WE>The topic was discussed on talk.origins back in 1995, with Walter
WE>ReMine himself pressing the difficulties of Haldane's dilemma.
WE>Andy Peters and Chris Colby asserted that Haldane's dilemma only
WE>arose with a "hard selection" scenario; under "soft selection"
WE>it just wasn't the problem that ReMine claimed. ReMine, of course,
WE>remained unconvinced.
WE>ReMine had also asserted that Haldane's dilemma could be seen in
WE>the operation of evolution simulations. Upon this point, I
WE>challenged ReMine to name the simulation which showed the difficulties
WE>of HD, to specify how such difficulties were recognized or quantified,
WE>and basically to have the whole issue put to empirical test. I offered
WE>to modify an existing program to use the "soft selection" described
WE>by Peters and Colby, and then we would see if the problems went
WE>away as Peters and Colby asserted, or remained as ReMine asserted.
WE>ReMine did not produce the name of the simulation that he had claimed
WE>showed the problem, nor any indication of how one recognized the
WE>problem in a simulation. It seems to me that he missed a great
WE>opportunity to try out his idea.
[...]
WE>I had checked out "The Biotic Message" once before, but I did not
WE>have it in hand during the talk.origins discussion. If the "weasel"
WE>simulation is really the one that Walter meant, then that should have
WE>been dead easy for him to name. I suspect it was the call for a
WE>definite and programmable criterion for detection of a "problem"
WE>due to Haldane's dilemma that caused the silence.
SJ>I would have thought it would be fairly obvious that "the `weasel'
SJ>simulation is really the one that Walter meant" since it is the one
SJ>that appears in his "Haldane's Dilemma" chapter.
If I had had his book in hand, probably so.
WE>I counted up to eight times that I asked for the name of the
WE>simulation and other associated support, and then presented
WE>it a few more times that I didn't count. The only response
WE>from Walter was "Hold your horses."
SJ>Presumably Walter thought you meant something over and above what he
SJ>wrote in "The Biotic Message"?
Well, according to this reply, he was just ignoring me.
[Walter's recall snipped]
WR>Many questions fly on talk.origins. My session did not get to
WR>Elsberry's particular one for several reasons: (1) His question
WR>was off topic, and on talk.o maintaining a focus is half your
WR>task. I was discussing the theoretical genetic issues of
WR>Haldane's Dilemma, but computer simulations raise a new range
WR>of issues. Even the theoretical issues bifurcate in lots of
WR>directions, and I was busy posting, night and day, to address
WR>those and keep the discussion focused. My talk.origins
WR>material ran for over a hundred pages. Elsberry has nothing to
WR>complain about.
The original set of claims which Walter made concerning Haldane's
dilemma specifically brought up computer simulation as an
indicator that the problem was "robust and firm". I fail to see
how a claim of "off-topic" discussion can be made when Walter
himself introduced the topic.
WRM> 4) The problem is robust and firm -- the phenomenon can even be
WRM> demonstrated in computer simulations, such as the same one Dawkins
WRM> used in his book _The Blind Watchmaker_.
[...]
WRM> Haldane's Dilemma is fundamentally simple. Anyone can understand it.
WRM> Anyone with a pencil can calculate it and see. Computer simulations
WRM> clearly demonstrate the problem. So evolutionists cannot claim they
WRM> were unaware. [...]
My challenge to ReMine...
If Walter doesn't want us to consider those claims to be abandoned,
he should come across with the following bits of information
immediately.
1) The name and reference for at least one of the already
existing computer simulations that he claimed demonstrated
Haldane's paradox.
(Walter has now responded with the answer to 1.)
2) The diagnostic means by which one determines that the
simulation identified in (1) actually does demonstrate
the problem.
(This is still missing.)
WR>(2) I will not re-type my book into the
WR>internet. I will not spoon feed it to each person who has a
WR>question. Elsberry knew where to get the answer to his
WR>question, he simply failed to pursue it.
I find this response to be odd, since Walter often told people
to buy or read his book for answers to questions, but did *not*
indicate that in response to me. His only response at the
time was to "hold your horses", with an indication that he would
eventually get around to making a reply. Such was not
forthcoming.
Does Walter's discussion of computer simulations in "The Biotic
Message" include a programmable description of recognizing or
quantifying a "problem" due to Haldane's Dilemma? If so, I'll
fire up Interlibarary Loan again.
WR>(3) My talk.origins
WR>session never intended to supply all the answers from my book,
WR>instead my intention, which I announced, was to serve notice
WR>about my book and its issues. My posts argued sufficiently
WR>well to establish them as live issues worthy of further
WR>consideration. My goal was to raise legitimate interest in my
WR>book. Since Elsberry continually blustered that he was
WR>interested, there was nothing further for me to do. The ball
WR>was entirely in his court.
