Through Kātyāyana’s vārttikas, we know that he interprets para in 1.4.2 vipratiṣedhe paraṁ kāryam as ‘the rule which comes later in the Aṣṭādhyāyī’s serial order’. For example, consider 3.1.67 sārvadhātuke yak which teaches that affix yaK occurs after a verbal root when a sārvadhātuka affix which denotes bhāva or karman follows. Consider vt. 47 on this rule: vipratiṣedhād dhi śapo balīyastvam ‘Given the vipratiṣedha8 [between yaK (cf. 3.1.67 sārvadhātuke yak) and ŚaP (cf. 3.1.68 kartari śap)], ŚaP is more powerful [and wins, because it is para i.e., taught later in the serial order of the Aṣṭādhyāyī].’
Note that this vārttika makes an incorrect statement. There is no conflict at all here: yaK is added to verbal roots followed by sārvadhātuka affixes denoting bhāva ‘action’ or karman ‘object’ whereas ŚaP is added when the sārvadhātuka affix denotes kartr̥ ‘agent’. In fact, we come across many such errors in Kātyāyana’s vārttikas.
But I think that it is unwarranted to look for ‘correct’ statements in the vārttikas. This is because, in my opinion, Kātyāyana’s vārttikas are often a medium for him to share all kinds of thoughts with fellow grammarians – not just the ‘correct’ ones. Very often, we find him use na vā ‘or rather not’ and ca ‘and’ in a series of consecutive vārttikas to discuss alternative or even contradicting possibilities and explanations. Let me give an example relevant to the topic of rule conflict. Consider vts. 3, 4 and 5 on 7.1.6 śīṅo ruṭ9 (Mbh III.243.12-21).
Vt. 3 jhādeśād āḍ leṭi
‘[It must be stated that, contrary to 1.4.2, the introduction of] āṬ, [which is taught by the pūrva rule 3.4.94 leṭo’ḍāṭau10 wins against] the substitution of jh [which is taught by the para rule 7.1.5 ātmanepadeṣv anataḥ11].’
204
Vt. 4 na vā nityatvād āṭaḥ
‘Or rather [this does] not [need to be stated] because [the rule teaching] āṬ is nitya [and thus defeats the other rule which is anitya].’
Vt. 5 antaraṅgalakṣaṇatvāc ca
‘And [also] because [the rule teaching] āṬ is antaraṅga [and thus defeats the other rule which is bahiraṅga].’
This style of discussing multiple possibilities without striving to always be correct, is very much akin to Patañjali’s style, which also involves a discussion about the pros and cons of various perspectives. In both Kātyāyana’s and Patañjali’s work, we find no rigidity or urgency to establish the truth. Instead, their work is characterized by curiosity and a willingness to critically examine a motley of ideas.
Coming back to the topic of para, suffice it to say that regardless of the correctness of its contents, vt. 4 on 3.1.67, which I have discussed above, buttresses the proposition that Kātyāyana interpreted para as ‘the rule which comes later in the serial order of the Aṣṭādhyāyī’. And while this interpretation of para taught by Kātyāyana – alongside tools like nitya, antaraṅga etc. discussed by him – has been fully endorsed and internalized by the later tradition, most traditional and modern scholars have almost entirely overlooked a very important idea about paratva that we find in a vārttika on 6.1.158 anudāttaṁ padam
ekavarjam.
6.1.158 teaches that a pada is entirely low-pitched (anudātta) with the exception of one syllable. But how should we decide which syllable is not low-pitched? Is it a syllable of the prakr̥ti ‘base’ or a syllable of the pratyaya ‘affix’? After discussing this topic in multiple vārttikas on this rule, Kātyāyana says, in vt. 1212: śāstraparavipratiṣedhāniyamād vā śabdaparavipratiṣedhāt siddham ‘[in the event of vipratiṣedha between two operations] because it has not been [explicitly] mandated that paratva of rules [alone should be used to resolve] vipratiṣedha, alternatively paratva of sounds [may also be used to] accomplish [the task of resolving] vipratiṣedha’.13 In other words, here, Kātyāyana suggests that alongside inferring that the rule that is para i.e., that comes later in the serial order of the Aṣṭādhyāyī
205
wins, we may also infer that the operation that is applicable to the para i.e., RHS sound or group of sounds wins.
This shows that Kātyāyana was either exposed to or himself thought about the possibility that para in 1.4.2 could stand for the RHS operation. If he had chosen to further develop this line of thought, this idea could potentially have reached its logical conclusion, namely the correct interpretation of para in 1.4.2. One could argue that, by choosing to focus on and subsequently by accepting the wrong interpretation from amongst the two possible interpretations of para discussed in the aforementioned vārttika, Kātyāyana completely changed the developmental trajectory of the Pāṇinian tradition. Kātyāyana’s successors too failed to recognize the sheer potential of this vārttika, and thus the key to the Aṣṭādhyāyī’s algorithm remained before everyone’s eyes and yet hidden from everyone’s mind.
