Natural Selection 113definition of fitness does not imply that there actually are heritable variations of
fitness, or the nature of these differences, or how they interact with other forces of
evolution, and so on. Even if the principle of natural selection is true by definition,
just as it might be true by definition that “bachelors are unmarried”, this doesn’t
mean that there can’t be empirical ways of finding out about fitness, just as there
are empirical ways of finding out who is a bachelor [Sober, 1984].
At any rate, one of the most important developments in responding to the
tautology problem was the development and defense of the so-calledpropensity
interpretationof fitness [Mills and Beatty, 1979; Brandon, 1990; Sober, 2000].
According to the propensity interpretation, we can think of fitness as a probabilistic
dispositional property analogous to the way we think about dispositions such as
solubility. Just as a cube of sugar might have a certain probabilistic disposition
to dissolve when immersed in a certain liquid, we can think of the fitness of an
organism as a probabilistic tendency to survive to adulthood (viability fitness)
or a disposition defined by a probability distribution with respect to how many
offspring the organism could have (fertility fitness). Suppose that one organism
has a .8 chance of having 3 offspring (and a .2 chance of having none) whereas
another organism merely has a .7 chance of having 3 offspring (and a .3 chance of
having none). This means that each organism has a fertility fitness that is defined
by a certain probabilistic propensity known as theexpectednumber of offspring.
The first organism is fitter than the second, in that it is more likely to have more
offspring, but because fitness is a probabilistic propensity, there is no guarantee
that the first will have more offspring than the second.
The propensity interpretation identifies fitness with an organism’sexpected,
rather than its actual, degree of reproductive success. If fitness were identified
with the actual reproductive success of an organism or trait type, it is hard to
see how fitness couldexplainthe reproductive success of the organism or trait
since nothing can explain itself. Defenders of the propensity interpretation argue
that the identification of fitness with actual reproductive success commits an op-
erationalist fallacy — defining a term by the way in which it is measured. By
thinking of fitness as an expectation, one could argue that fitness explains actual
reproductive success the same way that a coin’s expectation of heads and tails
explains the results of a series of tosses. This suggestion avoids the tautology
problem; however, it isn’t clear that such a dispositional account is particularly
explanatory. If we explain why a group of organisms with one trait survived while
another group with a different trait died by saying that the former trait was fitter
than the latter, this amounts to saying that what occurred did so because it was
more probable than the alternative, which doesn‘t seem especially illuminating
[Sober, 1984].
Other objections have been raised to the propensity interpretation [Rosenberg,
1985; Beatty and Finsen, 1989; Sober, 2001]. One problem is that there is not
just a single propensity interpretation of fitness; rather, there are several different
ways of understanding fitness as a propensity. For instance, there can be differences
between the short and long term fitness of a trait. A given trait might have a high