A New Philology: From Norm-bound Practice to Practice-bound Norm in Kannada Intellectual History
p. 389-406
Texte intégral
A number of regional languages in premodern South Asia possess philological traditions—whereby I refer to commentarial work as well as grammatical, metrical, rhetorical, and related disciplines—that constitute a vast, important, and scarcely examined domain. The Tamil embodiment of this tradition, thanks to the work of François Gros and scholars like him—though there are few like him—has actually been reasonably well served. For the history of Kannada philology, however, the situation is radically different. Modern scholarship written in Kannada itself is stunning; men such as R. Narasimhachar, D.L. Narasimhachar, B. M. Srikantia, M. V. Seetha Ramiah, M. Timmappaya, and M. G. Pai, and most recently T. V. Venkatachala Sastry, are the equal of any known to me from elsewhere, scholars endowed with authentic philological sensibilities, deep historical understanding, and keen critical intelligence. But almost nothing on the subject has been written outside of Kannada. Faced with what is virtually a blank slate for western readers, I aim here to present some brief reflections on several of the principal texts and persons of early Kannada philology; on what I believe was a crucial, but rarely noted conceptual revolution with great consequences for the history of consciousness and culture in South Asia, and to which that philology contributed importantly; and on the significance of all this intellectual activity for the political sphere of premodern Karnataka.
I.
The Kavirājamārga (c. 850) had already gone some way in establishing the groundwork for a systematic reflection on and disciplinary organization of literary Kannada.1 The philologization we find in the Mārga, which was to develop uninterruptedly for another four centuries or more, is not only precocious but, with respect to its relationship with other literary cultures of southern India, both autonomous and uncommon. Kannada grammatical science, to take just that discipline, appears to have originated in complete independence from that of Tamil, the only tradition of comparable antiquity.2 By the same token, Kannada philological scholarship seems to have exerted little influence on its neighbors. Marathi, for instance, the Kannada example at its doorstep notwithstanding, was not to be grammaticized in any serious way until Christian missionaries did so in the nineteenth century. Still, the study of southern Indian philology, and a fortiori its comparative study, like the comparative study of southern Indian literature more broadly conceived, is very much in its infancy at least in the West, and many questions concerning the larger conversation among Deccani and peninsular intellectuals that may have occurred in the early centuries of vernacularization remain quite open.
1Although the Mārga, given the parameters established by its Sanskrit models, refrains from offering a systematic study of grammar, prosody, or lexicon, it touches on all three areas and broaches a number of topics that, both because of their central importance to what I have called the cosmopolitan vernacular, and because it was the Mārga that had broached them, would continue long afterward to be addressed by grammarians, metricians, and lexicographers. More important than the particular problematics it bequeaths, however, is the metadiscursive framework within which these are situated. For, in effect, what everywhere conditions the Mārgás exposition is the specification of Kannada difference, and it is against the backdrop of the Sanskrit cultural episteme—defining what language, especially the language of literary culture, is supposed to be—that this difference is constituted. Every feature of the literary in Kannada is marked by a calculation of how the local responds to the global that seems ever copresent with it. Equally consequential is what the local had to do, philologically speaking, in order to respond at all.
2Part of what constitutes that philological episteme inherited by vernacular literary cultures is the transcendent authority from which it purports to derive. Various traditions associate the knowledge of Sanskrit grammar with an episode of divine revelation, a sacral relationship often perpetuated in (Śaiva) temple endowments for the study of grammar. While revelation is found to authorize cultural (and often political) practices in many parts of the world, the Sanskrit tradition perfected the argument for the transcendence of its own authority and the primacy of its changeless linguistic organization, from which all other languages could mark their difference only as deviation. In accordance with this ideologeme, moreover, grammar and the other forms of systematic philological thought were held to communicate a priori norms unaffected by history. Practices alone, it was thought, cannot establish rules, since rules always predate practices; they can only instantiate (or violate) rules. Moreover, the practices that, conceptually at least, formed the basis of Sanskrit grammatical attention, namely those of the Veda, had long been considered authorless and timeless.3
3As a result, a philology in the service of a cosmopolitan vernacular was compelled first of all to secure some kind of authority for its project to establish norms, discipline, and stability. This was all the more necessary given its object. After all, a vernacular was a language whose very essence, according to the dominant representation, consisted of abnormality, indiscipline, instability and above all, therefore, of untruth. Listen to Kumārila, the leading Brahman intellectual of the seventh century, on the scriptures of the Buddhists and Jains:
They are composed in overwhelmingly incorrect (asādhu) language—words of the Magadha or Dakshinatya languages or their even more dialectal forms (tadapabhraṃśa)—and they are therefore false compositions (asannibandhana) . . . When texts are composed of words that are false (asatyaśabda), how could they possibly communicate meaning that is true (arthasatyatā)? And how could they possibly be eternal [as true scripture must be] if we find in them forms that are corrupted (apabhraṃśa)?4
This was the problem confronting the author of the Mārga, and this is how he addresses it:
Among all herds of animals wild and domestic, and flocks of birds, there have been forever countless languages produced each for its own species. In the same way, there exists innately among men the uncultured use of languages. How can unlearned, common people know how to judge that one usage is good and another bad? Their behavior is indifferent, just as herds of animals will indifferently eat grass or grain or fodder. Therefore, one must completely master traditional scholarship (āgama). The man who has not first studied for himself the earlier literary compositions (pūrvakāvyaracanegaḷam) cannot possibly either possess knowledge with respect to words or attain beauty in a literary work. Even a dimwit can derive some knowledge straightway by instruction from a teacher—but there will be no real strength in his expression. Are not male and female parrots able immediately to repeat what they have learned? (KRM 1.7-10)
While the Mārga recognizes the multiplicity of languages and realizes that a certain linguistic competence is inborn, it is unwilling, in conformity to the dominant paradigm, to grant literary status to raw practice; correct usage must be knowledgeable usage. But for such usage in the languages of Place there exists no grammar created by god to which appeal can be made. If knowledgeability here is said to derive from mastery of some kind of systematic thought (āgama), this in turn is shown to be intimately, if paradoxically, dependent on antecedent literary practices, of a sort that have achieved some kind of canonicity. The infinite regress implicit here (or anyonyāśraya), and the criteria that constitute excellence are problems passed over in silence.
