Alex Rosenberg has an interesting debate going titled The Disenchanted Naturalist's Guide To Reality. If you keep up with the literature around naturalism, you should definitely read it. I don't agree with it though: I don't accept the reduction to physicalism - or at least, I can sort of accept it in the way that a compatibilist accepts the idea of determinism. I haven't quite shed the need for abstracta, and accept Armstrong's account of properties as being the least awful of some pretty unstomachable options. There are plenty of naturalists who think that the acceptance of anything other than a complete and total physicalism is heretical: I don't think that's so. The abstracta I accept are extremely modest, and they map to the subject matter of science rather than to Gods, faeries or anything magical or mystical. 
There's some parts of the naturalist system (if there is such a thing) that I still need to think through. Morality seems low on the agenda, but it's there. Much higher are questions of the mind, which I have intentionally neglected in order to focus on ontology. Being a naturalist isn't easy, you might say. If you have to do a few years of postgraduate study in philosophy in order to figure it out, surely there's something wrong. That may be so: being a naturalist may be hard - but just imagine the alternative! I tend to think that really we are quibbling over the details: fundamentally, you either believe we live in a world run by ghosts, ghouls and Gods or you don't. Either way, there are tough philosophical nuts you need to confront. Whether you get enraptured in the beauty of it all seems beside the point somehow.
One thing I find slightly puzzling in Rosenberg's essay (I'm sure I'll find more): Perhaps the most profound illusion introspection foists on us is the notion that our thoughts are actually recorded anywhere in the brain at all in the form introspection reports. This has to be the profoundest illusion of all, because neuroscience has been able to show that networks of human brain cells are no more capable of representing facts about the world the way conscious introspection reports than are the neural ganglia of sea slugs! The real challenge for neuroscience is to explain how the brain stores information when it can't do so in anything like the way introspection tells us it does - in sentences made up in a language of thought.
I'll try and reconstruct this as follows:
1. Neuroscience has shown us that human brain cells don't represent facts about the world.
2. Therefore, if one believes that one can access through introspection previously acquired beliefs or mental states that represent facts about the world, one is deluded.
There's a different reading too:
1'. Neuroscience has shown us that human brain cells don't represent facts about the world sententially.
2'. Therefore, [as (2)].
If (1) is true, (2) seems like a pretty good conclusion. But it feels like there isn't the whole story being told here. Think about a CD player or MP3 player or whatever. It stores streams of ones-and-zeros on some physical media that can be converted into audio playback in certain circumstances. Basically, you add layers of indirection to go from audio to ordered lists of ones and zeroes. Similarly, neuroscience may show us that human brain cells don't represent facts about the world - but there may be some process by which flows from the other. If we knew of one of these, we might have something a bit like a defeater for the argument from (1) to (2).
The modified argument seems worse though: there's obviously something like the first problem applicable to it. But there's a bigger problem: just because facts aren't represented sententially doesn't mean they aren't represented at all. If I said to you that there is a fact that there is a silver Jaguar parked on the road outside my window, that fact - let's grant some non-controversial form of being to the fact - is represented in natural, physical terms as a bunch of molecules arranged to form what we call a silver Jaguar and which happen to be spatially located in a street in central London and temporally located on the 28th of November 2009. Okay, it's also represented in terms of the words I used "there is a silver Jaguar parked on the road outside my window". It's also represented in some manner by my mind or brain - either as some mental concept or unit or just as a bunch of chemicals in some particular state. The physical representation of the fact - the car actually being parked on the street - is not a sentential representation of the fact. Why should we expect the brain to represent facts sententially? The brain being mental rather than linguistic hardly seems a big problem. The silver Jaguar out on the street isn't linguistic, but I seem to have the ability to tell you about it.
If science shows us that we don't actually represent facts sententially, that doesn't mean we should necessarily abandon that way of thinking. It may be a useful way of thinking about how we think, at least until a better model comes along.
(Disclaimer: These are not very well-considered thoughts. This is a blog, remember. I am allowed to use it to think out loud.)
