Uncommon Descent Serving The Intelligent Design Community

Part of the Discovery Institute’s secret research program uncovered

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

Through a little detective work, I found out where some of the Discovery Institute’s research funding has gone. It was an obscure comment in a paper that clued me in. The funding was for an exploration into the fundamental Speed Limits of Naturalistic Evolution. What was the plight of this exploration?

Here is what the principal researcher, Walter ReMine, had this to say regarding the ordeal his work endured (from a post at ARN):

this is about the evolutionist’s rejection of an entire field (Cost Theory), as well as their rejection of a central problem in evolutionary genetics — Haldane’s Dilemma. And it goes to additional issues, such as the evolutionist’s frequent claim that creationists/ID-ists “never publish in peer-reviewed science journals”. It also goes to the issue of negligence, because evolutionary leaders (apparently knowingly, according to their testimony) allowed the existing state of confusion to prevail for decades, and they rejected clarifications which they knew to be correct.

I wrote a paper clarifying the fundamentals of Haldane’s Dilemma, and submitted it to three mainstream science journals, where it was reviewed by evolutionary geneticists. At the first two journals, it was rejected as incorrect. At the third journal, it was acknowledged as “correct” by evolutionists Warren Ewens and James Crow. Nonetheless, they rejected it from publication on the astonishing grounds that my paper is unnecessary because they and their associates already “knew” my material “in the 1970s”.

Being that my paper is acknowledged as correct, yet rejected at three evolutionary journals (three strikes they’re out!), it is now published in the creationist journal, TJ, and available for free here.

The purpose of this present thread is to challenge evolutionists (or anyone else) to justify Ewens & Crow’s stated reason for rejecting my paper. They already acknowledge my paper as correct, so that’s not the issue here. In other words, I challenge evolutionists (or anyone else) to show: (1) that the clarifications given in my paper are already clearly and coherently given in the evolutionary literature, and (2) that no such clarification of the evolutionists’ cost-literature is needed or necessary.

I finally started looking through the papers again. There are two papers so far that I’ve found:

Cost theory and the cost of substitution—a
clarification

and

More Precise Calculations of the Cost of Substitution

Curiously, I found these statements in the papers:

This work was supported in part by a grant from Discovery Institute. Thanks also to Dr Paul Nelson.

I encourage the readers to take a peak at ReMine’s work. Ask yourself, “Does this work explore important scientific issues? Is this the sort of research that the scientific community should consider? Is his work accurate?”

Here is a footnote from one of the papers:

This paper was submitted previously to the journal Theoretical Population Biology, where renowned evolutionary geneticists Warren J. Ewens and James F. Crow reviewed it, along with Alexey Kondrashov and John Sanford. They all acknowledged this paper is essentially correct in all matters of substance. However, Ewens and Crow rejected it from publication on the grounds that it is not sufficiently new or different from what was known by themselves and some of their colleagues in the 1970s. However, they never communicated this knowledge to the greater scientific community, nor to the public at large. There were rare correct insights scattered sparsely in the literature, but those were incomplete, overwhelmed by confusion, and never communicated together in a coherent manner. This has all been very unfortunate, as there continues to be widespread misunderstanding within the scientific community regarding these important matters, even among those who have studied the cost literature for years. It is hoped that the clarifications presented in this paper, which are sound, will eventually reach the greater scientific community.

—Walter J. ReMine.

So the Discovery Institute has parceled out money to researchers in relevant fields. We also know about the Biologic Institute. I expect there is research going on elsewhere, and if not sponsored by the Discovery Institute, it will be pursued by individuals eager to discover Design in nature.

