The many voices of the Mahabharata.

Fitzgerald, James L.

INTRODUCTION

Alf Hiltebeitel has consistently been the single most open-minded and fearlessly imaginative Western reader the authors and editors of the Sanskrit Mahabharata (MBh) have ever had for their masterpiece. He renews his claim to that distinction with this book. In The Ritual of Battle (1) twenty-five years ago Hiltebeitel was "rethinking the Mahabharata" in terms of then recent (Wikander, Dumezil, et al.) and contemporary (Biardeau) discussions of the newly edited Sanskrit epic, and he has been doing so ever since. A long-time advocate of the intentional unity of the MBh (more or less the Pune text) and, especially, a fierce defender of the importance of the divine Krsna and Krsna-bhakti in that text, he has given us a wealth of insights into the Sanskrit MBh in a long series of articles over the past three decades. Hiltebeitel has also carried out groundbreaking cultural-anthropological investigation of the wider Mahabharata tradition (his work on the Draupadi cult Mahabharata in Tamil Nadu (2)) and has studied and reflected upon all of India's major oral epic traditions. (3)

This newest work, Hiltebeitel's second book focused exclusively on the Sanskrit MBh, offers a number of new ideas and interpretive leads that will play a role in future discussions of the MBh, but perhaps its single most important contribution is its theory of the MBh text and its composition. Hiltebeitel sketches here a unified theory of the MBh's creation and literary character that directly argues for the artistic unity of the MBh more forcibly than anyone since Joseph Dahlmann a century ago. As was Dahlmann's work, this book is, in part, a gauntlet thrown down before the "excavationists," (the "analysts" of Dahlmann's era), who, Hiltebeitel believes, approach the MBh with "tired generalizations" (4) and have produced several "learned misreadings" (5) but few reliable results in understanding it. (6) Though Dahlmann was stubbornly wrong about the historical facts, his "synthesizing" scholarship insisted upon principles worthy of high priority in the interpretation of the MBh--the presumption that the parts of the text were connected intelligently (7) and that the whole of the text made a meaningful contribution to its time and place--and those principles have had a significant legacy across the last century. Hiltebeitel has learned more, it seems, from Madeleine Biardeau, the twentieth century's most powerful exponent of Dahlmann's basic legacy, than from anyone else, and like Biardeau, Hiltebeitel is far more sophisticated and interesting than Dahlmann was. I do not find Hiltebeitel's view of the MBh's composition persuasive, but his laying out of his views on the matter is a valuable contribution to advancing our general discussion of the text. A presentation and discussion of the central aspects of this theory will form the major portion of this review article.

Though Hiltebeitel styles this work "A Reader's Guide to the Education of the Dharma King," that is, Yudhisthira, the book actually spends most of its time and energy constructing the internal images of the text's authoritative voices against a new set of reflections by Hiltebeitel on the text's actual authorship. (8) Obviously such matters are relevant to a depiction of a prince's or king's education, and the book does carry out an interesting and innovative tracing of part of Yudhisthira's epic figure. But the book's actual center of gravity is its theory of the epic, and every other argument advanced in the book depends upon its theory of the epic text. This review will first look at the book and its structure as a whole and comment upon a few of the larger themes Hiltebeitel takes up in it. I will then turn to his theory of the Mahabharata and the Vyasa-author fiction he sees in it.

A BIRD'S-EYE VIEW OF THE BOOK

Hiltebeitel provides a precis of the book in his introductory first chapter, and that is a good place from which to view the book as a whole. "Chapter by chapter, the book takes up a series of enigmas: empire, author, transmission, the stars, noncruelty, love, the wife, and writing. Threaded through them, but especially chapters 2 to 7, is the education of king Yudhisthira.... For the Mahabharata is Yudhisthira's education. And woven throughout is the question of the author, Vyasa, the ever-receding figure behind this hero's education and the ultimate enigma of his own text. These topics, laced through this book as they are through the epic itself, serve to invite a reading 'against the grain'--and, to be precise, against grains that are as much the product of the Mahabharata's own craft as they are the results of its learned misreadings" (p. 4). With the exception of writing, which is not really addressed or discussed in the book in a sustained way, these seven topics do loosely organize the progression of the whole.

But probing these enigmas is not as simple as it may sound. The book is actually a set of thirty-one essays grouped into eight chapters that (in the seven post-introductory chapters) roughly parallel the epic's movement from the first utterances of the outermost voice of its ultimate authorial agency (9) through the frame-story of the bard Ugrasravas' recitation before brahmins doing a sattra in the Naimisa Forest, (10) down to the story of the Pandavas and Kauravas told by Vyasa's pupil Vaisampayana to King Janamejaya at that king's genocidal sattra to eliminate all the world's snakes. In broad parallel to the way that the MBh itself presents its basic authorial information early (and parallel to the more fundamental reality that any author's voice necessarily stands unframed outside the text it speaks), Hiltebeitel makes reflections upon the epic's authorship the subject of chapter two; and just as the MBh presents much of the critical information about Vyasa toward the back of the MBh (in the Moksadharmaparvan, especially in the Suka stories in 12.310-20 and in the immediately subsequent Narayaniya), Hiltebeitel closes his book by mining the story of Suka for indications of some of the transcendent-author themes he believes the creators of the MBh wrote into their fictional seer Vyasa. Having thus framed his own exposition within discussion of Vyasa's authorship, Hiltebeitel next traverses the subjects of the construction and background of the MBh's Naimisa Forest frame (along with some discussion of the frame defined by Janamejaya's sarpasattra) in chapters three and four, and he develops a number of interesting new ideas as he goes. Finally, after treating the epic's Naimisa Forest frame and the sarpasattra frame contained within it, Hiltebeitel turns, in chapters five, six, and seven, to the "main story" of the epic and pursues the theme of "The Education of the Dharma King": Vyasa's series of existential challenges and lessons directed at Yudhisthira. The "author-teacher" is sometimes cruel to his pupil, who is charged with a cruel task (Yudhisthira was formerly the god Indra, as were the other sons of Pandu, and he has been born on earth as the son of Dharma-Yama to bring about the slaughter of the ksatra (11)), and the fundamental lesson Yudhisthira must learn in the context of these cruelties is kindness (anrsamsya) or 'noncruelty', as Hiltebeitel prefers to gloss it.

Hiltebeitel's examination of the main story of the MBh through this lens of "The Education of the Dharma King" presents three series of interrelated essays (twelve in all). Among a number of other topics he traces the word anrsamsya (non-cruelty), which he says pertains to delimited contexts, and contrasts it with (the ambiguously universal) ahimsa; lays out the "reflections and enigmas" posed to Yudhisthira (and all readers of the epic) by the story of Nala ("Nala is perhaps the exemplary subtale ... 'encapsulating' the epic narratively as the Gita does theologically" [p. 216]); and probes the mysteries and questions surrounding Yudhisthira's mysterious gambling away of his wife, the mysteries surrounding Yudhisthira's marital relationship with Draupadi, and finally, where the epic itself ends, the tests put to him by his father Dharma.

