Since a single action can have many different descriptions, e.g. 'sawing a plank', 'sawing oak', 'sawing one of Smith's planks', 'making a squeaky noise with the saw', 'making a great deal of sawdust', and so on and so on, it is important to notice that a man may know that he is doing a thing under one description and not another. Not every case of this is a case of his knowing that he is doing one part of what he is doing and not another (e.g. he knows that he is sawing, but not that he is making a squeaky noise with the saw). He may know that he is sawing a plank, but not that he is sawing an oak plank, or Smith's plank; but sawing an oak plank or Smith's plank is not something else he is doing besides just sawing the plank that he is sawing. For this reason, the statement that a man knows he is doing X does not imply the statement that, concerning anything which is also his doing X, he knows that he is doing that thing. So to say a man knows he is doing X is to give a description of what he is doing under which he knows it. Thus, when a man says 'I was not aware that I was doing X', and so claims that the question 'Why?' has no application, he cannot always be confuted by the fact that he was attentive to those of his own proceedings in which doing X consisted. (Intention Sec. 6, p. 11-12, emphasis in the original)
This, at any rate, is the citation offered by Anscombe in her later "Under a Description" in which she endeavors to clear up a large number of confusions that people had surrounding the notion of an action's being intentional under a description.
In my previous post on these issues I claimed that this "action under a description" business was in tension with Leibniz law. My argument was basically this:
1) If my flipping the light switch and my alerting the burglar are the same action, then for any property P, my flipping the light switch instantiates P if and only if my alerting the burglar instantiates P.
2) Suppose that my flipping the light switch was intentional, but that my alerting the burglar was not intentional, and that my flipping the light switch is the same action as my alerting the burglar.
3) Then, there is a property — the property of being intentional — instantiated by my flipping the light switch, but not by my alerting the burglar.
4) So, my flipping the light switch is not the same action as my alerting the burglar.
5) But, (from 2) they are the same action.
So, we have a contradiction following from the supposition in (2) and Leibniz Law. And for what it is worth, I think the argument is right: one should not have a view which commits them to all the elements of (2), unless one wishes to abandon Leibniz Law.
My mistake was in thinking that (2) correctly encapsulates the business about actions being "intentional under a description". As Anscombe makes very clear in the paper "Under a Description", the point of this under-a-description business was not to posit some weird entities, actions-under-descriptions and then take the stance that a-under-description-D1 and a-under-description-D2 (1) are the same thing, and (2) possess different properties. Rather, Anscombe points out that this "under-a-description" business is qua "in modern dress", and takes it to attach to the predicate, rather than the subject. So, it is not that A-under-description-D1 is intentional, and A-under-description-D2 is not intentional; rather, A is intentional-under-D1, but not intentional-under-D2.
It is clear that this is the way to structure the view, if one wants to say that the flipping of the switch is the same action as the alerting of the burglar. It is perfectly fine for there to be one action which has the feature of being (for lack of better phrasing) purposefully-switch-flippy while lacking the feature of being purposefully-burglar-alerty.
This Leibniz-law concern is just one of the issues that Anscombe discusses in "Under a Description". As I begin gearing up for the Intention reading group I'm organizing, I'll definitely be going carefully through that article as well, since it did a really nice job, I think, of clarifying this talk of actions being "intentional under a description".