Is Natural Selection a Population-Level Causal Process? - Mechanism and Causality in Biology and Economics - page 161

Fig. 9.1 Distinction between
selection-for, selection-of,
and drift in terms of
discriminate/indiscriminate
sampling processes

Selection-for

Discriminate

Selection-of

Sampling Process

Indiscriminate — Drift

Although Millstein succeeds in offering a right characterization of natural
selection- for , she does not offer a right characterization of natural selection tout
court. In contrast, Brandon's account of natural selection is right, but he fails to
make a distinction between selection-for and selection-of. In this regard, Millstein
and Brandon are half right and half wrong with respect to the characterization of
natural selection. In my view, a more complete picture would be like this (Fig. 9.1 ):
And a more satisfactory distinction would be as follows:

Natural selection-for: a discriminate sampling process whereby physical
differences between organisms are causally relevant to differences in reproduc-
tive success
Natural selection-of: a discriminate sampling process whereby physical differences
between organisms are causally irrelevant to differences in reproductive success
Drift: an indiscriminate sampling process whereby physical differences between
organisms are causally irrelevant to differences in reproductive success

One might wonder whether I use the term “discriminate” in an equivocal way: on
the one hand, “discriminate” means sampling processes in which differences in
reproductive success are connected to differences in fitness (i.e., discriminate
between fitter and less fit traits). On the other hand, “discriminate” means sampling
processes whereby differences in reproductive success are causally relevant to
physical differences (i.e., discriminate between causally more and causally less
accountable traits). Indeed, given my intention to use the term in a broader sense so
as to cover both selection-for and selection-of, it is not surprising that the problem
of equivocality arises.
To deal with the problem, let me draw a distinction between the two senses of
“discriminate” and call them, respectively, F -discriminate and C -discriminate. The
term “ F -discriminate” is intended to refer to discrimination between fitter and less fit
traits. In F -discriminate sampling processes, physical differences among organisms
are still statistically , though not causally , relevant to differences in reproductive
success. As for the term “ C -discriminate,” it refers to discrimination between
causally more and causally less accountable traits. In C -discriminate sampling
processes, physical differences among organisms are causally relevant to
differences in reproductive success. It turns out that selection-of is an F -discriminate
sampling process, whereas selection-for is a C -discriminate sampling process.