24. 116. Compare kārikā 9.

116. Compare kārikā 9.

117. That Bhoja is defining känti here is clear from the discussion beginning on p. 497 (compare also KAS 3.2.15). For dīptarasa, see also Ratneshvara ad SKĀ p. 77: "[This signifies] the highest development of rasa, that is, when passion or sense of self becomes intensified by means of the stable emotion when fully furnished with its aesthetic elements."

118. For this enlarged sense of "ornament," see n. 73 above and n. 122 below.

119. Bhoja has a special category of ornament wherein sense is inseparable from its expression (ubhaya or bāhyābhyantara), which comprises what are elsewhere considered pure figures of sense (see also Raghavan 1978: 379).

120. The sets of four in the similes before and after this require this addition; see next n.

121. The upameya, which also must be fourfold, consists of the gunas, śabdālankāras, arthālankāras, and ubhayālankāras.

122. Having defined rasa, Bhoja turns now to his main objective in the chapter, which is to show that language qualities, rasa, semblance of rasa, and emotions are all "ornaments" of a poem, and, in accordance with the above, that the "presence of rasa" consists in their "combination." The discussion is found also in SK beginning on p. 703.

123. The reference must be to p. 617.3.

124. Bhoja here cites KĀ 2.1, 3.

125. Bhoja quotes KA 2.273, above. Bhoja's purpose here is to show that when these three qualities are "intensified" they become rasa, or rather, three "stages" of rasa, while at the same time they are "listed as ornaments" (contrast Raghavan 1978: 297).

126. adrsta here refers to dharma ("property") in kārikā 4.

127. Bhoja is again glossing kārikā 4.

128. Source unknown, as is the next. Note that rasa is attributed to the speaker here and below.

129. See kārikā 9.

130. MM 1.29 (Madhava is speaking).

131. MM 5.28 (Madhava is speaking). In the two examples, the stable emotion, the transitory emotions, and the reactions are all shown to be present.

132. Bhoja is quoting kārikā 10. He may also have in mind Induraja on KASS p. 48, above.

133. I conjecture premarūpena for rasarūpena, or at least understand it as such, in view of the parallel passage, p. 636.11 (see n. 178 below). Bhoja's interpretation is confusing given Dandin's express differentiation between preyah as characterized by prīti or nonerotic affection, and rasavat as characterized by rati (etc.) (so Ratnashrijnana, Tarunavachaspati, and Hrdayamgama ad KĀ 2.281 [= 2.279]), but he may have in mind KASS 4.2, and perhaps even Induraja's remarks ad loc., p. 48, given above.

134. See SP p. 613, above (it cannot be confirmed whether this is intended as a direct quotation of a character).

135. URC 1.40.

136. See Pollock 1998a: 165.

137. After discussing the different meanings of the three terms when they are regarded as gunas and as alankāras, Bhoja returns to the question of rasa.

138. "Possessive," matup;"comparative," vati.

139. Compare Mahābhāșya 5.2.94, bhūmanindā-, etc.

140. P 5.1.115-16.

141. P 5.1.117.

142. Since any given verse would, given the definition of alankara, always consist of various "beautifying" features. The target of the critique is such a phrase as "mixture of various ornaments"; see p. 623.27, above.

143. That is, when he spoke earlier (p. 617.3) of the "presence of rasa" in a poem at the level of a passage (as distinct from its presence at the level of the work) as coming about especially by the "combination of various ornaments."

144. Bhoja has already shown that language qualities can also be considered ornaments, and one category of quality is that of the signification (arthaguna, p. 502).

145. Bhoja is only offering an analogy (although he accepts the principle he is enunciating; see p. 449.22). Nothing indicates an echo of Dhanika's vakyapadīya argument about the literary text (Avaloka p. 211, below). Contrast Raghavan 1978: 165, 522.

146. Bhoja will distinguish the first set as internalized reactions and the second as externalized, p. 638 (without any suggestion of serial correlation; see note there). See also SKĀ 5.38-40.

147. I suspect the correct reading here is -ottarottaro (for -ottarottara-).

148. In what follows we are given the entire series of rasas as they are produced, and the rasavad alankaras that arise from the rasa and embody it. All examples are borrowed from Dandin (KA 2. 278 onward), but the analysis of what rasa is and how it works in a literary text is Bhoja's own. In each case, it is the speaker of the verse (or, in the case of the fearful, the main character) who experiences the stable emotion, the conditions that enhance it, the predispositions that underlie it, and the rasa that results, and who "has" or "tastes" the rasa (compare SKĀ p. 728, cited below at ŚP p. 655.3). Bhoja reconsiders all eight examples on p. 636.

149. KĀ 2.278-79, above.

150. KĀ 2.287, above.

151. KA 2.282, above.

152. KĀ 2.288, above.

153. These are externalized transitory emotions for Bhoja (ŚP p. 637, above). 154. I omit the remaining examples (the violent, the fearful, the tragic, and the macabre).

