When Did Buddhism Become Anti-Brahmanical? The Case Of The Missing Soul

Joseph Walser* Many textbooks for Introduction to Buddhism or World Religions courses treat Buddhism as a competitor of either "Hinduism" or "Brahmanism" by asserting that Buddhism teaches that there is no eternal self or soul and Hinduism teaches that there is. I ask whether these assumptions hold up for one of the earliest sources about Buddhism, the Pali canon. Using statistical analysis of 5,126 suttas or "discourses," I argue that there is little evidence that the doctrine of soullessness was preached to "convert" representatives of the Brahmanical tradition to Buddhism. On the contrary, it would appear that Brahmin Buddhists had their own canon-within-acanon that simply avoided the topic of soullessness. Rather than seeing the canon as "what the Buddha taught," the argument here will present canonicity itself as one of the stakes in a nexus of power where different communities strove to assert their version of Buddhism to be "what the Buddha taught."
I WOULD LIKE to examine a claim that is occasionally made about Buddhism, sometimes in textbooks and sometimes in more advanced monographs, namely that Buddhism teaches that there is no self or soul, and that this doctrine marks the borderline between Buddhism and Hinduism (or Brahmanism). The latter, in this case, is assumed to comprise the teaching that the soul exists. Buddhism and Brahmanism are thereby asserted to be polar opposites with the doctrine of soullessness (Sanskrit anātman/ Pali anatta) held to be the refutation of the Brahmanical doctrine of the soul.

The latter assertion is, of course, rather crude and most works on Buddhism avoid putting the matter so baldly. Donald Mitchell's introductory textbook on Buddhism is typical:
…the Upaniṣadic teachers of the time held that lasting happiness
(ānanda) could, in fact, be found deep within one's self. They called this inner, permanent, and divine reality the "Self " (Ātman). The Buddha, on the other hand, presented a third characteristic of existence that rejects any such permanent Self. He said that all phenomena lack any underlying and permanent substance; they all have the characteristic of "no-self."
This also means that within human nature, there is no permanent self or soul. (Mitchell 2008, 37)
Notice that while Mitchell does not explicitly say that the doctrine of anātman was directed at followers of the Upaniṣadic doctrine, the juxtaposition of the statements certainly leaves the reader with that impression.

The default assumption here is that whatever the Buddha taught concerning the self and its liberation, it ran counter to that of the Upaniṣadic teachers. By extension, after reading the above account we would expect a
"Buddhist Upaniṣadic teacher" to simply be a contradiction in terms. One either follows the teachings of the Buddha or the Brahmanical teachings of the Upaniṣads. The doctrine of anātman is thought to form a rigid borderline between Buddhism and Brahmanism—a line that may only be crossed by conversion.

Although it is rarely mentioned in today's textbooks,1 it has long been debated whether the early Buddhist suttas ever actually deny the existence of a self or a soul in the first place. Prominent scholars of Buddhism are lined up on either side of this issue.2 While it is true that no sutta ever makes what should be a grammatically simple declaration to the effect that "there is no self or soul" or "the soul does not exist," the fact remains that the body of suttas discussing the doctrine of anatta were looked to by much later generations of logicians as warrant for arguments against the existence of the self. The question is whether the discourses dealing with the topic of anatta (however we want to interpret it) were intended to distinguish Buddhists from followers of Upaniṣadic teachings. To answer this question, I would like to examine whether and to what extent such a borderline can be found in early Buddhist canonical texts by examining whether the Buddha is ever represented as teaching the anatta doctrine to Brahmanical opponents in the Tipiṭaka. The simple answer is, not really. That short answer, however, is less interesting than the subsequent questions that the data raise, namely, what were the important religious distinctions at play in these texts prior to the distinction between Buddhism and Brahmanism, when was the borderline between Buddhism and Brahmanism drawn, who drew it, and why? Of course, if there were individuals interested in drawing a borderline between Buddhism and Brahmanism early on, we can assume that there were others who had no use for such a distinction. How are we to understand these two populations and the negotiations between them?

Some might want to argue that the doctrine of anātman does mark the borderline between Buddhism and other traditions for Vasubandhu in the fourth century CE. Vasubandhu rather famously opens the last chapter
(or appendix) to his Abhidharmakośabhāṣya with the question, "Is it the case that there is no liberation elsewhere?" (i.e., outside of Vasubandhu's own tradition). He answers, "There is not. Why? It is due to being invested in the false view of ātman."3 La Valée Poussin translates anyatra ("elsewhere") here as "apart from this doctrine (dharma)" and then inserts,
"apart from Buddhism," which was not in the original text (Pradhan 1975, 461).4 But Vasubandhu ascribes the ātman view to three opponents: the Vatsīputrīyas, the Grammarians, and the Vaiṣeśikas. Clearly, Vasubandhu understands the Vatsīputrīyas to be Buddhists, since all his arguments against them appeal to identifiable Buddhist sutras. Furthermore, he argues that if this opponent denies the authority of these teachings, he will cease to be, "a son of the Śākyan."5 This line of argument would be compelling only to someone who thought of themselves as Buddhist. So whatever line Vasubandhu is drawing here, it is not the same line we are trying to draw between Buddhism and Brahmanism. Sects such as the Vatsīputrīyas are often assumed to be a later development that are held to be heretical by the likes of Vasubandhu precisely because the borderline between Buddhism and Brahmanism had already been in the earliest Buddhist texts, the Tipiṭaka. I will be arguing that it is not—not even close.

A second objection to my argument might be that even if I am correct, the place for this argument is in a monograph on early Buddhism. Introductory textbooks on Buddhism should be allowed a degree of oversimplification to better introduce the admittedly tricky concept of selflessness. My problem with this particular oversimplification is that it presents "religions" as if they are armies battling one another. The narrative that I hear from undergraduates after they have read introductory textbooks goes something like this:
"Buddhism splits off from Brahmanism and fights with it for converts and patronage until it is finally defeated by Islam." This narrative is not only flawed historically,6 it naturalizes the idea that "world religions" are categorically incompatible and incommensurate, which in turn naturalizes the expectation that violence between religions is the natural, default state.

Shackled to the transcendence or at least the presumed integrity of our categories, it is only peaceful cohabitation and boundary blurring that requires explanation. In light of all the mischief precisely these assumptions have caused in colonial history, is this really the impression that we want to leave the next generation of religion scholars? A more nuanced attention to the way discursive traditions work will show that religious distinctions (I prefer "distinguishings") are actions: they are things people either do or don't do. They are not mere facticities or immutable properties of a religion, but are rather produced by specific human decisions made within nexuses of power in which these distinctions have very real and concrete stakes. Although this is no less true today, these stakes and their relation to the unfolding of tradition are perhaps more clearly interrogated in relation to the problem of traditions and their origins.7 The mistake is to treat the distinctions marking those nexuses of power as somehow natural to something called "Buddhism" opposed to "Brahmanism" (and now "Hinduism").

Here a parallel can be drawn to Christianity and its distinction from Judaism. Daniel Boyarin has argued that there was no real borderline between Judaism and Christianity until Justin Martyr wrote the Dialogue with Trypho in the second century. In the Dialogue, the main speaker states that the difference between Judaism and Christianity is that Christians believe in the Logos and Jews do not. Boyarin points out that this would have been news both to the Jews who had believed in the Logos for generations as well as to the Christians who had never adhered to the Logos doctrine.

Once Justin Martyr made this distinction, Judeans had to choose between identities: either Jewish or Christian—except of course for those who still (or even now) improperly recognized the borderline, and who were labeled "heretics." The first independent Christian communities distinguishing themselves from Judaism are found in Syria where Justin was writing. The idea that there was a borderline to be recognized between Judaism and Christianity spread to the rest of the Mediterranean only later. Of course, once the idea that there was a border reached a certain level of geographic and demographic saturation, subsequent generations would come to read the Jewish/Christian distinction back into much earlier sources, especially into the Gospels and the letters of Paul (see Boyarin 2004, esp. chapters 2 and 3).

