|
Author
|
Topic: Convergent evolution or Design signature?
|
Cre8ionist
Member
Member # 140
|
posted 25. December 2002 22:39
I'd like to address a couple of points on convergent evolution (CE) in this post.
1. Many cases of CE rise to the level of specified complexity which trigger a design inference.
2. There is a bit of a double standard. There are analogous cases in which similarities of much less complexity are attributed to ID without a problem. But when the far more complex cases are attributed to ID, it's labeled "unscientific."
CE is cited in cases ranging from similar enzymes to complete animals and just about everywhere in between. From the luminous creatures of the deep ocean to the luminous fireflys, from the marsupial wolf (a) to the placental wolf (b) (skulls pictured below, borrowed from A. Custance),
from the humming moth to the humming bird, striking similarities abound.
DE's would have us believe that environment rather than design is the key to the similarities. For example, the luminous creatures of the deep live in a dark place, so the environmental pressures drove the evolution of light on several different fronts. And so the story goes.....
I believe that using Dembski's criteria and mathematical calculations recently published by Spetner we can see that CE is not up to the task of creating many of these similarities.
Spetner conservatively calculated in "Not by Chance" that in a convergent transition of only 100 steps, the odds against it would be 10^600 to one. Far surpassing what's needed to validate a design signature. Further, he points out that thousands of steps would be needed for the evolution of compex organs. (pp 109 - 114)
But aside from the mathematical challenge, consider the issue of double standard.
Suppose we found replicas of Stonehenge on every continent, would we not conclude that these simple structures likely sprang from one design source?
Or (please indulge my fanciful imagination), take it further still, would you fault me if upon finding replicas of the White House, with its many rooms on every continent, I again held that they were causally connected in some fashion, unable to accept an explanation of their similarity which relied in large part on chance?
Finally, one last time, if I happened to find replicas of New York City, with its many buildings and streets, on every continent, would I not be justified in connecting them all to one designer?
If you agree that I would be justified then realize that there are not just examples of enzymes, not just examples of organs, but examples of whole creatures from different continents whose design is so similar that we call them by the same name even though they supposedly have completely different lines of descent. Calculate the probability of getting that through NS/RM as Spetner has done and you'll see why I agree with him, if variation arises largely from random errors convergent evolution is impossible!.............................................Cre8
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 26. December 2002 01:04
Creationist
While your contribution seems to lack any positive applications to ID I would like to address some of your claims
quote:
Calculate the probability of getting that through NS/RM as Spetner has done and you'll see why I agree with him, if variation arises largely from random errors convergent evolution is impossible!
And yet simulations show otherwise. I guess the problem may be with Spetner's arguments more than with the reality of scientific inquiry into evolutionary biology.
Some relevant links
The Evolution of Improved Fitness. Correspondence with Critic Lee Spetner
or
Lee Spetner/Edward Max Dialogue Dr. Lee Spetner continuing an exchange with Dr. Edward E. Max
or
"Neo-Darwinian Evolution Really Can't Work!"
Response
Review
So on one hand we have Spetner's claims and on the other hand the vaste amount of research and simulations that show otherwise.
But if Creationist believes that Dembski's method could be used to show evidence of design in these cases then by all means I would encourage him/her to make the argument with the supporting calculations. In fact I would argue that such an application of the design inference to a biological example beyond arguments of random chance would be a first.
I would also like to point out that Spetner seems to disagree with Dembski's approach to information and evolution
quote:
In Biology, information is nothing if it is not meaning.
Source
Probabilities tend so often to be used inappropriately when discussing issues related to evolution or origins of life. As a Christian scientist I often cringe at these 'examples'
I fail to understand your analogy with replicas of the White house. Are you saying that the White house procreates and mutates?
What am I missing here?
In Christ
F.