154. Bhoja's critique of Bharata's complex observations on the relations of rasas and bhāvas (NS pp. 286-88, above, and n. 22 there) is even more complex. Bharata is said to hold that, just as rasas can arise from bhāvas (postulate 1) and bhāvas from rasas (postulate 2), so a second rasa can arise from a first rasa (postulate 3), for which Bhoja gives the examples of hāsya from śrngāra, etc. (see NS 6.39). Bhoja refutes postulate 3 by saying (apparently) that the first rasa does not exist as such at the time the second arises, since it remains a bhava. Bharata's restriction to eight rasas is also false; some people add four more, śānta, preyah, uddhata, and ūrjasvin (wherein the same process of conjunction of factors is present), whereas others say all 49 bhāvas can become rasas. As for postulate 1, that rasas arise from bhāvas: rasas do not find their origin in bhāvas but rather are bhāvas in full development. As for postulate 2: bhāvas do not arise from rasas plural, because there exists in fact only a single rasa. As for postulate 3: for a second rasa to arise from a first entails that the first constitutes either the foundational factor for the second rasa or the prior state (prakrti) out of which it develops, but there is no causal or terminological regularity that shows this to be true.

155. Dwivedi and Raghavan have both arranged the lines so as to exclude the statement "it is only these eight stable emotions," etc., from Bhoja's critique (it is true that the statement echoes the end of Dandin's account of rasavad alankāra, but he is quoting Bharata too). The ensuing discussion requires that it be included, however; Bhoja will refute the idea on p. 632.19, and see kārikā 11.

156. Shiva, in his androgynous form.

157. Source unknown, as is the next poem.

158. The idiom (swearing an oath?) is unfamiliar to me.

159. Bhoja goes on to cite MVC 2.20 as an example of the tragic rasa coming from the violent, and MVC 1.35 as an example of the fearful rasa coming from the macabre.

160. Bhoja himself; see SKĀ 5.23, above.

162. Preyas ("affectionate") in the context of a rasa typically refers to nonsexual, especially parental, love (the example given below is not this, but neither is it erotic), and ūrjasvin to what Bhoja elsewhere calls udātta (e.g., SKĀ 5.23).

163. Pañcatantra (ed. Hertel), v. 79. Bhoja nowhere interprets śānta as the renunciant's peace.

164. ŚP p. 625, above.

165. VS 3.46.

166. KĀ 2.291, above.

167. prakṛti.

168. Bhoja here cites RKA 12.4, above.

169. Having refuted Bharata's view that there are only eight rasas, Bhoja now addresses the other two components of his theory.

170. NS 6.39. Bhoja goes on to show that no strict concomitance is found: other rasas arise, and sometimes the aforementioned

rasas do not.

171. Bhoja is speaking hypothetically, since he does not actually accept that rasas arise from other rasas.

172. The example here and those in the rest of the paragraph are omitted.

173. Raghavan observes that Bhoja here "mixes up" the foundational passion and the rasa (1978: 500 n. 2), but theirs is a kind of type-token relationship, and is constitutive of the ŚP.

174. NŚ 6.40.

175. Precisely Abhinava's position (see ABh 1.289, below).

176. NS 6.41 (perhaps understood by Bhoja as "the effect").

177. Which is position (1), and that has already been refuted.

178. Given that desire is the paradigmatic emotion. While for Bhoja rasa and prema ultimately signify the same thing, the logic of the argument indicates that we should read premarūpena for rasarūpena in the parallel passage, p. 625.21 (see n. 133 above).

179. Here and below, Bhoja paraphrases rather than quotes the earlier citations.

180. That is, we can see how the speaker's or character's sense of self constitutes the rasa principle and underlies the individual

rasas.

181. Bhoja adduces partial quotes for the violent, the fearful, the tragic, the macabre, the peaceful, the affectionate, and the

vainglorious.

182. Since for Bhoja rasa does not "arise" (let alone "rasas"), his interest in provisionally accepting Bharata's dictum (in the rasasūtra) is to demonstrate the singular rasa upon which the conventionally conceived plurality of rasas is epiphenomenal.

183. The three similes are for śrngāra, vīra, and karuņa respectively. For this last poetic convention I am unable to find a parallel.

184. The triadic grouping of the sthāyibhāvas (rati, etc., raudra, etc., śoka, etc.) throughout this passage derives from their correlation with sattva, rajah, and tamaḥ (discussed in Tattvaprakāśa v. 54). The categories are explained first by Hemadri, BhM p. 164, below.

185. MM 3.5.

186. The verse puns on the word atapa, which means both pain and sunlight.

187. Rāmacarita 19.89.

188. URC 1.28. The three poems cited both illustrate and corroborate the explanation of the mental transformation in the presence of the objective condition (Madhava seeing his beloved Malati; Sita reacting to Ravana's advances; Lakshmana's description of Rama's madness at the abduction of Sita). Bhoja's citations and brilliant analyses demonstrate throughout the deep

relevance of his analysis of rasa to actually existing literature.

189. I conjecture abhimānimanasaḥ (for manasaḥ; compare p. 636.23).

190. The three examples are the sudden shock of King Aja on seeing his dead wife; the increase of already existing feeling when Rati sees Vasanta while mourning for Kama; the emotional contrariety of Pururavas's feeling the south wind while longing for Urvashi.

191. RV 8.43.

192. KS 4.26.

193. VU 2.6.

194. The former set of items in each comparison is meant to indicate the interior and less distinct entities, the latter set the exterior and more distinct.

195. Compare NS 6.23.

196. KS 3.68. SKĀ p. 431 calls horripilation a psychophysical response (sāttvika), rather than an external transitory emotion as in ŚP here, and "dissimulation" (avahittha) a transitory emotion ("half-closed," paryasta-, in accordance with Bhoja's interpretation). ND (pp. 162-63) draws a similar distinction between external and internal reactions and feelings: " 'reactions,' i.e., firmness [or satisfaction, dhairya], etc., and perspiration, etc .; 'transitory emotions,' i.e., despair, etc., and sickness, etc. That is, each of these [including the vibhāvas] can be either mental or physical features."

