Re: Myths about evolution in "The Top 10 Myths about Evolution"

William Morse wrote:
John Harshman wrote:
William Morse wrote:
John Harshman wrote:

(snip)

That was a practical question, actually. How do you measure the Kolmogorov complexity of an animal?

By definition - the shortest program that can uniquely specify the
animal(you already knew that). We can't do this yet, but we are getting much closer thanks to the efforts of the evo-devo scientists. And unless one thinks that magic is involved (which neither you nor I do), it is clearly possible in principle to measure the Kolmogorov complexity of an animal.

I agree. But if you admit that we can't now measure the Kolmogorov complexity of an animal, how can you assert that this complexity has increased during the Phanerozoic?

Until we do know more about development, we can get a good idea of the relative complexities of _related_ species on the basis of comparing features to see the extent of derived vs primitive (I'm supposed to use the "morphic" terms but I forgot them) features.

Apomorphic and plesiomorphic. How? By making the unwarranted assumption that derived features are more complex than primitive ones? How could you justify that?

Which is more complex - a sugar cookie, or a chocolate chip cookie?

A chocolate chip cookie, sure. But the rest of your argument doesn't follow.

In
order to get to a derived feature we have to develop the primitive
feature, and then add a developmental step (like making sugar cookie dough and adding chocolate chips).

No we don't. This is Haeckel's old law of terminal addition. You should be aware that it just isn't true.

So my statement is rather easy to justify. In fact the unwarranted assumption is that an apomorphic feature would be less complex than the plesiomorphic feature from which it develope. It is possible (in theory one could find a route that skipped the primitive development) but you would have to prove it. The assumption in the absence of more detailed information is that the apomorphic feature is more complex.

That may be your assumption. It's certainly not the assumption any modern biologist would make, and for good reason. Terminal addition is a false model of evolution.

Admittedly, until we know a lot more about development it will not be practical to compare the Kolmogorov complexity of say Drosophila melanogaster and Acer rubrum.

Or of any two species, as far as I can see. If you disagree, please compare the Kolmogorov complexities of Drosophila melanogaster and Drosophila simulans.

Give me a detailed list of the differences between the two taxa, and I will make the comparison.

How exactly would you do that? But here you go:

http://genome.ucsc.edu/cgi-bin/hgGateway?hgsid=108730966&clade=insect&org=D.+melanogaster&db=0

http://genome.ucsc.edu/cgi-bin/hgGateway?hgsid=108730966&clade=insect&org=D.+simulans&db=0

Of course most of the differences have no effect on phenotype, if that matters.

You apparently are
arguing that the upper limit of complexity was already achieved by the
early Phanerozoic.

No. I'm saying that there is no general trend of increasing complexity in the Phanerozoic fossil record of animals. It's clear that even under a simple random walk model the most complex animal at any given time is expected to be more complex than the most complex animal at any previous time; only some kind of maximum complexity barrier would stop this from happening, and I'm not arguing that any such barrier has been reached (though I won't say that it hasn't either).

I understand your argument, and disagree if you insist on using the Phanerozoic. I actually would be open to an argument that there is a complexity "soft" barrier (at least regarding morphological as opposed to behavioural complexity) that might have been reached.

If you actually did mean that you disagree, can you give reasons for your disagreement.

I have done so and continue to do so. In the meantime, you haven't given
a single counterexample, e.g. a phylum in which there has been a decrease in complexity during the phanerozoic. In an exchange of pedantries, my expectation is that I will be informed.

You misunderstand what a counterexample would look like. It would be a phylum in which there has been *no increase* in complexity, not a decrease. And that would refer to no general increase. Some lineages could have become more complex, others less complex, as long as there was no general trend. And for that, I would suggest looking at any phylum at all, with the arguable excpetion of Chordata.

I did not make myself clear. My statement is that the increase in
complexity does exist as an observable fact. Gould I think might have accepted this view:

"And yet, undeniably (even for such curmudgeons as me), a basic fact of the history of life - the basic fact, one might well say - seems to cry out for progress as the central trend and defining feature of life's history." (Full House, p. 145) Gould goes on to reject the idea of a direction in evolution and offers the lengthening right tail distribution as an alternative.

This seems to me more like Darwin's rhetorical practice of making a claim only to refute it, though I don't have a copy of Full House handy to check. But if you're agreeing with Gould that any increase is a right-tail phenomenon, do you also agree with my point, that the fossil record doesn't display a pattern of increasing complexity during the Phanerozoic?

What didn't you understand about "the increase in complexity does exist
as an observable fact"?

I didn't understand what you meant by it. What do you mean by it? What increase in complexity? Complexity of what, exactly?

His explanation is the drunkard's walk, but he doesn't deny the observation of complexity increasing with time. So yes I want to defend that claim, and I want to see you attack that claim. I want you to show me mammals from before the K-T that are more complex than raccoons.

No, if anything you need to show me mammals from before the K-T that are less complex than raccoons.

