The spread of Sanskrit

A recent publication — Nicholas Ostler’s Empires of the Word (2005) — presents itself in its subtitle as A Language History of the World. Understandably, it deals extensively with what it calls “world languages”, languages which play or have played important roles in world history. An introductory chapter addresses, already in its title, the question “what it takes to be a world language”. The title also provides a provisional answer, viz. “you never can tell”, but the discussion goes beyond mere despair. It opposes the “pernicious belief” which finds expression in a quote from J. R. Firth, a leading British linguist of the mid-twentieth century (p. 20): “World powers make world languages [...] Men who have strong feelings directed towards the world and its affairs have done most. What the humble prophets of linguistic unity would have done without Hebrew, Arabic, Latin, Sanskrit and English, it it difficult to imagine. Statesmen, soldiers, sailors, and missionaries, men of action, men of strong feelings have made world languages. They are built on blood, money, sinews, and suffering in the pursuit of power.” Ostler is of the opinion that this belief does not stand up to criticism: “As soon as the careers of languages are seriously studied — even the ‘Hebrew, Arabic, Latin, Sanskrit and English’ that Firth explicitly mentions as examples — it becomes clear that this self-indulgently tough-minded view is no guide at all to what really makes a language capable of spreading.” He continues on the following page (p. 21): “Evidently, total conquest, military and even spiritual, is not always enough to effect a language change. [...] consider Sanskrit, taken up all over South-East Asia in the first millennium AD as the language of elite discourse, even though it came across the sea from India backed by not a single soldier.” What then, according to Ostler, made Sanskrit a world language? Still in the same introductory chapter, he makes the following observation (p. 21): “In that muscular quote, Firth had emphasised the religious dimension of power, and this is often important: perhaps, indeed, we should be talking not of language prestige but language charisma. Sanskrit, besides being the sacred language of Hinduism, has owed much to disciples of the Buddha [...]” But for more detailed information we must turn to the chapter dedicated to the career of Sanskrit (pp. 174- 226). The spread of Sanskrit across South and Southeast Asia is here presented in the following
words (pp. 176-178): A dialect of Indo-Iranian, [Sanskrit] is first heard of in the North-West Frontier area of Swat and the northern Panjab (now in Pakistan), spoken by peoples who have evidently come from farther north or west [...] Somehow their descendants, and even more their language, spread down over the vast Indo-Gangetic plain, as well as up into the southern reaches of the Himālaya (‘snow-abode’) mountains, so that by the beginning of the fifth century BC the language was spoken in an area extending as far east as Bihar, and as far south, perhaps, as the Narmada. [...] The result was the present-day situation, a northern Indian heartland, stretching from sea to sea, of languages more or less closely related to Sanskrit. [...] It also gained one offshoot in Śrī Lankā to the far south, creating the Siṃhala [...] community there: according to tradition, this group had come from Gujarat, on the north-western coast, in the fifth century BC. The advance of Aryan is continuing to this day in the northern regions of Assam and Nepal, where the official languages (Assamese, and Nepali or Gurkhali) are both Aryan, but have not yet become the vernaculars of large majorities of their populations. Not all the spread of Sanskrit was through full take-up of the language as a vernacular. Even when pre-existing languages, such as Telugu, Kannada and Tamil, held their own, they were usually permeated with terminology from Sanskrit. [...] The process of Sanskritisation did not stop at the boundaries of the subcontinent. Over the course of the first millennium AD, Indian seafaring traders or missionaries made landfall, not only in Śri Lanka, but also in many places along the coasts of SouthEast Asia. Here, the language spread above all as a language of elite civilisation and religion (whether Hindu or Buddhist), but the influence, and evidently the study made of Sanskrit as a vehicle of high culture, was profound. The region is known as Indo-China, quite rightly, for it became a crucible for the competing influences of India and China. This passage can easily give rise to confusion, for it speaks simultaneously of two altogether different phenomena: the spread of Sanskrit and the spread of the (other) Indo-Aryan languages. Sanskrit is an Indo-Aryan language, to be sure. The spread of the Indo-Aryan languages other than Sanskrit, however, is to be distinguished from the spread of Sanskrit. Both spread over large parts of the South Asian subcontinent, but the spread of the Indo-Aryan languages different from Sanskrit was not the result of the spread of Sanskrit, contrary to what the above passage suggests. We will see below that the two phenomena were largely independent of each other, and were of a different nature. Ostler also speaks of a spread of Sanskrit northward, round the Himalayas to Tibet, China, Korea and Japan (p. 178). We will not deal with this spread in this paper, because it is debatable whether it was one at all. We have no reason to think that Sanskrit established itself in any of these countries. There are no Sanskrit inscriptions, nor do we have any reason to believe that any Sanskrit texts were composed in them. We only know that these countries were interested in Buddhism, and to a lesser extent in Indian culture, so that efforts were made to translate texts from Sanskrit into regional languages. As a result there were some scholars in those countries who knew Sanskrit, but this is not to be confused with a supposed spread of Sanskrit, just as little as the Christianization of Europe is an indication of the spread of Hebrew. With regard to the Southeast Asian region, Ostler compares the widespread embrace of Indian culture with the enthusiasm for Americana that captured the whole world in the second half of the twentieth century (p. 179): “In that advance too the primary motives were the growth of profits through trade, and a sense that the globally connected and laissez-faire culture that came with the foreigners was going to raise the standard of life of all who adopted it. As with the ancient advance of Indianisation, there has been little or no use of the military to reinforce the advance of Microsoft, Michael Jackson or Mickey Mouse. There has been little