114 Christopher Stephens
short-term fitness but a low long-term fitness or vice versa. Another complication
is that when a certain type of organism has an offspring contribution that varies
solelybetweengenerations, then the geometrical mean is the appropriate measure
of fitness, but when there is variation solelywithingenerations, the arithmetical
mean is the better measure of evolutionary expectation [Beatty and Finsen, 1989].
It can be difficult to determine whether the variation is distributed one way or the
other, and it is also possible that a type will change reproductive strategies over
time. There are still many unresolved issues about how to interpret fitness.
What Does Natural Selection Explain?
Sober [1984; 1995] argues that natural selection can explain the frequencies of traits
— such as being long-necked — in a population, but doesn’t explain whyindividual
giraffes have long necks. Selection has a merely negative role of eliminating variants
and does not have a “creative” role of explaining why a particular individual has
the traits it does. Sober illustrates this with an analogy involving a selection
process for admission into a classroom. In order to be admitted to the next grade
level, a student must pass an admissions test, which involves being able to read.
Suppose childrenA,BandC pass the admissions test but the other children
do not. The selection process (the admissions test) explains why individuals in
the next grade level have traitP (being able to read), and why there are no
individuals with trait not-Pin the next grade level. However, the selection process
does not explain why childAhasP rather than not-P nor why childBhasP
rather than not-P, and so on. The selection process determines why the room
contains only readers (those with traitP) but does not explain whyAis able
to read. A’s ability to read is presumably explained by appeal to the history of
her particular experiences and mental capacities. Analogously, Sober argues that
natural selection explains why a population has thetypes of organisms that it
does, but it doesn’t explain why theparticularorganisms have the traits that they
do. Presumably, something like adevelopmentalexplanation would explain why
a particular organism has the traits that it does.
Some philosophers have argued,pace Sober, that selection can, in the right
circumstances, explain why individuals have the traits that they do [Neander,
1988; 1995a; 1995b; Matthen, 1999; 2002; 2003], while others have defended the
view that selection does not explain the traits of individuals [Sober, 1984; 1995;
Walsh, 1998; Lewens, 2001; Pust, 2001]. Neander, for example, argues that natural
selection can explain the traits of individuals (e.g., why each giraffe has a long
neck) by being part of acumulativeselection process. In the first stage, selection
explains why a certain ancestral population has the traits that it does — e.g., why
all of the individuals have genes that produce long necks. Those with genes that
coded for short necks died out before reproducing. In the second stage, the traits
of the descendents of these giraffes are explained by noting that they inherited
their traits from the members of this ancestral population, which was full of long
necked survivors. Hence, selection helps explain the traits of these individual