Comments
[...] I mentioned My retreat from public view, but I have a few loose ends to tie up. First, I had promised Walter ReMine a few weeks back that I would help advertise the first part of his response to Ian Musgrave. I point UD readers to Walter’s essay Haldane’s view of Haldane and also a discussion at ARN’s ID Observatory here. I indicated Walter’s work had been funded by the Discovery Institute and Walter’s work should not be ignored. [See: Part of the Discovery Institute’s secret research program uncovered.] [...]Tying up loose ends: ReMine, Behe, ERV, Illustra Media, Botnik | Uncommon Descent
September 3, 2007
September
09
Sep
3
03
2007
11:24 AM
11
11
24
AM
PDT
[...] I mentioned My retreat from public view, but I have a few loose ends to tie up. First, I had promised Walter ReMine a few weeks back that I would help advertise the first part of his response to Ian Musgrave. I point UD readers to Walter’s essay Haldane’s view of Haldane and also a discussion at ARN’s ID Observatory here. I indicated Walter’s work had been funded by the Discovery Institute and Walter’s work should not be ignored. [See: Part of the Discovery Institute’s secret research program uncovered.] [...]Tying up loose ends: ReMine, Behe, ERV, Illustra Media | Uncommon Descent
September 3, 2007
September
09
Sep
3
03
2007
11:02 AM
11
11
02
AM
PDT
Mike, Thank you for the fine work. We look forward to your other installments. The issues are important and our readers would be pleased to hear both sides presented on this important issue. Salscordova
January 25, 2007
January
01
Jan
25
25
2007
03:04 PM
3
03
04
PM
PDT
"Haldane's Dilemma" is really only relevant IF (big if) it can ever be demonstrated that any mutation/ selection process can account for the changes required if all of the diversity of living organisms owes its collective common ancestry to some population(s) of single-celled organisms. And then it would be applied to see how much time would be required to pull of such a thing. For in the end even if one mutation can become fixed in 1 generation or 1000, it does not matter if an accumulation of culled mutations just leads to the observed Wobbling Stability.Joseph
January 25, 2007
January
01
Jan
25
25
2007
02:40 PM
2
02
40
PM
PDT
First part of the example and response is now up at: http://scienceblogs.com/authority/2007/01/how_fast_can_evolution_work.phpMike Dunford
January 25, 2007
January
01
Jan
25
25
2007
01:43 PM
1
01
43
PM
PDT
lot of what I’m trying to say will come through more clearly if I take a simulation through a few steps. Unfortunately, I’m not a good enough programmer (read: not a programmer at all) to set up some process to do it by hand. I won’t have the time or inclination to do so today - I’ve already got several hours of math on the table, and am not voluntarily going to add more, but I’ll take a swing at it tomorrow. I’ll post it on my own blog - that will make it easier for me to do the formatting and save drafts - and post a link here when I’m done.
On behalf of our weblog, let me express our thanks for your willingness to participate. We may disagree, but I am appreciative of your time. Our readers need to be aware of the nuances in this difficult topic, and they should be aware of the arguments for and against what the ID-proponents have asserted. My sincere thanks.scordova
January 24, 2007
January
01
Jan
24
24
2007
12:03 PM
12
12
03
PM
PDT
Mike wrote: Hardly. The scenario that you sketched was neutral, because the trait was totally invisible in the population. That doesn’t mean that all such scenarios will be - think about dominant traits, for example. We can identify regions in the genome that are, or have recently been, under selection provided that we can find the right sorts of markers nearby because the gene frequencies in those areas deviate from the neutral expectation in specific ways.
Mike, Perhaps there is a misunderstanding here. What I was illustrating is that the only way soft selection can make a steady flow of simultaneous fixations per generation is when it behaves in a way that's almost indistinguishable from neutral theory. The only difference between pure neutral theory and a high speed, simulataneously fixating soft selection scenario is that traits finally fixed will be a small subset of the set of starting traits we deemed advantageous to begin with. Many of the advantageous traits will be lost just by random chance in this high speed scenario. The example of a population of 4 individuals (2 couples) illustrates something that can be generalized to other populations. It shows that the trade-off for simulataneous substitution is the sacrifice of a proportion of selectively advantaged traits in prior generations. Each time the soft-selection scenario is run a different outcome will result. If that is the case, one has to wonder whether natural selection is the appropriate metaphor to describe what's mostly a random walk.... Consider anti biotic resitance by bacteria. You can have several independent petri dishes, subject each to the same amount amoxicillin antibiotic. One will see the same traits appear in the survivors to antibiotic treatment in each petri dish. An analogous result happens for pesticide resistance and insect population dynamics. In contrast, trying to simulataneously fix traits via soft seleciton will result in different outcomes each time the scenario is played back. Chance will pull out some favored traits from the population by pure chance alone. One has to wonder then if invoking the word "selection" is even appropriate to decribe such random walks... The alternative is to accept a soft selection scenario which results in only sequential substitution of traits. But this is basically Haldane's dilemma...scordova
