free will and fitness

Wed, 1 Jun 1994 10:27:00 PDT

Blackwell writes:

"In reply to a message by
Free will can affect mean fitness. For example, if because of free will,
people in a population commit suicide (an extreme example admittedly),
this will affect the mean population fitness, simply because there
is a chance that at some of those people otherwise would have been
considered fit. A better example would be where numerous people over
the last few years have persisted in having unprotected sex knowing
the AIDs epidemic was a threat. This was certainly free will, but
undoubted has had and will continue to have drastic consequences for
the population as a whole, severely reducing its fitness."

Actually, Blackwell is referring to a repost that T. Riley make of a reply I
sent to him.

THere is a confusion here between ABSOLUTE fitness and RELATIVE fitness.
Evolution (biological, Darwinian) is driven by relative, not absolute
fitness. Blackwell's reply also hinges on what is meant by "free will."
Free will, in the strong sense of decision making whose outcome is not
predictable on the basis of the current state of the individual (as opposed
to determinism which says that the decision is predictable based on the
current state of the individual) has no effect on relative fitness of traits
since the decisions made under free will (strong sense) are unrelated to any
trait (otherwise, the behavior would be predictable). As someone earlier
noted, the legal system assumes free will in the strong sense (e.g., if you
know the difference between right and wrong you are then a free agent and
your decision to rob--or whatever--is not determined by your past, hence you
are guilty of having decided to rob and therefore must be punished), though
in practise that viewpoint becomes modified (mitigating circumstances).

Free will in the strong sense runs into the problem of statistical
regularities (e.g., Blackwell's example of persons continuing with
unprotected sex and increasing their risk of getting AIDS) which allow for
probablistic predictions, though this could be countered by saying that free
will does not require that decision making be UNRELATED to the state of the
person; only that the same person again in the same state could make a
contrary decision, so that the unexplained variance under the probablistic
model becomes justification for saying that there is free will. Of course,
Graber will properly reply that the unexplained variance could simply
represent our current lack of knowledge, which then brings it all back to:
if the brain acts nonlinearly, then the inescabable imprecision of any
representation by the brain would allow for different decisions ensuing from
what are perceived of as identical conditions. The latter makes some sense in
the two stage decision process I have outlined in that the second stage
(embedding even what could be deterministic decisions at the first stage into
a system of meanings) could be highly sensitive to small perturbations (or
inaccuracies) in initial conditions.

D. Read