This latter illustration perhaps refers to the statement £ * Give food to the one with white clothes.' The white clothes which are subordinate to the $r?*?qp»T«r (namely, the person wearing them ) are also understood in connection with the act of eating. In other words the white clothes are a necessary precondition for the act of receiving food. Only the man who wears white clothes is given food. Therefore the compound g^m*?^ is an example of ^gamfe^TR. Pat. has given and sftff afajfta as illustrations of 3. a ^ ^ a ^ m f e c ^ r a ^ ^ J T . The clue for understanding that a certain ^|ift% is 3 ^ y w f ^ H is the maxim <T3T*U% cr^*?T3: cF^n% ft^m^i ' It (the action ) takes place when it (the guT) is present, it (the action ) does not take place when it (the go?) is not present.' For instance, if the gm £f£ ^creRl (white clothes) is present, the act of giving food is possible, otherwise not. On the contrary the act of bringing a person possessed of variegated cows takes place in the absence of variegated cows (ferr *ft) which represent g^. Therefore, the latter compound ferg in the expression is an instance of s m ^ tft

The scope of ^ g o r ^ f ^ r R is here extended to such expression in which the gof is present at some other place and time. For instance, if the maxim ^5TT% ^ * ? F S : etc, is accepted, in the sentence 3: sp^rey ss^WH^, gipl?H3l cannot be a argarcrt^ra since the g«n 5 ^ <n&^ is connected with However, there are sthe past action of seeing ( S H : ) , not with the present. certain other expressions in which the gar associated with some other action, place or time is also connected with the present action. In the expression Sr f i ^ T f r ^ T SET^I^ ^Thf^, for instance, 7»^TT which is the g^ is connected with the past action of seeing. However, %qr^r is also connected with the conveyed by the verb




1. Even after the compound ^ r f t f t is understood as a n"^piHfit5n«T qggt%9 the word ^r# may not get the designation H ^ ? * ? ? . The objector now offers another interpretation of the rule and shows that the problem of non-inclusion of ^ in q^ffe list persists. According to him, the rule means that fw^ etc. are called *T?TOT^ when the word ^ is uttered near them. Thus the word T^ is meant to mark fw^J etc. by its proximity. The objector means to say that even after understanding the connection of go? with the action, it may not directly partake of the action. In order to support this statement he has offered two illustrations which will be explained below. 2. -*FF?#?€t?ft *?n*cra;i T h e w o r d Hmn^foftfNt perhaps refers to a kind of ascetic wandering with a snake on his neck worn as a sacred thread (for details, see V. P. Limaye, ^m^sff^f^Amrtadhara : Professor R.N. Dandekar Felicitation Volume, Ajanta publications, Delhi 1984, pp. 509-11). When it is said, ^iiFTsfftNfcft sfcysmi^ it is understood that the snake is present when the action of giving food takes place. However, it does not partake of the food offered to the mendicant. In other words the mtj is the characteristic of the person which qualifies him for getting food. 3. ^s?£r*pr *rtera. This s r ^ compound ^ ^ p 1 can be derived according to P. 6.3.9 f^?gn.rT STH*3T: tf^rrcm only if it is a name. The word is not attested elsewhere. From the content it appears to mean that a man with an infant on his shoulder is offered food by way of charity. When he is eating the food, his son does not partake of it, because he is an infant. The <p" who is the mark of the man and qualifies him for getting food, does not partake of the action. 4. If we read ^ ^r^p 1 * after the words ^pft g% the parallelism between the two examples becomes complete. For instance, the singular ending in the verbal form ?p% implies that the agent of eating is one person. However, the agents of g^% in sr^pft *ff% are two, whereas it is only one in the sentence

1. The objection stated above is refuted by again referring to the inherent nature of a word to convey different meanings in different contexts. The author now cites another example w$$Z *rtaq which is parallel with the earlier one, namely, ^ ? % ^ Wtm. The expression *T^p ^Tterr ' Give food to the person along with his son ' implies that the t p is not only the mark of the person; he also partakes of the action of taking food. The action of


Mahabhasija-Dipika, Ihnilca


[ 3. if

ta» : food conveyed by the root yp; thus involves both, the man and his SVJL .n this expression, whereas it Involves only the man in the earlier expression. This difference is attributed by the author to the nature of word itself. However, it is not directly conveyed by the word. 2. qsn g q»%3r- The partaking of food by both father and the son is not directly understood from the expression srf^w v}br*> since the object is mentioned in singular ( ^ j w ^ ) . However, this is a special feature of the meaning of g ^ itself. In an expression like sr^pft * p % the use of the singular form 3jf% is semantically understood as dual. The author gives an analogy of a person eating with two mouths. Imagine a person having two mouths. When he is eating with both the mouths, we still use the singular gryaatf g^T^f JJ^rK. In the same manner the use of the singular ¥jl% results in the dual sense, The word ^ (in the Ms) in the last line is redundant.

1. Here the author sums up the earlier discussion and establishes the conclusion that if ^ f ^ is understood as a cT^pmf^TFT 3§sftf|, the word srst gets the designation *T#^FT^ He has repeated the earlier illustrations* 2. 5Tc^rg??TWT«?f etc. The line in the Ms is difficult to read. It actually reads as follows: We suggest the reading as follows; ?#3T^sfq Jlforoilr I Here % and ^ appear to be redundant. Anyhow, there is a clear reference to 5T?^^r and SPJSTR. Does the author mean to say, as the reconstructed text given by us implies, that the designation spsNrsR; is directly applicable to fir^" etc. while it applies to ^ through direct knowledge as well as inference ? But this is not correct. If the rT^g<rofiRFffi' view is accepted ^ along with fqpsr, etc. is directly understood from the compound srgffi[. How does srgjRR come into the picture ? That is not clear.

After establishing the solution of 3^goKt&5TFT in the foregoing passages the author of the Dip. now proceeds to offer two alternative /solutions. One of them is dealt with in this passage. In consists in interpreting the word snf^ as ' near/ The compound *nrff^ would then mean 'that to which ^ is near.' Since the word sr# is near to the list as a whole, the list as a whole aud not the individual words will get the designation *r#wr^. But then the definition will be redundant, since no grammatical operation caused by ^t^rqf^^f will b^ applicable to the list as a whole.

3.25 ]



The definition being thus redundant is applied to each Individual word. Like eic. ^ also ift understood as applies to ?^f. This solution Is not attractive. It is not clear why the author lias indulged in offering such weak solutions after establishing the ftnal reply in the form of gnrgarofirarR.

Another equally weak solution offered in this passage is to understand the word snfr in the sense of * kind.' The author has not further elaborate ed this point. Perhaps he means that in this view j^j- becomes a qualifier of fihJ etc. and therefore the rule means that fip<j etc. which are ^ t f i | * like ^ ' are called H^m^L Since the word ^ # also is ' like H # * it gets the designation 2. spjn ^^infrq^n *HF?:. Apparently this is an illustration to explain the above interpretation of the rule. However, from the way it is explained here it seems to be more fitting in 1.7 where the purpose of the mention of ar^q^pjj1 by means of a pronominal form in the constituent analysis is discussed. The discussion of this compound ^ntJfimPRT: points out that the main meaning, the 3T?3q^p*, namely, ??^R: ' nights' is not understood by merely saying c^inft q^nfi" * 353*?Tnjt ( i s ) the fifth (night).' The word angr^ must be added to the constituent analysis, which would then mean * of these nights ^qiofr is the fifth ' ( ^ ^ I # q^fr STTSTT'O. However, if we compare this discussion with similar discussion in 1.7 we find that still the constituent analysis does not properly contain a reference to the sp^jq^]^. We have to add the pronominal form 3|: in order to understand the s^qq^sM^iuft q^4V 3THT W* (?HHJ: ) * those ( nights ) of which ^^uoilf is the fifth.' In the present context the illustration can be explained in a rather farfetched way as follows: Just as the ^^FJft which qualifies all nights is also understood from the compound, the word ^ which qualifies f ^ j etc. is also understood from the compound flcfif^- But in that case this interpretation of £R?f^ is not much different from that in the ^gtrwfiCsTH view. 3. w ^^TT^TT^T^T'- etc. This is another illustration to explain the alternative. It is also as vague as the earlier one. Perhaps what the author means to say is this: When an order is given to bring brahmins from Devadatta's house, Devadatta also is brought either because he is a brahmin or because he stays in the same house. The word qft is, in the same



Ahnika VI


manner, understood from the compound H^TT^, because it is also of the kind of H # . 4. srwjq^re... | The translation follows the emended Uxt , to be so read in the Text, p. 3,1. 25.

This passage contains the author's concluding remarks. 1. sq^jTit r ^ i % r ^ i ; etc. The author here refers to the objection raised against the meaning of *m\fe by assuming the meaning of mfk a s 5 first? and * near.' The earlier (osra^JT) meaning refers to the first objector who, assuming the ^s'qorrscf^rra view, argues that the word ^ r does not get the designation grsRifi^. This objection (^rf^T) Js refuted ( q $ p r ) by accepting rfTg«wi%sTR view. As to the second meaning ( s n f i q g ^ ) , the text of the Dip. discusses it only as an alternative solution (see 1.15). There is neither an objection ( TJY%F ) nor its refutation (<rf?pr) in this connection. Does it imply that the text of the Dtp. containing these details is missing ? 2. ^ g i T O f ^ r c i ^ - e t c . The Mbh. on Vat. 22 on P. 2.2.24 contains reference to the two views Hg^FT- ^RTWJ"- ' the meaning of a compound is the collection ( when a part is referred to ) ' and ^ f t sipftfT ^ g t n ^ f t ^ R l ? ^ fFor details see discussion on the compound sr^^ft^T: in the Mbh.) Although according to the author both these views have been mentioned by him, there is no reference to the view i *rg^T3: H^FTHPT.'' found in the present text of the Dip.

Summary : The objector in the Mbh, asks : ' Why is the rule of die* not applied to the word ^ N m ^ J ' P. 8.4.3: < s j h ^ *T^TRTO*n teaches that ^ in the second constituent is replaced by ^ if the cause of ur^" ( namely, either ^ or \) exists in the former constituent when the compound is a name. The word H#3TW3; fulfils all the conditions stated in this rule. Three arguments are first stated against this objection. Finally the objection is admitted and then refuted.

K ^ 3 ^ ^ R j J i ; . This sentence introduces the first objection. Grammarians have accepted the maxim : srsfrers^Ht ssprfM 3 T ? * the words 5 *f




used as proper names are not to be derived/ See the Vrtti on the Vak. I. 44 : ^r^Fr^Tsi%<J 3T^R"RWf TT^r^Tci;. In the case of proper names constituent analysis is not possible. Therefore the analysis of ^sfrsr^ into constituents cannot be made. Consequently, the m^ rule does not apply, 2. ^f%c£i?FR; ![% I We cannot deny the obvious constituents in. popular names like ^PJOMT (53?T and TJ?CJ), ^ ^ T H ( ^ * and *TH ) etc.

1. Correction I The ^*rg after =rsF5353; (in the Text p. 4,1. 6) should be removed. 2. #r% s r ^ r s T H * ^ . The Var. 1 ( Mblu, Paspasa): % | ; sa4"5rg% S T ^ s r ^ ^rr#3T v*7T#nOT * when the word, the meaning and the relation ( between word and meaning) is eternal and the usage of word is determined by meaning in accordance with popular practice, the science of grammar restricts (the usage) for the sake of religious merit,' The rules of grammar are thus applicable to the words, the meaning relation of which is already established. The relation between the word sHhrm^ and its meaning is not already established; it is to be established by a grammatical rule. Therefore, the rule teaching oi^r does not apply to the word 3. f f3<J]r=ci *T^rT SfTftfe^proir^T- E v e n if it is somehow admitted % that the rule of aicW is applicable to the word *NNm^, there are difficulties in applying it. There are two possibilities of m^ in the word *rf"?ro^: ( 1 ) cricW, which has already taken place in the name spfcrr??^, is simply justified by the rule. ( 2 ) or^c is effected in the name sHNr*r^ where it was not . there before. The former alternative does not stand, because u\^ cannot take place unless a word is a ^ T and the word *T#?TO3 is not a ^ T before the rule defines it. The second alternative also does not stand, because, effecting u i ^ in a word where it is absent would mean a tacit acceptance of the impermanence of words which is against the grammarians* principle of eternality of words. ( For details regarding the principle of eternality, seo note ( 2 ) on 2.3 below.)

1. &^$ ^ firfH^^: etc. Here the author tries to justify the absence of OT^ on two different grounds. A proper name is based on only ^^cr, . its own phonetic form. One is free to choose any word as a proper name. In the present context, whether the word ^Hfcfl^ should be regarded a compound or a single word depends upon the desire of the person who chooses to use it as a name.


Mahabhasya-ijipikZ, Ahnika Vi


2. qp?pT^?n. According to the philosophy of grammar, the sentence Is the meaningful unit from the point of view of ultimate reality. Thus even according to toe theory of ultimate reality accepted by grammarians the word ^#?n*?^ cannot be split into constituents since the word itself does not have an existence. The question of oj?^ therefore does not arise. The theory of ^ P R ^ t i is expounded in detail in the VP. 1. 73. Cf:

1. Three arguments were made against the objection of m^ in the three passages above. It is re-established in this passage. All the three arguments are subsequently refuted. The first argument that the constituent analysis of proper names is not possible is refuted in this passage. 2. ^n^fsq^fTT 1 ^f%cfx etc. In spite of the theory of grammarians have to split words into constituents wherever possible. Not only are words like n^Tpf, the constituents of which convey correlated meaning, analysed, but also words like ^TR^, the constituents in which convey a totally different meaning are analysed into them, jfr* and ^ ? convey meanings which are totally absent in the word ifiJ^R which means a wild monkey. Still they are stated as constituents of Words are analysed into four classes, describes them in the Mahabhasya as follows : ^rgs^ft 3 I ^ t f 5Tff%: jpiST^j: %^RI^T ^srers^Tsrgsfr: l ( B S n ^ p r * ^K^M ' P- 19 ). The ^ are different from the other three in that they are based on their own form. They have no $Tff%f?rftTTT other than their own form (i. e. phonetic sequence). Sometimes a ^ I P ? itself is constituted by two or more words which belong to other word-classes. For instance, the word rniis? which means a wild monkey contains the words rfl? * white ' and ^ ' hard' or * a donkey. ' Both these words are thus ^T^ra^pT, entirely different words with different meanings. Even then the constituent analysis of ?Ti?^ is given as shown above. [ Correction ! In the Trans., p. 39, 1. 6, read based on for conveying. ]

1. The division into <^tq^ and ^rTfq^ is possible not only in popular names, but also in grammatical names. The grammatical terms do not behave differently from the popular terms. In fact, the language of the Sutras must follow the rules of grammar taught by the Sutras themselves.




p, 2.3.11: ^f^f^ferf^T^" ^ zmtiri is a good example of the application of the rule to the language used in the rule itself. The rule which teaches the use of the ablative in connection with the words srfwf^Rf and srf^R, itself contains the use of the ablative in the word WF?cI> Cf. -dd^uVd on the %mi-

1. It was stated earlier (2.2) that the name-named relationship is artificial, that this goes against the theory of eternality of words. To this objection answer is given here. Another theory accepted by the grammarians is *?sr ^Tsrer^Sj: ' All words convey all meanings.' Accordingly, the word *HNr?Hv conveys all meanings. The rule ^jVrf^r ?T^RT*TTf*f, however, restricts its scope and allows it to convey only the 35 words listed in in the domain of grammar. Thus the relation between sraif^s and is already established (%$*;). Consequently, the rule of u\^( can apply to this word as any other word in popular usage. 2. ^ 5 T % ^ f t . . . Two views regarding homonyms are prevalent among grammarians : ( 1 ) CT-^JRH^^T, * the view of one power.' According to this view, a word like ^ conveys various meanings like hand, trunk, ray, tax etc. through one power. S T % is the power in the word to convey meaning. ( 2 ) sr^^rf^r^x^ $ view of multiple powers.' According to this view, there are as many powers as the meanings conveyed, by one single word. All these powers reside in that word. In the present context the restriction of the scope of meaning of the word H^Tffi^ by means of the rule srar^tfif *HfaT*nftr does not contradict either of the two views. Both the views tacitly accept the eternality of words. Since they are not contradicted, the view of eternality of words also remains undisturbed.

1. sr?^qf ^frn3> Here is a diversion. The author has now entered into the controversy regarding the eternality of words which was referred to in 2. 2 while arguing against the objection of nircC stated in the beginning of this section. In this passage he refers to the view which is similar to that of the ?T^^if^F>s. They speak of a twofold relation between words and meanings. The words that are in usage since Vedic times have an eternal relation with their meanings. It is known as f^sHfor 'divine convention/



Almilca VI

[ 4.25

'However the relation of the words like ft, g etc., which are newly coined along with their meaning, is artificial. This artificial relation which is termed as sngfira *T%3 (or sro?f) is not permanent. The line q§Nr...srf§rqr^riSai is difficult to construe. Perhaps the author compares the knowledge of the name-named relation with the knowledge by inference. Just as the knowledge of fire is dependent upon the knowledge of smoke, similarly the knowledge of the name-named relation is dependent upon H*T3 conveyed by rules like ^T^f^r ^FfcTOTft.

1. Here the author refers to the view of the *ft"*TO33s regarding -cternality of words. Unlike the ^rswttf^RSS the 7fcrfeT«&s hold that all words including H^TTS have eternal relation with their meanings. Ancient sages like Panini merely remind of and establish some of the already existing relations. Therefore their works are 2. ft^&fc % *icWra^. The purpose of the m^f rule is to discard the form without m^g. Once it is accepted that the function of the science of grammar is to explain the already existing words and their meanings, the question arises: If the words and their meanings already exist, what is the use of grammar? Kat. has answered this question as follows: fite*?^ ^H^fefrT % fa grammar %fecN5?WFi; fa^ ( Var. 10 on P. 1.1.1 ). The main function of g is thus to discard the incorrect forms by establishing the correct ones. 3, r3p^%tt5TTr3)5crf ST3Ji%S (*N* ^ o n t n e Pratyaharasutra rlk9 1.20). This Varttika is eked here to confirm the view of the iffcrf^s that the words are permanent. The Var. says that it is proper to use words like 3|^fcp (and not ^w%) as names as they are found in the usage. In the present context Panini also is expected to use sHfalH^ which is already established.

The Var. ^famH^VHT fkwwww&piiiR: explains the absence of #• replacement in the word on the basis of its being a f^qr^f^T * readymade form.* The objector, however, raises the same question suggesting that an incorrect ftq!^ should not have been made. The author explains the choice of the form without crr^r on the basis of usage. Since the word ssbnu^, and not H^Jira^, is used in popular usage also, the revered Teacher uses the same form. He does not coin a new word. He simply reproduces the one already existing in the usage.




1. Here the author puts the objection in a difTsrent way. The fttrmST 1 3HHT4-H ' can cancel the o ] ^ rule only when it is equally powerful. But it is not so. On the one hand we have a full-fledged rule ^ q ^ R ^ W W : which directly teaches cn^r, while on the other there is a 3?~rj rule srwrftfer HiffiTsnfer which is engaged in conveying name-named relation and only indirectly indicating absence of cn^. Thus due to the absence of <pgWr3T * state of having equal power' the ^ i f j ^ cannot cancel o i ^ simply by means of frpjHR. This objection is refuted in the following passage. 2. ornament, on a rule etc. (Mbh. ^pq^R^^TT- * Fully equipped.' The word 3nTC3n? literally means decoration etc. In the context of grammar, the ^ q ^ r c is effected by supplying words with sargff%, examples, counter examples, q^qsn- I. 11:

1. The above objection is refuted here by showing that a reference by f^xncHT is <pgN<a with a full-fledged f^f^r rule. It is accepted by grammarians that 3qft»Ctfl is a ^ J J % (cf. VP. L 141 ). Therefore whatever is taught by grammar is not an authority by itself. ^ % always follows ^fgr.. The statements of grammar have, therefore, status of SPJTTH ' inference ' rather than sr^T ' direct knowledge.' Therefore it may be a oi^" rule or a faq^R, both have equal status and, therefore, can be described as r^R^r. 2. The comma after v\v&i [p. 5, 1. 14) should be removed. Translated accordingly.

nf% ^ srftf^R 3^4 f?rqTW^l Panini's grammar contains both ftfRr rules aad srft^g rules. A srfit^vj rule directly prohibits the application of a fkfk rule in such a way that the i^f^ rule is cancelled once and for all. However, a fitqm^ cannot be regarded as equivalent to a negation, because it does not permanently cancel the application of a %Ry rule. When the word ^ f a r o ^ is mentioned, its correctness is conveyed but the optional status of Spshn*?^ is not totally discarded. The objector thus implies that a which is an indirect negation still allows the optional application of the rule (which is not desired).

1. *? ^ 5rf£nHrcq \*£: ... i The view of ' another' is that both and 5n%^r are equivalent. By f^qRH one infers non-application of a rule. In the same manner negation implies non-application.






l ' removing a nail by means of a nail '. In order to remove a. nail pinned into something another nail is pinned from the opposite side at the same spot. ( of. Mblu on P. 2. 2. 6 : ^repjT $V® 3TTfsrcrre: 5(m*it#> f^ffl?rT l) In the same manner ooe can say that a negation removes the application of a f^-fer rule which has already taken place. However, this Is not correct. Non-application of a liffsF rule is inferred from negation. A %% rule applies excluding the domain of negation. Therefore both negation and ftqTcR are on par. Thus there is no question of removing the application of a f^fvj rule that has already taken place. 3. In the Text (p. 5, /. 20 ), a ^ Translated accordingly. should be added after ft

The similarity between ftqicPT and d a ^ l is further established following Patanjali. STf^fftai fipfer^EfT etc. A fkfc* rule is a general statement regarding something, while a srfgftsr rule is a negation in a specific domain. P. 4. 1. 5 : sj^wfr # H teaches that the feminine suffix NiP is added after any stem ending in r and n. However P. 4, 1. 10 : ^ ^^l^lf^vq: cancels the application of this general statement in the domain of stems called sas and stems belonging to the svasradi list. A s r f g ^ is thus a restriction by prohibition of a fi?R|.

1. m&m *?teq?3T 4if^^t 5T- ... The 5n^T3i#if^^:qr^ which is often referred to in the grammatical commentaries is explained in this sentence. The sentence is an injunction : 'Give food to Brahmins but not to Kaundinya \ It is here assumed that Kaundinya is a Brahmin. There are two statements ( 1 ) Gi\'Q food to Brahmins. ( 2 ) Do not give food to Kaundinya. The former statement is a firfe whereas the latter is a srR*^. Both the statements are applicable to Kaundinya. According to the first he must get food, because he is a Brahmin; according to the other he must not get food. In fact, he is not given food according to the second statement. Now the objector asks the question : * Why does the second statement, i. e. sr%^r, become more powerful than the first one while interpreting the whole sentence ? * In the present context the rule ^q^T^... etc. is a fefer rule teaching &15-T, whereas the f^rqi^r s?5hn*T^ implies *n%%^ of ai^r. It has been therefore stated that the f^qirf^ cancels the application of airor rule. In other words the jrfir^ra in the form of Hq^ffi ]S more powerful. The objection is

6.6 ]


therefore against this preference for srfartter. This has been already observed in the case of ~C?TTF and sjq*r? by Pataiijali. Cf. Mbh. or> P 1.1. 63, 6 : i 4 ^fqcTi^f sTftTWfSr^r?^1 qanj|rfr*rii i * First the exceptions enter (In the domain of application ), then do the general rules \ 2. 5nj[ta^; I The reading ^fl^Rt is proposed by us. The Ms is not clear. The form, which is a denominative verb, form would mean * is treated as mild \ Prof. G. B. PALSULE has suggested another reading which is more acceptable. 3. In the text {p. 5, /. 26), the ^ Translated accordingly. after ^ # l # should be removed*

1. There are three other possibilities when both fgfir and become applicable in the same domain : ( 1) cancelling of srfa^vT by ( 2 ) application of both, and ( 3 ) optional application of each. The first is Cot possible, because if it is applied the g£&vr will be futile. It is the purpose of the ^ ^ R while stating the srfci^ that it should be applied. The second alternative is not feasible since one cannot apply both firfv? and srJ3%^ together. One cannot, for instance, give food to Kaundinya and at the same time abstain from giving food to him. The third possibility is equally out of question, because once the srf^H has cancelled the possibility of application the question of optional application does notarise. Therefore, finally we have to accept the remaining alternative, namely, cancelling

of ftfa rule by
etc. Two word safter %^ are missing. The line is difficult to interpret. This passage does not contain a clear-cut answer to the objection raised in the earlier passage. It is implied that since the three alternative possibilities fail, the remaining possibility, namely, cancelling of frfvy by means of g i n ^ stands. Further, the last Hoe, which makes better sense if it is read as $cTT«iaawGr ^ftsraf^fep suggests that the sf?r%w is more powerful because otherwise it has no scope, f^r^ absence of scope' is thus the basic reason behind the preference of Anyhow^ all this is simply speculative. 3. Correction ! Line 4 : *Hr should be replaced by <^r in the Text. Translated accordingly.

1. Here one more answer to the question about the preference for is given. When people utter one and the same word at different






times, places and in different contexts, it is the same word only so far as its phonological form is concerned. Really speaking, words in different sentences are different. The word wmw, for instance, is phonologically the same as any other utterance of the same word, but the word p?rpJ! in * mWR 'ifr^f?5li^? is different from any other word of the same phonology in another sentence. Because of the phonological sameness one understands the same meaning everywhere. However, the speaker intends a different meaning. For instance, the listener understands that Kaundinya also is included in the meaning of wiW^< However, the speaker means by mWi an Y ^W^ b u * not aprf^ssr. This being the case the statement sTl^f^T ^ s ^ T P ^ means * Give food to Brahmins other than 4ft*&*t%. The question of preference for 5 " % T thus does not arise. Tm v ^ ^ U ^ r ? etc. The word sn^TO in the sentence is not different from any other utterance of the word. It is just the same common word s t T T By word its phonological form is to be Hg O * understood. The word 5521 is attached to 5T%qg: by the scribe by wrong repetition. Perhaps jrf^rg: p*W%qT^ftqmT: would be a better reading. The passage is translated following this reading.

One more attempt at explaining away the priority of gfa^v? is made here. Even if it is accepted against the above mentioned maxim: sr%3!?*T fJren: 3T^T:, that the word is the same everywhere and it means the same, the application of sf^ter is automatic. For instance, when the speaker says STTfrou s f t ^ ^ T ^ the word sn^roT is used by him in the same general sense as It is used elsewhere. Therefore, it includes € f R l ^ . However, 4»lf*¥»l is meant to be excluded in the interpretation of the complete sentence 'sTTHtJir A ^ ^ n ^ *frfj^fr ¥f ^rfir^sq: i ' This is explained with a popular illustration already mentioned by Patanjali in his Mbh. on P. 1.2.39 (I. 212. 5 6 ) : 'just as a person intending (to buy) fish brings fish along with bones and scales, and he takes whatever is acceptable and leaves bones and scales.' The line in this passage of the Dip* ^re^: ^ ^ . . . etc. is reconstructed on the basis of this Mbh. When the speaker uses the word sTifTOT, he cannot use it without including #rrfe?q in it. However, in the complete sentence the word again means ?n^v\s other than

[2.16 ( / ) ]
1. *T%mHV cTfUoi^r. The knowledge of ^ r f ^ q comes from two Words in the sentence m%W\ vftofir^T:.- etc. : ( 1 ) mi* 0 ! a nd ( 2 )

6.22 ]



The koowledge from the word STTIRUT 1S v a § u e » doubtful, because the listnef is not sore whether the speaker intends to include ^rf**f?3 among STT^JIS or oot. However, his knowledge from the word ^ r f e ^ is direct and clear. This is therefore STR?^ while the former is snjsrnr, inference. Evidently, direct knowledge is more powerful than inference, therefore the negation first applies to ^rfe:?i. 2. ?rg ^ cFFlF^Tc!; Qtc An objection against the view stated above is first raised and then rejected. It was stated there that the knowledge about ^fn^Rf from the sentence snip^T *rfcrfe?«*n: is through inference, while it is direct from the sentence s ^ f ^ ^ t ?f sfrsFfircTosr:. Since there is no equivalence between these two kinds of knowledge, preference was given to the direct knowledge. Thus srf?r%W becomes more powerful. Here the objector points out that the recommendation of giving food to ^Tf^^l is due to his being a mw®i rather than being efrftrgrq. The meaning s^u\ is directly understood. Therefore the fsrfa ia which it is directly understood can also be regarded as more powerful. Both the statements thus being < p ^ w , option must be applied. snfir. is not word, ledge 3. The answer to the objection is according to the *ft??mT concept of srefifsqTE^rgr: * T O I ? ^ etc.: When a word denotes a class, the particular manifest in its meaning. Accordingly, the word ciTirai, which is a classdoes not expressly convey the particular, namely =fif*^?3?. The knowabout ^rf^?q from the word m^W is thus by inference, f is directly mentioned here in the

4. The word ft&pn^r cannot be construed with s r ^ f i r s ^ ^ : etc. Perhaps it is misplaced. Its proper place should be rather at the beginning of the next sentence, i. e. before

[2.16 ( u ) ]
1. The passage discusses the argument about improbability ( «KJ«TO ). Here the relation between firfe and gfgr<ta is compared to that between ScH^T and 3TTCT3- J u s t a s a n snreT^ is more powerful than 3c*PT> because otherwise its application is impossible, in the same manner a negation is more powerful than a positive statement, because otherwise it would be futile. aarci*ra is here equivalent to fTnr^jsr^ • absence of scope \

ftffa 2. qftffa fiffrnfacftft. w«fa is used here in the sense of s r ^ .


means that the negation automatically takes place first. Cf. Mbh. on P. 1.1.47 ( 1 . 1 1 4 ) : %^i «n3TTOfc sttfqr ^^TSPTTT: qs-c^^T^: s?mT m*&mn 1% I f ^fff ? ^rf%^Vs?t ^i?cr: I The priority of $rf?fqsr over f^f* is just like that % of WTOT^ over ^K&A, based on popular practice.


Arahabhasya-Djpika, Ahnika VI

{ 6. 23

One of the three alternatives in interpreting the sentence ^... tic, is, as is described earlier, flr^q or option. The argument in this passage is based on this * option '. If the whole sentence results in optional application of giving food, it should be regarded not as an ordinary option but a distributed option : osref^RrfirvnqT. It means that option operates in a decided manner. The domains of ftfw and S T ^ W are fixed and separate from each other. For instance, P. 3. 1. 143 : fir^TFn JTf. is treated as o^c^%rafg^TWT. The rule teaches optional uj suffix. The other alternative suffix is Mf%. Accordingly there will be two forms, m% and JT^. Since this is a o^%rrffgrsn3T, the form snf is used only in the sense of crocodile, while the form snj is used only in the sense of planet. In other words the forms derived by applying two options are in complementary distribution and not in free variation. In the same manner from the option that is understood from the sentence srniraiT ^fts^^T^ etc. two alternatives, eating and not eating, are understood. Of these the former is associated with STTIPTS whereas the latter is connected with ^ This way of interpreting the sentence as ssrwfercf£ft?M is not satisfactory. Strictly speaking, the alternative of f g ^ q results only in the case of cfifi^??. Giving food to cprfagsq is prescribed by the earlier part of the sentence, while it is rejected by the later. Thus two optional statements * give food to ^ f ^ ? ^ ' and ' do not give food to - f r f r ^ ' emerge. However, due to impossibility this option cannot be exercised. It has been already pointed out that food cannot be rejected to the same ^ i f r ^ i to whom it has been already given. o^feraf^WS! is possible if the domain of the application of two alternatives can be segregated into two watertight compartments. How to segregate sfifir^T who is the domain of the application of b*oth the alternatives ? The argument of s^femfevriqr is, therefore, questionable.

The whole passage is vague. If at all it makes any sense it may mean as follows : the srfo%v:i<=n=R (#rf^?ft ^ sfrsri^rsq:) is a sentence having a particular context. It is subordinate to the main sentence, namely, sniroiT *rt"iif£r3s:qi:. The meaning of the word efif^?^ in the subordinate sentence has no place in the meaning of the main sentence. Just as the constituents afrc and ^ convey no meaning in the interpretation of the word Tfr^nr, the meaning of the word cfrfj^sr also is absent in the interpretation of the main .sentence. Accordingly, the word 5rrp7iT: in the main sentence means all Brahmins except ^

7.12 ]



FrankH' speaking, we are not much satisfied with this interpretation
*»-*-" s "*•'•" .' " ' -"-~*^ -" J ."^^'H!^ it. firstly

never found to be used in the sen^c c>\' subordinate. However, through its association with zmm It has to be interpreted as subordinate, Secondly, the words actually occurring in the Ms are 4rpan; to make sense they are emended to *TtosTC« Finally, there cannot be the relation of main and subordinate between a firM and syftqg. The next line (in 2.17), however, contains the word %^F^f to describe the f^fw which indicates its superiority over the

The ways of interpreting fwfw and afa%W are different. This difference is illustrated in these two passages. The jftiqrai doctrine of sfffi=r is perhaps followed in this discussion. In a f^iRasn^q, although the whole class is understood from a word, there is no obligatory application of the fgrfsr to each member of the class. For instance, the f^fi? * mw\\ *fte3^rm' contains the word mw* which denotes the whole class of Brahmins. However, the speaker does not mean to feed all Brahmins. Even if only some are fed, the injunction is fulfilled. On the contrary, the negation c sffignjfr ^ f ? ^ j : f must be applicable to each Brahmin. If some Brahmins are not killed while some others are killed, the prohibition sniptnt *! fp^Rj: cannot be said to have been followed. In the same manner the statement *ijf^£?^ ^ vrfcrtSfceq: cannot be said to have been followed if one of the many e&rfn^pgs is denied food while his other namesakes are fed. wst^ : Cf. Mbh. on P. 1. 4, 23. ... This line is so far not traced. If ftf^ is not performed in some cases, there is no fault; but if by m%%^ total cancellation of action is not effected, it leads to fault. This maxim is not traced.

1. In this passage the author tries to show that when both an injunction and a negation with reference to a particular action are stated, the preference for negation by the listener is not his arbitrary choice. It is pointless to merely say that ^rf^?q is not understood by the word gr^Ji, because a person can accept either of the two alternatives. The choice of one of the two alternatives must be accounted for. For instance, first the master orders 1 go ' and then says * don't go.' The obedient servant, who is unable to carry out both the contradictory orders, prefers not to go, because he thinks that there must be some purpose in the mind of the master in cancelling the • order he had already given. Applying the same logic to the jrfSfaar here also


Mahabhazyct'DipikG, Mniha VI


one can say that the listener finds more meaning in the negation which follows an injunction. The passage contains two illustrations. In the first one there are two speakers, one saying * do this' and another saying s do not do this'. In the second illustration one and the same person gives both, positive and negative, orders. It is difficult to make out what is the difference between these two illustrations. Perhaps the author means to say that when there are two speakers speaking two opposite things the listener may accept either. However, when one and the same person gives two opposite orders the listener ponders over them and accepts the one given later, thinking that there must be some purpose in the mind of the speaker to contradict his own earlier statement. 3. ?p^ f^ref^n sf?RT%T?i;— It was earlier stated (2.16) that it is impossible to carry out both the injunction and its negation. However, so far it is nowhere found to be stated fe^fVr * in detail.*

[ 2.20 ]
It was pointed out in the previous passage why the listener accepts the negative statement of the speaker who has already made a positive statement. However, there Is still some scope for doubt. People may commit mistakes in their ordinary speech. Therefore, one cannot always apply logical rules and interpret their speech. For instance, when a speaker says two opposite things one cannot with certainty say that it is the latter that is intended by him. May be the speaker himself is in a confused state of mind or in mental disorder etc. However, there is no scope for such a doubt when the speaker is no less a person than the great sage Panini. U he uses the word sarvanaman which contradicts his rule of u\^, we have to admit that the non-application of m^ is intended by him in the term HsfaFR^. -For ^¥fo%RTft| (said of Panini) see the Mbh. I. 39.10-11.

1. It was stated in the previous passage that from Panini's fiftieth such as sHhn??^ we infer non-application of the rule about oir^f. However, one can equally infer cancellation of o\<% that has already applied. In order to clarify this point a distinction is made here between two kinds of sr%%v;j. When a sf3<N is made of what is called srste^T sjff%: ' natural inclination caused by some purpose ' it is a rejection of something which one is already prompted to do. For instance, the statement that a tame cock should not be eaten isTOW*f?i%w,because it is a prohibition of something which one is

prompted to do by nature. On the other hand, when a ^TS^^m JT^ftt: "inclination to do something upon somebody's words' is negated, one merely infers absence of doing it. Any sastric injunction which is stated in the form of a { fy]e is an Instance of ^ ^ m s p g f % . When the rule <i<tq^ic*?^R?m*Y: \ for instance, is taught, one understands that tu^ is to be effected. Before that one Is not naturally inclined to effect it. In such cases from a $fi=fq\* like the f^qTW^f of the word ^T^TO^ one simply infers non-application of the rule. He does $ot first effect u\^ rule and then cancels it. The inference of non-application (iT5nfT!i3'n^) is stressed here mainly to maintain the grammarians* theory of permanence of words. 2. ^nf^ 3Tc*lw?ff<T srf^qurrffiT etc. One is naturally inclined to do certain deeds. Sometimes he perpetrates such acts as killing a Brahmin even when he knows that it leads to bad consequences after death, because he is more concerned about the S€^3?, the immediate consequences. Thus the effect of negation in the case of ^^^irsr^frT can be very weak. This is not so with ^T^^^T^SPffrT which is prohibited by sfeVq" rules such as 3. We have emended the reading sr^ETOW.;^ of the Ms to ;

1. The objector is still not satisfied with the equal status of sri^HNr and . While a srfMrsr directly cancels the application of a fti^r, a forora simply records a correct word. It may not cancel the application of the fefer. The f^q^R ' S H R W ^ ' , therefore, may establish the correctness of the otherwise irregular word, yet it does not disallow the formation of ^^H^TOT^. Accordingly, both the words, H ^ T O ^ and ^ ^ H s r ^ , will be regarded as correct. 2. Wr*TOr*T^]rW£<f...

- °f the Ms is to be read as Otherwise the line is difficult to understand.

3. ^ ^ i £ m : ' restriction of its meaning as a proper name.' The function of firq^R is to restrict the scope of the word H ^ T F R ^ a technical term, by declaring it as a correct word.