4The founding of grammatical norms on literary practices in the vernacular world represents a truly radical break with antecedent conceptions. I explore this further below, but want now to call attention to the new model of cultural authority under construction here, to which is added a legislative power that the Mārga arrogates to itself. It is this that underwrites the project of theorizing, constituting, justifying, and safeguarding Kannada difference in every area of literary form. Consider for a moment the discussion of the selection of lexical items. The first chapter of the Mārga closes with a statement encapsulating its general position:
Words should enter [into a poem] in accordance with the thought [of the poet], and should not be permitted to counteract it. The beauty of the language of Kannada land [must be maintained] in [the use of] the Kannada words (nāl.nuḍiya beḍaṃgu kannaḍada mātinol). There must be propriety observed for Sanskrit words in due measure (pavaṇāgire) and there should be permitted to occur no stumbling over [Sanskrit] words with their harsh phonemes. The composition (pēl.vudu) thereby achieves sweetness and becomes strong, growing forth like the sprout of a vine. Such is the Way of the One Endowed with Consistent Political Wisdom. (KRM 1.149)
This admonition for solicitous attention to proportionality in the quantity of vernacular and cosmopolitan words—precisely the same issue underpins the many forms of maṇipravāla in South Asia, and it will be addressed as well in post-Renaissance defenses of vernacular European poetry5—finds specific application elsewhere, in the structure of compounds, for example, a topic that will be treated with great precision by later grammarians under the rubric of arisamāsa, or "compound of hostiles ":
One should form compositions in the Kannada language with the awareness that, if it is to be blended with appropriated Sanskrit words (samasaṃskṛta-), this must be done judiciously. Such is the Way enunciated by those conversant with traditional knowledge (āgama)… If however one intentionally decides to join in compound-expressions [the aforementioned] Sanskrit and Kannada words without understanding [the conditions of their combinability], the poem will be aesthetically displeasing (virasam), as when mixing drops of buttermilk with boiling milk. (KRM 1.51, 58)
A variety of other formal properties, beyond lexical choice and nominalization, required a defense of the local. One of these, to cite an example from the domain of metrics, concerns the violation of caesura. In the Kannada realization of cosmopolitan verse (both Sanskrit and Prakrit), the ignoring of word-boundary pause within and over the metrical line (and sometimes at half-stanza), in accordance with some other rhythmic sensibility, became obligatory from an early date. Examples are to be found in some of the oldest poems in the language, the undated epigraphs from Śravaṇabeḷgoḷa.6 The Mārga needs to justify and defend local practice and does so by appealing again to a new vernacular authority: "Earlier teachers (ācārya) explain this ´fault, ́ so to call it, to be a virtue in Kannada; in fact, they prefer it. They violate caesura on the grounds that it is superfluous, since in its stead is placed an initial alliteration that segments [the line] (khaṃḍaprāsa), and argue that this is in accordance with the aesthetic of Place (dēsi)" (KRM 1.75).
5All these might seem to modern readers to be the most trivial of concerns, but we will find them repeatedly, even compulsively, examined in the theorization of vernacularity everywhere, theorization that itself is taking place precisely at the centers of political power and among people who counted. And as one can easily observe in the case of Pampa, the opening verses of whose Vikramārjunavijaya echo the Mārga and whose poetry throughout enacts its prescriptions, the preservation of a proportionate difference in lexicon, nominalization, and the structure and selection of meters, was a core value in the practice of vernacularity as well. Observe, too, that while the technical vocabulary deployed in the Mārga for its discourse on metrics, nominalization, lexicon, and the rest is almost exclusively Sanskritic—a tendency that will continue unchecked in the entire later tradition—the ends these phenomena are described as serving, the aesthetics of the literary, are typically formulated in an idiom that is largely vernacular. Thus, in the passage cited above (KRM 1.7-10), highly Sanskritized language pertaining to systematic thought is complemented by dēsi words for "beauty" (beḍaṅgu, v. 9) or "force of expression" (nuḍivalme, v. 10). Here and elsewhere the localization of the aesthetic and the imaginative, in the face of the globalization of the ideational and the informational, seems to become part of the common sense of early vernacularity—as if it were possible to be local only in feeling the world, but not in knowing it.
II.
Given the powerful model of Sanskrit philology, with its full apparatus of grammars (produced with competitive ardor by courtly elites throughout the first millennium), dictionaries, and treatises on the arts of literature, especially poetics and versification, a comparable set of instruments for disciplining and dignifying a language of Place was clearly essential to Kannada if vernacularization was to be successful. (The only component almost entirely absent from the philological toolbox is commentaries on Kannada literary works, something especially puzzling in the Indian context in general and in comparison with the Telugu and Tamil traditions.) The philological works of Kannada, which grew out of a theory of vernacular aesthetics as presented in texts like the Mārga and its embodiment in poems like Pampás, not only shared the project of ensuring the cultural-political elevation of Kannada but like these earlier texts were again, and decidedly, a courtly enterprise. This imposing body of scholarship deserves a monograph of its own, for though it has some parallels in other south Indian vernacular cultures, none of these seems to have been as insistent on the production and defense of literary difference, or perhaps as accomplished in its scientific achievement. Here I can only glance at a few key texts to give a sense of the larger development, before going on to look in a little more detail at the masterpiece of Kannada philology, the Śabdamaṇidarpaṇa. Although the God of vernacular philology lies most definitely in the details, this review will concentrate on major themes and tendencies of the process by which Kannada was confirmed as cosmopolitan vernacular, and the political order in which this process was embedded.
6Late in the tenth century two new forms were developed for Kannada philology, the lexicon and the metrical treatise. The first dictionary, only fragmentarily preserved, is the so-called Rannanighaṇṭu (c. 990), ascribed to the epic poet Ranna, the "emperor-poet" to the kings of the Kalyāṇa Cāḷukya dynasty. We cannot get a very distinct sense of the scope of this text from the extant portion, but, in addition to offering synonyms of rarer Kannada words by way of local lexemes (dēsi) and Sanskrit derivatives (tadbhavas), Rannás procedure of defining in Sanskrit what in some cases are simple everyday Kannada words suggests in part an objective of disciplinary ennoblement, that is, of providing the appurtenances of scholarship whether practical or not, without which the vernacular cannot be literary.7 It is perhaps just this tendency that finds strengthening in the great Sanskrit-Kannada lexicons that succeeded Ranna, such as Nāgavarma’s Abhidhānavastukōśa (c. 1040) and Maṅgarāja’s Abhinavābhidhānam (1398). For their conceptual organization these works are wholly dependent on the models of Sanskrit lexicographers, as Nāgavarma acknowledges by citing his predecessors.8 Here the principal purpose appears to have been to make available to writers of Kannada poetry as wide a range as possible of Sanskrit vocables and their synonyms. Whether we are also to perceive in the rise of this genre of lexicon a gradual deterioration of Sanskrit competence among Kannada literati is unclear, for the evidence of long-term change in linguistic knowledge is somewhat confusing. While the later history of lexicography is decidedly Kannada-centric, with a half-dozen dictionaries produced from about 1400-1700 that define dēsi and tadbhava words to serve the reader of Kannada literature, the target idiom in use is often the far more Sanskritized form. At all events, one thing is as unambiguous in the history of Kannada lexicography as it is in Kannada literature, that Sanskrit had fully penetrated the language, and yet always remained a sign of something other than the local.9
7Around the same time that Ranna took the first steps in Kannada lexicography, Nāgavarma completed the first treatise on Kannada metrics, the Chandombudhi (Ocean of meters). There are substantial difficulties in sorting out the different Nāgavarmas of the first centuries of Kannada literary culture; there may be as many as five in the early period. According to scholarly consensus, the author of the Chandombudhi, normally identified as Nāgavarma I, was a Brahman of the Kauṇḍinya gotra and descendent of settlers in the very village of Veṅgi where Pampás father was born. He eventually relocated to Kannada land and became a client of Rakkasa, younger brother and later successor to Rācḷamalla, the Gaṅga king who ruled in the last quarter of the tenth century. Nāgavarma tells us he "learned from the learned" and wrote for them a treatise, a work "flowing with the nine rasas, new in diction, in which the ways of Place have become a thing of beauty," one that he knows full well to be an innovation, "an unprecedented work. "10
8This was in fact the time and place of a remarkable inventiveness in the area of metrics, a discipline of decisive importance for literary art and one that, as formulated in India, had enormous resonance across Asia, especially Java and even China. The first known commentary on the ancient Chandaḥsūtra of Piṅgala (along perhaps with much of the section on non-sacred meters) was composed by Halāyudha, the Sanskrit lexicographer who provided a model for later Kannada dictionaries and composed a remarkable grammatical śāstrakāvya, Kavirahasya (The Poet́s Secret), in praise of the Rāṣṭrakūṭa king Kṛṣṇa III.11 One of the earliest synthetic accounts of Sanskrit kāvya metrics was produced by Jayakīrti in the Chandonuśāsana (c. 1050), which includes a substantial section on Kannada meter.12 As the last-named work shows, although it is impossible for us now to reconstruct the conversation between Sanskrit and Kannada metricians, it was very likely to have been two-way. And yet the discursive universe was entirely Sanskritic. The whole structure of Nāgavarmás exposition, with respect to basic vocabulary, foundational concepts, techniques of scansion (guru, laghu, pāda, the eight gaṇas, and so on) will be familiar to the student of Sanskrit prosody. But it is of course Kannada that Nāgavarma is characterizing within this borrowed discursive apparatus, and upon which his eye is fixed. Second-consonant rhyme (prāsa), for example, is the first differentiator. Defined in vv. 31 ff., it is something essential for Kannada, without which poetry in the language is said to be unable to achieve beauty (v. 50). But it is the larger framework of his exposition of the metrical types themselves that will give us the clearest picture of how vernacular theorization understood the relationship of local to cosmopolitan forms.