Needless to say, there are Buddhist scholars who also assume that the borderline between Buddhism and Brahmanism dates back to the Buddha himself in a similar fashion to the way some Christian scholars read the break from Judaism back into the words of Jesus. Under this scenario, every Brahmin steeped in the Vedas and Upaniṣads who comes to follow the Buddha would have to be considered a "convert." But religious boundaries are like national boundaries. Sometimes individuals cross borders, but borders also cross people—especially in the early days of nation-building.

Buddhism And Brahmanism In Western Scholarship

The question of Buddhism's relation to Brahmanism has been debated by Western scholars since the middle of the eighteenth century. For over two hundred years, opinion has gone back and forth between assuming that Buddhism was opposed to Brahmanism and assuming that Buddhism was merely a flavor of Brahmanism. Gyanendra Pandey has written about the trend among nineteenth-century British colonial officials to see religion not only as a governing category of subject peoples but to assume therefore that religious difference was a rationale for violence that required no further explanation (Pandey 1990, 23–65). But more than two decades before the advent of the Hindu-Muslim "communal riot narrative" that Pandey documents, Sir William Jones addresses the Asiatic Society of Bengal with a similar observation about the relationship between Buddhism and Hinduism:8 The Brahmans universally speak of the Bauddhas with all the malignity of an intolerant spirit: yet the most orthodox among them consider Buddha himself as an incarnation of Vishnu. This is a contradiction hard to be reconciled, unless we cut the knot, instead of untying it, by supposing with Giorgi, that there were two Buddhas, the younger of whom established the new religion, which gave so great offence in India, and was introduced into China in the first century of our era. (Jones 1799, 123)
It would seem, then, that Jones encountered this enmity of Brahmins for Buddhists among the Brahmins he encountered in Calcutta. But if that is the case, who were the Buddhists that had so incensed those Brahmins and offended India? Jones doesn't say. It is possible that he encountered distain for the Buddhists only in the Nyāya texts that he was studying. Or perhaps he is referring to tales that were then circulating about past violence among these communities such as those recorded a decade later by Lieutenant Francis Wilford (Verardi 2011, 22). The only living Buddhists he refers to are Tartars, Tibetans, and Chinese. And yet Jones also notes that the same Brahmins also simply assume the Buddha to be an incarnation of Vishnu—so not only was there no enmity in certain contexts, but no borderline either. This situation seems to have puzzled Jones, though not the Brahmins he encountered.

From very early on, then, scholars had to choose between seeing the relation between Buddhism and Hinduism as one of enmity or identity. A little over fifty years later, Albrecht Weber takes up the thesis of identity in his History of Indian Literature (originally published in 1852). He sees Buddhism as "originally of purely philosophical tenor, identical with the system afterwards denominated the Sāṃkhya, and that it only gradually grew up into a religion in consequence of one of its representatives [presumably the Buddha] having turned with it to the people." He furthermore notes that, "The doctrines promulgated by Yājñavalkya in the VṛihadĀraṇyaka [sic] are in fact completely Buddhistic, as also are those of the later Atharvopanishads belonging to the Yoga system" (Weber [1875] 1882, 285). He argues that the identity between Buddhism and a number of Brahmanas and Upaniṣads is a result of the fact that both can be traced to the same region. Nevertheless, Weber (anticipating more recent work by Bronkhorst [2007]) also observes that while Buddhism had a number of philosophical links with Brahmanism, it originated in a kind of cultural frontier region. Postulating an indigenous culture separate from Āryan culture, he argues that cultural difference allowed enough perspective to develop a critique of Brahmanical norms such as the caste system (Weber
[1875] 1882, 286–87).

Yet if Weber saw Buddhism and the Brahmanical Upaniṣads to be philosophically identical, there were others who still assumed a natural hatred between the two. Just a few years after Weber's work was published in English, Rajendralala Mitra wrestles with the question of why early histories of Orissa tell us that Greeks and Mongols were in Orissa when there is no real evidence that they were—all the while omitting any mention of the Buddhists who were there. His sole explanation for this silence is communal hatred:
As already shown… during the four centuries preceding the Christianera, Orissa generally and the district of Puri in particular were under the domination of the Buddhists; but they abstain altogether from any reference to them.

It is impossible to suppose that they [the dynastic authors] knew nothing of the ascendancy of Buddhism, and the omission, therefore, can be attributed solely to religious hatred. They would do anything to avoid naming the Jains and the Buddhists; as the old adage has it "they would rather be eaten up by tigers than seek shelter in a Jaina temple." (Mitra 1880, 104)9 Of course, we might reasonably question whether silence is indeed an index of hatred, but it is also possible that Mitra's sources are not as silent as we might think. Jones is clearly talking to Brahmins who express antipathy toward Buddhists. Jones assumes that the Buddhists must be a separate group from the Brahmins who universally malign them. But what if they are not? Even as late as 1879, Leopold von Schroeder notes that, "…even now Maitrāyaṇīya Brahmana's who live at the foot of the Vindhyas at Bhaḍgaon, do not eat with those other Brahmaṇas; the reason may have been the early Buddhist tendencies of many of them." (von Schroeder 1879, 30).10 In other words, it is possible that Jones didn't see any Buddhists apart from Brahmins, because the only Buddhists around were Brahmins. This would certainly not be unheard of. At the beginning of the seventeenth century, a Jesuit, Roberto De Nobili, travels to Madurai and encounters a group of Buddhist Brahmins wearing their sacred threads:
A third reason to prove that the Brahmin cord is not tainted by superstition but is merely a badge of learning in any field whatsoever, is taken from the practice of the Buddhist Brahmins, alias the class of unbelievers, of whom we have spoken in chapters Three and Four. For these men are wont to wear the badge of the thread from time immemorial; and they flourished in these parts very long before those laws and books of the Idolaters were introduced. (Rajamanickam 1972, 91)11 Later, in 1889, Sir Monier Monier-Williams argues that the Buddha
"himself was a Hindu of the Hindus and he remained a Hindu to the end…. It is true he was a determined opponent of all Brāhmanical sacerdotalism and ceremonialism, and of all theories about the supernatural character of the Vedas… but being himself a Hindu he never required his adherents to make any formal renunciation of Hinduism, is if they had been converted to an entirely new faith…" (Monier-Williams 1889, 71). As for the doctrine of selflessness, Monier-Williams, like Weber, holds it to be one of the "mere modifications of the Sāṃkhya, Yoga and Vedānta systems of philosophy" (Monier-Williams 1889, 93) In this case, he argues that the denial of a self is merely the corollary to the Brahmanical thesis that nothing exists except for the "supreme Universal Spirit" (that is, only the Universal Spirit exists, therefore the self does not exist) (MonierWilliams 1889, 105–120).

Vincent Smith follows up on this work in his second volume of the Cambridge History of India generalizing the Sāṃkhya/Buddhist connection of his predecessors to state that, "When [the Buddha] died, about 487 B.C., Buddhism was merely a sect of Hinduism, unknown beyond very restricted limits, and with no better apparent chance of survival than that enjoyed by many other contemporary sects now long forgotten"
(Smith 1906, 170).12 What Smith adds to the discussion is the idea that Buddhism's spread lay in the Buddha's opposition to "bloody sacrifices," which happened to dovetail nicely with the policies of Aśoka Maurya. Thus, for Smith, Buddhism is basically a detour for "Hinduism," which eventually returns to the main highway with the advent of Mahāyāna Buddhism.

When we get to Charles Eliot's three-volume Hinduism and Buddhism
(1921), we find something of a disjointed picture of the relation between the two religions. On the one hand, apparently unaware of Weber's earlier discussion of the geography of Vedic schools, Eliot claims, "At [the time of the Buddha] the influence of the Brahmans had hardly permeated Bihar, though predominant to the west of it, and speculation there followed lines different from those laid down in the Upanishads, but of some antiquity, for we know that there were Buddhas before Gotama and that Mahāvīra, the founder of Jainism, reformed the doctrine of an older teacher called Parśva" (Eliot 1921, 8). Despite the apparent mutual isolation during the time of the Buddha, the image that Eliot uses most often is that of two streams running parallel down through history, which in the end (at least in India) reunite: "For twelve centuries or more after the death of this great genius Indian religion flows in two parallel streams, Buddhist and Brahmanic, which subsequently unite, Buddhism colouring the whole river but ceasing within India itself to have any important manifestations distinct from Brahmanism" (Eliot 1921, 292). But the contact for Eliot is not limited to the beginning and end of Indian Buddhism. The two cross-pollinate throughout their histories, and in cross-pollination, Mahāyāna Buddhism is especially singled out as marked by a Brahmanical infusion.

