From a post from Fall, 95:
An interesting drama has been playing itself out on talk.origins over the
last few months. A serious geneticist, Walter Remine, has demonstrated
a problem which utterly destroys the theory of evolution in anything
remotely like any of its present forms. This problem was first
described by the noted geneticist J.B.S. Haldane in the late fifties,
and is called Haldane's dilemma. The logical implications of the
problem are that even the gigantic time frames normally used by
evolutionists as an enabling mechanism for evolution, the 4 - 7 billion
year age of the Earth which you are familiar with, is nowhere even
remotely close to being enough.
The talk.origins crew has tried to use all of its normal tricks on
Remine, with the end result that they, Darwin, Darwinism, and all other
modern encarnations of evolution have been made to look very silly
indeed.
In ten years of watching and participating (as arch villian) in
talk.origins, this is the very worst beating I have ever seen these
people absorb. There is no way that anybody with any talent or brains
could have been reading all of this and still believe in evolution.
The talk.origins crew surely would prefer that viewers in other forums
did not see any of this material; naturally, I've been collecting as
much of it as possible, and hereby provide you with a sampler:
The action began with a couple of posts of mine and Robert Bass' on
talk.origins, and Remine's on other forums.
.............................................................
My post:
A number of the big theories in science are on the ropes today,
evolution and the big-bang theory being the most notable.
There are numerous problems with evolution, any one of which, were the
ideological stakes not so high, should have long since laid evolution to
rest. Aside from the impossible problem of abiogenesis, the problem of
catastrophism destroying the time-scales normally given out, or the
programmatic problems (such as with flying birds), you have the problem
of trying to get from something resembling an ape to modern man in ten
million years.
Walter ReMine <wjremine@mmm.com> notes:
>Evolutionary genetics has trade secrets too. The major one is Haldane's
>Dilemma, a problem discovered in the 1950s by the famous evolutionary
>geneticist, J.B.S. Haldane. Journals discussed it through the 60s, and
>ignored it thereafter. Evolutionists never publicly solved it, rather they
>brushed it aside. Here are my claims:
>1) Haldane's Dilemma is invisible in evolutionary genetics textbooks today,
>you will be lucky to find information on the problem. The little available
>information is cryptic and opaque, even to serious students.
>2) The standard model of evolutionary genetics -- prominently displayed in
>every evolution textbook -- is massively inadequate to solve the problem.
>Yet evolutionists continue to sell that model because it makes evolution
>seem easy and inevitable.
>3) Even if I arm you with information about the problem, you will find
>precious little in evolutionary textbooks that *might* be taken as a
>plausible solution.
>4) The problem is robust and firm -- the phenomenon can even be
>demonstrated in computer simulations, such as the same one Dawkins used in
>his book _The Blind Watchmaker_.
>In short, Haldane's Dilemma is a thorough trade secret of evolutionary
>geneticists.
>My book, _The Biotic Message_, has two chapters (and an appendix) detailing
>Haldane's Dilemma and rebuffing the many attempts to solve it. Here I'll
>draw from that material to describe the problem, and bring you up to speed.
>Then I'll answer your questions, and perhaps eventually we'll have our usual
>rip-snortin' debate. I'll keep my descriptions short and easy reading.
>Along the relevant primate line, our supposed pre-human ancestors had an
>effective generation time of 20 years. (I quote sources and details in my
>book, so I'll spare you here.) Imagine ten million years ago -- (that is
>two to three times the age of the alleged chimp-human split) -- that's
>enough time for 500,000 generations of our presumed ancestors.
>Imagine a population of 100,000 of those organisms quietly evolving their
>way to humanity. For easy visualization, I'll have you imagine a scenario
>that favors rapid evolution. Imagine evolution happens like this. Every
>generation, one male and one female receive a beneficial mutation so
>advantageous that the 999,998 others die off immediately, and the population
>is then replenished in one generation by the surviving couple. Imagine
>evolution happens like this, generation after generation, for ten million
>years. How many beneficial mutations could be substituted at this crashing
>pace? One per generation -- or 500,000 nucleotides. That's 0.014 percent
>of the genome. (That is a minuscule fraction of the 2 to 3 percent that
>separates us from chimpanzees).
>That's not a difficult calculation, yet it immediately reveals a problem.
>Is 500,000 beneficial nucleotides enough to explain the origin of humanity
>from some chimp-like ancestor?
>The problem gets worse. The scenario favored evolution in wildly
>unrealistic ways. I could name several, but one is simple: There is no
>possible way for a female primate to produce 100,000 offspring each
>generation!!! Here's the lesson:
> Evolution requires the substitution of old prevalent traits
> with new rare traits. But the substitution rate is limited by
> the species' reproductive capacity. If an evolutionary scenario
> requires an implausibly high level of reproductive capacity,
> then the scenario is not plausible.
>Haldane saw this problem and posed it within the framework of mathematical
>population genetics. We will discuss his calculations later, but his
>conclusion was easy to understand. He calculated that the higher
>vertebrates (such as mammals) have only enough reproductive capacity to
>sustain an average rate of 300 generations per substitution. The literature
>seldom states the figure, but when it does, that is the only one offered.
>Haldane's Dilemma is glaringly plain. Take the population we discussed
>above. In ten million years, it could substitute 1,667 beneficial
>nucleotides. That is less than 50 millionths of one percent of the genome.
>(And that is *before* we make deductions. For example, Gould says species
>typically spend *at least* 90% of their time in stasis, where little or no
>evolution occurs. There are other deductions we'll discuss later, but
>together they reduce the figure far below 1,667.) Is that enough to explain
>the origin of upright posture, speech, language, and appreciation of music,
>to name just a few of our uniquely human capacities? Is 1,667 beneficial
>nucleotides enough to make a sapien out of a simian?
>Haldane's Dilemma is fundamentally simple. Anyone can understand it.
>Anyone with a pencil can calculate it and see. Computer simulations clearly
>demonstrate the problem. So evolutionists cannot claim they were unaware.
>Nonetheless they were cryptic, effectively concealing the problem for nearly
>forty years. Few people have heard of it, and evolutionary geneticists
>offer no unified coherent solution. Haldane's Dilemma is a major scandal.
What Haldane is saying, in effect, is that starting from any ape-like or
<proto human> creature ten million years ago and using assumptions which
favor evolutionary change in the direction of modern man in more ways than
the evolutionists' fairy godmother could imagine, by 1995 you'd be lucky to
get to an ape with a slightly shorter tail.
The fact that anybody could try to claim, as they apparently do, that Haldane's
assumptions do not favor evolution enough, is an indication of the esxtent to
which these people (evolutionists) have succeeded in brainwashing themselves.
Let's look at one assumption which is implicit in just about everything
you see on evolution and which Haldane is also making: the idea that
while unfavorable mutations are known to occur far more often than
beneficial ones, it is ALWAYS the beneficial mutation which survives and
propogates through a population, just as we know from experience that it
is always the good which drives out the bad in government work,
academia, politics...
You only have to look at this a little bit to realize how stupid it
really is.
You are starting out with apes ten million years ago, in a world of fang
and claw with 1000+ lb. carnivores running amok all over the place, and
trying to evolve your way towards a more refined creature in modern man.
What's wrong with that?
Let's start from about ten million years back and assume we have our ape
ancestor, and two platonic ideals towards which this ape ancestor (call
him "Oop") can evolve: One is a sort of a composite of Mozart,
Beethoven, Thomas Jefferson, Shakespeare, i.e. your archetypal dead
white man, and the other platonic ideal, or evolutionary target, is
going to be a sort of an "apier" ape, fuzzier, smellier, meaner, bigger
Johnson, smaller brain, chews tobacco, drinks, gambles, gets into knife
fights...
Further, let's be generous and assume that for every one chance
mutation which is beneficial and leads towards the gentleman, you only
have 1000 adverse mutations which lead towards the other guy. None of
these mutations are going to be instantly fatal or anything like that at
all; Darwinism posits change by insensible degree, hence all of these
1000 guys are fully functional.
The assumption which is being made is that these 1000 guys (with the bad
mutation) are going to get together and decide something like:
"Hey, you know, the more I look at this thing, we're really
messed-up, so what we need to do is to all get on our motorcycles and
pack all our ole-ladies over to Dr. Jeckyll over there (the guy with
the beneficial mutation), and try to arrange for the next generation of
our kids to be in better genetic shape than we are..."
Now, it would be amazing enough if that were ever to happen once;
Darwinism, however, requires that this happen EVERY GENERATION from Oop
to us and, as far as I can tell, Haldane is assuming this also.
I haven't even mentioned the gambler's problem, which says that in
each generation the guy with the good mutation is starting out with one
dollar, the biker gang with 1000 dollars, and the house, i.e. the general
population, with $100,000 or something like that.
And yet, Colby and others claim that Haldane is making unreasonable
assumptions.
....................................................................
Phil Nichols responded with the usual talk.origins technique, calling ReMine
a liar and me an idiot:
Path: news3.digex.net!news2.digex.net!howland.reston.ans.net!news.sprintlink.net!uunet!in1.uu.net!library.erc.clarkson.edu!ub!newserve!rebecca!pn8886
From: pn8886@csc.albany.edu (Phil Nicholls)
Newsgroups: talk.origins
Subject: Re: Haldane's dilemma: the real story
Date: 30 Mar 1995 20:06:30 GMT
Organization: University at Albany, SUNY
Lines: 168
Message-ID: <3lf306$8gj@rebecca.albany.edu>
References: <3l949u$n0j@rebecca.albany.edu> <3ld9ep$dft@ra.lib.ucalgary.ca>
NNTP-Posting-Host: thor.albany.edu
>> Phil
> Ted
>> Haldane himself pointed out that the selective deaths would
>> be less detrimental to the population if they came in
>> embryonic or juvenile stages.
> I haven't really had time to try to totally comprehend that
> paper yet, but it appears obvious that Haldane realized he had
> just wiped the theory of evolution, and was trying in a number
> of ways to soften the blow.
Please don't project TEDOLUTION onto Haldane's work,
especially before you have "totally comprehended" it.
Haldane's dilemma is a dilemma because work by Dobzhansky and
others in the 50's had provided experimental confirmation of
natural selection, race formation and speciation.
>Haldane's dilemma applies to very large populations.
> Given standard assumptions, the human population would
> have always been "very large" during the time under discussion.
Ted, population has a specific meaning in biology. The human
species (Homo sapiens) or any species, for that matter, is
made up of many populations or demes. A population is a group
of organisms living belonging to the same species living in
the same place at the same time. Some human populations today
are large, some are small and relatively isolated.
Through much of the pleistocene, "human" (i.e. hominid)
populations were small. Large population concentrations did
not occur until the neolithic and the invention of argiculture.
> Haldane appears to be saying that you'd have to get the population
> down near zero before you could have some new trait spread
> through the surviving population as it rebuilt itself, but that
> a genetic change which would drive the population that low would
> destroy it.
Actually, that is not what Haldane is saying at all. You are
casting Haldane after Velikovsky.
>> Under punctuated equilibrium Gould and Eldredge have proposed
>> that the rate of evolution is best described as long periods
>> of relative stasis punctuated by periods of more rapid
>> evolution. The bursts of evolution take place in SMALL
>> PERIPHERAL POPULATIONS.
> Velikovsky's version of this whole thing assumes a flood or
> some other worldwide catastrophy, all animal groups either
> being exterminated or coming to some very low population
> point, and then old and mutated species reforming populations
> as they expand into a world all theirs. The populations would
> quickly grow and spread and become diverse. Whether change
> could be caused by chance processes in such a scenario is
> still an open question (I strongly doubt it), but at least
> inbreeding in a small space over a long time does not become a
> factor.
There is no evidence of a worldwide catastrophy of this
magnitude since the K/T event of 65 million years ago.
> Gould and Eldridge appear to be claiming that inbreeding is
> the source of all genetic progress. There are real reasons
> why cousins are not allowed to marry eachother in most
> nations. You only have to think about that one a tiny bit to
> understand how silly it is.
As I pointed out earlier, punctuated equilibrium does not
involve "inbreeding."
>> Hence the cost of selection is
> Again, anything can happen in a fairy-tale.
Your standard response when you can't deal with the facts. It
is a FACT that smaller populations reduce selection cost.
That is a built in assumption in Haldane's models.
>> It is not that Haldane's assumptions "do not favor
>> evolution." Haldane's conclusions are based on a mathematical
>> model and as we all know all models make certain assumptions.
> The assumptions which ReMine made favored evolution in evume, as did Haldane, that this is a population of
diploid organisms. That means that if we have 100,000
organisms we are talking about 200,000 alleles. Since ReMine
says a new allele (singular) appears in two individuals by
mutation we must assume that the individuals are heterozygous
at this locus. If A is the new allele and a is the old
allele. The in two individuals we have a mutation from
a ----> A.
Let p be the frequency of a and q be the frequency of A,
2
q = --------- = 0.00001 therefore p = 0.99999
200,000
The initial frequency of the new allele is 0.00001. A very
very low frequency and if you will recall the lower the
frequency of q the higher the cost of selection.
Therefore this assumption, contrary to ReMine's claim, is not
the best possible case. See Crow and Kimura's book if you
want to confirm this.
Now, the cost of selection per generation can be calculated as
follows:
Maximum Fitness - Mean Fitness
C = ---------------------------------
Mean Fitness
This is one of two models that Haldane bases his model on.
Now ReMine tells is that the two individuals with the new
allele survive, all the rest are killed off. The fitness of
the individuals with the new allele is 1.0, the fitness of
the individuals with the old allele is 0.
Mean Fitness is then 2/100,000 = .00002
Hence in ReMine's model, where Maximum fitness is 1.0:
1 - .00002
C = -------------------
.00002
C = 4999!!
Now look at you the article by Haldane you will find that he
estimated the Cost of selection should be between 20 and 100 per
generation, with an average of 30. Not 4999
In other words, ReMine has used an example of ultra-hard
selection and this was chose specifically to maximize C.
He chose the smallest possible frequency of q, which also
maximized C.
You've been HAD, Ted. You've been lied to. He sounds
impressive, but when to take apart his argument you find
deception.
Doesn't that bother you?
-----------------------------------------------
Marked as a lo"Why post evidence, when I can simply post lies,
and avoid all de facts, unlike dem t.o. guys."
So Ted boot up, and he post de posts,
and he write dem stories, "how can I b.s. de most?"
his mind be fried, he spew dat pollution,
an' all he keep talkin' 'bout is REMINEOLUTION...
Apologies to David Iain Greig
--
"To ask a question you must first know most of the answer."
- Robert Sheckely
Phil Nicholls (pn8886@thor.albany.edu)
..........................................................................
The arrogance, along with David Iain Greig's obvious lack of talent, might be
tolerated if Nicholls knew what he was talking about...
I replied:
Phil Nicholls writes:
>> Velikovsky's version of this whole thing assumes a flood or
>> some other worldwide catastrophy, all animal groups either
>> being exterminated or coming to some very low population
>> point, and then old and mutated species reforming populations
>> as they expand into a world all theirs. The populations would
>> quickly grow and spread and become diverse. Whether change
>> could be caused by chance processes in such a scenario is
>> still an open question (I strongly doubt it), but at least
>> inbreeding in a small space over a long time does not become a
>> factor.
>There is no evidence of a worldwide catastrophy of this
>magnitude since the K/T event of 65 million years ago.
There is a mountain of such evidence, and you can view a tiny fraction
of it at:
http://access.digex.com/~medved/Catastrophism.html
>> Gould and Eldridge appear to be claiming that inbreeding is
>> the source of all genetic progress. There are real reasons
>> why cousins are not allowed to marry eachother in most
>> nations. You only have to think about that one a tiny bit to
>> understand how silly it is.
>As I pointed out earlier, punctuated equilibrium does not
>involve "inbreeding."
Sounds like you need to read up on punc/eek...
Inbreeding is normally taken to be small isolated groups of individuals
breeding only amongst themselves over a period of time which needn't
be more than a century or two to cause major problems, at least amongst
humans, and we are talking about proto-humans and humans here.
If that isn't what Gould and Eldridge are talking about, then what ARE
they talking about? The peripheral population undergoing a very
substantial change in twenty years and then taking over and
replacing the main population???
>diploid organisms. That means that if we have 100,000
>organisms we are talking about 200,000 alleles. Since ReMine
>says a new allele (singular) appears in two individuals by
>mutation we must assume that the individuals are heterozygous
>at this locus. If A is the new allele and a is the old
>allele. The in two individuals we have a mutation from
>a ----> A.
>Let p be the frequency of a and q be the frequency of A,
2
>q = --------- = 0.00001 therefore p = 0.99999
> 200,000
>The initial frequency of the new allele is 0.00001. A very
>very low frequency and if you will recall the lower the
>frequency of q the higher the cost of selection.
>Therefore this assumption, contrary to ReMine's claim, is not
>the best possible case. See Crow and Kimura's book if you
>want to confirm this.
>Now, the cost of selection per generation can be calculated as
>follows:
> Maximum Fitness - Mean Fitness
> C = ---------------------------------
> Mean Fitness
>This is one of two models that Haldane bases his model on.
>Now ReMine tells is that the two individuals with the new
>allele survive, all the rest are killed off. The fitness of
>the individuals with the new allele is 1.0, the fitness of
>the individuals with the old allele is 0.
>Mean Fitness is then 2/100,000 = .00002
>Hence in ReMine's model, where Maximum fitness is 1.0:
> 1 - .00002
> C = -------------------
> .00002
> C = 4999!!
>Now look at you the article by Haldane you will find that he
>estimated the Cost of selection should be between 20 and 100 per
>generation, with an average of 30. Not 4999
>In other words, ReMine has used an example of ultra-hard
>selection and this was chose specifically to maximize C.
>He chose the smallest possible frequency of q, which also
>maximized C.
.......................................................................
Right...
Let's look at what ReMine is saying:
>Imagine a population of 100,000 of those organisms quietly evolving their
>way to humanity. For easy visualization, I'll have you imagine a scenario
>that favors rapid evolution. Imagine evolution happens like this. Every
>generation, one male and one female receive a beneficial mutation so
>advantageous that the 999,998 others die off immediately, and the population
>is then replenished in one generation by the surviving couple. Imagine
>evolution happens like this, generation after generation, for ten million
>years. How many beneficial mutations could be substituted at this crashing
>pace? One per generation -- or 500,000 nucleotides. That's 0.014 percent
>of the genome. (That is a minuscule fraction of the 2 to 3 percent that
>separates us from chimpanzees).
So far, he hasn't said a word about genetics, Haldane, or anything like that
at all. What he IS talking about is the population-replacement-in-toto
fairy doing her thing every twenty years, and he notes that even THAT
is not nearly enough to get from anything like an ape to a human in
10 million years.
He goes on to note that:
> Evolution requires the substitution of old prevalent traits
> with new rare traits. But the substitution rate is limited by
> the species' reproductive capacity. If an evolutionary scenario
> requires an implausibly high level of reproductive capacity,
> then the scenario is not plausible.
>Haldane saw this problem and posed it within the framework of mathematical
>population genetics. We will discuss his calculations later, but his
>conclusion was easy to understand. He calculated that the higher
>vertebrates (such as mammals) have only enough reproductive capacity to
>sustain an average rate of 300 generations per substitution. The literature
>seldom states the figure, but when it does, that is the only one offered.
I.e. he notes that the instant any sort of a real-world estimate of times for
a change due to mutation to spread over a large population, things get even
more ridiculous and out of hand. All of that strikes me as pretty simple.
I have emailed ReMine a copy of your post and will get back to you if he
adds anything beyond what appears obvious to me, which is that you are assuming
that gentic principals somehow enter into the scenario involving the
population-substitution fairy.
Understanding where you are coming from is further made difficult by the
observation that your comments on hard selection appear to undermine any case
for ape->human in 10 million years being possible rather than support
any such case, as one would assume you intended to do.
Beyond that, ReMine has recently been quoted as follows:
"Note of clarification: That formula is not the definition of cost. The
cost of a given situation is defined as its genetic deaths per survivor.
(And that cost must be "paid" by the reproduction of the survivors, in
births per survivor.) The actual cost depends on the specific situation.
The above formula gives the theoretical absolute minimum cost of any single
substitution under any and all situations. It gives an unrealistically low
estimate for real world situations.
