According to Stephen Stich, “natural selection does not care about truth; it cares only about reproductive success” (Stich Reference Stich1990, p. 62; quoted in the target article, sect. 9, epigram). Similarly, I suppose, natural selection “does not care about” digesting food, pumping blood, supplying oxygen to the blood, walking, talking, attracting mates, and so forth. For each of these activities can either be (biologically purposefully) set aside (the vomiting reflex, holding one's breath under water, sleeping) or simply fails to occur in many living things. Nonetheless, surely the main function for which the stomach was selected was the digestion of food, the lungs for supplying oxygen, and so forth, and a main function for which our cognitive systems were selected was the acquisition and use of knowledge – that is, true belief. As McKay & Dennett (M&D) observe, confusions about this arise from failing to take into account any of three fundamental facts about biological function, on each of which I would like, very briefly, to expand.

The first is that recurrent failure to perform the functions for which they were selected is completely normal for many biological structures and activities. The barnacle waves its little fan foot through the water once, twice, ten times, a hundred and ten times, and the hundred and eleventh time it picks up a microscopic lunch. One hundred ten failures for one success. (Bad statistics characterize almost all hunting and fleeing behaviors of animals.) A job of the gazelle's strong leg muscles is to allow it to outrun the lion, a job of the protective eye blink reflex is to keep sand out of the eye, and a job of various body membranes is to keep pathogens from entering; but, of course, none of these jobs always gets done. Our knowledge-making systems performed their functions less and less reliably, up to very recent times, I suppose, roughly as the objects of belief got further and further from in front of our noses. But, of course, it is getting straight about what is in front of our noses that is the first order of importance for us. Getting straight about things further away can be very helpful too, when we can manage it; and when we don't manage, false belief about distant things may, until recently, have been pretty much equivalent, in the simple randomness of the results, to having no belief at all about such matters. As with hunting, clearly it is better to try, even though one fairly often fails, than not to try at all. This is very different, of course, from saying that failures are helpful or that the organism has been designed to produce them. Failures are by-products of design for success. (On the inconstancy of Normal supporting conditions for proper operation of the cognitive systems, see Millikan Reference Millikan, Machamer and Carrier1998.)

The second observation is that biological systems may sometimes be designed to suspend or override the functions of certain of their parts, so as to avoid damage when Normal conditions for successful operation – operative conditions that helped account for past successes hence for selection of these parts – are conspicuously absent. Numerous animal species “play dead,” perhaps actually going unconscious, in circumstances where any manifestation of normal life will only raise their chances of injury or death. The fuse blows, the shear pin breaks, by design. Again, this does not imply that failures to function properly are helpful, but only that in some circumstances it is best not to attempt to function at all.

The third observation is that structures kept in place by natural selection primarily for one purpose are sometimes also utilized by piggyback mechanisms to help serve different functions, perhaps even interfering with their original functions on occasion. Certain kinds of beliefs might be useful to us for purposes other than their normal cognitive use regardless of truth or falsity (not, however, because of their falsity, as M&D emphasize). This is theoretically possible, but I do not think convincing evidence for it has been offered. If certain kinds of errors are common and also systematically useful, it does not follow that they are common because they are useful (compare our first observation above). It is also very hard to tell, given not our own current concerns but the concerns of natural selection itself, whether or not an error is useful. Those more hopeful of continuing life than is justified by the evidence may live a few months longer. But Dawkins has claimed that “[a]s soon as a runt becomes so small and weak that his expectation of life is reduced to the point where benefit to him due to parental investment is less than half the benefit that the same investment could potentially confer on other babies, the runt should die gracefully and willingly. He can benefit his genes most by doing so” (Dawkins Reference Dawkins1989, p.130). Something like this may also have been true, most places and times, for terminally ill adults being cared for by kin. People who are more confident that they can perform a certain task than justified by the evidence succeed more frequently as a result. Externally administered steroids – steroids above what the normal body usually manufactures – have a similar effect, but they are not good for you. Perhaps these people should be turning their attention to other activities just as rewarding but for them with a higher rate of success. The hypothesis that we have systems that (purposefully) override the normal belief-forming systems to create false beliefs that will motivate us more and make us more successful implies that our normal motivational systems are, for some reason, inadequately designed, hence need to be compensated for. Surely we should wonder what got in the way of better design for our motivational systems in the first place?