197. Source unknown. The reactions are the glances and the words, the transitory feeling is shame (compare SKĀ p. 440). Mourning is inappropriate for heroes so long as vengeance has not been exacted (see SP p. 363, with reference to this v.).

198. MM 9.20. The transitory feeling is faintness.

199. The three broad categories of emotional states (typified as above by desire, anger, and grief) produce rasa by three different kinds of processes, which the similes effectively suggest (contrast Raghavan 1978: 646, the process of producing rasa "cannot be said to have been sufficiently explained by these similes").

200. MM 5.28, cited above.

201. A key to this passage is provided by DR 4.34 (n.b., ātmabhāvam nayaty anyān [bhāvān] sa sthāyī and here, ātmarūpatām nayante), and the discussion of Dhanika and Bhatta Narasimha (a stable emotion is not interrupted by the presence of [other] emotions whether contradictory or otherwise; it assimilates them and thereby becomes a source of [even greater] beauty). Bhoja is showing how the erotic rasa can assimilate the transitory emotions and other factors associated with the tragic, which enhance it, as a salty mixture takes on a heightened intensity when a drop of sweetness is added (note the close parallel with his earlier example, p. 634.7).

202. Sattasai 4.69.

203. Bhoja here turns Bharata's formula upside down (or perhaps adopts the counterintuitive assessment; see NS p. 286, above, and n. 22): rasa precedes the elements that are conventionally taken to produce it through their conjunction, and transforms them into itself. The example he cites, however-the trees of the garden where King Aja had walked with his queen, which are a stimulant of the tragic rasa, are themselves affected by it-seems rather too literal for the large conceptual argument he is making.

204. RV 8.70.

205. Ratnāvalī 4.19. Bhoja appears close to Abhinava in his understanding of the higher-order unity of rasas (śrngāra and vīra, or at least the concerns of kāma and dharma) in the Ratnāvalī. See Ingalls et al. 1990: 511, as well as Avaloka on DR 1.33.

206. Bhoja illustrates the species of stable emotion by Shiva's first passionate look at Uma (KS 3.67; see above SKĀ 5.138); the species of a specific rasa by Uma's realization that she and Shiva are equally in love (KS 5.85, compare SKÃ p. 431, where the verse is cited as an example of the nispatti of rasa); the species of the semblance of emotion (not "the semblance of rasa," per Raghavan's

note ad loc .; see below on p. 644.27, where Bhoja repeats and comments on the verse) by Ravana's semblance of desire for Sita and semblance of anger at her (from the Udāttarāghava of Matraraja/Mayuraja).

207. Having established that language qualities, rasas, stable emotions, etc., are all "ornaments," that is, "factors of beauty in a poem" (compare the discussion beginning on p. 624), Bhoja goes on to identify six varieties of mixture of elements: mixture of language qualities, of rasas, of figures of speech, of qualities and rasas, of qualities and figures, and of rasas and figures. The selection picks up the discussion as Bhoja returns to the question of rasa.

208. sollekha, nirullekha (compare Raghavan 1978: 303).

209. That is, when full development of the emotion is not possible (in the case of the erotic, when desire is felt by the antagonist for the heroine, or by an animal for an animal).

210. Varieties of "mixture" are generally distinguished as to whether the elements remain "manifest" or not. Here the pertinent condition is whether the mixed elements are constitutive of some end product (painting, man-lion, mud) or not (rice-sesame, mirror-image, milk-water); see p. 648.

211. MVC 5.32.

212. See KP 7.65 v. 340, and Arjunavarmadeva, p. 4, both below (and see DhĀ p. 195).

213. The tragic is prakrta and hence dominant, so the two compounds are not yathāsāmkhya.

214. Udāttarāghava. I read -trāsa- (with SKĀ p. 717, for-tryaśra). Ravana is speaking and referring to himself in the third person.

215. Sūktimuktāvalī, p. 249 (attributed to Chandra, a legendary Kashmir dramatist; see p. 32). Cakravaka birds are condemned to be separated from their mates each night.

216. This would seem to overdetermine the semblance, since for Bhoja animals cannot experience true emotions anyway; see SKĀ 5.8, above, and ŚP p. 616, below (contrast Ekāvalī p. 106, below).

217. AmŚ 160.

218. This is technically and usually considered a transitory feeling, but (per kārikā 11 and elsewhere) Bhoja explodes all these categories.

219. Recall that horripilation is an externalized transitory emotion (see p. 638, above).

220. ananda. Referred to as a rasa only here in the SP (see also Raghavan 1975: 137).

221. URC 6.11.

222. svātantrya, identified a rasa only here, like pāravaśya in what follows.

223. I read praśamā[ḥ] ... samkīryamāņāḥ samupalabhyante (with SKĀ pp. 719-20, for praśame ...- āņaḥ ...- ate).

224. A discussion of mixture of figures follows.

225. This closely follows SKĀ p. 720.

226. That is, where qualities and rasas are mutually constitutive, or where the former is a natural byproduct of the poet's concentration on producing the latter (compare DhĀ 2.17).

227. Bhoja here cites KA and DhA to show that certain rasas and certain qualities always will be copresent if each is to exist as what it is (e.g., raudra and ojas).

