Home » Intelligent Design » Co-option, Berra’s Blunder, and Speculation Presented as Fact

Co-option, Berra’s Blunder, and Speculation Presented as Fact

In Bill Dembski’s thread, No Major Conceptual Leaps, I posted a comment about the evidential, logical, and probabilistic vacuity of the Darwinian co-option hypothesis. (I use the word hypothesis with reservation. A hypothesis in a domain such as this should at least be based on a minimal, mathematical probabilistic analysis.)

In response to my comment, another commenter offered this as a refutation.

This text from Deborah A. McLennan, of Evo Edu Outreach, is utterly embarrassing for her cause, because it makes the case for design, just as Tim Berra did with his infamous blunder.

Deborah writes:

The co-option of traits to serve new functions is not a difficult concept to understand. In fact, we ourselves do it all the time, which is why we speak about “new wine in old bottles” or “rebranding” for the repackaging of ideas, and more recently in keeping with the new management-speak, “repurposing”. We are forever finding new functions for old devices, using an old boot as a planter, a fishing rod to fly a kite, a magnifying glass to start a fire, a shell as currency, a berry or a root to dye cloth. The only difference between human and evolutionary co-option is that we purposefully change an object’s function, while evolution simply takes advantage of an opportunity with no direction, purpose, or forethought.

The only verifiable examples of co-option she presents involve agency. How about just one verifiable example of co-option that does not rely on agency? In addition, note the final sentence, which is pure speculation based on an assumed premise, but presented as fact. This is not how science works.

Berra’s Blunder:

If you compare a 1953 and a 1954 Corvette, side by side, then a 1954 and a 1955 model, and so on, the descent with modification is overwhelmingly obvious. This is what paleontologists do with fossils, and the evidence is so solid and comprehensive that it cannot be denied by reasonable people…

The point is that the Corvette evolved through a selection process acting on variations that resulted in a series of transitional forms and an endpoint rather distinct from the starting point. A similar process shapes the evolution of organisms.

Tim Berra, Evolution and the Myth of Creationism, 1990, pg 117-119

This is Phillip Johnson’s observation:

Of course, every one of those Corvettes was designed by engineers. The Corvette sequence — like the sequence of Beethoven’s symphonies to the opinions of the United States Supreme Court — does not illustrate naturalistic evolution at all. It illustrates how intelligent designers will typically achieve their purposes by adding variations to a basic design plan. Above all, such sequences have no tendency whatever to support the claim that there is no need for a creator, since blind natural forces can do the creating. On the contrary, they show that what biologists present as proof of “evolution” or “common ancestry” is just as likely to be evidence of common design.

Phillip Johnson, Defeating Darwinism by Opening Minds, 1997, pg 63.

It is illuminating that Darwinists, in many attempts to defend their blind-purposeless-undirected-chance-and-necessity hypothesis, draw on analogies from design and agency while trying to explain away agency and design.

It seems to me that this exercise should result in at least a modicum of cognitive dissonance.

  • Delicious
  • Facebook
  • Reddit
  • StumbleUpon
  • Twitter
  • RSS Feed

107 Responses to Co-option, Berra’s Blunder, and Speculation Presented as Fact

  1. 1

    Gil, do you know what precisely Berra was trying to illustrate in that passage? Does anybody have the context? Is he saying that Corvettes evolved naturalistically, or is the context about evidence for common descent?

    Unlike you, I’m not on a first-name basis with McLennan, so I’ll call her McLennan out of respect.

    Anyway, she is arguing by analogy. If arguments by analogy are to be forbidden, shouldn’t we discard every ID argument that extends a human experience of engineering or design to natural objects?

  2. The only examples of co-option she presents involve agency. How about just one example of co-option that does not rely on agency?

    This complaint is mystifying, as McLennan’s paper is chock full of examples of co-option in biology.

  3. Skeech,
    you have engaged in the logical fallacy called question begging. In other words, you have assumed the truth of the question we are trying to figure out. Are examples of co-option due to undirected, purposeless processes or by agency? The only things used to prove that co-option in biology must not be a result of agency, are examples that involve agency. Since no one saw biology do co-option, no one knows if it was done naturally or by artiface, so to answer the question we look to other examples of co-option, all of which seem to be a result of intelligent design.

  4. skeech,

    I foresaw your argument and edited the text. The only verifiable examples of co-option she offers are the result of agency. All of her biological examples are purely unsupported speculation, without any analysis concerning the variety of mutations required, the likelihood of their occurrence, the survival value of the transitional intermediates, the number of generations and individuals involved (i.e., the probabilistic resources) such that the transitional intermediates could be fixed in the population, etc. This is the kind of analysis one expects from a scientific hypothesis. Instead, we are offered an indoctrination piece in Darwinian speculation based on analogy to agent-driven design.

  5. 5

    I found it. Berra introduces the example with the sentence

    Everything evolves, in the sense of “descent with modification,” whether it be government policy, religion, sports cars, or organisms.”