[ 26. December 2002, 01:24: Message edited by: Frances ]
IP: Logged
|
|
Cre8ionist
Member
Member # 140
|
posted 26. December 2002 08:21
Hi Frances,
I'm comfortable with Spetner's position on convergent evolution, and reading the debate (again) with Max won't change that at all. In fact, I too encourage everyone to read the entire debate. It can be found at trueorigins.org.
The analogy should be clear enough. Yes life mutates, but as Dawkins has admitted, evolution, and in this case CE is arrived at largely through chance mutations. We wouldn't attribute a Stonehenge replica to chance, would we?
As to whether this is a contribution to brainstorms or not, I've seen worse, and I've seen better. What I've proposed here is that CE is another testing ground for Dembski's filter, therefore, it also provides DEs with an opportunity to demonstrate that it doesn't work.
Oh, what calculations do you have for CE based on chance mutations?............................Cre8
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 26. December 2002 13:57
Dear Creationist
You seem to still be confused about the mechanisms of evolution when you compare them to 'Stonehenge'. Unless the rocks at stonehenge mutate and procreate there is little reason for comparison. Evolution is based on variation and selection, something often missed when trying to show that 'evolution is not possible'.
I understand that you may be comfortable with Spetner's position but that does not mean that it is scientifically defensible. Looking at Korthof's email exchange with Spetner reveals a lot of problems in Spetner's arguments.
The most relevant issue is that despite Spetner's claims simulations by people like Tom Schneider have shown otherwise.
I am looking forward however for you to make your case. Spetner seems to believe that information requires meaning which seems to be at odds with Dembski's approach to information through the use of Shannon entropy.
Do you have any particular examples in mind with worked out numbers? Or were you proposing to see if applying Spetner's concepts and Dembski's concepts would be productive?
It would be interesting to hear from other ID proponents about Spetner's ideas.
Glenn Morton also explains his problems with Spetner's applications of information concepts
In fact the whole thread makes for interesting reading.
I find your proposal interesting and would love to see a worked out example integrating Spetner and Dembski.
IP: Logged
|
|
Cre8ionist
Member
Member # 140
|
posted 26. December 2002 18:16
There are a couple of different Spetner ideas which are being discussed here, the one I brought up and the one you keep referencing. Spetner's book does deal with the information question, but I'm talking about convergence and probability, which he also discusses. Computer models notwithstanding, nobody to my knowledge has overcome the mathematics on CE laid out by Spetner. If you can build two different complex machines via NS/RM that's one thing, but to hit on the same one over and over should raise one's eyebrows, IMHO.
Anyway, in all the cases I cited, it was the chance part that they shared. You simply cannot divorce NS/RM from chance or we'd call it NS/M. It is only the RM that I'm referring to here. Spetner himself believes in evolution, it's the kind of evolution that's in question. This is why I ended with this thought on CE:
[quote]Calculate the probability of getting that through NS/RM as Spetner has done and you'll see why I agree with him, if variation arises largely from random errors convergent evolution is
impossible!.............................................Cre8
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 26. December 2002 19:09
Creationist
quote:
Calculate the probability of getting that through NS/RM as Spetner has done and you'll see why I agree with him, if variation arises largely from random errors convergent evolution is impossible!
Which explains why the assumption of random errors makes for a perfect strawman and a poor rebuttal imho. As is appeal to incredulity imho
[ 26. December 2002, 19:17: Message edited by: Frances ]
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 26. December 2002 21:14
Hey guys,
This discussion is (1) unfocused, and (2) you guys seem to be talking past each other. It's time to get down to brass tacks--give some specific, focused ideas about what convergence tells us about the origin of these complex systems. Creationist, it would be good to see some ideas on how to apply probabilitistic reasoning to these systems. Can you do a calculation? That would help the discussion immensely; even if you can't do the calculation, show us how we could, in principle do one. If we can't apply some sort of mathematical analysis to your argument that convergence is a problem for evolution, this discussion can't really make progress. Frances, same goes for you--give us an idea of how to do a calculation that supports your scenario (after all, you keep insisting that Creationist is wrong without providing your own calculations to back up the claim). Also, try to give focused, on-topic, insightful replies, and avoid making short, quick, space-wasting posts like the one above.
