Speaking for Buddhas (Scriptural Commentary in Indian Buddhism)

Email

Whatsapp

Facebook

Pinterest

Twitter

Copy link

Buddhist intellectual discourse owes its development to a dynamic interplay between primary source materials and subsequent interpretation, yet scholarship on Indian Buddhism has long neglected to privilege one crucial series of texts. Commentaries on Buddhist scriptures, particularly the sutras, offer rich insights into the complex relationship between Buddhist intellectual practices and the norms that inform-and are informed by-them. Evaluating these commentaries in detail for the first time, Richard F. Nance revisits-and rewrites-the critical history of Buddhist thought, including its unique conception of doctrinal transmission.

Attributed to such luminaries as Nagarjuna, Vasubandhu, Dignaga, and Santideva, scriptural commentaries have long played an important role in the monastic and philosophical life of Indian Buddhism. Nance reads these texts against the social and cultural conditions of their making, establishing a solid historical basis for the interpretation of key beliefs and doctrines. He also underscores areas of contention, in which scholars debate what it means to speak for, and as, a Buddha.

Throughout these texts, Buddhist commentators struggle to deduce and characterize the speech of Buddhas and teach others how to convey and interpret its meaning. At the same time, they demonstrate the fundamental dilemma of trying to speak on behalf of Buddhas. Nance also investigates the notion of “right speech” as articulated by Buddhist texts and follows ideas about teaching as imagined through the common figure of a Buddhist preacher. He notes the use of epistemological concepts in scriptural interpretation and the protocols guiding the composition of scriptural commentary, and provides translations of three commentarial guides to better clarify the normative assumptions organizing these works.

Richard F. Nance is assistant professor of South Asian Buddhism in the Department of Religious Studies, Indiana University, Bloomington.

CLASSICAL INDIAN sastras (treatises) regularly enjoin aspiring authors to announce at least two things at the outset of their works, so as to secure the interest of judicious readers. The first of these is what in Sanskrit is termed the abhidheya, or, roughly, the work’s topic: what it’s about. According to Sanskrit commentators, authors should state this explicitly so that their audiences do not presume their texts to be meaningless or incoherent. Additionally, authors are told to make explicit their work’s prayojana: its point. Unless a point is specified up front, a prospective audience is likely to ignore a text, presuming that the text has no point, or that the point it does have is so recondite (or trivial) as to be practically useless. This is good advice, even today. In tribute to the authors whose texts are addressed here, and in the hope of securing the interest of today’s judicious readers, this introduction aims to delineate the abhidheya and the prayojana of the book to follow: what the work is about, and why it matters to the study of Buddhism.

In brief, the work to follow explores some of the ways in which successive generations of Buddhists-more specifically, Indian Buddhists-have transmitted Buddhist teachings through time. For more than a thousand years, over the course of the first millennium C.E., Buddhism flourished in India. The forms this flourishing assumed were, of course, varied; Robert Sharf is correct to note that from the perspective of a modern historian, “the term ‘Buddhism’ turns out to be a site of unremitting contestation, as a cacophony of voices-each averring privileged access to the essence of the tradition-lays claim to its authority"! Yet even in this cacophony, one finds a measure of consonance, for the cacophony itself testifies to a generalized need to appropriate authority via acts of speech. These attempts to speak authoritatively were complex discursive events through which Buddhists aimed to speak not only for themselves, but also for Buddhism more generally-and thus to speak for Buddhas.

This book explores what Indian Buddhist scholastic authors held such speaking to involve. In speaking for Buddhas, Buddhist scholastics were recapitulating an activity of the very Buddhas whose authority they assumed. Indian Buddhist literature portrays Buddhas as speakers par excellence, ceaselessly and patiently consenting to instruct others in the means by which they might be liberated from samsara. As speakers, Buddhas generate-or at least appear to generate-texts; in this respect, they are no different from Buddhist scholastic authors.' The words and sentences of these texts may subsequently be memorized, repeated, chanted, inscribed, collected, categorized, elaborated, debated, and paraphrased. Insofar as they are understood to instantiate "the speech of a Buddha" (buddhavacana), they are also held to manifest the characteristics of what the tradition terms "right speech" (samyagvac). In the Brahmajalasutta, the characteristics of right speech are explicitly and succinctly laid down by (and with reference to) the Buddha Gotama himself:

