At Uncommon Descent — though not only there! — one often come across the view that naturalism is inconsistent with rationality: if one accepts naturalism, then one ought not regard one’s own rational capacities as reliable. Some version of this view is ascribed to Darwin himself, and we can call it “Darwin’s Doubt” or simply “the Doubt.” Should we endorse the Doubt? Or are there reasons for doubting the Doubt?
Here are some positions I can think of about the relationship between naturalism and rationality:
(1) Endorse the Doubt: if naturalism is true, then rationality is undermined; but rationality is a basic presupposition of all thought and discourse; so we should reject naturalism. (C. S. Lewis seems to have held this view, and Alvin Plantinga has a really nice, very powerful version of it he calls the EAAN, or Evolutionary Argument Against Naturalism — we can get into that in the comments, or we can start another thread just about the EAAN);
(2) Endorse the Doubt (inverted): if naturalism is true, then rationality is undermined — but naturalism is true, and therefore so much the worse for rationality. (I would ascribe a sophisticated version of this view to Nietzsche.)
(3) Defuse the Doubt: naturalism and rationality are consistent, if we have a sufficiently emergentist (non-reductive, non-materialist) naturalism and a deflationary conception of what the philosophical tradition has called “rationality” — e,g. “intelligence”. (On my reading this is John Dewey’s position, and he’s very clear that we should reject the entire Plato-to-Kant/Hegel tradition — what he calls “the Quest for Certainty” — in favor of a more modest conception of human cognitive abilities.)
(4) Reject the Doubt: naturalism and rationality are consistent, if we think of rationality in terms of (a) discursive, inferential capacities that intrinsic to <I>language</I>; and (b) these capacities emerge from, and are grounded in the perceptual-practical abilities that we share with other animals (and that characterize each of us at a very early moment of our lives). (On my view, (a) was nicely worked out by Wilfrid Sellars and Robert Brandom, and (b) is implicit in the phenomenology of Merleau-Ponty.)
The key issue at stake between (3) and (4) is how much continuity or discontinuity there is between the cognitive abilities of non-sapient animals and the discursive-inferential abilities of sapient animals. But I take it that a satisfactory hybrid view of (3) and (4), one that locates correctly the different dimensions of continuity and discontinuity, would be sufficient to refute both (1) and (2).
I have some thoughts about how to sketch out such a view, but I thought it better to turn this over to comments before saying too much more.