228. While apparently just repeating the definitions of certain sabdagunas and arthagunas given previously, Bhoja has in fact selected those qualities characterized by a constitutive relationship with the expression of affect. This is clear from the kinds of examples adduced in the discussion of qualities in chapter 9, beginning on p. 498. 229. Examples of the mixture and relative dominance of language qualities and rasas follow.

229. Setubandha 15.66 (correct the Prakrit to -pphamsamahagghavia-).

231. The following discussion of the last three types, which alone have a bearing on Bhoja's conception of rasa, closely follows SKĀ (p. 728), where the analysis is introduced with the following general remark: "The type of mixture in which a rasa is predominant over a figure comes about when it is the person experiencing the emotions who describes the situation, for the actual sentence containing the figure is then in fact his or her reaction in the form of a verbalization." Contrast the example just cited ("Rama's arrows ... "), where the sentiment, insofar as it is not being expressed by the one who experiences it, is submerged by the figure of sense.

232. Source unknown, but likely Sarvasena's lost Harivijaya (Kukarni 1991: 8).

233. The anger of Satyabhama, a wife of Hari/Vishnu/Krishna, was provoked by seeing the bouquet of the divine coral tree he had given to Rumini, another of his wives.

234. Source unknown.

235. Bhoja defines this as the transference to one thing of the features of some other thing (p. 598). In the present case, the wasting away of the speaker through the woman's indifference to him is transferred to the woman's waist; the "reason supplied," the "obstruction" of her shapely breasts, applies in both cases.

236. This last condition explains why this is a semblance of rasa. 6. This las 237. My translation of the second clause is uncertain. Note that the man of passion (abhimanin, etc.) finds pleasure even in the experience of neglect and pain (see kārikā 8 and n. 66).

237. "A reason that is counteracted and fails to produce its effect despite the presence of all the causal conditions, is the figure 'nonreason"" (ŚP p. 571.21): the ability of the lover to produce decorations is in fact evidence of insufficient love between the couple; "a circumlocution is [effected by] the use of a pretext, a turn of speech, or a special occasion" (ŚP p. 607.10), here presumably of the second variety, where what is meant-"my lover and I actually share a deeper love than you and your lover"-is expressed in a roundabout way. (In SKÃ p. 77, where the verse is also cited, Bhoja observes that the two women are rivals for the same man, but his use of priyavayasyā here instead of sapatnī suggests he understood the verse slightly differently on this occasion.)

238. "Trembling" is typically categorized as a psychophysical response.

239. AN AESTHETICS REVOLUTION, 900-1000

240. This was already with case with Mahima Bhatta in the mid-eleventh century, who regrets "never having seen the Mirror" (VV v. 4). No commentator on the famous précis in KP 4 after Someshvara Bhatta (and Manikyachandra, who borrows from him) and Ruyyaka-all twelfth-century Kashmir-appears to have had direct access to the text. Jayaratha (1200, Kashmir) is the last writer to quote independently from the work, but probably through an intermediary. I include below a citation from the Kaumudī, Uttungodaya's thirteenth- or fourteenth-century commentary on the DhAL, but this can hardly be authentic. (I omit the one from Prabhakara Bhatta's sixteenth-century Kāvyapradīpa, p. 4, which clearly presupposes the definition of poetry in the KP, although he gracefully summarizes Bhatta Nayaka's system: kāvyārthaparyantābhidhā ratyādivișayā bhāvanā sahrdayavisayaś ca bhoga iti, Rasapradīpa, p. 26.)

241. RT 5.159. A verse is attributed to him in the Sarngadharapaddhati (#1020).

242. Vedantic leanings seem likely from his commentary on the Treatise on Drama (below), though missing from other fragments.

243. For details see Pollock 2010; also Rajendran 2004.

244. If I understand him correctly, Krishnamoorthy seems to be the first scholar to have appreciated this (1979a: 215). Bhatta Nayaka had important observations to make about poetic language too, which shows his abiding interest in literary form.

245. "Literal meaning" and "denotation" (Ingalls et al. 1990: 221, Gnoli 1968: 45, respectively) are imprecise; see also n. 30 below.

Chintamani's critique of De (1927: 271 n. 1) is also incorrect, and the true reading of the Locana passage is in fact abhidhā vilaksaņā eva (p. 183): literary language is "completely dissimilar" from the language of science [śāstra; or, just possibly, Veda] or everyday discourse, and it is the process of bhavakatva that renders it so. See also Someshvara Bhatta on KP p. 44.15, and n. 57 below.

7. Often tadbhogikrttva, i.e., of the "commonized" emotional complex (tad). Some later writers such as Jagannatha render this (oddly) as bhogakrttva.

8. Abhinava reappropriates this triad for his own view (DhAL 2.4; Ingalls et al. 1990: 225).

9. vyāpāra, one of Bhatta Nayaka's important terms (see vyāpāraprādhānye kavyagīr bhavet, #6 below), is often used in Mimamsa to define bhāvanā.

10. So Avaloka on 4.37, below.

11. bhāvanaiva hi vākyārthaḥ, Ślokavārttika Vākyādhikaraņa v. 330.

12. RG p. 28, below.

13. KM ed., pp. 29.9 and 30.2.

14. Manikyachandra (ed. Mysore, 1922), p. 147 (the verse is absent from Mysore 2nd ed. and the Ananadashrama edition; cited also by Bhattacharya 1964: 32 n., who misquotes and interprets differently).