"Remember, Haldane-style cost arguments use the term "cost"
in a very specific way. Anything going from 'few' to 'many' incurs a
reproductive cost. Nothing can go from 'few' to 'many' without paying that
cost. The reproduction rate limits the ability to pay, and thereby limits
the speed of substitutions. It's quite mechanical and unavoidable. It even
shows up clearly in computer simulations.
"When you say, "A beneficial trait was substituted into the population" you
are simultaneously saying, "All the organisms without the trait died off
without heirs" which means "They had to be replenished by the reproduction
of the survivors." The fact that the trait happens to be beneficial is
virtually irrelevant. *Any* trait that is substituted into a population
incurs a cost of substitution.
What this means, amongst other things, is that the population-replacement-
fairy scenarios, far having a cost of 5000, are free; one second after
the fairy does her thing, the max and mean fitness are the same and the cost
of the entire transaction is zero. Or, if that isn't obvious, think of it
in the terms noted above by ReMine: there is no population loss (cost) to
make up.
>You've been HAD, Ted. You've been lied to. He sounds
>impressive, but when to take apart his argument you find
>deception.
>Doesn't that bother you?
The arrogance here is sort of like icing; the cake is totaly misunderstanding
the nature of the non-cost involved in the scenario in which the population-
replacement fairy picks up the tab.
.......................................................................
And then the heavy cavalry arrived:
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 19 Apr 1995 12:03:17 -0500
Message-ID: <3n3fol$a04@dawn.mmm.com>
*** Haldane's Dilemma -- continued ***
Andy Peters <adpeters@bio.indiana.edu> offers what he feels are
"by far the strongest arguments against" Haldane's Dilemma:
>(1) Soft selection and truncation selection radically reduce the
>demographic cost (and therefore the rate limitation) of selection.
Andy says "soft selection" and "truncation selection" are important
solutions to Haldane's Dilemma. Okay, then these topics should be
especially prominent and clear in the textbooks. They should be front
and center, unavoidable. Right? No student should be able to miss it,
right? There should be empirical support for these, right?
These should be linked, indelibly, unmistakably to the model of
evolutionary genetics, right? ***But you won't find that.***
At least, it will be so exceedingly rare you won't find it on your own.
The typical evolutionary genetics textbook doesn't discuss them --
you'll be lucky to find a few pages devoted to it -- the description
will be cryptic -- the empirical support will be completely absent --
and they will be disjointed and disconnected, with little linking
them to a coherent model of evolutionary genetics. If you find these
topics at all, you will wonder: What the heck is it doing in this
book?
You rarely hear about truncation selection -- unless you challenge
evolutionists on Haldane's Dilemma. *Then* they will tell you about it!!!
And as always, they will put you down, saying YOU are ignorant, because
YOU didn't go to the libraries, because YOU didn't read all those books,
(because YOU didn't decipher the mysterious code-words, because YOU didn't
slog your way through all the misdirection and the half-stated arguments).
They will blame YOU!!! That is standard procedure for evolutionists, (and a
recurring theme of my exposes').
Truncation selection is difficult to sell, and evolutionists rarely sell it.
Rather, the standard model of evolutionary genetics -- prominent in
*all* evolution books -- has truncation selection nowhere in sight.
Let me make this point even clearer. John Maynard Smith was into
Haldane's Dilemma early, his 1968 paper offered a "solution" --
it was truncation selection. Does it (or "soft selection") show up in
his 1989 college textbook, _Evolutionary Genetics_? Nope.
Andy's post offers two supposedly "important" solutions -- but they are
both effectively absent from the textbooks. His post shows what a
scandal Haldane's Dilemma is. Neither the problem, nor its supposed
solutions are conveyed to the student.
>(2) Sexual reproduction allows several traits to spread through the
>population without detrimentally affecting the spread of other traits;
>i.e., replacements don't have to - and probably almost never do - occur
>sequentially.
Both I and Haldane allow that replacements can occur non-sequentially,
and that this does not solve Haldane's Dilemma. Sexual reproduction
combined with the standard version of evolutionary genetics --
the one model prominently displayed in all the textbooks --
cannot ease Haldane's Dilemma, because traits spread through the
population -- on average -- with no effect on the spread of other traits.
On average, they neither help nor harm the spread of other traits.
Each substitution incurs its own cost. The cost is paid separately,
and the payment for one beneficial trait cannot, on average,
pay for the substitution of another.
>(3) (To me, by far the most damaging argument) Haldane's predictions
>depended non-robustly on the assumption that there is a limit to the
>amount of variation a population can support at a given time. Low
>variation -> low rate of substitution. Less than ten years after Haldane
>pointed this out, however, it was shown that more variation than *anyone*
>suspected is in fact found in most natural populations. High variation ->
>high rate of substitution. ....
Andy's argument is mistaken on several levels. First, high genetic variation
does *not* inherently speed evolution (contrary to popular evolutionary
thought). That was an illusion created by mis-applying Fisher's
Fundamental Theorem of differential accrual (commonly mis-named as the
"Fundamental Theorem of Natural Selection"). That misapplication was
most prominently made by evolutionary geneticist James Crow. He used
the theorem to calculate the speed of evolution based on the measured
levels of high genetic variation, but he made several errors. One major
error was that the theorem does not apply (it is void) to situations of
real biological change where mutation, recombination, and migration are
injecting new variation into a population. When the theorem applies to a
situation it predicts the evolutionary process will grind to a halt --
it predicts precisely the opposite of what Crow was claiming.
Correctly understood, the theorem is no help whatever to
macro-evolutionists. (Again all details and documentation are in my book.)
Second, genetic variation does not solve the problem. Haldane's Dilemma
is about substitutions, their reproductive costs, and how they are paid --
and genetic variation cannot pay the costs. No amount of genetic variation
can pay the costs or reduce the costs. Genetic variation is not a solution.
Third, the observation of high genetic variation in natural populations only
deepened the problem. Selectionists tried to explain this as the result of
heterozygote advantage, and the neutralists quickly pointed out that this
mechanism -- all by itself -- would incur a reproductive cost higher than
mammals could plausibly pay. If a mammalian species' entire reproduction
couldn't pay the costs of heterozygote advantage, then it certainly couldn't
pay the additional costs of substitution -- thus this compounded Haldane's
Dilemma. The selectionists responded by de-emphasizing heterozygote
advantage, in favor of somewhat less costly balancing selection.
None of this eased Haldane's Dilemma.
>It's ReMine who's got it wrong. ... the cost of substitution is simply
>one form of genetic load (termed, among other things, "lag load").
Cost and load are not the same thing. They are quite different, but their
mechanics have similarities. For example, there are costs of: substitution,
segregation, and mutation. And those have parallel's in:
substitutional load, segregational load, and mutational load --
which are various forms of genetic load. Thus, substitutional load
is a form of genetic load. But "substitutional load" and
"cost of substitution" are different concepts -- that is Andy's mistake.
Such mixups between "load" and "cost" are common.
>[Haldane's Dilemma] hasn't been forgotten at all.
>It's just not called "Haldane's dilemma" anymore, ...
Thank you, Andy, for that admission. Haldane's Dilemma is the
trade secret of evolutionary genetics. Rarely can you find it even
mentioned, much less clearly described. Students get no vision,
no hint, no idea of the problem, and no coherent vision of a solution.
>.... Perhaps ReMine should look up "lag load" and "genetic load."
>I recommend a recent book by Bruce Wallace, "50 Years of
>Genetic Load."
Wallace did not solve the problem. His material is an extended
demonstration of how the concept of genetic load helped conceal,
rather than reveal, Haldane's Dilemma.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 20 Apr 1995 21:01:42 -0500
Message-ID: <3n73m6$llu@dawn.mmm.com>
*** More on Haldane's Dilemma ***
First a brief review. I claimed that both Haldane's Dilemma and its
supposed solutions (for example, soft selection and truncation selection)
are effectively absent from evolutionary genetics textbooks. Chris Colby
then claimed he found them in every textbook. The talk.origins crowd then
engaged in ad hominem attacks on me, claiming I am a "liar" and ignorant,
among other things. But Andy Peters now confirms my point. He claims it
is no longer called Haldane's Dilemma (therefore it is absent from
textbooks), and he admits that the solutions are typically absent from
textbooks.
Andy claims the solutions (soft selection and truncation selection) are
absent because they are too difficult and "it would be silly to spend a
lot of time talking about them." As justification he identifies several
specialized topics from evolutionary theory that are also seldom found in
textbooks. (He lists topics such as the modeling of evolutionarily stable
strategies, ESS.) But that does not justify the situation. Those topics
are narrow (and do not adversely impact the standard model of evolutionary
genetics), so they may legitimately be discussed separately. But
Haldane's Dilemma is not a narrow issue affecting some narrow arcane
ecological situation. Rather, it is broad and fundamental, and it
severely limits all macroevolution (especially for any species with low
reproduction, such as mammals). The standard model of evolutionary
genetics (the one model prominent in all the textbooks) is massively
inadequate to solve the problem, yet evolutionists teach that model anyway
-- without even mentioning the problem or its supposed solutions. This is
a scandal, and there is no excuse for it.
Andy shows confusion about what the solution is. One moment he claims
that soft selection and truncation selection are key solutions, the next
he claims the observed high levels of genetic variation "prove that the
problem doesn't exist." Though Andy didn't intend it, this is typical of
the run-around that evolutionists currently give on this issue. It is
difficult to get a straight answer, what with a dozen or so different
(erroneous) answers floating around. The confusion is evidence that the
problem was never actually solved. The confusion also helped obscure the
problem for decades. (More on this below.)
******
I wrote:
>> Truncation selection is difficult to sell, and evolutionists
>>rarely sell it. Rather, the standard model of evolutionary
>>genetics -- prominent in *all* evolution books -- has
>>truncation selection nowhere in sight.
Andy's latest post confirms my point above. Then he attempts to excuse
it. He says "This is primarily because the math of truncation selection
... is a real hassle, even for the best minds in the field." It is
unfortunate that the math is complicated, but that is no excuse for
leaving truncation selection out of the textbooks -- especially if
evolutionists claim it is an essential part of evolution and an essential
part of solving Haldane's Dilemma.
Truncation selection is difficult to sell, not because of the math, but
because of a lack of empirical support, and because of a lack of
theoretical justification that it is a plausible way for nature to behave.
Truncation selection, if it operated, would also have side effects harmful
to the evolutionary process. Moreover, teaching truncation selection
might raise student awareness of the problem it was intended to solve --
Haldane's Dilemma. All these issues (not just the math) are entirely
avoided in the textbooks. Students get no hint, not the foggiest idea, of
the problems.
Evolutionary genetics is a theoretical smorgasbord, where evolutionists
pick and choose models: The answer they give depends on the question you
ask. The problems of one model are ignored by conveniently choosing a
different model. Ask them about sexual reproduction, and they will
emphasize the critical importance of epistasis. Ask them about Haldane's
Dilemma, and they will emphasize truncation selection. But generally they
will emphasize their standard model, which is a lot easier to sell. This
smorgasbord approach shows up in Andy's post:
>If you look in the literature - particularly current literature
>in the maintenance of sexual reproduction and mating
>systems evolution - you'll see that epistasis is playing an
>extremely important part in explaining a number of
>phenomena. But, for a very large set of evolutionary
>questions, ... it's just a lot easier to use the simpler model.
******
Sexual reproduction, combined with the standard model of evolutionary
genetics, is massively inadequate to solve Haldane's Dilemma. Andy
disagrees. To support this he starts with an "extreme" situation, one
which he is "not claiming would ever happen in the real world" (his
words). He has us imagine that one individual has somehow (magically?)
acquired 100 new beneficial mutations. He says these are then substituted
altogether, all as one -- 100 beneficial substitutions for the cost of
one. However, Andy and I both agree that such a situation is "extreme,"
unrealistic, and cannot be relied on over the long haul to improve the
substitution rate. Individuals don't suddenly receive 100 new beneficial
mutations.
Moreover, sexual reproduction tears apart genomes and mixes them every
generation. This will rapidly split apart the 100 mutations so they are
no longer substituted as one unit. This way they would each incur the
full cost of substitution.
Andy insists there is something in sexual reproduction that lowers the
cost of substitution. But there is *nothing* in the standard genetic
model to allow that. Sex mixes traits all around the population. Sex
brings traits together, and breaks them apart with equal facility. The
traits are substituted, on average, without helping or hindering the
substitution of other traits. Sexual reproduction within the standard
model cannot solve Haldane's Dilemma. Evolutionary geneticists like John
Maynard Smith recognized that. That is why they proposed radically
different genetic models, such as truncation selection, as solutions to
Haldane's Dilemma.
>I'm saying that the *very same models* that predict a limit
>to the rate of evolution (Haldane's Dilemma) also predict that
>there should be very little genetic variation in natural populations.
That is incorrect. Haldane's Dilemma does not require or predict low
genetic variation. The problem remains for both high and low genetic
variation. However, the observed high levels of genetic variation
deepened the problem by increasing the cost incurred in maintaining the
variation -- which takes a bigger bite out of the available reproduction
-- which leaves less reproduction to pay the costs of substitution --
which ultimately slows the maximum plausible rate of substitution. In
this way the discovery of high genetic variation made Haldane's Dilemma
worse.
((NOTE #1: Andy explained his reasoning in only one obscure sentence:
>This is because the same demographic cost that limits substitution
>of alleles should manifest itself when there are large amounts
>of variation across the genome within populations.
His explanation makes no sense as it stands, much is left out, and I
suspect he is making the classic mistake of using the word
"cost" when he means "load".))
((NOTE #2: Haldane's concept of cost and the concept of genetic load
are different, with different meanings, and different formulas, leading to
two different approaches to analyzing Haldane's Dilemma and the
rate of substitution. Andy says cost and load are the same thing.
We are in total disagreement here.))
Creationist Paul Nelson recently brought the following substantial support
to my attention. George C. Williams is an expert evolutionist of
considerable renown. His latest book, _Natural Selection: Domains,
Levels, and Challenges_, has a chapter appropriately titled "Other
challenges and anomalies." In it, his first section is "Haldane's
dilemma" -- Yes, that is what he calls it. He writes:
"In my opinion the problem [of Haldane's dilemma]
was never solved, by [Bruce] Wallace or anyone else.
It merely faded away, because people got interested
in other things. They must have assumed that the true
resolution lay somewhere in the welter of suggestions
made by one or more of the distinguished population
geneticists who had participated in the discussion."
He briefly recounts the problem and the failed of attempts to solve it.
He concludes:
"I think the time has come for renewed discussion and
experimental attack on Haldane's dilemma."
Haldane's Dilemma is the trade secret of evolutionary genetics,
it was never solved.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 24 Apr 1995 12:57:15 -0500
Message-ID: <3ngopr$lng@dawn.mmm.com>
*** Haldane's Dilemma, and the textbooks ***
The latest batch of posts concentrates on the extent to which textbooks
do or don't deal with Haldane's Dilemma. My opponents claimed the issues
are discussed in every book they looked in, and they selected several
examples, but there were no surprises. We still have very different
perspectives. Let me summarize the literature in several stages.
First, there are the books aimed for wide public consumption. These are
generally books that sell evolution to the public, as a precise and
testable science. To do that, evolutionists inevitably reach for
population genetics as the convincer. The simple, easy-to-sell genetic
model is the one sold -- truncation selection and soft selection (the
alleged solutions to Haldane's Dilemma are nowhere in sight. All
non-college (and most college-level) evolution books and articles follow
that practice. I can bury you with examples, but will here offer only
Kitcher's anti-creation book, _Abusing Science_. He sells evolution via
the simple genetic model, and adds, "Mathematical population genetics
tells us *precisely* how the forces of immigration, emigration, mutation,
and natural selection produce evolutionary change." (Contrast that with
my opponents' admissions in this thread recently.)
All those books are wrong for doing that. If truncation and/or soft
selection are essential to solving Haldane's Dilemma, *then they should
be taught* (and defended) and any genetic model lacking them should have
appropriate disclaimers. No excuses.
This matter gets especially bad when you get to textbooks *specializing*
in evolutionary genetics. Any of these -- even mere introductory texts
-- should have the solutions to Haldane's dilemma indelibly connected to
the evolutionary model (not necessarily with all the math), and it should
be painfully clear to any student that these are not optional. On the
other hand, if a problem of such broad importance as Haldane's dilemma
were unsolved, then it would be inexcusable to leave it out of the
books. Either the solution, or the problem, or both, must be clearly
displayed in the book.
But that is not what we find. John Maynard Smith is a world authority on
Haldane's Dilemma and cost issues, and in the 1960s proposed truncation
selection as a solution. The preface for his recent textbook,
_Evolutionary Genetics_, says the book is intended for advanced
undergraduates, is hopefully useful to graduates, and that "proper
training in science requires that undergraduates are confronted by the
problems of contemporary science." The book discusses the usual standard
version of evolutionary genetics. But it doesn't discuss the following:
Haldane's problem or any version of it, cost of substitution,
substitution load, truncation selection, or soft selection. Everything
about the problem and its solutions is absent.
This is commonplace in textbooks on evolutionary genetics. Here is an
additional brief sampling of modern books, where everything is absent.
Molecular Evolutionary Genetics, Nei, 1987
Genetics of Populations, Hedrick, 1983 (***Mentions soft selection, but
doesn't tie it in as essential.)
Population Genetics and Molecular Evolution, edited by Ohta and Aoki,
1985 (***Mentions truncation selection in one paragraph, but doesn't
define it or tie it in as essential.)
Rates of Evolution, edited by Campbell and Day, 1987 (About half this
book is evolutionary genetics, but the stated topic of the book makes the
omission of both the Haldane problem and solution especially notable.)
Genetics, Paleontology, and Macroevolution, Levinton, 1988 (***Mentions
truncation in one paragraph, but doesn't tie it in as essential.)
Theoretical Population Genetics, Gale, 1990
Population and Evolutionary Genetics: A Primer, Ayala, 1982
Now we move to books that at least mention Haldane's problem.
Exceedingly few of these refer to it as "Haldane's dilemma". (Most
evolutionists will get angry if you call it a "dilemma", as you have
already seen in our ongoing discussion here.) The books dramatically
downplay it, insisting it's not a problem. So they often give
short-shrift to the real mechanics of the issue, with little more than
a few sentences, perhaps a paragraph. Even then they generally fail to
tie-in truncation selection and soft selection as essential to the
solution and essential to the evolutionary model. (Futuyma's text is an
example here.)
Finally we get to those few texts that actually discuss Haldane's
problem. But first, let me briefly restate what that problem was. In
1957, Haldane recognized that it takes excess reproduction to drive (or
"pay" for) a substitution, and that each substitution incurs a cost that
must be paid. He rigorously defined that cost, and estimated that, on
average, it is 30. In effect, 30 entire populations must selectively
perish during the course of one substitution. He estimated that, on
average, in each generation the typical mammalian population could
reproduce itself plus provide enough excess to pay for all sorts of
things (random death, and deaths caused by the influx of harmful
mutation, and deaths caused by other processes like heterozygote
advantage) and still have remaining a reproductive excess of ten percent
of its population size. This excess reproduction could pay for
substitutions. Thus, on average, the cost is 30, and it is paid in
installments of 0.1, so it takes 300 generations to pay for one
substitution.
People can understand reproduction "payments" and that there are firm
limits to it -- female primates cannot possibly average 50 births each!
So the key to Haldane's argument was understanding the reproductive costs
of substitution. Let me put it this way: Nothing can go from 'few' to
'many', without a reproductive cost being paid, and there are firm
theoretical limits to how low that cost can be. Understand that simple
concept, and you can overturn some of the commonplace pseudo-solutions to
Haldane's Dilemma. Put it all together right, and you have a broad
argument that many people can understand. But it requires a clear
explanation of cost and its unavoidability, otherwise Haldane's dilemma
is incomprehensible to the student. Haldane's argument made the costs
and payments explicit, and that increased clarity.
For example, tables showed how cost depends on dominance of the
mutation. The exceedingly high cost of beneficial recessive mutations
became clear, and that prompted some discussion about how to avoid it.
In effect, some evolutionists attempted to throw away *beneficial*
mutations -- when recessive -- because the cost is so high as to drag
down the entire evolutionary scenario into implausibility. That problem
was never really solved, but at least the problem was clear in a few
presentations.
