Thoughts on the Early Indian Yogācāra Understanding of Āgama- Pramāṇa

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

Thoughts on the Early Indian Yogacara Understanding of agama- Pramana . by Roy Tzohar

The Buddhist approach to testimony (aptavada, apt?gama) as a valid means of cognition (pramana) is far from univocal and involves an intricate and often also ambivalent attitude toward scriptural authority. The paper focuses on several early Yogacara Buddhist thinkers who accepted testimony as a reliable epistemic warrant, and offers an account of the sophisticated and highly reflective manner in which they approached the issue of scriptural meaning and authority. For this purpose, the paper first outlines the theoretical framework for considering scripture presented by the early Yogacara philosopher Vasubandhu’s Vy?khy?

yukti, focusing especially on his discussion of the criteria for canonicity and its implications for a system of hermeneutics based on the uncovering of authorial intent. The paper then examines in turn the way in which this framework and its internal tensions were worked out in the writings of Sthiramati (circa 6th century CE) and especially in his Madhy?ntavibh?ga-bh??ya-??k?, focusing on his definition of “treatise” (??stra) and his implied understanding of textual authority.

The Buddhist approach to testimony (?ptav?da, ?pt?gama) as a valid means of cognition (pram??a) is far from univocal – it varies across times, schools and sometime even within the thought of a single thinker1 – and involves an intricate and often also ambivalent attitude toward scriptural authority. Hence, for instance, we find on the one hand such Buddhist thinkers as the early Yog?c?ra philosophers Asanga and Vasubandhu, whose acceptance of testimony as a pram??a went hand in glove with a rather wary approach to the authority of scripture, and on the other, thinkers like Dign?ga and Dharmak?rti, who, while rejecting testimony as a pram??a, still held scripture to have a certain epistemic purchase.2

1 For instance, in his Abhidharmasamuccaya, Asa?ga takes testimony (?pt?gama) to be subservient to direct perception and inference, while in some sections of the Yog?c?rabh?mi corpus (the Bodhisattvabh?mi, and the section on the Hetuvidy? in the ?rutam?y?bh?mi) also traditionally ascribed to him, he seems to presents it as independent of these two. See, respectively: Nathmal Tatia (1976: 253, section iih); Dutt (1978: 25, line 17-19); Wayman (1999: 23).

2 While the epistemological school of Dign?ga and Dharmak?rti did not accept testimony as valid cognition (pram??a), it allowed for a type of inference that was based on scriptural sources, over and above the “ordinary” kind of inference (the latter defined as vastubalaprav?tt?num?na, literally – an inference that functions by the force of [real] entities, i.e., an inference that is evaluated on the basis of facts and states of affairs). These two types of inference were differentiated by a division of epistemic labor according to which ordinary “objective inference” (to follow Tillemans’ phrase) deals with imperceptible (parok?a) things, i.e., things that are inaccessible to direct perception but can be accessed through inference (for example, the “impermanence” of sound), while the latter kind of inference is limited to cases that involve extremely Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of agama- Pramana

In this essay I focus on several Yog?c?ra Buddhist thinkers who accepted testimony as a reliable epistemic warrant, and offer an account of the sophisticated and highly reflexive manner in which they approached the issue of scriptural meaning and authority. My aim is not to provide an exhaustive account of this vast and complicated topic but to highlight in broad strokes some of the basic presuppositions and tensions that underlie these thinkers’ conception of scriptural exegesis and authority. For this purpose I first outline the theoretical framework for considering scripture presented by the early Yog?c?ra philosopher Vasubandhu’s (est. 4th-5th century, CE) as laid out in his Principles of Proper Exegesis (Vy?khy?yukti, henceforth VY; Lee 2001). My focus in this first part is mostly on Vasubandhu’s analysis and discussion of the criteria for canonicity and its implications for a system of hermeneutics based on the uncovering of authorial intent.

In the second part of the paper I turn to examine the way in which this framework and its internal tensions were worked out in the writings of Sthiramati (Circ. 6th century CE), a later Yog?c?ra thinker renowned mostly for his commentarial work on Asa?ga and Vasubandhu’s treatises. Here I focus on Sthiramati’s intriguing remarks on what constitutes a treatise (??stra) and his implied understanding of textual authority, as presented in several key sections of his sub-commentary to Vasubandhu’s own commentary on Distinguishing the Middle from the Extremes (Madhy?ntaVibh?gabh??ya- ??k?, MVBh?; Yamaguchi and Lévi 1934).

1. Vasubandhu on the interpretation of scripture

The locus classicus of Vasubandhu’s understanding of scriptural exegesis is his VY, a work that belongs to the uncommon genre of manuals for the proper interpretation of scripture and composition of commentary. The work is not just prescriptive, however, but also demonstrates the application of its lessons by interpreting some one hundred sutra passages assembled in an appendix to the work (the Vy?khy?yuktis?traka??ha?ata), and as such, it is at once a commentary on Mahayana scripture and a valuable source of knowledge on the way in which the Buddhist scholastic tradition conceived of its role, aims, and limitations.3 In the work, Vasubandhu states that one of the characteristics of a good commentary is its ability to respond to various objections regarding particular interpretations of scriptural passages, and then demonstrates how this should be done in a inaccessible (atyantaparok?a) objects, such as the detailed workings of karma, which are both beyond direct perception and syllogisms of the first kind. See Tillemans (1999: 27-30).

3 Despite the works’ significance for both Indo and Tibetan Buddhist traditions (attested by its citation by various prominent commentaries), it has received relatively little scholarly attention to date, and has yet to be fully explored. For a survey of secondary literature on the work, see Tzohar (2013).