In the twentieth and now the twenty-first century, we notice that many of these issues are still alive and well. Arguing that Buddhism is really just a flavor of Hinduism seems to have become passé in Euro-American scholarship, although it is not hard to find many in India who assume this to be true. Recent scholarship tends to assume a natural antipathy between the two. To pick just one example, Sue Hamilton quotes G. P.

Malalasekera's 1957 The Buddha and His Teachings in the beginning of her 1996 book, Identity and Experience: The Constitution of Human Experience, as follows:
This is the one doctrine which separates Buddhism from all other religions, creeds, and systems of philosophy and which makes it unique in the world's history. All its other teachings … are found, more or less in similar forms, in one or other of the schools of thought or religions which have attempted to guide men through life and explain to them the unsatisfactoriness of the world. But in its denial of any real permanent Soul or Self, Buddhism stands alone. This teaching presents the utmost difficulty to many people and often provokes even violent antagonism towards the whole religion. Yet this doctrine of No-soul or Anatta is the bedrock of Buddhism and all the other Teachings of the Buddha are intimately connected with it… Buddhism denies and asserts that this belief in a permanent and a divine soul is the most dangerous and pernicious of all errors, the most deceitful of illusions, that it will inevitably mislead its victim into the deepest pit of sorrow and suffering. [emphasis mine] (Hamilton 1996, xv–xvi)
Hamilton cites Malalasekera, no doubt, because his statement as a
"distinguished modern Theravada Buddhist," lends an indigenous voice in support of the relevance of her project. But in citing him, she misses the context of Malalasekera's statement and merely takes for granted that both the boundary he draws with selflessness and the presumption of the violence that it (naturally?) engenders mark the entire category of Theravada Buddhism13 and not just the specific conjuncture that Malalasekera's 1957 work occupied (here Malalasekera was far more concerned with Buddhism's distinction from and conflict with Christianity than Hinduism).14 Not that Malalasekera was unique in stating this in the 1950s; just two years prior in India, T. R. V. Murti would write:
Indian philosophy must be interpreted as the flow of these two vital streams, one having its source in the ātma doctrine of the Upaniṣads and the other in the anātmavāda of Buddha. Each branched off into several sub-streams with a right and a left wing and several intermediary positions. There were lively sallies and skirmishes, but no commingling or synthesis of the two streams. Throughout the course of their development they remain true to their original inspirations. The Brāhmanical systems took the real as Being, Buddhism as Becoming; the former espoused the universal, existential and static view of Reality, the latter the particular, sequential and dynamic; for one space, for the other time, is the archetype. The Brāhmanical systems are relatively more categorical and positive in their attitude (vidhimukhena), while the Buddhists were more negative (niṣedhamukhena). Again, the former are more dogmatic and speculative, the Buddhists empirical and critical. Subjectively minded, Buddhism is little interested in cosmological speculations and constructive explanations of the universe. The Brāhmanical systems were bound to an original tradition; they all accepted the authoritarian character of the Veda. Buddhism derives its inspiration from a criticism of experience itself. (Murti 1955, 12)
In Murti's treatment, Buddhism and the teachings of the Upaniṣads are two sides that have "sallies and skirmishes, but no commingling or synthesis." They form polar opposites and can never be reconciled.

But if Malalasekera draws the boundary between Buddhism and Christianity with the selflessness doctrine, other scholars assume Buddhist doctrine to incite antipathy from a different "other": Brahmanism. Perhaps the most striking recent example is that of Giovanni Verardi, who takes this thesis so far as to say that the cultural differences between the Buddhist and Brahmanical cultures were so great in the first millennium CE that they actually motivated widespread pogroms of Buddhists by Hindus—leading to the eventual disappearance of Buddhism from the Indian subcontinent. For Verardi, the reason for the ongoing violence between Buddhism and Hinduism and for the eventual eclipse of the former lies largely in the contrasting natures of the two. He sums up his thesis as follows:
The historical domain covered by this book is thus one where an antinomial model takes the shape of a religious system, Buddhism, which is bound, by ideology and violence, from within and without, to renegotiate continuously and dramatically its own antinomial position…. The antinomial stance of early Buddhist thought and early Buddhist communities condemned them to the impossibility of emerging out of their subaltern positioning throughout the whole of ancient and medieval history.

(Verardi 2011, 12)
Verardi then rather awkwardly attempts to paint those who argue that Buddhism is somehow identical to Brahmanism as merely playing into the hands of the Hindu right wing (Verardi 2011, 44–54). Verardi's argument relies on identifying the transhistorical essence of Buddhism and Brahmanism to render their contrast as the ubiquitous condition for violence. Verardi, I think, nicely summarizes some of the differences between Buddhism and Brahmanism that are assumed by many scholars:
Many misconceptions formed over time on Buddhism and its position towards Brahmanism depend on assessments made on the relationship between early Buddhism and the world of the Upaniṣads. The phenomenological aspects common to the two systems conceal the profound difference that separates them.…
At the phenomenological level, we observe [in Buddhism] anti-cosmism (the phenomenal world is sorrow, and the causal chain keeps men bound to it) and dualism (saṃsāra is in sharp contrast with nirvāṇa, however the dynamics between the two came to be represented). The subordination of the divine is another typical feature of early Buddhism… The gods for the Buddhist are not relevant with respect to either the teaching, revelation and attainment of prajñā. Antinomianism is expressed by the refusal to recognize that caste division is sanctioned from above; it has, if ever, only a de facto validity. Destructuring the self, Buddhism tends to destructure society. Here then is an anti-clerical position: the brāhmanas, from whose ranks the priestly class comes… have no rights deriving from birth. The true brāhmaṇa is whoever enters the spiritual elite founded by the Buddha. The opposition to rituals, especially bloody ones, has the aim of emptying the sacerdotal functions. Another shared feature is individualism: while jñāna or gnosis is a means of individual salvation valid for the āryas… an individualist ethic informs as well the upāsakas… As for universalism… the Buddha's preaching is freed from social, ethnic, national and even religious ties. (Verardi 2011, 69–72)
After over two centuries of scholarship on Buddhism, then, we have a genealogy of two competing claims that find their way into general treatments of Buddhism: the first is that Buddhism is a form of Hinduism and the second is that Buddhism (and particularly its doctrine of anātman)
is pitted against Hinduism. But how do these theorists justify their picture of what Buddhism is in the first place? Verardi abstracts an essence of Buddhism and "the world of the Upaniṣads" without showing us how he comes up with this essence. Presumably, he has relied on the mountain of studies of early canonical material, which would seem natural enough. So where do scholars of early Buddhism get their understanding of what Buddhism is? Some, such as Richard Gombrich, ascribe to a kind of auteur theory, in which the figure of the Buddha becomes not only the (singular) "root or trunk of all subsequent Buddhist diversity," but a source who was "startlingly original" (Gombrich 2009, 2). This perceived originality is pitted against "the fact that the Buddha's main ideological opponents were Brahmins of Upaniṣadic views…." The coherence and uniformity of Buddhist doctrine is thus vouchsafed by being anchored in a single author, the Buddha. This Buddha lived in a thoroughly Brahmanical culture but resisted it either through straight out denunciation, or, when he appears to be supporting it, through "metaphor and irony, registers imperceptible to the literal-minded" (Gombrich 2009, 2). Gombrich, rather anachronistically, argues that all of the Buddha's apparent "capitulation" to Brahmanical ideas is merely evidence of his "skill in means" (a reference to the later Mahāyāna doctrine of upāya)—a term that he acknowledges is altogether absent from early Buddhist sources, even early Mahāyāna ones (Gombrich 2009, 7). But I don't single out Gombrich because he stands out, but because he represents a trend. The most famous example of this trend is that of T. W. Rhys Davids who describes the Buddha's "invariable method… when discussing a point on which he differs from his interlocutor":
When speaking on sacrifice to a sacrificial priest, on union with God to an adherent of the current theology, on Brahman claims to superior social rank to a proud Brahman, on mystic insight to a man who trusts in it, on the soul to one who believes in the soul theory, the method is always the same. Gotama puts himself as far as possible in the mental position of the questioner. He attacks none of his cherished convictions. He accepts as the starting-point of his own exposition the desirability of the act or condition prized by his opponent… He even adopts the very phraseology of his questioner. And then, partly by putting a new and (from the Buddhist point of view) a higher meaning into the words; partly by an appeal to such ethical conceptions as are common ground between them; he gradually leads his opponent up to his conclusion. This is, of course, always Arhatship…. (Rhys Davids 1899, 206)
With this hermeneutic, even if the Buddha appears to be saying something other than the party line, we know that he is not actually doing so.

