Re: Species diversity through time

Bill Morse wrote:

r norman wrote:


On Tue, 13 Nov 2007 14:23:19 -0800, John Harshman
<jharshman.diespamdie@xxxxxxxxxxx> wrote:


r norman wrote:


On Tue, 13 Nov 2007 12:03:55 -0800, John Harshman
<jharshman.diespamdie@xxxxxxxxxxx> wrote:



dkomo wrote:



John Harshman wrote:

<snip top>

dkomo wrote:





John Harshman wrote:




I wouldn't go examining the metaphor all that closely. He merely
means that life is not some grand progression toward some
predetermined goal. If the world is chaotic (in the mathematical
sense), then all you need would be an infinitesimal change in
initial conditions, or a little bit of randomness here and there, to
make the end result vary wildly. Where are you going with this? Do
you think that the history of life is indeed a series of inevitable
events?

Gould is wrong. There are basins of attraction in the state space of
life. Replay the tape of life and the end result will be close to
what it was before, as long as you don't vary the major historical
contingencies. A little bit of randomness here and there won't
perturb evolution's trajectory all that much.


Can you present any evidence for this interesting claim? And could you
clarify what you mean by "close"? Would we necessarily have humans, or
mammals, or chordates? Or is it just that, given animals, we would
almost certainly have many species with eyes and legs?


I have just as much evidence for my claim as Gould has for his, and I'm
just as likely to be right.

Or so you assert. Since you claim that Gould has no evidence, I will
take this as an admission that you have none yourself.


Well, both claims are based on a sample size of one. So I would say,
statistically speaking, that both have equal likelihood of being true.
Which is what dkomo said.

Nonsense. The sample size for some cases of convergence is in the
hundreds. And I have explained that evolution in isolated continents and
islands is a fair approximation of this "runnin the tape again" thing.

I've merely restated the idea of convergent
evolution in a more modern way. For instance, Simon Conway Morris
thinks that if you replay the tape of life convergent evolution will
lead to the evolution of pretty much the same set of plants and animals
as before.

Define "pretty much the same". Or let Conway Morris do it if you prefer.
What exactly does this mean? I have offered some notions, and you may
pick one if you like, or present your own.



Eyes and legs are attractors. Hairy homeotherms producing
live young are attractors. Large brains with thinking minds are
attractors.

Let's eliminate mumbo jumbo whenever possible. When you talk about
"attractors" you are talking about things that are commonly evolved as
responses to selection, yes? I don't see any other mechanism. I agree
that eyes seem to be like that in animals. But legs, or anything you
would want to call legs, seem to have evolved only twice, with a few
borderline cases. Hairy homeotherms only once. Large brains only once.
So what makes you think they're "attractors"? If they are, why haven't
they attracted more groups? I'll tell you what an attractor looks like:
a tree. The tree form has evolved independently some hundreds of times
in diverse groups of plants. Clearly, trees are going to happen if we
rewind the tape, assuming we first get land plants at all. But if you
think anything that's happened once during life's history is inevitable,
I don't think you can appeal to anything other than faith as a
justification.



You do think Morris has the same chance of being right as Gould does,
don't you?

No, I don't. (By the way, his last name is "Conway Morris" -- a
hyphenated name without the hyphen).



"Evolution shows an eerie predictability, leading to the direct
contradiction of the currently accepted wisdom that insists on
evolution being governed by the contingencies of circumstance."

http://en.wikipedia.org/wiki/Simon_Conway_Morris

Yes, and this is nonsense. For every example of convergence, usually in
ways you would find uninteresting, there are many examples of failure of
convergence.

Your argument appears to be mathematically untenable. Given the vast size of
the state space to be explored by variation and selection, there should be
essentially zero examples of convergence compared to the examples of
failure of convergence unless there is in fact a large element of necessity
in evolution as opposed to only chance.

Of course. Nobody is claiming that adaptive convergence can't happen.
We're only arguing about its power. And nobody is claiming that
convergence is due to chance either; obviously it's due to selection (or
obviously to me -- dkomo doesn't like selection).