Abandoning false speech, refraining from false speech, the ascetic Gotama dwells-a truth speaker, one to be relied on, trustworthy, dependable, not a de- ceiver of the world. Abandoning malicious speech, refraining from malicious speech, he does not repeat what he has heard here to the detriment of these, or repeat what he has heard there to the detriment of those. Thus, he is one who reconciles those at variance and one who encourages those in concord. Rejoicing in peace, loving it, delighting in it, he is one who speaks up for peace. Abandoning harsh speech, he refrains from it. He speaks whatever is blameless, pleasing to the ear, agreeable-that which reaches the heart, is urbane, pleasing, and attractive to the multitude. Abandoning idle chatter, he speaks at the right time, correctly, and to the point; he is a speaker of the teaching (dhamma), a speaker of the discipline (vinaya), a speaker whose words are to be treasured, [whose words are relevant, reasoned, well-defined, and connected with the goal.

Here as elsewhere, Buddhist agama portrays a Buddha as one who, having attained perfect awakening, has abandoned four ways of speaking: a Buddha no longer engages in false speech (musavada), malicious speech (pisunavaca), harsh speech (pharusavaca), and idle chatter (samphappalapa). His speech is thus seen to be pure; it is true, promotes concord, is pleasing, and is con- cerned with matters of import.

If "right speech" is to be a component of the Buddhist path, it must describe more than the purified utterances of a Buddha; those who aspire to awakening must likewise be capable-at least in principle-of engaging in it. The agamic literature recognizes this imperative. In the Samannaphalasutta, for example, the previously quoted passage from the Brahmajalasutta is repeated, but the figure of Gotama has vanished. In his place stands an idealized Buddhist monk whose moral conduct has been perfected. Such a monk speaks as a Buddha does." From the intersection of these two passages, a normative conception of the Buddhist monk as speaker begins to take shape. A monk's rhetoric is to be viewed as authoritative precisely insofar as it recapitulates (what is taken to be) the rhetoric of Buddhas. Like a Buddha, then, an ideal monk is a speaker of the teaching and of the discipline. He speaks in a manner that is timely, correct, and to the point; he knows when (and to whom) to speak, what to say, and how to say it. His discourse is, moreover, true, conciliatory, and pleasing.

As Buddhist literature acknowledges, any speech that stands a chance of being counted as true, conciliatory, and pleasing must accord with particular conventions-whether social, institutional, or linguistic. These conventions are held to apply even to Buddhas, insofar as Buddhas are seen to communicate effectively: Buddhist texts repeatedly emphasize the unfailing ability of awakened beings to employ the language and idiom appropriate to the particular circumstances under which they are called upon to teach.' According to a tale preserved in the Mulasarvastivada Vinayavastu, for example, the Buddha Gautama was once approached by four great kings-Dhrtarastra, Virudaka, Virupaksa, and Vaisravana-each hailing from a different region. Two-Dhrtarastra and Virudaka-are identified as aryan, while the other two are identified as dasyu. The two groups are unable to comprehend one another, which prompts the Buddha to reflect: "Were I to teach the dharma using aryan language, two would understand, two would not. Were I to teach the dharma using dasyu language, again, two would understand, two would not. So I should teach the dharma to two using aryan language, and to two using dasyu language." Opting to address the kings separately, the Buddha then teaches each using the particular language appropriate for him.

Comparable assumptions regarding right speech and its proper deployment affected-and were at times explicitly acknowledged by-subsequent Buddhist scholastic authors. These authors consciously endeavored to compose their works in terms appropriate to the audience(s) they sought to address. Such terms were not confined to the broad matter of language choice, but extended also to matters of style: Indian Buddhist scholastics were concerned to address their audience(s) in ways that would be seen as timely, correct, and to the point-ways informed by a network of normative presuppositions concerning what one should rightly say, how one should rightly say it, and when and where such speech is appropriate.