15. Dubious citations have been enclosed in brackets. For full annotations see Pollock 2010.

16. ABh 1.5, explaining the first verse of Bharata's NS.

17. I read -svālambanam (for -svālambana-); compare also Karnakagomin on Pramāņavārttika: svālambanam eveti vikalpapratibhāsam eva (p. 170).

18. The phrase "learning, meditating, and so on" references Vedantic interpretations of Brhadāranyaka Upanișad 2.4.5.

19. The quote is NS 1.329 (the reading here, śāntād utpadyate rasaḥ, disagrees with the printed text in chapter 7, sāntād bhāvaḥ pravartate). A corrupt citation from Bhatta Nayaka's NŚ commentary, NŚ 3.305 (Abhinava also quotes him briefly and obscurely in his refutation, pp. 309-10), suggests his acceptance of śāntarasa. Dhanika, Bhatta Nayaka's most fervid disciple, argues forcefully against the possibility of śāntarasa in drama (Avaloka, p. 202, below).

20. Abhinava adds that the foregoing interpretation underlies a verse in Bhatta Nayaka's Hrdayadarpana ("Homage to Shiva ... "), which he goes on to cite.

21. ABh 1.6. Probably the first verse of the treatise.

22. DhĀL pp. 91-92 (v. l. ekam for etam). "At the thirsty promptings of," -trsņayā; "in affection for" (Ramasaraka).

23. Spiritual adepts achieve "rasa"-in the sense of ultimate reality (Taittirīya Upanisad 27.2)-only with enormous effort; literature, however, provides her "calves," the lovers of art, with rasa unbidden.

24. Ruyyaka on VV p. 101 (I am uncertain about the translation of pādad, bhede 'pi hi tad asti yat).

25. DhAL p. 87. Abhinava cites this in support of his view that Valmiki, the first poet, did not create poetry directly out of experience but only once that experience was processed through creative genius.

26. DhÃL p. 87; ABh 2.298 (which reads arthe tattvena yukte tu; so too Someshvara KP p. 3]); KA p. 5; Ruyyaka, KP, p. 1, introduction; Jayaratha, AS, p. 9 (reading -dhīr for -gīr, "we have what we think of as literature"). Abhinava's criticism here turns on the meaning of "process" (vyāpāra): if this refers to the relishing of rasa based on implicature, then Bhatta Nayaka is saying nothing new; if it refers to signification or expression as such (abhidhaiva), then this has already been shown not to have primacy (DhĀL p. 87). Both objections seem odd: Bhatta Nayaka rejected implicature, as Abhinava well knew, and he also defined "process" far more broadly than expression, as the three-part unity expression, actualization, and experientialization (Abhinava, however, repeats the narrow interpretation in NS 2.298.9: bhattanāyakenāpi kuśalaśikșitābhidhāvyāpārapradhānam kāvyam; see also Kulkarni 2003:38).

27. DhĀL p. 39. See DR 1.6, which echoes this verse. Uttungodaya (and independently Ramasaraka) understands that the capacity to experience rasa constitutes the qualification for literature: "every person who experiences rasa is qualified for literature, every person who seeks knowledge ... is not" (p. 78). But Uttungodaya had no firsthand knowledge of the Hrdayadarpana and cannot be regarded as an unimpeachable interpreter (pace Ingalls et al. 1990: 73); the only independent "citation" of Bhatta Nayaka in the Kaumudī is #8 below, but unlikely to be his ipsissima verba. Another interpretation: every process in literature offers rasa (like a beloved); it does not offer knowledge (like a friend) or give commandments (like a master). Abhinava cites the verse to show that Bhatta Nayaka agrees that rasacarvanā is the principal matter in literary experience. But the point of contention between Anandavardhana and Bhatta Nayaka is not whether rasa is the telos of literature, but how rasa works.

28. Bhatta Nayaka may have been the first to fully formulate the idea concerning the three aims of formal discourse: enjoyment (literature, correlated with the beloved), advice (correlated with the friend), moral precepts (correlated with the master). (The full analogy is available only in DhĀL p. 40).

29. Kaumudī on DhĀL p. 79 (compare Krishnamoorthy 1979a: 221). Aside from the doctrinal issues (see next note), the art meter of the verse speaks against its authenticity.

30. That is, "all domains." But to the degree that the definition of abhidha in #11 demarcates literary language as a particular form of language, per n. 57 below, this definition would be too wide (though see also Samudrabandha in fragment #4a).

31. DhĀL p. 39; Jayaratha on AS p. 9. Abhinava cites the verse to indicate that the position represented by Bhatta Nayaka has been rejected by the DhA. (De 1988, vol. 2: 183 misunderstands the DhAL when he writes that Bhatta Nayaka "accepts the suggestion of rasa as the essence of poetry").

32. Presumably, distinct from literal and figurative meaning (not expression and actualization, per Ramasaraka, since implicature is not of the same order as actualization and so is not to be distinguished from it).

33. DhAL p. 88. I omit Bhatta Nayaka's specific criticisms ad loc. of Anandavardhana's analysis of two famous verses, "You're free to go wandering" (see Avaloka p. 206, below, and n. 170 there), and "Here sleep I" (DhAL pp. 68 and 72; see Ingalls et al. 1990: 90-91 and 99); and his downgrading of dhvani in DhĀL pp. 171 and 172 (see Ingalls et al. 1990: 209 and 211).

34. ABh 1.271; KA pp. 96-97 (vv. 1-2); Jayaratha on AS p. 9. In the original text the verses may not have been consecutive.

