Some Problems Concerning a Relational Concept of Mind

In discussing Looren de Jong's `Some Remarks on a Relational Concept of Mind' (1997), this paper signals some foundational problems for ecological psychology, indeed for any psychology trying to incorporate a `relational' concept of mind. Central to Gibson's theory is the notion of function. But how scientifically viable are (teleo)functional explanations? Specifically, the paper critically examines whether such ('broad') functional categories as sensitivity-for-affordances (a crucial notion in Gibson's theory) but also Marr's `solid-object detection' or Churchlands' `food-detection' are respectable psychological categories that are grounded in some causally explanatory theory; or whether they have no more than folk-psychological significance. A third, more intriguing, option would be that these categories are indeed causally inert, but nevertheless do generate unadulterated `functional' explanations that are scientifically respectable without being causally explanatory. To defend this third option the standard appeal to multiple realizability is shown to be inadequate. Instead the notion of multiple supervenience is developed. It is argued that this notion effectively justifies the use of teleofunctional explanations compatibly with naturalist constraints on psychological theory formation.