    It seems to me that Berra’s example of Corvettes is used to illustrate descent with modification, not naturalistic evolution. If that’s what it’s trying to illustrate with the example, there’s no blunder at all, and Johnson is wrong. Johnson also wrongly seems to equate “common ancestry” with “evolution” in the passage.

    I would say that Berra’s example does illustrate common ancestry. What’s the blunder if it illustrates what he claimed it does?

  6. Galton remarked on this pitfall back in the day, didn’t he? And I’ll mention the Phylogenetic Tree of Mixed Drinks again.

  7. 7

    Gil, that you edited the text after it was posted suggests that you didn’t foresee skeech’s comment.

  8. Collin:

    Since no one saw biology do co-option, no one knows if it was done naturally or by artiface…

    Since no one watched all of the molecules in the atmosphere to make sure they followed the laws of physics, we don’t know if today’s sunny skies over California are natural or the work of the Designer. Shame on those meteorologists for assuming that the weather is an unguided phenomenon.

    Gil, you wrote of the “logical vacuity” of the idea of co-option under the Darwinian paradigm. Could you explain what you find illogical about it?

    Phillip Johnson:

    On the contrary, they show that what biologists present as proof of “evolution” or “common ancestry” is just as likely to be evidence of common design.

    Johnson, like so many ID supporters and creationists, makes the mistake of thinking that common design and Darwinian common descent lead to the same pattern of distribution of features. They do not. There are many different ways of employing common design, most of which do not yield the amazingly consistent nested hierarchies, across multiple traits, that we see in nature. For some reason the Designer seems to have singled one the one method that does so, thus creating the illusion of common descent. How odd.

    Gil Dodgen:

    It is illuminating that Darwinists, in many attempts to defend their blind-purposeless-undirected-chance-and-necessity hypothesis, draw on analogies from design and agency while trying to explain away agency and design. It seems to me that this exercise should result in at least a modicum of cognitive dissonance.

    Yeah, Darwinists are like those goofy chemists who refer to ‘hydrophilic’ molecules — as if a molecule could love! And yet those chemists don’t feel even a modicum of cognitive dissonance.

  9. Gil, I think there’s something you’re missing here. Most people use analogies as illustrations, rather than evidence. Deborah starts off by saying that co-option of traits is not difficult to understand, then points to examples of co-option that her audience is familiar with to make sure they understand the general concept. The examples are not presented as evidence for co-option in biology. That, as you know, comes from other sources, like fin/legs or jawbone/earbone fossils, etc.

    Because I.D.ers so often attempt to use their analogies as actual evidence, perhaps it’s easy for them to be mistaken on this point.

  10. But Skeech, meterologists have seen the molecules in the sky behaving according to the laws of physics, many times before They can test it, and they often do, and it is always behaving according to those laws.

    No one saw the co-option take place. They can’t and don’t test it. I assert that it is reasonable that the molecules are behaving according to the laws of physics, but it is unreasonable to assume that an unproven process called undirected co-option takes place merely because co-option of some kind (directed or undirected) is taking place.

    It is a nuanced point, and I’m trying not to say “God-did-it” but I am trying to say that undirected co-option is more an ideological assumption than proven fact.

  11. Gil, you wrote of the “logical vacuity” of the idea of co-option under the Darwinian paradigm. Could you explain what you find illogical about it?

    Click on the second link in the OP.

    Please read the instruction manual before calling tech support.

  12. Gil,

    I read your comment when it first appeared, and I reread it just now to make sure I didn’t miss anything. I don’t see anything in it that identifies a fault in the logic of co-option — just a proclamation of personal incredulity.

  13. skeech,

    It seems to me that you are offering an argument from incredulity that my incredulity is not credible. This logically results in believing anything — even when evidence, logic, and probabilistic mathematical analysis weigh heavily against a speculative thesis.

  14. No, Gil, your incredulity is not my reason for thinking that co-option can happen without guidance.

  15. When a Darwinist is asked how something happened they will say it evolved.

    When pushed for how it evolved they will say it was selected.

    When they really get pressured and are in trouble as to how something happened, they reach into their back side and pull out that it was exapted.

    But if they really get in trouble about how something happened there is the old reliable cure for all problems, it emerged.

    Evolutionary biology, the only science where one’s imagination is evidence.

  16. Is that really as high as your “science” can aim, Skeech? Whether or not something *can* happen? Aren’t you a tiny bit interested in whether or not co-option *did* happen? If you’d read the linked post by Dembski regarding “no great conceptual leap,” you’d have realized that you are echoing precisely what it was meant to address.

    I can conceptualize cards falling in such a way that my poker opponent receives one royal flush after another purely as the result of natural forces, but you’d better believe that as early as the second one I’m going to be questioning more than whether or not such a thing *can* happen.

    Can’t we set the bar for science just a bit higher here? Instead of thoughts on what can happen, how about evidence on what actually did happen? Is that bar too high?

  17. skeech:

    No, Gil, your incredulity is not my reason for thinking that co-option can happen without guidance.

