Brahma: An Early and Ultimately Doomed Attempt at a Brahmanical Synthesis

In this paper, I argue that, by comparing certain passages from the early Buddhist su¯tras and the Maha¯bha¯rata, we can find evidence of a late- to post-Vedic ‘‘Brahmanical synthesis,’’ centered on the conception of Brahma¯ as both supreme Creator God and ultimate goal for transcending sam: sa¯ra, that for the most part did not become a part of the Brahmanical synthesis or syntheses that came to constitute classical Hinduism. By comparing the Buddhist response to this early conception of Brahma¯ with the way in which Brahma¯ is treated in certain sectarian portions of the Maha¯bha¯rata, I then argue further that the Buddhist critique of Brahma¯ as supreme deity was in part conceded by the Brahmanical tradition, and sectarian accounts of supreme godhead sought to reconcile pravr:tti and nivr: tti values more subtly than the crude juxtaposition offered by the earlier Brahmanical synthesis offered by Brahma¯. The result was that Brahma¯ was relegated to an inferior position as a fully sam: sa¯ric demiurge, a narrative found first in certain parts of the Maha¯bha¯rata and then continued throughout most of the Pura¯n: as.

Scholarship on early India has shown a renewed and increasing interest in recent years in the complex intersection of emerging religious institutions, literary developments, and political forces in the late first millennium BCE and early first millennium CE that served as the matrix out of which ‘‘classical’’ Hinduism and its Buddhist and Jain competitors emerged under the Guptas. In large part, the scholarly debate has centered on the ways in which older Brahmanical institutions tied to the transmission of the Vedas and s´rauta ritual adapted to socio-political changes such as urbanization and the rise of the Mauryan empire, and how these adaptations played out in newer Brahmanical texts such as the Gr: hya and Dharma Su¯tras; the Upanis:ads; and, especially, the Maha¯bha¯rata (MBh.). These texts and literary genres are associated with the rise of various new ideas and religious practices, including domestic rites,1 the a¯s´rama system,2 ideas about a¯tman and rebirth,3 and the S´aiva and Vais:n: ava bhakti movements.4 All of these, of course, would become important components of classical Hinduism, whose normative texts tended to shroud them in a veil of antiquity that has obscured, even for scholars, their relative novelty.5 There is another important religious development from this period that is easily overlooked, however, because, as I will argue here, it was ultimately doomed to failure and did not become a part of classical Hinduism. What I am referring to is the idea of Brahma¯ as the supreme deity, which developed out of older Vedic ideas concerning brahman, Purus:a, and Praja¯pati, but only found full expression in certain early Buddhist su¯tras and parts of the Maha¯bha¯rata, and was quickly superseded by the image of Brahma¯ as a hapless demiurge subordinated to one of the great bhakti gods,6 which emerged in the Maha¯bha¯rata and the Pura¯n: as. In an early Buddhist su¯tra, the Buddha’s Brahman interlocutors boast, Brahmans alone are the best varn: a; other varn: as are inferior. Brahmans alone are the white varn: a (lit., ‘‘color’’); other varn: as are black. Brahmans alone are purified, not non-Brahmans. Brahmans alone are Brahma¯’s own begotten sons, born from his mouth, born of Brahma¯, created by Brahma¯, heirs of Brahma¯.7 Likewise, (a)8 Brahma¯ is portrayed in another early Buddhist text as having the view—presumably reflecting the claims made about him by Brahmans at the time—that he was the was the Creator of all beings, and that his heavenly realm was  the highest realm that any being could hope to attain.9 These ideas are, unsurprisingly, roundly rejected in the Buddhist texts. What is surprising, however, is that the idea of Brahma¯ as supreme creator god whose heaven is the highest realm one can attain came to be rejected in classical Hinduism as well. How do we explain this radical change in Brahmanical attitudes toward Brahma¯ around the turn of the era? I will argue that the idea that Brahma¯ was both the supreme creator god and the highest religious goal, as found in the early Buddhist su¯tras, represented an early ‘‘Brahmanical synthesis’’ that was criticized by the Buddhists, and that much of the Buddhist critique was essentially conceded by the Brahmanical tradition, leading to new ways of conceiving of the nature of supreme deity in the bhakti traditions that came to dominate classical Hinduism. In order to make this argument, I will first look at the origins of Brahma¯ within the context of Vedic religion and the earliest valences of the term brahman. Then, I will outline what I call the first Brahmanical synthesis, an early vision of Brahma¯ as both the Supreme Creator of, and the ultimate goal transcending, sam: sa¯ra. Next, using the Pali Canon and Maha¯bha¯rata as counterpoints, I will argue that the early Buddhists successfully undermined this synthesis with their own account of multiple sam: sa¯ric Brahma¯s. This, in turn, I will argue, led to new Brahmanical syntheses to counter the Buddhist vision, in which Brahma¯ was no longer accorded supreme status, but incorporated in a diminished role10 that became dominant in classical Hinduism and endures to the present day.

