Michael Ruse has offered an interesting and insightful analysis of function statements in biology. The analysis he gives of statements of the form ‘The function of x in z is to do y‘ is :

1. (i) z does y by using x.

2. (ii) y is an adaptation ([4], p. 91).

The first thing to notice about this formulation is the peculiarity of step (ii). There are many cases in which we would naturally say that x was the adaptation, instead of y; or perhaps we might say that everything asserted in step (i) constituted the adaptation. For example, the function of the duck's webbed feet is to enable it to swim. But we would commonly and naturally say that the adaptation was the webbing, not the swimming. It seems to me that Ruse's insistence on the formulation in (ii) points to the major difficulty of his analysis : and understanding what drives him to this unnatural formulation suggests the way out of those difficulties.