My knee-jerk reply was all of them. But after doing a small amount of checking, I don't know that I can defend that statement. One for you.

Thanks. Is that a general acceptance of my point also regarding turtles, cephalopods, etc.?

Nope, sorry. After doing again a small amount of checking, it appears that complex cephalopods such as octopuses, and leatherback turtles (actually marine turtles in general) only appeared circa 120 million years ago. They may not be post K-T, but their predecessors do not exhibit their degree of complexity.

What is your evidence that octopuses are more complex than previous cephalopods, or that modern marine turtles are more complex than older turtles?

The octopuses are universally considered the most intelligent
invertebrates.

How intelligent were ammonites? Do you know? How intelligent were any shelless cephalopods that existed 500 million years ago? (We don't know whether there were any, of course, because they don't fossilize.)

Since marine turtles didn't exist before 120 million
years ago, they have to be more complex than older marine turtles. There mere existence shows an increase in complexity among turtles, since they have successfully invaded a niche they did not previously occupy.

This is not actually true. Some of the oldest known turtles were marine; Archelon, for example. And the increase in complexity you are talking about here is apparently the complexity of an ecosystem or community rather than of an organism.

And while we're on that subject, all land animals and plants show an
increase in complexity during the Phanerozoic, since there weren't any at the beginning of the Phanerozoic.

Again, you seem to be talking about ecological complexity, not organismal complexity. Unless you want to claim that land animals are universally more complex than aquatic animals.

I want you to show me Jurassic turtles that are more complex than leatherbacks. I want you to show me Devonian cephalopods that are more complex than octopuses.

Again, that's not the assertion. You need to show that past animals are less complex than living ones.

And then I want you to explain how it could be that the mechanisms of evolution by natural selection as explained by Darwin could have failed to produce an increase in complexity over time.

How would it be expected to produce an increase? Selection doesn't increase complexity; it optimizes (to the extent it can) organisms for their immediate environments. You seem to think this will result in a bias toward increased complexity. That's certainly not your boy Gould's view. Or mine. Gould is talking about an increase in variance resulting from a random walk, nothing more.

I agree that is Gould's point - but even that will produce an increase in complexity with time, which I thought was what you were disputing. I note that you yourself have cited the increase in diversity of taxa over time.

It would produce a slight and hardly measurable increase in the mean (or modal, or median) complexity. The only noticeable increase would be an increase in complexity of the maximally complex organism. And this would not produce a pattern in the fossil record.

Now I am willing to argue for a much more positive bias towards increasing complexity, and that the increasing complexity would be expected in both the ecosystems discussed by Ernest Major and the individual organisms that we have been talking about. One of the mechanisms that produces an increase in complexity is the Baldwin effect.

Explain why the Baldwin effect is expected to increase complexity rather than leave it the same or decrease it. Because I can't immediately see a rationale.

Let's look at beavers as an example. Some proto-beavers learn how to construct simple dams, enjoy much higher survival rates because of it, and the dam construction becomes hard-wired (or an increased ability to learn dam construction does). I suppose you could argue that some other ability was lost during the hard-wiring process, but it can't be one that is key to the dam building, and it seems much more likely that the hard-wiring increases the complexity of the organism.

I can see how the Baldwin effect could result in an increase in complexity. The question is how the Baldwin effect would be expected to result in an increase in complexity. Your faith that traits will never be lost, only gained, has no obvious justification.

I have no faith that traits will never be lost, but that does not affect the argument. The point of the Baldwin effect is that the organism is the same, save only that the behavior has been genetically added.

Better to say that behavior has been genetically modified. It's not as if some brand new gene encoding the behavior has just popped into existence. Existing genes have been modified to alter some genetically determined behavior. There is no reason to suppose that this makes the organism more complex.

Now traits may be lost, but the impact of the Baldwin effect is an increase in complexity. Unless you argue for an upper bound on complexity (and you have given no examples or argument for such an upper bound) so that a trait must be lost when one is gained, there is no reason for the Baldwin effect to do anything other than increase complexity. The trait loss is a separate effect that is random, while the Baldwin effect is directional.

I deny that assertion. It relies on a misunderstanding of genetic mechanisms.

By the way, I picked the beaver example because I recently encountered a spectacular beaver dam on a kayak trip. This was a stream that we could paddle down (well sort of - there was a lot of dragging involved), and we came on a beaver dam across the entire stream with about a two foot drop over the dam. Dang but those little rascals are good!

Perhaps a more fundamental mechanism is that predicted by control theory - adding additional feedback loops improves stability, and more stable systems are more likely to survive perturbations than less stable systems.

Remembering that we're supposed to be talking about organisms here, not ecosystems, what exactly is it the stability of that you're talking about? Do you contend that modern crustaceans have greater stability of something or other than Cambrian ones?

Yes. Horseshoe crabs have been around forever, but I would be astonished if modern limulus were not significantly different from ancient limulus. To suppose otherwise would be to suppose that the existence of sex is an accident, and that its prevalence is an aberration.

