>>>I'm not sure if anyone has mentioned this yet, but according
>>>to my understanding evolution is not a random process in the
>>>usual statistical sense of the word random. Here I don't have
>>>in mind the usual statement that Natural Selection is not
>>>random. I mean that the parts of evolution we usually refer
>>>to as "random" are not necessarily random. IOW, in evolution
>>>random usually just means that an event does not anticipate
>>>the needs of an organism. One can have highly deterministic
>>>processes that satisfy this definition of "randomness".
>>
Thanks, Brian. I should have been making this point a year or so ago when
I got into a wrangle with Steve Jones about the role of randomness in
evolution. In that discussion Steve quoted Dawkins, and Dawkins essentially
said what you did: that the genetic variations are not biased towards
increasing or decreasing fitness.
Garry wrote
>>Perhaps you are right about this. The randomness in evolutionary processes
>>is not the statistical randomness of a stochastic process. But certainly
>>"chance" or probability comes into play, for example, in whether a
>>particular mutation better enables a particular organism to survive a new
>>environmental pressure and reproduce offspring with the mutation. So
>>'random' may be the wrong term to use here. Nevertheless, the point I
>>tried to make about the counterfactuals of chance still applies, for I
>>think that indeterministic laws or causation, in which several outcomes are
>>possible with associated probabilities, gives rise to these troubling
>>counterfactuals and hence to God's ability logically to be able to create
>>an indeterministic process which has a guaranteed outcome (Glenn Morton's
>>example of a Sierpinski gasket notwithstanding).
When there's an element of chance, random is proper. However, I think
Garry's concerns come from looking at the randomness from a human point of
view. There are several possibilities which, if true, IMO make randomness
less of a concern. The first is that a process which satisfies any test of
randomness we wish to apply may still not appear random to an omnipresent,
omniscient, omnipotent God. Secondly, if God's desire was to design a
nature that was robust -- capable of responding to disturbances (and of
course He would have had to ordain at least the possibility of
disturbances. I don't try to explain that, but note that there seem to be
plenty of disturbances in nature.) then He might well have deliberately
introduced random processes into genetics to ensure that a rich enough
variety of variations would be produced to enable living populations to
adapt to disturbances. That's very like what some adaptive control systems
do, and very like what some optimization algorithms do.
Engineering is a discipline which frequently must obtain predictable
results from systems that are subject to various indeterminacies. I gave
the example last week of nonlinear control systems designers welcoming some
noise because it tends to help the system settle in the presence of effects
like stiction. A more prosaic example is determining what the gain of a
servo must be to ensure that anticipated disturbances don't cause the servo
to violate repeatibility requirements. There are also problems in game
theory in which the optimal strategy consists of selecting from two or more
strategies randomly, with the determination of the probability of each
strategy being an optimization problem. If we humans can (usually) do a
very credible job of compensating for random disturbances and in some cases
even using randomness to our advantage, imagine what God can do.
----------------------------------------------------------------------------
Bill Hamilton, Staff Research Engineer
Chassis and Vehicle Systems, GM R&D Center
Warren, MI
hamilton@predator.cs.gmr.com / whamilto@mich.com (home)