9Nāgavarma argues that there is a wide range of meters that, arising from the cosmopolitan languages, are universally available to languages of Place. A long-misunderstood passage at the beginning of his formal exposition makes this clear:
[adeṃteṃdoḍe] saṃskṛtaṃ prākṛtam apabhraṃśikaṃ paiśācikam emba mūr.uvare
bhāṣegaḷoḷ puṭṭuva [draviḍāndhrakarṇāṭakādiṣaṭpancāśat] sarvaviṣayabhāṣājātigaḷ akkuṃ.
If one were to ask, [we would say that] there are species of meter common to all the languages [of the fifty-six dominions, Drāviḍa, Āndhra, Karṇātaka and so on]. These metrical species have arisen from the three languages, Sanskrit, Prakrit, and Apabhramsha, and from the "half" language Paishachi.
It is not that the fifty-six languages themselves have arisen from the cosmopolitan languages, as all scholars who have dealt with the passage have suggested ("daughter languages," according to Kittel himself).13 No Kannada grammarian or scholar of other philological discipline ever conceived of Kannada as an evolute of some other language, while nonetheless fully acknowledging the limits of the vernacular relative to the transregional presence of Sanskrit, Prakrit, and Apabhramsha. Rather, Nāgavarmás point is that all literary languages make use of the cosmopolitan metrical forms, and these he proceeds to describe: the fixed syllabic meters (akṣaragaṇa) and certain moraic meters with or without fixed cadences (mātrāchandas, mātrāgaṇachandas) (chapters 2-4). It is these structures, as he reiterates several verses later on, that are borrowed directly from "both languages," Sanskrit and Prakrit, and are common to "the languages of all realms" (sarvaviṣayabhāṣā-, v. 44 [70]). Sharply to be distinguished from these, however, are the meters specific to the vernacular world, those "species belonging to the [language of the] Karṇāṭa region" (karnāṭakaviṣaya [bhāṣā] jāti) (described in chapter five). If the vernacular knows its place, it also knows its prerogatives.
III.
Instructive as these early works in lexicography and especially metrics are, the supreme achievement of Kannada philology is unquestionably grammar. It is here that all the powerful tendencies driving forward the process of vernacularization converge in a remarkable synthesis. The striving for the specification of the vernacular particular from within the dominating Sanskrit epistemological universal; the quest for discipline in the putatively lawless dialectal; the search for a new authority upon which this discipline could be founded; the royal court as the social site par excellence for the production of systematic vernacular knowledge—this entire culture-power complex of vernacularity finds its most condensed expression in the production of Kannada grammar.
10The Kavirājamārgam itself laid the groundwork for a philological science in Kannada (including grammatical science) and did so from a position at the center of the Rāṣṭrakūṭa court. This symbiosis of grammar and power remains everywhere in evidence in the Kannada world in the centuries following the Mārga. Consider only several of the most important works. The true grammatical organization of Kannada begins with two texts of the early eleventh century. The first, the Kāvyāvalōkanam (Light on Literature), is a literary treatise written in Kannada very similar in conception to the Mārga except that its first chapter, known as the Śabdasmṛti (Tradition of Words), comprises a brief systematic exposition of the rudiments of grammar. The second is the Karṇāṭakabhāṣābhūṣaṇa (Ornament of the Kannada Language), a full grammar written in Sanskrit. The author of both works is another Nāgavarma (Nāgavarma II), who is distinguished from his namesakes by a sobriquet, Kavitāguṇōdaya (Source of Literary Excellence), that appears in the colophons of his works. Although determining precisely when and where this scholar worked had been one of the thornier questions of Kannada literary history, it is now agreed that he was the kaṭakopādhyāya, or "Teacher of the kaṭaka," at the court of Jayasiṃha Jagadekamalla I of the Kalyāṇa Cāḷukya dynasty (r. 1015-42).14 The next significant work—and one of the greatest vernacular grammars of the premodern world—is the Śabdamaṇidarpaṇa (Jeweled Mirror of Language). This was composed by Kēśirāja in 1260 at the Yādava (that is, Hoysaḷa) court, where he, too, held the position of "Teacher of the kaṭaka," as he tells us at the end of the work:
The Śabdamaṇidarpaṇa of the noble Yādavakaṭakācārya Kēśava
will last as long as the sun and moon, Mount Meru and the ocean, spreading far and wide.15
Whatever the title kaṭakācārya may actually refer to (perhaps "head teacher of the royal capital"), it was evidently a position central to court culture and one that its occupants proudly advertised. And this makes it clear, too, that, like lexicography, metrics, and literature itself, vernacular grammar—precisely as the history of its Sanskrit counterpart would lead us to expect—was an enterprise promoted in the first instance by political elites and courtly intellectuals.
11As an epistemological object Kannada grammar too was profoundly shaped by the Sanskrit discipline. Note that the last premodern grammar of the language (Bhaṭṭa Akalaṅka Deva’s Karnāṭaśabdānuśāsana, 1604), like the first (Karṇāṭakabhāṣābhūṣaṇa) was written in Sanskrit. And yet a tension may everywhere be felt as this exogenous casing, capacious and flexible though it may be, was stretched over a language built to totally different specifications, and that constantly threatened to escape its enclosure. The points of deviation that result from this misfit are as significant as the points of convergence. But let us examine the latter first.
