Boundaries and Domains: Understanding Optionality in Buddhappiya's Rūpasiddhi

Aleix Ruiz-Falqués¹

ABSTRACT

Since the publication of Paul Kiparsky’s Pāṇini as a Variationist (1979), the discussion on optionality in Indian vyākaraṇa has been largely polarised into two views : Kiparsky’s thesis defends the idea that Pāṇini used different words to indicate different degrees of optionality ; others deny Kiparsky’s claim and maintain that the exact scope of grammatical options can only be determined by the context. Whereas Sanskrit grammarians do not substantiate Kiparsky’s claims, the Pali vyākaraṇa tradition of the Kaccāyana school recognises two different types of options. In this paper I will focus on the treatment of option markers in Buddhappiya’s Rūpasiddhi (South India, twelfth century). The Rūpasiddhi is based on the Kaccāyana grammar, but it takes an innovative approach towards the general structure of the rules. It also alters the mechanisms of optionality in a way that allows for higher accuracy in the process of word derivation. As a step forward in our understanding of the rich Pali grammatical tradition, this article provides a definition and classification of markers vā, kvaci, navā, and vibhāsā in the Rūpasiddhi. It also shows that the understanding of option markers in Kaccāyana classical scholarship differs significantly from Kiparsky’s understanding of linguistic variation.

1. INTRODUCTION

1.1. In the Pali language, one single word can adopt various correct forms. For instance, the ablative singular of the word buddha may be buddhā, buddhamhā, buddhasmā, or buddhato, and the third person singular optative can be care or careyya. In order to account for such multiplicity of derivations, Pali grammarians use a set of markers inherited from the Sanskrit tradition. The most prominent among them are option markers, such as vā or kvaci. Option markers delimit the domain of a rule ; they draw the boundary line between what is general and what is exceptional, thus distinguishing what is allowed in terms of derivation from what is not.² Therefore, understanding how option markers work is essential for the correct interpretation of classical Pali grammars.

1.2. The study of option markers has received considerable attention in Sanskrit scholarship, especially after Paul Kiparsky’s publication of Pāṇini as a Variationist (1979). In this monograph Kiparsky challenges the received opinion concerning the function of option markers in Pāṇini’s Aṣṭādhyāyī :

To indicate that a rule is to be applied optionally, Pāṇini uses 106 times vā, 112 times vibhāṣā, and 93 times anyatarasyām. Why this variety, when one word would do ? This question has a surprising answer, which has remained unknown for over two thousand years. Not even the oldest commentators Kātyāyana and Patañjali have any inkling of it.³

Kiparsky’s “surprising” answer is that these three technical terms express three different levels of optionality : vā “or rather”, “preferably” ; vibhāṣā “or rather not”, “preferably not” ; anyatarasyām “either way”.⁴ Kiparsky understands that preferred options reflect a wider linguistic usage, and less preferred options reflect marginal, or perhaps dialectal, usage. Among Sanskrit vyākaraṇa specialists there is a divide concerning the validity of Kiparsky’s thesis.⁵ But regardless of our judgement, the importance of his question seems undeniable. For economy of language is one of the cardinal principles of vyākaraṇa, and if all option markers mean the same : why use many and not one ?

1.3. The same question may be asked about option markers in Pali classical⁶ grammar, where virtually the same technical terms are used. So far, there has been little research in this area. Émile Senart, who published the first European integral edition and translation of the Kaccāyana grammar, openly condemns the manner in which Kacc uses option markers. The same goes for R.O. Franke.⁷ More constructively, Helmer Smith analyses option markers in the Saddanīti in a useful sketch in his Cospectus Terminorum (§ 7.3.2.3), although no further discussion is given. Ole Pind observes that, in the Kaccāyana and the Kaccāyana-vutti, the terms vā and kvaci “are apparently used interchangeably, thus making the exact scope of a rule difficult to interpret,”⁸ but he also notes that the Rūpasiddhi applies some corrections concerning optionalily of kvaci and na vā.⁹ Mahesh Deokar includes optionality in his thorough comparative study Technical Terms and Technique of the Pali and the Sanskrit Grammars (2008). Deokar points out that the minute distinctions that are observed in Sanskrit grammars “are not observed so rigorously by the Pali grammarians”.¹⁰ Some translators of the Kaccāyana do not reflect any difference of meaning between different option markers ;¹¹ others are more careful and try to distinguish them, e.g. Nandisena’s translation.¹² Among all translations and studies known to me, only those by Thiab Malai and Nandisena acknowledge the traditional system of reading option markers.¹³ The intricacies of this system, however, remain to be properly elucidated. The present article is intended as a contribution to the vastly unexplored field of Pāli byākaraṇa studies. It focuses on one of the major grammars of the Kaccāyana school, namely the Rūpasiddhi, composed in South India by Coḷa Buddhappiya¹⁴ around the twelfth century CE.¹⁵

1.4. Buddhappiya’s Rūpasiddhi is a recast of the Kaccāyana (Kacc) grammar.¹⁶ It uses practically all the suttas of Kacc, but their sequence is dramatically altered. Furthermore, the old Kaccāyana-vutti (Kacc-v) is replaced by Buddhappiya’s own commentary. In the introductory stanzas Buddhappiya briefly justifies his project. He states that he has composed a work that is byattaṃ “clear” and sukaṇḍaṃ “well-arranged”. According to the commentary, “clear (byattaṃ) means that it is particularly clear, easy to see, because of its clarity both in the delimitation of the domain of the rules and the derivation of what is prescribed by the rules”.¹⁷ This is a direct reference to Buddhappiya’s innovative system, which represents a significant improvement in the use of option markers. As for the adjective sukaṇḍaṃ it refers to a topic-wise arrangement that makes the Kacc easier to grasp, much like kaumudī grammars did with Pāṇini.¹⁸ Despite the obvious advantages of the new topic-wise arrangement, the alteration of the original sequence of rules inevitably disrupts the mechanisms of anuvutti (“recurrence”) that control option markers. Buddhappiya solves this problem by explicitly indicating which words, including option markers, recur in a sutta. And more importantly, he incorporates the concept vavatthitavibhāsā (“restricted option”) to the repertoire of option markers. In doing so, he achieves higher accuracy than his predecessor Vimalabuddhi when it comes to delimiting the domains and boundaries of options. In the following sections I will offer a preliminary examination of how this system of optionality works.

2. OPTIONALITY IN BUDDHAPPIYA’S RŪPASIDDHI

2.1. Classical Pali grammarians distinguish two main levels of optionality : one for an open “alternative” (vikappa) and one for exceptions to a larger rule. The general principle to interpret optional markers in the Kaccāyana Pali grammar is laid out in Vimalabuddhi’s Mukhamattadīpanī (Mmd), otherwise known as Nyāsa, composed around the tenth century CE).¹⁹ According to Vimalabuddhi, the option markers kvaci and navā (or na vā) have the same meaning ; they generally express the correctness of one form, dismissing the other. In contrast, the terms vā and vibhāsā, which also have the same meaning,²⁰ generally express an open option (vikappa) in which it is possible to derive two forms.²¹ The main issue at stake here is whether a word is correct in terms of derivation or not. In Pali, option markers have nothing to do with frequency of usage in the literature, although they are supposed to reflect forms that are attested in the literature (āgama). They are only concerned with grammatical correctness, what Sanskrit grammarians call sādhutva.²² Aside from the general principle that has been described, option markers do not work exactly in the same manner in every Pali grammatical treatise. But exactly how they differ, and why grammarians modify the system, is something that requires further research.