In his seminal work (Notes on Brahman) on the concept of brahman and all its cognates, including Brahma¯, Gonda (1950) reviews previous scholarship on the ‘‘original meaning’’ of the word brahman, much of which he deems overly speculative, and explores the relationship of the neuter term bráhman to the masculine term brahmán (nom. brahma¯). He concludes that brahman is derived from the root br: h, which means to grow, swell, or expand, and thus carries connotations of power (pp. 18–21). He also mentions that the Indo-European suffix -men- is used to ‘‘express the idea of powers manifesting themselves either in actions and processes or in beings or objects’’ (p. 73). Thus, we can interpret bráhman somewhere along the lines of ‘‘that which swells,’’ and brahmán as ‘‘he who is swollen.’’ On this basis, Gonda makes a cautious comparison (i.e., without intending to obviate obvious trans-cultural differences) to the possession of power by shamanistic figures in cultures around the world (pp. 53–54). In other words, he argues—and I think few would dispute this—that bráhman is a power that resides in a class of men called brahmán, or, more generally, bra¯hman: a (p. 56). Neither the neuter nor the masculine can be said to be ‘‘prior’’ or ‘‘more original,’’ for neither would make sense without the other. Gonda also argues, however, against the position that the god Brahma¯ is a personification of bráhman, on the basis that by taking such a position ‘‘we implicitly admit the chronological priority of the ‘impersonal brahman’’’ (p. 62). He rightly notes that ‘‘the ancient Indian thinkers did not draw a hard and fast line between the personal and the impersonal, the animate and the inanimate’’ (p. 62), but I would argue that this is beside the point. The question is not whether or not there were always both personal and impersonal conceptions of brahman—clearly there were, in the form of the brahmán priest and the bráhman he possessed and delivered through sacred speech. Rather, the question is whether or not there was always (i.e., as far back as our sources go) a god named Brahma¯. The answer to this question, I think, is quite obviously no, and is a matter of historical record. Gonda disputes this, saying, [S]cholars have often attached too much value to the well-known argumentum e silentio: in the oldest Vedic texts the god Brahma¯ does not occur. But does the fact, that the authors of the R: V. etc. do not or not clearly mention a personal Brahma¯ mean that to other circles and groups of ancient Indian society existing at the same time the personal aspect of brahman was entirely unknown? (p. 64) Gonda asks this question rhetorically, but I think we nonetheless can answer it with a confident yes: We can deduce from the evidence that Brahma¯—at least in the exalted form in which he came to be known by the time of the composition of the early Buddhist su¯tras—was not present in the early Vedic period. It is true that the sources we have are ‘‘only representative of part of their contemporaries’’ (p. 65)—i.e., they represent the limited perspective of elite bra¯hman: as—but they do constitute a rather voluminous tradition that presents ample opportunities for reference to a supreme Creator God Brahma¯,11 but does not actually refer to such a Brahma¯ until the late Vedic period. This is not, as Gonda suggests, an argument from silence: indeed, reference to a god named Brahma¯ is made as early12 as in the Atharva Veda (AV 19.9.12, 19.43.8) and in the S ´atapatha Bra¯hman: a (S´B 5.2.2.8, 10.1.3.8). The Atharva references are ambiguous at best, appearing as they do in the context of pithy charms, but they seem to amount to little more than appeals for protection, not just from Brahma¯, but from other gods as well. The Bra¯hman: a passages, on the other hand, refer to ‘‘brahmán Br:haspati’’; they use Brahma¯, in other words, as a synonym for Br:haspati in his role as brahmán priest of the gods. This connection between Brahma¯ and Br:haspati should not be overlooked, nor should it be dismissed as a mere linguistic coincidence. In his book on the Mythology of Brahma¯, Bailey (1983) argues that while the Brahma¯-Br:haspati connection has been overstressed, to the neglect of more extensive functional similarities between Brahma¯ and Praja¯pati, there nevertheless are some important functional similarities between the former pair of gods (pp. 76–82). Indeed, we find some passages supporting an equivalency of the two in the Maha¯bha¯rata. In MBh. 5.12–16, Br:haspati protects Indra¯n:¯ from the wicked ı Nahus:a, who has dethroned her husband Indra. Br:haspati is several times in this episode referred to as Brahma¯, and possibly even worshipped as such by Indra¯n:¯ı (MBh. 5.13.3).13 Moreover, in MBh. 8.24, Brahma¯ serves as Rudra’s chariot-driver in a battle to defeat demons that are oppressing the earth. This episode is clearly modeled after similar stories in which Br:haspati serves as the chariot-driver of Indra; indeed, Rudra and Brahma¯ are appealed to as the model warrior and war strategist (MBh. 24.127), just as elsewhere (e.g., MBh. 5.162.9) Indra and Br:haspati are appealed to as such. Taken together, this evidence demonstrates an early connection between Brahma¯ and Br:haspati, one that seems quite natural given that Br:haspati is the brahmán of the gods. What is not present in the early material is a conception of Brahma¯ as a supreme Creator God; this will not come until Brahma¯ is equated with Praja¯pati, which does not happen until the post-Vedic literature (Holdredge 1996, p. 65).14 If Brahma¯ as a reference to a god first appears as little more than a title of the god Br:haspati, then how did it come to refer to the supreme deity? Is it possible, as Gonda suggests, that Brahma¯ as supreme deity was worshipped from time immemorial, just not by the caretakers of the Brahmanical tradition until the late Vedic period? This, I think, is unlikely, for we can trace the development of Brahma¯ fully within the Brahmanical tradition itself. As I have already mentioned above, the early Buddhist su¯tras make it clear that the god Brahma¯ was considered by his adherents to be the everlasting and eternal Creator of the world (MN 49), and his abode, brahmaloka, was regarded as the highest goal (DN 13; MN 97.31). The Upanis:ads are entirely consistent with this account, except that they speak almost exclusively of the neuter brahman, while the Pali Canon refers almost exclusively to the masculine Brahma¯. Here, however, I think we can agree with Gonda that ‘‘Indian thinkers did not draw a hard and fast line between the personal and impersonal’’ (1950, p. 62) especially considering that the Kaus:¯ıtaki Upanis:ad (Kaus:U) describes brahman in vividly personal terms: He [the soul that has attained to brahmaloka] comes to the couch Amitaujas... On it brahman sits. Knowing this, he mounts it, at first only with one foot. Brahman says to him, ‘‘Who are you?’’ He should say, ... ‘‘I am who you are... You are all this.’’15 What follows is an extended dialogue between brahman and the soul in which, among other things, the masculine, neuter, and feminine names of brahman are discussed (Kaus:U 1.7). Grammatical distinctions between neuter brahman and masculine Brahma¯ are thus, as Gonda points out, largely irrelevant. But this does not mean that the god Brahma¯, in spite of his relative absence until the late Vedic period, has been ‘‘hiding’’ in the neuter brahman all along, for the brahman of the Upanis:ads is far more than the brahman of the early Vedic period. It is no longer simply the power of ritual speech, the power possessed by bra¯hman: as; it is the ultimate reality that lies at the basis of the cosmos. Indeed, one could say that the most important contribution of the philosophical speculations of the Upanis:ads to Brahmanical thought was the elevation of the power possessed by bra¯hman: as to ultimate reality, the equation of Brahmanical power with ultimate reality itself. It was this ideological shift, the equation of Brahmanical power with the ultimate, that made the idea of a supreme deity named Brahma¯ possible, for until then, Brahma¯ would be little more than a masculinized abstraction of the ritual power bráhman, suitable for use as an epithet of the divine brahmán Br:haspati, but certainly not as a supreme deity. But this is only one half of the equation—or, more properly speaking, only the first of a pair of equations. What of the equation of Brahma¯ with Praja¯pati and the concomitant absorption of his role as Creator? This appears to have arisen as a separate development in the Bra¯hman: as. Unlike the Upanis:ads, the Bra¯hman: as—at least in their earlier portions, prior to the development of the A¯ ran: yakas—were firmly rooted in pravr: tti values and involved speculations on the nature of the sacrifice and its equation with the cosmic Creator Praja¯pati. The Bra¯hman: as do not use Brahma¯ as a regular synonym for Praja¯pati as later sources do, but there are a couple of passages that equate Praja¯pati with the brahman: He makes [the brahman] the ultimate of this all; therefore, they say the brahman is the ultimate of this all. ... Praja¯pati is indeed the brahman, for Praja¯pati is of the nature of the brahman.16 As Brahma¯ rose to prominence in the transitional period we are studying here, his equivalency to Praja¯pati became a de facto reality insofar as he absorbed almost all of Praja¯pati’s functions (Bailey 1983, pp. 63–82), even when, in name, he and Praja¯pati were regarded as separate gods.17 In particular, both the early Buddhist su¯tras (MN 49, 93.5) and the Maha¯bha¯rata (1.36.22, 8.23.32) testify to the belief that Brahma¯ was the supreme Creator God and that he created the four varn: as from the parts of his body.18 So what are we to make of this new God Brahma¯ that emerges in the second half of the first millennium BC, the product of the equation of brahman on the one hand with ultimate reality and on the other hand with the Creator Praja¯pati? I suspect that what we are dealing with is a new Brahmanical synthesis meant to promote Brahmanical ideology as Brahmanical culture spread to new parts of India, in particular as it spread eastward into the middle Ganges basin in the late Vedic period. My reasons for this suspicion are twofold. First, the body of literature where Brahma¯ figures most prominently as supreme deity is the early Buddhist su¯tras, which originated in the eastern region around Magadha, and in a time not long after Brahmanism had spread to that area and bra¯hman: as had begun to be employed for the purposes of state legitimation.19 Second, the Upanis:ads, in which was developed the key ideological component that made Brahma¯ possible as a supreme deity— namely, the elevation of brahman to the level of ultimate reality—probably had their provenance in the central and eastern regions of Northern India and arose in the context of Brahmanical expansion and interaction with new ideas (Olivelle 1998, pp. 10–16; Witzel 1989, pp. 115–116). Indeed, in an era when Brahmanical culture was expanding into new areas and encountering new ideas and religious systems—some of which, like Jainism and Buddhism, would pose a formidable challenge to Brahmanical claims to supremacy—Brahma¯ would have provided a powerful synthesis of ritual values, caste ideology, and the latest philosophical speculations on ultimate reality into one, all-powerful deity. Brahma¯, I would argue, was the first major attempt at a ‘‘Brahmanical synthesis’’—a synthesis of the major values and ideologies being espoused by bra¯hman: as in the face of competing ideological systems—but one, as we will see below, that was ultimately destined to fail. Indeed, if it were not for the early Buddhist texts, we might not know with any certainty that Brahma¯ was ever regarded as supreme deity, for in the Maha¯bha¯rata, the signs of the breakdown of the Brahmanical synthesis Brahma¯ represented are already quite apparent


Although by the time of the Maha¯bha¯rata’s composition Buddhism had arguably gained the upper hand against Brahmanism in India, during its infancy only a few centuries before it had struggled to assert itself in a society where Brahmanical values and terminology were already well-established. The strategy of the Buddhists in the face of this competitor was, as evinced by passages throughout the early Buddhist su¯tras, to co-opt, redefine, and thus subvert the ideas and terms that lay at the center of the Brahmanical system.20 The centerpiece of this strategy is the bra¯hman: a himself, who is consistently presented throughout the early Buddhist texts21 as a person defined by his morality rather than his birth and thus, in the strictest sense, as equivalent to an arahat. Similarly, the Buddha22 redefines the ‘‘Triple Knowledge’’ (tevijja), which in his day would normally have referred to the three Vedas, by claiming to have gained the knowledge of past lives, the knowledge of the rebirths of others, and the knowledge of the destruction of a¯savas (An_guttara Nika¯ya III.58). He