Skip to main content



HINDUISM . Hinduism is the religion followed by about 70 percent of the roughly seven hundred million people of India. Elsewhere, with the exception of the Indonesian island of Bali, Hindus represent only minority populations. The geographical boundaries of today's India are not, however, adequate to contour a full account of this religion. Over different periods in the last four or five millennia, Hinduism and its antecedents have predominated in the adjacent areas of Pakistan and Bangladesh and have been influential in such other regions as Afghanistan, Sri Lanka, Southeast Asia, and Indonesia. But in these areas Hindu influences have been superseded or overshadowed by the influences of other religions, principally Buddhism and Islam. This account will treat only of Hinduism as it has taken shape historically in the "greater India" of the Indian subcontinent.

Indus Valley Religion

There are good reasons to suspect that a largely unknown quantity, the religion of the peoples of the Indus Valley, is an important source for determining the roots of Hinduism.

The Indus Valley civilization arose from Neolithic and Chalcolithic village foundations at about the middle of the third millennium bce as a late contemporary of Egyptian and Mesopotamian riverine civilizations. It engaged in trade with both, though mostly with Mesopotamia. Reaching its apogee around 2000 bce, it then suffered a long period of intermittent and multifactored decline culminating in its eclipse around 1600 bce, apparently before the coming of the Aryan peoples and their introduction of the Vedic religious current. At its peak, the Indus Valley civilization extended over most of present-day Pakistan, into India as far eastward as near Delhi, and southward as far as the estuaries of the Narmada River. It was apparently dominated by the two cities of Mohenjo-Daro, on the Indus River in Sind, and Harappa, about 350 miles to the northwest on a former course of the Ravi River, one of the tributaries to the Indus. Despite their distance from each other, the two cities show remarkable uniformity in material and design, and it has been supposed that they formed a pair of religious and administrative centers.

The determination of the nature of Indus Valley religion and of its residual impact upon Hinduism are, however, most problematic. Although archaeological sites have yielded many suggestive material remains, the interpretation of such finds is conjectural and has been thwarted especially by the continued resistance of the Indus Valley script, found on numerous steatite seals, to convincing decipherment. Until it is deciphered, little can be said with assurance. The content of the inscriptions may prove to be minimal, but if the language (most likely Dravidian) can be identified, much can be resolved.

At both Harappa and Mohenjo-Daro, the cities were dominated on the western side by an artificially elevated mound that housed a citadel-type complex of buildings. Though no temples or shrines can be identified, the complex probably served both sacred and administrative functions. A "great bath" within the Mohenjo-Daro citadel, plus elaborate bathing and drainage facilities in residences throughout the cities, suggests a strong concern for personal cleanliness, cultic bathing, and ritual purity such as resurface in later Hinduism. Indeed, the "great bath," a bitumen-lined tank with steps leading into and out of it from either end, suggests not only the temple tanks of later Hinduism but the notion of "crossing" associated with them through their Sanskrit name, tīrtha ("crossing place, ford").

A granary attached to the citadel may also have involved high officials in ceremonial supervision of harvests and other agricultural rituals. Terracotta female figurines with pedestal waists, found especially at village sites, reveal at least a popular cultic interest in fertility. They are probably linked with worship of a goddess under various aspects, for while some portray the figure in benign nurturing poses, others present pinched and grim features that have been likened to grinning skulls: These are likely foreshadowings of the Hindu Goddess in her benign and destructive aspects.

But most controversial are the depictions on the seals, whose inscriptions remain undeciphered. Most prominently figured are powerful male animals. They are often shown in cultic scenes, as before a sort of "sacred manger," or being led by a priestly ministrant before a figure (probably a deity and possibly a goddess) in a peepul tree, one of the most venerated trees in Hinduism. Male animals also frequently figure in combination with human males in composite animal-human forms. With female figures seemingly linked to the Goddess and males associated with animal power, it has been suggested that the two represent complementary aspects of a fertility cult with attendant sacrificial scenarios such as are found in the animal sacrifice to the Goddess in post-Vedic Hinduism. In such sacrifices the Goddess requires male offerings, and the animal represents the human male sacrificer. Most interesting and controveial in this connection is a figure in a yogic posture who is depicted on three seals and a faience sealing. Though features differ in the four portrayals, the most fully defined one shows him seated on a dais with an erect phallus. He has buffalo horns that enclose a treelike miter headdress, possibly a caricatured buffalo face, wears bangles and necklaces or torques, and is surrounded by four wild animals. Some of these associations (yoga, ithyphallicism, lordship of animals) have suggested an identification with the later Hindu god Śiva. Other traits (the buffalo-man composite form, association with wild animals, possible intimations of sacrifice) have suggested a foreshadowing of the buffalo demon Mashāsura, mythic antagonist and sacrificial victim of the later Hindu goddess Durgā. Possibly the image crystallizes traits that are later associated with both of these figures.

The notion that features of Indus Valley religion form a stream with later non-Aryan religious currents that percolate into Hinduism has somewhat dismissively been called the substratum theory by opponents who argue in favor of treating the development of Hinduism as derivable from within its own sacred literature. Though this "substratum" cannot be known except in the ways that it has been structured within Hinduism (and no doubt also within Jainism and Buddhism), it is clear that a two-way process was initiated as early as the Vedic period and has continued to the present.


The early sacred literature of Hinduism has the retrospective title of Veda ("knowledge") and is also known as śruti ("that which is heard"). Altogether it is a prodigious body of literature, originally oral in character (thus "heard"), that evolved into its present form over nine or ten centuries between about 1400 and 400 bce. In all, four types of texts fall under the Veda-śruti heading: Sahitās, Brāhamaas, rayakas, and Upaniads. At the fount of all later elaborations are the four Sahitās ("collections"): the gveda Sahitā (Veda of Chants, the oldest), the Sāmaveda and Yajurveda Sahitā s (Vedas of Melodies and Sacrificial Formulas, together known as the "liturgical" Sahitās), and the Atharvaveda Sahitā (the youngest, named after the sage Atharvan). These constitute the four Vedas, with some early sources referring to the "three Vedas" exclusive of the last. The material of the four was probably complete by 1000 bce, with younger parts of the older works overlapping older parts of the younger ones chronologically. The Sahitās, or portions of them, were preserved by different priestly schools or "branches" (śākhā s) through elaborate means of memorization. Many of these schools died out and their branches became lost, but others survived to preserve material for literary compilation and redaction. The subsequent works in the categories of Brāhamaa, rayaka, and Upaniad are all linked with one or another of the Vedic schools, and thus with a particular Vedic Sahitā, so that they represent the further literary output of the Vedic schools and also the interests of the four types of priests who came to be associated differentially with the ritual uses of the four Sahitās. It is from the gveda that Vedic religion in its earliest sense must be reconstructed.

Although the urban civilization of the Indus Valley had run its course by the time of the arrival of the Aryans in about 1500 bce, the newcomers met heirs of this civilization in settled agricultural communities. The contrast between cultures was striking to the Aryans, who described the indigenous population as having darker skin, defending themselves from forts, having no gods or religious rituals but nonetheless worshiping the phallus. As small stone phallic objects have been found at Indus Valley sites, this is probably an accurate description of a cult continued from pre-Vedic Indus Valley religion that prefigures the later veneration of the liga (phallus) in the worship of Śiva. In contrast to this predominantly agricultural population, the invading Aryans were a mobile, warlike people, unattached to cities or specific locations, entering Northwest India in tribal waves probably over a period of several centuries. Moreover, their society inherited an organizing principle from its Indo-European past that was to have great impact on later Indian civilization in the formation of the caste system. The ideal arrangement, which myths and ritual formulas propounded and society was to reflect, called for three social "functions": the priests, the warriors, and the agriculturalist-stockbreeders. Early Vedic hymns already speak of three such interacting social groups, plus a fourththe indigenous population of dāsa, or dasyu (literally, "slaves," first mythologized as demon foes of the Aryans and their gods). By the time of the late gveda, these peoples were recognized as a fourth "class" or "caste" in the total society and were known as śūdra s.

Most crucial to the inspiration of the early Vedic religion, however, was the interaction between the first two groups: the priesthood, organized around sacerdotal schools maintained through family and clan lines, and a warrior component, originally led by chieftains of the mobile tribal communities but from the beginning concerned with an ideal of kingship that soon took on more local forms. Whereas the priests served as repositories of sacred lore, poetry, ritual technique, and mystical speculation, the warriors served as patrons of the rites and ceremonies of the priests and as sponsors of their poetry. These two groups, ideally complementary but often having rival interests, crystallized by late Vedic and Brahmanic times into distinct "classes": the brāhamaa s (priests) and the katriya s (warriors).

Although the gveda alludes to numerous details of ritual that soon came to be systematized in the religion of the Brāhamaas, it brings ritual into relief only secondarily. The primary focus of the 1,028 hymns of the gveda is on praising the gods and the cosmic order (ta ), which they protect. But insofar as the hymns invoke the gods to attend the sacrifice, there is abundant interest in two deities of essentially ritual character: Agni and Soma. Agni (Fire) is more specifically the god of the sacrificial fire who receives offerings to the gods and conveys them heavenward through the smoke. And Soma is the divinized plant of "nondeath" (amta ), or immortality, whose juices are ritually extracted in the soma sacrifice, a central feature of many Vedic and Brahmanic rituals. These two gods, significantly close to humankind, are mediators between humans and other gods. But they are especially praised for their capacity to inspire in the poets the special "vision" (dhī) that stimulates the composition of the Vedic hymns. Agni, who as a god of fire and light is present in the three Vedic worlds (as fire on earth, lightning in the atmosphere, and the sun in heaven), bestows vision through "illumination" into the analogical connections and equivalences that compose the ta (which is itself said to have a luminous nature). Soma, the extracted and purified juice of the "plant of immortality," possibly the hallucinogenic fly agaric mushroom, yields a "purified" vision that is described as "enthused" or "intoxicated," tremulous or vibrant, again stimulating the inspiration for poetry. The Vedic poet (kavi, i, or vipra ) was thus a "seeer," or seer, who translated his vision into speech, thus producing the sacred mantra s, or verse-prayers, that comprise the Vedic hymns. Vedic utterance, itself hypostatized as the goddess Vāc (Speech), is thus the crystallization of this vision.

