1As a student of the Sanskritic disciplines of literary and dramatic theory, I am interested in better understanding the way that scholars writing in Tamil approached a subject of shared theoretical interest, the representation of literary emotion. In this essay, I attempt to trace the common conversation between the scholarship taking place in the two languages in roughly the same time and place, in the Tamil country over the course of the period that I call the long twelfth century.2 I begin by offering a new interpretation of the meaning of the opening three cūttirams of the Tŏlkāppiyam Pŏruḷatikāram’s Mĕyppāṭṭiyal (“section on mĕyppāṭu,” MI), as they were understood by the scholiast Iḷampūraṇar. This interpretation, which depends on a reconstruction of the sources explicitly and implicitly drawn upon by this scholar, occasions a further enquiry into the structure and compositional logic of several interrelated passages in Śāradātanaya’s Bhāvaprakāśana, a dramaturgical work composed in Sanskrit.

2I contend that it is through the juxtaposition of these works, near to each other in space and (as I will argue) in time, that we are able to better understand their individual arguments; in fact, that the adequate interpretation of the particular doctrines that I will review is only possible through the juxtaposition of these texts and the works that they draw upon. When I first undertook the research, the results of which are presented here, I presumed there to be an inherent methodological good in the practice of adducing parallels, understanding acknowledged sources, and tracking down unacknowledged ones — in short, what could broadly be called contextualization. As it turns out, contextualization did furnish useful information for understanding Iḷampūraṇar’s and Śāradātanaya’s ideas and the structure of their texts, as well as incidentally providing data to advance some text-historical hypotheses of a more general significance. But at the same time, in a self-critical vein, this exercise in trying to read just a few pages of two contemporaneous authors has produced, for me at least, some interesting complications about how we can and should read shastric texts from medieval Tamilnadu, and why it matters that we should do so. It is these final, broader questions that I will emphasize by way of conclusion.

1. Iḷampūraṇar and the question of mĕyppāṭu

3We know practically nothing about Iḷampūraṇar — the opening biographical eulogy to his text is a 19^th-century invention — although he can be reliably placed in the Coḻa country, and was active sometime before the 13^th century;3 as I will argue further below, this date can be further refined. Iḷampūraṇar’s commentary on the entire Tŏlkāppiyam is the earliest to come down to us. The MI, the section of his root-text that I will concern myself with here, has in recent times had its reputation impugned due to the quite obvious influence of Sanskrit sources on its design and conception. For contemporary scholarship interested in dividing out the text of the Tŏlkāppiyam into chronological layers, the MI, along with the section immediately following it, the uvamaiyiyal, on similes, has been unceremoniously dumped into the later fringes of the text, consisting of the less authentic materials cast (so goes the argument) under the deranging influence of Northern models. These sections are thus unfavorably contrasted to what is thought to be the authentic and indigenous literary theoretical scheme of the akattiṇai and puṟattiṇai, the celebrated poetics of love and war.4 Iḷampūraṇar’s interpretation of the MI has suffered in more recent times, but for rather different reasons: it has been effectively written out of the later history of Tamil poetics in favor of the altogether more straightforward presentation of the later commentator Perāciriyar.

4The theory of mĕyppāṭu, which attempts to account for the representation of emotional states as an essential element of literary enjoyment, allowing for eight basic emotional states, does in fact appear transparently to be indebted to the theory of rasa familiar from Sanskritic literary theory. It was P.S. Subrahmanya Sastri who first pointed out the parallels between the text of the MI and the received text of the sixth and seventh adhyāyas of Bharata’s Nāṭyaśāstra (NŚ). This discovery — an impressive piece of philological argument that Sastri first set out in his University of Madras thesis — has been reproduced in all subsequent scholarship on the subject, mostly without any further examination.5 Any attempts to understand the status of mĕyppāṭu are further hampered by late medieval scholastic efforts to reconcile this one set of critical vocabulary with that of the akappŏruḷand puṟappŏruḷ, reducing mĕyppāṭu to a watered-down and unsatisfying adjunct to the other theory. This effort at rationalization seems to have reached its apogee in the Vaittiyanāta Tecikar’s Ilakkaṇaviḷakkam (17^th century), and its legacy strongly colors, for instance, the recent survey of Indra Manuel (1997), whose synchronic view suppresses this centuries-long process of synthesis.

5There is, however, a terminological and hence conceptual difficulty at work in all of this. What Subrahmanya Sastri in fact demonstrated was that the definition sūtras for each of the eight mĕyppāṭus were indebted to the NŚ’s definition of the sthāyibhāvas, as detailed in the latter’s seventh adhyāya, the so-called bhāvādhyāya. There is, however, no terminological equivalent in the Tŏlkāppiyam’s text to the rasas that are the subject of the NŚ’s previous chapter: the term cuvai, which in later scholarly Tamil is the functional calque for rasa, strikingly occurs nowhere in the text of the Tŏlkāppiyam. As a result — and somewhat oddly, given the universal acceptance of the direction of influence from Sanskrit sources — what little scholarly discussion there has been of mĕyppāṭu has tended to argue for it as an index of the aesthetic difference of Tamil poetry as measured against Sanskrit.

6Norman Cutler’s judgment is representative (1987: 61), when he argues that “the Tamil poeticians seemed to be unaware of the essential difference between character and event in the purely literary context and in the ‘real’ world inhabited by the audience”, that is, in the putative difference that distinguishes (intratextual) bhāva and (extratextual) rasa.6 Extending Cutler’s thinking, Martha Selby has argued that for Tamil poetic theory “[a] Tamil reader’s response to a poem has been traditionally understood as empathetic and direct ... [it] did not require much mental reflection or intellectual articulation. It did not require a ‘supernormal relishing’ or, to put it more precisely, the ‘feeling of a feeling”’ (Selby 2000: 32–33). Like Cutler, Selby understands the status of mĕyppāṭu within the critical-interpretative tradition to be of only a subsidiary or auxiliary importance; by contrast, Monius (2001: 34–35), being explicitly interested in the affective contents of Tamil literary texts and faced with the relative theoretical poverty of the received theory of mĕyppāṭu, has to have recourse to the model of virtuoso interpretation that is (in her reading) the hallmark of the medieval Kashmirian literary theorists. Cutler and Selby’s interpretation of mĕyppāṭu understands the noun to mean “the conditions” [°pāṭu] “of the body” [mĕy°], while Monius opts for the parallel translation “occurring in the body”. This somatic understanding of the word has tended to reduce the status of mĕyppāṭu even further, suggesting from Subrahmanya Sastri’s time onwards an equation with the NŚ’s sāttvikabhāvas, the “natural” or involuntary reactions with which actors are meant to signal their represented emotions (defined and detailed in NŚ7.94–107).

7This idea of mĕyppāṭu as something exclusively somatic was already available in Iḷampūraṇar’s time. Commenting upon Puttamittiraṉār’s Vīracoḻiyam 3.5 (assigned to the reign of Vīrarājendra Coḻa and so completed ca. 1060–1068 CE), Puttamittiraṉār’s direct pupil Pěruntevaṉār, writing in verse (and so possibly citing another authority?), states měyppāṭṭiyalvakai mētaka virippiṉ / měykkaṭpaṭṭu viḷaṅkiya toṟṟañ / cěvvitil těrintu cěppal maṟṟ’ atuve, “To expand upon the variety of mĕyppāṭu: it is the manifestation that appears in the body, as well as the verbal expression [of it], when ill-health becomes apparent” (pp. 78–79). Besides the implied etymology evident in the phrase měykkaṭpaṭṭu (to which I return below), Pěruntevaṉār points towards another of the major problems of the interpretation of mĕyppāṭu, its status as both physical and verbal representation; this is a problem, as we shall see, that would also trouble the Tŏlkāppiyam commentators. In fact, this etymology of mĕyppāṭu as “that which occurs in the body” is not the only way to understand the word: there was, as we shall see, another competing interpretation that was certainly available to Iḷampūraṇar, which was quite plausibly the sense which the Tŏlkāppiyam author-compilers themselves intended.

8But understanding Tŏlkāppiyam’s original sense of mĕyppāṭu is no easy task, in large part due to the obscurity of its own account. The three opening cūttirams of the Mĕyppāṭṭiyal read:

paṇṇait toṉṟiya eṇṇāṉku pŏruḷum
kaṇṇiya puṟaṉe nāṉāṉk’ĕṉpa (MI 1 = PA 2457)

They say that that domain [consisting of] all of the eight times four elements which appear in the field amounts to four times four.

nāl iraṇṭ’ākum pālumār uṇṭe (MI 2 = PA 246)

And all of this set becomes four times two.

nakaiye aḻukai iḷivaral maruṭkai
accam pĕrumitam vĕkuḷi uvakaiy ĕṉṟu
appāl ĕṭṭām mĕyppāṭ’ĕṉpa (MI 3 = PA 247)

And indeed they say that this set are the eight mĕyppāṭus: nakai (laughter), aḻukai (weeping), iḷivaral (disgust), maruṭkai (wonder), accam (fear), pĕrumitam (boldness), vĕkuḷi (anger), and uvakai (desire).