I'll be watching this thread carefully.
[ 26. December 2002, 21:17: Message edited by: Moderator ]
IP: Logged
|
|
Cre8ionist
Member
Member # 140
|
posted 28. December 2002 11:08
Mod: quote:
Creationist, it would be good to see some ideas on how to apply probabilitistic reasoning to these systems. Can you do a calculation? That would help the discussion immensely; even if you can't do the calculation, show us how we could, in principle do one. If we can't apply some sort of mathematical analysis to your argument that convergence is a problem for evolution, this discussion can't really make progress.
NP, BTW, they're Spetner's actual calculations available in the aforementioned book. I'm just pointing out that Dembski's filter can potentially be applied here too. And if the filter's worth its salt, we can confirm what many already suspect, namely, there's more to the similarity we find between the squid's vision and ours than first meets the eye. ![[Wink]](wink.gif)
In order to mathematically evaluate the odds that the same RM's would occur in two different lines of descent Spetner applied reasonable values to a couple of parameters:
1. Number of potential adaptive mutation sites at each step (1,000,000)
This is figured at the nucleotide level.
quote:
Only if there are at least a million potential adaptive mutations will there be a chance of at least one in a million that a new species will evolve. That was the criterion we set up as the test of the theory. So we see that evolution will work only if there are at least a million potential adaptive mutations at each step. pg. 104
2. Number of steps in a convergent transition (100)
He utilizes the value 500 (for species transition) given by Stebbins, and then for arguments sake lowers it to 100.
quote:
The bat wing is not built exactly like the bird wing. So let's say that, in a convergent transition to a new species, not all the 500 choices had to be the same. If only 100 of them had to be the same, the odds against it would be 10^600 to one. pg. 110
Below is a reproduced picture from the book with Spetner's commentary. To me it's obvious that Spetner's been more than generous here, does anyone really think that the wolves' skulls (pictured above), actually came to look alike after 100 mutations? How many would you suggest? Still skeptical? Let's knock it down to fifty. Still works out to 10^300 to one against it. Anyway, play with the numbers all you like to try to make it work, I'd like to see it. Spetner uses evolutionary literature to secure his parameters, what will you use?
quote:
You can compare the path of evolution of the population with a path through a huge tree-like maze. The maze is a complex of pathways that diverge, and whose every path goes through 500 nodes. Each node is a fork, and every fork has a million branches from which to make a random choice. Fig 4.1 shows a simplified version of this tree. Instead of a million branches at each node, the figure shows only three. The figure also shows only three levels of branching. But imagine the tree going on for 500 levels. As the population moves through the maze it comes to a fork, or branch point, at each of its 500 steps where it has a million branches to choose from. pg. 108
...............................................................................Cre8
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 28. December 2002 12:42
Just a note on Dembski’s explanatory filter:
The introductory paragraph by Cre8ionist makes the EF and its mathematical calculation a primary issue of this thread.
In a later post, Cre8ionist said:
quote:
What I've proposed here is that CE is another testing ground for Dembski's filter, therefore, it also provides DEs with an opportunity to demonstrate that it doesn't work.
A very important point is that an argument that there is an evolutionary pathway which produces the given evolutionary result does not need to be substantiated as having necessarily occurred, in order for it to be a denial that there is a problem according to the EF. The reason is that the EF requires that there be a “sweeping clean” of all possible pathways that could produce that result in principle, so as to validate that the calculations shown of a chance explanation at a given significance level is required. (And this is by the standards that Dembski proposes for the EF, not some additional standard or argument made by scientists.)
The problem of course is that if the possibility of such pathways existing has not been “swept clean”, then the probability calculations that show that a particular class of chance explanations is less than the UPB of 10^-150 is far more likely to fail because of the failure to note an explanation that is not in the required class. So if there is a likelihood of better than 10^-150 of a pathway being a proper explanation of the evolutionary process, then the possibilities have indeed not been “swept clean”, and the EF does not apply.