2.2. As a grammar belonging to the Kaccāyana school,²³ the Rūpasiddhi uses virtually the same set of option markers : vā 154 times, kvaci 54 times, navā 6 times, vibhāsā 4 times (5 if we include saha vibhāsā in the vutti of Rūp § 351).²⁴ To this list we may also include other words that control optionality, e.g. ca (= vikappa Rūp § 117 = vā ; = aniyamattha = vā § 664), ṭhāne (= kvaci Rūp §§ 40, 42), tu (= kvaci Rūp §§ 125, 226), vikappena (Rūp § 117), niccaṃ “always” (Rūp §§ 135, 195, 196) and yebhuyyena “generally” (Rūp § 36). Keeping with the structure of Kacc and Kacc-v, the proper option markers remain vā, kvaci, navā, and vibhāsā.

2.3. Buddhappiya accepts Vimalabuddhi’s two levels of optionality, but he refines the system in the following manner :

vā “or” has two meanings :

vikappa “open option”
vavatthitavibhāsā²⁵ “restricted option” (used 42 times),²⁶ which involves three sub-domains :
- (a) nicca “mandatory” [Ø option]
- (b) anicca “not mandatory” [= vikappa]
- (c) asanta “inapplicable” [≈ kvaci/navā see below 4.6.]

kvaci “in some places” indicates exceptions
navā “or not” is glossed as kvaci (Rūp § 21)
vibhāsā “optionally” is glossed as vā (Rūp-ṭ ad Rūp § 360)

2.4. There are some exceptions : vā = kvaci in Rūp § 676 and probably in § 207 ; vā = sampiṇḍana in Rūp § 595 (alternatively interpreted as vikappa) and Rūp § 187, glosses by Rūp-ṭ as pakativikappana ; vā = katthacivikappanattha Rūp § 190 ; kvaci = vavatthitavibhāsā Rūp § 461. The relatively rare occurrence of anomalies shows that the general theory is quite consistently observed. In the following sections I am going to briefly examine of how these markers function.

3. VĀ AND VIBHĀSĀ

3.1. The words vā and vibhāsā express vikappa (Skt. vikalpa) “open option”. Whether these terms are explicitly glossed as vikappena or not, they indicate an option in which the alternation is unrestricted. For instance, in Rūp § 30 jhalānaṃ iyuvā sare vā “before a vowel, masc. neut. endings i/ī and u/ū may optionally be replaced with iy and uv respectively”. The word vā is not glossed, but a series of free alternatives are listed as examples, e.g. bhikkhuvāsane bhikkhu āsane vā (“[one can write/say] bhikkhvāsane or bhikkhu āsane [indistinctly]”). Similarly, in Rūp § 116 nāmhi raññā vā, we are given two possibilities for each word, e.g. rājena/raññā. Accordingly, we can form the sentences raññā kataṃ or rājena kataṃ.²⁷

The same type of option is expressed by the marker vibhāsā. For example, Rūp § 360 vibhāsā rukkha-tiṇa-pasu-dhana-dhañña-janapadādīnañ ca allows couplets of coordinative compounds that refer, for instance, to species of trees : assatthakapitthaṃ assatthakapitthā vā (“fig and wood-apple [trees]”). The word vā is used as the gloss of vibhāsā.

Eventually it may be possible to derive more than two forms, e.g. Rūp § 197 gives a triple option : kammunā, kammanā, kammena.

3.2. The word vā is often glossed as vavatthitavibhāsā (Skt. vyavasthitavibhāṣā), which means “restricted option”.²⁸ Buddhappiya is the earliest recorded Pali grammarian to use this term.²⁹ The usage may have been borrowed from the Cāndra tradition of Sanskrit grammar.³⁰ The standard expression of Rūp is vavatthitavibhāso’ yaṃ vāsaddo “this word vā represents a restricted option”. This means that the word vā is a vikappa only to a certain extent. This is explained in Rūp § 15 vā paro asarūpā :

In the same way that the preverb ā expresses an inclusive limit or an exclusive limit, the word vā in some places expresses an open option (vikappa) and in some places expresses the delimitation (pariggaha) of certain word-forms (rūpa) according to a specific restriction (yathāvavatthita). In the present sutta, the latter applies. Therefore the word vā here illustrates a grammatical operation (vidhi) which involves [simultaneously] a mandatory (nicca) domain, a non-mandatory (anicca) domain, and an inapplicable (asanta) domain.³¹

The option prescribed by Rūp § 15 is mandatory. For instance, if we apply sandhi to the expression yassa idāni, only the elision of the second vowel is allowed : yassa ’dāni. The option yass’ idāni is not contemplated.³² The same holds true for cakkhu indriyaṃ, which can only undergo the ligature cakkhundriyaṃ, never *cakkhindriyaṃ. This type of option is called nicca “mandatory”. Other cases fall under the category anicca “not mandatory”, equivalent to vā = vikappa, for example : moggallāno asi or moggallāno ’si. This is a case of “either … or …”, but note that the option is not between eliding the first vowel or eliding the second, but between eliding the second or not eliding at all. Finally, the category asanta “non-existent” or “inapplicable” is glossed as idha na bhavati (“here [the phenomenon of option 2] does not exist [i.e. does not apply]”),³³ for example : pañca indriyāni > pañc’ indriyāni, never *pañca’ndriyāni.

3.3. Following a method that reminds us of the Kātantra tradition,³⁴ Buddhappiya frequently synthesises the details of the restricted option in “versified summaries” (saṅgahagāthās), a sort of ślokavārttikas, supplementary rules in verse,³⁵ apt for easy memorisation along with the sutta. For instance, in Rūp § 15 we read :

bhavati ca vavatthitavibhāsāya :
avaṇṇato saro ’dānītīvevādiṃ vinā paro,
na luppat’ aññato dīgho āsevādivivajjito.

And there is also [the following stanza being the summary] of the restricted option :

After a-vaṇṇa [= a, ā], the vowel that follows is not elided, except in cases such as idāni, iti, iva, eva.