He is merely engaging in skillful means. Everything that does not fit can then be read as a strategy to lead the interlocutor to the opposite of the ostensive meaning of the text and the coherence of Buddhist doctrine across the canon is preserved. The problem, of course, with a thesis in which all counter-evidence is treated as evidence for the thesis is that the thesis is no longer falsifiable, rendering it suspect at the very least.

Now many authors avoid explicitly connecting a set of doctrines to the Buddha himself, but rather ascribe specific doctrines or sets of doctrines to generic "Buddhists" or occasionally to a reified "Buddhism." My point is that whether any given author asserts Buddhism to be the same as Brahmanism or a religion opposed to it, the disjunctive "or" requires us to take the two poles of the disjuncture as homogenous units.

In light of the fact that this debate has gone on for so long, what do I think I can contribute to the discussion? Whether one thinks that Buddhism and Brahmanism are the same or different, evidence is usually marshalled in the form of a close reading of a selection of sutras that are then synthesized to either characterize the whole ("Buddhism") or characterize an author ("the Buddha"). Murti explains it like this:
Passages must not be counted, but weighed. We must consider the entire body of texts together and evolve a synthesis, weighing all considerations.

We require a synoptic interpretation of the Buddhist scriptures. It is necessary to make a doctrinal analysis of the contents and assess philosophically their value. Such syntheses of doctrines and texts have been made from time to time by the Buddhist schools themselves. We need consider only three such important syntheses done by the Vaibhāṣika and the Sautrāntika, the second by the Madhyamika and the last by the Yogācāra. Each is an attempt to reconcile all the texts and doctrines from a definite point of view. In spite of the specific differences they exhibit, they have a generic affinity that is particularly Buddhistic. (Murti 1955, 12)
The problem, of course, is that there are many competing versions of Buddhism that one could synthesize from the collection of suttas as they stand. Scholars, both ancient and modern, who wish to cast Buddhism as
"anti-Brahmanism" must go to some effort to subordinate apparent contradictions to the dominant theme of their synthesis.

Of course, not everyone has been comfortable with this method or the synthetic unity that was the intended result of its hermeneutic. In 1935, Stanislaw Schayer asserts that the very idea that a text was not to be read at face value (that it was neyārtha instead of nītārtha) is a later scholastic invention, and that scholars should take each sutra at face value to properly evaluate what is going on in canonical texts (Schayer 1935, 124). This makes some sense, especially because the Tipiṭaka as a whole cannot be taken as the primary unit of authority before it was completed and understood to be an enclosed unit. In other words, maybe we should be counting sutras and eject our inherited academic synthesis of Buddhism—at least initially, so that we can begin to really listen to our Buddhist sources. Only then will we be able to recognize when, where, and why Buddhists do their own synthesizing of tradition to create coherent boundaries… or not to create them. I remind my students often that "Buddhism" doesn't teach anātman—or anything else for that matter. "Buddhists" do or "suttas" do, which also means that some Buddhists and some suttas don't.

What I want to argue here is that it matters which ones do and which ones don't, because when we look at the extant suttas in the Tipiṭaka, we are not looking at a homogenous pronouncement of an atemporal source, but rather the result of a long series of debates and negotiations between multiple communities over a period of some time. So instead of dismissing or interpreting contradictions away, I suggest we pay attention to them. It is in the conflicts in the canonical accounts that we can catch a glimpse of early debates that were not our own and allow for a critique of the very ideas of Buddhism and Brahmanism to entertain an "insurrection of subjugated knowledges" (Foucault 2003, 7).15

The Devil And The Details

So what do our early sources tell us about the borderline between Buddhism and Brahmanism? To get a better sense of the canonical terrain on this issue, I compiled a spreadsheet of all the discourses (sutta) of the Pali versions of the Dīgha, Majjhima, Saṃyutta, and Aṅguttara Nikāyas along with the Dhammapāda and the Suttanipāta and Udāna to see the distribution of discourses discussing or broaching the doctrine of anatta/anātman. If you count all the repetitions, this amounts to 5,126 discourses or potential discourses—an admittedly soft number considering the nature and number of repetitions involved. Although my sample does not comprise all the Pali canon, I believe that the distribution of doctrines in this sample as well as the caste of the interlocutors is representative enough of the whole.16 Of course, there are variants of most of these discourses preserved in other languages, but in order to not introduce possible errors or double-counting from what are essentially translation mistakes (or multiple translations of what may have been the same manuscript), I have kept to one language.

Counting discourses in this way was a bit tricky, since I had to make a judgment as to what counts as a discourse. Most discourses were obviously independent discourses, but some were either full repetitions of discourses found elsewhere (presenting the problem of double counting) or some were peyyālas or mere key words or phrases to discourses peppered with the Pali equivalent of an ellipsis (… pe…) to indicate that the rest of the discourse is to be filled in according to a previous paradigm, substituting the titular doctrine in the key place of the paradigm.17 Modern translators tend to ignore peyyālas, because they do not offer any new doctrinal content. However, the fact remains that they are there in the Tipiṭaka and someone not only thought they were important enough to insert into the canon, but generations of Buddhists for the past two thousand years have thought it important to keep them in there. On the one hand, an argument can be made to exclude these from my survey, because they are not really unique discourses and tend to inflate the importance of certain doctrines. On the other hand, it is precisely the fact that they inflate the appearance of certain doctrines in the Tipiṭaka that I chose to count them. Somebody at some point thought it was important to beef up, as it were, the number of discourses on things like anatta. Apparently, we are not the only ones who were counting discourses.

If we take each discourse at face value and include all repetitions, the doctrine of anatta (= Skt. Anātman) or some version thereof appears in about 378 of these 5,126 discourses (about 7.4%). This appears to be on par with other doctrines of liberation. By comparison, the Noble Eightfold Path occurs in 368 of these discourses, and the jhānas (or other states of absorption) are taught in 340 discourses. Surprisingly, however, only sixteen discourses18 contain both the doctrine of anatta and a discussion of the Eightfold Path, and only twenty-five19 contain a discussion of both jhāna and anatta.

The first thing to note here is that while the Tipitaka does teach the doctrine of anatta, it does so less than 10% of the time. Muslims eat burritos in North America, but the statistical frequency of this in other ethnic populations prevents us from jumping to the conclusion that burritos are iconically "a Muslim thing" in American culture. I will grant that the 7.4%
occurrence of anatta in the Tipiṭaka is significant in contrast to the relative absence of the term in non-Buddhist texts.20 Nevertheless, at 7.4% we can hardly say that the teaching of anatta is more prominent than other doctrines conducive to liberation (such as concentration). If 7.4% discuss anatta, then 92.6% of discourses in the sample never mention it at all. What this means is that potentially someone could listen to a great many Buddhist suttas and never hear a discussion about anatta. Furthermore, the doctrine of anatta could theoretically be a representative Buddhist doctrine even at 7.4% if a community of Buddhists had access to the whole Tipiṭaka. But this was assuredly not the case during the generations in which the canonical texts themselves were being composed. The question then is how much of this canonical material should we assume a given community would have access to? Might it be the case that different communities had access to and circulated different collections with different doctrinal compositions prior to the final redaction of the Tipiṭaka as a whole? If this were the case, then instead of seeing Buddhist doctrine as a whole marked by the doctrine of anatta in opposition to Brahmanism, we should be open to the idea that prior to the Tipiṭaka's final form, there were Brahmanical Buddhist communities and communities opposed to Brahmanism (as well as communities indifferent to it), each composing and circulating their own discourses as "Buddhism."
With this possibility in mind, let us return to the question of whether the suttas support the claim that the Buddhist doctrine of anatta was understood to exclude Brahmanical communities. Yes, but not really in the way one might expect. And the opposition that we do find certainly does not support the contention that there was a recognized borderline between Buddhism and Brahmanism in the Tipiṭaka, or that the doctrine of anatta was preached against Brahmanical or Upaniṣadic doctrines. The devil, of course, is in the details.

