Is the frication of an affricate considered part of its aspiration? Or does the aspiration start at the end of the frication?
And does voice onset time (VOT) measure aspiration (as defined by the answers to the questions above) or something different when it comes to affricates?
I ask because I've seen it both ways. Abramson & Whalen (2017) say:
Affricate VOT ... treats frication as (essentially) part of the closure, so that the end of the frication counts as the release.
Abramson is a co-author of the paper that proposed VOT, so I guess that settles it, but the definition of VOT that starts at the release of the occlusion is also encountered, e.g. in Ueta (2018). Is the latter just a fringe view (or perhaps a slip-up), or is it a definition that is (or was) used by a sizable portion of the field?
This has implications on the interpretation of reports of phonological inventories as well. If a language is reported to contrast /ts/ and /tsʰ/, for example, is it safe to assume the contrast lies in the interval between the frication and the following vowel, or could it plausibly mean the contrast may lie in the duration of the fricative component ([ts] vs [tsː]) rather than in the interval?
(In research for this question I learned via Clements & Khatiwada (2007) that there is a concept called ACT (after closure time), devised by Mikuteit & Reetz (2007), that encompasses both frication and aspiration. But that doesn't help interpret older literature more accurately.)