Hello, Stephen, I was beginning to think you'd forgottten about this old
thread. ;-)
>>>RW>With regard to your specific question about the creative power of
random
>>>>mutation and natural selection, the most impressive demonstration of
this
>>>>that I've seen is Nilsson and Pelger's computer simulation of the
>>>evolution of an eye.
>
>>SJ>...But there is however an immediately problem. If random mutation and
>>>*positive* natural selection is as pervasive in nature as it would have
to
>>>be if Neo-Darwinism is true, why is there a need to rely for one's best
>>>evidence of it on a computer *simulation*? Why not just go out and
>>>document it directly from nature?
>
>RW>I think that's rather obvious. Evolution in nature of a major structure
like
>>the eye takes hundreds of thousands of years.
>
>I did not say that Richard had to watch it happen! If the eye evolved by
>step-by-step random mutation and natural selection, as Nilsson & Pelger's
>simulation purports to show, then it should easily be able to be documented
>in nature. Where is that documentation in the *real* world?
Your conclusion "it should easily be able to be documented in nature"
doesn't follow from your premise "if the eye evolved by step-by-step random
mutation and natural selection".
Suppose, for the sake of argument, that the eye of some species is currently
in the process of evolving. Let's say we observe it for 100 years and,
during that time it changes by 0.1%. The difference may not even be
measurable, given that the variation within a population is probably many
times greater than that. Even if it can be measured, anti-evolutionists
would just say "that's only microevolution".
>>SJ>There are a number of other problems with the biological realism of
>>>Nilsson and Pelger's eye simulation, in their paper (Nilsson D.E. &
Pelger
>>>S., "A pessimistic estimate of the time required for an eye to evolve",
>>>Proceedings of the Royal Society B, 1994, v.256, pp.53-58) which I may
>>>deal with in a future post, time permitting. But the major problem that
is
>>>immediately evident is that it is a self-evident truism and therefore
>>>scientifically uninteresting.
>
>RW>I assume the "self-evident truism" that you're referring is the fact
that
>>complex structures can be created by an algorithmic process of random
>>variation and cumulative undirected selection.
>
>No. I was referring to the fact that to reverse engineer the eye unto
>1829 steps, as Nilsson D.E. & Pelger did, and then to run it through a
>computer algorithm which re-assembled those 1829 steps back into the eye
>again, is "a self-evident truism and therefore scientifically
>uninteresting".
Let me get this straight. Are you saying that they calculated all their 1829
steps, working back from the finished eye, and then presented those as the
steps produced by their algorithm?
If they did that, I would consider it an outrageous fraud. But I don't
believe it. You'll have to show me some very convincing evidence before I'll
accept it.
>The problem is that all these computer simulations of what purports to
>be evolution by random mutation and natural selection are *designed*
>unrealistically to *never fail*:
>
>"Yet the simulation not only succeeds in accomplishing its task, but it
>does so very rapidly and consistently It never fails, NEVER. The
>computer simulations use many unrealistic assumptions that favor
>evolution. First, the simulations assume away everything that could
>prevent evolution. They do not allow extinction, which normally would
>terminate all further evolution. They do not allow error catastrophe,
>which normally would cause a degradation away from any target
>sequence-no matter how severe the selection. They do not allow
>canyons and hills in the fitness terrain, which normally would prevent
>evolution. In short, they assume naive natural selection -that evolution
>is upward, ever upward. Having artificially disallowed all possible
>failure modes, it is not surprising the evolution simulations work.
>(ReMine W.J., "The Biotic Message," 1993, p232)
Of course these simulations are an enormous simplification, and I don't
suppose N & P claim otherwise.
The only claim I'm making about it is that it shows the power of an
algorithmic process (like random mutation and natural selection) to produce
complexity.
>If the program didn't work the first time (and no programs ever do),
>the programmer would redesign the program until it did what it was
>designed to do, show how random mutation and natural selection can
>build complex designs easily and quickly:
Even if that's true, it doesn't change the fact that it shows there *are*
undirected algorithms which produce complexity. But I'm interested to know,
how many times did they redesign the program? What results did they get the
other times?
>"How fast will the evolution happen? Here the simulations use
>additional assumptions to artificially increase the speed. Dawkins'
>simulation gives many unrealistic advantages to evolution:
>* The simulation disallows recessive mutations, ordinary epistasis,
>polygenic and pleiotropic effects-which would normally increase
>the cost of substitution and slow down evolution.
>* The simulation assumes perfect selection (s=1), which is not
>typical of nature. Evolutionists ordinarily acknowledge that the
>typical beneficial mutation has less than a one percent advantage
>(s<0.01).
