J Indian Philos (2007) 35:227251 DOI 10.

1007/s10781-007-9014-8

a: Jayadeva P Unraveling the Ka vyapraka s yu s .avars .as idiosyncratic sequence of topics in the Candra loka
David Mellins

Published online: 4 July 2007 Springer Science+Business Media B.V. 2007

a Abstract In his twelfth century alam ras stra, the Candra loka, Jayadeva : ka P yu s vyapr.avars .a reverses the sequence of topics found in Mammat :as Ka a, an earlier and immensely popular work. With such a structural aka s revisionism, Jayadeva asserts the autonomy of his own work and puts forth an ambitious critique of earlier approaches to literary analysis. Jayadeva investigates the technical and aesthetic components of poetry in the rst part of the Candra loka, prior to his formal semantic investigations in the latter half of the text, thus suggesting that aesthetic evaluations of poetry benecially inform scientic investigations of language. Jayadevas organization of his chapters on the semantic operations, moreover, intimates that the study of suggestive and metaphoric functions of language claries our understanding of denotation, which is conventionally understood to be the primary and direct path of verbal designation. a Keywords Jayadeva Candra loka Alam ras stra Sanskrit poetics : ka a Sabdavr Indian poetics Poetics Mammat vyapraka s . tti Indian :a Ka semantics Semantics Indian language theory Language theory abhidha Denotation Metaphor laks upaca ra .an .a

a Introduction: in the shadow of the Ka vyapraka s

D. Mellins (&) 720 Fort Washington Ave., Apt. 5G New York, NY 10040, USA e-mail: [email protected]

123

228

D. Mellins

pedagogical manual on alam sa ra stra (Sanskrit poetics), but rarely : ka acknowledged to be a serious or innovative work. The renowned historian of Sanskrit poetics, S. K. De, writing in the rst quarter of the twentieth century, surmises: Jayadevas Candra loka has been a deservedly popular manual, but in spite of its clearness and brevity of exposition and aptness of its illustrations, it is nothing more than a convenient epitome, its most remarkable feature being its detailed treatment of poetic gures, which occupy nearly half its bulk.2 a Na rayan str Khiste, the editor of the rst Chaukhamba edition of the .a S Candra loka, writes in 1929: There is no other work, as far as we know, which may be condently placed in the hands of ardent beginners as a safe guide to the study of Sanskrit Poetics on old traditional lines.3 Given these pronouncements, it is surprising to discover a statement regarding the Candra ma m loka by the formidable 17th century M : saka (Vedic Hermeneutician) Ga ga bhat t a: :: air yan na paricitam adya pi gahanam anekaih : kles : par ha sas tasmin mama bhavati yatnasya, tadapi oda alya yas kaus tanayapadabha va hitamana h : 4 a prasa dah na m : kles : phalati phalam atyutkat :am api. Even today, this difcult text is not properly understood despite many efforts. Some will thus laugh at my effort here. Still I will place my mind at the feet of Kr ma. .s .n . a and Ra Grace accomplishes the most arduous of goals. One might be tempted to attribute the disparity of these evaluations to the different critical standards of their respective eras. Yet even in pre-colonial periods, the Candra loka was rarely appreciated for its innovation, and instead assumed to follow the model of Mammat sa, an alam sa vyapraka  ra stra :as Ka : ka

1 An alias that distinguishes the author of the Candra loka from the author of the G tagovinda bearing the name same. At the conclusion of each chapter, the author of the Candra loka states his fathers and mothers names to be Maha deva and Sumitra , further distinguishing his personage from the more famous devotional poet who identies his parents as Bhojadeva and Ra ma dev ( sr sr sara sr bhojadevaprabhavasya ra ma dev suta jayadevakasya / para  dipriyavargakan g tago: :the  vindakavitvam astu // GG, 12.22). 2 De (1925: 248249). 3 Khiste (1929: 2). 4 (Prologue). RA

123

a Unraveling the Ka vyapraka s

229

treatise that had attained near canonical status by the middle of the second millennium.5 Structural distinctions In certain ways this assumption is understandable. Both the Ka sa vyapraka  and the Candra loka are divided into 10 chapters, which, excluding the rst, investigate a more or less distinct category of poetic analysis. The terms designating the chapters of both works refer to forms of lightthe Ka vyapraka sa is divided into ten ulla sas (ashes) and the Candra loka ten mayu sa, the Candra khas (rays). Like the Ka vyapraka  loka is highly systematic in format, detailing all the important topics in alam sa ra stra, and : ka like the Ka sa, the Candra vyapraka  loka in a departure from the Dhvanya loka and the Locana investigates both rasa and alam ra (gures : ka of speech) as independent sources of charm in poetry. sa and Still a careful comparison of the structures within the Ka vyapraka  the Candra loka reveals that Mammat :a and Jayadeva have distinct approaches to the study of poetry and different estimations of its most essential import. Examining the chart in Fig. 1, we can see that although the two texts cover more or less the same span of topics, the partitioning and sequencing of these subjects are quite different in the two texts. The rst chapters of both works are indeed remarkably similar, despite some additional topics in the Candra loka. Yet the parity stops here. In the subsequent four chapters, Mammat sab:a discusses topics related to the  dav:rttis, or semantic operations. In chapter two he denes abhidha ~ (denotation), laks a n a (metaphorical indication), and vya n jana (literary : : suggestion).6 In chapter three, he demonstrates that denoted, indicated

Vamanacharya Ramabhatta Jhalakikar, the 20th century authority on Mammat :a, lists 46 commentaries and sub-commentaries to the Ka sa in the introduction to his own, the vyapraka  Ba ga bhat labodhin . Ga :t :a, in his comment to Jayadevas statements of the purpose of poetry in m:rtapadena prabhusam Candra satraividhye loka 1.3, states atra nta sam : mitasuh:rtsam : mitaka : mitopade sva py ude sa(read upade sa)dvayasya  sabda sayatva sasya rthpradha nataya da vi t ka nta sam : mitopade 1.3), clearly super sva so vyajyate (RA rasapradha sayatya d am:rtatulyataya t:rtiyopade nataya davi imposing Mammat sa 2: vyapraka  :as analogy in Ka ase rthakr ivetaraks ka vyam ravide s : yas . te vyavaha .ataye pades ayuje sadyah nta sam : paranirvr : mitatayo . taye ka Ga ga bhat :tas argument is used to support his contention that Jayadeva truly intends poetry to have the six purposes enumerated by Mammat :a, rather than the three that Jayadeva lists in Candra loka 1.3. While this argument is not without merit, it requires a metaphorical interpretation of Jayadevas verse. 6 nandavardhana A tertiary power of speech, which many Sanskrit rhetoricians (most notably A and Abhinavagupta) contend is distinct from denotation and metaphorical indication. It was granted an exalted literary status on account of its unique capacity to evoke rasa in poetry.