9The explicit text of the Pŏruḷatikāram thus provides its readers with a series of diminishing equations, 32 = 16 = 8, with only the final cūttiram providing something to hang all of this on. Sūtras are of course supposed to be terse and to call out for commentary, but this opening — and indeed some of the other parts of the text’s outline of mĕyppāṭu — almost seems like a parody of the style. The unprepared reader has no idea to what these initial combinatoric exercises refer, or what the difference, for instance, really is between paṇṇai, puṟaṉ, or pāl (which I have respectively, if tentatively, translated as “field”, “domain”, and “set”). The first of these, paṇṇai, “field” is especially tricky, and occasioned strongly divergent interpretations. In his opening comments to the cūttiram, Iḷampūraṇar writes:

īṇṭuc cŏllappaṭukiṉṟa patiṉāṟu pŏruḷum kaṟṟu nallŏḻukku ŏḻukum aṟivuṭaiyār avaikkaṇtoṉṟāmaiyāṟ ‘paṇṇait toṉṟiya’ ĕṉṟār. ĕṉṉai? nakaikkuk kāraṇamākiya ĕḷḷal avarkaṇtoṉṟāmaiyiṉ. piṟavum aṉṉa.
[The author] said that all sixteen of the elements mentioned here “appear in the field”, as they do not appear in the assemblies of wise men, who possess proper conduct and learning. Why is this? Because ĕḷḷal (mocking laughter), which is a cause of nakai (or humor), does not appear among these men. And there are other cases that are similar to this.

10Here, Iḷampūraṇar evidently wished to widen the scope of literary representation beyond the confines of what we might call “high culture”, and accordingly saw “the field” (paṇṇai has strong agricultural resonances) as a reference to the broader world of human life. Perāciriyar’s understanding of paṇṇai is notably different (p. 2): paṇṇait toṉṟiya ... muṭiyuṭaiventaruṅ kuṟunilamaṉṉaru’ mutalāṉor nāṭakamakaḷir āṭalum pāṭalum kaṇṭuṅkeṭṭuṅ kāmanukarum iṉpaviḷaiyāṭṭiṉuḷ toṉṟiya, “‘which appear in the paṇṇai’ [means] ‘which appears in the pleasant entertainment in which such men as crowned kings and lesser rulers watch and listen to the dancing and singing of actresses, and have their desire excited’”. This strong delineation of a courtly performance context, however, seems exactly the opposite of Iḷampūraṇar’s interpretation, of which Perāciriyar was certainly aware. We must allow, then, for the fact that the brief testimony of the Tŏlkāppiyam was already subject to controversy in our earliest surviving scholiasts, and we must be prepared to see their efforts as constructive rather than simply expository.

11More than just an attempt at an adequate explanation of an underdetermined root-text, Iḷampūraṇar’s commentary gives the impression of an independent departure from the barest outline provided by the Tŏlkāppiyam. After first dealing with the first cūttiram’s rather tangled syntax, Iḷampūraṇar sets out to explain what the set of 32 consists of, in order to effect the first of the reductions. To summarize, these four sets of eight are understood by Iḷampūraṇar as

1. eight causes (kāraṇam or etu), for instance, eḷḷal, mockery, is a cause of nakai, laughter (here anticipating MI 4 [= PA 248], where Iḷampūraṇar refers to the four items beginning with eḷḷal as either kāraṇam or nakaippŏruḷ). Elsewhere, (p. 3) he refers to these as cuvaippaṭupǒruḷ, “materials that are tasted/experienced”;
2. the eight flavors (cuvai);

3. their markers (kuṟippu);

4. and the physical symptoms (cattuvam).

12Perāciriyar, in his later gloss of this same cūttiram, gives a parallel but distinct list replacing the first two terms with cuvaippǒruḷ (the represented “raw materials” of cuvai) and cuvaiyuṇarvu (“cuvai awareness”), retaining and indeed expanding on Iḷampūraṇar’s importation of the centrality of cuvai against the testimony of the Tŏlkāppiyam itself (pp. 3–6), while refining the cognitive processes at work in his understanding.

13The relationship of all of this to the classic theory set out in the Nāṭyaśāstra is clear enough, as the elements presented by Iḷampūraṇar seem to line up, respectively, with Bharata’s categories of vibhāva, rasa, sthāyibhāva, and sāttvikabhāva (the correspondence however is not exact, as there are more than eight vibhāvas in the Nāṭyaśāstra’s version of the theory). It is important to note, however, just how much Iḷampūraṇar is bringing into the discussion here, especially in his importation of the term cuvai itself, which, to repeat, does not occur in the Tŏlkāppiyam at all. Iḷampūraṇar is not actively trying to foist a perverse interpretation onto the text he is commenting upon; rather, he shows himself to participate in the confident interpretative theory — more familiar from thinkers working in Sanskrit — that denies innovation as a part of classical Indic systematic thought, the ideology that Pollock has detailed in a series of articles.8 In defence of his seemingly unwarranted introduction of the discussion of cuvai here, and of the larger apparatus of its communication, Iḷampūraṇar evidently wishes to reconcile his root-text with other systems of thought, working from the presumption that these other ideas, if valid, are always already present in an authoritative work like the Tŏlkāppiyam.

14He is explicit, furthermore, as to the source from which he derives this emphasis, or at least clear as to the authority he is recruiting to justify its inclusion. In the course of his initial unpacking of the sixteen-term typology that is the end result of the first cūttiram — and which he claims consists solely of the two middle terms of his fourfold set, the eight cuvais and their kuṟippus — Iḷampūraṇar introduces the possible existence of a ninth cuvai, naṭuvunilaimai, along with its respective kuṟippu, thereby disfiguring the Tŏlkāppiyam’s presentation with the possibility of eighteen instead of sixteen elements (p. 2). He goes on to exclude this ninth cuvai candidate from consideration, while still making reference to it in the rest of his presentation, including it innocuously enough in his series of equivalences between the pure Tamil vocabulary of the Tŏlkāppiyam and what we may perhaps presume to be the readily intelligible terms for the Sanskrit rasas, as given in tadbhava forms assimilated to Tamil phonology (ibid): uruttiram ěṉiṉum věkuḷiy ěṉiṉum ǒkkum. naṭuvunilaimai ěṉiṉum mattimam ěṉiṉum cāntam ěṉiṉum ǒkkum, “There is an agreement [in sense] between [the terms] uruttiram (= Skt. rudra, properly raudra) and věkuḷi. There is an agreement [in sense] between [the terms] naṭuvunilaimai (lit. “the state of standing in the middle”, ≈ Skt. mādhyasthyam/tāṭasthyam?), mattimam (= Skt. madhyama) and cāntam (= Skt. śānta)”. Evidently then, naṭuvunilaimai was understood to be identical with śāntarasa or the sentiment of meditative calm that is accepted by some as the ninth member of the class of rasas. This notion is strengthened by Iḷampūraṇar’s comment, a few lines later, that — in contrast to the other cuvais, which each arise due to a specific emotional trigger — “The absence of disturbance owing to any sort [of stimulus] whatsoever is called naṭuvunilaimai” (ibid.: naṭuvunilaimaiy ěṉpatu yāt’ǒṉṟāṉum vikārappaṭāmai). The reason for its excision from the Tŏlkāppiyam rests, he argues, on its otherworldly character: naṭuvunilaimai is not something that pertains to conventional behavior and so — we are led to infer — falls outside the scope of the Tŏlkāppiyam’s authority.9

15But in order to establish this, he proceeds to cite for the first time the Cĕyiṟṟiyam, his most important source throughout the discussion of mĕyppāṭu, and the vector for his introduction of the category of naṭuvunilaimai, and for much else. This was a work evidentially entirely on the drama, which is now only known through these quotations, several of which were subsequently taken up by Perāciriyar in his later discussion of the same passage, and by a handful of citations by other scholiasts.10 The rationale for naṭuvunilaimai’s introduction, and its anomalous status among the other cuvais are both broached by the first of his series of quotations:

mattimam ĕṉpatu mācaṟaṟ tĕriyiṟ
cŏllap paṭṭa ĕllāc cuvaiyŏṭu
pullātākiya pŏliviṟṟ’ ĕṉpatu

Should you wish to clearly know what mattimam is, it is said to be that which abounds in excellence [poliviṟṟu], untouched by all the other aforementioned cuvais.

16As can be seen already from this very brief quotation, the Cĕyiṟṟiyam possessed an explicit theory of cuvai (as opposed to the Tŏlkāppiyam itself), and it distinguished the workings of what it calls mattimam from the general operations of this theory. Iḷampūraṇar proceeds to give two more uncommented quotations in succession from this work, which set out its understanding in more detail, if by no means completely clearly (ibid.):

nayaṉuṭai marapiṉ itaṉpayam yāt’ĕṉiṟ
certtiyorkkuñ cārntupaṭuvorkkum
ŏppa niṟkum nilaiy iṟṟ’ ĕṉpatu

uyppor itaṉai yār ĕṉiṉ mikkatu
payakkun tāpatar cāraṇar camaṇar
kayakk’ aṟu muṉivar aṟivorŏṭu piṟaruṅ
kāmam vĕkuḷi mayakkam nīṅkiya
vāymaiyāḷar vakuttaṉar piṟarum
accuvai ĕṭṭum avarkk’ ila ātaliṉ
* iccuvai ŏrutalai ātaliṉ ataṉai
mĕyttalaip paṭukka itaṉ mikav aṟintore
iccuvai] conj.; accuvai Ed.