    Phinehas:

    Is that really as high as your “science” can aim, Skeech? Whether or not something *can* happen? Aren’t you a tiny bit interested in whether or not co-option *did* happen?

    Logically equivalent dialogue:

    skeech:

    No, Gil, your incredulity is not my reason for thinking that birds can fly.

    Phinehas:

    Is that really as high as your “science” can aim, Skeech? Whether or not something *can* happen? Aren’t you a tiny bit interested in whether or not birds *have* flown?

    skeech:

    Who invited you to this party?

  18. skeech:

    A few comments. You say:

    “Johnson, like so many ID supporters and creationists, makes the mistake of thinking that common design and Darwinian common descent lead to the same pattern of distribution of features. They do not. There are many different ways of employing common design, most of which do not yield the amazingly consistent nested hierarchies, across multiple traits, that we see in nature. For some reason the Designer seems to have singled one the one method that does so, thus creating the illusion of common descent. How odd.”

    That’s correct, although IMO not definitive. That’s why most of us in ID believe in common descent: designed common descent. Pure common design is always a possibility, but I am well aware that there are arguments against it, such as those you cite.

    “Since no one watched all of the molecules in the atmosphere to make sure they followed the laws of physics, we don’t know if today’s sunny skies over California are natural or the work of the Designer.”

    This point, instead, is IMO unacceptable. We have very credible scientific models for the events you describe: they may not be perfect, but good models they are, with explicit reasoning about mechanical laws, probability and chaotic systems, and a reaslistic mathematical treatment of theose factors. The same is absolutely not true of darwinian theory, least of all of the cooption “hypothesis” (I fully share Gil’s reservations about the term).

    Just as an example, I quote here form the cooption paper:

    “In the first event one of the copies of a heat shock protein (copy 1) retained its plesiomorphic function (protection against damage) and plesiomorphic location (throughout the body), while the other (copy 2) added a change in
    function (focusing light) to the old function of protection from degradation in the plesiomorphic location. The second duplication event involved copy 2; as before, one copy retained the (new) plesiomorphic function3 (focusing light, protection against damage) and the plesiomorphic location
    (many tissues), while the other copy retained the plesiomorphic
    function coupled with a change in location (restricted to the lens).”

    That’s what we in ID call, deservedly, a “just so story”. There is no discussion of causal factors, no analysis of probabilities, no credible detaile path. Just blind faith that darwinian evolution can do everything we demand of it. For instance, how was the “change in location” achieved? By a random mutation in the rehulatory, and vastly unknown, procedures which determine the emergence of specific transcriptomes in different cells? How credinle is that assumption? How credible are all the other assumtpions made in that single paragraph? Indeed, is there anything else than gratuitous assumtpions in that paragraph? and in the whole paper? Or in the whole concept of cooption?

    Well, maybe the evolution of Corvettes is a fact. Ah, but I forgot that it was just an analogy…

  19. @gpuccio:
    I totally agree with you about the “just so stories”. When you hear a darwinist talk about the evolution of the eye for instance they are not describing darwinian pathways, but describing design steps or brilliant optimisations of the eye without even knowing it. The “change in location” problem has always fascinated me, because the systems that determine location(both in space and time in development) are very complex.

  20. 20

    gpuccio [15]:

    I quote here form the cooption paper:
    “In the first event one of the copies of a heat shock protein (copy 1) retained its plesiomorphic function (protection against damage) and plesiomorphic location (throughout the body), while the other (copy 2) added a change in
    function (focusing light) to the old function of protection from degradation in the plesiomorphic location. The second duplication event involved copy 2; as before, one copy retained the (new) plesiomorphic function3 (focusing light, protection against damage) and the plesiomorphic location (many tissues), while the other copy retained the plesiomorphicfunction coupled with a change in location (restricted to the lens).”
    That’s what we in ID call, deservedly, a “just so story”. There is no discussion of causal factors, no analysis of probabilities, no credible detaile path. Just blind faith that darwinian evolution can do everything we demand of it. For instance, how was the “change in location” achieved? By a random mutation in the rehulatory, and vastly unknown, procedures which determine the emergence of specific transcriptomes in different cells? How credinle is that assumption? How credible are all the other assumtpions made in that single paragraph? Indeed, is there anything else than gratuitous assumtpions in that paragraph? and in the whole paper? Or in the whole concept of cooption?

    Someone with a scientific background should know the differences between (a) a review written for a general audience (the paper in Evo. Edu. Outreach by Deborah A. McLennan), (b) a review written for a specialist audience, and the primary peer-reviewed literature.
    Perhaps you missed the opening sentence of the paragraph from which your quotation was taken: “Crystallins are soluble proteins found in the lens of all vertebrates examined to date and some invertebrates (the following summary is taken from (True and Carroll 2002) and references therein).”

    The True and Carroll paper (Annu. Rev. Cell Dev. Biol. 2002. 18:53–80) is an example of (b), a specialist review. If you peruse it, you will find discussions of crystallin and lens evolution on pages 57-61 supported by numerous citations to the primary literature.

