Emendation Of Some Verses In Bhrtharis Trikandi

Author(s): Ashok Aklujkar
Publisher: Ashok Aklujkar
Catalog link: https://jainqq.org/explore/269725/1

Page #1

EMENDATION OF SOME VERSES IN BHARTṚHARI'S TRIKĀNDI*
By Ashok Aklujkar, Vancouver
1.1 Anyone who has tried to edit an ancient or medieval text systematically is usually aware that collection of manuscripts, preparing a genealogy or stemma codicum, collation of relevant manuscripts, and evaluation of available readings amount only to apară vidya or primary textual criticism. A good editor must go beyond even these demanding and time-consuming procedures. After he has proceeded as far as the manuscripts will take him, he must at least try to recover what is not in the manuscripts. Such secondary textual criticism becomes necessary in the case of ancient and medieval works mainly because we can rarely go back all the way to the autograph or the author's own copy. Almost always even the oldest or most reliable of available manuscripts is considerably removed from the original copy. It has usually a long line of predecessors and is seldom better than, say, an inattentively prepared press copy of a modern author.
1.2 As is evident from many of my earlier publications, I have been engaged in a textual study of the works of Bhartṛhari and his ancient commentators for the last few years. In the present paper, I shall confine myself to only one work from that complex of works and to only one aspect of secondary textual criticism; I shall concentrate on the kārikā or main verse text of the Trikandi (or Vakyapadiya) from the point
* This paper is a revised version of a part of the paper I read in the South Asia section of the 30th International Congress of Human Sciences in Asia and North Africa, which met in Mexico City in August 1976. I wish to express my thanks to all the individuals because of whose cooperation it became possible for me to study the Trikāṇḍī (or Vakyapadiya) kārikā text in its three traditions: (a) Kārikā manuscripts, (b) Vṛtti mss., and (c) Tikā mss. Professor RAU's generosity in allowing me free access to the typescript of his edition of the karikās (1977) helped me considerably in finalising my conclusions at an early date. I am grateful to him also for providing several other study facilities. The findings contained in the paper were made in 1973-74 when I was a research fellow of the American Council of Learned Societies. They were given the present form while I was engaged in Trikāṇḍi research as a grantee of the Shastri Indo-Canadian Institute.

Page #2

64
ASHOK AKLUJKAR
of view of emendation. A further restriction I shall observe is that I shall not discuss any acceptable emendation that is somehow present in the printed texts and manuscripts pertaining to the Trikāņdi. In the beginning of my study, I thought of many more emendations than I present below. Gradually I discovered that some of the readings I preferred were available in isolated manuscripts (e. g. akşyādinām in the place of akşādinām in 2.409), or were implicit in the commentaries of the Trikāņdi (e. g. nirbhāgam in the place of nirbhāge in 3.11.14), or could be attested in works quoting the Trikāņdi (e. g. ākhyātam sabda-samghāto in the place of ākhyāta-sabdaḥ samghāto in 2.1 as can be gathered from Pārthasārathi-miśra, Nyāya-ratnākara p. 860), or had been suggested by earlier editors (e. g. manvate in the place of manyate in 2.121 as is the suggestion in Raghunātha Sharma's Ambākartri). Here I shall refrain from discussing all such cases. Emendations based entirely on different splitting and different punctuation of the text matter will also be left out of consideration. They amount to correction of previous editions, not necessarily of the more basic manuscripts. Anyone wishing to know them should go through Professor WILHELM RAU's (1977) edition of the kārikā text, in which the editor has kindly recorded my preferences in sentence segmentation implicitly in some instances (e. g. 2.144c, 2.241 c, 3.1.28b, 3.14.271 ab, and 3.14.307 b) and explicitly with the abbreviation “AA” in some other (e. g. under 1.28d, 2.20 a).
2.1 It is well-known in Mimāņsā that a person's eligibility to perform a certain rite depends on three factors: (a) He must have desire to perform the particular rite. (b) He must be capable of performing what he desires. (c) He must not have been declared ineligible by the scriptures for the specific sphere of action. Verse 2.79, occurring in a group of verses devoted to listing some mimāmsā-nyāyas, seeks to state this relatively simple observation in Mimāmsā: 2 arthitvam 3 atra4 sāmarthyam asminn
1 The numbering of Trikāņdi verses varies considerably in the available editions. I follow RAU's (1977) enumeration.
2 Nonsensical variant readings have been ignored in the following lines. Readings that are grammatical in themselves (and to that extent sensible) are specified in the footnotes irrespective of their syntactic and metrical acceptability. Their sources and their relative worth according to the objective criteria of textual criticism will be known from my proposed edition of the Trikāņdi. Any inclusion in the present paper of that mass of information would have made some of my arguments even more difficult to follow.
3 v. 1. arthitvād. 4 v. l. asya.

