that tradition stems from textual, written revelation, and that a religion is nec-
essarily coherent.
But in spite of these criticisms there are nevertheless Hindu analogues to cat-
egories of revelation, tradition, theology, and practice, although these arguably
do not point to a unitary referent. We might say, then, that Hinduism contains
both uniting and dispersing tendencies that we might, borrowing from Bakhtin,
call centripetal and centrifugal forces. On the one hand, there is the Sanskritic
tradition of brahmanical orthodoxy, flowing from the ancient revelation of
the Veda, concerned with correct ritual procedures, the maintenance of caste
boundaries, and the interpretation of scripture. This is a decisive constraint on
the traditions that comprise Hinduism. On the other hand, there is a great
proliferation of decentered traditions, often founded by a charismatic teacher
or guru, and communities expressed in vernacular languages that cannot be
defined by a central, brahmanical tradition and which are often set against that
tradition. The teyyamtradition of Kerala or the Sant devotional tradition of
northern India would be examples here (see the essays by Freeman and Martin).
We can trace the history of the fairly recent term “Hindu” and “Hinduism”
from its initial coinage by those outside of the Hindu fold to its appropriation as
a term of self-description by “Hindus” themselves. Much before the nineteenth
century, people of South Asia did not consider themselves to belong to a wider,
united religious identity, but would rather be members of a tradition and com-
munity whose focus was a particular deity or practice. One would be a Vais.n.ava
whose focus is the deity Vis.n.u, a S ́aiva whose focus is S ́iva, a follower of partic-
ular Tantras (ta ̄ntrika) and so on (Oberhammer 1997: 19). But there were by the
medieval period if not earlier uniting features that cut across these diachronic
processes, such as pilgrimage to sacred centers, particularly great regional
temples, ritual offerings to deities in concrete form (mu ̄rti), devotion (bhakti), and
the practice of textual exegesis by scholars in centers of learning.
South Asian cultures are highly textualized in the widest sense of the term
with many oral traditions, some of which stretch back thousands of years. There
are traditions of vedic recitation in several regions of India that function, as
Michael Witzel says in this volume, as “three thousand year old tape recordings.”
This revelation of the Veda, verses believed to have been revealed to and heard
by (s ́ruti) the ancient sages (r.s.i), as symbol and legitimizing reference if not
actual text, is central as a constraining influence on later traditions, providing
the authority for tradition (Oberhammer 1997: 21–31). Some would argue that
this is a defining feature of Hinduism (Smith 1988: 40). As constraining force
the Veda has been used to legitimize different philosophical positions, as the basis
of Hindu law and power structures, and has provided a reference point against
which some traditions and charismatic teachers have reacted. Whether accepted
or rejected, whether traditions are indifferent to its injunctions, it is seldom
ignored as symbol. As Heesterman observes, the hold of “Vedism” on Indian
thought and imagination has persisted not withstanding the cult of the temple,
popular devotion, and tantric texts and practices (Heesterman 1993: 43). Given
this reference point, we might say that both centripetal and centrifugal tenden-
4 gavin flood