Vedic religion is decidedly polytheistic, there being far more than the so-called thirty-three gods, the number to which they are sometimes reduced. Though the point is controveial, for the sake of simplification one can say that at the core or "axis" of the pantheon there are certain deities with clear Indo-European or at least Indo-Iranian backgrounds: the liturgical gods Agni and Soma (cf. the Avestan deity Haoma) and the deities who oversee the three "functions" on the cosmic scale: the cosmic sovereign gods Varua and Mitra, the warrior god Indra, and the Aśvins, twin horsemen concerned with pastoralism, among other things. Intersecting this structure is an opposition of Indo-Iranian background between deva s and asura s. In the gveda both terms may refer to ranks among the gods, with asura being higher and more primal. But asura also has the Vedic meaning of "demon," which it retains in later Hinduism, so that the devaasura opposition also takes on dualistic overtones. Varua is the asura par excellence, whereas Indra is the leader of the deva s. These two deities are thus sometimes in opposition and sometimes in complementary roles: Varua being the remote overseer of the cosmic order (ta ) and punisher of individual human sins that violate it; Indra being the dynamic creator and upholder of that order, leader of the perennial fight against the collective demonic forces, both human and divine, that oppose it. It is particularly his conquest of the asura Vtra ("encloser")whose name suggests ambiguous etymological connections with Varuathat creates order or being (sat, analogous to ta ) out of chaos or nonbeing (asat ) and opens cosmic and earthly space for "freedom of movement" (varivas ) by gods and humans. Considerable attention is also devoted to three solar deities whose freedom of movement, thus secured, is a manifestation of the ta, a prominent analogy for which is the solar wheel: Sūrya and Savit (the Sun under different aspects) and Uas (charming goddess of the dawn). Other highly significant deities are Yama, god of the dead, and Vāyu, god of wind and breath. It is often pointed out that the gods who become most important in later HinduismViu, Śiva (Vedic Rudra), and the Goddessare statistically rather insignificant in the Veda, for few hymns are devoted to them. But the content rather than the quantity of the references hints at their significance. Viu's centrality and cosmological ultimacy, Rudra's destructive power and outsiderhood, and the this-worldly dynamic aspects of several goddesses are traits that assume great proportions in later characterizations of these deities.

Although it is thus possible to outline certain structural and historical features that go into the makeup of the Vedic pantheon, it is important to recognize that these are obscured by certain features of the hymns that arise from the type of religious "vision" that inspired them, and that provide the basis for speculative and philosophical trends that emerge in the late Veda and continue into the early Brahmanic tradition. The hymns glorify the god they address in terms generally applicable to other gods (brilliance, power, beneficence, wisdom) and often endow him or her with mythical traits and actions particular to other gods (supporting heaven, preparing the sun's path, slaying Vtra, and so on). Thus, while homologies and "connections" between the gods are envisioned, essential distinctions between them are implicitly denied. Speculation on what is essentialnot only as concerns the gods, but the ritual and the mantra s that invoke themis thus initiated in the poetic process of the early hymns and gains in urgency and refinement in late portions of the gveda and the subsequent "Vedic" speculative-philosophical literature that culminates in the Upaniads. Most important of these speculations historically were those concerning the cosmogonic sacrifices of Purua in gveda 10.90 (the Puruasūkta, accounting for, among other things, the origin of the four castes) and of Prajāpati in the Brāhamaas. Each must be discussed further. In addition, speculations on brahman as the power inherent in holy speech and on the ātman ("self") as the irreducible element of personal experience are both traceable to Vedic writings (the latter to the Atharvaveda only). This article shall observe the convergence of all these lines of speculation in the Upaniads and classical Hinduism.

Religion of the BrĀhamaas

The elaboration of Vedic religion into the sacrificial religion of the Brāhamaas is largely a result of systematization. The first indication of this trend is the compilation of the liturgical Sahitās and the development of the distinctive priestly schools and interests that produced these compendiums. Thus, while the gveda became the province of the hot priest, the pourer of oblations and invoker of gods through the mantra s (the term hot, "pourer," figures often in the gveda and has Indo-Iranian origins), the newer collections developed around the concerns of specialist priests barely alluded to in the gveda and serving originally in subordinate ritual roles. The Sāmaveda was a collection of verses taken mostly from the gveda, set to various melodies (sāman s) for use mainly in the soma sacrifice, and sung primarily by the udgāt priest, who thus came to surpass the hot as a specialist in the sound and articulation of the mantras. And the Yajurveda was a collection of yaju s, selected sacrificial mantra s, again mostly from the gveda, plus certain complete sentences, to be murmured by the adhvaryu priest, who concerned himself not so much with their sound as with their appropriateness in the ritual, in which he became effectively the master of ceremonies, responsible for carrying out all the basic manual operations, even replacing the hot priest as pourer of oblations. A fourth group of priests, the brāhamaa s, then claimed affiliation with the Atharvaveda and assumed the responsibility for overseeing the entire ritual performance of the other priests and counteracting any of their mistakes (they were supposed to know the other three Vedas as well as their own) by silent recitation of mantra s from the Atharvaveda. As specialization increased, each priest of these four main classes took on three main assistants.

The Brāhamaasexpositions of brahman, the sacred power inherent in mantra and more specifically now in the ritualare the outgrowth of the concerns of these distinctive priestly schools and the first articulation of their religion. Each class of priests developed its own Brāhamaas, the most important and comprehensive being the Śatapatha Brāhamaa of one of the Yajurveda schools. The ritual system was also further refined in additional manuals: the Śrautasūtras, concerned with "solemn" rites, first described in the Brāhamaas and thus called śrauta because of their provenance in these śruti texts, and the Ghyasūtras, concerned with domestic rites (from gha, "home"), justified by "tradition" (smti ) but still having much of Vedic origins. The Śrautasūtras were compiled over the period, roughly, from the Brāhamaas to the Upaniads, and the Ghyasūtras were probably compiled during Upaniadic times.

The domestic rites take place at a single offering fire and usually involve offerings of only grain or ghee (clarified butter). Along with the maintenance of the household fire and the performance of the so-called Five Great Sacrificesto brahman (in the form of Vedic recitation), to ancestors, to gods, to other "beings," and to humans (hospitality rites)the most prominent ghya ceremonies are the sacraments or life-cycle rites (saskāra s). Of these, the most important are the rites of conception and birth of a male child; the Upanayana, or "introduction," of boys to a brāhamaa preceptor or guru for initiation; marriage; and death by cremation (Antyei, "final offering"). The Upanayana, involving the investiture of boys of the upper three social classes (vara s) with a sacred thread, conferred on them the status of "twice-born" (dvija, a term first used in the Atharvaveda ), and their "second birth" permitted them to hear the Veda and thereby participate in the śrauta rites that, according to the emerging Brahmanic orthodoxy, would make it possible to obtain immortality.

The śrauta rites are more elaborate and are representative of the sacrificial system in its full complexity, involving ceremonies that lasted up to two years and enlisted as many as seventeen priests. Through the continued performance of daily, bimonthly, and seasonal śrauta rites one gains the year, which is itself identified with the sacrificial life-death-regeneration round and its divine personification, Prajāpati. In surpassing the year by the Agnicayana, the "piling of the fire altar," one gains immortality and needs no more nourishment in the otherworld (see Śatapatha Brāhamaa

Śrauta rites required a sacrificial terrain near the home of the sacrificer (yajamāna ), with three sacred fires (representing, among other things, the three worlds) and an upraised altar, or vedī. Nonanimal sacrifices of the first varieties mentioned involved offerings of milk and vegetable substances or even of mantra s. Animal sacrifices (paśubandhu )which required a more elaborate sacrificial area with a supplemental altar and a sacrificial stake (yūpā )entailed primarily the sacrifice of a goat. Five male animalsman, horse, bull, ram, and goatare declared suitable for sacrifice. It is likely, however, that human sacrifice existed only on the "ideal" plane, where it was personified in the cosmic sacrifices of Purua and Prajāpati. The animal (paśu ) was to be immolated by strangulation, and its omentum, rich in fat, offered into the fire. Soma sacrifices, which would normally incorporate animal sacrifices within them plus a vast number of other subrites, involved the pressing and offering of soma. The most basic of these was the annual Agnioma, "in praise of Agni," a four-day rite culminating in morning, afternoon, and evening soma pressings on the final day and including two goat sacrifices. Three of the most ambitious soma sacrifices were royal rites: the Aśvamedha, the horse sacrifice; the Rājasūya, royal consecration; and the Vājapeya, a soma sacrifice of the "drink of strength." But the most complex of all was the aforementioned Agnicayana.

A thread that runs through most śrauta rituals, however, is that they must begin with the "faith" or "confidence" (śrāddha) of the sacrificer in the efficacy of the rite and the capacity of the officiating priests to perform it correctly. This prepares the sacrificer for the consecration (dīkā) in which, through acts of asceticism (tapas), he takes on the aspect of an embryo to be reborn through the rite. As dīkita (one undergoing the dīkā), he makes an offering of himself (his ātman ). This then prepares him to make the sacrificial offering proper (the yajña, "sacrifice") as a means to redeem or ransom this self by the substance (animal or otherwise) offered. Then, reveing the concentration of power that he has amassed in the dīkā, he disperses wealth in the form of dakinā s (honoraria) to the priests. Finally, the rite is disassembled (the ritual analogue to the repeated death of Prajāpati before his reconstitution in another rite), and the sacrificer and his wife bathe to disengage themselves from the sacrifice and reenter the profane world.

In the elaboration of such ceremonies and the speculative explanation of them in the Brāhamaas, the earlier Vedic religion seems to have been much altered. In the religion of the Brāhamaas, the priests, as "those who know thus" (evamvid s), view themselves as more powerful than the gods. Meanwhile, the gods and the demons (asura s) are reduced to representing in their endless conflicts the recurrent interplay between agonistic forces in the sacrifice. It is their father, Prajāpati, who crystallizes the concerns of Brahmanic thought by representing the sacrifice in all its aspects and processes. Most notable of these is the notion of the assembly or fabrication of an immortal self (ātman ) through ritual action (karman ), a self constructed for the sacrificer by which he identifies with the immortal essence of Prajāpati as the sacrifice personified. And by the same token, the recurrent death (punarmtyu, "redeath") of Prajāpati's transitory nature (the elements of the sacrifice that are assembled and disassembled) figures in the Brāhamaas as the object to be avoided for the sacrificer by the correct ritual performance. This Brahmanic concept of Prajāpati's redeath, along with speculation on the ancestral ghya rites (śrāddha s) focused on feeding deceased relatives to sustain them in the afterlife, must have been factors in the thinking that gave rise to the Upaniadic concept of reincarnation (punarjanman, "rebirth"). The emphasis on the morbid and transitory aspects of Prajāpati and the sacrifice, and the insistence that asceticism within the sacrifice is the main means to overcome them, are most vigorously propounded in connection with the Agnicayana.

In the Brāhamaas' recasting of the primal once- and for-all sacrifice of Purua into the recurrent life-death-regeneration mythology of Prajāpati, a different theology was introduced. Though sometimes Purua was identified with Prajāpati, the latter, bound to the round of creation and destruction, became the prototype for the classical god Brahmā, personification of the Absolute (brahman ) as it is oriented toward the world. The concept of a transcendent Purua, however, was not forgotten in the Brāhamaas. Śatapatha Brāhamaa 13.6 mentions Purua-Nārāyaa, a being who seeks to surpass all others through sacrifice and thereby become the universe. In classical Hinduism, Nārāyaa and Purua are both names for Viu as the supreme divinity. This Brāhamaa passage neither authorizes nor disallows an identification with Viu, but other Brāhamaa passages leave no doubt that sacrificial formulations have given Viu and Rudra-Śiva a new status. Whereas the Brāhamaas repeatedly assert that "Viu is the sacrifice"principally in terms of the organization of sacrificial space that is brought about through Viu's three steps through the cosmos, and his promotion of the order and prosperity that thus accruethey portray Rudra as the essential outsider to this sacrificial order, the one who neutralizes the impure forces that threaten it from outside as well as the violence that is inherent within. Biardeau (1976) has been able to show that the later elevation of Viu and Śiva through yoga and bhakti is rooted in oppositional complementarities first formulated in the context of the Brahmanic sacrifice.