123

230

D. Mellins

and suggested meanings each have the capacity to suggest subsequent meanings. In chapter four he describes the varieties of dhvani as determined by different types of vya~ njana ; and in chapter ve he discusses _ gya, poetry where suggested meaning exists but is subordigun bhu tavyan : nate to the literal meaning, and proceeds to describes its sub-varieties. Jayadeva, on the other hand, does not discuss the  sabdav. rttis until the second half of the Candra loka. In the rst half of the treatise he describes topics related to the technical evaluation of language, topics that were the principle concern in treatises from the early period of alam sa ra stra : ka dos : a (poetic blemish) in chapter two, laks : an : a (poetic distinction, to be distinguished from the  sabdav:rtti or semantic power termed laks ) in : an :a chapter three, gun ra (gures : a (poetic virtue) in chapter four, and alam : ka of speech) in chapter ve.7 Mammat :a does not discuss these topics until the second half of the Ka sa. He discusses dos vyapraka  : a in the seventh chapter, gun sabda lam ra (acoustic gures of speech) in : a in the eighth,  : ka the ninth, and artha lam ra (semantic gures of speech) in the 10th. He : ka does not discuss laks sa . vyapraka  : an : a (poetic distinction) at all in the Ka Jayadeva seems to be the only a lam rika to discuss laks : ka : an : a, a distinction of language elaborated in the Na sa :tya stra and other dramaturgical works but disregarded in other alam ra treatises.8 Another peculiar distinction is : ka found in the sixth chapters of these two works. Mammat :a discusses citraka vya, a type of poetry that lacks appreciable suggestion and rasa, in the
7

Jayadeva characterizes these features of poetry as follows: sya c ceto vis yena saks raman yata .ata .ata . , 2.1). abde rthe ca kr s : dos . tonmes .am .am udghos .ayanti tam (CA When it enters the heart, the charm in poetry is damaged. Manifesting in sound and meaning, the experts call it poetic blemish. hur mahars itya di laks ri ka vyasya : bhu : .an . am .ayah 3.11). svarn jis latvaprabhr va mah bhujah :h :n : (CA . abhra .h . ubha . t Great sages say that such features are the distinctions (laks : an : a) of poetry, Just like signs such as a brilliant golden forehead are the distinctions of a king. a gun aurya am das h vye pum dayo yatha : ka : si s .a _ tilaka dyam iva str n m : vidagdhahr . dayangamam .a , 4.10cd 11). _ ram vyatiriktam alanka s m (CA : prakr : gira . ter bhu .an . am These ten are the virtues (gun : a) within poetry just like courage and other qualities are virtues within man. The gures of speech (alam ra) are the ornament of speech and are distinguished from its : ka natural state. These are thrilling, like the forehead ornaments of women, to the connoisseur.

The Na sa san _ gha sobha :tya stra enumerates 36 lak: ta,  , uda haran : as: bhus : an : a, aks : arasan : a, hetu, saya, tulsam saya, d:rs san ses tipa ta, ati nta, pra pti, abhipra ya, nidar :a : ::ta : a, nirukta, siddhi, vi : an : a, gun yatarka, padoccaya, d:rs sa, anunaya, ma ra, viparyaya, bhram la , da ks ::ta, upadis : :ta, vica : : in : ya, garhan : a, artha sa, ks rtana, anuktasiddhi, and patti, prasiddhi, p:rccha , sa rupya, manoratha, le : ak : obha, gun , 13070). Some of these ati priyokti (NS saya, le sa and hetu are categorized as alam ras in : ka early alam sa mahas Ka ra stra treatises such as Bha vya lam ra. : ka : ka

123

a Unraveling the Ka vyapraka s

231

Fig. 1 Chapter topics in the Ka sa and the Candra vyapraka  loka

sixth chapter of the Ka sa. Jayadeva, on the other hand, discusses vyapraka  rasa itself in the sixth chapter of the Candra loka. a Organization of the Ka vyapraka s To a considerable extent, the organization of topics in the Ka sa is vyapraka  over-determined by Mammat :as threefold gradation of poetry at the end of the rst chapter, which states:

123

232

D. Mellins

_ ayini vyangye idam uttamam atis va cya d dhvanir budhaih : kathitah : _ _ i gun ata dr bhu tavyangyam tu madhyamam .s . : vyangye 9 _ abdacitram s cyacitram avyangyam . tam : va : tv avaram : smr Poetry is of the superlative grade when the suggested meaning predominates over the literal. Scholars call such poetry dhvani. Poetry of the middle grade, where the suggested meaning is otherwise (subordinate), is _ gya. The lowest grade of poetry, which lacks called gun bhu tavyan : appreciable suggested meaning, is called either acoustic citra or semantic citra. Mammat sabdav:rttis in the second and third chapters of the :a discusses the  Ka sa because it is essential that his audience gain a preliminary vyapraka  understanding of abhidha (denotation), laks (metaphorical indication) : an :a and vya~ njana (literary suggestion) prior to his discussion of the highest grade of suggestive poetry, dhvani, in the fourth chapter. This discussion of the  sabdav:rttis similarly informs his explanation of the middle grade _ gya) in chapter ve. However, of suggestive poetry (gun bhu tavyan : Mammat vya (the lowest grade of Ka vya, lacking any :as discussion of citraka signicant trace of suggested meaning) in the sixth chapter precipitates a shift in the orientation of his discourse. In subsequent chapters, he analyzes technical properties of speech such as dos : a (blemish), gun : a (poetic virtue) and alam ra (gures of speech), utilizing categories and methods estab: ka lished during the pre-dhvani era of Sanskrit poetics. In his auto-comment to the nal verse of the introductory chapter, Mammat :a had stated: citram iti gun lam rayuktam.10 .a : ka citraka vya is suitable on account of gun ra : a and alam : ka indicating that the charm of citraka vya results from the auspicious choice of words and skillful applications of gures of speech, rather than suggestive and emotive properties in the meaning. The sequencing of topics in the Ka sa to a certain degree reects vyapraka  miri the dhvani-centric vision of poetry Mammat :a had inherited from his Kas nandavardhana and Abhinavagupta. Herein, the function of predecessors, A properties such as gun ra is oriented and indeed stipulated by the : a and alam : ka more essential functions of rasa and dhvani. Thus Abhinavagupta states in the Dhvanya loka: ritya vinives anam rasabha va dita tparyam a s alam na m sa m alam ratvasa dhanam11 . t : kr : sarva : ka All gures of speech gain their authority from the fact that their guration is established in relation to the rasas, the sentiments and the like.
9

KP (4cd 5). KP (vr . tti to verse 5). (vr DhvA . tti to 2.5).

10 11

123

a Unraveling the Ka vyapraka s

233

Similarly, he explains that certain poetic blemishes are faulty in relation to certain rasas: rutidus ita s dayo dos anitya ye ca dars h .t :a .a : 12 _ re te heya . dhvanya tmany eva s rnga ity uda hr h . ta : The faults such as indelicacy in sound are shown to be relative.13 These are to be abandoned only in the case of the erotic rasa when this is brought to life by dhvani.14 nandavardhanas view, gures of speech are literary devices that ornaIn A ment the function of rasa; blemishes are the unfortunate features obstructing this. Their gurativeness or faultiness does not independently pertain to the sound or meaning constituting the body of poetry, but rather their capacity to expedite or damage specic rasas. nandavardhana, Mammat Emulating A :a also correlates his discussion of gun a , do s a and ala m ka ra to the function of rasa and poetic suggestion. He : : : acknowledges, however, that such poetic attributes to a limited degree have an independent literary signicance. Regarding dos : a (blemish) he states: ca mukhyas tada raya mukhya rthahatir dos s d va cyah .o rasas : 15 ubhayopayoginah syuh s abda dya s tena tes v api sah . : : : Blemish is that which damages the main purport. The purport may be a rasa, but it can also be a literal meaning associated with rasa. Blemish can also exist in the words and phonemes etc., as these are what expedite the rasas and the literal meaning. While dos : a (blemish) principally pertains to a rasa, it may secondarily be analyzed with respect to the literal meaning associated with a rasa or with respect to the words and sentences constituting a rasa or associated literal meaning. Similarly, in regard to alam ra, Mammat : ka :a pronounces: _ upakurvanti tam ren tucid . a ja : santam : ye ngadva 16 ha ra divad alam ka ra s te nupra sopama dayah : :

12 13

(2.11). DhvA

anitya h : . Literally, impermanent. In this context, it conveys that certain faults are damaging only to specic rasas, and are not destructive in all cases of poetry.