Today, not only is the name "Haldane's dilemma" almost completely absent
from the modern textbooks -- so is Haldane's problem, as he formulated it.
I mean two things here. His cost/payment problem is either absent
altogether, or it is substantially replaced by a recasting of the problem
into a total genetic load argument. But either way, Haldane's original
cost/payment analysis is underdiscussed, and this leaves students
ill-prepared to grasp the genetic load argument. Most textbooks don't
adequately explain what cost is, why it is unavoidable, and that there are
firm limits it can't go below. They give little feel for cost, and there
is even less on details like the high cost of beneficial recessives.
Within a few years, evolutionists had recast Haldane's problem in a
different direction, whose focus is the total genetic load of all the
substitutions together. It is called a genetic load argument, and though
much of the underlying mechanics is the same as before, its presentation
is dramatically different: with different formulas, and many new
unnecessary confusion factors. The presentations argue in the opposite
direction from Haldane, and tend to focus almost exclusively on matters
that don't show up in the cost/payment analysis. That is why I say the
cost and load arguments are different.
The load argument goes something like this. You start by assuming
evolution is possible via the substitution of distinct mutations, and by
assuming some overall substitution rate that would make evolution
plausible to you. Typically, those assumptions are only vaguely stated,
and can often go unexamined. The numbers are typically given in terms of
"gene" or "amino-acid" substitutions, (rather than "nucleotides") and in
substitutions "per year" (instead of "per generation"), this makes their
interpretation less transparent, less immediate, and less transferable to
other situations where the generation time may be different.
Some presentations use estimates of the genome size and the number of
normal-sized genes. Some presentations then add an estimate of the
beneficial mutation rate. These factors take the student ever further
away from things that are known, and increase confusion, which is
unfortunate because the argument can be stated without them and thus made
more compelling. On the other hand, the issue of the high cost of
beneficial recessives is totally ignored, and many students are unaware of
it.
The total genetic load is then calculated. Actually, it is just the
total substitutional load, but this detail is often not pointed out,
which means all other types of load are implicitly ignored. Those
omissions dramatically tilt the presentation in favor of the evolutionary
scenario.
The analysis results in a number for the load, and the final task it to
interpret it to see if it makes sense for the species under consideration.
If it does seem plausible, then the underlying substitution scenario seems
plausible. But the immediate results are way too implausible, and the
load is far too high. This has resulted in a number of approaches
intended to explain away the situation. These generally take the form
"The load only seems way too high, in reality perhaps we are substantially
incapable of seeing the effects of load in natural populations. Perhaps
the populations are actually carrying a high load and it's just that we
have difficulty seeing it." The issue then becomes, how much can such
rationalizations be stretched?
Thus, the literature moved from Haldane's original cost argument to a
load argument, and the underdiscussion of the former left students
ill-prepared to understand the latter. Key relevant issues were left
out, and many unnecessary confusion factors were brought in. That was
not intentional, but in my view it is the result of the evolutionists'
lack of pursuit to truly teach and truly solve the problem.
Let me summarize the literature: the material on Haldane's problem is
generally entirely absent, or too little and too cryptic for the intended
audience. The textbooks taught it poorly. The ultimate proof is
two-fold: (1) Evolutionists claim the problem is solved, but they do not
remotely agree on what that solution is. (2) The problem was thought
solved, when it wasn't,
One last point. Haldane laid at our doorstep the ability to calculate
the number of substitutions available for human evolution. Using his
result, in ten million years an ape-human species with a 20 year
generation time could substitute no more than 1,667 beneficial
nucleotides. It's an easy calculation, yet in nearly forty years no
evolutionist saw fit to publish such a figure. Is it possible they all
thought it uninteresting and irrelevant? We can understand that
scientists have debates, and we can understand that any given scientist
sometimes says things that are later overturned and withdrawn, that is
how science progresses. But in this case it is the widespread,
persistent, long-term silence that is so inexplicable.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 27 Apr 1995 12:33:54 -0500
Message-ID: <3noki2$s7f@dawn.mmm.com>
*** Haldane's Dilemma, issues on sexual reproduction ***
Andy Peters says sexual reproduction (ie. recombination) can reduce the
cost of substitution and thereby speed up evolution. Andy feels this can
at least help solve Haldane's Dilemma. It can't, because sex ***slows***
evolution. "Populations would have to be in the millions for
recombination [ie. sex] to be important in increasing the spread of
favorable mutations." (G.C. Williams) Sex cannot make evolution go
faster, unless the population sizes are quite large, at which point the
cost of substitution is quite large and Haldane's Dilemma especially bad.
Andy mistakenly cited John Maynard Smith as supporting his position.
Maynard Smith agrees with what I said above, and discusses it, albeit
briefly, in his textbook (pages 240-1). Moreover, researchers now agree
that if there is epistasis (fitness interactions among the various
mutations), then sex dramatically slows down evolution. This is a very
robust result.
Once again I emphasize that evolutionary theory (and evolutionary
genetics) is a smorgasbord -- the answer that evolutionists give you
depends on the question you ask. Ask them a question, say, about how
populations get rid of harmful mutations fast enough to avoid error
catastrophe -- and they will usually tell you the critical importance of
epistasis. Ask them about Haldane's Dilemma, and they will emphasize the
importance of sex and small population sizes. But in small populations,
sex slows evolution. Moreover, sex and epistasis together make evolution
vastly slower, under the widest of circumstances. Evolutionists cannot
have it both ways, yet they continue the double-speak.
Under our current understanding, sex cannot solve Haldane's Dilemma.
Therefore, if you have trouble visualizing Haldane's Dilemma in sexual
species, then switch to asexual species, where evolution is faster yet the
dilemma is clearer and even more compelling.
A historical note here. The origin (and maintenance) of sex is a major
problem for evolutionary theory (traditionally one of its trade secrets).
Continually evolutionists tell their stories about how an advantage of a
minuscule fraction of one percent is sufficient to cause the eventual
elimination of all who lack that trait. But sex has a whopping
disadvantage, well over a 50 percent disadvantage, it's an immovable tough
hurdle. Evolutionists claimed it evolved anyway. For decades they said
sex was an advantage because it helped species evolve faster. They felt
this advantage was self-evident, and sufficient to overcome that whopping
50+ percent hurdle. Then the population geneticists showed that sex
actually slows evolution -- and this immediately caused a 180 degree
turn-around. Evolutionists now claimed, with straight-faced conviction,
that the advantage of sex is that it made species evolve slowly. This 180
degree turn-around was observed and lamented by evolutionist Graham Bell
in his fine encyclopedic book, _The Masterpiece of Nature_. The incident
is now a memorial to the incredible flexibility of evolutionary
storytelling. (For additional details and documentation, see my book,
_The Biotic Message_.)
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 28 Apr 1995 12:17:31 -0500
Message-ID: <3nr7vb$c1n@dawn.mmm.com>
*** Haldane's Dilemma and genetic variation ***
<Andy Peters, your posts are consistently the most intelligent and civil
of the bunch. Thank you for your patient replies detailing your position.
Without those I could not answer your issues on your terms. I hope the
following is helpful.>
Andy claimed that the observation of high genetic variation is
incompatible with, and therefore disqualifies, Haldane's Dilemma. Major
evolutionists have published the same claims. It took several posts
before Andy revealed his full line of reasoning, and the truth turns out
to be dramatically opposite what he was claiming. But before I show that,
let me first document some versions of what he was saying, so you can see
the difference between what he said and what he actually meant.
>I'm saying that the *very same models* that predict a
>limit to the rate of evolution (Haldane's Dilemma) also
>predict that there should be very little genetic variation
>in natural populations. This is because the same
>demographic cost that limits substitution of alleles should
>manifest itself when there are large amounts of variation
>across the genome within populations. Thus, the large
>amount of variation we see shows that the assumptions
>which predicted less variation (and, therefore, the
>assumptions which drive Haldane's Dilemma) must be wrong.
...
>the genetic variation doesn't solve [Haldane's] problem
>so much as prove that the problem doesn't exist.
...
>The existence of the variation was (and, in fact is)
>indeed a dilemma. The question still remains, "how is
>that variation maintained without incurring a demographic
>cost and wiping out populations left and right."
>However, as I've said above, far from "deepening the
>problem" of Haldane's Dilemma itself, it erased it:
>clearly, the assumptions which led to both the prediction
>of low variation and the prediction of low rates of
>evolution must be wrong.
...
>the concept of genetic load exposed the truly interesting
>phenomena behind the concept of Haldane's Dilemma, once
>the existence of large amounts of genetic variation in
>natural populations showed Haldane's Dilemma to be nonexistent.
...
>Which of the numerous assumption(s) are invalid is still
>unclear. That's the real current problem pertaining to
>this stuff: exactly what is it that we don't understand
>about selection that allows this variation to exist (and, by
>extension, allows relatively rapid evolution by selection)?
>... that particular problem hasn't yet been solved.
There is no such link between Haldane's argument and a requirement for
low variation. Haldane's argument is compatible with both low and high
levels of variation. I can refute Andy's claim several ways.
First, let me momentarily assume Andy's claim is true, and then show that
it leads to a contradiction. Begin with a situation where there is low
genetic variation, in full agreement with Andy's requirement for the
validity of Haldane's argument. Haldane's argument then places a valid
limit on the maximum plausible rate of evolution, and the population must
slowly evolve within those limits. Now give the population a large supply
of harmful mutations, say with x-rays, chemicals, or some special genetic
technique. The population is worse off than before. But it now has a
high level of genetic variation, and according to Andy's thinking this
sets the population free to evolve faster. That is the contradiction. In
effect, Andy claims the harmful mutations allow evolution to go faster.
I say the contrary. The harmful mutations are an extra burden for the
population. They slow down evolution because the increased harmful
mutation causes an increased level of genetic death, and more of the
population's reproductive capacity must be devoted to replacing (or
"paying" for) those genetic deaths. Therefore less of the reproduction
can go to paying for substitutions.
On the other hand, take the same example above, and instead of injecting
harmful mutations, inject neutral mutations instead. This likewise
creates a measurable "high level of genetic variation" but there is no
reason whatever that this should impact Haldane's argument. The neutral
mutations do no harm, they do not have to be eliminated, so they do not
increase the genetic deaths. They take no bite out of the available
reproduction, so the reproduction can go on paying for substitutions at
the same rate as before.
Thus, I have refuted Andy's claim by showing it leads to contradiction,
and by giving two counter-examples.
Andy's original claim was confusing, and it didn't become clear until he
gave the details of his argument. Now it is clear that when he said "low
genetic variation" he actually meant low harmful mutation load, which is a
different thing. His argument is about harmful mutations and their
drastic consequences on populations. Also he inadvertently created
confusion by focusing on how it is "maintained."
>The question still remains, "how is that variation
>maintained without incurring a demographic cost and
>wiping out populations left and right."
You don't want to "maintain" a high load of harmful mutation -- you want
to *get rid of it.* To speak of "maintaining it" just adds confusion, and
it took me a while to decode what he meant. (There are times when
evolutionists legitimately seek to maintain high levels of genetic
variation, such as by heterozygote advantage or balancing selection, but
that is a different issue altogether from harmful mutation, which you
just want to get rid of.)
My next post will examine the innards of Andy's argument.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 28 Apr 1995 13:26:42 -0500
Message-ID: <3nrc12$h50@dawn.mmm.com>
*** Haldane's Dilemma and genetic variation -- continued ***
Andy Peters claimed that the observation of high genetic variation is
incompatible with, and therefore disqualifies, Haldane's Dilemma. My
previous post refuted his claim by showing it leads to contradiction, and
by giving two counter-examples. This post deals with the specifics of his
argument, which are even more revealing. Here is Andy's argument again
for the sake of documentation. I'll explain his argument down below.
>as a result of the exact same assumptions as those for
>substitution load - there should be a limit on the amount
>of variation that can be maintained in a population.
>
>Yet the amounts of variation in natural populations are
>far beyond what could be expected under such a limit.
>For illustration, imagine a single individual with 100
>mutations across its entire genome. Even if each
>mutation by itself would only decrease fitness by 1%,
>the individual's fitness is .99^100 =~.36 . Estimates
>of the numbers of loci at which real-world organisms
>might be expected to have deleterious alleles range
>from 1000 to 5000, and a 1% decrease in [fitness] is very
>likely. That leads to mean fitnesses of 4 e -5 and
>1 e -22. If we apply the same criterion to this result
>as Haldane applied to get the substitution rate
>limitation - asserting a maximum proportion of selective
>deaths which should be allowable to the population every
>generation (Haldane used 10%) - we can see real-world
>populations should all be extinct if the assumptions
>behind Haldane's model hold (Proportion of selective
>deaths should be 99.996% for 1000 mutated loci,
>99.99999999999999999999% for 5000 mutated loci - the
>average individual would have to produce 100000 or 10^22
>offspring to expect only one to survive, and that
>doesn't even begin to take into account deaths due to
>ecological and accidental occurrences). In other words,
>under Haldane's model, *the amount of variation that
>exists in the real world, *can't* exist*!
>
>There's clearly something wrong with the assumptions
>leading to the predictions of both [Haldane's] rate
>limitation and variation limitation.
Andy begins with the standard model of evolutionary genetics. He calls
this "Haldane's model" but there is nothing about it that is unique to
Haldane's model. It's just the standard model of evolutionary genetics,
no epistasis. It is the one model prominently displayed in all evolution
books.
Then Andy says there are very good empirical grounds to believe the
typical organism has from 1000 to 5000 harmful mutations, each causing
roughly a 1% decrease in selective value.
Then, Andy happens to choose a level of selective death that is the same
as Haldane used, ten percent. But there is no connection, he could just
as well have arrived at a different figure and justified it separately
from Haldane. But it's all a red-herring anyway, because Andy never uses
the figure. In fact, even a figure near 100% would still not affect his
conclusion.
Andy then calculates the harmful mutation load based on the standard
model. The result dramatically conflicts with reality. The result
indicates that such species would rapidly go extinct, but since they
aren't extinct there seems to be a contradiction.
Here is Andy's mistake. He pinned the discrepancy on Haldane's argument
(just as many evolutionists do) and left it at that, as though Haldane's
argument was the thing at fault, and the existence of "high genetic
variation" (really high mutation load) is just a research problem yet to
be solved. The truth is a bit different. Andy's calculations leave
three, and only three, options:
(1) Perhaps the data is wrong. Andy employed measurements of data (the
number of harmful mutations in each individual, and their selective
disadvantage) and those measurements might be somewhat in error. But Andy
expressed strong confidence in the data. Also Andy is using his argument
as a strongly reliable reason to throw out Haldane's Dilemma, this is
additional evidence that Andy feels the data is strongly reliable. So
Andy would probably consider option #1 as invalid, and so continue to
option #2.
(2) The many observed species (and presumably all species) are in
error-catastrophe, and are deteriorating genetically, as Andy concluded
they are. Andy rejected that possibility because he assumes evolution
must be true, and the situation he described is evidence against
evolution. (Note: That all species are genetically deteriorating is a
possibility acceptable to the worldviews of many creationists.) Also,
option #2 means that Haldane's Dilemma remains valid because the model on
which it is based remains valid. So, option #2 provides an additional
independent line of evidence against evolution.
The problems don't just add, they magnify each other. All by itself the
mutation load seems too high, and the population close to, if not within,
error catastrophe. And all by itself the cost of substitution is too high
for macroevolution to be plausible. These problems aggravate each other,
because evolution must somehow accomplish these tasks simultaneously, but
the costs are paid separately out of the limited supply of reproduction.
Thus, making macroevolution especially problematic. Andy may find option
#2 unacceptable, and so continue to option #3.
(3) The standard model of population genetics is wrong. This is precisely
what I have been saying in my book and my posts: It is the one model
prominently taught in every evolution book, yet it is seriously inadequate
to justify macroevolution. And Andy's argument shows that evolutionary
geneticists *know it.* So why are they teaching that model universally,
almost exclusively, with none of the appropriate disclaimers, for decades
after they knew the discrepancy? It is because the standard model is
easier to sell, and population genetics is central to selling evolution.
That is the scandal I have been talking about.
Phil Nicholls writes:
>If reality and the model are wrong, toss the model, not reality.
Precisely my point. Why haven't evolutionists tossed the standard model?
They've known for decades that it can't justify evolution. Moreover,
there are several, independent, simple ways of showing that, and the
necessary data and arguments were available decades ago. (See my book,
_The Biotic Message_, for details.)
Andy garbled the truth the same way evolutionists have for decades. Like
this: "Haldane's Dilemma is a non-problem -- the observed high level of
genetic variation is an arcane puzzle -- sell the public on the standard
model of evolutionary genetics -- evolution is a fact." Instead, Andy
should be agreeing with me right now, and shouting -- right along side me
-- that the standard model of evolutionary genetics is manifestly
inadequate to justify macroevolution. If Andy believes his own argument,
that is what he must do.
The critical pressure point is not genetic variation, but rather the
standard model. There is irony in this. As this situation unfolds we may
well see creationists defending, and evolutionists attacking, the standard
model displayed in all those textbooks.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 29 Apr 1995 18:49:25 -0500
Message-ID: <3nuja5$bpf@dawn.mmm.com>
I made a clear claim: "Haldane laid at our doorstep the ability to
calculate the number of substitutions available for human evolution.
.... It's an easy calculation, yet in nearly forty years no evolutionist
saw fit to publish such a figure."
Tim Ikeda claimed to have a specific counter-example. I called his bluff,
and he lost. He shrugs it off, saying I am "technically correct." But my
claim is completely correct. Lewontin (and all evolutionists) gave no
figure, not even to an accuracy of one significant digit. So Tim attempts
to change the issue.
Evolutionists completely ignored the simple, obvious, readily
understandable, highly relevant issue of total substitutions, and instead
spoke obscurely in terms of substitution rate. Let me put this in greater
focus. Of the few evolution books that mention Haldane's problem, less
than a third state Haldane's substitution rate. Of those, nearly all
state it cryptically, by speaking of "gene" or "allele" substitutions, as
in "one gene substitution per 300 generations" or some variant of that
wording. That is how Lewontin did it (page 224). That practice is nearly
universal, and it effectively removed the magnitude of Haldane's Dilemma
from understanding by students and the public. A figure like "1,667
nucleotides" draws immediate attention and consternation. (It is roughly
50 millionths of one percent of the human genome.) Then even the ordinary
person sees the importance. But "one gene substitution per 300
generations" is effectively opaque, even to most evolutionary
specialists. The problem is effectively concealed.
One renowned evolutionary geneticist/professor chose to argue with me
recently. He insisted the substitutions are new 'genes' not
'nucleotides,' and therefore the number of beneficial nucleotides
available for human evolution is vastly higher than 1,667. It took a
while to convince him of his error, he had been so habitually used to
thinking of it as "gene substitutions." In fact, I cannot find even one
evolutionary textbook that disabuses students of that fallacy about
Haldane's substitution rate.
Haldane's dilemma is most severe for species with long generation times
and low reproduction, such as humans and simians -- species of key
interest. No evolutionary specialist could have overlooked that or found
it irrelevant. Yet for nearly forty years they never once took the
obvious step of publishing the number of beneficial substitutions
available for human evolution.
There are no excuses for that, though the talk.origins crowd surely tries.
Tim Ikeda says Lewontin "understood the implications" of Haldane's
argument. Fine, then why didn't evolutionists publish the implications
completely and clearly -- and truly share them with the world? Why did
they universally fail to do that? Chris Colby flatly asserts that a limit
of 1,667 nucleotides would not be a problem. Fine, then why didn't
evolutionists publish that? But no evolutionist published it -- not one.
That is now history, and it cannot be changed.
We are faced with a puzzle -- the evolutionists' widespread, persistent,
long-term silence on Haldane's dilemma. Evolutionary experts did not
reveal the magnitude of the problem. Instead, at every turn they
downplayed it and discussed it cryptically. Their silence allowed key
aspects of the problem to fall into obscurity. Their silence also allowed
the unchallenged decades-long promulgation of over a dozen phony
"solutions." Their silence also allowed their internal disagreement to go
unnoticed -- so that virtually all evolutionists agreed the problem was
solved (and few recognized they didn't remotely agree on what the solution
was). Lastly, most evolutionary textbooks are silent on both the problem
and its alleged solutions. Again I emphasize, it is the evolutionists
peculiar pattern of silence that is so damning.