sense that the advance is planned or coordinated by political powers in the centre of innovation, whether in India then, or in the USA today. And the linguistic effects are similar too: English, like Sanskrit, has advanced as a lingua franca for trade, international business and cultural promotion.” Here, then, Ostler gives his opinion about the role of Sanskrit in Southeast Asia, and the reason of its success. It was, he thinks, a lingua franca for trade, international business and cultural promotion. Unfortunately he does not tell us why he thinks so, and we will see that there are good reasons to reject this opinion as not corresponding to historical reality. As a whole, the picture presented by Ostler is seriously misleading. Contrary to what he suggests, the spread of Sanskrit in Northern India did not precede the spread of other IndoAryan languages; the opposite is true in many parts. This is not so because Sanskrit developed out of those other Indo-Aryan languages; it did not. The spread of Sanskrit is rather to be looked upon as an altogether different phenomenon. Indeed, the spread of Sanskrit into the southern regions of the subcontinent and into Southeast Asia was not accompanied by any preceding or subsequent spread of other Indo-Aryan languages. The epigraphical evidence illustrates the above. The earliest inscriptions in Indo-Aryan languages do not use Sanskrit. For three or four centuries, from the time of emperor Aśoka (3rd cent. BCE) onward, they used only Indo-Aryan languages other than Sanskrit. Sanskrit does not make its appearance in inscriptions until the early centuries of the Common Era. Then it gradually takes over and becomes the inscriptional language par excellence in the whole of the South Asian subcontinent and much of Southeast Asia. For almost a thousand years Sanskrit “rules” in this enormous domain. Sheldon Pollock (1996; 2006) speaks for this reason of the “Sanskrit cosmopolis”, which he dates approximately between 300 and 1300 CE. How do we explain the strange vicissitudes of the Sanskrit language? Is Ostler right in thinking that it owes its remarkable spread to being a lingua franca for trade, international business and cultural promotion? Does this make sense at all, once we realize that the spread of Sanskrit is to be distinguished from the spread of vernacular languages? And is the spread of Sanskrit into Southeast Asia to be explained in the same manner as its spread within the Indian subcontinent? Pollock puts the emphasis elsewhere. By introducing the expression “Sanskrit cosmopolis”, he draws attention to the political dimension of the phenomenon. One defining feature of the Sanskrit cosmopolis, he states (1996: 197), “is that Sanskrit became the premiere instrument of political expression in the polities that comprised it, those of most of South and much of Southeast Asia.” He rightly points out that Sanskrit was not a lingua franca of the kind proposed by Ostler:1 “Sanskrit’s spread was effected by traditional intellectuals and religious professionals, often following in the train of scattered groups of traders and adventurers, and carrying with them disparate and decidedly uncanonized texts of a wide variety of competing religious orders, Śaiva, Buddhist, Vai∑ṇava, and others. [...] There is little to suggest [...] that Sanskrit was an everyday medium of communication in South let alone Southeast Asia, or that [it] ever functioned as a language-of-trade, a bridge-, link-, or koiné language or lingua franca (except among those traditional intellectuals) [...]” Pollock continues: “We have little direct evidence that Sanskrit actually functioned as a language of practical imperium — the medium of chancellery communication or revenue accounting, for example — certainly not in Southeast Asia, almost certainly not in peninsular India or the Deccan [...]” The hypothesis he then proposes (pp. 198-99) is “that Sanskrit articulated politics not as material power — the power embodied in languages-of-state for purposes of boundary regulation or taxation, for example, for which so-called vernacular idioms typically remained the vehicle — but politics as aesthetic power. To some degree the Sanskrit ‘cosmopolis’ I [i.e., Pollock, JB] shall describe consists precisely in this common aesthetics of political culture, a kind of poetry of politics.”2 Further explanation follows on p. 199: “Constituted by no imperial power or church but in large part by a communicative system and its political aesthetic, the Sanskrit ecumene is characterized by a transregionally shared set of assumptions about the basics of power, or at least about the ways in which power is reproduced at the level of representation in language, and Sanskrit’s unique suitability for this task.” Having discussed the epigraphical and related evidence from a number of regions, Pollock then depicts the situation around 1000 CE in the following passage (pp. 229-30): A traveler around the year 1000 [...] would have seen, from the plain of Kedu in central Java to the basin of Tonlé Sap in Cambodia, from Gangaikoṇdacoḻapuram in Tamil Nadu to Patan in Gujarat and beyond, imperial formations that had many features in common. The material and social ones I have ignored here: their largely hierarchized societies, administered by a corps of functionaries, scribes, tax collectors, living in grand agrarian cities geometrically planned in orientation to the cardinal points and set within imaginary geographies that with their local mountains, rivers, and springs recapitulated the geography of India, urban structures “freighted with cosmic symbolism, helping one to visualize the order of things” [...] It is their common political-cultural, especially literary-cultural, features I have emphasized: the existence of cultural and political élites assiduously mastering the intricate codes and protocols of Sanskrit poetry, and the publication of their works throughout these cities, in varying degrees of density and grandeur — stately public poems in Sanskrit engraved on the obiquitous copper-plates recording gifts and donations, or on stone pillars looming up from gigantic architectural  wonders. There was thus, I think, a certain concrete reality to the ‘Sanskrit cosmopolis’, one that does not exist only in the retrospective gaze of the historian. For a millennium, and across half the world, élites participated in a peculiar supralocal ecumene. This was a form of shared life very different from that produced by common subjecthood or fealty to a central power, even by shared religious liturgy or credo. It was instead a symbolic network created in the first instance by the presence of a similar kind of discourse in a similar language deploying a similar idiom and style to make similar