If the doctrine of no-soul was supposed by the authors of the suttas to challenge the Brahmanical doctrine of the soul, then we should expect to see this challenge represented somewhere in the canon. Somewhere we should find a representation of the Buddha or some paradigmatic non-Brahmin Buddhist refuting the doctrine of an eternal soul espoused by someone coded as Brahmanical. Now, what scholars today mean by "Brahmanical" or "Brahmanism" appears to mean anyone influenced by the Vedas, and more importantly their ancillary literature: the Brahmanas, Āraņyakas, Upaniṣads, Gṛha- and Śrautasūtras. One did not have to be a Brahmin to learn these works or to practice them.21 Nevertheless, when issues of the Vedas or Upaniṣadic teachings are addressed in the Tipiṭaka, the chosen representative of that position is invariably someone who is in fact a Brahmin by birth. Hence, if the doctrine of anatta was targeting Brahmanism, we should expect to find some representative Brahmin abandoning this belief and accepting the doctrine of anatta, thereby "converting" to Buddhism.

With this in mind, I set out looking for a sutta in which somebody tries to convince a Brahmin that there is no soul. Generally speaking, Brahmins are not too hard to spot in the Tipitaka, and I think we can safely assume that Brahmanical status would have been even more obvious for the intended audiences of early canonical texts than it is for us. A character named Mahākassapa or Bhāradvāja would have been a walking Brahmanical signifier in much the same way as a character named
"Moishe Finklestein" might index Judaism for us today. Of course, some interlocutors have names that are less easy for us to recognize, but often there are other indications that the character was meant to be read as Brahmin. Sometimes Brahmanical status will feature as a narrative element in the sutta or else the Brahmanical status of the interlocutor will be reflected in forms of address (e.g., only Brahmins and devas address the Buddha with the vocative "Gotama").

With this in mind, I went through the set of 5,126 sutras that discuss anatta four times, recording the name and caste of every interlocutor in every discourse. What I discovered was interesting.22 First of all, while there are Brahmins represented as holding the doctrine that the soul is eternal, no sutra claims Brahmins to have a special relation to this doctrine, and in fact there are a few sutras that depict Brahmins as holding a wide range of beliefs concerning the soul, including the belief that when the body dies there is nothing left (e.g., Brahmajālasutta of the D.N. and the Pañcattayasutta of the M.N.). But to say that the doctrine of anatta has nothing to do with Brahmins is not quite right either. When we single out discourses in which the interlocutor actually is a Brahmin, the situation gets more complicated. Of the 5,126 discourses, there are 648 discourses delivered to recognizable Brahmins (roughly 12.6% of the sample). Of these, 195 discourses are delivered to Brahmin householders or to Jaṭila ascetics performing sacrifices in the woods. In these 195 discourses the doctrine of anatta is never mentioned or broached even once. Not that there are many discourses in which the doctrine is taught to a non-brahmin householder: there are 194 discourses to named non-brahmin lay interlocutors of which seven contain a teaching that could be understood as anatta.

23 Of course, the authors of these suttas were well acquainted with their nonfictional Brahmin friends and neighbors in a way that most modern scholars are not. Despite considerable literature on the Buddha's so-called
"redefinition of the Brahmin,"24 most Brahmins in these suttas are simply assumed to have been born as Brahmins. Just as there is nothing to prevent a contemporary Brahmin from being a devout bhaktin or staunch Marxist, our early Buddhist authors knew that individual Brahmins could and did believe many things. There was certainly nothing that prevented a Brahmin from embracing the doctrine of anātman or nirātman as a variety of Brahmanical belief.25 Nevertheless, if we can assume (and I think that we can) that the social dynamics represented in the suttas reflect the social assumptions of their presumed audiences, this absence of the anatta doctrine in the presence of Brahmin laity is remarkable and suggests some kind of avoidance regarding Brahmin laity when it comes to the doctrine of anatta. So, yes, it does appear that the doctrine of anatta was understood to be confrontational to Brahmins, but also that there was some effort devoted among our sutta writers to not being confrontational. If anything, the tacit message to Buddhist preachers is: don't bring up that doctrine when preaching to Brahmins. Why? They might not become followers of the Buddha if they thought "Buddhism is the doctrine of anatta." And with only 7.4% of sutras even mentioning the doctrine, there were lots of other options for what to preach to Brahmins.

But if Buddhist preachers avoided the doctrine of anatta to accommodate Brahmins within an emerging Buddhism (or, alternately, if Brahmin preachers of Buddhism avoided the doctrine of anatta so as to not to contradict the Upaniṣads circulating within their communities),
then this avoidance just as surely does not support the idea that the doctrine of anatta formed the assumed borderline between Buddhism and Brahmanism. On the contrary, the assumption seems to be that a special place was to be carved out within Buddhism for Brahmins as Brahmins. As Ryutaro Tsuchida (1991) pointed out long ago, the fact of the matter is that the Buddha does teach to Brahmins, most of whom take refuge in him or otherwise express appreciation for his teachings. And even when he tangles with a cranky Brahmin, the Brahmin is usually depicted as being a young disciple of a Mahāsāla Brahmin who, in turn, chastises his disciple for being rude. Some groups of Brahmins were even granted special status due to the perceived propinquity between jaṭila who perform the agnihotra and Buddhist monks. For example, in the Pali Vinaya, there is a passage discussing the probationary period (parivāsa) that a member of another sect must wait before ordination. Toward the end of the section it states:
If fire-worshippers and Jaṭilas [alt. "fire worshipping jaṭilas"] come to you, O Bhikkhus, they are to receive the upasampadā ordination (directly), and no parivāsa is to be imposed on them. And for what reason? These, O Bhikkhus, hold the doctrine that actions receive their reward, and that our deeds have their result according to their moral merit. If a Sakya by birth, O Bhikkhus, who has belonged to a Titthiya school, comes to you, he is to receive the upasampadā ordination (directly), and no parivāsa is to be imposed on him. This exceptional privilege, O Bhikkhus, I grant to my kinsmen. (Rhys Davids and Oldenberg 1899, 190–1)26 This passage, which at first appears only in Theravādin texts but begins to be known in northern texts by about the fifth century,27 asserts the distinction between Buddhists and agnihotrins to be one of close affinity based on a common belief in karma. Apparently, in some Buddhist schools at least, certain schools of Brahmanic ascetics were granted privileged status due to a perceived propinquity to Buddhist asceticism.

So what are lay Brahmins taught if not anatta? Among the discourses involving Brahmin laity, there are twenty-five that discuss the jhānas28 (12.3%), five that discuss the Eightfold path29 (2.5%), one that discusses the formless meditations,30 and six that present the state of nothingness as the highest attainment31 (3%). If we compare this to the 194 discourses involving named non-Brahmin laity, we have the six32 suttas discussed above dealing with the doctrine of anatta (3%), seven teaching the Noble Eightfold Path33 (3%), ten discussing the formless spheres34 (2%), seventeen discussing the jhānas35 (5%), and only one36 that teaches the state of nothingness as the highest state.

Not all Brahmins were lay-brahmins. There are 130 discourses in which the interlocutor is neither a householder nor a Buddhist monk but a "wanderer" or paribbājaka. Of these discourses, there are 115 in which the paribbājaka is identifiable as a Brahmin. Of the 115, there are 60 or roughly 53% that either invoke or have been interpreted to invoke the doctrine of anatta. However, the statistics here are deceiving, since 59 of these discourses are variations of the same discourse delivered to the same person: Vacchagotta. If we factor out the Vacchagotta discourses that broach the subject of anatta, we are left with only two other discourses delivered to Brahmin paribbājakas that discuss anatta (the Susīmasutta [SN 12.70]
and the Anattalakkhanasutta [S.N. 22.59]; N.B. Susīma's Brahmanical status itself is uncertain37). Thus, if we factor out the Vacchagotta repetitions, and count the number of independent sutras touching on anatta delivered to Brahmin paribbājakas we get only two (three if you count Koṇḍañña in the Anattalakkhanasutta).38 This is pretty much on par with what we find among non-Brahmin paribbājakas. There are thirty-six discourses in which the interlocutor is a non-Brahmin paribbājaka and there are only one or two39 in which the doctrine of anatta is broached.