I encourage each of you to explain that for the public -- for those many
ordinary work-a-day folks who want to understand how science operates.
Tell them what Tim, Chris, and others said here. And watch the reaction.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 2 May 1995 01:18:34 -0500
Message-ID: <3o4irq$8oi@dawn.mmm.com>
*** Haldane's Dilemma and sexual reproduction ***
Andy Peters writes:
> I'm conceding the point that, under most reasonable conditions,
> recombination can't erase - or even dramatically reduce -
> the cost of substitution.
Andy concedes that sexual reproduction (ie. recombination) cannot reduce
the cost of substitution or solve Haldane's dilemma. I emphasize that his
previous misunderstanding was not his fault. In fact, such
misunderstandings about recombination are rather common, especially in
discussions of Haldane's dilemma. Yet Andy has had extensive training in
this field, and it reinforces my point that the textbooks do not teach
Haldane's problem adequately (or the problems of the maintenance of sex
adequately).
Andy asked for more references documenting the evolutionary difficulties
with sex, recombination, and epistasis. I'll offer a few here, (there's
more in my book.) We've already accomplished what we need concerning sex
in this thread, so I won't be pursuing it further:
"[T]he occasional production of extremely fit genotypes in the sexual
population will have no permanent significance, as long as fitness depends
at all on heterosis and epistasis. Given any plausible level of heterosis
and complex interactions among loci [ie. epistasis], almost all of a
sexual population will have suboptimal genotypes, no matter how long
selection continues. All that this means is what everybody knows, that
sexual reproduction generates recombinational load, but I suggest that
this mundane fact may be the primary significance of sexuality in
evolution. It greatly retards the final stages of multilocus adaptive
change and severely limits the attainable precision of adaptation." (G.C.
Williams, 1975, p 150)
"To save the situation then, we must perform a complete volte-face: just
as it was self-evident to Weismann, Fisher and Muller that a faster rate
of evolution would benefit a population, so we must now contrive to
believe in the self-evident desirability of evolving slowly." (Graham
Bell, 1982, p 100)
"[I]t has proven difficult to produce explicit models that show that mixis
[ie. the genetic mixing caused by sexual reproduction] accelerates
adaptive evolution." (Michod and Levin, 1988, p 2)
Epistasis not only slows down evolution in sexual species, it also makes
the origin and maintenance of sex much harder to explain:
"[S]election will always favour a reduction in the rate of recombination,
provided that there is any degree of epistatic interaction between loci,
.... there is remarkable unanimity among theorists that the ineluctable
suppression of recombination is not only a very general but also a very
robust result." (Graham Bell, 1982, p 407)
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 2 May 1995 01:38:00 -0500
Message-ID: <3o4k08$91d@dawn.mmm.com>
Andy Peters writes:
>ReMine keeps implying that the selective use of
>assumptions in model building is a bad thing.
>It's not.
....
>This is, of course, the crux of the biscuit. ReMine seems to advocate an
>all-or-nothing approach to this, however: the "standard model of
>population genetics" is "manifestly inadequate," so it should be thrown
>out altogether. The fact, of course, is that the standard model of
>selection does predict very precisely lots of population phenomena. It's
>clear that selection works in natural populations, so throwing away the
>most general model of selection entirely would be throwing the baby out
>with the bath water (I've always wanted to use that expression). Tweaking
>various assumptions of the model (and thereby generating ReMine's
>"Smorgasbord") is the *appropriate* response to such a situation - though
>ReMine insinuates that it's dishonest.
Andy says that selective use of assumptions in model building is not a
bad thing. I say evolutionary genetics texts fail that on four counts:
1) The textbooks must *say* when they employ the selective use of
assumptions, and why.
2) The textbooks must *identify* the assumptions in each model.
3) The textbooks must delineate the various models, their strengths *and
weaknesses,* and their role (or not) in evolutionary theory.
4) The textbooks must not give contradictory models and pass them off as
one unified solution to problems.
I say evolutionary genetics textbooks routinely fail on these four
points, (especially #3 and #4). The following examples are macroscopic
problems that apply broadly, they are not narrow ecological issues.
1) Error catastrophe (also mutation load)
2) The origin and maintenance of sexual reproduction (recombination)
3) Haldane's Dilemma -- puts limits on the rate of macroevolution
4) Kimura's problem with the limitation on the rate of beneficial
mutation, which also puts a limits on the rate of macroevolution.
It is a contradiction to embrace epistasis to solve one problem (such as
error catastrophe), and then renounce epistasis to solve another problem
(such as sexual reproduction).
It is a contradiction to embrace "small population sizes" as a solution
to Haldane's Dilemma, and also embrace "large population sizes" as a
solution to Kimura's problem.
It is a contradiction to embrace "sex and small population sizes" as a
solution to Haldane's dilemma, when those two together make evolution
slower, especially when combined with epistasis.
The textbooks routinely employ those contradictions in a smorgasbord
fashion, where they fail to identify the contradictions. The answer they
give you depends on the question you ask.
The problem is not the selective use of models, per se. Rather, the
problem is the hidden contradictions, the absence of disclaimers, and the
absence of disclosure. The textbooks present evolutionary genetics as a
uniform, coherent, testable theory. It is not. Andy claims I am
"throwing the baby out with the bath water." I am not. I have
consistently said things are *absent* from evolutionary textbooks that
ought not be absent. I'm not throwing out anything. I say the textbooks
should *bring in* some things -- things they left out for decades.
<I'll handle Andy's latest technical issues in another post.>
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 4 May 1995 19:40:57 -0500
Message-ID: <3obs6p$gu@dawn.mmm.com>
*** Haldane's Dilemma and genetic variation ***
This is my response (part 1 of 3) to Andy Peter's latest post on genetic
variation.
Andy asserts that the observed high mutation load invalidates Haldane's
Dilemma. Andy's post often re-states his assertion (in slightly different
words), over and over again, without justifying it or explaining it. It
is a repetitious and unilluminating post in that sense. Also, it departs
from Andy's normally more patient style and engages in microscopic
sniping, taking every opportunity to be contrary and to merely re-state
his assertion yet another time, as if that will drum it into us better.
That is unfortunate for several reasons. While I can sort Andy's post, I
fear many readers here cannot and are left more confused than enlightened.
Moreover, evolution textbooks don't remotely teach Haldane's Dilemma
adequately -- even evolutionary specialists are confused -- and I've
documented that point repeatedly. Even Andy now offers the
self-incriminating excuse, "I don't give a flying fajita whether or not
Haldane's Dilemma is dealt with to [the student's] satisfaction in
textbooks and popular literature." To make matters worse, evolutionists
frequently accuse their opponents of ignorance. (We've seen that often in
this thread.) But they have no high ground for any such charge.
Evolutionists themselves are responsible for the rampant ignorance. For
decades they literally sold ignorance in their textbooks.
Finally, I cannot find Andy's refutation of Haldane's dilemma detailed
anywhere in the literature, and have encouraged him to cite it if he knows
examples. Since he has not done that, it appears Andy's refutation is
being given here for the first time anywhere, so it is especially
important that he teach it clearly and patiently, as it is the only
example of its kind.
Altogether, Andy has no grounds for sniping, for shortness, or for
charging me with ignorance.
******
Let me now eliminate some confusion factors.
Andy claims the observed "high level of genetic variation" invalidates
Haldane's Dilemma. To refute his claim, I provided two counter-examples.
In one counter-example, the high level of genetic variation is due to
neutral mutations, which do not affect the beneficial substitution rate or
Haldane's argument. Andy accepts my point as "valid," thus we both agree
that some kinds of genetic variation (variations that have neutral effect)
do not invalidate Haldane's Dilemma.
Then Andy appears to contradict himself:
>Under Haldane's model, however, genetic variation and
>mutation load are in fact exactly the same thing.
No, Haldane's model does not require genetic variation and mutation load
to be the same thing. I just gave an counter-example and Andy already
acknowledged it as "valid."
Perhaps we are arguing over the term "genetic variation," but I continue
to emphasize that not all genetic variation is the same. Harmful,
neutral, and beneficial mutations create different types of "variation"
that are all rather different in their effects and in their problems.
Many evolutionary arguments achieve their illusory power by using the
ambiguous term "variation" and then shifting back and forth between the
various meanings. To end such mischief we must identify the type of
variation and use that identification consistently. If the "variation"
is actually "harmful mutations" or "mutational load", then use these latter
terms instead, as they are less ambiguous. I will make that replacement
when I quote Andy, as that seems to be his intended meaning throughout.
That replacement makes Andy's mistakes more visible. For example:
>The [load of harmful mutations] doesn't *allow* the
>population to evolve faster (well, actually it does ...
That is nonsense. Harmful mutations are a burden to be eliminated. They
do not allow a population to evolve faster in any sense whatever. They
slow evolution.
Another example is when Andy discusses "the problem of maintaining high
genetic variation." Unfortunately, his confusing wording is *identical*
to legitimate statements about an altogether different problem involving
an altogether different type of genetic variation. (eg. Mechanisms such
as the neutral theory, heterozygote advantage, or balancing selection are
legitimately used to explain the maintenance of special types of genetic
variation.) So I thought that's what Andy meant. But I later realized
Andy meant something different -- "the problem of maintaining [high levels
of harmful mutations]." Even that was confusing, because there is no such
maintenance problem. The problem is not "maintaining" harmful mutations,
rather the problem is *getting rid of them* "without wiping out
populations left and right." Or to put it differently, the problem is
"maintaining the population" not the harmful mutations. I recommend
clearer wording, such as "the problem of eliminating harmful mutation."
Again the nexus of confusion is the term "genetic variation" and its many
ambiguous meanings. I'm not picking on Andy here. The terms he used are
commonplace in the literature.
(continued in part 2)
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 4 May 1995 19:56:34 -0500
Message-ID: <3obt42$n9@dawn.mmm.com>
*** Haldane's Dilemma and genetic variation ***
(This is the continuation (part 2 of 3) of my response to Andy Peter's
latest post on genetic variation.)
Let me call this Andy's assertion: The observed high mutation load
invalidates Haldane's Dilemma.
An earlier post of mine refuted Andy's assertion. My argument momentarily
assumed Andy's assertion is true, and then shows it leads to a
contradiction. Andy responded to my argument, and I now show where his
response errs.
I had Andy imagine a population (any kind he wishes) that meets *his*
criteria for the validity of Haldane's limitation on the rate of
evolution. According to Andy this would require low genetic variation --
low mutation load. Thus my argument begins by accepting Andy's assertion
on Andy's terms. What is there for him to complain about?
But Andy bails out of my argument at that point, apparently because he
cannot even conceive of any population being subject to Haldane's
limitation under any circumstances whatever. He bails out by inventing
yet another reason for rejecting Haldane's limitation. Here it is in action:
>[J]ust because our hypothetical population has low
>genetic variation, ***that doesn't mean that the
>selective mechanisms that reduce the cost of
>selection don't exist.***
[wjr: my emphasis]
In other words, Andy spontaneously invents unspecified mechanisms for
removing Haldane's limitation. (He lamely calls them "selective mechanisms
that reduce the cost," but that is just an empty name.) He never
identifies what those mechanisms are, but they are central to his
argument, for it is no longer "high genetic variation" that overturns
Haldane's limitation, rather it is now Andy's unknown mechanisms. In that
way, Andy dodges my argument an avoids having his position revealed as a
self-contradiction. He then merely repeats his assertion several more
times.
In summary, Andy asserts that high mutation load "shows Haldane's
assumptions to be false." I just disproved that. You start by giving Andy
everything he could want (including Haldane's assumptions and Andy's
requirement for low variation), then Andy bails out and invents another
reason to reject Haldane's conclusion.
For sake of documentation, here is the entire quote from Andy's post,
starting with my original argument, alternating with his responses:
<Begin quote>
> First, let me momentarily assume Andy's claim is true, and then show that
> it leads to a contradiction. Begin with a situation where there is low
> genetic variation, in full agreement with Andy's requirement for the
> validity of Haldane's argument. Haldane's argument then places a valid
> limit on the maximum plausible rate of evolution, and the population must
> slowly evolve within those limits.
Not necessarily. The existence, or lack thereof, of genetic variation in
any given population doesn't have any direct effect on the cost of
substitution in that population. In other words, just because our
hypothetical population has low genetic variation, that doesn't mean that
the selective mechanisms that reduce the cost of selection don't exist.
All the existence of large amounts of variation in natural populations
does is to *demonstrate* that the cost of substitution doesn't exist - it
doesn't actively reduce the cost itself.
> Now give the population a large supply
> of harmful mutations, say with x-rays, chemicals, or some special genetic
> technique. The population is worse off than before. But it now has a
> high level of genetic variation, and according to Andy's thinking this
> sets the population free to evolve faster.
Nope. The variation doesn't *allow* the population to evolve faster
(well, actually it does, .... <Andy repeats his assertion here.>
> That is the contradiction. In
> effect, Andy claims the harmful mutations allow evolution to go faster.
.... <Andy repeats his assertion again here.>
<End quote from Andy's post>
(continued in part 3)
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 4 May 1995 20:02:31 -0500
Message-ID: <3obtf7$se@dawn.mmm.com>
*** Haldane's Dilemma and genetic variation ***
This is the continuation (part 3 of 3) of my response to Andy Peter's
latest post on genetic variation.
Here are three models of evolutionary genetics:
1) The model that supports Haldane's dilemma.
2) The model that supports the concept of mutational load.
3) The standard model, the one in all the books, the one used to sell
evolution to the public.
I say all three are the same model. In fact, merely use "standard model"
whenever Andy uses "Haldane's model," and his post still reads as true.
Andy agrees that 1 and 2 are the same model, and most of his post labors
to establish that. (I was already convinced a long time ago.) But Andy
didn't establish that models 1 and 3 are different, or that 2 and 3 are
different. ***He didn't even try.*** Instead he takes the problem of
mutational load, and blames it on "Haldane's model", citing its huge
failure to avoid error catastrophe and extinction. But he tars the
standard model with the same brush, because they're the same model.
So I cheer him forward! Andy is making my case for me. If Andy believes
his own argument, the standard model of evolutionary genetics massively
contradicts macroevolution. And one of them must go.
There is much that Andy and I agree on here. The difference is: (1) Andy
wants to scapegoat the problems onto "Haldane's model." (2) Andy wants to
continue selling the standard model of genetic evolution without revealing
the massive contradiction that Andy himself identified.
******
There was an option that Andy arbitrarily rejected, the option that many
species (presumably all species) actually are in error-catastrophe and are
deteriorating genetically.
Andy explainss:
>I rejected [that] possibility because it goes against the evidence.
>Natural selection happens, whether or not all life shares a common
>ancestor. It's been demonstrated countless times. It's not an
>"assumption" on my part - it's a conclusion, based on evidence.
That's a classic move. When your argument gets in trouble just say,
"Natural selection happens!" and we're supposed to all fall over. When it
comes to evolutionary illusions, there is none bigger or more intricate
than natural selection.
Let me make it clear. If every species on the planet were genetically
deteriorating, then that would be totally compatible with natural
selection. The moths, the finches, and the laboratory flies could take
two steps up the hill, and slide three steps back down the hill. The
species could "survival of the fittest" all the way down to extinction.
Let me broaden that. Take Haldane's argument, mutation load, the
standard model of evolutionary genetics, and natural selection -- They
are all totally compatible with genetic deterioration, even on a global
scale. None of them predicts against it.
Yet Andy rejects genetic deterioration. He claims (mistakenly) that
Haldane's model makes a "prediction" against genetic deterioration. In
truth, he gets that notion from outside -- from somewhere else. He gets
it from his worldview. Because global genetic deterioration is
incompatible with macroevolution.
******
Andy writes:
>Under Haldane's assumptions, however, almost every
>population in existence should have been wiped out
>centuries ago, given the amount of variation we've
>seen in them.
Let me clarify Andy's statement:
1) "Haldane's assumptions" = "the standard model of evolutionary genetics"
2) You can't make measurements on organisms **today** and conclude they
"should have been wiped out centuries ago." That's hyperbolic (as well
as impossible). Change it to "should be extinct sometime in the near
future."
3) Change "variation" to "harmful mutation", which is what Andy means.
Then the statement is accurate and true to Andy's argument.
Like this:
>Under [the standard model of evolutionary genetics],
>however, almost every population in existence [should
>be extinct sometime in the near future], given the amount
>of [harmful mutation] we've seen in them.
That statement ought to be in the textbooks. It is simple and clear,
anyone can interpret its meaning for themselves. So why does Andy garble
it beyond the public's understanding? Why did it take so long for Andy to
fully reveal his key argument? Why did it take so many posts to strip
away layer after layer of misdirection and confusion? Andy totally
believes his own argument, so why didn't he just shout it in plain text?
Throughout my posts I have shown that evolutionists persistently avoid key
issues inconvenient to them. On such matters they speak cryptically, if
at all. Andy has just given a demonstration of it.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: ReMine statistics
Date: 6 May 1995 16:41:21 -0500
Message-ID: <3ogqe1$g00@dawn.mmm.com>
In 1957, the renowned evolutionary geneticist J.B.S. Haldane published a
problem that eventually bore his name "Haldane's Dilemma." He saw that
reproductive capacity is limited, and this puts a limit on the speed of
evolution. He showed that in species with low reproductive capacity
(such as higher vertebrates, mammals and man), the long-term average
beneficial substitution rate cannot be faster than one substitution per
300 generations. These substitutions are single nucleotides.
I wrote (in my April 24 post):
"Haldane laid at our doorstep the ability to calculate
the number of substitutions available for human evolution. Using his
result, in ten million years an ape-human species with a 20 year
generation time could substitute no more than 1,667 beneficial
nucleotides. It's an easy calculation, yet in nearly forty years no
evolutionist saw fit to publish such a figure. Is it possible they all
thought it uninteresting and irrelevant? We can understand that
scientists have debates, and we can understand that any given scientist
sometimes says things that are later overturned and withdrawn, that is
how science progresses. But in this case it is the widespread,
persistent, long-term silence that is so inexplicable."
1,667 beneficial nucleotide substitutions amounts to about 50 millionths
of one percent of the human genome. And that is *before* making several
deductions, such as the 90% reduction due to the 90% of the time spent in
stasis, according to Gould. Is a maximum of 1,667 beneficial nucleotides
enough to make a sapien out of a simian?
Chris Colby responds:
>What ReMine hasn't done is discuss
>his assertion that 1667 adaptive
>[nucleotide] substitutions is too few
>to explain our divergence from the chimp/
>human ancestor. Even if we accept his number
>*****(and I'll admit that it is within the ballpark),*****
[wjr: my emphasis above]
>there is no indication that it is inadequate.
>.... a limit of 1600 nucleotide changes may
>not be a problem.
Chris should defend that view in a journal. He will be the first
evolutionist *ever* to even try.
******
Chris Colby repeatedly misrepresents my posts to the thread "Haldane's
Dilemma." For example, he writes:
>[ReMine] claims that Haldane's dilemma is absent from evolutionary texts.
In my defense, I will merely quote portions of my original post:
<<Begin quote from my post of April 24>>
Let me summarize the literature in several stages.
First, there are the books aimed for wide public consumption. These are
generally books that sell evolution to the public, as a precise and
testable science. To do that, evolutionists inevitably reach for
population genetics as the convincer. The simple, easy-to-sell genetic
model is the one sold -- truncation selection and soft selection (the
alleged solutions to Haldane's Dilemma are nowhere in sight. All
non-college (and most college-level) evolution books and articles follow
that practice. I can bury you with examples, but will here offer only
Kitcher's anti-creation book, _Abusing Science_. He sells evolution via
the simple genetic model, and adds, "Mathematical population genetics
tells us *precisely* how the forces of immigration, emigration, mutation,
and natural selection produce evolutionary change." (Contrast that with
my opponents' admissions in this thread recently.)
All those books are wrong for doing that. If truncation and/or soft
selection are essential to solving Haldane's Dilemma, *then they should
be taught* (and defended) and any genetic model lacking them should have
appropriate disclaimers. No excuses.