You will have to explain this somehow. Is sex supposed to be an engine of increasing complexity? If so, why? And why should modern Limulus being different from ancient Limulus mean that they were also more complex? How would you tell? By the way, Limulus isn't a crustacean.

I never said they were. I said that I contend that modern crustaceans have greater complexity than Cambrian ones, and then I gave a separate example of an ancient species that shows little morphological change but is still likely to be different now than when it first appeared. To respond to your first question, AFAIK the current consensus on sex is that it persists because it maintains diversity as a response to parasitism or environmental change or both, If it were possible for an organism to be perfectly adapted so that no change were necessary (ancient Limulus being identical to modern Limulus) then sex would have no purpose.

So what does that have to do with increasing complexity? And (to repeat the question you ignored) why should modern Limulus being different from ancient Limulus mean that they were also more complex?

Really, there is no chance of a general phenomenon here, based on the history of life. If there were one, it would affect all groups, and modern jellyfish, e.g., would be more complex than Paleozoic jellyfish. Do you claim that to be the case?

I didn't note this initially, but the claim that a general phenomenon would have to affect all groups is a non sequitur. There only has to be a general increase in complexity among most groups, with no countervailing decrease in complexity among the remaining groups. Are you claiming that the jellyfish have decreased in complexity?

No. And I agree. Not all groups have to increase in complexity for there to be a general phenomenon, just most of them. Now I would contend that very few groups have demonstrably increased in complexity since the Cambrian. As examples of this I cite...everything. I will agree that Chordata can be argued to have increased in general complexity. But what else?

Remember that there is a well documented increase in biodiversity with time.

How is that relevant?

Unless you want to state that the average complexity per species has decreased, an increase in number of species times the same average complexity per species equals an increase in overall complexity.

I deny that claim. Overall complexity cannot reasonably be computed by adding up the complexities of individual organisms. 500 species of Drosophila are not more complex than one species of Drosophila. And if you measure complexity this way, it doesn't show up as an increase in complexity of fossils.

There is also an increase in symbiosis and sociality with time ( or are you going to argue that ants and naked mole rats were present in the early Phanerozoic).

No. We don't actually know much about eusociality or lack thereof in the Cambrian. There were certainly no ants, termites, or mammals; or land life of any sort beyond bacteria, as far as I know. At any rate, I'm sure that the right tail of the eusociality curve has extended since the Cambrian. So?

So there has been an increase in complexity.

If that's how you want to define it, I agree. But would you agree that this is not a general increase in complexity, nor would it produce a pattern in the fossil record?

I can't say whether modern jellyfish are more complex than Paleozoic jellyfish ( although I would not be surprised if they were) - but there is still a trend towards increasing complexity.

What trend? And what exactly do you mean when you say "trend"?

Well, let's see. There is an increase in skeletal complexity, there is
an increase in individual behavioral complexity (intelligence), there is
an increase in social behavioral complexity, there is an increase in
ecosystem complexity (at least according to Ernest Major), there is an
increase in species diversity, there is an increase in technological
complexity (which we haven't previously discussed). Geez, I don't see any trend, do you?

No. What I see is an inability on your part to keep focused. If you want to call right-tail phenomena increases, fine. But right-tail phenomena don't produce patterns in the fossil record.

I note that you still haven't given any counterargument to the prediction of increased complexity due to the better control added by additional feedback loops.

What prediction? What feedback loops? I don't remember any such prediction, and I can't find one in this post.

.

Relevant Pages

  • Re: Stephen Wolfram vs. Charles Darwin on natural selection
    ... =A0respective animals 'down to the same level of detail'. ... organisms is more complex than an individual organism. ... So if I see an increase in complexity, is it aa increase in complexity ... all of life, from bacteria to highly specialized animals? ...
    (sci.bio.evolution)
  • Re: Evolution increases the computational ability of organisms.
    ... far removed from single genes. ... pattern *is* a reduction in complexity, though, even if all of the genes ... features) that is retained and well developed in us. ... fact that virtually all animals that have ancestors that acquired ...
    (talk.origins)
  • Re: Evolution increases the computational ability of organisms.
    ... far removed from single genes. ... pattern *is* a reduction in complexity, though, even if all of the genes ... features) that is retained and well developed in us. ... fact that virtually all animals that have ancestors that acquired ...
    (talk.origins)
  • Re: Innoivationm and the Curse of Knowledge, etc
    ... its complexity" which was the whole point Glen. ... that they don't study animals in all their ... conditioning was trivial. ... stimuli - as in so-called concept formation etc. etc. ...
    (comp.ai.philosophy)
  • Re: Myths about evolution in "The Top 10 Myths about Evolution"
    ... And unless one thinks that magic is involved, it is clearly possible in principle to measure the Kolmogorov complexity of an animal. ... feature, and then add a developmental step. ... Or of any two species, as far as I can see. ... bookshelf I found this from Matt Ridley in _Genome_ "The longer animals ...
    (talk.origins)
Loading