12The most striking fact is that the structure of grammatical exposition itself is entirely Sanskrit-derived. Consider the formal organization in the Kannada-language Śabdasmṛti. Its six chapters concern, respectively: technical terms (saṃjñā), euphonic combination (sandhi), nouns (nāma), nominal compounds (samāsa), secondary derivatives (taddhita), and verbs (ākhyāta). This very closely resembles the structure of Śarvavarma’s Kātantra, though Nāgavarma adds technical terms eliminated in the Kātantra and eliminates the case relation section (kāraka) that the ancient grammar included. Indeed, it was clearly in tribute to Nāgavarmás skill in having adapted this grammar to Kannada that he was adorned with the title "The New Śarvavarma (abhinavaśarvavarma), as it was clearly in tribute to Śarvavarmás own post-Vedic, this-worldly understanding of Sanskrit grammar itself that his work (and not Pāṇinís) was chosen as the model for vernacularization in the first place.16 Precisely the same structure is preserved in the Darpaṇa, too, though it is enlarged to include verbal roots (dhātu), secondary derivatives from Sanskrit (here termed apabhraṃśa) and indeclinables (avyaya).
13Not only is the structure of Kannada grammar derived from Sanskrit, but the entire technical vocabulary for the description of grammatical phenomena in Kannada is Sanskritic. Thus in the Darpaṇa, in addition to what has already been mentioned, we find kriyā for verbal action, bhūta, bhaviṣyat, saṃprati for past, future, and present tense; prathama [puruṣa], madhya, uttama for third, second, and first persons of the verb; kāraka for case relations, vibhakti for case ending, ekavacana and so on for number, guṇavacana for adjective, sarvanāma for pronoun, bhāvavacana for abstract noun. And for those cases where Kannada possesses a grammatical function unavailable in Sanskrit, Sanskrit terminology is invented. Thus gamakasamāsa is used to refer to certain kinds of so-called consecutive compounds unknown to Sanskrit or not considered compounds (though sometimes resembling the aluksamāsas), while liṅga is used to refer to nominal themes including declinable verbal bases. That the eyes for which this work was intended seem sometimes to be predominantly Sanskritic eyes is suggested by explanations such as the one provided for the dual number: although dual morphemes do not exist in Kannada, says Kēśirāja, the dual can be inferred from context (sūtra 94). The evidence from lexicography noted earlier for the pervasion of the vernacular by Sanskrit during this period is further confirmed in the Sanskrit glosses that Kēśirāja provides for his list of roots, and in his final chapter on "obscure usages" of earlier writers (gūḍhapadaprayogas), which are again defined by Sanskrit terms.
14Yet, even if the discursive foundations of the grammar are clearly appropriated from Sanskrit, the conceptual orientation of the grammar as a whole, as the example of the dual or the gamakasamāsa shows, is, quite the contrary, to constitute its object by way of a range of differentia from Sanskrit, in phonology, sandhi, syntax, vocabulary, and the rest. The premodern grammarians of Kannada fully understood that their object of analysis was an order of language different from that whose expository structure they adopted in order to describe it, and stood in some tension with it. No grammarian, or lexicographer or metrician (as noted already in reference to Nāgavarmás Chandombudhi) believed the language to be a mere derivative of Sanskrit. It is never considered in itself to be Prakrit (let alone Apabhramsha), in contrast to some north Indian regional languages (Gujarati, for example, was conceived of as such as late as the eighteenth century).17 Further evidence for this nonequivalence lies in the fact that no grammarian ever conformed to the analytical model available to them from all previous Prakrit grammars and sought to derive the forms of Kannada by transfer rules from Sanskrit. The one exception to this autonomy is found in the analysis of tadbhava words, but to understand this we require a more general account of the grammarianś conception of the Kannada lexicon.
15The Darpaṇa analyzes the vocabulary of Kannada according to four categories: dēśīya, "words of Place," often also termed accagannaḍa, "pure Kanna- da "; apabhraṃśa or tadbhava (the two terms are used synonymously), or "corrupted words," that is to say, those derived from Sanskrit or Prakrit; samasaṃskṛta, "words equal with Sanskrit," borrowed directly with virtually no phonological change (in other South Asian grammatical systems these are called tatsama); and tatsama, which in this text refers exclusively to twenty-one vocables (maṇi, mañca, etc.) that Kannada and Sanskrit share but where the question of provenance is undecidable.18 It is only for the apabhraṃśa lexemes that we encounter the use of transformational phonological rules that seem to presuppose the primacy of Sanskrit. But Kēśirājā presents these at once as "tadbhavas of Sanskrit" (sakkadada tadbhavaṃgaḷ") and as "Kannada [words] that have arisen for [i.e., in place of] Sanskrit (saṃskṛtakke puṭṭida kannaḍa). And indeed, his purpose is anything but to fetishize their Sanskrit origins. Quite the opposite: such words provide precisely a means for avoiding the use of Sanskrit:
For those who want to employ pure Kannada (accagannaḍa) with unadulterated expression (cokkaḷikeyim) without resorting to Sanskrit, apabhraṃśa words provide a handy treasury. They are permitted to form compounds with dēśīya words. (ŚMD 314)
The analysis of tadbhava and dēśīya words is meant to help frame Kannadás stringent rules on nominalization (one of them is given in the above sūtra), which themselves serve well to index the heterogeneity of Kannada and Sanskrit: Unlike the two other classes of words, the tadbhava and dēśīya (along of course with the twenty-one tatsamas), which can freely compound with each other, Sanskrit cannot enter into compounds with "pure Kannada words" except in such rare circumstances as archaisms and lists of courtly titles. Other such combinations are considered "conflicted" or "hostile" compounds (viruddhasamāsa, arisamāsa).19
16The concern with the specification of Kannada difference, found in the discussion of lexicography and in fact throughout the Darpaṇa, is consolidated at the very end of the text in a single memorial verse. This seeks to summarize what makes Kannada distinctive with respect to everything from phonology, sandhi, and syntax to lexicon and prosody:
The uniqueness of Kannada [lies in nine features]: compounds that are intelligible [even though they do not conform to the rules of Sanskrit compounding] (gamakasamāsa); the phonemes /l./ [r.aḷa], /r./ [śakaṭarepha], /ḷ/ [kṣaḷa]; harmonious sandhi [not exceeding two phonemes]; locative absolutes that are appropriate [to Kannada, i.e., even given the absence of strictly locative forms]; vocables that are identical to Sanskrit [but used with slight phonological change] (samasaṃskṛta); v/m and h/p functioning as allophones [as they do not in Sanskrit]; the ban on using Sanskrit indeclinables as nominal themes; the fact that [certain] conjunct consonants are prosodically weak [as they are not in Sanskrit]; and "violation" of caesura. (sūtra 342)
The verse is thought to be an interpolation, and surprisingly it seems to get a few things wrong, regarding the phoneme /ḷ/, for example.20 Nor does Kēśirāja ever use the phrase sati saptami (locative absolute) or discuss the syntagma as such (though absolute constructions certainly do exist in Kannada). Nor, lastly, is the "violation" of caesura anywhere discussed in the Darpaṇa. Nonetheless, the verse is an old one and, in its attempt to specify not only the principal Kannada distinctions but the very difference those distinctions make, it expresses what readers were likely to have felt about the grammaŕs procedures and Kannadás very character: if Kannada was constructed as a conceptual object from the perspective of a Sanskrit that defined literary language, the construction itself was intended to demonstrate heterogeneity and not homogeneity. Indeed, if all these admittedly abstruse reflections share any larger goal it may lie in forging a grammatical weapon from the materials offered by Sanskrit in order to defeat Sanskrit and preserve the local particular. And observe that this heterogeneity is intended transregionally, too: while a number of the linguistic features listed in the memorial verse, for example, are common to other Dravidian languages, nowhere does Kēśirāja ever comment on what makes Kannada similar to Tamil or Telugu, only how it relates, differentially, to Sanskrit.