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

section titled “Objections and Responses” (’gal len, *codyaparih?ra). Here I would like to focus on a subsection in which Vasubandhu deals with “objections to meaning,”4 that is, cases in which Mah?y?na scriptural claims appear to be inconsistent or incompatible with other claims, primarily with what is normally regarded as authoritative speech (yid ches pa’i gsung, ?ptav?da), in this case, the canonical works of early Buddhism.

Vasubandhu begins by attending to the phenomenon of interpretative difficulty posed by internal contradictions within a given Buddhist scriptural corpus – for example, the Buddha’s various and seemingly inconsistent claims about “truths.” This is settled by an appeal to a common Buddhist hermeneutical device, namely treating all the claims as true insofar as they express different implicit intents (abhipr?ya) of the Buddha in different contexts (Nance 2012: 257-258). For our purpose it should be noted that according to this account, the meaning of scripture is not always self-evident but requires hermeneutical mediation to

maintain consistency and coherency. Moreover, meaning is understood here in terms of the speakers’ (implicit) meaning, which can however be made explicit by an interpretative reconstruction. So, while scripture is capable of serving as an authoritative source for knowledge, this ability depends on its reliance on a ‘pre-textual’ authorial intent, which can be successfully reconstructed through interpretative efforts.

The second kind of “objections to meaning” presented by Vasubandhu are cases in which scriptural statement stand in contradiction with reason (rigs, yukti), and by extension, with one of the three means of valid cognition (pram??a), namely perception, inference, and authoritative speech (yid ches pa’i gsung, ?ptav?da). Below I focus solely on the latter case, which Vasubandhu presents through the following objection placed in the mouth of a hypothetical ?r?vaka opponent, translated by Richard Nance:

‘The Vaipulya group [of texts] is Mah?y?na’ is a statement that is in contradiction with scripture (lung, ?gama). Some say that books of s?tra containing lengthy expositions are Vaipulya, but not Mah?y?na. Why? [Because] that [class of texts] is not the speech of a Buddha. Why is it not the speech of a Buddha? Because of contradiction — [i.e.] it contradicts what all groups acknowledge to be the speech of a Buddha (Nance 2012: 120). Here the opponent undermines the canonical status of the Mah?y?na scriptures because they are allegedly contradictory to what is already canonically accepted as the word of the Buddha and hence considered to be authoritative speech and a source for knowledge. The opponent then 4 This section was analyzed at length in Cabezón (1992), Nance (2004) and Nance (2012). The following analysis draws from both authors.

Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

proceeds to point out various features of the Mah?y?na scriptures that undermine their authority and canonicity, for instance, their manifestation of internal contradictions, their incompatibility with fundamental Buddhist tenants, etc. (Cabezón 1992: 226-33; Verhagen 2005: 587-93).

Vasubandhu’s line of defense is rather exceptional insofar as it does not seek to buttress this or other Mah?y?na claims but, rather, to undermine the opponent’s very conception of canonicity, and by extension to question the latter’s understanding of authoritative speech. In his reply, Vasubandhu points out that contradictions and inconsistencies are rife even in the s?tras of the early canon, which is itself incomplete (as some of the texts referred to there were not in existence already in Vasubandhu’s time) and inhomogeneous even among the ?r?vaka schools (hence its authority cannot be derived from its past acceptance by a certain school or lineage). In light of all of this, he argues, the opponent’s own criteria for scriptural authority render canonicity itself inadequate

to serve as the standard for scriptural truth. As noted by Cabezon, Vasubandhu’s aim here is neither to refute the authenticity of the early Buddhist canon nor to undermine the epistemic role of testimony but only to demonstrate the practical impossibility of establishing the authenticity of any Buddhist text (the Mah?y?na s?tras included) solely on the basis of a clear-cut philological or historical criterion (Cabezón 1992: 228). Canonical indeterminacy and textual inconsistencies, he seems to suggest, are not themselves markers of textual inauthenticity, but are a call for the interpreter to step in and recover the implicit intent underlying the imperfect text, thereby restoring its connection to the authoritative words of the Buddha.

While the hermeneutical framework described above is by no means exclusive to the Mah?y?na,5 it does seem to assume an exceptionally significant role in this context. Seen as constituting a response of sorts to doubts regarding the Mah?y?na scriptural authenticity, this framework emerges as not merely a hermeneutical device but a constitutive feature of the Mah?y?na textual and commentarial identity. This very understanding seems to underlie, for instance, the hermeneutical agenda of such seminal texts as the Sa?dhinirmocanas?tra, and is openly acknowledged by various other Yog?c?ra scriptural and commentarial

passages. An example that epitomizes this approach is found in Sthiramati’s Abhidharmasamuccayabh??ya, where he states, with respect to the definition of 5 Ronald Davidson has argued that an earlier hermeneutical apologetics was applied by the Abhidharma — being the first new class of texts in Buddhist India that claimed the status of scripture, and laid the ground for the later Mah?y?na apologetics. See Davidson (1990: 303-06).