January 24, 2007
January
01
Jan
24
24
2007
11:59 AM
11
11
59
AM
PDT
DaveScot:
Haldane used the maximum amount of hard selection practical to arrive at the highest speed of fixation possible (300 generations for humans).
I'm sorry, but that just isn't correct. First, and on a more minor count, Haldane's calculations were not human-specific, and 300 generations was a general approximation. Second, that figure is not the highest speed of fixation possible. It is the highest speed of fixation practical in a hard selection scenario. Under soft selection, fixation can take place more rapidly than 300 generations. Sal:
But let’s look at it briefly, and one will see it is only a tautologous version of Kimura’s neutral theory, with the “beneficial” label retrodictively (after-the-fact) added to the subsitutions that make it to fixation!
Hardly. The scenario that you sketched was neutral, because the trait was totally invisible in the population. That doesn't mean that all such scenarios will be - think about dominant traits, for example. We can identify regions in the genome that are, or have recently been, under selection provided that we can find the right sorts of markers nearby because the gene frequencies in those areas deviate from the neutral expectation in specific ways. DaveScot:
What doesn’t make much sense is trying to say Haldane is wrong about the cost and speed of beneficial substitutions.
I'm not sure that anyone is making that argument. Haldane was right about the cost and speed of beneficial substitutions in the specific case of that sort of hard selection. The error only comes in extrapolating the results obtained using Haldane's specific set of assumptions to cover all cases, including those which do not meet those assumptions. A lot of what I'm trying to say will come through more clearly if I take a simulation through a few steps. Unfortunately, I'm not a good enough programmer (read: not a programmer at all) to set up some process to do it by hand. I won't have the time or inclination to do so today - I've already got several hours of math on the table, and am not voluntarily going to add more, but I'll take a swing at it tomorrow. I'll post it on my own blog - that will make it easier for me to do the formatting and save drafts - and post a link here when I'm done.Mike Dunford
January 24, 2007
January
01
Jan
24
24
2007
09:37 AM
9
09
37
AM
PDT
DaveScot wrote: I think natural selection is the principle mechanism but it doesn’t act on random mutations. It acts to conserve planned species not evolve unplanned new species
Another possibility after a planned mutational event is that species simply radiate into 2 or more seperate species. One gets fixation without even selection. And the speed limit problem is solved. For example the ancestor of dogs, wolves, foxes, and jackals simply have radiation events. Same for all other species. Some in the Baraminology group have been tracking attempts to reverse-radiate some animals and see how quickly the reverse-radiated animal re-radiates. The most well-known are llama-camel, dog-wolf, etc.. If there were pre-programmed radiations, there will be a certain resistance to reverse-radiation breeding. We need to engineer such a creature, and then it will be convincing as a mechanism, I suppose. I disucssed such an experimental exploration here: Marsupials and Placentals: a case of front-loaded, pre-programmed, designed evolution?scordova
January 24, 2007
January
01
Jan
24
24
2007
06:40 AM
6
06
40
AM
PDT
I think natural selection is the principle mechanism but it doesn't act on random mutations. It acts to conserve planned species not evolve unplanned new species.DaveScot
January 23, 2007
January
01
Jan
23
23
2007
08:38 PM
8
08
38
PM
PDT
It doesn’t come anywhere near disproving common ancestry.
I assume common ancestry in this case as a working assumption, (although you are aware of my personal views). I wish to simply argue that: 1. Natural Selection cannot be the principle mechanism 2. Neutral evolution cannot be the principle mechanism (although it has far less flaws than Natural Selction from a substitution standpoint) There are some other mechanisms which may work, one of them being front-loaded evolution and/or special creation. I don't delve much into the arguments over front-loaded evolution vs. special creation. I'm content to officially allow some degree of irresolution over the topic for now, even though it is quite evident what my personal views are. Salscordova
January 23, 2007
January
01
Jan
23
23
2007
08:31 PM
8
08
31
PM
PDT
P.S. I was just watching a Discovery Science program on evolution and saw that the same mutation in drosophila that gives them white eyes also causes extraordinary changes in behavior. The normal red eyed flies walk about the glass wall of their containers randomly but the white eyed mutants march around in lines one behind the other like soldiers. I think illustrates that small changes in single genes can have quite far reaching and disparate effects. I think this is the lesson that Haldane's Dilemma teaches - it doesn't take a lot of genetic change to have a lot of phenotypic effect. And conversely we also know (google "ultraconserved phenotype" a term I coined) that large genetic changes can have virtually no effect on phenotype.DaveScot
January 23, 2007
January
01
Jan
23
23
2007
08:27 PM
8
08
27
PM
PDT