>* The simulation assumes extreme truncation selection, which is
>not like nature.48 Nature has no means to count mutations, rank
>the population by the count, and keep only the highest ranking
>individual. Truncation selection gives an unrealistic advantage to
>evolution by substantially lowering the cost of mutation.
>* The simulation disallows any costs due to random death,
>balancing selection, heterosis, or segregation.
>* The simulation assumes a high probability of beneficial mutation.
>It assumes that a mutation has one chance in 27 of being so
>beneficial that the rest of the population is completely wiped out.
>This high rate of beneficial mutation reduces the cost of harmful
>mutation and increases the speed of evolution
>* Dawkins did not say in his book, but his simulation must have
>assumed a high reproduction rate (N=100 or higher) This is higher
>than real species can produce For a sexual species to accomplish
>this, the females must give at least 200 progeny each.
>* Dawkins did not say, but he must have chosen the mutation rate
>to optimize the speed of evolution. If he had chosen a low mutation
>rate, (such as 10^-8 as in humans) then the simulation would
>require roughly 50 million generations. On the other hand, if he had
>chosen too high a mutation rate, then it would cause error
>catastrophe and the target phrase would never be reached.
>Dawkins picked the mutation rate that produced the fastest
>evolution.
>Dawkins' readers got the impression he casually threw the
>computer simulation together and speedy evolution just happened
>automatically. In reality, Dawkins carefully designed his simulation
>to favor rapid evolution. One could hardly design a simple, easily
>understood simulation that is faster. His computer simulation aids
>the illusion that evolution is simple in concept, inevitable, and fast."
>(ReMine W.J., "The Biotic Message," 1993, p.233)
I thought we were talking about P & E's simulation, not Dawkins's. Anyway,
I'm not qualified to judge how realistic their assumptions were. What
impressed me most was that the simulation did produce an eye-like structure,
not the estimated time that they calculated.
>RW>I think the phrase "self-evident truism" may be overstating the case,
but I
>>would agree that it's hardly a surprising result. Isn't it amazing, then,
>>that so many anti-evolutionists refuse to accept it? And whether their
estimated time is reasonable or not, they still demonstrated that there
*are* undirected algorithms which produce complexity.
>By "self-evident truism" I was referring to Nilsson & Pelger's
>simulation.
>
>And I note Richard's attempt to use words, ie. "*refuse* to accept it"
>(my emphasis) to make out that those who are unconvinced that
>Nilsson & Pelger's simulation is a realistic model of what happened
>in the real word, are being unreasonable.
>
>But it is not only "anti-evolutionists" who are not convinced by it.
>The well-known anthropologist Jeffrey Schwartz, in his book
>"Sudden Origins" does not accept it either:
>
>"More recently, the sociobiologist and promoter of the notion of
>"selfish genes," Richard Dawkins, has championed a proposal that
>estimated how many generations it would take to go from having
>no eye almost to having the kind of eye that vertebrates and
>octopuses have.
>
>The scenario that Dawkins embraced had been formulated by two
>biologists, Dan Nilsson and Susanne Pelger. They began with their
>hypothetical eye being already in a formative stage that consisted
>of a patch of light-receptive cells at the skin's surface that was
>sandwiched between a transparent protective layer of cells above
>and a layer of darkly pigmented cells below By way of
>mathematical modeling, Nilsson and Pelger calculated,
>conservatively, that it would take only 400,000 generations for this
>non-eye region of skin to be transformed into an organ of sight.
>Even when only a 1 percent change per generation is invoked over
>what is really a relatively short period of geologic time, Nilsson and
>Pelger predicted that selection would be able to cause the skin to
>invaginate, bringing the presumptive retinal layer down with it, and
>then to fill with fluid of a very low refractive index Still no functional
>eye, but then a lens begins to emerge and eventually it achieves a
>refractive index sufficient to provide sight. The maximum number of
>tiny steps required to go from the flat to the invaginated structure
>was estimated at 1,033. It took only 529 more steps to make a lens
>and put the eye into its final, semi-flattened shape.
>
>Absent from this simulation was consideration of the origin of the
>patch of layered cells in the right place, the development of a
>variable iris and controlled focusing, the creation of the nervous
>optic chiasma that is the region in which the optic nerves cross
>over, and the innervation of the eyeballs by the optic nerves, which
>constitute one of the twelve pairs of primary nerves that emanate
>from the brain itself. Curiously lacking, as well, was any discussion
>by Dawkins of the selection pressure that would have set the
>process in motion and of the selective advantage members of more
>than 399,000 generations of their species would have enjoyed as
>they served as conduits for this ever-invaginating, liquid-filled pair
>of pockets in their head region. But, once the process had taken off
>on this trajectory, there was, as Dawkins saw it, no turning back.
