Re: Evolution increases the computational ability of organisms.

John Harshman wrote:
dkomo wrote:


What does this mean? In Roger Lewin's _Complexity: Life at the Edge of Chaos_, in Chapter 7 "Complexity and the Reality of Progress", he interviews Norman Packard and there is the following exchange:

---------------------------------------------------------------------
I can see what might be meant by computation in organisms that have a brain of reasonable size, I said to Norman, but what about clams and trees? "Survival has to do with gathering information about the environment and responding appropriately," Norman answered, clearly echoing Ayala. "Bacteria do that, by responding to the presence or absence of certain chemicals and by moving. Tree communicate chemically too. Computation is a fundamental property of complex adaptive systems, which, you'll remember, is optimized at the edge of chaos. Any complex adaptive system can compute; that's the key point. You don't have to have a brain to process information in the way I'm talking about it." But it helps? "It's higher on the scale of computational ability, if you like."

When you say "higher on the scale", I asked Norman, are you suggesting a history of successive increases in computational ability in evolution? "That's how it looks to me," he replied. "Intuitively, it seems reasonable that the task of survival requires computation. If that's true, then selection among organisms will lead to an increase in computational abilities. That creates an arrow of change, not just a drift upward."

Most species on Earth today are single-celled organisms, as in the pre-Cambrian, and much of the rest are insects, I said. That doesn't look like inexorable progress to greater computational ability, does it? "We're talking about survival, " said Norman. "And, yes, there are countless niches out there in which species do very well with certain levels of computational abilities. But where survival is contested, mostly you will see an increase. Think of it as a constant exploration of the utility of increased computational complexity in evolution. Sometimes it gives an advantage, and that gives you the arrow."
-----------------------------------------------------------------------

--p. 138 & 139


Of course, this isn't exactly a new idea. Lewin writes:

---------------------------------------------------------------------
For example, in a classic text on evolution by Theodosius Dobzhansky, Francisco Ayala, G. Ledyard Stebbins, and James Valentine, the "ability to gather and process information" is said to have increased through evolutionary history, and, indeed, to be a mark of progress. A few years ago I attended a conference at the Field Museum in Chicago, where the topic was "evolutionary progress." Francisco Ayala was one of the first to speak. "The ability to obtain and process information about the environment, and to react accordingly, is an important adaptation because it allows that organism to seek out suitable environments and resources and avoid unsuitable ones," he said.


Hey, I went to that conference. All the other participants (I remember
Gould in particular) thought Ayala was out to lunch on that notion. He
caught heavy flak from both speakers and audience. And I agree that he
deserved it.

Funny that Roger Lewin didn't mention this at all, eh? And he was there also. Perhaps one's memory of what went on and how significant it was is colored by one's biases?

If I remember, there was a symposium volume, and his
contribution to it bore no resemblance to his talk, probably for that
reason.


Edward O. Wilson also considers information processing as a measure of complexity. "No question about it," he told me. "There's been a gradual increase in information processing over the last 550 million years, and particularly in the last 150 million years." If at least some biologists and the dynamical systems people collectively point to information processing as a mark complexity, we may be getting somewhere.


No, what we are getting is that not enough biologists are thinking about
the whole picture. They're concentrating on the outliers

You're already presupposing a statistical model by using such a term. I don't think of multicellular organisms, and mammals in particular as "outliers".

and deriving
universal principles. In order to claim that there's an increase in
information processing in the last 150 million years, don't you have to
claim that bacteria are getting more complex,

No.

that bacteria now are more
complex than bacteria 150 million years ago?

No. Bacteria are optimized for rapid *reproduction*, so they have to stay simple, which means a small genome and a limit on their "computational ability". They solve the challenges in their environment by mutation and swapping genes, and reproducing as fast as possible.

And we have zero evidence
for this.

So what?

What we have evidence for is that the extreme right tail of
the complexity or information processing distribution has been extending
fitfully over time. (It's now occupied by humans, presumably, with
mammals and birds in general being toward that tail.)


Once again you're bringing in your pet statistical model and that colors your whole view.


----------------------------------------------------------------------

--p. 137 & 138


Here's an example of information processing inside E. coli:

"But how can an individual bacterium, locked into the genome it has inherited, cope with environmental fluctuation?"

"Think, for instance, of an E. coli living in the erratic environment of a human colon, dependent for its nutrients on the whimsical eating habits of its host. If the bacterium is deprived of the amino acid tryptophan, which it needs to survey, it responds by activating a metabolic pathway to make its own tryptophan from another compound. Later, if the human eats a tryptophan-rich meal, the bacterial cell stops producing tryptophan for itself, thus saving the cell from squandering it resources to produce a substance that is available from the surrounding solution in prefabricated form. This is just one example of how bacteria tune their metabolism to changing environments."


Campbell & Reese, _Biology_, 6th ed., p. 347


Is there any evidence at all that this sort of thing has increased
within the last couple of billion years? I think not.

No? Cambrian Explosion? The progression from prokaryotes to eukaryotes to fish to amphibians to reptiles to birds and mammals? That's not evidence that "this sort of thing has increased."

Bacterial
communities have changed very little as far as we can tell. All they've
done is invade a few new environments, like the guts of metazoans for
example. But these environments closely resemble some of their previous
environments, and I doubt that much new "information processing" was
necessary. In any case this was all worked through several hundred
million years ago.


If I may use genetic algorithms as an example here, a GA may be progressing quite nicely toward finding an optimal solution to a problem by evolving a few members of a large population, but most of the members of that population will have poor poor fitness. Gradually though, the overall average fitness of the population increases, and a few members will have the sought for high fitness.

It's the same principle when you look at information processing capabilities across the entire range of living species and across evolutionary time.


--dkomo@xxxxxxxx


.

Relevant Pages

  • Re: Evolution increases the computational ability of organisms.
    ... interviews Norman Packard and there is the following exchange: ... history of successive increases in computational ability in evolution? ... of the utility of increased computational complexity in evolution. ... considers information processing as a measure of complexity. ...
    (sci.bio.evolution)
  • Evolution increases the computational ability of organisms.
    ... interviews Norman Packard and there is the following exchange: ... history of successive increases in computational ability in evolution? ... of the utility of increased computational complexity in evolution. ... considers information processing as a measure of complexity. ...
    (comp.ai.philosophy)
  • Evolution increases the computational ability of organisms.
    ... interviews Norman Packard and there is the following exchange: ... history of successive increases in computational ability in evolution? ... of the utility of increased computational complexity in evolution. ... considers information processing as a measure of complexity. ...
    (comp.ai.philosophy)
  • Re: Evolution increases the computational ability of organisms.
    ... they mean by "computational ability" in this context. ... of the utility of increased computational complexity in evolution. ... the environment, and to react accordingly, is an important adaptation ... considers information processing as a measure of complexity. ...
    (sci.bio.evolution)
  • Re: Evolution increases the computational ability of organisms.
    ... of the utility of increased computational complexity in evolution. ... considers information processing as a measure of complexity. ... claim that bacteria are getting more complex, ... example of how bacteria tune their metabolism to changing environments." ...
    (talk.origins)