( 1 ) ^ j ^ after <£% (in line 2) should be removed. Translated accordingly. Read

( 2 ) P. 6 . 1 . 77 : f^t wf% teaches that f, 3-, 5^ and <£ are replaced \* \ and ^ respectively before a (dissimilar) vowel, in euphonic






combination. This rule implies that f^ vowels do not remain before a vowel. Since this rule is an indirect srfgTcf, we have two optional forms in cases like ^ r ^ # The form ^ ^ is correct in euphonic combination with following vowels while the form ^fw is correct elsewhere. This rule which is an indirect srfgfqvj is thus comparable to the f^rr^r ' ^ i i ^ 1 . 3. fN>lftf^f.rnrra:- The relation between original and modification. Because of the principle of permanence of words ?>^q; cannot be regarded as a fip^R ' modification' of ^fer. 4. s r ^ f ^ s i F * — The maxim that ashes of ?g%" wood are the result of the Tgf^j tree. In other words, the maxim of cause-effect relationship. The m^kl wood and its ashes do not exist simultaneously. One disappears and the other comes into existence. This principle of cause-effect relationship is not accepted in grammar since it contradicts

[ 2.25 ]
rfsi:... According P. 1. 1. 49 a word in the genitive in a %%>% is connected with the word ^r?i^. The word "%w- in P. 6.1.77 is therefore connected with ^n^* and means * in place of ^ . ' The genitive thus establishes the relationship of pjrifir^ and sn%3T between ^ and q q respectively and consequently f^ is cancelled by 3 % There is thus no optional application of ^ ^ as maintained in the previous passage. It should be noted here that the approach of the objector towards P. 6. 1. 77 is different from that of the %^lf?cT^. It is true, as the objector says, that f^s are not permanently cancelled by ^ » because in the absence of euphonic combination f^s remain. However, according to the %^Tfy^ this can be a correct illustration if both the options are possible in the same situation. «[^>s are retained in the absence of euphonic combination while take place when it is there. Therefore, we cannot say that the rule f *Gt produces two optional forms as is the case with 2. q^i a r f ^ L . e t c . While the ^ i ^ i t f i r l ^ T is helpful in discarding the possibility of optbn in the case of the rule %^t w f e , it raises another problem, namely, contradiction with the theory of srs^f^R^. WheQ it is said that ^ s are replaced by ^njs, it implies that one word like ^fif is destroyed and another, ^y^, is produced. Katyayana also raises the same objection in his Var. 12 on P. 1. 1.56: argqqw ^l?9T%5Tc# f^TcWI^. Therefore here the author interprets the relation between *2?if^ and 3TT%5r m a different way. The word %>JJH does not mean * place.' It is traditionally paraphrased as srtf^ ' occasion.' Therefore, the rule f^t ^ r f e wou!4




©ean that ^ s are not allowed to stand In H%?TT. The srfaw of f^ in g-f^-^fT is thus uesrulve, of the nattire of absolute non-appeasj.w~*e. Thvio being thus no physical removal of ^ and its replacement by sp*!, the ^ i ^ ftc^c«f *s n o t contradicted. But still the question can be asked : How can iiuf replace %^ which is not there ? The very concept of t^qi^^T^TP" !*s unjustified it it is assumed that the p-nfei does not exist. Katylyana lias answered this question by adducing illustrations of a dlscipie coming in the place of his teacher or puiika trna being used in the place of soma. In both. these cases ^ f i j ^ is not actually present. In other words, the word p-n^1 means * instead o f rather than 4 in the place of \ 3. ^%^q"?cn;a! etc. Or one can say that the replacement takes place only in thought. When we say w;e derive a particular word from stem and suffix etc. ihc whole process is only mental. The actual stern and suffix do not undergo modifications. Thus even when ^ H means the * in place o f , there is no problem, because the place of ff» is only mental. 4. 3f¥Rifererfq 3&3". 3TPi refers to both the views stated above : ( 1) assuming that ^ s do cot appear in fHf cFT, and ( 2 ) the replacement is a mental process.

J. Correction I 2.26 starts with the sentence ^ ? f t sew *?#!"... which is wrongly included in 2.25. 2. q^t IT S R % is to be read as ^ i t H f fgr *ref?ir * there is a genitive «*:.' The rule ^q^Tf^sTRT^... also contains a genitive, namely ??: which will be likewise connected with p-jfi^r. Interpreting the ^?RTOt in either of the two wavs shown above one Onally arrives at absolute non-existence of ^ indicated by the rule i^q^T^ •• However, since the ^j rule ^R^frf^1... does not contain genitive, absolute non-existence of ^ is not indicated by it. Therefore, still there is scope for option. To sum up : The passages 2.23-26 discuss the possibility of optional &c in the word sHbrnr^* which is an objection raised by the *^h"2J. Our !W author follows q^Hffe in rejecting the analogy of f^ ^f^However, finally he comes to the point and justifies the objection.

1. ST3T%?Tr: ^ s n ^ * : ^ s ^ a etc. Since the illustration of ^sjft qaife was discarded, the ^ q g j offers another illustration to show the possibility ofoptional application of two contradictory rules. P. 3. 1. 68 : ^#f^ ^rq; teaches that the f^Ruj sr^ is added after any root when a HT^ng^s ending convey1


Mahabhasya-Dipikat Ihmka



ing agent follows. P. 3 . 1 . 69 : %uf%¥qr: ^ teaches that the fk^m %$\ is added after roots "%; etc. in the same situation. Since P. 3.1.69 is an indirect srfgfo of P. 3, 1.68, both the rules will be optionally applicable by alternating with each other. Just as optional application is possible in the case of 5f^ and 3 3 ^ , it can also be possible in case of the f$r<JT?PT ' ^ b r o ^ ' and ui?t taught by wi^fcr... etc. It should be noted here that in both the cases the optional application is not desirable also for the ^ q % ^ . However, he simply points out the possibility of such a wrong application, also in another case, sraj^nT: ^RFT^T. etc. is Pat.'s answer to this objection. He says that ^W and other fir^crfs will be stated as substitutes in place of ^ . Therefore! just as in the earlier illustration (<pjst qffrf%), here also the 3*r%3r will permanently replace the P4lft^. la this way Pat. discards also the second Illustration. $xf^ etc. Ms. fq^Tf^r does not make any sense. It has to be read as, spipf 3RT?ftft fq^iferl^: ^TcTJ Here an objection is raised against Pat.'s reply that 3 ^ etc. are regarded as substitutes of W{» The form ZR\% is derived from root fj by adding the third person singular ending % and the fzp&yr\ 3- by P. 3. 1. 79: ^RT%pl>%si 3": t If the f^cn 3 T is regarded a substitute of ^\\ as suggested by Pat. it will become fqg; because the ftfii of ^ will be imposed upon it by ^nf^^^TR. Consequently, it will be unaccented by P. 3 . 1 . 4 : arg^rf gfaRfl". However, the which is attested is desired. The assumption of P-n?3T^T*n^ between and other f ^ p n s thus results in undesirable forms. 3. ^ R T few ^rrq"^ etc. The author's reply to the above objection is that the addition of the anubandha ^ to the suffix ^ r ^ indicates that the anubandhas of pjnfir^ are not imposed upon STI^T. The anubandha ^ is meant to account for the accent by P. 6. 1. 197 : fe^rilf^Rc^ in verb forms such as f^sq% where the actual accent falls on fir- The "fin^ of srq; also leads to the same accentuation. It was possible for Panini to drop ^ and impose fq^r of ^T^ on 3 ^ by *3jT^R?vrm. Yet he mentions ^ in ^r^. This fact indicates that according to Panini the anubandha properties are not inherited by an 4. w=3f%*3 %xcra%r- Other f ^ R S such as ?T, ^TT, ^TB; etc. contain the anubandha ^ even when it was possible to continue it from the ^nffif^ ^T^- These f^^nns contain ^ and so are fiFt. Therefore, they get the designation m^W^b by P. 3.4.113: %^%^[ tfHbn<p^« At the same time fqv=r by ^rri^^^TT^ is n o t imposed upon them as already pointed out. Therefore by P. 1, 2. 4 : HT^^n3^?f^^ all these suffixes become




The line fo % ^ ?T *re% is to be corrected as ^ fe# *R%. Otherwise it cannot be construed with the earlier context.

1. f f ^ Ig^ft. etc. Another difficulty arising out of the assumption of ^rn^cTscTW ^ etw een ^ and other f^^^ors is illustrated by the derivation of the form ^hft, a present participle in feminine from the root fj. It is derived as follows: f ^ P. 3.2. 323 $ ^ # 1 P . 3.2. 124, P. 4. 1.6 f^^#^P.3.1. 6 8 $ T wg;#KP. 3.1.79 § * ^ tf^t P. 6. 4. 110 P. 6. 1.77 In this derivation if 3" is regarded as a substitute for sn^, the property ^ 3 will be imposed upon it by P-JlfN^TH*. A s a result, P. 7.1. 81 : ^n^T^R^nj; which teaches the augment «JJ^ after ^ will be applied and the undesirable form *ffeft will be derived. This objection is again replied by referring to a ^T?$. The addition of the anubandha ^ to the f^sf^n ^ r ^ even when it was available by ^rrf^^pTFT indicates that according to Panini, the property wi$^ is not inherited by the substitutes of ^ . 2. i&gs: ^% I The verb formfeg^:is the present third person dual of the root % % It is derived as follows : 3 *n* P. 3.1.80 T crac P. 6. 4. 48 P. 8.2.66, P. 8. 3.15 Here P. 6.4. 48 : sr^t &tT- causes elision of 3T, because the following suffix g- is 3rrv:n=rTcp3 by the rule su^hnp* ^ * ( P. 3. 4.114). However, if ^ is regarded a substitute of ^ 0 the HR^Tp^T of ^ will be imposed on 3 by ^nfir^^n^. Therefore, application of P. 6. 4. 48 cannot take place as T 3" becomes HT^hn<p5- The derivation of the form ferg^n is thus another difficulty raised by the objector. Here also another ^T^J5 is adduced. The purpose of the anubandha ^ is to ascribe the property of srHbrjc^rS to a suffix ( P . 3. 4. 113: %f%cj; Accordingly, wi is **T^II3$- Its substitute ^r also is ^ j ^ It could inherit the property of srishnip^ from ^ simply by




Ah-nikn VI


*nw» Tt was therefore not necessary to attach the anubandha ^ to it. T^e f;:-;t 'hat PSnitii attaches it indicates that the property of ^•f^r?f^c^ is no i lone riled by S^fTT^Rrvn^*

1. nnnn W*i etc. So far the author discussed three objections against the view of p-iifir^wiw» Here he discusses one more illustration which contradicts the said view. The fipfj^JT ^ ^ consists of the anubandha ^ . According to P. 1. 1. 47 : fe^^ts?^!"!?*, a grammatical element containing the anubandha ^ is added after the last vowel of the element to which it Is prescribed. Therefore, ^ ^ will be an augment - a n infix of the r o o t rather than a substitute of w\ and so it will be added after s n \ The character of 3TO; as an augment goes against Pat.'s statement 2. f^ cTR^T TO^r: etc. After introducing the objection the author, instead of giving a reply, puts a counter-question to the objector. It is essential to understand the background before stating the counter-question. P. 3 . 4 . 7 8 : fanjergftr... etc. teaches different substitutes for ^ ^ , the middle first person singular ending, is one of them. According to P. 1.1.46 : sn^F^T ZW'ih a grammatical clement having the anubandha ^ is added at the beginning of the element to which it is prescribed. Therefore, ^ being possessed of the anubandha z will be added at the beginning of ^ ( ^ , %^ etc.). However, it does not happen so. In spite of ^, ^ is substituted in place of ?§;. The author asks: Why is f^ treated as a substitute and not as an augment ? 3, sra +??TO etc. If the objector answers that ^ is treated as substitute, because all the other elements enumerated in P. 3.4.78 are substitutes, the same answer holds true also for the question regarding ^ ^ . Since all other ft^^^ are substitutes of sr^, ^ ^ also has to be regarded as a substitute.

[ 2.30 ]
The passage cannot be interpreted satisfactorily, because there is a confusion of i^q^T and f^iyrT. We have tried to interpret it in such a way that it leads to the concluding remarks in the next passage. 1. rr^( q^jt ^ q " % I The objector tries to solve the problem of ^ by resorting to the view that the genitive ^ q is one. He means that it conveys one single meaning, namely, pjn^T^TOre anc * so ^ has to be understood as a substitute only.

9,1] 2.

Notes vmrz$T ^=ri?T etc. The %^Tf??H replies by resor!in::: o

ill another

v i e w w h i c h i s t h e o p p o c * t ? -^r */••• • . - V v - ' d - ' i ••: : b y (•:.* ^ L \ > . . _:. , ^ r . v . ; _ ; . ^ .•.*

him tile genitive is separately connected with each substitu-:.*. Th.refoie, it conveys different relation in the context of each substitute. The- statement fiffQ g[jT could therefore mean, on the basis of the anubandha % : ?:, which is undesirable. 3. qjfeqTfetTfaflf^f... etc. In order to support the view adopted by him the f^rf^R; cites one illustration from the unadis. The unadi rule (1.18) ^feqiferfimfir^^t *pj qfbTTfe^-ar J The genitive word fijsif^TSTF^ is connected in partitive relationship with *p» and in relationship with qiferfif?^*. Thus one and the same ^ff is construed with different elements with different meanings. Pat. has also made a similar remark in his Mbh. on Var. 4 on P. 1. 1.49.

It is difficult to interpret the passage literally since the handwriting in the Ms. is very bad. Further, it is also difficult to link the thought in this passage with the previous passage. We propose a possible interpretation of the passage as follows : It has been already pointed out by the % ^ f r ? ^ the genitive can be interpreted differently in connection with different elements. However, if this view is accepted the problem of %% remains as it is. The f^Rnf?cPj; does not provide any answer to the problem. However, the discussion in the present passage implies that the expression gr^q 1^ has to be interpreted as ^ r p:rr^ f^ 1 (In fact, the view JT^n^r $ H ^ " ^ T ^€t adopted by the f^r^rf?^ is not favourable for this interpretation. We fail to understand how the f^ncif^f^ might have solved the problem of ^ r . ) The objector argues that if the genitive &$$ everywhere conveys > the genitive srq: also must be understood as conveying pjrpmConsequently, 3 ^ must be substituted in place of w{> We have, therefore, translated the first line assuming =pr*rT instead of Tjqj In the text. The %^if?3^ replies that 3 ^ should be compared to ^ taught in P. 6. 4.47 : ^ t i t * t q w T5RF^CR^T^« The rule teaches that ^ is substituted in the place of ^ and the penultimate ^ of s r ^ and placed after the final vowel (i. e. ar) before an S T T ^ ^ suffix. ?n has thus a double character. Since it is ftrcj; it is placed after the last vowel i.e. ^ in the root $ r ^ . Therefore, it is an snrn}. At the same time it removes ^ and ^ and is supposed to take their place. Therefore it is an srfrsr. Finally, the form derived is *pt.. (Why Panini did not teach the substitution of ^?^ in the place of scs^ instead of this complex procedure is not clear.) Similadyf

Mahahhasija-Dipika, Ahnika VI


^ as a substitute removes ^ and as an augment conies after the last vowel of a root ( w % for example ). [ Correction ! In the Text (p. 8, /. 26 )
re-ad ( f mm ) for ( f *g%\). ]

2. q^f^r^^r etc. After it is established that foqr^f is equivalent to srfi^hr, now the author gives his final answer to the problem of srf^Nr of cfi^c by f^q^R which is regarded as more powerful than the rule teaching trifg\ A cross reference is here given to the Mbh, on P. 1. I. 47 and the Dtp. on it as follows: Mbh. %^^ sfr^tqsrfct ^rrfq i r% mi

The question of optional application of both u\^ and q^T^T^ is thus finally answered by invoking popular practice. 3. In the Text (p. 9, L 1), ^nc<^t srefir, should be read instead of ^HT^d" *rafir ? Translated accordingly.

[ 2.32 ]
1. ^f| ^-WTf^fft^ etc. Pat.'s suggestion to assume =pr^ etc. as substitutes of ^p?s is an answer to the objector's idea that ^ and ^r^ are mutually options. However, there is no direct statement in the rules of Panini that 3^1 etc. are substituted in place of ^\, In order to make a statement like this Panini must read the wcrd ^rq: in every rule, for instance, *"f^nf^n 5Tq: ¥$%. According to Pat. the additional mention of the word ^rq: in each , rule is not necessary. The word ^ from the rule $if? ^{\ ( P. 3.1. 68 ) is continued in all the rules teaching ^q^ etc. The objector's problem in this connection is that the word ^r^ is in the nominative whereas the genitive ^rq: is required. Pat/s answer to this problem is as follows: ^if^vq fc^^T q^*ft 5Tf«rfe rfWTCT' ^3T ^^qftr^rft gR3Tif^rg^^% : 'According to the rule crsn%^TlW (P. 1.1.67) the ablative f^rif^r- will cause the nominative sn^ to change itself into genitive.' However, the discussion in the Dip. does not take note of this answer. The author seems to raise an objection assuming that the nominative ^\ is continued in the subsequent rules. 2. ^^^ ^ 3T3f%: ... The whole passage is difficult to interpret. However, with the help of some emendations (e.g. 3T«pr%£ is read in place of ^ T ^ l ) w e ^ e v e tr*ec* t 0 S et *° t n e import as follows : The objector argues that even after ^JJ; etc. are looked upon as substitutes of ^ their optional application cannot be avoided. When ^ is, for instance, continued in the




rule %rrfHsi: ^ the rule would read i^if^**?: 3 ^ W{* The simple interpretation of the rule would be that ^ ^ and ^i\ are added after f^ etc. This objection obviously disregards Pat.'s solution involving as cited above. 3. STCTKT^RT etc. QWWJ is a suggested emendation. The sentence is a reply to the above objection, ft is stated here that option stands even before the anuvrtti of ^TT^, because both wi^ taught by the general rule ^#f| $r^ and ^ ^ taught by f^gif^wj: ^$\* are equally applicable to f^ etc. On the other hand, after anuvttti is effected option does not stand, because if it is accepted, after anuvrtti the earlier rule (sp#f* srqQ becomes redundant and this results in the fault of repeated statement of ^ . Thus the objection that the annuvrtti of ^ \ also results in optional application is refuted.

1. f f mi etc. Another illustration of possibility of optional application of two rules contradicting each other is referred to here. P. 5,3.70: teaches the suffix ^ conveying no specific meaning. P. 5.3.71 : n* h teaches the suffix snji^ after indeclinables and pronouns. This latter is, however, added before the last vowel of the stem. According to the objector both these suffixes would be alternately added Rafter an indeclinable or a pronoun. 2. %5TftRNft ^fifig etc. The author makes clear the meaning intended by the objector. He tells us that when the objector shows the undesired position of optional application of 35 and m^%, it is not due to the fact that they occupy separate places. In other words, the objector does not mean that option is prompted by the fact that the suffix 3s is added after the stem, whereas s r $ ^ is added before the last vowel of the stem. The option is prompted, according to the objector, by the fact that both the suffixes conveying the same meaning cannot be simultaneously added after one and the same stem. Therefore, it is not possible to derive the form % by adding both the suffixes simultaneously. (cf. snriter: ar^TcTT^: cr%5T sftRtac3T^ ^ %qt: srg^q ffar *nT- 1). However, the optional forms, 3^g%: and 3 % ^ are possible by alternative application of sp and It must be noted that all the illustrations offered so far by the objector presuppose his ignorance of the fundamental relationship of 3"c*rf and 3TT3i^ between general and specific statements leading to the 3T«TOT>-F$ relation between the two. This is given as a reply by Pat. in connection with this illustration which is explained in the following passage.


Mahabhasija-Dlpika, Ahnika VI


1. mm*B f& %S5fi3TR¥^ir l Pat.'s lepiy to the above objection that there cannot be optional application between =s and srar^, because %s^% is an exception ( *iqsn^) of ^, U is applicable wherever sp is already applicable.' According to the Pan. S R mmB $t firfem?*^ * ?THT 3TO^fr vrafif (read T under Pari. No. 57 ) a grammatical rule A cancels another rule B which is necessarily applicable wherever the rule A applies. The relation between % and m^% is thus that of 3"r*?n and srqgi^. Pat. further slates : {^qT3«T?F^sNjrfFPfsiN l mm® ^?% f?F?!tR??R^?t ^ W\^4 tifkwfa I T u u s according to Pat. ftqTcR is an exception to u\r% and therefore cancels it. The words after sr^rfTSPffrr., which are read in the Ms. as do not make sense. The translation is based on common-sense understanding of the statement. 3. 5?Rt qifiiftr^r etc. The whole passage beginning with this line is extremely difficult to construe. Some words are not legible in the Ms, What we have understood from the passage is this: When Panini teaches prohibition of something it is not cancellation of something which is already effected; it indicates just non-application. G. B. Palsule has suggested the following reading: jfjjfif: f^9Tc%S¥sp*T*qin3r T f%vTR ^ fil!% ^f%^: ' Panini states this, namely, that when the permanence of word is accepted, there is neither a positive nor a negative statement of a grammatical operation anywhere in grammar.' Once it is accepted that words are permanent, no modification can be introduced in them either by effecting some change or discarding it* Anyhow, the passage remains vague. It is difficult to connect the purport of these lines with the last two lines of the passage beginning with s m w ^ ^ff... etc. which are equally difficult to interpret.

1. It is finally established that a firTFR is equivalent to $rft?%^ and consequently it cancels the 37JFPT rule permanently. This statement cancels the possibility of the other option, namely **Hfam^. However, this leads to some other problem. Some grammarians speak of optional elision of the final nasal of sam before c T Thus both the forms, ^cT^Jj; and ^RTC^ are R« derived. P. 6. 1. 144: srq'RTO: fe^PfTlcf^r contains a ftqicR ' H T ^ ' which will discard the form H T ^ ^ - However, both ^ 3 3 and ^ 3 are found. Since f^qj^nr is supposed to cancel the other form, the form HcTcf will be regarded as incorrect.

9-. 23] 2.



? r ^ 7 i % f f ^ g r . . . T h i s line m a k e s b e t t e r s e n s e '(it is r e a d a^ 7iri|i?%«

and does not allow option ? ' 3. irg- ^ ?cf% etc. Three letters after srfar are not clear in the Ms. We propose to read ^^T in that place instead of ^ q^TT (printed ). rr# xf sg% ^vil STSRTTE^ would mean 'this being Vac case the twofold derivation is not possible'. 4. ^ft^fts? 3??TT: etc. The difficulty stated above is genuine also according to Pat. Therefore he proposes that a special effort has to be made to justify the optional forms ??c!!TE[ and #=frr?i\ However, at the same time he again asserts that a fsrqirR always cancels a firfw rule. The special effort intended by Pat. in this connection is the inclusion of the two forms grafcT and HcRT in the ^qH^HWIRT^ which accounts for similar other optional forms. 5. s*i^fir^?p?Fi;etc. Although the forms ^^?r and #gr?r are sanctioned by the Var. mentioned above, there is restriction regarding the use of their derivatives. The suffix syaTi by P. 5. 1. 123 : goi^^R3ro«JU%.7i: ^f^T ^ can be added to both the words ^<=m and *RRT when *rrw ' state' is to be conveyed. However, the author tells us that only v-ud<^ is a correct form, Srhr^r i s n o t accepted as a correct form. As a reason for this the author refers to ^oqf^pg-R?. Pat. refers to it as srf^ffgr^?^ and explains It as follows : Mbh. on P. 4,1.88 : era sprh srs^it: ^r?TRTWT%?r f%5rfrs^^rTpq!%^fw^i?rf^5T^Tcp?qf^r j c ^ ^ T ^%5?!effl% fw*JlT 3ffg^^TS^r|rqf%VR% 3Tlfe^fi?f^ 1 What TTT Pat. means is this : The word srifw^ is derived from 3T%£ with the addition of the suffix &^ by P. 4. 3. 154 : srrfiiTsrarf^wits^. Although the grammatical analysis of the word arif^^ contains the word srfw^i its semantic analysis is always given as sr%iTT^ * flesh of a goat.' In the semantic analysis the word srf^r a ° d not 3rf^g5 is used. The restriction regarding the use of different synonyms in different contexts is thus at the basis of The same applies to **Ucrei and



Summary. Kat. shows that if the name H#?rm^ is given to a word like ^ r irrespective of its meaning, the definition suffers from the fault of overapplication. For instance, if ^ is called H#^T?T^ even when it is used as a ^;TT * proper name' or as an ^ T O ^ T * subordinate member of a compound,' all the grammatical operations conditioned by the term ^«TTO3 as well as the operations stated with regard to the sFrnTxrs (17171s within Jj) such as ^ ^ T ^ m : (P. 7.2. 102), ar^[ ^Rl^w?:... ( P. 7. 1. 25)


3ffihabha.vja-Dtjnkat Ihnika


[ 9. 23

will he applicable to m even in the domain of *tsTT and 3<W&T. This overapplication of the designation ^ ^ N T O ^ m t n e domain of ^ l and i q ^ i i results in undesirable forms such as *^WT, *«rf^H#^» *3TfefS3?T, etc. instead of - F R H '*TT%?wk, sf%*^?^, etc. respectively* Therefore, K a t recommends that a statement prohibiting the application of the name in the domain of ^ r and 3q$Hra should be made. Pat. agrees with Kat. on the problem of overapplication. However, he does not agree with him on the point of solution. In his usual style he tries to do away with the difficulty by employing various devices instead of adding a statement to Papini's grammar. He offers four different solutions to the problem as follows : i p. 1.1.34 : ^tq^^%oThRTqrrTOfir ^ w i r s m ^ ^ i s s P ! i t i n t 0 two rules : ( 1 ) ^#TO^?%^TT?FRTvTOffT s^P-TFtt*^ and (2 ) ST^f^rRT?^. The latter rule is interpreted as a general rule teaching that the words q$ etc. are called e*=!?*FI except when they are used as ^ s . Thus by using this device o[ qfrlftvn*T * split of a rule ' the problem about ^ r is solved. For zq®-§m he offers two solutions: ( a() P. 4 . 1 . 1 4 : ) r zq®§m which is already rejected as being redundant in the section dealing with feminine suffixes, is transferred to the section dealing with H # ^ T ^ ^ T . Read as srgt?H*T(^) * ar*fc the rule means that the rules of substitution of « and an* aPP!>' n o t t 0 3TO&T3. S i n c e t h c « ^ ^ operations consist only in substitution of sr and ^ , this rule in its new disguise prevents their application in the domain of sq^OT. ( b ) The paribhasa ar^lfelT ^ | ^ ^ ^imiaift¥f%^ ^ f e teaches that the rules in the section headed by the mfy^R ' 3TP3' { P- 6.4.1) apply when case-ending is added after the stem which Is mentioned in rules. Since the case-endings like # (dative) are not added after n t when it is an ^qn^H in a compound such as arfitspf, there is no question of the application of ^ r # ^ ? : ^ ( P. 7. 1. 14) in the situation «rfyj^ + i s because of this parihhasa. 2. The rule srafcftft «*TT^rft is not a definition rule. is an adjective of *Rrff£r. The rule thus means ^ etc. are « names for all.' When ^ is used as a proper name or an s w ^ k it ceases to be a name for all. Therefore, the operations conditioned by H ^ n ^ do not apply in thc domain of ^ r and ^qn*T. Another difficulty arises, however, in this view. The grammatical operations conditioned by the designation ^ m ^ do not apply to sr# which is s ^ r m ^ ' n a m e f o r a 1 1 ' since it does not have the designation *nfara%. Therefore, Pat. proposes another solution.




Each word in the rule *rer£fir ^Hhwfo is regarded as *rai^fSr ^ *ra^Vfii i sHfaunfir ^ nw^T^nfii ^ *Tg-?mnfa i This device of q^Hfir'i^ results in two statements : ( I ) words in the ?T3TE[ list are *pNn*j^s * names for all'. (2) The ?r^if%s (which are names for all) are called ^rsro^s. The rule thus achieves both the purposes : The qualification of ^r^tf%s by the term H ^ H ^anc * tho designation JCTSNTO^- In this way the word %$ is called *?^NT?H only when it stands for all. Since ^rt used as a H^T or 3"q*ribr is not a name for all, the undesirable application of rules is prevented. 4. The technical term *?sbrra^ is a significant term (spgafchjr). The basic purpose behind coining a technical term is to secure brevity. Panini has coined a number of technical terms such as ti, ghu, bha etc. which express a lot of meaning in the briefest possible manner. These terms have no meaning in themselv2S. However, sometimes Panini uses lengthy terms such as ^sfcnrj^. His use of such lengthy terms even when it was possible for him to use shorter terms in their place, indicates that the terms stand not only for their forms but also for their literal meaning. The term sHhfFR^ therefore signifies two things : (1 ) designation ^#?TTTT^ and (2 ) being a name for all. Thus wherever the term is applicable it will represent both these meanings. Since a *j^j or an 3 W ^ T cannot be a name for all, it cannot get the designation ^ f o j j ^ even though the word signifying it belongs to H^Tf^ list. This is the last and the final reply of Pat. to the problem of ^ r and s ^ s p f . The author of the Dip. deals with each one of these solutions in an elaborate manner. His discussion is further diffused with many diversions taking the reader sometimes far away from the main topic.

1. ^ ^ T O ^ r s f i ^ i : . This is a JTfoq^mrf^ prohibiting the application of sHRmst^T in the domain of H^T and ^T^rlFT (for details, see the summary above). 2. 5Trtfird<c3rr^ e t c - The author sees two opposite principles of homonymy in the two statements of Kat. and Pat. respectively. The srfirN1§7T^ implies, according to him, that the word ^ meaning a proper name is different from the word H*T meaning * all'. This implies the principle On the other hand, Pat.'s paraphrase of the 5rf^%^r^nf%^, namely, lcnfo ST^Pftfir * tft^mTfai implies that one and the same word *?# meaning ' all' also can function as a proper name. Kaiyata on the Mbh. above :


MahabhSsya-Dipika, Ihnika


[ 9.23

There i s however, no enough evidence to support this observation of the author, who has squeezed out these two meanings from the respective statements iYi the 4 great sages.' Nagojibhatta tries to reconcile both the statements in his Uddyota in the following manner: : \

1. Two alternatives for the srf^qvrrsnfi^ in respect of ^"T are proposed and discarded in this and the next passage. The first alternative suggested here consists in making use of the paribhasa, jfpn (No. 15). The objector argues that if we accept this paribhasd, the prohibition with regard to ^j need not be stated. This alternative is, however, discarded by pointing out that the paribhasd operates only in the domain of q^ftfa. Since sNfaraH^r and H^TW$T4S are not q ^ f i r s , the paribhasd cannot be useful in this context. 2. sfrtrf^^pifrr etc. The paribhasd iirorgwft: etc. teaches that when a word can convey both, a primary and a secondary meaning, the grammatical operation should be understood in connection with that word when it denotes the primary meaning only. The author cites 3*?ftqftfT mora^n1 as an illustration of this paribhasd. *xm and HT£T are proper names of two deities. A proper name is, therefore, the primary character of each of them. When the two words are compounded in a ^ (traditionally known as ^eRn^r) two grammatical operations, namely ^-substitution for the final of sffjr ( P. 6. 3. 27 ) and q^ in *?m ( P. 3. 3. 82 ), are effected and the form 3{#fn?r is derived. Tradition gives sffjienfr *?T*ra«PT as a counter-example. When the words arf?i and fjfa convey secondary meanings namely, * the boys similar to and £?ir?' the ^ compound (which cannot be described as ^Rrnjr£) is derived, for the above two grammatical operations do not take place here in accordance with the paribhasd. { For details, see 4 below ). 3. q ^ r ^ p ? " mm ztc* This paribhasd cannot serve as a substitute for the srf^TsRiffoj, because it operates in the domain of q^fiHc, 'operation with regard to finished words.' When a grammatical operation has a q^[ as its basic unit, it is termed q ^ R , e. g. *mm> 3%3 etc. The definition rule ^r^frftf etc. as well as rules like P. 7. 1. 14 ^ s n * ^ : $T are not q ^ p f s , since the basic unit for these rules is a snfSrqf^fi or sr^ * stem \ 4. 3T5^T3r% xr f^7XiT% etc. This line is quoted from the Mbh. on P. 1.1.15. The full text reads as follows:




i *mns#f% \ wibm &?&& «r^fir' ^r^^i^r ^ %%m^ \ (The same passage occurs also in the Mbh. on P. 8. 3. 82 ). What Pat. means is this. The primary meaning of the word T\X is bull or cow, while its secondary meaning is m^t^ ' a man from the country of ^Tffc&s' ( a m^t^ is called rft, because of his dullness : irft^Vsg:). When one hears an injunction THT^F^' *a qt is to be fastened \ QI\Q fastens a bull and not a ^T![ta>> because the primary meaning is understood in connection with the action. But then the question arises, how to explain the usage such as 3fr%njfir, STTSTFP? etc. where the word ait stands for griftas and not for bull. The f% taught by P. 7. 1. 90 and P. 7.2.115 in the nominative singular 5 f and the sm^ taught by P. 6.1.93 nn in the accusative singular *TT*£ cannot be applied to *ft when it conveys secondary meaning, according to Tfroi^q^TR. The reply to this question given by Pat. is that the ff% and 3Tir=T are concerned only with the phonetic form of the word (^T^Tsre )> they have no concern with the meaning The author of the Dip. cites this line from the Mbh. to point out that grammatical operations such as ^ ^ j ? ^ : ^ will be applicable to sarva irrespective of its meaning and so a new prohibitive rule is necessary. For further discussion on the Mbh. above see Dip. Fas. II, p. 17.

[ 3.3 ]
1. Another alternative for HsTT^f^ra is the acceptance of ^EHSTTfcfcR. P. 1-4.1 : 3 T ^^T?T^«SI H^T is an srfeEH" teaching that only one of T the two or more H^TS, taught under this srfv}^, is given to an element when all of them are simultaneously applicable. The one term that is taught in a later rule cancels all the earlier terms (P. 1. 4. 2). The term ^r#^j?^ is referred to in the Var, 41 on P. 1.4.1, whereas the term ^ r j itself is defined as ^^jfafir%%^T S s T ' a name is that which resides in one object' TT in the Var. 44 on P. 1.4. 1. Since the term H^TT is stated in a later Var. it cancels the earlier term ^ ^ j ? ^ , when both are applicable to ^ # for instance. Thus the grammatical operations caused by 5R^TI^a^I will be automatically prevented when ^ is a 2. %-r^firfo. The author gives % r ^ n as a counter-example to explain the working of TT^jH^if^r^R- Kat. has arranged different tfjfjs in his Var. on P. 1. 4. 1 in a sequence so that the latter H ^ 7 cancels the earlier one(s). Just as he has enumerated sNfan?^ in Var. 41, he mentions the term g u R ^ in Var. 40. The term Tpici^r * qualifier,' though not defined by Panini is used by him in rules like P.5. 1. 124. A word like ^ ^ T T is a *ynsr3?T as well as £37. According to ^ H ^ l f e ^ R , however, it will get only the designation H ^ I . Therefore, the suffix oqsj; taught by P. 5. 1. 124 to a


Mahabhamja-Dipika, Ahnika VI



In the sense of * state • or « action ' is not added after %^TT. Consequeniiy, u - undesired form * ^ w r which would have been arrived at by applying ^ is prevented. Pat. on Var. 44 on P. 1.4. 1 W mi *q?%% | However, Pat. does not cite *%f<^? as an example. 3. if^nffk^r rnftrej-pm. This alternative is discarded simply because the term * % as defined by Kat. has not been taken into consideration by taken into b sideration by hilf Th tf^refSPTO iis stated by him disregarding the ^Wf^r and tffS himself. The therefore, the ^ n ^ f ^ r ? . This statement of the author implies that there Is no ground for Kat. to regect his own Far.'s in r r ^ n p f ^ and state fe% here.

1. srfim#? is given as an example of grq*T^wflr%*. If the ^wi^ is not stated, the dative singular of arforf1 will be **%nMr instead of the desired form $r%refa> because the rule spt^n^: & (P. 7. 1. 14) will be wrongly applicable there. Therefore to secure the correct form 3?%*?#j the TTO^fsrfcl^r is necessary. The author discusses two questions about this word : ( 1) the type of compound and ( 2 ) the applicability of the rules caused by ^JfTTO^r to f f The word 9ffhm can be regarded as ( 1 ) : e#: ^ ST.' * he whom all have surpassed \ or ( 2) s r r f ^ p q ' one who has surpassed all'. The latter compound is not taught by Panini. Far. 4 on P. 2.2.18: srr^: fEraf and the statement of Saunagas cited by Pat. in his Bhasya on the same Far.; sr^T^q: ^ f%$mi accounts for this type of compound. It is stated here that 1%T# is to be regarded as a sn%mrcr (although 3|pWt also is possible) for the simple reason that P. 1.1. 29 : * aifsftfir already prohibits s^Trotf^l to m etc. in a ^gwtf| compound.

Finally, Kat.'s proposal for prohibition is accepted. But how to state the prohibition ? Here the author points out different possibilities of stating it. {1 ) It can be stated with reference to the ^ ^fi
, * *?# etc. are called ^ # ^ I T T ^ except in the case of sr^r and T T O ^ T '. A or ^qgr^T is thus excluded from the scope of the H^rmsfgT. ( 2 ) T h e prohibition can also be stated with reference to the ^ f ^ list. In this case there are again two possibilities: ( a ) prohibition with reference to the *Firif§[ Hst mentioned in the H ^ T ^ r (<ira), or ( b ) with reference to the $&\f% list in the ganapalha ( q u j ) . In the former case s n i f ^ s which may Stand for ^ j and ^ q ^ ? r as they are listed in the ganapatha are specified




as srasffasT^Tlft and only these *rgftf^s get nffsfimpF. In the la tier case, ho wevei\ ^ etc. denolincr *?*?* :m«i are not included in the list Itsdf. ( 3 ) One more possibility pointed out by the author here is to state the prohibition with reference to ^ i % s at both (he places, in the ^srj rule and in the ganapatka.