This matter gets especially bad when you get to textbooks *specializing*
in evolutionary genetics. Any of these -- even mere introductory texts
-- should have the solutions to Haldane's dilemma indelibly connected to
the evolutionary model (not necessarily with all the math), and it should
be painfully clear to any student that these are not optional. On the
other hand, if a problem of such broad importance as Haldane's dilemma
were unsolved, then it would be inexcusable to leave it out of the
books. Either the solution, or the problem, or both, must be clearly
displayed in the book.
But that is not what we find. John Maynard Smith is a world authority on
Haldane's Dilemma and cost issues, and in the 1960s proposed truncation
selection as a solution. The preface for his recent textbook,
_Evolutionary Genetics_, says the book is intended for advanced
undergraduates, is hopefully useful to graduates, and that "proper
training in science requires that undergraduates are confronted by the
problems of contemporary science." The book discusses the usual standard
version of evolutionary genetics. But it doesn't discuss the following:
Haldane's problem or any version of it, cost of substitution,
substitution load, truncation selection, or soft selection. Everything
about the problem and its solutions is absent.
This is commonplace in textbooks on evolutionary genetics. Here is an
additional brief sampling of modern books, where everything is absent.
Molecular Evolutionary Genetics, Nei, 1987
Genetics of Populations, Hedrick, 1983 (***Mentions soft selection, but
doesn't tie it in as essential.)
Population Genetics and Molecular Evolution, edited by Ohta and Aoki,
1985 (***Mentions truncation selection in one paragraph, but doesn't
define it or tie it in as essential.)
Rates of Evolution, edited by Campbell and Day, 1987 (About half this
book is evolutionary genetics, but the stated topic of the book makes the
omission of both the Haldane problem and solution especially notable.)
Genetics, Paleontology, and Macroevolution, Levinton, 1988 (***Mentions
truncation in one paragraph, but doesn't tie it in as essential.)
Theoretical Population Genetics, Gale, 1990
Population and Evolutionary Genetics: A Primer, Ayala, 1982
Now we move to books that at least mention Haldane's problem.
Exceedingly few of these refer to it as "Haldane's dilemma". (Most
evolutionists will get angry if you call it a "dilemma", as you have
already seen in our ongoing discussion here.) The books dramatically
downplay it, insisting it's not a problem. So they often give
short-shrift to the real mechanics of the issue, with little more than
a few sentences, perhaps a paragraph. Even then they generally fail to
tie-in truncation selection and soft selection as essential to the
solution and essential to the evolutionary model. (Futuyma's text is an
example here.)
<<End quote from my April 24 post>>
My original post has some additional discussion. Then I concluded:
"Let me summarize the literature: the material on Haldane's problem is
generally entirely absent, or too little and too cryptic for the intended
audience. The textbooks taught it poorly. The ultimate proof is
two-fold: (1) Evolutionists claim the problem is solved, but they do not
remotely agree on what that solution is. (2) The problem was thought
solved, when it wasn't."
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 6 May 1995 18:24:45 -0500
Message-ID: <3oh0ft$h2f@dawn.mmm.com>
Tim Ikeda writes:
>Perhaps one way to examine whether it is inevitable that genetic
>deterioration eventually leads to extinction would be to look at
>organisms with rapid doubling times, relatively high mutation
>rates, or ones which seem to have been around for a very long
>time. Viruses and bacteria could be useful subjects in the
>first two instances, and perhaps, horseshoe crabs in the latter.
>To a first approximation, could we suppose that bacteria and
>viruses would approach "genetic oblivion" at a higher rate than,
>say, birds? The coding and efficiency of bacterial chromosomes
>seem tighter than that found in vertebrates.
Genetic deterioration depends on the genome size, particularly that
portion that is non-inert. (Leaving ploidy out of the discussion for the
moment.) Organisms with large genomes will tend to reach error
catastrophe sooner.
Also organisms such as mammals have many genome duplications that occur
in the interval between egg fertilization and the fertilizable eggs of
the next generation. (The same with sperm.) Each of these duplications
can introduce errors. The process of recombination (ie. sexual
reproduction) adds additional chances for error. These tend to increase
the per generation mutation rate in multicellular organisms.
Also, species with low reproductive capacity (such as mammals) cannot
fend off error catastrophe for as long as bacteria, which have very high
reproductive capacity. Reproductive capacity is a critical factor.
Using the standard model of evolutionary genetics (ie. no epistasis), we
can make a simple calculation with data already available to Motoo Kimura
and others in the 70s. Assume 97 percent of the human genome is totally
inert, and mutation to this portion of the genome does absolutely no
harm in this generation or any generation to come. This assumption
dramatically favors the standard model, as you will see, because it
dramatically reduces the number of harmful mutations experienced by the
organism. Then only 3% of the genome is available to suffer harmful
mutation. Using the genome size and the mutation rates (data that Kimura
and others regularly used in their discussions, and which are still
consistent with today's figures) the following can be shown. To avoid
error catastrophe, human females would have to conceive more than 16.3
children. That is the cost of harmful mutation. And that is before
adding in the additional costs of substitution, segregation, and random
death, which place additional demands on reproductive capacity. (See my
book for details.)
This argument, and Andy Peter's mutation load argument, both address the
problem of harmful mutation. And the results point in the same
direction. (1) Either the data is wrong (which is getting progressively
more unlikely). (2) Or, many organisms are in error catastrophe. (3)
Or, the standard model is wrong.
The subject of error catastrophe is interesting. But it gets us too far
afield from our subject ... which is Haldane's Dilemma. So I won't be
addressing it further in this thread. Let me just point out again, that
several independent lines of evidence indicate that the standard model of
evolutionary genetics contradicts macroevolution. And everything
necessary to show this ... was available decades ago.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
--Kurt Vonnegut from "Galapagos"
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 8 May 1995 14:06:00 -0500
Message-ID: <3olq2o$k0j@dawn.mmm.com>
Andy Peter's latest post has little new from his previous efforts, and
much confusing thrashing about.
Andy continues to flog the standard model of evolutionary genetics (a.k.a.
"Haldane's model") -- with even greater vigor now than before. And I
continue to cheer him on. The more he flogs, the more ridiculous his
position becomes. I could say several things about the details of his
mutation load argument ... but will restrain myself. He is
self-destructing quite enough already.
******
There is a cost of substitution, it is unavoidable, it must be paid from
a species' limited reproductive capacity, and this places a limit on the
rate of substitution. The discovery of "high levels of genetic
variation" *aggravates* that problem, no matter what evolutionary genetic
model you choose. Higher levels of harmful mutation, or heterosis, or
balancing selection, or random death, all take higher bites out of the
available reproduction, leaving less reproduction to pay the costs of
substitution -- thus slowing evolution. Start with *any* of the
evolutionary models, and higher levels of harmful mutation (or higher
levels of mutation load) make the problem worse. Andy is wrong to even
hint otherwise.
The observed high levels of harmful mutation doesn't invalidate the
problem, as Andy suggested. Rather, it invalidates "solutions." For
example, Andy's mutation load argument massively invalidates the standard
model of evolutionary genetics.
And the problems don't go away. The costs still remain no matter what
genetic/selection model you choose. The different models impact the
analysis in different ways. But the problems must still be solved, not
just brushed aside, as Andy attempts to do.
I have before seen evolutionists *claim* that genetic variation
invalidates Haldane's conclusions. (eg. Mayr, 1982, p 594) But Andy
offered the first example of actually filling in that argument. I
recognize Andy's argument as new (and have asked him for references if it
is not).
I also recognize that Andy's interpretation is the ***opposite*** of G.C.
Williams (evolutionist and author of several well received books).
Though Williams uses the exact same example that Andy used.
<Begin quote from Williams' book>
"There are perhaps 100,000 loci in a mammalian genome, and we might
expect at least 10,000 of them to be polymorphic, ... Perhaps an average
individual has a less than ideal genotype at 10% of these loci, or 1000
in all. In a comparable example, Wallace (1989) gives an average
individual a suboptimal genotype at 5000 loci. If the average fitness
loss per locus is 1% at 1000 loci, the average individual will only be
0.99^1000, or about 10e-5 of the fitness of the best possible genotype.
The variance in fitness would be such that almost no individual would be
of even 1% of the maximum. Since maximum fitness is no assurance of
survival, the population envisioned would have an utterly inadequate gene
pool for survival beyond the current generation." (G.C. Williams, 1992,
p 145)
<End quote>
Williams' argument is identical to Andy's argument, right down to the
numbers. But Williams (correctly) presents it the opposite of Andy.
Williams' immediately preceding sentence is: "A bit of arithmetic shows
the ***severity*** of Haldane's dilemma." (my emphasis) Williams
doesn't try to cover it up, or brush it aside, or misinterpret it, as
Andy does.
******
Andy cites one of Haldane's papers on variation, dated 1937, a full 20
years before Haldane's paper on the limitations to substitution rate.
They are separate papers, dealing with separate issues, in separate parts
of Haldane's career. Yet Andy reaches into that (irrelevant?) paper and
pulls out an (irrelevant?) quote:
"In a species in equilibrium, variation is ***mainly***
due to two causes. ... [heterosis; and mutation-
selection balance]." [my emphasis]
Andy concludes:
>It's clear: under Haldane's assumptions, variation = mutation or
segregation load.
Wrong. Even in that old paper it is clear that Haldane leaves room for
other types of variation, such as neutral mutations.
That example constitutes the *entirety* of Andy's attempt to detach the
standard model away from Haldane's Dilemma. Andy is wandering around in
history, trying to pin the problem on Haldane ... and bury it with him.
I say it again -- the cost of substitution and the cost of harmful
mutation are each serious problems for the standard model. (Even Andy's
mutation load argument shows that.) The existence of neutral variation
neither helps, nor harms, nor invalidates, the problems, and Andy has
already acknowledged this point as "valid." (his word)
******
>The standard model (which is, indeed, largely Haldane's) does not
>"massively contradict" macroevolution. In its explanation of allele
>substitution, for instance, it describes the spread of an allele through a
>population. ReMine's position is, "No, it doesn't! The assumptions of
>the model are wrong - Andy's shown that himself."
By Andy's own mutation load argument, the standard model massively
contradicts macroevolution. So Andy changes the subject. He wants to
talk about "the spread of an allele", never mind that the population
"should already be extinct"!!! (His words.) He blames the problems on
Haldane, and he wants to keep selling the standard model to the public,
without revealing the contradictions.
******
>The fact is, however, that *the very same math* describes
>precisely the spread of an allele through the population
>under soft selection - it's merely the assumption about the
>way selection acts on the population that changes, and that
>doesn't change the genetic dynamics.
Andy makes an interesting point: soft selection uses "the very same math"
as the standard model. This is because soft selection takes the standard
model and merely emphasizes the role of frequency-dependent and
density-dependent selection. That provides nothing to eliminate harmful
mutations more efficiently, and is just as prone to error catastrophe.
Thus Andy's mutation load argument is a serious problem for soft
selection too.
Combine that with the fact that soft selection doesn't even solve
Haldane's dilemma, the cost of substitution. (As pointed out by G.C.
Williams, 1992, p 143-148)
******
>I'm not sure about how well the standard model describes
>the spread of an allele under truncation selection,
>but it's certainly at least a fair approximation.
I'd challenge that point. Truncation selection is under-discussed and
'selectively' discussed. It's part of the evolutionary smorgasbord.
Evolutionists discuss truncation rarely, and only then to "solve" cost
issues, like Haldane's dilemma. Then they ignore it thereafter. They
pick its effects that they like. And they ignore its effects that they
don't like. (Truncation selection does have bad effects for
evolutionists, but I've never seen them discuss those.) It's not enough
to say "truncation selection!" and expect us to fall over. If
evolutionists are serious about it as a solution, then they must hold it
high, defend it, and pursue it as a major feature of evolution. ... They
haven't done that.
******
Andy injects yet another "model" into our discussion. He calls it the
"general model." He praises it glowingly for its precision and science.
I call it a structureless smorgasbord of conflicting and contradictory
mechanisms, instantly ready to be shaped by the evolutionists' whims of
the moment.
>The model isn't perfect - such is the burden of science. It remains,
>however, the most successful general model to explain the phenomena of
>population genetics: because it still offers a good to excellent
>approximation of real world events; and because it's still unclear to what
>extent soft selection, truncation selection, and other selective
>mechanisms act in nature, the general model is the best model. Indeed, if
>it turns out (as it very probably will) that these mechanisms are of
>varying importance under different conditions, the general model is still
>the best model.
>
>That's the way science works.
I disagree. That is not how Science works. Science does not work by
obscuring the problems, hiding the contradictions, and misrepresenting
the models -- which is what evolutionary geneticists have done for
decades.
******
I previously pointed out to Andy that his use of the term "genetic
variation" is ambiguous and confusing about key issues -- that it serves
to conceal the serious problem of harmful mutation. Nonetheless, Andy's
post continues using his ambiguous term -- now more resolutely than ever,
without even once clarifying it. So my concern now is for those many
ordinary folks who struggle to understand the issues. Though Andy
doesn't intend it, his ambiguous language places hurdles in their path.
Evolutionists have no right to accuse anyone of ignorance -- until they
take seriously their obligation to stop selling it.
Perhaps evolutionists are just lazy about communication? Not so! They
can be emphatically clear, when they really want to. They are loud and
unequivocal ... when attacking any fleck of creationism in the schools.
So why are they persistently cryptic about the deeper problems with
evolution?
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 8 May 1995 19:18:23 -0500
Message-ID: <3omccf$507@dawn.mmm.com>
Chris Colby writes:
>[I]f Williams is an evolutionary biologist and doesn't
>cover up [Haldane's dilemma]; how can you claim evolutionary
>biologists have covered it up for decades?
I claim it, because it's true. For decades, the problem was pronounced
solved. The mechanics of the problem were taught cryptically, and rarely,
and fell into obscurity. Well over a dozen phony "solutions" lasted in
the technical literature for decades, with little effort to remove them,
even though many of the solutions were silly. Even the major solutions
(soft selection and truncation selection) were mentioned fleetingly in the
textbooks and were not merged into a coherent evolutionary model. That is
my claim, and the recent quote from Williams confirms it:
"In my opinion [Haldane's dilemma] was never solved, by [Bruce] Wallace
[and soft selection] or anyone else. It merely faded away, because people
got interested in other things. They must have assumed that the true
resolution lay somewhere in the welter of suggestions made by one or more
of the distinguished population geneticists who had participated in the
discussion."
Williams is the exception to the rule. He is the *only* leading
evolutionist to state that Haldane's dilemma remains unsolved. That is
not a ringing vindication of evolutionists. On the contrary, it is an
admission of a four-decade *failure* to teach and pursue the truth.
Chris uses the words "cover up", which I did not use. I do not think
the obfuscation of Haldane's dilemma was intentional, but then again it
was not a random accident of nature. Evolutionists did it, and it is for
them to explain.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 9 May 1995 13:43:42 -0500
Message-ID: <3ood4u$krr@dawn.mmm.com>
*** Haldane's Dilemma is not invalidated by genetic variation ***
Andy Peter's writes:
>In his latest post, ReMine offers nothing but flat nay-saying of my
>points: he continues to make no effort to address any specific argument
>I've made.
Wrong. Andy has (somehow) missed the fact that I turned his own argument
against him. His mutation load argument is ***his*** argument. He finds
it ***totally*** convincing. Therefore, he cannot dodge its conclusion:
The standard model of genetic evolution -- the one featured in all
evolution books -- is clearly false and massively contradicts
macroevolution. Andy keeps trying to dodge his own conclusion.
By Andy's own argument -- using the standard model of genetic evolution --
species' are so full of harmful mutations that they could not survive even
two generations, and would go extinct. Andy challenges me to overturn his
argument. Why should I? Andy dug himself a hole. Why should I dig him
out?
I repeatedly asked Andy for literature references that make the same
argument he does. I know of none, and he has not supplied any.
I also supplied a reference (G.C. Williams, 1992) where Andy's identical
same mutation load analysis is employed to reach a conclusion opposite
from Andy's. It is used to show the "severity of Haldane's dilemma." To
that, Andy offers no comment.
Andy writes:
>ReMine's arguments ... have basically been comprised
>of insinuations that: [1] all variation is neutral
Wrong. I never said all variation is neutral. I said that neutral
variation neither helps, nor harms, nor invalidates Haldane's dilemma (or
Andy's mutation load problem for that matter). Moreover, I have *agreed*
with Andy that the level of harmful mutation is quite high. This is a
serious problem for *any* model of genetic evolution.
>... and [ReMine makes] insinuations that selection doesn't
>happen in the real world.
Wrong. I never said that selection (ie. differential survival) doesn't
happen. I said that Andy's "Natural selection happens!" argument hid the
central issues.
******
Andy's post is another repetitious re-statement of his previous posts,
while being completely non-responsive toward my arguments. It seems the
discussion has reached an impasse, so I will state the conclusion again,
for those readers who may have arrived late.
Andy's "genetic variation" argument does not invalidate Haldane's Dilemma,
as Andy once claimed it had. Rather it invalidates the standard model of
genetic evolution. The cost of substitution and the cost of harmful
mutation are unavoidable, and they remain serious problems no matter what
genetic model you choose. These problems did not die with Haldane, or
with Andy's refutation of the standard model. The problems remain
unsolved.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 10 May 1995 01:11:53 -0500
Message-ID: <3oplf9$ma8@dawn.mmm.com>
*** Haldane's Dilemma and soft selection ***
Chris Colby offers soft selection as "at least part of the solution" to
Haldane's Dilemma. (He's already hedging away from previous statements.)
Chris's discussion follows that of Bruce Wallace, the originator of the
"soft selection" concept. While the following criticisms are directed at
Chris's text, they apply to Wallace's material also, as they share the
same errors.
******
The cost of substitution is often equated (dubiously) with genetic load,
and Chris follows that tradition. Genetic load focuses on selective
value, also known as "fitness". Thus, discussions of cost are diverted
into discussions of load, which are diverted in discussions of fitness,
which are diverted into discussions of:
"optimal" versus "average" fitness;
"absolute" versus "relative" fitness;
scaling of fitness;
normalizing fitness values to 1;
whether the most fit (or the least fit) organism
should be normalized to a fitness of 1;
and the normalizing of population sizes.
(And that's just for starters. That's before we even
try to physically interpret the concept of load.)
Because of these confusion factors, discussions of load are frequently
opaque, even to evolutionary specialists. Chris follows that tradition
also. For sake of documentation, here are examples. Notice the
frequent, confusing focus on "fitness." (Notice also the incorrect use
of the term "cost.") Don't get lost, I'll explain things down below:
<Begin quote from Chris Colby's post.>
In [hard selection], the presence of relatively more fit alleles decreases
the absolute fitness of other alleles. ....
[In hard selection], the appearance of more relatively fit alleles lowered
the _absolute_ fitness of other alleles. Under [soft selection], +
alleles raise mean fitness because it is not scaled. Put another way, in
[hard selection], the appearance of more relatively fit alleles lowered
the _absolute_ fitness of other alleles. In [soft selection], this is not
the case. A less fit allele does not induce a cost to more fit alleles.
This is a more realistic way of modeling evolution (IMHO). Let's say a
type of invertebrate normally lays 500 eggs; would the old type that lays
500 eggs impose a cost that had to be paid by the 550 egg layers? Would
the presence of 550 egg layers lower the amount of eggs 500 egg layers
laid? It would have to in order to accord to the assumptions of [hard
selection]. But, you would not expect it to.
Formulated this way, one can see that the details of how fitness differs
at a loci is important to seeing if there is a cost imposed by a new
allele. Specifically, how much does the new allele lower the absolute
fitness of the other allele(s) at that locus.
<End quote>
******
I prefer the term "cost of substitution" because it is clear, and it
immediately focuses your attention on the central issue -- substitution.
Substitution has a cost. If a species is to substitute a trait, then
there is a cost.
*
* Here is the rule:
*
* If a trait is be substituted into a population,
* then the trait must go from 'few' to 'many',
* then there is a cost,
* and it must be "paid"
* with reproductive excess of the survivors.
* There is no way around it.
* Absolutely, ...
* positively, ...
* no exceptions.
*
The issues are: substitution; growth "from few to many"; and reproductive
excess. "Selection" and "fitness" are virtually irrelevant by
comparison.
