I offered my thoughts in the comments on Vic Reppert's blog, and repeat them here:
The conclusion that rationality is *undermined* doesn't follow--at best the conclusion is that the connection between the physical causes and the rational inferences is at best a contingent one that is in need of explanation, which I think is a valid conclusion. But it's one that is in the process of being answered as we learn about how the brain and perceptual systems work, how language develops, and how the mind evolved.Naturalists and supernaturalists agree that we do engage in rational inferences. The supernaturalists think we do so using magical non-physical properties; many of them think that our minds are completely independent of our brains, though I think this is a position that is untenable in the face of empirical evidence from neuroscience (evidence which I have yet to see a substance dualist even attempt to address). In the face of arguments about the fact that computers are physical devices which engage in computation and inference, they respond that this is not real computation and inference, but only a derived computation and inference that is fully dependent upon human computation and inference.
If the fact that the brain operates in accordance with physical law undermined rationality, then the fact that computers operate in accordance with physical law would undermine their ability to perform logical inferences and computations.
The real question is *how* brains came to be able to engage in rational inferences in virtue of the way that they physically operate, not *whether* they do. Gilson (and Victor) argue that they could only have this ability by being divinely designed to do so--a thesis that doesn't seem to be particularly fruitful for scientific exploration.
Naturalists, by contrast, think that our abilities to engage in rational inference and language have evolved, and that they are both dependent on natural causes and productive in generating additional natural causes of reasoning and action. They are far from perfect--we can identify systematic failures of reasoning that occur (e.g., examples of the sort in Kahneman & Tversky's classic Judgment Under Uncertainty). And our understanding of our own abilities is far from complete--but is growing rapidly.
Scientific examination of our cognitive capabilities has been extremely productive, while the supernatural thesis has been moribund.