Dennett, Pragmatism, and Animal Mind

Today's post is some sketching for a discussion of Dennett, both his basic philosophy of mind and his position on the minds of animals.
The "meta" stuff first: Dennett's basic view (The Intentional Stance) is that intentional (belief/desire) descriptions constitute a kind of universal hermeneutic that intelligent interpreters use to explain any system (including, say, paramecia and thermostats) that the interpreter does not understand at a lower level (the "design" level or the yet-lower "machine" level). "My car doesn't want to start." These intentional descriptions can be replaced as the design (the functional organization) of the system becomes better understood. "My car needs a new starter." This is an instrumentalist theory of mind: "mind" is a useful set of concepts for dealing with complex systems that are not (yet) understood in purely materialist terms. This pragmatist position appears to have the benefit, then, of buttressing materialism: when we appreciate that the intentional stance can be usefully applied to any dynamic system (the thermostat is an important example), we should also appreciate that there is no need of any transcendent source of intelligent organization, what Dennett has called "skyhooks" (Darwin's Dangerous Idea, one of Dennett's better books by the way).
I'm not sure that this grand materialist strategy works. The pragmatist component is key here, and thus we should think about standard problems for pragmatism. Pragmatist epistemology holds that we build up a conception (by this I just mean a system of description and explanation) of the world by trial and error, in the same way that evolution builds up our physical traits by trial and error (one of Dennett's best papers; "Why the Law of Effect Won't Go Away"). Thus the particulars of our conception are strictly contingent (underdetermined by the selection process): there are undoubtedly equivalently useful, or even more useful, possible conceptions that are very different from the one we happen to have, but we just have the haphazard selection history that we do. The beliefs are fixed by utility value, not by any traditional notion of "correspondence" with the world.
The problem here is that pragmatist epistemology is question-begging: why is it that some conceptions have more utility value than others? I'd guess they do because the world is more like this conception than it is like that conception. That is, the very fact of utility reflects environmental constraints on what we can and cannot (usefully) believe. And so intentional descriptions are ubiquitous because they pick out some deep feature of the world.
I think that what intentional descriptions pick out are relationships. I have been calling these "relational properties" but I'm not now sure whether these are properties or not (John Heil has got me thinking about the metaphysics of properties). But think of any formalizable relationship (modus ponens will do I think). All sorts of sets of things can come under these relational descriptions. I think that these relational forms are the "multiply realizable" things that persuaded functionalists to reject reductive materialism. That is, I think that materialism may indeed be false, because the mathematical/logical structure of the world is an a priori fact about the world (a forbidden "skyhook").
Well that's the first line I wanted to sketch out today. The other one is about Dennett's position on animal minds. A basic argument of mine is that both behaviorists and evolutionary psychologists do not have any principled reason to make distinctions between the semantics of psychological predicates applied to humans and those same words applied to many non-human animals. Dennett is squarely in this position. If it turns out that he doesn't think that intentional descriptions of human beings can be cashed out like they can for thermostats, then his entire project is a failure: the whole point was to naturalize psychology by showing that the intentional stance was a (mere) hermeneutic. Like all instrumentalist views, his only succeeds if it applies to human beings. Dennett suggests (Kinds of Minds) that the intentional hermeneutic in fact masks deep differences between the mental lives of humans and those of other animals, but one more time: either human intentionality can be cashed out into functional and physical descriptions, or he fails to naturalize psychology. Can't have it both ways.