kinds of claims about the nature and aesthetics of polity — about kingly virtue and learning; the dharma of rule; the universality of dominion. A network, accordingy, wherein the élite shared “a broadly based communallity of outlook”, and could perceive “ubiquitous signs of its beliefs”. Readers may be surprised to see that this passage makes no reference to Brahmins. Isn’t there an old and well-established link between Sanskrit and Brahmins? Can one speak about the spread of Sanskrit without speaking about Brahmins that presumably introduced and cultivated it? Pollock speaks very little of Brahmins in his article.3 Where he does so, his aim appears to be to weaken or even to deny the link between the two. He does so, for example, where he criticizes the notion of ‘legitimation’. He cites (p. 236) in this connection the following passage from an article by Hermann Kulke (1990: 20 ff.): At a certain stage of this development Brahmins ‘came hither’ [to mainland Southeast Asia] in order to legitimize the new status and wealth of these chiefs. Obviously there existed a tremendous need of additional legitimation which obviously no other traditional institution was able to provide fully ... Brahmins appear to have been invited particularly as a sort of ‘extra’ legitimators of a new and more advanced type of authority which was not sanctioned by the traditional societies of South-East Asia ... Obviously in both [South India and Southeast Asia] there had existed the same or at least similar socio-political needs for a new type of legitimation.”4 Pollock is very critical about the notion of ‘legitimation’, and he argues that “there is no reason to accept legitimation theory”.5 However, he seems to think that the rejection of “legitimation theory” also does away with the question of the connection between Brahmins and Sanskrit in South India and Southeast Asia, for he does not return to it. And yet, there is ample evidence to show that there were Brahmins in virtually all the regions that were affected by the spread of Sanskrit. Even if one were to accept that legitimation theory does not explain their presence in all those regions, this hardly justifies leaving this presence out of consideration. Innumerable Sanskrit inscriptions, both in India and in Southeast Asia, testify to the presence of Brahmins. It is a fair question to ask whether the users of Sanskrit in all these regions were not preponderantly Brahmins. Even if one were to admit that ‘legitimation’ was not the reason why these Brahmins were, there is no reason to deny that they were there, and that their presence was intimately connected with the use of Sanskrit in those regions. Elsewhere in his article Pollock suggests that there was no specific link between Sanskrit and Brahmanism during the period he considers. He does so while discussing the first appearances of Sanskrit in inscriptions in South Asia. In short, his argument is that ruling dynasties with a clear penchant for Brahmanical religion did not necessarily use Sanskrit in their inscriptions, and that the first Sanskrit inscriptions we have were commissioned by rulers who had no special links with Brahmanism. The Sātavāhanas — whose rule lasted from the last quarter of the third century BCE to about the middle of the third century CE — constitute the most important example of the former. As Pollock puts it (p. 202): “From the multitude of inscriptional and numismatic evidence available to us now [...], something very striking emerges: Although this was a decidedly vaidika dynasty, as evidenced both by their continual performance of śrauta rites and by explicit self-identification (e.g., ekabamhaṇa [...]), there is no evidence for their use of Sanskrit in any non-liturgical context [...]”.6 The first public political text in Sanskrit of importance is no doubt the celebrated inscription of the K∑atrapa king Rudradāman, which dates from shortly after 150 CE and sets a new trend.7 Pollock comments in the following manner (p. 205-06): “The appropriation of Sanskrit for public political purposes at the end of the first century C.E., is an event symptomatic or causative of a radical transformation of the historical sociology of Sanskrit, comparable, and no doubt related, to the Buddhist appropriation of Sanskrit [...] In this process newly settled immigrants from the northwest seem to participate centrally. [...] What is historically important is not so much that newcomers from Iran and central Asia should begin to participate in the prestige economy of Sanskrit [...] but rather that Śakas, Kuśānas [sic] and the Buddhist poets and intellectuals they patronized begin to turn Sanskrit into an instrument of polity and the mastery of Sanskrit into a source of personal charisma.” This development should not, according to Pollock, be interpreted as essentially linked to traditional Brahmanism (p.207):8 “We may [...] wish to rethink the received account that imagines a ‘resurgence of Brahmanism’ leading to a ‘re-assertion of Sanskrit’ as the language of literature and administration after the Maurya period [...], and consider instead the possibility that a new cultural formation, a Sanskrit cosmopolitan formation, was on the point of being invented.” Indeed, “[t]he radical reinvention of Sanskrit culture seems to have occurred — at least, it is here that we can actually watch it occurring — [...] in a social world where the presuppositions and conventions of vaidika culture were weakest: among newly immigrant peoples from the far northwest of the subcontinent (and ultimately from Iran and Central Asia), most importantly the Śakas (the so-called Indo-Scythians), especially a branch of the Śakas known as the Western K∑atrapas, and the Ku∑āṇas” (2006: 67). By disconnecting Sanskrit from Brahmanism and from Brahmins, Pollock can then formulate questions relating to the spread of Sanskrit in terms of the language itself rather than in terms of its users. This allows him to propose his hypothesis of “politics as aesthetic power”. A consequence of this disconnection is that “we cannot simply read off automatically from the choice to express political will in Sanskrit any particular social consequences (e.g., hierarchization, hegemony; the production of false belief)” (p. 245). No, the qualities of the language itself have to account — if not fully, then at least to a large extent — for its extraordinary expansion: “This had to be a language of transethnic attraction; a language capable of making translocal claims [...]; one powerful not so much because of its numinous qualities [...], but because of its aesthetic qualities, its ability somehow to make reality more real. [...] These aesthetic qualities, moreover, are authenticated by the language’s