Sal A big misunderstanding seems to be confusing cost of substitution with speed of substitution. Hard selection has the highest cost of substitution and the greatest speed of substitution. Soft selection has the lowest cost of substitution and the slowest speed of substitution. Haldane used the maximum amount of hard selection practical to arrive at the highest speed of fixation possible (300 generations for humans). I'm far from convinced this is a problem for human evolution from the last common ancestor with chimps. All it says is that the number of substitutions required to get from point A to point B is fewer than what you might think. The number of fixed differences between human and ape could simply be an artifact of a number of extremely close calls with extinction in both species. Fixation can occur very quickly through hard selection in very small populations. A lot of unimportant mutations could be fixed in that manner. Given that we're the only surviving hominid line and our numbers and geographic distribution were quite small at many points in time this seems to be a reasonable explanation. What doesn't make much sense is trying to say Haldane is wrong about the cost and speed of beneficial substitutions. That seems to be one of the rare cases in evolutionary theory where there is a mathematical proof behind it. But all it proves is there aren't more than a couple thousand beneficial mutations required to account for all the differences between man and chimp. It doesn't come anywhere near disproving common ancestry.DaveScot
January 23, 2007
January
01
Jan
23
23
2007
08:13 PM
8
08
13
PM
PDT
For the reader's benefit, here is an excellent link that dicusses soft-selection and other supposed solutions to Haldane's dilemma: Haldane's Dilemma (creation wiki) For the readers, I recognize the issues are complex and hard to understand, but it is at this level of meticulous detail the truth will prevail.
Soft selection remains ambiguous and confused in evolutionary literature. Most arguments surrounding it are based on genetic death or genetic load, which cause error. For example, the cost of substitution was often defined in terms of “genetic death” or elimination of the previous-type individuals. So, soft selection was claimed to reduce the problem by reducing or delaying the elimination. But that is wrong-headed, because the issue is not ‘elimination’ of the previous-type individuals, rather the issue is the ‘increase’ of the new-type individuals (and the reproduction rate necessary to make that happen) – and this version of soft selection does nothing whatever for that increase. (On the contrary, it keeps the previous-type individuals around longer, taking up resources, and slowing evolution down.) By focusing on the wrong issue, genetic death fostered this false solution to Haldane’s Dilemma. Soft selection is most commonly defined as frequency-dependent selection, or density-dependent selection. But that is not sufficient to reduce Haldane’s Dilemma, much less solve it. The total cost of substitution has an absolute minimum (that occurs when the selection coefficient approaches 0+), and rises as the selection coefficient rises. Haldane’s equations already favored evolution by granting the absolute minimum total cost of substitution – through assuming consistently tiny selection coefficients (approaching 0+). If the selection coefficient varies (as in frequency- or density-dependent selection), then the total cost of substitution cannot possibly go lower than Haldane’s equations, but will surely go higher – making the problem worse than before. This point is basic, yet still widely confused today, which is further evidence of evolutionist negligence. Soft selection is sometimes defined as “rank-order selection,” which is effectively just another version of truncation selection – and is not realistic, (see below). To give that an air of realism, soft selection is typically equivocated to mean “frequency-dependent or density-dependent selection.” Such equivocation (or bait-and-switch) should be rejected. The prevalence and sufficiency of soft selection are disputed (e.g., by evolutionary geneticist, G. C. Williams).
scordova
January 23, 2007
January
01
Jan
23
23
2007
07:09 PM
7
07
09
PM
PDT
I didn’t add the other pathways, but when blending all the probabilities, I think it’s not going to be greater than 1 per generation.
To complete the calculation one needs to blend the other possible scenarios. For example there is one scenario that in 2 generations ALL novel traits are lost. this happens 1/16 of the time. But I argue this is sort of a moot point because, one can take Kimura's neutral model an retrodictively add the label "beneficial" to the substitutions that reached fixation, and voila, "soft selection".scordova
January 23, 2007
January
01
Jan
23
23
2007
06:08 PM
6
06
08
PM
PDT
This changes the “speed limit” part of the calculation, and, even more importantly, totally eliminates the need for substitutions to take place sequentially. (The calculations in the reducto assume sequential substitution.)
(The calculations in the reducto assume sequential substitution.)