The Upaniads

Several trends contributed to the emergence of the Upaniadic outlook. Earlier speculations on the irreducible essence of the cosmos, the sacrifice, and individual experience have been mentioned. Pre-Upaniadic texts also refer to various forms of asceticism as performed by types of people who in one way or another rejected or inverted conventional social norms: the Vedic muni, vrātya, and brahmacārin, to each of whom is ascribed ecstatic capacities, and, at the very heart of the Brahmanic sacrifice, the dīkita (the sacrificer who performs tapas while undergoing the dīkā, or consecration). These speculative and ascetic trends all make contributions to a class of texts generally regarded as intermediary between the Brāhamaas and Upaniads: the rayakas, or "Forest Books." The rayakas do not differ markedly from the works that precede and succeed them (the Bhadārayaka Upaniad is both an rayaka and an Upaniad), but their transitional character is marked by a shift in the sacrificial setting from domestic surroundings to the forest and a focus not so much on the details of ritual as on its interiorization and universalization. Sacrifice, for instance, is likened to the alternation that takes place between breathing and speaking. Thus correspondences are established between aspects of sacrifice and the life continuum of the meditator.

An upaniad is literally a mysticaloften "secret""connection," interpreted as the teaching of mystical homologies. Or, in a more conventional etymology, it is the "sitting down" of a disciple "near to" (upa, "near"; ni, "down"; sad, "sit") his spiritual master, or guru. Each Upaniad reflects the Vedic orientation of its priestly school. There are also regional orientations, for Upaniadic geography registers the further eastern settlement of the Vedic tradition into areas of the Ganges Basin. But the Upaniads do share certain fundamental points of outlook that are more basic than their differences. Vedic polytheism is demythologized, for all gods are reducible to one. Brahmanic ritualism is reassessed and its understanding of ritual action (karman ) thoroughly reinterpreted. Karman can no longer be regarded as a positive means to the constitution of a permanent self. Rather, it is ultimately negative: "the world that is won by work (karman )" and "the world that is won by merit (puya )" only perish (Chāndogya Upaniad 8.1.6). The "law of karma" (karman ) or "law of causality" represents a strict and universal cause-effect continuum that affects any action that is motivated by desire (kāma ), whether it be desire for good or for ill. Thus even meritorious actions that lead to the Vedic heaven "perish," leaving a momentum that carries the individual to additional births or reincarnations. The result is perpetual bondage to the universal flow-continuum of all karman, or saśara (from sa, "together" and s, "flow"), a term that the Upaniads introduce into the Vedic tradition but that is shared with Jainism and Buddhism. As with these religions, the Upaniads and Hinduism henceforth conceive their soteriological goal as liberation from this cycle of saśara: that is, moka or mukti ("release").

Moka cannot be achieved by action alone, because action only leads to further action. Thus, though ritual action is not generally rejected and is often still encouraged in the Upaniads, it can only be subordinated to pursuit of the higher moka ideal. Rather, the new emphasis is on knowledge (vidyā, jñāna ) and the overcoming of ignorance (avidyā ). The knowledge sought, however, is not that of ritual technique or even of ritual-based homologies, but a graspable, revelatory, and experiential knowledge of the self as one with ultimate reality. In the early Upaniads this experience is formulated as the realization of the ultimate "connection," the oneness of ātman-brahman, a connection knowable only in the context of communication from guru to disciple. (Herein can be seen the basis of the parable context and vivid, immediate imagery of many Upaniadic teachings.) The experience thus achieved is variously described as one of unified consciousness, fearlessness, bliss, and tranquillity.

Beyond these common themes, however, and despite the fact that Upaniadic thought is resistant to systematization, certain different strains can be identified. Of the thirteen Upaniads usually counted as śruti, the earliest (c. 700500 bce) are those in prose, headed by the Bhadārayaka and the Chāndogya. Generally, it may be said that these Upaniads introduce the formulations that later Hinduism will develop into the sanyāsa ideal of renunciation (not yet defined in the Upaniads as a fourth stage of life) and the knowledge-path outlook of nondualistic (advaita ) Vedānta. Even within these early Upaniads, two approaches to realization can be distinguished. One refers to an all-excluding Absolute; the self that is identified with brahman, characterized as neti neti ("not this, not this"), is reached through a paring away of the psychomental continuum and its links with karman. Such an approach dominates the Bhadā-rayaka Upaniad. Avidyā here results from regarding the name and form of things as real and forming attachment to them. The other approach involves an all-comprehensive Absolute, brahman-ātman, which penetrates the world so that all forms are modifications of the one; ignorance results from the failure to experience this immediacy. In the Chāndogya Upaniad this second approach is epitomized in the persistent formula "Tat tvam asi" ("That thou art").

The later Vedic Upaniads (c. 600400 bce) register the first impact of theistic devotional formulations, and of early Sākhya and Yoga. Most important of these historically are two "yogic" Upaniads, the Śvetāśvatara and the Kaha, the first focused on Rudra-Śiva and the second on Viu. Each incorporates into its terminology for the absolute deity the earlier term purua. As Biardeau has shown in L'hindouisme (1981), they thus draw on an alternate term for the Absolute from that made current in the brahman-ātman equation. The Purua of gveda 10.90 (the Puruasūkta ) is sacrificed to create the ordered and integrated sociocosmic world of Vedic man. But only one quarter of this Purua is "all beings"; three quarters are "the immortal in heaven" (RV 10.90.3). This transcendent aspect of Purua, and also a certain "personal" dimension, are traits that were retained in the characterization of Purua-Nārāyaa in the Śatapatha Brāhamaa and reinforced in the yogic characterizations of Rudra-Śiva and Viu in the previously mentioned Upaniads. The Upaniadic texts do not restrict the usage of the term Purua to mean "soul," as classical Sākhya later does; rather, it is used to refer to both the soul and the supreme divinity. The relation between the soul and the Absolute is thus doubly defined: on the one hand as ātman-brahman, on the other as purua -Purua. In the latter case, the Kaha Upaniad describes a spiritual itinerary of the soul's ascent through yogic states to the supreme Purua, Viu. This synthesis of yoga and bhakti will be carried forward into the devotional formulations of the epics and the Purāas. But one must note that the two vocabularies are used concurrently and interrelatedly in the Upaniads, as they will be in the later bhakti formulations.

The Consolidation of Classical Hinduism

A period of consolidation, sometimes identified as one of "Hindu synthesis," "Brahmanic synthesis," or "orthodox synthesis," takes place between the time of the late Vedic Upaniads (c. 500 bce) and the period of Gupta imperial ascendancy (c. 320467 ce). Discussion of this consolidation, however, is initially complicated by a lack of historiographical categories adequate to the task of integrating the diverse textual, inscriptional, and archaeological data of this long formative period. The attempt to cover as much of this span as possible with the name "epic period," because it coincides with the dates that are usually assigned to the formation and completion of the Hindu epics (particularly the Mahābhārata ), is misleading, because so much of what transpires can hardly be labeled "epic." On the other hand, attempts to define the period in terms of heterogeneous forces operating upon Hinduism from within (assimilation of local deities and cults, geographical spread) and without (heterodox and foreign challenges) either have failed to register or have misrepresented the implications of the apparent fact that the epics were "works in progress" during the whole period. The view one takes of the epics is, in fact, crucial for the interpretation of Hinduism during this period. Here, assuming that the epics already incorporated a bhakti cosmology and theology from an early point in this formative period, this article shall try to place them in relation to other works and formulations that contributed to the consolidation of classical Hinduism.

The overall history can be broken down into four periods characterized by an oscillation from disunity (rival regional kingdoms and tribal confederacies on the Ganges Plain) to unity (Mauryan ascendancy, c. 324184 bce, including the imperial patronage of Buddhism by Aśoka) to disunity (rival foreign kingdoms in Northwest India and regional kingdoms elsewhere) back to unity (Gupta ascendancy, c. 320467 ce). The emerging self-definitions of Hinduism were forged in the context of continued interaction with heterodox religions (Buddhists, Jains, jīvikas) throughout this whole period, and with foreign peoples (Yavanas, or Greeks; Śakas, or Scythians; Pahlavas, or Parthians; and Kūāas, or Kushans) from the third phase on. In this climate the ideal of centralized Hindu rule attained no practical realization until the rise of the Guptas. That this ideal preceded its realization is evident in the rituals of royal paramountcy (Aśvamedha and Rājasūya) that were set out in the Brāhamaas and the Śrautasūtras, and actually performed by post-Mauryan regional Hindu kings.

When one looks to the component facets of the overall consolidation, these four periods must be kept in mind, but with the proviso that datings continue to be problematic: not only datings of texts, but especially of religious movements and processes reflected in them, and in surviving inscriptions. Most scholars ordinarily assume that when a process is referred to in a text or other document, it has gone on for some time.

Śruti and smti

Fundamental to the self-definition of Hinduism during this period of its consolidation is the distinction it makes between two classes of its literature: śruti and smti. Śruti is "what is heard," and refers to the whole corpus of Vedic literature (also called Veda) from the four Vedas to the Upaniads. Smti, "what is remembered" or "tradition," includes all that falls outside this literature. Exactly when this distinction was made is not certain, but it is noteworthy that the six Vedāgas or "limbs of the Veda" (writings on phonetics, metrics, grammar, etymology, astronomy, and ritual) are smti texts that were composed at least in part during the latter half of the Vedic or śruti period. The ritual texts (Kalpasūtras) are subdivided into three categories: Śrautasūtras, Ghyasūtras, and Dharmasūtras. Whereas the first two (discussed above under Brahmanic ritual) pertain to concerns developed in the Vedic period, the Dharmasūtras focus on issues of law (dharma ) that become characteristic of the period now under discussion. Dates given for the composition of these texts run from 600 to 300 bce for the earliest (Gautama Dharmasūtra ) to 400 ce for the more recent works. Both Ghyasūtras and Dharmasūtras were sometimes called Smārtasūtras (i.e., sūtra s based on smti ), so it seems that their authors regarded them as representative of the prolongation of Vedic orthodoxy (and orthopraxy) that the smti category was designed to achieve. As the term smti was extended in its use, however, it also came to cover numerous other texts composed in the post-Upaniadic period.

This śruti/smti distinction thus marks off the earlier literature as a unique corpus that, once the distinction was made, was retrospectively sanctified. By the time of the Manāva Dharmaśāstra, or Laws of Manu (c. 200 bce100 ce; see Manu 1.23), and probably before this, śruti had come to be regarded as "eternal." Its components were thus not works of history. The Vedic i s had "heard" truths that are eternal, and not only in contentthe words of the Vedas are stated to have eternal connection with their meaningsbut also in form. The works thus bear no stamp of the i s' individuality. Such thinking crystallized in the further doctrine that the Vedas (i.e., śruti ) are apaurueya, not of personal authorship (literally, "not by a purua "). They thus have no human imperfection. Further, it was argued that they are even beyond the authorship of a divine "person" (Purua ). Though myths of the period assert that the Vedas spring from Brahmā at the beginning of each creation (as the three Vedas spring from Purua in the Puruasūkta ), the deity is not their author. Merely reborn with him, they are a self-revelation of the impersonal brahman. In contrast to śruti, smti texts were seen as historical or "traditional," passed on by "memory" (smti ), and as works of individual authors (paurueya ), even though mythical authorsboth human and divineoften had to be invented for them.