Page #3

Emendation of some verses in Bharthari's Trikāņdi
65
artho5 na bhidyateśāstrāli) prāptādhikāro 'yam vyudāso 'sya kriyāntare. Il In the form known at present, this verse succeeds in making the intended statement as far as the quarters (pādas) 'a', 'c', and 'd' are concerned. It is only quarter 'b', syntactically related to sāmarthyam of the preceding quarter, that does not make any contextually acceptable sense. This is true irrespective of the reading one adopts: sāmarthyam asmin artho/arthesarthi nibadhyatenibhidyatená bhidyate/na vidyate. Therefore, I would like to suggest that 'b' should be read as: asminn arthini vidyate. There is only a slight departure from manuscripts in reading thus. We are required to assume that artho/arthe arthi ni/na is a corruption of arthini, which is in fact attested in one of the versions of the kārikā manuscripts (see RAU 1977:64). The other assumption implicit in the emendation, namely that bhidyate resulted from vidyate, is justified by the textual evidence available for 2.136 (see 2.4 below), 3.6.13 (vidhiyate for bhidhiyate), 3.14.417 (°bhāvanā for Ovāsanā°), and 3.14.536 ("hidhāne for vidhāne). As for the meaning expressed by the suggested reading, it is easy to see that a remark like sāmarthyam asminn arthini vidyate ('This person seeking (to perform the specific rite) has [the necessary] capability') is both contextually proper and agreeable to the positive tone of quarters 'a' and 'o'.
2.2 Ever since reading Trikāņdi 2.91 (aprasiddham tu yad bhāgam adrstam anupaśyati stāvat tv asamvidam mūdhaḥ sarvatra prati padyate II) for the first time, I have been puzzled by its wording. At first I thought that what I suspected to be an error existed only in the printed editions .- that the manuscripts would dissolve my problem. However, when I consulted the manuscripts, my hope gradually evaporated. Choice existed only between yam and yat and between tāvati and tāvat tu. But these were not the words that gave rise to the problem. Nor did they help in solving the problem by precisely determining what the other words of the verse should be. From all indications the problem was to be located in the words asamvidam 'non-cognition, non-awareness', sarvatra 'everywhere, in the whole' and prati padyate "comes to, com
5 v. l. arthe, arthi. 6 v. 1. na vidyate, nibhidyate, nibadhyate.
7 As far as I know, Nandi-nāgari is the only script in which v and bh are so close as to cause confusion. Such a confusion is likely, although not equally likely, also in old Telugu and modern Udiya. There is no evidence that the former was used to transmit the Trikāndi text. The use of the latter is ruled out by the very age of the manuscripts concerned.
8 v. l. yam. 9 v. 1. tavaty.

