Actually, I think Carrier's passage does make this distinction. I just didn't quote all of it:Which does seem to discuss hypotheses and distinguish their nature, separately, from the nature of sources of evidence, much along the same lines that you describe (although this passage could be more clear on the issue of distinguishing the meaning of e.g. "Lucian's book," whether that's referring to a known manuscript, a reconstructed text, etc.).
By contrast, a medieval manuscript of Lucian’s account of his interactions with Proteus Peregrinus (in Lucian’s Death of Peregrinus) is evidence of what happened to Peregrinus, because more than plausibly it is causally connected to that: the best explanation (or at the very least a definitely plausible explanation) of the manuscript’s existence is that it sits at the end of a chain of causally connected copies of an original book written by Lucian (either literally, in his own hand, or by a scribe at his dictation and correction), which in turn was causally connected to what happened to Peregrinus by the fact that Lucian actually saw what happened to Peregrinus.
Carrier then discusses how this isn't the only possible explanation, but it is certainly a plausible one, so the medieval manuscript is evidence for the hypothesis. Evidence, of course, does not establish a certainty; it increases a probability.
Statistics: Posted by Peter Kirby — Thu Jan 30, 2025 3:17 pm