even, in one sutta (DN 5), convinces a Brahman not to hold a Vedic sacrifice by presenting the act of giving food to the san_gha as a superior alternative. On the basis of such evidence, Bailey and Mabbett (2003) have argued that the whole ideology of merit-making in the early Buddhist Canon is modeled on the ideology of Vedic sacrifice (ch. 11, pp. 232–256), and White (1986) has argued that the Buddhist concept of merit transfer, found in later texts of the Pali Canon such as the Petavatthu, is modeled upon and presented as an alternative to the Brahmanical s´ra¯ddha rites. The early Buddhists did not always take this accommodating approach to Brahmanical ideological constructs and terminology, however; in some cases, they openly repudiated them. We see this most clearly in the trope of the delusion of Brahma¯.23 In one rather well-known sutta, the Buddha visits the abode of a certain Baka Brahma¯ to refute his pernicious view that ‘‘[t]his is permanent, this is fixed, this is eternal, this is complete, this is not subject to passing away, for this is not born, does not become old, does not die, does not pass away, does not arise again, and beyond here there is no further escape.’’24 In an article that examines this sutta in detail, Gombrich argues that the choice of Baka, which means ‘‘heron,’’ as the name of Brahma¯ is a joke25 and that the sutta as a whole is a parody meant to portray Brahma¯ as a ‘‘wily rogue but not wily enough: he overreaches himself, and in the Buddha more than meets his match’’ (2001, p. 107). Brahma¯’s ‘‘reach,’’ Gombrich says, is to offer to make the Buddha his heir (through a sort of yogic transference of faculties!), but the Buddha realizes this for what it is: an invitation to remain within sam: sa¯ra (pp. 101–102). Other suttas deliver a similar message. In the humorous Kevaddha Sutta (DN 11), a monk asks Maha¯brahma¯ where the four elements end—clearly a question about the nivr: tti goal—but the latter, after bombastically praising himself for the benefit of his na retinue, takes the monk aside and tells him to ask the Buddha, as he himself does not know the answer. In another sutta (DN 1.2.1–6), the Buddha explains that Maha¯brahma¯ develops the false view that he is the eternal, all-powerful Creator because he is the first being to fall from the A¯ bhassara world to brahmaloka when the cosmos expands at the beginning of a cosmic cycle. Indeed, the import of all these suttas, as well as of the fully developed Buddhist cosmology derived from them,26 is that Brahma¯ is neither all-powerful nor unique. The status of ‘‘Brahma¯,’’ while it is the fruit of lifetimes’ worth of meritorious karma, is entirely sam: sa¯ric and therefore temporary, and it is only through ignorance and delusions of grandeur that any one Brahma¯ can claim to be the eternal, omnipotent Creator. Of course, none of this discussion would be of particular interest, except that the various claims made about Brahma¯ by the Buddha, his followers, and their Brahmanical interlocutors in the early Buddhist su¯tras are paralleled by certain passages in the Maha¯bha¯rata. As one might expect, we find in some passages depictions of Brahma¯ that accord quite well with the claims made by the Buddha’s Brahmanical interlocutors in the early Buddhist texts. A good example of this can be found in the first book of the epic, in which Brahma¯ sends the other gods to Earth to fight the asuras: Being oppressed, World-Protector, by the great asuras thus overflowing with strength and power, this Earth supplicated Brahma¯. …….. Therefore the world, World-Protector, afflicted by the burden, oppressed with fear, went for refuge to the god who is the grandfather of all creatures.She saw the god Brahma¯, the imperishable Creator of the World, surrounded by very auspicious gods, twice-borns, and great seers. [Brahma¯] was being joyfully venerated by gandharvas and asparases, devoted to the acts of a herald, who had approached him, and she went to him and venerated him. And so Earth, seeking refuge, informed him in the presence of all the guardians of the world, Bha¯rata. But Earth’s business was already known, King, to him, the Supreme Self, the Self-Created, the Supreme. For how, Bha¯rata, could the Creator of the World not know, without remainder, the thoughts of the worlds of gods and asuras? O Great King, the Lord of Earth, the Source of All Creatures, the Lord, the Beneficent, the Lord of Creatures said to Earth, ‘‘I will entrust every single deity with that purpose for which you have come into my presence, Bearer of Wealth.’’ Having said this, King, and having dismissed Earth, the god Brahma¯, the Creator of Creatures himself, then ordered all the gods. ‘‘In order to cast off Earth’s burden, you each individually will be born on her for the conflict,’’ he said. ……... For the sake of the destruction of the enemies of the immortals and the welfare of the whole world, those deities descended in succession from heaven to this Earth. Thereupon, Tiger Among Kings, the deities were born according to their desire among the lineages of Brahman seers and families of royal seers. (MBh. 1.58.35, 37–46; 1.59.3–4)
 The account we see here is very much what we would expect given the assertions made by Brahmans in the early Buddhist texts. Brahma¯ here is clearly a singular, all-powerful Creator god; he is described as ‘‘imperishable creator of the world’’ (lokakarta¯ram avyayam), the ‘‘Supreme Self’’ (pradha¯na¯tmanas), the ‘‘Source of All Creatures’’ (prabhavah: sarvabhu¯ta¯na¯m), and the ‘‘Lord of Creatures’’ (praja¯- pati). All other gods obey his command, including, explicitly, Na¯ra¯yan: a (MBh. 1.59.1). Likewise, elsewhere in the epic, Brahma¯ is described as ‘‘indestructible and eternal’’ (MBh. 5.95.2: aks:ayas´ ca¯vyayas´ caiva brahma¯ lokapita¯mahah:) and as having created Brahmans from his mouth (MBh. 8.23.32: bra¯hman: a¯ brahman: a¯ sr: s:t:a¯ mukha¯t). Surprisingly, however, in other places in the epic, statements are made that seem very much in line with the Buddhist response to the Brahmanical synthesis outlined above. In the Gı¯ta¯, for example, Kr:s:n: a asserts that all the worlds, including Brahma¯’s, are subject to rebirth, but those who are devoted Kr:s:n: a are not reborn (MBh. 6.30.16). Similarly, in another place the seer Durva¯sas is told by the Envoy of the Gods, in a description of the ‘‘faults of heaven,’’ that The consciousnesses of those falling (from heaven) are upset and they are afflicted by emotions