1. Kat. and Pat. recommend that the prohibition should be made with reference to the sraffe list in the H5n*£W. P a t explains this as follows : ewi^fr ssbfTOlfir H^wlRTf?! * *raKtf*> ' *r# etc. arc ( named) sarvanaman. Those which are *?=TT and 3W^«* do not belong to the sarvadi (list).' The author of the Dip. here points out faults in the other alternatives referred to in the preceding passage. 2. H5TRT % srfire§r etc. If tisfl and 3W^?T are excluded from the domain of ^ s f a n r ^ r , the prohibition will not apply to srenrn^FU* The jyifs such as w^rf^, ^a^f^ etc. are described as srs^fcns, because they are lists within the sg^ff* list. Whenever a grammatical operation with reference to such w^rir&js is stated, for instance, ?i^T^fauT* ( P* 7. 2.102 ), it will be applicable to H ^ I and ^ q ^ ^ . For instance, the pronoun ^ T is 3 W ^ T in the compound srf&^ : srfeaFT?cr: ^ * one who has surpassed Wm.' While deriving the nominative singular of the word 3rf^?r^5 S T ^ t o W will be applicable ( even though ^ is ^ T H ^ ) and the undesirable form *3rfifH: instead of srfgr^ will be derived. The ^TW^$r#f%v:r must cover also s^rffoi^rls such as ^ i ^ w : . They will not be covered if the prohibition is stated with reference to the xr^TTTO^Tj because in that case it will be applicable only to the operations caused by sHhrrarcTsTT. 3. 7M ^ feqyifur etc. If the prohibition is stated with reference to the &'3H^ list in the ganapatha (( b ) above), it will not be applicable to other words like $r«m etc. which are not included in the H ^ % list but also get (P. 1. 1. 33). The ^tq^HSffir^T being thus restricted to the list, grammatical operation caused by applicable to STO*? etc. even when they are *??TT or ^ q ^ ^ . For instance, the dative singular of 5T«re * person called 5m?r' will be *jj?j3T^r instead of by the wrong application of 3n$3T*«T: ^ ( P. 7. 1. 14 ). In view of these faults in the two alternatives the prohibitiou is stated with reference tc^^Jf^ list in the ST^Trcj^ itself. 4. sra^tewtflw^^nfir. The word H^qH^^lf^ has to be supplied here before ?r Hfef§rSlfH to arrive at the proper meaning. ^ etc. standing for H^-f and TqH^^T are not at all included in the H^tf^ list which is being


Mahabha^ya-Dipika, Ahnika



defined as s?#?rm^. ^~rrqH^?I% is thus a qualifier of H^ which are included in ^gn%s are thus already specified as before they get the name H^^TSR^ Therefore even 3fc?cRPH3^fs will not apply to qsTT and It is, however, not clear how the difficulty regarding sm?? etc. as shown above can be got rid of by accepting this alternative. Perhaps, if the word 3TH^1W§RTf^i is continued in all the subsequent rules defining ^^TWI it will qualify TOST etc. in P. 1.1.33 and thus the prohibition will also be connected with spjji? etc. 5. H^S^HRf^gc^frT.. The word e f * name of a person » is phone* tically similar to the word T^ meaning * all'. But since the former is not included in the ^^if^ list which is specified as s r ^ f f a H ^ n ^ i it cannot get the name 6. %w f t ^ R T ^ ^ r r : ... P. 8. 1, 27 : fe^t tfnn^C fcH^^ft^PTT: is cited here as an illustration to show how a list is specified in a rule instead of in TpjfqH* The rule teaches that the words *rfo etc. following a finite verb-form are unaccented when they convey contempt or repetition. Here the restriction of the scope ofwords qrft^Tj etc. with reference to their denotation is already stated in the rule. Consequently, the words ntsT etc. conveying only contempt and repetition are listed in the im* Both Kat* and Pat, further state that the ^mrfi[ list is always specified with these meanings whenever it is mentioned by Panini: ftr^t *Ttert ^WTTmT^cri^^T^H^rT^J ( P . 8. 1. 57) (Mbh. on Var* 1 and 2 on P. 8.1. 27).

1. The e ^ T O ^ S f f o w is meant to prohibit, in the domain of and 3"TO^r, grammatical operations stated with regard to srfwFJ. There are various kinds of grammatical operations with regard to ^Hrwr^. Here the author makes clear the scope of the prohibition. The grammatical operations applicable to g&TTO^ c a n ^>e classified under three groups as follows: ( i ) those which are stated with reference to e3^RT^ in general (arf^ftm), e.g. sHhn^T: ^ ( P . 7.1.14), arrftr srfw^n § ^ ( P . 7.1.52), etc.; ( 2 ) operations stated with reference to specific groups of H^WT^S, such as c^rftem: { P. 7.2. 102), s r ^ ^cRjf^^:.. (P. 7.1.25), etc. and ( 3 ) operations stated with reference to specific words which are named H'lwT^ ( f ^ i r w ) , for instance, g^^^T^v^f ^ f t s ^ . (P. 7. 1.27), rr^7*t; ^ q j \ . , ( P . 6. 1. 132), etc. The author tells us that




the prohibition applies to all these kinds of grammatical operations except ibe last one. 2. ^ g J ' m ^ T T * ^ ff^" etc. 3?^rmHsTT for gsjy^ and s r ^ r is already taught by P. I. 1. 27. Therefore the ^rq*F§Rsrf^rw raiist be applicable also to the operations stated with reference to these two words. For Instance, P. 7.1. 27 : ^r^g-yTjf -^i^ts^ should not be applicable to the formation of the genitive singular of the compound srf^fs^rr * one who ha3 surpassed you \ However, its application is desired to account for correct forms like srfar^cr and to avoid wrong forms like *3T%g^r?:. The scope of the H^q5^7fjn%%^r is, therefore, not extended to such operations. The author accounts for this non-extention by saying H f ^ s n ^ c R J ^ j ^ . He means that the reference to ?F%P^ and «rare[ in P. 7. 1.27 is totally different from the reference to them in $P%\f% list. In other words, s r q ^ and srar^ are mentioned in this rule simply as snf^nf^s, not as sdRTH^s. Therefore, there is no question of application or otherwise of any rule caused by srw-

1. Kat. enumerates two purposes of the Hsfrqsn&Wsrf^N' in ^ e following two Vdr-s: ( 1 ) ^rcfcrR^cf^ptf^Nt * for the sake of grammatical Operations which are directly applicable to *rS etc. and ( 2 ) spftsR ^3*T%H1+H^*nw * the purpose ( of the prohibition is found in operations such as ) sflcj; substitution in the place of (the singular case added after) ^n[ etc. \ la this passage the author of the Dip. deals with the second Var, leaving the first unnoticed. The illustration for the first one can be given as follows : The word &$ used as a proper name or a subordinate member in a compound such as ar%H^ is not liable to undergo grammatical operations directly stated with regard to H^RW^- For instance, while deriving the dative singular form of ^ ( proper name) or srfctH^. P« 7. 1.14 : tfj«n?«T: ^ does not apply, owing to the f^M^^gfiftre, and we get correct forms, s^fa and , instead of *^J4& and * s * f ^ M l respectively. . The author cites srf^r*^ as an illustration of the second Var, The compound means ST%PFCT^ ^CR*[ * one that has surpassed which ( of the two ) ' . While deriving the nominative singular of arft^rc 1 P. 7. 1. 25, which teaches sfcj substitution in the place of the nominative singular case-ending added after a word ending in the suffix ^?, is applicable, because which precedes the case-ending is derived from %f£ by adding the suffix ( P. 5. 3 92). However, P. 7. 1. 25 does not take effect, because of the £ Therefore, the form will be 3T%$3TO^ not

Mahabhasya-Dipika, etc.
c,>,n[.:••••.. J • 7717 =* ^ fioig; ^




The word q w w r r ^ is a
Therefore, the

the absolute which ( ? ).'

word ^aTr| which is a ^ S R W ^ is not an s r s r f o (in ^^fT^i compound the latter constituent is the main member). Consequently, the does not apply in the derivation of TOT^?. Therefore by applying P. 7.1,25 the nominative singular will be q?*R$cn[f as against ^cfepgR^.

1. T^ftsTT 'qfariirsrm: etc. Pat. rejects KSt.'s proposal for the prohibition of HSFTO^r in the domain of H^F and ^q^^f. He suggests that instead of making a special statement of prohibition, it can be achieved through some other device. For instance, he introduces the device of sjfaTflfWFT to obtain exclusion of sr^TWHsn in the domain of proper names. P. 1. 1.33: #qTf^^%ortTfnq?M?TM *TO9Tram£^!*m teaches that ijf, qf, etc. are called $ ^ r o ^ when each one of them refers to fixed direction except when they are used as proper names. This rule, says Pat. is split into two : ( 1) ^^r^T^%^tTfnTO v ITrf^ ^^??*TR; and ( 2 ) scH^m^. ™ s artificial splitting of one rule into two is known as ^nfir^i^T. A qtofirvrFT is always resorted to in order to obtain something desirable which cannot be achieved by means of the single rule (^rrftrvrTTR^%^^i^^T^). In the present case the second rule ( s r ^ r a r a O IS interpreted as a general statement meaning that the name s^Nm^ is given to *H* etc. when they are not used as proper names. 2. BrB^mTftr^HgrEf^^Tnci; etc. The rule arcr^W^ after it is split from the single rule is disconnected from the earlier context. It, therefore, becomes a distinct statement ( sTHJ^fir^^). However, it expects some subject item to connect itself with, in order to complete the meaning. The subject item is provided by sr^rff^ etc. 3. crer ^r £fir%3T. etc. ^fSr%^r here means a list. Different lists such as H^5f^» I^TOWT... etc, which are given in roles dealing with s^nsw?!! are all continued in this rule and the word «RJ3TPim is connected with them. The rule thus means that all the words mentioned in the rule defining *W1-H^ get the designation sNNn?T3 except when they are used as proper names. Why the author says ^ HsfT * the name (?3WW^ is) not (expected ) ' is not clear. G. B. Palsule suggests that * ^ H^T ' implies that the proposed ^rrf^vrtiT is to be done in the sutra in the gampotha rather than in the sutrapatha. If it is introduced in the siltrapatha, the prohibition arHSTRT^ also will be restrictecj to the nominative plural as the

ii. i]

Notes [3.10]


1. hi this and the following passage the author discusses Pat.'s rejection of 3^*?^!srfif%sr. Pat. uses another device for P. 4. 1. 14: 3TgTO$*n?t, which is an sifeiR rule governing a group of rules teaching feminine suffixes, is rejected by Kat. in Var. I on P. 4. 1. 14 : a^f?3^TTOfWCT%H; 5TT%7T%^T <T^?crf^fafiPTsrPT. ^Q rule thus rejected in the domain of ^frreros is interpreted in a different manner by Pat. so as to avoid the grammatical operations due to e ^ O T ^ to an 2. qR"?WE5i^?i: etc. The objector argues that the rule iT^r^R!^ is necessary in the domain of "®\%&m to derive forms like CTftTOT^T. The form ^fbsr^STT^ *s oa of a female potter *, for instance, is derived in the following way : ( P . 3.2. 1,4.1.15,4.1. 120) (P. 2.4.71, 1.3.3, 9, 7. 1.2, 7.2.115) ( P . 6. 4.148) ( P . 7. 2.118) The snf^f% is effected for the initial 3" by assuming the stem Ipvith reference to the suffix ^ . However, according to the Pari. ^ nm?$ fe%r^cr^T%: JTf^ ( See Mblu below ), ^if^, and not ffnroft, has to be regarded as the stem. The q%TW teaches that a mention of a suffix should be understood as a reference to that part which begins with the element to which the suffix is added. Consequently, the mention of w\ in P. 4. 1. 15 will be understood as a reference to ^r^f and not to ^ w ^ I ? (since the suffix sn^ is added to $, not to J**T-J-f). mft being thus the stem ending in #H» the suffix ^ is added after gpift, not f5WRft. Therefore, the an%f% will be effected for the initial 3?r and the wrong form instead of the desired one ^«T^5TT^ will be derived. This difficulty can be avoided by accepting the rule -w^ Being an srfir^rc the rule is continued in P. 4. 1. 15. As a result the suffix s r ^ stands for that form derived by adding 3T^ which is the main member of the compound. In the compound fWPT*, ^T^ is a r g w ^ i (main member). Therefore, on the basis of the arfira^ rule srgq^sTT^ the above qftvn^T is discarded and in the absence of ^T"Rrf^i*r ®rt^ stands for Mbh. on Var. 5 on P . 4. 1.14 : ^ efff

Mahabhasya-jDtpina, Mnika



This statement of justification of the ^ f e R £ arjq^fsPTT^' is not quite acceptable. Firstly, there is no good reason to discard the ri^Tf^ft^H by merely admitting the srRr$rc. Commentators like Kaiyata explain this as due to mw^ ' the force of the statement of the affirm • ( otherwise it would be redundant). Secondly, the un-Paninian compounding of ^T^JT? as *§&xm %H: Is accepted in order to prove that cf5R is the main member. 3. «r %^ f T ^ f ^ etc. One may argue that the assumption of the stem ^?¥P5R is possible by another qft^nqF, namely, ( No. 28}. It teaches that the mention of a krt stands for its mention along with the preceding n% °r qsir^s. Accordingly, the mention of the krt «TT in P. 4. 1. 15 stands for the mention of 351^ alongwith the preceding ^ 1 ^ viz. §>wr. IF^TOI being thos the stem, the problem of 3TT%rf^ in the derivation of ^w$Tf*? is solved. The author replies this argument by saying c f f ^ s ^ W ^ * &\ is also . What he means is this: The above qf^sn^T applies only when the krt is krt and only krt. If a suffix is both krt and not krt at the same time, the Pari. does not apply. The suffix ar^ is taught by Panini both as a krt and as a taddhita ( P . 4. 1.83). Therefore, the qft*mT does not apply to 3?R3\ The Mbh. (ibid) t f^srf*t *Tfa^=?^=fe?Tfo sf0}

1. 3T3Jtar*T^fir etc. The author shows how the words can be derived even without the rule ar^q^brffiC. The feminine suffixes are exempted from cf^if^f^nTi stated by the qfoTTfr mentioned above, by the addition of the word sr^r^Rfcr to the qf^vfiqj ( Mbh. on Var. 3 on P. 6. 1. 13 srmJTf^r 3**?T?H c r ^ % o t ^ ^ ^ # 5 ^ ^ % l). P. 4. 1. 120 : # y ^ t &§ contains reference to ^%w%. Therefore, the gr^jf^nw is not applicable there. Consequently, ^VT^RV ( and not only %jft) will be regarded as the stem with reference to the suffix ^ . (See Pari. No. 26, 5 iff in the Paribhasendusekhara). 2. ^ % : m m ^ $ [ etc. Due to the restriction stated above the statement of 3^Tf??f?R*? is not expected in thQ rule 3, qiregqsR^r et.\ The exemption of feminine suffixes from rT^rf^leads to another problem. In the word q w ^ r ^ W P ^ P " , samprasdrana in the place of ^ o f q^R35Rta*T?^ is effected by P. 6. 1. 13 : ^q^: STJTErrcot 5^* Tf a word ending in the feminine suffix nq^ ( P. 4 . 1 , 78 ),




is compounded with q ^ j , in ^dNn?^ Since rr^ifeftq?? is not applicable to feminine suffixes, ssj^ in P. 6. 1, 3 can stand for not only ^JTffPF^ ^ u ^ also q^T^rf^JF^lT. In the same manner samprasamna will be wrongly applicable to the word sifasBTTfcrnrangTf: and the undesirable form n will be derived. 4. 3^1% sn%f^%ni; etc. The wrong application of P . 6. I. 13 in the derivation srf^Tft^F^Tipr* is avoided by introducing one more amendment In the qf^?qj teaching acpf^ftqq. The Mbh. on Var. 3 on P. 6. 1. 13 mentions this amendment as 3T^forr!$Hijqf^«f* It further reads : zft % rer qf**?rqr Sc^^BIfot Wile*? rT^jNpr *ra?ftfal Thus by adding one more word, namely ggjrqfrjfo> the difficulty is solved. The rfTif^fiTOr which is not applicable to feminine suffixes in general is now made applicable to those feminine suffixes which are added after 3W3R. In the compound 8Tf^3f5i^cpTr'<^n, $Tfttniran is an ^qn^fa ( because the compound is dissolved as srft^i?^: ^pfam«9T*0 ending in the feminine suffix <s?f. Therefore, gf^lf^#R5T is applicable to it and so the samprasarana by P. 6. 1. 13 does not apply.

1. qfon^T firWT^ etc. The rejection of the s r f ^ R ^q^T^TT^' by Kat. flfc thus finally accepted. However, the rule is not discarded. It is regarded as a q%nqr * a rule of interpretation \ The status of the rule gy JHM>3»UC{ is thus shifted from arf^lT to qf^^n^T. An 5^ft^R is a section-heading rule without meaning anything by itself. It is mainly meant for its continuation throughout the particular section headed by it. On the other hand aq%TNf is a rule of interpretation. However, Pat. describes it as one of the three kinds of s r f w * . The Mbh on Var. 4 on P. 1. 1. 49 : srferrV ?rm & W : I ^fk%^%*TW *?£ ^ ^ f £ r e ^ 9 % * O 5 # ? : g s r s ^ c r : m ^flrfiri^^f^ I Nagojibhatta T? in his Paribhasendusekhara (Pari. 2-3 ) remarks : a r f q ^ H ^ ^ qi^^rfcqf^TOI T&m I A q"R*TF?T is thus present in all the rules wherever it finds a scope for is application. 2. ^STTTr^srcfen- ^ n e ^ i n d °^ ^ ^ ^ T is traditionally known as qitqP-nq^-qf^Ti^T. When a q^ri^r recommends use of a certain q^ in certain rules it is described as q^q^TTq^j. There is some indication in the rule itself which invites the presence of that q^. Such indication is called %^ of that qfT*nsrT« In the present case, when the rule a r g q ^ ^ r ^ is regarded as q%nqf it means ** in the rules where sr or «r^ is taught as a substitute, the word 3Tgq*n?;r is present \ This is a q^tqp:nq33-qfwTT since it prescribes the presence of the q^ ' ^ q ^ ^ '. The %fj? for the employment of this is ^ or 3Tr£.


Mahahhasya-Dipika, Ahniha VI


3. fgqT n\ etc. The rule sf^q^^Trj; which basically consists of one word is here regarded as containing three words : STgro&waJ^ when it is understood as a qf^TTOT* ST^qrereft is understood either as (I) genitive singular argqspffi^J w*th ^ e case-ending dropped (<p?fws?%*fi) by P. 7. 1.39, or (2) a reference as a nominal base. In the latter case the suitable case ending is added to the nominal base in the rule in which the word is continued according to the requirements of interpretation. Thus the Pan*, s r g w ^ ! ^ {?*rgqH^f-if-W^) means that the word icgqsdbr is present in the rule where m or anj is taught. The word ergqfr&i stands for either »i j'MH^«n£3 or for any other inflected form of the term In this reioterpretation of the rule sfgqsribfFt ^a*# n a s assumed that among all the grammatical operations due to gr^Mi^ only «r and ^ s u b s t i t u tions result in wrong forms if they are effected in the domain of Such rules are, for instance, P. 7.2.102 : c^rj^rsr, P. 7.1.25 : «re etc. It is, however, not clear how the application of rules like ^% (P. 7. I. 14) to 7qo*r«rcan be avoided. In the next passage the author sums up the problems arising out of the acceptance of this

1. Although the rule ^gqe^rn^in its new appearance achieves certain desirable results, it also leads to some unwanted forms. Pat. puts forward one such difficulty which is discussed in the present passage, 3rfd^<4|^ is the ablative plural form of the word asff^**?^ * one who has surpassed you \ It is a srif^cfrP^ according to the statement 3rc?n^ sPFcn^ ••• (Mhh.). The word g 1 ^ in the compound is 3W&;*. Therefore, the sn^ substitution in the place of the ablative taught by P. 7.1.31 : q^Pij a ^ does not take place in the derivation of arfepr^. For, the word snjq^rfo is present in P. 7. 1. 31 according to the Pari. ^. In order to solve the above difficulty Pat. proposes delimitation of the rule aflr-gq^ircT^to derivations in which ar or 3 ^ substitution takes place after a stem ending in sr. This he achieves by assuming the presence of one more a r in the rule, which according to him originally e reads snjq^^T sr m s ^ . The additional a* assumed thus is interpreted as sr^TfTr^r and taken as a qualifier of srgqwH. The presence of this additional m in the wording of the rule srgq^^rirj; is ascribed by Pat. to the device of sr%sfir^r ' fused reference '. It means that 3 in the rule is a fusion of two T ITS. This is a popular device of grammarians employed in order to squeeze the desired meaning out of Panini's rule. While i r ^ q is a fused reference of two or more words, $r%q is a fused reference of two or more phonemes.




3. m\ < ^ ! S T ^ etc. The device of sr%ffsr%T does not help. It creates some other problems. The author has enumerated a few such problems implying his dissent over this solution to the problem of ^ as an alternative for ^ l f i ^ ( I ) P. 7. 2. 102 : c*^pffcm?: teaches w- substitution. But it is not conditioned by a stem ending in sf. Therefore, the Pari. $f«fq^«{i3; is not present In this rule. Consequently, it wrongly applies to nominative singular form a r f i l ^ ' o n e who has surpassed that' in which thesg^TO^ *<i^f issw&sf. The desired form issrfareR: whereas the wrong form by the undesirable application of ^ p » f a p ^ would be scfiter: ( see 3 below ). ( 2 ) fsra% is a srfsftff meaning fi^r £T ®W ' One to whom two are dear \ The pronoun % at the end of the compound f&q% is 3 r q $ r ^ Still the ST substitution taught by P. 7. 2. 102 wrongly applies to the nominative singular formation and the undesirable form *fswgr results. The Pari. arjqH ! R T ^ is not present in P. 7. 2. 102 since «r substitution in this case does aot take place after a stem ending in ST (but in ^ ) . ( 3 ) In the nominative singular form of srfecrer; not only does P. 7, 2. 102, but also P. 7. 2. 106 wrongly applies due to the non-effect of the qf^iraT. P. 7. 2.106 : w%: w- STTWr^Rfr: teaches ^ substitution in the place of g; and % of t*?3[Tf% pronouns before the nominative singular ending. This substitution also is not conditioned by stem ending in sr Therefore, the Pari. db is not present in P. 7.2.106. ( 4 ) P. 7. 1.14 : ^<f?s?: ^tr does not teach S T or sjrj; substitution. R Therefore the restrictive Pari. aijrM^vJHH *s n o t present in the rule. Consequently, in the dative singular of wrfiispt * one who has surpassed all \ where the pronoun qA is ffqffjR, P. 7.1. 74 wrongly applies yielding the undesirable form *«Tftr$rt^ instead of the correct form To sum u p : The rule anfqsriraT^ in a new disguise and additional equipment furnished by Pat. is still unable to check incorrect application of rules to the stems in which flfHW^ is

1. The device of the Pari. s r j q H ^ T ^ ^ a ^ s t 0 achieve the as it is clearly pointed out at the end of the previous passage. Therefore, another solution is introduced in the form of a maxim a m f ^ i r 3 ^ 3 % ^mPlf^WTE^igr *ra% I This is stated by Pat. as an alternative solution. It teaches that the grammatical operations stated in the STft( P . 6. 4. 1 : 3j^T) are effected when case endings are added aftey


Mahabha^ya-DlpikSt Mnika



the stems that are mentioned in the rules, la other words, the rules in the w p f e p f apply when the connection between the stem and the following smii\ is noc mere juxtaposition but semantic. Accordingly, since a caseending added afttr a compound is semantically connected with the mala member of the compound and not with ^ H ^ T , the subordinate member, the rules in the ST^if^R do not become applicable to 3W^ffi- ^ T O ^ r f i w r Is then automatically achieved. Before interpreting the above Pari. and comments of the Mbh. thereon,the author of the Dip. enters into a lengthy discussion on various ways In which the rules in the ST^ife^R can be applicable in the inflection of compound words ending in a s r t ^ n ^ . Finally, he points out that all these ways account for some forms but suffer from either non-application or over-application and have to be therefore abandoned. In this way he prepares a background for this Pari. and then explains it following Pat. 2. sr^tsg^f^STrSf etc. The relation between the ST%ER word sr^^r which is continued in each rule in that section and a word representing stem in the rule of application ( JT^T ) is that of f§r$Nuift^nrre * qualifierqualified relation*. The case-concord which is the usual feature of f^Ffcrris sometimes absent in this case. For instance, words like which are used in ablative hold the flrctaorf^fcTOleT with which is in genitive. Therefore, the author says 3rg^?fiRT£} <psrf^T etc. The reading given here is an emendation. In the Ms. the word 3T^R%TTRT is missing. However, the reading suggested here can be easily reconstructed from the context by looking upon the omission as haplology. The fiiretaGifHr^rara being thus established between s r ^ T on the one hand and the words representing stems in the sr^T rules on the other, the author further suggests two alternatives : ( i ) 3 * 1 ^ is the fgrlr^? while the other word is its fgrtrq*! or ( i i ) sn^sj is the f%$r^oi of the other word. Here the author discusses the results of accepting the first alternative, Gamely, mw*% as the fi£i«a< and the other word as its fi Now, the very notion of firctq-ai introduces another Pari. while interpreting rules in the srgnfsrqBTC. That is the cr^fM^rfonsri ( P. 1. 1. 72, see the earlier discussion on this Pari. in III. 10, 11 ). According to this Pari. a f^faoi stands for that which ends in it. Applying the firer^oif^p^^Tre and this Pari. to P. 7. 1. 2 5 : ar^|cRTf^¥q: q^¥^:> one arrives at the following meaning: The substitute s r ^ replaces ^ (nominative singular ending) added after a stem ending in a ^cRT*3T word. (Incidentally, it should be noted here that the word ^r<[ in the rule stands for ^rn?cT by the Part.


" Notes


inn?:} 3 W 3 [ [ r ^ ^ T ^ f ] being a ikm^l of «Cf further o stands for ^cRFcCPcf b y the a^rrfsfa ( P . 1. !. 72), For instance, =p?r? is ScPTT^ while <ffier$?nr or q^T33cl? is ^rffpcfi?^. Here, following the usual practice of commentators ^ ^ T F s r is expressed by the word ^ ^ n . :5?rt, a word, comes at the end of the compound q**r^r. Therefore, meaning i ^ i w ? ! ^ ! can cover such words. Accordingly, the sre^can take place at the stage q**npcTT-t-^ an( ^ * n e ^ es * re ^ form ^ derived. However, this alternative opens the way for some undesirable forms sush as *$jrfissRR3" instead of ^rf^^ci^r where the m^ substitution is not desired, since ^y is ^qjg^f. ( a r f ^ R ^ - a r f i i a B i ^ m?*i * that which has surpassed, which of the two' is a srrf^ compound in which the meaning QI 8ff?T is the main meaning while H^CR is subordinate.) The application of ?r^??Tf^fif and firstirof^FSTOre 3'n this alternative thus fails to achieve the desired ^q<g&Tsrferfcr.

1. The other alternative is discussed here. According to it ^%% is. fg^Ifori while ^^rf^ is f%%^. Here the cf^cff^f^q'fwrqr ^ o e s n o t apply, because of incompatibility ( s r i T ^ Mfti * ^cR ending in s r ^ ' is an impossibility). The rule P. 7. 1. 25 then means that arg^ replaces ^ added after ^cTTI^cT stems which are srf. The rule thus applies only to stems SSKIT e t c and not to stems ending in *&^ etc. The difficulty of the wrong application of P. 7. 1. 25 to srfarqj^f + ^ is thus avoided and the 3rqsHfcra%^ is achieved. However, the fault of non-application (srsqrf^) results in this alternative. The application of P. 7. 1. 25 which is desired for the derivation of the word q^fr^cf?^; is prohibited as it is not 2. ^*TTfcfcJcTC[ a?Tr^^ etc. 3T%gs*T3[ and sm^T^r are the forms o f ablative plural of the words srf^RjsJT^ and ara^r^ respectively. The ablative plural ending s ^ added after g^T^ and ST^R^ is replaced by ST^ by P. 7. 1. 31. Due to the reversed fwtn^ffw^r^Tvrr^ and consequent non-application of ^TT^ffirf^T, P. 7. 1. 31 is not applied to srf^pqcf and s T r ^ j ^ . The present alternative is thus unable to explain the derivation of three forms, qfsr^pcRT 8f%g*JT3[ and s r ^ ? ^ .

1. Both the alternatives yielded by farw^fetr'^re 1 thus suffer from faults of over-application and non-application. Here the author views at the relation between the 3T%$R '3Tf?T and the words representing stem in a


Mahabhasya-Plpikci, Ahnihi VI


different way. He suggests pari-and-whole relationship. P. 7. 2. 102 • si^tetfr- can be, for instance, interpreted as follows : sr replaces (the final of) ^ r etc. which are parts of a stem ( m^) when a case-ending immediately follows. The rule therefore applies wherever ? ^ etc. are part of wif. The nominative singular of qRffiRtsr, namely, qRUH', is correctly derived because of this relationship between rfcf/ and the sri7 s q*Hcra[ \ However, here also the fault of over-appiication entails. The nominative singular of srfe^; issrf^rct. But according to the above interpretation of the rule, it applies at the stage fric!^ -f ^ and the wrong form sniws: is derived. In short, the new relationship also is unable to avoid application of the grammatical operations to is regarded as a qualifier. Therefore, naturally the relation between m^ and the stem-representing word is partand-whole. The genitive added after m^ can be, therefore, correctly described as partitive. The author describes it as fsrcfaoregV in a broader sense. Thus 3T^3 is a f^i^ai of ^^nf^vq: in spite of the absence of concord of case and number. The word fkwtw has to be understood here as a differentiating mark. The KaL on P. 2.1.37 says : $^=fc fgrffa^ ' a f^fao? is that which differentiates9.

It was already staled that the application of the ^r^tfirfw in the rules under the STfjfsrcpR" does not solve the problem of s w ^ R - Now the author suggests the application of cr^rrftf^r not in the 5R[$T rules but in the ^sp rule itself as another solution. This he describes as a 5n=Rfoi3> 37rftfef^r. The definition of the term sdfaW^ forms the sr^oi 'topic '. This alternative is based on two assumptions : ( i ) the ^ P F T T ^ SHR1*! means that Panini defines the names which stand for all; in other words ^ ^ R ^ is by itself a significant term; ( i i ) the ^?cff^f^ is applied in the definition ofe^nTO^; The word is regarded as a qualifier of ^^mif^T. Therefore, by employing it means Hwfep^fo ' ending in m etc. \ Thus the rule means that the words ending in ^ etc. and standing for all are called ^ r m ^ . (Of course it is assumed that the application of the ^rcrfefw does not prevent the sHbriJTCNn from being applicable to single words ^ etc.). When a word like &4 etc. coming at the end of a word is either a ^^ri cr 3:^33^, it cannot be called *HFU*I3;» because it is not a name for all. In this way, the ^ H ^ s r f ^ N r is obtained by means of 1. ifaTOT*r£OT1% srf^qft. w r c r a i s a word ending in a Tfie word *j<f in this compound stands for all. The word q**TH^ also stands.




for all. Therefore q ^ ^ can be a ^4^HHS according to Therefore, the dative singular # added after qq4jjf4 is replaced by ^ by P. 7. 1. 14. Here the word *$ is neither a #^r nor an : J T O ^ . Therefore, the application of P. 1, I. 14 correctly derives the form <r*i?sr^r. 2. ^jfasi^fo etc. The counter-example of srmfot^ ^PrrfscRr is srfirn the compound srfg^l which ends in a s r ^ m ^ the word *Hr is The word 3TRTH4 ' one who has surpassed all' does not stand for all. Therefore, it is not a gfj^m^, and so P. 7. 1. 14 does not apply. The dative singular suf^rafo is thus correctly derived. ^ R T ^ T ^ cj 5? ^r^m etc. Although the application of properly accounts for the non-application of P. 7. 1. 14 etc. to words like affHsHr* it cannot prevent application of rules like ®&-^ 33?rf^33T' ( P. 7. 1. 25 ) to ^TTO^Ffs, because the P. 7. 1. 25 is not s r ^ f e e . (It is not stated with reference to ^ ^ n ^ which is the sr^T^T.) Therefore, rules such as P. 7. 1. 25 are wrongly applicable to srft^TT resulting in the undesirable forms *3T%3*R^. In this way sn^fii^-cf^rfirBr only partially succeeds in obtaining l 4. 3T5TT^?fti% 3 ^?ctfiiRr etc. <?^ffev:fr in the rules in the is 3T!3wfeE, since the rules are not stated with reference to any * topic'. This kind of ^?crf^fi| which is suggested by Kat. in Vat. 8 on P. 1.1. 72 rirqhR gsRTOW*0 etc. is already treated by the author (see III. 14,15). 5. ^rvnrer spfftRsrffrfar. It has been already pointed out that the faults of over-application and non-application prevail in both the alternatives issuing from the application of cr^reff^rf^ and fwtr^lfirlpaww (See III. 14,15). 3.

1. Having pointed out that the various possible ways of interpreting the rules in the sr^jfv^TC and their application to compounds ending in sHfafH^faslt0 achieve the purpose of ^qg*fa5n%%*sr, the author now turns to the solution offered by Pat. Pat. deals with the ^jimmcrf^^T quite extensively and finally accepts it as one of the solutions to the problem of 2. H«J^ar * fw^r^^T^- etc. While explaining the meaning of the ^ w the author speaks of two kinds of relation between the stem and the case-ending following i t : ( 1 ) fanSR^oi: 'characterised by enjoining ', in other words, mere juxtaposition, ( 2 ) 3TshJVcRf7tf*rrr:, * meant for the sake of indicating the meaning (inherent in the s t e m ) ' or semantic.





Ft is the latter which is intended by Pat. in the word Tpgirrarf^wfife, literally meaning ' a case ending which is added after .1 stem that is directly mentioned \ The Pan. thus teaches that the rules in the sffffeF? apply when Che relation between the stem and the following case-ending is semantic. For instance, the case-endings added after the word ST^NKR, in which the H^^TW^ 3KR is 3-qxr^r, are semantically connected with the meaning of the whole compound and not with ^ ^ . Therefore, rules such as P. 7. 1. 14 do not apply to this word. The ^ r o ^ r s r f t w is thus achieved by a dopting this Fan. 3. M^srhmfiffinT:. Tradition speaks of five meanings understood from a case-ending when it is added after a stem : ^ p j « the meaning of the stem % 3p? 4 substance ', f§s^ s gender ', %w*\ *-number* and ^ R ^ 'relation with action '. There are two views regarding the denotation of meaning by caseendings as pointed out by Bhartrhari In VP ( If. 164) : ^lfe$T €rfcr$T ^T sg^fcnftof fif¥Hf>q: \ The case-endings either directly convey these meanings or they merely indicate the meanings already existing in the stem just as a lamp illuminates the objects already existing in darkness. The author here refers to the second view by the word 4. ^ 5 f^f^* j The ablative plural forms of the words srf^pj?^ and ^ are srfgrgcjT^ and 3Tc9Wfs respectively. ( For their derivation see Note 2 in III. 15.) If the *rimmf?*uqT is accepted, the rule cannot be applied, because the case-ending is not semantically connected with and The author has pointed out this difficulty without offering any solution. However, Kaiyata deals with this problem. His reply to this is as follows: g ^ I Kaiyata means that if the relation between the stems 3 ^ and -^m\ and the case-endings is assumed to be that of juxtaposition ( ^ q ^ ? g r ) , there is no difficulty in applying P. 7. 1 . 3 ! to » r ^ p T ^ 4 - *q?£ and

J. The MbJu brings forward an objection by the i ^ q % ^ who has not properly understood the Pari. qft??q^ is a nominative plural form of the stem q ? * ? q ^ . The case-ending is dropped by P. 7. 1.22 : q^psfr g^ f and after the elision of the final ^ by P. 8. 2. 7 the form mm^ is derived, According to the ajinqfoiqf^nqi, P. 7. i. 22 is applicable only when the case fading ^ i s added after a q\stem (P. 1. 1. 2 4 : f » i p a i ^ l a numeral ending

12. 14]




in %or ? is called <f^'). Since it h added after q w r a ^ ( which Is not **%, and not after q ^ j which Is q\) ?, 7, I. 22 cannot be applied. This objection is raised by the t£fqf§R by assuming the flrst kind of relation, namely.. f^TR^^rfj]^^ (see note III. 18.2). The author here refers to this as one of the two possible assumptions by the ^ q " % ^ . Another assumption is explained in the following note. 2. sra WT c^T^faro; Qtc. The author dwells upon the example a little longer and thinks of another possible assumption by the ^ q % ^ . P. 1. 1. 68 : s^r ^ t ^s^TreTs^g'T is a Pari. teaching that throughout the grammar word's own form is understood by a word. When the Pan. applies to rules in the sr^fRp&R, the word 3jira!*ifirs?% in the ^jwiif^qf^TFfT has to be interpreted as that case-ending which is semantically connected with a stem which occurs in its own form. The author here refers to the rule t ^ i ^ R i n : ( P . 7. 2. 102). The rule applies when ^R[ etc. appear in their owe forms before a case-ending. Similarly, in the present case qr^zjt <g^ ( P» 7. 1. 22) applies only when* the case-ending 3T^ is semantically connected with a q\ in its own form, q ^ , not T ^ T T ^ is a ^ . At the stage q w r ^ + ^ r ^ , st^is semantically connected not with the qf^r (q^?Q in its own form, but with qRHW^. Therefore, the rule will noc apply. This second assumption based on s ^ q r s n r * adoption of word's own form ' is rather farfetched. Pat.'s answer in the next passage suggests that he considers the objection only on the basis of the first assumption,.

1. *f^mF? 1<S *mw* Pat.'s reply to the above objection is that the case-ending is connected with f \ (i. e. q ^ in the compound because q r ^ i s the main member of the compound ( i H + i q ^ is a and hence ^TRcr^rT^T )• This reply implies the second kind of relation,' namely sr^^tcT^firftTT ( See Note 2 in III. 18 above). As there is a semantic relation between . qm\ (in WTsJ«0 and the case-ending added after it, P. 7. 1. 22 can apply. 2. rT^Rrftr% etc. Now tbe author replies the objection based on the second assumption. The Part, ^f ^f... etc. does not mean that a word's own form only is understood in grammatical rules. In fact, the form is just an indicator of the meaning. Therefore, ^f ^qrq; means ' a word's own form conveying its meaning \ Therefore, even according to the Pari ^ %y•- the relation between q ^ a n d the following case-ending can be estabr lished. Although the case-ending is not connected with ^\( i, e. q ^ ) in


Mahabhasya-Dtpika, Ahnika VI


Its own form, it is certainly conoected with the meaning of n\ ( because the meaning of q ^ i s the main meaning of the compound ). 3. ?TOT Q ^ 5 *T^rtJ... In order to emphasize the point that the caseending is connected with q^f which is the main meaning of the compound, the author supplies an Illustration. The Pari. STrnF^fcr sq<T%sn *rafe? I teaches that sometimes when a reference is to be made to many items together only the main item is mentioned. For instance, when a king is going along a road along with his servants, elephants, horses etc., people say, 'the king is going1. Since the king is the main figure in the picture, he is only mentioned. In the same manner, although the meaning of the word q*?c also is understood along with the meaning of q^f from the compound q w m ^ f it is the meanlog of q ^ which is primary and therefore considered while effecting grammatical operations. As a matter of fact, the whole discussion dealing with the argument introducing the Pari. # ^ . . . etc. is not much happy. m

[ 3.21 ]
The author of the Dip. introduces another objection against the Tpw?oiqfc*nqT. He is reminded of another word containing the word q ^ ( , namely q ^ . The word is derived by adding the suffix ^ to q ^ by P. 5.3.70. While deriving the nominative plural form, at the stage q ^ $ + 3R^ P. 7. 1,22 : sr^fr ^ becomes wrongly applicable, because q^f> is rfrB In the sense previously explained. The suffix ^ being a s suffix, the word q ^ h a s the primary meaning of the derivation and so the case-ending ^ is semantically connected with q i i ^ , a q^. That is why P. 7,1. 22 becomes wrongly applicable. Although the author leaves the objection unreplied, Kaiyata answers it in the following manner: ^ft f & ^H: i What Kaiyata means is this: After the suffix ^ is added after it ceases to be a q^ since it does not end in ^ . Therefore the application of P. 7.1. 22 is automatically avoided.