Overall, these essays form less a "reader's guide" to a clear thesis about the MBh than a set of commentaries in progress, moving through the passages as Hiltebeitel marshals them, tenaciously building support for the author's insights as the author sifts etymologies and translations of particular words, seizes upon nuances in particular turns of phrase, infers the underlying assumptions of a passage by reading it in the light of passages far removed within the epic, insisting often that the text's solutions of surface continuity were deliberate artifice, much as they are in some contemporary prose. These commentaries are sensitive to the same kinds of existential, affective, and symbolic issues raised by such scholars as Wendy Doniger and David Shulman--rereading, in fact, some of the same episodes they have treated--and they are deeply influenced by Madeleine Biardeau's readings of the epic. These complex essays overflow with ideas and they often repay the close reading they require. But at the same time, Hiltebeitel's arguments are often intricate and subtle, and at times they seem to sprawl across the pages while the author is too often understated and oblique.

THE EDUCATION OF YUDHISTHIRA

Hiltebeitel's reflections upon Yudhisthira, his role in the epic, and his relations to Draupadi tap into especially rich epic veins and join his voice to mine and others that have claimed Yudhisthira is more central to the epic than Arjuna. (12) On the other hand, the claim that "the Mahabharata is Yudhisthira's education" is so strong and so fundamental to the argument of this book that some systematic and critical discussion of it is called for. The notion is a useful conceit for examining the epic, and it is obviously an accurate description of certain elements of the epic that are critically important to its overall construction (e.g., the episode I call "The Persuasion of Yudhisthira" [12.1-38], Dharma's disguised testing and applauding Yudhisthira three times, and Bhisma's long instructions of Yudhisthira in books twelve and thirteen). The idea can also embrace some of what are, at the least, the significant plot devices of Vyasa's (and others') frequent appearances to offer advice or make suggestions to Yudhisthira, and it can also embrace the rich opportunities for reflection upon the main narrative that occur when Yudhisthira is told such instructive stories as the Nalopakhyana, the Ramopakhyana, and Savitryupakhyana. But construing the whole epic as Vyasa's (and the divine Krsna's) education of Yudhisthira requires a fuller statement of who the epic supposes Yudhisthira to be and what sort of "education" he needs than Hiltebeitel gives us. Yudhisthira pursues dominion through warfare for brahminic and divine purposes, and he is to be some kind of "king." The formation and stabilization of his character and intellect are critically important to this action and this role. Hiltebeitel says he sees Yudhisthira's kingship as important, and he does begin to address this aspect of him in considering "empire and epic" in the introduction and in his entree into the main story in chapter five. But after Hiltebeitel notes that epic and puranic representations of political history make the category of "ksatriya" problematic in a fundamental way, he moves on too quickly to a focus upon the insanely genocidal aspects of the MBh's representation of its anti-ksatriya violence. That is, he moves away from the politics of the MBh too quickly to the universal human issues of violence (and non-violence) and kindness and then onto the conundrum of Yudhisthira's failing fully to possess himself in his wagering of Draupadi. While this probe of the Dharmaraja's soul bears some valuable fruit in its consideration of the relation of Yudhisthira and Draupadi, it fails to pay sufficient attention to the particular cruelties of war and rule that the epic poets impose upon their Yudhisthira.

THE MAHABHARATA AS FICTION

Some of Hiltebeitel's recent reflections (13) have helped us to focus upon the characters of the MBh as artistic constructions, and one of the contributions of the current book is Hiltebeitel's alert use of recent arguments exploring the idea of "fiction" as a possible authorial intention in the composition of some ancient literature and of the MBh in particular. Inspired by Michel Foucault, (14) J. L. Mehta (15) (who was inspired by Jacques Derrida), and Mikhail Bakhtin, (16) and relying somewhat on Annette Mangels, (17) Hiltebeitel's reflections in chapter two on the genuine mysteries and powers of authorship as these bear upon the MBh are mostly welcome trains of thought and many of these reflections are helpful in thinking about the MBh. Hiltebeitel is partly right to argue that Mahabharata criticism should shift its focus from oral epic theory and mythology to theories of fiction and self-conscious authorship. But Hiltebeitel's dismissal of "oral epic theory" (my italics) is curious and is more a matter of his relatively circumscribed use of the term "epic" than a denial of prior oral precursors of the MBh (see below). The MBh has a number of different discourse styles within it, some of which are doing more than merely transmitting received tales or wisdom in the fashion of oral tradition. (18) It is not a simple matter of either "fiction" or "oral epic." Hiltebeitel overgeneralizes the role of self-conscious fiction-making behind the MBh and bypasses all discussion of the quite different narrative and discursive styles and textures in the epic. It could be fruitful to approach every aspect of the text as being, possibly, a contingent invention designed for some specific artistic purpose, but not all elements of the MBh were designed with the same degree of artistic purpose and freedom; and at the other end of the spectrum, some portions of the text (such as the Moksadharmaparvan) seem clearly to have been used as convenient containers for the preservation and transmission of texts and passages deemed important independently of and prior to their introduction into the MBh. (19) Some parts of the text Hiltebeitel examines closely in this book (e.g., Nala, Dharma's tests of Yudhisthira, the Suka story at 12.310-20) are good examples of passages that do exhibit an inventive freedom suggestive of "fiction."

THE VEDA IN THE MAHABHARATA

Quoting Madeleine Biardeau's view that the epic utilizes Vedic texts in a way that is "at once both very free and very savant" (p. 131, n. 1), Hiltebeitel declares at the beginning of chapter four, "There is no end to the maze one could trace between the Veda and the Mahabharata, or more particularly between Vedic precedent and epic sattras" (p. 131). If the MBh was composed at basically a single point in time sometime after the demise of Mauryans, as a work of fiction (as Hiltebeitel argues, see below), then any use or representation it makes of past traditions, Vedic traditions in particular, is a deliberately constructed element of its artistic design. "The frame stories [of the MBh] link with other conventions and allusions to make the whole appear Vedic" (p. 131, n. 2; italics of AH) is a statement that expresses well one of the fruits of this valuable line of inquiry. (20) Hiltebeitel pursues various themes in the epic utilization of the Vedic past at length and to some interesting results, though what "Vedic" means is often not sufficiently clear, and at times this reader had trouble following Hiltebeitel's threads through some of the twists and turns of the epic-Vedic maze. (21) In a good extension of Jan Heestermann's, Harry Falk's, and Christopher Minkowski's work on MBh sattras (22) and Hans-Georg Turstig's work on abhicara rites, (23) Hiltebeitel charts and discusses the occurrence in the MBh of various non-philosophical themes originating in Vedic literature, in particular that of sattras and abhicara rites ("black magic" intended to wreak harm upon an enemy). He also makes good use of some of Michael Witzel's work mapping the Vedic skies, places, peoples, and political trends in Vedic texts. (24) Some recent works on the MBh have paid close attention to the geography and cosmography represented in the epic, (25) and Hiltebeitel continues and develops this important and fruitful trend.