For example, suppose the population is all red, and there is a mutation
that turns one individual blue, and that blue-trait is substituted into
the population. The fate (and fitness) of the reds is almost irrelevant.
It is the reproductive excess of the survivors (in this case the blues)
that counts. The blues must go from 'few' to 'many', and the speed is
limited by their reproductive excess. The rate of substitution is
limited, because the reproductive excess is limited.
******
Chris's post uses the alternative term "cost of selection" which
misdirects our attention onto selection. That takes our eyes off the
central issues. He discusses "fitness" abundantly, (and confusingly).
***But he never once talks about actual substitutions.*** He talks about
the "presence" or "absence" of alleles, but never their substitutions!
In fact, his post implicitly *disallows* them. Here is an example:
>Would the (absolute) fitness of the less fit type increase
>if the more fit type went away (i.e. would it produce more
>offspring that survive to reproduce)? If yes, selection
>is soft and Haldane's cost of selection goes away.
To show Chris's mistake let me fill in the missing pieces. Imagine a
population, and a beneficial mutation occurs that creates a more fit
individual. The more fit type then starts surviving better than the
original type. In other words, a substitution is underway. This is the
situation that Chris is charged to explain. But Chris says, "if the more
fit type went away" then the "cost of selection goes away." Of course
the cost goes away. But so does the substitution!!! No substitution =
"no cost of substitution." Chris reaches the conclusion of "no cost" by
implicitly *disallowing* substitution.
Here is another example:
>The key is reproductive excess. If, in the absence of the more fit
>genotype, the less fit genotype can produce enough offspring to
>replace itself; selection is soft and there is no cost to selection.
Chris says, "in the absence of the more fit genotype ... there is no cost
to selection." Chris implicitly *disallowed* substitution in order to
conclude "no cost."
Chris writes:
>Under soft selection, Haldane's cost is an artifact ...
>not a real biological phenomena.
Chris is mistaken. There is absolutely no escaping the cost of
substitution, there is always a cost and it must be paid. That mistake
is common, even among evolutionary specialists, because they do not
understand the physical *meaning* of cost. The concept was garbled by
interpreting it as a genetic load.
******
That explains how soft selection was mis-sold. Next I explain why soft
selection cannot lower the cost of substitution.
Soft selection takes the standard model of genetic evolution and merely
emphasizes frequency-dependent and density-dependent selection. These
can alter the selective value of a given beneficial mutation depending on
its frequency or the population density. In other words, soft selection
allows the selective advantage of a mutation to *vary* as the
substitution proceeds.
However, the cost of substitution is fairly constant for selective
advantages of 10 percent or less, which covers the vast majority of cases
in evolution. So the selective advantage of a mutation can vary all over
the place, as long as it stays below 10 percent the cost of substitution
remains substantially ***unaffected.***
What happens if the selective advantage goes above 10 percent? Then the
cost rises *rapidly,* which aggravates Haldane's Dilemma.
Thus, soft selection might vary the selective value, but that cannot
plausibly reduce the cost of substitution. Altogether, soft selection
cannot reduce Haldane's Dilemma.
******
Two more points I've made in other posts:
(1) Evolutionist G.C. Williams recently denounced soft selection as a
non-solution to Haldane's dilemma. He discusses that in his book.
(2) Andy Peter's mutation load argument invalidates soft selection just
as thoroughly as it invalidates the standard model of evolutionary
genetics. There is precious little difference between the two models,
and *no difference* in their inability to eliminate harmful mutation. By
Andy's own argument, soft selection would lead to error catastrophe and
extinction. By Andy's own argument, soft selection contradicts
macroevolution.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 10 May 1995 02:26:28 -0500
Message-ID: <3oppr4$n6v@dawn.mmm.com>
*** Haldane's Dilemma and punctuated equilibria ***
Gould feels the punk eek model applies to hominid and human evolution.
He also estimates that the typical species spends at least 90 percent of
its time in stasis where little or no change occurs. Such morphological
stasis probably indicates genetic stasis.
(Thus, if species are in stasis 90% of the time, we must deduct 90% off
of the number of substitutions. This leaves 167 beneficial nucleotide
substitutions to explain the origin of humans from 10 million years ago.)
Gould acknowledged that stasis put a constraint on modes of genetic
evolution. He wrote:
"[Punctuated equilibria] exerts some constraint upon
genetic modes. Gradual and sequential substitution
of genes will not be a good model for the origin of
higher taxa if stasis be prevalent." (Gould, *New
Scientist*, April 15, 1982, p 138)
A few years later, punctuationists made a subtle attempt to 'decouple'
genetic evolution from morphological evolution, as though the two are not
closely related. (eg. Gould, *Paleobiology*, Vol. 11, 1985, p 7)
Possibly a geneticist had made them aware of Haldane's Dilemma, and the
decoupling was suggested as a way to avoid amplifying the problem. If
that is true, punctuationists have not actually said so. Haldane's
Dilemma has remained the trade secret of evolutionary geneticists.
As I point out in my book, there is a squeeze taking place in the theory
of evolutionary genetics. Many different factors are all causing problems
for the evolutionary model. The plausible beneficial substitution rate is
too low. Even the expressed neutral substitution rate is too low.
Organisms are close to (if not within) error catastrophe. The mutation
load is too high. Beneficial mutations are too rare. The genome is too
inert. Etcetera. To save the situation, evolutionists are reaching ever
farther into their theoretical smorgasbord, and pulling out ever more
far-fetched fixes. For example, they now attempt to ***convert*** neutral
mutations to beneficial mutations, without paying any theoretical price.
The attempt to decouple morphological evolution from genetic evolution --
as though the two are not closely related -- is another glaring example.
******
Andy Peters writes:
>I'd also like to point out another problem with
>ReMine's "Punk Eek makes the problem worse" schtick.
>The most convincing proposed driving forces behind the
>peripheral isolate mechanism for punk eek are not
>individual-level selection (which is what ReMine says
>imposes a demographic cost) but drift followed by
>deme- (subpopulation-) level selection. Neither
>of these mechanisms leads, under any "standard
>models", to a demographic cost.
Andy emphasizes the importance of 1) genetic drift, 2) group selection,
and 3) small population sizes. Let me take these in order.
1) Genetic drift is an opponent, not a friend, to selection. Genetic
drift (especially in small populations) can rapidly overpower and
eliminate beneficial mutations (in fact this is by far the most frequent
outcome). Also it can substitute harmful mutations into a population (in
fact this is the frequent outcome of inbreeding in small populations or
demes). You might get "rapid change" in small populations, but overall
it won't be beneficial.
2) Group selection (or deme-selection) -- which is different from
individual-level selection -- is extremely costly. That's why it fell
into disfavor in the sixties.
John Maynard Smith writes:
"Thus the maintenance of a characteristic favourable only to the group
requires N times as many selective deaths as the maintenance of an
individually favourable characteristic, where N is the number of
individuals in a reproductively isolated group." (1972, p 116)
3) Rapid evolution within demes or small populations, is a disguised way
of claiming that the beneficial mutation rate is exceedingly, implausibly
high -- a rate even higher than the rate of lethal mutation. This
argument was used by evolutionist Motoo Kimura as a limitation to the
speed of beneficial substitution.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 11 May 1995 00:31:57 -0500
Message-ID: <3os7gd$dru@dawn.mmm.com>
Andy Peter's latest post repeatedly misrepresents me and our debate.
Andy still continually tries to pin the problems on Haldane and bury
them. He keeps referring to the standard model of genetic evolution as
"Haldane's assumptions," yet he has never shown any difference between
the two. To clarify Andy's statements just plug in "standard model" at
the appropriate places. I will uniformly make that replacement below.
******
Andy now focuses even more intently on "selection occurs," and "selection
happens."
>[My mutation load argument showed that] selection occurs.
>In other words, the data show unambiguously that
>natural selection does not impose a demographic cost.
>Since [Walter] has failed to specifically address either
>the premises or the conclusions of my argument, we have
>to conclude that he agrees with it.
>The evidence shows that selection happens; therefore it's [the
>standard model], rather than selection itself, which are invalid.
>ReMine has repeatedly tried to turn my argument around: he
>wishes to imply that invalidating [the standard model]
>invalidates selection and macroevolution altogether.
Andy keeps diverting this into a debate over whether "selection
happens." (his words) That is misdirection. It's like shouting, "There
are differences in survival and reproduction!" It may be true, but it's
not the issue, it never was, and Andy's focus on "selection happens"
merely evades the real issues -- which are reproductive cost, error
catastrophe, and extinction.
Andy misrepresented me. I never said anything about invalidating
selection. Also, I never said anything about invalidating
macroevolution. I said the standard model of genetic evolution
*contradicts* macroevolution, and one of them must go.
Substitution, segregation, harmful mutation, and random death all have
reproductive costs that are *unavoidable*. If a species cannot plausibly
pay the reproductive costs required by a given evolutionary scenario,
then that scenario is not plausible. Andy's focus on "selection happens"
is both silly and evasive.
******
>Now, [ReMine's] point is that [the standard] model of genetic
>evolution leads to the prediction of a cost and - hold on to your hats -
>that the evil evolutionists still promote that model in the textbooks.
>I've contended that that's the way science works: the standard model is
>the most general, and therefore it's perfectly legitimate to teach it as
>such.
Andy continues to misrepresent the textbook issue. The problem is not
that they taught the standard model. The problem is that evolutionary
geneticists knew for decades of serious deep contradictions with that
model ***and did not reveal them***. That is the scandal I keep pointing
out. If Andy believes his own mutation load argument, then he must shout
those contradictions -- clearly -- right along side me. And stop trying
to bury them with Haldane.
******
>The fact is, that the data show that selection
>doesn't impose a demographic cost.
The data does not show that. The cost of harmful mutation -- the central
focus of Andy's mutation load argument -- is ***absolutely
unavoidable***, no matter what genetic/selection model you choose.
Harmful mutation ***always*** imposes a cost. And that cost can
potentially become overwhelming in *any* genetic/selection model, thereby
leading to error catastrophe and eventually extinction. The cost exists,
it is unavoidable, and it cannot be brushed aside, as Andy attempts to
do. Andy has not once addressed this issue. He just keeps asserting,
over and over, that the data proves there is no cost. He doesn't
understand the concept of reproductive cost.
>[D]espite ReMine's trumpeting of the fact that he's failed
>to refute my conclusion that selection doesn't impose a cost, he's back to
>insisting that it does. What amazing feats of logical acrobatics! ReMine
>should join a rhetorical circus!
Again Andy misrepresents the situation. His mutation load argument shows
that harmful mutation *does impose* a cost -- a huge cost. In fact, the
cost is so high that -- under the standard model -- species couldn't bear
the cost for even two generations, and would go extinct. The cost is so
high that the standard model of evolutionary genetics is shown to
massively contradict the requirements of macroevolution. The cost is so
high that evolutionists must find themselves another model that can
reduce the costs better. The cost problems haven't gone away. They must
still be solved.
******
I repeatedly asked Andy for literature references that document his
argument against Haldane's Dilemma. He never supplied one. Then I
supplied an example, G.C. Williams, who uses Andy's same mutation load
analysis to show "the severity of Haldane's Dilemma." Andy now lamely
claims his own argument "is the argument put forth by Williams, and it's
the argument I've been putting forth all along." That is outrageously
false. Williams and Andy reached *opposite* conclusions.
******
To summarize: Andy's mutation load argument shows that the standard
model of genetic evolution leads to error catastrophe and rapid
extinction. This massively contradicts the requirements of
macroevolution. And evolutionists have yet to settle on an alternative
model that could solve the problem.
Our debate on "genetic variation versus Haldane's Dilemma" has now
solidified, and become repetitious. I am happy to engage in honest
discourse. But Andy is not offering anything new, is not engaging my
arguments, and now continually seems content to misrepresent me and our
situation. I cannot spend my life correcting Andy's many
misrepresentations and evasions. It is time to move on.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 11 May 1995 17:49:41 -0500
Message-ID: <3ou4a5$mi2@dawn.mmm.com>
*** Haldane's Dilemma versus soft selection ***
This is my response to Chris Colby's latest post.
I wrote:
>Chris Colby offers soft selection as "at least part of the solution" to
>Haldane's Dilemma. (He's already hedging away from previous statements.)
Chris responded:
>Uh, which statements have I been hedging away from, Walt?
Chris just admitted something. A while back I began our discussion with
the following claim. Haldane's Dilemma -- both the problem and its
supposed solutions -- are substantially under-discussed in evolutionary
literature. The problem is stated cryptically, and rarely. The supposed
solutions (such as soft selection and truncation selection) are also
discussed little; are not tied-in as essential to evolution; and are not
joined to a coherent model of evolution. As examples I listed a variety
of population genetics textbooks where neither the problem nor the
solutions are discussed. A few of those books obliquely 'mention' soft
selection in one or two sentences in the entire book, and I identified
those at the time. Chris Colby then claimed soft selection as *a*
solution (if not *the* solution) to Haldane's dilemma. And he actually
quibbled over whether "mentioning" soft selection amounts to a textbook
discussion.
Now Chris has backed off. He says soft selection is at least "part of
the solution". That means at least "part of the problem" still remains.
But ***the textbooks didn't say that.*** On the few occasions when
textbooks discuss Haldane's Dilemma, they present it as a non-problem, a
solved problem. But if even a "part" of the problem remains, then the
textbooks should have said so. That is the scandal I keep documenting.
Either Haldane's Dilemma was solved or it wasn't. Chris now admits *he
knew* that soft selection may not have been a complete solution. I keep
pointing out, it is the persistent, long-term disparity -- between what
evolutionists *knew* and what they put in the textbooks -- that is so
damning.
******
Chris Colby writes:
>{In hard selection] the less fit type had a fixed
>mortality rate with a probability of survival of less
>than one. One way of looking at this is the less fit
>type, by itself, could not "fill" its environment up to
>the carrying capacity. The more fit type had to fill in
>the extra "slots" in the environment -- _that_ was the cost.
No, not quite. The cost of substitution is not from "filling the extra
slots" in the environment. Rather, the cost is from going 'few' to
'many'. It doesn't matter whether there are many small 'extra slots', or
whether there is one big wide gapping ecological hole. The cost of
substitution is the same. It's the growth from 'few' to 'many' that's
important.
Let us examine a single substitution. If the newly mutated trait starts
out with N_start copies in the population and spreads until there are
N_end copies (N_start < N_end), then the absolute minimum cost, under any
and all conditions, under any and all genetic/selection models, is given by:
Cost of substitution = Ln(N_end / N_start)
Where Ln() is the natural logarithm.
That is the formula Haldane derived. I'll come back to that formula in a
moment.
The cost depends on the *speed* of the substitution. The faster the
speed, the higher the cost. The substituted trait can be "beneficial,"
"neutral," (or conceivably even "harmful") but that makes no difference
to the cost of the substitution. The speed is what really counts. As
the trait goes from 'few' to 'many': the lower the reproductive excess,
the lower the speed. The two are linked together. Think of a new rare
trait in a population:
lower reproductive excess = lower speed of growth
That relationship is direct, tight, and unavoidable.
Haldane's formula can even be derived without any reference to selective
values or specific genetic models -- and I do just that in my book, _The
Biotic Message_. I show how the cost varies with the *speed* of the
substitution. [Note: Because the derivation is free of confusion
factors (such as selective values, fitness normalization, population size
normalization, and specific genetic models) the derivation is especially
transparent and convincing. I highly recommend such a derivation for
giving insight into the meaning of cost, and its unavoidability. That is
missing from evolutionary genetics texts today.]
Of course, the speed of substitution is ordinarily highly correlated with
selective advantage. So we ordinarily think the 'selective advantage'
directly determines the cost, and that's a pretty fair way of thinking
about it. But I point out, in passing, that *every* substitution incurs
a reproductive cost -- beneficial, neutral, or harmful -- no exceptions.
The selective value is not the focus. Rather 'few' to 'many' is the
primary focus, and speed is the secondary focus, (in nature selective
value is highly correlated with speed).
Obtaining the minimum cost (given in Haldane's formula above) requires
that the substitution proceed infinitely slowly. However, the cost is up
slightly for speeds corresponding to a selective advantage of ten
percent, which covers the vast majority of cases in evolution. For
faster speeds (ie. when the selective advantage is greater than ten
percent), the cost goes up *rapidly*.
In other words, the cost remains relatively constant for selective
advantages up to ten percent. Thus, even if soft selection could
actually vary the selective value, it would leave the cost substantially
***unaffected.*** I remind Chris of this point, as he avoided it in his
latest post. Soft selection, even if it operated, could not reduce the
cost noticeably.
Let's take an even closer look. Chris's discussion of soft selection
focused on dominant mutations -- which are the least costly to
substitute. When evolutionists (including Haldane) calculated the cost
of a dominant substitution they always used the above formula which
already gave them the absolute lowest cost for their calculations. In
effect, they calculated the lowest conceivable cost -- the cost for a
dominant substitution of infinitely small selective advantage. For a
higher selective advantage the cost can only go up. Despite using the
theoretically lowest cost, evolutionists were still confronted with
Haldane's Dilemma. Now here's the point. Varying the selective
values can only make the costs go up *above* what originally caused
Haldane's Dilemma. When soft selection is added to the picture, it can
only aggravate the problem.
******
Let me clarify something. Various evolutionary processes causes genetic
death -- or terminations of a genetic line. If the population is not to
shrink to extinction, the genetic deaths must be replenished by genetic
births. And these are supplied by excess reproduction.
A given species has a limited total reproductive excess, this is the
reproduction over and above what is required for sustaining the genetic
lines from generation to generation.
The total reproductive excess pays all the costs of doing evolutionary
business. Some of it pays the cost of eliminating harmful mutation.
Some pays the cost of segregation and heterosis. Some pays the cost of
random death. Some pays for balancing selection. Etcetera. A portion
pays the cost of substitution. The literature on Haldane's Dilemma often
calls this the reproductive excess (and is how I used the term above).
It is the portion of the total reproductive excess that pays for
substitutions. Haldane estimated that in a typical mammalian species
this is ten percent. That is, the species could reproduce an excess of
ten percent of its total population size each generation, and devote it
specifically, and with perfect efficiency, to paying the costs of
substitution.
Let me briefly summarize from my previous posts. Haldane estimated an
average cost of 30, paid in installments of 0.1 per generation. Which
means it takes 300 generations to pay for one substitution. Which is too
slow to explain human evolution.
******
Here is another argument against soft selection.
(1) Chris claims that under hard selection "the less fit type, by itself,
could not 'fill' its environment up to the carrying capacity." therefore
the cost is high and the substitution rate is slow.
(2) Chris also claims that "Under soft selection, ... the less fit types
could fill the environment to carrying capacity" therefore there is lower
cost (no cost?) and the substitution rate is faster, according to Chris.
(3) Thus, Chris is claiming the absurd. He claims that when the
environment is ***fuller*** of the less fit type -- all using up the
limited resources -- then the more fit type can excel and be substituted
***faster.*** That is absurd. In effect, Chris claims it is the
abundant presence of the less fit type that makes the more fit type grow
faster. There is no physical basis for such a speed up. The new fit
type must go from 'few' to 'many' -- and the less fit type can only get
in the way. Soft selection cannot speed up the substitutions.
******
Andy Peters (a leading evolutionist here on talk.origins) makes the
following mutation load argument. Based on the standard model of genetic
evolution, and the available data -- all species have too many harmful
mutations and would go extinct within two generations. Since that hasn't
happened, it invalidates the standard model -- the one model prominently
displayed in all evolution books.
Andy's argument invalidates soft selection just as effectively as it
invalidates the standard model. Soft selection cannot eliminate harmful
mutation any better than the standard model. The two are identical in
that respect, and the mutation load argument impinges on both with equal
force. No one, neither Chris Colby nor Andy Peters, has addressed this
issue yet.
******
>[T]here is still a limit to the rate of evolution. And, as
>Walt correctly points out, it is tied to reproductive excess.
>The rate of evolution is limited by how many excess organisms
>are produced by the more fit type and how competitive they are.
>I think it should be easy to see that a type that produces many
>"excess" and highly competitive offspring would increase in
>frequency faster than a type that produced fewer (and less
>competitively superior) offspring.