[ 3.22 J
In his reply to the question regarding the derivation of q^q-af Pat. explaines that in a compound a i^mn represents the main meaning of tie compound. Upon this the objector raises another objection. The word ft€f$ff2I: fsH srffsf z($q * one to whom a shaft of a cart is dear * is a compound. The instrumental singular form of the word is derived in following way r *

12.21 ]



(P. 7. 1.75: -f an (P. 6. 4.134 : ff^tqrs^r: i )

Here according to the objector, P. 7. 1. 75 : which teaches substitution of S R ^ in the place of the final f of ^rf^r etc. does not take effect if the sjirsnii is that which represents the main meaning of the compound. Since SFW^TST (viz. brahmaiia) is the main meaning of a 3|ptff[ compound, ??fip:r cannot be said to be the main meaning of the compound fsRSff^T. Since the case-ending STT is semantically connected with the m^l^J^ ( one to whom shaft of a cart is dear) and not with P. 7. 1.75 cannot apply. Incidentally, it should be noted that in the example fsrrofipjr the word does not mean a thigh. Otherwise by the affixation of the girraFft q ^ taught by P. 5.4, 113 : sfcftfT OTWWfft: S^TfFrai <f^, the form would become fsrqH^r. A noteworthy feature of this and many other examples such as srfir^f?, qw$3T etc. is that they are meant only for the sake of argumentation since they are hardly used in reality.

[ 3.23 ]
Pat.'s simple reply to the above objection is to make such cases aa exception to the 3jin?T»JHTfwF>rT. He therefore states another Part. *rsnfrftff& 3f^% Sf afa*rar ^ STcfft. This Part, teaches that when a grammatical operation is taught by a rule containing a locative ( denoting the following condition ) that operation is to be effected when the case-ending is connected with the sr^n. In the present case, P . 7. 1.75 contains the locative condition scRr fih*T$ which is continued from the preceding rule. Therefore, by applying the Pan. we can effect the S R ^ substitution in place of the final % of nfer when the case-ending is added after fejHf^r (and not The word xrfcrf^TW is rendered in the Pradipa as The Uddyota further remarks sreKcW^HTsH^l The word sif^r in this Part, thus seems to mean s ^ according to tradition. This Pan, is an exception to 32^rju&rafii*rr*T in the sense that in the iTf**n^f the connection between the case-ending and the stem directly mentioned in the rule is considered, whereas in this P a n . the relation between the case-ending and the ^ as a whole is taken into consideration. The difference between the two Is observable only in the case of compound stems like fflrarcrf^r. According to ^ u i W ^ T P * ! , SR^- substitution will be






effected only when the case-ending is connected with THPHTOI i. e. the other hand, according to the sr^fetofoqf^OTTL the sr^f substitution will be effected also when the case-ending is connected with the whole stem

It is interesting to note in this connection that in the same passage in the Mbh. Pat. rejects both these Pari.s and justifies the desired forms by means of some other devices. We shall come to that later in 3.27 and the following passages. Till then we have to follow the author of the Dip* along the diverted track,

[ 3.24 ]
Here the author lingers behind and takes delight in further toying with the new F a n . ^nfrt7[f^%... etc. He offers three alternatives implied by the word 5§c?%*?%. The threefold relation between srpT and f^r% is illustrated by means of the rule s r f ^ f e . . . etc. ( P . 7. 1.75) as follows: ( 1) s??Tf replaces mfei etc. which immediately precede a case-ending that is added after an ajf; ( 2 ) ara^ replaces sr%r etc. when a case-ending is added after an sr^ ending in srflsj etc.; and ( 3 ) %mw replaces srfp.7 etc. which are wf. It seems that according to the author thus the term srfcrfwfifi covers all the three kinds of relation and therefore the Part, has a more extensive scope than the ^i?T*Jiqf^TT^T. All this discussion on g$3fgnT% does not add much to our understanding of the Pan. which is unfortunately doomed to be rejected.

1. Coming back to the Mbh. the author introduces here an objection against the P a n . 3T3*ftftf^% etc. The rules P . 7. 1. 102 : ? 3 ^ t a r a : and P . 7. 2. 106 : <f^: W* srra*T?c*J*ft: contain the locative word fir*?^ which is continued from the preceding context. Therefore, the Pari. *fsnfVftf^% etc, applies in both the rules. When the nominative case-ending is added after the stem s r f e ^ both the above rules will apply because of the Pan. and the wrong forms *3rfOT:, *arfaRJT will be derived. This objection is stated in the Mbh.

1. SWTf'T ^ s q * i \ . . The acceptance of the Pari. leads not only to undesirable forms but also to the fault of prolixity. The statement of this additional Pari. leads to jfr^l- For both these reasons the Pari. is, questionable.





The word p r e f e r s , according to Kaiyata, to both the Paris s stated, above. Pradipa : ^^f\% \ iggmvnfkx^^fa *?8ifrfeff§ *?J^?ir ffir ^ 1 > Nagojibhatta reiterates the same remark. The author of the Dip. on the other hand very clearly says that <nj refers only to the Pan. ^nftfirf|%. His interpretation is logically more acceptable since ^ usually refers to the facts immediately preceding. The two commentators, Kaiyata and Nagojibhaita extend the meaning of w? in view of the proposed rejection of both* these Pari.s by Pat.

[ 3.27 ]
1. Since it was pointed out by the objector in the preceding passage that the additional statement of the Pari. HH#^rfi% etc. leads to afrcsr, P a t rejects the Pari. by saying ^ n^s^T^ \ However, the following discussion, dealing with the derivation of 3 ? % ^ etc. implies that he rejects the *pi-; TTWf^Trer also. When the nominative singular ending ^ is added after thei compound aT%cgg^, P. 7. 1. 25 becomes applicable. Its application results, in a wrong form *3rfd<$ff^. As it has been already explained, it is avoided, by means of the ajiHTT^T^n^rT (See I If. 18). Here Pat. suggests another alternative; ^^ crrw^ 3f^^ '^rrn'^^" "TIBT^ ipn

His argument can be summed up as follows : When the srfspgTC word^ ^ continued in the rule ST^I^RT... etc., the genitive s r ^ r cannot be construed with the ablative ^mf^*:. The relation of qualifier-and-qualified , Is supposed to be there between the two. It cannot be there in the absence ~ J of case-concord between the two. Since ^^TTB.^: cannot qualify s i ^ q it has to be construed differently. The only alternative is to construe it with the word conveying suffix (^TR\: in the present context). Therefore, by interpreting the ablative according to P. 1. 1. 67 as ^reif^W ^ R F I and ; further by interpreting STTC^J as f§rf|flF( we arrive at the meaning of the : rule as follows : srfFl ^=RTf^ft f^f|cTF* ( 5 5 ^ ? ^ ? ^-5Tc^^^ ^ ^ r ^ ) ; 3T^ * 3T^ replaces ^ a n d srq; which are added after ^ ? and which are : connected with &% \ On the word firf|^ cf. the Pradipa:

When ^ is added after srfef^R it is not replaced by a r ^ , because according to this interpretation of the rule it is not "ft%r after ^r* ( semantically connected with ^rR) although it is connected with srf ( ^ ^ - ^ q ' ) / The ar^; substitution being thus avoided, there is no possibility of the derivation of the wrong form *3rfecfoH^ . In this way by means of a simple;


Mahahhasya-Dipika, Almika VI

[ 13. 3-

Interpretation of the ablative and making it f^%ffir#P! the desired result Is achieved. Pat. thus implies that the JjsmToi^^n^T *s no more required. 2. sff ?mm fiftr ... etc. The claim that f^%ffiRT^ is the only device to remove the abseoce of case-concord between the genitive 373^3 and the ablative ^cRif^vq: is questioned by the objector. Here the author introduces two more alternatives suggested by the objector. One of them is the devic? of f^T%f^ra$oiTO 'modification of case-ending\ This device is occasionally resorted to by grammarians while interpreting Panici's rules. In order to remove an apparent inconsistency the case of a certain word in a rule is replaced by another suitable case. The author illustrates with the help of an example how this device helps achieve the same result P . 7. 1. 9 : amr fire ~TO contains the ablative s=RT:f and the genitive srf^f is continued. So this is another instance of absence of case-concord. By means offitvi%fiprf^JiT?T»ST^PJ is changed into the ablative 3 ^ 3 ; . The rule thus would read sr^r: sf^R[ $%W-TO' TO replaces fi?^ added after an «rf ending in a short & \ Due to the concord between «RT: and s T i ^ the relation of qualifier-and-qualifled can be established and consequently, ^r^T&f^r can be employed. Similarly, m^ ^RTfi[^: can be interpreted as ^ a t f ^ * : &Wl?l ( s * ^ replaces ^ a n d ar^ added after an sr^ ending in ^ R '. The objector does not, of course, show how f^%finTiT'JiT*? can be employed in the rule s r ^ . . . etc. In fact, it cannot solve the problem of Another alternative according to the objector is to construe both the genitive and the ablative separately with the word conveying suffix : $r*^3 nt firerCTfr* ^ ^ TO * ^ replaces fa^ which is connected with srf and comes after short er'. Apparently, there is no difference between this second alternative and the f^f|[3firfTCpJi view of Pat. The difference lies, perhaps, in the fact that according to the fgrf^rfft^oi view semantic connection between the suffix and the preceding ( ^ R e t c . ) is intended, whereas in the second alternative such connection is intended between the suffix and the a?f. The two alternatives proposed by the objector above can serve as alternative ways of removing inconsistency or discord in cases. However, they cannot lead to the same result. Both the alternatives suggested by the objector cannot, for instance, avoid the form *3rfg^R^. The objection should be therefore understood in a limited sense. 3. wi$a?f%fgr *{ etc. The objector points out another defect in the view. It has been already explained that according to this view

13.131 the suffix is prescribed after the item expressed in the ablative, fo other words, there is a semantic connection between the suffix and the item preceding it. Since such connection does not exist between sprf (in the word 3?%3KTO and the immediately following ^ , the rule mT^ ^nrjr...etc. does not apply and the undesirable form is avoided. Now the objector points out that the rule m^ ^m etc, will not apply also at the stage q^sR^;-f-^ for exactly the same reason, namely, that 5^ is not prescribed after ^f?, but after q^sr^qr. This will result in another undesirable form *q^i^w^ instead of the desired q^i^f^. In the same manner, owing to the non-application of P. 7. 1,22: q^vtfr <p and P. 7. 1. 31 : *m*m sr^ .respectively in the derivation of q ^ ^ and ari&g**^ these two forms cannot be derived.

[ 3.28 ]
1. This passage deals with the author's reply to the objections raised above. The author dispenses with the first alternative suggested by the objector, namely,fin*ftEf3"q%n;??by saying that it is the last resort which is meant to be adopted only if a rule as it is cannot lead to the desired result. Since the rule «rar^[ icm... etc. can be properly interpreted as shown by Pat, it is not necessary to resort to fiwRlifitaf^inT. The same holds true also of the second alternative. In fact, the author does not explicitly comment oa it. However, his words STPSJT m ^ R r indirectly refer to the second alternative. He further remarks that since the ablative ^^n^^> is construed with the word f^f|ft, there is no question of ^?afeflr and consequent wrong derivations such as * 2. 3*4 q**F$cR^ etc. The author's reply to the question regarding the form qrs^TC? is that the form is not known to Pat, and hence need not be justified somehow. He further adds that strictly speaking qpn^cTf^ would be the correct form. It is interesting to note the observations of Kaiyata and Na*gojibhatta on the same issue. Kaiyata while commenting earlier on the Mbh. S T O T S ^ ^ fcf|<n etc. remarks: «r%f# fw^TTOsrf^^ I iN" qrir^rf* %^^Tfq" *ra% i Thus according to Kaiyata, since fw%cf implies semantic connection, there is no problem in applying ^ substitution rule when ^ is added after <TTO$3?, because s^is semantically connected with ^ R ( which is the main meaning of the qri^nrq' compound q^n^r^ ). The Uddyoia observes : q?ri7^R%^ri^T%% ^ W^IFcTC HRJife 1 Nagojibhatta in this way shows his disagreement with Kaiyata. The author again takes up the issue of at the end (see Note 1 on 3.29). 3. 5p-t qw?? etc. The problem of the derivation of the form is solved by the author in a funny way. According to thefitf|^rfMta<*iview





[ Ill3

( P. 7. 1. 22) applies only wheo the suffix sn^ or sr^L is prescribed after *p*;. Since in the derivation of q^i^x it is prescribed after ifffl^ and not f n \ the rule does not apply. This difficulty is removed by the author by suggesting that *§^m should be understood as the dative plural rather than .ablative plural The rule ^fpit <pr will then mean : The suffixes sr^and w®. which are meant for sf^are dropped. The suffixes 3F£ and ST^ added after meant for q\( since the qp^ is the main member of the compound Therefore, the rule applies and the desired form is obtained. . In order to support his interpretation of as a dative plural form, the author cites another rule in which the apparent ablative is to be understood as the dative : P. 4. 3.91 : sng^ftfif^*^*tfifff teaches that the suffix & is added after a name of mountain in the sense * it is his ancestral habitation', when the word so derived conveys the people earning their livelihjod by weapons (i. e. mercenary soldiers). Strictly speaking this meaning cannot be obtained from the present wording of the rule.. Literally the rule would mean, the suffix ^ is added after a word denoting people earning their livelihood by weapons in the sense ' it is his habitation % when the word so derived refers to a mountain. In order to arrive at the desired meaning the commentators change the meaning of the ablative ^Tgw^f^T-. They understand it as a dative meaning sng^snftf^si: W ' ^ o r the sake of ang^sftfiir^s '• This interpretation is missing in the Mbh. However the KaL gives it for the first time as follows: (For details see Siddh. Kau. on this rule%) Really speaking, this is all fqg^fim- The problem of q^qq^ has been already discussed by Pat. ( see 3. 19). There although he does not refer to the word q ^ r i ^ he mentions similar other words like qnr£RT£T and answers the problem simply by saving q^TOH' ... ^rr^ 1 :1 All the discussion regarding in this passage is therefore redundant. Equally redundant is the discussion in the same passage regarding the derivation of srfa^rq;* since it also has been already dealt with earlier ( 3 . 13). The solutions to the problem mentioned here are already referred to (see 3. 13-20).

1- *ta( 3 ^InT^etc. The difference of opinion regarding been already recorded m 3. 28. 2, Here the author justifies the form ^ from the view-point of those grammarians who regard it as correct. The {^%Hf^^rTrjT is interpreted by these grammarians in a different way. The expression ^rRTf^W firffp'n does not simply mean added after the words ^ ^ etc. It means * added in the sense of 3 ^ etc \ The suffixes sra




etc. convey the meanings singularity, plurality etc. already existing in the stem according to this view. Though the fir%F relationship is on phonological \QVQ\ between <%wt and 5TF[, it is on semantic level between ^ ? ? ^ R and -s!^ because the suffix srgr conveys plurality not of ^r?, but of qfsj^ri?. Thus through this extension of the meaning of fgrffcT on semantic level the application of ^ f ^?fff ... etc. becomes possible in case of ^SRKR" and the form is derived. In order to support this interpretation the author refers to a passage in the Mbh. The Var, I. on P. 4. I. 1 reads * (The srf^&ft) isrj^nfip?i^F£' is required for the sake of the names 3Tf, *r, tj<* etc. While discussing this Var. Pat. makes one observation in his Mbh. on the above Var.: trqp^nrf^^ s^l^ft flrafa?^ 1 ^ ^ This observation is relevant in the present context only so far as it suggests that the meanings singularity etc. are inherent in the stem and the suffixes are added to such stems. If we cast a glance at all these solutions mentioned in this passage we find that the author has not contributed any new thought. Just the same old wine in a new bottle !

1. ?m^iT%^r: etc. This line from the Mbh. is an objection against a statement made in the preceding passage. The contents of the passage can be summed up as follows : In some rules in the ar^rf^rapK, in addition to the ' some other genitive words are mentioned, for instance, 3T%i( P . 7. 1.75), c^nftm**; ( P . 7.2.102) etc. In such rules it is possible to establish the relation of qualifier-and-qualified ( fe^WT^r) between the srfv^rc word er^FT or the genitive word ^ and the word denoting case-ending etc. Two options are available in this situation : (1 ) s r ^ r is the qualifier of the case-ending suffix, or .( 2 ) the other word in the genitive is the qualifier of the suffix. The example of the former alternative would be : «rf^ ftrvrrfr aT*»3?T#Tro*r^ ?re% • «nr^ replaces etc. when a case-ending is added after sr^f \ This explains the form as it has been already shown. The example for the second alternative would be : r ^ T ^ f t f^*rtfJR: vmfcf ®W& ' * replaces ^ etc. after which a case-ending, which is connected with a?f, is added '. This explains forms like qr*?*T*. Now in connection with this second alternative the objector says: *rera*rfcTCT5 etc. He means that if the genitive word r ^ i ^ h n ^ qualifies the (case-ending), the a* substitution cannot take place in the form : which is derived from «rfira^+^ because the case-ending is added not


Mahabhasya-Dipika, Mnifa



after a?, but after the whole «ff viz. «f%af. The author explains this obi-JCiiLUi by giving another illastratlon, namely, «rfirR> Apparently, all this is confusing. The word m%^ can be understood in two different ways : { I ) $n%frgW " ^%^F^* cn^4 one who has surpassed him % ( 2 ) ^HW W* 'beautiful he' (The word ?*far is a ^ S C T C T N conveying praise). In the former case the nominative singular is srfaasr; whereas in the latter case it is m?m*« According to the objector, if ^ F ? > u ^ is connected with f^rai, «• substitution cannot take place in the derivation of «ricta: as in the derivation of iscfif^. Pat. gives the same answer which he has already given in connection with the form wrcw ( 3. 19). He says : c^rf^rer* T* mm-1 This statement needs no explanation since the similar expression has been already discussed ( see 3. 19 )• Our author adds a few more funny compounds like ww «f^ etc. which can be explained the same way.

1. am S T k^ etc. After so much of diversion both Pat. and ouf C author are back to the basic p problem, namely, the prohibition of s^N author are back d P t solti in the domain of TOT and vqm*. Pat. suggests a new solution to the problem of * @ and ^ r ^ f . h n He makes a radical suggestion that the rule ..33^ft— should not be understood as a definition. snfom^ is not a TOE it is just a predicative of *reff$fa. The rule therefore means that the words listed in irffem are g*Wfl(s ' names for alt \ Each word In the nJjfa list is thus a name for all. Thus all grammatical operations caused by ?rf*tm*lw i l 1 b e applicable only when the word concerned is a name for all. When *rt, for instance, is used as a proper name or 3W&* it ceases to be a ^ m ^ * n a m e f o r all> * Therefore grammatical operations caused by do not apply to it. The same holds true of compounds such as ^ ^ R etc. which are not names for all. 2. ^ f ^ ^ r m ? w ^ n n etc. The word m^m^ ^ has to be inter* 2. ^f^^rm?w nn preted in a restricted sense here. Each and every word in the *prffi[ list is not a «name for all' in the literal sense. Only a few words like srt, fire etc. convey the meaning « all' and therefore can be literally called a «4<tmv However, words like ^ , mr etc. are not *$*m\ in the literal sense. Therefore the author explains the concept of ^NNTOC^ as ^ f i r a w r - s ^ w c * , with the illustration of the concept of ^ f o . A ^rfir 'genus' is a feature which is common to all. There are some features such as 'objectness' (tprfosf), * being • ( srai) which are common to all the objects in the world. They can be therefore called universal. However, there are some features which are common not to all objects in the universe but to all objects in a

specific domain, i f ^ j s > for instance, a m% (universal) in this restricted sense, It is found in each and every cow or boll that is fouud in this wo;Id. in the same manner words like ?rr» ^T e t C e s t a n ( ^ ^ or a ^ objects within specific domains restricted by contexts.

1. The <5$q%^ asks, if the rule s ^ f t - is not regarded as teaching the name q#^r^; f grammatical operations caused by *f#Hm#17, such as iS*Nw: ^T (P. 7.1. 14 ) will not operate. If the e t ^ m W ! does not exisr, such rules will have scope nowhere.
etc. If it is argued that the word spfcro^ is an adjective of each word listed in ^niif^TM* the enumeration of ^ # , fk*% etc. f^ list will be redundant, because by their very nature they are *?#names for all \ This is an additional fault pointed out by the author with regard to the new solution proposed by Pat,

3. w^^^h ?r^ f^TTFJ^ I Pat's reply to the above objection is that wherever the word g^nm^ is mentioned in a rule, it will be simply understood in its literal sense • a name for all \ The concerned rules would, therefore, apply to such words which are names for all. Since the words 3?# etc. listed in m ^ i m are names for a)l there is no difficulty in applying such rules to them. The fault of the redundant enumeration of fipf, ^pf etc. remains unanswered.

Here is again a little diversion. The author is found recording various denotations of the word *r#. According to him, the word srf" literally meaning ' all' is used in three different contexts : (i) number, (ii) kind, and (iii) a whole composed of parts. He explains the use of the word ^ with illustrations in all these three contexts. The use of the word *?# referring to number is illustrated by two examples : (1 ) ^ mmm'.~ »ft^r *T*F?*r if • Ail the sixteen or seventeen ( people) came \ ( 2) H^fa*rerr§r ?gmfk 'The whole food was eaten without anything being left over \

Perhaps the slight difference between the two examples is that the quantity can be quantified in number in the former case whereas it is not possible to do so in the latter case. In the latter case the word ^ means * whole • rather than ' all \ The word ^ referring to kind is illustrated by the example ^ f ? # * j ^ * (He ) eats every food \ The word *p| here does not denote quantity. The sentence means that he eats whatever kinds of food are (here. In the example ^ sfr^Tt igftt m *n>W?Fri%n ' Whatever rice


Mahabhasya-Bipika, Mniha VI


was served in the dish was eaten • reference is to the part of the whole rice* that was cookm. Thus m refers even to the part and denotes the whole* part By showing this analysis of denotations of the word m the author suggests that if the word mmmy'is understood as ^ 'significant', it. will cover all these contexts.

1. m^ ^ ^ 1 ^ etc. Fat. points out the fault of over-application in this new interpretation of ^hm^. For instance, the words * ^ , ^ ^ etc. also denote all and hence could be described as ST^ra^. As a result che rules containing the word H * W T ^ will be applicable also to these words.' The author describes different grammatical operations being applicable to different words, the so called m*im$. Thus he tells us that the operationslike & substitutions will be effected In the derivations of words ending in short ar, such as ^ ^ while the suffix ®w% will be added after all * ^ m ^ s ; (Although the author says *&m*3ir&'-, **% can be added, as a matter of fact, after any s ^ b u ^ ) ^ T ^ . . . The fault of over-application is extended still further and It h stated that not only the words ^ r , $?** etc. but even the words qualified by these words will become *r*srm^s. Mbh.: fi[ mtitm : *Each one of these words (^3Ry, f ^ r etc.) hag its own domain. Whatever word denotes that domain will, being ( denotative ) of that (domain), be due for operation caused by H ^ m ' . The author gives an illustration: ^ : afcp: 'all rice'. The expression *rf: m$^: taken together conveys two notions, allness and riceness. It can be argued that allness is understood from a f t ^ as well as q$. In that case m^T also can be described as d b r o ^ . Thus grammatical operations like %?^ substitution ( P. 7. 1. 15) become wrongly applicable and the undesired form f*Y^f%nr will be derived. In fact, Pat. does not directly refer to the form sfhprferq;. However, Kaiyata and his followers interpret the Mbh. in this way suggesting that Pat. implied it. This argument is explained by Kaiyata as follows: ^ ^ rf


etc. The author strikes a note of discord with the above argument which extends the,spope of ?&*TWq; to words like




The deliberate use of the word sairaf^: * well-versed in logic f referring to the Bhasyakara is very sarcastic. The author clearly explains,irora • a-layman's point of view, how the word aff^r cannot convey the notion of* allness. Neither is the meaning of snf restricted to rft^f, nor is the meaniog of .sft^r restricted to ^rw. Both these words when put together convey, the notion { all rice.' But the notion of allness is understood neither from H^ nor from rft^r. It is understood from the sentence. 2. TT%^ fc^fcfe etc. The Mbh. on Var. 2 on P. 2. 3. 46 : : W* supports the above view. It states that whatever meaning over and : above tht meaning of individual words is understood in a sentence, that is the meaning of the sentence. From the sentence *pr: stf^r: neither simply allness nor merely riceness is understood; aliness in the content of riceness is understood from the sentence. This is the meaning of neither of the two words, s r t ^ therefore cannot be described as covering the notion of allness.

[ 3.36 ]
1. The same argument is elaborated here by another illustration. In the expression ^fe?j; 37W*5; c a blue lotus' blueness as well as lotusness are understood. For the common man blueness is conveyed by the word sften^ whereas lotusness is conveyed by 3xW^. In the same manner, of the meaning * all rice f the part * all' is conveyed by the word srltf while the part ' rice ' is conveyed by the word * sjj^f: *. The idea of £t%% conveying •allness' does not fit in this simple logic. 2. ^sra^Kfg*T3^f ^T f^qq q^Rt:. This line is difficult to explain. Literally it means that, these two words (?fte^ and ^rcq^r^) denote a meaning which is either devoid of both the meanings or composed of both.

1. s?T3[ 33RJ5;. Here the author tries to justify Pat.'s argument. Although the ' allness ' being conveyed by aft^i: could not be justified from the listener's point of view, it could be explained from the speaker's point of view. Taking up the same illustration sftsnj; T^l^rq;, the author explains that when the speaker uses the word ^fte^ he uses it in the sense in which he has already understood it. He knows that the word ^te^can be connected with any object (like srz, qz, ^Tcq^ Qtc.). Therefore, * blueness' conveyed by. the word ^"V^ can be understood in the context of >dcM^<. When the word s f t ^ is used by the speaker to denote 3^q * object' it can denote ^qgsr as well as any other object. In the same manner, the word sft^r being used to convey rice which is s all' or ' not all' can denote * all rice ' and . hence allness.






2. wm Wf% w$ etc. The line from the Ms. Is difficult to reeoa* struct The import can be given as follows: both the words * *p|:' and 1 fit^f:' are connected with each other in such a manner that each conveys the meaning of both. Such being the case, why is H 4 a *?^TT?r^ but not #^5f ? This point has been already clarified by Kaiyata quoted above.

L The Bhasyakara has to abandon the earlier solution since it suffers from the fault of over-application. It is clear from the foregoing discussion that the word 3Hhrm3 is required as a H^i in order to avoid its over-application to words like %?%*$, f ^ C etc. It must at the same time convey its literal sense g name for all' in order to avoid its over-application to ^Tf and 3W^T- Here Pat. suggests a solution which helps achieve both these purposes. The word ^FHNR^ is a H^T? of H3" etc. and at the same time it means * a name for all \ In other words, the word sHhfl?H meaning 4 name for a l l ' qualifies the H3tf%m, and the word *nbm7^ is a m^t of H# etc. 2. s p | tpr\%5| sfi^if etc. P a t . suggests that the rule is interpreted in such a way that it serves both the purposes mentioned above. The objector asks, « How can one effort lead to two results 1 • One aod the same rule H^rftf^r etc. cannot make two statements. Each word has a power to convey one meaning. After it is exhausted the word cannot convey any other meaning. The word JTf % is used here in the sense of 5P|%#rftrrT ' cause for the use of a word * ( For details see discussion on ^gs^ft WZptt JTf %: in the Mbh.2hn.n). 3. w $T3S iTcf jp|^|: etc. In order to justifiy his statement that one effort leads to one result the author gives an illustration. Since some words in this line are not legible in the Ms., our interpretation is only a guesswork. An effort is required to push a thing upwards. The same effort is not enough to sustain it in that position. For sustaining, another effort is required. The same holds true of words and meanings.

1. Pat. explains how to achieve two results with ore effort. The r u k ff^Htfsr etc. is to be regarded as ffipjtaftqfer ' a fused reference \ It is an abbreviation of two rules sref^tfr *pforarfiir and smf^ft sHbmnfif. According to P. 1. 2. 64 ? ^^qnoiT^t^ ^ f i n * ^ , the two words sref^ft and qrefffifa which are ^^q* * having the same form f are fused into one. so are the two words ^rsRmrft and ^TT7TT%. The rule which appears on th,e surface as

IS. 20]



one single rule thus represents twin rules each having a separate identity ( A reference has been already made to the vn$m sprMt% in the Note T. 4). The two rules can be further interpreted as follows I


(ii) ^m *rft wmfk mfk *
Rule No. { i ) excludes words like f^w, H ^ > etc. from the scope of , whereas the rule ( i i ) excludes H^T and OTST^R". This is how P a t tries to achieve both the results by one single rule. 2. ?PI gprNtnft "T*3T etc. First the author does not approve of this solution viz. ^ ^ m f t ^ r mainly on the ground of its un-Paninian character* In this passage he points out, giving different reasons why the present rale cannot be accepted as i^mfepfer . Jn the following passages he tries to justify Pat. by intepreting QffiR in a number of ways. The presence of many ^ q objects and the occasion of their independent mention are the two prerequisites for ig^ra according to the author. He first argues that both these are absent in the present case. In the first place each one of the words in ^ f i [ list, nameJy *rw etc., is not supposed to have many sr^qs, Nor is the meaning understood from the word ^wfR multiple. sreffg[ signifies a list, and one and the same list is understood in both the proposed rules. There are thus not many entities understood from the word wffi% although it can be said that there are two separate occasions for its use. 3. f %3ftft3rA:'" etc * Another prerequisite for the formation of % tj$%q is that all the H^q words must be connected with the same action at the same time. (The word i t ^ literally means commencement, the word t££tq3>*f means having the same purpose of mention or use ). In the present context the two ^ q words H^T^tft are used in two different sentences conveying two ideas. The same holds true of the pair of sHhmnfa. Moreover, the ^ i rule presupposes the other rule in which srar^fir are qulified as ft^rnnft' Thus the two rules do not belong to the action at one and the same time. For the reasons given above the rule cannot be regarded as

[ 3.40 ]
1. s t W b 5HW ^c. The author now looks at Pat.'s proposal through a considerate eye and tries to justify it in various ways. The first way introduced here is to consider this as not a strictly Paninian ^ ^ . It fa something similar to t£f£ra. According to the author it is ^ rather than



Ahniha, VI


cr^%^, which is meant by Pat. when he says frsf^fiff*?rrs^. ^ ^ is a device, •bv men •> of which a word used once conveys different meanings in different context., : ^ | ^ f % ^ % s t ^ r ??*^R Mmm sffi^sisrfinn^^ &%K( Afyayakoia, p. 318}). The device is used by the Mimanisakas while interpreting the Vedic texts. 2. 2iwi^: %$*%: etc. The anther further clarifies the device of 373T by'means of an illustration. One and the same word assumes different forms when it is utterred by different speakers. Or a musical note produced by a wind-instrument multiplies itself into many owing to resonance caused by the peculiar place in which it is produced. Both, the word and the note, are not actually many. ( Correction I In the Translation ( p . 57, //. -20-21), for Indeed it is...word ghata, read : There is no plurality for this (second) reason also : since one who quotes ( i t ) , utters one ( word) only, saying * (they) utter this ( word ) ghata \ ) / Cf. VP. '11.111: In all these cases one is found to assume many appearances. The jother way round it can also be said that many appearances are reduced to one through the device of cf^r and that thus one mention accounts for many, ^ T f ^ w t ^ ^ ^ l f e ^ literally means the pipelike part of a wind-instrumenf.

1. 5 1 ^ W'^m, etc. While elucidating further the device of ?rar, the author points out that (w^l) words can be mentioned only in two ways : ( i ) either separately in a sequence like ^graj ign* f^nsr or (ii.) collectively in a group, like f^n:. The rule s n K t f t belongs to the second category. It is a collective reference. ; Cf. VP. II. 462-63 :

' * 2. nf% ^m^wi etc. The meaning of this sentence is not quite clear. Perhaps the author distinguishes between the rule as a whole, understood. £s a collective mention, and each part (araraw) of the rule, ^ K t f i r a n d ^sbrmTf^ being separately understood as a collective mention. In the latter case each part can convey many words and it can be construed with the other part in a ^variety of ways. This can lead to a birth of even a hundred sentences. ^However, when the rule as a whole is understood as a collective mention 4here is no such freedom to construe one part with another in a number of




ways. This is just an attempt to find something sensible out of the words. Perhaps some words are missing nft^r the wor %

1. Here the author proposes another alternative fonrsp^r. It is * repetition '. According to the author repetition is inherent in the nature of •words. A word can be therefore repeated conveniently in order to g®t the desired meaning. The rule *refe}fif thus being repeated can convey the two desired meanings. One and the same rule in this way serves a double purpose. 2. 3«n <?5^r^l" g% e f c. It is explained with the help of an illustration how repetition is in the very nature of a word. In the sentence, qr^fr^t ^% '(He) eats five times', the repetition of the act of eating five times is conveyed by the suffix f^rc^ taught by P. 5. 4. 17 : w ^ ^ H fk^l^m^mn^t fcWg^. The suffix ^ R i does not convey five different acts of eating but the repetition of the same act of eating. The root ^ i t s e l f conveys the repetitive action of eating. When the repetition is counted in number the suffix is added after the word denoting the number.

3. mm srafafft etc. P. 3. 1.22:


teaches that the suffix q^ is added after a root with one vowel and beginning with a consonant in the sense of repetition of action. P. 6. 1.9: s p ^ r : teaches reduplication of the root followed by the suffix 3 ^ or q^. Here the reduplication (repetition) is effected of the root which already conveys repeated action. These two illustrations are supplied by the author to prove that repetition is the nature of a word. He concludes that the word which naturally repeats itself is used by the teacher in the rule.


. ;



The same point viz. anffo or repetition is a natural process of a word is again emphasised here by taking an illustration from ritual context. ' In a sacrificial performance known as f fi* recitation of seventeen *nf^%?fir ^ks is prescribed in Samhitas as well as Brahmams (for instance, see the Taittiriya Samhita 2. 5. 7. 1, Satapatha Brahmana 1.3. 5 etc.). In actuality, however, only eleven rks are mentioned. The ritual texts further state that both the first and the last rks are to be recited thrice ( See TS. 2. 5.7). Again after recitirg the last one two times more one gets seventeen Samidhents, Thug (repetition ) plays an important role ritual performance.


Mahabhamja-bipika, Ahniha VI


L Yet another way to get a twofold meaning from the word e^scm^ is to assume that the word has two powers to convey two meanings. (It should be noted that while discussing this and the preceding alternative for mn^$qf the author has mentioned only the word ^ b r o ^ , because in two inter* preta tions of the rule ^pf]^| shown above, it is only the word H^fPR^ which conveys two different meanings and is therefore responsible for two different statements.) .. The author gives an example to explain two powers residing io one and the same word. A person wounded in the darkness lights a lamp with two purposes, to see the wound and to burn ( or to warm ) if. One and the same lamp thus serves two purposes, because it is invested with two powers: to illuminate and to burn. In the same way the word sHhmH. possesses two different powers to convey two meanings. Here is a tacit reference to the principle of homonymy. 3. mi '%$ fifWfflNs?&' etc. The author again takes a dive into ritual and finds out an instance which can explain his idea of one word invested with many powers. The word fgsgj in the rk ^ fe^f^gflfr (Rgveda I. 22. 17) means (s) Self in the context of metaphysics, ( i i ) the God Narayana in the mythological context and (iii) the wooden ring on the top of the sacrificial post in the context of sacrifice. 48 tg% ^ $c3T..- etc. One more case of homonymy is supplied by the word msiWZ in the rk portion %$r nm^ (RV. I. 105. 18 ). Bhar means to say that a difference in the pause (sremf*!^:) in the padapatha can convey difference of meaning. If instead of ?n l H^C l as in the present papapatha of the RV. ( 1 . 105, 18 ) 3akaiya had analysed as qrerst^ it would mean • the moon ' (lit. the month-maker ). Incidentally, the Pp. *TT I H f ^ l means * not once»). — It is rather strange that Bhar does not just mention the actual pp« ( m i HfKf-J )• Also the actual Vedic sequence is : (ar^ift) m $J^G[ f $ : ( not fsiif ?rori?^ a s Bhar quotes ). ( Correction ! In the Translation ( p. 58t //. 20-21), for The word masa used in the sense of moon is separated, read : When there is a reference to the moon, masa would be separated.)