WHAT IS THE MAHABHARATA AND HOW SHOULD WE READ IT?

Hiltebeitel's View of the Text and Its Composition

Hiltebeitel's introductory first chapter situates his work within the general ambit of MBh scholarship and then moves to his views of the historical formation of the MBh. Citing David Quint's Epic and Empire: Politics and Generic Form from Virgil to Milton, Hiltebeitel states, though without any discussion, that "epic" is a genre that articulates "imperial visions" of rule. He proceeds quickly to the idea that the MBh reflects the imperial themes pointed out by Michael Witzel (26) as being in some of the latest and most easterly Vedic texts, and its main narratives seem to be "creations of Brahman poets who developed their own variants on the epic genre, centering their stories on kingdoms that had been eclipsed by the time of Magadha's metropolitan states and celebrating these former kingdoms as empires transposed into a deep and glorious double past that is simply beyond the reach of the historian" (p. 7). These are interesting ideas, but without more details regarding "epic" and "empire" (27) than he offers here or at the beginning of chapter five, it is hard to understand what these "visions" are supposed to amount to, or how they relate to some of the text's other stated political concerns--mainly focused in Yudhisthira's rule--which often seem to be firmly rooted in the affairs and perspectives of local kingdoms. As I have already mentioned, Hiltebeitel diverts his gaze from Yudhisthira the king to focus upon Yudhisthira's relations to his father, the god Dharma, and to his wife, Draupadi. Hiltebeitel alludes often enough to Biardeau's and his own sense that the MBh truly is about the incarnation of the supreme god in Krsna Vasudeva at a juncture of yugas, but for the most part these politically important themes remain the concealed dark matter of this book.

Hiltebeitel says these epic projections of empire onto the ancient Kuru kingdom occurred as a written effort on the part of a group of "'out of sorts' Brahmans" (p. 19) for whom "the rise of the imperial Mauryas [was] a negatively evaluated historical backdrop" (p. 17). Hiltebeitel gives us more particulars on these brahmins--if not their politics. "The inner core [of the composing committee, jlf] ... would no doubt have had a philosopher and a dharmasastra connoisseur among them, and perhaps a retired Brahman general (a senapati), while the master of the house [the grhapati of sattras, jlf] kept them all to a common purpose" (p. 169). Having thus laid the theoretical foundation for explaining much of the surface heterogeneity of the MBh, he immediately goes on to characterize their style of composition in a way that purports to account for many of the solutions of continuity on the surface of the text. "The design with its parvans ('joins, knots') and subparvans suggests that different units would have been 'joined' together, leaving awkward fits and provocative contradictions among the endless riches to ponder" (p. 169).

Hiltebeitel reconstructs the basic image of this symposium from the epic itself, by viewing the Naimisa Forest setting of the bard's retelling of Vyasa's tale as a self-portrait of the composing committee. That is, the actual authorial committee saw themselves, or at least fictively represented themselves, as brahmins carrying out a sattra in the Naimisa Forest, "a Naimisa Forest that seems to be celestial" (p. 285), doing what such brahmins typically did in the intervals of their rites, listening to marvelous stories that collapse time and space and action into 'single moments' (nimisa-s, the etymological base of 'Naimisa'). (28) "The 'bard' and all the others who figure in the epic's three frames (29) are fictions of the text: fictions, let me propose, of real Brahman authors who must have enjoyed creating them in some complex image of themselves" (p. 101). "No doubt the bard represents oral tradition--in particular narrative skill and ... perhaps even a gift for improvisation" (p. 169). Finally Hiltebeitel, following a suggestion of Biardeau's, believes the women of these poets were within earshot and "probably sometimes listened well" (pp. 166-67) and exerted an indirect influence on the text, for "so many of the epic's women, and in particular Draupadi, come to life" (p. 167).

Hiltebeitel believes these poets finished their work within a relatively short time span ("at most through a couple of generations" [p. 20]), between "the mid-second century B.C. and the year zero" (p. 18), and he believes that the particular kind of sattra in which the Naimisa brahmins were depicted was the twelve-year sitting kind intended to "set the world in motion" (pp. 157-59). Very interestingly, Hiltebeitel sees this compositional committee as operating in parallel to the authors of some Mahayana Buddhist texts. Quoting Donald Lopez, (30) Hiltebeitel writes, "[A]s Lopez observes, one purpose of [some] Mahayana sutras is to 'wrest ... authority' from multiple 'Hinayana' texts and arhats, 'and restore it solely to the absent Buddha.' In the Mahabharata, however, authority remains in a multisigniferous Veda"--relocated, however, into a "'fifth Veda' that is its author's 'entire thought,' never spoken by him, but only transmitted by others" (p. 169). Regarding one of their motivations Hiltebeitel writes, "... it would appear that the Mahabharata is designed to set in motion the entire history of the universe down to and just past the Mahabharata war, to the 'dawn' (samdhya) of 'our times,' the Kali yuga" (p. 158). And furthermore, reiterating the profound claim about the construction of this vast text above, Hiltebeitel states that the poets' motivations were "extraordinarily subtle" and that they combined "bold instructive teachings with a delight in concealment; [were] not averse to rough joins, repetitions and reiterations, multiple and deepening causalities, overdeterminations, and intriguing contradictions ..." (p. 164).

This work was not a reworking of oral epic traditions, nor did it grow appreciably once it had come into being. But though Hiltebeitel seems to recognize prior oral traditions of (non epic) "Mahabharata" that have some kind of presence in the written epic, it is only "a written dynamism ... that must explain this epic's composition both in its inception and in its northern and southern redactorial variations" (p. 23). And this written epic does not require "two or more written redactions, an erasure by a 'normative redaction,' or the Guptas to account for the history, diversity, and complexity of what I would rather simply call a single written archetype" (p. 26), though "one must grant that there are indeed major and minor interpolations into the text, and also probably losses from it" (p. 28). Knowing full well how much "against the grain" (p. 4) this view is, Hiltebeitel adds, "The real challenges will continue to come from those who find one or another reason to argue that some portion or passage within the Critical Edition is late, such as the highly devotional Narayaniya ... or the entire Anusasana Parvan. Maybe so. But since no one is close to proving anything, let us be all the more cautious about what we try to disprove. I would only argue that even these axiomatically late portions must be looked at with an eye fresh to the possibility that they are not any later--or at least much later: hours, weeks, or months rather than centuries--than the rest, once the rest, and its principles of composition and design, are better understood" (pp. 28-29).

But the Mahabharata Cannot Be Resolved into One Nimesa (31)

So, according to Hiltebeitel, virtually the whole of our MBh was composed by a deliberately interdisciplinary committee of brahmins in a matter of a few years between thirty and 180 years of the fall of the Mauryans. These brahmins were negatively disposed toward the Mauryans and had high ambitions to launch a new ideological beginning under an aegis of Vedic authority. They composed their grand narrative with "loose joins" and in a deliberately baffling way.

