If you believe that naturalism is correct and you intend to apportion your beliefs to the evidence, then this question is particularly relevant for you.
In slightly clearer English, it is this: is the probability (P) that we have reliable cognitive faculties (R), given (/) background beliefs of naturalism (N) and evolution (E), actually (and counter-intuitively!) low? Alternatively stated, does the conjunction of belief in ‘naturalism’ and ‘evolution’, with all that this conjunction entails, serve as a defeater for our belief that our cognitive faculties such as memory and perception are reliable? The claim is not that our cognitive faculties are in fact unreliable, but that if naturalism and evolution are true, it is not rational to believe that these faculties are reliable, for we are without independent reason to think they are. If naturalism is true but undercuts the assumed reliability of our cognitive faculties, we are not justified in trusting our cognitive faculties when they deliver a conclusion of naturalism (or anything else!) [This whole discussion is going with the standard view of knowledge as justified true belief – the mere fact that a belief is true is not considered sufficient for it being ‘knowledge’, or a reasonable belief. I’m not an epistemologist, so I’m not aware of the details of the disputes around this, but as far as I know, noone claims that mere true belief is a sufficient criterion for knowledge.] This is the heart of Alvin Plantinga’s Evolutionary Argument Against Naturalism (EAAN) and I’m real sorry Al if I butcher it in what follows 😦
Translated into the clear logical form which (incidentally) the reviewer rather strangely dismisses as “quasi-mathematical formulae”, I would pose his question like this: sure, P(R/N&E) considered by itself may be low, but P(R/ N&E&X), where ‘X‘ is the success of science, is much higher!
This is very interesting and is one example of an attempted ‘defeater deflector’ which is in the realm of what Plantinga calls the “conditionalization problem”; the question gets quite technical when we consider what would count as a legitimate deflector of the defeater – for instance, one which had R built in to it (where X = R or entails R) would just be question-begging. More substantially, as Plantinga says (bearing in mind that ‘S’ is the subject of the beliefs concerned, i.e. the ‘believer’), “no proposition P that is evidentially dependent upon R for S – that is, such that S believes P only on the evidential basis of R – is a defeater-deflector for R.
As an aside, the sceptical reviewer Callahan implies that Plantinga believes that cognitive faculties are reliable based on the science of cognition. I doubt that Plantinga reasons in this way – I expect he would instead say that the science of cognition itself relies upon our cognitive faculties (is evidentially dependent upon R) and as such our faith in our cognitive faculties is more fundamental than that; their accuracy is directly perceived and hence assumed, in the absence of defeaters. Further, Plantinga’s argument is not that naturalism contradicts science (though if science is indeed an independent source of warrant for our cognitive faculties, one could say that what we would expect on naturalism is not what science shows us, so naturalism is to some extent disconfirmed by science) but rather that belief in naturalism undercuts itself (and in so doing, also incidentally defeats science and all other beliefs!)
The complex question which Callahan offers, which, cutting through the various misconceptions, seems legitimate to me is, first: does the success of science show that scientific beliefs are true (or more likely to be true than in the absence of such evidence)? It is plausible that continued progress in the scientific project does make scientific beliefs more plausible in some way (I am a scientific realist and a major argument for realism is the success of scientific predictions). But, if this is the case, the second question, broadly, is: to what extent does this rescue the reliability of our cognitive properties tout court? This is unclear. Perhaps the success of science only increases the likelihood of scientific claims being true if we first grant, i.e. have background beliefs of, the reliability of our cognitive faculties. But perhaps more can be conceded; maybe it is the case that naturalism is true and also scientific beliefs (and a consequent limited subset of our cognitive faculties) receive some warrant due to the success of science – but that our non-scientific beliefs (such as a belief in naturalism) are not warranted. In such a case, belief in naturalism would not be rational, even if it were true; we couldn’t rationally know that it were true.
Perhaps someone can help me clarify that last question; I think there are more distinctions within it to be made. Philosophy is all about making lots of distinctions; but not too many! Anyway, I think that the fact that science is successful is itself indirect evidence for theism, in that it fits better in an ordered than an unordered universe. In general I’m inclined to take the EAAN in a more restricted sense, as applying to non-empirical beliefs (rather than all beliefs due to our cognitive faculties), as Plantinga says, “metaphysical beliefs”, which aren’t too relevant to survival and reproduction, well; “it is only the occasional member of the Young Atheist’s Club whose reproductive prospects are enhanced by holding the belief that naturalism is true.” (p. 349, Where the Conflict Really Lies)
Even if EAAN is not an entirely successful defeater for naturalism, due to some kind of independent warrant for our cognitive faculties followed by some argument from, e.g. our perceptions, to naturalism, I think it does help to shift the burden somewhat – it is one of a collection of arguments clustered around the reliability of the scientific project itself that show that naturalism is not clearly the natural home of rigorous theorizing about and systematizing of the world as we apprehend it.
 In case you’re wondering, I personally accept evolution, in terms of common descent, (at least for the most part – I’m still working out the full relationship between teleology and the history of the biosphere) but I don’t accept naturalism, construed as the denial of the existence of God or anything like God.