Abstract: In the cognitive sciences, representations play a central role in explaining behaviour. Characteristically, representing correctly explains successful behaviour and misrepresentation explains failure. The standards of success and failure at play here are often tacit. What are they based on? Teleosemantics offers one answer: success is a matter of performing an evolutionary function. This paper uses case studies from cognitive neuroscience to argue that a system can have functions, at which representational explanations are directed, that do not depend on evolutionary history.

A broadly Aristotelian approach to understanding the functions of behaviour has two aspects. The first is that it is behaviour directed at a goal in the sense that the behaviour tends to achieve a certain outcome in a range of circumstances and in the face of a range of obstacles, and tends to pursue the outcome until it is achieved. (This does not presuppose that a goal is represented.) The second aspect is that achieving this outcome is a good thing for the organism in some sense. Evolutionary functions are one way of spelling out this second aspect of Aristotelian teleology. This paper will argue that both aspects of the Aristotelian approach are in place in many cases of representational explanation, and for good reason. The behaviour is goal-directed, and the outcome at which it is directed is good for the organism. In suitable cases being good for the organism depends only on recent contributions to the persistence of the organism, irrespective of its evolutionary history.