After a vowel other than a-vaṇṇa [i.e. after i, ī, u, ū, e, or o], the following long vowel is not elided. Exception is made in cases such as āsi eva [where no elision at all takes place].³⁶

We may turn now to Rūp-ṭ to clarify the purport of the stanza :

In the summary stanza, the syntax (sambandha) is the following : “after a-vaṇṇa [= a/ā], when it is not homogeneous [with the next vowel], and when it is the cause (hetu) of the restricted option (vavatthitavibhāsā), the next vowel — i.e. i-vaṇṇa [i/ī], u-vaṇṇa [u/ū], etc. — is not elided, except when [it is the first vowel] of words such as idāni, iti, iva, eva, etc.” Herein, the word ādi includes examples such as evaṃ kira me [cf. M III 25,2]. The long vowel that follows a non-homogenous vowel like i-vaṇṇa, etc., is not elided, except in cases such as āsi eva [cf. Vin IV 74 ( ?)].³⁷

Another example of a summary stanza that clarifies the scope of a restricted option is found in Rūp § 69 sabbayonīnam ā e, “All forms yo (nom./acc. pl.) and ni (nom./acc. n. pl.) [can be optionally] replaced with ā and e [respectively].” Buddhappiya indicates in the vutti that vā recurs. Then he continues :

The word vā here has the meaning of a restricted option, that is why, with regard to this rule, it is stated :

“The word vā illustrates the operational rule in this manner : It is mandatory (nicca) in the masculine ; not mandatory (anicca) in the neuter ; inapplicable (asanta) in masculine/neuter i/ī stems when they operate as a-endings.”³⁸

Thus we always find nom. pl. purisā and acc. pl. purise, but never nom./acc. pl. *purisayo. Neuter words replace ni with ā and e, but not mandatorily, thus : nom./acc. pl. cittāni, but also nom. pl. cittā, acc. pl. citte. Finally, the rule is not applicable to the masculine/neuter stems in i/ī such as aggi “fire” : nom./acc. pl. aggayo, but never aggā and agge.

3.4. The option marker vā³⁹ is used sometimes “for the purpose of blocking” (nivattanatthaṃ) other option markers. See for instance Rūp § 190, where the term vā is not explicit in the rule, but its governance is invoked : vādhikāro katthacinivattanattho “the governance of vā is in order to prevent the application of katthaci”.⁴⁰ In its equivalent rule in the Kaccāyana grammar (Kacc § 239), the governance of vā is missing because the author of Kacc/Kacc-v understands that the effect of optionality has already ceased, and the rule should be understood as a simple injunction.

3.5. Furthermore, in Rūp §§ 155, 226, and 440, ca is used also as katthacinivattanatthaṃ ; in § 201 as katthacipaṭisedhanivattanatthaṃ ; and in § 141 as kvacinivattanatthaṃ. The words katthaci° and kvacinivattanatthaṃ appear to be synonymous. In § 155, for instance, the effect of ca, glossed as katthacinivatthanatthaṃ,⁴¹ is precisely to cancel the exceptional nature of the rule and make it open in some cases. The same rule is affected by the option marker vā, recurring from the vutti of Rūp § 153. This vā yields the following vavatthitavibhāsā :

bhikkhuppabhutito niccaṃ ; voyonaṃ hetu-ādito
vibhāsā ; na ca vo no ca amuppabhutito bhave.

[The ending yo for nom./acc. pl.] is mandatory in words such as bhikkhu, etc.
[The ending yo for nom./acc. pl.] can optionally (vibhāsā) take vo or yo in words such as hetu, etc.
[The endings] vo and no [for nom./acc. pl.] do not apply to words such as the [pronoun] amu, etc.⁴²

Note that the role of ca as blocker of an exception gives, as a result, the vibhāsā described in the summary stanza : one can form the plural of hetu : hetuyo or hetuno. Thus, one and the same rule can have more than one option marker in the interpretation, and these option markers have different roles.

4. kvaci and navā “exceptionally”⁴³

4.1. In the Rūpasiddhi there are fifty-six rules involving kvaci, of which forty-five, the large majority, express exceptions of some sort. Seven more rules could be called exceptions too, but the role of kvaci is not entirely transparent (§§ 43, 45, 109, 111, 260, 350, 588) ; in three cases kvaci seems to mark vikappa (§§ 39, 266, 502) ; and once it is glossed as vavatthitavibhāsā (§ 461).

4.2. The function of kvaci as exception marker is illustrated in Rūpasiddhi-ṭīkā ad Rūp § 466. The commentary explains that the insertion of i is allowed in asabbadhātuka⁴⁴ verbal roots when the following vibhatti begins with a consonant, but not when it begins with the vowels a or i, and he adds : “It should be understood that this distinction (viseso) is established by the governance of the word kvaci” (ayaṃ pana viseso kvacī ti adhikārato siddho ti veditabbo).⁴⁵ According to this statement, the term kvaci indicates a case that is special or particular (visesa) as opposed to a general state of affairs. Thus, kvaci is often found as a blocker of a more general option marked with vā. It should not come as a surprise, then, that the function of the word kvaci can also be carried out by its synonym navā, i.e. “the vā [open option] does not (na) [apply]” (see below 4.5.).⁴⁶

4.3. As stated above (3.5), the terms kvaci and vā (= vikappa) may both affect one and the same rule. Even in such cases they are explicitly described as performing different functions. Consider for instance Rūp § 336 :

na pañcamyāyam ambhāvo kvacī ti adhikārato,
tatiyāsattamīchaṭṭhīnan tu hoti vikappato.

Due to the governance of kvaci, the ending aṃ is not there before the fifth case ending,
but it may be there optionally (vikappato) before the third, seventh, and sixth case endings.⁴⁷

The stanza makes clear that there are two different poles of optionality, one is kvaci, the other is vikappa. The commentary explains : “Because of the governance of the word kvaci, there is no aṃ for the fifth vibhatti, but aṃ is optional for the third, seventh and sixth vibhattis.”⁴⁸ From this statement we understand that the function of kvaci is to isolate an exception from the main rule, which may or may not be a vikappa.⁴⁹ This makes kvaci different from nicca “mandatory”, as kvaci could probably be translated as “mandatorily no option is allowed”. The term nicca does not presuppose any previous option, whereas kvaci does. In the end, however, both terms lead to one single derivation pathway.