If it is asked, “what is the nature of this [mattimam], according to propriety and tradition?” [They] say, “It is that enduring state that can be likened to that of those possessing the right qualities and of those who are favorably inclined”.

Who are the characters [who manifest] this?11 Those who practice great tapas, those who have attained magical power, śramaṇas, sages who cut away ignorance, and others, such as the Buddhists;12 men of truth, who renounce desire, anger, and delusion, the devout [vakuttaṉar] and [still] others. For them, none of these eight cuvais [truly] exist and so, when this cuvai [ex conj] being of a different sort, makes that [other] one appear real, these are [the kind of men who] truly comprehend this.

17Inasmuch as it can be gathered from such a short series of excerpts, we can see that the Cĕyiṟṟiyam is simple and clichéd in style, and is largely given over to enumerative versified lists.13 In this, the text can be compared to the great many enumerative sections that pepper the Tŏlkāppiyam itself, as it can be contrasted with the very different Tamil register — that of discursive, expository, forensic prose — that is on display in Iḷampūraṇar’s own urai.

18And much of the content of the list given here does not surprise: if mattimam or naṭuvunilaimai is the sentiment of beatific calm, then it stands to reason that the figures mentioned here — sages, mendicants, etc. — would be the sort of people connected to its production. It is the final, and most opaque part of this quote that is the most significant. As I understand the passage–and as I have conjecturally emended it — Cĕyiṟṟiyaṉār seems to claim that the regular set of eight cuvais, being grounded in ordinary emotions, do not exist for these kinds of adepts. The ninth cuvai, mattimam, inasmuch as it is qualitatively different from the rest (ǒrutalai, literally, “on one side”14) can only be experienced by these kinds of men, insofar as they alone are able to genuinely comprehend that mattimam manifests (paṭukka) the true nature or reality (měyttal) of any other cuvai.

19I propose that this is based on the Abhinavabhāratī of Abhinavagupta, especially on its celebrated and idiosyncratic interpretation of śāntarasa as the underlying state upon which all experience of art (and indeed, experience as such) is grounded. That mattimam/naṭuvunilaimai is clearly meant to supply a conceptual analog to śāntarasa is not in itself enough to connect the Cĕyiṟṟiyam with AG’s theory, as there are other, and indeed other southern, attestations of the śāntarasa-concept prior to this time.15 The decisive borrowing is rather the claim that certain sorts of adepts are capable of realizing the heightened emotional awareness of cuvai/rasa while remaining dissociated from the content of these experiences.

20Cĕyiṟṟiyaṉār was not entirely successful in either understanding or in translating Abhinava’s theory: anyone who has tried to read the corrupt text of the Abhinavabhāratī here, and to make sense of its argument — or anyone who has surveyed the large and often heated scholarly controversies that its interpretation has provoked16 — may feel some sympathy with his efforts. He was evidently trying to maintain the conventional notion of śāntarasa — that it is possible to successfully depict the spiritual exercises of literary characters in a way that is emotionally satisfying, as in the celebrated case of Harṣa’s play Nāgānanda, or perhaps as in Cāttaṉār’s Maṇimekalai (Richman 1988, Monius 2001) — while incorporating the central innovation of Abhinava’s doctrine of śāntarasa, its radical involution of dramatic theory into a description of the operations of a fully self-aware consciousness.

21In this revision, it is tattvajñāna or awareness of reality (which, in a Pratyabhijñā turn, is said to constitute the self’s non-dual awareness of itself) that acts as the basis upon which all of the content of subjective experience is projected, and it is that which supplies the necessary condition for the realization of śāntarasa. The conclusion to the Cĕyiṟṟiyam passage includes what I understand to be a direct echo of Abhinava’s language, drawn from his summarizing statement of the newly argued centrality of tattvajñāna:17

ratyādayo hi tattatkāraṇāntarodayapralayotpadyamānanirudhyamānavṛttayaḥ kaṃcit kālam āpekṣikatayā sthāyirūpātmabhittisaṃ śrayāḥ santaḥ sthāyina ity ucyante. tattvajñānaṃ tu sakalabhāvāntarabhittisthānīyaṃ sarvasthāyibhyaḥ sthāyitamaṃ sarvā ratyādikāḥ cittavṛttīr vyabhicārībhāvayan nisargata eva siddhasthāyibhāvam iti tan na vacanīyam.

Now, [the eight stable emotions] such as rati, desire, only function or cease to function due to the presence or absence of certain specific causes. Insofar as these thus only exist in a relative sense for a given period of time, they come to be called “stable” [solely on account of their] connection with the backdrop that is the Self, which is itself [actually] stable. By contrast, knowledge of reality, which can be likened to the backdrop for any and all emotional states, is far more stable than any of the [so-called] “stable” emotions; insofar as it transforms all of the mental phenomena that are the stable emotions such as desire into [its own] transitory emotional states, its stability is established as a matter of course. This cannot be denied.

22There are several levels of deep wordplay here, much of it depending on Abhinava’s manipulation of the technical vocabulary of the Nāṭyaśāstra, notably in his use of the term sthāyin. Whereas in the NŚ this has the value of a technical term (as one of the several kinds of bhāva, it is the “thematic” emotion that is the prime prerequisite for the emergence of rasa), Abhinavagupta reactivates its basic adjectival meanings of something “stable” or enduring.

23This revivification of a dead metaphor is used in the service of a further etymological gambit, seen in Abhinava’s culminating flourish of the inchoative participle vyabhicārībhāvayat. Outside of this particular context, the form means something like “destabilizing”; in the present case, at first glance, this means that tattvajñāna renders the other bhāvas less stable (than itself, it being explicitly described as sthāyitama, the “most stable”, relative to them). Set within the technical language of dramaturgy, however, vyabhicārībhāvayat suggests (and is clearly meant to suggest) “transforms ... into [its own] transitory emotional state” or vyabhicāribhāva. Thus the supposedly stable emotions that are the central catalysts of the rasas bear only a temporary, adventitious relationship to the enduring core of self-awareness, in the same way as erotic passion might be adventitiously associated with a moment of anxiety.

24This turn of phrase appears to have caught the eye of Cĕyiṟṟiyaṉār; as his own language could not encode the same grammatical category, he had to resort to periphrasis, resulting in the culminating predicate phrase of the quotation, ataṉai [/] mĕyttalaip paṭukka itaṉ mikav aṟintor. The crucial phrase here, mĕyttalaip paṭukka, reproduces Abhinava’s participle as a verbal noun (mĕyttal “being true”, in the second case) and an imperfective participle or so-called “infinitive” (paṭukka, “to bring about”) based on the effective (or “transitive”/“causative”) stem of the root paṭu.18 It is here where I believe the influence of Abhinava’s text is most clear, as this phrase mĕyttalaip paṭukka relies on an argument through etymological analysis, clearly embedding within it an interpretation of the term mĕyppāṭu itself in a way that parallels Abhinavagupta’s manipulation of the sense of vyabhicāribhāva: as this latter term describes the way in which all of the canonical sthāyibhāvas are relativized when set against the non-dual awareness that is constitutive of śānta, so those figures who manifest mattimam understand how it lends a seeming reality to all of the canonical cuvais, so “making them real” in an implicit etymological analysis of mĕyppāṭu.19 It should be said that Cĕyiṟṟiyaṉār’s presentation seems confused on the critical question of rasāśraya: the various figures detailed here appear certainly to be literary characters, and to claim that they are the locus of mattimam would represent a fundamental misunderstanding of Abhinavagupta’s aesthetics of reception. This is an error, however, that is symptomatic of the wider concerns of the literary theory in the far South, as we shall see.