1. "fe%^fe%^. The word ssbrm^ thus being interpreted as f a
mme for a l l ' does not apply to isf^I and STO&T because both of them refer to specific objects and not all objects. Pat. uses the word fi^Ff in contrast with $$ in mwH*a a compound like ftw!:=ft<*fj[?cr; 3 % j | < • One who has left all' the word ^ is i w & T , In the constituent analysis where the form




rfCTftt s used, the word &$ h &$*{*&; There it does not refer to a specific object. However, in the compound ficwf, it abandons its own meaning and denotes the meaning which is denoted by the compound. Therefore it can be said to denote a specific object, when it is an T I ^ J T . This is how the line ?fi|| ^ n f a * ^ — etc. could be best interpreted. Some words in this sentence are not clear in the Ms. 2. nfox m^^^m^^ etc. The author points out a fault if this view regarding sHNm^ is accepted. Since according to this view a s^Nfam^ is always ' a name for all f the term cannot apply to words such of f^, ^ etc* The word f ^ refers, for instance, to objects which are being perceived. The word ^ refers, on the other hand, to objects which are out of sight (*?ft$r). In this way the words f^, ^ etc. are used in restricted domains. They cannot be, therefore, described as * names for all'. 3. creHc£33f£(^3 trqf H*RRTc3^etc« The solution to this problem is suggested by the author himself. According to him the words f^, tf^ etc. can be described as names for all in their restricted domains. For instance, the word f ^ refers to all that is being perceived (sr^r) while ^ refers to all that is out of sight. The author further explains this restricted notion of flEfawqi by means of an illustration. He refers to the concept of snJTfl^ferewr of the ffrft^s. The one universal property common to all the objects in this world is Srir • being \ It is known as q^HTJTie*?. There are again many general properties which, though not universal, are all-pervading in restricted domains, for instance, qtc^ * cowness', SPSF^W ' horseness etc. This latter, kind of snui5^, also known as ®rqTsn*rc?3j, is described as HTHF^fe^1 in the (see%?rf^jjr No. 1.2. 10, 14 etc.). The ^TTOI^fetre and the qrjvu*T^ a r e similar to each other in that both are all-pervading in restricted areas. For instance, the word f ^ i s used to refer to any object which is being perceived. Therefore, it is a

1. If the notion of H^^TR^ is accepted in this restricted sense, it would be fatal to the word ^ itself. The word ^ is a name for all in the literal sense. It always stands for totality of all kinds. Here the author discusses the case of the word *r|. According to him the word *Hr does not cease from conveying totality even when it is construed with another word. For instance, in the expression ^ : aft^: ' whole rice' although the totality conveyed by the word ^ is restricted to the domain of rice, the restricted meaning is the meaning of the whole construction, not of the word &$. The word *rf conveys totality only. In this sense it is *$^f *im and hence


[ i*o. 24

Thus for the sake of the word *pt the notion of i?#^m^ has to be accepted in its broader sense. ^ ; Words in a sentence convey their own specific meaning. This view is exactly opposite to the view expressed in 3. 38 (see notes on 3. 37 and 38 ).

[ 3.47 ]
1. The author points out here that it is the srffofeftrrT * cause of the use of a word ' which decides whether it is a ^ s r m ^ or nor. Since the srfi%£Tf?rtr of *?#, namely 35T^4, remains the same even when it is construed with another word such as m^: as shown in the previous passage/ ^ # is a 3t$«fw^ in the construction *?#: aft^:. However when it is used as a proper name it no more conveys ^ T ^ ^ I 4 totality \ Its sr^f^fiffcrr! changes and so :! ceases to be a 2. frsj ^ fr^t^j^ etc. Here is a reference to the Mbh. on P. 4. 2. 251

In a sacrificial performance the hotr priest gives atfa to the adhvaryu priest to recite some hymns in praise of the deity and the oblation. A Ifa contains the name of the deity in the dative. The Mbh, stated above deals with the question as to what will be the correct dative form of the name ag signifying, the God srsnqfir: ^|?r or 3^5r ? The question arises, because ^ can be looked upon as a substitute of % ^ which is a sNNrW^. Pat. records two views. According to one view it should be ^ f because 33 or f%i^, whatever it is, signifies VFmfih who represents all the objects in the world. SHnqf^r being thus the name for all, the dative form would be ^If. According to another view the dative must be 331*1, because = or f^;, whatever it is, is a proper name p aod h?-vje not a 3P$TOT«^. The author refers to the second view simply because it supports the statement under discussion. 3. gurftftTTTFTr ••• The author explains the use of the dative sp^r as correct when the name %q; is based on the property (*pT). In other words the original meaning of %q^ is extended to the proper name.

ftr Just as the word &$ abandons its own wljen it is used as a sjg-j, it has to assume the meaning of the ma jo




constituent of the compound when it is used as an s w l r a . For instance, ill the compound aff?r^J: the word 7$i being subordinate to the word si%(pfiFcr:) conveys the same meaning that is couvced hy the *vord --rfWi?^ However, that meaning is not its own ^frnsrftfrr. In this sense it is described as fe^sw-aRi? 'convening specific meaning*. However, in the constituent analysis 8T%3)];<T: ®^K ^ i e WOI*d ^Hr conveys it? own meaning! 1 totality \

• 1. *r5TT ^ sns? etc. Pat/s final answer to the problem of fbri and STO^T is to regard ihe name s^PTT^as a significant term. The rule H#ftPf defines ^cbwR^ which is in itself a significant term. The term -H^iUH, thus signifies two meanings simultaneously : ( 1 ) a name for all, aod ( 2 ) n f etc. Putting both the meanings together we have to interpret the rule ^Rf^tfil etc. as follows: ^ f etc. are called ^"TTR^S * the names for all f . The term snfai*^ is described as s??cPTO5T? * significant name \ Pat.' explains why this particular term has to be regarded as i R ^ H f j . The basic purpose of a 333T] is to signify an object in the briefest possible manner. Therefore, a 3?^ is always supposed to be a short expression. la fact,' Panini, who pays much attention to the principle o( brevity makes use of short sn?Tis such as fe, n, 3? etc. Why does he then choose a lengthy term S^bfW^simply to signify a group of words?. There must be some purpose in Panini's mind when he prefers a lengthy terra, even disregarding the principle of brevity. That purpose is according to Pat. this viz, that the name does not stand for the object signified by it, but also for its literal meaning, ST^TO^ being thus an sry^rsNTf will signify only those which are names for all. Since snrif^s used as proper names or ^ caonot be names for all, ^ r ^ u i ^ n does not apply to them. 2. ^rs^sq^rV s^faHoitafar- Cf. the Nirukta (1.2):

3. nw srrg^" etc. The author supplies a few illustrations to show that sometimes a Hlff is not merely ^ * conventional' but ^ V J T ^ * etymological and.conventional f . The name §TOTis given to the god mimm because he is ^m * black '. Similarly, the mythical hero sn|q[ is given the name ;g^fwn| because he is gf^?sm| ' possessor of a thousand arms'. The nameH^RTO" is also likewise a ^ito^a name given to the words H etc, when they are names for alL W



1. sr^ xrter%. The author introduces the last attempt of the objector to attack this final solution proposed by Pat. and finally refutes it. The

Ilahabhasya-Dipika, Mnifci VI

{17. !3

objector admits that according to this solution of operations caused by &&n*r^ such as ^#tff^r: *% ( P . 7 . 1 . 1 4 ) do not apply lo ^rsrr and ^ e ^ T , simply because they are not names for all. IJpf operations applicable to WFcHTQls such as ^ T ^ R O T ( P . 7. 2. 102), m%$ wimikmi... { P. 7. 1. 25) etc. do apply to ^ etc. even in the domain of mm and 3TO&T, because the fact whether they are names for all or not does aof matter in the case of such operations. In other words, 3T?cl*W$l4s arc not conditioned by *HNra?3 * the property pf being a name for all \ Therefore, they can apply even to ^j and 2. q$5rcaffe&. While introducing the refutation of this objection the author first reminds the objector of the earlier solution in which a provision was made to avoid the application of ss^ifoi^rsi's to *H***H» It consists in the new interpretation of the rule ^qH^Tf^ (See 3.10-13 above). However, this solution concerns only ^qgH«T. Therefore, the author introduces the answer given by Pat.', The word %%^^i * th$ honourable one • most probably refers to the author's teacher. If is occasionally used while referring to a different view. According to the view of $5«T37cn, the name g<fam^ is given to H # etc. because they are qams? of all. The S T ^ ^ ^ T is thus caused by the 3T?g«T*f%^. f ^ l is so c^.llfdf because he has dark complexion. In the same manner *r# etc. are called ^ ^ I H ^ s , because they already are names for all. The property of SfflfJBff thus already being existent in ^ etc. is only highlighted by the designation ^rihnii^. All this means that every word enumerated in the sraH^C list is by its very nature a name of all. Since a word signifying a H5JT o f s w J f a I? not enumerated in the ^rf% list, the question of the application of the does not arise. This reply is similar to the one already given by Pat., (See 3. 6 ) . In fact, the same old argument is steted here in a different manner. In this way, finally the author establishes the last solution, namely, «firgs}^T3vF*?» suggested by P a t which implies refutation of the VUr* teaching prohibition of the name sarvanaman in the case of q%\ and




Summary : Kat. points out in the two Var-s ( 8 and 9 ) that the words ^^r and 3"¥R are in complementary distribution. The word grvr is used in dual as well as in feminine whereas elsewhere the word ??m is used. la its.

ft. 6)




restricted domain of usage the forms of •&* do not evince any morphological changes peculiar to ^trn??^. The only advantage of its inclusion in the sraftf^ IiSt is, according to Kit., the addition of the suffix ar^ % However, in the last Vat. (No. 10 ) it is suggested that the addition of either eg or $&%% makes no difference. Thus ultimately the inclusion of w in ^ f ? list serves no purpose. Pat, explains the meaning of the three Far.s and agrees with the

The author of the Dip, refers to the views of other grammarians also regarding the use of :y*r and wm» He clearly points out the difference between the two words on semantic ground. Finally, he shows at least) one purpose for which the inclusion of ;j¥r in H^fi? list Is accessary.

1. *?$% $Rtenr^ etc. According to Bhar., Pat. understands the VSr. in a somewhat different way from the one in which it was intended by Kat. Bhar, however, does not give any reasons why he thinks so. He suggests that according to Kat. a r ^ is the only purpose of the inclusion of ^ in the *rotf^ list, while Pat. introduces it simply as one of the purposes. This reading too much between the lines is an occasional feature of the style of the Dip. 2. srar ^ 3 ^ ^ f r ^r^r:. The example s*fr ^ f r ^ # r SRS: • two hands are facing each other • shows that the word 3~sj conveys not only two things but also their separateness, distinctness from each other. 3. ^ q$j zifi^qq: etc. This line is difficult to interpret. 4. % J^nsr^Ri:. The word OT^ in the Var. is understood by the author as standing for aiq^R, other suffixes similar to OT^. Such siffixes are the *3T$fe suffixes taught in the section P. 5. 3.1 onwards. These Suffixes do not convey any additional meaning other than that of the stem \Wt\L Although ?\\ is not strictly a ^ C T ^ suffix, it also does not convey additional meaning other than the meaning of the stem.

Both tjvr and z^tn are read in the 3fttf^ list. According to P. 5. 2.44t ^ t | ^ t ftc*m> the word znw is derived from ;J*T by adding thfc suffix m?r% in the sense * composed of two parts \ Although these two






words differ in their meaning the tradition beginning with Kat. describes them :: being in complementary distribution. For instance, the Yar* 8 : OTWTrer %^^S?^sRccfI3C and the Far. 9 : OTrarswisr suggest that gr*T is used only in dual and before femiaine suffix while the word ^r*nT is used elsewhere. These facts about the usage of gir and ^vm are explained differently by different grammarians. Here the author sums up the statements of various grammarians as follows: ( 1 ) 31? is replaced by 3FW? when it is followed by an 3ri?T? or a suffix other than a dual ending or JT^. (,2) When a dual ending or zi\ is added after 31m the suffix m\ is dropped. ( 3 ) 3"?n? is used except before, a dual ending and z\\. ( 4 ) gn? isreplaced by s w r when it is not followed by a dual ending, All these views state more or less the same facts. While TTTT^ is an' additional factor conditioning the use of 3 1 ^ according to the first view," the factor zt\ which is found in all the views is absent in the last view. The second view teaching the dropping of the suffix ^^\ is recorded also in the VP.



In order to illustrate the complementary distribution of 317 and &*TQt. Pat. gives examples such as 3iq% %3?njs3T: 'consisting of [both, the gods and men ' and zsmt Sffsn ' a two-fold gem \ In these expressions a whole composed of parts is denoted. It has been already pointed out that the word 31T stands for two separate objects. Separateness or difference is therefore the meaning conveyed by 31?. In the above two expressions the difference disappears. Therefore, there is no scope for the use of .^r. Consequently, the other word which is nearer to 3"*?, namely zwi, is used. This is how the author shows the reasoning behind the usage of these two words.

1. It was already stated in 4. 1 that both the words, 3"*r and %,r convey ^ . Here it is argued that just as ^ cannot be accepted in. <j%;, since its meaning, namely, ^ , is abandoned (rather, it becomes subordinate}** in. the same manner the word f^ also should not be accepted in fftr for-; the same reason. In other words just as we have usages like 3Uq i^cjErgeqi:, why cannot we have usages like # ^ q t Trfq:, *3 jg etc. ? Similarly in compounds like fgnpr: the word % should not exist, its meaning, namely %^, being given up. The real answer to this question regarding the difference of usage between 3"*T and % would be : srs^rcrra i ^ a i ^ 1 This is the nature of the word itself. However, the author tries t o account for this inconsistency on semantic basis.



. •;.



-2. mf* «*%*?t£t ^ : etc. Although'$3 i s abandoned in ff% ^uch as • %3W,, the spffwrftrrf * cause of ihe use ' of the word l£ ^ u ~-i a --:. u.:ieu. Perhaps according to the author ^mj ' number J is the ST^KTI^T! or %. As long as the srif^f^^rf exists the word most be used. The word <nr has no other than ^ itself. Therefore, in ffer, after the % has disappeared gr*r automatically disappears. ; 3. ^ f%%% etc. The author has already ( 4. 1 ) pointed out the" difference between ^ and % in the followiag words : grarais^1 ff snsrafe**' v n w . * ! * : •. %^T# ^ T R R V f : i -The author means that the word % is used when ^ may be mainly or subordinated conveyed. However, ZW can be used to convey; $n* only as a main meaning. As soon as ^ is. subordinate to some other meaning ^^r ceases to be there. ^ Implies parts( « r w k ) and it is subordinate to the whole (w^Rfar^). Therefore, w? used in a reference to «RRf^. Oo the other hand, % conveys parts even when they become whole. In fact, according to the author, in the case of % the relation of identity exists between the-, parts and the whole in ff%. For instance, saying * these two ' is equivalent to--saying 'this group of two*. The Sanskrit equivalent for the latter expression would be 3^33;. However, it is not used because of this identity, says the author.



J. It was stated earlier that 3^r is replaced by 31?*? except before dual endings and feminine suffix srqr. Here the author justifies the complex mentary position of ^w and ^vw with reference to each other. 2. f^srioi^etc.. p - 5 - z 4 4 : ^ ^ ^ fiten* teaches that m%-issubstituted in place of cPHL added after ^T in the sense 'This is its part* and this substitute Is obligatory and accented. This is how g-^R is derived^ from ^vr. The word ftc^r 'obligatory ' in the rule indicates that according to Panini 3 ^ must always replace 31? whenever the whole (anrofir^) is to be denoted.

3. The reading mtffo in ^ over the original srirftftr.

^ ast

l i n e IS a

suggested improvement
• •

1. ^ r s ^ ^ T ^ e t c . ' Here the author sums up the grounds for the use of the word T*f before dual endings and before the feminine- suffix ZH,P The three reasons enumerated here are : ( i ) STS^^r^T.*, because of the 1 nature of- the word (see 4. 4 above), and (ii) v^r^, because of tke


MahSbliBwa-Dlpika, Ahnika VI ^

statement to this effect by Kat. ( Far. 8 %<p?^f^W*f?*t)» a n d («» ) because of the logic behind it as shown by the author ( set; 4. 4 ) ,

2, sp: !i^j%^ etc. Var, 1: ^TT^I 3r^;nHc%S3"f& has pointed out that tine only purpose served by the inclusion of 31? in the <FF3rtfS[ list Is the addition of the affix a p ^ after 3"*T by P. 5. 3. 7 1 : areraHtarcrPCT^ *JH5 ^ : } s * ^ is an smT^ of the m$m suffix ^ taught in various meanings in different rules under the srfg^rc ' snfit^T^F: * ( P . 5. 3.70). If ¥*? is not listed in the l, the general suffix $ will be added after it. 3S e$&5Hf ^Rv€tf^%w: etc. Although the suffixes % and w ^ d i f f e f In their phonetic form and are added at different places (^5 is added aftdr thi stem, whole wtw^, is added before the last vowel of the stem), the same forfit OTfgp is derived by means of either. There is thus no formal difference between the two saffixes. Nor is there any difference of accentuation^ While $ is accented by P. 3. I. 3,$$^ is accented oh the final by P. 6. 1. 163. Therefore, Pat. asks the question, " what is the difference between o» g 4. « r ^ ^ % ^ i ^ : I The suffixes s and m^%differ from each other only & in one respect, namely, sf^sj^. In the q^qis of a #f|tlT text q^s are isolated. Sometimes elements within a tr^ are isolated by means of aflrasrf ' the sign s *• The author here considers the possibility of writing the word w^tx in the q^« qrs. In the q^qTS, W*V$h will be mentioned as s*rs3>t, if the suffix ^ is added. On the contrary, it will remain as it is when afeg^is added. Since a r ^ i s insert* ed within ww, ^r^f cannot be isolated from sr$. This difference between ^ and B ^ ^ i s , however, not to be taken into consideration according to the author, because Pat. says, ^ ^ ^ 5 R q^JRT 3Tg^?qt: * The authors of the q^crre are iiot to be followed by ^ ( ^ R ) * (Mbh. on Var. 2 on P. 3. 1. 109). Pat. means that while formulating rules of grammar a grammarian need not scr formulate his rules as to conform to the practice of the q^rre. tiri rife* other hand the author of the q^qre has to consult grammar while isolating tf^s from the 0%3f.

1. The sentence ^ m?$% ^ n ^ - J c f e m ^ is read twice in the Ms. In our text we have retained its occurrence after jrftqsr:, while the earlier occurrence after 2. ^ m ^ X ^ l ^ ^ t c . The author here introduces the difference between i$ and m^%* When 3Praf^is added after ;J*T, it is inserted before the final «f of 3 ^ (^^+«T5p-f-«r) as it is laid down $v$\: l immediately before the part beginning with the last vowel \ According to the Part.

{ No. 90) the suffix mw% which is inserted in the stem becomes om with the stem. As a result, ^ s g is regarded as the same as ^vf. When the cjuaf ending is added after ^?&> (?mm + $fr) w e ^ n stl]l s a v t b a t * * i s Immediately followed by a dual ending. The basic condition for the survival of the form ^ 7 is the existence of an immediately following dual ending of tfee feminine zi\. This condition is fulfilled even after the addition of M®% and the consequent change in the form of 3 ^ as ^?m> due to tbs reasons stated above : The addition of $F$\ thus causes no intervention between ^n? and the dual ending, the fact which provides sufficient ground for ^ r to exist.


1. |) 5p: ^fir etc. While ^rq»=^ does not cause intervention between W*f and the dual ending, ^ does, because it is not ^wr^rqfgrar * inserted within it \ When ^ is added after the q;» s ^ + sr?, the sequence ^ S + sfr-ff TT being cT%tflFcT gets the designation 5m%q%5. According to P. 2. 4. 71 : gift ^f3snf^lT%fcsfc the dual case-ending is dropped. To derive the dual form another dual ending is added after 3;. Thus we can speak of two dual case ladings in the sequence ^ r ( + afr) + 9 + ^ . However, both of them ff§ unable to provide ground for ^ r to exist. The former is dropped while ttw latter is intervened by ^ , 2. srHTgr^Tfrs^T etc. In the situation described above ^ r caaa#{ f, because it is np more followed by dual ending. Its use in the forna inus nas to be regarded as incorrect. Consequently the word must be used instead of HTT. 3. G% f^c^q^cH etc. In order to account for the form where the suffix ^ is added after 31?, two solutions are given. Pat. suggests that a special statement, viz. that gir is to be regarded as immediately follow* ed by dual ending when 3; is added after it, should be made. The author suggests another solution, namely, that it should be stated that the word svr followed by ^ which is in its turn followed by dual ending should be regarded as correct. 4. «riii%$r. «rfrr%3T * extended application • is a kind of statement in which the properties of one object are extended to another. In the present context, fgqNHMKdl 'the state of being immediately followed by dual ending*, which is a feature of the word ^TT, is extended to the word ^

[ 4.9 ]
1. It was argued in the previous passage that a special statement that T*T is followed by dual case-ending should be made when the


Mahabhasya-Dipitca, Ahnika



suffix m is added after \t. This need of a special statement is based on the assumption that T*T can exist only when followed by a dual case-ending, Pau now raises another objection on the basis of the same assumption. He says that just as a special state-meat i« needed in the case of ^ it is also needed when the feminine sufiix z\\ is added after g^r. For, at the stage ^ + £1^ + wh ^vr cannot survive since it is not immediately followed by the dual ending. However, the author of the Dip. does not approve of this argument-. ; He discusses it in this passage. 2. srjq 3T?%^T^ etc. The Var. 8 : contains reference to sn*^ (i. e. zi\) side by side with f§[3p3T. This fact clearly points out that T*1 can survive also when followed by zn^ Therefore, the argument that a special statement should b~ made in the case of z\\ is not acceptable. Occasionally the author is found to be differing from Pat. This is one such instance where he docs not hesitate rinding fault with him. :

It was said in the preceding passage that PaL's argument regarding Zj\ was inconsistent with Kat.'s statement. In th's passage the author cites two views justifying the argument. 1. crer enfofor etc. According to some scholars the statement trarff? f ^ ^ R q ^ T ^Es^iT is an objection against those grammarians who read : 3r*TF*t¥Rrs%^7r ' srr becomes S ^ R except when followed by dual ending*. Since zi\ is not mentioned along with %-^gvf by these grammarians, a special statement of 3rfe%^T, namely, that TVT should be understood as being followed by dual ending when it is in fact followed by zi% is required. Thus there is no question of inconsistency between the Mhh. and the Var. since Pat.'s objection is levelled against these othei grammarians. 2. ff^R?cT^T^:. The opinion of * the honourable one' represents another attempt to justify the above statement. It is difficult to identify the ^¥T3F3:. Does it mean that the teacher of the author of the Dip. is respectfully mentioned by him in plural? Whosoever the ihahhavantah may be, the argument runs as follows: Kat. specifically mentions zi^ in his Var., because by his earlier assumption tgi* is a correct word only when followed by dual ending. The mention of ^r^ in the Var. thus provides for the correctness of the usage of 3RT followed by zi% However, this provision is insufficient. It also must be stated that srvr is always declined in dual even when z\\ is added after it. In other words, the feminine suffix zi\ added after 31T is alwavs followed by dual, neither singular nor plural. This


m m


additional information about the feminine form of TO is supplied by Pat. la this way in the opinion of ffSR??r: the statement ffwifir %sr35fq?m n^m conveys a different meaning from what is usually understood and therefore, it is neither redundant nor inconsistent. Curiously enough, the author does not comment on either of these two views. Perhaps he admits the the latter view that Kat. has accounted for the correctness of :j*r before z\\ while Pat. has recommended the use of the feminine form of TO only in dual.

1. Here the author points out, following Pat., that a special statement of %gf^?q^i in the case of a?r?^ is not required. While Pat. explains this by resorting to 3Fcrw^)rra, the author of the Dip. explains it in quite a different manner. This difference is due to the difference in understanding the statement. Pat. argues as follows: In the sequence TO + OTC+4r, 3?T is not immediately followed by the dual ending due to the intervention of m\* However, when zi^ is combined with the final sr of TO» the resultant single substitute in jhe place of the both is an. According to P. 1. 1.21 : mw®* ^%$%T^, a single substitute (for both) is treated as the final of the previous unit or as the initial of the following one. Consequently, here atf is treated as the final of TO- Therefore, in the stage TO + STI, "J*T can be described as %^^^q?. Read the Mbh. on the Var. 9 : 2. 7 %^?TO3f etc. The author of the Dip., on the other hand, dls* T cards the necessity of a special statement by interpreting the word %ww?rfic^l in the Var. 9 in a different manner. Here the word %^3W does not mean, according to him, a dual ending as defined by P. 1. 4. 102 : cn?^^^^%g^T* but si.riply duality. f ^ ^ f ^ W or % ^ ^ n [ therefore means" that which is able to convey duality. The word ^i^ conveys 3 ^ or duality as long as it does not become subordinate to another meaning. In a ff% such as STJTO or crflr^r its meaning is subordinate. Therefore, it cannot convey duality. In that case it does not survive. However, when it is followed by Zl\ its meaning is not subordinate to another meaning. Being a feminine suffix, zt\ conveys ^r«r ' own meaning ' of TO. Therefore, TO survives. 3. *r% q*nrtftr^ra*3; etc. The Mbh. on the Var. 2 on P. 2. I. ! defines ffe? as TOsrrftr-TH ff%: ' expressing the meaning of another word '. When the meaning of the word TO is a part of another meaning, e. g. when it conveys a whole (3re??filr^) rather than parts, then duality, which Is its own meaning, becomes subordinate to the main meaning i.e. whole, Therefore, is used instead of TO,




4, f^rftcrgTOrsrrcR;. In a f % , the TO?J is the main meaning while the mjn of a constituent becomes subordinate to it. However, when a i g r f ^ suffix is added, ^ $ itself becomes the main meaning. Thus the position of ^rrsr is reverted wheo ^ ^ is added after snT- Since its owff meaning is the main meaning, the use of ^w before Z[\ is regarded as correct. 5. zm$tii ^ p k ^ C F i ; . According to the view •the five, namely, ^ s j , ^ g , %^r, ^ ^ a n d ^ 5 5 are conveyed by a nominal base * feminity is the meaning of ^ itself. The feminine suffix Zl\ is therefore described as 6. spffl'sft w%: The usage of the word IM is correct not only before the feminine suffix zi\ but also before the suffix q», because just like £1%, 3£ also is ^np-N>. Therefore, when ^ is added tjsr does not abandon its own meaning i. e. $ ? . Consequently, it can survive. The far. 10 : sra Rferf^ ^ c f e f q - j ^ n states that if a special statement about is not required for zr^i it is also not required for ^ t Pat. makes the point clear by saying that just like z\\> 35 also is a ^r$N> suffix; therefore ^Tvr does not abandon its own meaning: The difference of interpretation between Pat. and our author is quite evident in this case.

1. The correctness of the usage of 3 ^ before the suffixes zi\ and 3; was established by showing that they are *crffif$s. However, the interpretation of the word %^3iTqf referring to its semantic rather than formal aspect as stated by the author in the above passage leads to another difficulty. In this passage he discusses one such problem. In the words ^ T l ^ n ^ ^2T*Rt: etc. the use of the word T^T cannot be accounted for since the dual endings ^ r ^ and arl^ are not ^npfe. They convey m$m and srfawJi respectively. Both of them thus refer to some ^ 5 which is outside the ^m$. The meaning of ^sr, namely %^, is thus subordinate to the m¥% meaning understood from the case-endings v?irq; and rft^. The existence of 3-*T in the forms mentioned above is thus at stake. 2. ^ arfq ^ ^Tfr$r etc. The author replies to this objection by pointing out that even ^|^» meaning is ^rp} since it is included in the q^sp, the group of five denoted by a nominal base (see 4. 11, note 5, above). 3. gRT^y 3TSRH ^ ^ r etc. If case-endings like ssira; and afr^ are treated as ^HPN», the suffixes such as 3 ^ , ^ which convey meanings of cases




{See P. 5. 3. 7 : q ^ ^ f ^ a n d P. 5. 3. 10 ; *ra*TO<5;; also will be treated as ^f?njsp. Consequently, the undesiraoie forms *&**: a.m ' .^tt will ho derived. For the word t3H does not abandon its meaning- i. e. ^ before these suffixes. The treatment of tae case-endings as ^m^ thu, leads to the problem of wrong derivation. 4. 3*sr 3qp4% ... etc. One may try to solve the above problem by arguing that the suffix ^ does not denote *gm, because the 3rf^^?ui-^R55 conveyed by it is outside. For instance, the compound srf^f^ ' in connection with woman' cannot be formed of mfy and ^ft unless it is assumed that the scf^^uj conveyed by the word srfw is separate from the meaning of ^ t . A compound can be formed of two words conveying two different meanings. Just as srfef^ar is outside the meaning of ^$t in arf^%f, the rn^Rm meaning denoted by ^ ^ is outside the meaning of ^ ? . Therefore the meaning of 31? being subordinate to that meaning, there is no ground for ^ T to exist before sr$. 5. ^ H % - - T h e above argument is answered in the following manner: If the 3*1^*0! conveyed by ^ ^ is understood as outside the meaning of &%, the q$ft35 meanings conveyed by case-endings such as * ^ > sft^also will be outside. As a result the usage of ^ in forms such assTTPSH*^, 3"spft: cannot be explained. The same law cannot therefore be applied to the suffixes <Hi» 3c5^etc. on the one hand and the case-endings on the other. 6. cre*rpH <*ftfT*: etc * Therefore the problem of the wrong derivation of *^?rer:, *3^3T etc. is solved by referring to the concept of %{&m. The forms ending in g ^ , fr%^etc. are defined as snsR ( P. 1. 1. 38). An affo^ is that which remains unchanged in all genders, case-endings and numbers. According to this definition of sraapj, the meaning conveyed by 3 % ^ etc. is not only srf^ra^ui, but meanings of all case-endings in general. Since the suffixes sr^k, etc. thus convey a meaning which is different from ^%\$$ &% has to abandon its own meaning and be subordinate to the meaning conveyed by sr^etc. Therefore, it is replaced by T*W and we get correct forms like w*m% etc.

1. It is concluded from the foregoing discussion that the use of the word^T is incorrect when, being subordinate to some other meaning, it loses its own meaning. At such a place the word ^T*J is used. Here the objector raises an interesting question. Assuming that the word 3r?re is derived; from iw, he asks: How can the word 31T exist in the word 3 ^ in which it has to sacrifice its own meaning in that of the whole ? In the derivational





sequence snr-f-ara^s ^ Is neither followed by f ^ ^ n o r by zmj> nor does It convey Its own meaning, namely *§%, which is subordinate to the meaning of the derivation, which is whole. Therefore the suffix vw% must be added after T*W rather than after sir. etc. The answer to the above objection is equally interesting. The author here says that the derivation 3PT«? is its first occurrence. If T*f is not allowed to stay before mn%, the word 3 ^ will not be derived. The complementary distribution between 31? and ^irq starts only after °^% is derived. Therefore, the rule which presupposes existence of T*W cannot be applicable before its existence. Once it receives its form the word 3T*rq replaces ^r( elsewhere.

I. %^ ^«T|f^5 etc. P. 5. 4.128 : fg^i^rrf^s*?^1 contains a reference to f f ^ i s r ? ^ 1 3 ! which consists of irregular compounds. The words a e *" enumerated in this list. 31T in the compound 3^tt3f% is to be regarded as correct, only because of its direct mention ( ftqTcR ) in the list. The reading of 3"¥RT3f% however suggests the optional use of 31T. 2. *^r?el" •••• While summing up the discussion on the inclusion of ^•*T in the ^ftf% list, the author gives his own view. Here he clearly points out that the mention of ^ in the H 3 ^ list is not meant for a * ^ . Although the author has not pointed out in his commentary, Pat. also has finally rejected this purpose of the mention of 3"*r in the sreW^ list. However, Pat has not made any explicit remark to that effect. The author of the Dtp. however has found out one purpose of the inclusion of ^ in the snrff^" list. P. 2. 3. 27 i T ^ teaches that when the word %rj is used to denote reason the instrumental or the genitive is added after a pronoun. Since 313* is a *f«Mm^ the constructions ^rf¥<ir %v5*?Tr^ and irnqf: %?sr>: are derived according to this rule.

Correction ! In the Text and the Trans, for 4. 15, read 5. 1. 1. In the last Var, ( N o . 1 1 ) : ^ ^ ^ s ^ ^ ^ i c ^ l f i r the Varttikakara explains the purpose of the inclusion of the word *nj^ in the H^f^ list. l a tnis Var., he has listed three purposes : a c ^ , # * , and sncST. Here the author discusses these purposes and Pat.'s comments thereupon. 2. ^j?arer ^resfer ^ ^ 1 ^ : etc. The author points out that the usual purpose served by ^ f ^ m ^ T , namely, the replacement of case-endings




by substitutes such as ^ etc. ( P. 7. 1.14 etc.), is out of question in the case of *rar£, because the substitution takes place only when the stem ends in the short vowel *% {the word ssttti4 after short sc' is continued in the rules teaching substitution). This fact leads us to the three purposes enumerated by Kit. Pat. gives three examples : *ra$T^, *R?m and S R T ^ of the three purposes respectively. Of these the form s r e ^ clearly shows the difference. If the suffix ^ is added, the form would be ^Rc^:. It would be vm*®^ if a r ^ is added. Since the word s ^ ^ i s desired, it is necessary to include sraq; in list. 3. ^ J t * *nfcnfl39roi:. By $m Kat. means »^hr. P. 1. 2.12 : s j ^ l ^ l ^ teaches that when a sarvanaman in the ^ r 1 % list is mentioned in the same case-ending along with some other word, the sarvanaman is retained for both, for instance, the q^TO °f %^T^* *rafs* * Devadatta and you ' will be vra?!}" * you two \ The author remarks in this connection that the q&sm here is not caused by s^Rm^but by the inclusion of s r ^ in the list. Although the author's remark is apparently correct, is not the list itself meant for 4. £3l5WJn3rqtetc* decoding to Pat. the enumeration in the Var* does not exhaust all the purposes. It just shows some purposes by way of example. There are some more purposes. As an illustration Pat. refers to the rule ^ n ^ r ^ p f t q i ^ ( See no. 2 in Note IV. 14 above ). According to that rule the construction VRCTT |g*n and xm$t %#n both meaning * because of you ' are derived. 5. wareTCFnft ^ ^ w n ^ - The author of the Dip, well acquainted with Pat.'s usual tendency of finding fault with Kat. pays him in the same coin. By the remark T^i^nrnn^ Pat. suggests that Kat's enumeration is incomplete. By using the same word, the author points out the same fault in Pat*s treatment. He states that H ^ n ^ ^ l ^ T ^ is not the only additional case. There is one more purpose served by the et^rm^^TT of v p ^ . P. 5. 1.124: groraTOitmTf^&r. ^ f a ^ teaches that the suffix <aj^ is added after a jjufrHH • qualifier' and the words STUM etc - *n t n e sense ' a state of that • or • his activity \ When *r^t is regarded as a guM^-^ the rule becomes applicable and the wrong form *vn^T is derived. However, as it has been already pointed out (Note 3. 3 ) the H't^rraH^I cancels the juT^^Fr^^rr due to its being m * later \ Thus the 3$3r*ftT5?r of ^f^^ serves the purpose of avoiding the wrong derivation.


Mahabhamja-Dipika, Ahnika VI vihluisa diksamase [bahuvrihau] / P. 1.1 • 28


in a gfsffft compound containing words expressing direction ( H*T-

**fl--':i-«i, ^ optional*.

1. Here author discusses the purpose of the word ff^ in the rule* In the absence of the word f^ the rule would read fgrOTF snnSr snpftff. Consequently there will be a conflict between this rule teaching optional ^fwmWT and the next rule ^ wpftfr p g ex w f prohibiting it in the domain *pft% pt f p hi hibiti Since the option includes prohibition, the rule teaching prohibition will be redundant. The mention of the word % resolves the conflict by restricting the scope of optional srfsnro^l to that « § * $ which consists of wordsdenoting directions (fe^rj. 2. m$ fRr^T etc. The whole remark of the author points out that Pat. has borrowed argument on this point from the ffofp and incorporated it in his Bhasya in such a manner that the homogeneity of his text is not at all disturbed. s?*r^rar literally meaning • shadow of the text' implies that even a slightest disturbance in the IQXI is not caused by the insertion of the argument of the ff^pr. It can also be looked upon as an indirect taunt to Pal for borrowing the whole discussion from his predecessor. ff^jsr was perhaps some old lost commentary on the Astadhyayi. 3. fcfP-H f| ftNs^ etc* If the word f^ is absent, the rule reads finmT im% vgfti*aad t h u s c o m e s d i r e c t l y i n c o n f l i c t w i t h t h e D e x t r u l e * ^x^^r. While the latter prohibits ^msinfT in *f£ft%, the former allows ifoptionally in the same context. Since option involves both the alternatives application and prohibition, prohibition of g*rw*ftT can be achieved by the statement of option. Therefore the rule * *pft$, teaching prohibition, becomes redundant. 4. IMta**; etc. An exposition of science aims at achieving benefit for mankind. If it contains contradictory statements (%qftOTC) the desired aim cannot be achieved. The contradictory statements, finTT^T W*m Vgftp (•frfrorckn m%) and ^ ( ^ f t ) qgAft ( ^ f T ^ ^ T ) which would appear in the Astadhyayi would go against the basic aim of teaching science. 5. 5CRJ5T *Ssqy etc. It is difficult to understand the implication of this sentence. Pat. uses it sometimes in the Mbh. (e. g. 2. 356.1-2 ) to warn the objector not to be overcriticai with Panini. Perhaps the author has used this remark of Pat. with the same purpose in his mind. He means that what purpose the word f^g in the rule serves is clear by reading the next rule; before reading it the objector is not supposed to cast aspersions on Paninu


!• The purpose of the word *T*n*T in the rule is questioned. answers this question by distinguishing between two kinds of f ( 1 ) W|#ff^rnT^ and ( 2 ) an expression treated as €iipff%. The word ^ % ^ f i p l F ? is given by Pat. as an example of the latter. P. 8. 1. 10 : BfFtl% ^ teaches that a word is used twice when the speaker Intends to convey harm or trouble and the word twice used is treated as g^cftf^ The expression ^ S m i ^ ^%oi^r is accordingly treated as sjfcftff, as a result of which the dropping of case-ending by P. 2. 4. 71 and ^rsritf by the Mbh, on P8 2. 2.28 Is effected and the form ^%ar*%a?^|' is derived, ^%OIT is a f^. Therefore, by the present rule (^f^rrp*? %3|Wt|l\ i. £• when the word CTHCT is absent) -optional H^^fTffW^f is applicable to ^%jri^%iireq. Because of the option, in the absence of ^"?rmH^'f the wrong form *^%*n[%J!Hr also is derived. la order to avoid this difficulty the word H?ftxT is mentioned in the rule. Owing to the mention of *T?irH, the option taught by the present rule applies only to the first kind of ^|pft% mentioned above, because the second is ^cft%, not . In fact, apparently there is no difference between wfwWf and However, if a distinction between the two is not mentioned* the use of the word *r*rw in the present rule will be redundant. The distinction can be mentioned by assuming the above two kinds of ^J#f%. Therefore, Pat. argues that because of the mention of snrnFT in the rule, it applies only to such expressions which can be regarded as STOTH and ^m® only. The expression ^%q^%oi# cannot be regarded as such, since it is srfsftftw^ and not wjsftff. Therefore the option ( o r prohibition resulting out of option ) does not apply there. Here the Ms. reads only ^ Instead of H*n*f» In fact ^ as the short form for ^RT^FT is used by ancient grammarians (e. g. ^T^n%, a *TiR|W on P . 5. 2. 128). 2. q^br *fHmg%3r etc. srerasrspJf here should be understood as ^Jr. It refers to ^ f t % ^ in P. 8. 1. 10. The author means that just as by ascribing the status of «rg^ff| to repetitive expression such as 3%&I3%G1FT grammatical operations like j p ^ r w , fpf^fa etc. (see the preceding note ) become applicable, in the same manner the optional H^bfnTHslT taught by this rule also would be applicable. In order to avoid its application the mention of the word STHTH is necessary. 3. q*n z&fe q: 3Kii In order to support Pat's argument based on the assumption that repetition of a word is meant for emphasis, the author gives another illustration which is already discussed by Pat. P. 1. 3.78 :

Mahabhawa-Dipika, Ihnika VI


contains an apparently redundant mention of the word $nft, since the word sj^rft Is already continued from P. 1. 3. 14. The repeated mention is meant to emphasize that q ^ t f ^ is effected when g % and only «B$ is p conveyed. The qR?lrq? taught by P. 1.3. 79 : ^qryyqf W%: *s n o t effected in constructions such as 3r«f%q% tW^if^ * ( he) imitates himself since the % f ending here conveys no\ apjj only but $*fer£ ( P. 3. I. 87 ). The Mbh. on the

Var. 4 on ?. 1. 3. 78 : TO?^ ^ f o | % ^ ^ ^afrfo ? ^fc *: **rf a* m ^ ^rw *n SJ^ft I Kaiyata on this Mbh. remarks :

4. am ^r ??HIT? fr^ren; ^ 4 H The author explains the purpose of the word ^mrH »n a different way. The word %mm in the rule is, according to him, uttered with ^rffc accent. P. 1. 3. 11 : ^f^^nfiteR: states that a word with a ^f^?r accent should be understood as an STRT^JR. Accordingly, the word WRWt with ^f$% in the present rule indicates irf^T*. P. 2 . 1 . 3 : S an ^ ^ ^ containing the word ^ r w : - The mention of the word CTJTTO in the rule therefore implies that the rule applies to that which is taught under the s r f ^ R * ^nTRT-% ( P. 2. 1. 3 ). The rule does not apply to cases like ^%7r^%ai^%, because the ^gsftff understood there is not taught by any rule under the srarHTf^SR. The words smtSTtsfq cRT *n *J?J are rewritten In the Ms. after f ^ f ^ ^r4^ft They are, however, not required there.