The moderator has suggested that calculations be given so as to substantiate the argument. But several posters have given possible pathways for various CE structures that have been suggested. I suggest that what is needed is not simply a “calculation”, but also an argument that denies that the possible pathways are indeed “possible”, or else the EF requirement of “sweeping clean” the possibilities has not been met. I don’t think that just a presentation of a calculation is sufficient to support the argument.
Indeed, the presentation of possible pathways for the suggested “CE” evolutionary events has already invalidated the argument that the EF applies, and the argument is already complete in my opinion.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 28. December 2002 14:02
Creationist raises some very interesting claims, some of which I will try to address.
First of all it is obvious that Spetner's approach is purely probabilistic, based on variations but does not deal with the main aspect of natural selection. Any approach which ignores this part of evolutionary theory can be easily rejected as 'attacking a strawman argument'.
Creationist
quote:
To me it's obvious that Spetner's been more than generous here, does anyone really think that the wolves' skulls (pictured above), actually came to look alike after 100 mutations? How many would you suggest?
Without further investigations, it will be hard to determine how many 'mutations' are needed in fact it may very well be that a few mutations in delays or extentions to developmental pathways may be enough to cause these morphological changes.
A great example of evolutionary change can be observed in the transition from reptile to mammal
as noted Here and hereand here
[ 28. December 2002, 14:47: Message edited by: Frances ]
IP: Logged
|
|
Art
Member
Member # 179
|
posted 28. December 2002 14:14
Hi cre8ionist,
Thanks for spelling out some of Spetner’s ideas here. I think that the numbers may serve as a good starting point to improve on Spetner’s efforts.
quote:
In order to mathematically evaluate the odds that the same RM's would occur in two different lines of descent Spetner applied reasonable values to a couple of parameters:
1. Number of potential adaptive mutation sites at each step (1,000,000)
This is figured at the nucleotide level.
Correct me if I’m wrong, but does this number come from an estimate of the number of evolutionarily-important genes, and the approximate size of each gene?
From Spetner:
quote:
Only if there are at least a million potential adaptive mutations will there be a chance of at least one in a million that a new species will evolve. That was the criterion we set up as the test of the theory. So we see that evolution will work only if there are at least a million potential adaptive mutations at each step. pg. 104
This is where Spetner seems to lose me. In a reasonably-sized population, each and every one of these potential adaptive mutations will occur each and every generation. This is a simple matter of mutation frequency and population size.
This consideration aside, the number Spetner uses may, in fact, be much smaller. Consider the statement by Doebley and Lukens (Plant Cell 10, 1075-1082, 1998):
quote:
In this review, we have presented some inferences that we and others have drawn from the modern understanding of gene structure, function, and interactions. These include (1) that development is modular in organization, (2) that protein function tends to be conserved over long evolutionary periods, (3) that plant promoters are labile in nature, (4) that signaling genes tend to have strongly pleiotropic effects, and (5) that transcriptional regulators act as key switches in plant development. The conclusion we draw from these inferences is that the evolution of plant form will be most readily accomplished by changes in the cis-regulatory regions of transcriptional regulators. The challenge now is to test this model by identifying the molecular mechanisms underlying specific events in the evolution of plant form among natural species. Genetics and transgenic plants will provide the necessary tools.
The implications are that the average “target size” (e.g., gene size) that Spetner seems to be using may be more than a factor of ten too large.