I am in general agreement with Hiltebeitel concerning the general time frame of the creation of "the main Mahabharata." (32) As I have argued elsewhere, I think "the main Mahabharata" is concerned to provide ideological and narrative grounding for a brahminical conception of kingly rule and hierarchical society in the wake of the Mauryan empire and that government's cosmopolitanism and its insufficient recognition of the uniqueness of brahminic authority. On the other hand, I do not find persuasive Hiltebeitel's arguments for this Mahabharata being the first, last, and only "Mahabharata." He makes an interesting argument, and it is useful that he makes this argument. But Hiltebeitel's attempts to explain the variations and discontinuities of the text in terms of the intellectual and artistic variations of a composing committee working together synchronously are a bit far-fetched in some regards and do not provide an adequate explanation of those variations and discontinuities. Two of the main features of Hiltebeitel's theory--the interdisciplinary nature of the compositional committee and the argument that solutions of continuity on the surface of the narrative are deliberate artifice--are designed expressly to account for the surface heterogeneity of the text and the large and small breaches of artistic continuity on its face. The amount and kind of variation in the MBh could be the result of the sort of symposium Hiltebeitel imagines, but seeing it that way requires postulating a highly unusual and otherwise unknown event to account for variations that may be explained more economically by imagining the authorial agents responsible for the "epic" to be separated in time and interest and location. And why does Hiltebeitel imagine such a symposium? I believe it is for the purpose of grounding absolutely a reading of the text in terms of themes and motives that diachronic "excavation" of the text often marginalizes as "late." Only occasionally does Hiltebeitel make observations (usually incidentally, often only in the notes) regarding the central importance for the MBh of such themes as Krsna's being the supreme god incarnate, bhakti, the four-yuga theme, the avatara theme, the soteriological worldview of yoga, and so on; but these asides serve to remind readers of those arguments of Hiltebeitel and Madeleine Biardeau that do depend on a rigorously synchronous reading of the text. (33) For Hiltebeitel these are governing themes in the text, and any apparent lack of rigor in their connection to other parts of the text is the result of the committee's lax process and its love of concealment. I too think concealment and paradox are basic intellectual pleasures of some of the epic's authors--I believe an appreciation for mystery and concealment (and its opposite, apocalypse, that is, disclosure) is a fundamental element in their invention and use of the name(s) Krsna/Krsna--but I have difficulty imagining a committee of poets jointly inventing such a complex and ingenious connected narrative and at the same time allowing itself such "loose joins." Everything we know about brahminic and old Indian textual traditions tells us that editors and compilers amalgamate texts and do not at all mind "loose joins" or having no "join" other than physical contiguity; but individual authors in the Sanskrit tradition--when we have them: Panini, Patanjali, Asvaghosa, Kalidasa--enforce exceptionally tight connections in the texts they fashion.

My disagreement with Hiltebeitel on this is not a matter of principle, far from it. I think interpreters should work with the regulative principle that all texts are "presumed one until proven many." And though I am convinced that the MBh definitely has a diachronic developmental history, the larger discourse of ancient Indian studies should be grateful for the work of scholars like Biardeau and Hiltebeitel, whose conviction of the text's unity has led them to many profound insights into the MBh over the years, insights which those of us who do not believe in the text's oneness would likely never have achieved. But it is also worth saying that "excavationist" approaches to the MBh have not arisen mainly out of some scholars' prejudices against Indian artistic sensibilities, nor mainly from the lack of imagination of some scholars, nor mainly because it is intellectually easier and more economical to take something apart than it is to find the often subtle connections that hold it together. "Excavationist" approaches persist for these reasons, but also because there is no synchronous explanation or reading of the text that allays the suspicions many of us have about apparent heterogeneities on the surface of the text and apparent breaches of meaning (which often trouble the traditional commentators too, as much as we can get at them). (34)

The ultimate value of an assumption of synchronicity lies in the plausibility or value of the resulting reading, a criterion which is, of course, more than a little subjective and culturally determined, but not, I would argue, hopelessly so. Some of Hiltebeitel's readings are persuasive and do point out the value of trying to read a text synchronously before assuming that it cannot be done. His arguments in this book should slow down anyone who might try to argue that the god Dharma's three tests of Yudhisthira are only late additions to the MBh, and his discussion of the Nala story adds significantly to the arguments that that gem did not just find its way into the MBh by the Brownian movement of anonymous Sanskrit literature. Others, like the Vyasa-author reading (see below), are, in my judgment, misguided.

Nor Was the Mahabharata Composed in One Sattra

Several other specific historical arguments Hiltebeitel has made need to be addressed. The first of them concerns the mention in the MBh of the names of foreign people and places known in India only in later times, during the Christian era: the cities antakhi (Antioch) and roma (Rome) at 2.28.49 and the Huns (hunas and harahunas) at 2.47.19 and elsewhere. Hiltebeitel suggests that rather than being evidence of direct knowledge of the people referred to, these are references to "foreign histories with which the epic poets were familiar" (p. 31), which could thus be sufficiently early for the composition of his first and last MBh. This argument could be true, but the references to Antioch and Rome are associated with southern coastal realms (where India gained direct knowledge at least of Rome through overseas trade in the first centuries of the Christian era), which makes direct knowledge more likely than a pre-Christian-era history lesson, which presumably would not have located them to the southwest of Indraprastha. The references to the Huns, which occur more frequently in the epic, locate them in the north and mention them frequently along with the Chinese, and once with Tokharians (tukharas, 3.48.21). Hiltebeitel's argument that reference to them does not necessarily presume their presence in India seems quite plausible in this case.

Hiltebeitel's thesis about the text's composition requires that he make certain arguments regarding matters of prosody that are untenable. He wishes to maintain that all metric variation in the text derives simply from the heterogeneous styles or archaizing abilities of the poets of the purported symposium. There is a striking range of diversity in the tristubh verses of the MBh. Some passages exhibit a great deal of freedom in the construction of the stanza, in ways similar to the freedom of Vedic tristubhs (the least regular epic tristubhs, however, are much less variable than their Vedic ancestors). At the other end of the spectrum are stanzas that exhibit the perfect uniformity characteristic of classical Sanskrit poetry. It is fair to see these differences in epic tristubhs as earlier and later tristubh forms, for one of the major axes of the diachronic development of Sanskrit prosody is that from non-uniformity in the stanza to rigid uniformity in the stanza. At the same time, there is no obvious justification for insisting that an older, less uniform style of versification in a tristubh is a necessary indication of that stanza's older provenance, though in most instances I suspect it is. On the other hand, tristubhs that are nearly perfect upajatis (or jagatis), salinis, and vatormis are definitely near the chronologically later end of the epic's range of versification styles. Misled by Mary Carroll Smith's imprecise comparisons of epic and Vedic tristubhs, (35) Hiltebeitel wrongly argues that irregular epic tristubhs and slokas are both Vedic meters and suggests that the non-uniform tristubhs of the epic (Smith's "irregular" or "Vedic type" tristubhs) are the result of deliberate archaization on the part of some members of the compositional symposium. This suggestion has no plausibility, for, since no one would ever mistake the epic's "Vedic style" tristubhs for vaidika tristubhs, there is no obvious motive for "archaization." If, however, one postulated that the freer tristubhs might be a feature of a prior oral tradition (with the corollary that more uniform tristubhs and other uniform quatrain stanzas signify the use of writing), a motive for literate poets to compose such tristubhs would appear.