1. It is shown that the word snTR? is necessary to avoid the application of the rule to the second kind of 3i|^Vff (3gaft%^)» The author now asks the basic question regarding the acceptance of the second kind as <9pft%. P. 8. 1.9 and a few more rules teach that certain words are treated as *f|pft% in certain situations. The suffix gp^ is added after a word ending in instrumental, locative or genitive indicating a standard of comparison (¥WFf) in connection with some action ( P. 5.1. 115 : fo <p4 %qy -k$fk: I P. 5. 1. 116: a* a*3ra). The word g | g f t f | ^ therefore means that the concerned word undergoes grammatical operations that are due to 3gpft%« However, it does not mean that it is called «r§cft%. The srfaf^n1 * extended application * conveyed by the suffix ^^ is always with reference to action. It is not meant to extend the name of one object to another. The second kind of ^||pftfi|> named as such does not, therefore, exist. Thus it canaot be argued that the word *WT?? is required to exclude it. The author gives two replies to this question as shown below. 2. *m ^ WllT^: e t c - Although the suffix ^ does not confer the of H^n (i. e. s q q ^ r ) , it is understood here. For instance, the very

21. 101


meaning of the suffix mt% namely ^fanjsr, is understood even without it.< Pat. has said, for example, that the sentence rrq- ww^* ' ^ is Brahmadatta* means s r s i ^ c r ^ ^ ^ « he is like Brahmadatta \ Here the meaning of ^C isunderstood in the absence of w^t- ( The Mbh. on the Var> 4 on P. I . I . 23 : i) Fn the same manner here srgcftf^^ can be understood as s*?q^5T 'extension of name' without any cause. Here what the author means is not quite clear. snj cnf^V «rq^nfeffi[ etc. Tradition speaks of five kinds of srania^r, ^m\ foftn*T,° °m\?iV, and ^ q j \ sqq^rf consists in extending the name ( 5^7 ) of one object to another. The word be regarded as sqq^nfir^sT without any reason. Both these replies are not much satisfactory. It could be simply stated that since cpf^f^^r is most important in grammar, the optional application of P. 7. 1. 15 which is caused by optional H^TTJTH^F by the present rule is avoided by the mention of HTTTH in the rule. However, it cannot be avoided without restricting the scope of 3T#3T*RT5Tr* The two kinds of 3rjr^r% mentioned above should be, therefore, understood in a loose sense. 3.

1. In this section Pat. discusses the purpose of the use of the word snpftfl in the present ruie. He suggests different purposes, discards them and finally accepts one purpose. They can be summed up as follows: ( 1 ) . To avoid the application of this rule to jg compound of f ^ words; ( 2) to continue the word snpftff in the subsequent rule; and ( 3) to extend the scope of this rule to that ^sftff which is a constituent of another compound. After rejecting the first two purposes, Pat. finally states the last one. Pat. deals with the first purpose at greater length. He resorts to various qfrsrreis to prove and disprove certain arguments. The author of the Dtp. further deals with each argument in an elaborate manner. In the present passage he introduces the first purpose. Just as the words denoting f ^ such as *£§•, ^%m etc. can be compounded into ^cftf%, they can also be compounded into {*§r. Now if the word ^|pft% is absent in the present rufe, it would read as finrNr %FST*TT^ and therefore would mean that in a ~* compound of f ^ words ^ ^ u r ^ f is optional. The scope of the rule thus ' being widened, it can be applicable to ^ r compounds such as ^f^rorlTW'i^ constituted of %r, words. CKving to the optional ^[#^T?T^rf the augment ; g ^ before the genitive plural ending srr^ taught by P. 7. 1, 5 2 :


Mahmasya-Dipika, Ahnika VI


becomes optionally applicable as a result of which two forms and *3f%9itTTr^TOT^ are derived. The latter however is not desired. II can be avoided only by prohibiting s?gh??iTOr5ff to these words ia f f compound. By the use of the word ^ir^tff the rule is restricted to and the application of sHbrm^T to g% is checked. etc. The question of •£% is further discussed and finally solved by resorting to ^oisrfaq^frrKqf^nsT (See 3. 7 ) . Before that certain other issues are raised on the basis of different assumptions. The author suggests the ad hoc character of these issues by making the above remark. The following lines of the text are not clear. From Kaiyata's commentary on this portion of the Mbh. the purport of these lines can be understood. The word 3f§rotrnni<3rfaTq[ is genitive plural of the f^ ^%sjTfTin?r. The constituents of the compound are ^%i?T, 3TR? and ^stf. According to the Mbh. statement on P. 2. 2. 28, rr^nT^r !%*?!% S^fW:, all the three feminine ewwr^s are reduced to their masculine counterparts. Although the compound ^ r o i f c l ^ i as such cannot be called *ptarm^, since is prohibited in £% compound by P. 1. 1. 29 : c %% ^ ' , of each constituent cannot be rejected. Therefore the augment g ^ before the case-ending STT^ following the sr^RTtft * ^ ' cannot be avoided. In order to avoid it the word ^|f^ff is mentioned in the rule so that its scope is restricted to

The first reply to the question, what is the purpose of the mention of the word sr|cftft ?, is cited here from the Mbh. However, the author neither discusses the purpose nor does he explain the example 3 % o f t i Taking a diversion he enters into the issue of 5^"*fl^r in this compound. He records two views about the ^ ^ V T R stated by the Mbh. ^HbfR^t ^ H m ^ gr^ilpn ( on P. 2. 2. 28 and P. 5. 3. 28 ). In one view it is regarded as an independent statement of 5^*nsr. Therefore, in the present case jcT^m^ is effected in connection with all the three constituents 3%STT, 3TRT and < ^ . In another view the statement of Pat. is to be understood in the section dealing with 5^T*?H. In that case the condition 3TT!T?\ being continued also in this statement, only that ^ q ^ which is immediately followed by ^tRq^, will undergo S ^ « T R . STTCT is such a i j b ^ , because in the popular concept tfsjr is the SrRq^. Therefore, 5^^¥TT^ will be effected only in connection with 3"rRT, not with Really speaking the form of the compound ^%oihTrifrt^T^ will remain the same, no matter which of the two views mentioned above is accepted, because be it ^%on or ^ % q , its # * with TTR will be the same, namelyf




Therefore, the author gives another illustration, where the absence of <pr**r3r is clearly visible. Anyhow* all this discussion about two views regarding *pr*THf carries no meaning in the context of the present example.

1. The discussion over the issue of gg^vrra *s carded further. A difficulty is raised in this passage. If the view that the *f^r*fr^ is effected in connection with all the constituents is accepted, it will contradict the statement of PSnini : q ^ % # £ f Sr^Wtf: ( P. 2. 4. 26 ), * the gender of the i£3[ and ^^W^ compound is according to the gender of the fioal constituent *s After ^ ^ S T I W the final constituent, <j^ft for instance in the present example, Is no more feminine. Therefore, the compound ^^Jftrf? 1 !^ cannot be feminine. It is however desired. This difficulty is solved by the author in a number of ways : ( 1) 3rrf3jTsrp-iFn^ etc. The question of gender arises only in a sentence i.e. constituent analysis. At that lime the gender of the final constituent will be extended to the compound. Thus there will be no conflict between *pr^*TR and 2. arsf 3T < p r ^ m 5l*<tii*q, etc. The process of pFrsnsr applies only to the phonetic form of a constituent while its meaning is Intact. In other words, jpr^scra" of ^ means that the feminine suffix an is removeds not the feminine meaning which is inherent in the word. Thus even after it becomes ^ it still conveys feminity. In order to express that feminity the suffix zi\ is addde after the compound % ^ I 3. ^C fljffiqiq; etc. Although the feminine.suffix z\\ added after <j^ is cll^d uwiii£ to flic j^vn^fi another sjqMs added after the whole compound IB consonance with P. 2. 4. 26. 4. zv&m%t % # m n e t c \ The other view about <pc^TH! is accepted and g ^ T R is affected in connection with *j#qR[ when T^VT% follows. Since it is not effected in connection with 4 i H ^ (e. g. vgft) the feminine ^n^ [ and therefore the gender of the TrRq^ ] is retained. One more zrt is added after the whole compound to convey its gender again in consonance with P. 2.4. 26. The worth of each one of these solutions is too evident to need any further explanation. The addition of one more 2r% cannot be semantically justified although it apparently gives scope to P. 2. 4, 26. 5. «Fcn%**TT^rai%f etc. After the addition of the feminine suffix zt\ after the whole compound, the compound is treated as z\^F^ * ending in z\\*


Mahahhasya-.Dipika, Ihnika



and also as a q-iwm^ The author unnecessarily tries to account for these too obvious features of the compound by referring to different q%nqrs. One of them is arrcF^vrrc. P- 6. I. 85 : sn^ff^a* teaches that one single substitute in place of two phonemes Is treated as the final part of the preceding element or ihe initial part of the following element. The £ f a of the feminine suffix Tl\ with the final sr ( or WT ) of the ^ ? q ^ * ^# ' ( or *£=$?) is an- It is treated 3S the final part of the word ^ # ( o r ^wf). Therefore, it can be described as ZHrrf. By another qflsrrqr, namely that of rR^fkfk which is stated by Kit., the whole compound (^%OIVTR^T ) is called y ^ m ^ . It is difficult to understand why the author indulges in these trifle issues when the of such compounds in undesirable. 6. «ra mfTRF? etc. The author is not satisfied with the above discussion regarding ihe gender etc. of the compounds such as s^roftrrrj^He raises another difficulty in adopting ^ ^ f ^ n r f o ? N T - When the feminine suffix zi\ added after ^ is retained according to the views regarding ^W^^TR, ^ \ can be described 2isj?<T, but the whole compound cannot be described as zm^. The Pari. ( Pari* No. 23 ) does not allow the whole compound to be described as z&r%' This difficulty is solved by referring to the amendment made In the above qft^nqr by insertng the word sr^sTc^Sfa ( see Notes, p. 132, lection 3. 11). All the discussion in this and the previous passage seems to be verbose, redundaat and out of place.

1. It was already stated that the rule suffers from the fault of overapplication. If the word «rfcftff is not mentioned there, it applies also in the domain of jjrj. Pat. introduces an objection that since the rule $fc ^ ( P. 1. 1. 31) already prohibits sHfomH^TT in £% compound there is no scope for this rule to teach optional H ^ P % ' In reply to this objection Pat. refers to the maxim, 5fa ;rmT§r 3R3" srrsR ^ f i r (see Nag. on the Pari. -No. 5 7 ) . The qftvnqT leaches that if a grammatical operation A is taught with reference to certain environment when another grammatical operation B is indispensably applicable in the same environment, A cancels B. The optional application of ^HRUTST^TT is taught by this rule in the environment of %FHi7ra while its prohibition is applicable in the same environment by P. 1. 1. 31. Therefore, this qr%nqT cancels the rule. In the same manner it also cancels P. 1. 1. 29 : q sjgsftfr which teaches prohibition of snbrraHsTT in «rpftf|. It is interesting lo note that finally Pat. discards this purpose *qf the use of the word i|sft% by referring to the same


2. ;?g ^ etc. Here the author introduces an objection against the very employment of this qf^TT*TT in the present situation. The <£#q%^ argues that the employment of the 5fa WTma^^T^I IS irrelevant in the present context, because neither P. 1. 1. 29 nor P. 1. 1. 31 is indispensably applicable in the same environment in which the present rule (1%wm f&®m%) applies. For instance in the *r|sftf| compound SrRi^rf, P. 1. 1.31 does not apply, whereas in the ^ r compound ^%oftrrr!^t ^* *• *• ^ ^ o e s n o * aPP^y« Since the application of either of the two rules cannot be described as indispensable, the qf**T[S(T has no scops to apply. This argument is obviously based on the wrong notion about the 5

1. The objection stated in the preceding passage is removed by explaining the qf**nqr in a different way. The author paraphrases the word ?nsiH as 31^18" ' absolutely covered or pervaded \ He further states that although P. 1. 1. 29 is not applicable in the domain of P. 1. i. 31 and vice versa, each one of them pervades its own domain. The present rule teaching .optional .^HRRHSTT also applies in the respective domains of both these rules. Therefore, it can cancel them. 2. <pr *T^. We have to understand this question in accordance •with the subsequent answer rather than with the preceding discussion. The 'question means, why does not this rule (finTNT f ^ n r i % ) cancel P. 1.1.31? (See the Mbh. ). The reply to this question is as follows : The rule already applies in the domain of ^Jcftff where prohibition is stated by P. 1. 1.29. After it has served its purpose of cancelling P. 1. I. 29 it applies in the domain of P. 1. 1.31. However, according to the % r f c ^ iffOTT^T ( P . 1. 4. 2 ) it is cancelled by P. 1. 1. 31 since it is a later rule.

[ 3.6 ]

' -

1. Pat. proposes another solution to the problem of ^ . It is clear •from the foregoing discussion that P. 1. 1.28 can be looked upon as an w^K o f b o t h P- l- l- 2 9 a n d p - L L 3 K (While P. 1. 1. 29 and 31 refer to 3|pftfi[ and ig in general, P. 1. 1. 28 refers to f^^rrn?T, which is a specific kind of both. In this sense P. 1. 1. 28 can be regarded ah STT^T^O There are two views regarding the relation between srq^rr^ and ^c^T when one 3TqgrTQ[ is related to more than one 3 r ^ . They are ( 1.) ^TOmT^f^ST * consideration of all Scfltfs together' and ( 2 ) «[«q&tf3&?3T * consideration of a specific &?wi \ According to the first view, when one wm\^ is related to many ^Hrfs, it cancels all of them. Ace-*


MahabhSqya-Dlpika, Mnika



ording to the second veiw, the mzm% cancels only one out of many Different Pari.-s are stated to point out which specific <rc?r*T is to be cancelled by wrcr? i° different situations. One of such Parirs cited by Pat. in the present context is : ^^rr^qw^T 3fW?cRR firvfH W W% iftU\ (No, 59). T ? Referring to the order of rules in the sutrapatha the qfr«TT*fT states that if an gpfcm is stated earlier to 3r*FTs, it cancels only that sm^ which is imme* diately subsequent to it, not the later ones. Here the g^tft P« 1- !• 29 is nearer to the srqwr^ P. 1. 1. 28. Therefore, the mm^ cancels P. 1. 1. 29, not P. 1. l. 31. Since this rule cannot thus cancel the prohibition taught by f% ^ , it cannot be stated that the purpose of the use of the word s^pftfir in the rule is to prevent the over-application to %%.

1. Pat. proposes another solution to this problem of over-application to f%. He refers to the ^arirfirq^qflwFTT a ad argues that since a *rfsft% type of % * w m is sfgrq^TtE whereas a g$ f&mw is STOflte, the rule applies to the former, not to the latter. Before applying the qftron1 to $£ and sr|cftf| P a t - applies it to two examples of q§3tft- The Paru ^ p i s t E m ^ t a * srfafqr^^^ 5?f^ (No. 105) teaches that when a rule is applicable to an element directly mentioned in a rule as well as to an element implied by a rule, the rule applies to the former, not to the latter. For instance, ignm is a gipH|- compound containing words expressing f^. It means * (a mad person ) who understands east as north ' and is derived by the general rule 3 ^ w ^ ( P. 2. 2. 24). There is another word < p f o containing fff^roif^, meaning * the space between the east and the north \ This latter also is a gfcftf! derived by the rule f^rmF3^TCT% ( p - 2. 2. 26). The rule f^TT^T f^RTOi^1 srpftfr is applicable to both these words. According to the ^^Toiaf^q^hEqf^n^T, however, it applies only to the latter which Is sfre^tE, since the rule 2.2. 26 directly mentions %fmmfa. The rule i p t f T B W ^ d ° e s n o t c o n t a i n a direct reference to words expressing f ^ Therefore the compound 'pfrR derived by this rule is ^TS^t^In the same manner a £% constituted by f^Tmrfir such as f^t is <STSrfoR5 whereas a ^fsftff constituted by the same f^ words, namely ^ r o i t ^ l ? , is srfifq^hE, because the former is taught by a general rule : *m £g: ( P. 2. 2. 29 ), while the latter is derived by the rule containing direct reference to ffo;words r f ^ n ^ ^ O T ^ ( P. 2. 2. 26 ). Therefore, in accordance with the qft^rmT, the present rule, I^Tiqfl $?CTnft *gftit applies only to ^f^i^» not to f%. The defect of over-application being thus removed, there is no necessity of mentioning the word *ijeftf^ in tira rule.




2. iff ^ ^mm^^mt^-" etc. The objector argues that the § compound taught by the rule f^srraT^OT^ eaonot be described as ^trfi. When a ^ord or a grammatical element is mentioned io its-'jorapute, accurately defined form in a rule it can be described as rrftrq^tTE. For instance, the constituents Mm etc. and p f etc. are directly mentioned in their accurate form in the rule ^m^: ftflf^ftr: ( P . 2. 1. 59). Since the constituents of a f ^ ^ r s r a are not mentioned in the compound it cannot be described as belonging to the category of s 3. s^q%W%q^Nvcn e t c - ^ s a reP^y t o the above objection the author describes the kind of srfoq^E^i assumed in the present case. He calls it ^q^srf^Pi^t^Kf?. ssrqlffir is name or designation. f^3F3TOT7T is a designation of such compounds as contain a £ j ^ word. The rule f^«TFF?TR?g^y% which mentions f ^ words contains therefore srfgq^taKfT. Thus although the constituents as such are not directly stated in the rule, a reference to the word f^ to which the compound owes its designation is found in the rule.

[ 3.8 ]
1. An objection is raised in this passage against the application of the 5^nmf§[q^^q^riq| to the example of 3g^fr%, namely, ^ f r l ? as shown before. The objector here points out the contradictory nature of the Mbh, itself. In his Bhasya on P . 6. 3.46 Pat, first raises an objection that the art substitution in the place of the final phoneme of q^ taught by this rule applies to compounds like TrfRFST which are gfaq^Hif; ( P . 2. 1. 61 ) and not to compounds like 5R^iirf which are $rT$rfn«5 ( P. 2, 2. 24). However, there also the substitute is desired. To this objection Pat. replies that even the l p f t f | taught by the general rule is to be treated as srfeq^tTfi on the basis of the repeated statement of OTtTq^ro. Any kind of 3fsftf| is thus, according to Par., srfoq^ftK ( m m % * ifcft ^fcftff:' ?% # r t * 4 & (Mbh. III. 160.24-25 ) | Kaiyata remarks: ^T 1). If all 3|pftff compounds are srf^rq^txK the earlier argument made by Pat. regarding the compound <£-TTTR cannot be justified on the basis of the ^ T O I ^ 2. ^T5Rn^T?cni^^rRT: ^ ^ ^ « The author replies to the above objectioa that Pat. described every ^fcft% as sfg^fcE in comparison with other types of compound. However, within the domain of 3|^t% itself the and srftrq^rrEc^ may differ from compound to compound. It isf


3Iakabham-Dipikat Ihnika VI

{22. [2

therefore., possible to describe one • g ^ f f

as sjfarq^tTfi and another as

3. *TOT SITO^KHT.... The author gives an illustration from section to show distinction within one and the same domain. While describe iog the purpose of the suffix TO in the word 3Tv73»tTO ^n ^ l c definition of ( P . 1. 4.42} Pat. s a y s : ? N r ^"^% l ( Mbh. I. 331. 11—13 ). He means that each and every is HIW^ 'productive in generating action*. Howsver, ^R^I is the most productive, therefore, it is described as HRT^OT. * a other words,; among the ^if^> there arc degrees of

Here the author points out that this solution by can be accepted only if the rule f^Trmif^... is accepted. However, K3t. rejects the rule in his Vai\ 1 on P. 2. 2. 28 which is cited by the author in this passage. The Var, f^f^?^} 1 ... etc. states that the rules f ^ ^ ( P. 2. 2. 26 ) and fo s%fe g^sM 1 ( P. 2. 2. 23) need not be stated, because the meanings sTr^t^" * space * and ST^H£ tnain ' to be understood from the respective compounds can be subsumed under w&yq^n and the general rule wStamwnr^r can account for these compounds. If this rejection is accepted, the solution which is based oa the rule f^FTOTf^t etc. is out of question.

L Pat.'s final answer to the question regarding the use of the word sgStfif 'a this rule is that it is meant for its continuation in the following rule. A strange answer indeed ! Is it not then possible to mention the word ^sfH? in the subsequent rule itself? Pat. answers that the subsequent rule *?! snpftll' already contains the word srfeftfr. However, it requires one more word 5jgcft|V. This can be achieved by sT*Jff% from this rule. The Mhh. further explains the purpose of the double mention of ^ f t f r in P. 1. 1. 29. It is meant for emphasis and for extention of the scope of the rule to all varieties of sqpfrff. The prohibition stated by P. 1. 1. 29 is thus, for instance, applicable also to that sripfVff which is a part of a larger compound. Pat,, gives the following illustration : tat u: ( - ^ ^ - ^ 9 0 - 24-26) i and 3r^RT«^R are two ^"^^% compounds. They seem to be synonyms, bcth meaning '(those) having cloth or clothes as their outer g a r m e n t \ The word 3 ? ^ is called HSRHFJ;, according to P. 1. 1.36 : %{??£ 3fip?mftrV:, since it means outer garment in these two words. By P. 1.1,29,



however, ^ T O H ^ J is denied to It since it is a part of the 7 f | TO*ns?r*r.' ( Nora. p|.) Is a £% of the two ggsTlff words. After a £ 5 is formed, the word ^ y ^ becomes part of a £§[. Therefore, it gets optional ^ T O m ^ n before nominative plural by P. I. 1. 32 : ft*n*T 3T% I Accordingly the nominative plural forms will be ^ F r ^ ^ R T ^ ? ] ; : and *gnsn?3T3rHTF^\ The latter one, however, is not desired. It can be avoided only by avoiding to sr?^r. The rule T ^sftfr denies ffWISRT^T in «npftf|. The T j of wgitfj from this rule ( 1. 1.23) into the rule * g p t f T ( I. *• 29) implies the rejection of HWWF^r? even for that ^g-^% which is a part of another compound. Therefore ^r^TTiT^HI is prohibited in the case of the word s r ? ^ which is basically a part of the ^=flrf| which is further a part of ^gr. The mention of the word ^rgsftff in this rule thus serves a purpose, according to Pat., by its continuation in the subsequent rule. 2. « r f | m f| ^fo etc. Tn this passage the author explains the function of the snfff^ of the word «rgcftfr in P. 1. 1. 29. At the stage WT^cR^^ffil^ •4- sn=r, two rules are applicable: ( i ) prohibition of H^TTTO^T in the case of W*w by if sigsft^T ( P. 1. 1. 29 ) and ( i i ) optional £P=NT*TH^-[ by finnan ^ T » since it is a part of jjgr. The former is 3T?cT^ with reference to the latter which requires an outside element, namely sr^ as a condition. According to the Pan. srfoi; *f|?fjpcrc# ( N o . 50) optional srtwm^rT ^ e J n g ^ % ^ cannot be cancelled by the 3T?cfir^ prohibition. However, on the basis of the sr«p[f% of 3^tffr from the previous rule into P. 1.1. 29, which would otherwise be redundant, it is understood that the prohibition cancels the optional ^ # srrasUn. Therefore, the desired form Wl?3^H*FcRT: is derived and the undesired form is avoided.

The author of the Dip. does not approve of this statement of the Mbh. regarding the purpose of the word ^jcftfT iQ this rule. According to him, ths optional ^bttsrc?3n taught by the rule f^r*nqi ^T% is applicable neither to WT'cR" nor to ^RT?3^ since they are not read in the * p ^ { list. If it is argued that it is obtained by ^r^^rf^f^r, the author replies that even c r ^ f e f a cannot help in his case, because the word srrcHf is ^qrerlffi in this compound and is not allowed to an Finally, in his resume to the whole section the author concludes that the word irgsftfr does not serve any purpose. Kaiyata seems to have taken cognizance of this objection, for he remarks i-^rg^JTSTfeft^1 grf%$T?s?? * *piWPf, * ^ft|Tf^*re«TTf^% ?r«R^r^ T If it is assumed that the prohibition of ^FTWHWT *n t n e c a s e


Mahabhasya-Dlpika, Mnika



T^R is not taught by Panini, this purpose of the mention of the word ^fsffrfr stands. Nagojibhatta further observes that this Mhh. is only an' elcadesyukti.

na bahiivrihau / P. 1. 1. 2 3
Summary : In the inventory of examples the objector rejects the first example, namely fmxk*$m> on the ground of availability of another synonymous alternative. The second example, namely •sx&im or sqrqre» is finally established by Pat. rejecting the arguments of the ^ q % ^ . Further, Pat« re-establishes the first example finrf^P?, and in order to justify the use of this prohibition within «n|5ftff, adduces a few more examples like r^crj%<p;: etc. The first Var, on this rule proposes an additional statement about the prohibition of ^HRR^rT in the compound ^jjrjjl. However, the second Var* rejects the proprosa! on the ground that ^ is not a ^HhlTO^' So the question of prohibition does not arise. The author of the Dip. mainly indulges in hair-splitting discussions on certain words of minor importance in the Mbh. On a few occasions, however, he adds a few bits of information showing an improvement over the Mbh, in his usual manner.

1. The present rule prohibiting sHbrFRH^T in Wfcft% needs, by way of example, a ^jsftff which ends in a ^ R T O ^ SO that the application of rules likely to apply to a case-ending following a ^ j n w ^ c a n be prohibited by means of it. Therefore, Pat. offers the example firaf^iTST, the compound ending in a jKisfqpffi. Because of the prohibition of ^T^ITO^TT by this rule, P. 7. 1. 14 does not apply and fStaf^rra instead of the undesirable *fjtofgpar*H is derived. In the Mbh. this example is rejected by the i4<J%^on the basis of the Var. on P. 2. 2. 35 : ggsftft ff4Hl^H^^^T^qr^ which teaches <$£rqrcl ' earlier placing ' of ^HNTR^ and &m\ ' numeral * in a ^Jsfrft1. According to the Var. a H I R I ^ must be placed first in a srf sftf| compound. Therefore, the form of the compound will be flpsfsRI rather than f i f 2. =T3r ^ F ^ T ^f^TT^- etc. The author goes a step further and anticipating the reply to this objection given by Pat. later in his Bhasya raises another objection. Var. 2 on P. 2. 2. 35, m fsfcnE** teaches optional i l f t r r a ' of the word fita in «r§cfti| compound. Accordingly, both

23. 5 ]



the forms, t^rfstq and fsrof^re will be correct. The objector now argues that since the other alternative fi^iftra is available, it will be used instead of ft«r%^. Therefore, the prohibition for the sake of fsjsrf^par is unnecessary.

1. After the first example is thus dispensed with, Pat. offers another example, rather a pair of examples, namely spqre and s^wn*. Both the ^pft% compounds sp?? and ssjwi end in the ? i k ^ 8 W?^f' and therefore are liable to grammatical operations doe to sHr«*rarchn since each one of them becomes a ST^RTSP^ by cr^erfirfif. The prohibition stated by this rule is required to avoid the undesirable application of rules caused by leading to a wrong form such as 2. ^ f ^ q T ^ s ^ i T ! ^ etc. The author first justifies the sptfsrw of the numerals % and fir in the compounds SF^I and sq?q, the fact which prepares the ground for the next objection. He tells us that since the numerals are qualifiers of the objects to be counted, they are always placed earlier according to P. 2. 2. 35 : KOTtf^tOT Wf itiu The q*firqm of the 3^=Rji?^ 5 ««j * being thus fully justified in both these compounds, the prohibition can be said to be required for them. 3. «i«3 ^ n f l r etc. Pat. mentions the objection in this connection that according to the Var, stated above, here also the ^wf^TRf of the word 3F3? is expected. The author points out an interesting fact about these two compounds, namely, that each one cf them contains both a wm and a *|4«IN-H> The Var. stated above teaches ^SjtfijpjFT of both w^\\ and ^Hb?m^« These two examples, each containing both y??^n and ^^{m\t therefore create a dilemma. The objector, taking advantage of this problematic situation proposes a solution that in the case of a conflict like this the ^ra?fT3R^ must claim Tj^fSfq^j, because its mention in the Var. is meant to suggest that wherever a conjunction of a w$m and a n ^ n H ^ occurs, the HSRra^must be placed earlier. The mention of sntHPFPi; in the Var. must suggest this, because there is no other rule teaching ^ f ^ q i s of a *r4'^'P4^ and if this Var, is not understood as suggesting the fact stated above, the mention of HWHTO^ will be redundant. In short, according to the objector, the ilfiiqrra of a S H ^ W ^ is established beyond doubt by the Var. as a result of which instead of sr?q and ?q?q, af?i?% and 3C?srf^ will be correct forais. Consequently, the prohibition is not required, ^

1. The objection is turned down by Pat. by referring to his own interpretation of the above Var. He himself remarks on the Var. as follows » '


Jlahabhdsya-Dipikai Ihnihi



In a w|pff% e x i s t i n g of both a STJS^TT and a sr^rTO^i what is placed earlier ? A wmi { is placed earlier)* because it is later 4 Accordingly, the forms and ^p^r are correct in his interpretation of the Var. etc. Pat. reestablishes the first example by invoking the Var. 2 : if? fsra^r (see I. 1.2) which sanctions both the forms tk^kn and firef^R*. The author of the Dtp, who has already referred to this Var. interpreting it in a way advantageous to the ^fqfaj^now states that the view o( the <^q»r based on this Var. is not correct. The Var* does not imply that when one of the alternatives is available the other one should not be used. Oe the contrary it implies that the speaker is free to use either of the two. Moreover, the Var* itself is an indication of the fact that the word faqfe*! also is available in usage.


i. The examples given so far were selected on the basis of the assumption that the purpose of the prohibition is perceptible only in that 3lpftf| which ends in a ^|#^m^. Now Pat. claims that the rule serves purpose even in that ^|pWf which contains a *?#HT??^ a s a l^Hft* The Mbh. : T T


1 f aff ? %

2. fife" 5Rt^r^ etc. The author here anticipates the question in the mind of the ^ q % ^ who is unable to understand how Pat. is going to explain the purpose of the prohibition even in an STH^fel?^ «rfsftff- Three grammatical operations due to the name $Hfcn;T«^are enumerated and it is shown that all the three of them are ineffective in a ^jpftfir with a H^Tra^as *j|q^. The ^4q%^ then argues that since none of these grammatical operations is effective, the prohibition which is meant to avoid their application is redundant. The three operations are as follows : ( I ) STTCQT. P. 6. 3. 91 : sn JC^HIW teaches substition of sn in place of the final of a H ^ F ^ before the elements v$, CT and ^ 5 . This rule is obviously not concerned with the «fg^t% containing a n W i T ^ a s a ^ q ^ , since the said elements do not occur as ^xim^ in a «n|cft%. ( 2) <ptaT. P. 2. 3. 27 : H ^ T I ^ T ^ ^ T ^ teaches the use of the instrumental after a ^ ^ T H ^ to denote cause. This rule also is inapplicable in the context of «lgcft%, bcecause no case-ending can follow a in a compound. (3) ^?iqi^q:. P. 7. 1.14 : H^TT**' ?&, P . 7.3. 1 1 4 : ^ ^ T ^ r - etc. teach different substitutes in the place of different case-endings added after ^T^HTir^* The question of the application of these rules does not at all arise, because the case-endings after the <£jq^ in a compound are dropped by P. 2. 4, 71.




3. s r ^ m ^ . After all the operations have proved ineffective as the author has shown above, Pat. has found out at least one operation which can be shown to be effective. P. 5. 3. 71 teaches the addition of the suffix *m^ after a pronoun. The suffix s r ^ i s an exception to ^ taught by P . 5. 3, 70. The example will be discussed later. 4. s q ^ R T ^ fe%q;. Although anu^ is an exception to ^ the difference between the two is observable only in the stems ending in a consonant. In the case of stems ending in a vowel the addition of either^ or aar^ results in the same form. The difference is therefore illusrated by Pat. with the help of the examples consisting of words ending in consonants : c^r^f^J^» and T^%~ fq<£3f>:. The constituent analysis of these compounds is, respectively, g^J^ ^. fq?2 ^ and arergr ^ fq<j ^ ( ^ fqm w H* ?n?%fW5> * He whose father you are', srf fqrcfT TFT *?: STF^ft"^: ' he whose father I a m ' ) . If sHtaTJTSnjI of gq^" and ar^T^is retained in ^gsft%, s r ^ w i l l be added before the final vowel of g«*T^ and s r ^ s ( the rule says, $n^> ^:) resulting in the forms g*?T3^[ and aan3l353*. By the application of other rules of substitution the wrong forms will be *c^5^fqtj^: and **RS^fq?£3$:. However, if the H ^ W ^ ( is rejected in the suffix 35 will be added after gspra; and sr^Tci; resulting in the forms and s T ^ r ^ Here also after the application of substitution of rules the forms f3r$f^P>: a n ( * T r U f a ^ will be derived. It is clear from the above pair of examples that the prohibition of ^g4tff is desired in order to arrive at correct forms by avoiding the addition of arq^. 5. aToq^F%c^fq etc. The author of the Dip. shows an improvement over Pat.'s remark by saying that the difference between ^ and sF^^is observable not only in the stems ending in a consonant but also in the stems ending in a vowel. He illustrates this point with the help of the example % ^ 5 ^ . In the compound f^sj^T (^f < T 3FT), ' one who possesses two sons ' the pT constituents of derivation are % SHT JpT ^ . If % is regarded as a ST^RTO^; 3 ^ ^ w i l l be added before the final vowel of % and consequently the form of the compound will be ^^f^^^. Oa the contrary if f|r is not regarded as S^TO3; in ^jfsftff compound, simply ^ will be added after % and the form will be %&3^T:. The difference between the forms gf^pr and %s<pr is caused by the addition of sp and sr^r^ In this way even in the case of a stem ending in a vowel ( % in the present context ) the difference is visible.