Page #4

66
ASHOK AKLUJKAR
prehends, thinks'. The context is this: Bhartphari is an advocate of the indivisible sentence-meaning theory. On the strength of several insightful observations, he argues that, although one can, and one does, divide a sentence into certain constituents while engaged in linguistic analysis, on the level of cognition and actual communication a sentence or a sentence-meaning is a whole - something that transcends division. To bolster this thesis, Bhartshari states and refutes almost all the objections one can think of. One apparently powerful objection runs thus: Sometimes a person hears a sentence like vanāt pika āniyatām 'Bring a cuckoo from the forest'. It so happens that he does not know the meaning of the word pika. Hence he says, 'I understand vanāt and ānīyatām, but what does pika mean?' Now, is such a question possible if the sentence were an undivided whole ? Obviously, the person involved understands the meanings of vanāt, aniyatām, and pika separately. In other words, there are divisions in sentence-meaning and hence in the sentence. Bhartshari counters this objection by maintaining that actually the person concerned did not understand the whole sentence, although his question gives us the impression that he did not understand only a part thereof. Someone who does not know what a wild bull is and who sees a wild bull for the first time actually does not cognise the entire creature, although it may seem to him at a later moment – on reflection – that a certain part was familiar and a certain other part was previously unseen10. Now, it is such a statement that verse 2.91 should contextually contain. However, one cannot derive any such statement from it unless one switches its words around in a very unnatural way or supplied certain crucial expressions that a careful author like Bhartphari is not likely to have left out. If one accepts the reading tāvat tu, the natural meaning of the second half is: 'Until then the dullwitted one accepts non-cognition with respect to the whole.' On the other hand, if one accepts the reading tāvati, the natural meaning of the second half is: "The dull-witted one accepts non-cognition of that much (of a part) as non-cognition of the whole.' Both the meanings are contrary to the context, as what the dull-witted one does in Bhartshari's view is exactly the opposite; although he has not comprehended the entire object, he thinks he has comprehended a part thereof; he accepts non-cognition in part, whereas, in Bhartphari's view, he should accept
10 What happens at the first moment of the encounter is relevant in understanding the grammarian's position, not what happens in the later stage of introspection, since the grammarian does not deny that sentence and sentence-meaning are divisible on the level of analysis.

Page #5

Emendation of some verses in Bhartshari's Trikāņdi
67
it in totality. The commentatorll senses this difficulty and implicitly construes the line as mudhah tu tāvat sarvatra asamvidam prati padyate to derive the meaning: 'But (tu), on the other hand (tāvat = punaḥ), the dull-witted one (mūdhah), in truth, comes to non-cognition (i. e. does not cognise) with respect to the whole'.12 But this construal is extremely forced. It ignores the relative yat or yam of the preceding half. It arbitrarily tears asamvidam from tāvat tu or tāvati, and places it between sarvatra and pratipadyate. It cannot stand unless the commentator writes a sentence in between telling us what the reality is (paramārthatah). It overlooks the fact that asamvidam prati padyate is a strange construction, since what it should contextually convey can be stated simply by na prati padyate 'does not cognise’13.
2.3 The considerations noted above led me to believe that textual damage had taken place and that the commentator of the second book of the Trikāņdi had simply offered a desperate interpretation out of his strict adherence to manuscript readings (sthitasya gatiś cintanīyā). I was confirmed in my belief when I chanced upon Bhoja's (Śộngāraprakāśa vol. II, p. 333) citation of the verse: aprasiddham tu yam bhāgam adrstam anupaśyati sarvatrāsamvidam mūdhastāvati prati padyate. From the fact that Bhoja transposes the words tāvati and sarvatra it is clear that he felt the same contextual and syntactic difficulty as I did. Of course, this does not mean that I must accept Bhoja's solution. Instead of effecting a dainty little emendation, Bhoja almost rewrites the second half of the verse; he acts like a mechanic who prefers replacement of parts to repairing them. A less radical solution is called for and is possible. The problematic verse should be read as: aprasiddham tu
11 The commentary (Tikā) to the second book (Vākya-kānda) of the Trikāndi is generally ascribed to Punyarāja. In a 1974 article I have argued that it should be ascribed to Helārāja.
12 Tika: tatra ca gavayādau bhāgam aprasiddham adrstam aniścitam anupaśyati mudhah. paramārthatas tu samvit tasya sarvatra nirvibhāge tasmin kutaścit käranān notpanna. sarvatrāyam asamvidam [ asaṁvidan ?] tāvat punah prati padyata iti.
13 The phrase asamvidam prati padyate is plausible if the intended sense is 'advocates non-cognition, holds non-cognition to be the case. But then one would be speaking of a theoretician, of someone consciously putting forward a view, not of a common man as mūdha indicates. In the obviously related preceding verse, prati padyate is used to express what the person perceiving a wild bull thinks, not what deliberate judgement or view he develops.