4.4. Another clear example of kvaci functioning as a marker of exception is found in Rūp § 35 sare kvaci. Buddhappiya explains that kvaci (“in some places”), means “in those places where prosody is broken, or pronunciation is difficult, and also in those instances where sandhi is neglected. The point is that elision (lopa), replacement (ādesa), and modification (vikāra) do not obtain [in those cases].”⁵⁰

4.5. The term navā appears six times in the Rūpasiddhi (§§ 21, 28, 246, 247, 256). In § 21 Buddhappiya explicitly glosses it as a synonym (pariyāya) of the word kvaci.⁵¹ This is in line with the interpretation of Vimalabuddhi (see 2.1). The other five cases of navā do not clearly attest to this equivalence. If it is true that navā and kvaci have the same meaning, the question naturally arises as to why Rūp § 21 ivaṇṇo yaṃ navā (= Kacc § 21) does not use the word kvaci. According to Vimalabuddhi, the use of navā in Kacc § 21 serves the purpose of blocking a former kvaci. It thus marks an exception to an exception. It seems that Buddhappiya has silently accepted Vimalabuddhi’s solution, although we should keep in mind that Rūp-ṭ is quite critical of Vimalabuddhi’s methods regarding optionality.⁵²

4.6. The recurrence of navā can only be studied in a string of three suttas beginning in Rūp § 234. This rule allows the pronominal replacements tavaṃ mamaṃ for the acc.sg. It is followed, with “a frog’s leap” (maṇḍūkapluti), by § 246, which allows vo no for acc./dat./gen. pl., and subsequently by § 247, which allows te, me for acc./dat./gen. sg. In this case the word navā is not blocking any other option marker. In fact, if we look at the Kaccāyana parallel (Kacc § 144), navā is glossed by Vimalabuddhi as “vikappena”.⁵³ There are reasons to believe, however, that Vimalabuddhi has missed the mark, and here navā is not equivalent to vikappa, but to kvaci, in accordance with the general criterion postulated by Vimalabuddhi himself. The Rūpasiddhi, once more, clarifies the point :

Why [does the sutta state] navā ? [Consider the counter examples :⁵⁴] idaṃ cīvaraṃ tuyhaṃ vikappanatthāya dammi “I give these robes to you for the sake of assigning” ; and : suṇātha vacanaṃ mama “listen to my words”.

Here [Rūp §§ 246–47], due to the governance of navā [from Rūp § 234],
these [pronominal] forms, namely vo, no, te, and me,
do not apply at the beginning of a verse (pāda),
or when they precede particles such as ca, vā, eva, etc.⁵⁵

The expression na honti te (lit. “those [instances] do not exist”) tells us that navā is here a marker of an asanta-type exception. What the word navā means in this string of rules, then, is that the cited pronominal forms are forbidden in certain contexts.⁵⁶

4.7. The last occurrence of navā is found in Rūp § 256 catūpapadassa lopo t’ uttarapadādi cassa cu-co pi navā. This rule addresses the behaviour of the numeral catu (“four”) in compounds that express other numerals, such as cuddasa (“fourteen”). Let us look at Buddhappiya’s paraphrase in order to unpack the full meaning of the rule :

When it expresses a numeral and precedes the word dasa (“ten”) in a compound, the word catu (“four”) suffers the elision of the syllable tu, and the remaining ca may exceptionally (navā) be replaced with cu or co. [Examples :] cuddasa, coddasa, catuddasa (“fourteen”).

On account of the mention of the word api, even when catu is not the first word of a compound, if it expresses a numeral, the initial syllable ca is elided ; and, exceptionally (navā), cu or co are there [i.e. replace ca]. For instance : tālīsaṃ, cuttālīsaṃ, cottālīsaṃ, cattālīsaṃ (“forty”).⁵⁷

The idea here is that navā expresses an exception to the main rule formulated in the first half of the sutta. The procedure seems to have two stages : First, a mandatory elision of tu, and second, an exceptional replacement of the remaining ca with cu or co. The interpretation of api is not relevant to our discussion, but it confirms the use of navā as a marker of a special case within a general state of affairs.

5. CONCLUSION

5.1. A preliminary examination of option markers in Buddhappiya’s Rūpasiddhi has shown that the system of two levels of optionality, originally postulated by Vimalabuddhi in ca. the tenth century CE, seems to be functional on the suttas of Kaccāyana and suffers from few exceptions. The claim that Kaccāyana grammarians use option markers randomly (see 1.3), therefore, needs to be corrected. It is important to note, too, that each grammarian may tweak the system according to his needs. Thus, Buddhappiya’s system is an improved version of Vimalabuddhi’s, and the improvements have to do with the radically new approach that Buddhappiya adopted towards the sequencing of the rules. Later grammarians of the same school do not hesitate to apply Buddhappiya’s refined system to the old sequence of rules in Kacc.⁵⁸ Future studies on vernacular commentaries, such as the Rūpasiddhisannaya, will no doubt amplify and correct what has been stated in this paper.

5.2 To conclude, let us go back to Kiparsky’s innovative approach of option markers in Pāṇini, and how it may be related to Pali grammarians. Unlike the Pāṇinian vyākaraṇa tradition in Sanskrit, Pali grammarians openly discuss the role of different option markers. Their description of these technical terms, however, does not fit Kiparsky’s thesis, which posits different degrees of “preference”. Rather, Kaccāyana grammarians seem to favour a type of analysis that is more in line with Kiparsky’s critics.⁵⁹ Indeed, the distinction of option levels does not seem to be related to frequency of use, or any preference of the users of the language, but simply to a hierarchy of domains. It has nothing to do with the outside world, so to say. It has to do with the very structure of a grammatical treatise composed in the sūtra style. For the sake of parsimony, a basic contraposition is established between the general rule (utsarga) and the particular or exceptional rule (apavāda).⁶⁰ In-between we find what is generally called an alternative or open option. To put it in other words : a general rule will state that only X form is possible. An open option will posit that both X and Y forms are possible. A particular or exception rule will state that only Y is possible. Thus, we have three different possible domains. These domains are not related to statistical occurrence, or to any preference of usage, but simply to the sheer possibility of achieving the “perfection of a [word-] form” (rūpasiddhi). Accordingly, Pali grammarians use vā and vibhāsā to derive an optional form that is as correct as a form given in a general rule previously promulgated, and they use kvaci and navā to mark an exception to the option marked by vā/vibhāsā. It is not by chance, perhaps, that Buddhappiya compares the option marker par excellence, namely vā, with the preverb ā in the sense of “limit” or “boundary”, both inclusive and exclusive. This comparison subtly suggests a parallel with the inclusive option-boundary (vibhāsā) and the exclusive option-boundary (kvaci). For, as we have shown in this article, option markers are not only capable of indicating variation, but they can also indicate whether the variation occurs within the general domain of a rule, or beyond its boundaries.

ABBREVIATIONS

For Pali texts, I adopted the list of abbreviations of the Epilegomena to A Critical Pāli Dictionary : http://cpd.uni-koeln.de/intro/

Be = Burmese edition

Ce = Cinghalese edition

DSG = A Dictionary of Sanskrit Grammar, Abhyankar, 1961

GRETIL = Göttingen Register of Electronic Texts in Indian Languages (http://gretil.sub.uni-goettingen.de/gretil.html)

