122 Christopher Stephens
(“apparent design”) as the primary explanandum, whereas developmental biolo-
gists often view the diversity of organic forms as the primary explanandum. On
the developmentalist picture, the diversity of organic forms is best explained by
constraints-on-form, rather than constraints-on-adaptation.
Amundson points out that there are two different kinds of dispute here — one for
each part of the two-stage process involving variation and selection. The first stage
concerns the production of heritable variation, whereas the second concerns the
sorting of that variation. Natural selection only operates at the second part of this
process. The debate between advocates of selection and advocates of drift occurs
in this second part of the process, whereas advocates of developmentalism claim
that it is at the first part of this process where constraints play a significant role.
In other words, development may or may not constrain what the actual variations
are in a population, whereas other sort of constraints (such as the genetic ones
mentioned previously) operate on the already existing variation.
Amundson then uses this to draw a distinction between “soft” and “hard”
adaptationism in which the soft adaptationists allow for the significance of de-
velopmental constraints on the process of determining what the variants will be.
According tosoft adaptationism, developmental constrains-on-form must result in
a constraint-on-adaptation. Hard adaptationism, on the other hand, claims that
developmental constraints can be overcome by the power of natural selection, and
so constraints-on-form are not constrains-on-adaptation.
Amundson also notes that some constraints-on-form can actually increase cer-
tain adaptive possibilities. For instance, because of certain developmental con-
straints there will be certain correlated changes in form — it is not necessary
that both the left hand and the right hand have to independently be the target
of selection for them both to change. This means that adaptationists such as
Stephens and Krebs (1986) are wrong in thinking that constraints must be defined
as restricting adaptations.
Reeve and Sherman [1993] claim that adaptationism is consistent with the exis-
tence and importance of constraints, though Amundson points out that there are
different kinds of constraints, some of which conflict with adaptationism. Whether
adaptationism is in tension with constraints depends on the kind of constraints –
adaptationism is genuinely in conflict with there being a large number of genetic
constraints. It is not, however, necessarily in conflict with mechanical constraints,
such as those that occur when leg length must be traded off with leg strength
[Maynard Smith, 1978].
Although Amundson and Godfrey-Smith are raising interesting issues here, it
isn’t clear how central they are to the adaptationism debate. It is hard, for exam-
ple, to make sense of the claim that natural selection is more powerful or “does
more” in a case where it chooses among many actual variations compared to a
case where, due to developmental constraints, there are only a couple of alternate
phenotypes to act on. This is precisely because, as Amundson and others point
out, natural selection operates at a different point in the process than the develop-
mental constraints. To resolve this issue would require determining the extent to