Page #6

68
ASHOK AKLUJKAR
yad bhāgam adrstam anupaśyati | tāvat tv asamvidan 14 mūdhaḥ sarvam na prati padyate. || It is then to be understood thus: 'As he regards the notwell-known part as uncomprehended, the dull-witted one, not cognising that much, to the contrary (actually, in fact), does not comprehend the whole'. In this interpretation, the words of the original are not forced out of their natural affiliations, in order not to violate the context. All that one is required to suppose is that asamvidam mudhah is a corruption of asamvidan müdhaḥ (see fn. 14) and that sarvatra is a corruption of sarvam na. Both the corruptions are highly probable in older Sanskrit manuscripts, in which the words of a sentence, as is well-known, are rarely written separately; "dan mū0 written together as 'danmūO can be mistakenly read as "dammūo and later written as dammu (mis). leading the editors to separate it for the modern readers as Odam mū. Similarly, as an anusvāra is pronounced like the nasal of the class of the following consonant, sarvam na, pronounced as sarvan na, can be written in the manuscripts as sarvanna. It can then easily be mistaken for sarvatra.
14 When CHARU DEVA SHASTRI's copy of the Vākyapadiya text was collated by the pandits of the Government Oriental Manuscripts Library at Madras with a manuscript in that library (probably A [18] of Rau's (1971) account; see CHARU DEVA SHASTRI 1934: Preface p. 6 for the identification of his source ga 4), he came to know the existence of a reading tävaty asam.. vidan. In his incomplete edition of the Vakyakända (1941 [?]: 63 fn. 2). he noted that reading and rejected it as ungrammatical. According to Pānini 1.3.29, sam + vid, when used intransitively, should take ätmane-pada endings, so that tāvaty asamvidānah would alone be acceptable in the present context. However, if we read tāvat tv instead of tāvaty, the use of the parasmaipada present participle asamvidan will not be ungrammatical, as then sam + vid will have an object associated with it in the form of tāvat. As for tävat tv, CHARU DEVA SHASTRI objects to its acceptance as follows: atra tu-sabdena kim vyāvartyata iti śrutyà noktam. kim ca. tavataḥ punar-arthe 'vyayasya parigrahe yad iti pūrvärdha-gatam kena sambadhyeta. That this objection is not valid can be seen from my translation of the verse. The indeclinable tu has been used in the verse to bring out the contrast between (a) thinking that one has not comprehended only a part and (b) the actuality of not having comprehended the whole, i. e...the contrast between an impression and a fact. Secondly, yad is not used in the verse as a relative pronoun standing for any specific entity, so that it would need a form of tad in the remainder of the verse; it rather stands for an observation (a judgement or a proposition, if you will), and is to be translated by a clause-introducing “that" or "as" (cf. the usage of yad in kārikās 1.87 and 2.22, and English constructions of the type: "That he is well-versed in Sanskrit need not be doubted').

Page #7

Emendation of some verses in Bhartphari's Trikāņdi
69
2.4 The difficulty and the solution are of a considerably simpler nature in the case of 2.136: ekasminn api drøye 'rthe darśanam bhidyate prthak kālāntareņa caiko15 'pi tam paśyaty anyathā punaḥ ll. Here the expression bhidyate prthak contains redundancy; either prthak or bhidyate would be sufficient to convey the intended sense. The author of the Tīkā was probably aware of this, as he shows no explicit recognition of pệthak16. In my opinion, the solution lies in changing bhidyate to vidyate. Such a change, besides being minimal, conforms to what is noticed elsewhere in the Trikāņdi manuscripts (see 2.1 above). No equally elegant or justifiable emendation seems possible for prthak.
2.5 Verse 2.230 (yam' artham āhatur bhinnau pratyayāv eka eva tam 17 | kvacid āha pacantiti dhātus18 tābhyām vinā kvacit It appears in the discussion of non-uniqueness of grammatical description. Bhartshari stresses that the same linguistic reality can be analysed and described differently in different grammars; grammars are merely means or upāyas, and hence should not be mistaken for the reality itself. In this connection he points out the absence of a one-to-one relationship between the elements recognised by a grammar and the components of meaning conveyed. In pacati, two suffixes, sap (= a) and tip (= ti)19, convey agent, voice, tense, etc. However, the same information may be conveyed by only tip (= ti) in yāti etc., and in spite of the absence of both śap (= a) and tip (= ti) in ahan etc. Now, 2.230 expresses this observation quite clearly. It can be easily understood as: "The meaning which is expressed by two different suffixes is in some cases expressed by only one suffix. In some (other) cases, the root is bereft of both the suffixes (and yet conveys all the information that is associated with the two suffixes)'. As will be clear from the omission in my translation, the problematic part of the verse is pacantiti. The form pacanti, containing both śap and tip, is hardly appropriate as an example after the statement, 'The meaning which is expressed by two different suffixes is in some cases expressed by only one suffix.' If pacantiti is to be accommodated, it must be understood in the sense of pacanti ity atra and placed
15 v. I. vaiko 'pi.
16 Țikā: vastuta ekasminn eka-rūpe dréye 'rthe śāstra-vāsanā-bhedad darśanam jñānam asmin bhidyate. eko 'pi ca puruṣaḥ sugata-darsana-samskrtamatir anyathädhyavasyaty artham, kālāntarena vaiseşika-sästra-śravanad anyatheti.
17 v. 1. tām. 18 v. l. yās tu.
19 The reference here is, of course, to the grammatical system of Pāņini.

