Natural Selection and Causal Productivity - Mechanism and Causality in Biology and Economics - page 154

interaction from occurring at the population level. This case illustrates how even in
non-frequency-dependent situations, natural selection exhibits causal production at
the population level.
These three points taken together clarify the way in which populations can be
seen as causally productive via Salmon's Mark Transmission account. However, I
must make a few caveats. I am not endorsing Salmon's account over other accounts
of causation or mechanisms; other accounts may be needed to supplement Salmon's
account or may handle other sorts of cases better. 13 Moreover, I do not think this
discussion of Salmon's account captures everything there is to be said about natural
selection as a population-level causal process (or, to be consistent with Salmon's
terminology, perhaps I should say “natural selection as a population-level causal
nexus ,” since there are many interactions between populations and their
environments and populations and other organisms); for example, I have not said
anything about the way in which natural selection can be distinguished from other,
similar causal processes, such as sexual selection and artificial selection. But I do
think that Salmon's views on causal propagation and causal production can capture
some important aspects of the role of populations in natural selection. His views
help elucidate the ways in which populations propagate their influence through
space and time as well as the ways in which populations' interactions with various
other entities in their environment produce changes in those populations.
6 Conclusions
My main goal in this chapter has been to respond to Glennan; Glennan argues that
entities like populations can only give rise to causally relevant causes in the process
of natural selection, but as I have sought to show, populations can be causally
13 One worry that has been raised by a number of recent authors, including Glennan (Glennan
2009 ; see also Hitchcock 1995 and Craver 2007 ), is that Salmon's account fails to pinpoint which
of the causal processes that produce an effect are explanatorily relevant. In one version of an
example which purports to illustrate the problem, Ms. Slims chalks her cue stick with blue chalk
and deftly hits the cue ball, which hits the eight ball, which proceeds to the corner pocket. The
claim seems to be that, while the blue “mark” has been transmitted (perhaps even to the eight ball),
it is not explanatorily relevant to the effect. However, I think we need to be clear on what the effect
is; if we are talking about a token chain of events (and not a type of chain of events), then the effect
that occurred is that an eight ball with a blue mark dropped into a corner pocket. And the blue mark
is explanatorily relevant to that token event, just as the momentum of the cue ball is. We still might
be worried that Salmon wanted his account to be able to give an explanation for the event type
“ball in the corner pocket” and that the blue mark is not relevant to that. Here, I think three possible
responses are open. One is that explanatory relevance and causal relevance come apart; the blue
mark is always causally relevant, but it simply is not explanatorily relevant to the event type.
Second is to insist that in explaining why an eight ball with a blue mark has gone into the corner
pocket, we have already explained why the eight ball has gone into the corner pocket. Third is to
give up on using Salmon's account to explain event types and only use it to explain event tokens.
(Thanks to Christopher Hitchcock for helpful discussion).