Something's been troubling me about the activity reading. Since this is a blog, I'll put down this half-formulated worry, and either I'll follow up on it later, or someone else will do so for me.
My worry is whether the activity reading can say what Caston needs it to say. Consider the opening lines:
'Epei\ d' ai0sqano&meqa o3ti o(rw~men kai\ a)kou&omen, a)na&gkh h2 th|~ o1yei ai0sqa&nesqai o3ti o(ra~|, h2 e9te/ra.Caston construes this as:
Since we perceive that we see and hear, it is necessary either by means of the seeing that one perceives that one sees or by another [perception].Attend to Caston's 'by means of the seeing' (not 'by means of seeing', or 'by means of a seeing'.) That is, Caston takes th|~ o1yei to indicate the very same reality indicated by o(ra|=: "Which seeing? --Namely, the very same seeing that one perceives, when one perceives that one sees."
Caston needs to read the phrase in that way, in order to have the line express his view that one and the same act of perception (a 'token identical act') is both of a perceptible and of itself. (If it were just 'a seeing', then Aristotle would have to provide an additional argument, that that seeing was one and the same seeing as the original one. And one might naturally take 'by means of seeing' to mean the faculty.)
But can the definite article alone do that work, referring forward to something not yet said? (What would be an analogue of this elsewhere in Aristotle?) And if Aristotle meant to say 'by this act of seeing, namely, the seeing (of the color)' wouldn't we expect a qualification like that to come after o(ra|=? Moreover, as we saw, if Caston were correct about this, then a few lines later the very same phrase, th|~ o1yei, would be used, not simply with a different meaning (to indicate now a capacity rather than an actuality) but also in an entirely different grammatical role (as an ordinary definite article).
I'll have to investigate this more. Comments from philologists are welcome.