1. 3>5-I Jjfri^^arfifr-etc. An objection is raised here against the counter-examples ^ ^ ^ f q ^ and **r$f?qr£S by invoking the «r?cn (No. 50). The y$v[\q\ teaches that an s n ^ f m$ is more powerful than a % one and therefore is effected first. In the present case ${*$% is ^ ? ^ f with






reference to the prohibition of ^ehiJTH^fs since it requires only ^ ^ m ^ ^ f , ,viiert,i-s the prohibition is ^fff^ since n is dependent upon the compound. Therefore, mw\ must be effected first. Once it Is effected it cannot be cancelled by the prohibition. In this way c^jrjfqsjsp etc. cannot be counterexamples of the prohibition. ^ literally means the basis of ^ # % . The constituents and &m% are the basis of ^ p i ) f | in the above two compounds ? w ^ f t ^ 2 and *Tc$fr<i#' *n ord*r to emphasize the fact that &%\ is ^7^W with respect to the prohibition the author describes it in this manner. He means to say that s r ^ i s added to the very basis of wjltft which is a cause of the prohibition. The prohibition is therefore

1. The objection based on the ^cf^qf^Ti^i is discarded by Pat. by introducing another qf^rrqi which is exactly opposite to the The Part. s*?3?ffHfq $r$\\ ^ % # t f^firefe% teaches that sometimes a % operation is more powerful and cancels an STrcFff operation. Pat. illustrates this qfc*n<*! with the help of an example. The compound JTterfcST*?: c one to whom the possessor of cows is dear' is derived in the following way : P. 2. 2.24 . 4.71 Here the dropping of the case ending after the word ^Tr^ can be effected also by another rule, namely, f^qTS¥*iV. (P. 6. 1.68). The sftq taught by P. 6. 1. 68 is s^cRf with respect to the ^ taught by P. 2. 4. 71, because the former is dependent only upon the stem and the suffix while the latter requires the whole compound. According to the ^ H l ^ f t v n ^ , therefore, theformer deserves priority to apply at the stage *Ttaci; ^ fira *£. If it is applied it leads to the derivation of the wrong form ?rfaTF$to i n t n e following way. ifoq?H fePJ, 1. 6S. Even after the dropping of ^ by P. 6. h 68, grammatical operations caused by ^ take effect because of the rule ( P. 1. 1. 62 ). Therefore, - 7. 1.70 P . 6 . 4 . 14 P. 8. 2. 23

In order to avoid this undesirable form the $%?f-$pj has to be regarded as more powerful than the ^?cTTf-^tq. This is established by the



Park «Fcf?fFRf^ % f t ^ *%?ft fefiraNk Owing to this Pan. the 75^ (P. 2. 4. 71), although 1%:^ takes effect. Other grammatical -.^rations such as P. 6. 4. 48 do not take place after • 3*5, because P. 1. I. 63 prohibits ?. !. 1.62. 2. STr^w^q^mi etc. P. 7,2. 98 ; H ^ ^ ^ I W - S J is regarded as a * indicator' of the above qf^rftT* When a certain qff^rrqr is suggested by a word in a rule or by the whole rule, it is trditionally known as a f?iw?%lTWI. The tradition speaks of four stages leading to a sjjq^qf^rei s (1) sf?f| i t w N ^ ' first redundancy \ a certain part of a rule or the whole rule is found to be futile, (2) fEi«t3rrq?rn 'suggestion of the desired meaning' a statement, the acceptance of which justifies the rule is assumed. ( 3 ) &\$ ^rfi^rp:4n * fruitfulness in its own part % the statement is applied in the rule itself to justify the questionable part of it. ( 4 ) s r ? ^ q j ^ ' use elsewhere (in the Astadhydyi) \ The statement is made a general statment by showing its application to some other rules. A ^riqss srftsriqi which is hinted at the second stage is thus finally established as a general The rule 5Tc^Nr?q^Rt^ teaches that ^r and 1$ are substituted in the place of 3^r^ and $\IM<X respectively when a suffix or an ^r^qa[ follows. Apparently, the word ^rtrcq^ is redundant in this rule, because whenever ^rffq^ follows a suffix also follows 3 ^ and ST^RT. It is therefore sufficient to state * sr^Rfr ^ \ For instance, in the compound r*%R: the constituents are 5 ^ ^ ^ fel ^ . Here it can be said that ^ ^ T is replaced by rW^ because it is followed by the suffix \ . The substitution can take place also because it is followed by W3Vf% when % is dropped by P. 2. 4. 71. Of these two causes for the substitution, sre^nr is ^ f ? ^ ^ with reference to 3rRq^[. Therefore, by a n ^ f R f ^ W , substitution caused by the following VRW will take place. The mention of the word ^^VVI in the rule being thus redundant implies that the 5[%^ operation is more powerful than w^rw. If the ^flTF operation is regarded as more powerful, after HSTW^T is given to the constituents, P. 2. 4. 71 applies and after the elision of the case-ending, the substitution caused by w^V^, takes place. In this way the F a n . ( sf^cr^Hfa f^ft^sj%Cf m^ is implied by the mention of the word in the rule. In the case of the examples c^c33fap$: etc. the sameqf^iqt is applied. The 3T?cRf * 3F$=^' is cancelled by the 3%?f prohibition and the form is derived. 3. % q ^ ^ r *Jc*n. In the context of 3iTO»qf^rrqT grammarians speak of aOT=rF?3!^at(-3nq^and a fgr^qr^^T-^rq^. The former means an indication leading to a general qf^vnqrr, whereas the latter m^ans an indication leading

Mahabhasya-Dlpika, Ihnika VI


to a particular statement. The objector now argues that the rule q^JV«T can be looked upon as a festar^-griq^ leading to the restricted Pari.: ^ r ^ R K r fl"4t^ g % # t ^?i ^TOH, rather than a general statement ^cf?ifR% f%^5r%#rftRl"?rT^. Theqf^vfi^f being thus restricted to <g% does not apply to Pat.'s reply to this objection is that the rule ^ ^%^% i t s e ^ w*^ o e redundant if the qft*rr^T is not applied to s r ^ . Therefore, in order to justify the prohibition the *?T*n?2tT^r-^rTq3J should be accepted.

[ 1.7 ]
"STtr^q^r!^... The objector again points out that the prohibition already has another purpase to serve. In examples such as fsfaf^l the ^ l ^ f H ^ f is applicable to the word fq^fw*i by ^??ff^Rr. The prohibition also applies to the whole word. Both the H3Tf and its prohibition are thus fr*?RT?ri 'having the same substratum'. On the other hand, in examples such as f^^fq-q the H^TWHWT *s concerned with only f^sr whereas the prohibition ?r ^rgiftfT is concerned with the whole compound. The ^ T and the prohibition are thus not ^ R r s r e in this latter case. The prohibition applies to the former case ( i . e. ^JTHTST** ) and having thus fulfilled its purpose does not require any other examples. It is, therefore, not necessary to extend ths scope of the 3rfpfqf^Tj<?? to the prohibition in order to justify its mention. 2, s w ^ r f ^ ^ R etc. Pat. answers the objector that the purpose of the prohibition as pointed out by the objector is already achieved by some other means. In the compound f^jRto * One to whom alt are dear ' the word firu cannot get 3T#3TOH3?r, simply because it is :3w3?j ( See Var. 2 oo P. L I. 27). The prohibition^ «f|pftft has therefore no scope in such examples. It cannot be thus described as a rule which already has served its purpose. !.

1. *TSTO STS^IST ^ n \ etc. In the foregoing discussion Pat. tried to show that the prohibition is required to avoid the addition of sr=s^ after ^rorr, sn??^ etc. when they are former constituents of a «npftff. Now he himself proposes another solution to the problem of sr^v^. The author explains that this solution is based on the meaning of the word ^sft% itself. Thus according to Pat. 3f£ftf| means not only the compound so named, but also the constituents which form themselves into a 3^cft%. The *?#?n*r^r |s rejected in the case of fair not in the compound f^fita but even before the




compound is formed, i.e. in the constituent analysis: fqra ^ fef ^ . The 3pt3rra&^? being thus already prohibited to fk^} the question of the addition of s r ^ d o e s not arise in 2. 8i% m ^ m u ^ ^ i i ^ e t c - R u t k ° w c a n t h e worc ^ ^f%tfk m e a n t n e constituents which form a ^gitff? To this question P a t ' s answer i s : 31^'^rf^ a r ^ s ^ r ^ . A word also conveys that which is meant for the object literally denoted by the word. Here Pat. refers to one of the semantic features of words. Sometimes words convey, over and above their literal meaning, also some other meaning which is related to the literal meaning. This other meaning is implied or suggested according to later theorists who call it ^ p j p j . The relation between the ^^ptrTO * implied meaning' and 3T^3T%T * directly expressed meaning' is of various kinds. In the present case the relation is ^n^dr. Since the constituents are meant for wgsft%» they are related through the relation of m?M with the literal meaning of «r|py%, namely, a kind of compound. Because of this relation constituents also are understood from ths word ^Jsftff* Mbh.: ^rgsffirarift I^rf^r ^fsft%f*% i

1. &%4 fqrgrr etc. The solution to the problem of m^\ as given above by Pat. creates another problem. It is stated that since the word ^J=ft% stands also for the constituents of «r^sft% the prohibition applies to constituents. The sentence a * ^ fqar IF? (sr:) * He whose father I am • can also be described as the constituent analysis of the «r^=fVf| 'JTr^fi?^:'. Therefore, the prohibition will be effected and due to the absence of the form s r ^ q ; ending in sr^^will not be derived. The acceptance of the principle ai^wrrex ^rr^ssrn thus leads t o the defect of over-application of the prohibition where it is not desired. 2. %^f ^fa:. This difficulty is solved by pointing out the difference between two kinds of constituent analysis. Every compound has a 3?r%35fk%% and an 3T^r%$fTOf. ^rf^^5T^ consists of constituents which are found in usage. For instance, srf$ f w 3^? ^: is a ^n%sfejifcTR*?. On the other hand, constituents in the 3*wf%$firsrf are purely grammatical elements. For instance, sfTSRT 3T3S\^ fq*^ ^ is an ar^rfsfofro^rap? of the same compound mentioned above. The constituents of the 3?^l%3jfer! are never used in the language. The author tells us that the principle of crr^^T is adopted only in the case of ar^rf^f^iTf, because the ^rfe$%Tf cannot be described as meant only for 3 5 ^ % , it can also be used as a sentence.





Finally, Pat. cites the opinion of qfa^fq that the suffix ^ ^ \ and the accent caused by *ra«nwafr should be effected In the case of the examples mentioned above. In other words, prohibition does not apply to mw\ and the forms cWr^sf*?^ and ??r^fq?pj: as examples of prohibition are dismissed. The author of the Dip. now sums up the whele discussion on this rule. In all, three examples were supplied to justify the rule : ( 1 ) f^fiWfil and fihri^TFT ( 2 ) g p ^ and ?q&m and ( 3 ) ^r?s%p5- a " d STc^f^p;:- The first pair of examples was discarded by referring to 3WWr%%^" and the statement of ^ f ^ m r . The second pair also was first rejected by referring to the statement of ^ftrrra 1 but then it was accepted by giving a different interpretation of the statement about ^irf^qrc?. The third pair of examples was rejected by simply accepting the counter-examples as correct according to TT^Ffra. (Traditionally, this nfrrfra is either Pat. himself or Kat. However, there is no clearcut evidence to prove either.) Now, the author expresses his own view. First referring to the opinion of some grammarians he states that the rule of prohibition is required to justify compounds like WiscR^nTT^T: ( s e e discussion on the previous rule). Once this purpose is accepted it is not necessary to accept the unusual 3%?f<?^*Tr?T or the principle of ?rr^r to justify the rule. There* fore, all the examples except WT?3^HHT?rRT: are discarded. At the end, however, the author rejects even this example by referring to 3W5R5rft^r. He argues that since the word 3T?a* is OT^T in the compound WRcR^TPcRT:, it does not get the designation s p b r o ^ . The question of s#?n*rensrfir^r therefore does not arise. To sum up : The rule serves no purpose at all. The verbal form in the last line of this passage is not clear in the Ms, It is read as sqH% meaning 'rejects ( ? ) ' •

1. This section deals with the prohibition of H ^ T H ^ T to a nonsnpftff compound ending in ^#. In his first Var. Kat. proposes a special statement to this effect, as the Paninian rule is unable to cover it. However* in the second Var. he dismisses the earlier statement by pointing out that since ^fcfWSTsTT itself is not applicable to ^ in such compounds, there is no question of its prohibition. The two Far.s run as follows ;





Pa?, simply explains these two F a n s with an illustration with no comment of his own. 2. « n ^ n ^ ? %%. 8TO?n? meaning «fiOTt W&: *one w h o was formerly rich * is traditionally a ^l^l or %^B*rrcr. It is none of those compounds that are explicitly taught by Panini. (To justify compounds such as this one, the commentators of Panini split the rule ^rf igm ( P. 2. 1. 4 ) . For details, see Siddh. Kau. on P. 2 . 1 . 4 and P. 2.4.71 in the m$m section). In the sentence s n ^ T ^ f a %% * Give to one who was formerly rich' the dative singular an^T^fcr c a n ^ e derived ooly if HWifWH^T of *$ is prohibited. ^ is a ^risTO5!; according to P. 1. 1. 34 : ^ n p i By ^^cffirRr, snssrsjt becomes a Consequently, substitution rules like ^^so*^: w ( P. 7. L 14) also become applicable and the undesirable form *3rranjMr is derived. In order to avoid it, it is necessary to prohibit ^ptw?#*n for tjjf in a compound. As the word s n ^ i ^ is not a «rgcftf| compound, the present rule ( ^ Wg^til) cannot be helpful. Therefore, a special statement prohibiting tf^prah^ to ^ meaning vgs^ in a compound is recommended by Kat. 3. sgra^pnfiTcff^fr"* T n e ^rTO^WT 'additional statement' with regard to Kjrrj^ is rejected by arguing that the word f^i in such compounds cannot be a spsmm^. According to the rule '^qfi^f.., ( P. 1. 1. 34) ^ is a f W ; * ^ only when it conveys ^ ^ ^ n ' restriction of limit f either of time or space ) \ ( For the explanation of the meaning of ssre^n see notes on P. 1.1.34.) Since the word ^ in compounds like sfi^T1^" does not convey «ra^TT, it is not called sHfaTO^. This is the argument put forth in the second Var. and explained in the Mbh. on it. The author of the Dtp. carries the point a bit further and remarks that s q ^ n is, indeed, conveyed by fir also in compounds sech as STRq^f* (The word <$<t denotes the time prior to the time of the use of the word «TO?IiW. Hence ^m^TI is conveyed.) However, it is not the main meaning of the compound, because the word **<£ is a subordinate member. The word *£=} being ^PTH^T cannot get ^HbnHHWT although it conveys sq^^rr. Thus the author points out that ^ is not a H^TTff^, not because it does not convey SSR^STT, but because it is ^rrsHbr.

1. 3i?% ^T^f?cf etc. After giving his own interpretation of the Mbh*: S ^T3 ^^P-TI *T*^> the author now records two opinions of other gramT marians on the same remark of the Mbh. One of them is discussed in this






passage. Some grammarians agree with Pat. that oqsrt^n i s flot understood by ^ # in the compound ^ r ^ ? . However, they explain the meaning -of tzra^rr in a different way. According to them ezra^n is a fixed limit. It does not change. When the words <|<f etc. denote something which is not fixed, they cannot be said to convey «r^?n. For instance, in the expression ^%W! TP-ra>T: * expert singers' the word ^%oi denotes skill or expertise, which is not a permanent feature of the singers. Since it is liable to change, the word ^ % J I , which conveys it, cannot be said to have conveyed
S2r**P-Ti, Accordingly, it is not a ^?#JTW^. The sentence 3[%<J?T TW3>T: is, in

fact, cited as a counter-example for s q ^ r f in the commentaries on P. 1.1.34. In the same manner in the expression gfi^T^: ( or srf^TT ^ ^ ) t n e richness as a transitory feature of the person (denoted by the word s r r a t r ^ ) l s conveyed by the word ^#. Since the word ^-f thus does not convey something of a fixed nature, it does not mean o^pjn. Consequently, is not applicable to it. This concept of sq^arr as contrary to 3TTFT ' transitoriness • is rather strange. The author also rejects it in his comment on this view. etc. This interpretation of &H$*w does not stand according to the author. He argues that the <$#r^ does not undergo perishability,, Although the richness conveyed by the word sn^T is perishable the limit of time, namely, the past, conveyed by the word ^ is fixed. Therefore it cannot be argued that o q ^ n , meaning a fixed point, is not understood from the word i # in the compound 3 T 3. 5?ra3?n ^ ^ f R I * ^ etc. The rule ^q*ra*... etc. means that ^ etc. are called *?^Nwrs when they convey oq-^rr. Thus O^^^IT is connected with ^# etc. not with the other constituents like a T ^ - Therefore the R TT above argument referring to the transitoriness conveyed by sn^T is false.

1. ^ r ^ r s r j ^ i f # etc. The view of some other grammarians is that the word ^ in the compound ^ m ^ does not convey sqsrp-TT, because it gives up its own meaning. According to the view of ^r^P-JT ff%:, the constituents of an integration abandon their own meaning. ff% is q^isrffinsTHH * expression of other meaning'. The compound m^q^ also expresses a meaning other than the meaning of its constituents. The word i|# being thus meaningless is described as not conveying However, this view is questionable. If everywhere the constituents abandon their meaning, the forms q?prcri*lir, *?w$cf^ etc. cannot be derived

24. 19 ]


191 Anyhow, the

as the words ^r#, ^ ^ r etc. can no more be called H#*TTOR^ author of the Dip. is silent over this point.

trtiyasamase / P. 1.1. 30
Summary : This is a rule prohibiting the sHNrmH^n in By ?pfcn$j£rTH is here meant a <=Tc!p*7 with a word in instrumental as its former constituent. For instance, 7H31^ 1S a ^FffaWFWW formed according to P. 2. 1. 31 : ^TOSsiwfcnsr... etc. The word ^ is a H^ffiT*?^ according to P. 1. 1.34: %4q?T3T... etc. Therefore, by <?3[?crfqrfa, JTW^ will be a ^HcHm^ and by the wrong application of the substitution rule sr~:fcri«T- **T the undesirable form **TT$ri<l^" will be derived. The present rule avoids it by prohibiting sHtaroHsTT in Kat, does not take note of this rule. However, Pat. discussest just as in the case of the earlier rule, the purpose of the word *mi% in this rule. The word sr*n% can be continued from the earlier rule (ftrarff f^rarci 1 — )• Therefore, the mention of the word <gjTf% ia this rule is apparently redundant. The wording of the rule could be merely Pat. justifies the use of the word H^I^T by pointing out two purposes : ( 1) The word extends the scope of the prohibition to sentences containing an instrumental connected with a ^Hhrm^i and ( 2) It helps to split the rule into two. In fact, both these purposes are meant to achieve the same result viz. prohibition of ;a4*nus(3fl in a sentence also. They can therefore be rather described as two alternative interpretations of the rule made possible by the use of the word The author of the Dip. does not contribute much to the discussion already found in the Mbh. except misinterpreting Pat. at one place.

1. According to Pat. the mention of the word ^TTH in the rule is meant to extend the scope of the prohibition. The word <jRftqflFT*lTSr does not only mean a ^ 5 ^ T with instrumental, it also denotes the constituents of a r^ffaren?T*T» This additional meaning is obtained through the principle cTf^sifci; crrs^ST^ (See Note 1. 8, 2 on the previous rule ) on the basis of the mention of the word *n?T% which would otherwise be redundant. Thus the rule means that ^ etc. are not called SHRHT^ in r^ffaTsnrrsr as well as in a sentence containing a ^ w i ^ construed with a word in (j^qj. As a result


Mahabham-Dlpika, Mnilea VI

[24. 19

the word ^ ceases to be ftehnsr^in both, a compound and a sentence, and the desired forms like ^ ^ J T R etc. on the one hand and m%^ ^ p k era on the other are derived. This Is the achievement of the mention of the word in the rule. The author of the Dip., however, seems to misinterpret Pat. The whole passage containing the discussion summed up above is understood by the author as an assertive statement by Pat. that the word ^ T I ? is redundant We would like to draw the attention of the reader to the following passage In the Mbh. in order to judge the correctness or otherwise of its interpretation in the Dip. :


As commonsense would have it, the sentences beginning with af4 are to be interpreted as Pat.'s reply to the question stated in the previous sentence. (In this case a similar passage in the Mhh. on the previous rule beginning with «pi m*n$ 3]pft%. ••• etc. is comparable with this one )• However, it appears from the comment on this passage in the Dtp. that the author understands Pat. as trying to obtain both the results - prohibition in rfcifoTOTTH and in sentence - even without the word H*TT&. The author's argument runs as follows: The rule is read as tjcfl^T. The word snn% is continued from the preceding rule. Thus ultimately, the direct ( s p j ) plus the continuation (sTglirfii) result in the Su. ^cft^TH5?T^' The rule means that in a <JcffaTSrara as well as in a sentence leading to qRfon*T3RTCT ' Spl' etc. are not called SHFTTST^- Here also by employing the principle of m^ssfat gri^ssn^ the word tjatarcraiH is understood as the sentence meant for <£ftem?*mr. In this way ^WTTTH^T is prohibited in the compound as well as in the sentence by means of the continued word ^rrf% and the principle of The direct mention of H*TT% in the rule is thus absolutely useless. 2. 3^srR^ etc. The author further explains that in the previous rule the principle of <n^2?fe[ aress^n^ ne ^P e d to prohibit sr^rwstsTr in 3T^r%^fils^r « grammatical constituent analysis • ( e. g. f^yr ^ f s R ^ ) , whereas in the present rule the same principle helps to prohibit sHNnsr^TT in s?r%$f^nrf * semantic constituent analysis • ( e. g. *rrcfa ^ ^ r ) . Now the question arises: How can the principle of grf^^fej; g r r ^ g s ^ lead to ^r%^%Tf ? Just as •T^tfeP^Tf 1S nesessarily meant for a compound, ^ r f ^ ^ J T f is not necessarily meant for a compound. It can be used also as a sentence. The author's reply to this question is that cn^^T * the state of being meant for t h a t ' can be understood secondarily in the case of certain sentences which




evince some similarity with the ^ % ^ 5 l f of a compound. In other words, s^tftaftirareT also means a sentence containing a. sf^W 5 ^ .'^-^uuc*.: • :t : . ».* • • • sirumental which is similar to ^rf^f^JT^ of a compound. 3. i ^ % ^ t p f i ^ etc. The similarity with ^rfiFf$i%flf thus being the criterion, sentences like % ^ T R ^ f e f ^ d o not come under the jurisdiction of the second meaning of the word tjtftfWWH- For %^rW ^ f ^ cannot be a fe of a compound, since the two words are not mutually semanticaily connected (<|crf??^ is connected with

1. Pat. suggests an alternative explanation of the mention of the word ^111% to serve the same purpose. It is, according to him, meant to be a part of the rule so that the rule can be split up into two ; ( i ) <2<ifar and ( 2 ) 3T7?m%. The first rule (<jcfter) teaches prohibition of p ? ? ! ^ ^ ? In > while the second rule (3T*mi%) teaches it elsewhere, i. e. in a sentence. In the first rule the word *T*?T% is continued from the preceding rule whereas in the second rule the word <pffaT is continued. The two rules together prohibit *nhn?THtrr in both a compound and a sentence containing an instrumental. 2. ^ 3 ^j^TrsT ... The author raises an objection against this alternative explanation by pointing out the contradiction : While on the one hand the purpose of the mention of the word *r*TCH is questioned, on the other hand the author refers to the word s^TTO- How can the mention of lead to the word 3. <$3T5rft ST^f^ft etc. The reply to the objection given by tbe author is this : although the word HT S is heard, it is actually w^HTR? which, ? TT when combined with the earlier word r j ^ T loses the initial air because of the one single substitute STT in the place of both sn and a*. The contradiction is, therefore, only apparent. ^HifihTfTt ^ etc. After the inclusion of the word ?iwt^ ( or rather which facilitates the qfrrf^sTFT the extension of the prohibition to sentences like JTI^T ^ ? T is made possible. The second rule ^ r s u ^ is interpreted as teaching the extension. But how can the word ^ S T R ? mean 5T|f**f^iI^ ? arsreTST literally means non-compound and non-compound means anything that is not a compound, may be a word, a phrase or a sentence. The author therefore cites a qf^nqT which leads to the meaning of srsmra as ^ i % $ % r s . The Pari. ^ f t r e g ^ p ^ ^ f e R W a*H i m f ? n (No. 74) teaches that a word with a negative particle denotes that which is ciifferent from but similar to the object denoted by the negatived word. The 4.


Mahabhasya-Dtpika, Mnilca VI

[25; I

word w??J?R? therefore denotes that which is not a $mm but Is similar to it. Since ^rT%3p%rf is similar to ^ H it is understood from the word


# vibhasa jasi / P. 1 . 1 . 32


Summary: The present rule optionally allows spNfTOsf! f-° W$ ctc » in a fif compound when the nominative plural ending u ^ follows. It is thus an exception to. P. 1.1. 31 : £% = - which prohibits ^shn^^r? in a ^ q compound. According to this rule, for instance, at the stage 33733J? + ^ , P. 7. 1. 17 : SFH: ^ft is optionally applicable and the two forms and 33733ft are derived. There is, however, one more rule caused by WI?T^^T which also is optionally applicable owing to this exception viz. the addition of the suffix s r 3 ^ by P. 5.3.71 : « r « i W f a ! « i m ^ *n$ ^ : ! T n e optional application of this rule again results in two forms 33733ft and *3373TO^. The latter form is, however, not desired. Therefore Pat. explains the rule as follows: ^ H : 3 R S% f^rvTTOl 5Rf3% * ^?^% : ' T h e T option refers to the grammatical operation with regard to 3T^. ( The suffix ) of course is not ( added) \ There is no Var. on this rule. Pat. also had made just one observation. The author of the Dip, records different opinions of grammarians regarding the issue of

1. 3sTFfif% sr%f§%-.. If Pat.'s interpretation of the rule in accepted, there is no possibility of the addition of W3^. as it is not a mw 3FT^ 'grammatical operation with regard to 3T^\ When S P S ^ is not added, the general suffix 3 is added after 3?f73cT?T. At the stage $3?$3*T-f $ + src^ P. 7. i. 17 (m®: itt) does not apply because ^ is intervened by 3 from the H^ISR; ^af^. Thus only the form 33733TOT: will be derived. 2. qa^qH*!Sfa%% etc. The non-application of m^\ is obtained by P^t. by restricting the scope of H # W T H ^ I taught by the present rule to the operation caused by 3?^. The author now cites views of different grammarians regarding the obtainment of the desired result. According to one view, the rule f^n*cr ^ffn also allows optional application of ST3^ due to optional sHmraWT- Therefore, a special statement to prohibit the application of sins^ in this situation is needed. 3. 3fc53T& % ITcr etc. Some other grammarians make use of another device and interpret the rule as follows: H^1e *c. are optionally




called H#3fWl[io a %% compound when f is to be substituted in the place of 5R£. This interpretation automatically cancels the applicability of g r ^ . However, this interpretation raises a few questions: How is It achieved? The author only tells us that some split the word STIST into srsr f. Perhaps he means that mf^[ h a genitive compound meaning SRJ: f: '% substitution in place of s ? ^ \ Strictly speaking it is f and not f which is taught as a substitute. It is to be understood here on the basis of srradt (i. e. otherwise * < r r ^ ' will be meaningless ). Further, srfiqr itself is to be =$ regarded as a locative form by assuming dropping of the locative ending which is traditionally known as ^fe^fepftfisr * a mention without case-ending', srf*? thus stands for ^ r and finally, the rule means, H ^ etc. are optionally called ^f;rj?j^ in a fj compound when the \ (§•) substitution for cR^ is to be effected. This is no doubt a farfeched way to avoid nonapplication of The third and the last view is based on the Pari. (No. 3) which teaches that a rule of definition is present in every rule of application where the definition in question is used. In other words, a rule of definition is not independently interpreted. Accordingly, this rule will be construed with the rule srsr: $ft as the items 5R£ and $pfaro«^ are mentioned there. The two rules together would mean * sr^ which is added after a HI^TW^ which is optionally so called in a %% compound before sr^, is replaced by sft'. Therefore, naturally option will be applicable only in the grammatical operation concerned with 5H-£. is thus left unnoticed by the present rule. etc.

purvaparavaradaksinottaraparadharani vyavasthayam asainjndyam / P. 1 . 1 . 34

1. The rule occurs in the same wording also in the garuapatha. The rule in the ganapatha is meant to teach srcRm^^rr for ^ etc. in a specific meaning, whereas the rule in the sutrapatha allows it optionally before the nominative plural ending s r ^ . The rule in the sutrapatha is thus a 5TTg* an optional statement of what has been already applicable \ 2. sf^f 3fFR3$Rt H ^ T R ^ T^gg%- The author makes clear the position of the author of the Var. who rejects the mention of purva etc. in the above rule in his statement are^T^rar ^ JR1-


Mahabhawa-Blpifca, Ahmka VI

{ 25. 14

The Vakyakara means the VSrttikakaiu. Kat. does not mean that the whole rule is redundant, because he knows that it Is required to teach optional e^TO^WT before s?*r. He also knows that the rule la the gana* patha is meant to teach SHFTOH^TT for ^ etc. In the specific sense 533$^ etc. To put It in other words, his objection is against the repeated enumeration of the words r p etc. According to him Panini could simply state ^T^IT «ra instead of making the three rules P. 1. 1. 34-36, which are already found verbatim in the gayapatha.

Cf. Kaiyata: f W I $r%q? ^ t ^

^% firareHr qftercfir.

The word vrttikara Is perhaps a reference to the author of the lost pre-Kat. commentary on the Astadhyayl. (See also Dip. 1.1 on P. 1. 1. 28.)

1. sT^RlfFm?; etc. The words ^ f i ? and srenf^ have caused a lot of confusion throughout the Mbh. as well as the Dip. on this rule as it will be observed later on. While the list begins with *$ and hence can be correctly called I ^ T % the Varttikakara refers to it as srerfft- ^ n e commentators try to remove the inconsistency by simply describing the word snfi[ in sr^nr% as conveying kind. Thus Kaiyata remarks: Nagojibhatta says : The Uddyotakara thus feels that Kat. suggests by saying areTT^FF?^ instead of <£5iT^Rm that the first occurrence of the word sra? in the £% * ^ c r f i ^ . . . ' etc. would have been grammatically more correct. To a modern student of Panini, however, it appears that perhaps the list known to Kat. began with a=r^r rather than with <jcf. 2. The reading mmk ftf%I emended on the basis of the Pradipa. ^ JTr^firWRTT ^rf^i^fticr... ( See Kaiyata.) etc. is

1. sre fir^rJT^-T... etc. Kat.'s objection is based on the assumption that the gaiiapatha was composed earlier than the sutrapatha. Pat. asks the basic question regarding the assumption itself: How do you know that the ganapatha is earlier than the sutrapatha 1 The fault of jfr^r in Panini's repeated mention of ^ etc. as pointed out by Kat. can be removed if the chronological order of the sutrapatha and the ganapatha is decided. This point will be clear in trhe following discussion.

26.6 j




2, cf?FTf*rR etc. The author here introduces >\n interpretation of P a t ' s question mentioned above by some grammarians. The interpretation is mainly, based on the meaning of the word snf^. It has been already pointed out in the discussion on P. I. 1. 27 ( See Note on P 1. 1. 27) that the word $$\f\ conveys two different meanings when it denotes a garta : {1) ^?WFIT * fixed order ' and ( 2) J T ^ ' kind \ For instance, P. 2, i. 59 : famzpv ^cfif^ftr: contains reference to two ganas w^ffif and p n % Since the words in the 3nnnfi[ Trot occur in a fixed order and number, the word wff% here is said to convey sq^p-jl. On the other hand, tht number and order of words in the f^nf^-Tun is not fixed. Therefore, there the word m\T% conveys" SiqpH". It is assumed that in ganas like ficrrff, words are added by later grammarians after observing the usage. It can be, therefore, argued that if a certain gana mentioned in a rule contains the wordanf^ conveying s^rron, the ganapatha exists prior to the sutrapatha; if the word snf» conveys JT^T?»* the concerned ganapatha is later than the sutrapatha. The meaning of the'word m% is thus a clue to decide the relative chronology of the sutmpaiha and the ganapatha.

1. ^ ^ %3 *TI**HJ^3 etc. It is clear from the discussion in the, previous passage that if the word ${jf% in a gana conveys o^^sjf then only. the sutrapatha suffers from the fault of TTR^, because the ganapatha already exists. However, if the word STT1% is understood in the sense of sjcpi^ * kind % the question of 5jr^ does not arise. In such cases the enumeration in the. sutrapatha which is based on the usage of the learned, is later made precise by adding more examples, counter-examples etc. in the gayapatha. The repetition of words in this case is not to be questioned according to the author. He cites a line from the Mbh. to support his statment. Pat. says I 3 ^Tsfiwi^nfr ... etc. ( I . 12.9-10). What he means is this. The SutraT kara first makes a statement. Later he introduces some modifications in his earlier statement. He does not bother to return to the earlier statement and change or cancel it. He retains it as it is. After all, he is the Revered Teacher ! The word ^TT^r^rif^i perhaps refers to the ganasTdras which sup* posediy complete the earlier enumeration given in the sutrapatha. 2. CT«TI ^ f^r^TrBt^i^... The author gives a few illustrations of inconsistency evinced in the composition of the grammar of the Revered Teacher. The first illustration, namely, fgr*nEtai ^THfir^firirtfT: etc. is not quite clear; Perhaps it refers to the inconsistent use of case. Sometimes the Sutrakara takes care of changing the case of a word so as to suit the contxt. F o r

Mahabhdsya-bipika, Ahnika VI


instance, in P. !. 2. 45 the word srrtfNfe^ is used in the nom. sg. The gen. sg. form of the same word is required in P. !. 2. 47. The Sutrakara changes it accordingly and says : ^eft" ?f^r% anfirmf^'R ! On the other hand, he uses the word ^ N r o ^ n ^ in P. 7. I. 86 : f^rsc[ ^#?TW^I^", which is continued in P. 7. i. 93 : lifcit % r without bothering to change the form i Secondly, when the ganas are borrowed from other grammatical systems some of them are modified, while some are retained intact. For instance, the sfdras f?^??r: ( P . 1. 1. 39 ) and sr^€mrew ( P. 1. 1. 41 ) are a modified form of the ganasutra ^Mmi^^BTl^&^TXl^'S I The author suggests that in such cases one should not blame the Sutrakara for inconsistency.

1. This passage contains a repetition of the same argument. Here the author again makes clear the position of Kat. who rejects the repeated mention of ^# etc. in the sTitra. He says that since Kat. has understood the the word anf^ (in wzxfk, list) in the sense of 5q33OT* he has rejected the enumeration of ^# etc. in the three sutras ( P . i. 1. 34-35 ). 2. 'sjf^ g S c T T etc. Here again the author gives some more TC T O illustrations to show that the use of the word snf^ in the sense of $ZR? does mot lead to the fault of iff?^ The mention of a gana containing the word «r?fe conveying %?JRJ implies that the sutrapatha does not give an exhaustive list. The list is later modified by additions. The author then enlists a number of additional statements (3"q*t^H"s) found in the ganapathas in the passage, ^m^ £Flf§H . . etc.

1. ^ r . <pc:... etc. In the foregoing discussion the author explained that Kat. rejected the lists in the three sutras on the assumption that the word snf^ conveys «??3^n;. Here he further says that it has also been made clear that the word sni% in the word H^lf^ does convey s^gf^r. Now, if the word snf^ does not convey 5T^TT, why should an objection be raised by assuming that it conveys sr^i^ ? The reference is to Pat.'s question, ^ f yr=rfcn=f ••• etc. which is interpreted by some grammarians as being based on this two-fold meaning of s?ff% (see No. 2 in Note I. 3 ). If it is already admitted that iTff^ in the word H^rif^ conveys s^ra^T, Pat.'s question is unjustified. 2. 3"*n*2n; 3n^3f%:... The author replies the above question by pointing out that the usage shows that the word snf^ can be understood as conveying sm*. In the r/£f?^Tift q?Tre3i sfij^j^etc. ( RV[. 108. 10 ) as also In the statement *TTTIW •• l ^ t f t ( MS 1. 4. 12 ) the forms




% and araroq; imply that the words qro, J^W^ ^T?H and «nr(O are since they evince the ^ ^ s u b s t i t u t i o n and the suffix ct^swhich are caused by S ^ T S F ! . These words are not found in the ??^rrf^ list. In order to include them in the *?^rf^ list for the sake of *p=hfim?^T it is necessary to understand 3n% in the word *r=crf^ ^ s conveying sr^r*.

1. Kat.'s rejection of the enumeration is thus justified only if it is assumed that the ganapatha is earlier than the sutrapatha. But what is the proof to show that the ganapatha is earlier than the sutrapatha ? (Mbh* fo 'SH 3H:qre: ?) Pat.'s reply to this question is : cfj^r % i This sentence has proved a hard nut to crack. It can be literally translated as follows: * Because, those ( words) begin with *$% ( and ) these ( words ) begin with sr^r '• Does Pat. mean to say that the ganapatha is earlier, because it begins with «^ and the sutrapatha is later because it begins with sra? ? In other words, does he imply that the sutrapatha which begins with sra? is a partial reference (q^s^rfiftsr) to the already existing ganapatha which includes ^m but begins with ^ ? Although all the available texts of the sutrapatha - even the one included in the Mbh. - begin with ^ , Kat.'s reference to the sutrapatha as sren^tal^ in his Var. implies the possibility of some text of the sutrapatha beginning with Later commentators including our author understand Pat.'s reply differently as will be clear from the following remarks. 2. ^rlf^^V ^ n s j : etc. The author records different interpretations of Pat.'s statement by other grammarians. Before mentioning them he explains what is exactly meant by Pat. Later on in his Bhasya Pat. establishes the priority of the ganapatha on the basis of a ^rq^. P. 7. 1. 16 : ^rff^^ft Jf^v^t 3 ? teaches the substitution of ^rr^ and fer^in the place of the ablative and T the locative singular endings ( nasi and ni) optionally after the nine words ^ etc. This sTdra describes i^W^s as *ra\ ' nine' instead of actually listing them. In fact, ^rf^s mentioned in the stUrapatha ( P . I . 1.34-36) are nine. If P. 7. 1. 16 is understood as referring to the ^ f ^ s mentioned in the sutrapatha (i. e. P. 1. 1. 34-36 ) the word 7f^:q: is purposeless, because there are not more than nine words in the sutrapatha. The word IRVHX being thus purposeless, indicates the presence of a list containing more than nine words. Such a list is obviously the H^Tf^ list. The rule <£=fif^jV ?nn^: etc. thus refers to the nine words from the wi\f% list which begin with the word ^ . In this way the word ?&w. is a ^iq^ indicating the existence of the ganapatha earlier than the sutrapatha.



Ahniha VI


Here the author says that Pat.'s reply mf% I^T^tfir ®tc. assumes this In other words, Pat. means by this reply that the ganapatha is earlier, because the order of the words *p etc. is already fixed there, and the sutrapatha is later, because it refers only to one group of words, sr^t etc., from the ganapatha* 3. f a ^ c M w%n etc. Another disputant who is not aware of the adduced later by Pat. understands the Mbh. in a different way. By means of an ordinary illustration the author shows that the ' | ^ ; ' takes the words t^sn^tft and WciST^H^ to mean <j#?f^i% and sra^rerrfir. Of the two brothers one is described as elder, because he is born earlier ( ^ ) while the other one is called younger, because he is born later (afcR). * n * a e same manner, Pat. argues, the words ^# etc. mentioned in the ganapatha are earlier, because they are read eariier, whereas the sutrapatha containing the words s?^- etc. is later, because they are read later. The sentence f Jflfir si^Rlfrf?! would be thus paraphrased as ^ J J I ^ R ^ etc. Although apparently the words *jg^ftr and refer to two different lists, they denote, in fact, the same group of words. Thus the word <|# in the sutrapatha denotes *j# in the ganapatha v/hereas the word wm in s r a ^ t f t also refers to *$ in the sutrapatha. Now, if the words t p ^ f i t and 3RTr^H% are thus understood as referring to the two lists, how to understand the meaning ^$i<$ift and 3re?^T^lfiT from the same two words respectively ? The view of the ^ r ? cited in this passage is thus ambiguous in this respect. In order to remove the ambiguity we have to interpret ^ t ^ f r as <£f$raTfif ^ ^ r f i r and s r ^ T ^ T as ^T^^T^Tf^ ^ i ^ t f t - However, this is not a happy solution. 4.