    It is inaccurate to disparage the scientific status of evolutionary co-option on the basis of a paragraph taken from a type (a) review.

  21. 21

    Sorry, I was referring to gpuccio [18]

  22. I wrote:

    Johnson, like so many ID supporters and creationists, makes the mistake of thinking that common design and Darwinian common descent lead to the same pattern of distribution of features. They do not. There are many different ways of employing common design, most of which do not yield the amazingly consistent nested hierarchies, across multiple traits, that we see in nature. For some reason the Designer seems to have singled one the one method that does so, thus creating the illusion of common descent. How odd.

    gpuccio replies:

    That’s correct, although IMO not definitive. That’s why most of us in ID believe in common descent: designed common descent. Pure common design is always a possibility, but I am well aware that there are arguments against it, such as those you cite.

    Actually, the problem is even worse than that for the ID proponent, and conceding the truth of common descent doesn’t solve it.

    “Designed common descent”, as you call it, is also at odds with the evidence. This is because phylogenetic trees are inferred using a principle of parsimony: when two possible trees are being considered, the one involving the least evolutionary change is preferred.

    The fact that this heuristic yields consistent nested hierarchies across traits indicates not only that common descent is true, as you concede, but also that evolutionary changes are incremental.

    Darwinian theory predicts small changes, and is therefore compatible with the evidence. A designer, however, can make changes of any size he likes.

    Conclusion: If common design is true, then the designer specifically chose a method of design that both conforms to a nested hierarchy and limits itself to incremental changes. In other words, out of a huge number of possible design strategies, he chose the only one that makes it appear as if he’s not there at all.

    This is a huge problem for ID, and one that deserves to be better known among ID supporters.

  23. Adel:

    thank you for the references, I will look at them. Anyway, I have many doubts that the “just so stories” in that paper will be solved. we will see. In the meantime, there were no specific references in the paper to support the affirmations I quoted. But let’s postpone the discussion to later, when I have read your references (if I can find them in the internet).

  24. skeech:

    “Darwinian theory predicts small changes, and is therefore compatible with the evidence. A designer, however, can make changes of any size he likes.”

    Wrong. And for two different reasons.

    First of all, you seem to have too much faith in gradual changes, even for an orthodox darwinist. Have you entirely forgotten OOL, punctuated equilibriums, ediacara explosions, cambrian explosions, flower explosions, and so on?

    Second, and most important, I am always amazed at how much darwinists seem to know of the designer, his modalities, his purposes, and his nature, when we in ID are explicitly ignorant about that. What do you know of what a designer of biological information can do, wants to do, likes to do, and so on?

    Where does your strange security comes from, that a designer should implement his design according to your wishes, according to your perception of time and space, and according to you ideas of his purposes?

    Personally, I do believe, and have always believed, that a designer, even if he be an omnipotent god, can well act according to specific constraints, once he is acting in a specific context. Any context spontaneously generates constraints. And modalities. And forms. Biological information is a form (indeed, an incredibly rich set of different forms) inside other forms. Any design of biological information, and any implementation of that design, is bound to have specific forms and times and modalities.

    But among those constraints, there is one that I firmly believe has never worried the designer, whoever or whatever he is: the necessity to please darwinists and to comply with their expectations.

    So, I see no huge problem for ID (indeed, not even a small problem), but I am grateful for your willingness to help.

  25. Adel:

    I would have like to comment the paper by True and Carroll in detail, but unfortunately I have no access to it. The best I could find was a discussion on discovery.org about it (“Neo-Darwinism’s Unsolved Problem of the Origin of Morphological Novelty”), where a passage of it is quoted:

    “The mechanisms by which duplication and transposition bring about co-option of novel gene functions have thus far been hidden from view because functionally important polymorphisms [i.e., variations] involving these events are difficult to identify. The next phase of evolutionary developmental biology research must address this paradox by investigating the levels, causes, and consequences of microevolutionary variation in developmental systems.” (p. 74)

    But I understand that it’s not much.

    Anyway, I eally don’t see how True and Carroll could prove a detailed mechanism by which, through credible random variation, events like “the other copy retained the plesiomorphicfunction coupled with a change in location (restricted to the lens)” took place.

    If you have access to the paper, and if you can sum up some of the evidence, I would be very much interested in discussing it.

  26. skeech at 22

    Darwinian theory predicts small changes, and is therefore compatible with the evidence. A designer, however, can make changes of any size he likes.
    . . .In other words, out of a huge number of possible design strategies, he chose the only one that makes it appear as if he’s not there at all.

    The “minor” problem with your statement is that it is diametrically at odds with facts. See:

    Intelligent Design: The Origin of Biological Information and the Higher Taxonomic Categories
    By: Stephen C. Meyer, Proceedings of the Biological Society of Washington May 18, 2007

    There is no credible possibility that small incremental changes could have developed such massive increases in biological information in such a short time – followed by stasis.
    That burst with stasis is commonly seen in engineering design.
    Applying actual facts to your logic results in the opposite conclusion.