14

More literally, they are to be abandoned only in the cases of erotic rasa, when this has dhvani as its soul (a tma). KP (49). KP (67).

15 16

123

234

D. Mellins

Sometimes gures of speech such as alliteration and simile serve a dominant rasa17 by ornamenting the body of poetry (speech and meaning), in the manner of necklaces and other jewelry. Even in cases when alam ra assists a dominant rasa, Mammat : ka :a acknowledges that alam ka ras may still ornament the body of poetry as constituted by : speech and meaning. In his auto-comment to this verse, Mammat :a species that gures of speech do at times function independently from rasa: ktivaicitryama yatra tu na sti rasas tatro traparyavasa yinah : . kvacit tu san pakurvanti.18 tam api no But where a dominant rasa does not exist, gures of speech bestow beauty to a poetic expression. Even in some cases where a dominant rasa exists, there can be gures of speech, which do not come to its service. Such statements clearly depart from the uncompromisingly dhvani-centric ideology espoused in the Dhvanya loka. Mammat :a implies that in superior grades of poetry, aesthetic properties such as alam ra and gun : ka : a do expedite the function of a predominant rasa, but additionally recognizes the utility of analyzing these technical features as components of speech and meaning. Such an understanding is supported by the organization of his work. Investigating _ gya he dos ra subsequently to dhvani and gun tavyan : a, gun : a and alam : ka : ibhu clearly suggests the primacy and centrality of rasa and dhvani in poetry. However, by isolating the technical investigations of dos ra : a, gun : a and alam : ka in the latter half of the Ka sa, after his explication of citraka vyapraka  vya, Mammat :a implies that these investigations serve an independent if secondary purpose within the analysis of poetry.

Structure of the Candra loka Jayadeva, in his own distinct sequencing of topics in the Candra loka, rebels more forcefully against the hegemony of rasa/dhvani theory, striving to resuscitate a poetic discourse centered on the analysis of alam ra. Investi: ka gating gun a , do s a and ala m ka ra in the rst half of his treatise, prior to the : : : mention of dhvani and rasa, Jayadeva indicates that these constitute
17 Mammat tra claries that the pronoun tam refers to mukhyam :as auto-comment to this su : rasam : (dominant or primary rasa):

_ tis ayamukhena mukhyam ye va cakava cyalaks nga : rasam : sam : bhavinam upakurvanti te .an .a _ na ar lamka kan dyanga m utkars dha nadva ren rin raka ha ra daya iva ra h :ha : .t .a .a s . o py upaka (KP, vr . tti to 67). Those things, which come to the service of a dominant rasa by bestowing an excellence to its body, i.e. its literal speech and meaning, are gures of speech. They are similar to necklaces and other ornaments, which become servants to the self by bestowing excellence to limbs such as the neck.
18

KP (vr . tti to 67).

123

a Unraveling the Ka vyapraka s

235

a Fig. 2 Thematic sequences in the Candra loka and the Ka vyapraka s

independent properties of speech and meaning, rather than subsidiary properties of rasa or dhvani. While this prerogative indicates the primary motivation for the unusual structure of the Candra loka, I would like to suggest a more theoretically driven purpose to Jayadevas topical organization. Before I present my hypothesis, I would like to re-examine the structure of the Candra loka, applying more general categories, as exhibited in the chart in Fig. 2. If we exclude the introductory subjects of the rst chapter, we can divide the topics of the Candra loka into three general classes: the rst constitutes the technical examination of poetry which includes dos ra; the second constitutes the : a, laks : an : a, gun : a, and alam : ka sentimental, psychological and cognitive analysis of poetry which includes _ gya; and the third constitutes the rasa, vya~ njana bhu and gun tavyan : _ gya semantic analysis of poetry which shares vya~ njana bhu and gun tavyan : with the previous class, and additionally includes laks a n a and abhidha . : : The chapter topics in the Ka sa can be categorized into three vyapraka  more or less identical classes, which are ordered in exactly the opposite sequence as those in the Candra loka (see Fig. 2): the semantic investigation taking place in chapters 2 and 3, followed by the sentimental and psychological investigation in chapters 46 and the technical examination in chapters 610. The rst question that arises is whether this retro-gradation in sequence is purely adventitious or whether Jayadeva intentionally structures the Candra sa. I would strongly advocate loka in a contrary fashion to the Ka vyapraka  the latter view since there is signicant evidence suggesting Jayadeva was critical of Mammat loka, Jayadeva :a and his texts. In verse 1.7 of the Candra states: _ karoti yah abda ang vyam rtha v analam . t : ka : s : kr asau na manyate kasma d anus n am analam kr 19 .. : . t If he concedes that poetry can be speech and meaning lacking gures of speech, should not that wise scholar also believe in a re without heat.
19

(1.8) CA

123

236

D. Mellins

This verse clearly mocks Mammat sa vyapraka  :as denition of poetry in Ka 4ab, which states: abda tad ados rthau sagun v analam punah pi.20 .au s .a : kr . t : kva Poetry is a complex of speech and meaning that lacks blemish (dos : a), possesses virtue (gun a ), and that sometimes is anala m k r t , i.e. devoid of : : : distinct gures of speech. Jayadeva puts forth a more subtle, though perhaps even more deprecating jest in verse 1.2: ham n : ho cinmayacittacandraman . ayah : sam : vardhayadhvam : rasa re re svairin rakavite ma sma praka s bhavah . i nirvica : ulla sa ya vica rav cinicaya lam rava ra m : ka : nidhes candra lokam ayam yu s 21 .avars .ah . t : svayam : vitanute p : kr Oh you with moonstone hearts!22 Expand the rasas! but you, unfaithful wench! Undiscerning poesy, do not reveal yourself! The skillful P yu s .avars .a creates his own moonlight in order to expand the ocean of alam ra that has discriminating wisdom : ka as its waves. Here, I direct the readers attention to the cvi formation at the end of pada bma s bhavah sma praka  : (do not reveal yourself). Through double entendre ( sles sa. On its vyapraka  : a), this expression may also refer to the Ka own, such a lexical congruence might simply be a case of ghuna ks ya, or : aranya random coincidence, but if we look at the beginning of the pada, we nd the word ulla sa ya, which intimates the designation of the chapters in the Ka saulla vyapraka  sa. Similarly, the very next word is a compound beginning with the term vica ra, which may or may not call to mind Mammat :as treatise abdavya on semantics, the S paravica ra. Given the concurrence of these words, there is good reason to believe this verse is yet another jest at Mammat :a, in which case Jayadeva, perhaps envious of Mammat :as popularity, identies him with the svairin : i (loose woman) and requests that he not reveal himself in the Candra loka (moonlight).23

Semantic analysis in poetics: foundation or consequence? One of the more unusual and striking features of Jayadevas organization is the placement of his discussion on sabdav:rtti (semantic operations) in the nal
20 21 22

KP (4). (1.2) CA

cinmayacittacandraman hi compound that might be literally translated you who : ayah : . A bahuvr possess moonstones that are hearts of pure consciousness.

23 A common trope in Sanskrit poetry involves a female lover having to cancel a rendezvous on account of the full moon, which would likely reveal her surreptitious journey.