Given these different points, I would argue that Spetner has “cooked the books” with estimates that are not very accurate representations of real-life situations. Indeed, direct empirical study suggests that this “one-in-a-million” number is closer to unity. From Lauter and Doebley (Genetics 160, 333–342, 2002):
quote:
ABSTRACT
How new discrete states of morphological traits evolve is poorly understood. One possibility is that single-gene changes underlie the evolution of new discrete character states and that evolution is dependent on the occurrence of new single-gene mutations. Another possibility is that multiple-gene changes are required to elevate an individual or population above a threshold required to produce the new character state. A prediction of the latter model is that genetic variation for the traits should exist in natural populations in the absence of phenotypic variation. To test this idea, we studied traits that are phenotypically invariant within teosinte and for which teosinte is discretely different from its near relative, maize. By employing a QTL mapping strategy to analyze the progeny of a testcross between an F1 of two teosintes and a maize inbred line, we identified cryptic genetic variation in teosinte for traits that are invariant in teosinte. We argue that such cryptic genetic variation can contribute to the evolution of novelty when reconfigured to exceed the threshold necessary for phenotypic expression or by acting to modify or stabilize the effects of major mutations.
Moving on:
quote:
2. Number of steps in a convergent transition (100)
He utilizes the value 500 (for species transition) given by Stebbins, and then for arguments sake lowers it to 100.
I don’t know about Stebbins, but other empirical work suggests that this number can be as low as 10-20. The paper by Doebley and Lukens is instructive in this regard.
From Spetner:
quote:
The bat wing is not built exactly like the bird wing. So let's say that, in a convergent transition to a new species, not all the 500 choices had to be the same. If only 100 of them had to be the same, the odds against it would be 10^600 to one. pg. 110
The 10^600 estimate is very different from one derived with more reasonable estimates of “target size” and “target number”.
quote:
Below is a reproduced picture from the book with Spetner's commentary. To me it's obvious that Spetner's been more than generous here, does anyone really think that the wolves' skulls (pictured above), actually came to look alike after 100 mutations?
This suggestion seems reasonable enough, given the direct experimental studies that bear on the matter. (Others in this thread have added support for my suggestion.)
quote:
How many would you suggest? Still skeptical? Let's knock it down to fifty. Still works out to 10^300 to one against it. Anyway, play with the numbers all you like to try to make it work, I'd like to see it. Spetner uses evolutionary literature to secure his parameters, what will you use?
I would argue that Spetner has ignored almost all of the relevant literature. From the above considerations, and many, many others, I can see no probablistic issues here.
One more abstract, and then a closing remark:
quote:
From White and Doebley (Trends in Genetics 14, 327-332, 1998):
To begin to investigate the molecular basis of the morphological evolution of maize, Doebley et al. used quantitative trait locus (QTL) mapping to determine the number, location and relative level of effects of genes controlling the morphological differences between maize and its wild ancestor, teosinte. Two different maize-teosinte F2 populations were analysed in these mapping experiments, both involving crosses of teosinte and primitive varieties of maize (26,2). The results from these two populations were generally congruent, revealing that traits distinguishing maize and teosinte are controlled primarily by five chromosomal segments. Each of the segments identified has an effect on several traits, suggesting that they could carry either a single gene with pleiotropic effects, or several linked genes.
In closing, I would add that the reflexive response is likely to be that studies of plant evolution are not relevant in this thread. But I would argue otherwise. The specific cases I have pointed to involve morphological changes that exceed most changes that are seen in the evolution of vertebrates. It follows that, if such vast, vast changes are within probabilistic certainty, the much smaller changes being discussed here are also going to be much too likely (probabilistically) to pass through Dembski’s EF.
As importantly, while I have only cited studies of plants, the ideas have been shown to be relevant to animals as well.
Thus, given the broad sweep of Spetners’ remarks, I would argue that my points are highly relevant, and issues that Spetner must pay attention to.
(An added suggestion - participants and lurkers alike can do a bit of synthesizing with this post and those in the thread that deal with vertebrate examples, with the goal of identifying stages that may be relevant to the matter of regulatory evolution. This takes one down the path of testable hypotheses - asking which genes are important in these processes, and asking if regulation, as opposed to structural changes, are involved in differentiating different species.)
IP: Logged
|
|
|
Printer-friendly view of this topic