  27. Skeech, in what way are those logically equivalent? Are you implying that you think that co-option can happen in the same way you think that birds can fly?

    Instead of two-stepping around the issue, maybe you could directly address:

    Have you seen co-option in action the same way you’ve seen birds flying?

    Do you have any evidence whatsoever that co-option did happen?

  28. They’re equivalent because in both cases, ‘skeech’ says that something ‘can happen’, and ‘Phinehas’ jumps to the absurd conclusion that this means that skeech is unwilling to claim that it has happened.

    Do you have any evidence whatsoever that co-option did happen?

    Of course. Didn’t you read the McLennan paper?

    Here McLennan describes one of the best-known examples of co-option:

    All amphibians, mammals and reptiles have this middle ear bone (now called the stapes) thanks to the modification of the hyomandibula, which itself was a modification of the epihyoid, which itself was one part of the original acellular cartilaginous rod in the two pharyngeal slits that arose in the ancestor of the deuterostomes. No new structures had to appear to support this part of the transition from filter feeding to respiration to predation to sound detection. All that was required was the co-option, followed by modification, of the basic building blocks that had been laid down 900 mya ((for a thought provoking, incredibly readable discussion of the importance of co-option in the evolution of animal bodies, see (Shubin 2008)).

  29. DLH wrote:

    There is no credible possibility that small incremental changes could have developed such massive increases in biological information in such a short time – followed by stasis.

    According to Stephen Meyer and the Discovery Institute, no. According to biologists, yes.

  30. gpuccio wrote:

    First of all, you seem to have too much faith in gradual changes, even for an orthodox darwinist. Have you entirely forgotten OOL, punctuated equilibriums, ediacara explosions, cambrian explosions, flower explosions, and so on?

    gpuccio,

    You’ve fallen prey to a common misconception about these phenomena. This excerpt from the Wikipedia article on punctuated equilibrium explains your error:

    Punctuated equilibrium is often confused with George Gaylord Simpson’s quantum evolution,[10] Richard Goldschmidt’s saltationism,[11] pre-Lyellian catastrophism, and the phenomenon of mass extinction. Punctuated equilibrium is therefore mistakenly thought to oppose the concept of gradualism, when it is actually a form of gradualism, in the ecological sense of biological continuity.[3] This is because even though evolutionary change appears instantaneous between geological sediments, change is still occurring incrementally, with no great change from one generation to the next.

    gpuccio then fulminates:

    Second, and most important, I am always amazed at how much darwinists seem to know of the designer, his modalities, his purposes, and his nature, when we in ID are explicitly ignorant about that. What do you know of what a designer of biological information can do, wants to do, likes to do, and so on?

    Where does your strange security comes from, that a designer should implement his design according to your wishes, according to your perception of time and space, and according to you ideas of his purposes?

    …But among those constraints, there is one that I firmly believe has never worried the designer, whoever or whatever he is: the necessity to please darwinists and to comply with their expectations.

    Your tirade is odd, considering that I’ve said absolutely nothing about the designer’s desires or purposes. Nor have I expressed any expectations regarding the designer.

    Here are my words, verbatim:

    Conclusion: If common design is true, then the designer specifically chose a method of design that both conforms to a nested hierarchy and limits itself to incremental changes. In other words, out of a huge number of possible design strategies, he chose the only one that makes it appear as if he’s not there at all.

    I have commented only on what the putative designer did, not on his purposes. You did exactly the same thing when you suggested that the designer worked by “designed common descent.”

    ID proponents like to talk about “following the evidence where it leads,” so let’s do that now. There are two possibilities:

    1. There is a designer, but he chose to design in a way that makes it appear as if he’s not there.

    2. There is no designer.

    Note that in both cases the evidence points to the absence of a designer. If you accept #1, you’re doing so in spite of the evidence, not because of it. In other words, to “follow the evidence wherever it leads” is to conclude that there is no designer. To insist that there is a designer is to abandon science and resort to faith.

    You’re free to do so, of course — but please don’t claim that you’re doing it for scientific reasons.

  31. skeech:

    Wikipedia is not necessarily truth, you know, and we in ID have a slight feeling that it is a little bit biased against us. And you did not comment about all the other examples of “saltation” i mentioned. Will you offer us even for them a comforting reassurance from wikipedia that they are perfectly compatible with pure gradualism?

    Regarding the second point, you may notice that my argument started with the following:

    “I am always amazed at how much darwinists seem to know of the designer,”

    So, it should be evident that I was rather widening the argument, including other similar affirmations made by other darwinists in the past. If you felt that I was including you in all the examples I offered, I apologize. But you did make a similar statement about the designer, about his methods and capacities if not about his purposes. Here it is:

    “A designer, however, can make changes of any size he likes.”

    and then:

    “the designer specifically chose a method of design that both conforms to a nested hierarchy and limits itself to incremental changes”

    So, my answer remains perfectly valid: what do you know about what the designer can do or likes to do? what do you know about his choices?