possessing a tradition of literary texts that embody and realize them.” (p. 239-40). Indeed, “the unique expressive capabilities of Sanskrit poetry allow the poet to make statements about political power that could be made in no other way” (Pollock, 2006: 139).9 All this is interesting and deserves careful consideration. It yet leaves one with the apprehension that the traditional connection between Sanskrit and Brahmins has been too hastily disposed of. Pollock is no doubt right in rejecting “the received account that imagines a ‘resurgence of Brahmanism’ leading to a ‘re-assertion of Sanskrit’ as the language of literature and administration after the Maurya period”. Indeed, one of the main arguments of my book Greater Magadha (2007) is that Brahmanism did not resurge after the Maurya period, but commenced at that time its spread over the subcontinent and beyond for the first time. We are, as a matter of fact, confronted with two remarkable instantiations of spread: the spread of Brahmanism and the spread of Sanskrit. And the question that cannot be avoided is: Were these two really unconnected? Is it not more likely that they had something to do with each other? In order to answer these questions we must be clear what we are talking about. Pollock’s observations about the spread of Sanskrit are enlightening and, by and large, sufficient for our present purpose. But what is meant by “spread of Brahmanism”? The expression Brahmanism can be used to designate the religion and culture of the Veda, but it is only in a very limited sense that these can be said to have spread during the period following the Mauryas. No, the spread of Brahmanism was primarily the spread of Brahmins as Brahmins. That is to say, a region is brahmanized when its population, or its rulers, accept Brahmins as the by right most eminent members of society. This population, or these rulers, are not so much converted to a different religion: no converts are made to Vedic religion, or to any other specific religion promulgated by the Brahmins. No, these populations or rulers are made to accept a different vision of society, in which Brahmins are highest because they have access to the supernatural. An important instrument in the hands of the Brahmins is their knowledge of the Veda, a collection of texts which the vast majority of the population is not even allowed to hear recited, much less study.10 It is their often secret knowledge that gives them the power to work for the good of a kingdom, its ruler and its population. It also allows them to do the contrary, and this is an important reason to humor them. For reasons that are not at all clear at present, Brahmins succeeded in the course of time to convince many rulers that it was a good thing to provide them with what they needed to carry out their rites and do whatever else would benefit the kingdom. The growing presence of Brahmins all over South Asia is well documented,11 but they also came to be present in Southeast Asia, even in countries that became Buddhist: “even in states where Hinayana Buddhism prevailed, Brahmans played an important ceremonial part, especially at Court, and still do so in Burma, Siam and Cambodia, though themselves strikingly different from their counterparts in India.”12 The oldest known inscriptions in Indonesia — we read in The Economic and Administrative History of Early Indonesia (van Naerssen & de Iongh, 1977: 18) — are those of East Borneo. Here there are seven stone sacrificial posts, called yūpas by archeologists, that date from around 400 CE. What is written on them is described in the following terms:13 In clear, well written Sanskrit verses Mūlavarman ‘the lord of kings’, his father —Aśvavarman, ‘the founder of a noble race’ — and his grandfather, ‘the great Kuṇ∂ungga, the lord of men’ — are mentioned on the occasion of a sacrifice. ‘For that sacrifice’, we read on one of the stone poles, ‘this sacrificial post has been prepared by the chief amongst the twice-born [dvija, JB]’. (‘Twice-borns’ is applied to the members of the brahmaṇical or priestly caste.) Apparently these “priests [vipra, JB] who had come hither” (as is written on the second pole) were rewarded by king Mūlavarman for their religious services. Thus the third inscription sounds: “Let the foremost amongst the priests and whatsoever other pious men hear of the meritorious deed of Mūlavarman, the king of illustrious and resplendent fame — (let him hear) of his great gift, his gift of cattle, of a wonder-tree [...], his gift of land. For this multitude of pious deeds this sacrificial post has been set up by the priests.” A Sanskrit rock inscription in West Java dating from about 450 CE deals with an occasion on which the Brahmins were presented with 1000 cows.14 About Cambodia we read the following:15 In Cambodia the Brahmans for many centuries maintained a powerful hierarchy. They were the only one of the four castes that was really organized, this caste having taken form in the fifth century and been constantly augmented by immigrants from India.16 In the days when Yaśovarman was king (acceded A.D. 889), Śaivism was predominant, and we learn from the following inscription that the Brahmans still enjoyed a position similar to that which was theirs in India: “This king, well-versed (in kingly duties), performed the Koṭi-homa and the Yajñas (Vedic sacrifices), for which he gave the priests magnificent presents of jewels, gold, etc.”17 The cult of the Royal God, though founded by Jayavarman II (A.D. 802), did not reach the height of its development until some two centuries afterwards, and was especially associated with Vai∑ṇavism and the temple of A∫kor Vat. This cult led to the Brahmans enjoying an even more exalted position. The Cambodian hierarchy was established by Jayavarman II, and the priesthood became hereditary in the family of Śivakaivalya, who enjoyed immense power; indeed, this sacerdotal dynasty almost threw the royal dynasty into the shade.18 Brahmans were depicted on the reliefs of A∫kor Vat and Coedès has identified Droṇa and Viśvāmitra amongst them.19 In one of the reliefs which illustrates a royal procession, it is interesting to note that the Brahmans are the only onlookers who do not prostrate themselves before the king, as was also the case in India.20 [...] Another point of interest that we learn from the reliefs of A∫kor Vat and A∫kor Thom is that not only the Brahmans, but also the aristocracy wore the chignon, the lower classes having short hair.21 One very remarkable sign of the power of the Brahmans during the A∫kor period is that, contrary to the modern custom, by which princesses of the royal blood rarely marry, formerly alliances were common with the Brahmans;22 and up to the present day there is a