Posted by
glen e. p. ropella-2 on
Feb 26, 2010; 12:01am
URL: http://friam.383.s1.nabble.com/speaking-of-evolution-and-self-organization-tp4550631p4636257.html
Thus spake Miles Parker circa 10-02-25 02:10 PM:
> So to be more clear I'm thinking of Comte, LaPlace, etc... and the
> whole idea of social physics..and the extreme consequences of that
> POV. And with that I can't help but mention that it struck me
> recently that this is precisely why the ascendancy of physics at SFI
> in the last how many years has been so off-putting. There is some
> kind of inherent hubris at work in that kind of approach.
Thanks for the clarification, although for me epistemology dominates
ontology. It seems useless to talk about what _is_ unless we are
excruciatingly clear about how we know what we think we know.
As for the dominance of physics, LdF actually carries your point all the
way to a conclusion in his book. I don't have it with me now; but he
explicitly states that we do not have to reform biology to adhere to
physical methods, rather the way we do physics must change in order to
elucidate biology. So, I suspect he would have a similar reaction
despite his ontologically positivist approach.
I can try to ferret out precisely what he means if anyone cares.
> It sounds like a really interesting point if I understood it, but can
> you say a bit more about how selection vs. auto-evolution are bound
> to different views WRT to determinism? Sorry if I'm being obtuse.
OK. Let's consider something simple like bit strings where * means
"don't care". A selection perspective starts with something like:
10*1*
And then, regardless of where the individual bit strings come from,
their history, ontogenesis, etc., they'll either fit that scheme or they
won't. Any given individual that pops out of thin air can be selected
for or against by applying the above pattern. There's no requirement for
randomness in this perspective. But there is a requirement for "don't
care". It's determined, but ambiguous.
Then let's take an autoevolutionary perspective. We have to start from
a kernel bit string, say 10011, and we have to have rules for the
transformation of that bit string. They don't have to be random rules
like mutation. They can be determined. Let's say our only applicable
rule is "flip the 3rd bit". We apply the rule and get 10111. Now we
have 2 bit strings.
>From the autoevolution perspective the only individuals that _can_ exist
are those produced by the only rules that apply. From the selectionist
perspective, the only individuals that _can_ exist are the one's that
fit through the filter. They're duals of each other. Neither are
stochastic, both produce the same result.
> Right, which to my view is like saying "determined, ad absurdium" :D.
> So, as I probably have made boorishly clear, I actually think that
> we cannot determine in principal, or in practice, or indeed in any
> lasting useful way.
That's a great skeptical approach. And I agree to a large extent. But
we have to be careful that our skepticism about the positive claim does
NOT become an absolute claim of the negative. [grin] I.e. It's fine for
you to _believe_ that determination is impossible, in principle and in
practice. But it is not fine to be convinced that determination is
impossible (and that you know that for a fact).
I think it's cool when someone like LdF claims that we will one day
be able to predict the products of evolution based on the theory of
autoevolution. If autoevolution ever matures to the point that it can
make a concrete prediction, then we may be able to falsify at least that
part of the theory!
But we can't jump in before autoevolution is mature and immediately say
that it's a fool's errand because we know determination in advance is
impossible.
To be clear, though, LdF is careful to avoid asserting that there is NO
randomness in nature. I did find at least one spot where he mentions
chaos (not by name) and several spots where he talks about asymmetry.
So, it's reasonable that he would admit stochasticity to some degree....
just not to the degree that neo-Darwinism relies upon it as a central
tenet. That, together with his use of "canalization" over
"determination", lead me to believe that his fundamental gist is that
autoevolution will facilitate prediction BETTER than neo-Darwinism
(which is softer than his rhetorical foil that neo-Darwinism isn't
scientific at all).
> I don't think the real me would agree, but as I spend a lot of time
> working on the realization that such a thing doesn't exist there may
> be some fraction of what I perceive to be me that does. OK, I promise
> I'll stop now...
Ouch. I've never practiced yoga.
--
glen e. p. ropella, 971-222-9095,
http://agent-based-modeling.com============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at
http://www.friam.org