This book excavates some of the nodes in this network. It explores a set of normative protocols and the impact of those protocols on a number of au- thors who labored during the latter part of the first millennium and worked in or around the great monastic complexes (mahavihara) of north India.'? The book focuses principally on Buddhist texts and practices, though it should be kept in mind that such texts and practices did not appear in a vacuum: Indian Buddhism was not simply Buddhist; it was also Indian, and Buddhist authors of the late first millennium clearly drew on their understanding of non-Buddhist texts and practices when they composed their own works. A sustained exploration of these non-Buddhist sources and their impacts is a desideratum, but it is a project for another time. The ambition of the present work is more limited: I aim to sketch a preliminary map of the territory, in order to call attention to normative aspects of Indian Buddhist textual production that have, as yet, been largely ignored, and also to a huge and neglected corpus of Indian Buddhist literature: commentaries on Buddhist sutra texts.

1. Indian Buddhist Sutra Commentaries Commentary has long been recognized as the preeminent vehicle for the dispensation of knowledge in India." Throughout the first millennium C.E., learned Indian Buddhists appear to have held the study of scriptural com- mentaries to be an important resource for understanding the speech of Buddhas, and the composition of sutra commentaries to be a fitting means for the communication of their understanding.

The extent to which scriptural commentaries were valued by Indian Buddhist intellectuals of the late first millennium is evidenced by the extant Tibetan imperial-period textual catalogues (dkar chag)-texts that afford us a window into Buddhist learning as it was transmitted to Tibet during the so-called early propagation period (i.e., the snga 'dar, usually dated from the seventh through the early ninth centuries)." Two such catalogues are extant. The earliest and best known of these is the Catalogue of Lhankar (Lhan kar ma dkar chag; hereafter Catalogue). Compiled early in the ninth century following the reign of the Tibetan king Khri srong lde'u btsan, the Catalogue inventories a collection of texts housed at the palace of Ldan/Lhan (d)kar. It counts and categorizes more than seven hundred texts translated from Indian and Chinese sources. These translations were typically collaborative undertakings in which Tibetan translators were assisted by learned representatives of Indian and Central Asian Buddhist traditions. The texts inventoried in the Catalogue were clearly valorized by those who translated them, and it is not unreasonable to assume that this valorization was learned from non-Tibetan Buddhist scholastics. If this assumption is correct, then the Catalogue can be seen as casting light-albeit indirect light-on Indian and Central Asian Buddhist scholastic presuppositions current during the eighth century.

With very few exceptions, the Catalogue inventories each of its texts as follows. The text is categorized under one of thirty textual categories; its title-and sometimes its author-is noted; and its length is then given (this length is measured in terms of slokas-here referring to a textual unit of sixteen syllables of verse or prose-and Tibetan bam po; the latter is an unstable unit of measurement but typically approximates 300 slokas). Some of the information contained in the Catalogue is summarized in table 0.1. The left-hand column lists the thirty categories in the order in which they are given in Herrmann-Pfandt's edition of the Catalogue. The central column gives the number of discrete texts categorized under each heading, and the right-hand column provides an approximate total aggregated length (in bam po) for the texts attested in each category.

The categories that structure the Catalogue's inventory are somewhat Borgesian, and they sometimes make it difficult to judge the nature of the particular texts that they include and exclude. What seems clear, however, is that not all categories commanded equal attention. To judge from the sheer bulk of the translated material (see the third column of table 0.1), the translation teams of this period appear to have devoted most of their energies to translating what they marked as three kinds of text: Buddhist sutras-particularly those identified as Mahayana; disciplinary literature (vinaya); and commentaries on Mahayana sutras.

The attention lavished on sutra and vinaya texts is hardly surprising. Sutra texts were held to constitute the speech of a Buddha-utterances of paramount authority-and the translation of Buddhist vinaya literature was likely to have been seen as a sine qua non for the establishment of Buddhist monasticism in Tibet. The considerable attention that the translation teams granted to Mahayana sutra commentaries may, however, come as a bit of a shock. Modern scholars who study Indian Buddhism scholastics appear to have put them in a privileged position.

If we hope to gain a more nuanced understanding of the complexity of Indian Buddhist tradition, and the impacts on doctrine wrought by successive contributions to that tradition, we would do well to investigate these works. We are, however, ill-equipped to read them, and approaching them for the first time can be a disorienting and frustrating experience. Buddhist sutra commentaries were composed according to rhetorical protocols that we do not share. In contrast to the smooth narrative flow of a sutra text, the commentaries contain a jumble of statements, phrases, and single terms, accompanied by lengthy digressions on issues that may appear only peripherally related to the sutra under discussion. If we are to begin to under- stand Buddhist scriptural commentaries, we need to ask what normative conventions shaped-and were shaped by-their production. The present work charts these conventions, at least in part, and aims thereby to encourage further study of this neglected body of literature.

