Religion in Human Evolution: From the Paleolithic to the Axial Age

by Robert N. Bellah

  The Middle Vedic Transformation

  Michael Witzel has argued that there is a gap in time between the society we have tried to reconstruct from the older parts of the RV and the quite different society that emerges from the Brahmanas. Exact dates are impossible to assign, but if we view the older parts of the RV as dating from the late centuries of the second millennium BCE, then perhaps the earliest Brahmanas might date from a century or two into the first millennium BCE. Witzel finds some texts that represent a stage in the development of early Sanskrit that makes it likely that they come from this gap period between the two major text collections. These include mantra texts and other fragments from later collections, as well as late parts of the RV itself, book 10 in particular. These texts give us clues to what seem to have been major political changes that would lead to the remarkably different society and culture of the Middle Vedic period, in which the Brahmanas became central.

  Geographically there was a shift from the Panjab to a region further to the east, on and just beyond the divide between the Panjab and the upper Gangetic plain, a region known as Kuruksetra. Whether this indicates a movement of population, or, just as likely, a shift in the area of cultural focus, we cannot know, but political changes were at the heart of this shift, even if we can discern them only vaguely. We have noted above that in the late RV first the Purus and then the Bharatas came to prominence among the thirty or more “tribes” (on the difficulties involved in the use of this term, see Chapter 3) of the Aryans, but were unable to establish any lasting rule in a constantly unstable situation. In Witzel’s gap period, however, the Kurus, holding themselves to be the legitimate descendants of the Bharatas, established a stable regime that would set the pattern for all subsequent Indic political history.37

  The Kuru leaders continued to call themselves rajan, which, as we have seen, should in the Rgvedic period be translated as “chief” rather than “king.” Witzel uses both “chief” and “king” for the Kuru leaders, but at one point, in trying to pin down the Kuru polity, he refers to the leader as “great chief,” and points to a paramount chiefdom, that is, a regime in which a great chief has many lesser chiefs owing allegiance to him. Witzel notes as an indicator of his new strength the capacity of the Kuru great chief not only to recycle the booty of military raids but to exact tribute (bali) from his subordinates, something that lesser chiefs were never able to do.38 All of this is strongly reminiscent of the paramount chiefdoms in Hawaii just at the time of Western contact when one of them seemed on the verge of creating an early state. My suspicion was confirmed when I found Witzel himself contrasting the later true Indic kingdoms with the Kuru regime: “Absolute power was realized only in the first great states with aspirations of empire, such as Magadha about 500 B.C. The Vedic Kuru realm still resembles that of a large Polynesian chieftainship such as that of Hawaii-and with a similar ideology.“39 But, as with Hawaii, the Kuru realm was probably in transition to an early state, so ambiguity in terminology may reflect the social ambiguities of the

  There is both archaeological and literary evidence that Kuruksetra was at some point in the first half of the first millennium BCE attaining a status that no Aryan society had achieved before. Erdosy, in his survey of settlements in the Indo-Gangetic divide and the Ganga valley through most of that period, found that in many places there were only two levels of settlement in terms of size, and large areas where there was only one (small) settlement level, implying that chiefdoms were the main social structure (chiefdoms would normally involve two settlement levels) beyond the still common tribal level (one settlement level would imply a tribal society). However, there was one area with a three-tiered settlement level, namely Kuruksetra, and three levels implies paramount chieftainship at On the basis of literary evidence Erdosy believes that perhaps by the sixth century, the term janapada “acquires its classical meaning of `realm,”’ and that “Kuruksetra, home of the most famous of all Late Vedic tribes, may have been the first region to be clearly delineated [as a janapada].“42

  Witzel notes that “an important, if not the chief one among the religious developments is that the new royal center in Kuruksetra (‘the land of the Kurus’) gave rise to a new mythology of the region.“43 He describes the mythology that gave religious expression to this new level of political development as follows:

