[ExI] Evolution is not random,...

[Damien Broderick]

At 11:21 AM 11/26/2007 -0500, Dan wrote:

>I think the better way of putting it is it's blind -- regardless of
>whether the processes underlying it are stochastic or deterministic or a
>little of both.  This means simply that it isn't forward-looking.

As Robert Picone noted, the original article wasn't talking about 
*evolution* (whatever that means), but made a drastically more 
limited claim, namely:

<The actual study made no such claims, they said that developmental 
evolution in the nematode vulva occurred  primarily through 
deterministic mechanisms.

<Basically, it was a claim that when ways to form a single, already 
defined structure are evolving, only two of forty factors were 
random, while the others were at least somewhat convergent.>

But suppose the question is regarded more generally. How do we *know* 
evolutionary change does not sometimes contain a forward-looking 
component, a kind of foresight or preemptive adaptation? The  answer 
seems obvious: firstly, because the altered future to which any 
critter would have to be pre-adapted is unknown and unknowable; and 
even if it were in principle calculable by an intelligence, the genes 
don't have direct knowledge of the world, nor any means to calculate 
probabilities, nor any means to modify themselves except via random 
shufflings.

But is all of this true? And how do we know it is? Has anyone put it 
to the test? Might it explain certain aspects of life without doing 
violence to what's already known? The proposition seems so outrageous 
that I'd bet very few scientists in the last half century have ever 
tried to model the question.

Suppose some of the spare clock cycles of any complex brain were used 
to model (i.e. imagine counterfactually) the benefits of certain 
phenotypic changes in worlds somewhat different from here&now. 
Absurd, of course, but if that were possible (say, if the Penrose 
hypothesis is right, and brains are quantum computers), and if some 
kinds of bio-feedback allowed microchanges at the cellular and 
genomic levels that conduced to favoured phenotypes, might we see a 
kind of Lamarckian aspect to evolution?

Suppose further that some kinds of psi are real, that a sheaf of most 
probably futures and their weightings can be glimpsed in advance, and 
that a kind of anomalous perturbation or psychokinesis at the 
microlevel can deform genomes to create phenotypes better fitted to 
such imminent environments... Absurd, of course, everyone knows that, 
but still--suppose it were the case...

What sort of experiments have been done that would reveal such a 
feature of evolution?

Cutting off the foreskins of umpteen Jewish generations is NOT a 
test, disproving any such effect by the lack of babies born without 
foreskins. There's no life-and-death crisis connected with being born 
with a foreskin; indeed, it's possible that in a culture where 
circumcision is mandatory and numinous, being born without a foreskin 
would make a boy freakish and disadvantaged.

But one might make several preparations of a known bacterial or 
murine population into which, at random or by preordained choice, 
certain poisons or currently-undigestible-nutrients will be 
introduced in two or three generations' time. Might there be an 
anticipatory genomic shift? This sort of "precognitive" advance 
adaptation would distinguish the eventual genome distribution from 
that of control groups. (I'm no experimentalist; this might not be 
the best way to test the idea.)

Is there any cultural difference between human or other complex 
mammal groups that might display such an effect unambiguously? The 
question, to say it yet again, is absurd, of course--but it's 
entertaining to think about how one might *test* such unthinkable 
ideas, rather than simply dismissing them ex cathedra.

Damien Broderick