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

Vaipulya, that those inclined to understand only the literal meaning (yath?rut?rth?bhinive?a) cannot “ride” the great vehicle, because the Mah?y?na doctrines are expressed only via implicit meaning.6 It should be emphasized that this understanding of meaning as inherently layered was not merely a feature of the interpretations and commentaries on the Mah?y?na scripture, but traces back to the very nature of the Buddha’s word. This is manifested in the polysemic quality of the Buddha or a Bodhisattva’s speech, seen as capable of simultaneously teaching beings of diverse levels of understanding. Consider, for instance, the following paragraph from the Da?abh?mikas?tra, a text traditionally identified with the Yog?c?ra (and for which Vasubandhu is said to have written a commentary):

If all the living beings, involved in the triple thousand great thousands world regions, approaching him, would ask questions in a moment and each of them would ask with the variety of immeasurable sounds and the second living being would not ask what the first one asks, a Bodhisattva will comprehend (all) the words and syllables of voices of all the living beings and after having comprehended he will completely satisfy the intentions of minds of all living beings by

uttering one voice (Honda and Rahder 1968: 249). The polysemic quality of the Bodhisattva’s voice is clearly supposed to explain his ability to relieve the suffering of all beings, but it may also, against the backdrop of the Mah?y?na apologetics discussed above, stand to explain the apparent diversity of Buddhist teachings and constitute an argument of sorts for the Mah?y?na s?tra’s own authenticity (Davidson 1990: 309).

While this passage seems to imply that a Bodhisattva’s authorial intentions are as multifarious as his interlocutors’ understandings of them (and that the hermeneutical recovery of these intentions is necessarily contextual and relational),7 it should not be understood as abandoning the possibility of interpretive closure. Indeed, while the Mah?y?na notion of meaning as inherently layered opens up a hermeneutical space in which the commentarial endeavor can materialize creatively and unapologetically, this hermeneutical openness is far from being all-permissive. The very idea of the Buddha’s teaching as revealing

the true nature of reality requires meaning to be decisively and univocally determined, and to this end various Mah?y?na discourses establish a set of hermeneutical tools and criteria designed in theory to provide a definitive understanding and interpretation of the Buddha’s words (the most important of these tools being, of course, the distinction between 6 Tatia (1976: 112, lines 8-19). On the issue of non-literality in the Mah?y?na scriptures with respect to the La?k?vat?ras?tra, see Lamotte (1988: 15).

7 Or alternatively, that interpretative closure is not to be found in the speech acts of the Buddha. See, for instance, Lugli 2010. Lugli points out an underlying dichotomy – common to various Mah?y?na s?tras – between the mere words (ruta) and the meaning (artha), and understands the latter to be inconceivable within the sheer realm of language. Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

interpretable and definitive meaning – ney?rtha and n?t?rtha respectively – and their respective conditions, which apply to both discrete claims and entire texts; see Lopez 1988: 65-67 and Lamotte 1988: 16-20).

The main difficulty with these criteria for achieving interpretative objectivity and conclusiveness, however, is that in light of Vasubandhu’s critique described above, we must regard them as inevitably circular: definitive teachings are the mark of authoritative texts, and authoritative texts, in turn, determine which teachings may be considered definitive. Thus, despite the rhetoric of definitive meaning and authority, in praxis these criteria were often applied within Mah?y?na literature more through commentarial ingenuity than through rule following so as to confirm sectarian loyalties and pre-existing theoretical assumptions.

To recap: the hermeneutical framework described above presents a notion of scriptural authority marked by an internal tension – between the inherent openness and proliferation of meaning underlying the Mah?y?na view of scripture as the Buddha’s word, on the one hand, and its need for a clear-cut objective hermeneutical criteria for the determination of meaning, on the other; and between the rhetoric pronouncing such a definitive criteria and the inevitable circularity that in fact underlies its application.

In the following section I turn to examine how and to what extent this hermeneutical framework, which has thus far appeared to be largely normative and prescriptive, along with its underlying tensions and inconsistencies, was worked out by the “normal science” of commentarial discourse. For this purpose I turn to Sthiramati’s sub-commentary on Vasubandhu’s commentary on the Madhy?ntavibh?ga, the former commentary being particularly fruitful for this kind of inquiry 8 See Lopez (1988: 52) and Lamotte (1988: 19).

9 It is hard to determine the extent to which this tension was explicitly acknowledged by the Yog?c?ra thinkers, though it is unlikely that Vasubandhu could produce in the VY such a nuanced textual critique of early Buddhist canonicity without seeing its broader implications for his own criteria for establishing the status of the Mah?y?na texts – and in particular, the circularity of these criteria. An indication that Buddhist thinkers were indeed aware of this tension can perhaps be found in their professed and recurrent expressions of concern regarding scriptural misinterpretation, even at their own hands. Viewed in this light, Vasubandhu’s famous statement in the concluding section of his Abhidharmako?abh??ya is particularly telling:

‘Is the catechism [[[Abhidharma]]. R.T] expounded in the present treatise the same as the one that the Teacher set forth? For the most part I have expounded the catechism established in the teachings of the Vaibh??ika school of Ka?mir.... Whatever I have misunderstood here is my own fault. For only Buddhas and their direct disciples are authoritative in teaching the true religion. Now that our Teacher is dead, the eyes of the world are closed, and now that the majority of those who had firsthand experience [of the truths he taught] have met their ends, his teaching, which is being transmitted by those who have not seen reality and have not gained freedom [from their passions and misconceptions] and are inept at reasoning, has gotten all mixed up.... So, those who desire salvation, seeing that the Buddha's teaching is gasping its last breaths . . . must not become distracted.’ (Hayes 1984: 654). Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

because of its heightened hermeneutical self-awareness, a trait that no doubt reflects Sthiramati’s particular position within the Yog?c?ra’s textual development, as he is facing the task of further synthesizing a vast corpus into a single coherent and consistent worldview that fits under the doxographical heading of Yog?c?ra.