Here I must object as to the reason why the calculations assume sequential substitution. If Male A and Male B have 2 traits, even if they mated with the same female, they cannot "fix" those traits into the next population. Hence, by way of extension, soft selection is still limited to at most 1 substitution per generation in the reducto case. Whether this is true in the constant population case is worth visiting. But let's look at it briefly, and one will see it is only a tautologous version of Kimura's neutral theory, with the "beneficial" label retrodictively (after-the-fact) added to the subsitutions that make it to fixation! Consider a constant population of 4 diploid individuals (2 males and 2 females) where each has 1 novel "beneficial" trait and subsequent generations do not have novel beneficial traits for some time. The mean nucleotide fixation rate is less than 1 nucleotide per generation [check my numbers please] for the next 4 generations when the probablity that each couple passes both novel, favored traits to each of their 2 kids (probability is 1/4). The most optimistic scenario is that by generation 2, all kids have the 4 novel traits of their grandparents, but that happens only 1/16 of the time!!!! 4 traits divided by 2 generations times the likelihood of the scenario: [4 / 2 ] * 1/16 = 1/8 nucleotide on average I didn't add the other pathways, but when blending all the probabilities, I think it's not going to be greater than 1 per generation. The alternative scenario is to imagagine every generation having 1 "beneficial" mutation per individual per generation. But one will see, mathematically this is equivalent to neutral theory with only the label "beneficial" retrodictively added.scordova
January 23, 2007
January
01
Jan
23
23
2007
05:54 PM
5
05
54
PM
PDT
Sal: To be precise, Haldane assumed (a) that the population size could not decrease too much without causing extinction AND (b) that the selection against non-favored traits would reduce population sizes. The 300 generation figure AND the requirement that substitutions be sequential rather than simultaneous stem from both assumptions. Under soft selection, assumption b does not apply. This changes the "speed limit" part of the calculation, and, even more importantly, totally eliminates the need for substitutions to take place sequentially. (The calculations in the reducto assume sequential substitution.) In terms of rates in the wild, I really can't comment. The organisms that I work with are invertebrates, and I'm mostly looking at neutral markers.Mike Dunford
January 23, 2007
January
01
Jan
23
23
2007
12:20 PM
12
12
20
PM
PDT
Haldane's main argument was that selection must operate more slowly than the theoretical maximum because more rapid rates would reduce the population size to the point where extinction becomes virtually impossible. Therefore, he argued, the actual "speed limit" would be slower than the theoretical maximum. (In automotive terms, that's like having street tires on a formula one car - the tires are going to blow well before you hit the car's maximum.) This is true, but only for cases where the selection results in a decreased overall population size - hard selection. In soft selection, the selection does not decrease the population size. This means that although the theoretical time to fixation is longer, the actual speed can be much higher than is the case in hard selection. Think NASCAR with the right tires - not as fast as F1 could be, but much faster than F1 with highway tires. Moving back to ReMine's paper: The first issue with using ReMine's calculations to determine time to fixation is that his cost calculation does not calculate proportions or time to fixation. It calculates time to reach a certain number of individuals. Setting that aside, the second issue is that the wonders of geometric increase mean that even slow reproducers can reach very high numbers very quickly - for example, if you start with one individual with the trait, and if each individual with the trait produces 1.1 offspring who reproduce, you hit the 1,000,000 mark within the first 150 generations. If the rate is 1.5 (that's 3 offspring per breeding pair), it only takes about 35 generations to get there. If the rate is 2 (family of four), it's only going to take about 20 generations. Put (as closely as I can) in the terms that ReMine uses, that means that if the population can "afford" the "cost" of one extra offspring per parent per generation, a trait can reach fixation within 20 generations. Where the hard versus soft selection issue comes into play here is in addressing whether the organisms will need to produce more offspring than they already are in order to pay that cost. ReMine assumes (or seems to assume) that they will - that the cost of the substitution is in addition to all other reproductive costs. However, that is not necessarily (or always, or usually) the case. If the trait increases the chances that offspring will successfully reproduce (for example, by decreasing infant mortality in some way) then some or all of ReMine's substitution cost is covered by a reduction in some of the other reproductive costs. In such cases, the increased cost of substitution is compensated for by a decreased cost of other factors, and there is no need to produce more offspring than are already being produced. In practical terms, this means that a trait under strong selection can spread through the population relatively quickly without decreasing population size, and the trait can spread through the population quite quickly with even relatively small changes in the number of offspring that survive to reproduce. Finally, soft selection does something that ReMine does not address at all: it eliminates the need for substitutions to be sequential instead of simultaneous. The need for sequential change was entirely driven by the need to keep the population size from decreasing too far. As long as the population size remains relatively stable, as is the case with soft selection, multiple substitutions can spread at the same time. ReMine's paper does not address this at all.Mike Dunford