>For, in his view of evolution, "[u]nlike human designers, natural
>selection can't go downhill, not even if there is a tempting higher
>hill on the other side of the valley." It is, however, one thing to
>model how such changes might have occurred seamlessly and
>gradually, and another to have a basis for doing so.
>
>(Schwartz J.H., "Sudden Origins: Fossils, Genes, and the
>Emergence of Species," John Wiley & Sons: New York NY, 1999,
>pp.361-362)
Schwartz, like you, seems to be under the impression that, to be of any use,
a simulation must simulate *everything*. I accept that P&E's simulation has
it's limits.
>RW>That the pessimistic estimate of the time required for an eye to evolve
>>turned out to be fairly low (less than half a million years) is certainly
>>not a self-evident truism, but an interesting discovery.
>
>It would only be "an interesting discovery" if it faithfully modelled
>biological reality. Where is the evidence from the *real* world that the
>vertebrate camera eye evolved by random mutation and natural selection in
>"less than half a million years"?
That's another question. I was just giving you the single most impressive
(to me) piece of evidence of the power of natural selection to produce
complexity.
>>SJ>That is, given that the existing vertebrate camera eye is: a) comprised
of
>>>three layers of cells; and b) optimal for sensing light; then it follows
>>>deductively that one can, at least *in theory*, reverse engineer it on a
>>>computer by breaking it down into a series of small stages such that are
>>>each lower stage is slightly less optimal for light-sensing than the next
>>>higher one in the series, and stopping the series when the model starts
to
>>>break down (ie. when one has to explain where the three layers of cells
>>>came from). Then one simply re-runs the program in reverse, and hey-
>>>presto, the camera eye that one started with, arises of *necessity*!
>
>RW>What does this mean? How can you run a computer program in reverse?
>
>I meant to run the *steps* in reverse.
>
>>SJ>But whether this exercise in *virtual* reality bears any more than a
>>>superficial resemblance to the *real* world of biology is another matter
>>>entirely.
>
>RW>True, but you only asked me for evidence of the power of random mutation
and
>>natural selection. That's what I gave you.
>
>What Richard gave was a computer *simulation* of random mutation and
>natural selection". No one denies that one can program a computer to
>appear to *simulate* random mutation and natural selection. Dawkins had
>already done this with his METHINKS IT IS LIKE A WEASEL program.
>But few biologists were impressed.
P&E's simulation is significantly different from the weasel program -- it
doesn't have a built-in goal.
>What Richard needs to do is provide "evidence of the power of random
mutation
>and natural selection" to build the *real* vertebrate camera eye (not a
grossly
>oversimplified model of it), in the *real* world of biology.
I mentioned one piece of evidence. I'm not going to give you a whole list. I
expect other people have mentioned other items of evidence to you in the
past, and they were probably much better qualified to talk about them than I
am.
>>SJ>That is probably why this `evidence' has never, AFAIK, been taken up
and
>>>published by leading scientific journals, like NATURE and SCIENCE, but
>>>has remained buried in a relatively obscure like the "Proceedings of the
>>>Royal Society", and is only used by Darwinist propagandisers like
>>>Dawkins.
>
>RW>Well, I'm not familiar with the Proceedings so I don't know whether
they're
>>obscure, but I rather doubt it. And I assume that articles in it are
>>peer-reviewed (correct me if I'm wrong).The Royal Society is a prestigious
>>organization, our nearest equivalent to your National Academy of Sciences.
>>If a paper has already been published there, why should some other journal
>>republish it?
>
>I am interested that Richard regards this simulation by Nilsson and Pelger
>as his best "evidence of the power of random mutation and natural
>selection" yet he has apparently not even read the article?
Do you have to read the original paper to know anything about a scientific
study? I've seen the results summarized. I take the time calculation with a
pinch of salt, because I know it's based on very broad estimates. But I'm
impressed by the fact that they did get an eye-like structure.
>And if Nilsson and Pelger's simulation was among the best evidence for
>the creative power of random mutation and natural selection, one would
>expect that at least one of the leading journals like SCIENCE or
>NATURE (or even a lesser one like New Scientist or Scientific
>American) would have republished it. That does not seem to have
>been the case?
I said this is the evidence which most impressed *me*. I can't speak for
anyone else.
And are these journals in the habit of republishing papers from other
journals? I don't know, but I suspect not.
Richard Wein (Tich)
See my web pages for various games at http://homepages.primex.co.uk/~tich/
This archive was generated by hypermail 2b29 : Wed Mar 22 2000 - 20:16:03 EST