  Now we are able to understand the importance of Kuruksetra. It was deliberately turned into the land of the gods, their devayajana [the place where the gods sacrifice], where even the heavenly river, the Milky Way, touched down on earth and continued to flow through Kuruksetra as the Sarasvati and Drsadvati rivers and where stood the world tree Plaksa Prasravana at the centre of the world and of heaven. While quickly becoming a peripheral area of India, this land of Kuruksetra remained its holy land to this very day. Here the Mahabharata war was fought, here one made pilgrimages, in fact to this very day, along the banks of the sacred rivers, here one could gain immediate access to heaven, here the purest Vedic language was spoken, and from here even the medieval kings of Eastern India brought their Brahmins, the famous Saravatas.44

  Kuruksetra was, of course, a region, not a city: there was no capital city, indeed no cities at all. Witzel says, “Note that the kings roam about in their territory because of their comparative lack of centralized power, in order to control the various parts of their realm.“45 The development of a unique mythology of place regarding Kuruksetra is only one aspect of a major reorganization of religious practice that will be described below, a change intimately related to changes in social structure, as we will see.

  Given that early states have emphasized hierarchy perhaps more strongly than any societies before or after, we may note that it is in the Middle Vedic period that the varna system, which divided the society into four orders, comes into view in its mature first full description of it is to be found in one of the latest of the Rgvedic hymns, 10.90. Here we find that the human hierarchy is embedded in a cosmological hierarchy, so characteristic of archaic societies, as we saw in Chapter 5. The hymn is referred to as Purusasukta, “the Hymn of Man,” and here purusa is the androgynous primordial man or world giant from whom the universe, the gods, and humans come. Here are some selected verses:

  1. Thousand headed is Purusa, thousand eyed, thousand footed. He covered the earth on all sides and stood above it the space of ten fingers.

  2. Purusa alone is all this, what has been and what is to be, and he is the lord of the immortals, who grow further by means of

  6. When with Purusa as oblation the gods offered sacrifice, the spring was its clarified butter, the summer the fuel, the autumn the

  11. When they portioned out Purusa, in how many ways did they distribute him? What is his mouth called, what his arms, what his thighs, what are his feet called?

  12. His mouth was the Brahmana, his arms were made the Rajanya, what was his thighs was made the Vaisya, from his feet the Sudra was born.

  13. The moon from his mind was born; from his eye the sun was born; from his mouth both Indra and Agni; from his breath the wind was born.

  16. The gods sacrificed with the sacrifice to the sacrifice. These were the first

  To give a full explication of even the verses quoted above would take the rest of this chapter, but certain things can be noted. This famous hymn has clearly moved beyond myth to mythospeculation.50 Purusa, the ordinary word for (usually male) “human being,” is here in transfigured form elevated above the usual gods of the Rgveda and seen as their creator, or their source, as with Indra and Agni in verse 13. Speculation has raised the question of a higher order of ultimate reality than the gods. Further, the final verse, 16, offers a new speculative idea of sacrifice. Wendy Doniger O’Flaherty explains: “The meaning is that Purusa was both the victim that the gods sacrificed and the divinity to whom the sacrifice was dedicated; that is, he was both the subject and the object of the sacrifice. Through a typical Vedic paradox, the sacrifice itself creates the sacrifice.“5’ Typical, however, of Middle Vedic specu
lation, not of the older RV thinking as evidenced in the hymn to Indra, 3.45, quoted above. O’Flaherty also comments on “first rites” in verse 16: The word Maurer here translates as “rites,” she tells us, is dharmas, which she translates as “ritual laws.” O’Flaherty recognizes that dharma is a “protean word,” but here designates “the archetypal patterns of behavior established during this first sacrifice to serve as models for all future sacrifices.“52

  The whole hymn is archetypal in O’Flaherty’s sense, and most particularly in verses 11 and 12, which Paul Mus called “the first constitution of India,” because for the first time it described the varna system, the basic structure of Indic society up to recent times. In verse 12, O’Flaherty translates Brahmana as “the Brahmin,” Rajanya as “the Warrior” (also called Ksatriya), Vaisya as “the People” (which earlier translated vise), and Sudra as “the Servants.“53 Although this is the first time a system of four orders is ranked in a cosmological context, we still need to ask if this is only the systematization of a long-standing practice and whether what we know later as the four varnas is really what is being described here. In other words, what is described in RV 10.90 is only one moment in an evolving social system, important though that moment may be, and we must try to understand it as such.