Chris and I are in substantial agreement there.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 11 May 1995 18:54:32 -0500
Message-ID: <3ou83o$p9s@dawn.mmm.com>
I've have repeatedly shown a great disparity between what evolutionary
geneticists knew and what they put in the textbooks. Deep problems with
evolutionary genetics were persistently obscured for decades.
But there are several steps in the acceptance of a new idea.
Step 1: Fight it like mad. That's what the talk.origins crowd
predominantly did with this issue, especially at first.
Step 2: Makes excuses. We've heard a bunch of those. Here is the latest
one.
>[Evolutionists] cannot spend [their] life correcting
>what I'm sure seem to you to be many misrepresentations
>and evasions when new knowledge is arriving by the
>freightcar on a daily basis." <Andrew Delling>
(Note: In the original, it was a backhanded attempt
at making an excuse: evolutionists are just too busy
dealing with the data, why waste time.)
That gradually gave way to
Step 3: Misrepresent it, and ridicule it. For example, misrepresent it
as an "conspiracy theory." (See my earlier post, where I explicitly
denied that charge.)
>Do you think that your pathetic attempts to suggest a
>conspiracy theory that has apparently covered decades
>and thousands of biologists is going to sound any more
>cogent to the general public than the average militia
>loony shouting about the Trilateral Commission? <Andrew Delling>
The leads to
Step 4: Poo-poo it as unimportant. Trivialize it.
>But I fear that what we have been witnessing is a descent
>into not madness but compulsion. Walter, in the
>great scheme of things whether or not Haldane's Dilemma has
>received adequate press coverage just doesn't make any
>difference. You can continue to ride your hobby horse,
>but you begin to look more and more like the old woman
>with the 16 foot wide string ball made from spare threads
>recovered during her job stuffing "Inspected by #16" into
>Levis. Interesting, but bizarre.
>....
>It is time to move on. <Andrew Delling>
The final step is more awkward, and will take much longer
Step 4: Proclaim that "We evolutionists knew it all along!"
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 12 May 1995 18:19:47 -0500
Message-ID: <3p0qej$5kb@dawn.mmm.com>
Andy Peters made a new argument, a fair one. As it is not merely a
repeat of our previous debate I will answer it.
I had made this claim: By Andy's own mutation load argument, the
standard model of genetic evolution contradicts macroevolution, and one
of them must go.
I did not discuss the option that 'BOTH of them might go', because:
1) That option was sufficiently obvious.
2) Andy's argument does not force (or even suggest) that option, and
Andy would inherently reject that option anyway. So why waste time with
it.
3) I have always said: ***It is Andy's argument.*** It is not mine, it
is Andy's. So my perspective and options are not pertinent. Why waste
time with them when they are a diversion from the central issues? My
points are much more forceful when I state them cleanly in Andy's terms
on Andy's turf. He already finds his argument totally compelling, so he
cannot deny its conclusion. I have merely worked to make his argument
-- and its conclusion -- explicit and clear. He has no right for
complaint.
Our disagreement is that Andy tends to leave out the key part, and state
the conclusion in a manner that obscures the source and depth of the
problem. Andy's conclusion should be given like this: The available
data on harmful mutation, together with the standard model of genetic
evolution, predicts that all species would go extinct within two
generations. Anyone can understand that, and interpret it for
themselves. Evolutionists should take the full truth of Andy's argument
*and teach it.*
******
Andy claims the data shows that harmful mutation does not impose a cost.
That is the same as saying that harmful mutation doesn't cause any
deaths. That is clearly untrue. Harmful mutation causes deaths. Can
Andy show otherwise? The cost is real and unavoidable -- so long as
there is harmful mutation, the cost is never zero.
Also, higher levels of harmful mutation cause greater numbers of deaths.
Can Andy show otherwise? The cost increases with higher levels of
harmful mutation.
Also, those points are true for *every* genetic model. Pick one.
Harmful mutation causes genetic deaths. As the harmful mutation rate
increases, more of the species' reproductive excess is required simply to
replace the genetic deaths. Right at the brink of error catastrophe
*all* of the reproductive excess is required, just to fend off
deterioration. With a yet higher harmful mutation rate, the population
goes into error catastrophe, a steady genetic deterioration. The rate of
deterioration depends on how far the harmful mutation goes beyond the
error catastrophe threshold. Can Andy show any genetic model where these
points aren't true? The cost of harmful mutation is an issue for *every*
genetic model.
Andy's mutation load argument did not show there is no cost. Rather, it
showed a *contradiction* between the standard model and the data. So we
reject the standard model, in favor of the data. But that doesn't
eliminate the cost of harmful mutation. A cost still remains. However
the mathematics of the standard model (such as the usual mutation load
calculations) are rejected along with the standard model -- so Andy no
longer has a way to calculate the cost. A cost still exists, it's just
that (for the moment, until he finds a new model) we have lost the
theoretical structure with which to calculate the cost. In other words,
Andy calculated a cost of harmful mutation, the cost is too high for
species to pay, and that specific cost calculation gets thrown out the
door, right along with the standard model. But a cost still exists. (I
can't calculate the cost for him, because he hasn't identified an
evolutionary model to base it on. So I am left saying "The cost still
exists! It's unavoidable!")
Without a model, Andy's estimate of the cost is very crude. Like this:
Selection happens, species exist, they aren't extinct, so obviously there
is no cost. He hasn't shown there is no cost, and he hasn't hardly begun
to measure it.
For example, species might be in a slow genetic deterioration -- and
nothing stated in Andy's argument or data indicates otherwise. He sees
the deterioration is not fast. Fine, it might be slow.
(Note: This is where Andy drags the standard model back in again, despite
his having thrown it out already. He uses it to show that the extinction
*must* be exceedingly rapid, within two generations. And he challenges me
to refute his argument. But he already accepts the invalidity of the
standard model, so the 'rapidity of extinction' gets rejected right along
with the standard model. This means that populations might actually be in
a *slow* genetic deterioration. Andy's argument and data have not
eliminated this possibility.)
Even if it turns out that species are not in genetic deterioration, there
is still a cost of harmful mutation. How large is that cost? How much
of the species' reproductive excess must go to paying it? Those are key
questions (and readily answerable within the standard model). But Andy
can hardly begin to answer them. He has no idea how large the cost is,
he has not measured it, and without a model he lacks the ability to
calculate it. He merely observes that species are not in rapid
deterioration.
Andy's mutation load argument invalidates the standard model. So
evolutionists must seek another model that can cope with harmful mutation
better. Once they have it, then we can compare it with the data on
harmful mutation, just like Andy did. After that, we can check its
ability to reduce the cost of substitution and solve Haldane's Dilemma.
In the mean time we must teach the truth as best we can see it. The best
evolutionary model (by Andy's own admission) is the standard model --
which predicts all species would go extinct within two generations.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 12 May 1995 19:40:29 -0500
Message-ID: <3p0v5t$8oj@dawn.mmm.com>
*** Haldane's Dilemma and genetic drift ***
I said previously that cost and load are substantially different. They
use very different formulas. Also they have different ranges. Haldane
estimated the average cost of substitution is 30. A load, however, is
always a number between 0 and 1 inclusive. They also have different
units. Load is a unitless number, and cost is not. And they have
different physical interpretations.
Those interpretations are especially different for the case of neutral
substitutions. Nothing can go from 'few' to 'many' without paying a
reproductive cost. This is the cost of substitution, and it applies even
to the substitution of neutral mutations. The cost is never zero.
What is the genetic load of a neutral substitution? Since a neutral
substitution does not affect selective value, many evolutionists
(especially neutralists) claimed that neutral substitutions have zero
load, and therefore zero cost. Which is untrue. Neutral substitutions
have zero load, but *non-zero* cost. Cost and load are different.
The cost of a neutral substitution is an issue that evolutionists have
traditionally confused. In this case, it merely decreased understanding,
and did not adversely affect our view of evolution. The neutralists
concluded (falsely) that neutral substitutions have no cost -- and
therefore their rate of substitution is not cost limited, and is limited
instead by the neutral mutation rate. They happened to reach the correct
conclusion. But they left the cost issue confused.
It turns out that neutral substitutions are very costly. But they have
sources of payment and sources of cost reduction that are not available
to beneficial substitutions. This whole matter is unraveled and
explained in my book. In all it gives a uniform picture of cost/payment
analysis and how it applies. It clears up several aspects of the neutral
theory that historically remained mysterious.
I won't go into it further in this thread, as neutral substitutions are
not the issue.
******
Andy Peters writes:
>[Genetic drift] doesn't directly decrease the cost of
>selection, though; It only does so indirectly, by causing (drift-induced)
>increases (or, unfortunately, decreases) in frequency that don't carry a
>demographic cost. Essentially, the population "cheats" by covering a
>bunch of allele-frequency ground by drift (no cost) rather than selection
Genetic drift *decreases* the beneficial substitution rate in several ways:
1) Beneficial mutations are almost always ***eliminated*** by genetic
drift. For example, a mutation with a half-percent selective advantage is
eliminated 99 times out of a hundred, and doesn't participate in
evolution. (Most people don't know that.)
2) In small populations, genetic drift is stronger, and it overpowers
selection more easily. Population geneticists, like Motoo Kimura, showed
the following. Let the effective population size be N. When a trait has
a selective value of less than 1/N, then the trait appears to be
neutral. In effect, it no longer seems to have a beneficial advantage.
Mutations with a half-percent selective advantage appear to be neutral in
a population of 200. Their "beneficial" effect is overpowered by the
genetic 'noise' and sampling error that drives genetic drift. The problem
gets worse for mutations of lower selective advantage.
3) Though beneficial mutations are predominantly overpowered and
eliminated by drift, they don't exit peaceably. They put up a "fight".
They cause the genetic death of less endowed individuals, but they don't
successfully lead to a substitution. Thus, they incur a partial cost,
but no substitution. This effect *raises* the average cost of
substitution, and therefore slows the substitution rate. Kimura
calculated the size of this effect for a dominant mutation. (See my book
for details.)
(Note: Since someone asked, Haldane left for India in 1957, the year of
his paper on Haldane's Dilemma. He died there in 1964.)
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 14 May 1995 14:36:45 -0500
Message-ID: <3p5m4d$noj@dawn.mmm.com>
*** Haldane's Dilemma and soft selection ***
This is my response to Chris Colby's latest post on soft selection. He
has not addressed any of the arguments raised in my previous post, so I
would merely repeat my post again here, and add only new comments.
Chris writes:
>A new mutation can sweep to fixation in a single
>generation with lethal selection (a subset of hard selection.)
>Imagine a bacterial population of 1 million cells -- 999,999
>of these were Lac- (unable to utilize lactose) and one was Lac+.
>Now, if they were placed in an environment with lactose as the only
>carbon source, only the single Lac+ cell would survive (neglecting
>new mutants to Lac+). Substitution in a single step. But, the pop-
>ulation size would be low for a few generations until it grew back up
>to carrying capacity.
Using examples like Chris's, evolutionists frequently confuse Haldane's
Dilemma. They imagine a large population of bacteria placed in an
adverse environment (eg. penicillin, or lactose) where only one cell
survives. Evolutionists then say, "This was a substitution in a single
generation!" But they falsely used a single generation to represent all
evolution. Evolution cannot continue that way generation after
generation. Instead, it will take many generations for the population to
grow back to normal size, to a size where it is likely to receive another
beneficial mutation. This growth takes a while, ... and as always it is
limited by the species' reproductive excess. The substitution cycle
takes many generations. The cost of the one substitution takes many
generations to pay.
******
>The gist of Haldane's dilemma is that (hard) selection at numerous
>loci will drive a population to extinction. .... This is where
>Haldane got his estimate of the max rate of evolution from --
>too many substitutions and the population goes extinct.
Evolutionists frequently confuse Haldane's Dilemma by misrepresenting it
as a theory of extinction. With that confusion in place, they then
obscure the issue further, by saying, "Haldane's argument indicates that
if a population receives too many beneficial mutations, then it will go
extinct. But how can the substitution of beneficial mutations cause
extinction? Obviously there is something wrong with Haldane's
analysis." That false argument is quite old, yet still unchallenged by
evolutionists, and now almost universally propagated in those few
textbooks that discuss Haldane's Dilemma for more that a few paragraphs.
Haldane's Dilemma is not a theory of extinction. Rather, it is a
criterion of plausibility. If an evolutionary scenario requires a
substitution cost that species cannot plausibly pay, then the scenario is
not plausible. The cost of substitution places a limit on the rate of
evolution.
******
For sake of documentation, I include Chris's entire explanation of how
soft selection supposedly makes substitutions faster. (All emphasis and
notes are mine, to cue you to the key points.) Don't get lost. I'll
explain it down below.
<Begin quote from Chris's post.>
But, under soft selection, the rate of substitutions is higher. Imagine
a species of bird that lives on an island. In order to reproduce, males
must maintain a territory, of which there are 1 million on the island.
(We'll assume monogamy here just to keep the math simple.)
Initially, all males are equally competitive. Reproducing birds, on
average, leave four offspring, but only two ever reproduce because
there are only enough territories for 1 million males (and the
corresponding 1 million females).
But, then a mutation arises at one locus (we'll call the locus
A and the new mutant allele the +, old allele -). Let's say this
bird type (+/+ or +/- at that locus (i.e. + is dominant)) is
equal in every respect except to other birds except it is slightly
better at holding a territory.
***The + allele will sweep to fixation.***
[NOTE: Pardon my pun, but Chris is 'sweeping' the issue
under the rug. The *rate* at which the new mutation moves to
fixation IS THE CENTRAL ISSUE of the cost of substitution. There
is a cost of going from 'few' to 'many', and there are limits to
how fast it can be paid.]
In addition, ***the population size will not dip***
[NOTE: Chris makes much of the population size "not dipping",
but it does absolutely nothing to resolve Haldane's Dilemma. It
doesn't lower the cost of substitution.]
because -/- types will
fill any territory not filled by +/X types. Now imagine a second
locus, with a + and - allele. If this is again soft selected, it
can sweep to fixation with no drop in population size.
[NOTE: Next comes Chris's key argument, so watch closely here.]
This can be extended to any number of loci. Thus, soft selection
increases the number of substitutions that can occur (although each
individual substitution may take awhile.) This is what I have
been trying to get across to Walter.
<End quote>
Chris's argument is a disguised way of claiming that sexual reproduction
makes evolution go faster. But that notion is untrue. (Also, Andy
Peters and I previously refuted that notion to the satisfaction of all
present here.) Let me state the key part of Chris's argument again.
>This can be extended to any number of loci. Thus, soft selection
>increases the number of substitutions that can occur (although each
>individual substitution may take awhile.)
Chris acknowledges that "each individual substitution may take awhile" --
which is the key issue, and Chris never showed that it would occur faster
under soft selection. He didn't even try to show that.
The entire focus of his argument is that the substitutions "can be
extended to any number of loci." -- as though the substitutions are
occurring *in parallel.* That is where he sneaks in a reliance on sexual
reproduction. Chris's defense of soft selection has nothing to do with
soft selection. His defense is nothing more than relying on the false
illusion that sex speeds evolution. Sex actually ***slows*** evolution.
(See my book or my earlier posts for more details.)
In summary, Chris did not show how soft selection speeds an individual
substitution. (Which is the central issue.) Instead, he secretly
shifted his discussion over to the role of sex.
******
Chris makes an argument concerning a different issue -- fluctuations of
population size. For documentation sake, here it is with my notes inserted.
<Begin quote>
Let's say the population size is fluctuating. Every X years a big
hurricane roars through and kills 80% (or whatever) of the birds
at random. Would this effect the overall rate of substitution? I
think not.
[NOTE: Random death slows the substitution rate. It reduces the
population size, thereby directly reducing the rate at which the
*population* receives beneficial mutations. Also, random death imposes a
cost. That cost, in addition to the costs of segregation, harmful
mutation, balancing selection, and substitution, must be paid out of the
species reproductive excess. Because the available reproductive excess
is limited, higher levels of random death *slows* the substitution rate.
This is visible in computer simulations. Increased random death slows
the substitution rate, and brings on error catastrophe sooner.]
I haven't worked the math (I'll try once I have some
time), but here's a verbal description of why I think not. Take
a look at a single locus with two alleles (+ and -) again. Right
before the hurricane hits, they will be a certain frequency (let's
say .50/.50 for simplicity sake). The knocking down of pop. size
may alter this frequency due to sampling error. But, on average,
the amount of change will be zero. So, due to drift, the rate of
substitution will not change.
[NOTE: No, just because the *frequency* of that *particular*
allele hardly changes, that doesn't mean the overall
substitution rate remains unchanged. See my notes above and
below.]
Indeed, Kimura has shown that neutral substitutions occur at a
rate independent of pop. size.
[NOTE: Chris's reference to Kimura is a red-herring. It's
irrelevant. Kimura was talking about *neutral* substitutions,
which are rather different from the 'beneficial substitutions' Chris
is discussing.]
(The variance in amount of allele frequency change is greater
at smaller population sizes, although the average is zero.)
<End quote>
Chris chose his example to favor evolution in two subtle ways. (1) His
left out the population size, in effect implying an extremely large
population (approaching infinite). Random effects, such as random death
and genetic drift have less effect in large populations, where selection
tends to dominate the process. (2) Chris chose as his example a
beneficial mutation with a 50 percent frequency in the population. This
made the beneficial mutation substantially safe from the effects of
random death (except in the very smallest of populations).
Let me alter Chris's example to show what I mean. Let's say the
population is 1,000 (instead of infinite) and the beneficial mutation has
a frequency of half percent (instead of 50%). Now when the hurricane
randomly kills 80 percent of the population, there is a fair chance (33%)
it will completely eliminate the fledgling new beneficial mutation. Even
if not, then it will leave the mutation in such low numbers that ordinary
genetic drift will almost always (with a probability greater than 98%)
finish the elimination soon afterward.
Random events, such as random death and genetic drift work against evolution, especially in smaller populations where these
effects dominate over selection. Ask evolutionists about Haldane's
Dilemma and the cost of substitution, and they'll emphasize 'small
populations.' Ask them about the problems from genetic drift and random
death, and they'll emphasize 'large populations.' It's a smorgasbord.
The answer you get, depends on the question you ask. There is no
coherent model of how evolution is supposed to occur.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 14 May 1995 15:40:25 -0500
Message-ID: <3p5prp$q1c@dawn.mmm.com>
*** The cost of harmful mutation ***
Andy Peters writes:
>I apologize for the awkward wording; more precise wording would
>be, "rates of deleterious mutation have been demonstrated to be
>between 0.25 and 1.0 per genome per generation;" this is a rate,
>not a ratio. Thus, the average genome suffers between 0.25
>and 1 deleterious mutation per generation. Since in a diploid
>population each individual has two genomes, this means that
>each individual has on average 0.5 to 2 new [deleterious]
>mutations that its parents didn't have.
Andy says "each individual has on average 0.5 to 2 new [deleterious]
mutations that its parents didn't have." That would place many
populations precariously near or within error catastrophe, where harmful
mutations accumulate generation-to-generation faster than they can be
eliminated.
For example, with a rate of 2 new harmful mutations per progeny, a female
would need to conceive 14.8 progeny just to replace herself and her mate
with progeny that are defect free. And that is *before* random death,
non-genetic developmental malformations, segregation, balancing
selection, and substitutions take their toll on reproductive. Altogether
females would need to supply a superabundance of progeny just to remain
in the same evolutionary place and fend off error catastrophe. It is
questionable whether humans, primates, and many mammals have enough
reproductive excess to accomplish that.
The standard model of evolutionary genetics has no provisions for
efficiently removing harmful mutation. If the species cannot supply the
above reproductive excess, then error catastrophe is inevitable, and a
steady genetic deterioration occurs generation to generation. The
addition of sexual reproduction cannot solve the problem, because all it
does is mix the harmful mutations around the population.
For a harmful mutation rate of 1 per progeny, the above figure drops to
5.4 progeny, which is still a sizable bite out of human reproduction.
For a rate of 0.5 harmful mutations per progeny the figure drops to 2.56
progeny.
This argument reaches the same conclusion as Andy Peter's mutation load
argument. The data, together with the standard model of genetic
evolution, indicates that many species are within (or very near) error
catastrophe.