123

a Unraveling the Ka vyapraka s

237

four chapters of the Candra loka. Here, Jayadevas strategy is contra-posed to that of Mammat :a, who discusses the sabdav:rttis in the second chapter of the Ka sa, in order to more clearly articulate his subsequent explanations vyapraka  _ gya and citraka of dhvani, gun bhu tavyan vya. In point of fact, Mammat : :as organization provides the model for most of the alam sa ra stra treatises : ka composed during the second millennium. Hence we nd the  sabdav:rttis dis: n _ ga cussed in the 8th of the 36 chapters in Bhojas 11th century Sr sa, in rapraka  the rst of the eight chapters of Hemacandras 12th century Ka sa vya nu sana, in the second of the eight chapters in Vidya dharas 14th century Eka val and vana in the second of the 10 chapters in Kavira ja Vis thas Sa hityadarpan : a. While one might be inclined to dismiss Jayadevas discussion of the sabdav:rttis at the end of his treatises as an arbitrary decision or at best a vain and haphazard innovation, we should consider his expressed ambition to integrate aesthetic and logical methodologies within the Candra loka. As he states in verse 1.3: yuktya sva dyalasadrasaikavasatih hityasa rasvata: sa r ks ra mbhodhir aga dhata m upadadhat sevyah s yata m24 . : sama The milk ocean of poetics is the single locus of reasoning and aesthetic savor (rasa). As its depth is unfathomable, it should be approached with great care. While it is common enough for a lam rikas to apply logical methodologies in : ka their analysis of poetry, traditionally logicians saw little utility in applying poetic methodologies within logical analysis. Jayantabhat :t :a, the ninth century miri Naiya Kas njar yika, writes in the Nyayama~ : dr obhate athava ne carca kavibhih .s : saha s vidva m so pi vimuhyanti va kya rthagahane dhvani25 : Furthermore, discussions of this sort with poets lead to little benet. Even authoritative scholars become perplexed on the mysterious path of semantics. Jayadeva, envisioning an aesthetically informed reasoning, turns such a judgment on its head. In Candra loka 6.27, he castigates scholars who would exclude aesthetics in their pursuit of logic _ _ angabha ngollasall la tarun smaratoran . . am apu tarkakarkas rn ptotkat m 26 . oktipra . tha :adhiya : vr This young woman whose beauty shines in her undulating limbs is the gateway to love.

24 25 26

(1.3). CA NyM (1st a hnika, refutation of dhvani) 1:78. (6.27). CA

123

238

D. Mellins

She is useless, however, to those whose minds have calcied from studying too much logic. In Candra loka 1.2, Jayadeva identies poetics as vica rav cinicaya lam rava : ka ra nidhi the ocean of ala m ka ra that has vica ra as its waves. Understanding m : : vica sabdavya ra to be  pa ravica ra, poetics in the Candra loka becomes an ocean of gures of speech that produces waves of semantic investigation. According to such a model, the technical and aesthetic investigation that Jayadeva conducts in the rst two thirds of the Candra loka becomes the ideal foundation for the semantic investigation in the nal third of the work. That is, the investigation of how language functions in poetry informs the understanding of how words operate in their most fundamental sense. Poetic principles, rather than being secondary and articial features of language, are integral to the function of all language. Jayadevas retrograde presentation of the  sabdavr . ttis Along these lines, there is a nal peculiarity in the structure of the Candra loka that remains to be discussed the anomalous sequence within Jayadevas discussion of the three  sabdav:rttis (denotation, metaphorical indication, and literary suggestion). As is the case in the Ka sa, in practically all alam sa vyapraka  ra stras sympathetic to : ka rasadhvani theory, the analysis of  sabdav:rtti is seen to examine abhidha (denotation) rst, laks njana (metaphorical indication) second and then only vya~ (sug: an :a gestion), in contradistinction with the two more primary semantic operations. In nandavardhana justies this typical sequence: Dhvanya loka 1.9, A ikha janah a loka rth yatha d pas ya m n : yatnava : 27 tadupa yataya tadvad arthe va cye tada dr . tah : Just as a person desiring light will endeavor to ignite the wick of a lamp, this being a means to his goal, one who is devoted to dhvani (literary suggestion) will rst direct his attention to the literal meaning. nandavardhanas analogy, even if a rhetorician is more According to A enamored of suggested meaning he should rst investigate the literal meaning, which is prerequisite for the manifestation of the suggested sense. Once again, we should consider whether Jayadevas contrary sequence is arbitrary or purposeful. According to Jayadevas commentator, Ga ga bhat :t :a, Jayadevas sequence is simply a matter of causal retrogression: aktim laks m pya taddhetubhu ta m payati28 .an .a : niru : s : niru Having explained indication, he proceeds to explain its source: denotation.29
27 28

(1.9). DhvA (10.1). RA

29 Ga ga bhat sakti synonymously in his commentary to this (denotation) and  :t :a uses abhidha chapter.

123

a Unraveling the Ka vyapraka s

239

~jana In Ga ga bhat and laks after vyan :t :as view, Jayadeva discusses abhidha : ana because these are causes for its manifestation. In the same way, Jayadeva discusses abhidha after laks , because denotation is the precipitating cause : an :a of metaphorical indication.30 There may still be, however, a deeper purpose for this organization. The role of Vyan jana (suggestion) in Laks (indication) .an .a According to Kunjunni Ra ja , the recently deceased authority on Indian semantics, M sakas and logicians of the earlier period (Pra ma m cinanaiya yikas) : typically state two conditions for indication. The rst of these is a precipitating condition, alternately described as mukhya rthaba dha (a contradiction in the primary meaning) or anvaya (an unsuitability with respect to syn nupapattih : tactical unity); the second condition is something of an instrumental condition known as a sam : bandha, a relational property that orients the presumption of the intended secondary meaning with respect to the unsatisfactory literal _ ga meaning. In the stock example gan ya m : ghos : ah : (a hamlet on the Ganges), the precipitating condition is the illogicality of the meaning on the Ganges the fact that such a hamlet would sink. The instrumental condition is the relation of proximity that allows one to presume the secondary meaning on the bank. Abhinavagupta and the a lam rikas who are his successors attribute a third : ka condition to laks , namely a prayojana (purpose) sometimes referred to as : an :a a vya~ njana (suggestion). In most literary metaphors, an author uses an indirect expression for a specic reason rather than simply expressing the purport of this _ ga secondary sense as a literal meaning. An author would say gan ya m : ghos : ah : (a _ gatire ghos hamlet on the Ganges) rather than gan : ah : (a hamlet on the bank of the Ganges) because the former expression presents a more intimate suggestion of coolness and purity than the latter, as well as a greater range of polyvalent associations. Since a njana lam rikas hold vya~ to be an essential : ka condition of literary indication, a full understanding of laks would then : an :a require a proper understanding of vya~ njana . This may, at one level, explain why Jayadeva discusses vya~ njana prior to laks . : an :a The gurative basis of abhidha (denotation) We still, however, need to consider why Jayadeva discusses abhidha (denotation) as his nal topic. In recent decades, a number of prominent scholars in the disciplines of linguistics, cognitive psychology and comparative literaturemost notably George Lakoff, Mark Johnson and Mark Turnerhave argued that what we typically consider to be the literal meanings of words are in reality constituted by an elaborate chain of metaphorical associations. According to these theories, the conceptual processes that shape and motivate language are fundamentally metaphoric in nature,
30

To the extent that denotation is necessary to reveal a contradictory literal meaning motivating the secondary sense, denotation is understood to be a cause of metaphorical indication.