tradition amongst the Bakus, who are the descendants of the ancient Brahmans, that in the event of the royal family failing, a successor would be chosen from amongst them.23 As early as the reign of Jayavarman V (A.D. 968) we find evidence of the admixture of Mahāyāna Buddhism with the cult of the Royal God. “The purohita should be versed in Buddhist learning and rites. He should bathe on the days of the festivals the image of the Buddha and should recite Buddhist prayers.”24 And the rites and duties of the purohitas remained a mixture of Hinduism and Mahāyānism until the introduction of Pāli Buddhism in the thirteenth century,25 after which this powerful sacerdotal caste degenerated with their religion to the position occupied by the modern Bakus.26 But the Brahmans of Cambodia perhaps never sank so low as did those of Campā, where “In the Po Nagar Inscription (No. 30) we read that the king’s feet were worshipped, even by Brāhmaṇas and priests”.27 King Yaśovarman of Cambodia created numerous āśramas, among them some that were specifically meant for Vai∑ṇavas, Śaivas and Buddhists. Interestingly, in all three, including the Buddhist āśrama, Brahmins had to be honoured more than anyone else: “In the Saugatāśrama, too, the learned Brāhmaṇa should be honoured a little more than the āchārya versed in Buddhist doctrine [...]”.28 The situation in Thailand was not independent from the one prevailing in Cambodia:29 Though the Thai were Buddhists, their kings surrounded themselves with the appurtenances of Khmer royalty, and recruited their Court Brahmans from Cambodia.For centuries, indeed, Brahmanism enjoyed quite an important position; for although Buddhism was the religion of the people, and was protected by the kings, Hinduism was still considered as essential to the monarchy, and so received a great share of royal favour. The famous inscription (about A.D. 1361) of King Dharmarāja I mentions the king’s knowledge of the Vedas and of astronomy;30 while the inscription on the Śiva statue found at Kāṃbèn Bejra records the desire of King Dharmaśokarāja to exalt both Hinduism and Buddhism. And this is as late as A.D. 1510

It would be a mistake to think of the Brahmins in Southeast Asia as an endogamous group of people, as they were in India. Indeed, G. Coedès (1964: 54) cites a Chinese text from the 5th century which states that “dans le royaume de Touen-siun il y a plus de mille brahmanes de l’Inde. Les gens de Touen-siun pratiquent leur doctrine et leur donnent leurs filles en mariage; aussi beaucoup de ces brahmanes ne s’en vont-ils pas”.32 de Casparis & Mabbett (1992: 287) sum up present knowledge about the role of Brahmins in Southeast Asia: Brahmins had great influence in the Southeast Asian courts in various capacities. As they had access to the sacred texts, the lawbooks and other literature in Sanskrit, they were employed as priests, teachers, ministers and counsellors: the principal advisers of the kings. Government, particularly in early centuries, depended upon such men, who were the chief available sources of literacy and administrative talent and experience. As in the early Indian kingdoms, an important office was that of the purohita, a chief priest with ritual and governmental functions. The epigraphic record of the mainland kingdoms demonstrates the powerful influence of purohitas, notably in Burma and Cambodia, where they often served under several successive rulers and provided continuity to the government in troubled times. In ninth-century Angkor, for example, Indravarman I had the services of Śivasoma, who was a relative of the earlier king Jayavarman II and was said to have studied in India under the celebrated Vedānta teacher Śa∫kara. About the origins of these Brahmins — where they Indians or not? — de Casparis and Mabbett have the following to say:33 If such brahmins were Indians (the Indian brahmins are indeed occasionally mentioned in Southeast Asian inscriptions), one wonders how or why they should have left India. This is the more surprising since Indian lawbooks contain prohibitions for brahmins against overseas travel, which was regarded as ritually polluting. These prohibitions may have had little practical effect, and would not have deterred ambitious men lured by the hope of honour and fortune in a distant land. It has been suggested that some learned brahmins were invited by Southeast Asian rulers at a time when commercial relations between Indian and Southeast Asian ports had spread the fame of such brahmins to the courts. It is indeed likely that this happened sometimes, but probably not on a large scale. It is, for example, striking that the Indian gotra names, never omitted in Indian inscriptions, are not normally mentioned in Southeast Asia. On the other hand, in the few cases where they are mentioned it is likely that they refer to Indian brahmins. It therefore follows that the great majority of Southeast Asian brahmins would have been Southeast Asians, many of whom had acquired their knowledge of the Sanskrit texts and of brahmanic ritual in Indian ashrams. The services of the Southeast Asian Brahmins extended beyond the limits of any single religion:34 Not only in the ‘Hindu’ courts, such as Angkor, but also in the Buddhist courts, such as those of Pagan in Burma and Sukothai in Thailand, the brahmins conducted the great ceremonies, such as the royal consecration, and functioned as ministers and counsellors, but had to share their influence with that of the Buddhist monks. By its very nature Buddhism was concerned with the acquisition of spiritual merit and moral perfection rather than with the rites and ceremonies of a royal court, which were left to the brahmins.35 The grand ceremonies in Pagan [...] required the services of numerous brahmins, although Theravāda was then well established. In Cambodia, as late as the thirteenth century [...], Jayavarman VIII built a temple for the scholar-priest Jayama∫galārtha, and likewise for the brahmin Vidyeśavid, who became court sacrificial priest. The Chinese visitor Chou Ta-kuan refers to the presence of brahmins wearing the traditional sacred thread. de Casparis and Mabbett (1992: 288) draw the following conclusion: What is shown by the role of such brahmins is that it is appropriate to speak of Brahmanism as distinct from the specific cults of Śiva or Visṇu, or any of their innumerable kin: the priests stood for a social order and for the rituals that gave to the political or local community a sense of its unity and its place in the world. The part of this conclusion which must be emphasized is that Brahmanism is distinct from the specific cults of Śiva or Visṇu, or any of their innumerable kin, and that the Brahmins stood for a social order.36 This seems obvious and