FITNESS AND FUNCTION

According to historical theories of biological function, a trait's fun ction is determined by natural selection in the past. I argue that, in addition to historical functions, a historical functions ought to be recognized. I propose a theory of biological function which accommodat es both. The function of a trait is the way it contributes to fitness and fitness can only be determined relative to a selective regime. The refore, the function of a trait can only be specified relative to a se lective regime. Apart from its desirable pluralism, only this view of relational function can support the function/accident and function/mal function distinctions commonly thought to be part of the concept of fu nction. Furthermore, only relational function correctly characterizes the explanatory consequences of function attributions in evolutionary biology.