Hiltebeitel is vague about the use of writing in the development of the MBh (he believes the authors deliberately concealed the activity of their writing [p. 316]), and when it comes to the issue of the purported symposium's relation to prior traditions of "Mahabharata," Hiltebeitel is clear only about one thing: whatever oral tradition of Mahabharata there may have been prior to the symposium, it was not a tradition of epic poetry. He acknowledges that some kind of "Mahabharata" existed before the second c. B.C. when, discussing the fact that Panini had mentioned the names Mahabharata, and Arjuna, Vasudeva, and Yudhisthira, he says that "[b]y 'Mahabharata' he probably refers to a story known in one or more genres" (p. 18). (36) Regarding this problem he concludes "[o]ne cannot infer from such minimal information that he knew of a pre-second century epic, much less an oral one" (p. 18). Hiltebeitel is here mincing some sense of the word epic which he has not previously conveyed to his readers directly. (37) This point becomes clear as he continues: "Similarly, if we know anything of a pre-Valmiki Rama story, it is that what we know is not epic" (pp. 18-19). So when Hiltebeitel writes shortly afterward "[w]ith the creation of the [written] Mahabharata and [written] Ramayana, epic is something new on the Indian scene" (p. 19), it is not clear exactly what he means to say is new. (38) And when he goes on to write "[p]rior oral epic[italics of AH] versions of these texts are, for now at least, a creation of modern scholarship, and oral theory [italics of AH] another Western fashion with which to dress them up in the emperor's new clothes" (p. 19), he is saying very dramatically something rather smaller than it seems at first glance. (39) He admits there was some prior oral tradition that was some kind of precursor of the MBh, it just was not "epic." I would have no trouble with this had Hiltebeitel discussed exactly what the parameters of the genre of "epic" are in connection with the text under discussion. (40) So Hiltebeitel does not dismiss altogether the large amount of work Grintser, Brockington, Vassilkov, and others have done charting linguistic forms and poetic formulas in the epics, (41) but these issues complicate and burden his theory of the text. Hiltebeitel takes some cognizance of the survival of oral poetic traits in the epics when, quoting V. Narayana Rao, (42) he writes vaguely of "Mahabharata composition" as drawing on "an originally oral manner of composition distinctive for being 'a kind of oral literacy' or a 'literate orality'" (p. 22). Hiltebeitel thus recognizes some kind of oral textual tradition which has some sort of ambiguous existence in the written "epic," but he does not say clearly what that tradition may have been.

As I have said, I do think there was a significant MBh compositional effort at some point in the first century or two after the demise of the Mauryans, and I agree that writing was likely used in the creation of this text. Hiltebeitel does not really explore the issue of the text's being written in a focused and sustained way, but the notion that this "main Mahabharata" involved writing seems likely to me for two basic reasons that are impressionistic rather than demonstrable: the intricacy of the narrative would have been easier to develop with writing, and some highly refined elements of the text, such as the perfectly regular classical meters, suggest the likelihood of writing being used in their development. If we bear in mind the occasional gatherings of early Buddhists and, later and half a world away, early Christians, to address institutional concerns, debate doctrines, and argue about the canons of their common texts, there is nothing inherently implausible in the idea of occasional conclaves of some brahmins for similar sorts of purposes, including, perhaps, composing, redacting, or commissioning new texts. It is not inconceivable that a range of skills might have been recruited for some such compositional effort on some occasions. But what is implausible is Hiltebeitel's argument that the vast tale of the MBh that we have is the result of one such ambitious, far-seeing, internally diverse symposium. The implausibility I find in this suggestion, however, is basically only a recurrence of my prior conviction that the heterogeneity of the epic requires more than one significant redactorial effort at different times. I myself am inclined to imagine the development of the Mahabharata more along the lines of the gradual building, (43) modification, and occasional refurbishment of a great cathedral under the direction of different architects and master-builders at different points in time. I think the "gradualist" models of the epic's development that have prevailed in Western scholarship are obviously more plausible than Hiltebeitel's one-time symposium. And more plausible too would be envisioning the gradual development of the new kinds of institutions involved in the creation of written narratives in ancient India. Hiltebeitel complained recently that many Western scholars credit ancient Indian poets with virtually all skills except writing, (44) but that lacuna has been well nurtured by the larger brahminic tradition itself (Hiltebeitel might argue writing has been deliberately "concealed" by the larger tradition), which, in ancient times, virtually never presents itself as a written tradition. And if we are going to imagine a written MBh, then we must entertain the kind of mundane issues involved in the production of a gigantic written text, such as who supplied the implements and the material receptacle of the text (the palm leaves or the tree-bark) and who fed and housed the scribes or authors while they worked? These matters are much less pressing in the case of an oral tradition, for which the main comparable question is how an apprentice reciter was supported while learning the text. It is considerations such as these that led me earlier (45) to suggest that significant levels of royal or imperial support must have been involved in the production and propagation of any written version of the whole text.

The Search for the Author Vyasa in and out of the Text

As I said earlier, some of Hiltebeitel's reflections upon the abstract functions of authorship bring interesting new ideas to our thinking about the possible ways the MBh was created and propagated. But I find Hiltebeitel's determined construction of the "ever-receding" author Vyasa as a deeply knowing fiction of authorship on the part of the alleged symposium a rather forced reading of the text, one that well illustrates the danger of invoking deliberate authorial concealment or silence as a hermeneutic principle.

According to Hiltebeitel, the composing committee cloaked its ultimate authorial voice in the persona of a shadowy seer, Veda-Vyasa, whose "thought entire" (46) the epic portrays being unfolded to the sattra-sitting brahmins of the Naimisa Forest. (47) They portray this seer as having been an important agent in the epic events, as coming and going mysteriously from epic scenes, as regularly vanishing to who knows where after imparting a few choice words. From his deliberately concealed location he created a record of these events, which is now transmitted through his pupil Vaisampayana. The subject of the MBh's claim to authority--and similarities between those claims and those of the Mahayana Buddhists--is quite interesting (see above). But Hiltebeitel's reading of Vyasa as an analogue of the absent Buddha, as the author typically concealed beyond the text (and beyond most of the world, "on the other side of the mountain"), seems to me a tendentious misreading of the Suka episodes of MBh 12.310-20 in his chapter eight ("Vyasa and Suka: An Allegory of Writing") and a consequent reading of too much into the rest of the Vyasa record of the epic. Finally, when Hiltebeitel claims that the Suka episodes describe the specifically "literary" nature of the MBh, and that the story of Suka and Vyasa should be taken to be the fundamental vantage point for the reading of the "main story" of the MBh (pp. 280, 317), as one of the deepest principles of Vyasa's "thought entire," and so, actually, the outermost frame story of the entire epic (p. 279), he has stretched the motif of fiction much farther than is warranted.