When their garlands wither, they fear that they are about to fall from there. Ah, up to the Brahma-world, Maudgalya, there are these severe deficiencies!27 Indeed, the whole motif of Brahma¯ representing pravr: tti values, and being subordinated as such to another god, that Bailey delineates in his book is found frequently in the Maha¯bha¯rata and becomes the dominant way of depicting the erstwhile supreme deity in later texts such as the Pura¯n: as. Although a direct causal link would be difficult to prove, the rhetorical similarity of the relegation of Brahma¯ to sam: sa¯ra that is found in Brahmanical texts starting with the Maha¯bha¯rata to the critiques of the Brahmanical synthesis found in early Buddhist texts is impossible to ignore. What I would like to suggest, therefore, is that in the complex web of inter-sectarian polemic, philosophical debate, and cultural exchange at the turn of the era—a debate in which the Buddhists surely had a fairly dominant voice—the idea that Brahma¯ was not the ultimate, that brahmalokawas just another realm into which beings are reborn, won out in the end. The Brahmanical synthesis of the late Vedic period, in other words, disintegrated.

Many previous scholars have, of course, attempted to draw links between the Maha¯bha¯rata and Buddhism. As Hiltebeitel notes in a recent article (2005a) that surveys and engages with the major theories on the epic’s response to Buddhism, such attempts date to Adolf Holzmann’s late nineteenth century ‘‘inversion theory’’—namely, that the Maha¯bha¯rata originally celebrated As´oka in the character Duryodhana, but that the epic was later appropriated by Brahmans and ‘‘inverted’’ to favor Kr:s:n: a (p. 108). More recent scholars have also seen a special role specifically for As´oka in the Maha¯bha¯rata. In her monumental work on the epic, Biardeau (2002) argues that the Maha¯bha¯rata was written as a direct response to As´oka’s rule, and that we must ‘‘decode’’ it in order to find hidden messages that it bears regarding Buddhism and its greatest patron, As´oka. In particular, she reads As´oka in the figure of Jara¯sandha, whom she shows to have interesting ties to Kalinga, the country conquered in real life in a bloody war by As _ ´oka; in addition, his name contains the word jara¯ (‘‘old age’’), which plays an important role in Buddhist doctrine, and he is killed by Kr:s:n: a and the Pa¯n: d: avas in Girivraja (a.k.a. Ra¯jagr:ha), a city closely associated with the Buddha’s life-story. Nicholas Sutton and James L. Fitzgerald, on the other hand, have read As´oka in one of the epic’s main characters, Yudhis:t:hira—although Sutton sees Yudhis:t:hira as a positive depiction of the emperor, while Fitzgerald sees him as a negative depiction thereof (Hiltebeitel 2005a, p. 114). In both cases, they see Yudhis:t:hira as reminiscent of As´oka in his remorse over the violence that had been required to attain kingship and his desire to abandon said kingship and retire to life in the forest. Fitzgerald (2001) argues (rightly, as I see it) that this cannot be a positive depiction of As´oka since Yudhis:t:hira is ultimately convinced that violence is proper to the dharma of a ks:atriya and that he should not abandon kingship (pp. 64–65). I tend to agree with Hiltebeitel (2005a), however, that we should not be so quick to focus on the Maha¯bha¯rata as a response to Buddhism or As´oka to the exclusion of other (to the minds of Brahmanical orthodoxy) problematic religious movements and kings; nor should we read individual characters in the epic as unidimensional images of particular historical figures, to the exclusion of more ‘‘polysemic’’ readings (p. 129). Although As´oka and Buddhism are undeniably important to the history of India in the late first millennium BCE, they represent by no means the only king and the only rival religious movement of that era that the epic authors would have found objectionable. Our tendency to attribute singular importance to them is based in part on what historical artifacts the vagaries of time have bequeathed to us and even simply the way that we think about historical causality (e.g., in terms of ‘‘great historical figures’’). In other words, there is no a priori reason to assume that the epic authors, if they wished to make a political statement with their work (which nearly all scholars agree they did), must have done so by identifying a particular figure from the past (As´oka) and a particular ‘‘rival religion’’ (Buddhism) against which to formulate their own political vision; concomitantly, if their ‘‘real world’’ rivals appear to be absent or nearly absent from the epic, we need not assume that they are actually secretly ‘‘encoded’’ into it. Indeed, it is not clear that a historical–political discourse even existed at the time that would have made such a pointed critique of As´oka or ‘‘Buddhism’’28 intelligible

The type of discourse we do have evidence for is precisely what we find in the epic text—the articulation of a political vision through an apocalyptic narrative, over and against an alternative that is represented not by historical actors or real-world institutions, but by a particular set of values, in particular the valorization of ahim: sa¯. It is within such a methodological framework that I wish to situate my argument that there was an early Brahmanical synthesis centered on Brahma¯ that disintegrated under the Buddhist critique that I have outlined above. I am specifically not trying to argue that the authors of the Maha¯bha¯rata necessarily had access to any Buddhist texts that attacked the earlier Brahmanical synthesis or formulated a ‘‘response’’ to that attack in a conscious way. Indeed, as I have hinted above and will argue in more detail below, the sections of the epic that accommodate the Buddhist critique of the earlier Brahmanical synthesis appear to be later additions to the original apocalyptic narrative. What I am arguing is that Buddhist critiques of the first Brahmanical synthesis—perhaps among similar critiques by others that are no longer extant in the historical record—led to the abandonment of that synthesis in Brahmanical circles and the development of newer, more robust syntheses.30 The Maha¯bha¯rata, then, does not represent a ‘‘response’’ to the Buddhist critique; rather, it reflects the impact that the Buddhist critique of the first Brahmanical synthesis had on ancient Indian discourses on Brahma¯. According to my hypothesis, the stage in this shift in the discourse that the Maha¯bha¯rata would belong to is a transitional stage between the first