January 23, 2007
January
01
Jan
23
23
2007
12:01 PM
12
12
01
PM
PDT
Mike wrote: In cases like that, a trait can move to fixation relatively rapidly without the population size ever changing. In those cases, if we use the standard definitions of cost, there is no cost for this substitution. The population maintains a fixed size, while the gene frequency within the population changes.
Mike, I appreciate your visit an input. Haldane assumed constant population size. The categorization of soft versus hard adds only a confusion factor, and I don't think it cures the dilemma away. The figure of 1 "trait" per 300 generations was given by Haldane and corroborated by Kimura and Ohta. Until there are some hard nubmers on: 1. improvement in speed by soft selection 2. probability of soft selection A convincing case agaist Haldane's dilemma will still remain in the minds of many. ReMine gave a very simple Reductio ad absurdum Consider the extreme case of selection (somewhat like pesticide resistance) for a human-like population. A couple within a population of 100,000 individuals has the favored trait, and all 99,998 of the others die without leaving offspring, and only the couple survive and have inbreeding childred. Of the inbred children that survive and reproduce, how long will it take to restore the population size to 100,000? It will take 16.6 generations assuming the population doubles every generation. This is a very optimistic scenario. Clearly a fixation of 1 nucleotide per 16 generations is still too small. 180,000,000 indels plus about 30,000,000 point mutations will take 67 Billion years. How about the most optimistic scenario, 1 nucleotide trait fixed per generation, that would still be 210,000,000 years. The figure of "1 per 300" tries to factor in a mix of dominant and recessive traits. And what if the trait is not visible to selection (I pointed out several times at UD that there are could be substantial amounts of biological functionality and robustness almost invisible to selection), what hope is there of fixing those traits via selection? Finally, what direct empirical evidence do we have in the wild for AVERAGE nucleotide fixation rates of novel mutation rates for species alive today? It's a little disturbing that claims these problems are solved regarding the distant past, yet I would wager there is a scarce amount of data on novel nucleotide fixation rates in existing mammalian populations. I you are aware of any, I would be appreciative in learning of it. Welcome, by the way to our weblog. Salvadorscordova
January 23, 2007
January
01
Jan
23
23
2007
10:55 AM
10
10
55
AM
PDT
Actually it's Haldane who points out that hard selection is the quickest way to fixation (one generation) but that hard selection taken to extremes will result in a breeding population too small to be viable in slow reproducers. Trading off hard selection for soft selection will only slow down the process. In slow reproducers there's a cap on how much excess reproductive capacity is available. Remine restates the problem in terms of excess reproduction. There's a limit on how much excess a species can produce. A human female might produce and raise 10 or maybe even 20 viable offspring but she can't produce and raise 50. Therein lies the speed limit. Haldane alludes to this by saying the Dilemma isn't a problem for fast reproducers where one individual can produce thousands or millions of offspring. The problem is that soft selection operates without decreasing population sizes, so there is no cost (sensu Haldane). True, if that's how you're measuring cost. Remine isn't measuring cost that way. He's measuring it by differential reproduction which includes both soft and hard selection. Keep in mind the point of this thread isn't how many individuals die to fix a mutation but how many generations it takes to fix a mutation. The cost of substitution and the speed of substitution aren't the same thing. DaveScot
January 23, 2007
January
01
Jan
23
23
2007
10:07 AM
10
10
07
AM
PDT
First, and separately from everything else, I'd be curious to see exactly what in ReMine's paper makes you think that he demonstrated that soft selection increases the cost of substitution. As far as I can tell, ReMine claims that the type of selection is irrelevant to calculating the cost using his methods. Second: If that is the whole point of ReMine's paper, then I think I understand why it was rejected - it really doesn't add anything new, and it effectively fails to address anything more recent than Haldane. The rationale for looking at the cost in deaths rather than in reproduction, as Haldane did and everyone else working in the field still does, is because that's the more restrictive cost. Haldane was trying to ensure that the population numbers did not drop too far, because of concern over extinction risk. The need for excess reproduction breaks down to a simple geometric growth question - and it's easy to see (ReMine's work notwithstanding) that geometric growth reaches very high numbers very quickly. For example, if each organism produces three survivors in the next generation, you can hit 1,000,000 plus individuals in 14 generations, and 1,000,000,000 just four generations after that. The problem is that soft selection operates without decreasing population sizes, so there is no cost (sensu Haldane). In addition, most organisms are already producing more offspring than will successfully reproduce - if every mouse born survived to reproduce, we'd be ass deep in mice within a few years. If the selection operates to increase the proportion of offspring that survive to reproduce, then the cost (sensu ReMine) is offset by a reduction in the need for excess reproduction that is driven by other sources. This means that while there still might be a "cost," there does not necessarily need to be any increase in reproductive output to meet that cost.Mike Dunford