Finally, ordained monks and nuns could also be coded as Brahmins.

The majority of the main disciples of the Buddha discussed in the Tipiṭaka are in fact Brahmins by birth. Sāriputta was one, but Mahākassapa, Mahākaccāna, and Piṇḍola Bhāradvāja in particular are coded as Brahmins by birth. Moggallāna is even addressed by the Buddha with the vocative
"brāhmaṇa" in a few discourses.40 When the Buddha refers to a monk as Brahmin,41 it is unlikely that he is referring to an acquired brahmanical status, since he never addresses a monk born of some other caste with this vocative. Indeed the only monks other than Moggallāna with whom the Buddha uses this form of address are the monk Dhammika42 (who is thoroughly unpleasant) and Aṅgulimāla, the [Gargīya] Brahmin serial killer (albeit reformed).43 Hence, it was not considered unusual for the authors of these discourses that a Buddhist monk would still be known and addressed as "Brahmin" long after ordination,44 while to my knowledge the Buddha never addresses as "Brahmin" any monk who was not born Brahmin, no matter how accomplished.

In contrast to Brahmin laity, there are quite a few discourses involving the doctrine of anatta taught to Brahmin Buddhist monks. Of 712 discourses involving named monks or nuns, 394 discourses involve named Brahmin monks or nuns (153 of which involve Sariputta alone), whereas 309 involve named non-Brahmin monks. Of course, many of the dialogues involving Brahmin monks also include non-Brahmin monks, so there is quite a bit of overlap between these two sets of discourses. If we filter out the discourses in which Brahmins are present (in order not to double count), we get 283 discourses in which a non-Brahmin monk is present when no Brahmin (of any sort) is present. Of the discourses to Brahmin monastics, only twenty-nine45 (or 7.4%) deal with anatta, as opposed to fifty-five anatta discourses (or 18.7%) delivered to named non-Brahmin monks or nuns when no Brahmin is in sight. Of these twenty-nine Brahmin monastic discourses, there are none in the Dīgha Nikāya, one in the Udāna, three in the Majjhima, three in the Aṅguttara,
and a whopping nineteen in the Saṃyutta Nikāya. Of the nineteen in the Saṃyutta Nikāya, nine are iterations of the same discourse delivered to a single monk, Rādha. In all, only ten Brahmin Buddhist monks are ever depicted as being aware of the doctrine of anatta,
46 and only one of these discourses registers any awareness that the soul might be a doctrine outside the pale of Buddhism.47 By contrast, there are eighteen named non-Brahmin monks who are treated to discourses on anatta (Rahula, the Buddha's son hears thirty of these discourses alone).48 That leaves fifteen Brahmin Buddhist monks and nuns inhabiting a total of twenty-nine discourses who are never depicted as having been anywhere around when anatta is taught.49 So, for instance, monks coded as über Brahmins such as Mahākassapa and Piṇḍola Bhāradvāja apparently never heard of the doctrine—at least in the current sample. By contrast there are fifty-eight discourses (14.7%) delivered to Brahmin monks about the jhānas and forty-five (11.4 %) about the Noble Eightfold Path, as opposed to eighteen discourses (6.4%) delivered to or by named non-Brahmin monks about the jhānas when no Brahmin is around50 and sixteen (5.7%) that mention the Noble Eightfold Path delivered exclusively to non-Brahmin monks. Finally, if we group together all the different types of concentration from the jhānas to the formless meditations to the cessation of perception and feeling (and thus avoid double counting), we come up with a total of 116 (29.4%) discourses containing a discussion of absorption delivered to or by a Brahmin monastic while only thirty-two (11%) are delivered to or by a non-Brahmin monastic.

All the above suggests that discourses depicted as involving interlocutors coded as Brahmin lean significantly more toward teachings involving mental absorption and the Noble Eightfold Path than toward the doctrine of anatta, while discourses involving ordained non-Brahmins skew toward the doctrine of anatta, at the expense of mental absorption or the Noble Eightfold Path. The numbers are, needless to say, quite small, and so the usual Chi-square test is not valid (many of the values are less than 5). However, if we do a Fisher's Exact test, we find that the table probability is less than .0001 for the association between Brahmins and jhānas and non-Brahmins and anatta.

51 In other words, the apparent skewing is, in fact, statistically significant. All of this suggests a general attitude among our sutta authors that the absorptions and the Noble Eightfold Path were considered more appropriate for Brahmin Buddhist monks than the doctrine of anatta, and that the doctrine of anatta was generally more appropriate for non-Brahmin Buddhist monks.

In this regard, it may be significant that, as mentioned above, there is very little overlap between discourses that discuss some kind of concentration or absorption and those discourses that discuss anatta. Between the 340 discourses discussing samādhi and the 378 discussing anatta, there are only twenty-five discourses discussing both.52 If we put this together with the fact that discourses including Brahmins skew more toward samādhi practices, then our data suggest that the tension between concentration and insight across the canon generally may in fact index a tension between Brahmin Buddhist communities and non-Brahmin Buddhist communities. In other words, I believe that the trend we see here probably has something to do with the fact that meditations akin to the jhānas and the formless meditations were understood to be appropriate for Brahmins because they were understood to already be Brahmanical.

53 Whether the discourses on anatta actually contradict the teachings of any particular upaniṣadic tradition,54 it appears that some of our sutta authors did not want to take the chance. Every tradition has a functional canon within a canon. It would appear then that there was a functional canon of discourses circulating (though not exclusively) within Brahmin communities and another functional canon circulating among non-Brahmins, and that the distinction between the two canons fell along the lines of samādhi and insight into selflessness. The discourses that deal with both selflessness and samādhi may have been written either at a time and for a community when there was an interest in bringing the two canons together, or else they may have been written in an environment in which the two parts had already been brought together and there was an interest in subordinating one part of the canon to the other.

The Interpretation?

What do we make of all of this? A key point here is that in the sutras favoring Brahmins that appear interested in having Brahmins contribute to Buddhism as Brahmins, there is little evidence for any kind of solid boundary between Buddhism and Brahmanism. This is not to say that there are no anti-Brahmanical (or better yet, "anti-Brahmin") texts, but not many. Presumably the communities that wrote and circulated these discourses, like Justin Martyr's Syrians, had a vested interest in distinguishing themselves from (and denigrating) Brahmins. But also like Justin's Syrians, we cannot assume the texts of this community to stand for some pan-Buddhist perspective on caste. Indeed, if a data driven approach shows us anything, it is that claims that "Buddhism teaches x" are merely a metonymic mask for what should be a set of demographic claims (this community tended to talk about and circulate x doctrines and practices in contrast to that community). It is highly unlikely, for instance, that Brahmins would nod their heads in agreement to the statement made in the Soṇasutta of the Aṅguttara Nikāya that, unlike dogs, Brahmins will have sex with anything that moves.

In most Brahmin-favorable texts, the Buddha's teaching is presented as the fulfillment of the path that his Brahmin interlocutors were already on (despite the fact that some youthful Brahmins may not have recognized that fact), and in discourses delivered to Brahmin audiences, what the Buddha teaches is usually confirmed by Brahmin celebrities as being a correct interpretation of the Vedas. If we couple this information with the fact that discourses delivered to audiences coded as Brahmin are statistically less likely to have included a teaching of anatta and far more likely to have preached an absorption series (albeit new and improved) that Brahmins were already using, then we come to the conclusion that there was, in effect, a canon-within-a-canon that appears to have circulated in Brahmin communities.