2. Sthiramati’s definition of ??stra in the Madhy?ntavibh?ga-bh??ya-??k?

In the introductory part of his sub-commentary, commenting on Vasubandhu’s salutary verses, Sthiramati provides an overview of the text’s transmission and accounts for its authoritative status. According to this account, the “author” or “promulgator” (pra?et?) of the text was Maitreya, said to be a Boddhisattva removed from complete Buddhahood by only one birth, who revealed the work in verse form10 to the expounder (vakt?) Asa?ga, who in turn made it available to Vasubandhu, who composed the commentary.11 According to Sthiramati, one of the aims of the salutary verses is to inspire general reverence for the text and buttress its authority:

Reverence is generated towards the commentary because the meaning of the S?tra is unerringly stated in it because these two [Asa?ga and Vasubandhu. R.T], since they possess the highest wisdom, are able to understand, retain and explain [it], without erring. In this way, reverence arises towards both the S?tra and the commentary on the part of those who rely on the authority of individual people. Also, on the part of those who rely on the Dharma, reverence arises towards

both the author and the expounder because after the true meaning of S?tra and the commentary has been understood, when a positive determination occurs, it is brought about through the understanding of the author and the expounder, but it is not accomplished through just speculation and scriptural tradition – thus reverence is generated towards the author and the expounder (Stanley 1988: 3).12

10 In the dedication, Vasubandhu refers to “the author of the treatise” (??stra). Sthiramati’s commentary clarifies that this means the k?rik?s authored by the Bodhisattva Maitreya, and hence he consistently refers to the work not as a ??stra but as “the Madhyant?vibh?ga s?tra.” See Yamaguchi and Lévi (1934: 2, line 4). The expounder he explains to be Asa?ga. 11 Sthiramati seems to distinguish between two commentaries – the hypothetical v?tti of the expounder Asa?ga, and the bh??ya of Vasubandhu, which is the work before us. See Yamaguchi and Lévi (1934: 1, lines 11–12; 3, lines 1–2; and 2, lines 8– 9), respectively. Apart from these instances, Sthiramati makes no further use of this distinction, which was perhaps intended to indicate that the text delivered to Vasubandhu was already accompanied by Asa?ga’s elucidation. 12 Here I follow Stanley’s suggested corrections (based on the Tibetan) to Yamaguchi’s edition. The passage thus reads as follows:

t?v api uttamaprajñ?vantau ‘bhr?nti prativedhadh?ra?opade?as?marthy?d atra s?tr?rtham abhr?ntam upadi?ata iti v?ttiy?m gauravam utpadyate /eva? ye pudgalapr?m??ik?s te??? s?trav?ttigauravotpatti? / ye ca dharm?nus?rinas te??? s?trav?ttyo? ?ubh?rthe ‘vabodha? / ni?caya? ced utpadyate sa pra?et?vaktr?vabodh?d api prabh?vito bhavati na tu tark?gam?tre?a prabh?vito bhavat?ti pra?et?vakt?gauravotpatti? /

See Yamaguchi and Lévi (1934: 3, ns. 6, 7). Unless otherwise indicated, all further translations of ??k? passages are my own. Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

Here the authority of the text seems quite evident, firmly established by the double criteria of the reliability of its source (attested by the level of accomplishment of the author and expounder) and the text’s compatibility with the doctrine (once the author’s and expounder’s intentions are properly understood). Sthiramati, however, goes on to further problematize this issue by complementing it with a rather intriguing definition of ??stra:

Now this should be discussed: what is the nature of a treatise (??stra), and why is it [called] ??stra? A treatise consists of representations appearing as a collation of names, words, and syllables. Or alternately, a treatise consists of representations appearing as specific linguistic expressions that procure supramundane direct knowledge. But how can representations be articulated or expounded upon? There is no fault in here since the representations [obtained] by hearing follow from the representations of the author and expounder [[[Maitreya]] and Asa?ga].

Sthiramati’s definition of ??stra (alongside an etymological definition drawn from Vasubandhu’s Vy?khy?yukti)15 is obviously made to fit the Yog?c?ra understanding that all phenomena – including all types of discourse – can be either known or discussed as mere mental representations (vijñapti). But this, says the opponent, poses a difficulty for its function as a discursive communicative activity. There are several ways of understanding what it is that Sthiramati sees as a possible difficulty. On one interpretation, the difficulty concerns, as in the case of the epistemological problem of the knowledge of other-minds, the epistemic access or lack thereof of a commentator to the mental representations of the expounder and author.16 An alternative interpretation,

presented recently by Sonam Kachru, understands Sthiramati as referring here to the fundamental difficulty of attributing intention ascriptions – which are presupposed by any communicative discursive acts (and all the more so by commentarial activity) – to mere mental events (i.e., independently of any intentional agent; Kachru 2014). Drawing from Buddhist philosophy of action, Kachru suggests that Sthiramati’s solution to this difficulty is to point out that intentional content, much like public action, is not 13 As mentioned above (n. 19), the term ??stra stands in this context for the entire MAV corpus, including both revealed verses and commentary.

14 idam id?n?? vaktavya? k?d??a? ??strar?pam | ??stra? ki? ceti n?mapadavyañjanak?yaprabh?s? vijñaptaya? ??stram |atha v? lokottarajñ?napr?paka?abdavi?e?aprabh?s? vijñaptaya? ??stram | katha? vijñaptaya? pra??yanta ucyante v? | pra?et?vakt?vijñaptiprabhavatv?t ?rava?avij?apt?n?? n?tra do?a?|. The Sanskrit is from Yamaguchi and Lévi (1934: 2, lines 16- 20). I have accepted Stanley’s suggested correction of the Sanksrit, based on the Tibetan (D190a5), of ?rava?avij?apt?n?? in place of prajñapt?n??; see Stanley (1988: 3, n. 9).