25Leaving aside this thorny conceptual problem, however, if this passage does provide evidence for the sequence Abhinavagupta→Cĕyiṟṟiyaṉār→ Iḷampūraṇar, there are some significant consequences for the literary history of medieval Tamil commentarial writing. If Iḷampūraṇar is drawing on a work that is itself demonstrably based on Abhinavagupta, then the composition of the earliest surviving commentary on the Tŏlkāppiyam must date to at least two generations after the Kashmirian’s career in the early decades of the 11^th century. This, however, depends on the generous presumption that the transmission, translation, and adaptation of Abhinava’s difficult text occurred as rapidly as is humanly possible: it is more likely that we should consider dating Iḷampūraṇar at some point in the 12^th century, if not later.20 But the correlative also holds true: whatever be the temporal gap that lies between Abhinava and Iḷampūraṇar, we have in the Cĕyiṟṟiyam roughly dateable evidence for the reception and assimilation of the Abhinavabhāratīin the far South.21

26Iḷampūraṇar also evidently derives his emphasis, and even his definition, of cuvai directly from Cĕyiṟṟiyaṉār. This can be seen in comparing Iḷampūraṇar’s own definition with what we can see of the earlier author’s language. The commentator defines cuvai rather schematically and in a way that readily accords with the conventional understanding of rasa familiar from the post-Abhinavagupta consensus in alaṃkāraśāstra (p. 2): iṉi cuvai ěṉpatu kāṇappaṭu pŏruḷāṟ kāṇporakattiṉ varuvat’ or vikāram, “Now, cuvai is the name for the change that occurs in the awareness of the spectators, which arises due to some perceived element”. He then adduces a series of quotations, seemingly from the Cĕyiṟṟiyam,22 as what he terms cārpǒruḷor “supporting material”. The first of these is a cryptic definition of cuvai (iruvakai nilattiṉiyalvatu cuvaiye, “cuvai occurs in two types of locus”); the second, partial quotation expands upon this (niṉṟa cuvaiye [...] ǒṉṟiya nikaḻcci cattuvam eṉpa, “[there is] the existing cuvai...the corresponding occurrence is called the cattuvam”); the third and final quotation fills out the list of ten types of cattuvams, in its enumeration following the NŚ definition against that belonging to the Tŏlkāppiyam. Iḷampūraṇar thus seemingly adduces these to aid in his reduction of the thirty-two element typology to that of sixteen elements: those elements that are apparent to others (piṟarkkum pulaṉāvatu) are excluded in favor of the strictly cognitive phenomena (maṉanikaḻcci). He is thus able to enlist the authority of the Cĕyiṟṟiyam for his own exegetical purposes, as the earlier work was not beholden to the task of maintaining the Tŏlkāppiyam’s taxonomy.

27Expanding on this, however, leads Iḷampūraṇar to an exposition on cuvai that is deeply eccentric from the point of view of the received theory of rasa. He takes up the example (later adapted by Perāciriyar) of a man who sees a tiger or a ghost: the latter supply the motive or cause, while the fear the unfortunate man feels “which endures for the entire time that he sees these” is the cuvai (puliyum peyum cuvaippaṭuppǒruḷ. avaṟṟaik kaṇṭa kālantǒṭṭu nīṅkāta niṉṟa accam cuvai). This example, which seems to entirely collapse the distinction between artistic representation (the domain of bhāva and rasa for the Sanskrit theorists) and real life, seems to place the categorial status of the whole discussion in hazard.

28To make sense of Iḷampūraṇar’s conceptualization of cuvai, and of the relationship between mĕyppāṭu and cuvai, we need to rely on a further quotation from the Cĕyiṟṟiyam, which he cites in his comments on the third cūttiram:

mĕypāṭ’ eṉpatu yātov ĕṉṉiṉ
uyppoṉ cĕytatu kāṇpporkk’ ĕytutal
mĕypāṭ’ ĕṉpa mĕyyuṇarttore
ĕnac cĕyiṟṟiyaṉār otaliṉ, accam uṟṟān māṭṭu nikaḻum accam
avaṉmāṭṭuccattuvattiṉāṟ puṟapaṭṭuk kāṇporkkup pulaṉākun taṉmai
mĕyppāṭ’ ĕṉak kŏḷḷappaṭum.

Should it be asked, “What indeed is this thing called ‘ mĕyppāṭu’?” [one may reply] on the basis of the statement of Cĕyiṟṟiyaṉār, that “Those who understand the truth of the matter say that mĕyppāṭu is the taking up by the spectators of the actions of the leading character [uyppoṉ],” and so when a man experiences fear, and that fear, as represented by his words, is made manifest through his [further?] words and physical reactions, and is thereby made visible to the spectators, the nature of this is what we should understand by mĕyppāṭu.

29The whole language of visuality that Iḷampūraṇar freely adopts in his own definition of cuvai appears to be another conceptual importation from Sanskrit dramaturgical sources: kāṇpor in both definitions seems certainly calqued upon prekṣakāḥ, as seen in the Nāṭyaśāstra.23 This is distant from the language of the Tŏlkāppiyam, which only details the keṭpor or listeners who exist within the narrated world of the Caṅkam poems. Iḷampūraṇar then turns to the rival somatic interpretation of mĕyppāṭu, and while the rasa-congruent model of the Cĕyiṟṟiyam is still influential in his thinking, here he relies on the testimony of the Tŏlkāppiyam itself (p. 4, with vv. ll.):

mĕyyiṉkaṇ tǒṉṟutaliṉ mĕypāṭ’ āyiṟṟu aḥtel, ivvilakkaṇam kūttiṉuṭ
payaṉpaṭal uṇṭātaliṉ īṇṭu veṇṭuv ĕṉiṉ īṇṭun cĕyyuṭceyyuṅkāṟ
cuvaipaṭac cĕyya veṇṭutaliṉ īṇṭuṅ kūṟa veṇṭum eṉka
uytt’ uṇarv’ iṉṟit talaivaru pǒruḷiṉ24
měyppaṭa muṭippatu25 měyppāṭ’ ākum
ěṉa ivvāciriyar měyppāṭuñ cěyyuḷuṟupp’eṉa otiṉamai26 uṇarka

It might [be objected that] “it has been [called] mĕyppāṭu because it occurs in the body”; since this definition is applicable in the case of dramatic performance, should it be accepted here [i.e. when we are concerned with non-dramatic genres]? [In response to this] one should reply that in this case too it ought to be accepted, since when we are formulating rules of poetic composition it is accepted that [mĕyppāṭu] is something that can be savored: bear in mind that this author [i.e. Tǒlkāppiyaṉār] declares mĕyppāṭu also to be an element of verse composition, [when he teaches in Cěyyuḷiyal, cū. 192:].
“Something that is represented which, without any conscious reflection, succeeds in becoming real through [the depiction of] its subject matter, becomes [known as] mĕyppāṭu”.

30As we can see here, Iḷampūraṇar is aware of a somatic interpretation of mĕyppāṭu, one which is in its wording practically identical to that advanced by Pěruntevaṉār.27 Iḷampūraṇar is thus apparently willing to grant some validity to this understanding of the word. Nevertheless, the tenor of his comments here, his reliance throughout upon the Cĕyiṟṟiyam, and finally his invocation of the testimony of the Tŏlkāppiyam itself (containing yet another implicit etymology of mĕyppāṭu in its phrase měyppaṭa muṭippatu) all point to a different etymological understanding of the word. On this view, instead of “occurring” (pāṭu, as derived from paṭutal) in “the body” (měy), the word denotes the “making” (pāṭu as derived from paṭuttal) “real” (měy). Just as this same etymological understanding enabled Cĕyiṟṟiyaṉār to venture his equation between mĕyppāṭu and (vyabhicāri-) bhāva, so this understanding of mĕyppāṭu as that which “makes real” is reminiscent of the Nāṭyaśāstra’s own nirukti-etymologies of bhāva.28 Indeed, there are strong grounds to believe that this sense of mĕyppāṭu — and not the somatic understanding of the word — was that intended by the authorcompilers of the Tŏlkāppiyam. Beside the quotation from the Cěyyuḷiyal adduced by Iḷampūraṇar, there is both internal and external evidence to suggest this. The MI in fact contains two lists of mĕyppāṭus, the eight listed in cūttiram 3 (evidently equivalent to the sthāyibhāvas) and another list of thirty-two (cūttiram 12 (= PA 256), almost exactly equivalent to the thirtythree (N.B.) vyabhicāribhāvas). Similarly, the anonymous Taṇṭiyalaṅkāram (likely composed in the mid-12^th century, and in some senses a translation of Daṇḍin’s Kāvyādarśa) in its definition of the figure of speech called cuvai (based on Daṇḍin’s rasavat, KĀ2: 281–292) is completely explicit on the functional identity between bhāva and mĕyppāṭu.29

31If the Tŏlkāppiyam’s term mĕyppāṭu is a direct calque of the Nāṭyaśāstra’s bhāva, this lends some perspective on the long-term history of the translatability of technical terms from Sanskrit into Tamil, a perspective that enables us to see the Cĕyiṟṟiyam’s and Iḷampūraṇar’s own habits of translation and equivalence in a better way. Both medieval texts largely avoid an obviously Sanskrit-derived lexis (that is, both tend to avoid tatsama and tadbhava words). But if in fact the Tŏlkāppiyam’s theory of literary emotions is based solely on a conception of bhāva rather than rasa, there are a number of interesting consequences that follow. This might suggest that the bhāva descriptions found in the seventh chapter of the received text of the NŚ were the only materials available to the Tŏlkāppiyam’s compilers or their Tamil sources, or that the absence of an explicit theoretical counterpart to rasa points towards another, earlier system of genres, where certain works (ex hypothesi Tamil poetic works among them) were thought competent to produce bhāva, but not rasa (the latter perhaps being the domain of the theatre alone). These are only speculations; it is worth registering that the model of mĕyppāṭu at work here partly vindicates Cutler’s observation, mentioned earlier, that the poetics of early Tamil are largely concerned with text-internal phenomena, rather than with the text as something that provokes a reaction in its audience. But if anything, this reflects the earlier internal or formalist preoccupations with alaṃkāraśāstra in Sanskrit as well, and not some special theoretical agenda uniquely (or perhaps deficiently) present in Tamil poetics.30