    First of all there is no evidence that design is implemented “only” in incremental changes. Indeed, there is no evidence at all that it is implemented in incremental changes. For all supposed transitions (always vague) from one protein to another, the supposed intermediate forms are not known, and probably never existed. Incremental pathways of evolution at the molecular level (let’s say, through selectable changes of one-two aminoacids at a time) are exactly what darwinists never detail, and IDists always ask for. There is no known incremental molecular pathway for OOL. There is no incremental explanation for ediacara explosion and cambrian explosion, whatever wikipedia, you, or anybody else may imagine.

    So, was OOL incremental? We don’t know, but there is absolutely no evidence of that. Was the emergence of the bacterial flagellum incremental? there is no evidence of that. Was the emergence of the fundamental proteins which are present even in the most simple organisms incremental? Please show the evidence. Was the emergence of the genetic code incremental? And where are the intermediate codes, other than in the fantasy of darwinists?

    So, excuse me, but I maintain my statement: you seem to have too much faith in gradual changes, even for an orthodox darwinist.

    About the designer, you say a lot of arbitrary and wrong things. You say:

    “he chose to design in a way that makes it appear as if he’s not there”

    What does that mean? Even if the designer acted only gradually (which is not proven at all), why should that make appear as if he’s not there? You seem to be very confused about what proves the existence of the designer according to ID: it is not certainly the chronological modality of the implementation, but rather the formal properties of the designed product and the absence of any credible non design explanation of biological information. Time modalities have no role in that argument. So, why do you attribute so much importance to them?

    Even if the designer acted gradually, his presence is obvious because of the nature of the things he has designed: he may have designed them gradually in one billion years, or in one day, nothing changes: designed they are just the same. Moreover, if the designer is a God, he is probably not in a hurry, and his ideas of time may differ a little from ours.

    Do you usually judge the value of a work of art according to the time it required to be created? Or the graduality of the process? Does a painter paint gradually, or does he usually paint for one day, then stops for ten years, then paint again for ten minutes, then stop again for one year? Is he more a designer if he acts that way? Or is he a mechanical random machine if he paints regularly throughout the time he has available?

    You say: “out of a huge number of possible design strategies, he chose the only one that makes it appear as if he’s not there at all”. Again, what do you know of the “huge number of possible design strategies”? Are you an expert of the available strategies in the context where biological information emerged? Could you have done better? And what do you know about choices available to the designer, and about the possible motives why a choice could have been made? You seem to imply that, were you the designer, you would have made your presence more obvious. But you know, to many of us that presence is very obvious, although not arrogantly self-emphasized.

    So “the evidence points to the absence of a designer” only in your very biased and superficial reasoning. Please don’t ask that I comply with it. And I maintain that I make my scientific choices for scientific reasons. You are free to disagree, and I will not probably be too unhappy for that.

  32. All amphibians, mammals and reptiles have this middle ear bone (now called the stapes) thanks to the modification of the hyomandibula, which itself was a modification of the epihyoid, which itself was one part of the original acellular cartilaginous rod in the two pharyngeal slits that arose in the ancestor of the deuterostomes.

    Just so. :/

    This qualifies as evidence how?

    I have not read the paper, but I’m curious to know whether it goes beyond telling stories to, at the least, giving some sort of scientific description of these modifications. Was it modified through selection, exaptation, or emergence?

    Or maybe I just don’t have the requisite imagination?

  33. 33

    skeech,

    “ID proponents like to talk about “following the evidence where it leads,” so let’s do that now. There are two possibilities:

    1. There is a designer, but he chose to design in a way that makes it appear as if he’s not there.

    2. There is no designer.

    Note that in both cases the evidence points to the absence of a designer. If you accept #1, you’re doing so in spite of the evidence, not because of it. In other words, to “follow the evidence wherever it leads” is to conclude that there is no designer. To insist that there is a designer is to abandon science and resort to faith.

    You’re free to do so, of course — but please don’t claim that you’re doing it for scientific reasons.”

    There are more than those two possibilities.

    1. The Designer chose to design in such a way that we cannot discern his intentions of being known.

    2. The Designer designed in such a way that we can discern whether or not he intended on making himself known.

    3. The Designer designed in such a way that we can know that he intended not to be known.

    4. The Designer designed in such a way that we can know his design and know that he intended on us knowing his design and him being known through it.

    5. The Designer lacks the capacity to show himself through the design.

    6. The Designer lacks the capacity to hide himself through the design.

    7. The Designer doesn’t care one way or the other to show himself through the design and hasn’t done so.

    8. The Designer does care one way or the other to show himself through the design and has done so.

    This could go on……………

    Number three looks like a contradiction, because if he really were to make things look like he weren’t there, and if he succeeded, we wouldn’t discern him. But, nevertheless, it would still be designed, we just wouldn’t know it. But the evidence wouldn’t speak one way or the other about it, not for it or against it. So to follow the evidence would never tell you that there wasn’t a Designer–it would only tell you that you don’t really know one way or the other. In your #1, there is a Designer, and the conclusion that he doesn’t exist would be false, even though the evidence led you to that conclusion. But your #1 is really an inference, at all costs; and the inference could, realistically, go the other way, and evidence that the Designer is present, and the continuing inference that he is making himself known.