15 See TD. 4061 shi 123a, in Lee (2001: 227). 16 See Kochumuttom (1982: 210: 212 n. 7). Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

external to the perception of a communicative act but rather constitutes, phenomenologically speaking, its experience as such. In this respect, ascribing intention to a mental event would not be false because intentions are regarded as an intrinsic feature of the experience of the mental event as a discursive act. As regards our present concern, this interpretation carries important implications for the understanding of authorial intent. As noted by Kachru, while this shift into the phenomenological realm sustains the necessary alterity of the author’s intent (that is, it is different than the intent of the interpreter), confining intent to our experience also implies that it is inevitably mediated and accessible only via our interpretation.

While I find this argument largely convincing, it should be noted that its view of intent as necessarily mediated (and hence inevitably reconstructed) by interpretation arguably undermines the Buddhist hermeneutical approach described above and in which Sthiramati partakes. This is because treating intention ascription as an intrinsic feature of our experience would appear to be incompatible with treating the conclusive determination of authorial intent as a criterion for correct interpretation, and thus as extrinsic to interpretation. This incompatibility can be reconciled, it seems to me, by considering that

Sthiramati is referring in this context not just to any discursive acts but to those prompted by the exceptionally accomplished personas, namely the author, expounder and “listener” (Vasubandhu).17 So that whereas under the framework of mere mental representations intentions become themselves a matter of mental representations (insofar as they are necessarily mediated by interpretation), it is these accomplished personas’ utmost wisdom, Sthiramati seems to tell us, that guarantees that the intention interpreted is indeed the one intended. It is therefore the textual authority that stems from possessing true knowledge of reality that rescues interpretation from becoming mere self-interpretation.

Finally, there is yet another way to understand the difficulty that arises from Sthiramati’s account of ??stra. According to Sthiramati’s definition – and it should be noted that this is his own emphasis, and not Vasubandhu’s – ??stra, insofar as it is carried by a line of second order mental representations, is fully immersed in the conventional realm. As such, he tells us, ??stra can be understood to operate in two distinct ways. First, it can be understood simply in terms of its assembled linguistic parts and their meanings, like any other communicative act. In addition to this, however, and unlike ordinary communicative acts, the discursive elements of ??stra can be 17 Most telling in this respect is Sthiramati’s omission of his own commentarial project from this tally. 18 This framework seems to be on par with Sthiramati’s additional gloss of ??stra as an edifying discourse, which is evidently intended for Buddhist “insiders.” See Yamaguchi and Lévi (1934: 2, line 20; 3, line 2). Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

understood as efficacious in procuring supramundane gnosis, implying that the genre has a performative and transformative function over and above its informative role.19 The main difficulty that arises from such a view is that defining discourse as mere mental representations of mental representations can imply a kind of mentalism, and as such, threatens to deprive discourse of any extra-discursive referential foundation.20 An explanation is therefore required as to: 1) how discourse can maintain its meaningfulness – how it can “be articulated or expounded upon,” in the words of the objection above– and 2) how, if

discourse is self-referential, it can still maintain its pretense to “reach” reality so as to be able to procure true knowledge. Seen in this light, Sthiramati’s discussion of the nature of ??stra seems to go beyond mere local textual concerns to question the meaningfulness and efficacy of the full range of the Yog?c?ra discourse. A possible solution to this difficulty may be found in other sections of Sthiramati’s subcommentary – most notably the third, “Reality”

chapter (tattvapariccheda?)21 – which propose a general understniang of how reference and ordinary discourse map onto the Yog?c?ra three natures scheme (trisvabh?va). As we will see shortly, this understanding allows Sthiramati both to ground discourse in causal mental descriptions, enabling him to avoid the traps of mentalism, and to view ??stra as a vehicle for transformative knowledge.

As a preface to my exploration of this account of the referential process, a brief exposition of the three nature scheme is in order. Stated in the most general terms, this important Yog?c?ra tenant provides an account of reality as an interplay between three different aspects or points of view – the deluded way in which we ordinarily conceive phenomena (the imagined nature), its interdependent causal foundation (the dependent nature), and the true understanding of the ultimate state (the perfected nature). The presentation of this scheme varies across Yog?c?ra texts, perhaps reflecting a gradual development of this

doctrine, and the question of its proper interpretation and the nature of the relations among the three natures is a matter of contention among scholars, both ancient and modern. One useful way of approaching this broad and complex issue is through Alan Sponberg’s distinction between a “pivotal model” and a “progressive model” of the three natures scheme (Sponberg 1982).

19 Here too Sthiramati seems to follow Vasubandhu’s VY, where the Buddha’s speech is described as the ultimate ??stra. See text quoted in fn. . 20 Insofar as mentalism treats meaning as a function of the speaker's mental content. 21 In particular the subsections on the “subtle and gross reality” (aud?rikas?k?matattvam) and the “skillful reality” (kau?alyatattva?), the relevant sections roughly commenting on MAV verses 16 c-d and 10 b-c respectively. Sthiramati’s comments are invariably consistent and coherent, but their presentation is not ordered or linear. In what follows I will try to reconstruct their overall argument.

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

In brief, the pivotal model takes both the imagined and the perfected nature to be aspects or modes of the dependent nature. The imagined nature is regarded as the deluded way in which we ordinarily conceive phenomena (in terms of a dualist distinction between subject and object and through conceptual differentiations), and the dependent is understood as the ontological interdependent causal reality underlying the imagined. In other words, while the imagined is the mistaken way in which reality appears to us, the dependent stands for the real causal nexus that brings about these false appearances. The perfected is understood within this framework as simply the absence of the imagined from the dependent – i.e., the dependent as seen once our misconceptions of it have been understood and removed.

