In response to Glenn Morton's post, David Campbell wrote:
>Probably much more important is that natural selection makes evolution
>highly non-random. A favorable mutation has greatly increased probability
>of becoming established; a detrimental one has a greatly decreased chance
>of becoming established. The pattern of mutation does appear random in
>some senses of the word, but superimposed on that is the directional
>component of selection. If anyone wants more details, any population
>genetics text will give formulas.
>
>The illustration of the puzzled codebreaker does illustrate another aspect
>overlooked in the quotes disparaging the effectiveness of random
>processes. Lots of random arrangements will give coherent sentences; many
>more will give comprehensible approximations of sentences. The
>codebreaker by providing the selective pressure of a search for a military
>meaning eliminates numerous unsuccessful interpretations of the
>message. As long as the codebreaker picks an interpretation that pleases
>his superior officer, even if it is not exactly the message sent, he will
>keep his job. Likewise, there are generally numerous ways to get a
>protein to do a particular job.
I think an important component to add to discussions of random processes is
the use of qualifiers that make explicit their randomness *with respect to
x*. Without such a qualifier, in some sense nothing is random; even point
mutations are *determined* by some actual physical/chemical environment
(such as the availability of nucleotides, the conformation of the extending
strand, etc.). In this sense, the environment that causes genetic mutation
is as nonrandom as the environment which causes natural selection in
populations of organisms. The important point is that most genetic
mutations are clearly random with respect to subsequent fitness, whether of
the DNA molecule itself or the organisms it ends up in.
Such qualifiers are critical in discussions relating natural and
theological sciences, since the meaning of words in these two spheres are
so very easily muddled together and confused. By contrast, within
population genetics or evolutionary biology alone, the qualifiers are not
usually necessary since references to random events are generally
understand by the context.
Doug
This archive was generated by hypermail 2b29 : Thu Sep 21 2000 - 13:00:08 EDT