Brahmanical synthesis and later syntheses that came to replace it. As the examples given in the previous section indicate, however, the ‘‘transition’’ reflected by the epic does not consist of descriptions of Brahma¯ somewhere halfway between those offered by Brahmans in the early Buddhist su¯tras and those offered by the Buddha; rather, it consists of a mixture of mutually incompatible descriptions, some of which are consistent with the old Brahmanical synthesis and others of which reject it. Such a state of affairs would appear to be consistent with the argument made by James Fitzgerald and others before him that the epic developed over time, and that not all sections of it were authored by the same person or persons. In particular, the Bhagavad Gı¯ta¯ (BG) has convincingly been demonstrated to be a late interpolation,31 and the existence of both blatantly S´aiva and blatantly Vais:n: ava passages within the epic begs for at least some sort of explanation. It is this issue of competing sectarian viewpoints32 within the Maha¯bha¯rata that I will focus on here, because I believe it holds the key for understanding the process by which the Brahmanical tradition dismantled the older synthesis that centered on Brahma¯ and developed new syntheses to deal in a better way with many of the same problems the earlier one had. Alf Hiltebeitel has argued in recent years, however, that the Maha¯bha¯rata can be understood as a coherent narrative written more-or-less at one time by a single author or, more likely, ‘‘committee’’ (Hiltebeitel 2001).33 In doing so, Hiltebeitel argues quite convincingly, contra previous theories of a ‘‘heroic’’ Kr:s:n: a that later became divinized, that Kr:s:n: a’s divinity is indispensable to the epic.34 This is of particular import given the account, already cited above, of Brahma¯ sending the gods—including Kr:s:n: a—to incarnate themselves as human beings (MBh. 1.58.35– 44). Indeed, this episode is crucial for interpreting the narrative as not only an account of an epic battle, but also a full-blown apocalypse in which the primordial and thus paradigmatic battle between devas and asuras is replayed on Earth. But while Hiltebeitel’s theory of a unified epic works for much of the epic narrative, it is very difficult to reconcile with certain parts that I would call irrefutably sectarian—first and foremost the Gı¯ta¯. One simply cannot reconcile, for example, the Gı¯ta¯, in which Arjuna praises Kr:s:n: a in his supreme form as ‘‘infinite lord of the gods, the dwelling-place of the world, imperishable, being, non-being, and that which is beyond (both)’’35 with the end of the Sauptika Parvan, in which Kr:s:n: a nonchalantly explains that the massive slaughter of the last remnants of the Pa¯n: d: ava and Pa¯n˜ca¯la armies was possible because ‘‘the son of Dron: a went to the imperishable god of gods, lord of lords, for refuge,’’36 adding, without any apparent irony, ‘‘For I know the Great God as he is ... For he is the beginning, the middle, and the end of beings.’’37 Such discrepancies can only be explained by an appeal to multiple sectarian and proto-sectarian additions to the epic text as we receive it today. This in no way undermines Hiltebeitel’s contention that there is an underlying unity and authorial intent to the Maha¯bha¯rata; rather, it is a recognition that the unified narrative Hiltebeitel sees does not correspond, as Fitzgerald puts it, to ‘‘the first, last, and only ‘Maha¯bha¯rata’’’ (Fitzgerald 2003, p. 811).38 Having taken all this evidence into account—evidence both for an underlying unity to the text and for a diachronic development of it—I am persuaded that, as Hiltebeitel argues, the Maha¯bha¯rata was first composed with a definite authorial intent as a story about an apocalyptic battle to restore Dharma in which gods and demons instantiate themselves in human actors to essentially replay the Vedic battle between devas and asuras, but that later on, as argued by Fitzgerald and others, sectarian additions were made. In the first stage of the epic text, as exemplified by the account of Brahma¯ sending the gods to Earth discussed in the last section, the late Vedic Brahmanical synthesis has not yet broken down: Brahma¯ is head of the gods, but otherwise all of the gods—even and especially Kr:s:n: a/Na¯ra¯yan: a—are, much like ordinary human actors, finite, multi-faceted, and often morally ambiguous. This early version of the text would correspond to Hiltebeitel’s epic and vindicate his most important insights about it—that it is a cohesive work with authorial intent, that it is non-sectarian, and that the divine (if not supreme) Kr:s:n: a is integral to the narrative. Fitzgerald argues that the Critical Edition corresponds more or less to a version of the epic that was disseminated under the Guptas; he argues that the massiveness of the undertaking and the wide dissemination of the text that followed must have been done with imperial support (2004, p. xvi n. 2). This hypothesis—as well as, for that matter, Hiltebeitel’s theory that the Maha¯bha¯rata as a whole was composed at a much earlier date—is, of course, based on the assumption that the Critical Edition does in fact correspond to some sort of historical artifact. But as Le´vi (1929) objected almost immediately after it began to be published, the editors of the Critical Edition had to make so many arbitrary decisions to reconcile differences between the various manuscript traditions that are found in nearly every line that it is far from clear that they have indeed discovered an ‘‘Ur-Text,’’ however far removed from the ‘‘original epic’’ that might or might not be; indeed, Le´vi admits, ‘‘I think that [Sukthankar] has simply created another recension, the Poona recension’’ (p. 347). Biardeau (1968; cf. 2002, pp. 19–20) has gone further and argued that Sukthankar et al., out of a rather traditional Brahmanical desire to find the most authoritative version of the Maha¯bha¯rata possible, were attracted to the aura of scientific accuracy offered by Western text-critical methods and applied them to an object of study they were never designed for—namely, oral texts, which she argues preserve the meaning of stories rather than their exact wording. Both Fitzgerald and Hiltebeitel reject this criticism of the Critical Edition, arguing that, as Fitzgerald puts it (1980), ‘‘[t]he amount of unity, both petty and general, that exists among the MBh manuscripts, ... can be explained only on the assumption of a text antecedent to those mss., an archetype’’ (p. 57 n.).