Smti texts of this period thus proclaim the authority of the Veda in many ways, and nonrejection of the Veda comes to be one of the most important touchstones for defining Hinduism over and against the heterodoxies, which rejected the Veda. In fact, it is quite likely that the doctrines of the eternality and impersonality of the Veda were in part designed to assert the superiority of the Veda over the "authored" and "historical" works of the heterodoxies, whose teachings would thus be on a par with smti rather than śruti. But it is also likely that the apaurueya doctrine is designed to relativize the "personal" god of bhakti. In any case, these doctrines served to place a considerable ideological distance between śruti and smti, and to allow smti authors great latitude in interpreting śruti and extending Hindu teachings into new areas. Smti thus supposedly functioned to clarify the obscurities of the Veda. But the claim that smti texts need only not contradict the Veda left their authors great freedom in pursuing new formulations.

Varāśramadharma ("caste and life-stage law")

The most representative corpus of smti literature, and the most closely tied to the continued unfolding orthodox interests of the Vedic priestly schools, is that concerned with dharma ("law" or "duty"). As a literary corpus, it consists of two kinds of texts: the Dharmasūtras (600/300 bce400 ce), already mentioned in connection with the śruti/smti distinction, and the Dharmaśāstras. The most important and earliest of the latter are the Mānava Dharmaśāstra, or Laws of Manu (c. 200 bce100 ce), and the Yājñavalkya Smti (c. 100300 ce). But other Dharmaśāstras were composed late into the first millennium, to be followed by important commentaries on all such texts. The main focus of these two classes of texts is fundamentally identical: the articulation of norms for all forms of social interaction, thus including but going far beyond the earlier Sūtras' concern for ritual. Four differences, however, are noteworthy: (1) Whereas the Dharmasūtras are in prose, the Dharmaśāstras are in the same poetic meter as the epics, Manu in particular having much material in common with the Mahābhārata. (2) Whereas the Sūtras are still linked with the Vedic schools, the Śāstras are not, showing that study and teaching of dharma had come to be an independent discipline of its own. (3) The Śāstra legislation is more extended and comprehensive. (4) The Śāstras are more integrated into a mythic and cosmological vision akin to that in bhakti texts, but usually ignoring bhakti as such, with references to duties appropriate to different yuga s (ages), and the identification of north central India as the "middle region" (madhyadeśa ) where the dharma is (and is to be kept) the purest.

The theory of varāśramadharma, the law of castes and life stages, was worked out in these texts as a model for the whole of Hindu society. There is little doubt that it was stimulated by the alternate lay/monastic social models of the heterodoxies, and no doubt that it was spurred on by the incuions of barbarian peoplesfrequently named in these texts as mleccha s (those who "jabber")into the Northwest. The model involves the working out of the correlations between two ideals: first, that society conform to four hierarchical castes, and second, that a person should pass through four life stages (āśrama s): student (brahmacārin ), householder (ghasthin ), forest dweller (vānaprasthin ), and renunciant (sanyāsin ). The first ideal is rooted in the Puruasūkta. The second presupposes the śruti corpus, because the four life stages are correlated with the four classes of śruti texts. Thus the student learns one of the Vedas, the householder performs domestic and optimally also śrauta rituals of the Brāhamaas, the forest dweller follows the teachings of the rayakas, and the sanyāsin follows a path of renunciation toward the Upaniadic goal of moka. But although all the life stages are either mentioned (as are the first two) or implied in the śruti corpus, the theory that they should govern the ideal course of individual life is new to the Dharmasūtras. Together, the vara and āśrama ideals take on tremendous complexity, because a person's duties vary according to caste and stage of life, not to mention other factors like sex, family, region, and the quality of the times. Also, whereas a person's development through one life ideally is regulated by the āśrama ideal, the passage through many reincarnations would involve birth into different castes, the caste of one's birth being the result of previous karman. A further implication is that the life stages can be properly pursued only by male members of the three twice-born vara s, as they alone can undergo the Upanayana ritual that begins the student stage and allows the performance of the rites pertinent to succeeding stages.

Each of these formulations has peisted more on the ideal plane than the real. In the case of the four āśrama s, most people never went beyond the householder stage, which the Sūtras and Śāstras actually exalt as the most important of the four, because it is the support of the other three and, in more general terms, the mainstay of the society. The forest-dweller stage may soon have become more legendary than real: In epic stories it was projected onto the Vedic i s. The main tension, however, that peists in orthodox Hinduism is that between the householder and the renunciant, the challenge being for anyone to integrate into one lifetime these two ideals, which the heterodoxies set out for separate lay and monastic communities.

As to the four vara s, the ideal represents society as working to the reciprocal advantage of all the castes, each one having duties necessary to the proper functioning of the whole and the perpetuation of the hierarchical principle that defines the whole. Thus Brāhamaa s are at the top, distinguished by three duties that they share with no other caste: teaching the Veda, assisting in sacrifice, and accepting gifts. They are said to have no king but Soma, god of the sacrifice. In actual fact the traditional śrauta sacrifice counted for less and less in the brāhamaa householder life, and increasing attention was given to the maintenance of brāhamaa purity for the purpose of domestic and eventually temple rituals that, in effect, universalized sacrifice as the brāhamaa 's dharma, but a sacrifice that required only the minimum of impure violence. This quest for purity was reinforced by brāhamaa s' adoption into their householder life of aspects of the sanyāsa ideal of renunciation. This was focused especially on increasing espousal of the doctrine of ahisā (nonviolence, or, more literally, "not desiring to kill") and was applied practically to vegetarianism, which becomes during this period the brāhamaa norm. Brāhamaa s thus retain higher rank than katriya s, even though the latter wield temporal power (katra ) and have the specific and potentially impure duties of bearing weapons and protecting and punishing with the royal staff (daa ). The subordination of king to brāhamaa involves a subordination of power to hierarchy that is duplicated in contemporary rural and regional terms in the practice of ranking brāhamaa s above locally dominant castes whose power lies in their landed wealth and numbers. Vaiśya s have the duties of stock breeding, agriculture, and commerce (including money lending). Certain duties then distinguished the three twice-born castes as a group from the śūdra s. All three upper vara s thus study the Veda, perform sacrifices, and make gifts, whereas śūdra s are permitted only lesser sacrifices (pākayajña s) and simplified domestic rituals that do not require Vedic recitation.

Actual conditions, however, were (and still are) much more complex. The four-vara model provided the authors of the dharma texts with Vedic "categories" within which to assign a basically unlimited variety of heterogeneous social entities including indigenous tribes, barbarian invaders, artisan communities and guilds (śrei s), and specialists in various services. Susceptible to further refinement in ranking and regional nomenclature, all such groups were called jāti s, a term meaning "birth" and in functional terms the proper word to be translated "caste." Thus, although they are frequently called subcastes, the jāti s are the castes proper that the law books classified into the "categories" of vara.

To account for this proliferation of jāti s, the authors asserted that they arose from cross-breeding of the vara s. Two possibilities were thus presented: anuloma ("with the grain") unions, in which the husband's vara was the same as his wife's or higher (in anthropological terms, hypergamous, in which women are "married up"), and pratiloma ("against the grain") unions, in which the wife's vara would be higher than the husband's (hypogamous, in which women are "married down"). Endogamous marriage (marriage within one's own vara ) set the highest standard and was according to some authorities the only true marriage. But of the other two, whereas anuloma marriages were permitted, pratiloma unions brought disgrace. Thus the jāti s supposedly born from anuloma unions were less disgraced than those born from pratiloma unions. Significantly, two of the most problematic jāti s were said to have been born from the most debased pratiloma connections: the Yavanas (Greeks) from śūdra males and katriya females (similar origins were ascribed to other "barbarians") and the caāla s (lowest of the low, mentioned already in the Upaniads, and early Buddhist literature, as a "fifth caste" of untouchables) from the polluting contact of śūdra males and brāhamaa females. It should be noted that a major implication of the prohibition of pratiloma marriage is the limitation for brāhamaa women to marriages with only brāhamaa men. This established at the highest rank an association of caste purity with caste endogamy (and the purity of a caste's women) and thus initiated an endogamous standard that was adopted by all castesnot just vara s but jāti sby the end of the first millennium.

This accounting of the emergence of jāti s was integrated with further explanations of how society had departed from its ideal. One is that "mixing of caste"the great abomination of the dharma texts and also of the Bhagavad-gītā increases with the decline of dharma from yuga to yuga, and is especially pernicious in this Kali age. Another is the doctrine of āpad dharma, "duties for times of distress" such as permit inversion of caste roles when life is threatened. A third doctrine developed in the Dharmaśāstras identifies certain duties (kalivarjya s) as once allowed but now prohibited in the kaliyuga because people are no longer capable of performing them purely. Through all this, however, the ideal peists as one that embraces a whole society despite variations over time and space.

The four puruārthas (goals of humankind)

The theory that the integrated life involves the pursuit of four goals (artha s) is first presented in the Dharmaśāstras and the epics, in the latter cases through repeated narrative illustrations. The development of distinctive technical interpretations of each artha, or facets thereof, can also be followed during the period in separate manuals: the Arthaśāstra, a manual on statecraft attributed to Candragupta Maurya's minister Kauilya but probably dating from several centuries later, on artha (in the sense now of "material pursuits"); the Kāmasūtras, most notably that of Vātsyāyana (c. 400 ce), on kāma ("love, desire"); the already discussed Dharmasūtras and Dharmaśāstras on dharma; and the Sūtras of the "philosophical schools" (darśana s) insofar as they are concerned with the fourth goal, moka. Early sources often refer to the first three goals as the trivarga, the "three categories," but this need not imply that the fourth goal is added later. The Dharmaśāstra and epic texts that mention the trivarga are focused on the concerns of the householderand, in the epics, particularly of the royal householderthese being the context for the pursuit of the trivarga. The fourth goal, moka, is to be pursued throughout lifeindeed, throughout all livesbut is especially the goal of those who have entered the fourth life stage of the sanyāsin. The trivarga-moka opposition thus replicates the householder-renunciant opposition. But the overall purpose of the puruārtha formulation is integrative and complementary to the varāśramadharma theory. From the angle of the householder, it is dharma that integrates the trivarga as a basis for moka. But from the angle of the sanyāsin, it is kāma that lies at the root of the trivarga, representing attachment in all forms, even to dharma. Paths to liberation will thus focus on detachment from desire, or its transformation into love of God.

Philosophical "viewpoints" (darśanas) and paths to salvation

As an expression of Hinduism's increasing concern to systematize its teachings, the fourth goal of life (moka ) was made the subject of efforts to develop distinctly Hindu philosophical "viewpoints" (darśana s, from the root dś, "see") on the nature of reality and to recommend paths to its apprehension and the release from bondage to karman. Six Hindu darśana s were defined, and during the period in question each produced fundamental textsin most cases sūtra sthat served as the bases for later commentaries.

In terms of mainstream developments within Hinduism, only two schools have ongoing continuity into the present: the Mīmāsā and the Vedānta. And of these, only the latter has unfolded in important ways in the postsynthesis period. Nonetheless, all six have made important contributions to later Hinduism. It must thus suffice to discuss them all briefly at this point in terms of their basic features and major impact, and reserve fuller discussion of the Vedānta alone for the period of its later unfolding.

Of the six schools, twoMīmāsā and Vedāntaare rooted primarily in the Vedic śruti tradition and are thus sometimes called smārta schools in the sense that they develop smārta orthodox currents of thought that are based, like smti, directly on śruti. The other fourNyāya, Vaiśeika, Sākhya, and Yogaclaim loyalty to the Veda, yet are quite independent of it, their focus instead being on rational or causal explanation. They are thus sometimes called haituka schools (from hetu, "cause, reason").