undeniable, and yet it is often overlooked by scholars who wish to assign Brahmanism to the category ‘religion’. In reality, Brahmanism is primarily a  social order. Only this way can we make sense of the evidence from Southeast Asia,37 as well as of the evidence from South Asia.38 It appears, then, that some of the proposals made already in 1934 (in Dutch) by J. C. van Leur still hold good.39 About South Asia he said (van Leur, 1955: 97):40 The chief disseminator of the process of ‘Indianization’ was the Brahman priesthood; the aim of the ‘Brahman mission’ was not the preaching of any revealed doctrine of salvation, but the ritualistic and bureaucratic subjugation and organization of the newly entered regions. Wherever the process of ‘Indianization’ took place, ‘religious’ organization was accompanied by social organization — division in castes, legitimation of the ruling groups, assurance of the supremacy of the Brahmins. The colossal magical, ritualistic power of the Brahman priesthood was the most characteristic feature of early Indian history. The rationalistic, bureaucratic schooling of the priesthood as the intellectual group, which went to make up its great worth, its indispensability even, for any comprehensive governmental organization, was [...] interwoven with the sacerdotal function. The Brahman priesthood developed high qualities in that field as well, but its decisive influence came from the magical, ritualistic power of domestication it in the absoluteness of its power was able to develop. The spread of Brahmanical institutions to Southeast Asia was hardly more than a continuation of this process (pp. 103-04): The Indian priesthood was called eastward — certainly because of its wide renown — for the magical, sacral legitimation of dynastic interests and the domestication of subjects, and probably for the organization of the ruler’s territory into a state. Pollock may object to the word legitimation in these two passages. Nothing much is lost by removing it.41 The factual situation remains the same. Brahmins were called to Southeast Asia (or were found in Southeast Asia; there is no reason to insist on the Indian origin of all of them), and these Brahmins brought with them their sacred language, Sanskrit.42 It will be interesting to draw also Sri Lanka into the discussion. This country was Buddhist for most of its history, and it had to be governed and needed organization. Where did the Singhalese rulers find information about these matters? Lingat (1989: 152) has the following to say about this: Où les rois de Ceylan vont-ils chercher des conseils sur l’organisation administrative du royaume, sur le choix et les attributions des ministres, sur la police des villes et des villages, sur les impôts, leur taux, leur perception, sur l’organisation de la justice et comment elle doit être rendue, et sur les questions multiples que pose sur le plan interne le gouvernement d’un Etat? Ce n’est certes pas dans les Ecritures, que ces problèmes ne concernent pas, mais dans une littérature indépendante. [...] il semble qu’il n’y ait eu à Ceylan aucun ouvrage consacré à la politique; du moins aucun traité de littérature pâlie n’en mentionne. [...] Il est dès lors naturel que les rois de Ceylan se soient tournés vers la littérature sanskrite qui est fort riche en ouvrages de ce genre. Outre l’Arthaśāstra de Kauṭilya qui paraît bien être mentionné dans les chroniques sous le nom de Koṭalla (LXIV, 3; LXX, 56) mais surtout comme traité militaire, une référence semble nettement être faite au Code de Manu sous le nom de Manunītivisārada (LXXXIV, 1). Mais, le plus souvent, il est seulement fait allusion à l’art de la politique, nīti, ou rājanīti, l’art de la politique à l’usage des rois. On trouve aussi Manunīti (LXXX, 9: Vijayabāhu II), l’art de la politique selon Manu, expression dans laquelle le mot Manu, croyons-nous, ne vise pas nécessairement l’auteur mythique du Mānavadharmaśāstra mais plutôt le prototype du législateur humain. Manunīti, comme rājanīti, désigne l’ensemble des règles et des principes suivant lesquels la société humaine, entendez la société laïque, doit être organisée, administrée, pour fonctionner convenablement. Le Bouddha n’a touché à ce sujet qu’incidemment [...] Mais ce sont les ouvrages brahmaniques qui contiennent l’exposé le plus détaillé des institutions nécessaires à une bonne organisation de la société. [...] pour les rois de Ceylan, la société indienne, telle que la décrivent les dharmaśāstra et les ouvrages d’arthaśāstra, reste le modèle même de la société; ils ne peuvent la concevoir autrement. D’ailleurs, ils sont élevés dans une ambiance purement indienne. Les rites brahmaniques sont pratiqués à la cour. Le couronnement est une institution brahmanique. Pour le futur Duṭṭhagāmaṇī, on célèbre la cérémonie de la dation du nom et celle de la première bouchée de riz (nāmadheya [...]; annaprāśana: XXII, 65 et 74), qui sont des saṃskāra. A la nouvelle de la naissance prochaine de Parakkamabāhu, Mānābharaṇa fit non seulement réciter sans cesse le paritta par la communauté des bhikkhu, mais, dit la chronique (LXII, 33), à la naissance furent célébrés les divers rites prescrits par le Veda (LXII, 45) et, durant son enfance, tous les autres rites (LXII, 53), y compris le cū∂ākaraṇa correspondant au sikhāmaha (LXIII, 5) et l’upanayana (LXIV, 13) qui est célébré avec une grande solennité. Il fit également accomplir par des purohita et des brahmanes versés dans le Veda et le Vedanta des sacrifices tels que le homa et d’autres rites tenus pour salutaires. Il est donc naturel que les rois de Ceylan se soient tournés vers les ouvrages brahmaniques pour y puiser des conseils sur l’art de gouverner leur royaume. Les réformes accomplies par Parakkamabāhu dans le Dakkhiṇa-dēsa quand il fut devenu roi de cette partie de l’île — réformes qui sont décrites dans le Cūlavaṃsa (LXIX) … et qui  donnent un rare aperçu de l’administration des provinces — sont évidemment inspirées des ouvrages de nīti, l’Arthaśāstra de Kauṭilya et le Code de Manu … . The comparison with Sri Lanka is interesting in that Sanskrit never predominated there.43 Something like the Indian caste system44 survives until today but, as Ryan (1953: 8) points out, “the most significant factor for an understanding of Sinhalese caste structure is not, as is commonly supposed, that the Sinhalese preserved Buddhism, but that the Sinhalese did not preserve the Brahmin”.45 The subsidiary role of Sanskrit, which is no doubt connected with a feeble presence of Brahmins in historical times,46 did not prevent the Brahmanical vision of how to run a state from exerting a strong influence on the island.47 If Lingat is right, the reason in this case is not the need for legitimation but the absence