1. %fei| •• It was stated earlier that according to the view of one disputant ( ^ R ) , by the word s r a t f ^ f a reference is made to the word ^ in the sutrapatha. This can be obtained by assuming the word sru^ in vnrcrf^t as conveying JHSTR. Here the author records an opinion of some other grammarians who think that the word ztm itself denotes the word <$jS\ To them aararcr^f?t means ^ f ^ t f t . It is difficult to understand how are* can stand foj ^#. The author tries to explain this with the help of an illustration. The whole life span of a living being is divided into various states (^TT^%qpR). (snw*?) H^T 'Being' (pure) is prior to all states. Therefore, it can be described as i^vrf^ with reference to all states. However, when one is born, then he exists. In other words the state of birth (SFIT^) is prior to the

27.23 ]

Notes 9. I i


state of being (f^tt^rtU) of that particular being {Nirukta


Thus the state mv$ preceds the state s r % . Thus the ( f ^ m ^ f T ) is nearer to the last state Trrer * destruction * than the first state sraq^ In this way the same state of sjrTT can be described as <£f*ug[ as well as mwWZ* (The siraFSgrfr is ijtsTT^ while the fsntfTOTTt is 3nf?s?ra). Tn the same manner the same word <j# is referred to by two words <j<f and srar in accordance with its earlier and later existence in the ganapatha and the sutrapatha respectively. Really speaking, the illustration given by the author is not very clear. If we assume two kinds of *T^?T viz. ^srpq^rTT an< 3 f^PTOrTf, the illustration can be explained as shown above. Anyhow, this is just one way to explain the text. The whole argument trying to prove that both the words ^ and S R ? refer to the same word i<# in two different lists is neither quite satisfactory nor is it required.

1. Here the author explains the meaning of the sentence <n& <£=n#fo etc * according to the grammarian fg?, already mentioned before in a more precise way. It appears that the words < < and mm are understood as $J adjectives of the word ^ . Thus ^ ^ t f a means * beginning with (the word ^ which is) earlier (mentioned in the gariapaiha)', whereas the word «R?l^rft means • beginning with (the word ^ which is ) later ( mentioned in the sutrapatha ) ' , 2. ^^i^JT^RTr^i^T^ra; etc. This line is not properly understood. Our translation represents a desparate attempt at interpreting it. Further, the scribe has added to the confusion by writing ^ in place of i ^ . This confusion is often met with in the Ms. portion belonging to this rule.

1. f*n??rRT ^ ^ r f i r etc. The objector who is not aware of the ^m^j (i. e. P . 7. 1. 16) argues that the reply given by Pat. is merely begging the question. The question, viz. how to know that the ganapatha is earlier and the sutrapatha is later is answered by merely saying * the ganapatha is earlier ( ^ H t f t ) and the sutrapatha is later (argf^fir). Just as one can say that the ganapatha is earlier, one can also say that the sutrapatha is earlier. 2. srf^r?u%aTfl?re3?Tf?r etc. The priority of the ganapatha cannot be established by merely proclaiming ($rf^3rT*n%<JT) that it is earlier $ f ) . Some reason must be given.


Mahabha^ya-Dtpika, Ahniki



3. 3TRSJT mm^t £f%: etc. The objector further points out that Pat's reply Is based on the assumption that snf? in the word w%\w conveys 5333*11. It Is, however, equally possible that k conveys gcf&f. If it is assumed that it conveys 5PFHT, the question of repetition and the consequent rejection of the sutrapatha does not arise. I. The above objection is replied by Par. by adducing the ^r^^ in P, 7. 1. 16 ( See No. 2 in Note 1.7). The author explains with the help of a few more illustrations of Paninian rules that a reference to the number of words presupposes the existence of a list containing words more than the number referred to. For instance, the reference ^ c M ^ in ^orf ^ HHHT^ ( P. 6. 4. 125 ) presupposes the existence o( the dhatupatha containing more than the seven roots beginning with ^ ^ . In the same manner, the reference wmm: in the rule "j^Tf^y^r etc. becomes significant only if it is assumed that the ganapatha containing more than these nine words existed prior to the sutra. The reference *vz?m thus can be understood as meant for the exclusion of other words such as ^ etc. in the <FFftT^[ list. 2. Sf3rafir%ffrnft? etc. The author further remarks that sometimes the rule refers only to the gana mentioned in a Sutra rather than the whole list. For instance, P. 8. 1. 58 : ^ f ^ j ^ is a reference to the ^T1% gana which is already mentioned in P, 1.4. 57. This gana with its full enumeration occurs only in the gatiapaiha. This is another clue to prove priority of the ganapatha. 3. ^^"f^nic^T^ etc. Finally, the author points out that srif^ in the word <cr^f^ need not be understood as conveying irspi^. The examples cited in order to show the necessity of this meaning are all Vedic and Vedic formations do not require a special grammatical rule for their derivation.

1. %m sn^ etc. The objection ^RTF^fq ^ f ^ r can be interpreted in a different way. One who understands the words <j^r^fw and sf^T^fif in their literal senses, namely, * beginning with the word ^ ' and * beginning with the word 3R?' respectively (see No. 1 in Note I. 7), argues that if the priority of the ganapatha is decided on the basis of the fact that it begins with the word <£#, whereas the sutrapatha begins with the word sra?, the sutrapatha also can be regarded as prior because it begins with the word ^#. The point is elaborated by the author in the following manner: The words ^ t etc. are enumerated in the ganapatha in a fixed order. The group of nine words beginning with ^# forms a homogeneous group on

28, 16 ]



the basis of the common meaning, namely sqsr^T. A group like this Is mentioned in a s~tra i-^m-:^:1;^ by referring to it-; ";;rt» For instance, P . 2. 2. 12: %if ^r i^rrer^ mentions <£sTT3T^ which is a part oi the group 7?f^ff|;i3Tf?T^: ( P. 3. 2. 188). In the same manner, the mmtpafka Is a partial reference to all the nine words in the ganapatha including and m which are read earlier than ^m. Therefore, <snrcif?s °f ^AC sut patha also are virtually igjf^s of the ganapatha. The state of being thus cannot be looked upon as a criterion to decide the earlier existence of the ganapatha. The author dees not reply to this objection. However, the same reply given above by Pat. in the form of the 3iq-^ is to be understood here also.

1. The ^rjq^ brought forward by Pat. thus establishes that the ganapatha is earlier than the sutrapaiha. Consequently, the word-to-word enumeration of ^ # etc. in the sutrapatha proves to be redundant. Then Pat, offers one more purpose of the sfiT^TO : ^» efff S R I H ^ — oqsr^rraTJTH5THTjftr% ?"$pn?ftf§r I The present passage explains this statement of Pat. as follows : The words ^# etc. listed in the ganapatha convey specific meanings like o q ^ i f etc. Tn the sutrapatha also tj^ etc. conveying the same meanings are intended. Therefore the Sutrakara thinks it necessary to mention their respective meanings also in the sutras. Thus in order to show their association with the respective meanings the Sutrakara repeats the whole list in the sutrapatha. The srf^rq^qr^ is thus required for the statement of meaning in the sutrapatha. 2. fip^flFcft % etc. One may argue that the respective meanings of ^ etc. are already stated in the ganapatha and so, when ^ | etc. are referred to in the sutrapatha, naturally their meanings stated in the ganapatha also are implied. It is thus not necessary to repeat the whole list. The author replies this objection with the help of an illustration by referring to the restricted application of qualifiers. The qualifiers such as giprren^' wearing white clothes \ *(f&l l possessed of matted locks of hair' are relevant only so far as the action of bringing the person they qualify, which is conveyed by the sentence 5pprrH*RTFT3 nf^R?; etc., is concerned. These qualifiers are not supposed to qualify the same person always. When the same person is referred to in connection with some other action he need not necessarily be qualified by the same qualifiers. The association of the qualifier with the qualified is thus restricted to the domain in connection with which it is stated. In the same manner the meanings ssnrpjn: etc. qualifying *£# etc, in


Mahabhasya-Dipika, Ahnika VI


rlie ganapdfha cannot be continued outside the gavapatha in the Sutra* patha where T # etc. are referred to. It Is therefore necessary for the Sutrakara to refer to the qualifiers along with the words wji etc. again in the sutrapatha so that It is understood that the same words »JJS} etc. which are mentioned in the ganapatha are intended in the sutrapatha. In spite of this justification we find that Pat. himself has rejected this purpose of repeated mention. He says: c ^ f t ?Tlfe ^ r^r "This is also not the purpose. These (words *$ etc.) are read In the ganapatha as qualified by this ( meaning, namely strain). '> Thus Pat. is of the opinion that the words *%% etc. need not be repeated in order to connect them with (heir meaning s^qrwn, because they are already connected with it in the ganapatha and the same words are mentioned in the sutrapat ha*

1. fenft spsN etc. Pat.'s reply as stated above is ascribed to the other disputant ( f ^ ) by the author. He explains it further in such a way that the justification of the purpose stated earlier (see No. 2 in the previous note) is cancelled. He refers to the principle of semantics : vt^R: srqx: * meanings are distinctive \ According to this principle there are as many words as there are meanings. The word ^ conveying sq^JT, for instance, is different from the word ^ conveying any other meaning. The words ^ etc. which are mentioned in the sutrapatha refer back to the words ^ etc. in the ganapatha which are qualified with a specific meaning. Due to this principle they cannot staud for other words tj<} etc. conveying different meanings. 2. src^T'Fll'swfirf^WT ... etc. It is already pointed out in the discussion on *i5rrqs?^5rfrnNr under P. 1. 1.27: ^T^tft sHbroifit, that the word sre which is used as a proper name is not a sHfcro^ although it is phonetically identical with ^ meaning • all', because it is 3TH{%%H ' not included ' in the ^mxf?* list. The same argument is extended here. It implies that the word ^ which is included in the ganapatha conveys sargrFTT* The word i ^ mentioned in the sutrapatha also conveys ihe same meaning because the word ^5 conveying other meaning is not included in the ganapatha. 3. %3ERi?rT etc. The author cites an illustration to explain the point. In fact it is already cited in the Mbh> on P. 1. 1. 44 and 56, There the word :pn*T^ in P. 4. I. 15 : f ^ ? r ^ ^ - ^ ^ ^ - ^ ^ - H r a ^ ^ ^ - ^ - 5 ^ - ^ ^ - ^ q : , which teaches the feminize suffix ^t\ after words ending in the suffixes with the




anuhandha^ suffixes «, sr^etc., is interpreted as a SI^T^R beginning of the suffix"^m^and ending in ^ of the suffix ara^. The objector argues that if mw% thus- stands for a srniTfR and not for the sufffx Jt&% alone, the rule P . 4. 1. 15 will be applicable also 10 ^rorai s a drop of ghee ', "l^n^T •a drop of oil • and the undesirable forms **pmnft and *%WTTT^ft will be derived. This objection is replied by Pat. by saying that although the TOH*.contains *T5T, it is the suffix m%%, not the noun m%. The noun does not come under the jurisdiction of P. 4. 1. 15, because it is not though it is essT. & e a d t h e MhJu ( ! - l 0 6 - 5 ) : ^ ^

4. <nr qlraT«J3F^ra etc. The anubandha ^attached to the suffix *&% performs the function of a qualifier distinguishing the suffix from the noun ?n*?. Therefore by m^ in the S c ^ R , the noun m^ cannot be understood.

1. ^^3^rif^qfe%7l etc. Another purpose of the repetition of ^ etc. ia the sutrapatha is suggested in the Mbh. The sequence of the words listed in a ganapatha is not always the same in all the systems of grammar. For instance, in the case of the SFTtf^ list the author has pointed out that the three gamsutras y^wxm •••, ^ ^ ! % -land s r ? ^ srff - are read before the c^if^ list in the Paninian school, whereas they are read after the x?F%f% list in the school of ApiSali. P. 5. 3. 2 : f%H#^RT^f **ftssrif^: and the subsequent rules teach various suffixes such as <sfg^, sr^, etc. after the word ft^ and a H ^ W ^ °^lQT t n a n t n e g r o u P beginning with f^ in the ganapatha. Now, if the order of words in the mtfk l i s t according to the school of Apisali is adopted, the three ganasutras ^ # T O ° etc. being included in the group beginning with the word % (which is included in ^ i f ^ list), the suffixes g%^» 3o£> etc. will not be applicable to them as indicated by the word « r ^ i % q : in P. 5. 3. 2. The addition of the suffixes d ^ etc. after *jsf etc. is however, desired in order to arrive at forms like ^cT: and ijbr. This is achieved by the repetition of ^q*i Q etc. in the present rule. The rule repeats sHfamH^T for ^ etc. which, being thus doubly confirmed, sets aside the prohibition of the addition of 3 ^ 5 ; e t c . indicated by the word wrsqif^^:. The repetition thus serves the purpose of allowing the addition of d s ^ e t c . after ^ , q* etc. which was blocked by the q | ^ r 2. a r j ^ n i r ' ^ ^ ^ T H 5 ^ ^ etc The author gives an illustration to show how the repetition of a certain statement results in cancelling a grammatical operation which is applicable in the absence of the statement. P. 1.4. 84: s n j ^ n t teaches the ^tsra^^qH^T f ° r t n e w o r c * ^ 3 m t f l C






sense of indication. Another rule P. 1. 4. 90 : ^<frc ? ^n'?3r ufknm^i, which is a more general statement, also teaches the #5Ti iur srg in ihe same sense. The repetnicn oi ^^qreRT^sTT for P. 1.4. 84 thus becomes redundant. This is stated in the Var. etc. Pat. removes this redundancy by arguing that the repetition is meant to re-establish the use of the accusative by P. 2. 3. S and cancel the application of P. 1. 3. 23 which teaches the employment of instrumental when a cause is to be conveyed. P. 2. 3. 8 : ^ K ^ ^ ^ t ^ r % fgcffcn teaches the addition of the accusative ending after the word which is semantically connected with a ^sreH^fni. P- 2. 3. 23 : %^r cancels P. 2. 3. 8 since it is later ( m ) if both are simultaneously applicable. However, when 3T*j is used in a sentence, due to its w^W^t^t^l which is repeated by P. 1. 4. 847 the application of P. 2. 3. 23 is cancelled and P. 2. 3. 8 applies. ( The Mbh. on Var. 2 on P. l. 4. 84: ^hf aff sranR^i %rta? w ^?ct 3a|sra^ftog%> f^dmfif s terc^i % ? T ^ %?fNi ^ s ^ R ^ g ^ ^ r i * %^% \ *w^wr$pMi l )• This results -in the derivation of correct constructions such as fir§^ scg TO[ * the lightning (flashes) in the direction of the forest ? instead of the undesired One may argue that P. 2. 3. 8 and P. 2. 3. 23 cannot simultaneously apply since they are caused by two different conditions, P. 2, 3. 8 by =f^jpr^rftar*. and P. 2. 3. 23 by %gp^. Pat.'s statement ®%Qm %g?T^ z%W* (the Mbh. on Var. 2 on P. 1. 4. 84) is a reply to this objection. He means that in the case of the use of SPJ both the rules mentioned above are applicable, because 3T«f is called ^i|5fqR*?N" when it denotes ^roi and a includes %g. 3. e n s R ^ i ^ e t c - T l i e t h r e e r u I e s » r i^TO^.« e t c - ( p - 1- 1. 34-36) are meant for teaching optional *Hfam^r before ^ as will be clear from the concluding statement of Pitt, on this rule (see i. 18 below). However, if the repetition in these three rules is meant to serve some other purpose as pointed out above, how can they achieve this purpose ? The author's solution to this problem is the device of «Ti^iTv According to this device, a rule is supposed to contain two ^pp?s or statements. Each of the rules t^q^j... etc. accordingly consists of two statements. One statement simply teaches the stwm^fT for ^ etc. whereas the other teaches optional s r ^ m ^5jT to <^c etc. before the nominative plural (suffix jas). This ^Tf^^ or understanding two sentences from a rule is obtained by imagining that all the three rules are marked with a 3^fer. According to P. 1. J. 11: ^f^fcnfv^RJ, which is interpreted by Pat. as ^ftl^rf^^: 35R: (see the Mbh* on Var% 7 on P. 1.1.11), the ^fter accent indicates ihat a certain rule teaches




something additional which is not explicit in the ruk its.If. In th 3RT^: 'the additional function' perform, d by :re ~-*>\ : f '- : second statement which implies tiie jOii^iUauoii oi uiv w •._':«:> <iff*T i.^*,,'".^. In this way the repetition serves a double purpose: it repeats the of *jir, etc. which results in cancelling the non-application of ^"N etc, and it also helps g r ^ q i ^ leading to another statement by connecting itself with the words ^f% fcfrrrrr.

1. The import of the passage is not clear. Ft can be conjectured from the last line that the passage deals with the refutation of the view that the sutrapatka is earlier. However, it has no relevance with the previous context which contains justification of the repetition of r$ etc. in the sutrapaiha based on the assumption that the ganapatha is earlier. etc. P. 1. 2. 38 : V^sj^ofnrg^TTn teaches that the words \m £Tfn&i: in the ^ p n ^ i chant are uttered with ST^TTT (grave) accent. Pat.'s remark en this rule, ^ ^ o ? n ^ T T c ^ % suggests that the is optional. Kaiyata further explains this as follows: If the words are understood as preferential, the srg^jTf is already applicable by P. 6. 1. 198 as well as P. 80 1. 19 ( both the rules reading w m f ^ i ^ i ^ i ) . In that case the rule becomes redundant. The statement of sr^ixf by this rule thus indicates that the words ifa] s^\m: should not be understood as coreferential. The rule thus indicates S^PJHPCR * another arrangement' of^^r: and The above example is thus understood. what it implies. However, it is not clear

3. snf^rfo^sriw ^ 3rc*n?3R^ e t c - The word % i ^ is included in f^ list. P. 5. 3. 2 : r % ^ ^ ^ etc. contains a special mention of the word % * j ; , because, although it could be subsumed under sfsnTO^ the q | ^ m '«isnfi[W would not allow it to be included under e ^ i s p ^ . The separate mention of % ^ implies the fixed ordering of words in the gana« pallia. Such references to fixed sequences lead us to conclude that in the section dealing with H ^ T H ^ I the earlier existence of the ganapaiha is presupposed.

1. 1^bn%^T3% etc. Pat. rejects the purpose of the repetition as described in 1.15 on the basis of a ^"[q^. Panini makes use of the word in P. 8. 2. 1 : ^ 4 ^ ! % ^ . It is derived by adding the suffix R a f t e r ^ .


Mahabhasya-JDlpika, Ahnilca VI


This nipatana implies that the prohibition of 3T<^ etc. made by the t%^?:' does not apply to tjcf etc. ^ ^ etc. Pat. adduces one more ^rrq^ from [Cat's Var, ^^rmit%% %f ^WWWiWT?qsrr^IW: ( Var. ?. on P. 8.3. 13), which contains the word ^rf^w. According to the author this another ^w& is mentioned by Pat. in order to avoid the misunderstanding that the earlier =TPT35, namely, ^•^i is restricted only to the word ^ .

i. Finally, Pat. gives the purpose of the snlrq^qH m the three sutras as follows: %i m% SRtSR — sr% fimiqi w$m?$t^\ The three rules are meant to teach optional ^^RifW^f for ^ etc. in their respective meanings before 3fCT. The repetition of all the words along with their meanings in three separate rules is meant to avoid intermixture of the three meanings. If one rule, *<^T3*?V ^ were stated instead of the three, it would mean that the cine words ^ , etc. are optionally called sPFns^ before* ^n^ when they convey s ^ ^ r r , ^lf?HFf etc - * n t n a t c a s e a n v meaning will be associated with any word. In order to avoid this confusion the three rules, separately dealing with each group of words connected with their respective meanings are stated by Panini. Kaiyata remarks : srfifrq^n5 Tsft1 etc. One may argue that even then the length of the rules could be reduced although their number cannot be reduced. For instance, instead of saying, ^q^[i^^"%nTtTRiqinvTOfiT Panini could say tj^ft. ^tf*T HH' The author replies that although in this manner the length of the sutra could be reduced, the separate statement in the form of the rule could not be avoided. The existence of the three sldras with the same wording as the three ganasutras remains unquestioned. 3. qsjT ar^re tftfa etc. When the wording <J^T^I: ^^r replaces T e t c - nothing much is gained by way of brevity. The same three — groups are retained there also. In the same manner, the wording ^ T T I ^ R . ^ does not cause much nR=*. Both the alternatives are almost of equal status since the number of the rules remains the same in both.

1, Pat.'s statement : srfa f^*n«rt sr^U*ft%, which describes the final purpose of the s r f ^ q r e , implies that the rule tjsfansR... etc. is understood as a 5nsrf^*TTCT« The sr^T*ra*CT which is already applicable to t p etc. by the general definition STsfrftH *HNnwrft is rendered optional by the present




rule before ^ . The rule thus becomes a sngf^m? -vhkh is desirable. On the other hand, if it is understood as 3TaTaf§RTTTr, it would mean that the ^TiTO^f which is not at all applicable 10 v$ etc. is made optionally applicable to them before *n^, while elsewhere they are not *?SRT*FJS a t a^» The author now asks : £ Why is the rule not understood as sngrfstaTOT m the absence of uiirq^iH? * He himself answers the question by saying that it is more proper to understand a rule as arsrrsfiraTqT. because by understanding a rule as STRrfifarreT, its importance as a statement is reduced. A ^TH^^NI does not teach anything new, it reiterates the same statement made earlier. Further, the words are interpreted in the light of the preceding context. Thus they are not interpreted in their natural way. This leads to iffT^ On the contrary, If It is understood as snrr^fwvn^T it is accepted as makiog a new statement. Therefore, a preference is naturally given to 2. 3 % ^T5*te 5??^T Qtc. The author supports his answer by citing two statements from the Mbh. The statement 3T^IT 3r^R S ^T *Tf crVs^fe; ( Mbh, Tc T on P. 2. 1, 69 ) suggests that one should always try to get more and more results with less efforts. In other words, each rule should be understood as teaching something new. The second statement rT3T^T f^f^T^cJ-.. etc. (Mbh. on P. 1.4. 3) also suggests that when a choice is to be made between two kinds of statement - a new statement and a restriction - the former should be preferred. Cf. f3r%fiFTOH*T% ikfefrn *mm^(Pari. 109 ). Both these statements imply that srqT^f^r^rmT should always be preferred to STTH^T^T.

[ 1.20 ]
1. Here the author raises another question. It has been pointed out that in the absence of srf^ro^n; the rule would be understood as srsnsrfwsTN?, because the expression ^ 7 ^ t ?m does not denote the same words ^ etc. characterized by specific meanings like sq^^n (see Note I. 18). Here the author asks, why are the same words *j# etc., which are read in the ganapatha along with their meanings, not understood in the sutrapatha? To put it in another way, why are the respe;tive meanings of *$ etc. not understood in the sTitrapatha ? Why is it necessary to mention them again ? The author replies this objection in the same manner in which he has already dealt with it (see Note 1. 1 3). He works out the same argument by citing a few more illustrations. His argument can be summed up in the following manner : The qualifiers of an object which are stated in a particular context are not necessarily continued in another context. For instance, in the sentence * Quickly bring that goat which is tied' the goat which is tied cannot be brought unless he is untied. Therefore the qualifier ' tied' of the goat cannot be understood in connection with the action of bringing.






In the statement ' Give food to him who is practising pranayama in water' food cannot be offered to him as long as lie is practising pranayama, Another injunction stated as an illustration is * Employ a person who has restrained his speech., who is firm and motionless, a; a witness '. How can a person who has restrained his speech become a witness? He has to open his mouth for acting as a witness. The qualifiers which are incompatible are thus not continued. On the other hand, the qualifiers which are compatible with the new context are continued. For example, when It is said * Give food to him who was seen wearing white clothes', white clothes can qualify the person in the context of feeding. To sum u p : The continuation of qualifiers in another context is uncertain. Therefore, the meanings stated in the ganapatha may or may not continue in the sulrapatha. In order to remove this uncertainty the sutrapatha contains srftq^qis along with the meanings. 2. %n SFf?%%% ^•'• The 5jfit<r?qre confirms the nature of the rule as STFB^rifl which does not allow the words to be used in a different sense.

1. The objector wants to retain the reading ^I^^V W% proposed by him and tries to attain the desired result by means of some other devices. The main objection against this wording was that it resulted in undesirable 3TSTOH'SflfT. Now the objector suggests two ways out of this difficulty : ( i ) continuation of the word &w\f% from the earlier context and (ii) understanding the word H^NTO^* whi;h is also continued, as a qualifier of ^ etc. Let us see how the author deals with both these solutions. The author states that if the word sr^ftfsr is continued here, it will be continued also in P. 1. 1. 33 : TO^roTcTOT5qTsMcrc^*nai I It is true that the srgffrT of the word sra^tft would result in rendering P. 1.1.34 a JTOfw^n. However, it will equally render P. 1, 1.33 a srurfinTTCT, which is undesirable as far as cR is concerned. The word cR in P. 1.1.33 stands for words derived by adding the suffix zwi ( P - 5 - 2- 4 2 )• I f p - *• *•3 3 i s regarded a STTH^^NTJ it will lead to the pairs of forms 3 - ^ and *^rn?T'» whereas only 3"*FI is desired. It will, on the other hand, not be applicable to words such as %<rq, f^ni etc. since they are not ^cfif^s. In this way, since the sr^frf of H^rrftf?* results in the derivation of undesired forms and non-derivation of desired forms, it has to be discarded. The word srefcrra^ which is continued in all the subsequent rules stands for a ^ r . Therefore, it represents its own form (s^r^q) rather than meaning. It cannot function both as a ^ \ and a qualifier of H % ^ simultaneously. Therefore, it is not proper to say that the word H3*rW3; is understood as qualifying ^ f ^ in order to render the rule




2. 3j%n H^T^rs^: sTcfffrsn^T^ etc. The objector cites an illustration to prove that a HsTT word can function as a H % ? N word when it is continued in a subsequent rule. P. 1.1.23: ggnorag^far w$m defines #?srr. The word J ^ T T which is a ^*rr is continued In the subsequent rule, P. 1. 1.24 * ^IFTf q^ which means that a ?t*??T[ which ends in % or ^ is called ?%%. It is quite obvious that the word ^m which is originally a ^m\ becomes a # % ^ in this rule. In the same manner the word ^HhwHrfSf which is a ^ r i can function as a H % ^ o r a qualifier of # % ^ i n the subsequent rules where it Is continued. The reply to this objection is already given in the preceding Note. \

3. qrj ^PcpWfcg— The author repeats the same observation regarding the understanding of qualifiers in a different context (see 1. 13, and 1.20). He reiterates the point that a special effort to continue the qualifiers of objects mentioned in a different context is always preferrable. Therefore it Is better to directly mention the relevant meanings o^c^n etc. also in the sutrapatha, rather than to imagine their continuation in this context.

svam ajnatidhanakhyayam / P. 1.1. 35
*(The word) sva (is optionally called sarvanaman except when it conveys a kinsman or wealth *, before jas)

Summary: There is no Vdrttika on this rule. Patanjali raises a question regarding the mention of the word an^m in the rule and himself answers that it is required to prohibit the optional H^TRHsTT to ^ only when it is used as a synonym for a kinsman or wealth. The author of the Dtp. first illustrates the four meanings of the word ^ . Then he explains the objection against the mention of the word srr^i as based on a certain view viz. s?£fqf ^reqHr: and refutes it by adopting the opposite view viz. ^ t 3T*FIT$|: t In his usual style he tries to enrich the whole discussion with illustrations.

1. The author enumerates different meanings of the word ^ as ' s e l f ' o w n ' , ' k i n s m a n ' and 'wealth*. These meanings are illustrated' by citing examples from Vedic literature. It appears that one or two sentences between the words ^ sr#% and su^Fcr ^ I ^ T ... etc. are missing, because the sentence 3H3?cf q ^ 9 ^m. etc. contains the word ^R*. in the sense of kinsmen rather than wealth, and so it cannot be regarded as an.


Mahabh58ya--Dipikai Mnika



Illustration of ^ meaning vm. The examples cited here do not appear In the respective &equence of the meanings mentioned. 2. snSrfirsri^r etc. P. 5. 2. 126 : ^ B f W i teaches the word ^fift^ as a nipatana conveying th^ i mastery \ Tradition explains the word as derived from m by adding the suffix anftsrq, ( ^as. on P. 5. 2. 126 : ^f^TS^^pJ^T%fr mm m^^mx §m?r^% I ). The author records this word as conveying an additional meaning of ^ , namely \*%4. It is however not clear why he refers to w^ and f^j? separately as meanings of sw• Further, while the other grammarians such as the KaSikakara mention the suffix mfk% (or infiRX) the author here mentions anfirfN^where t n e v o w c l I d o e s not seem to serve any purpose.

1. Ws^lilifcui %?j$?f\ Pat. raises the question about the use of the word tsmm in the rule. According to the objector, since Panini could simply say ^ i j g r f e R ^ : and achieve the same result, the mention of 3TF?*n in the rule is redundant. The author explains the objection in the following manner : Some grammarians hold that the words in a sentence convey their respective meanings while the additional meaning is understood from the connection of the words with each other and it is the meaning of the sentence as a whole rather than that of the words. This view is summed up in the statement ^*Flf WR^T^: \ The author gives an illustration to explain it. q g m^ (srrewr) is an injunction from a ritual text. It means ' ( One should sacrifice) an animal which is a goat,' The word q g denotes animal in general. However, when it is connected with the word ^rq as in the above sentence, it denotes only a sacrificial animal. The restriction of its meaning is thus the result of its connection with the word ^rn* Consequently, q^j in the sentence does not convey any other animal such as a donkey etc. The universal qg^cr that is understood from the word q g used in the sentence does not associate itself with a goat as well as a donkey at one and the same time. In the same manner, in expressions such as ^% Trig*: ' one's own bulls• and^rg^T: 'one's own sons' the word ^ simply conveys the meaning • own \ whereas the meanings * wealth f and « kinsmen ' are understood from the connection of ^ with TUT- and 5^1: respectively. Thus the word ^ is not an w?*rr ' name ' of a kinsman ' or * wealth ' as the two meanings are understood only from the connection ( HH'T ). This being the case it is not necessary to say si^ifingFn^qi^T^ since the word ^ always conveys the meaning • own ' and it can never be an STT^TT * name ' of gTifgr or VR. The whole argument as explained above is the result of the author's ushering in the view HOT?f 3FRP5:! Pat.'s argument regarding both, th$




objection and its refutation, is very simple as will be clear from the following Note.

!. Pat.'s answer to ihe question, effi^rT^rfif %TPT*^? is : sni^PTTOferm^t m ^ s r s ^ r a w *m^ *n ^ ^ %%v- twi* j^n: # m^> ^ n *TR: I This can be explained as follows: The word ^ conveying 5ff% or\3Rcanbe found in two types of constructions : (1 ) <rpr t^n * his own ( kinsmen; or wealth) * ( 2 ) && ^% *Tf^: ' his own bulls \ In ( 1) the word tW is a synonym of kinsmen, whereas in ( 2) it denotes the meaning property only indirectly, i. e. through its association with TIJW- which it qualifies. The word scp^if in the rule implies that the prohibition * 3T^i^TtTP?3PfflRt * should cover cases like ( 1 ) only, where *w is an sn^TT ( synonym) of ^7% or ^f. Since the word ^ in cases like ( 2 ) does not denote g-ffif or ^ directly, it is not covered by the prohibition. The author goes further and makes the simple reply given by Pat* more complicated by bringing in another view, namely, ^ t WF?wl; I According to this view the specific meaning understood from a word as a result of its association with another is the meaning of the word itself, not of the sentence. Thus the meaning * kinsman ' understood from the word ^ through its association with 537: is the meaning not of the sentence but of the word itself. According to this view, which is opposed to the view $wif ^r&WT-, the mention of 3TT^i is necessary to restrict the scope of the prohibition to the cases where ^ directly conveys 5f[% or ^s|. In the expression %%j TII^:, or ^y. 537:, ^ does not directly convey the meanings ' wealth' or f kinsman'. The view %^t ^pflsr* differs from the view *teFTT ^T^TT^.* only in that according to it the specific meaning understood after connection of a word with another word is the meaning of the word itself. As explained by the author in this passage, according to this view every word has a capacity to convey many meanings.

antaram bahiryogopasamvyanayoh / P. 1 . 1 . 36
Summary x The Varttikakara questions the mention of the word in the rule since it can be subsumed under the broader meaning In the second Varttika he justifies the separate mention of ^qsh^W* because, according to him the meaning ^ffWT cannot denote a pair of garments which are of equal size. Further he states that the word w%% is not called a H#HT?Hswhen it is construed with the word 5^.


Mahabhasya~Dipika, Aluiika


[ 31. 6

Pat. does not approve of the justification of the separate mention of the word ^ T H ^ R given by Kat, He contradicts Kat.'s statement, giving his reasons. Curiously enough, he does not give any solution, nor does be expressly show agreement with the first Var. questioning the mention of the word

The author of the Dip. tries to defend Panini. He elaborates upon the second Var. of justification and taking a cue from the word srff^ used by Kat. shows how Kat.'s justification is correct by discovering some other Illustrations which Pat. had not noticed. 1. $BC V^W\ srPrqfrT:. The word gffprfrT is interpreted by the author in two senses: ( 1 ) gffqy €t*T: ' connection with outside ' and ( 2) SITJR sftrr: 1 connection with something which is outside \ The examples of the word 81?^ conveying both these meanings are given. According to the author, the word 3r??r* gets the designation STsfcro^in both these senses. Thus sr?^ JJ^T-* outer houses' means the houses of outcaste people which lie at the outskirts of a town, if the first meaning of ^%^T?T stated above is adopted. If the second meaning is adopted, the same expression 3T?erc ^ T : means ' inside houses * or interiors of a house, because they are connected with the exterior of the house ( ^ I S R *\\T\\). 2. q^gTO^i^sfir etc. The word ^ W Q ^ R also is understood in two different senses: ( I ) gronfolH ^ ^ q H ^ T ^ ' 3fq*T^FT is that which is covered * 1. e. an under-garment which is covered by an outer garment and ( 2 ) ^TOWRHSTR1^! SWoqjq^ € 3 " q ^ H is t n a t b y which ( something) is covered ' i. e. an outer garment by which an under-garment is covered. The word s - w ^ r a is thus derived from sq + H ^ + *% ' to cover closely', by adding the suffix ^g^ ( P, 3. 3. 117) in the sense of 3 ^ and ^ o r respectively resulting in the above two meanings. The sentence * s w ^ P H ^ ^ [ q ^ T H H ; ' should be supplied in the text of the Dip. as it is required to explain the meaning

1. The Var. 1 : 3q£s?n?n*ITOR«Hj g%wfal fcfr^I^ raises the objection that the separate mention of 3W3qrH is redundant since the word lf|*n*T includes that meaning. As pointed out in the preceding passage each of the two words, *I%?JFT and 3"q*r^H, conveys two meanings. The word means either the under-garment or the outer garment. If the word is understood in its first meaning, namely, connection with outside, it




22 i

can include in itself ^qw^jsf meaning an outer garment (because it is connected with outside). If, on the other hand, the ^%rfiT is understood as conveying connection with something lying outside, it can cover ^ " q ^ n ^ meaning outer garment (because it is connected with the outer garment that lies outside). Thus the meaning of the word ^q^o^H can be included in the word ^%^fn in either case. It is therefore not necessary to separately mention the word 2. ^TTS^m^r^ etc.. Kat. justifies the mention of 3"q*?oqH as meant for a pair of garments. His statement is explained by the author as follows : Usually the under-garment is shorter than the outer garment. However, when two garments are of equal size, it is difficult to know which is the outer garment and which is the under-garment. Consequently, neither of them can be included in either of the two meanings of ^f|qnr, since there is ambiguity. Thus the word gfifqfrT cannot cover such a pair of garments which, being of equal size, cannot be separated into under-garment and outer garment. 3. sfqffffcrerft' etc. One may argue that when a pair of garment is of equal size it can be understood from the way one wears, which is the undergarment and which is the outer garment and thus the doubt does not exist even when both are of the same size. To this the author replies that in the case of some garments like a cloth with flowers painted on it or a loin-cloth one can easily understand even without seeing how they are worn, whether they are under-garments or outer garments. However, in the case of garments of equal size one cannot recognize until they are worn. Thus the doubt exists when they are not put on.

1. snnft ^ ^ ?Tg^: etc. Pat. disagrees with the above justification given by Kat. He says that an intelligent man can distinguish between an under-garment and an outer garment even though both are of the same size. In this connection the author refers to the interpretation of this statement by some grammarians. According to them Pat. means that when the pair of clothing is of equal size, an intelligent man knows that it is going to be used as a clothing and therefore it will be described a s ^ q ^ ^ H . It is not necessary to separately mention the word 3"q^n^T only for such pairs, because being understood as would-be s w ^ H S they can be well included in 2. sr?^ g Ttrvffi etc. The author records the view of some other grammarians who want to re-establish Kat.'s position. They argue in the following manner : The Var. reads ^ m ^Tre^TTjsrsfo; i The word


Mahabhasya'Dlpika, Ahnika VI


after ^Tf^i^l implies that the mention of ^ q # f R is meant for some more things. For instance, when three or four garments are put on one upoa another, the first inner garment cannot have ^%^TH connection with the outer garment which is third or fourth. Therefore, it cannot be denoted by 3f%Jf?|. Similarly, the first and the second layers of inner-garments are not connected with the outside, because they are covered by third and the fourth ones.. The ^4^lWsfT, is however, required in all these cases. There* fore a separate mention of 3W**?H is necessary. We leave the whole discussion to our readers without our comments !

1. The Var. ac^ftf^ = r ^ ^ teaches that it should be stated that when the word 3f*cR conveying 3%qt*T (in either of the two senses ) is used as an adjective of $^'city *, it is not called *rfwr^« Therefore, the locative of «F?<rcT in sr?cRT3i 3 $ ' in the city which is at the outskirts s is correct and *3C?cr^n^ which is incorrect is not derived. 2. snfeM^r^St etc. The above Var. teaches prohibition of <gjl t 0 *?^rc and the example srer?rcf gf? contains the word This is apparently inconsistent. ^TI^H^I is applicable to sPcl*, not So there is no question of the wrong derivation STFcR^m. The author removes this inconsistency by referring to the Pari. (Pari. No. 71) which teaches that the mention of a jrrfiRf^ implies also the mention of its forms in specific genders. Accordingly, «r?^f also stands for WftZT* which is its feminine form.

ABBREVIATIONS Bhar, Dip. Kaly. Kat. Kas. Mbh. Ms. NSg. P. = Bhartrhari = Dipika = Kaiyata = KatyHyana = Kaiika = Mahabhasya — Manuscript = Nage§af Nagojibhatta

= Panini = Paribhasa = Paru = Patanjali Pat. Sid-kau, = Siddhanta-Kaumudi Var. «= Varitika



Sign up to vote on this title
UsefulNot useful