It is not my intention here, however, to enter into a debate over the merits or flaws of the Critical Edition. I have already explained why I find it unlikely that the Critical Edition (or the ‘‘whole Maha¯bha¯rata’’ in any other way construed) was composed all at once. This leaves two other possibilities: that a version corresponding more or less to the Critical Edition was edited and disseminated at some time after the original composition (possibly, as Fitzgerald [2001, p. 68] suggests, under the Guptas), or that the various recensions that have come down to us are the product of a long and fluid process of written and oral composition and transmission that still are not, and perhaps cannot be, fully understood. If Fitzgerald is correct that the Critical Edition corresponds to a version that was disseminated under the Guptas, then that would leave quite a bit of time—about three centuries—during which the Maha¯bha¯rata would have been a rather free-floating text, and it is in this period that various elaborations were made upon the text—in different places, in different times, and by different people—some of them invisible to us now because of their innocuousness, others clearly belied by their sectarian bent.40 A more skeptical approach to the Critical Edition, such as has been advanced by Le´vi and Biardeau, on the other hand, directly presupposes that the Maha¯bha¯rata was for centuries a fluid, oral text and thus would have been subject to elaborations in different places and times and by different people. What I would like to show now is that those elaborations that are most clearly sectarian are attempts to deal with the failed Brahmanical synthesis represented by Brahma¯. If Brahmans did indeed accept the Buddhist critique that a single god cannot in an unproblematic way both stand outside of sam: sa¯ra and be the Creator and sustainer of sam: sa¯ra, then they must have found some way to respond.

One of the most influential Vais:n: ava responses to the collapse of the Brahmanical synthesis, embodied in the Gı¯ta¯, was to save svadharma by replacing renunciation of action with renunciation of the fruits of action—that is, by advancing a new theory of yoga called karmayoga. As Fitzgerald explains, The Gı¯ta¯’s main argument is that human beings can and should effect the yogic religious transformation of themselves into fully perfected beings, enjoying eternal union with the ultimate entity of all being, God, within the context of the normatively proper four-caste society, and through the activities ordained as proper for one as a member of one of the four general castes. (Fitzgerald 1983, p. 616) Kr:s:n: a unambiguously advances this thesis from the beginning of his sermon, asserting that Ritual action arises from brahman...brahman arises from the imperishable. Therefore the all-pervading brahman is eternally established in sacrifice,42 and rhetorically asking Arjuna, ‘‘This world does not belong to those who do not sacrifice; how about the other...?’’43 He then advances his own theory of yoga, saying, ‘‘The one yoked, having abandoned the fruit of action, attains firm peace,’’44 and, ‘‘Renouncing all actions with his mind, the master (of the passions) sits happily.’’45 This discipline clearly borrows a philosophical move first advanced by the Buddhists, that is, a mentalization of karma, and thus the transformation of renunciation into a primarily mental act. Unlike the Buddhists, however, the Gı¯ta¯ does not use this relegation of renunciation to the mental sphere to reject ritual action as irrelevant; rather, in a clever subversion of Buddhist doctrine, it uses the delimited nature of renunciation to argue that renunciation is fully compatible with—indeed, impossible without—ritual action, and all of svadharma along with it.46 But how does Brahma¯ fit into all this? Kr:s:n: a asserts to Arjuna that ‘‘people return again from the Brahma-world,’’ thus relegating the erstwhile nivr: tti goal of brahmaloka to sam: sa¯ra, but promises that ‘‘having reached me, ... there is no rebirth.’’47 Nevertheless, while Kr:s:n: a is now the new nivr: tti goal, he is not simply a replacement for the old Brahma¯, nor is Vais:n: avism—or at least the brand of Vais:n: avism advanced by the Gı¯ta¯—a mere recapitulation of the old Brahmanical synthesis with a cosmetic name-change. Unlike Brahma¯—who as nivr: tti goal must stand outside of sam: sa¯ra, yet also inexplicably is the Creator of sam: sa¯ra—Kr:s:n: a-Vis:n: u is the metahistorical principle48 that at once transcends sam: sa¯ra and subsumes it. As such, he does not himself create (which would imply a dichotomy between himself and his Creation49), but contains both Creator and Creation. This is illustrated by the standard Vais:n: ava cosmogony in which Brahma¯ emerges from a lotus growing from Vis:n: u’s navel to create the universe (MBh. 3.194.11–12)50 and even more graphically in the Gı¯ta¯ when Kr:s:n: a reveals his supreme form (paramam: ru¯pam: ) and Arjuna perceives all the worlds and the Creator himself in the person of the Lord (BG 11.9–49). The Vais:n: ava response to the Buddhist critique, therefore, is to reject the dualism on which it and the old Brahmanical synthesis were based. Where Brahma¯ failed in trying to straddle the pravr: tti and nivr: tti as two radically separate realms, Kr:s:n: a succeeds by embodying a nivr: tti that subsumes the pravr: tti within it

