Tony is continuing his habit of running away from threads and declaring
victory - does he see Comical Ali as his role model?
This missing context is that Tony has asserted, and totally failed to
support, Dembski's claim, abusing the No Free Lunch theorems, that the
evolutionary processes can be no better than blind guessing.
The No Free Lunch theorems state that averaged over all mathematically
possible problem spaces no algorithm can outperform a blind search
(random guessing). This is an unsurprising result, but no doubt requires
some skill to invent a formal proof.
They don't state that no algorithm can outperform a blind search when
applied to specific classes of problems.
For Dembski's assertion that the No Free Lunch theorems entail that
evolutionary processes can't outperform blind guesswork (in spite of
experimental evidence to the contrary) to be valid the set of problems
encountered by evolution would have to be an unbiased subset of the set
of all mathematically possible problems.
The burden of proof that this is the case lies on you (and Dembski).
However it can readily be seen that it is not the case. We observe that
small changes in genotype often result in small changes in phenotype. We
observe that small changes in phenotype often result in small changes in
fitness. We observe that small changes in environment often result in
small changes in phenotype. We observe that small changes in environment
often result in small changes in fitness. All these observations that
the problems encounted by evolution are a biased subset of all possible
problems, and that Dembski's conclusion is fallacious, based on the
failure of its unstated second axiom.
Most members of the set of mathematically possible problems spaces have
no or rare correlation between function values at neighbouring points -
if this held for living organisms we would be living in a world in which
a dog is as likely to give birth to a cat (or catfish, or coconut) as to
another dog (and in which life rapidly goes extinct due to successful
reproduction being vanishingly rare). You can't avoid this by going to
asexual reproduction and eliminating mutations, because the animals
would die anyway when the temperature changes by a fraction of a degree,
or the atmospheric oxygen concentration changes by a fraction of a percent.
Tony might argue about whether the efficacy of evolutionary processes is
sufficent, but he can't use the No Free Lunch theorems to dismiss them
preemptively - whether they are sufficient is an empirical question. He
might take some comfort from the relatively limited amount of
observations of the capability of evolutionary processes, but he has to
recognise that
1) the voluminous evidence of the factuality of evolution from several
diverse but mutually supporting lines of evidence means that it is
perfectly reasonable to infer sufficent efficacy from the fact that life
has evolved.
2) experimental measurement of rates of evolutionary change finds rates
in nature far in excess of the rates needed to account for changes seen
in the fossil record.
3) the diversity of domestic dogs, domestic pigeons and Brassicas
obtained over periods of a few hundred years demonstrates that these
rates are not a flash in the pan dependent on standing variation.
--
alias Ernest Major