A bit is told about Vyasa and his authorship of the MBh in Ugrasravas' and Vaisampayana's framing of their recitals, (48) but it is only the account of Vyasa's son Suka's attaining moksa at MBh 12.310-20 that gives any details of Vyasa's existence apart from his interventions among the epic heroes. (49) Far from regarding the Suka story as a peripheral and relatively late fiction about the otherwise famous seer Vyasa (which I think is the conclusion one will draw if one does not have an a priori conviction that the story must be synchronous with the rest of the MBh), Hiltebeitel sees the Suka episode as an integral part of the artistic whole. For him the depiction here of Vyasa together with his son is a brief tipping of their hand by the epic's actual authors. They give us here a glimpse of the normally concealed author in his literally "out-of-this-world" home base. (50) And Hiltebeitel sees in the fictional author's extraordinary boy Suka an evocative representation of the text and story of the MBh itself (p. 289). The boy, a dazzling gift to Vyasa from the great god, Siva, emblem of the written composition of the MBh, goes on to outshine his more world-bound father by gaining moksa as a mere boy, leaving his forlorn father with only a shadow. And in this poignant story of father and son, Hiltebeitel says, the epic authors offer us a key to the reading of their otherwise inscrutable text: "The main story [of the MBh], we are told, is a story Vyasa tells to Suka and his four other disciples. For one thing, the tale the father tells his son is often about father-son tales. Suka is the author's son who learns the sorrowful story from which he is allowed to escape. Or alternately, the story is the shadow and echo of what the father once told the son, who is now gone.... We should "read the 'main story' [of the MBh] from the vantage point of this story" (p. 317).

Parts of the story of Vyasa and Suka certainly do form a masterful fiction of intergenerational anxiety, and (in his closing discussion of these matters) Hiltebeitel's comparison of Vyasa's grief to that of the author of the Ramayana, Valmiki, is moving and thought provoking. But to argue--perhaps under the influence of the Ramayana's poignant depiction of poetry's being rooted in grief--that we should read the MBh through the keyhole of the Suka story highlights an extravagance of Hiltebeitel's hermeneutic method. To what extent does the MBh know or feel Vyasa to be its author (whatever the MBh itself might mean by that) in the first place? As I mentioned above, the Vaisampayana framing passages and the Ugrasravas framing passages know Vyasa as the author and tell us something about his authorship (precious little in fact, but all of it interesting in its own ways), and the couple of passages already discussed or mentioned that are at the back of the MBh amplify and reinterpret the idea of his authorship of the epic. (51) But while Vyasa's frequent incursions into the MBh narrative often are couched in terms of his parental relationship to various of the characters, and while he might be described in contemporary terms as one of the main directors of the MBh drama, his eventual authorship of an ambitious record of his descendents' deeds is never hinted at apart from the passages just itemized that lie on the periphery of the text. This fact is quite evident in Hiltebeitel's catalog of all Vyasa's appearances in the main story (in part 3 of chapter 2). As this catalog indicates, Vyasa's eventual authorship of the MBh is not even mentioned in the passage at MBh 15.26, where various seers, including Vyasa and his pupils, tell Dhrtarastra various dharmyah kathah. Hiltebeitel's continual description of Vyasa here as "the author" simply calls attention to how little the MBh itself seems to regard him as such. The implicit argument that the few passages that explicitly speak of Vyasa as the author of the MBh actually inform the entire text with that notion is just too tenuous; it gives too much leverage to too few passages that are located too thoroughly on the perimeter of the text. (52) And the suggestion that the evanescence of Vyasa's authorship--as seen from within the text itself--is an indication of the actual authors' sophistication and/or playfulness merely illustrates that Hiltebeitel's theory of the text provides a hermeneutic lever that is too powerful and easy to use.

Hiltebeitel's theory of the Mahabharata and its creation attempts a kind of reverse "Copernican revolution" in our perception of the text, but his theory of "the author" and "his works" only reaffirms my judgment that we have in the MBh numerous different "star systems" that might, if seen from far enough away, vaguely resemble a multi-armed spiral galaxy.

CLOSING

With very few exceptions I have ignored here the numerous translations of and commentaries on specific passages which make up the main substance of the book and call for much discussion.

In spite of my criticisms and dissents regarding aspects of this book and its premises, I view it as a very helpful step forward in our discussion of the Mahabharata. Hiltebeitel has done us a great good by the boldness with which he has laid out a definite vision of the great epic's creation and by the resourcefulness and imagination with which he reads the text and highlights many of its literary marvels, such as the poignant fact of Yudhisthira's never getting to ask Draupadi that question at the end of it all. (53) Like none before him, Hiltebeitel has charged us with a store of reflections that deepens our sense of that question and its very real enigmas.

This is a review article of Rethinking the Mahabharata: A Reader's Guide to the Education of the Dharma King. By ALF HILTEBEITEL. Chicago: UNIVERSITY OF CHICAGO PRESS, 2001. Pp. x + 365. $50 (cloth); $25 (paper).

1. Hiltebeitel, The Ritual of Battle: Krishna in the Mahabharata (Ithaca: Cornell Univ. Press, 1976).

2. The Cult of Draupadi, 2 vols. (Chicago: Univ. of Chicago Press, 1988 and 1991).

3. Rethinking India's Oral and Classical Epics: Draupadi among Rajputs, Muslims, and Dalits (Chicago: Univ. of Chicago Press, 1999).

4. Ibid., 20.

5. See p. 4 of the work under review.

6. See pp. 1-3, and also "Reconsidering Bhrguization," in Composing a Tradition, ed. Mary Brockington and Peter Schreiner (Zagreb: Croatian Academy of Sciences and Arts, 1999), 155-57.

7. I do not think all the parts of the MBh are connected with the same kinds or degrees of intelligence or meaningfulness--at some times the synchronous unity of the text is purely a mechanical fact based on factors extrinsic to meaning or art--but I think the presumption of meaningfulness is a necessary starting point.

8. Five of the book's eight chapters, 220 of its 322 pages, are devoted to these matters.

9. The ultimate authorial agency is represented by the actual creators of the text; and by "outermost voice" is meant the voice that begins every work and performs any and all later introductions of speakers.

10. Hiltebeitel argues in chapters three and four that these Naimiseya brahmins constitute a literary self-portrait of the epic's actual authorial committee.

11. Hiltebeitel is relying here on The Story of the Five Indras, MBh 1.189, which I too think is a very important passage in the epic.