2. Normativity and Positivist Presuppositions
in the Study of Indian Buddhism
Normative conventions are not confined to the authors whose texts this book will discuss-and, in fact, the concern with the normative evident throughout this book may well be seen to violate a normative convention that characterizes much of current Buddhist Studies scholarship. According to this convention, to concern oneself with the normative dimensions of texts is to capitulate to a model of scholarship whose problems have been extensively charted over the past two decades, a model that privileges doctrine over practice, the ideal over the actual, and what one finds in texts over what one finds on (or under) the ground. Gregory Schopen, Todd Lewis, Donald Lopez, and others have repeatedly reminded us of the scholarly con- fusion this model has generated. The present work neither represents nor recommends a return to it.

The terms in which the model has been critiqued were largely set by Schopen's probing and seminal 1991 paper titled "Archaeology and Protestant Presuppositions in the Study of Indian Buddhism." This paper has attained the status of a classic. It manifests exceptional philological acumen, and its argument is presented with tremendous rhetorical skill. Yet the argument is flawed, and flawed in instructive ways.

Schopen begins by distinguishing between two bodies of data: archaeo- logical/ epigraphic material (hereafter "inscriptions"), on the one hand, and texts, on the other. Unlike the corpus of texts, the corpus of inscriptions "records or reflects at least a part of what Buddhists-both lay people and monks-actually practiced and believed." Inscriptions, in other words, are typically descriptive as opposed to normative. The distinction between the descriptive and the normative can be thought of as a distinction that marks how particular accounts are used and understood by those who make use of them. Descriptive accounts are those that concern (or are taken to concern) what has been, what is, or what will be so. To the extent that an account fails to do this, it will have failed as a descriptive account. On the other hand, normative accounts are those that concern (or are taken to concern) what should be so. Normative accounts are used prescriptively and proscriptively; their value lies in presenting idealized paradigms-models-to guide belief and action, regardless of whether those models ever conform to what is so.

We need not draw a hard and fast line between the descriptive and the normative. It is true that every successful descriptive account instantiates- even if it does not itself describe-certain norms. Likewise, every normative account is, insofar as it is an account, descriptive in at least some sense. But even if the line between the normative and descriptive is not always perfectly clear, it is clear enough to allow us to apprehend a crucial point: the normative and the descriptive rarely coincide perfectly. What happens "on the ground" can diverge considerably from what is held to be worth happening.

Schopen proceeds, then, by delineating two dichotomies-inscriptions vs. texts, on the one hand, and the descriptive vs. the normative, on the other. He then collapses the two dichotomies into one another: inscriptions describe, whereas texts enshrine norms. This collapse enables Schopen to leverage a claim against antecedent Buddhist Studies scholarship: scholars of Buddhism, relying almost exclusively on texts, have failed to recognize the normative nature of their sources. As a result, they-we-have too often unwittingly constructed normative accounts, rather than descriptive accounts, of what Indian Buddhism was. We have propounded doctrine and called it history.

Having diagnosed this scholarly malaise, Schopen proceeds to trace its etiology: we have proceeded in this way because we have inherited a certain set of unquestioned "Protestant presuppositions" about where "real religion" is to be found. This, in Schopen's view, is a mistake. But it is important to be clear about the nature of this mistake: when Schopen tells us that "there is a remarkable similarity between the value assigned literary sources in modern historical and archeological studies and the argument of Protestant reformers concerning the location of true religion," he is not thereby questioning the idea that something called "true religion" may exist. Rather, he's implying that if it does exist, we will not find it by looking at texts. To think otherwise is to embrace a Protestant presupposition. The solution is simple: we should stop relying exclusively on texts and start looking carefully at inscriptions. Only then will we begin to discover "what actual Buddhists did and believed," and hence, in his view, only then will we recover what Indian Buddhism truly was.