Now we can't do the Gould experiment directly -- "rewinding the tape of
life". The closest we come is various natural experiments comparing
evolution on isolated islands and continents. There are of course
caveats -- starting conditions are never quite the same, including input
species. But if all is so stereotyped as you seem to claim, shouldn't we
see the same groups evolving multiple times? Take Australia vs. most of
the rest of the world. Sure, we see marsupial "wolves". But instead of
bovids we get kangaroos. The similarities are interesting, though mostly
simple and obvious, there being not all that many ways to be a mammalian
carnivore. But the differences are profound. Or take Madagascar. In
Africa we have woodpeckers, in Madagascar we have aye-ayes. Convergent
on some level, sure, but again profound differences.

On the kangaroo question, before you insist on the similarities being
simple, I suggest you take a look at some pictures of the heads of
kangaroos and deer. Just looking at the heads, you would be hard pressed to
tell the difference. Now you may define differences in locomotion as
profound, but what if I don't?

Why wouldn't you? Sure, dentition of mammalian herbivores tends to
converge. Archosaurian herbivores have/had different solutions.
Convergence always works better when it starts from similar places.

I would also like to point out that starting point is important. Many
forms have evolved multiple times within some group, but never or seldom
outside it. Is eusociality inevitable? Perhaps it is in hymenopterans,
where it's happened dozens of times. But how many times has it happened
outside of hymenopterans? Just a few. So the argument that eusociality
is inevitable would seem to transform largely into some argument that
hymenopterans are inevitable, or whatever it is (perhaps haplodiploidy?)
that makes them evolve that character so much. Hence, contingency.

Well certainly there is much contingency in evolution. But one might note
that Alexander predicted naked mole rats without knowing they existed.
Eusocial mammals might not exist, but if they do they are likely to look
like naked mole rats.

I have no problem with that.

"Attractors" are mathematics devices to describe the space in which
solutions to differential equations reside. In evolutionary terms I
guess it would just mean an 'upland' region on the fitness landscape,
a region toward which evolution must necessarily be drawn. Yes, it is
really just an appropriated buzz-word in this context but a reasonable
one in my opinion.

I think it mystifies a fitness peak to call it an "attractor". And I
still wonder if dkomo is talking about selective landscapes at all.
Remember, he doesn't think much of selection.


Big brains have evolved more than once. the entire vertebrate line
being one with mammals being the end of a progressive line (just
kidding!). The cephalopod molluscs, especially Octopus, are the
other. I would add dipteran and hymenopteran insects. Although not
large in human terms, they are pretty impressive on their own terms.

Depends on what you mean by "big". I took dkomo as talking about the
inevitability of human-level intelligence.

I don't know that dkomo is claiming that, but he can speak for himself. I am
a big advocate of necessity in evolution, but am not about to claim that
human-level intelligence is necessary.

Then we have no argument.

'Hairy' homeotherms developed twice, counting feathers as being
sufficiently hairlike to be a functional equivalent.

Organisms need a surface covering, and most of the possibilities appear to
have been evolved. I would also note that tunas and some sharks show some
degree of homeothermy, and bee colonies keep the internal temperature
fairly constant, so we are up to four for homeothermy.

Sure, if you expand the definition enough.

Now that's another question. If feathers are close enough to hair, we
can expand the number of convergences, but convergence becomes a looser
thing. One might suppose that given Amniota, some kind of homeotherm
became likely, and given a homeotherm, some kind of insulative covering
became likely. But in the first case, why only two? If what we have to
go on is the number of evolutionary events, two in 300 million years
isn't all that much.

To turn your argument on its head, all of the terrestrial homeotherms have
evolved insulating coverings. Obviously necessity and not chance.

I agree. In order to be a terrestrial homeotherm, you need either an
insulative covering or large size. Are you arguing with something I said?

Decent legs may have evolved only twice but something similar would be
needed for animal-like active locomotion in a terrestrial environment.
I would guess that the cephalopod molluscs would be capable of
developing some internal strengthening system, if not a real skeleton,
to enable terrestrial life in a different replay of the tape and then
they, too, would have 'legs'.


You may guess. But then again, legs appear to have developed, and only
twice, in aquatic animals. Nobody gained legs after becoming
terrestrial. So cephalopods have had as much opportunity as anyone.
Becoming terrestrial also seems to be an unlikely event, except perhaps
in arthropods.

Oh please. Do you really want to maintain that invading terrestrial
environments was unlikely?

Yes, for anything other than an arthropod. Why else has it happened so
few times outside of arthropods?

Do you really want to maintain that legs are an
unlikely method of locomotion?

If they're not unlikely, why have they only happened twice in all of
earth history?