35. I read -krtir with Jayaratha (for Hemachandra's -krtam).

36. I read matah (with Jayaratha, for hi yat with Abhinavagupta and Hemachandra, which would seem to take tad- in c as correlative, when it is part of the compound that is a technical term, as v. 1 indicates); and again with Jayaratha -krti- for -krta -.

37. Jayaratha includes hereafter a half-verse that is difficult to interpret without further context: "Or, just by virtue of being perceived, this clearly becomes an aid to spiritual release."

38. KA p. 97. Hemachandra adds that the view expressed here has (Abhinava's? KLV: Bhatta Tota's) full endorsement. In DR, the idea of rasa as a vakya is clearly part and parcel of the argument against vyañjana and in favor of the tatparya position (Avaloka p. 211, below; Bahurupa Mishra's avataraņikā seems to allude to this verse of Bhatta Nayaka's), a linkage that will be misunderstood later in the tradition (Kumarasvamin on PR 4.137 and RG p. 30.2, below; see Pollock 2012b).

39. "Veda": śastre (compare KLV p. 307.24, who otherwise misunderstands the verse). I conjecture samsargad dhi (for KA's samsargadir; note that Dhanika on DR 4.37 offers samsrsta, and refers to no other linguistic operations). For the idea here compare PMS 2.1.47: arthaikatvād ekam vākyam (and see Vākyārthamātrkā 11.5, p. 378: ekakāryatvād vākyam apy ekam). The Bhatta Mimamsakas hold that sentence meaning is a samsarga, or conglomeration, of word meanings, or more precisely, that the class properties denoted by words take on specific meanings through conjunction with other words in sentences (see TV [Chowkhamba ed.] p. 429). See also ŚP p. 418: vākyārthaś ca kriyākārakasamsargarūpaḥ.

40. ABh 1.271. For variants and the problems of ascription and interpretation, see Pollock 2012b: 241-44.

41. SD p. 11, and ascribed to the HD by a sixteenth-century author who cannot possibly have had access to it (Prabhakara,

Rasapradīpa, p. 4). (This certainly looks like yet another attack on Mammata's definition of poetry as adosa, and thus long postdating Bhatta Nayaka, pace Kane 1971: 224).

42. ABh, pp. 270-71 and KA, pp. 96-97, supplemented by the version in DhĀL pp. 180-83, the précis of Someshvara (S) (RORI ed. v. 1, pp. 43-44) and of Manikyachandra (M) (Mysore ed. vol. 1, pp. 216-21), and KLV (p. 306). As in the case of Shri Shankuka, the terseness of the ABh account requires exegesis based on these supplements, which I have incorporated into the translation.

43. In DhAL (p. 182.1), this criticism is linked with the utpattipaksa, not as here with the pratītipaksa.

44. Reason 3 seems dependent on 4, but they are independent in all versions (they are reversed in the DhAL, but still unconnected). Both ABh and KA read sādhāraņīkaraņāyogyatvāt (notably missing in DhĀL), but taking it in Bhatta Nayaka's technical sense is awkward, for precisely in such circumstances is "commonization" supposed to work (as Dhanika clearly shows, on DR 4.40-41ab). DhĀL's pūrvapaksin gives the right sense: kāntātvam sādhāraņam, implying something like sādhāraņyāyogyatvāt in the siddhanta, which is what I adopt in the translation (this also construes well with asādharanyat in the next line).

45. An act typically ascribed to Hanuman; DhAL replaces with "Rama's bridging the ocean." The odd variation is found elsewhere as well.

46. The clause pratyaksad iva, etc., is found here in all sources, but the argument belongs at the end of the following paragraph, since it concerns the problem with perceiving rasa externally (this is how DR, which undoubtedly had direct access to the original, presents it, 4.39, pratīti-[= tātasthya ]vrīdersyā-, etc .; things are not helped by reading, as does Gnoli, pratyaksād iva. nāyaka-, which is anyway contradicted by DhAL, p. 182.1). Its current position is mystifying, and suggests an old corruption dutifully reproduced.

47. I read tan na (with Someshvara, for tatra in both texts).

48. On "manifestation" as a linguistic phenomenon, which Bhatta Nayaka rejects (see fragment #8), as distinct from the psychological phenomenon that it will later be held to be, see the introduction, section 4.2, and Pollock 2012b.

49. So that one would never be said to have a full rasa experience (see Ruyyaka below, Pollock 2012b: 250-51, and contrast Ingalls et al. 1990: 221).

50. What follows are features of the expression process, not of actualization. Ruyyaka on KP below assimilates them to actualization as contributory elements, but I suspect he is making the best of a corrupt text (expect abhidhāvyāpāreņa abhidhīyamāno after -abhinayarūpeņa).

51. Like the play's music, literary language has a transformative mental impact ( (see ABh 1.285.16: svagatakrodhaśokādisankața hrdayagranthibhañjanāya gītādiprakriyā, and compare DhĀL p. 189.6).

52. druti, vistāra, vikāsa (H reads vistara, which is authenticated by DR 4.43; see also Mallinatha on Ekāvalī p. 96; Ingalls et al. 1990: 36, like KLV p. 307.3, seem to have misread vikāsa as vikāśa and hence translated as "radiance"). The fourfold analysis given by Dhanika (on DR 4.43-45ab, below, and n. 230) is irreconcilable with what we find here, and indubitably authentic. Abhinava's criticism-that limiting the "apprehensions" or experiences of rasa to these three alone makes no sense, since there should be as many responses as there as rasas (ABh 1.271.15-18)-is actually preempted by Dhanika's exposition of what seems to be Bhatta Nayaka's actual position.