One key repercussion of Kātyāyana’s belief that para in 1.4.2 stands for ‘the rule that comes later in the Aṣṭādhyāyī’s serial order’ must have been that he likely got numerous incorrect forms at the end of derivations where he solved SSRI using his interpretation of 1.4.2. Perhaps it is to avoid these undesirable outcomes - wherever possible - that he decided to reduce the jurisdiction of 1.4.2. For example, in vt. 1 on 1.4.2, he defines vipratiṣedha in a way that allows him to exclude anavakāśa-sāvakāśa pairs from the jurisdiction of 1.4.2: dvau prasaṅgāv anyārthāv ekasmin sa vipratiṣedhaḥ (1)14 ‘[When] two rules [which are] applicable elsewhere [become applicable] to the same place, this [is called] vipratiṣedha’. Thus, an SSRI between two sāvakāśa rules (i.e., rules which are applicable elsewhere) is called vipratiṣedha. We know that an SSRI can be either a conflict scenario or a non-conflict one. But as I have said in previous chapters, Kātyāyana is, for the most part, interested in conflict. Thus, I will take the liberty, for the sake of this chapter, to translate the traditional interpretation of vipratiṣedha as ‘conflict between sāvakāśa rules’.
In vt. 2 on 1.4.2, he says: ekasmin yugapat asaṁbhavāt pūrvaparaprāpter ubhayaprasaṅgaḥ “[Given the] impossibility [of] co-application at one [i.e., the same step, there arises] the undesirable scenario of both pūrva and para being applicable.” In vt. 5, Kātyāyana says: apratipattir vobhayos tulyabalatvāt ‘Or [maybe this results in] the failure of both [rules] to apply because of [their] equal strength’. In vt. 6 he says: tatra pratipattyartham etad vacanam ‘So, this [sūtra] has been formulated in order to instruct us about this [i.e., the decision regarding which rule should apply]’. From vts. 1, 2, 5 and 6 on 1.4.2, we can conclude that,
206
according to Kātyāyana, the conflict between two sāvakāśa rules is called vipratiṣedha, and that these two rules are treated as tulyabala ‘of equal strength’. Note that this is the only occasion on which Kātyāyana uses the term tulyabala. Patañjali too uses the word tulyabala only once – when commenting on vt. 5 on 1.4.2.15
Kātyāyana has composed several vārttikas discussing terms like nitya, anitya, antaraṅga, bahiraṅga, apavāda, utsarga, anavakāśa and sāvakāśa. This indicates that he was familiar with or himself constructed these concepts and established relationships between nitya and anitya rules, between antaraṅga and bahiraṅga rules, between apavāda and utsarga rules,
and between anavakāśa and sāvakāśa rules. While Patañjali does not always agree with Kātyāyana, he has embraced all these concepts wholeheartedly in his commentary. We get no evidence of Kātyāyana connecting these concepts directly with tulyabalatva, and only one piece of evidence of him establishing a direct link between one of these tools and vipratiṣedha, which is as follows. On 6.1.135 suṭ kāt pūrvaḥ ‘The augment sUṬ is added before k16’, Kātyāyana says, in vt. 717: avipratiṣedho vā bahiraṅgalakṣaṇatvāt ‘[Alternatively, one can argue that this is] not a case of vipratiṣedha because [sUṬ is] bahiraṅga’. This shows that he excludes antaraṅga-bahiraṅga pairs from the domain of 1.4.2.
But even after inventing tools like nitya, antaraṅga, apavāda and anavakāśa, Kātyāyana was unable to resolve certain conflicts, especially those involving DOI mutual blocking, using any of the aforementioned tools. On many occasions, solving such conflicts using 1.4.2 too led to an incorrect answer at the end of the derivation.18 Thus, he wrote the ‘pūrvavipratiṣiddha’
vārttikas. By using the expression ‘pūrvavipratiṣiddha’, Kātyāyana points out that instead of the para sūtra, which should win as per his interpretation of 1.4.2 vipratiṣedhe paraṁ kāryam, it is the pūrva sūtra which emerges victorious. We have already looked at some such vārttikas in chapter 2, so I will simply mention one of them here. On 7.1.96 striyāṁ ca, vt. 1019 reads: guṇavr̥ddhyauttvatr̥jvadbhāvebhyo num pūrvavipratiṣiddham ‘In case of vipratiṣedha, the pūrva sūtra, which teaches the insertion of the augment nUM, takes
207
precedence over para sūtras which teach (i) guṇa, (ii) vr̥ddhi, (iii) auttva, (iv) tr̥jvadbhāva’. By writing this and other pūrvavipratiṣiddha vārttikas, Kātyāyana draws attention to the perceived failures of / loopholes in / exceptions to the rule 1.4.2.