32That the Cĕyiṟṟiyam is committed to a position of the reception-centered aesthetics of the Kashmirian rasa-theorists is obvious enough; that it evinces direct contact with the verbal matter of the Abhinavabhāratī is more debatable but, I believe, likely. What is then surprising is the way in which Cĕyiṟṟiyaṉār appears to fumble his presentation of Abhinava’s reception-based theory in the case of mattimam and, in turn, how Iḷampūraṇar proceeds to apply this only partly receptive understanding of cuvai to his own explanation of the Tŏlkāppiyam. While importing the model of cuvai and the awareness of the kāṇpor-spectator as its locus wholesale into his comments, Iḷampūraṇar goes on to interpret the relationship between mĕyppāṭu and cuvai in a way that seems profoundly eccentric from the post-Abhinavaguptan understanding of rasa, as he explains that each of the mĕyppāṭu enumerated in the text arise (piṟappatu) from that which he had earlier declared to be equivalent to the rasas. So aḻukai, weeping, arises in avalam, the pathetic (aḻukai ěṉpatu avalattiṟ piṟappatu; avalam had earlier been equated with karuṇai), pěrumitam, boldness arises in vīram, and uvakai, desire, arises in ciṅkāram (= śṛṅgāra).31 This appears similar to a question raised and summarily dismissed in the Nāṭyaśāstra itself, as to whether bhāvas may be said to arise from rasas, or whether they may be said to be mutually constitutive. Abhinavagupta defends the NŚ’s unidirectional bhāva-leads-to-rasa doctrine, but seems somewhat sheepishly attracted to the idea of parasparasambandha (the bhāvas only acquire their significative ability in the wider context of a dramatic performance, and so derive from their innate connection with rasa).32 But I suspect here that Iḷampūraṇar is not entering polemically into a debate found in his sources (or his sources’ sources), but that he is dealing with the consequences of his attempt to retrofit the new, affective theory of literary communication to his mūla, itself a part of an earlier dispensation.

33All the explicit testimony to be found in the Pŏruḷatikāram, whether here or in the later Cěyyuḷiyal, suggests that mĕyppāṭu was only ever meant to account for the represented display of emotions within the literary text itself. The Tŏlkāppiyam thus provided no account of literary reception because it was not an issue for it: it was enough that emotion could be vividly depicted in Tamil poetry, and that these representations could be typologically recognized by the educated reader. As a legatee of the new, post-11^th century model of literary reception, Iḷampūraṇar attempted to reconcile the perhaps newly expanded critical vocabulary of Tamil poetics. He did so centrally by incorporating cuvai itself with the underdetermined theoretical matrix of the Pŏruḷatikāram. The resulting synthesis is awkward — most notably in its potential collapse of the distinction between art and life — in order to describe within the Tŏlkāppiyam’s available terms how it is that the feelings of others (whether actual or represented) excite feelings of our own. But it would be a mistake to understand Iḷampūraṇar’s project as a failure: instead, it attempted to mediate between the existing terms of Tamil criticism and the new conceptual innovations from the North, using mĕyppāṭu as a mediating hinge between the two different poetics. It is the very awkwardness of Iḷampūraṇar’s effort, however, that makes his testimony especially valuable: in the hands of the later Perāciriyar, this is neatly harmonized into a theory with no intellectual rough edges, a suave presentation that effectively consigned mĕyppāṭu to irrelevance in later Tamil poetic theory.

2. Śāradātanaya’s confected authorities

34As a southern author working to integrate the new post-Abhinav-aguptan theoretical model, Iḷampūraṇar was not alone. Śāradātanaya’s Bhāvaprakāśana (BhP, “On the Display of Emotion”) evinces strong points of comparison, and equally striking differences, with the Tamil grammarian-critic. The BhP is a lengthy Sanskrit verse text on literary and dramatic theory that was composed probably sometime over the course of the 13^th century in Uttaramerūr, the famous brahmadeya in what is now Kanchipuram district. Śāradātanaya tells his readers that he was attached to a theatrical troupe in a local Sarasvatītemple as a sort of staffpandit. Under the tutelage of this company’s master, one Divākara, he claims to have been instructed in all the different theories said to make up the nāṭyaveda, including those associated with such mythical figures as the sage Agastya, Hanūmān, and Sarasvatīherself (BhP, p. 2.14–1933). The BhP, he tells his readers, is a digest of all of these. And in fact the BhP is rife with quotations and recastings both acknowledged and unacknowledged, beginning with the Nāṭyaśāstra and extending up to Mammaṭa’s Kāvyaprakāśa. Because of these admitted borrowings, Śāradātanaya has often been understood as merely a legatee of other, more original thinkers, a description that obscures much of interest in the text. Much of what is original to the work, as I will show, can be seen at the level of compositional technique instead of argument.

35In light of the fraught concept-history of mĕyppāṭu that we have seen in the Tamil sources, the BhP’s explicit focus on bhāva is noteworthy, as signalled already in its title. This contrasts markedly with the way in which other (especially later) works interested in questions of dramatic or literary theory tend to topicalize rasa in their titles — for instance, the Rasārṇavasudhākara, the Rasakalikā, or the Rasagaṅgādhara. This is something further borne out in the structure of Śāradātanaya’s presentation: he begins his text with a lengthy and largely original etymological analysis of the bhāvas, and he develops his explanation of the ways that artistic language functions through constant reference to these elements rather than their endpoint. Though his analysis is ultimately directed towards the culmination of rasa, Śāradātanaya tends to bypass the questions of aesthetic psychology that had come to dominate large sectors of Sanskrit alaṃkāra writing to focus on the creation of dramatic works, both as verbal art and as performance, circumscribing these within the several typologies of bhāva.

36Śāradātanaya’s text, from its title to core features of its structure, thus suggests that this focus on bhāva as the central element of literary affect was neither limited to works in Tamil, nor was it purely the artifact of an earlier and superseded theoretical dispensation. Like the Cĕyiṟṟiyam and Iḷampūraṇar’s urai, the BhP sought to salvage this emphasis while rationalizing it in light of the innovations in literary theory emanating from Kashmir. This suggests that medieval Tamil authors like Iḷampūraṇar and Śāradātanaya might be best understood as operating within a common horizon of problematics and shared sources, and that the best way to further our understanding of the Tamil and Sanskrit theories of literature would be to read them side by side.

37The BhP, however, introduces a useful complication into this argument. Much of it is an admitted compilation of earlier texts, drawing especially but by no means exclusively on works from Kashmir and from the literary salon of the Paramāra court at Dhārā (especially the Daśarūpaka and Bhoja’s Sṛṅgāraprakāśa). Many of these have been usefully tabulated by Ramaswami Sastri in his edition of the work.34 But there is more than meets the eye here: many of the quotations are unacknowledged, or have been subtly rewritten by Śāradātanaya to serve his own particular interests, or result from blending two different unrelated sources together into a new idea.35

38Distortive quotation or quotation from memory is not unusual in Sanskrit; in Śāradātanaya’s hands, however, this sort of recasting rises to the level of a major compositional principle. Sometimes this is to the detriment of the BhP’s coherence as a work of śāstra: Śāradātanaya tends to pile up discordant statements from multiple sources without bothering to say how they relate to each other, or to offer what he thinks is the final understanding. But this is not the result, I would argue, of authorial inattention, or of argumentative lapses. Śāradātanaya’s recastings are instead ultimately directed towards his effort at reconciling the received priorities of southern poetics with the wide spectrum of avant-garde theory finding its place in the Tamil country.

39Perhaps the clearest case of Śāradātanaya’s self-consciously synthetic project can be seen in his unusual habit of pseudepigraphical quotation, where he attributes a text or a concept to someone other than its acknowledged author. I will provide three examples. Śāradātanaya informs his readers on two occasions early in the BhP’s sixth adhikāra, which largely consists of an extended survey of different theories of sentence-meaning, that it is composed kalpavallyanusārataḥ, following the Kalpavallī, and he returns late in the same chapter to this supposed source, though there referring to it with the synonymous name Kalpalatā (pp. 131.4; 142.6; 175.18). The chapter is, however, largely a palimpsest of quotations from Bhoja, Bhartṛhari, Dhanañjaya, and others with some light overwriting, and consists in the main of a précis of the first four chapters of Mammaṭa’s Kāvyaprakāśa. While there is an alaṃkāra text called the Kalpalatā attributed to the 12^th century author Ambaprasāda (now only extant in the pratīkas found in its surviving commentary), and this text does seem to contain a section on sentence-meaning oddly folded into its discussion of śabdālaṃkāras,36 it seems unlikely that this is Śāradātanaya’s source here, given that the authorities he reproduces or recasts are works explicitly cited elsewhere in the BhP, and which he actually cites throughout the body of the chapter.37 His one explicit reference to the Kalpavallī/Kalpalatāwould seem in fact to imply that he is not referring to Ambaprasāda’s (in any case deservedly obscure) work: “The set of four verbal meanings, beginning with the directly denoted, and the [corresponding] four kinds of expressive language, beginning with the directly-denotative, have been authoritatively explained in this way in the Kalpalatā, and have been illustrated in the Kāvyaprakāśa and by myself in the current work”.38 Evidently, Śāradātanaya wishes his readers to believe that he is in possession of a work that is the source of Mammaṭa’s celebrated and much commented upon textbook. The only problem is that there is no warrant to believe such a work ever existed outside of Śāradātanaya’s own citations.