53. KP p. 560.

54. I read -satattvena (with Shridhara, for -sattvena).

55. From Ruyyaka's commentary on KP p. 563.7 (ed. Bhattacharya p. 16). Confusions in this précis lead one to doubt the manuscript or the ascription to Ruyyaka.

56. The specific traits of a character like Sita.

57. śabdavyāpāra. Ruyyaka is using this as a synonym of abhidhā (this more expansive sense is found also in Mukula Bhatta, mukhyena lākşaņikena vābhidhāvyāpāreņa, Abhidhāvrttamātrkā, p. 1, and comes from Udbhata; see DhĀL p. 32 and Bronner ms.).

58. Since for Bhatta Nayaka, per the above extract, it is the figures of sound and sense that are involved rather than direct and indirect meaning, we must assume that "indirect meaning" (santarartha) is another way of characterizing figures of speech (recall that for thinkers from Dandin to Bhoja, vakrokti was a synonym of alankāra as such). Vishveshvara (Rasacandrikā p. 45) takes the two terms to refer to śakti and laksanā respectively.

59. In the standard distinction, perhaps introduced by Bhatta Nayaka himself, scripture is like a master in that it commands; historical discourse is like a friend in that it advises; literature is like a beloved in that it seduces.

60. That is, for anyone other than her husband Rama, such as the spectator.

61. The reading bhavad bhavanam is dubious, but is found in both editions and in Shridhara's KP commentary, which essentially

reproduces Ruyyaka.

62. I conjecture guņālankārayogāt (for guņālankāratvād, see guņālankārayoga- in Shridhara KP p. 565; compare also guņālankāramayatva- in ABh 1.271.1 and KA p. 96.19 cited above).

63. AS of Ruyyaka (R), p. 9, with the commentaries of Jayaratha (J) and Samudrabandha (S).

64. The text has been imperfectly transmitted, with Jayaratha alone preserving the truth: kāvyāmśatvam bruvatā (Vidyachakravartin falsely reads kāvyātmatva, for kāvyāmśatva, ed. Janaki pp. 11-12; Samudrabandha is adrift). The "brazen" admission refers to #9 above (and/or possibly to #13).

65. -uttīrna, borrowed from DhAL p. 52. Since bhoga is the result of bhavakatva, it cannot be "far beyond" it (Ingalls et al. 1990: 81).

66. The reference is to DhA 1.1c.

67. Fragment #9. I conjecture rupatā (for rūpitā).

68. kavikarmanah. This is confusing, since vyāpāra is not itself the literary work (the normal meaning of the term kavikarma), but the work's capacity to produce an experience of it.

69. For this mischaracterization, compare n. 247 on RG 29.10-11, below.

70. From Mallinatha's "Central Gem" (c. 1400), p. 85.

71. I conjecture -krta- (for -krtena).

72. That is, most importantly, the character.

73. Here Mallinatha draws a distinction not previously noticed: the final process, perhaps the process as such (which Bhatta Nayaka has in mind when he says literature is vyāpārapradhāna, #5) is distinguished here as bhāvanā, whereas the second component, the subordinate element of commonization, is called bhāvakatva. Bhatta Nayaka himself above uses anyā bhāvanā for bhāvaka (tva).

74. RAS, p. 251.

75. See Avaloka p. 168, below.

76. I read -vyāpāreņa (for -paryāyeņa). For Bhatta Nayaka, bhāvanā/bhāvakatva is a second function, different from abhidhā, see #11 (and abhidhāto dvitīyena, ABh 1.271.2-3).

77. viparivartamāna -. See n. 89.

78. The impropriety would stem from the fact that Sita is another man's wife.

79. DR 4.87.

80. Munja's first extant charter was issued in 974-75 (Corpus Inscriptionum Indicarum vol. 7, part 2, pp. 10ff.). The last of his predecessor, Siyaka, was issued in 969 (pp. 8 ff.), however, and thus Munja could well have been king as early as then.

81. De 1988: 125, for example.

82. Noted first, I believe, in Kane 1971: 248.

83. See also Raghavan 1978: 667-68 (correcting his earlier statement, p. 89). I suppose some of Bhoja's thinking about tātparya (SP pp. 366-90) could have been informed by Dhanika's restatement of Bhatta Nayaka (Avaloka on 4.37, below, and the selections from his Kāvyanirņaya, p. 212, below), but the case is hardly decisive (pace Venkatacharya 1969: lviii).

84. No doubt the process could be seen as the reverse one too: what must rasa be (the ontological question) in order for our experience of it to make sense (the epistmological question)?

85. Bhatta Nayaka, the source of Dhanamjaya's and Dhanika's theory, is also unknown to Bhoja; see Pollock 1998a n. 57 (pace Raghavan 1978: 469) and Pollock 2009.

