Davidson suggests that we need to supply ourselves with enabling conditions for desire in order to understand how it issues in action. For, he tells us, we can perfectly easily imagine cases in which, whilst yet wanting something, we fail to act so as to satisfy our want. This is supposed to knock Wittgensteinian anti-causalism on the head since what, he asks, with a confidence which makes the question come across as unassailably rhetorical, could take the place of such an enabling condition if not a causal relation?
To which the Wittgensteinian reply (that I'm recalling here from Tanney) is (in my own words) that, leaving aside the in-any-case-rather-too-dialectically-blunt-to-be-philosophically-interesting causalism versus anti-causalism debate, Davidson's argument gets arse-over-tit the logic of the relation between desire and action. For the cases of unsatisfied desire which he imagines are in truth not ones which require us to supply enabling conditions to explain the efficacy of desire. They are rather ones which require us to supply defeating conditions if we are even to make intelligible the suggestion that we do yet genuinely have to do with desire even in the absence of the kinds of action which, in whatever sense we give the term, typically 'flow' from them.
Bandy about (or not) however you like terms like 'flow' and 'cause' and 'explain' and 'disposition'; the fact remains that we can all make a clear enough distinction between cases of failure which are, and cases of failure which are not, even intelligible in the absence of defeating conditions. And that when we apply it to the present case, we find - don't we? I know I do - it hard to sustain the appearance of intelligibility for a dogged ascription of a desire if here we meet with an absence of the relevant action despite not being able to locate any relevant defeating conditions. I really wanted an ice-cream and had the money and was in the well-stocked shop - but didn't buy one. Perfectly intelligible, I grant you. But perfectly intelligible even if nothing we could offer (e.g. wish to lose weight, allergy, hatred of the shop-keeper, recent ice-cream poisoning scandal) by way of explanation as to the non-procurement of ice-cream is found acceptable? I think not.
This little conceptual distinction between defeating and enabling conditions - this 'drop of grammar' - is, I believe, that down to which a huge amount of fuss in philosophical anthropology may be boiled. Mentalists are those who proffer enabling conditions just where constitutivists would proffer defeating conditions. The topic is neat, it seems to me, because, amongst other matters, it enables us to immediately see the connection between metaphysical doctrine and metaphilosophical inclination. The mentalists - who reckon on mind as extrinsic to behaviour and to world, now propose identity theories to give some kind of reality to the now otherwise dangerously unanchored mental domain, and also use their philosophy to offer putatively substantive explanations to link the domains (mind, body, world). The constitutivists - who reckon on mind as a constitutive dimension of behaviour - have no need to accept the explanatory burden / nice job opportunity met with in the mentalist's program. Start off on the constitutivist's foot and there will be no explanatory work to do to anchor mind in body or to show how we get from desire to action or more generally from mind to world - for we're always-already there (to use a Heideggerian phrase which, despite being (used and yet) described as 'annoying' by Rowan Williams, is surely rather helpful (and, yeah, not annoying at all! Jesus!)). Start off always-already in the world and you don't need to go about looking for necessary and sufficient enabling conditions. As with knowledge-first epistemologists: start off with knowledge and subtract from it (by providing defeating conditions) to get to mere belief; please don't start off with belief and then try and collect enabling conditions to help you securely make your way out to the world.
Something that interests me here is the work the enabling/defeating distinction as deployed by a constitutivist can do for us when considering the nature of the dynamic unconscious. For the distinction between conscious and unconscious can (to draw on some helpful conceptual formulation from Finkelstein) be said to be one regarding which Freud's treatment mirrors Davidson's. Freud tells us that the clinical facts he's discovered regarding our unconscious emotional life oblige us to offer an explanatory account of our having of regular-style conscious emotions. These, he says, we are obliged to comprehend as the upshot of our exercise of a faculty in inner perception of items - emotions - which in themselves are unconscious. (An unconscious emotion is, then, to be understood as an emotion of which we have yet to become conscious.) But what Finkelstein implies is that, instead of this putatively explanatory introspectionist-inspired ontological-myth-making, we do better to provide not enabling conditions which render conscious the otherwise unconscious, but instead defeating conditions which render the otherwise conscious unconscious. This, he claims - and I think it an irresistible consideration when rightly grasped - is what an unconscious desire is all about: it is a desire which expresses itself in our verbal and non-verbal behaviour yet which cannot be expressed through a verbal self-ascription. Contra Freud, a conscious desire is not an (intrinsically unconscious) desire plus an act of inner desire-directed consciousness. Rather, an unconscious desire is a(n otherwise conscious) desire the ability to express which in self-ascription is defeated through the presence of defence mechanisms. Contra Freud, unconsciousness is not, as it were, the resting state of nature onto which consciousness is bolted. Rather, talk of 'conscious' desire only makes sense as an antonym to unconscious desire (animals have neither; they just have desire punkt), and unconscious desire only makes sense as desire which is partly defeated in particular ways.
Desires burn with their own light, so you don't need a torch to see them; what you do need is a concept of a defensive occlusion to make sense of why sometimes we can't.