  Gregory Nagy, in an interesting effort to get at the Indo-European background of both Greek and Indic societies, has turned to the work of Georges Dumezil, particularly in connection with Dumezil’s theory of the three functions that are supposed to characterize all Indo-European societies: the first function is sovereignty/priesthood, the second is the warrior class, and the third is agriculture/herding.54 Emile Benveniste, Dumezil’s follower, according to Nagy, “shows clearly that the basis of Indo-European social organization was the tribe.“55 However, most of our evidence for Indo-European societies comes from early states, where what were originally “functions” could have become more firmly differentiated and institutionalized, with the Indic varna system being a case in point (and with the addition of the Sudra as the fourth varna, “servants,” who are included but excluded in that they could not fully participate in the sacrifices and festivals that defined Aryan culture, thus indicating that they were not part of the original tripartite inheritance).

  Nagy is aware of the difficulty involved in the use of the term “tribe” but he provisionally defines it, borrowing from Montgomery Watt, as “a body of people linked together by kinship, whether in the male or in the female We have seen the term vis translated as “people” in contrast to “ruler” and “nobles” in early Vedic society. Nagy, drawing on Benveniste, translates vis as “tribe, people,” and speaks of it as referring to a “social whole,” again drawing on Benveniste in relating vis, “tribe,” to visva, “all.“57 He finds interesting parallels between Greek phule, “tribe,” and Indic vis, in that both terms relate both to the social “all” and to a division within the all, indeed the lowest of the three Indo-European divisions:

  The semantic relationship between the name of the lowest in the order of three phulai [plural of phule], the Pamphuloi, and the word phule itself, corresponds to the semantic relationship between the name of the lowest in the order of the three leading social classes or varna-s in Indic traditions, the vaisya, and the word from which it is derived, vis “tribe”: just as the word Pamphuloi implies the whole community while designating the lowest of three parts, so also the word vaisya, by virtue of its derivation, implies the whole community, the vis, while specifically designating again the lowest of three

  What I make of all this is that there was a degree of tribal egalitarianism underlying the differentiations that were developing with the gradual emergence of an early state in Greece and RV 10.124.8 says of the gods in relation to Indra: “Choosing him as all the people choose a O’Flaherty here translates viiah (plural of vise, and thus, literally “peoples” or “tribes”) as “all the people” in this passage, but Nagy prefers to translate this late RV passage, in effect, as “Choosing him as the tribes choose an overking [paramount chief?].“61 Even if such a choice were largely symbolic, there was, if we can take the verse as reflecting reality, still an expression of popular consent to rulership.62

  What we can perhaps see in RV 10.90.11-12, the earliest clear formulation of the hierarchy of the four varnas, is a movement away from a loose conception of a people with chiefs and priests above and followers below, and all linked by kinship, to a society of orders, differentiated roles, that, though often inherited, in principle transcend kinship and cross tribal boundaries. Even in the Middle Vedic period, however, there was apparently more fluidity and movement between varnas than there would be later. Erdosy notes a degree of mobility in that intermarriage between varnas was possible and status could still be earned rather than inherited: that is, stories of youths of uncertain birth, who, through intensive study became recognized as Brahmins, and, on the other hand, the idea that one of Brahmin birth who didn’t know the rituals wasn’t really a Brahmin.63 Patrick Olivelle finds in texts older than the Upanisads the question, “Why do you enquire about the father or the mother of a Brahmin? When you find learning in someone, that is his father, that is his grandfather.“64 And of course, all through history conquerors of whatever background could claim Ksatriya status. As in Hawaii, there were always those who could come up with convenient genealogies. Lineages and families remained important, as they would throughout Indic history, but the varna system brought solidarities-and antagonismsthat transcended the primary focus on kinship.

