Re: Species diversity through time

John Harshman wrote:

dkomo wrote:


r norman wrote:



On Thu, 08 Nov 2007 08:49:10 -0700, dkomo <dkomo871@xxxxxxxxxxx>
wrote:




The Last Conformist wrote:




On 8 Nov, 15:57, dkomo <dkomo...@xxxxxxxxxxx> wrote:




r norman wrote:




On Wed, 07 Nov 2007 17:52:44 -0700, dkomo <dkomo...@xxxxxxxxxxx>
wrote:

r norman wrote:

On Wed, 07 Nov 2007 15:24:06 -0700, dkomo <dkomo...@xxxxxxxxxxx>
wrote:

One man's random walk is another's increase in complexity.

More accurately one man's random walk is also known as a gambler's
ruin.

W Ross Ashby, a half century ago in "Design for a Brain: The Origin of
Adaptive Behavior", described how to build a problem solving system.
If my 40+ year memory serves correctly, it ran something like this:
Connect the system so that the unsolved problem is a state of low
entropy and the solution is a state of higher entropy. Then let it
operate and the problem will be solved!

I think that generating complexity by a random walk is about the same
level of sophistication and value.

Nevertheless that is the Gould/Harshman vision of how evolution produces
complex organisms. Mine was an ironic response to that. I don't
believe, of course, that that is all there is to producing complexity.

To clarify, evolution *appears* to follow a random walk after all the
mutation, genetic drift and adaptation across deep time have taken
place. This is not the same as saying that evolution *is* a random
walk. That's what a creationist/IDiot would say, and therefore
evolution could never produce complex life.

Yes, you can use random processes to develop complexity, but these
must be combined with sophisticated selection mechanisms and a ratchet
system first to recognize and then to accumulate complexity.

Is that all? This is all that's needed to develop complexity? If you
say yes, then you side with Gould and Harshman.

Yes, that is all there is to evolution, a process that certainly is
_capable_ of producing complexity.

This is only an assertion.





The most important point is that
complexity is a possible but by no means an inevitable result of
evolution. What you fail to accept properly is that a random process
plus a selection process is completely different from a random process
without selection. How do you define a random walk?

I'm using the term in the same sense Gould did. Replay the tape of
evolution, he said. If you do you end up with a different set of living
things all along the evolutionary time line. Replay it again. You end
up with yet another set. Each time you replay it, you get a different
set. That's because there are two significant random processes at play:
mutation and historical contingency (i.e. random environmental events).
This constitutes evolution's random walk.


No, it constitutes an assertion that evolution has a random element.
Unless Gould simply redefined "random walk", in defiance of prevailing
usage,

And what, pray tell, is this "prevailing usage".


See Wikipedia which says "In population genetics, random walk
describes the statistical properties of genetic drift "
http://en.wikipedia.org/wiki/Random_walk



Consider the state space of life. Evolution follows a trajectory through this space as a function of time. At a given t the "step" to the next point is random just as in an ordinary random walk, and it is a memoryless, Markov process because the next point depends only on the current point, again like an ordinary random walk. Every time Gould's "tape of life" is rewound and played again, a different trajectory through the state space results. That's why I say evolution is a random walk in Gould's view. Please keep in mind, this is not a view I agree with!


Please keep in mind that it's not a view Gould would have agreed with
either. Gould didn't think evolution was a random walk. He didn't think
that the next point on the typical evolutionary trajectory was random.
He knew that there was such a thing as natural selection, for example.


I was going to address the selection issue but ran out of steam. But I have just enough steam left to do it here.

Point being, selection itself *fluctuates*. That is, the environment is continually changing. Gould believed strongly in historical contingency did he not? He believed it was a major factor in evolution. If a huge comet hits the earth, that's one hell of a selection event! If there is a sudden increase in volcanic activity, which is one theory of what caused the Permian extinction, that also is a major change in selection.

Given that throughout deep time there have been many such unpredictable enviromental upheavals, evolution can't help but to be a random walk, expect maybe on very short time scales, where you could claim that if selection is fairly constant, evolution will then be approximately deterministic.

Now, this part I'm not sure of, but in Gould's replay of the tape of life metaphor, didn't he write that the historical contingencies would also be different? So it wasn't just play the tape through the same set of historical contingencies and depend on mutation to produce different results.

<snip rest for later -- steam low again>


--dkomo@xxxxxxxx

.

Relevant Pages

  • Re: Evolution increases the computational ability of organisms.
    ... >> Fluctuating selection would satisfy Gould equally well. ... >> Probably the major theme is that complexity follows a random walk ... You were trying to justify a claim that genome length increases through ...
    (talk.origins)
  • Re: Evolution increases the computational ability of organisms.
    ... is progressive in more than the Gouldian random walk way. ... natural selection. ... Probably the major theme is that complexity follows a random walk ... refers to ancient genome duplications which occurred at least several ...
    (talk.origins)
  • Re: Species diversity through time
    ... believe, of course, that that is all there is to producing complexity. ... This is not the same as saying that evolution *is* a random ... I'm using the term in the same sense Gould did. ... This constitutes evolution's random walk. ...
    (talk.origins)
  • Re: Species diversity through time
    ... believe, of course, that that is all there is to producing complexity. ... This is not the same as saying that evolution *is* a random ... I'm using the term in the same sense Gould did. ... This constitutes evolution's random walk. ...
    (talk.origins)
  • Re: Species diversity through time
    ... believe, of course, that that is all there is to producing complexity. ... This is not the same as saying that evolution *is* a random ... must be combined with sophisticated selection mechanisms and a ratchet ... This constitutes evolution's random walk. ...
    (talk.origins)