17The Darpaṇás treatment of tadbhava and apabhraṃśa words, which specifies and organizes a wide range of sound changes, is a notable if imperfect attempt to find lawlike processes in the apparently lawless phonological behavior of "dialectal" or "corrupted" words. It also raises a set of fundamental questions about the method of vernacular philology as such. For when the proper use of apabhraṃśa words is said to depend on the observation of norms (lakṣaṇa) and due regard for idiom (lōkarūḍhi),21 the text is directing our attention toward the search for regulation and the basis of normativity presupposed by regulation. And it was the challenge of precisely this search that brought about a major transformation in the relationship between literature and grammatical theory that had dominated for a thousand years or more. To appreciate the startling discontinuity effected by the quest for vernacular normativity requires revisiting some main themes of the theory and practice of Sanskrit philology.
18Let us start by restating a fundamental postulate of Sanskrit grammatical theory: language norms are always-already given, and rules therefore precede any given instantiation: as Mīmāṃsā argues, language knowledge depends on abhiyoga, which is "knowledge derived from mastery of grammar," but the origins of grammar itself lie not in usage; until as late as the eighteenth century they were hold to be found in other "Vedic texts remembered" stretching infinitely back in time (smṛtiparaṃparā).22 Sanskrit grammars, accordingly, do not proceed by way of literary exemplification; in all of Patañjali’s vast Mahābhāṣya, literary works are cited only a dozen or so times, and never as standards of usage or proof-texts.23 In the Sanskrit thought-world, literature does not authorize grammar, nor does grammar adduce literature in order to establish usage. On the contrary, grammar itself authorizes literature, and literature was often produced—however contrary this may seem to modern readers—to illustrate grammar.
19We should guard against thinking that the priority of grammar to literature is some late sign of decadence, or that the texts illustrating the subservience of literature to grammar are mere scholastic exercises. Grammatical poetry marks Sanskrit literature from its very commencement, and continues unabated through the course of the cosmopolitan epoch. Such texts are available from across the historical spectrum. One of our earliest courtly epics, Aśvaghoṣa’s Saundarananda (c. 150 C.E.), has an entire chapter structured in part as a conjugational exercise of rare aorist forms. The Bhaṭṭikāvya (mid-seventh century), the most celebrated Sanskrit Rāmāyaṇa in Southeast Asia, is designed to illustrate the grammatical rules of Pāṇini (as well as the rhetorical rules of an as yet unidentified alaṅkāraśāstra). One of the most important historical texts of medieval Gujarat is Hemacandrás Dvyāśrayakāvya, the "Two-Purpose Poem" (c. 1200). One purpose is to provide a history of the Chaulukya (or Solanki) dynasty (especially Jayasiṃha Siddharāja and Kumārapāla), another is to illustrate the rules of his grammar, the Siddhahemacandra. Nor is it clear which of the two is primary.
20In the second place, the study of Sanskrit was never coterminous with the study of literature, as it was, for example, in Latin antiquity. Whereas the study of literature could scarcely forgo the study of grammar, the study of grammar could easily forgo the study of literature. In Sanskrit treatises on literary theory, the relationship between grammar and poetry was probed already in Bhāmahás work in the seventh century and exhaustively considered in Bhojás Śṛṅgāraprakāśa, which devotes the first 350 pages to the examination of signifiers and signifieds. Yet the consensus even among such literary theorists is not that poetry produces grammatical correctness, but rather that grammatical correctness (in part) produces poetry: "Words and meanings endowed with expression-forms and without faults" are the first words in perhaps the most famous definition of kāvya, Mammaṭás Kāvyaprakāśa (Light on Literature, c. 1050). To express the matter in Sanskrit terms, the lakṣaṇa, or norm, precedes the lakṣya, its embodiment, in large part because the norms themselves were eternal. Indeed, in explaining the very binary itself, "rule and instantiation," that was first introduced by Kātyāyana in defining language analysis (lakṣyalakṣaṇe vyākaraṇam), Patañjali largely dismisses the role of instantiations: "From the rules themselves people can grasp the words [that instantiate them]. "24
21Completely contrary to this is the empiricist procedure of the Kannadiga grammarian, and so far as I can tell, of all South Asian vernacular philologists (as well as the Apabhramsha grammar of Hemacandra). In the Śabdasmṛti chapter of the Kāvyāvalōkanam, the first Kannada-language grammatical work, literary examples are cited repeatedly (in the same authoŕs Sanskrit-language Karṇāṭakabhāṣābhūṣaṇa by a perhaps significant contrast, quotations are comparatively few and fragmentary).25 But the real force of the vernacular reversal in the vector of grammatical authority is felt in the Darpaṇa. Think of its opening invocation of the great poets of previous generations who are to be Kēśirājā́s authorities:
The expert Ways (sumārgam) of Gajaga, Guṇanandi, Manasija, Asaga, Candrabhaṭṭa, Guṇavarma, Śrīvijaya, Honna [= Ponna], Hampa [= Pampa], Sujanōttaṃsa —these provide the illustrative instances (lakṣya) in this work (v. 5).
The text is in fact as much an anthology of poetry as it is a grammar; some twenty poets and thirty different works are cited (unfortunately about half of them no longer extant), and virtually every rule in the grammar is illustrated with quotations from earlier poetry. However simple may be the feature under discussion, euphonic combination, for example, or case terminations, however undisputed the grammariańs judgment, a basis in literary usage must be adduced. In reference to domains beyond the descriptive reach of the grammar, as in the possible meanings of roots and verbal themes, the reader is advised to come to a determination after consulting the works of those who have achieved exemplary status (lakṣyasiddhiviḍiduvan ar.idu, sūtra 262). At the same time, it is quite possible that some of the proof-texts cited by Kēśirāja are his own; after all, he emphatically announces his status as poet at the beginning of the text, and he ends the work with an account of his literary productivity. The ultimate source of normativity, accordingly—and here we are at the farthest remove from Sanskrit—can in fact be located in the poet-grammarian himself. Kēśirāja can speak to this for himself: "Wherever he proceeds is the Way; however he undertakes to plant his step is the proper stance. What is inexplicable to Kēśava?… He alone is master of language norms in the world. "26
22The consequence of the imposition of norms through the new authority claimed by those in possession of literary excellence was to produce a literary language of notable conservatism and great uniformity. By the middle of the thirteenth century when Kēśirāja wrote, the Kannada language had already begun to change dramatically. At the most basic level, the distinctive phoneme /l./ (r.aḷa) had become obsolete and indistinguishable from the /r./ (śakaṭarepha). Even though the grammarian shows himself to be fully aware of this change by allowing the phonemes to function as rhyming consonants, he is insistent on preserving their individuality. (The great phoneme shift from /p/ to /h/, however, was already too far advanced to be reversed; as we have seen, Kēśirājā actually writes Hampa instead of Pampa.)27 As for uniformity, whereas undoubtedly a vast variety of dialects of caste and status must have been in use, and a certain amount of what is apparently dialectal variation is permitted in the grammar (such as optional lengthening of /a/ in the genitive, or in the accusative, or before vōl), these are, after all, very minor. The literary works upon which the Darpaṇa is based and the grammar itself by and large promulgate a literary Kannada that had become a regional, supra-dialectal code. Poets and intellectuals at the southern courts of the Gaṅgas and Hoysaḷas wrote according to precisely the same linguistic standards as their northern peers. Ranna, before releasing his Kannada campū, Sāhasabhīmavijaya, which he composed for a king at the northernmost Karnataka court of the Cāḷukyas, could have the work "evaluated by the leading men in the metropolis of the king of the Gaṅgas" in the southernmost Kannada kingdom.28
IV.