The progressive model, by contrast, takes each of the three natures to represent stages of understanding, progressively higher, of the true nature of reality. Within this framework, one first conceives phenomena, mistakenly, as dualist appearances (the imagined), then understands them in causal and mental terms (the dependent), and eventually transcends this understanding (which is still pervaded by dualistic distinctions) in favor of a true vision of ineffable reality. With respect to the MAV, Mario D’Amato has argued that while the pivotal model appears to be the more prevalent, there seems to be at least one instance that

conforms to the progressive model, suggesting perhaps that these are not after all incompatible perspectives.23 For present purposes, however, suffice it to point out the common features in the theoretical work performed by the two models of the three natures scheme: both enable the Yog?c?ra to supplement the two-truths distinction with a more dynamic model of reality understood in terms of a shift between three epistemic points of view, and this dynamic model, in turn, can readily accommodate the school’s causal description of the workings of the mind and its role in the construction of phenomena. Bearing this in mind, let us now return to Sthiramati’s explanation of the workings of language and reference through their mapping onto the three natures scheme. According to Sthiramati,

the realm of the imagined can be correlated to a realist and essentialist understanding of meaning in 22 Still largely within the framework of the pivotal model, one can further distinguish between roughly two common interpretative strands that are tightly connected to the question of the nature of the Yog?c?ra Idealism – an approach that sees the three natures doctrine as providing an ontological account of reality, and one that considers it a description of reality from three different epistemic points of view. A lucid discussion of these interpretative approaches and the philosophical role of the three natures scheme is provided in a recent polemic between Jay Garfield and Jonathan Gold held at the 2011 Annual Meeting of the American Academy of Religion. See Garfield and Gold

(2011). 23 ‘But perhaps the pivotal and progressive models may be seen as consistent with one another, each providing an interpretation of the three-nature theory in a different key: while the pivotal model interprets the three natures in an ontological epistemological key (mind in relation to world), the progressive model interprets the three natures in a semiotic-epistemological key (mind in relation to signs, language, concepts)’ (D’Amato 2012: 18). Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

which designations are ordinarily (and falsely) understood to refer to existent essences (Yamaguchi and Lévi 1934: 140, lines 5-14). The dependent correlates to the understanding of these referential processes in terms of the causal-mental activity that underlies them. The perfected, as the absence of the imagined from the dependent, is understood as the ultimate nonexistence of such designated essences. Sthiramati further notes that while the dependent serves as the causal support for the process of designating, this is not evident in ordinary discourse because words are erroneously understood as referring to essences that

are superimposed (adhy?ropita) on the causal nexus.24 So, according to Sthiramati, words are ordinarily understood to refer to (imagined) existent essential objects that as a matter of fact are superimposed on the dependent, the latter conceived of as the causal nexus that brings about the mental appearances of these objects (and their corresponding designations).

In another subsection of this chapter, the “subtle and gross reality” (aud?rikas?k?matattvam), Sthiramati further develops this theory of meaning to differentiate between three modes of conventional language use (which once again map onto the three natures scheme).26 According to Sthiramati, when a person uses a word to refer to what he considers to be real existent entities, this mode is termed “the conventional as designation” (prajñapti-sa?v?tti?). Here designation is understood in terms of its bare denotative (abhidh?) function, subsumed within the category of the imagined. The second mode is the “conventional as comprehension” (pratipatti), which stands for the deep structure of linguistic use, namely attachment (abhinive?a) to objects that are taken to be externally existent because of conceptual differentiation (vikalpa). This mode is seen as corresponding to the dependent nature.

The final mode, “the conventional as expression” or “proclamation” (udbh?van?), is conventional language use with respect to true reality (dharmadhat?), despite the latter’s ultimate inexpressibility. This mode corresponds to the perfected. It should be emphasized that for Sthiramati, these three modes do not represent three different ways of talking 24 Which he elsewhere seems to refer to as consciousness. See Yamaguchi and Lévi (1934: 139, lines 3-9). 25 Sthiramati’s account of how reference and meaning work according to the Yog?c?ra (that is, along the three natures scheme) is then supplemented, and to some

extent supported, by a devastating critique of a realist-essentialist theory of meaning. This consists in demonstrating that any attempt to understand linguistic meaning as the outcome of relations between the triad of designation, an object, and its essence, is logically untenable. To this end, Sthiramati embarks on a set of arguments that also consider other referential theories – such as the view that designations manifest or reveal the object – and which are a summarized synthesis of similar arguments from a variety of sources, including the Bodhisattvabh?mi, the Vini?cayasa?graha??, and the Mah?y?nasa?graha. See Yamaguchi and Lévi (1934: 140, line 10; 141, line 21). For a survey of similar arguments in the Yog?c?ra texts mentioned above, see Tzohar (forthcoming). 26 The text section under discussion is Yamaguchi and Lévi (1934: 123, line 19; 125, line 3). Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

but, rather, three ways of conceiving of the meaning of the same language use27 – the first merely in its denotative-semantic function, the second as indicating the various causal processes underlying it, and finally in terms of its soteriological efficacy, that is, its capacity to delineate its own limitations. This layering of the meaning of conventional discourse brings us back full circle to Sthiramati’s initial definition of ??stra and its attendant difficulties. Recall that these concerned, first, the meaningfulness of discourse, and second, its ability to procure true knowledge under a framework of mere mental representation (which threatens to undercut discourse form any extra-discursive referential basis). As a possible response to these difficulties,

Sthiramati seems to tell us that conventional discourse, when properly understood, reveals itself to be inherently polyphonic, simultaneously speaking in different voices that embody different points of view on the same reality (and which correspond and are delimited by the three natures scheme). It is therefore ??stra’s embrace of this polyphony that renders it a vehicle for obtaining super-mundane knowledge: the polyphony reflexively facilitates a shift away from the understanding of discourse as reifying linguistic phenomena toward its understanding as a line of second-order representations understood in terms of their underlying causal and mental reality. In this respect, ??stra, while still conventional insofar as it reflexively points out the limits of ordinary expression and facilitates a gradual understanding of the true causal reality underlying language, is seen to be denotative of the ultimate.