  34. On Berra’s Blunder:

    Clearly, there is a need to realise that the only descent involved with the Corvette was in the minds of designers [in the various depts of GM over the years from the 1950's on], and that the only modifications were in the minds of designers. Common design, not reproductive descent with modifications driven by spontaneous random variation and environmental natural selection.

    GEM of TKI

  35. 35

    kairosfocus, no, the descent is not just in the minds of designer but in the alterations of function. You have misunderstood the point of Berra’s example just as Johnson did.

  36. 36

    This is strange. At 7:50 AM today, I posted a comment and it appeared as number 35 in this thread.

    It has now, at 3:30 PM, disappeared.

    Moderators, what is going on?

  37. 37

    Clive Hayden [101 on another thread]:

    03/29/2009
    3:06 pm
    madsen,

    I personally have not deleted any comments at all except from David Kellog asking to be taken out of moderation. We switched UD to a new server, so that may explain these strange occurrences.

    Ah, maybe I’ve been another victim of server exchange. I’ll try again later.

  38. 38

    One more time.

    gpuccio [23]:

    thank you for the references, I will look at them. Anyway, I have many doubts that the “just so stories” in that paper will be solved. we will see. In the meantime, there were no specific references in the paper to support the affirmations I quoted. But let’s postpone the discussion to later, when I have read your references (if I can find them in the internet).

    You say this after I pointed out that the first sentence in the paragraph from which you quoted cited True and Carroll (2002). I didn’t mention it, but there is another supporting reference in that paragraph: Graw (1997). So your statement that there were no specific references in the paper is incorrect. Both of these references can be found on the Internet, but each will cost you about $20/15 Euros for online access.

    gpuccio [25]

    I would have like to comment the paper by True and Carroll in detail, but unfortunately I have no access to it.

    I sympathize with your desire for rapid (and free) access to supporting documents. My original point was that due diligence for a critic should (depending on sincerity) involve some effort at exploring a topic beyond ridiculing (and misrepresenting) a quotation from a general review. Do you concede that point collegially?

  39. Adel:

    I had answered your #38, but my post too has disappeared. I had not saved it, so I cannot just paste it again.

    In brief, I wanted to say that, while agreeing that we must do our best to deepen our analysis of the topics we discuss, still your point seems a little too formal: my comment about the paper in question wanted only (and still wants) to point out that in the paper in wuestion there were no explicit arguments in support of cooption, but only just-so stories (please remember that the paper had been given here as evidence in favour of the concept). So, when I said that “there were no specific references in the paper to support the affirmations I quoted” I was not using references in the technical sense of literature quotations, but rather in the sense of “discussion” (I apologize for that, I am not an english speaking contributor, and sometimes I miss possible shadows of meaning when I write in a hurry).

    So, I maintain that the paper in question did not contribute with any evidence to our understanding of cooption, and even if it was only a review, a good review should have given at least some glimpse of the arguments, and not only of conclusions which are given as established (“Phylogenetic and functional analyses indicated that
    ? crystallins originated following two duplication events”). And if the review was aimed to the general public, that is even less correct, because the general public will probably accept anything it says as established knowledge, without checking the literature support.

    The pint is more general than it may seem. We in ID do disagree on many assumptions which are usually considered established truth in evolutionary literature. In particular, we do disagree on the interpretations of many common observations made in biology, and on the causal inferences usually made about them. Just to give an example, I do disagree that the simple observation of homology between two proteins, even if statistically significant, is any proof of causal derivation of one from the other, in the absence of a credible causal pathway. Homology is present at various levels, at various degrees, and can mean various things; moreover, it should always be interpreted in reference to how much we know (or do not know) of the function.

    For instance, the discussion I quoted from the discussed paper, as it is in that paper, and probably even with the supporting arguments in the original paper I have not read, could be acceptable as a very speculative theory about the “natural history” of the appearance of a protein, but it is not in any way an established causal pathway of it. The same generic events summed up there could well be the description of a designed pathway, while an implicit assumption is usually made in evolutionary literature that any description of a possible formal pathway is proof of its causal consistency and that the causal mechanisms implied can only be the officially recognized ones, that is RV and NS.

    That is what we are discussing here: we are not discussing if cooption is something which can generically happen, but rather if cooption is a credible causal mechanism which can help explain biological information. The difference between “just so stories” and science is exactly that: to pass from a just so story to a credible natural history you need to give evidence that not only things could have happened that way, but that they probably happened that way; and to pass from a likely natural history to a credible casual explanations you have to give evidence that they could and did happen that way for the reasons you imagine (in other ways, that your causal model is consistent and empirically credible).