"Archaeology and Protestant Presuppositions in the Study of Indian Buddhism" can be (and should be) taken as a corrective to a one-sided model of Buddhist Studies that opts to focus exclusively on texts while ignoring other available data. Read in this way, Schopen is encouraging us to broaden the body of data from which we draw our conclusions-a suggestion that is eminently reasonable. However, the paper can also be read as implying that if one happens to be interested in what Buddhism was, as opposed to what certain Buddhists have thought Buddhism ought to be, one need not bother with the investigation of texts. This was almost certainly not Schopen's intention in composing the essay; indeed, to judge from his subsequent extensive-and superb-work on the Mulasarvastivada Vinayavastu, it is not a reading that he would endorse. But it is not difficult to find traces of such a reading in recent Buddhist Studies scholarship. Consider, for example, the following programmatic passage, drawn from the preface to Martin Mills's 2003 volume on Gelukpa monasticism in Tibet:

I have attempted to build a picture of Tibetan Buddhist life that begins by asking what it itself is, rather than what it should be. In particular, I have attempted to move away from a logocentric approach which either sees Tibetan Buddhism as defined by explicit written teachings, or as centered around them. Rather ... I have sought to examine the practice of Tibetan Buddhism in a particular con- text ... in order to see what' it can tell us about how we should interpret the intellectual content of Tibetan Buddhist texts.

Here, again, two dichotomies are delineated. The normative is opposed to the descriptive, and texts (and scholarly studies that focus principally on texts) are opposed to practices (and scholarly studies that focus principally on practices). Once delineated, the dichotomies are again collapsed: if we want to know what Tibetan Buddhism is, then we should focus on practices. Mills elsewhere notes that he sees a text-based approach to be inadequate because such an approach derives from "a thinly-veiled post-Reformation bias against ritual and clericalism that is arguably inappropriate to the study of non-theistic religious traditions." Like Schopen, then, Mills sees Protestant biases at work in Buddhist Studies. Moreover, like Schopen, Mills has not discounted the heuristic utility of postulating (or the possibility of locating) "true Buddhism": he has simply shifted its location from texts to practices.

I have emphasized the way in which both Schopen and Mills attempt to read the distinction between the normative and the descriptive as tracking the distinction between bodies of data in order to encourage further reflection on whether the collapse of such dichotomies is legitimate. We might ask, for example, whether it is true that inscriptions do not express, inculcate, or both express and inculcate normative ideals. This seems a problematic assumption, given that inscriptions are typically formulaic (a fact pointed out by Gerard Fussman in an article that Schopen himself cites approvingly): i.e., they are influenced by and reinforce normative assumptions as to proper form and content." Moreover, there are well-known inscriptions (e.g., Asoka's fourth pillar edict) that are clearly prescriptive in nature. We might also ask whether it is true that Buddhist texts do not generally express what Buddhists really practiced or believed. This is surely true in some cases, yet it seems to me a mistake to draw from these cases the sweeping conclusion that Buddhist texts are constitutively normative (though they undoubtedly contain many claims that do function normatively). The Buddhist textual corpus is more heterogeneous than Schopen suggests; some of the ideas he associates with the "inscriptional story"-a story he reads as descriptive-can also be found articulated in Buddhist sutra texts." The upshot is clear: it is a mistake to think that these two dichotomies-inscriptions vs. texts and descriptive data vs. normative data- can be cleanly collapsed. We should not assume that Buddhist inscriptions are constitutively descriptive, nor should we assume that Buddhist texts are constitutively normative.

It may still be true, of course, that in looking principally at texts to figure out what Buddhism has been, scholars have been gazing through a lens of presuppositions that is serving to distort our picture of where pertinent information for "true religion"-in this case, "true Buddhism"-is rightly to be found. If we assume that the notion of "true religion" is not itself a manifestation of the very presuppositions that Schopen aims to critique, we might wonder where we should look for a more appropriate heuristic.

Schopen notes that "our picture of Indian Buddhism may reflect more of our own religious history and values than the history and values of Indian Buddhism," This suggests that he would hold Buddhist presuppositions regarding the location of "true Indian Buddhism" to be an improvement over the Protestant presuppositions he excoriates. When we turn to Buddhist sources in search of these presuppositions, however, we find that the sources routinely direct us to texts and textual practices: practices of memorization, reflection, discussion, debate, teaching, preaching, and so on. The Chinese pilgrim Xuanzang (600?-664 C.E.), who visited many Buddhist monasteries during his lengthy sojourn to India, explicitly notes a range of material benefits to be gained as a consequence of demonstrating familiarity with the content of Buddhist texts. Although the sources that direct us to texts are often textual themselves, they are not always so: the Nalanda stone inscription of Yasovarm(m)adeva, for example, describes the residents of Nalanda as "scholars well known for their [knowledge of the] sacred texts and arts" (sadagamakalavikhyatavidvajjana).