Kacc = Kaccāyana

Kacc-v = Kaccāyana-vutti

Kāt = Kātantra

MBh = Mahābhāṣya

Mmd = Mukhamattadīpanī

Mmd-pṭ = Mukhamattadīpanī-pūrāṇa-ṭīkā

Mmd-sāra = Mukhamattasāra

Mogg = Moggallāna

Mogg-v = Moggallāna-vutti

Mogg-p = Moggallāna-pañcīkā

MW = Sanskrit–English Dictionary, Monier Williams, 1899

Pāṇ = Pāṇini’s Aṣṭādhyāyī

Pas = Paspaśāhṇika, ed. Joshi and Roodbergen, 1969

PED = (Pali Text Society’s) Pali–English Dictionary, Rhys Davids and Stede, 1921

Rūp = Rūpasiddhi

Rūp-ṭ = Rūpasiddhi-ṭīkā

Rūp-bh-ṭ = Rūpasiddhi-bhāsā-ṭīkā

Sadd = Saddanīti

REFERENCES

PRIMARY SOURCES

Aṣṭādhyāyī = Böhtlingk, 1887

Kaccāyana and Kaccāyanavutti = Pind, 2013

Kaccāyanasuttaniddesa =

Be Suttaniddesapāṭh, Rangoon : Zabu Meit Swe Press, 1912

Ce The Kaccāyanasuttaniddesa by the Venerable Neruttikācariya Chappada Mahāthere, Revised and Edited by the Rev. Mabopitiye Medhankera Bhikkhu, Colombo : Vidyabhusana Press, 1915

Kātantra = Eggeling, 1874

Kāśikāvṛtti = Sharma et al., 1969

Mahābhāṣaya = Kielhorn, 1880

Mukhamattadīpanī = Nyāsapāṭh. Rangoon : Sudhammavati Press, 1913

Mukhamattadīpanī-pūrāṇa-ṭīkā = Saṃpyaṅ-ṭīkā-pāṭh, Yangon : Kavi Myat Hman Press, 1914

Mukhamattasāra = Forthcoming edition by A. Ruiz-Falqués, Pune : Pune Indological Series.

Rūpasiddhi = Padarūpasiddhi, Saccamaṇḍaiṅ Press, Yangon, 2006 (3rd reprint)

Rūpasiddhi-ṭīkā = Padarūpasiddhiṭīkā, Mandalay : Padesā Piṭakat-cā-pe, undated reprint of 1964 (CS 1326) edition

Rūpasiddhi-bhāsā-ṭīkā = Ashin Janakābhivaṃsa, Rūpasiddhi-bhāsā-ṭīkā, Mandalay : First Book Press, 2001

Saddanīti = Smith, 1930

SECONDARY SOURCES

Abhyankar, Kashinath Vasudev. 1961. A Dictionary of Sanskrit Grammar. Baroda : Oriental Institute of Baroda.

Böhtlingk, Otto. 1887. Pāṇini’s Grammatik. Leipzig : H. Haessel.

Bronkhorst, Johannes. 1982. “The variationist Pāṇini and Vedic : a review article,” Indo-Iranian Journal 24 : 273–82.

D’Alwis, James. 1863. An Introduction to Kacchàyana’s Grammar of the Pali Language. Colombo : Williams and Norgate.

Cardona, George. 1997. Pāṇini : His Work and Its Traditions. Volume 1 : Background and Introduction. Second Edition, Revised and Enlarged. Delhi : Motilal Banarsidass.

Cardona, George. 2004. Recent Research in Pāṇinian Studies. Delhi : Motilal Banarsidass.

Deokar, Mahesh A. 2008. Technical Terms and Technique of the Pali and Sanskrit Grammars. Varanasi : Central Institute of Higher Tibetan Studies.

Deshpande, Madhav M. 1984 “Review : Pāṇini as a Variationist by Paul Kiparsky and S.D. Joshi,” Language, Vol. 60, No. 1 (March) : 161–64.

Devasthali, G.V. “Pāṇini and Vedic — a Critique,” Annals of the Bhandarkar Oriental Research Institute, 1983, Vol. 64, No. 1/4 : 137–48.

Eggeling, Julius. 1874. The Kātantra with the Commentary of Durgasiṃha. Calcutta : Royal Asiatic Society.

Franke, Rudolf O. 1902. Geschichte und Kritik der einheimischem Pali-Grammatik und -Lexicographie. Strassburg : Trübner.

Geiger, Wilhelm. 2000. A Pāli Grammar. Translated into English by Batakrishna Ghosh. Revised and edited by K.R. Norman. Oxford : Pali Text Society.

Gornall, Alastair. 2017. “Ratnamati et ses oeuvres [Review : Dragomir Dimitrov, The Legacy of the Jewel Mind. Naples : Università degli studi di Napoli ‘L’Orientale’, 2016],” BEFEO, 103 : 475–91.

Gornall, Alastair. 2020. Rewriting Buddhism. Pali literature and monastic reform in Sri Lanka, 1157–1270. London : UCL Press.

Gornall, A. and Gunasena A. 2018. “A History of the Pali Grammatical Traditions of South and Southeast Asia by Vaskaḍuvē Subhūti (1876), Part One : The Kaccāyana-vyākaraṇa, Its Commentaries and Major Handbooks,” Journal of the Pali Text Society 33 : 1–53.

Gornall A. and Ruiz-Falqués, A. 2019. “Scholars of Premodern Pali Buddhism.” In Brill’s Encyclopedia of Buddhism, vol. II. Lives, edited by Jonathan Silk, Richard Bowring, Vincent Eltschinger, and Michael Radich. Leiden : Brill : 420–36.

Grierson, G.A. “The Prakrit vibhāṣās,” Journal of the Royal Asiatic Society of Great Britain and Ireland, Jul. 1918 : 489–517.

Grünwedel, Albert. 1883. Das sechste Kapitel der Rūpasiddhi, nach drei singhalesischen Pāli-Handschriften herausgegeben. Berlin : A.W. Sachade Buchdruckerei.

von Hinüber, O. 1983. Notes on the Pāli Tradition in Burma (Beiträge zur Überlieferungsgeschichte des Buddhismus in Birma, I), Nachrichten der Akademie der Wissenschaften in Göttingen, I. Philologisch-historische Klasse, No. 3, 68.

von Hinüber, O. 2001. Das ältere Mittelindisch im Überblick. 2. erweiterte Auflage. Wien : Verlag der österreichischen Akademie der Wissenschaften.

Joshi S.D. & Bhate, S. 1984. The Fundamentals of anuvṛtti. Pune : University of Poona.

Joshi S.D. & Roodbergen, J.A.F. 1986. Patañjali’s Vyākaraṇa-Mahābhāṣya : Paspaśāhṇika. Introduction, Text, Translation and Notes. Pune : University of Poona.

Kahrs, E.G. 1992. “What Is a ‘tadbhava’ Word ?,” Indo-Iranian Journal, 35 : 225–49.

Kielhorn, Franz. 1880. The Vyākaraṇamahābhāṣya of Patañjali, Vol. 1. Bombay : Government Central Book Depot (Reprinted 1970, Osnabrück : Otto Zeller Verlag. Second edition revised 1891. Third edition, revised by K.V. Abhyankar, Pune : Bhandarkar Oriental Research Institute 1962).

Kiparsky, Paul. 1979. Pāṇini as a Variationist. Pune : Centre of Advanced Studies in Sanskrit, University of Poona.

Malai, Phramaha Thiab. 1997. “Kaccāyana-vyākaraṇa. A Critical Study,” unpublished PhD thesis, Department of Sanskrit and Prakrit Languages, University of Pune.