January 23, 2007
January
01
Jan
23
23
2007
08:53 AM
8
08
53
AM
PDT
Mike Dunford I'm afraid you might be misunderstanding differential reproduction. There's need to talk about natural selection, either hard or soft, if we instead call both hard and soft selection "differential reproduction". All substitution takes place through differential reproduction. The cost of substitution is excess reproduction of the mutant and/or deficient reproduction of the non-mutant. If excess and deficiency are in balance then population remains constant. If unbalanced then rise or decline occurs. The inescapable fact remains that differential reproduction is the only means of substitution and whether the seleciton is hard or soft makes no difference at all because differential reproduction encompasses both. The flaw in your description is: Soft selection operates whenever the organisms with the trait are less likely to be in the group that dies from “other” causes. There is no implicit or explicit reason for the organism with the beneficial trait to be less likely to be in the group that dies from other causes except for the fact that it has the trait. If the trait is the reason then it becomes an excess producer, and the trait moves closer to being fixed through differential reproduction. This was the whole point of Remine's clarification. Haldane was working the problem through what Remine referred to as the "traditional" teaching model where the cost is death. Remine reverses that and says excess reproduction is the cost. However, Remine still gives an example of fixation being you have 5 individuals, 1 with a beneficial mutation. 4 of them die without reproducing and the beneficial mutation is thus fixed in the surviving population. I think Remine should have simply said that substitution cannot take place without differential reproduction instead of cannot take place without excess reproduction. It can take place with any combination of excess mortality in organisms without the trait or excess reproduction of those with the trait. There IS NO OTHER WAY. You can obfuscate the terms and change the description any way you please but the bottom line remains there is no way other than differential reproduction (absent horizontal gene transfer, endosymbiosis, or the like) for a beneficial mutation to move towards fixation. The speed limit is thus set by how much excess reproduction is possible in addition to how much death can be tolerated. Remine shows that the math works out the same. He presents a general case whereas Haldane's was a specific case that falls under Remine's general case. Haldane used hard selection, a fixed population, and a maximum tolerable mortality rate, because that is the most efficient means of fixation and it simplifies the problem. Soft selection only increases the cost of substitution. Remine demonstrates this in his paper. DaveScot
January 23, 2007
January
01
Jan
23
23
2007
02:57 AM
2
02
57
AM
PDT
I'm afraid that you might be misunderstanding soft selection a bit. Let me try a clearer outline: 1) More organisms are born than reproduce. 2) Some organisms become successful reproducers. In hard selection, the selection acts to reduce the number of successful reproducers - it acts after step 2, above. In soft selection, the selection acts to influence which organisms reach step 2, but does not act to decrease the number of successful reproducers. Soft selection does not operate because organisms without the trait are dying from "other" causes. Soft selection operates whenever the organisms with the trait are less likely to be in the group that dies from "other" causes. That's why the two do not cancel - the organisms without the favored trait wind up with an increased chance of dying (relative to the group with the favored trait). Let me try another example. We'll start with a pool of 100 individuals. Ten will be able to "find territories" and reproduce, and ten have a mutation that makes them twice as likely to find a territory. (10% chance for the "unfavored," 20% for the favored.) I'll use a random number table to determine the order in which the individuals get to "attempt" to find a territory, and we'll assume that the trait breeds true. In cases like that, a trait can move to fixation relatively rapidly without the population size ever changing. In those cases, if we use the standard definitions of cost, there is no cost for this substitution. The population maintains a fixed size, while the gene frequency within the population changes.Mike Dunford
January 23, 2007
January
01
Jan
23
23
2007
02:12 AM
2
02
12
AM
PDT
Soft selection is self cancelling. If a creature without a beneficial mutation in the process of being fixed dies young from "other" causes and thus "helps" pay the cost of fixation then one must also consider the cases where creatures with the beneficial mutation die young from other causes. The "other causes" being equal, one cancels the other. If the differential reproduction is contingent upon the beneficial mutation then the selection isn't soft. You can't have your cake and eat it too. Nice try but soft selection as a partial solution to Haldane's Dilemma is without merit.DaveScot