Of the smārta schools, the Mīmāsā is most concerned with ritual traditions rooted in the Vedas and the Brāhamaas, whereas the Vedānta is focused on the Upaniads. It is notable that both sustain Vedic orientations that reject (Mīmāsā) or subordinate (Vedānta) bhakti until the Vedānta is devotionalized in its post-Śakara forms. Beginning with Jaimini's Mīmāsā Sūtra (c. 300100 bce), Mīmāsā ("reflection, interpretation") provides exegesis of Vedic injunctive speech, in particular as it concerns the relationship between intentions and rewards of sacrifice. Great refinement is brought to bear on issues relating to the authority and eternalness of the Veda and the relationship between its sounds, words, and meanings. Vedic injunctions are taken literally, the many Vedic gods are seen as real although superfluous to salvation (there is an anti-bhakti stance here), and it is maintained that the proper use of injunctions is alone enough to secure the attainment of heaven (not a higher release, or moka, as propounded by all the other systems, including bhakti ). Mīmāsā persists in two subschools, but only in small numbers among brahman ritualists.

As to the Vedānta ("end of the Veda," a term also used for the Upaniads), the foundational work is Bādarāyaa's Vedānta Sūtra, or Brahma Sūtra (c. 300100 bce), an exegesis of various Upaniadic passages in aphoristic style easily susceptible to divergent interpretations. These it received in the hands of later Vedantic thinkers.

The haituka schools are notable for their development, for the first time within Hinduism, of what may be called maps and paths: that is, maps of the constituent features of the cosmos, and paths to deliverance from bondage. Emerging within Hindusim at this period, and particularly in the schools least affiliated with the Vedic tradition, such concerns no doubt represent an effort to counter the proliferation of maps and paths set forth by the heterodoxies (not only Buddhism and Jainism, but the jīvikas). They allow for a somewhat more open recognition of the deity of bhakti (Sākhya excepted) than do the smārta schools, though none of the haituka schools makes it truly central.

Nyāya and Vaiśeika, systems first propounded in Gautama's Nyāya Sūtra (c. 200 bce150 ce) and Kaāda's Vaiśeika Sūtra (c. 200 bce100 ce), were quickly recognized as a hyphenated pair: Nyāya-Vaiśeika. Nyāya ("rule, logic, analysis"), emphasizing logic and methods of argumentation as means to liberation, was viewed as complementary to Vaiśeika ("school of distinct characteristics"), which advanced a theory of atomism and posited seven categories to explain such things as atomic aggregation and dualistic distinction between soul and matter. At least by about the fifth century, when the two schools had conjoined, Nyāya logic and Vaiśeika cosmology served to provide influential arguments from design for the existence of God as the efficient cause of the creation and destruction of the universe and liberator of the soul from karman.

Far more influential, however, were the pair Sākhya ("enumeration") and Yoga. The foundational texts of these schools may be later than those of the others, but they are clearly distillations of long-continuing traditions, datable at least to the middle Upaniads, that had already undergone considerable systematization. Thus Patañjali's Yoga Sūtra is from either about 200 bce or 300500 ce, depending on whether or not one identifies the author with the grammarian who lived at the earlier date. And Īśvaraka's Sākhyakārikās probably date from the fourth century ce. Even though Sākhya's "atheism" and its soteriology of the isolation (kaivalya) of the soul (purua) from matter (prakti) have been modified or rejected in other forms of Hinduism (both doctrines may link Sākhya with Jainism), Sākhya's cosmology and basic terminology have become definitive for Hinduism at many levels: not only in the Vedānta, but in bhakti and Tantric formulations as well. In fact, given the preclassical forms of theistic Sākhya founded in the Upaniads and the Mahābhārata and their use in bhakti cosmologies, it may well be that the atheism of the classical Sākhya results from a rejection of bhakti elements from a fundamentally theistic system. Sākhya thus posits purua without a transcendent, divine Purua, and its prakti is also abstract and impersonal.

In any case, a number of Sākhya concepts became basic to the Hindu vocabulary, only to be integrated and reinterpreted from different theological and soteriological perspectives by other schools. These include the concepts of the evolution and devolution of prakti, the sexual polarity of purua as male and prakti as female, the enumeration of twenty-three substances that evolve from and devolve back into the prakti "matrix," the concept of matter as a continuum from subtle psychomental "substances" to gross physical ones (in particular the five elements), and the notion of the three "strands" or "qualities" called gua s (sattva, goodness, lucidity; rajas, dynamism; tamas, entropy), which are "braided" together through all matter from the subtle to the gross.

Meanwhile, whereas Sākhya provides the map to be "known," Yoga defines the path by which purua can extricate itself from prakti. The "eight limbs" of Yoga (an answer to the Eightfold Path of Buddhism?) represent the most important Hindu formulation of a step-by-step (though also cumulative) path to liberation. The first two "limbs" involve forms of restraint (yama ) and observance (niyama ). The next three involve integration of the body and senses: posture (āsana ), breath control (prāāyama ), and withdrawal of the senses from the dominance of sense objects (pratyāhāra ). The last three achieve the integration of the mind or the "cessation of the mental turmoil" that is rooted in the effects of karman: "holding" (dhāraā ) to a meditative support, meditative fluency (dhyāna ), and integrative concentration (samādhī ) through which the freedom of purua can be experienced.

The classical Yoga of Patañjali, known as rājayoga ("royal yoga"), diverges from the Sākhya in acknowledging the existence of God (Īśvara). But Īśvara is a focus of meditation, not an agent in the process of liberation. The use of the term rājayoga, however, suggests that by Patañjali's time the term yoga had already been used to describe other disciplines or paths, resulting in a situation where the terms yoga ("yoke") and mārga ("path") had become interchangeable. One will thus find rājayoga mentioned later along with the more generalized "yogas," or "paths," that become definitive for Hinduism through their exposition in the Bhagavadgītā (c. 200 bce): the paths (or yogas) of karman ("action"), jñāna ("knowledge"), and bhakti ("devotion").

Classical bhakti Hinduism

The consolidation of Hinduism takes place under the sign of bhakti. And though Mīmāsā ritualism and Vedantic and other "knowledge" trends continue to affiliate with an "orthodox" strain that resists this synthesis, or attempts to improve upon it, classical bhakti emerges as constitutive henceforth of mainstream Hinduism, including forms of devotional sectarianism.

Intimations of bhakti developments are registered as early as the late Vedic Upaniads, and in inscriptions and other records of syncretistic worship of Hindu deities (Viu and Śiva) alongside foreign and heterodox figures in the early centuries of the common era. However, the heterogeneity and scattered nature of the nontextual information available on the emergence of bhakti during this period have allowed for conflicting interpretations of the salient features of the process. But rather than reweave a fragile developmental web from supposedly separate sectarian and popular strands, it is better to look at the texts themselves to see what they attempted and achieved. It should be noted, however, that to the best of existing knowledge it was achieved relatively early in the period of consolidation, for the Bhagavadgītā the text that seals the achievementseems to be from no later than the first or second century bce (it is cited by Bārdarāyaa in the Vedānta Sūtra ), and possibly earlier. Of course, continued unfolding occurred after that.

The achievement itself is a universal Hinduism that, following Biardeau's discussion of bhakti in "Études de mythologie hindoue" (1976), one may designate as smārta. It inherits from the Brahmanic sacrificial tradition a conception wherein Viu and Śiva are recognized as complementary in their functions but ontologically identical. The fundamental texts of this devotional smārta vision are the two epicsthe Mahābhārata (c. 500 bce400 ce) and the Rāmāyaa (c. 400200 bce)and the Harivaśa (c. 300400 ce?). These works integrate much Puranic mythic and cosmological material, which later is spun out at greater length in the classical Purāas ("ancient lore"), of which there are said to be eighteen major and eighteen minor texts. The epics and Purāas are thus necessarily discussed together. But it should be recognized that whereas the smārta vision of the epics and the Harivaśa is fundamentally integrative and universal in intent, the Purāas are frequently dominated by regional and particularistic interests, including in some cases the strong advocacy of the worship of one deity (Śiva, Viu, or the Goddess) over all others. It is thus tempting to think of the period of Purāa composition (c. 4001200 ce?) as one that extends the integrative vision of the fundamental texts but develops it in varied directions. Still, as it is not clear that instances of Puranic theological favoritism are motivated by distinct sects, it is misleading to speak of "sectarian" Purāas.

Taken together, then, the Harivaśa and the Mahābhārata (which includes the Bhagavadgītā ) present the full biography of Ka, and the Rāmāyaa that of Rāma. The Harivaśa (Genealogy of Harii. e., Ka), the more recent of the texts concerning Ka, presents the stories of his birth and youth, in which he and his brother Balarāma take on the "disguise" (vea ) of cowherds. Thus they engage in divine "sport" (līlā ) with the cowherd women (gopī s), until finally they are drawn away to avenge themselves against their demonic uncle Kasa, who had caused their exile. The Mahābhārata (Story of the Great Bhārata Dynasty) focuses on Ka's assistance to the five Pāava brothers in their conflicts with their cousins, the hundred Kauravas, over the "central kingdom" of the lunar dynasty (the Bhārata dynasty) at Hāstinapura and Indraprastha near modern Delhi. Both texts incorporate telling allusions to the other "cycle," and because both stories must have circulated orally together before reaching their present literary forms, any notions of their separate origins are purely conjectural. The Rāmāyaa (Exploits of Rāma) tells the story of Rāma, scion of the solar dynasty and embodiment of dharma, who must rescue his wife Śita from the demon (rākasa ) Rāvaa. Though each of these texts has its special flavor and distinctive background, they become in their completed forms effectively a complementary triad. Indeed, in the "conservative" South, popular performances of Hindu mythology in dramas and temple recitations are still dominated by three corresponding specializations: Mahābhārata, Rāmāyaa, and Bhāgavata Purāa, the latter (c. 800900 ce?) enriching the devotional themes of the Harivaśa in its tenth and eleventh books and in effect replacing it as representing the early life of Ka.

The smārta universe in these texts is structured around Viu, and more particularly around his two heroic incarnations, Rāma and Ka. Thus other deities are frequently represented as subordinated to or subsumed by these figures. But there is also recognition of Viu's complementarity with Śiva: some passages that stress mutual acknowledgment of their ontological unity, others that work out the interplay between them through stories about heroic characters who incarnate them, and scenes in which Viu's incarnations do homage to Śiva. It should be clear that efforts to find "tendencies toward monotheism" in such texts involve the reduction of a very complex theology to distinctly Western terms. The same applies to those Purāas that are structured around Śiva or the Goddess rather than Viu but are still framed within the same cosmology and the same principles of theological complementarity and subordination.

This smārta vision is not, however, limited to one theological conundrum, for it extends to encompass Śiva and Viu's interaction with other major figures: the god Brahmā, masculine form of the impersonal Absolute (brahman ), now subordinated to the higher "personal" deities; the Goddess in her many forms; Indra and other deva s (now "demigods"); their still perennial foes, the demons (asura s); and of course humans, animals, and so on. It also presents an overarching bhakti cosmology in which the yogic supreme divinity (Śiva or Viu) encompasses the religious values of sanyāsa, tapas, knowledge, and sacrifice, and introduces the view that taken by themselves, without bhakti, these values may be incomplete or even extreme "paths." Further, it incorporates the smārta social theory of the Dharmasūtras and the Dharmaśāstras, and works out its implications within the cosmology. The details of this smārta vision are best discussed, however, in relation to the Hindu chronometric theory that is presumed and first articulated in these texts and then further developed in the Purāas.