Furthermore, this framework also enables Sthiramati to avoid the kind of mentalism that would readily attach itself to a view of discourse as second-order mental representations. This is because while there are no real external objects to serve as referents for words, meaning and reference are nonetheless grounded in the causal mental reality that produces them. So with respect to the first difficulty, while this account undermines a notion of meaning as a function of a realist correspondence theory of truth, it does not render all discourse meaningless. Rather, what we find here is an understanding of discourse that, by reflexively uncovering its very conditions of becoming, broadens its semantic range to include causal and mental descriptions.

27 The way these three modes being analogous to the three natures scheme as three “modes” of viewing reality seems selfevident. 28 It should be emphasized that this does not mean merely that ??stra is capable of demonstrating the conventionality of all discourse and the limits of language (and hence of gesturing in the direction of an inexpressible absolute); it suggests that ??stra also facilitates the turning of one’s gaze upon the ontological and causal reality that lie at the root of conventional discourse, and in this case, of ??stric discourse itself. ??stra shows the ontological conditions that underlie discourse, including the ones underlying itself.

29 Hence accounting for referential relations in terms of causal and mental descriptions rather than under a realist correspondence theory of truth. This link between causal ontology and the realm of language and meaning is found also in Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

3. Conclusion

In the first part of this essay I outlined several features of Vasubandhu’s account of authoritative speech as grounded in a broader understanding of scriptural interpretation. I began with his critique of the attempt to establish canonicity and textual authority using philological or historical criteria, and his alternative proposal of a hermeneutical approach that identifies meaning with implicit authorial intent, excavated from the text by the application of hermeneutical criteria that ensure the definitive status of such meaning. I argued that this non-literalism and the attendant view of meaning as inherently layered becomes something of an interpretative strategy for the Mah?y?na and an important part of its textual identity, which is often in conflict with a no less important practical need for final and definitive interpretative closure. Thus, this scheme was shown to be marked by an unresolved tension between the inherent openness and proliferation of meaning underlying the Mah?y?na view of scripture, on the one hand, and the school’s need for clear-cut objective hermeneutical criteria for the determination of meaning (and the inevitable circularity that in fact underlies its application), on the other.

In the second part I examined how the inherent tensions of this framework are worked out in Sthiramati’s MVBh?. Considered in light of the general interpretative framework described above, Sthiramati’s account of ??stra appeared to be engaged primarily with the problem of conclusively determining the meaning of a text under a “representation only” account, once both the agent and external objects are taken out of the equation. This is because the “representation only” view threatens to undermine the understanding of meaning as a function of either intention ascriptions or an extra-discursive referential ground, which is required for establishing the text’s authority and its efficacy in procuring knowledge, respectively. Sthiramati’s solution to this problem, I

suggested, is multifaceted. First, he suggests that the correct determination of meaning is guaranteed by the exceptional expertise of the persona involved (which serves as an extra-textual criterion for its meaning, authority, and relatedness to reality). This appeal to extra-textual authority, however, is not at odds with the understanding of the text as inherently layered, a feature that in the hands of Sthiramati is identified with the polyphonic quality of ??stric discourse – its ability to speak in multiple “voices” simultaneously. These ‘voices,’ it was demonstrated, are systematically and the unique theory of meaning explicated in Sthiramati’s bh??ya to Vasubandhu’s Trimsika, in which the causal underpinnings of language are developed into something akin to a (non-realist) figurative causal theory of reference. See Tzohar (2017).

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017) 275 hierarchically ordered in accordance to the degree to which they can embody and express the underlying causal and mental reality – degrees outlined by the three natures scheme. Moreover, this framework, insofar it serves to evaluate discourse in terms of its insight into its underlying mental and causal conditions, provides an additional hermeneutical criterion – and this time intra-textual – for determining the meaning and truth of discourse. In this respect,

the framework offered by Sthiramati was shown to strike a certain balance between the need for a decisive interpretative reduction, on the one hand, and the Mahayanic conception of the layerdness of meaning on the other. By incorporating both strands into his understanding of the nature of ??stra, Sthiramati is therefore able to guarantee the kind of interpretative closure necessary for upholding the authority of scripture, while understanding the ingrained ambiguity of the text not as an obstruction of truth but as means toward a transformative end.30

Abbreviations

MV Madhy?ntavibh?ga. [[[Maitreya-Asanga]]]. MVBh Madhy?ntavibh?gabh??ya. [[[Vasubandhu]]]. MVBh? Madhy?ntaVibh?gaBh??ya??k?. [[[Sthiramati]]]. Vy Vy?khy?yukti (rnam par bshad pa'i rigs pa). [[[Vasubandhu]]].

References

Cabezón, Jose Ignacio. 1992. “Vasubandhu's Vy?khy?yukti on the Authenticity of the Mah?y?na S?tras.” In Texts in Context: Traditional Hermeneutics in South Asia, edited by Jeffrey R. Timm, 221- 243. Albany: State University of New York Press. D’Amato, Mario. 2012. Distinguishing the Middle from the Extremes: A Study and Annotated Translation of the Madhy?ntavibh?ga, Along with Its Commentary, the Madhy?ntavibh?ga-Bh??ya. New York: American Institute of Buddhist Studies.