Page #8

70
ASHOK AKLUJKAR
at the beginning of the sentence: pacanti ity atra yam artham bhinnau pratyayau āhatuħ tam ekaḥ eva (pratyayaḥ) kvacid āha. The Tīkā takes this course20. However, a close look at the Vștti helps one realize that the strained construing suggested by the sīkā need not be followed. As the Vrtti21 says, tathā bhinna-pratyaya-samnidhāno 'rthah kenacid eva pratyayenābhidhiyate. tad yathā odanam paca, atti, atsāḥ iti, one can see that the reading it presupposes in the Kārikā is pacāttiti, although the Kārikā portion even in the Vrtti transcript does not contain that reading. With pacāttiti we need not indulge in any transposing. Furthermore, we get a doubly appropriate example; paca illustrates absence of tipstin, and atti illustrates absence of sap. That pacāttīti is likely to change into pacantiti should be acceptable to anyone knowing the Indian scripts.
2.6 Verse 2.409 (teşām atyanta-nānātvam nānātva-vyavahāriņaḥ akşyādinām 22 iva prāhur eka-jāti-samanvayāt 23 II) is about expressions that are put to different uses without any change in form. When a Vedareciter recites the words of a Veda in daily repetition or review, he does not utter them to convey either their form or content. But when he teaches them to a disciple, they stand for their own forms. Further, when the same words are employed in a ritual they refer to objects that is, they convey meanings other than their own forms. Now the question is: Are these words, which are identical in form, three different entities because of the difference in intention – because of what they do or do not signify – or are they one and the same entity simply assuming three different roles ? Verse 2.409 is written to state the first view. Its first three quarters can be translated thus: "Those who talk of difference say that they (the Veda expressions in the three situations) are utterly different [from each other) just as expressions like akşi [the
20 Tikā: tatra pacantityādau bhinnau pratyayau sap-tinar yam artham kartr-laksanam āhatus tam evaiko 'py aha atti juhotityādau. kvacid dhätur eva tam artham tabhyām vinaivāha ahan vstram indra ityādāv iti.
21 Bhartphari's Vrtti of the Vākya-kānda or the second book of the Vākyapadiya is not available in its entirety either in the Trikāndi editions printed so far or in the manuscripts discovered so far. I propose to publish whatever of it is available in my forthcoming edition of the Trikāņdi. Until that edition is published, my reference to the Vrtti will naturally be based on the manuscript material in my possession.
22 v. l. aśvādinām, aksādinām. 23 v. 1.° samanvayam.