Commenting on most pūrvavipratiṣiddha vārttikas, Patañjali says that they are not required at all. He gives various reasons for this, of which the following one is used by him on multiple occasions. On vt. 10 on 7.1.96 stated above, he says: na vaktavyaḥ. iṣṭavācī paraśabdaḥ. vipratiṣedhe paraṁ yad iṣṭaṁ tad bhavati ‘[This] should not be said. The word para means desirable. In [the event of] vipratiṣedha, the para i.e., desirable [rule] applies.’ It is evident that in this context Patañjali tries to defend 1.4.2 against Kātyāyana’s criticism. In fact, this is anything but an isolated instance: scholars like Goldstücker (1861: 119-121) and Weber (1872: 297-298) were amongst the earliest modern scholars to argue that Kātyāyana was severely critical of Pāṇini’s sūtras, and that Patañjali invested significant effort in countering such negative remarks. While many scholars, starting with Kielhorn, have presented rebuttals to this, even Kielhorn (1876: 50) cannot deny “that Patañjali has refuted some of the (i.e., Kātyāyana’s)20 objections, that he has rejected some of the additional rules of Kātyāyana.”
Coming back to vt. 10 on 7.1.96, I would argue that by hurrying to dismiss Kātyāyana’s pūrvavipratiṣiddha vārttikas using a rather feeble argument, namely that para means iṣṭa, Patañjali missed the opportunity to discover the truth of 1.4.2. Instead, if he had accepted Kātyāyana’s statement as valid and had pondered over the cause of this phenomenon, he could possibly have realized that Kātyāyana’s interpretation of para itself was incorrect, and that it was this misinterpretation which had led him to write the pūrvavipratiṣiddha vārttikas. This would certainly have been a far superior defence of Pāṇini’s rule 1.4.2 against Kātyāyana’s criticism than the one mounted by Patañjali.
After the composition of the Mahābhāṣya, ideas about the terms vipratiṣedha, para, tulyabala, and the various tools of conflict resolution discussed above began to take more concrete shape. Direct links and relationships between these concepts came to be established. For example, on 1.4.2, the Kāśikā, which was written in the 7th century AD, says:
yatra dvau prasaṅgāv anyārthāv ekasmin yugapat prāpnutaḥ sa tulyabalavirodho vipratiṣedhaḥ. tasmin vipratiṣedhe paraṁ kāryaṁ bhavati.
208
utsargāpavādanityānityāntaraṅgabahiraṅgeṣu tulyabalatā nāstīti nāyam asya yogasya viṣayaḥ, balavataiva tatra bhavitavyam.
‘When two operations which can be applied at other sites become simultaneously applicable at one [and the same site], this is called a conflict of equal strength or vipratiṣedha. In the event of vipratiṣedha, the rule that comes later [in the serial order of the Aṣṭādhyāyī] prevails. A general rule (utsarga) and its exception (apavāda), or a nitya rule and an anitya rule, or an antaraṅga and a bahiraṅga rule, are not rules of equal strength. These pairs do not fall under the jurisdiction of this rule. In these cases, the stronger rule wins.’
Notice that, unlike Kātyāyana and Patañjali, the authors of the Kāśikā explicitly exclude nitya-anitya, antaraṅga-bahiraṅga and apavāda-utsarga pairs from the ambit of vipratiṣedha by calling them ‘not tulyabala’. Thereafter, in both Pāṇinian and non-Pāṇinian paribhāṣā literature, we find multiple versions of the same paribhāṣā which compares the ‘strengths’ of the tools mentioned above. The earliest Pāṇinian paribhāṣā treatise to include it is the Paribhāṣāpāṭha of Puruṣottamadeva written in the 12th century. It reads: pūrvaparanityāntaraṅgāpavādānām uttarottaraṁ balīyaḥ (Pbh. 39). ‘Of [these five kinds of rules, - viz.] a preceding [rule], a subsequent [rule], a nitya [rule], an antaraṅga [rule], and an apavāda [rule], - each following [rule] possesses greater force [than any one of, or all, the rules which are mentioned before it].’21
In sum, the relationships between tulyabala, vipratiṣedha, nitya, antaraṅga, para, apavāda etc. were fully and concretely established by the twelfth century. Alongside the paribhāṣās teaching these tools, dozens of paribhāṣās teaching exceptions to these tools were also written by the paribhāṣākāras. On this account, given its unwieldy and complicated nature, the traditional solution completely fails the Occam’s razor test. Additionally, the flexibility of ideas, free thinking, willingness to consider a wide variety of possibilities and alternatives, which, as stated earlier, are so characteristic of the early tradition i.e., Kātyāyana’s and Patañjali’s work, came to be replaced by a willing acceptance of rigid, ossified, established and widely-accepted ‘facts’ and ‘truths’ in the later tradition – in particular, in paribhāṣā
literature. It is noteworthy that many of these paribhāṣās are anitya ‘not always applicable’ by the tradition’s own admission!