123

240

D. Mellins

and hence metaphorical speech temporally and functionally precedes literal speech. This is to say that the most basic and natural expression of human language is metaphoric rather than literal, and that literal designation, somewhat ironically, is a secondary and dependent function of language. While this theory might seem far removed from the linguistic suppositions of early second millennium South Asia, I should like to point out that such speculations are not limited to the modern era. The Roman rhetorician Quintilian writes in the rst century C.E: Quare illo intellectu priore et communi nihil non guratum est.31 Therefore in the rst and common sense of the word everything is expressed by gures. Almost 2000 years prior to Lakoff and Johnson, Quintilian determines that the most primary function of language is gurative. As it turns out, ancient Indian scholars were equally aware that metaphor pervades the most common and basic of speech processes. In the second adhya ya of the Nya yasu tra, Gautama defends the Nya ya position that the denotative meaning of a word is a combination of the generic property (ja ti), the shape (a k:rti) and the individual entity (vyakti) against the challenge (pu rvapaks : a) of hypothetical realists who would claim that the denotative meaning is precisely the individual entity.32 The pu rvapaks : a argues that because words most commonly signify individual entities, these must be regarded as the true denotative meanings.33 Gautama counters this claim, pointing to the fact that words are commonly used to signify meanings they do not actually denote:
31 Quintilian, The Institutio Oratoria of Quintilian / with an English Translation, trans H. E. Butler (Cambridge: Loeb Classical Library, 1921) book 3, 353355. 32 33

NyS (2.2.62).

vyaktya sayah sabdasamu k:rtija tisamnidha v upaca ra d sam  hatya gaparigrahasam v:rddh: : // ya : khya yapacayavarn sa nubandha na m v upaca ra d vyaktih : asama : vyakta : // NyS (2.2.6061). There is uncertainty [regarding your contention] because words would signify the individual, the shape and the generic property which are inextricably connected. The meaning of a word is the individual entity because the relative pronoun, clusters, abandoning, acquiring, enumeration, growth, diminution, color, verbal compounding, and succession apply to the individual. For some two and a half millennia, Indian philosophers have deliberated what constitutes the denotative meaning of words. The primary contenders, according to most authorities, are (1) the universal or generic property (ja ti), (2) the individual entity (vyakti), (3) the shape (a krti). The dilemma, which can be inferred from the seemingly contradictory implications of Pa n tras 1.2.58 and 1.2.64 (see . inis su Scharf 1996), was rst systematically articulated in the second century BCE by the grammarian jali (Scharf 1966, 2), who postulated the denoted meaning of words to be a combination of Patan jalis terminology) and the individual entity (dravya in the class characteristic (a k:rti in Patan jalis terminology). Early M Patan ma m : sakas assumed the most formally orthodox view that the denotative meaning is the universal or class property (ja ti) alone, while Gautama and early Nya ya proponents adopted the most pragmatic position that the denotative meaning is a combination of the universal or class property (ja ti), the individual entity (vyakti), and the shape (a krti). Later logicians such as Jayantabhat :t :a contended that the denotative meaning of the word is the individual entity qualied by the generic property (Ra ja 1963; Scharf 1996; Ganeri 1996; Deshpande 1992). One would do well to consult Ra ja (1963) for a lucid introduction to this debate, and turn to Deshpande (1992) and Sharf (1996) for more comprehensive treatments of this topic. Tiwari (1994) presents an excellent account of early Nya ya approaches to semantic theory.

123

a Unraveling the Ka vyapraka s

241

sahacaran nata darthyavr nadha ran m pyayogasa dhana dhi. astha . ttama . asa _ s a cakat patyebhyo bra hman jasaktucandanaganga t nnapu. aman :ara :aka rus ve pi tadupaca rah .es .v atadbha : In [the examples involving] 1) a Brahman, 2) a cot, 3) a mat, 4) a king, 5) barley, 6) sandalwood, 7) the Ganges river, 8) a cloth, 9) food, and 10) a man, there is semantic application34 of a word even though the primary referent is not present35 because of 1) co-presence, 2) placement, 3) a purpose, 4) a behavior, 5) measure, 6) containment, 7) proximity, 8) connection, 9) means, and 10) sovereignty. Va tsya yanas commentary to the su tra elucidates the embedded illustrations: abdena bhidha atadbha ve pi tadupaca ra ity atacchabdasya tena s nam iti sahacaran a d yas t ika m bhojaye ti yas t ika sahacarito bra hman o . .: .: . bhid: ca h yata iti. stha na d man h kros ant ti man castha h purus a abhidh yante. . : : ta darthya t kat a rthes u v ran es u vyuhyama nes u kat am karot ti bhavati. . . . . : : : ti tadvad vartata iti. ma vr d yamo ra ja kubero ra je na d a d . tta : hakena hakasaktava iti. dha ran a t tula ya m dhr tam candanam mita h saktavah a d . : . : : : : : _ ya carant ti des o bhidh tula candanam iti. sa m pya d ganga m vas yate : ga a sannikr t kr gen t yate. .s .t :ah : . yoga .s .n . ena ra . a yuktah :s :akah : kr .s .n . a iti abhidh sa dhana d annam pra n a iti. a dhipatya d ayam purus ah kulam ayam . . : : : : yam gotram iti. tatra sahacaran a d yoga d va ja tis abdo vyaktau prayujyata . : iti.36 The statement there is semantic application of a word even though the actual referent is not there means by the word, there is denotation of another word. Because of co-presence, the expression feed the staff means feed the Brahman; because of position, the expression the cots are crying means the people lying in the cots are crying; because of purpose, he is weaving a mat means straw is woven for the sake of a mat; because of behavior, King Yama37 means King Kubera,38 i.e., 39 he acts in the same way; because of measure, a barley a d : haka means an amount of pounded barley that measures one a d : haka; because of containment, balanced sandalwood means the sandalwood
34

upaca ra. While in a previous context (NyS 2.2.60), Va tsya yana had glossed the word upaca ra with the synonym prayoga (use, application), in the current context the association with the expression atadbha ve (the absence of that [referent]) claries that upaca ra signies secondary or metaphorical usage (see Tiwari, 157). Many thanks to Walter Slaje, Bogdan Diaconescu, Madhav Deshpande and Gary Tubb for their advice and assistance in interpreting this su tra. One might alternatively translate atadbha ve pi tadupaca ra to mean there is application to a meaning even though this is something other [than the referent of the word] NyBh (2.2.62). The god of death. The god of wealth.

35

36 37 38 39

A measure of weight approximating 7 lbs. 11 oz. (Apte, The Practical Sanskrit-English Dictionary {Kyoto: Rinsen Book Company, 2003} 321).

123

242

D. Mellins

contained in the balance; because of proximity the cows roam on the Ganges means a location proximate [to the Ganges]; because of connection, the cloth is black means the cloth that possesses the color black; because of means [there is the expression] life is food; because of sovereignty, he is the family means he is the head of the family. In the present consideration, because of co-presence or connection, the word for a universal is applied in the sense of the individual entity. It should be acknowledged that this su tra as well as Va tsya yanas commentary represents a provisional and preliminary argument within Gautamas greater deliberation on the meaning of a word (pada rtha), and hence does not present Gautamas conclusive view; here a hypothetical proponent of the view that a word denotes the generic property (ja ti) claries his rebuttal of a hypothetical proponent of the view that a word denotes the individual entity (vyakti). Nevertheless, Gautamas su tra provides evidence of an ancient Indian view contending that metaphorical indication (laks or upaca ra) is responsible : an :a for one of the most fundamental linguistic processes: the use of words to identify individual entities. While all the illustrative expressions in the commentarial passage cited above present some degree of metaphorical attribution, the third exampleka:tam : karoti (he makes a mat)is a seemingly direct grammatical construction, which most people, if questioned, would deem to be a literal expression. In fact, the grammarian Pa n . inis (fth century BCE) explanation 40 of the function of the verbal object (karma)kartur psitatamam : karma (karma is the thing most desired by the agent [with respect to the action]41)justies the syntactical logic in this expression,42 thus avoiding the need for application (upaca ra) to a secondary meaning. According to the Nya ya view, however, the action is not performed on the direct object (the mat), which corresponds to the nished product, but rather on the raw material (straw) for the purpose of a mat. On the basis of this discrepancy, Gautama advocates the transference of a secondary meaning, which Va tsya yana determines to be ka:ta rthes ran nes : u v : es : u vyuhyama : u (straw is woven for the sake of a mat). A similar evaluation can be made of Va tsya yanas eighth example:  sa :takah : k:rs :n : ah : (a black cloth). While the word k:rs :n : ah : (the color black) might appear to be an adjective qualifying the noun  sa se s :takah : (cloth), according to the Vai : ika categories adopted by early logicians, these correspond to separate strata in the ontological hierarchy. K:rs :n : a, being a quality (gun : a), cannot be literally equated with a substance (dravya) such as  sa:taka, and therefore a relation of inherence is metaphorically indicated. In this way, the expression  sa :takah : k:rs :n : ah : is understood to have the meaning a cloth conjoined with the color black
Pa n : (1.4.49). The Ka sika  v:rtti supplies here the ellipsis kriyaya (with respect to the action): kartuh yad : kriyaya a nj~ nam ptum is rakam ::tatamam : tat ka : karmasa~ : bhavati (KV, 1.4.49). (The nominal relation known as karma is that which the agent most desires to attain with respect to the action).
41 42 40