What would these Brahmin communities have looked like, and at what point were the texts of these communities composed? Oskar Von Hinüber (2006), in a delightful article, goes a long way toward answering these questions. He begins with the observation that while most Buddhist sutras begin with the standard, "Thus have I heard, at one time the Blessed One was staying at X…," there are a number of suttas beginning with a slightly different introduction, which he calls a "parenthetical." For example, at the beginning of the Nagaravindeyyasuttanta [M.N. 150], we find:
"Thus I have heard. At one time, the Lord, walking on tour in Kosala together with a large group of monks, where there is the Brahmin village of Kosala named Nagaravinda, there he went." (Hinüber 2006, 200)
Drawing on the work of Karl Hoffmann, Hinüber points out that this alternative introduction most likely represents the oldest textual strata of the Pali Canon, since it replicates an Old Persian formula that had been in use by the Acaemenians (Hinüber 2006, 198–99). If we isolate those sutras having this parenthetical formula, a number of interesting questions present themselves. First of all, none of the standard cities (nāgara)
such as Savatthī or Rājagaha are mentioned. Instead, we have the Buddha travelling to and preaching at "market-places" (nigāma) and "Brahmin villages" (brāhmaṇagāma). We know, then, that even in the very earliest phases of Buddhist writing, Brahmin villages were assumed to be one of the important venues for Buddhist preaching. Looking at the locations of the Brahmin villages found in this earliest formula, Hinüber notes that their distribution conforms to what we would expect at a time when Brahmanism was just beginning to spread into eastern India.

Interestingly, nine of the altogether fourteen Brahmin villages mentioned in the Theravāda-Tipiṭaka are situated in Kosala, four in Magadha, and only one in the Malla country. This compares well with the evidence gathered from Vedic literature on the history and geographical distribution of the Vedic schools. As research by M. Witzel has shown, Kosala was at the eastern fringe of later Vedic literature, and the Brahmins there used to study the Kāṇva Śākhā of the Śatapathabrāhmaṇa. These then could well be the very Brahmins traced in ancient Buddhist literature. (Hinüber 2006, 200)
But why is there no mention of a monastery either in the market places or in the Brahmin villages? (Hinüber 2006, 201) If the Acaemenid formula does indeed reflect an early date of composition, is it possible that these sutras were composed at a time before Brahmin Buddhist monks lived in monasteries? Could these Brahmin Buddhist monks have been itinerant or perhaps hosted in the homes of laity? Two suttas in the Saṃyutta Nikāya (S.N 47.4 and 48.51) taking place in the Brahmin village of Sālā bear an introductory formula that is intermediary between the old Persianesque formula and the later standard formula. What is interesting about these discourses is that they take place in a Brahmin village where there is no mention of a monastery, and yet the Buddha's audience is not a group of householders, but "monks." Was the assumption here simply that these Brahmin Buddhist monks lived as Buddhist monks in their own homes—presumably still tending to their household fire?55 The questions are provocative, even if we don't have answers.

While it has become a scholarly commonplace to state that Buddhism was an urban phenomenon, with Brahmanism rooted in the countryside, the archaeological evidence actually points in the opposite direction.

Hinüber's data has the earliest sutras placing the Buddha's discourses in marketplaces, not cities. The closest thing to a city for these early venues would be precisely the Brahmin villages. Indeed, while there are stūpas scattered across the subcontinent from the time of Aśoka, there is no real evidence that Buddhism had any significant presence in urban centers until at least the first century CE and the second or third centuries CE in Gandhāra. According to Pierfrancesco Callieri:
When we consider the sum of archaeological evidence indicating the Buddhist presence in urban settlements of the Northwest, it is striking that there is an almost complete absence in pre-Kuṣāṇa layers and a weak presence during the Kuṣāṇa period. It is worth noting that at Mathurā, evidence from the German excavations at Sonkh also shows a remarkable prevalence of local cults and a limited presence of Brahmanical sculptures. (Callieri 2006, 77)
As Buddhism was being established and circulated among Brahmin communities in the earliest phases of its development, it is significant that unlike many other groups at the time, Brahmin communities are represented as homogenous, tight-knit, and set apart both physically and by rules of commensality from other groups. In a brāhmaṇagama, everyone is assumed to be a Brahmin. In cities, we find a similar balkanization of Brahmins. According to Megasthenes:
The philosophers have their abode in a grove in front of the city within a moderate-sized enclosure. They live in a simple style, and lie on beds of rushes or deer skins. They abstain from animal food and sexual pleasures and spend their time listening to serious discourse, and in imparting their knowledge to such as will listen to them…. After living in this manner for seven-and-thirty years, each individual retires to his own property, where he lives for the rest of his days in ease and security. They then array themselves in fine muslin, and wear a few trinkets of gold on their fingers and in their ears. They eat flesh, but not that of animals employed in labour. They abstain from hot and highly seasoned food. They marry as many wives as they please, with a view to have numerous children. (McCrindle 1877, 99–100, as discussed by Lubin 2013, 31–33)
Although Strabo's term "philosopher" includes both Śramaņas and Brahmanas, his description here probably does not refer to Buddhists, but as Timothy Lubin has argued, probably to Brahmins. For our purposes, what is important about Megasthenes's observations is the fact that these Brahmins lived in a separate Gemeinschaften immediately outside the city walls but with clear ties and functions within the city. The strong network ties engendered by commensality, cohabitation, and consanguinity, not to mention the teacher-disciple relationship of Vedic learning, would have easily accommodated and incorporated the Buddhist teachings on absorption and nirvāṇa.

While it is certainly true that Buddhism came to be distinct from Brahmanism in certain communities at some point, the borderline was never universally recognized. It certainly appears that there were always Brahmins who continued to be recognized for both their caste status as well as their ritual prestige while also self-identifying as Buddhist. In the seventh century Harṣacārita of Baṇa, King Harṣa tells us, "I have heard there is a gentleman, a childhood friend of the deceased Grahavarman of auspicious name, who was a Maitrāyanī and who, in spite of the threefold
(knowledge), was a descendent of Brahmins who dwelled in the unarisen wisdom in the doctrine of the Well-Gone One and who as a youth donned the Buddhist robes."56 Whether one translates vihāya as "distancing himself from" as Cowell and Thomas do, or "in spite of " as I have, we have in Baṇa's work the depiction of an ordained Maitrāyanī Brahmin Buddhist who wears Buddhist robes and practices a form of Buddhism that is coded as a form of Perfection of Wisdom (its samādhi being the
"unarisen wisdom" vidva-anutpanna-samādhi). What is more important here is that Divākaramitra, the Maitrāyanī brahmin, continues to be referred to as a brahmin even after ordination (just like Aṅgulimāla and Moggallāna in the canon).57 Closer to our time, Christiaan Hooykaas
(1973), Gananath Obeyesekere (2015), Justin McDaniel (2013), and Jacques Leider (2005/6) have each written on the ritual role of Brahmins in the courts, temples, and monasteries of Indonesia, Sri Lanka, Thailand, and Burma, respectively.

Conclusion

In the sutras favoring Brahmins that appear interested in having Brahmins contribute to Buddhism as Brahmins, there is little evidence for any kind of solid boundary between Buddhism and Brahmanism marking the former as orthodox and the latter as heterodox. For what it is worth, the work of Vincent Eltschinger (2014, 35–92) shows the (non-Buddhist) Brahmanical authors similarly did not make a strong distinction between orthodoxy and heterodoxy until much later, during the Gupta Dynasty (ca. 320 CE–550 CE). Thus, the idea that one or a handful of doctrines like selflessness distinguish "Buddhism" from other "religions" is less helpful than paying attention to the way doctrines were used and distinguished (or not) among particular communities. Statistically speaking, no one set of doctrines really rises to the top as defining of the genre of Buddhist literature. We therefore need to be careful when we teach surveys of Buddhism not to present Buddhism as if it was ever simply identifiable with a single doctrine—especially anatta.

The second point is that our survey does not show that the doctrine of anatta had anything to do with the border between Buddhism and Brahmanism during the time of the composition of the Tipiṭaka.

Later on it did, for some sects,58 but for reasons that were specific to the nexus of power in those times and places—not the Buddha's. In fact, early evidence for such a border is tendentious at best. Finally, depending on how loosely one wants to define "conversion," I see no evidence in the Tipiṭaka of anyone abandoning one religion to embrace another.