At issue in all this fervid philological activity is precisely this unification, standardization, and status-elevation of a literary language, something far more common across premodern vernacular traditions that is commonly realized. The new logic of language-boundedness at work here has something at its core akin to the new logic of spatial boundedness, as it is proclaimed in the Kavirājamārgam and adroitly reproduced in Pampás localized epic, though each has technologies specific to it.29 The former deploys grammars and dictionaries and metrics to discipline, purify, and correct, but above all include, exclude, and de-fine. Whereas in non-literary usage only Language, not languages, may be perceived to exist, in a sort of undifferentiated communicative continuum, segmenting that continuum is precisely one important function of the literary cultures of the vernacular millennium. In a very similar way, homogeneous space comes to be segmented and turned into places by such cultural-political practices as the issuing of inscriptions.30 Unifying—that is to say, creating—a language appears to comprise, in some importantly related ways, the unification and creation of a new type of space, where a particular kind of culture is in Place, as well as a particular, socio-textual community and, perhaps, a particular polity.
23Some four to five centuries from the moment it was first reduced to writing, Kannada embarked on a course of ever-accelerating "literarization," or regulation according to the norms of a superposed literary culture—as is fully visible in the epigraphical record—whereby a whole new set of texts and practices was brought into being. The Kavirājamārgam appropriated the discourse of the cosmopolitan Ways (vaidharbhimārga, etc.) for the vernacular sphere (in what, or so I have tried to suggest elsewhere, may in fact have been the reappropriation of a southern contribution to the discourse of cultural cosmopolitanism).31 It described the elementary forms of a vernacular philology, and established Kannada as a language of science even as it demonstrated that it could function as a language of literature. It turned space into place by mapping out the domain within which the new literature would circulate, and projected something of a community of readers/listeners. All these concerns are elaborated and refined in the following three or four centuries by a wide range of new developments: the local epicization of Kannada polity by Pampa (and Ranna and others I have scarcely mentioned), and the localization of the Sanskrit literary global through a wide range of works;32 the maturation of an ennobling philology in texts on metrics like Chandombudhi, and the continuing refinement of grammar from the Karṇāṭakabhāṣābhūṣaṇa to the Śabdamaṇidarpaṇa, which sought above all to identify and preserve Kannada difference. In asserting at once the regionality of Kannada and its literary value by associating the discourse on Kannada closely with that of Sanskrit, by localizing the Sanskrit global (of lexicon, meter, theme), and at the same time articulating what was thereby marked as Kannada distinctiveness, these discourses produce what I have called cosmopolitan vernacularism. This is a phenomenon that sometimes seems as much to index a new aesthetic and cultural sense of being as to refer to anything in the world. But the cosmopolitan vernacular certainly exists in the world, and has its own conditions of possibility.
24If the Kavirājamārgam and all the great works that followed in its wake can give us a vivid sense of the discursive and literary strategies by which a high-culture vernacular is produced, how are we to make sense of the time and the place of this transformation? Why did vernacular poets starting in the ninth-tenth century renounce what was not only a potential but an actual translocal, quasi-global audience for Sanskrit and for the first time begin to speak locally? Why did vernacular intellectuals, from within some of the most powerful courts in India at the time, decide to constitute their language as a new cognitive object and target of normative management? What, historically speaking, is the social and political content of the new cultural forms they were all creating, how were these shaped by, and how did they in turn shape, the polities to which they make constant reference and from the very centers of which they emerged?
25These questions are very complex in themselves, but any response to them has to take account of the larger environment in which these cultural-historical changes were occurring. For the transformation of vernacularity of which Kannada provides so dramatic an instance was itself a quasi-global phenomenon—new literatures were about to be called into existence everywhere, in South and Southeast Asia as well as in Europe. Of course, nothing dictates that all instances of vernacularization must share a single logic and fall under a single explanatory model. But this is something we cannot know prior to reconstructing something of its macrohistory, but one based, self-evidently, on the kinds of microhistories that South Asia preeminently enables us to discover.33
Bibliographie
References
Primary Sources
Adhvaramīmāṃsākutūhalavṛtti of Vāsudeva Dīkṣita. Edited by P. N. Pattabhirama Sastry. New Delhi: LalBahadurShastri Vidyapeeth, 1968-69.
Anantanāthapurāṇam of Janna. Edited by E. C. Devirappa and M. C. Padmanabhasarma. Mysore: Kannada Adhyayana Samsthe, 1972.
Chandombudhi of Nāgavarma. Edited by Kukkila Krishna Bhatta. Mysore: D.V. K. Murthy, 1975. Also edited by F. Kittel. Reprint: New Delhi: Asian Educational Services 1988 (first published 1875).
Karṇāṭakabhāṣābhūṣaṇa. Edited with an introduction by Lewis Rice. New Delhi: Asian Educational Services, 1985 (first published 1884).
Karṇāṭakabhāṣābhūṣaṇam of Nāgavarma. Edited by R. Narasimhachar. Mysore: Sarada Mandira, 1975. Original edition included in the Kāvyāvalōkanam, Mysore, 1902.
Kavirājamārgam. Edited by M.V. Seetha Ramiah. Bangalore: Karnataka Sangha, 1968, reprint 1994; Also edited by K. Krishnamoorthy. Bangalore: IBH, 1983.
Kāvyāvalōkanam. Edited by R. Narasimhachar. Mysore: University of Mysore, 1967. Third edition.
Mahābhāṣya of Patañjali. Edited by Franz Kielhorn. Pune: Bhadarkar Oriental Research Institute, 1962-72. Third edition, three volumes.
Mīmāṃsādarśana. 1970-76. Edited by K.V. Abhyankar and G. Joshi. Pune: Anandashrama Press. Five volumes.
Śabdamaṇidarpaṇa of Kēśirāja. Edited by T.V. Venkatachala Sastry. Bangalore: Kannada Sahitya Parishad, 1996. (All references in the text are to this edition). Also edited by A. Shanker Kedilaya, Madras: University of Madras, 1973. Madras University Kannada Series No. 5.
Vijayanagara Inscriptions. Edited by B.R. Gopal. Mysore: Directorate of Archaeology and Museums, Government of Karnataka, Four volumes. Centenary Publication. no. 3, 1985-.
Secondary Sources
Kahrs, Eivind G. 1992. "What is a Tadbhava Word?" Indo-Iranian Journal 35: 225-49.
