the more “authentic”? Is the correction of grammatical forms legitimate, assum-
ing as it does a fixed grammar? And so on.
A text in South Asia does not, of course, necessarily mean a written text. It
might well mean an oral one, although an oral text, particularly the Veda, is not
necessarily less stable through time than a written text. This compounds the dif-
ficulty for the philologist, and the kind of work needed on oral traditions has
hardly started, although important work has been done (for example, Blackburn
1988 on Tamil sources; Smith 1991 on the epic of Pabuji; Staal 1961 on vedic
recitation). The principle assumed by philology is that we need to establish the
text, as Witzel says, to find out what a thinker such as S ́am.kara actually taught
(Witzel 1997: vi). If an apparent theological contradiction within a text is really
a contradiction, for example, we need to first establish the text and whether con-
tradictory elements are later additions or perhaps have entered from other texts
in the intertextuality of oral tradition. Let us look more closely at this problem,
taking our examples from the large group of medieval texts called Tantras.
The vast body of tantric material presents us with difficult questions. What is
the function of these texts? Who composed them and for whom? What are the
procedural difficulties of the outsider in approaching these texts? Is it possible to
establish an original text? And so on. Certainly the Tantras were regarded as
revelation and treated as words of authority, and certainly they developed in a
social context that fostered their dissemination (see Colas’s and Flood’s essays).
But the difficulties of establishing critical editions, along with reading and
making informed comment on this material, are considerable. The texts them-
selves often use forms of Sanskrit at variance with “correct” usage; a form of
language known as “divine” or “belonging to S ́iva” (ais ́a), which, Goodall notes,
seems to cause commentators some embarrassment (Goodall 1998: lxvi).
Reading tantric texts we need to be aware that the context of their reception
would have involved oral comment by a teacher, practice, especially supereroga-
tory forms of ritual beyond those required by vedic orthopraxy, as well intellec-
tual speculation about their meaning by commentators and their audiences. The
aim of the commentaries is to establish a text within a particular field of inter-
pretation and, presumably, to explicate meanings to an educated, brahmanical
audience for whom the text was alive and important. These commentaries are
aware of themselves as establishing a particular interpretation over and against
other possible readings. Thus Ks.emara ̄ja (ca. 1000–1050 ce) composed his com-
mentary on the Svacchanda Tantrato defeat adherence to the dualist interpreta-
tion of the rival tradition, the S ́aiva Siddha ̄nta (Kahrs 1998: 60). That these texts
require commentary is itself an indication of their openness and their non-
transparent nature. The “contextualizing practice” (Lemke 1995) of Ks.emara ̄ja
certainly relates these texts to the practices of his culture and the texts of his
tradition. Indeed, the commentators highlight the texts’ intertextuality through
quoting other scriptural authorities, sometimes of rival traditions.
Almost from their inception, then, it was not possible to establish originary
texts. Even by the time of the commentator on the text, the dualist theologian
introduction: establishing the boundaries 11