12. See my "Making Yudhisthira the King: The Dialectics and the Politics of Violence in the Mahabharata," Rocznik Orientalistyczny 54 (2001): 63-92; Nick Sutton's "Asoka and Yudhisthira: A Historical Setting for the Ideological Tensions of the Mahabharata?" Religion 27 (1997): 333-41; and R. C. Zaehner's discussion of Yudhisthira in his textbook Hinduism (Oxford: Oxford Univ. Press, 1962). While Hiltebeitel and I share a high regard for Mukund Lath's paper "The Concept of anrsamsya in the Mahabharata" (in The Mahabharata Revisited, ed. R. N. Dandekar [New Delhi: Sahitya Akademi, 1990], 113-19) and while we traverse a few of the same themes in the figure of Yudhisthira, our understandings of Yudhisthira turn out to be complementary. Hiltebeitel may end up a little closer to Sutton's idealist interpretation of Yudhisthira, with which my own darker view of the king collides.

13. "Reconsidering Bhrguization," 156, 161.

14. "What is an Author?" in Textual Strategies: Perspectives in Post-Structuralist Criticism, ed. Josue Harari (Ithaca: Cornell Univ. Press, 1979), 141-60.

15. "Dvaipayana, Poet of Being and Becoming," in The Mahabharata Revisited, 101-11.

16. The Dialogic Imagination, tr. Caryl Emerson and Michael Holquist (Austin: Univ. of Texas Press, 1981).

17. Zur Erzahltechnik im Mahabharata (Hamburg: Verlag Dr. Kovac, 1994).

18. See John Brockington, "Issues Involved in the Shift from Oral to Written Transmission of the Epics: A Workshop Report," in Composing a Tradition, 132-38, for a recent discussion of style and discourse analysis in the epics.

19. I argued this as a then unwelcome conclusion in my Ph.D. dissertation, "The Moksa Anthology of the Great Bharata: An Initial Survey of Structural Issues, Themes, and Rhetorical Strategies" (Univ. of Chicago, 1980).

20. Of course it needs to be observed that if one does not postulate the MBh to be a synchronous artistic unity, then the maze of connections between the Veda and the epic reverts to being the matter of the older question concerning the epic's actual developmental history.

21. Some elements of this complicated discussion that are not sufficiently clear are the precise significance of the yatsattra to the epic authors, the exact connections seen between the earth and sky in the yatsattra and the MBh, and the exact senses of the important and frequently invoked name Naimisa. (I have done my best to reconstruct Hiltebeitel's tracing of this last idea in n. 28 below.)

22. Heestermann, "Vratya and Sacrifice," Indo-Iranian Journal 6 (1963): 1-37; Falk, Bruderschaft und Wurfelspiel (Freiberg: Hedwig Falk, 1986); Minkowski, "Snakes, Sattras and the Mahabharata," in Essays on the Mahabharata, ed. Arvind Sharma (Leiden: E. J. Brill, 1991), 384-400.

23. "The Indian Sorcery Called abhicara," WZKS 29 (1985): 69-117.

24. Especially "Sur le chemin du ciel," Bulletin des Etudes Indiennes 2 (1984): 213-79, and "The Development of the Vedic Canon and Its Schools: The Social and Political Milieu," in Inside the Texts--Beyond the Texts: New Approaches to the Study of the Veda, Proceedings of the International Vedic Workshop, Harvard University, June, 1989, ed. Michael Witzel (Cambridge, Mass.: Dept. of Sanskrit and Indian Studies, Harvard Univ., 1997), 257-345.

25. Andreas Bigger, "Wege und Umwege zum Himmel: Die Pilgerfahrten im Mahabharata," Journal asiatique 289 (2001): 147-66; and Thomas Oberlies, "Arjunas Himmelreise und die Tirthayatra der Pandavas: Zur Struktur des Tirthayatraparvan des Mahabharata," Acta Orientalia 1995: 106-24.

26. Witzel, "The Development of the Vedic Canon," 257-345.

27. Hiltebeitel does discuss C. V. Vaidya's views of an "Epic Age," but without engaging the issue of how the genre term "epic" is best construed. Likewise, the issues of empire and the "projection of empire" onto the distant Kuru past is never really pursued.

28. Hiltebeitel explores, in chapters 2 and 3, several aspects of the sense of nimisa (and nimesa), Naimisa-Forest (since it is always naimisaranya in the epic, and never a vana [though, as Hiltebeitel points out, we sometimes have naimisakunjas 'bowers'], we should wonder if "forest" is the best translation), and Naimisiya, or Naimiseya, rsis. Basing himself on frequent epic and puranic associations of the naimisaranya with the telling of stories, Hiltebeitel starts with nimesa as a measure of time (an 'instant') based on the concrete sense of 'blink of the eye' and develops this into the idea of "the 'Momentous Forest', or even the 'Forest of Literary Imagination', the forest where bards like [Lomaharsana or Ugrasravas] can enchant Brahman rsis: a 'momentous' forest where stories, to put it simply, transcend time and defy ordinary conceptions of space" (p. 96). After discussing several passages in which this idea or locale occurs, Hiltebeitel comes to the conclusion that "[David] White [Myths of the Dog-Man (Chicago: Univ. of Chicago Press, 1991)] is probably right that as the site for early sattras ... Naimisa Forest begins its etymological history as 'the twinkling forest,' the place where the stars would shine especially during darkening winter nights ...," though it is neither a specific constellation or a precise geographical location. "Rather, I believe it is the entire ever-changing visible night sky, which gods. Brahmans, rsis (Brahman and Royal), warriors who die in battle, Fathers, and also Nagas fill to all its reaches. There the heavenly Sarasvati can still appear, retaining her Rg Vedic associations with the Milky Way, to symposia of Naimiseya Rsis virtually anywhere they happen to gather" (p. 158). Finally, in setting the epic's outer frame with Saunaka's sattra in the Naimisa, the authors of the MBh "transformed the 'Twinkling Forest' of the Vedic heavens [i.e., the night sky described above, jlf] into the 'Momentous Forest (no less twinkling, but now more winking) of the Literary Imagination' ..." (p. 166).

29. The three frames are the Vaisampayana-Janamejaya frame of the snake-sattra, the Ugrasravas-Saunaka frame in the Naimisa Forest, and the "outermost frame" of the author Vyasa who composed the text way up north in the celestial mountains.

30. "The Institution of Fiction in Mahayana Buddhism," in Myths and Fictions, ed. Schlomo Biderman and Ben-ami Scharfstein (Leiden: E. J. Brill, 1993), 355-88.