January 22, 2007
January
01
Jan
22
22
2007
11:48 PM
11
11
48
PM
PDT
I just read the "clarification" paper, twice, and I can understand why it was rejected. As far as I can tell, the paper does restate the situation as of about 40 years ago, without covering any of the work that's been done since then. In particular, I think ReMine really does fail to address "soft" selection, and I don't think he comes close to adequately justifying that. I think that's most apparent near the bottom of the right hand column of page 114, where he lists the various "costs." Here, he lists the substitutional cost as an additional cost, over and beyond all of the other costs that already apply. That is the way that Haldane framed the problem, and it assumes hard selection. In soft selection, the substitutional cost is absorbed in whole or in part by the already existing "reproductive excess." Remember, most species really do already produce more (in some cases many more) young then will survive to maturity. If the selection operates to determine which of the young survive to reproduce, without reducing the number that survive to reproduce, then there is no "cost" in the way that ReMine uses the term. To see how the point in the life cycle where selection operates matters, here's a quick and dirty example. I'm using artificial selection and a lab just to make things clear and easy to understand: Scenario A: Start with a pool of 1000 insects. Select randomly 100 to reproduce. Before reproduction, go through the randomly selected reproducers and keep all those who do have the "favored" trait, and eliminate some fraction of those who do not. Scenario B: Start with the same pool. Instead of randomly selecting individuals to reproduce, inspect the flies and select those who do have the "favored" trait at a higher rate than those who do not. In Scenario A, which corresponds both to Haldane's original formulation and ReMine's restatement, the "cost" is in addition to all other reproductive costs. In Scenario B, which correpsonds to "soft selection," some of the "cost" (as ReMine uses the term) is at least partially offset by changes in other sources of pre-reproduction mortality.Mike Dunford
January 22, 2007
January
01
Jan
22
22
2007
09:42 PM
9
09
42
PM
PDT
Unfortunately the "solution" is to either ignore the dilemma or keep it under wraps.Joseph
January 21, 2007
January
01
Jan
21
21
2007
06:05 AM
6
06
05
AM
PDT
ajl, "I once had a paper rejected..." I wonder if there is a difference between "I once had a paper rejected ..." and "this paper was rejected three times." I wonder if the experience of Richard Sternberg would cause ID based researchers to question whether there was an ulterior reason for the rejection of his work. What I notice about this topic is the details made public wicipedia. In the wicipedia article the fact that Haldane's dilemma "is raised mostly by those opposed to evolution", the article does not provide even the vaguest solution to the dilemma. Either Haldane's dilemma is a big deal, or there is a good solution to the dilemma. If it is a big deal, works like ReMine's deserve publication. If there is a good solution, bring it on.bFast
January 20, 2007
January
01
Jan
20
20
2007
12:45 PM
12
12
45
PM
PDT
“First they ignore you, then they laugh at you, then they fight you, then you win” - Ghandi "The acceptance of radical ideas that challenge the status quo (and Darwinism is as status quo as it gets) typically runs through several stages. According to Arthur Schopenhauer, “All truth passes through three stages. First, it is ridiculed. Second, it is violently opposed. Third, it is accepted as being self-evident.” Similarly, evolutionist J. B. S. Haldane remarked, “Theories pass through four stages of acceptance: i) this is worthless nonsense; ii) this is an interesting, but perverse, point of view; iii) this is true, but quite unimportant; iv) I always said so.”" - Dembski ReMine's paper is excellent. No doubt. And it helped me a lot in understanding Cost-theory. I haven't seen nowhere a clarification like this. First they ignored ReMine, then they said his paper were incorrect, then they fought by rejecting it from publication,then ... you will see. But it is funky, that it was Haldane ( --> topic of ReMine's paper) who "predicted" the "I always said so" from Warren Ewens and James Crow.ChristopherSaint
January 20, 2007
January
01
Jan
20
20
2007
07:07 AM
7
07
07
AM
PDT
I once had a paper rejected where the referee indicated: it was an interesting read, but there wasn'tit didn't add anything really new to the field. I disagreed, thinking we had identified a number of new twists on an old idea. So, I submitted it to another journal and it got published. But, I didn't go public with my whining about it - and name the name of the editor of the journal! I think it is certainly acceptable for a peer review board to say that a paper does not add anything new, that isn't already in the literature. Now, I didn't read the article, so I don't know what is in it. But, if it was just a rehash of a paper in the 1970's, then I would agree with the peer review. Perhaps ReMine could have resubmitted it as a "Review Article", or "Short Note", rather than a research article. That said, I agree with him that most evolutionist probably aren't aware of Haldenes Dilemma, so it is a shame that an article like this doesn't get entered into the literature. But, perhaps it would have been better as a review article.ajl
January 20, 2007
January
01
Jan
20
20
2007
06:02 AM
6
06
02
AM
PDT

Leave a Reply