Monier Williams, Monier. 1899. Sanskrit–English Dictionary. Oxford : Clarendon Press.

Nandisena, Bhikkhu. 2005. Kaccāyanabyākaraṇaṃ. Translated into English by U Nandisena. Yangon : ITBMU (International Theravada Buddhist Missionary University).

Oberlies, Thomas. 2019. Pali Grammar. The Language of the Canonical Texts of Theravāda Buddhism. Part I : Phonology and Morphology. Bristol : Pali Text Society.

Palsule, Gajanan B. 1982. “Review : Pāṇini as a Variationist by Paul Kiparsky and S.D. Joshi,” Annals of the Bhandarkar Oriental Research Institute, vol. 63, No. 1/4 : 340–42.

Pind, Ole. 2012. “Pāli Grammar and Grammarians from Buddhaghosa to Vajirabuddhi : A Survey,” Journal of the Pāli Text Society XXXI : 57–124.

Pind, Ole H. 2013. Kaccāyana and Kaccāyanavutti. Bristol : Pali Text Society.

Rachiwong, Phramaha Sriporn. 1995. “Rūpasiddhi : A Study of Some Aspects,” unpublished PhD thesis, Department of Sanskrit and Prakrit Languages, University of Pune.

Rhys Davids, T.W. and Wilhelm Stede. 1921. Pali–English Dictionary. London : Pali Text Society.

Ruiz-Falqués, A. 2015. “A Firefly in the Bamboo Reed”, unpublished PhD dissertation, University of Cambridge, UK.

Ruiz-Falqués, A. 2018. “Review : Thitzana, Ashin U. Kaccāyana Pāli Grammar, Translated into English with Additional Notes. Pariyatti Press, Onalaska, 2016,” Journal of the Ñāṇasaṃvara Centre for Buddhist Studies I (2018) : 279–304.

Senart, Émile. 1871. Kaccāyana et la littérature grammaticale du Pāli. Paris : Journal Asiatique.

Sharma, Rama Nath. 1983. “Review : Pāṇini as a variationist,” Language and Society, vol. 12, No. 3 (September) : 361–69.

Sharma, A., Deshpande Kh., Padhye D.G. 1969. Kāśikāvṛtti, Vol.1. Hyderabad : Osmania University.

Smith, Helmer. 1930. Saddanīti, la grammaire palie d’Aggavaṃsa. Vol. III : Suttamālā. Lund : Glerup. (Reprint Oxford : Pali Text Society, 2001)

Smith, Helmer. 1949. Saddanīti, la grammaire palie d’Aggavaṃsa. Vol. IV : Tables 1e partie : Textes cités, Racines, Morphèmes, Système Grammatical et Métrique. Lund : Glerup. (Reprint Oxford : Pali Text Society, 2001)

Smith, John D. 1982. “Review : Pāṇini as a Variationist by Paul Kiparsky,” Bulletin of the School of Oriental and African Studies, University of London, vol. 45, No. 1 : 185–86.

Thitzana, Ashin U. 2016. Kaccāyana Pāli Grammar : Translated into English with Additional Notes, Sample Explanations and Tables. Vol. 2. Pariyatti Press, Onalaska.

Tin Lwin. 1991. “The Saddanīti,” in : Studies in Honour of Mingun Sayadaw. Yangon : Tipitaka Nikaya Ministrative Organization : 117–26.

