Do We 'Mean' What We Say?

The difficulty in Hursthouse's way of dismissing the Doctrine of the Mean is the following.

She holds that the doctrine was accepted by Aristotle as a truism and a common way of analyzing the natural world, and that he, as it were, unreflectively imports it into his Ethics. Someone who has great respect for Aristotle's judgment in ethical matters, then, need not worry too long about whether he in fact meant something important by the Doctrine. It's as misguided as most of Aristotelian science. Rather, we should investigate what true things Aristotle was talking about, using this language, which he inherits from his environment, of a 'mean'. Aristotle didn't understand what he was doing here; we however can disentangle his true remarks from the packaging in which he presents them to us.

But isn't the Ethics marked by a great sensitivity as to which principles apply properly to the explanation of human character and action, and which principles, in contrast, perhaps because they are taken directly from natural science, 'belong to another inquiry'? A very good example of this (there are others) is VIII.1, where Aristotle insists than any general principle of 'like being attracted to like' (or unlike to unlike), taken from natural science, should not be employed in ethics, except perhaps in some analogous sense.

It's in this very context that, in fact, Aristotle brings in a physical notion of the mean (1159b19-20), and he's quite clear that this has a significance different from what would apply in ethics, and that it is a principle foreign to the explanation at hand:

But presumably opposite is drawn to an opposite not in itself, but incidentally, and its impulse is for a mean (meson). Why? This is good, for instance, that the dry become not moist, but that it arrive at a mean condition; likewise for hot and the other opposites similarly. Yet let's put these considerations aside. Of course (kai gar) they belong to another subject (allotriwtera).

One might mention also the great fanfare with which Aristotle introduces the Doctrine of the Mean; his use of it to organize the virtues and vices (an innovation); and his claim that it is verified by an examination of cases (1107a30-31). It's a well-entrenched thesis that cannot, I think, plausibly be dismissed in the manner that Hursthouse attempts, at least.