In the sentence ka:tam : karoti (he makes a mat), ka:tam : (mat) is the most desired object.

123

a Unraveling the Ka vyapraka s

243

(k:rs :n : ena yuktah : ). It is the very pervasiveness of such metaphors in ordinary speech that lends support to the argument that the use of words to identify individual entities is due to a metaphorical indication (upaca ra) based on the denotation of the generic property.

Jayadevas denition of denotation At this point, it will be useful to consider Jayadevas explanation of abhidha or literal designation: abdah cit puraskr dharmam yah . tya pra : kan : s : pravartate abdas ta yatha [read yaya ]43rtham cas m abhidha m .t .t : spas :am a :e s : viduh : aya ja tya gun vastuyogena sam . ena kriyaya : jn 44 ena ca manyante s nirdes m abhidha m h .ad : vidha : budha : Generally a word will signify a meaning by bringing a specic property to the forefront.45 Scholars regard denotation to be the [semantic operation] by which words clearly express meanings. They understand denotation to have six types: denotation by way of 1) a universal, 2) a quality, 3) an action, 4) a connection to a thing, 5) a name, and 6) a self reference.46
Pradyotanabhat loka, comments : tatra yaya :t :a, the earliest known commentator to the Candra bhidha v:rttya sabdah s:te sa  ks d a ca: (here the operation by which a word directly signies : spas ::tam : sa :a is abhidha ga bhat tha Pa yagun ). Two subsequent commentators, Ga :t :a and Vaidyana : a cite Pra.d dyotanas explanation in their comments to the same verse. It seems likely, given the similarity between the letters y and th in devana gari script, that manuscripts may have interpolated the reading yatha rtham rtham refers to : in the place of yaya : . All three commentators explain that yaya v:rttya (by [this] semantic operation). According to this reading, yaya (by which) becomes a suitable relative pronoun to the correlative pronoun ta m (that). While it might be tempting to accept the reading yatha rtham, translating this to mean a suitable meaning, this would result in the loss of a corresponding relative pronoun for the correlative pronoun (ta m). 44 (10.12). CA
45 46 43

purask:rtya. Less guratively, having utilized. While the word nirde sa literally means specication, Jayadeva explains the denotation of this type of word in the following way: _ na yoga der a yatanam : na sanketaniketanam , 10.4). as abdo yam vr nirdes sva bhidheyaya (CA : mukhyaya . ttya Being neither the locus of en etymology nor the site of a verbal convention, The [nirde sa] word signies by way of the primary semantic operation, which [in this case] targets the form of the word itself. For this reason, I have translated nirde sa as self-reference. The reader will notice that in this verse, Jayadeva appends two additional types of denotation, vastuyogena (connection to a thing) jal and nirde sa (self reference), to the four enumerated by Patan in the Maha bha s ti, gun : ya: ja : a, kriya na and yad:rccha . Jayadeva also uses the synonym sam in the place of yadrccha . : j~

123

244

D. Mellins

Jayadevas denition of denotation differs ever so subtly from that of Mammat :a, which describes the semantic operation in more conservative terminology:47 sa ks t sam cakah .a : ketitam : yo rtham abhidhatte sa va : sam ketitas caturbhedo ja tya dir ja tir eva va : bhidho cyate48 sa mukhyo rthas tatra mukhyo vya pa ro sya A denotative expresser (va caka) signies a meaning directly established by convention. A conventionally established meaning may have four varieties universals, etc. (qualities, actions and adventitious names) but in another way of thinking it is simply a universal (ja ti). This is the primary meaning, and the primary semantic operation signifying this is called denotation (abhidha ). In a certain respect, Mammat :a and Jayadeva have the same principal concerns: (1) establishing the directness or primacy of the denotative operations and (2) explaining how a single word (e.g. gauh : or cow) can efcaciously identify countless individual entities (anantavyaktayah : ) belonging to the proper conceptual class. If a word is indeed understood to signify the individual entity (vyakti), then it would necessarily signify a single entity each time the word is used, or it would need to signify each and every entity at all times. If the word signies a single entity, however, then each time a word signies a different entity, the meaning of the word will change and denotation will succumb to the pitfall of variability (vyabhica ra). If, on the other hand, the word signies all the entities within a class, then the meaning of the word will acquire the pitfall of innitude (a nantya) to accommodate the innity of individual entities, and hence an ordinary person will never be able to comprehend the meaning of a word. In resolving this semantic dilemma, the two a lam rikas resort to different strategies. : ka Mammat a denes denotation as the operation revealing a single meaning : invariably joined (sam : ketita) to the word by force of verbal convention. These denoted meanings take the forms of four properties: generic properties, qualities, actions and names. As he succinctly explains in his auto-comment to the aforementioned verses: yady apy arthakriya ka ritaya pravr vyaktir eva, tatha py . ttinivr . ttiyogya a nantya d vyabhica ra c ca tatra sam : ketah : kartum : na yujyata iti, gauh :
47

The preceding one and a half verses state: abdo tra vyan jakas tridha sya d va cako la ks : s .an . ikah va cya dayas tadartha h tparya rtho pi kes : syus ta .ucit o rtha s jakatvam ap sarves m yas na m : pra : vyan .a .yate Since words are threefolddenotative, metaphoric and suggestive The meaning of words are denotated, metaphorically indicated and suggested, even if in some cases they are determined by the intended import of a sentence. In general, all of these meanings become suggestors.

48

KP (7cd 8).

123

a Unraveling the Ka vyapraka s

245

uklas calo d ti ca tadupa s d na m go na prapnot dha v : ittha itya : vis .ayavibha 49 eva sam : ketah : It is not proper to establish a verbal convention with respect to the individual entity, even though this alone is the suitable [object] of engagement and disengagement on account of objective efcacy (arthakriya ), because [this would lead to the unwarranted consequences] of innitude and variability. Thus words such as cow, white, moving, and the name D : ittha do not have multiple referents and the verbal convention [is to be established] only with the distinctive property (upa dhi). In equating the denoted meaning exclusively with a distinguishing property (upa dhi), Mammat :a emphatically excludes the individual entity (vyakti) from the range of denotation proper. According to Jayadevas denition, however, a word does not unequivocally denote the distinguishing property (dharma), but rather utilizes this property (purask:rtya literally having brought to the forefront) to expedite the denotative operation.50 In stipulating that denotation brings a property to the forefront (purask:rtya), Jayadeva implies that a substrate remains in the background. As the commentator Pradyotanabhat :t :a (16th century) explains: _ akyah akyo] vyaktima tina kevalavyaktau sanketagrahah trasya : s : [read s es prasaktatva t vyaktau vis asya vis es akam dharmam vina nupapatteh . . . : : : api _ kavyaktau sankete ca vyaktyantaraprat tir na sya t.51 It is not proper to understand a verbal convention solely in relation to the mere entity, because this would lead to undesirable consequence.52 It is impossible to distinguish the mere entity in the absence of a distinguishing property. Moreover, if the verbal convention were to pertain to a single individual entity, this would not allow for the comprehension of other individual entities [in the same class].