Kulli, J. S. 1984. "Kannada Grammars and Sanskrit Tradition." International Journal of Dravidian Linguistics 13,1: 40-47.
Masica, Colin. 1991. The Indo-Aryan Languages. Cambridge, England: Cambridge University Press.
Master, Alfred. 1943. "The Mysterious Paiśācī." Journal of the Royal Asiatic Society: 34-45, 217- 33.
Nayak, H. M., and T. V. Venkatachala Sastry, eds. 1974-. Kannada Sāhitya Caritre. Mysore: Kannada Adhyayana Samsthe. Five volumes published to date.
Pancamukhi, R. S. 1941. Karnatak Inscriptions. Dharwar: Kannada Research Institute. Two volumes.
Pollock, Sheldon. 1985. "The Theory of Practice and the Practice of Theory in Indian Intellectual History." Journal of the American Oriental Society 105: 499-519.
______. 1989. "Playing By the Rules: Śāstra and Sanskrit Literature." In A. L. Dallapiccola and S. Zingel-Avé Lallemant, eds., The Śāstric Tradition in the Indian Arts. Wiesbaden: Steiner.
______. 1997. "́Traditioń as ´Revelatioń: Śruti, Smṛti, and the Sanskrit Discourse of Power." In S. Lienhard and I. Piovano, eds. Lex et Litterae: Essays on Ancient Indian Law and Literature in Honour of Oscar Botto. Turin: Edizioni delĺOrso.
______. 1998. "The Cosmopolitan Vernacular." Journal of Asian Studies 57: 6-37.
______. 2003. Viśvātmaka Deśabhāṣe, tr. Akshara K.V. Heggodu (Karnataka): Akshara Prakashana. DR Nagaraj Nenapina Akshara Chintana Maale.
______, ed. 2003. Literary Cultures in History: Reconstructions from South Asia. Berkeley, London: University of California Press.
Ronsard, Pierre de. 1993. Oeuvres complètes. Paris: Gallimard
Rao, Velcheru Narayana. 2003. "Multiple Literary Cultures in Telugu: Court, Temple, and Public." In Pollock, ed. Literary Cultures in History.
Scharfe, Hartmut. 1977. Grammatical Literature. Wiesbaden: Harrassowitz. A History of Indian Literature vol. 5, fasc. 2.
Velankar, H. D. 1949. Jayadāman. Bombay: Haritosha Samiti.
Venkatachala Sastry, T. V. 1992. "A Historical Survey of Dictionaries Sanskrit and Kannada" Arts Journal [University of Mysore] 54: 163-76.
Yashaschandra, Sitanshu. 2003. "From Hemacandra to Hind Svarāj: Region and Power in Gujarati Literary Culture." In Pollock, ed. Literary Cultures in History.
Notes de bas de page
1 See Pollock1998 for details.
2 Possible linkages have been suggested between the Śabdamaṇidarpaṇa (henceforth ŚMD) and the Tamil Naṉṉūl, or between the Karṇāṭakabhāṣābhūṣaṇa and the Āndhraśabdacintāmaṇi (though the latter work, long ascribed to the eleventh-century poet Nannaya, is now viewed by some scholars as a mid-seventeenth-century text, see Rao 2003), but few in-depth historical studies have been undertaken.
3 See Pollock 1985, 1989, 1997.
4 Tantravārttika in Mīmāṃsādarśana 1.3.12, p. 164 lines 8-15, rearranging slightly the verse and the prose gloss. Already by Kumārila’s time Dākṣinātyā (assuming this refers to Maharashtri) and perhaps even Māgadhī had been grammaticized, as in the version of the Prākṛtaprakāśa commented on by Bhāmaha (Scharfe 1977: 192). Accordingly, there would be no paradox in Kumārilás speaking of their "corrupt forms."
5 Compare for example Ronsard: "Je te veux encores advertir de n’escorcher point le Latin, comme noz devanciers, qui ont trop sottement tiré des Romains, une infinité de vocables estrangers, veu qu’il y en avoit d’aussi bons en nostre propre langue" (I want to advise you again not to overdo Latin, like our predecessors, who quite recklessly adopted an infinite number of foreign words from the Romans. For there are perfectly good words in our own language) (1993: 1187).
6 These are collected in Epigraphia carnatika, vol. 2.
7 The extant portion actually begins, kuḷir ene śaityam, "The [Kannada word] kuḷir means [Sanskrit] śaityam [cold]," similarly v. 5, "beḷaku means dīpa [lamp]," v. 6, baḷi means vaṃśa [bamboo/lineage]."
8 These are: Amarasiṃha, Bhāguri, Śāśvata and the Halāyudha (a scholar at the court of Rāṣṭrakūṭa king Kṛṣṇa III [r. 939-67], on whom more below).
9 On Kannada lexicons see Nayak and Venkatachala Sastry 1974 ff., vol. 3: 733-35; Venkatachala Sastry 1992.
10 Chandombudhi vv. 12 [4] (the name of the village is Veṅgipal.u); vv. 27-8. See also Kāvyāvalōkanam ed. Narasimhachar 1967: 14-15. Krishnabhatta’s edition restores the opening verses in praise of Śrī, Śiva, Vināyaka, Durgā, the Sun, and Bhāratī, which had been intentionally suppressed by Kittel in his editio princeps. Nāgavarma refers to the learned in v. 8 [2- (ballar) and his work as apūrvam āge kṛtiyam (v. 8 [2]) and aśeśavidvajjanahitamam (v. 29 [Kittel 11 differs]). "In which the ways of Place… ," dēsiye dēsevettudem… prabandhamam (v. 11 [16].).
11 Halāyudha’s commentary on the Chandaḥsūtra was written at the court of Muñja of the Paramāras, where the metrician emigrated perhaps in consequence of the weakening of Rāṣṭrakūṭa power with the death of Kṛṣṇa.
12 For specifics on Jayakīrti see Velankar 1949.
13 See Chandombudhi p. 10 in the edition of Krishnabhatta (the bracketed portion is available in a number of mss.), Kittel 67, p. 22; cf. p. 21 for "daughter languages." In the introduction to his Karṇāṭakabhāṣābhūṣaṇa (p. iv) Rice translates, "Born in the three and a half languages… are the Dravida (Tamil), Andhra (Telugu), Karnataka (Kannada) and others." So Master: "There will be the fifty-six varieties of language, Tamil, Telugu, and Kanarese, etc., which originate from the three and a half languages" (1943: 44). The correct analysis of this passage is owing to T. V. Venkatachala Sastry (personal communication).
14 This is based on the evidence of his recently discovered campū, the Vīravardhamānapurāṇa, which is dated precisely to 1042. In his Anantanāthapurāṇam (1230) Janna refers to Kavitāguṇōdaya Nāgavarma as Jagadekamalla’s kaṭakopādhyāya. (The second half of Janna’s verse is to be understood as referring to "the kaṭakopādhyāya of the present day "—that is, under the Hoysaḷa king Narasiṃha II, r. 1220-35—i.e., Sumanōbāṇa, the teacher of Janna (and father-in-law of Mallikārjuna, see the following note). Nāgavarma II is also author of the Abhidhānavastukōśam referred to above.
