Natural Selection and Causal Productivity - Mechanism and Causality in Biology and Economics - page 144

turn off the house is causally relevant to the basement flooding. However, failure to
turn off the hose is not an event (there is no object doing something), and as a result,
locality is not satisfied, either; thus, the “Mom” example fails to exhibit causal
production, according to Glennan. That is, we cannot, Glennan asserts, say that the
fact that his mother did not turn off the hose produced her flooded basement. On the
other hand, Glennan maintains that there are some cases of apparent causation that
fit causal production but not causal relevance, such as cases of overdetermination
(Glennan 2010b ). In overdetermination cases, each putative cause is sufficient to
produce the effect, but neither is necessary, so that one cannot say that if the cause
had not occurred, the effect would not have occurred (i.e., the counterfactual is not
satisfied).
Glennan claims that full understanding of the causal basis of an event requires
both the causally productive causes and the causally relevant causes and can be
expressed in the form: event c causes event e in virtue of fact f . I myself am not fully
convinced that there are two types of causes; indeed, I suspect that accounts of
causal relevance and causal production reveal different aspects of the same phe-
nomenon and that there are ways of handling the omission and overdetermination
cases. However, as nothing I intend on arguing for in this chapter turns on causal
monism, I will assume, for the sake of argument, that causal pluralism of the type
that Glennan endorses is true. Moreover, I will mainly focus on causal production,
since the question I am examining is whether natural selection exhibits causal
production at the population level.
3 Glennan's Arguments Against Population-Level Causal
Production in Natural Selection
To try to show that natural selection fails to exhibit causal production at the
population level, Glennan gives an example of frequency-dependent selection,
which seems like it would exhibit population-level causation if any kind of selec-
tion does (Millstein 2006 ). He asks us to imagine a population of light and dark
water bugs whose survival depends on not being seen by a predator fish. The rarer
form is always fitter than the more common form because the predator fish form a
stereotypic searching image associated with the more common color. Thus, when
the light-colored bugs are rarer, they are fitter, but once the light bugs come to
predominate in the population, the dark bugs become rarer and thus fitter.
Glennan says that the water bug example shows how and why the frequency of a
color form (a population-level property) is causally relevant to that form's fitness as
well as to changes in the distribution of forms within the population (a population-
level effect). Indeed, I have argued that natural selection in general (i.e., not just
frequency-dependent selection) satisfies counterfactual accounts of causation; if
there were no heritable differences in physical characteristics among the organisms
in a population (a population-level property), then there would be no differences in
reproductive success. In other words, there would be nothing to be selectively