The only transition in identity that I see is laity deciding to follow a particular teacher and some individuals shifting from lay to ordained life.59

References

Abeysekara, Ananda. Forthcoming. "Religious Studies' Mishandling of Origin and Change: Time, Tradition, and Form-of-Life in Buddhism." Cultural Critique. Attwood, Jayarava. 2013. "Siddhārtha Gautama: What's in a Name?" Available at https://www.academia.edu/4866512/Siddh%C4%81rtha_Gautama_Whats_ in_a_Name. Accessed August 22, 2016. Bloomfield, Maurice. 1908. The Religion of the Veda: The Ancient Religion of India (From Rig-Veda to Upanishads). New York and London: G.P. Putnam's Sons. Boyarin, Daniel. 2004. Border Lines: The Partition of Judaeo-Christianity. Philadelphia: University of Pennsylvania Press. Bronkhorst, Johannes. 2007. Greater Magadha: Studies in the Culture of Early India. Leiden: Brill. _ . 2009. Buddhist Teaching in India. Boston: Wisdom Publications. _ . 2011. Buddhism in the Shadow of Brahmanism. Leiden: Brill. Callieri, Pierfrancesco. 2006. "Buddhist Presence in the Urban Settlements of Swāt, Second Century BCE to Fourth Century CE." In Gandhāran Buddhism: Archaeology, Art, Texts, edited by P. Brancaccio and K. A. Behrendt, 60–82. Vancouver: UBC Press.

Chattha Sangayana, Tipiṭaka 4.0. 1995. Igatpuri, India: Vipassana Research Institute. Collins, Steven. 1982. Selfless Persons: Imagery and Thought in Theravāda Buddhism. Cambridge: Cambridge University Press.

Cowell, Edward Byles, and Frederick William Thomas. [1897] 1993. The HarṣaCarita of Bāņa. Delhi: Motilal Banarsidass.

Eliot, Charles. 1921. Hinduism and Buddhism An Historical Sketch. Vol. 1. London: Routledge & Kegan. Eltschinger, Vincent. 2014. Buddhist Epistemology as Apologetics: Studies on the History, Self-Understanding and Dogmatic Foundations of Late Indian Buddhist Philosophy. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Elverskog, Johan. 2010. Buddhism and Islam on the Silk Road. Philadelphia: University of Pennsylvania Press. Foucault, Michel, Mauro Bertani, Alessandro Fontana, François Ewald, and David Macey. 2003. Society Must Be Defended: Lectures at the Colleége de France, 1975-76. New York: Picador. Frauwallner, Erich. 1953. Geschichte der indischen Philosophie. Vol. 1. Salzburg: Müller. Gethin, Rupert. 2007. "What's in a Repetition? On Counting the Suttas of the Saṃyutta-nikāya." Journal of the Pali Text Society 29:365–87.

Gombrich, Richard F. 1988. Theravāda Buddhism: A Social History from Ancient Benares to Modern Colombo. London: Routledge & Kegan Paul. _ . 2009. What the Buddha Thought. London: Equinox. Hamilton, Sue. 1996. Identity and Experience: The Constitution of the Human Being According to Early Buddhism. London: Luzac Oriental.

Harvey, Peter. 1995. The Selfless Mind: Personality, Consciousness and Nirvāṇa in Early Buddhism. Surrey: Curzon Press. Hooykaas, C. 1973. Balinese Bauddha Brahmans. Amsterdam: North-Holland Publishing Company. Jones, Sir William. 1799. "On the Chronology of the Hindus." Asiatic Researches 2:111–47. La Vallée Poussin, Louis de. 1980. L'Abhidharmakośa de Vasubandhu. Brussels: Institut Belge des Hautes Études Chinoises. Leider, Jacques. 2005–2006. "Specialists for Ritual, Magic, and Devotion: The Court Brahmins (Punna) of the Konbaung Kings (1752-1885)." The Journal of Burma Studies 10:159–202. Lubin, Timothy. 2013. "Aśoka's Disparagement of Domestic Ritual and Its Validation by the Brahmins." Journal of Indian Philosophy 41:29–41. Malalasekera, G. P. Dictionary of Pali Proper Names. 1960. London: Published for the Pali Text Society by Luzac & Co. McCrindle, John W. 1877. Ancient India as Described by Megasthenes and Arrian. London: Trübner and Co. McDaniel, Justin. 2013. "This Hindu Holy Man Is a Thai Buddhist." Southeast Asia Research 21 (2): 191–210. Mitchell, Donald W. 2008. Buddhism: Introducing the Buddhist Experience. New York: Oxford University Press. Mitra, Rajendralala. 1880. Antiquities of Orissa. 2nd ed. Calcutta: W. Newman & Co. Monier-Williams, Monier. 1889. Buddhism: In Its Connection with Brahmanism and Hinduism and in Its Contrast with Christianity. London: John Murray. Murti, T. R. V. 1955. The Central Philosophy of Buddhism: A Study of the Mādhyamika System. London: George Allen and Unwin. Obeyesekere, Gananath. 2015. "The Coming of Brahmin Migrants: The Śudra Fate of an Indian Elite in Sri Lanka." Society and Culture in South Asia 1(1): 1–32.

Oetke, Claus. 1988. "Ich" und das Ich: Analytische Untersuchungen zur buddhistich-brahmanischen Ātmankontroverse. Stuttgart: Franz Steiner Verlag Wiesbaden. Pandey, Gyanendra. 1990. The Construction of Communalism in Colonial North India. Delhi: Oxford University Press. Pérez Remón, Joaquín. 1980. Self and Non-Self in Early Buddhism. The Hague: Mouton.

Pradhan, P., ed. 1975. Abhidharmakośabhāṣyam of Vasubandhu. Rev. 2nd ed.

Patna: K.P. Jayaswal Research Center. Rajamanickam, Svarimuthu. 1972. Roberto De Nobili on Indian Customs. Palayamkottai: De Nobili Research Institute. Rhys Davids, Caroline A. F. 1941. Wayfarer's Words. Vol. 2. London: Luzac. Rhys Davids, T. W. 1899. Dialogues of the Buddha. Vol. 1. London: Flowed. Rhys Davids, T. W., Hermann Oldenberg. 1899. Vinaya Texts. The Sacred Books of the East. American ed. vol. 4. New York: Scribner's. Sayers, Matthew. 2013. Feeding the Dead: Ancestor Worship in Ancient India. New York: Oxford University Press. Schayer, Stanislaw. 1935. "Precanonical Buddhism." Archiv Orientalni 7:121–32. Schmithausen, Lambert. 1969. "Ich und Erlösung im Buddhismus" Zeitschrift für Missionswissenschaft und Religionswissenschaft 53:157–70. Smith, Vincent A. 1906. The Early History of India from 600 B.C. to the Muhammadan Conquest, Including the Invasion of Alexander the Great. New York: The Grolier Society, Edinboro Press.

Stuart, Daniel M. 2013. Thinking about Cessation: the Pṛṣṭhapālasūtra of the Dīrghāgama in Context. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien, Universität Wien. Tambiah, Stanley J. 1992. Buddhism Betrayed? Religion, Politics, and Violence in Sri Lanka. Chicago: University of Chicago Press. Tsuchida, Ryūtarō. 1991. "Two Categories of Brahmins in the Early Buddhist Period." Memoirs of the Research Department of the Toyo Bunko 49:51–95. Verardi, Giovanni. 2011. Hardships and Downfall of Buddhism in India. New Delhi: Manohar Publishers. Von Hinüber, Oskar. 2006. "Hoary Past and Hazy Memory: On the History of Early Buddhist Texts." Journal of the International Association of Buddhist Studies 29 (2): 193–210. von Schroeder, Leopold. 1879. Ueber die Mâitrayâņī Samhitā, ihr Alter, ihr Verhältniss zu den verwandten Câkhâ's, ihre sprachliche und historische Bedeutung. Dorpat. Weber, Albrecht., et al. [1875] 1882. The History of Indian Literature. London: Trübner & Co. Wynne, Alexander. 2007. The Origin of Buddhist Meditation. New York: Routledge. _ . 2010–2011. "The Atman and its Negation: A Conceptual and Chronological Analysis of Early Buddhist Thought." Journal of the International Association of Buddhist Studies 33 (1–2): 103–171.