209
Here, one may ask: why do the Kāśikā and the paribhāṣā texts not question the correctness of Kātyāyana’s interpretation of the term para in 1.4.2? I think the first broad reason is that, along with Pāṇini, who composed the foundational treatise of the tradition, Kātyāyana and Patañjali too came to be worshipped in the tradition, which might have made it almost unthinkable for subsequent scholars to disagree with Kātyāyana or Patañjali over such fundamental aspects of the grammar as the meaning of para in 1.4.222. It must be noted that even though the Kāśikā does present an alternative viewpoint to that of the Mahābhāṣya on many occasions, it completely embraces Patañjali’s ideas on this subject. Secondly, even amongst the three munis, Patañjali’s word superseded Kātyāyana’s and Kātyāyana’s word superseded Pāṇini’s, right from the time of Kaiyaṭa, who famously stated: yathottaraṁ hi munitrayasya prāmāṇyam23 ‘Among the three munis, the authority of later muni supersedes that of his predecessor(s)’24. Thus, Patañjali became the most important person in the tradition, surpassing Pāṇini himself, whose work he had set out to expound on. So, hypothetically speaking, even if a traditional scholar had discovered that Patañjali had misinterpreted para in 1.4.2, he would have preferred Patañjali’s interpretation to Pāṇini’s in all likelihood!
One would have expected the tradition to start paying ever closer attention to the topic of rule conflict with the writing of the Kaumudī texts, the main goal of which was to teach students how to perform derivations. To achieve this goal, the Kaumudī texts took the radical decision to reorder the rules of the Aṣṭādhyāyī so that a rule would be taught in the Kaumudī only when it applied at some step in a certain derivation. However, unfortunately, these texts did not challenge the existing interpretation of para in 1.4.2 and, like previous texts, performed derivations using the traditional tools for conflict resolution. In fact, not only did the Kaumudī texts fail to discover the correct meaning of 1.4.2, but they also unwittingly ensured that coming generations would not decipher the same.
They did this by shifting the focus of the tradition from the comprehensive functioning of the Pāṇinian machine to the many individual products of the machine, namely, individual derivations of various forms. Over time, students of the Kaumudī got so familiar with these derivations that now, they do not have to and, consequently, do not, stop at most steps of the
210derivation to ask themselves: which rules are applicable at this step? Which of these rules should I apply? And why? And if pupils do apply conflict resolution tools of their own accord and end up getting the wrong form, they are not encouraged by their teachers to ask why. Instead, they are advised to consult the Kaumudī texts to ‘correct’ themselves i.e., to memorize the explanation offered by their authors.
This chain of accepting what previous scholars have said was finally broken by many modern Indologists, including Houben (2003), who asked if Pāṇini’s grammar is meant to function like a machine at all25, and Bronkhorst (2004) who questioned the ‘linearity’ of Pāṇinian derivations.26 Others have tried to make changes in some parts of the traditional conflict resolution mechanism. For example, multiple scholars, starting with Faddegon (1936), have advocated restricting the jurisdiction of 1.4.2 to 1.4.2-2.2.38. Cardona (1970: 57-58) has proposed limited blocking, which essentially deals with more complex cases of SOI, even though he does not state this explicitly.27
Joshi and Kiparsky interpret vipratiṣedha as ‘mutual blocking’ and state that “for…so-called vipratiṣedha, no general solution has been found”28 by them.29 However, they do propose a solution for those cases which involve unidirectional blocking, namely the siddha principle. What it essentially does is resort to the nitya principle30 to solve not only these cases which the tradition solves using nityatva, but also those which it solves using antaraṅgatva.
Bronkhorst (1984: 310-313) and Cardona (1999: 154-161) have correctly criticized the reasoning behind this principle.31
211
Even though none of these scholars have been able to offer a radically different interpretation of 1.4.2, their willingness to ask questions, to propose new ideas and to challenge the traditional method of conflict resolution inspired me to do the same, eventually leading me to the interpretation of 1.4.2 I have presented in this thesis.
In the following section, I will discuss how my findings can help us better understand other aspects of the Aṣṭādhyāyī and linguistics in the future.