These data suggest that although Schopen and Mills may well be correct in tracing the origins of many of the presuppositions that inform Buddhist Studies scholarship to Protestantism, similar presuppositions were held by at least some Indian Buddhists. The whole issue of "Protestantism" may, in this specific context, turn out to be something of a red herring. To dismiss the careful study of Buddhist texts as evidence that one is in the grip of a "Protestant presupposition" is to do a disservice to the complexity of the Indian Buddhist tradition-a tradition in which practices aimed at the meticulous exegesis and analysis of claims existed alongside (and, at times, were thoroughly interwoven with) practices of devotion, sorcery, and ritual.

3. The Structure of the Book
I have organized this study so as to facilitate its use by scholars interested in issues that extend beyond the relatively narrow domain of Indian Buddhist sutra commentary. The book's final chapters trace normative assumptions that pertain specifically to commentarial texts, but this point is reached only after an extended exploration of the wider normative context within which these works-and many others-were composed. The book thus commences by taking a wide-angle approach to context; as it proceeds, this context narrows.

Buddhist commentaries on scripture were-and were seen by their authors to be-paradigmatically (though not exclusively) linguistic utterances. These utterances were also pedagogical: like the texts on which they commented, commentarial texts constituted, and were taken to constitute, teachings. We may thus hypothesize that commentaries were informed by specific (if sometimes implicit) normative assumptions concerning correct speaking and correct teaching. These assumptions are explored in the first two chapters of the book.

Chapter 1 begins by investigating the notion of "right speech" articulated in Buddhist texts. Two normative models are discussed. The first is prescriptive (it provides a guide to how, what, when, and where one should speak); the second, proscriptive (it provides a guide to how, what, when, and where one should not speak). The two models are developed via close reading of portions of texts likely to have been well known to Buddhist commentators working in large Indian monastic complexes during the mid-to- late first millennium C.E.: Matrceta's Satapancasatka and Catuhsataka, and the (Mulasarvastivada) Pratimoksasutra. These texts appear to have been among the first memorized by Buddhist monastic novices, and they likely served as resources from which monks consciously or unconsciously drew many of their assumptions regarding proper modes of speech.

Chapter 2 focuses on the activity of teaching as it has been portrayed in Buddhist texts and as it has been undertaken by those who have aimed to transmit those texts authoritatively. The chapter is principally concerned with investigating the normative and historical dimensions of a figure regularly associated with the transmission of Buddhist teaching from generation to generation: the Buddhist preacher. In the first of two sections, I examine portraits of this figure sketched in Buddhist sutras and sastras. These portraits regularly explain the preacher's pedagogical skill in terms of four discriminations, which I investigate in some detail. The second section of the chapter aims to show the ways in which idealized portraits of teachers painted in texts may have conformed-and sometimes failed to conform-to the practices of those outside those texts who would shape and be shaped by them.

In chapters 3 and 4, I turn to investigate issues of textual interpretation and composition. What specific resources were acknowledged by Indian Buddhists of the mid-to-late first millennium C.E. to be tools by which the meaning of sutra texts could be comprehended and communicated? Chapter 3 investigates epistemological discourse and asks whether and how the formalized tools for argument that this discourse provided were presumed to aid in interpretive work. Chapter 4 looks specifically at protocols guiding the composition of scriptural commentary, as set down explicitly in what I term "commentarial guides." These guides provide explicit guidelines for how one should undertake to comment on scriptural texts. The chapter focuses especially on Vasubandhu's Vyakhyayukti, comparing it to other extant guides, including the *Vivaranasamgrahani, attributed to Asanga, and a short section of the Abhidharmasamuccayabhasya, sometimes attributed to Sthiramati. To facilitate further study of these texts, reference translations are provided in the book's three appendixes.

The conclusion of the volume begins by recapitulating the findings of the preceding chapters. These findings are then further elaborated and deepened by briefly examining a final exemplary commentary. Through this examination, and through the work as a whole, I hope to illuminate traditional responses to a question that I would argue is crucial for anyone who endeavors to understand Indian Buddhist commentaries and those who composed them: what is it to speak for Buddhas?