of a workable alternative.48 Perhaps this applied to the countries of Southeast Asia as well, thus contributing to the explanation of the presence of Sanskrit and of Brahmins there. We see that it will be hard to separate Sanskrit from Brahmins, both in South and Southeast Asia. The one complicating factor is Buddhism. During most of the period of the Sanskrit cosmopolis, Buddhists appear to have coexisted successfully with Brahmins at the royal courts of Southeast Asia. The question why Buddhists in South Asia adopted Sanskrit for their texts is essentially different, and will be discussed below. Pollock does not deny the presence of Brahmins in the different regions of Southeast Asia. The growth of a class of Khmer Brahmins, he states on p. 222 of his article (1996), is perfectly reasonable, for precisely such a development occurred in Java and Bali. On the same page he notices that Indian Brahmins were occasionally imported, “as for example for the lustration of the Khmer domain in the ninth century”. Is it, in view of all this, correct to disconnect the spread of Sanskrit from the spread of Brahmins? At first sight one would think not, but there are some issues that need to be dealt with before a decision can be taken. It cannot be denied that the first political use of Sanskrit did not take place under the Sātavāhanas, who had a strong (though not exclusive) connection with Brahmanism, but under the western immigrant kings known as Ksatrapas, whose Brahmanical connection was less strong.49 What conclusion can be drawn from this? A look at the inscription of the Ksatrapa Rudradāman from the second century CE, already referred to above, may be useful. This inscription records the restoration of a lake, called Sudarśana, which had been constructed during the reign of Candragupta Maurya.50 Rudradāman’s inscription refers back to these earlier events, recalling that the lake had been dug by the governor of Candragupta Maurya and embellished for Aśoka Maurya by the Yavana king Tusāspha. What strikes us most in the context of our investigaton is that the governor of Candragupta is referred to as “the Vaiśya Pusyagupta” (vaiśyena pusyaguptena).51 Here, then, there is an explicit reference to a Vaiśya. There is no need to recall that Vaiśyas constitute the third of the four Brahmanical varṇas: Brahmin, Ksatriya, Vaiśya, Śūdra. They have their place in society as conceived of in Brahmanism. This Brahmanical conception of society became very popular in India, but owed this popularity to Brahmins. The vision of society as being thus hierarchically layered spread with Brahmanism and was an integral part of it. This raises a puzzling question: Rudradāman’s inscription claims that the governor of Candragupta, who ruled more than four centuries before him, had been a Vaiśya. Brahmanism and the Brahmanical vision of society had little or no influence in the realm of Aśoka, even less in that of his grandfather Candragupta. How can we believe that the usual Brahmanical division of society played any role at the court of these rulers? Kielhorn, the editor of the Junāga∂h inscription, was obviously aware of the problem. He made a feeble attempt to solve it in a footnote (1906: 41 n. 8), stating: “The Vaiśyas according to Varāhamihira are a people of the western division”. But a much simpler solution would be to assume that Rudradāman, though not a “Brahmanical” ruler in any strict sense, had adopted the Brahmanical vision of society. He may not have been the first to do so; perhaps he simply inherited it from his father and grandfather, both mentioned in the original inscription. Having adopted this view, he retroactively assigned previous rulers and their collaborators a place in the Brahmanical order of varṇas.52 Indeed, he refers to “all the varṇas” (p. 43 l. 9), and takes care to specify that King Tusāspha, who does not fit well into the system, is a Yavana, i.e., presumably a Greek.53 It follows that the contrast between Rudradāman and the Sātavāhanas with regard to their Brahmanical connections should not be exaggerated. We may have no records of Vedic sacrifices being carried out by Rudradāman,54 but his inscription does reveal a certain partiality toward Brahmins where it says:55 “he, the Mahāk∑atrapa Rudradāman, in order to [benefit]56 cows and Brahmins for a thousand of years, and to increase his religious merit and fame, — without oppressing the inhabitants of the towns and country by taxes, forced labour and acts of affection — by [the expenditure of] a vast amount of money from his own treasury and in not too long a time made the dam three times as strong in breadth and length ...” It is clear, then, that Rudradāman knew and honoured Brahmins. He also knew and respected their vision of society as consisting of a number of varṇas, one of them being that of the Vaiśyas.57 There is therefore no reason to disagree with the following general appreciation:58 “It appears that the use of Sanskrit for inscriptions was promoted, though not originated, by the Scythian rulers of northern and western India in the first two centuries of the Christian era. Their motivation in promoting Sanskrit was presumably a desire to establish themselves as legitimate Indian or at least Indianized rulers, and to curry the favor of the educated Brahmanical elite.”59 Indeed, “the shift to using Sanskrit, the Brahmins’ liturgical language, for the business of state was primarily the initiative of foreign rulers — Scythians and Kusāṇas — anxious to align themselves with a priestly class firmly rooted in Āryāvarta, the ‘Land of the Āryas’ [...] Once introduced by arrivistes, this policy was fully established as the royal standard by the imperial Guptas.”60 Respect for Brahmins in South and Southeast Asia should not be confused with “conversion” to Brahmanism. This is still true at the time of the Pāla rulers of north-east India. We have, for example, a copper-plate grant of the end of the ninth century, in which King Devapāladeva appears as a devout worshipper of the Buddha. In spite of this, he gives a village to a Brahmin of the Aupamanyava gotra and Āśvalāyana śākhā. 