  The difficulties of consolidating what had been numerous lineages of poets/priests and chiefs/subchiefs in many small chiefdoms into a relatively large paramount chiefdom were major. Even the establishment of stable chiefdoms had not been possible without conflict. Hartmut Scharfe points out that rajan was probably at first a temporary term, meaning a “war chief” who would function only during a campaign. In the RV Indra is frequently called rajan when he intervenes to fight a particular enemy, but then withdraws. “If rajan did not originally denote a position held in permanency, ddmpati `master [father?] of the house/family’ and viipdti `master of the clan/ settlement’ surely did.“65 Hermann Kulke refers to another important term, grama, which originally meant a “trek” of migrating Aryans, which later came to be used for villages, “settled treks.” And whereas “the early gramas were led by the gramanis the `trek leaders’ who always belonged to the vii population of the grama,” the settled village “witnessed the emergence of the gramin, the `village owner,’ who appears to have always come from the rajanyas or ksatriyas.” 66 Kulke suggests that the viipopulation of the grama did not always take kindly to their would-be overlords, who could be thrown out, or, if worst came to worst, abandoned as the village population simply moved away. In other words, settlement and hierarchy leading to ever more centralization did not necessarily come smoothly. Strongly organized subordinate groups could resist those who sought to dominate them. Chiefdoms and paramount chiefdoms everywhere are notoriously fragile: as chiefs attempt to dominate villages, and paramount chiefs to dominate subchiefs, there is always the possibility that subordinate groups will break away. An early state develops structures and practices that make this increasingly difficult, but the state in India never quite transcended the fragility of the paramount chiefdom.

  The varna system was only one effort to create larger solidarities in a society still divided by many subgroup loyalties. Lacking a powerful administrative apparatus, the Kuru rulers, with the help of a much more clearly defined Brahmin class, developed a ritual system far more complex than what had preceded it and closely related to the developing varna system. Under the direction of the Kuru king the hymns that had been created continuously for generations by many lineages in many tribes/chiefdoms were now gathered into one collection, what we know as the Rgveda, to be shared by all of the “newly formed Brahmana class,” even though each hymn was still marked by the name and lineage of the original poet. Under Kuru pressure the “copyright,” jealou
sly guarded by earlier lineages, was now no longer effective as the hymns became the common resource of the newly established Brahmin priests.67

  Most remarkably, the canon of the Rgveda was now closed. While cherishing and imitating the archaic features of the inherited material, the new priestly class was devoted to the development of a new and much more complex ritual system, one that focused on the Kuru king and his court but had other significant functions as well. In connection with this new ritual system, additional collections of texts were developed over time: the Samaveda and the Yajurveda, providing the ritual chants and the mantras, respectively, mostly drawn from the Rgveda, and the Atharvaveda, which does not supply material for the new solemn (irauta) rituals but for smaller and more private rituals. The srauta rituals required priests specializing in each of these four Vedas and the complex commentarial literature that developed around them.

  One of the earliest of these new srauta rituals has been, on linguistic grounds, dated from soon after the closing of the RV and is associated with the Kuru court, possibly with the reign of the great Kuru king, Pariksit, the Agnicayana or fire ritual.68 Given the number of officiants required and the fact that it took nearly a year to complete the ritual, it must have been enormously expensive, such that only someone of very high status could have had it performed. Theodore Proferes writes, “The unction ceremony for the sacrificer (abhiseka) that is performed in the Agnicayana connects this rite, too, with the most powerful of leaders.“69 Proferes notes on linguistic grounds that the Asvamedha rite, so closely linked as we will see to the institution of kingship, and the Sautramani rite, “which, judging by its focus on the figure of Indra, may well also have been originally a royal rite,” also date from this gap period and are the “Ur-liturgies” coming between the Rgveda and the earliest of the other srauta texts.70 Proferes sums up what he thinks was happening at that critical moment in Indic history: “As part of their programme to consolidate power, the Kuru kings sought to overcome the divisive tendencies inherent in the clan-based organization of their priestly elite by encouraging the development of what we might call an `ecumenical’ ritual system, one which did not rely upon or perpetuate the clan divisions characteristic of the