15 ŚMD v. 341. He attributes the same title to his maternal grandfather: "I am the poet Kēśava, the glorious grandson of Sumanōbāṇa, who was a poet and kaṭakācārya of the Yādavas; and the son of Cidānanda Mallikārjuna, supreme master of yoga" (v. 2). Mallikārjuna was a praśasti poet of the Hoysaḷa dynasty and literary anthologist. His grand eulogy of the Hoysaḷavaṃśa is carved on the walls of the Mallikārjuna temple in Basarāḷu (EC 7: 211 ff. nos. 29 and 30, dated 1234 and 1237). His Sūktisudhārṇava (Nectar Ocean of Poetry), prepared for the Hoysaḷa king Vīrasōmēśvara (r. 1234-54) (see 1.24 and colophon) is the foremost anthology of early Kannada literature; it also reproduces a number of the praśasti verses.
16 See Kāvyāvalōkanam 4.23, and Anantanāthapurāṇa 1.34, where Nāgavarma is called "a present-day Śarvavarma" (idānīṃtana śarvavarma; a similar title was given to Śākaṭāyana two centuries earlier). Note that 34 of the 280 sūtras of the Bhāṣābhūṣaṇa (and 17 of the 97 sūtras of the Śabdasmṛti) are borrowed or translated from the Kātantra (Kulli 1984: 41).
17 So Akho (Yashaschandra 2003: 581). I am familiar with no reference to Kannada as "Prakrit" aside from a fragmentary inscription of 700 (ataḥ paraṃ prākṛtabhāṣayā padyāny etāni dattāni suṅka <…> bādāvi <…>, IA 10: 60), but nothing about the identity of this "Prakrit" language can safely be inferred from the passage (though Panchamukhi 1942, vol. 1: 3 believes it to be Kannada).
18 Neither Kēśirāja nor anyone else in the premodern period ever phrases the matter in precisely this fashion, however. At any event, the words in the tatsama category are all clearly Dravidian in origin.
19 The strictures on compounds of incompatibles such as mukhatāvare (in contrast to mukhapadma) are found at ŚMD sūtra 185 (where in the prayoga the term arisamāsa is used) and cf. 90 prayoga (where KRM 1.59 is also cited). (The apparently general requirement of arisamāsa in Karnāṭakaśabdabhūṣaṇa sūtra 132, p. 5, has to be understood as a restrictive one applying only to those Sanskrit words that do not undergo tadbhava transformation.) On tadbhavas see also ŚMD 266-67 and vṛtti; tatsama is described in 312, samasaṃskṛta in 90. From the analysis in the Darpaṇa one may deduce that, pace Kahrs 1992 (especially p. 245), who implies it to be Jain dogma that tadbhava "really" means "existing [eternally] in that [i.e., Sanskrit]," this was certainly not the case for the Kannada grammatical tradition, Jain though some of its representatives (including Kēśirāja) were. Nor were such words considered prākṛta "only insofar as they are not subject to the regularizing rules that govern the eternally existing forms of Sanskrit." On the contrary, implicit in the Kannada tradition, as the very preference for the term apabhraṃśa implies, is a conception of some sort of historical change; in addition, the ruleboundedness of the derivation of tadbhava lexemes is entirely obvious in Kēśirāja’s exposition.
20 If I understand matters correctly, this /ḷ/, called kṣaḷa, which in Kannada represents Sanskrit /l/, is not in fact included among the five phonemes of Place belonging to "pure Kannada" (accagannaḍa) that are introduced by Kēśirāja when describing the phonemic difference of Kannada over against Sanskrit (these five are /r./, /l./, /ḷ/ [the kuḷa], /ĕ/ and /ŏ/). Rather it is one of the ten that are peculiar to Sanskrit (along with vocalic /r/ and /l/ both light and heavy, palatal and retroflex /s/, the three types of visarga; see sūtras 29, 43).
21 ŚMD sūtra 266: īkṣisi śikṣāsūtrada lakṣaṇamaṃ lōkarūḍhi kiḍadavol.
22 Pūrvamīmāṃsāsūtra 1.3.24 and Adhvaramīmāṃsākutūhalavṛtti vol. 1: 93.
23 Literary usage seems not to be cited as authoritative for Sanskrit grammar until very late, perhaps not before the early seventeenth-century Prakriyāsarvasva of Nārāyaṇa Bhaṭṭa (Scharfe 1977: 174 n.).
24 sūtrata eva śabdān pratipadyante, Mahābhāṣya vol. 1: 12.15 ff.
25 Cf. Nayak and Venkatachala Sastry 1974 ff., vol. 3: 718 and n. 509 The seventeenth-century commentary on Bhāṣābhūṣaṇa, however, cites profusely.
26 ŚMD 338: naḍedude mārgam padaviḍal oḍarisidude bhaṅgi kēśavaṅgariduṇṭe… |… tān e lōkadoḷ lākṣaṇikam | | See also sūtra 2 and for his works sūtra 339.
27 See sūtra 170 (the shift was already flagged in the Bhāṣābhūṣaṇa, sūtra 115).
28 Sāhasabhīmavijaya 1.40 (vastupuruṣar belemade).
29 See Pollock 1998.
30 An additional and much longer article would be required to analyze the differentiating functions of regionalized scripts. A distinctive Kannada form was being developed already by the seventh century and quickly became dominant in most areas, although practices in border zones remained fluid. As late as the sixteenth century Kannada could occasionally be written in Telugu or even Nagari (and Telugu in Kannada or Nagari), as the corpus of Vijayanagara inscriptions shows (Gopal 1985 ff.). The quest for difference seems to me far older than my colleague Colin Masica, for example, suggested (1991: 144).
31 Pollock 1998: 21 ff.
32 These include such works as the Karṇāṭaka Kumāra saṃbhava of Asaga in the late ninth century and the Karṇāṭaka Mālatīmādhava of Kannamayya in the eleventh (neither of which is extant), Nāgavarma’s Karṇāṭaka Kādambarī in the eleventh, Durgasiṃha’s Kārṇāṭaka Pañcatantra in the twelfth.
33 This is the subject of part 2 of my forthcoming book The Language of the Gods in the World of Men, from which the present essay is extracted (a somewhat longer version of the essay itself appears in my Kannada collection, Pollock 2003).
Auteur
Sheldon Pollock (b. 1948) is George V. Bobrinskoy Professor of Sanskrit and Indic Studies at the University of Chicago. He was educated at Harvard (ABmcl Classics, 1971; Ph.D., Sanskrit and Indian Studies, 1975) and in India (Sanskrit with P.N. Pattabhirama Sastry, among others; Kannada with T. V. Venkatachala Sastry). He recently edited Literary Cultures in History: Reconstructions from South Asia (Berkeley 2003), and, with Homi Bhabha et al., Cosmopolitanism (Duke 2002). His Language of the Gods in the World of Men is due out in late 2004 from Berkeley. Currently he directs the international research project "Indian Knowledge-Systems on the Eve of Colonialism."
Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
La création d'une iconographie sivaïte narrative
Incarnations du dieu dans les temples pallava construits
Valérie Gillet
2010
Bibliotheca Malabarica
Bartholomäus Ziegenbalg's Tamil Library
Bartholomaus Will Sweetman et R. Ilakkuvan (éd.) Will Sweetman et R. Ilakkuvan (trad.)
2012