40This could be explained away as an anomaly, were it not for the evidence of the BhP’s second adhikāra. This chapter provides the central statement of Śāradātanaya’s understanding of the emergence of rasa from (sthāyi-) bhāva. Towards the chapter’s end, after introducing all the terminological machinery familiar from the earlier dramatic tradition, Śāradātanaya describes the emergence of rasa from the bhāvas as the outcome of a set of cosmological and psychological processes deriving ultimately from Sāṅkhya, in which rasa is generated out of an interaction between the workings of individual’s sense of self and the external world. This basic, so to say ontological, condition is only made available to the human subject in the presence of the appropriate elements of stagecraft (performance, gesture, costume, etc.). This, Śāradātanaya informs his readers, is Śiva’s teaching to Vivasvat in a text called the Yogamālāsaṃhitā, along with particular details of several modes of dance.39 But the theory on display here clearly closely resembles that of Bhoja’s Śṛṅgāraprakāśa, and this again is a work elsewhere openly and repeatedly cited by Śāradātanaya. This uncredited reworking was earlier noticed by Raghavan (1978: 485–86) who, with evident irritation, denied that such a work as the Yogamālā could have existed. He was equally dismissive of a second, closely allied explanation of rasa-etiology that Śāradātanaya offers next (pp. 47.11–48.6), this time said to be the teaching of Vāsuki the serpent to the sage Nārada, and differing only in that Vāsuki accepts the possibility of śāntarasa. Here once again it seems that Śāradātanaya claims a textual warrant for his theory of rasa that is in all likelihood an invention.40

41The third and final example of Śāradātanaya’s confection of textual authorities is by far the most striking, in the way that he approaches the disordered textual bulk of the Nāṭyaśāstra itself. Early in the second adhikāra, Śāradātanaya divides the text into several authorial voices, running directly counter to Abhinava’s assertion of the unitary authorship of the authentic core of the text.41 He does so at a critical point in his argument, when turning from one of the several etymological excursuses that pepper the text,42 he gives his major statement of the thematic emotion’s transformation into rasa (pp. 36.7–37.2, with my added sigla, a–g):

vibhāvādyair yathāsthānapraviṣṭaiḥ sthāyinaḥ smṛtāḥ|
caturbhiś cāpy abhinayaiḥ prapadyante rasātmatām || (a)
vibhāvaiś cānubhāvaiś ca sāttvikair vyabhicāribhiḥ|
ānīyamānaḥ svādutvaṃ sthāyī bhāvo rasaḥ smṛtaḥ || (b)
vyañjanauṣadhisaṃyogo yathānnaṃ svādutāṃ nayet |
evaṃ nayanti rasatām itare sthāyinaṃś ritāḥ || (c)
evaṃhi nāṭyavede ‘smin bharatenocyate rasaḥ |
tathā bharatavṛddhena kathitaṃ gadyam īdṛśam || (d)
yathānānāprakārair vyañjanauṣadhaiḥpākaviśeṣaiś ca saṃskṛtāni vyañjanāni madhurādirasānām anyatamenātmanāpariṇamanti tadbhoktṝnāṃmanobhis tādṛśātmatayāsvādyante tathānānāprakārair vibhāvādibhāvair abhinayaiḥsaha yathārham abhivardhitāḥ sthāyino bhāvāḥ sāmājikānāṃ manasi rasātmanā pariṇamantas tādātvikamanovṛttibhedabhinnāḥ tattadrūpeṇa tai rasyante. (e)

nānādravyauṣadhaiḥ pākair vyañjanaṃ bhāvyate yathā |
evaṃ bhāvā bhāvayanti rasān abhinayaiḥ saha || (f)
iti vāsukināpy ukto bhāvebhyo rasasambhavaḥ |
tasmād rasās tu bhāvebhyo niṣpadyante yathārhataḥ || (g)

The thematic emotions already described attain the status of rasa through the triggers [and other components] when these are properly deployed along with the four types of performance techniques.
The thematic emotion is considered a rasa when it is made delightful [svādu°] through the triggers, the consequents, the bodily manifestations, and the incidental emotions.
Just as the combination of curries and herbs will make rice delicious [svādu°], so these others, [properly] arranged, transform the thematic emotion into rasa.
This is the way that Bharata has explained rasa in the Nāṭyaveda; so too this has been taught, in prose, by the elder Bharata:
Just as curries, prepared with cooking herbs of various kinds and out of [other] ingredients, transform into one of the [six] flavors, beginning with sweet, and are enjoyed by those who eat them, thinking them to have that flavor, so too the thematic emotions, as they are appositely augmented by the various kinds of triggers and other bhāvas and accompanied by performance techniques, transform into rasa in the mind of the spectators and, as they are differentiated by the workings of the spectators’ minds at that particular moment, are savored by them in whichever form [that they have thus assumed].
As a curry is produced through different methods of cooking with various herbs and [other] ingredients, so bhāvas, together with the techniques of performance, produces the rasas.
Thus it was taught by Vāsuki, that the rasas are born from the bhāvas; thus, the rasas are produced from the bhāvas, as according to their capacity.

42From the perspective of propositional content, this is mostly unambitious stuff, repetitively told: the NŚ position that the rasas proceed from the bhāvas, and not the other way around, is repeatedly emphasized, and the general sense of the passage seems merely to repeat the axiomatic doctrine of dramatic rasa arising due to the presence of the various contributory factors. But just beneath the surface of Śāradātanaya’s presentation, there are multiple levels of textual legerdemain at work. Three of the verses (a, d, g) are entirely from Śāradātanaya’s own hand; these are essentially linking passages, introducing topics, naming authorities, and supplying conclusions. One (b) is a direct borrowing of the Daśarūpaka’s leading statement about the nature of rasa, with a single albeit significant variant,43 while the remaining two verses c and f and the prose passage e are based wholly or in part on material gleaned from the NŚ. The two verses are close adaptations of Bharata’s defense of the unidirectional bhāva-to-rasa model: f is almost identical to NŚ 6.35 (which reads nānādravyair bahuvidhair in its first quarter44), while c’s first half is identical to 6.37ab.45 Both of these NŚ verses are given there as the preexisting views of another authority, introduced by the prose tag bhavanti cātra ślokāḥ (“there are some verses on this matter”).46 Yet Śāradātanaya explicitly assigns them to two different authors, with the earlier of the two — half of which consists of Śāradātanaya’s own words — given as Bharata’s siddhānta, and the other (almost exactly as transmitted in the NŚ) given, once again, as the teaching of the mythical serpent Vāsuki.

43Set within this complex web of borrowings and recastings, however, the gadya passage (one of the very few pieces of prose found in the BhP) is particularly striking. Though pointedly attributed to “the elder Bharata”, I find no reason not to think these words are Śāradātanaya’s own, based on the model of the celebrated prose dṛṣṭānta found in NŚ6, immediately before the verses that provide the raw material for c and f. This reads (with common or approximately shared phrases in bold):

ko dṛṣṭāntaḥ. atrāha yathā hi nānāvyañjanauṣadhidravyasaṃyog ād rasaniṣpattiḥ tathā nānābhāvopagamād rasaniṣpattiḥ. yathā hi guḍādibhir dravyavyañjanair auṣadhibhiś ca ṣāḍavādayo rasāḥ nirvartyante tathā nānābhāvopetā api sthāyino bhāvā rasatām āpnuvanti. atrāha rasa iti kaḥ padārthaḥ. ucyate. āsvādyatvāt. katham āsvādyaterasaḥ. yathāhinānāvyañjanasaṃskṛtamannaṃbhuñjānā rasān āsvādayanti sumanasaḥ puruṣā harṣādīṃś cādhigacchanti tathā nānābhāvābhinayavyañjitān vāgaṅgasattvopetān sthāyibhāvān āsvādayanti sumanasaḥ prekṣakā harṣādīṃś cādhigacchanti. tasmān nāṭyarasā ity abhivyākhyātāḥ.47

What could serve as an example of this? He replies: just as flavor arises due to the combination of substances such as various kinds of curries and herbs, so rasa arises due to the coming together of various kinds of bhāvas. Further: the flavors of, for instance, a sweet are developed by ingredients and condiments like treacle as well as by herbs; and so it is that the thematic emotions, replete with the various bhāvas, attain the status of rasa. What sort of thing is rasa? He answers: [It is the way that it is] because it can be savored. How is rasa savored? Just as thoughtful men, when they eat rice that has been prepared with various curries, savor the flavors and so feel happiness and other [pleasant sensations], so thoughtful spectators savor the thematic emotions as they are manifested by various bhāvas and performance techniques and accompanied by reactions in the voice and the body [of the performer] and feel happiness and other [pleasant sensations]. Thus the rasas of the theater are exhaustively described.