My guess, based on absolutely zero evidence, is that if you replayed
the tape from the start, we would have ecologically similar beasties
wandering around including something like the size distribution we now
see but nothing at all like the present set.

That would be my guess, depending on what "replay the tape" means and
what "start" means. In the first case, a tape is a poor metaphor, since
nothing actually changes when you replay a tape. So what is allowed to
vary? But never mind that. In the second case, I'm willing to accept
your notion if we rerun, as Gould did, from the Middle Cambrian. But my
guess is that the majority runs of the tape "from the beginning", if we
mean the start of the earth, would leave you with a world of
single-celled organisms. That is, life is likely, but multicellularity
is fairly unlikely, as are several of the steps that lead to big fierce
animals.

I agree that multicellularity, along with abiogenesis, is not well
explained at present, and may well be a matter of chance rather than
necessity.

You say that as if it exhausts the possibilities. Contingency is a nice
concept. Selection provides the necessity, but the selective regime in
which a given feature is promoted may be a rare and chancy thing.

As to the steps that lead to big fierce animals, you probably
have to look well before the Middle Cambrian in terms of the roots of
diversity in multicellular animals, at least based on what Dawkins has to
say in Ancestor's Tale.

As you will note, I was starting with the origin of life.

Still, some sort of
digestive, respiratory, circulatory, skeleto-muscular, regulatory, and
reproductive systems would be seen as would some sort of ability to
sense and respond to the environment including the ability to alter
behavior based on past experience and even to anticipate the
consequences of behavior and work out alternative strategies without
actually having to execute them.

True from a Eumetazoan-centric perspective, at least.


Some sense of one's own identity as
an actor in this process would also be something I would expect in the
most independently active organisms. You might have to wait rather a
long time for all this, but I think it would eventually happen.

Who knows? I have no idea how many times "a sense of identity" has
evolved. I'm not even sure how to tell if one exists.

My main attempted point was that we do have some data that at least
suggests we can tell Gould's and Conway Morris's ideas apart, and that
if we look at that data, Gould seems closer to being correct than Conway
Morris.

Given that I think Hymenoptera have big brains, you can guess what my
definition of 'big' is.

Hear, hear. As I understand it, hymenoptera brains are similar in complexity
to the computer I am using to type this on, which would be capable of
extraordinary feats of legerdemain if it's programming had undergone
sufficient selection pressure.

You can define anything however you like. It's not clear to me, given a
definition of "big brain" that includes Hymenoptera, how many times big
brains have evolved independently. The criterion is at present not well
defined.

I don't think counting evolutionary instances the way you do is
reasonable. One instance followed by massive radiation will fill
niche space pretty well impeding other attempts. The question is: had
insects not occupied terrestrial habitats, would something else come
along to do the job? In the end, you still would have only a small
number of such developments. The question is only who can come up
with that development first or most successfully.

And, yes, much as I dislike Gould's style, in this case he was right.

No, as I said above, in this case we can't say if he was right, since our
sample size is one.

In the most naively literal sense of an actual "rewind of the tape",
sure. But we have a great many cases of convergence and missed
opportunities for convergence in earth history, and that's a sample size
considerably greater than one.

.

Relevant Pages

  • Re: Evolution is NOT random
    ... that planet will be different. ... The species of life will be different because that planet and its sun ... evolution is random. ... Meteors striking the earth are part of the tape of life. ...
    (sci.bio.evolution)
  • Re: Species diversity through time
    ... you think that the history of life is indeed a series of inevitable ... Replay the tape of life and the end result will be close to ... evolution in a more modern way. ... so we are up to four for homeothermy. ...
    (talk.origins)
  • Re: Species diversity through time
    ... you think that the history of life is indeed a series of inevitable ... Replay the tape of life and the end result will be close to ... evolution in a more modern way. ... that makes them evolve that character so much. ...
    (talk.origins)
  • Re: Species diversity through time
    ... that life is not some grand progression toward some predetermined goal. ... Replay the tape of life and the end result will be close to what ... evolution in a more modern way. ... Large brains with thinking minds are attractors. ...
    (talk.origins)
  • Christof Koch
    ... Simpler Origin for Life" from earlier this year by Prof. Robert Shapiro, ... Life began with the appearance of the first RNA molecule. ... before proteins and DNA in the evolution of life. ...
    (talk.origins)
Loading