WR>In other words, Elsberry's complaining is typical talk.o antics and
WR>posturing. Anything but deal with my book and the issues it raises.
Walter seems to have forgotten that I offered to do more than
simply read sources.
[Quote]
Produce the computer simulation that you claimed above
already exists. Show us how you determine that cost of
selection is an issue. The programmers on t.o. will then
try applying the concepts that have been forwarded by Andy
Peters and Chris Colby. We'll then see whether the
program shows no change (your claim) or changed behavior
(Andy and Chris's claim).
[End quote]
[Quote]
*** ReMine's Dilemma ***
This is Remine's Dilemma: To accept his own challenge means
that he accepts the significant chance of being demonstrated
to be wrong in a way not easily amenable to simple dismissal,
as has worked previously. To not accept the challenge means
that he would be, in effect, abandoning his prior claim that
the problem of Haldane's paradox was so pervasive that it
caused difficulties in computer simulations of evolutionary
processes. From a cost/benefit standpoint, if I were Walter,
I would quietly ignore the challenge, figuring that the claim
concerning computer simulation is not such a strong support
that losing it costs much, whereas pursuing that claim could
jeopardize the whole argument. Ignoring it also leaves open
the possibility that some will not recognize that the original
claim concerning simulation has, in fact, been abandoned.
Perhaps not surprisingly, this is what Walter has thus far
done.
[End quote]
[...]
WR>2) According to Motoo Kimura's theory of neutral evolution --
WR>Neutral evolution (during the same time as above) could
WR>substitute no more than 25,000 *expressed* neutral mutations.
WR>That amounts to 0.0007 percent of the human genome.
"Extremely interesting from almost every point of view." -- BotR
That's if one uses total base pairs to work the numbers, and it
definitely does not square with the emphasized "expressed"
statement. If one works from the number of loci, then Walter's
number is much too small, since changes in 25,000 loci would
mean about 1/4th of all loci could be affected, a significant
amount by any standard. The 1/4th figure does *not* represent
an upper limit, BTW. Not all mutations are point mutations.
Let's work it out with these assumptions:
1) 100,000 expressed loci in the human genome
2) Every mutation is a point mutation
3) 25,000 point mutations are distributed over the expressed loci
4) The chance of occurrence in any locus is equiprobable
This short Perl script gives us a numerical approximation to the
number of loci likely to be affected.
srand; # Seed random number generator
for ($jj=0; $jj < 100; $jj++) { # Repeat 100 times, Monte Carlo
for ($ii=0; $ii < 100000; $ii++) { # Clear the loci
$loci[$ii] = 0;
}
for ($ii=0; $ii < 25000; $ii++) { # Add the point mutations to the loci
$loci[int(100000 * rand)]++;
}
$cnt=0;
for ($ii=0; $ii < 100000; $ii++) { # Count the loci with mutations
if (0 < $loci[$ii]) {
$cnt++;
}
}
print ("Loci count = $cnt\n");
$sum += $cnt; # Sum up the number of loci with mutations
}
$ave = $sum / 100; # Find the average number of loci with mutations
print ("Ave. = $ave\n");
The result?
Ave. = 22125.889999999999418
The average number of loci affected is just over 22,000. That
means that even taking into account multiple mutations at the
same locus, over 1/5th of all loci will be likely to have
changed via neutral mutation, given the stated assumptions.
Some numbers from the program run:
Loci count = 22052
Loci count = 22131
Loci count = 21986
Loci count = 22110
Loci count = 22072
Loci count = 22096
Loci count = 22159
Loci count = 22178
Loci count = 22098
Loci count = 22142
Loci count = 22199
Loci count = 22162
Loci count = 22177
Loci count = 22082
Loci count = 22061
Loci count = 22164
Loci count = 22128
Loci count = 22196
Loci count = 22139
Loci count = 22142
Loci count = 22041
Loci count = 22133
Loci count = 22076
Loci count = 22159
Loci count = 22054
Loci count = 22202
Loci count = 22070
Loci count = 22105
[...]
WR>4) The phenomena of error catastrophe (from harmful mutations),
WR>and low substitution rates (of beneficial mutations), show up
WR>even on computer simulations (if not arbitrarily prevented by
WR>the programmer), and give support to my claims above.
I'm still willing to try putting Walter's claims about computer
simulation to the test. Now that I have the name of the
program, all that's needed is Walter's description of the
diagnostic function that indicates the presence or severity of
a problem. Of the two, "severity" would be much better, since
that would imply a quantification rather than mere
categorization.
Wesley