Page #9

Emendation of some verses in Bhartṛhari's Trikāṇḍi
71
neuter noun and aksi the finite verb 24] are.' However, the fourth quarter, eka-jāti-samanvayāt, meaning 'because of the presence of an identical universe', hardly agrees with this statement. In fact, it is contradictory, as it invests the verse with the meaning: 'Vedic words in the three utterances are different from each other because they belong to the same class.' Obviously, the original text must have read something different. I think it can be restored by inserting a visarga after eka-jāti. The last quarter then contains two words instead of one and becomes a separate sentence having the meaning: (There is) one universal [of such entities that are actually different] because of continued presence (i. e. because some elements such as form persist because some features are common).' Thus, whereas the first three quarters state the view of the advocates of non-identity, the fourth explains how the advocates of non-identity account for the common acceptance of homophonous expressions as one and the same entity25.
2.7 In 3.14.297 (ekārthe vartamānābhyām asatā brāhmaṇena ca | yadā 26 jatyantaram bähyam kṣatriyādy apadiéyate ||), the reading yada is as old as Helārāja's time27. It forces the reader to understand at least as much
24 (a) The derivation of aksi as a verbal form is possible in more than one way according to Vṛsabha (p. 136.9-10): akşater aśer vā apratyayena sapi luki ca sati sipi rūpam. ader vă luni ghasl-bhave karmany uttamaikavacane mantratvac cer luki.
(b) The following are the examples given by Bhartṛhari of the expressions that can be derived either as a nominal form (naman) or as a verb (akhyāta): aksi, aśvaḥ, vayuḥ, tena, khyātam, aruṇaḥ, śyāmaḥ, asyāḥ, ācitam [?], samaḥ [?], duhitā and bhāvaḥ gāvaḥ (see fn. 25a below). See Tripādi p. 38.13, p. 138.18-19; Vrtti 1.72; Vṛtti 2.268, Vṛtti 2.315-316, Vṛtti 2.409; cf. Tikā 2.318 (pp. 218-219), which adds ajapayaḥ, yasya, and tasya to the above list; also Parthasarathi-miśra, Nyaya-ratnākara, p. 897, which elucidates aśvaḥ as a verb.
25 (a) Vṛtti: yathaiva gavo ['ksy a]sv[as t]ena vayur [see fn. 24b above] iti tulya-rūpatvān nāmākhyātayor atyanta-nānātve 'py avisiṣṭe buddhi-samanvayad ekatva-vyavahāramātram evam artha-prakaraṇādibhir atyanta-bhede sva-bhava-siddha eva bhedo gamyate.
ha
(b) Tikā: teṣām sabdanām bheda-vādino nānātvam āhuḥ. kevalam ekajati-samanvayas teṣām sādṛśya-rupo 'sti. udaharanam aha akṣādinām [→akṣyādīnām?] iveti. yathākṣādayaḥ [→yathākṣyādayaḥ ?] śabdā bhinnā eva sthitāḥ. sādṛśyat tv abhedenaiva pratiyanta iti.
26 v. 1. yatha.
27 Prakirna-prakāsa: idänim anya padartha-pradhana-paksa-vicārāyātadavakalpate 'nya-padartha-pradhano nañ-samasa iti prakaraṇād avaseyam... yadā tv asat-sāmanya-vṛttir apy ayam nañ sad-viseṣa-nibandhanayā brāhmaṇa-śrutyā sāmānya-viścșa-bhāvenaikadhikaranyena sambandham anubhaya tat-sambandha-präpitärthātikayam bāhyam padāntaropādānam kṇa