An excellent example of a S´aiva attempt to deal with the fractured Brahmanical synthesis can be found in the Stha¯n: u myths analyzed by David Shulman. We have already seen a part of the most important myth Shulman utilizes; it is found in a passage at the end of the Sauptika Parvan in which Kr:s:n: a describes Maha¯deva as the highest god.51 Kr:s:n: a then continues by recounting the creation of all beings within a S´aiva framework: Thus desiring to create beings, the Existent, the Grandfather, saw the First And said to him, ‘‘Create beings, quickly!’’ Harikes´a said, ‘‘So be it!’’ Seeing the defects of beings, The one of great tapas generated tapas for a long time while plunged in water. Having waited for him for a very long time, the Grandfather Created another creator of all beings using his mind.52 This ‘‘other creator’’ is Praja¯pati, and he in turn creates other Praja¯patis to carry out the creation. Unfortunately, these Praja¯patis run amok: They get hungry; turn on their father, the original Praja¯pati; and try to eat him. Brahma¯ now has to intervene to give them food, and it is at this point that the Great God reenters the scene: But when the world-guru was pleased with the multiplication of the collection of beings, The Most Excellent arose from the water and he saw these creatures. Having seen the multifarious creatures that had multiplied through their own power, Lord Rudra became angry and struck his lingam

S´iva is clearly upset because, while he was busy generating the tapas to create perfect creatures, his subordinate Brahma¯ got impatient and did the job imperfectly. What is done is done, however, so S´iva makes the most of the situation and dedicates his tapas to the eternal renewal of food for his creatures (MBh. 10.17.25). Unlike in the Gı¯ta¯, where the unequivocal sine qua non for any discussion of rapprochement between pravr: tti and nivr: tti values is the affirmation of the primacy of svadharma, the acceptance of nivr: tti values seems at first to be straightforward and complete in this S´aiva myth—the high god here, after all, is himself a tapasvin. The potential nivr: tti utopia is turned entirely on its head, however, when S´iva’s attempt to create perfect creatures through austerity leads ironically to the flawed Creation of Brahma¯ and to its sustenance. As Shulman (1986) writes, ‘‘[N]ivr: tti has a definite primacy, though it always leads, ironically, to pravr: tti. To deny the world is to fuel its flawed unfolding’’ (p. 108). The S´aivas here are less concerned than the Vais:n: avas with the preservation of svadharma, but they do deal with the failure of the Brahmanical synthesis head-on, in quite a clever way. First, they acknowledge the critique of the old Brahmanical synthesis by separating the nivr: tti goal (S´iva) from the pravr: tti Creator (Brahma¯). Then, through a subtle mythic interweaving of austerity and fertility, other-worldly dreams and this-worldly realities, they undercut the radically dichotomizing logic, used so successfully by the Buddhists against Brahma¯, that made the separation of nivr: tti goal from Creator necessary in the first place

As we have seen over the course of this article, Brahma¯ is by no means a static fixture within the Indian tradition; he is, instead, both the product of and the impetus for centuries of religious change and contestation, especially between the Buddhist and Brahmanical traditions. After most of the relevant myths in the Maha¯bha¯rata as it comes down to us were composed, Brahma¯’s mythology continued to be elaborated upon in the Pura¯n: as, but, for the most part, he never returned to his erstwhile status of supreme deity, this position having been permanently usurped by the up-and-coming S´iva and Vis:n: u. On the contrary, these classical Hindu texts continued to subordinate Brahma¯ to one or the other of the sectarian High Gods, and in so doing produced the image of the hapless, four-faced  demiurge by which Brahma¯ is usually known today.57 This image of Brahma¯ is obviously quite different not only from that found in what I have called the first Brahmanical synthesis, as attributed to the Buddha’s Brahmanical interlocutors in the early Buddhist su¯tras, but also from the pantheon of sam: sa¯rically entrapped Brahma¯s presented in those same texts as the Buddhist alternative. That is, while the sectarian architects of the new Brahmanical syntheses that came to dominate classical Hinduism appear to have accepted the Buddhist critique that the combination of pravr: tti and nivr: tti values—of both Creation and liberation—found in the older conception of Brahma¯ was untenable, they did not simply reproduce the Buddhist pantheon of multiple Brahma¯s connected to a putative Creation only by their own delusions of grandeur. This, however, is easily understandable given the acute differences between the ideological agendas of the early Buddhists and the early Brahmanical bhakti movements. The early Buddhists, opposed as they were to the very idea of Creation, sought to defuse the claim of a single Maha¯brahma¯ to have performed that act by making him no different from the other denizens of brahmaloka, just one Brahma¯ among many. In contrast, various Hindu traditions preserved a place for the world’s Creation, simply relegating it to a secondary position by making Brahma¯ the demiurge of a higher God. Nonetheless, both articulations of Brahma¯, the early Buddhist and the Pura¯n:ic Hindu, draw on a common Brahmanical tradition in which Brahma¯, synthesized from several Vedic sources, was at least for a time considered the supreme deity, and both deviate from and elaborate upon that early Brahmanical tradition to cope with, respond to, and ultimately shape the ever-changing contours of the Indian religious world.