Re: Adaptationism's Lessons (was Re: Evolution, "adaptation")

Paul Gallagher (
14 Sep 1996 13:20:58 -0400 (John Wilkins) writes:

>"Any unit of hereditary information for which there is a selection bias
>equal to several or many times its rate of endogenous change." 24

>What I think the Williams/Dawkins concept of adaptation requires is that,
>if there is a selection bias for a unit of hereditary change (including
>cultural units), *then* it must be an adaptation by definition, and what

My impression is that Williams does not accept Dawkins' ideas about the
selection of "cultural units." Check his 1992 book.

I'll briefly summarize our discussion on about the problems
with the model, or definition, you propose.

First, the selection coefficient at a particular locus is difficult to
measure. At the time of Lewontin's The Genetic Basis of Evolutionary Change,
no selection coefficient of any locus in any wild population had ever been
measured. Thus, your definition makes adaptation testable in principle, but
in practice almost impossible to test.

Second, by not providing any definition of fitness or adaptation independent
of selection, the statement, "adaptation is the result of selection" becomes
a tautology, and also rather uninteresting, since it makes no reference to
"good design" in nature, which adaptation through selection was originally
intended to explain.

Third, Dawkins, if not Wilkins, has implicitly reduced the study of evolution to
the problem of adaptation, moreover adaptation solely as a matter of gene
frequencies. It may be the case that selection is the sole cause of adaptation,
but is relatively trivial in the evolution of the genotype, for example, if
most alleles are selectively neutral. Plus, some people simply have other
interests. Selection may be the sole cause of adaptation, but the constraints
on the path of selection may be more interesting as determinants of form.
Plus, there is more to study of evolution than the frequencies of genes
with positive selection coefficients. This level of analysis is not privileged
over any other, (especially when it requires simplifying assumptions to keep
the mathematics simple enough for Dawkins to understand!)

A related problem pertains to the definition of the gene, and its relation
to the phenotype. Dawkins writes that if we were to find a gene for dyslexia,
we then would have found a gene for reading. The implicit definition of a
gene for a trait is then a series of codons that, if absent, preclude the
appearance of the phenotype. There are several problems with this: a confusion
of necessary and sufficient causes, and a confusion of identity, supervenience,
and causation. For example, taking an aspirin relives a tooth ache. It does not
therefore follow that the production of prostaglandins is the cause of the
tooth ache. If we were to find an abnormal allele that prevented prostaglandin
production, and hence the symptoms of the toothache, we could not call that the
gene that causes of the tooth ache, at least not without confusing a lot of

By redefining such terms as adaptation and gene to suit oneself, it becomes
possible to speak of "jealousy is as adaptation" as a testable hypothesis.
But it becomes one that is very difficult to test, and one whose meaning in
Dawkinsian context (a series of codons without which jealousy would
not exist that have a positive selection coefficient) differs from its ordinary
language meaning.

Of course, I'm probably just wasting my time here! Bryant or Wilkins will just
reply with:

>I find it frustraiting that neo-Darwinism's critics here are apparently
>limited to their reading of Gould, and don't even understand his
>"alternatives" to natural selection (such as his vaulted 'allometric
>rules of development') well enough to defend them against adaptationist
>attack. Describing is not explaining.