Page #10

72
ASHOK AKLUJKAR
as tada nañ-samāsaḥ (cf. Helārāja p. 287.11, 18) if the verse is to be syntactically complete. Although it is possible that Bhartrhari could have intended to indicate such a completion in his planned but unwritten Vștti28, the probability of this being the case is negligible. Firstly, in the extant Vrtti (i. e. in the first two books), we find instances of a proper beginning being supplied to a verse (AKLUJKAR 1972: 190-193), but not a single instance in which a necessary concluding portion is supplied. In other words, the syntactic supplementation made by the Vstti always pertains to the subject part, never to the predicate part, of the sentence contained in a Kārikā. Secondly, 3.14.297, as Helārāja correctly states, introduces the anya-padārtha-prādhānya-paksa concerning the nañ-samāsa. In that respect, it is similar to 3.14.260 and 3.14.306, which respectively introduce the uttara-padārtha-prādhānya-paksa and the pūrva-padārtha-prādhānya-pakşa. These latter verses state how a nañsamāsa conveys its meaning or how the meaning understood from a nañ-samāsa can be characterized. They do not have the form: 'If the semantic situation happens to be such, then a pūrva-padārtha-pradhāna nan-samāsa (or an uttara-padārtha-pradhāna nan-samāsa) comes into existence'. It would be strange to have that form of expression only for 3.14.297. Thus, the addition of tadā nan-samāsaḥ or tadā anya-padārthapradhānah nan-samāsaḥ cannot be justified. Consequently, the reading yadā which occasions that addition is suspect. I think yada has come about through a corruption of yad vā (mistaking dvä for dā is not improbable). If one reads yad vā ('or', 'to come to another alternative') not only will the syntactic incompleteness be removed, but the verse will also contain the necessary indication to the effect that the discussion of the uttara-padārtha-prādhānya-paksa is over with 3.14.296 and that 297 takes up the second paksa for consideration.
2.8 The last passage I wish to discuss here is 3.14.605: ekasyāpi 29 pratiyeta 30 bhinnā pratikytiḥ31 saha kāśya pasyeti 32 tenāyaṁ pratyekam avatişthate 11. The context is that of the expressions garuda iveme hastinaḥ
triyady anupatati, na tavaty eva viérāmyati tadānya-padārtha-pradhāno nansamāsaḥ. tathā hy avidyamāno brāhmano brāhmaṇa-vyaktir asraya-bhūtā yasya ksatriyasya ksatriya-jäter ity asyā anya-padārtha-bhūtäyä varti-padārthā. bhyām abhäva-brahmanabhyām avaccheda iti jātyantaram bahyam bhavati.
28 I shall discuss the extent of the Väkyapadiya-vịtti in a separate paper.
29 v. 1. ekasyätra. 30 v. l. pramiyeta. 31 v. 1. prakertibhih, pravikrtih. 32 v. l. kasyao.

Page #11

Emendation of some verses in Bhartṛhari's Trikāṇḍi
and kasyapā ivemäḥ pratikṛtayaḥ. The first is likely to be employed in speaking of the arrangement of an army for a battle. If the elephantriders in the army are arranged in an eaglelike formation, an observer may say, "These elephants are like an eagle'. But suppose the observer is viewing images or statues of the members of the Kasyapa family. Will it be proper if he says käsyapa ivamāḥ pratikṛtayaḥ instead of kāśyapāḥ ivemāḥ pratikṛtayah? If the singular garuḍah is acceptable in the other expression, why not use a singular here as well? Bharthari points out that in the first expression a group of elephants is compared with an eagle on account of the specific shape of the group. However, images are not compared through some specific feature with the group of Kasyapas in the second expression. Therefore, the only proper usage will be the one with the plural kadyapäḥ. Helārāja explains this point accurately, but, to derive it from the verse, he takes saha in the unattested sense tulya 'equal' and construes the verse as ekasya api kāśyapasya bhinnä pratikṛtiḥ saha (= tulya) pratiyeta33. This is clearly strained. I would like to suggest that saha and kayapasyeti should be replaced by sadrk and käsyapas ceti. The k of sadrk could have been dropped in dictation or uninterrupted writing, as another follows. Then the remaining sady, being an ungrammatical expression, could have been easily mistaken for saha. The emendation suggested for kalyapasyeti is also transcriptionally probable. Thus, spurred by the awkwardness of Helārāja's explanation, we can guess that the first sentence in the verse is to be read as kasyapab iti ca35 ekasya api bhinnä pratikṛtih sadrk pratiyeta, and to be understood as: 'And, in the case of [the singular
73
33 Prakirna-prakãán: yadi kalyapa ivemaḥ pratikrtayaḥ käéyapa ity eka-vacanäntam upamanam prayujyate tada garuḍasyeva hastino vyuhena tulyaḥ [...] kāśyapasya nānā-bhūtā pratikṛtiḥ kenacit samniveśādinā dharmeņa saha tulya vijñāyeta. na tu pratyekam upamanam gamyeta. sahasabdas tulyarthaḥ. tad-yoge kasyapasyeti şaşṭhi. tasmat pratyekam upamanabhāva-pratipattyartham atra kasyapa-sabdaḥ pratyupameyam bhedenaivāva
tisthate.
34 The iti in kasyapasyeti cannot possibly be indicative of quotation. A construal of the form 'kasyapasya' iti ekasya api bhinna pratikṛtiḥ saha/ sadrk pratiyeta is out of question, since the form used in the intended usage (cf. Helārāja p. 420.13) is the nominative kasyapaḥ, not the genitive kāśyapasya. Therefore, if one reads kāśyapasya, one must understand iti in the sense 'therefore' and construe the verse quarters 'abc' as ekasya api kāśyapasya bhinna pratikṛtiḥ saha/sadrk pratiyeta iti. However, such a construing not only throws together ekasya and kasyapasya that are far removed from each other, but also makes the following tena 'therefore' redundant.
35 Insertion of ca before iti (for the sake of the metre) is not strange, as it is noticed in other verses.

