Andrews et al. subscribe to the view that distinguishing selectionist from nonselectionist hypotheses – or, distinguishing adaptations from mere spandrels or exaptations – is important to the study of psychology. I offer three reasons for thinking that this view is false; that considerations of past selective efficacy have little to contribute to inquiry in psychology.

Teleosemantics asserts that mental content is determined by natural selection. The thesis is that content is fixed by the historical conditions under which certain cognitive mechanisms – those that produce and those that interpret (respond to) representational states – were selectively successful. Content is fixed by conditions of selective success. The thesis of this paper is that teleosemantics is mistaken, that content cannot be fixed by conditions of selective success, because those conditions typically outnumber the intentional objects within a given representational state. To defend against this excess, advocates of teleosemantics must attempt to privilege some conditions of success while ignoring others. This results in selective explanations that are ad hoc, thereby depriving teleosemantics of the virtues it hoped to inherit from the theory of evolution by natural selection, including its alleged naturalistic credentials.

Sober (1992) has recently evaluated Brandon's (1982, 1990; see also 1985, 1988) use of Salmon's (1971) concept of screening-off in the philosophy of biology. He critiques three particular issues, each of which will be considered in this discussion.