44The relationship between these two passages is thin, but telling. The words attributed to Bharatavṛddha are an original departure from the NŚ dṛṣṭānta, not a variation upon it. Yet we can see clearly enough where Śāradātanaya’s words (for they are surely his) derive from their model, consciously reworking a locus classicus to produce something similar, yet distinct enough to claim an independent authority. Śāradātanaya imputes these views to a long-lost precursor or expanded recension of the NŚ. It may fairly be asked: why? Why bother with this elaborate confection of works like this, or like the Yogamālāsaṃhitāor the Kalpavallī? Was he acting with the deliberate intent to deceive his readers, to foist his own views on the etiology of rasa upon them?

45To write off Śāradātanaya as simply a dishonest scholar or a forger would, I think, miss the point.48 Within the context of the BhP itself, these inventions possess an integral logic: Śāradātanaya invents when he wants to rationalize his inherited models, both within his own literary and theoretical ecology and in light of the disparate materials he is bringing together. When, in his Bharatavṛddha passage, he substitutes his source’s sumanasaḥ prekṣakāḥ with the practically identical but lexically distinct sāmājikānāṃ manasi, he is drawing into higher relief the emphases of the new reception-oriented scholarship while offering it a would-be ancient pedigree.49 But besides smoothing over possible theoretical anachronisms, Śāradātanaya’s inventions also provide scope for his own innovations: in the same Bharatavṛddha passage, his introduction of the idea that the rasa experience varies depending upon the mental state of the spectator at the moment of reception (i.e. that the rasas are experienced tādātvikamanovṛttibhedabhinnāḥ) appears to be his own contribution to the theory. It is this problem — Śāradātanaya’s awareness of it presents a noteworthy contrast to the Kashmirian theorists, for whom all sahṛdayas experience the same thing50 — that occasions his Sāṅkhya-based digressions on the difference of the audience’s subjective constitution later in the same chapter which he attributes to the spurious Yogamālā and the teachings of Vāsuki. The assignment of the doctrines of sentence-meaning to the Kalpavallī seems to share a similar logic: bringing together so many disparate sources — and rewriting many of them in the process of composition — may have motivated Śāradātanaya to posit a single locus of attribution for the resulting work of bricolage. Throughout, he manages to maintain the southern preoccupation with bhāva that evidently exerted such an influence on the Tŏlkāppiyam tradition, while introducing the concerns with the hermeneutics of reception that are characteristic of the rasa theory synthesized by Abhinavagupta. He is in fact able to advance these in the unanticipated direction of the psychophysical mechanisms that he proposes as underlying the aesthetic capacities of a reader or viewer, while retaining his own focus on text-making as the central concern of the BhP.

3. Map and territory

46For all the many differences of language, style, and argument that we can see in Śāradātanaya and Iḷampūraṇar, their texts operated within a unified textual and intellectual horizon, and did so with an intelligible and shared rhetoric. Neither wished to be seen as explicitly innovating upon a pre-given theoretical matrix, but both were participants in a textual order that was flush with new ideas, new texts, and new articulations of old ideas. Faced with this, they adopted different but really complementary strategies: Iḷampūraṇar, invoking a belief in the plenitude of his root-text that we might call “classicism”, held that new ideas could be harmoniously integrated into its understanding; Śāradātanaya, working without a mūlagrantha, allowed himself the possibility of a proliferating canon of authoritative sources, both real and invented. Further, both authors were faced with the shared problem of integrating the new theoretical dispensation emanating from Kashmir with their inherited model of literary and dramatic scholarship. In this earlier model — which was not peculiar to the Tamil country, but which possessed an ancient and prestigious authority in the Tŏlkāppiyam — the overriding theoretical concern was with the text-internal effects of literary and dramatic art, with how beautiful literary representation could be successfully achieved. Seen from this perspective, bhāva, the represented tenor of literary affect, had primacy: its culmination in rasa was a question that the earlier theory could essentially take for granted, to the extent that the Tŏlkāppiyam compilers could ignore it altogether.

47In the wake of the dissemination of the new northern poetics, with its focus on the receptive dimensions of art, this older model must have appeared insufficient. Yet its terms and its priorities, especially its focus on text-making as opposed to textual criticism, continued to compel the adherence of southerners like Iḷampūraṇar and Śāradātanaya. This seems to have been the predicament which led to the apparent oddities of their presentations from the perspective of our received understanding of the history of alaṃkāraśāstra. In navigating between the necessity of unpacking his root-text and the evident attraction of the Cĕyiṟṟiyam’s Abhinavaguptan gloss, Iḷampūraṇar effaces the categorical distinction between art and life, while binding cuvai and mĕyppāṭu together into a mutually constituting complex. The difficulties he thereby creates for himself should not be overstated, but the clear demarcating lines that the Sanskrit theorists had limned since Bharata’s time seem to become blurred in the process. Śāradātanaya broaches a different question: working firmly from the text-centric viewpoint, and so understanding the workings of the dramatic text as an engine of affect to be the primary point of interest, when he turns his attention to the effect of the work on the audience, Śāradātanaya admits the possibility that their individual reaction to the common textual stimulus might differ. The question of this heterogeneity of emotional response is something that Sanskrit literary theory never asks elsewhere, and it leads Śāradātanaya to think about (and in turn to invent earlier thinkers about) the metaphysical preconditions of aesthetic response, the ways in which our reaction to art might be hard-wired into our selves. The two variations on the normative theory, while thus differing in their details, evidently share a common logic. It seems not to be the case, as Selby earlier argued, that for southern readers the emotional response to poetry did not require “mental reflection or intellectual articulation”; for these two authors at least, reflection on literary art is constantly overflowing its disciplinary boundaries, and spilling over into more fundamental questions of subjectivity and ontology.51

48It bears emphasis that these variations or innovations upon received literary theory were only possible, and can only be interpreted, from within the patterns of inheritance, indebtedness, and resistance that the two authors stake out with their textual precursors. To better see this, the relationship between Iḷampūraṇar and Śāradātanaya’s works and their sources may be set out tabularly (see Fig. 1). It is a tangled picture, even if we were to exclude the various fictive authorities on which the BhP allegedly depends. This stemma-like table of relationships is to some degree interesting in itself, allowing us to discern patterns of potential significance, and prompting possible new avenues for research.52 What it emphatically suggests is the bibliographical complexity at play among medieval Tamilnadu-based literary scholarship, whether composed in Tamil or Sanskrit. Lines tend to blur between languages, as indeed between disciplines, and everyone seems to have read (and to have cribbed from) everything. And this is a very incomplete picture: with the meager few quotes we have from the Cĕyiṟṟiyam we cannot even begin to chart out the horizons of its author’s reading.

49But there are other very prominent features that this schematic representation elides: for example, the differing literary and forensic strategies seen in the Cĕyiṟṟiyam, in the BhP, and in Iḷampūraṇar’s urai; the motivation for undertaking such projects of translation, comparison, and recombinant invention; and — a question all but unexplored here — the effect that these works exerted on subsequent scholarship and on the creation of literary works. The effort to gain some sort of analytical purchase on these and related questions suggests that we consider the difference between the map (as laid out in the diagram) and the territory, both bibliographic and conceptual, that was shared by thinkers interested in the theory of literature in this time and place.

50For thinking through this, I have found some of Quentin Skinner’s writings on the method of intellectual history to be helpful. Very schematically, Skinner repurposes select parts of Austin’s speech-act theory for the intellectual historian, in an effort to recover the illocutionary force (to use the speech-act jargon) underlying the complex propositions of works of systematic thought. This commits Skinner to two linked presumptions:53 a strong argument for the coherence and legibility of authorial intentions, and an equally strong notion of the informing context necessary for the realizations of such an intentional project. The historian needs to understand, Skinner compellingly argues, the tacit as well as explicit sets of rules, and of inherited and contemporaneous habits of language use that regiment the uptake of such complex speech-acts; only in doing so can the historian recover the conditions of possibility and of intelligibility of such an intervention. This in turn necessitates a notion of context that is as capacious as possible; presuming that the complex speech-acts of systematic thought “can never be viewed simply as a string of propositions; they must be viewed at the same time as arguments” (2002, I: 115), it becomes incumbent upon the historian to reconstruct precisely the set of prior and contemporaneous other speech-acts that make these arguments meaningful. Inherent in this is an even more radical possibility: when he argues that in order to “recover that context in any particular case, we need to engage in extremely wide-ranging as well as detailed historical research”, Skinner is led to “challenge the categorical distinction between texts and contexts” itself (2002, I: 116, 117). This Skinner contrasts with a rival approach, one that understands the task of the intellectual historian to consist of reading “over and over” the arguments of the classic texts in some given field, in order to capture the nuance of the particular author’s contribution to a lading-list of long-standing questions and problems.54 In Skinner’s own historical (as opposed to theoretical) writing, one can see how this might be put into practice: how, for example, Machiavelli’s elevation of virtù to the central quality of rulership relies upon his subversion of this category as employed by earlier republican authors, or how freshly circulating translations of Cicero and Livy inflected the terms of debate in the parliamentary maneuvering prior to the outbreak of the English Civil War.55

51Now, I am not the first person to engage with Skinner’s ideas within Indological scholarship: most recently and significantly, Jonardan Ganeri (2008) has critically questioned the applicability of Skinner’s method to premodern Indian intellectual culture. While I disagree with much of Ganeri’s argument, both in terms of particulars and in principle, his thoughtful pūrvapakṣa has much to recommend it, presenting a very welcome effort to think in a genuinely methodological way about the practice of premodern Indian intellectual history. Ganeri’s central claim is that “Skinner’s conception of ‘context’ is both too rich and too poor” (2008: 557) to be of use in Indic materials: “too rich” because (and he admits this point to be “rather banal” [2008: 553]) we Indianists lack access to the sort of fine-grained level of details available to Skinner for the biographies of early modern European authors, and for the histories of their societies; and “too poor” because it is the long-term trajectory of a particular śāstra (or set of śāstras) that forms the adequate scope for our studies, and that these “create broader contexts of intellectual intervention than Skinner considers” (2008: 557).