86. See headnote to chapter 2. Until recently it was the only known commentary on that chapter.

87. Pollock 1998a: 132-33; Raghavan 1978: 412.

88. As I have earlier observed (1998: 137 n. 40); contrast Venkatacharya 1969: lxv. See also Kamimura 1976.

89. viparivartamāna -. For the use of this word in apparent reference to Bhatta Nayaka, see ABh 1.273.8, and RAS (Bhatta Nayaka #1b). The term including the nominal form viparivrtti is used by earlier writers with reference to the psychological process whereby words form sentences in the mind (for Prabhakara Mimamsa it is a synonym of samnidhi; see Prakaranapañcikā p. 389: atha sannidhiḥ kaḥ? yasyārthasya śravaņānantaram ākānksāyogyatābhyām arthāntare buddhiviparivrttiḥ). Compare Śābarabhāsya 4.3.11 [BI ed. p. 498].

90. The literary text, which is described as "rasa-laden," only causes rasa to arise in the viewer/reader.

91. Prabhā brings out the implication of this vague formulation: "As, for example, when a poet says, 'This man is the god of love himself,' or 'The woman is Rati herself, wife of the god of love." It is literary language as such that, for the viewer/reader, transforms quotidian things into aesthetic elements; see Dhanika on 4.46-47ab, below, and Bhatta Nayaka fragment #10.

92. NŚ 1.340.

93. -bāhyasattva -; compare Helaraja: cirakālātītatvāt kamsādinām bahirasattvād ... (Vākyapadīya v. 3 part 2, ed. Sharma, p. 99). As Helaraja makes clear, the problem at issue is verbal narrative, where in contrast to painting, sculpture, or drama, there is no actual physical image or likeness (pratikrti). Presumably, in the case of the poem about Urvasi (the illustration cited from VU p. 170, omitted here), while the physical "nymph" has an existence on the stage, she only can be said to become a foundational factor for the viewer when she is described in literary language.

94. śabdopadhānāt. Given Bhatta Nayaka's new response theory of rasa, Dhanika must ask how the factors work for viewers/readers, for whom they do not exist in the same way they do for the actual characters. (Improbable is Bhatta Nrisimha's notion that the question is how figuratively constituted factors can be efficacious, since they correspond to nothing in the real world.)

95. bhavaka, an early (if not the first) use of the term. It refers exclusively to the viewer/reader who "actualizes" the experience of the main character, the term referencing bhāvanā, actualization, as described here and in 4.38.

96. As Dhanika's earlier restatement shows, he thinks the subject of the verb is the listener, not Patanjali (so Helaraja; this indicates that, n. 93 above aside, no conclusions can be drawn regarding Dhanika's knowledge of Helaraja).

97. Krishna's evil uncle and king of Mathura, whom he eventually slays.

98. Vākyapadīya 3.7.5. The point of the verse seems to be the purely mental status of syntactical relations, of which storytelling is a sort of limit case: storytellers create a reality of their own, and when in their story they "have Krishna kill Kamsa," those narrated characters can seem to have an agency of their own. See also Filliozat 1963: 344-45.

99. Bharata's NS is traditionally said to contain six thousand verses (Dhanika's reference appears to be the first). The citation is NŚ 1.342.10.

100. sāmānyagunena. While the original usage in NS is unclear, it would appear that Dhanika's understanding of the phrase is an anachronism, meant to provide a deeper historical pedigree to Bhatta Nayaka's concept of "commonization." Similarly in the case of Abhinava; see ABh 1.275.6, below.

101. I omit the examples of the foundational and stimulant factors given by Dhanika (which also show how hyperbole and

metaphor respectively create the factors, as he notes earlier).

102. bhava-, which Bahurupa Mishra takes to refer to any of the 49 emotions (the position of Induraja, on KASS 4.2, above).

103. The novelty of Dhanika's thought has engendered much confusion in the manuscripts, and as a result the precise reading is uncertain. Presumably the reactions enable viewers/readers to experience their own stable emotions as well as the character's. Dhanika's remarks on the "sensitivities" in what follows suggests that the reactions might even be those of the viewer/reader (so ND, see p. 160, below and n. 23 there, and RG p. 55).

104. "Reactions" are not only the visible physical effects in the character of his rasa but also causes of rasa in the viewer/reader, and can therefore pertain to narrative no less than performance. (Bhatta Nrisimha comments, rather unhelpfully: "Things that enhance the rasa can be considered reactions, and since both poetry and dramatic representation enhance the rasa, reactions can be said to belong to both.")

105. I.e., of what the characters experience on the stage/in the poem.

106. "Real-world" (laukika), that is, of the storyworld. This is Dhanika's term for the rasa experience of the character (see Bahurupa Mishra's introduction to 4.46), and is differentiated from "the rasa of drama" (nātyarasa) and "the rasa of poetry" (kāvyarasa, on 4.42cd), that is, the rasa experience of the viewer/reader (Dhanika on 4.41cd; he can thus differentiate "real-world tragedy" and "literary tragedy," see 4.43-45ab), though for Dhanika the "real-world" variety is no longer rasa in the full sense. See also ND pp. 160-61, below (which makes the exact same distinctions); PR 4.138, p. 210, below. Note also Nagesha's gloss of "character under description" as "the substratum of [i.e., the character who experiences] real-world rasa" (varņanīyaśabdena ca laukikarasāśrayādir ucyate), KP p. 141. The use of the term in NS 1.283.2 and 6.33 may be related (and suggests further the revisionist nature of the passage).

107. "Here," that is, in the realm of aesthetic response (the commentator's remark below notwithstanding). In the storyworld, reactions are an effect in the main character of his own emotion; in the world of aesthetic response, the character's reactions are a cause of rasa in the audience. The same distinction is drawn in ND p. 160.8-9, below.

108. Dhanika illustrates Dhanamjaya's definition with a poem of his own (omitted).