Is Natural Selection a Population-Level Causal Process? - Mechanism and Causality in Biology and Economics - page 173

4 Conclusion
Selection-of and selection-for, arguably, are one and the same sampling process.
Thus, an adequate account of natural selection should be broad enough to include
both selection-for and selection-of. Millstein's causalist account, however, defines
discriminate sampling process (natural selection) in terms of selection-for only. Her
view would be more complete if the notion of discriminate sampling process
includes both F -discriminate (selection-of) and C -discriminate (selection-for) sam-
pling processes.
A second problem with Millstein's causalist account is that it does not meet a
plausible requirement: namely, whenever natural selection and drift operate
together, a change in the strength of natural selection implies an inverse change
in the strength of drift, and vice versa. In contrast, Brandon's account makes perfect
sense of the trade-off between the two factors' strengths. The problem with
Millstein's account arises from the view that, ontologically speaking, natural
selection and drift are two separate sampling processes. To avoid the problem, it
suffices for Millstein to de-unify the concept of drift in such a way that drift
comprises both unrepresentative and indiscriminate sampling processes. Since
drift is an umbrella concept supposed to cover all cases where chance plays a role
in a sampling process, such an un-unifying (even un-unifiable) drift concept, of
which Brandon's and Millstein's account of drift are a part, turns out to be credible
rather than ad hoc.
The view that natural selection is a population-level cause seems to face a
dilemma: it is undermined either by the epiphenomenon problem or by the
overdetermination problem. The prospects of the view thus turn on a satisfactory
solution to both the problems. With the help of an analogy, I show how the two
problems can be dealt with: natural selection is not a shadow process, for it is part of
the cause of a change in the relative frequency of traits. On the other hand, the
overdetermination problem can be avoided, for although many individual-level
causes suffice to sum up and calculate how the relative frequency changes, they
do not suffice to causally explain why it changes in the way that it does.
Millstein's causalist account of natural selection and drift is instructive and
persuasive as well. Still, one may wonder whether it is too incomplete to be
adequate. It is noteworthy that the idea of fitness plays no role in Millstein's
characterization of natural selection and drift. To be sure, Millstein has good reason
to refrain from using the term “fitness”: it is contentious, and she does not need it to
make the points she wants to make. Nonetheless, it seems undeniable that biologists
routinely use (variation in) fitness to account for evolutionary events. Though it
remains unclear whether an explanation in terms of (variation in) fitness would
properly be called causal, it seems safe to say that any account of natural selection
and drift, in which fitness plays no significant role, would hardly be adequate.
Indeed, any such account would seem to throw out the baby with the bath water.
How to put fitness into the causalist account and give it due weight? An answer to
this question must await, however, another paper.