52As for the first of Ganeri’s objections, for the medieval South Indian materials which I have reviewed here, there is a strictly local counter-objection that can be offered. There is in this case considerably more historical reality which we can hope to latch onto, thanks to the great successes of the social history of the medieval Tamil country by the likes of Karashima (2001, 2009), Subbarayalu (2001), Heitzman (1997), Veluthat (2009), and Orr (forthcoming). From this work, we possess a gathering sense of the massive transformations underway in precisely the period of the long twelfth century. This period witnessed the rise of a locally modulated notion of private property, for example, and widespread, fundamental changes in the landed order, in political culture, and in forms of collective life in the wake of the winding down of imperial Coḻa hegemony. It needs to be said that this scholarship is completely indifferent to intellectual history, but there is no reason why this lack of application needs to stay the case forever.56

53Ganeri’s second objection — that the Indic materials operated according to a fundamentally different temporality and concomitantly with a radically different notion of philosophical problematic — is much more interesting, but I think is also misguided. When Ganeri suggests that Indic texts are concerned with intervening within “a literary/intellectual rather than a physical/socio-political context” (2008: 553), he seems to confuse some of the preoccupations of Skinner’s sources — he is after all a historian of political thought — with the larger aims and interests they evince within their own fields. Beyond this minor point, Ganeri’s criticism appears based on either a misunderstanding of Skinner’s ideas or an unacknowledged adoption of them. When Skinner explicitly argues that the complex communications that make up works of systematic thought amount to position-takings “in relation to some pre-existing conversation or argument” (2002, I: 115), and when he proceeds in his substantive writing to demonstrate the time-depth of these arguments and of the problematics they embody, he sounds very much like he is arguing the position that Ganeri claims to be an India-specific rebuttal to his method. This general flexibility attributed to what potentially counts as contextually relevant, as well as the particulars of the case I am interested in here, suggest that we need to look for these pre-existing contextual frames of relevance both in long-term disciplinary problems and in the conjunctural cases of local intellectual cultures, and that indeed the two cannot meaningfully be separated.

54In sum, I am willing to take the programmatic side of Skinner’s method on board, and I am not convinced by Ganeri’s attempt to offer a particularist critique of this method. Nevertheless, I willingly concede there to be local dimensions and dynamics for which we need to account in order to really get our money’s worth out of Skinner. But I understand this to be an extension, rather than a fundamental revision, of the priorities of his method. Skinner claims that, once we adopt his historicist view of the way works of systematic argument are made to cohere, it becomes a matter of “a combination of inference and scholarship” (2002, I: 143) to reconstruct the intentional project underlying a particular proposition: that, say, bhāva in certain dramatic or literary contexts gives rise to rasa. He is, however, unclear as to what form this scholarship should take. Skinner’s own method tends towards the strictly semantic: he seeks to account for the sphere of consensually assented-to meanings of a term or set of terms. This allows him, in a Nietzschean or Weberian vein, to explain conceptual change as a matter of rhetorical innovation: the way that a particular author or group of authors repurpose a given lexical item is what enables conceptual change. And this is all well and good.

55In the materials that interest me, however — in Iḷampūraṇar’s reliance on the Cěyiṟṟiyam, or in the spectrum of acknowledged, unacknowledged, and invented source-texts brought together by Śāradātanaya — we need to look beyond the semantic realm of denotation and focus on the way in which these complex works of language work: that is to say, it is philology that provides the necessary condition for this sort of intellectual history; and it is through philological scholarship that we can hope to frame as well as answer the sort of questions that Skinner urges. The very term mĕyppāṭu — where the reference of the lexeme is itself unstable and can only be determined through philological argument — presents a limit case of this. But beyond the sort of historicaletymological problems that the mĕyppāṭu-cuvai complex presents, it is further necessary to attend to the generic or textural features of texts like Śāradātanaya’s and Iḷampūraṇar’s, as revelatory both of their authors’ projects and of the wider textual and argumentative context. And it is through attention to these intentional features that we can best account for what the texts were meant to accomplish and what makes them more widely significant.

56There is, for example, the noteworthy difference between Iḷampūraṇar’s writing and Cěyiṟṟiyaṉār’s. Iḷampūraṇar writes in a formal prose style that is obviously calqued upon Sanskrit śāstric writing, while the lost source text is an independent work written in a classical verse-form, largely devoid of a Sanskritic lexis or a syntactic structure borrowed from Sanskrit, and one which seems, above all, grounded in the stylistic model of the Tǒlkāppiyam itself. We can with some confidence argue that the time of composition of the Cěyiṟṟiyam dates to the late 11^th or 12^th century (after Abhinavagupta but prior to the earliest citations of Iḷampūraṇar), and so it may be located within a still wider world of śāstric Tamil writing, which was evidently far more heterogeneous than those works to which we still have access, as can be gathered from the sources adduced by Aṭiyārkkunallār in the dramaturgical essay he appends to his gloss on Cilappatikāram 3.12.57 These other texts (though this is difficult to gather from the surviving references) may have more closely resembled the Cĕyiṟṟiyam in their form and style of presentation, leaving it to authors like Iḷampūraṇar and Aṭiyārkkunallār to fashion a new style of scholarly Tamil prose closely following in their wake.

57While Śāradātanaya’s habit of pseudepigraphical citation might seem to problematize such a picture of textual profusion (might we wonder if Iḷampūraṇar and Aṭiyārkkunallār are just making up their sources, too?), this is less of a problem than it appears to be. In fact this permits us access to another side of this same bibliographic proliferation, the sea in which our authors swam. What we can see in the BhP if anything casts the picture of this proliferation into even higher relief: faced with an already wide spectrum of sources and argumentative warrants, Śāradātanaya seems to have been prepared to conceal or to downplay his own innovations and unprecedented combinations by displacing these onto other, invented works. This suggests to me something of the sheer quantity of possibilities that were then available: because new texts were circulating and gaining acceptance so rapidly, this sort of compositional (and so argumentative) move was not only possible but acceptable. These newly confected citations tend to be displaced not onto properly “authorial” works of śāstra, but instead onto works ascribed to supernatural figures like Śiva or Vāsuki, or onto absent ur-texts such as Bharatavṛddha or Mammaṭa’s putative source, the Kalpalatā.

58All of this accords well with the bigger picture that text-historical scholarship has begun to work out for the far South in the long twelfth century. From the early 1100s there is the well-known if all but unanalyzed creative explosion of literary Tamil, the seemingly miraculous generations of Cayaṅkǒṇṭār, Ǒṭṭakkūttar, Cekkiḻār, and Kampaṉ. There is also the enormous growth industry of new and usually anonymous Sanskrit verse composition, resulting in whole new canons for different Śaiva, Vaiṣṇava, and Śākta religious orders,58 along with an historically localizable new production of Sanskrit purāṇas (to be soon followed by the earliest Tamil talapurāṇams), often with greater literary or canonical ambition than earlier works of the genre.59 A new or newly invigorated regime of textual creation appears to have been in progress, a cultural-historical transformation for which literary theory plausibly served as an auxiliary, while the compositional practices of the literary theorists themselves help us to sharpen our attention towards what is really distinctive about it. That all of this occurs in the midst of the transformative social-historical processes that historians have reconstructed is, needless to say, significant: a rapidly changing world provides the occasion for the creation of new texts to describe, to regiment, and to beguile.

59Poetics or literary theory was the discipline that is perhaps most attuned to these sorts of cultural transformations, and so potentially at least the śāstra from the perspective of which we are best able to venture questions about the relation to wider social transformations, as well. Alaṃkāraśāstra, as McCrea has convincingly argued (2011), was the only classical knowledge system that was constitutionally inclined towards instability and to conceptual change, and this observation can be generalized, to include Tamil literary theory as well as the documentable exchange that occurred between theorists working in either of the two languages. Above and beyond their individual theories, arguments, and debates — which we certainly should read “over and over” — the works on the nature of poetics and literary language that emerged in the medieval Tamil country provide a powerful evidentiary resource for tracking wider patterns of change, and so at once giving these interwoven disciplines an intellectual history that is, so to say, inside of history.