But do they understand the brain at all?
I definitely see a problem in how some people consider the structure of the brain. It seems to me that they have this strawman idea that a physicalist/materialist brain model is like a single computer (maybe an Apple IIe or something), and they laugh because of all the things this Apple cannot do that a mind can.
But, as I said, this is a strawman at best.
A more reasonable concept of what the brain is like, in materialist/physicalist terms, is a closed set of computers - maybe four or five, maybe a dozen or more - which altogether share inputs and outputs with each other, and communicate with each other in different ways. Inputs come in from optic nerves, auditory sensors, etc., finding their ways to different computers in the system; processing occurs and output sent from one computer to another to another. No single computer is aware of its own processing, maybe; but they're sending data back and forth to each other, and are certainly aware of information they're receiving. This provides the whole system with internal awareness like that in our brains, and like this hypothetical system, no one part of the brain is self-aware, but the brain as a whole is aware of itself as a being.
Each individual computer works in a deterministic manner, but is composed of circuit boards of the type I proposed earlier in this conversation: the results of millions of years of plugging in random components and discarding the boards that do nothing or fail in odd ways. These boards work, and they work deterministically; but they don't always work like we would expect them to. As for the programs running on these computers, they're making up the programs as they go along. It's just as likely for a computer to come up with a program that causes 1+1 to equal 3 or aquamarine as it does to equal 2; it isn't until they interface with other sets of computers that they can get appropriate standards of programming together. And this interface is shoddy at best - sensory inputs interpreted and translated over time, and outputs via the meat-sack of our bodies.
Imagine if computers could only talk to each other by semaphor, for example, and could only use image processors with very low frame rates? Now, imagine if this is how we programmed our computers?
It seems to me that some people have this oversimplified concept of what deterministic brain action means. They envision inputs going in one end, being processed as part of a simple mathematical formula, and producing predictable outputs on the other side. What they're failing to realize is that the middle portion - the 'mathematical formula' - is actually billions of formulae, each loaded with hundreds of variables that most of us cannot account for consciously, each modifying every other variable and formula in some unknown way.
It's like dropping a mole of sulphur in a lake, and trying to guess what it will become when it reaches bottom.
I don't think anyone is arguing that oceans have free will; but if I put a plastic boat into the water, can I tell where it's going to wash ashore? Of course not. Does this mean it has free will? Of course not.
ST may not like 'complexity', but complexity accounts for most of the 'problems' that non-materialists seem to have with materialism.
...
Sorry for the long post... just needed to dump some nonsense out of my system.