Time is structured according to three main rhythms, hierarchically defined, the longer encompassing the lesser. Most down-to-earth is the series of four yuga s named after four dice throws, which define a theory of the "decline of the dharma ": first a ktayuga ("perfect age"), then a tretāyuga and a dvāparayuga, and finally a degenerate kaliyuga ("age of discord"). A ktayuga lasts 4,000 years, a tretāyuga 3,000, a dvāparayuga 2,000, and a kaliyuga 1,000, each supplemented by a dawn and twilight of one-tenth its total. A full four-yuga cycle thus lasts 12,000 years and is called a mahāyuga ("great yuga "). These are not human years, however, but divine years, which are 360 times as long as human years. Thus a mahāyuga equals 360 times 12,000, or 4,320,000 human years, and a kaliyuga is one-tenth of that total. A thousand mahāyuga s (4,320 million human years) is a kalpa, the second major time unit, which is also called a "day of Brahmā." Brahmā's days are followed by nights of equal duration. Brahmā lives a hundred years of 360 such days and nights, or 311,040 billion human years, all of which are sometimes said to pass in a wink of the eye of Viu. The period of a life of Brahmā, called a mahākalpa, is the third major temporal rhythm.

Working backward now, one may observe the modus operandi of Viu and Śiva (and of course others) as it is envisioned in the smārta Hinduism of the texts.

First, at the highest level, Viu and Śiva are great yogins, interacting with the rhythms of the universe in terms of their own oscillations between activity and yogic concentration (samādhi ). At the mahāpralaya ("great dissolution"), the deity (usually Viu in these early texts, but just as often Śiva or the Goddess in later Puranic ones) oversees the dissolution of the universe into the primal prakti in accord with the cosmological theory of Sākhya-Yoga. This ends the life of Brahmā, but it is also to be noted that it marks the restoration to its primordial unity of prakti, whichas feminineis regarded mythologically as the ultimate form of the Goddess. From a Śaiva standpoint, the male (the deity as Purua) and the female (the Goddess as Prakti) are reunited at the great dissolution of the universe, a theme that is depicted in representations of the deity as Ardhānarīśvara, "the Lord who is half female." Their union is nonprocreative and represents the unitive experience of the bliss of brahman. Creation then occurs when the deity (whether Śiva or Viu) emerges from this samādhi and instigates the renewed active unfolding of prakti.

The coincidence of the death of Brahmā with not only the dissolution of the universe but the reintegration of the Goddess and her reunion with Śiva is highly significant. The Goddess is an eternal being, worthy of worship becauselike Viu and Śivashe outlasts the universe and can bestow moka. Brahmā, ultimately mortal and bound to temporality, is worshiped not for moka but ratherand mostly by demonsfor earthly power and lordship. Stories that portray Śiva's severing of Brahmā's fifth head and refer to the "head of Brahmā" (brahmaśiras ) as the weapon of doomsday, are perhaps mythic echoes of this ultimate cosmological situation wherein the coming together of Purua and Prakti coincide with his death.

The primary creation has as its result the constitution of a "cosmic egg," the brahmāa ("egg of Brahmā"). Further creation, and periodic re-creations, will be carried out by Brahmā, the personalized form of the Absolute (brahman ). Insofar as the brahman is personalized and oriented toward the world, it is thus subordinated to the yogin Purua, the ultimate as defined through bhakti. Moreover, the activity of Brahmāheir in his cosmogenic role of the earlier Prajāpatiis conceived in terms of sacrificial themes that are further encompassed by bhakti.

It is at this level that the three male gods cooperate as the trimūrti, the "three forms" of the Absolute: Brahmā the creator, Śiva the destroyer, and Viu the preserver. Within the brahmāa, Brahmā thus creates the Vedic triple world of earth, atmosphere, and heaven (or alternatively heaven, earth, and underworld). These three samsaric worlds are surrounded by four ulterior worlds, still within the brahmāa, for beings who achieve release from saśara but still must await their ultimate liberation. These ulterior worlds are not henceforth created or destroyed in the occasional creations or destructions. As to the triple world, Brahmā creates it by becoming the sacrificial boar (yajñavarāha ) who retrieves the Vedas and the earth from the cosmic ocean. The destruction of the triple world is achieved by Śiva. As the "fire of the end of time," he reduces it to ashes, thus effecting a cosmic funerary sacrifice. And Viu, the god whom the Brāhamaas identify as "the sacrifice," maintains the triple world while it is sustained by sacrifices, and also preserves what is left of it after the dissolution when he lies on the serpent Śea ("remainder") whose name indicates that he is formed of the remnant of the previous cosmos, or more exactly of the "remainder" of the cosmic sacrifice. This form of Viu, sleeping on Śea, is called Nārāyaa, a name that the Śatapatha Brāhamaa already connects with the Vedic Purua, the "male" source of all beings. When Viu-Nārāyaa awakens, Brahmāwho in some fashion awakens with himre-creates the universe. Through all these myths the earth is a form of the Goddess, indeed the most concretized form she takes as a result of the evolution of prakti (earth being the last of the evolutes emitted and the first to dissolve).

Thus the greater universe whose rhythms are integrated within the divine yoga of Viu and Śiva encompasses an egg of Brahmā, which encloses a triple world whose rhythms form a round sustained by the divine sacrificial acts of the trimūrti. This pattern is transposed onto the third temporal rhythm, that of the yuga s. Thus the characteristic religious virtues of the yuga s are as follows: dhyāna ("meditation") or tapas ("asceticism") in the ktayuga ; jñāna ("knowledge") in the tretāyuga ; yajña ("sacrifice") in the dvāparayuga ; and dāna ("the gift") in the kaliyuga. Thus the two śruti -based ideals of knowledge and sacrifice are enclosed within a frame-work that begins with yogic meditation as a divine ktayuga activity and ends in the kaliyuga with the devotional gift. Bhakti thus encompasses knowledge and sacrifice.

The distinctive feature of the rhythm of the yuga cycle is that it is calibrated by the rise and fall of dharma in the triple world. Beings who have achieved release from the triple world oscillate between the four higher worlds, enduring periodic destructions of the triple world and awaiting the great dissolution of the universe that will dissolve the egg of Brahmā (coincident with his death) and result in a vast collective ultimate liberation of reabsorption into the supreme Purua. Needless to say, this is to occur only after an almost incalculable wait. But beings who have attained these ulterior worlds are no more affected by dharma than the yogic deity beyond them. The maintenance of dharma within the triple world thus engages the deities in their third level of activity, that of "descent." In classical terms this is the theory of the avatāra. Though the term is not used in the epics or the Harivaśa in its later, specialized sense, these texts are suffused by the concept and its bhakti implications, which include narrative situations wherein the divinity looks to all concerned, and sometimes even to himself, as a mere human. The programmatic statement of the avatāra concept (without mention of the term itself) is thus stated by Ka in the Bhagavadgītā: "For whenever the Law [dharma ] languishes, Bhārata, and lawlessness flourishes I create myself. I take on existence from eon to eon [yuga to yuga ], for the rescue of the good and the destruction of evil, in order to establish the Law" (4.78; van Buitenen, trans.).

The classical theory of the ten avatāra smost of whom are mentioned in the epics and the Harivaśa, but not in a single listis worked out in relation to Viu. One thus has the following "descents" of Viu in order of appearance: Fish (Matsya), Tortoise (Kūrma), Boar (Varāha), Man-Lion (Narasiha), Dwarf (Vāmana), Rāma with the Ax (Paraśurāma), Rāma of the Rāmāyaa, Ka, the Buddha, and the future avatāra Kalki, who will rid the earth of barbarian kings and reestablish the dharma at the end of the kaliyuga. There are various attempts to correlate appearances of the avatāra s with distinct yuga s and even kalpa s, but the one feature that is consistently mentioned in these formative texts is that Ka appeared at the interval between the last dvaparayuga and kaliyuga, and thus at the beginning of the present age. It is likely that the theory was first formulated around Ka and Rāma along with the Dwarf (the only form to be associated with Viu in the śruti literature) and the apocalyptic Kalki. But in actuality, the avatāra theory is more complex. In the epics and in living Hinduism, Viu does not descend alone. In the literature, his incarnations take place alongside those of other deities, including most centrally Vāyu, Indra, Sūrya, the Goddess, andat least in the Mahābhārata Śiva. And in localized temple mythologies throughout India, one hears of avatāra s of Śiva and the Goddess as well as of Viu. In devotional terms, the avatāra is thus a form taken on earth (or, better, in the three worlds) by any one of the three deities found at the ultimate level of cosmic absorption, where all that remains beside the liberated beings who join them are the eternal yogic deities Viu and Śiva and the primal Goddess.

The classical concept of the avatāra, structured around Viu, remains, however, the chief Hindu use of the term. Its formulation in the epics and the Harivaśa is thus constitutive for succeeding eras of Hinduism, in which it will only be enriched but not essentially changed by later bhakti theologies. Looking at these texts comprehensively, then, with the Gitā as the main guide, one can outline its main contours. Against the background of the vast, all-embracing bhakti cosmology, the involvement of the yogic divinity on earth takes place completely freely, as "sport" or "play" (līlā ). Still, the god takes birth to uphold the dharma and to keep the earth from being unseasonably inundated in the waters of dissolution under the weight of adharmic kings. The avatāra thus intercedes to uphold the system of varāśramadharma and to promote the proper pursuit of the four puruārtha s. Because he appears in times of crisis, a central concern in the texts is with the resolution of the conflicts between ideals: renunciation versus householdership, brāhamaa versus katriya, killing versus "not desiring to kill" (ahisā ), dharma versus moka, dharma versus kāma and artha, and conflicts between different dharma s (duties) such as royal duty and filial duty. But though the texts focus primarily on the two upper castes, the full society is represented by singular depictions of figures who evoke the lowest castes and tribal groups. It is also filled in with figures of real and reputed mixed caste.

Confusion of caste is a particularly prominent issue in the Mahābhārata, where it is raised by Ka in the Gitā as the worst of ills. Most significantly, the Mahābhārata and the Harivaśa identify a particularly pernicious form of caste confusion among the barbarian (mleccha ) peoples of the Northwest (the Punjab), mentioning Yavanas, Śakas, and Pahlavas among others as enemies of the dharma and causes for such "mixing." The fact that events of the period from 300 bce to 300 ce are projected into the distant past indicates that part of the bhakti synthesis was the articulation of a mythical theory of historical events. One may thus look at these smti texts as posing a model for the revival of Hinduism in accord with "eternal" Vedic models, with the descent of the avatāra and indeed of much of the Vedic pantheon along with himguaranteeing the periodic adjustment of the sociocosmic world to these eternal norms. Furthermore, the tracing of all Hindu dynastic lines back to the defunct if not mythical "lunar" and "solar" dynasties provided the model for the spatial extension of this ideal beyond the central lands of Aryavata where the dharma, according to both Manu and the Mahābhārata, was the purest.