Page #12
74 ASHOK AKLUJKAR expression] kasyapah, the varied images would be comprehended as similar to even one (Kasyapa) [which is a self-contradictory, undesirable proposition]'. BIBLIOGRAPHY AKLUJKAR, ASHOK. 1972. "The authorship of the Vakyapadiya-vitti." Wiener Zeitschrift fur die Kunde Sudasiens, 16:181--198. 1974. "The authorship of the Vakya-kanda-tika. Charu Deva Shastri Felicitation Volume, pp. 165--188. New Delhi. Bhartphari. Trikanai: (a) Kanda 1: (Ed.) SUBRAMANIA IYER, K. A. Vakya padiya of Bhartihari with the Vstti and the Paddhati of Vrsabhadeva. Deccan College Monograph Series, no. 32. Poona: Deccan College. 1966. (b) Kanda 2: (Ed.) MANAVALLI, GANGADHARA SASTRI. Vakyapadiyam... Sri-Bhartphari-viracitam Sri-Punyaraja-ksta-prakasakhya-tika-yutam. Benares Sanskrit Series no. 11, 19, 24. Benares: Braj B. Das & Co. 1880. (c) Kanda 3: (Ed.) SUBRAMANIA IYER, K. A. Vakyapadiya with the commentary of Helaraja, Deccan College Monograph Series no. 21. Poona: Deccan College. 1963. Vakyapadiya with the Prakirnaka prakasa of Helaraja. Poona Deccan College. 1973. See fn. 1 above. . Bhoja. Sengara-prakasa. (Ed.) JOSYER, G. L. Mysore: International Aca demy of Sanskrit Culture. 4 volumes. 1960--1974. CHARU DEVA SHASTRI. 1934. (Ed.) Vakyapadiyam prathamam kandam. Lahore: Ramlal Kapoor Trust. - 1941 [?] (Ed.) Vakyapadiyam Bhartsharyupajna-vitti-sanatham Punya raja-tika-samyutam dvitiyam kandam. Lahore: Ramalal Kapoor Trust. [Incomplete] Helaraja. See Bhartphari (c). Parthasarathi-misra. Nyaya-ratnakara commentary on Kumarila-bhatta's Tantravarttika. (Ed.) TAILANGA, RAMA-SASTRI MANAVALLI. Chau khamba Samsksta Granthamala No. 3. Kasi. 1898. Prakirna-prakasa. See Bhartshari (c). RAU, WILHELM. 1971. Die handschriftliche Uberlieferung des Vakyapadiya und seiner Kommentare. Abhandlungen der Marburger Gesellschaft, no. 1. Munchen: Wilhelm Fink Verlag. - 1977. (Ed.) Bhartpharis Vakyapadiya (mula-karikas). Monograph series of the Deutsche Morgenlandische Gesellschaft, no. XLII, 4. Wies baden: Franz Steiner Verlag. SHARMA, RAGHUNATHA. 1968. (Ed., comm.) Vakyapadiyam Part II (Vakya kandam). Sarasvati Bhavana Grantha-mala, no. 91. Varanasi: Varanaseya Sanskrit Vishvavidyalaya. Tika. Usually ascribed to Punyaraja. Available in MANAVALLI's (see under Bhartshari above) and RAGHUNATHA SHARMA's editions. See AKLUJKAR 1974 for the view that Helaraja is probably the real author. Trikandi. See Bhartrhari. Tripadi. (Eds.) ABHYANKAR, K. V. and LIMAYE, V. P. Mahabhasya-dipika of Bhartphari. Post-graduate and Research Department Series, no. 8. Poona: Bhandarkar Oriental Research Institute. 1967--1970. Visabha. See Bhartphari (a). Vstti. See Bhartihari (a) and fn. 21 above.