Now, Humphreys has suggested in his most recent posting that our ability to discriminate causal from non-causal sequences is pre-theoretical, perhaps even pre-linguistic, and this seems quite plausible. But then we need to ask: why is such knowledge thought to be non-propositional? What exactly is the argument here? A language-of-thought person would still seem to be in business at this point. Yes, causal processes are going on out in the world, but surely so is everything else that's not mental (or so us realists would argue). So one question is: what makes causal knowledge different from other kinds of perceptual knowledge, given our ideas on non-propositional signs?
Jim Fetzer has suggested that minds are sign-users, and that icons and indices are non-propositional signs. But all that could be true even if the signs are external - outside the mind. Now, he has also suggested that sensations are icons, and this must mean, e.g., that my current image of the coffee cup resembles the coffee cup. (If only we could assume this resemblance! The radical skeptics would be largely out of business.)