KP (7ab, vr . tti). In this respect, Mammat :as and Jayadevas denitions demonstrate separate philosophical inclinations. In dening the denoted meaning to be a meaning established by convention, Mammat sakti (inherent (denotation) otherwise known as  :a more or less equates abhidha designative capacity) with sam : keta (semantic convention), thus demonstrating his sympathy with traditional M ma m : saka presentations of semantic theory. In contrast, Jayadeva appears to concur with the Naiya sakti or abhidha yika position that  amounts to the relationship between a distinguishing property and the thing designated. 51  S A (10.1)
50

49

tiprasaktatva vyaktima trasya t. Literally, because the mere entity is stretched unwarrantedly. In the current context, the qualication atiprasakta has the sense leading to unwarranted con_ ga). In a manner of speaking, even the etymological meaning here is suitable in sequences (prasan that the mere entity is stretched unwarrantedly in the role of a verbal convention, because such a convention does not enable the signication of the multiplicity of individual entities within a given species.

52

123

246

D. Mellins

In Pradyotanas opinion, Jayadevas denition denies the possibility that the pure individual entity can be the denotative meaning, since this can never be the object of a single invariable and efcacious verbal convention. However, _ ketagrahah in stating na kevalavyaktau san sakyah :  : (it is not possible to understand the verbal convention with respect to the mere entity) rather _ ketagrahah than na vyaktau san sakyah :  : (it is not possible to understand the verbal convention with respect to the entity), Pradyotana implies Jayadevas consent that a word may in part signify the individual entity, since this becomes the substrate for a distinguishing property. In Pradyotanas view, Jayadevas denition allows that a word, utilizing a distinguishing property, produces an understanding of an individual entity in the process of denotation itself. Jayadevas explanation is unique among denitions of abhidha in that it consciously avoids specifying the denoted meaning of the word (abhidheya rtha). Instead, it draws attention to the different types of properties (dharma), which serve as the instrumental means of denotation (prav:rttinimitta). By their nature, these properties are understood to inhere in substrates (dharmin), which are no other than the individual entities implicated in most verbal expression. In a certain way, Jayadevas explanation is noncommittal since it does not stipulate whether (1) a word denotes only the property and the listener must cognize the individual entity by way of the secondary power of speech (laks ) and a separate cognitive function such : an :a as inference (anuma na) or presumption (artha patti), in which case denotation does not produce objective efcacy in the absence of metaphorical indication, or (2) denotation itself produces the cognition of the individual entity as qualied by a property (dharmavi sis : :tavyakti). If we accept this latter view, then there might be the objection that a word cannot simultaneously denotes two things, a property and a particular, required to convey a relationship. This objection is countered by the understanding that, apart from abstract contemplation, there can never be a determinate perception of a general property in the absence of a particular entity constituting its substrate, just as there can never be a determinate perception of a particular entity in the absence of an inhering property. A determinate perception, which provides the basis for verbal signication (prav:rttinimitta), is based on this indissoluble relationship. The denotative meaning of a word, which mirrors such a determinate perception, inherits this relational character as well, and in this subtle way functions through the trope of inherence (a substrate possessing a property).

The analogy to Jayantabhat n :t :as doctrine: denoted meaning as tadva If we are to interpret Jayadevas explanation in the former way, then Jayadevas denition, like that of Mammat :a, reiterates the traditional M sa view that the word designates only the generic property. If, in ma m : accordance with the latter interpretation, Jayadevas explanation stipulates

123

a Unraveling the Ka vyapraka s

247

that a word designates the individual entity as qualied by the generic property (ja sis tivi : :tavyakti), than his explanation resonates most closely with that of the ninth century logician, Jayantabhat :t :a, who states: tyevama anyes m dis .u tu prayoges .u ga .u : deh va huh tadvato rthakriya yoga t tasya pada rthata m53 : But because of the objective efcacy of other usages such as bring the cow, Scholars say that the meaning of the word is just that possessing this (the generic property).54 With such a doctrine, Jayanta recongures Gautamas doctrine so as account for the relationship of the generic property and individual entity. Where Jayadevas explanation of denotation intimates that the denotation of a word aims at something beyond a delimiting property, namely the individual entity possessing this property, Jayantas doctrine of tadva n emphatically species such a relationship. Still, dening the denotative meaning of a word to be that possessing the generic property does not completely eliminate the need for attribution in denotation. Even if one assumes that the meaning of the word is something possessing a specied property, a listener will still need to attribute the possession to the particular entity intended by the speaker.

The synedochic structure of Jayadevas explanation of denotation However we chose to interpret Jayadevas model of denotation, this presents the verbal signication of the individual entity as a type of synecdoche, whereby a property or attribute is used to signify its locus. Hereby, denotation utilizes a property a generic property (ja ti), a quality (gun : a), an action (kriya na ), a relationship to a thing (vastuyoga), a name (sam ), or a self: j~ reference (nirde sa) to identify the entity possessing this property. The reader may consult Figs. 3 and 4, which schematize the quasi-gurative nature of two types of denotation presented by Jayadeva. When examined closely, Jayadevas model of denotation embodies features remarkably similar to the three causal properties found in metaphorical indication (laks ): (1) a direct sense that is somehow unsatis: an :a factory (mukhya rthaba dha or anva ya nupapatti), (2) a relationship between the ultimately targeted meaning and the unsatisfactory direct meaning (sambandha), and (3) a purpose for the indirect expression (prayojana). By way of Jayadevas explanation, when a speaker uses a word to designate a particular entity, the word species a qualifying property, which allows the
Jayantabhat :t :a (2:37). This passage is cited in Ganeri (1996). I am indebted to this article for both its explanations and its provision of a number of valuable sources.
54 53

Jayanta later claries the term tadva n: kim srayah scid anullikhita sa ses ma nya  baleya divi n ity ucyate (NyM, 2:38). : tu sa : ka : as tadva

123

248

D. Mellins

Fig. 3 Denotation by way of a Generic Property (ja ti)

Fig. 4 Denotation by Way of a Connection to a Thing (vastuyoga)

listener to identify the proper type of particular entity, but which does not ultimately satisfy the speakers intention. A parched traveler asking for some water, cannot not quench his thirst with the generic property water-ness, even if this is said to expedite his communication. His intention is to receive a measure of a specic type of liquid substance, i.e., water. The relationship between the immediately specied quality (waterness) and the water itself is one of inherence (samava ya). The purpose for this indirection is that it allows for the objective efcacy (artha kriya ) of a single word in multiple contexts. Such an indirect pathway insures that the speaker can use the same word, water, to successfully communicate the same type of substance, water, in any given instance or circumstance. It is functionally impossible to directly signify the individual entity alone because this would require a unique word for each and every individual entity. In order to avoid this predicament, denotation depends upon a subtle form of semantic indirection. What distinguishes most cases of metaphorical indication from the attribution I have just described is type of purpose. The semantic indirection within metaphor allows a speaker or writer to suggest a quality or feeling that would not be conveyed through direct expression, such as the coolness and holiness suggested by the expression a hamlet on the river Ganges. The subtle indirection within denotation, in contrast, simply provides the single word the capacity to communicate particular entities in multiple contexts.