It also shows, once again, that harmful mutation has a cost, and it is
never zero. The cost must be paid out of the species' limited
reproductive excess.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 14 May 1995 17:26:05 -0500
Message-ID: <3p601t$r3@dawn.mmm.com>
Jamie Schrumph writes:
>Why is it only in the field of evolutionary biology do unanswered
>"dilemmas" cause such problems in the lay population? Fermat's Last
>Theorem took nearly 200 years to finally solve (at least I THINK it's
>been solved), yet no one claimed that mathematics was false.
One issue is that Haldane's Dilemma remains unsolved.
Another issue is that evolutionists said it was solved when it wasn't.
Yet another issue is that the supposed solutions were not held high, they
were not integrated into a coherent model of how genetic evolution is
supposed to occur. Instead, the entire problem (and its supposed
solutions) were obscured for decades.
We can understand that Fermat's Last Theorem remained unsolved for
centuries -- but at least mathematicians admitted it all along. What sets
evolutionists apart is that they did not truly teach and truly pursue the
truth. It wasn't intentional, but that doesn't change the facts.
That is commonplace in evolutionary theory. For example, Stephen Gould
revealed the large morphological gaps between fossil species as the "trade
secret of paleontology". But that fact had been known, among evolutionary
paleontologists, since nearly Darwin's time.
Or take Haeckel's recapitulation theory. Evolutionary embryologists knew
it was false by the 1920s, but it was still taught (even in public
schools) until 1977, when Gould sought to clear the matter once and for
all with his book, _Ontogeny and Phylogeny_. The disinformation still
even remains today as common folklore. We still frequently hear the false
claim that humans have "gill slits".
I continue to point out the broad, persistent, long-term disparity between
what professional evolutionists *knew* and how evolution was sold to the
public.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 14 May 1995 17:34:43 -0500
Message-ID: <3p60i3$15d@dawn.mmm.com>
*** Haldane's Dilemma and soft selection ***
This is my response to Andy Peters latest post on soft selection.
I challenged Andy with his own mutation load argument. I showed that
his own argument invalidates soft selection right along with the
standard model of evolution. Here are my words:
>Andy Peters (a leading evolutionist here on talk.origins)
>makes the following mutation load argument. Based on the
>standard model of genetic evolution, and the available
>data -- all species have too many harmful mutations and
>would go extinct within two generations. Since that hasn't
>happened, it invalidates the standard model -- the one
>model prominently displayed in all evolution books.
>
>Andy's argument invalidates soft selection just as
>effectively as it invalidates the standard model. Soft
>selection cannot eliminate harmful mutation any better
>than the standard model. The two are identical in that
>respect, and the mutation load argument impinges on
>both with equal force. No one, neither Chris Colby nor
>Andy Peters, has addressed this issue yet.
Andy responds:
>....
>If soft selection, rather than hard selection, is acting at
>every locus in a population in mutation-selection balance,
>then the population is replacing itself, by definition.
>Hence, soft selection doesn't impose a demographic cost,
>and fits with the observed genetics and population dynamics.
Andy's mutation load argument indicates that species are in error
catastrophe, and cannot get rid of harmful mutation fast enough. But he
now *assumes* the problem out of existence by assuming the population is
"in mutation-selection balance." That is just other wording to claim the
population is not in error catastrophe. Andy secretly assumed the
problem away, he didn't actually solve it.
Also, he says "the population is replacing itself, by definition.
Hence, soft selection doesn't impose a demographic cost..." Andy is
continuing with his false interpretation of cost. He believes (wrongly)
that if the population is merely "replacing itself" then there is no
cost. The cost of harmful mutation is over and above mere
self-replacement, and can be quite large. And its payment requires
*reproductive excess*, not mere self-replacement.
First Andy falsely defended the ability of soft selection to
eliminate harmful mutation. Next he turns around and dramatically
*de-emphasizes* its role in eliminating harmful mutation:
>my feeling is that soft selection is more likely to be
>important in cases of substitution than it is in
>mutation-selection balance.
Andy misses the point. Soft selection is *inherently* unable to solve his
mutation load argument. Combine soft selection with the standard model
in any manner you please. Emphasize one over the other, or take one to
the exclusion of the other, I don't care. They are all invalidated by
Andy's own mutation load argument. Andy completely avoided this issue in
his post.
******
Andy writes:
>Despite ReMine's assumptions to the contrary, I don't
>claim to know exactly what mechanisms reduce the cost.
Despite Andy's misleading wording, he is agreeing with me. These cost
problems remain unsolved, they are the trade secrets of evolutionary
genetics.
******
For the remainder of Andy's post he wanders around in the evolutionists'
theoretical smorgasbord, casually mentioning various mechanisms, such as
epistasis, "And, who knows, there may be other selective mechanisms out
there as well. [I]t's very unclear which selective mechanisms are
important in which situations (and the fact that the relative importances
are likely to be different under different circumstances, life-history
stages, etc.)"
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 17 May 1995 11:22:42 -0500
Message-ID: <3pd7si$504@dawn.mmm.com>
This is my response to Chris Colby's post titled "Haldane Speculations."
Chris writes:
>I thought about linkage. In Haldane's model, all the alleles were
>unlinked (either on different chromosomes or far apart on the
>same chromosome). Two beneficial alleles, if they were tightly
>linked, would only incur the cost of one substitution.
That scenario requires the fortuitous tight linking of two separate
beneficial mutations, so they are substituted as one, for the cost of one.
The link would have to be tight, otherwise sexual reproduction would
rapidly split the two mutations apart. We may call that idea the
"super-gene scenario", and it was debunked by Haldane himself. He showed
it cannot plausibly reduce the cost of substitution. I discuss the
debunking, and expand it slightly, in the appendix of my book.
******
>Also, what about spatial structure in a population? Or population
>subdivision? Imagine a population that fluctuated in size. When the
>population was large, it existed as a single population. When small,
>it existed as isolated clusters. (I think a lot of populations
>exist like this over time.) Haldane's limit (if we chose to
>accept it) only applies to a single population. If the population
>was split into N subpopulations, you could get N times as many
>substitutions amongst all the sub pops. When they came back
>together as a single population, the substituted alleles in
>each sub-pop. would be at intermediate frequency. Haldane's
>cost was greatest when alleles were very rare. The cost for
>substituting alleles that start out at a reasonable frequency
>was negligible.
Evolutionists are desperate for a solution to Haldane's Dilemma. They
simultaneously want the advantages of 'large' AND 'small' populations. So
they divide one large population in many small sub-populations. The
ingredient that turns "one large" into "many small" is lack of migration.
(And vice versa.) That transforms the problem from large/small to
no-migration/migration. Now evolutionists are poised to answer your
questions -- in true smorgasbord style. How can the small sub-populations
acquire a high rate of new beneficial mutation? Answer: Gene migration!
What can reduce the cost of substitution? Answer: Lack of gene
migration! And so forth. Ask a question, and the evolutionist will dodge
in the appropriate direction. The answer you get, depends on the question
you ask.
The cost of substitution remains unaffected by all that. A new mutation
must still go from 'few' to 'many' -- now with waiting stops midway -- but
that doesn't lower the cost. The cost remains the same, and the midway
waiting stops just slow things down. In effect, evolutionists say the
sporadic ***absence*** of gene migration makes new mutations spread
faster. That contradiction is difficult to sell, once you see it clearly.
I discuss this matter in my book.
******
>Haldane assumed the cost of selection had to be payed
>by reproductive excess; couldn't it also be payed by
>sexual excess? .... Reproductive excess of the more
>fit type would be manifest as more successful sperm in
>the females reproductive tract rather than more successful
>juvenile organisms in the environment. The excess would
>simply come into play at an earlier stage in the life
>cycle. And, we know that the number of excess sperm
>far exceeds the reproductive excess as measured in
>juveniles.
Chris proposes that the cost of substitution can be paid by "sexual
excess" rather than reproductive excess. He is mistaken. A trait cannot
go from 'few' to 'many' without reproductive excess -- it is absolutely
required, and nothing else will do. It doesn't matter how much sex there
is. It doesn't matter how many millions of excess sperm lay dying in the
females' reproductive tract. It doesn't matter how many progeny there
'might have been' if only all the sperm had succeeded. What matters is
the actual, realized reproductive excess (in particular that portion that
can be applied to substitution).
******
Evolutionists like storytelling. Ask a question, and they'll manufacture
a story special to answer your question. Never mind that the special
story brings with it a whole raft of its own special problems. Ask about
those problems, and the evolutionist will answer with yet another story,
about yet another special situation. Round and round you go, chasing
evolutionists around the bush, trying to get them to say how evolution
occurs.
Chris does that here:
>Also, Haldane's model assumed equal mating success (all
>selection was based on mortality). .... [However,]
>In most sexual species, a few males get most of the
>matings. .... A male with one mutant "super allele"
>could drive the frequency from 1/2N to 0.5 in a single
>generation if he mated with every female and ensured
>other males did not mate successfully. That's
>unrealistic of course, but I think you can see how
>variance in one sexs' mating success could take the
>place of reproductive excess.
Chris tells a story about how a substitution can happen rapidly. His
story requires an extremely small population where only one male sires all
the progeny. That would be a population in the few tens. Also, since
only one male breeds, the effective breeding size of the population is cut
roughly in half. Thus, the effective population size is very small
indeed. The population therefore receives beneficial mutations at an
exceedingly low rate, and almost half the time they are wasted because
they occur in non-breeding males. Such a small population cannot have a
fast substitution rate. (Remember this was Motoo Kimura's argument
against selective evolution.) The substitution rate 'starves' for lack of
beneficial mutations.
Chris assumes his way around that problem. He starts his story by
assuming the one breeding male already possesses a beneficial mutation.
He conveniently ignores the abundance of time when that doesn't happen.
[Note: Chris's story also wastes nearly half the reproduction of the
species, right off the top. The species produces half males, and all
(save one) die without heirs. Such males are born, they live and use up
resources, they die, but they do not participate in the evolutionary
process.]
******
>All this Haldane's dilemma stuff has been reasonably fun.
>In the course of thinking about it, I wondered how a few
>other considerations would effect the cost of selection.
When the talk.origins crowd "wonders" and "speculates" about the solution
to Haldane's Dilemma, then I have successfully accomplished my purpose
here. I have shown that Haldane's Dilemma truly is the trade secret of
evolutionary genetics. If a clear, documented solution existed, then my
opponents would have found it by now.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: colby@bio.bu.edu (Chris Colby)
Newsgroups: talk.origins
Subject: Haldane summary
Date: 19 May 1995 01:33:27 GMT
Message-ID: <3pgsh7$g0g@news.bu.edu>
OK, I'm bored. Unless Walter brings up anything new, or interesting, I'm
going to quit responding to the Haldane thread -- I think both sides have
laid out their arguments. Also, Wesley Elsberry has offered to help
write a computer program to help settle the issue. Actually, I think writing
a brief pop. gen. program for hard vs. soft selection would be cool
for reasons other than the Haldane argument. Alternately, maybe
we could take a stab at an analytical approach (although I should
check the literature more thoroughly to see if there is one. Initially,
I assumed there must be; but now I'm not so sure. Unfortunately, I'm stuck
at BU which has a lousy library... )
In any case, here is a brief condensation of my arguments.
Walter basically made two claims. 1.) That Haldane's dilemma was
the "trade secret" of population genetics and 2.) Haldane's dilemma
was not solved.
The "trade secret" assertion was demolished nearly instantly. Haldane's
dilemma is in nearly every intro text (for example, Futuyma's and Ridley's).
One exception was found (J.M. Smith's text), but the remainder of the
books ReMine cited to support his argument were specialized books, so
one wouldn't expect them to contain Haldane's dilemma. In addition,
some advanced books treat Haldane's dilemma at length. A full (more
full than Haldane's original) treatment can be found in Crow and Kimura's
introductory population genetics text. Both Lewontin and Wallace wrote
books that discussed Haldane's dilemma at length. ReMine has simply
ignored these facts and continues to assert that it is a trade secret.
Apparently, it's a special class of trade secret that permeates the
literature of a field.
Walter's assertion that Haldane's dilemma is unsolved is a bit more
interesting. One thing I should have been more clear about in the
first place is; what is a "solution" going to look like? Haldane's
cost of selection gives a maximum rate (in substitutions per generation)
of evolution. It's not really anything that can be "solved" in the
standard mathematical usage of the term. One can, however, assess if
it is a good estimate of the maximum theoretical rate of evolution.
Also, one can attempt to assess if the rate is too low to correspond
to observed macroevolutionary transitions.
I think several factors could raise the estimate of max. number of
substitutions per generation. The two commonly cited factors are
truncation selection (which Andy argued for) and soft selection
(which I argued for). I also mentioned population subdivision and
paying the cost of substitution in sperm rather than juvenile organisms.
There was not much progress in arguing these. Walter's counter arguments
were convincing proof that a little knowledge is a dangerous things; he
got so many things half-right that he was almost persuasive. But, this
part was somewhat fun, and it did stimulate me to think about somein pop. gen. that I haven't in awhile.
Finally, one needs to ask how Haldane's limit would effect evolution
even if it were true. ReMine has repeated asserted (without support)
that Haldane's cost of selection indicates that there could not be
enough substitutions for human evolution to occur. I claim that we
don't know enough to really say one way or the other; but, one can make a
good case for it not being a problem. Human evolution (from the chimp/
human ancestor on) is thought to be the result of a modification of
existing primate developmental processes. This coupled with the fact
that changes in regulatory sequences can effect many genes involved
in development suggest that few changes are needed. (The genetic
similarity of chimps and humans indicates that indeed few changes
are present.) ReMine has just plain ignored this line of argumentation.
Well, that's my brief summary; here's my ultra-brief recap:
Haldane's theoretical maximum rate of evolution can potentially be
elevated, but this hypotheticial maximum may not be a problem even
if it is correct.
I now return you to your regularly scheduled flamefest.
Chris Colby --- email: colby@biology.bu.edu or colby@acs.bu.edu ---
"'My boy,' he said, 'you are descended from a long line of determined,
resourceful, microscopic tadpoles--champions every one.'"
--Kurt Vonnegut from "Galapagos"
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 21 May 1995 17:25:36 -0500
Message-ID: <3poel0$9fp@dawn.mmm.com>
This is my latest response to Chris Colby. He *still* hasn't addressed my
earlier refutations of his position on soft selection. (See my previous
posts for details.) Briefly those refutations are: 1) Soft selection is
rare, according to G.C. Williams, it cannot solve Haldane's Dilemma. 2)
Andy Peter's and G.C. Williams' mutation load argument. 3) Varying the
selective values (as in soft selection) cannot significantly reduce the
cost of substitution, but will most likely raise it. 4) The *presence* of
the less fit organisms, all using up resources, (as in soft selection)
cannot speed the spread of the more fit organisms.
******
Now for his latest post. Chris again accuses me of "lying." (His words.)
And he issues me a "challenge".
Chris writes:
>I challenge you to produce even one quote where an evolutionist
>claims numerous beneficial mutations cause extinction. You are
>just plain lying now. If this statement is quite old and "universally
>propagated" (as you claim_ you should be able to back up what you are
>saying. Book title and page number, Walt.
I answer Chris Colby with evidence #1 ... namely, Chris Colby.
In that same post he wrote:
>One way Haldane's dilemma can be presented is as a theory of
>extinction; there's more than one way to skin a cat.
[NOTE: I previously emphasized that Haldane's Dilemma is *not* a theory
of extinction (rather it is a criterion of plausibility). In the above
quote from Chris, he emphatically rejected my point, and he (for the second
time) fervently *re-embraced* the (mistaken) notion that Haldane's dilemma
is a theory of extinction. Thus, he was told about the error, but he
re-embraced it nonetheless. So his example is an especially sweet answer
to the challenge he made.]
******
I wrote in my previous post:
>Chris's argument is a disguised way of claiming that sexual reproduction
>makes evolution go faster. But that notion is untrue.
Chris responds, first by saying I am wrong:
>I never made any claims about sexual reproduction and Haldane's
>dilemma. The example I used was of a sexually reproducing species'
>but that was not the crux of my argument.
Then Chris takes it all back, and (sarcastically) agrees with me.
>Gee Walter, you _can_ read. Haldane's cost of selection refers to
>how many total substitutions can occur genome wide, not at a single
>locus. So, yes, [sexual reproduction] was the focus of my
>argument (give yourself a lollipop, Walter) BTW, substitutions
>don't have to all occur strictly in parallel, but it's easier to
>explain that way, ...
Then Chris writes:
>Oops, I spoke to soon. I specifically said that [a soft selection]
>substitution at a single locus may be slower,
[NOTE: In other words, soft selection may well be ***slower***.]
>but [due to sexual reproduction] many substitutions can be going
>on (across the genome and/or over time.)
In other words, Chris completely confirmed my point. He did not show how
soft selection speeds an individual substitution, or the substitution rate
in general. His *entire* argument relied on sexual reproduction alone.
It was merely a disguised way of claiming that sex speeds evolution --
which is untrue. (See my book or previous posts for details.)
Chris totally failed to show how soft selection can even partially reduce
Haldane' Dilemma.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119
From: us236454@dawn.mmm.com (Walter Remine)
Newsgroups: talk.origins
Subject: Haldane's Dilemma
Date: 21 May 1995 18:19:39 -0500
Message-ID: <3pohqb$bbk@dawn.mmm.com>
*** Haldane's dilemma and dominant "alpha" males ***
Chris Colby previously gave an example of a single dominant "alpha" male
being the *only* male to breed. Chris felt this would speed up the
substitution rate. I previously dismantled Chris's example. James Acker
now reponds to me.
James writes:
>As in interested observer, this rings positively dead
>false in light of the real world, Walter. The pattern
>of dominant male mating with all the females in a
>pride/pack/tribe (for lack of a better word for groups
>of chimps and gorillas) /herd bull, etc. is WELL-KNOWN.
>Why would it persist if it's not a sound evolutionary
>strategy? Those males get there by demonstrating
>fitness -- they're stronger, faster, smarter, you name
>it, but they have to work to be Alpha and get the
>privilege of the vast majority of matings.
>....
>Walter, you need to get out more. I'm not an evolutionary
>biologist, but lay literature shows you're wrong. I read a
>recent article on the mating strategies of dunnocks (English
>sparrow species) that easily refutes your argument above.
>It was in Natural History, March or April. I'm pretty sure
>it was the issue with a snake on the cover.
>
>Put briefly, the basic pattern is this: an Alpha male usually
>gets the majority of the matings, but other males get a
>percentage of matings and offspring.
James missed the point. The *existence* of populations with a dominant
"alpha" male is not the issue. "Existence" is not the issue,
substitution rate is the issue. James' examples are irrelevant. Chris
Colby brought up the alpha male scenario, and I showed why it doesn't
increase the substitution rate.
>Furthermore, the females do occasionally subvert the Alpha
>with a little promiscuity, to mix up the gene pool and
>increase the fitness of their offspring. Thus, the
>scenario Chris notes is _real-world_. It appears to work.
James now *changed* Chris's example, by having females "occasionally
subvert the Alpha [male] with a little promiscuity, to mix up the gene
pool". That was not part of Chris's example. (In fact, it controverts
the very point Chris was trying to make.) This is the usual razzle-dazzle
that evolutionists give you. You point out problems with their model, so
they change the model. That happens over and over. Round and round the
bush you go, with evolutionists continually contradicting themselves.
They won't give you an agreed, coherent model of how evolution is supposed
to occur -- because there isn't one. Evolutionists themselves cannot
agree on an answer. So they just hope to wear you out.
>Haldane's "dilemma" DOESN'T REFLECT REALITY. It's a model
>that doesn't work.
>....
>In summary, the arguments have come to this: Haldane's
>Dilemma will not produce a realistic simulation of nature.
>The assumptions are flawed.
>
>
>You've shown this: Haldane's Dilemma was an interesting
>little problem that isn't very relevant to evolutionary
>biology because the assumptions that allow it to be
>formulated are too simplistic.
That's the classic move I've been telling you about: Try to blame the
problems on Haldane, and bury them with him. The problems don't go away
with Haldane, or with the demise of soft selection and the standard model
of genetic evolution. The problems must be solved.
Walter ReMine Standard disclaimer: These thoughts are mine.
P.O. Box 19600
Saint Paul, MN 55119