Thanks EB, that is a fascinating discussion of how the Suttas may have been constructed.
Although I have attempted to show that the early Buddhist sutta texts
were, in the words of R.Gombrich, "deliberate compositions which were
then committed to memory, I would certainly agree that accounts of
what the Buddha is supposed to have said and discourses on his teaching
would have been given by the monks and nuns after the Buddha's death in
an improvisatory manner, at times drawing heavily on memorised material,
or as R. Gethin (1992) has argued, by using lists as a foundation. Such
discourses may then have become the basis of later fixed texts. But these
accounts and discourses were fundamentally different from the essentially
fixed, memorised texts transmitted by the community, however imperfectly.
Finally, the Parry-Lord model does not exhaust the oral or literary/perfor-
mance dimension of oral cultures. In ancient, pre-literate India there was a
strong tradition of composing fixed, religious texts which were designed to
be memorised and transmitted verbatim.