[From Bill Powers (940915.0830 MDT)]
Avery Andrews (9404124.1237) --
I'm not sure about the historical connections, but the philosophers
distinguish between 'intentions' (purposes) and 'intensions' (an aspect
of meaning). The treatment of both is certainly a mishmash, however!
Back in the General Semantics days there was talk of "intensional" and
"extensional" meanings, which I never bothered to understand. I haven't
seen that terminology more than once or twice since, which probably
shows only how little I read of philosophy. Doesn't Searle speak of
"intention" with a t? I really think that the usages I have seen reflect
a linguistic phenomenon: a word that had a meaning suddenly had the
meaning removed by scientific scoffers, leaving the word orphaned.
People who still felt that there was something to the concept then
looked around for something else it could mean that was something like
the original meaning. They didn't come very close.
ยทยทยท
----------------------------------------------------------------------
Martin Taylor (940914.1300)--
RE: the beak of the finch
You won't see this until you get back, but I'll put it on the net anyway
in case anyone else wants to take it up.
I'm glad that I'm not the only one subject to making wrong judgments on
the basis of faulty algebra.
Right. Sorry. All I can say is that typing is bad for algebra.
p+(q/(1+D))
p = 1 - q
1 - q + q/(1+D)
1 - q(1 - 1/(1+D))
1 - q(1+D - 1)/(1+D)
1 - qD/(1+D). Voila. You are right.
------------------------------------
Actually, the whole thing would work much more straightforwardly if
instead of having two functions there were just one, representing the
probable number of organisms starting iteration N at each point where
there was one organism at the end of iteration (N-1). The probability
that there are no organisms would increase with distance from the
target, and the probability there are two would decrease with distance
from the target. There would still have to be a factor corresponding
to resource limitation pressure,
With your abstract concept of the "target," we have lost some detail
that makes a difference. You're positing right away that there is a
continuum of effects; that small changes in the position of the organism
in this abstract space will have small effects on reproduction rate. By
leaving out the details, you don't bring up the case where there are n
things the organism might do (or places it might go), with only one or a
small number of them being correct for improving fitness. This is the
case where either the organism stumbles across the critical mutation or
it doesn't. If it doesn't, it dies. Then we want to compare this case
with the one where the organism has some sensory indication of its
closeness to a critical location, the sensory information itself not
contributing to survival (only actually getting to one of the target
states would do that). This information provides the basis for E. coli
control of mutation rate, which steers the successive generations toward
the target state. Fitness is improved only when organisms get TO the
target state. In E. coli, its progress is organized around the smell of
food, but it has to get to the food to actually eat.
If you don't include this mechanism explicitly in the model, you won't
be able to tell the difference between directed and undirected
evolution. A highly efficient steering mechanism working off a basically
low probability of improvement per generation could be modeled without
the steering mechanism, but with a much higher probability of
improvement per generation being assumed. In a large population, the
rate of overall evolution could be exactly the same. There would be
nothing to indicate that the assumed probability of improvement was far
too large in comparison with the actual probability.
I'll keep working at this.
-----------------------------
One thing that should be mentioned: the higher levels of control that
are developed from reorganization, once the genome can supply such a
mechanism, can greatly influence evolution. The finches, during the
worst of the drought years, simply stopped mating. Neither males nor
females developed the typical color changes signifying readiness to
mate. This reduction in the reproduction rate, while apparently
amounting to a drastic reduction in fitness, was actually adaptive,
because birds trying to support fledglings under such extreme conditions
would probably starve themselves and the fledglings as well. This
adaptive change implies a rather sophisticated mechanism evolved to put
a long-term benefit ahead of a short-term one. This behavioral mechanism
would surely skew the results expected from blind selection.
Also, when mating was going on during high-stress conditions when larger
body and beak size were beneficial, and males greatly outnumbered
females, females selected as mates the larger males with the larger
beaks. So the reference signals in the female birds had a large effect
on which crossovers would be most likely to occur (I am told by a
biochemist to say "crossovers" rather than "recombination"). It was only
under conditions of plenty that hybrids began to appear.
There's an even more striking behavioral factor that may have had an
extreme effect on the progress of evolution. In one of the drought years
(maybe more than one, I don't recall), the scientists could see rain
falling day after day on another island, but they were in the rain
shadow and not a drop fell where the finches were being studied. The
finches were perfectly capable of flying the distance to the other
island, but very few did. Being bird-brained, most of them stayed where
they were and died by the hundreds.
But birds _did_ fly to the other island, and one can only guess how
their genes fared over there. From a larger point of view, the birds
that escaped to where the rain was falling amounted to a behavioral
variation that must have had an enormous effect on the future
propagation of the species. If the birds had been more venturesome and
had been able to communicate the location of better feeding grounds as
bees communicate where the nectar is, the forces of natural selection on
the dry island would have been swamped by the behavioral effects; the
island soon would have been deserted.
-------------------------------------
I think it's possible to be too simplistic in modeling evolution. The
simpler the model, and the less detail there is in it, the more it
amounts to a put-up job. If you just say that x percent of each
generation will reproduce itself by a factor k, the outcome is
foreordained. What makes such models interesting is trying to guess what
must be going on in order that x percent of each generation will
reproduce itself by a factor k. The hidden details contain all the
critical assumptions.
-----------------------------------------------------------------------
Jeff Vancouver (940913) --
Welcome back. I hope I can persuade Mary to give a report on her
attempts to communicate with Locke; you should find them interesting.
-----------------------------------------------------------------------
Tom Bourbon (direct) --
Thanks for the update; I'm still waiting for the next reel. Time for a
report to the net pretty soon?
-----------------------------------------------------------------------
Best to all,
Bill P.