But the focus of the avatāra is not solely on the renovation of the dharma. He also brings to the triple world the divine grace that makes possible the presence, imagery, and teachings that confer moka. The epics and the Harivaśa are full of bhakti tableaux: moments that crystallize the realization by one character or another of the liberating vision (darśana ) of the divine. Most central, however, is the Bhagavadgītā, which is both a darśana and a teaching.

The Bhagavadgītā (Song of the Lord) takes place as a dialogue between Ka and Arjuna just before the outbreak of the Mahābhārata war. Although he is the third oldest of the five Pāavas, Arjuna is their greatest warrior, and Ka's task in the Gitā is to persuade him to overcome his reluctance to fight in the battle. Fundamental to the argument is Arjuna's requirement to fulfill his dharma as a katriya rather than adopt the idealunsuitable for him in his present life stageof the renouncer. Thus the Gitā champions the theory of varāśramadharma as upholding the sociocosmic order.

Ka presents his teaching to Arjuna by revealing a sequence of "royal" and "divine" mysteries that culminate in his granting a vision of his "All-Form" (Viśvarūpa-darśana ) as God, creator and destroyer of the universe. In this grand cosmic perspective, Arjuna is told that he will be but the "mere instrument" of the deaths of his foes, their destruction having now come to ripeness through Viu's own agency in his form as cosmic time, or kāla (Bhagavadgītā 11.3233). Arjuna thus recognizes this omniform deity as Viu in this climactic scene.

On the way to this revelation, however, Ka acknowledges the three paths (yoga s) to salvation: action, knowledge, and devotion. These are presented as instructions by which Arjuna can gain the resolute clarity of insight (buddhi ) and yogic discipline by which to recognize the distinctions between soul and body, action and inaction, and thus perform actionsincluding killingthat are unaffected by desire. Ritual action and knowledge are set forth as legitimate and mutually reinforcing paths, but incomplete unless integrated within and subordinated to bhakti. Ka thus presents himself as the ultimate karmayogin, acting to benefit the worlds out of no personal desire. He thus bids his devotees (bhakta s) to surrender all actions to him as in a sacrifice, but a sacrifice (karman ) no longer defined in Vedic-Mīmāsā terms as a means to fulfill some personal desire. Ka also presents himself as the object of all religious knowledge, the highest Purua (uttamapurua ) and supreme self (paramātman ), beyond the perishable and the imperishable, yet pervading and supporting all worlds (15.1617).

One other facet of the bhakti synthesis to which the Gitā alludes is the transition from traditional Vedic sacrifice (yajña ) to new forms of offering to the deity (pūjā, literally, "honoring"). This corresponds to the theory that the "gift" is the particularly appropriate religious practice for the kaliyuga. Thus Ka says: "If one disciplined soul proffers to me with love [bhakti ] a leaf, a flower, fruit, or water, I accept this offering of love from him. Whatever you do, or eat, or offer, or give, or mortify, make it an offering to me, and I shall undo the bonds of karman " (9.2627; van Buitenen, trans.). The passage probably refers to domestic worship of the "deity of one's choice" (iadevata ). But it is also likely to allude to temple worship, for it is known from inscriptions and literary sources from the third to first century bce that sanctuaries existed for Vāsudeva and Keśava (presumably as names for Ka and Viu), as well as for other deities. By the beginning of the Gupta period, around 320 ce, temple building was in full swing, with inscriptions showing construction of temples for Viu, Śiva, and the Goddess. Temples were built at sites within cities, as well as at remote holy places, and sanctuaries at both such locations became objectives along pilgrimage routes that are first mentioned in the Mahābhārata. From very early if not from the beginning of such temple worship, the deities were represented by symbols and/or iconic images.

Certain aspects of temple construction and worship draw inspiration from the Vedic sacrifice. The plan of the edifice is designed on the ground as the Vastu-puruamaala, a geometric figure of the "Purua of the Site" (vāstu ), from whom the universe takes form. The donor, ideally a king, is the yajamāna. The sanctum sanctorum, called the garbhagha ("womb house"), continues the symbolism of the Vedic dīkā hut: Here again the yajamāna becomes an embryo so as to achieve a new birth, now taking into his own being the higher self of the deity that he installs there in the form of an image. The temple as a whole is thus a Vedic altar comprising the triple world, but also an expanded image of the cosmos through which the deity manifests himself from within, radiating energy to the outer walls where his (or her) activities and interactions with the world are represented.

But the use of the temple for ordinary daily worship involves radically non-Vedic objectives. The Vedic sacrifice is a means for gods and humansbasically equalsto fulfill reciprocal desires. Pūjā rites are means for God and humankind to interact on a level beyond desire: for humans to give without expectation of reward, or, more exactly, to get back nothing tangible other than what they have offered but with the paradoxical conviction that the deity "shares" (from the root meaning of bhakti ) what is given and returns it as an embodiment of his or her grace (prasāda ). God is thus fully superior, served as a royal guest with rites of hospitality. Basically four moments are involved: offerings, taking sight (darśana ) of the deity, receiving this prasāda, and leave-taking by circumambulation of the garbhagha and the image within. The offerings are the pūjā proper and comprise a great variety of devotional acts designed to please the deity, some of which may be worked into a daily round by the temple priests, who offer on behalf of others.

Finally, one last element of the consolidation of Hinduism achieved by early Gupta times is the emergence of the Goddess as a figure whose worship is recognized alongside that of Viu and Śiva and is performed with the same basic rites. Indeed, it is possible that aspects of pūjā ceremonialism are derived from non-Vedic śūdra and village rites in which female deities no doubt figured highly, as they do in such cults today. The two epics, the Mahābhārata and the Rāmāyaa, reflect themes associated with the Goddess in the portrayals of their chief heroines, Draupadī and Śita, but the Harivaśa is probably the first text to acknowledge the Goddess as such. There she takes birth as Ka and Balarāma's "sister" (actually she and Ka exchange mothers). Some of her future demon enemies are mentioned, and there is also reference to her having numerous places of worship and a cult that apparently included animal sacrifice. Thus the Goddess is integrated even within the texts of the early smārta Hinduism that are centered on Viu. But the text that registers her full emergence is the Devimāhātmyam (Glorification of the Goddess). Probably from about 400600 ce, it was included in the Mārkaeya Purāa. Here the Goddess is recognized under all her major aspects, as primal matter embodied in the universe yet beyond it, incarnate in many forms, cause of the joys and miseries of this world and of liberation from it, the power (śakti ) enabling the roles of the trimūrti, yet higher than the gods and their last resort in the face of certain demons, most notably the buffalo demon Mashāsura, her most dedicated and persistent foe through cults and myths both ancient and current. This emergence of the Goddess is registered more fully in the development of Tantric Hinduism.

Tantric Hinduism

Tantra is literally "what extends." In its Hindu form it may be taken, according to its name, as a movement that sought to extend the Veda (whose pedigree it loosely claimed) and more particularly to extend the universalistic implications of bhakti Hinduism. However, although it was quick to integrate bhakti elements and to influence bhakti in nearly all its forms (late Puranic, popular, and sectarian), its earliest and most enduring forms "extend" Hinduism in ways that were directly opposed to the epic-Puranic bhakti synthesis. Nonetheless, it is still formulated within the same cosmology.

Early Tantrism developed most vigorously, from the fourth to sixth centuries ce, in areas where Brahmanic penetration had been weakest: in the Northwest, in Bengal and Assam in the East, and in the Andhra area of the South. These are areas where one must assume non-Aryan influences in general, and more particularly probably also tribal and folk practices involving shamanism, witchcraft, and sorcery, and, at least in the East and South, a cult of the Goddess. As Tantrism gained currency in succeeding centuries throughout India, the shamanistic and magical features were assimilated to yogic disciplines, while the elevation of the Goddess gave full projection on a pan-Indian scale to roles and images of the Goddess that had been incorporated, but allowed only minimal scope, in the early orthodox bhakti and even earlier Vedic sacrificial traditions. The earliest extant Tantric texts are Buddhist, from about the fourth to sixth centuries. Hindu Tantric texts include Vaiava Sahitās, Śaivāgamas from a slightly later period, and Śākta Tantras (exalting the Goddess as Śakti, or Power) from perhaps the eleventh century on. But from its start Tantrism represented a style and outlook that placed the Goddess at the center of its "extensions" and to a certain extent cut across sectarian and religious distinctions, whether Hindu, Buddhist, or even Jain.

Though Hindu Tantra thus asserts its Vedic legitimacy, its stance is intentionally anti-Brahmanic. It was especially critical of Brahmanic concepts of hierarchy, purity, and sexual status, all of which had been reinforced by the orthodox bhakti synthesis and which were in particular bound up with a theology that viewed the supreme divinity as a male (a Purua, whether Śiva or Viu) whose ultimate form was accessible only beyond the rhythms of the cosmos and its hierarchy of impure and pure, gross and subtle worlds. For Tantrics, dualities were artificial and their experience was the result of delusion. On the analogy of the union between Śiva and Śakti, which in Puranic devotional terms is conceivable only at the end of the mahāpralaya, or great dissolution of the universe, Tantric practice (sādhana ) addresses itself to experiencing the unity of purua and prakti (purua being both "soul" and deity, prakti being both "matter" and Goddess), male and female, pure and impure, knowledge and action, and so on. Most important, all this takes place here and now, not only in this world, where prakti and purua on the macrocosmic scale are one, but in the human body, where their microcosmic embodiments can be experienced. The body thus becomes the ultimate vehicle for liberation, the dissolution of opposites taking place within the psychophysical continuum of the experience of the living adept, who realizes beyond duality the oneness of brahman.

In terms of practice, Tantra's rejection of Hindu orthopraxy is even more decisive. And practice is clearly exalted above theological or philosophical formulation. Two types of Tantra are mentioned: "left-hand" and "right-hand." The Tantric rejection and indeed inveion of orthopraxy is most pronounced in the former, as the right-hand Tantra interprets the most anti-Brahmanic practices of the left metaphorically, and also includes under its heading a wide variety of ceremonial rituals assimilated into bhakti Hinduism that are simply non-Vedic. These include the use of non-Vedic mantra s as well as yantra s and ma ala s, aniconic and non-Vedic geometric devices used for visualization and integration of divine-cosmic forces. Adepts come from all castes, but low-caste and even tribal practitioners and teachers are especially revered. The goal of liberation within the body takes the specific form of seeking magical powers (siddhī s), which in orthodox forms of Hinduism are regarded as hindrances to spiritual achievement. Under the tutelage of a guru, who embodies the fulfillment sought and its transmission and who is thus all-important, the siddhī s are sought through yoga disciplines that show the impact of Tantra through their anatomical analysis of the "subtle body" (liga śarīra ). First practiced is hahayoga, the "yoga of exertion or violence," that is, rigorous physical discipline geared to coordinating the body's "ducts" or "channels" (nāī s) and "energy centers" (cakra s). This is followed by kualiniyoga, which awakens the dormant śakti, conceived as a coiled-up "serpent power" in the lowest cakra between the genitals and the anus, so that it (or she) can pierce and transform all the cakra s (usually six) and unite with Śiva in the "thousand-petaled cakra " in the region of the brain.