123

a Unraveling the Ka vyapraka s

249

The representative nature of speech prevents denotation from ever attaining the status of perfectly direct signication. The word re does not burn our tongues; the word water will not quench our thirsts. Semantic directness is further compromised by the fact that a single word must have the capacity to identify countless individual entities. In their effort to establish the directness of verbal designation in statements involving objective efcacy, e.g., bring the cow, Indian language theorists postulated a relationship between a word and a general property, which is to be then attributed to the appropriate particular entities. No school of thought perfectly succeeded in explaining non-variable verbal denotation fully apart from this attribution. In truth, denotation is direct only to the extent that it is the most direct of the semantic operations. Jayadevas explanation that generally a word signies by bringing some property to the forefront acknowledges this attributive nature more openly than do other classical Indian models. In the face of this limitation, he allows his analysis of gures of speech (alam ra) and metaphorical indication (laks ) to : ka : an :a inform and contextualize his explanation of denotation. In reversing the conventional sequence within his own discussion of the  sabdav:rttis, Jayadeva emphasizes that investigations into the separate semantic operations denotation, indication and suggestion mutually reinforce one another, that by cultivating an understanding of the indirect functions of speech, one gains greater insight into the direct semantic capacity. Inasmuch as it is impossible to understand indication and literary suggestion in the absence of a rudimentary comprehension of literal designation, the elusive nature of literal designation, disputed for over two millennia in India, may most effectively be understood in opposition to its more indirect semantic counterparts.

Concluding remarks To summarize the main points of the paper: Jayadevas idiosyncratic sequence of topics in the Candra loka has a threefold purpose. First, this organization differentiates his treatise from the Ka sa, an act of vyapraka  deance that literally unravels the structure of Mammat as text. Second, this : organization intimates that aesthetic explorations into language inform scientic investigation into the same. Lastly and somewhat paradoxically, his organization suggests that the investigation of more indirect functions of language (verbal suggestion and metaphorical indication) claries our understanding of the most primary semantic operation (denotation). Through this unusual structure, Jayadeva presents evidence for a position that becomes more forcefully articulated by Jaganna tha Pan ja in the .d : itara lam _ ga Rasagan ka rikas (literary rhetoricians), as the true con dhara: that A : noisseurs of gurative and suggestive language, are sovereign epistemological authorities, unbeholden to other philosophical traditions. Here, I

123

250

D. Mellins

conclude with a verse from Jayadeva P yu s .avars .as drama the Prasannara ghava: alakala yes m vyakaus l la vat bha rat .a : komalaka tes a m karkas atarkavakravacanodga re pi kim yate . : : h yaih ka nta kucaman d ale kararuha h sa nandam a ropita s .: : : 55 ara taih kim mattakar ndrakumbhas ikhare na ropan ya h h . : : : s : Those whose speech possesses skill, technique and beauty suitable for tender poetry, What would they suffer, even in the articulation of complex circuitous logical statements?56 Those whose ngernails have blissfully pierced the breasts of beautiful women Why shouldnt their arrows pierce the eshy temples of rutting elephants?

References
: Candra CA yu s loka of Jayadeva P loka of Shri Jayadeva with the Ra ka gama .avara. In Candra Commentary of Ga ga bha:t:ta. Ed. Ram Ananta Shastri Vetal. Benares: Chaukhamba Sanskrit Series, 1938. De, S. K. (1925). A history of Sanskrit poetics. 2nd rev. ed. Calcutta: Mukhopadhyay, 1960. Deshpande, M. M. (1992). The meaning of nouns: Semantic theory in classical and medieval India. Dordrecht: Kluwer Academic Publishers. ri A : Dhvanya nandavardhana with the nandavardhana. In The Dhvanya DhvA loka of A loka of S ri Abhinavagupta. Varanasi: Chowkhamba, 1965. Locana of S Ganeri, J. (1996). Meaning and reference in classical India. Journal of Indian Philosophy, 24(1), 119. GG: G tagovinda in G tagovinda of Jayadeva: Love Song of the Dark Lord. Ed. and Trans. Barbara Stoler Miller. Delhi: Motilal Banarsidass, 1977. Jhalakikar, Vamanacharya Ramabhatta. Benediction in Ka sa of Mamma:ta with the vyapraka  Ba labodhin of the Late Vamanacharya Ramabhatta Jhalakikar. 7th ed. Ed. Raghunath Damodar Karmarkar. Poona: Bhandarkar Oriental Research Institute, 1965. str (1929). Introduction to The Candra loka, by P Khiste, Na rayan : yu s : avars : a Jayadeva with . a Sa arada the Commentary, S sa by Padmana sra alias gama alias Candra lokapraka  bhami str Pradyotana Bha:t:ta ra yan Khiste. Benares: Jai Krishnadas-Haridas ca rya. Ed. Na . a Sa Gupta. KP: Ka sa of Mammat sa of Mamma:ta with the Ba vyapraka  vyapraka  labodhin of the :a. In Ka Late Vamanacharya Ramabhatta Jhalakikar. 7th ed. Ed. Raghunath Damodar Karmarkar. Poona: Bhandarkar Oriental Research Institute, 1965. KV: Ka sika mana and Jaya ditya.  v:rtti of Va a njar njar th S str . Varanasi: NyM: Nya yama~ of Jayantabhat yama~ . 3 vols. Ed. Gaurina :t :a. In Nya vavidya Sam rn nanda-Sam laya, 1982. : pu : skrita-Vis .a NyS: Nya sanam with the Bha yasu tra of Gautama. In Gautam yam yadar s : Nya : ya of Va ts yana. Ed. Anantalal Thakur. New Delhi: Indian Council of Philosophical Research, : ya 1997.

55 56

PR (1.18).

vakravacana. A likely pun alluding to vakrokti, the poetic indirection of meaning which a number of a lam rikas in particular Kuntaka held to be the basis of literary charm. : ka

123

a Unraveling the Ka vyapraka s

251

NyBh: Nya sya of Va tsya yana. In Gautam sanam with the Bha yabha : yam yadar s : Nya : ya of Va ts yana. Ed. Anantalal Thakur. New Delhi: Indian Council of Philosophical Research, : ya 1997. Pa n n dhya y of Pa : : As ::ta . ini. Varanasi: Chowkhamba Sanskrit Series Ofce, 1965. ri Jayadeva with the the PR: Prasannara ghava of Jayadeva. In Prasannara ghava of Maha kavi S esara arma Regmi. ca Chandrakakala Sanskrit and Hindi Commentaries. Ed. A rya S ja S Varanasi: Chowkhamba Vidyabhavan, 1963. Quintilian. Institutio Oratoria. In The Institutio Oratoria of Quintilian with an English Translation. Trans H. E. Butler. Cambridge: Loeb Classical Library, 1921. Book 3, 353355. : Ra ga bhat RA loka of Shri Jayadeva with the Ra ka gama Com ka gama of Ga :t :a. In Candra mentary of Ga ga bha:t:ta. Ed. Ram Ananta Shastri Vetal. Benares: Chaukhamba Sanskrit Series, 1938. Ra ja, K. K. (1963). Indian theories of meaning. (2nd ed.). Reprint. Madras: The Adyar Library and Research Centre, 1977. A arada : S S gama of Pradyotanabhat loka of Jayadeva along with the Commentary :t :a. In Candra arada S gama of Pradyotana Bha:t:taca rya. Ed. Ganga Sagar Rai. Varanasi: Chaukhambha Sanskrit Sansthan, 1985. Scharf, P. (1996). The denotation of generic terms in ancient Indian philosophy: Grammar, Nya ya and M ma m . Philadelphia: American Philosophical Society. : sa Tiwari, H. (1994). One and many: The early Naiya yikas and the problem of the universals. Journal of Indian Philosophy, 22(4), 137170.

123