My position on the issue can be described as "compatibilist" -- the at-first-sight odd idea that free will and determinism can go hand-in-hand. Determinism, of course, is pretty much required by a naturalistic or physicalist account of consciousness and culture (leaving aside quantum indeterminacy, which wouldn't affect the argument here in any case). But free will is another matter -- it's bound up with notions of agency, culture, and that cultural Indeterminacy Principle that was the topic of the previous post. As we'll see, I think free will and determinism are compatible simply in the sense that both are indispensable in their particular ways.
Compatibilism is a philosophical position that goes back a fair way, but it's never really surmounted the common intuition that the two are fundamentally incompatible, an intuition well captured by the so-called "Consequence Argument" (quoting from SEP):
1. No one has power over the facts of the past and the laws of nature.
2. No one has power over the fact that the facts of the past and the laws of nature entail that only one future is possible (i.e., determinism is true).
3. Therefore, no one has power over the facts of the future.
But the tack I want to suggest here is that, in situating an agent in the midst of a causal sequence, this argument is in fact conflating two properly distinct orientations or "modes of discourse", one that deals with causation and one that deals with agency.
This might be seen as a version of "Multiple Viewpoints Compatibilism", for philosophical categorizers. Within the perspective of causation -- what I'll call the mechanistic orientation -- the notion of "freedom" is either meaningless or pointless, since the only alternative to saying that our behavior and our will is caused is to say that they're uncaused or random. And within the perspective of will -- i.e., the agency orientation -- the notion of cause appears simply as reason or intention, which is always present no matter how minutely we examine ourselves (or others, to the extent we're able). In the latter orientation, freedom does indeed have meaning, but it means precisely that our will, our behavior, and the facts that result from that, are determined by our own reasons/intentions, rather than by some other force or agency (e.g., "fate", as Sartre suggested some while back*) that might be manipulating them. Given different facts (that is, counterfactually), from the mechanistic orientation our present behavior would be different and so, therefore, would the future facts -- and from the agency orientation, our reasons/intentions would be different, and so, therefore, again, would the future facts.
Now, Dennett has put forth a version of "Multiple Viewpoints", with "Intentional" and "Personal" Stances contrasted with a Deterministic one, and arguing that the former are simply more pragmatic when dealing with certain complex systems. My feeling is that, while the idea of the "stance" is a good one, this isn't quite right, nor sufficiently far-reaching. The "Intentional", in fact, isn't really a "stance" at all, but applies literally to all conscious or aware entities (and is only metaphorically, or, worse, sentimentally applied to non-conscious mechanisms like thermostats). And the "Personal Stance", or what I'm calling the agency orientation, hasn't to do with the complexity of an entity, but rather simply with the fact that we're in communication with it. If we encountered an alien species that was without language, for example, it wouldn't make sense (and certainly wouldn't be practical) to adopt a Personal Stance toward it regardless of how "intelligent" or complex individual organisms appeared to be (they might, in fact, even be manufactured rather than evolved entities). If we're able to establish communication, on the other hand, then, regardless of the "natural" status of the beings involved, the agency orientation comes into play and a moral dimension comes into being. And what enforces this is precisely that cultural Indeterminacy Principle mentioned above -- because they cannot deal with one another in a purely instrumental fashion, beings in mutual communication are in an inherently moral as opposed to instrumental relationship with one another (and the attempt to deal with one another instrumentally or manipulatively is itself widely viewed as an immoral act).
Thus, the notion of an "agent" -- the "one" assumed in the Consequence Argument above -- has meaning only within a particular orientation, in which "causation" has, at best, only a secondary significance, after "will" or "purpose". Within the mechanistic orientation, on the other hand, the notion of an "agent", in the true sense of the word, simply vanishes, to be replaced by causal sequences. Both orientations are needed --mechanism because of its obvious practical benefits, and agency because of the inescapable moral dimension. But trying to conflate the two, as the Consequence Argument does, is just a mistake -- and the result is often the sort of confusion and mystification that we find in the unfortunate, homunculus-like image of the "ghost in the machine".
* "So, contrary to what could be believed, the imaginary world occurs as world without freedom: nor is it determined, it is the opposite of freedom, it is fatal." The Psychology of Imagination, Washington Square Press, 1966 (1948), p. 221.