I would like to thank Anuja Ajotikar, Paolo Visigalli, Alastair Gornall, Petra Kieffer-Pülz, Professor George Cardona, and the anonymous reviewers for their precious feedback. I am also very grateful to Dr William Pruitt for revising the English, and to Professor Mahesh Deokar for his crucial support in my Kaccāyana studies. All errors in this article remain my own responsibility. I would also like to express my deepest gratitude to the Venerable Dr Khammai Dhammasāmi, Dr Pyi Phyo Kyaw, and all my colleagues and sponsors at SSBU, Taunggyi, for supporting this research. This article is dedicated to my teacher, Dr Eivind Kahrs, who first encouraged me to study the mysterious mechanisms of optionality in Pali grammars. ↩
Kahrs (1992 : 233) compares Pāṇini’s grammar with a map describing a territory that is the language : “If we have a map — and I think it is justified to call the linguistic descriptions of the ancient Indian grammarians a map — it will tell us a great deal about those who made the map. A basic question is this : What features of the territory are represented on the map ? If the territory is absolutely uniform, nothing would be represented on the map except the borders of the territory. Otherwise, what will be represented on the map is really differences of various kinds.” ↩
Kiparsky 1979 : 1. ↩
Kiparsky 1979 : 1. ↩
Strong reservations regarding Kiparsky’s thesis are presented in Palsule 1982, Devasthali 1983, and Cardona 2004 : 162ff. Other distinguished reviewers, however, have accepted the thesis : Bronkhorst 1982 : 273 ; Smith 1982 : 185. For a middle-way assessment, see Deshpande 1984. ↩
By “classical” Pali grammar I mean the Kaccāyana, the Moggallāna, and the Saddanīti, along with their commentaries and offshoots. I use the term classical in its meaning “exemplary”. For overviews of classical grammatical scholarship in Pali, I refer to Franke 1902, Pind 2012, Deokar 2008, Gornall and Gunasena 2018, and see also von Hinüber 1983, and Gornall and Ruiz-Falqués 2019 ; for an overview of the Kaccāyana tradition, see Ruiz-Falqués 2016. ↩
For Kacc sutta as lacking a systematical approach, see Senart 1871 :14 ; see also Franke 1902 : 14 ; for the wrong use of vā, see Senart 1871 : 93. ↩
Pind 2012 : 83. ↩
Ibid. ↩
Deokar 2008 : 367, 369. ↩
For instance, in his notes to the ākhyāta chapter of the Rūpasiddhi, Grünwedel translates both kvaci and vā as “arbiträr”, but he is obviously aware of the levels of optionality of the vavatthitavibhāsā “bestimmte vibhāshā” (Grünwedel 1883 : 52) ; Ashin Thitzana (2016), translates the four main option markers in Kacc vā, kvaci, na vā, vibhāsā as “sometimes”, cf. Thitzana 2016 : 129 n.11. ↩
D’Alwis (1863 : 25–26), for instance, translates vā as “optionally” and kvaci as “sometimes.” ↩
Malai 1997 : 105 ; Nandisena 2005 : 48. ↩
For the most up-to-date biographical sketch of Buddhappiya, see Gornall 2020 : 69ff. ↩
Rachiwong 1995 : 10 ; Gornall and Gunasena 2018 : 33 ; Gornall 2020 : 24. ↩
For a critical edition of Kaccāyana and Kaccāyanavutti, see Pind 2013. ↩
Rūp-ṭ Be 4,3–5 byattaṃ lakkhaṇavisayavavatthānassa, lakkhiyābhinipphattiyā ca paribyattito abhibyattaṃ supākaṭaṃ. ↩
Gornall and Gunasena 2018 : 34. ↩
Pind 2012 : 71. ↩
For the unclear relationship between the terms bhāṣā “language” and vibhāṣā “dialect” see von Hinüber 2001 : 102–103. Von Hinüber provisionally accepts Kiparsky’s conjecture that vibhāṣā could mean “zu meiden (als Alternative” (i.e. an alternative to be avoided) versus vā “zu bevorzugen” (i.e. to be preferred). ↩
Vimalabuddhi’s discussion is in the context of Kacc § 21 ivaṇṇo yaṃ navā and it is about reading navā as one word (Mmd 32,24ff.), see also Malai 1997 : 105. Note that in his critical edition Pind reads na vā always. The principle of the two levels was synthesised by Guṇasāgara of Pagan (c. thirteenth century) in the Mukhamattasāra (unedited work), and is quoted in Kacc-nidd Be 154,10–11 ; Ce 150,13–14, for the context, or lack of context, of this quotation, see Ruiz-Falqués 2015 : 142. Guṇasāgara’s stanza is quoted by Nandisena (2005 : 48) via a quotation found in the Kaccāyana-vaṇṇanā (16th c. CE, Burma). ↩
Pas III §§ 46–55 ; Joshi and Roodbergen 1968 : 70–78. In the discussion on 1.1.44 naveti vibhāṣā, Patañjali quotes the iṣṭhi : sādhvanuśāsane asmin yasya vibhāṣā tasya sādhutvam “In this teaching of correct [words], correctness applies to that which is optional” (MBh I 104,8). ↩
This is the interpretation of the Rūpasiddhi-ṭīkā, problematically ascribed to Buddhappiya himself (see 5.4). ↩
The numbers include instances in the sutta and the gloss (vutti) properly speaking, that is to say the first line of the commentary in which the expanded formulation of the sutta is given, including the anuvutti (information recurring from previous suttas). Option markers in further sections of the vutti are excluded from this count. These sections are no doubt relevant to the study of optionality and variation, e.g. the word ca introduces a kvaci option in Rūp § 27, or Rūp § 488, invoked a number of times in the elaborations of the vutti, e.g. Rūp §§ 470, 488, 489, 492, 500. A more detailed study on option markers, including all the sections of the Rūp commentary and other Kaccāyana commentaries, is in preparation. ↩
This term is known in Sanskrit grammatical literature. Patañjali uses it in his Mahābhāṣya (ad Pāṇini 2.4.56, 3.2.124 and 7.1.56). It becomes part of the system of interpretation in later works such as the Kāśikāvṛtti (cf. Kāś ad Pāṇ 1.3.70 tad asmin viṣaye nityaṃ, anyatra vikalpaḥ. vyavasthitavibhāṣā hi sā ; or Kāś ad Pāṇ 2.3.17 vyavasthitavibhāṣāvijñānād eva na bhavati). The Kātantravṛtti also uses this term, especially in relation with navā : Kāt-v 1.4.2, 1.5.5 ; and see Kāt-v 4,6.83 : svariṣyatīti paratvād vikalpo na syād vyavasthitavāvacanāt ; in Kāt-v 4.1.72 we can observe the triple domain used in Rūp : nityaṃ, na syāt (= na bhavati) and vā (= vibhāṣā). ↩
In cases such as Rūp § 207 or § 259, the vutti suggests that vā is being used in the sense of vavatthitavibhāsā, but that is not explicitly stated in the main paraphrase of the sutta ; therefore, we exclude such cases from the present study (see n. 24). ↩
In reality, rājena only appears as part of a compound (dhammarājena, devarājena, etc.), never in a sentence such as rājena kataṃ (cf. Oberlies § 45, 2b). ↩
“Option limitée” (Smith 1949 : 1146, § 7.3.2.3), “bestimmte vibhāsā” (Grünwedel 1983 : 52). ↩
For the relative date of the Rūpasiddhi, older than the Moggallāna-byākaraṇa (1165 CE), see Gornall and Gunasena 2018 : 34, n.1. As for the relative dates of the Rūpasiddhi and the Saddanīti, they are problematic. Some scholars assume that the Sāsanavaṃsa’s dating of Sadd, 1154 CE, is correct (e.g. Franke 1902 : 25ff. ; von Hinüber 2001, § 62). Other Burmese sources, as little reliable as the Sāsanavaṃsa, point to the thirteenth century CE (cf. Tin Lwin 1991). Aggavaṃsa seems to rely on his predecessors for the understanding of the term : vavatthitavibhāsāyaṃ vāsaddappayogo ācariye payirupāsitvā gahetabbo “The usage of the word vā in the sense of restricted option (vavatthitavibhāsā) should be adopted respecting the teachers” (Sadd 889,2–3 and n. 2). For a detailed examination of the relative chronology : Rūpasiddhi > Nyāsappadīpa > Saddanīti, I refer to my forthcoming monograph on Guṇasāgara’s Mukhamattasāra. ↩
