:-) Perhaps I shouldn't ask what you mean by 'sex in itself'
:-), but what I was thinking of was what biologists often
term 'recombination' (which is probably the most sexless word
in the language ;-)). Recombination is most definitely a
non-adaptive evolutionary process, as it reshuffles genetic
material.
> reproduction (ideally with a mechanism for trying new
> gene combinations) for evolution to proceed. I think you
Well, sex is often thought to fill the reproductive function,
at least for many animals. :-) <just joking here, I know what
you mean; the wording just caught my attention :-)>
> Even that concept seems questionable, though. Every organism
> lives in its own world, its own niche. If I'm a male peacock,
> I have to adapt to a world in which females look for that
> big feather display. Fitness is determined by the existing
> environment, not by abstract criteria. It can always be said
> of a species that its strategies are not optimal as seen from
> our position. Group selection, even ecosystem selection, can
> come into play, fostering apparently non-adaptive strategies
> among individual species.
Sexual selection is usually classed as an adaptive process (as
a subset of natural selection). Whether that's appropriate or
not is, I suppose, open to debate. I guess biologists typically
argue that the sex biases of the species are part of the fitness
landscape, even thought they are "artificial" in some sense, and
can even be "anti-adaptive" in the sense of other 'natural'
criteria. I'm not too familiar with the argument, though.
> Perhaps someone could itemize the non-adaptive processes
> Dawkins (or anyone else) envisages? I guess I'm just wondering
> how fuzzy the concept is.
There are five evolutionary processes commonly mentioned, although
depending on who's talking, sex selection gets added in, or others
are proposed. Mutation, recombination, genetic drift, and gene
flow (or horizontal transfer) are all non-adaptive evolutionary
processes. Selection is an adaptive process. Selection is probably
the fuzziest of the bunch, simply because the details of how it
works with natural circumstances, the sexual biases of the species
as we discussed, and so on are really complicated. Mutations of
their various sorts, recombinations, and gene flow are very well-
defined. Genetic drift is easy to model quite precisely, but it
is unclear as yet what importance its role is in evolutionary
history, so perhaps it is next-fuzziest. (We should note that
'fuzzy' is a relative term, and there are quite sophisticated and
un-fuzzy models of NS using adaptive landscapes and such which
exist.)
-Greg