The idea of authorship as understood here contrasts with conceiving of the texts mainly as attempts at recollection—Anālayo’s “textual memories”. Shulman & others’ point is that the texts do more than try to record history: they create new texts out of combining ‘approved’ templates, stock passages, etc., without the pretension that they refer or correspond to a historical teaching event. And yet, of course, they’re meant to convey the true teachings of the Buddha.
If I may, I think there are two misunderstandings here:
(1) Authorship does not refer to who is preaching according to the sutta, but to the person who crafted the sutta (in which X or Y is preaching this or that teaching). In Anālayo’s view there are no authors in that sense, only reciters/preservers/transmitters…
(2) As far as I understand, this is not what these scholars mean by versions of each other. What you write conforms to Anālayo’s perspective, which treats the teaching events as versions of the same teaching/doctrine, and parallel texts as different records of each of those events. Instead, these scholars see the parallel texts themselves as versions of each other. They call them literature not in the sense that they contain literary devices or are beautiful, but in the sense that they are, in the main, religious literature rather than history.
For example, in their view parallels of the Satipaṭṭhāna Sutta are not alternative descendants of an original memory or teaching occasion(s), which may or may not be reconstructable; but alternative formulations of a certain teaching—and they were considered valid because they conveyed that teaching well, not because the Buddha or a disciple of his actually taught it in that way, on a certain day & in a certain village. From Anālayo’s conception, it makes sense to compare those alternatives and conclude which one is closer to the ‘original’, which one is corrupt, etc. From that of Shulman, Gethin (and less clearly McGovern), it does not—Allon is a bit more in the middle.
The Satipaṭṭhāna Sutta is a good example because, while the reviewed works don’t mention it, Anālayo is famous for pointing out that under dhammānupassanā, the hindrances & awakening factors are common to all versions while other elements are not, and that this means something relevant. It might indicate that
- the original, even if un-reconstructable, is less likely to have included the elements not shared across all versions
- the more un-shared elements are in one version, the more removed that version is from the original, the more corrupt, the less reliable, etc.
The other scholarly approach does not support any of this. It’s quite different: differences among versions need not imply ‘errors’ or lack of accuracy.
You’re quite right: Gethin should give more examples from the 10 instances of what Anālayo considers errors. I take him to say simply that there are other explanations to account for differences, such as seeing them as legitimate variations, as a feature of the composition of suttas rather than a bug in its transmission. So he writes that
labelling them ‘errors’ is to prejudge the matter (p. 47)
because, in a way, that judgement is prior to and drives the quest to explain differences in a particular way. One could just as well start from a different premise and not consider them errors at all—I think that’s his point.