31. A single instant, a simultaneous unity.

32. By this vague expression I mean to refer to the MBh with most of its familiar narrative elements, including the "Persuasion of Yudhisthira" after the war, the incognito in Virata's kingdom, the frame of the tirthayatra, etc. I believe some kernel of Bhisma's instruction of Yudhisthira (perhaps 12.67 through 12.90, or 12.59 through 12.108, or even 12.56 through 12.128) was likely present and the basic Vaisampayana frame with its amsavatarana listing. Most of the material in Bhisma's instructions probably came later, as did the Bhagavad Gita, as did all episodes that elaborate some theme of devotion to Visnu, Siva, or Krsna (such as, for example, the Sisupalavadha in Sabha Parvan; Arjuna's and Duryodhana's attending upon Krsna and choosing, respectively, for aid in the war, Krsna and his Narayana warriors at Udyoga Parvan 7, and several highly polished expressions of Krsna bhakti in the narrative wake of Yudhisthira's abhiseka in 12.40 up through the initiation of Bhisma's instruction of Yudhisthira in 12.56). How much and which elements of the Rama Jamadagnya material were in this "main Mahabharata" is still a puzzling question; see my "The Rama Jamadagnya Thread of the Mahabharata: A New Survey of Rama Jamadagnya in the Pune Text," in Stages and Transitions: Temporal and Historical Frameworks in Epic and Puranic Literature, Proceedings of the Second Dubrovnik International Conference on the Sanskrit Epics and Puranas, August, 1999, ed. Mary Brockington (Zagreb: Croatian Academy of Sciences and Arts, 2002), 89-132.

33. For the most part these observations do not contribute anything new to those arguments, but an exception is Hiltebeitel's discussion of the questions surrounding Draupadi's call to Govinda when Duhsasana tried to strip her in the cleverly entitled second section of the seventh chapter, "Disrobing Draupadi, Redressing the Text," 246-59.

34. That said, I would agree broadly with Hiltebeitel's grudging summons: "To the extent that there is value in some ... excavative projects ... it would be in asking less about the text's pre-history, and more about what it does with some of the matters raised" (p. 4). But it is only a matter of degree, emphasis, or tone. No matter what metaphor of disconnection is chosen--"excavation," "analysis," "surgical excision"--any scholarship that separates elements of the text is, at least implicitly, arguing about the prior history of the text.

35. The Warrior Code of India's Sacred Song (New York: Garland Publishing, 1992).

36. Hiltebeitel continues: "And with Arjuna and Vasudeva [Panini] probably alludes to a local cult."

37. It is implicit in his early assertion that "epic" involves "visions of empire," but he does not flesh out this special sense of epic, nor is it appropriate that he rely upon Quint's specialized treatments of many Western epics without directly explaining how and why.

38. And when he writes "Grintser (Pavel Grintser, Drevneindijski epos, genezis i tipologiya [Moscow: Nauka, 1974], 29) is not convincing that before they were written, 'both epics existed already as fully completed poems in the oral stage of their composition'" (p. 23), it is not clear exactly what he means to deny.

39. For all the scorn of "modern scholarship" and "Western fashion" in this sentence, he concedes something to it three pages later when he writes "Whatever 'oral dynamism' may precede the Mahabharata's written composition and continue after it to prompt modifications in the written text...."

40. The same mincing of terms occurs in the note to the last point, when he writes "Contrary to Vassilkov ("The Mahabharata's Typological Definition Reconsidered," Indo-Iranian Journal 38 [1995]: 249, 255), an 'archaic stage' of 'living oral [Mbh] epic' is not 'an established fact'" (p. 19, n. 74). And likewise, he engages in unwarranted paradox when he says, "Nowhere has oral epic been found to have emerged in a literary vacuum, such as now posited for Vedic India" (p. 19), for what I believe he means to say is simply that "oral epic" arises only as a derivative of written epic, and since writing was unknown in Vedic times there could not have been any "oral epic," whatever oral stories might have been circulated by whatever institutions. Again he is invoking his own undefined, but more specialized, sense of "epic," rather than our usual undefined use of the word.

41. See John Brockington, The Sanskrit Epics (Leiden: Brill, 1998), 103ff, for a review of much of this research and see his detailed discussion in "Issues Involved in the Shift from Oral to Written Transmission of the Epics: A Workshop Report," in Composing a Tradition, 132-38.

42. "Purana as Brahminic Ideology," in Purana Perennis, ed. Wendy Doniger (Albany: State Univ. of New York Press, 1993): 85-100, p. 95.

43. Frequently constructed around smaller churches that were often built over Roman shrines (to neutralize or occlude them), as these Roman shrines had overbuilt earlier pagan shrines. (I am grateful to my colleague Thomas Heffernan, a scholar of European Christian antiquity, for his observations on these matters.)

44. "Reconsidering Bhrguization," 158.

45. "The Great Epic of India As Religious Rhetoric: A Fresh Look at the Mahabharata," Journal of the American Academy of Religion 51 (1983): 611-30. (That paper has the unfortunate typo of "warfare" for "welfare" in line fifteen from the bottom on p. 618.)

46. For Hiltebeitel this phrase is a veiled allusion to the whole mystery of Vyasa's authorial activity: "The MBh, however, fuses yoga with bhakti, making its author and deity [Vyasa and Krsna--one wonders if Hiltebeitel's occasional yoking of these two is a concealed allusion to the Narayaniya's assertion that Vyasa is, like Krsna, an instance on earth of the supreme God Visnu Narayana; see 12.337.4-5 and 55-56] masters of yoga who unify Time, save souls, and can envision their 'entire thought' as a divine plan in which they can intervene" (97, n. 19).

47. The ultimate authorial voice of the text also shows up from time to time not only as Vyasa, but also as forms of Visnu and as Narada, Markandeya, and others, also taking over at times, or at least commissioning, the voices of Vaisampayana, Samjaya, and Bhisma. See p. 34, where Hiltebeitel quotes Michel Foucault concerning an author's dispersal of himself into a plurality of simultaneous selves.

48. In Ugrasravas' account of Vyasa's attendance at Janamejaya's sarpasattra at MBh 1.54 and in Vaisampayana's account of Vyasa's birth near the beginning of that pupil's recital to Janamejaya in MBh 1.57.

49. It is a remarkable piece in its own right, though, from the point of view of the MBh as a whole, it is comprised of two items near the back of the large and complex anthology of the Moksadharmaparvan, immediately preceding the Narayaniya, and its ultimate intent is to glorify the pursuit of moksa at the expense of garhasthya and the caturasramya, by having the son given Vyasa by Siva show up the duty-bound and earthbound father. (See the commentary upon it in the Narayaniya at 12.337.45-46.)

50. Hiltebeitel throws out hints from time to time in the book that Vyasa is located somewhere out of this world.

51. The interconnected Suka stories of 12.310-15 and 316-20 and the Narayaniya, 12.321-39, especially 337.

52. Among various interesting things that this survey reveals is that Vyasa becomes less clandestine, more public, and more frequent in his interventions toward the end of the MBh story and is several times accompanied by "pupils" and is located in an asrama. Do we witness in these later passages the actual birth of the idea that Vyasa amitatejas should be the author of the tale, when the idea that it needed an author actually arose?

53. See pp. 276-77.

JAMES L. FITZGERALD

UNIVERSITY OF TENNESSEE