Beyond these practices, "left-handed" Tantrics pursue in literal fashion the ceremonial of the "five m 's" (pañcamākarapūjā ). That is, they incorporate into their cultic practice five "sacraments" beginning with the syllable ma: fish (matsya ), meat (māsa ), parched grain (mudrā, regarded as an aphrodisiac), wine (madya ), and finally sexual intercourse (maithuna ). It is likely that most if not all of these practices involve the incorporation of elements of the cult and mythology of the Goddess, who already in the Devimāhātmyam delights in meat and wine and is approached by lustful demons for sexual intercourse. Tantric texts stress that these practices are to be carried out within a circle of adepts and supervised by a male and female pair of "lords of the circle" who insist on strict ritual conventions that guard against an orgiastic interpretation. Classically, the male is to retain his semen at the point of orgasm, this being a sign not only of profound dispassion but an actualization of the nonprocreative union of Śiva and Śakti at the dissolution of the universe of dualities.

It is interesting to note that, although their historical validity is debated by scholars, there are strong Indian traditions suggesting that Śakara's philosophical nondualism had practical Tantric repercussions.

Śakara's Advaita VedĀnta and SmĀrta Orthodoxy

The Advaita (nondualist) interpretation of the Vedānta can be traced back at least to Gauapāda (c. 600 ce), but it is Śakara (c. 788820) who established this viewpoint as the touchstone of a revived smārta orthodoxy. Born in a small Kerala village, Śakara spent his alleged thirty-two years as a vigorous champion of the unity of Hinduism over and against intra-Hindu divisions and the inroads of Buddhism and Jainism. He toured India, setting up monasteries (maha s) near famous temples or holy places at each of the four compass directions, and appointed a disciple at each center to begin a line of renunciant "pontiffs." And he wrote works of great subtlety and persuasiveness, including commentaries on the Upaniads, the Brahma Sūtra, and the Bhagavadgītā that inspired contemporaries, disciples, and authors of later generations to write additional important works from the perspective that he developed.

An essential feature of Śakara's argumentation is that lower views of reality must be rejected as they are contradicted or "sublated" by higher experiences of the real. Finally, all dichotomous formulations must be abandoned upon the nondual experience of the self (ātman ) as brahman. The world of appearance is sustained by ignorance (avidyā ), which "superimposes" limitations on reality. Māyā ("illusion" or "fabrication"), itself neither real nor unreal, is indescribable in terms of being or nonbeing. It appears real only so long as brahman is not experienced. But it is empirically real relative to things that can be shown false from the standpoint of empirical observation. Māyā is thus said to be more mysterious and unknowable than brahman, which is experienced as being, consciousness, and bliss (sat-cit-ānanda ).

As philosophy, Advaita is thus a guide to moka, which is experienced when the ignorance that results from superimposing māyā on brahman is overcome. Liberation arises with knowledge (jñāna ), but from a perspective that recognizes relative truth in the paths of both action and bhakti. Practically, Śakara fostered a rapprochement between Advaita and smārta orthodoxy, which by his time had not only continued to defend the varāśramadharma theory as defining the path of karman, but had developed the practice of pañcāyatanapūjā ("five-shrine worship") as a solution to varied and conflicting devotional practices. Thus one could worship any one of five deities (Viu, Śiva, Durgā, Sūrya, Gaea) as one's iadevatā ("deity of choice"). As far as varāśramadharma was concerned, Śakara left householder issues largely aside and focused instead on founding ten orders of sanyāsi s (the daśanāmi, "ten names"), each affiliated with one of the four principle matha s he founded. But traditional orthodox views of caste were maintained. According to Śakara, as śūdra s are not entitled to hear the Veda, they cannot pursue knowledge of brahman as sanyāsi s; rather they may seek moka through hearing the Mahābhārata and the Purāas. Four of the ten sanyāsi orders were thus restricted to brāhamaas, and it does not seem that any accepted śūdra s until long after Śakara's death. Bhakti sectarian reformers were generally more liberal on this point. As to the god (or gods) of bhakti, Śakara views the deity (Īśvara) as essentially identical with brahman and real relative to empirical experience. But by being identified "with qualities" (sagua ), God can be no more than an approach to the experience of brahman "without qualities" (nirgua ). Viewed from the experience of the self as nirgua brahman, which "sublates" all other experiences, the deity is but the highest form of māyā. Clearly, bhakti traditions could not rest with this solution. But it should be noted that in opposing Śakara and abandoning the universalist vision of the epic-Puranic devotional synthesis, the sects turned their backs on the main impulses that had attempted to sustain the unity of Hinduism.

Sectarian Hinduism

The elaboration of bhakti Hinduism continued to unfold in the later Purāas, linking up with the temple and pilgrimage cultus and with local and regional forms of worship. It thus established itself until the time of Śakara as the main expression of Brahmanic orthodoxy and the main shaping force of popular Hinduism. But though it proclaimed a universal Hinduism, it gave little weight to the problem of the immediate accessibility of salvation. While caste hierarchy was to remain in effect on earth to assure, among other things, the pure temple worship of the gods by the brāhamaas, the ultimate release that the Purāas promised was almost infinitely postponed. It is possible that their postponement of a collective liberation was a kind of purification process for liberated souls and thus a prolongation of the concern for brāhamaa purity on earth. In any case, the remoteness of salvation and the defense of caste purity and hierarchy in the Puranic devotionalism of Brahmanic orthodoxy were probably incentives for the development of alternate forms of bhakti. These emerged in sectarian traditions, in movements led by saint-singers who inspired vernacular forms of bhakti revivalism, and more generally in local and regional forms of Hinduism.

Sectarian traditions

Sectarianism and bhakti revivalism are movements of separate origins that converge for the first time in the eleventh and twelfth centuries in the Tamil-speaking area of South India. There the fusion was accomplished in the traditions of the Śrī Vaiavas and the Śaiva Siddhānta, sects whose names indicate their distinctive theological preferences for Viu and Śiva. Henceforth, sectarianism and bhakti revivalism continued to interact and produce hybrid forms as they spread over all of India.

Generally speaking, sects followed a reformist impulse, and in most of them one can identify the emergence of the guru as a new type of figure: not the transmitter of an "impersonal" Vedic teaching, but one who takes inspiration from the personal deity of the sect, with whom he may even be identified. Traditional hierarchy was generally respected, but with the proviso that within the sect divine grace was not limited by caste boundaries. Nonetheless, as groups formed around masters and their teachings, they took on many of the characteristics and functions of castes (endogamy, interior ranking), and certain sects formulated their stands with particularly positive attitudes (the northern school of Śrī Vaiavas) or negative attitudes (Ligāyats and Vīraśaivas) toward brāhamaa s. Sects distinguish themselves over and against each other by many means, and often quite passionately: by bodily markings, forms of yoga discipline, worship, theology, and in particular by their choice of supreme deity, whether Śiva, Viu, Śakti, or, in the North, Ka or Rāma. Nonetheless, they generally participate in wider Hindu activities such as pilgrimage, festival, and temple worship (the Ligāyats are an exception) and draw upon fundamental Hindu belief structures. Thus most sects acknowledge other deities as subordinate to the supreme deity of the sect. In particular, most have worked out ways of encompassing the relation of the God and the Goddess at some fundamental theological level. Persistently the supreme deity is identified both as the ultimate brahman a nd also as in some way personal. The sects also frequently define various stages of divine descent or interaction with the world, various stages of the soul's ascent, and various types of relation between the soul and God. Thus the sects elaborate upon the epic-Puranic cosmology while modifying and refining the theological and soteriological terms. It is only against this background that their formulations are intelligible.

From the historical vantage point, one may note that the consolidation of the separate strands of sectarianism and bhakti revivalism occurs after, and is no doubt in part a response to, the growing success of Śakara's Advaita Vedānta. Prior to Śakara, sectarian groups had centered primarily around distinctive ritual traditions that were increasingly influenced by Tantrism: not only in forms of worship and theological formulation, but also, in some Śaiva sects, in actual practice. Thus the Vaiava Pañcarātras and Vaikhānasas and the Śaiva Pāśupatas (all mentioned first in the late Mahābhārata ) between the fifth and tenth centuries produced their Sahitās and gamas to regularize the construction of temples, iconography, and pūjā ceremonialism. Some Pāśupatas and Kāpālikas (a Tantric Śaiva sect) also incorporated forms of abrupt anticonventional behavior modeled on Śiva's character as the great yogin ascetic. With the exception of the Pañcarātras, who elaborated an influential doctrine of the emanations (vyūha s) of Viu that paralleled the cosmogonic theory of evolution in the Sākhya system, the theological formulations of these movements were apparently among their secondary concerns.

Saint-singer tradition

Whereas the early sectarian movements were able to spread their impact from north to south using Sanskrit as their medium, the bhakti revivalist movement began in the South, drawing on Tamil. Like the sectarian movements, the saint-singers developed their traditions along Vaiava and Śaiva lines. The sixty-three Nāyamār (or Nāyaārs) promoted the worship of Śiva, while the twelve vārs similarly honored Viu. Part of the revivalist motivation was provided by the earlier spread of Buddhism and Jainism in the South, both of which lost considerable following as a result of the efforts of the Nāyamār and vārs, as well as those of their contemporary Śakara.

Some of the most renowned among these two companies of saint-singers have left songs that they composed at the temples of Viu and Śiva, praising the form and presence of the deity therein, the place itself as his manifestation, and the communal attitude of worship generated there through pilgrimage and festival. Though they honor the deities in terms familiar from Puranic myths, the stories are set in the local terrain. The emotional side of bhakti thus draws from deep Tamil traditions, including a revival of classical Tamil poetic conventions involving the correlations between different types of landscape, different divinities, and different types of male-female love. In the hands of the saint-singers, erotic love in particular was drawn on as a metaphor for devotional feelings that stressed the feminine character of the soul in relation to the deity and idealized a softening of the mind or heart that could take the forms of "melting" into the divine, ecstatic rapture, divine madness, and possession.

Following the advent of Śakara, most of the sectarian and revivalist movements found common cause in their devotionalist stance against Advaita nondualism and continued to develop for the most part interdependently. Thus, most formatively, the songs of the vārs were collected in the ninth century for eventual use by the Śrī Vaiavas. And the poems of the Nāyamārsupplemented by the songs of Māikkavācakar, who apparently lived just after the list of sixty-three Nāyamār had been set (ninth century)were collected to form parts of the canon of the Śaiva Siddhānta. However, the revivalist and sectarian strains could also at times follow somewhat independent courses. The saint-singer tradition continued to take Śaiva and Vaiava forms among the Ligāyats and the Haridāsas of Karnataka, and also to be associated there with sects (the Ligāyats themselves and the Brāhma Sapradāya or Dvaita Vedānta tradition of Madhva, respectively). But its spread through Maharashtra, the Hindi-speaking areas of North India, and through Bengal was most focused on Viu, or more accurately on his forms as Rāma and Ka, who in turn, in the Hindi and Bengali areas, became the deities of different sects. In the case of Ka, erotic devotional poetry opened new dimensions on the theme of Ka's love-play with his "new" consort, Rādhā (her name does not appear before the twelfth-century Sanskrit Gitāgovinda by the Bengali court poet Jayadeva). In Hindi and Bengali poems, not only are the emotions of motherly love for the baby Ka and erotic love for the youthful Ka explored, but they are tied in with a classical theory of aesthetic appreciation (rasa ).

As to the sects, the impact of Śakara's Advaita is evident at many points. Although Śaiva monasticism may predate Śakara by about a century, his establishment of ma