Davidson, Ronald M. 1990. “An Introduction to the Standards of Scriptual Authority in Indian Buddhism.” In Chinese Buddhist Apocrypha, edited by Jr. Robert E. Buswell, 291-326. Honolulu: University of Hawai’i Press.

Dutt, Nalinaksha. 1978. Boddhisattvabh?mi?. Patna: K.P. Jayaswal Research Institute. 1st ed. 1966. Garfield, Jay, and Jonathan Gold. 2011. “On the Reality of the Mind in Yog?c?ra: A Constructive Debate on Vasubandhu’s Trisvabh?vanirde?a.”

http://religion.princeton.edu/events/2011/aarpanel/AAR_Conference_Panel_2011.html. 30 It is transformative in the sense that the polysemy of the text (understood in terms of a polyphony of voices) when mapped unto the three natures scheme, facilitate a progressional understanding of reality. Roy Tzohar – Thoughts on the Early Indian Yog?c?ra Understanding of ?gama- Pram??a

Hayes, Richard P. 1984. “The Question of Doctrinalism in the Buddhist Epistemologists.” Journal of the American Academy of Religion 52.4: 645-70.

Honda, Megumu and Johannes Rahder. 1968. Annotated Translation of the Dasabhumika-Sutra. New Delhi: International Academy of Indian Culture.

kachru, sonam. 2014. “Text, Language, Mind(S):A Footnote on Sthiramati’s Definition of Textual Instruction (??stra) as a Mental Phenomenon (Vijñapti).” Unpublished paper. Kochumuttom, Thomas A. 1982. A Buddhist Doctrine of Experience: A New Translation and Interpretation of the Works of Vasubandhu, the Yog?c?rin. Delhi: Motilal Banarsidass.

Lamotte, Etienne. 1988. 2The Assessment of Textual Interpretation in Buddhism.” In Buddhist Hermeneutics, edited by Donald Lopez Jr., 11–27. Honolulu: Kuroda Institute. Lee, Jong-Cheol. 2001. The Tibetan Text of the Vyakhyayukti of Vasubandhu: Critically Edited from the Cone, Derge, Narthang and Peking Editions, Bibliotheca Indologica Et Buddhologica. Tokyo: Sankibo Press.

Lopez, Donald S. 1988. “On the Interpretation of the Mahayana Sutras.” In Buddhist Hermeneutics, edited by Donald S. Lopez, 47-70. Honolulu: University of Hawaii Press.

Lugli, Ligeia. 2010. “Meaning without Words: The Contrast between Artha and Ruta in the Mah?y?na S?tras.” Buddhist Studies Review 27.2: 139-76.

Nance, Richard F. 2004. “Models of Teaching and the Teaching of Models: Contextualizing Indian Buddhist Commentary.” PhD diss., University of Chicago.

Nance, Richard F. 2012. Speaking for Buddhas: Scriptural Commentary in Indian Buddhism. New York: Columbia University Press.

Sponberg, Alan. 1982. “The Trisvabh?va Doctrine in India and China: A Study of Three Exegetical Models.” Bukky? Bunka Kenky?-jo Kiy? 21: 97–119.

Stanley, Richard. 1988. “A Study of the Madhyantavibhaga-Bh??ya-?ik?.” PhD diss., Australian National University. Tatia, Nathmal. 1976. Abhidharmasamuccayabh??yam. Patna: K.P. Jayaswal Research Institute. Tillemans, Tom J. F. 1999. Scripture, Logic, Language: Essays on Dharmakirti and His Tibetan Successors, Studies in Indian and Tibetan Buddhism. Boston, Mass.: Wisdom Publications.

Tzohar, Roy. 2013. “Vasubandhu.” Oxford Bibliographies in Buddhism, edited by Richard Payne. New York: Oxford University Press. http://www.oxfordbibliographies.com/obo/page/buddhism.

Tzohar, Roy. 2017. “Does Early Yogacara Have a Theory of Meaning? Sthiramati’s Arguments on Metaphor in the Tri??ik?-Bh??ya.” Journal of Indian Philosophy 45.1: 99-120. Tzohar, Roy. forthcoming. Meaning in the World and in Texts: A Buddhist Yog?c?ra Theory of Metaphor. Oxford, New York: Oxford University Press.

Verhagen, Peter C. 2005. “Studies in Indo-Tibetan Buddhist Hermeneutics (4): The Vy?khy?yukti by Vasubandhu.” Journal Asiatique 293.2: 559-602.

Kervan – International Journal of Afro-Asiatic Studies n. 21 (2017)

Wayman, Alex. 1999. A Millenium of Buddhist Logic. 1. Buddhist Tradition Series. Dehli [u.a.]: Motilal Banarsidass.

Yamaguchi, Susumu and Sylvain Lévi. 1934. Madhyantavibhagatika: Exposition Systématique Du Yogacaravijñaptivada. Nagoya: Librairie Hajinkaku. Roy Tzohar (Tel Aviv University) specializes in the history of philosophy with a focus on Buddhist and Brahmanical philosophical and poetical traditions in India. He is currently a tenured professor in the East Asian Studies Department at Tel Aviv University. Roy Tzohar holds a PhD from the Religion

Department at Columbia University (New York, 2011), and an M.A. in philosophy from Tel Aviv University’s Interdisciplinary Program for Outstanding Students (Tel Aviv, 2004). His monograph on Buddhist philosophy of language, Meaning in the World and in Texts: A Buddhist Theory of Metaphor is forthcoming from Oxford University Press.