61 The Brahmanical vision of society is absent in South Asian inscriptions that are not in Sanskrit and whose makers or instigators have no association with Brahmanism. This is true, to be sure, of the inscriptions of Aśoka.62 They refer to none of the four varṇas except theBrahmins, nor to the system as a whole.63 It is also true of the early Tamil inscriptions, edited and studied by Iravatham Mahadevan (2003), which concern Jainas but not Brahmins, and depict a society with an “absence of a priestly hierarchy” (p. 162). It is equally true of other inscriptions in Kharosṭhī, judging by Konow’s index,64 and of all non-Sanskrit inscriptions in Brāhmī that precede 150 CE.65 An exception has to be made, not surprisingly, for Nāsik Cave Inscription no. 2 of the Sātavāhanas, the one which also contains the expression ekabamhaṇa: it has the term khatiya, refers to the four varṇas (cātuvaṇa), to the twice-born (dija), and even to the (Brahmanical) three objects of human activity (tivaga).66 Rudradāman, one of the first to refer to the Brahmanical (di-)vision of society, is also one of the first to use Sanskrit. Is this coincidence? The obvious answer to this question must be: no. Numerous questions surround the Buddhist appropriation of Sanskrit in South Asia (after half a millennium of rejection), which cannot all be dealt with in this paper.67 Pollock (2006: 56-57) sums up the present situation in the following words: “What exactly prompted the Buddhists to abandon their hostility to the language after half a millennium [...] and finally adopt it for scripture, philosophy, and a wide range of other textual forms, some of which they would help to invent, is a question for which no convincing arguments have yet been offered.” He may be right in stating that in this process, too, “newly settled immigrants from the northwest seem to participate centrally” (1996: 205-06).68 Did these Buddhists simply follow a trend at and around the royal court, adjusting themselves to a fait accompli?69 “The adoption of Sanskrit by Buddhists after centuries of resistance is often explained by its being ‘the language of learning’ or possessing ‘technical precision’. We are never told why, after five centuries, it suddenly became necessary or desirable for Buddhists to begin to participate in such learning, or indeed why the precision of the local languages of Buddhism (Gandhari, Tocharian, and so on), which had often been vehicles for liturgy, metaphysical doctrine, and moral discourse, had suddenly failed. The fact that the cosmopolitan transformation of Sanskrit occurred concomitantly with the consolidation of Ku∑āṇa power, however, suggests, if not a clear cause, at least a clear context for the momentous change, and one fully in keeping with the abundant evidence over a very long term of a specific kind of politicization of literary culture in premodern South Asia.” (Pollock, 2006: 513) In spite of the lack of clarity, I would suggest that here, too, we should not separate the Sanskrit language from its real “owners”, the Brahmins. We know that the Buddhists left a number of intellectual and practical domains to the Brahmins, from an early date onward. Asita, the seer who predicted that the just born Bodhisattva was going to become a Buddha, was a Brahmin, and this was no coincidence. The activity of predicting the future, whether on the basis of bodily signs or by any other means, remained a prerogative of the Brahmins which the Buddhists never contested. Indeed, Buddhists in India excluded themselves from the field of astrology-astronomy-mathematics, to which they have not contributed a single known author. It appears that the Buddhists left other areas, too, to the Brahmins, including sophisticated political advice,70 and the language in which courtly matters had to be dealt with, viz., Sanskrit. Since the Buddhists yet needed support from the royal courts, the adoption of Sanskrit became a matter of adjusting to the exigencies of the day. To what extent did Buddhists contribute to the consequent spread of Sanskrit? A widely discussed Vinaya passage creates the impression that the temptation to recite the canonical texts in Sanskrit rather than in various regional languages had been resisted before the Sarvāstivādins, presumably, changed their minds.71 It is difficult to put exact dates to the resisted temptation and to the final surrender.72 The questions asked above remain for the time being difficult to answer. The situation in Southeast Asia may be even more complex. Some scholars are of the opinion that Buddhism reached Southeast Asia before Brahmanism.73 But if so, did it contribute to the spread of Sanskrit in those regions? The inscriptional evidence suggests otherwise. While postponing a detailed discussion of the question of the Buddhist adoption of Sanskrit to a future occasion, a few words must be dedicated to another issue. How is it that the old Buddhist canon is familiar with the Brahmanical vision of society? If inscriptions remain largely silent about it until the first centuries of the Common Era, is it not remarkable that the Buddha, who may have died around the year 400 BCE, was acquainted with it? Formulated in this manner, these questions are misleading. It is true that certain Sūtras of the ancient canon are familiar with the Brahmanical division of society, but this does not allow us to conclude that the historical Buddha lived in a society where this division held sway. Let us, by way of conclusion, quickly recall the evidence. Wagle, in his analysis of the Buddhist Sūtras with an eye on extracting from them information about social relationships (1966), recapitulates on p. 74 that “there are three religious orders, each fighting against the other for superiority. To the brāhmaṇa order belong the brāhmaṇa, the paribbājaka, some gahapatis and others; to the Buddhist order belong the Buddha, the monks (brāhmaṇa and others) and the upāsakas. The [...] third is the Jain order, comparatively less significant but no less hostile to the others. Their following consists of Nigaṇṭhas such as Dīgha Tappasi, Saccaka and so on, and lay disciples. All the three groups contend for superiority in the eyes of the gahapati who represents the bulk of society.” The bulk of society, then, consists of “house-holders” (Pāli gahapati), without internal distinctions.74 This category is not even distinguished in principle from that of the Brahmins, because Brahmins, too, can be house-holders; they are then referred to as brāhmaṇa-gahapati. 75 Society is therefore not divided into Ksatriyas, Vaiśyas and Śūdras in most of the Sūtras. That is to say, the Sūtras depict a society in which there were Brahmins, to be sure, but not one in which the Brahmanical model of society had been accepted. To my knowledge, the Buddha is never depicted as having a discussion with this or that Ksatriya, Vaiśya or Śūdra. Some Sūtras are acquainted with the Brahmanical vision of society. They know the system of varṇas, and point out that this system is not in use among the Greeks and the Kambojas. They are also concerned with the superior status of Brahmins, a claim not accepted by the Buddha. These Sūtras contain indications that belong to a period well after the Buddha. They suggest in this manner that the Brahmanical vision of society made its entry in these texts a considerable time after the Buddha.