Gornall 2017 : 479 : “Au début de sa discussion du chapitre sur les kāraka, le commentaire singhalais sur la Rūpasiddhi affirme que Buddhappiya, auteur de la Rūpasiddhi, s’est appuyé sur la Cāndrapañcikā de Ratnamati dans son interprétation du mot ‘vā’ (‘facultativement’) dans les Sūtra de Kaccāyana.” I thank Alastair Gornall for calling my attention to this important point. It is possible, too, that the term vavatthitavibhāsā/vyavasthitavibhāṣā was borrowed from the Kātantra tradition, where it is widely used, see below § 3.3. ↩
Rūp 9,1–4 : yasmā pana mariyādāyaṃ abhividhimhi ca vattamāno ā-upasaggo viya vāsaddo dvidhā vattate, katthaci vikappe, katthaci yathāvavatthitarūpapariggahe. idha pana pacchime. tato niccam aniccam asantañ ca vidhim ettha vāsaddo dīpeti. The word asanta literally means “non-existent”, i.e. instances of the rule are not found. See also Rūp-ṭ 14,21–15,6 : yasmā ā-upasaggo viya vāsaddo dvidhā vattate, kvaci vikappe, yathā vā ṇapacce ti ādīsu, katthaci yathāvavatthitānaṃ udāharaṇarūpānaṃ pariggahe, yathā vaggantaṃ vā vagge ti ādīsu imasmiṃ sutte paranayena yathāvavatthitarūpapariggahe vattati, tato niccañ ca aniccañ ca asantañ ca vidhim atra vāsaddo dīpetī ti yojanā “The connection (yojanā) is the following : Because like the preverb ā, the word vā operates in two ways : somewhere in the sense of option, as in vā ṇapacce [Rūp § 361, Kacc § 346], and in some places as the adoption (pariggahe) of the example forms as they have been determined (i.e. restricted), as in vaggantaṃ vā vagge [Rūp § 49, Kacc § 31], in this sutta it operates as adopting the restricted form by taking the following [speech-sound] (paranayena), therefore the word vā here illustrates an operation that is mandatory, and also not mandatory, and also non-existent.” ↩
I have only been able to find one exception to this rule in the Burmese edition of Sāriputta’s sub-commentary on the Aṅguttara Nikāya, twelfth century CE (cf. A-ṭ Be II 179). I could not find any instance in canonical or aṭṭhakathā texts. Rūp GRETIL ed. p.15 yassadāni yassidāni, is a wrong reading for yassa dāni, yassa idāni. ↩
In Sanskrit grammars we find an equivalent expression in na syāt. ↩
Cf. Grünwedel 1883 : 69. ↩
Cf. DSG s.v. ślokavārttika. ↩
Rūp Be 9,18–20. ↩
Rūp-ṭ Be 15,1–6 : saṅgahagāthāya vavatthitavibhāsāya hetunā asarūpabhūtā avaṇṇato paro ivaṇṇuvaṇṇādiko saro ṭhapetvā idāni iti iva eva icc evam ādikaṃ na luppatī ti sambandho. etth’ ādisaddena evaṃ kira me ti ādi saṅgahyati. aññasmā asarūpabhūtā ivaṇṇādito paro dīgho saro āsi eva icc ādivivajjito na luppati. ↩
Rūp Be 45,23ff. : vāsaddo ’yaṃ vavatthitavibhāsattho, tena c’ ettha : niccam eva ca pulliṅge, aniccañ ca napuṃsake ; asantaṃ jhe katatte tu, vidhiṃ dīpeti vāsuti. ↩
The different meanings of vā are stated in an appendix on nipātapadas at the end of the Nāmakaṇḍa, cf. Rūp Be 132–36. ↩
Rūp Be 88,10–14. ↩
With the alternative gloss noggahaṇānuvattanatthaṃ (“for the sake of triggering the recurrence of the ending no [from § 151 yonaṃ no]”). This is simply a different way of stating the same thing. ↩
Rūp Be 72,7ff. ↩
Deokar 2008 : 368–69 navā interpreted as “rarely”. ↩
That is to say, “non-sabbadhātuka”. In Pāṇinian grammar, sārvadhātuka refers to a specific set of affixes marked with the letter ś (DSG s.v. sārvadhātuka). In the Kaccāyana school the term is defined in Kacc § 433 (= Rūp § 458) hīyattanīsattamīpañcamīvattamānā sabbadhātukaṃ “the technical term sabbadhātuka denotes the finite verbal endings of hiyattanī (imperfect), sattamī (optative), pañcamī (imperative), and vattamānā (present)” (trans. Deokar 2008 : 205) ; its negation, asabbadhātuka, is used in Kacc § 518 (= Rūp § 466) ikārāgamo asabbadhātukamhi “in non-sabbadhātukas, [there is] augment i [obtains] ;” see Deokar 2008 : 206 : “In Pali grammars, the main function of these two terms [sabbadhātuka and asabbadhātuka] is to distinguish between the finite verbal endings which take an augment i from those which do not.” ↩
Rūp-ṭ Be 191,4ff. ↩
See Kiparsky 1979 : 3 : “Vibhāṣā is defined by na vā in 1.1.44 na veti vibhāṣā. Na vā is not used otherwise, apart from the cases when it arises implicitly from the combination of a na continued by anuvṛtti with an overt vā stated in a rule.” ↩
Be 180,22ff. ↩
Rūp-ṭ Be 135,24–26 : kvacī ti adhikārato aṃbhāvo pañcamyā vibhattiyā na hoti, tatiyāsattamīchaṭṭhīnaṃ aṃbhāvo pana vikappena hotī ti yojanā. ↩
See also the stanza in Rūp § 466 (Rūp 276,19ff.), where vā and kvaci control different domains : asabbadhātuke byañjanādimhe vā ’yam āgamo kvacādhikārato byañjanādo pi kvaci no siyā. ↩
Rūp Be 22,5–7 : kvaci chandabhedāsukhuccāraṇaṭṭhāne sandhicchārahitaṭthāne ca, na lopādesavikāram āpajjante ti attho. Similarly, euphony and metre mark kvaci-type exceptions in Rūp § 35 (Rūp 22,18ff. : kvacī ti kiṃ ? itismiṃ chandānurakkhaṇe sandhi hoti), Rūp § 37 (24,18ff.) sukhuccāraṇachandarakkhaṇaṭṭhānesu and Rūp § 38 (25,11ff.) chandānurakkhaṇe. And see also Rūp §§ 53, 54, 57. ↩
Rūp § 21 (Be 14,6ff.) : navāsaddo kvacisaddapariyāyo ; see also Malai 1997 : 105. Note that it does not say ayaṃ navāsaddo “this particular navā”, but simply navāsaddo, “the word navā [in general]”. ↩
Cf. Rūp-ṭ Be 21,5–16. Furthermore, in Rūp-ṭ ad Rūp § 35, the author of the commentary disparages Mmd for understanding kvaci as a synonym of vā in the commentary upon Kacc § 24 sare kvaci. ↩
Mmd Be 138,28. ↩
It is customary in Kacc and Rūp to give counter-examples by showing how the desired result would not be there if a word was missing. The counter examples are introduced by the question tag kiṃ, and thus they are called kimudāharaṇa “examples of why [a certain word needs to be used in the sutta]”. The question could be also rephrased : “What would happen if the word X was missing in the sutta ?” The first attestation of the term kimudāharaṇa in the sense of “counter-example” is in Vimalabuddhi’s Mmd, cf. Mmd 26,23 ; 34,27 ; 35,20 ; 39,26, and passim. ↩
Rūp Be 116,3–8 : navā ti kiṃ ? idaṃ cīvaraṃ tuyhaṃ vikappanatthāya dammi, suṇātha vacanaṃ mama. navādhikārato c’ ettha, vo no te me ti ye ime ; pādādo ca ca-vā-evā-diyoge ca na honti te. ↩
Cf. Rūp-bh-ṭ I 276. ↩
Rūp Be 119,6–12 : gaṇane dasassādimhi ṭhitassa catu icc etassa upapadassa tusaddo lopo hoti, uttarapadādimhi ṭhitassa catūpapadassa cakārassa cu-coādesā honti navā. cuddasa, coddasa, catuddasa. apiggahaṇena anupapadassāpi gaṇane padādicakārassa lopo, cu-co honti navā, yathā tālīsaṃ, cuttālīsaṃ, cottālīsaṃ, cattālīsaṃ. ↩
Chapaṭa Saddhammajotipāla, commenting on the old suttapāṭha of Kaccāyana, uses the terms nicca, anicca, and asanta, cf. Kacc-nidd Be 12,5–10 ; Ce 10,24–11,6. ↩
Notably Palsule 1982 and Cardona 2004, cf. n.3. ↩
For a detailed introduction to the dialectics between general rules and exceptions in Pāṇinian vyākaraṇa, see Cardona 1997 : 404–12 ; for the descriptive model that combines general and particular rules in Pāṇini, see Kahrs 1992 : 232–33. ↩