  40. 40

    DLH, David L Hagen, in message #26

    I noted that the paper by Stephen C. Meyers – “Intelligent Design: The Origin of Biological Information and the Higher Biological Categories” 2007, fails to mention my paper -

    “Ontogeny, Phylogeny and the Origin of Biological Information” Rivista di Biologia 93: 513-524, 2000.

    I mention this just to establish priority.

  41. 41

    gpuccio [39]:

    Thank you for your detailed and helpful response. I find your openness, honesty and civility delightful, and this post of yours is exemplary in those respects. This encourages me to ask if you will provide some further assistance to my understanding.

    In brief, I wanted to say that, while agreeing that we must do our best to deepen our analysis of the topics we discuss, still your point seems a little too formal: my comment about the paper in question wanted only (and still wants) to point out that in the paper in wuestion there were no explicit arguments in support of cooption, but only just-so stories (please remember that the paper had been given here as evidence in favour of the concept). So, when I said that “there were no specific references in the paper to support the affirmations I quoted” I was not using references in the technical sense of literature quotations, but rather in the sense of “discussion” (I apologize for that, I am not an english speaking contributor, and sometimes I miss possible shadows of meaning when I write in a hurry).

    No need to apologize, Dear Fellow. Your English is excellent for a speaker of a different language, and I now understand why we were miscommunicating on the issue of references.

    So, I maintain that the paper in question did not contribute with any evidence to our understanding of cooption, and even if it was only a review, a good review should have given at least some glimpse of the arguments, and not only of conclusions which are given as established (”Phylogenetic and functional analyses indicated that ? crystallins originated following two duplication events”). And if the review was aimed to the general public, that is even less correct, because the general public will probably accept anything it says as established knowledge, without checking the literature support.

    When I used the term “general” for the McLennan paper, I assumed that you would have surmised from the name of the journal in which it was published, Evolution and Education Outreach, that its intended audience was not the general public, but teachers at the secondary and baccalaureate levels. The distinction I was making was between the intended audience of such a review and the intended scholarly and specialist audience of the True and Carroll review.

    In any case, how would you have dealt with the shortcomings you perceive in the McLennan review? Please help me understand what kinds of arguments McLennan should have given by providing an example of what you consider to be a more correct approach.

    (The statement that you quoted: “Phylogenetic and functional analyses indicated that ? [alpha] crystallins originated following two duplication events.” is, incidentally, perfectly accurate.)

    The pint is more general than it may seem. We in ID do disagree on many assumptions which are usually considered established truth in evolutionary literature. In particular, we do disagree on the interpretations of many common observations made in biology, and on the causal inferences usually made about them. Just to give an example, I do disagree that the simple observation of homology between two proteins, even if statistically significant, is any proof of causal derivation of one from the other, in the absence of a credible causal pathway. Homology is present at various levels, at various degrees, and can mean various things; moreover, it should always be interpreted in reference to how much we know (or do not know) of the function.

    I thought you understood that there is no “established truth” in science. Correct me if I’m wrong, but I thought you wrote that very thing on another thread. All scientific inferences are hypotheses, are they not? So, given the observation of sequence similarity between two genes, how would you go about disproving that they both derived from a common ancestor? Or, alternatively, how would you construct a “credible causal pathway” leading to the opposite conclusion? Again, an example of your work in either case would be most helpful and appreciated.

    For instance, the discussion I quoted from the discussed paper, as it is in that paper, and probably even with the supporting arguments in the original paper I have not read, could be acceptable as a very speculative theory about the “natural history” of the appearance of a protein, but it is not in any way an established causal pathway of it. The same generic events summed up there could well be the description of a designed pathway, while an implicit assumption is usually made in evolutionary literature that any description of a possible formal pathway is proof of its causal consistency and that the causal mechanisms implied can only be the officially recognized ones, that is RV and NS.

    Excellent. You have highlighted the competition between evolutionary theory and design theory with concise eloquence. Therefore, I ask you now to make the distinction clear by describing the alternative causal mechanisms established or envisioned by design theory. And it would be especially helpful if you would stipulate how one would go about deciding which explanation is the better one.

    That is what we are discussing here: we are not discussing if cooption is something which can generically happen, but rather if cooption is a credible causal mechanism which can help explain biological information. The difference between “just so stories” and science is exactly that: to pass from a just so story to a credible natural history you need to give evidence that not only things could have happened that way, but that they probably happened that way; and to pass from a likely natural history to a credible casual explanations you have to give evidence that they could and did happen that way for the reasons you imagine (in other ways, that your causal model is consistent and empirically credible).

    I am pleased to agree that it is a goal of science to provide credible causal explanations. Therefore, I ask you to present the alternative credible causal explanation of the origin of alpha crystallins based on design theory or any alternative theory you may prefer. (If my estimation of your character is correct, I am confident that you would not require evolutionary theory to meet a standard that your own theory does not meet.)

  42. Greetings Mr. Kellogg,

    I’d like to take a stab at answering your questions:

    “Do you [Gil?] know what precisely Berra was trying to illustrate in that passage?”

    I do not know what he was trying to illustrate, but I do know what he illustrated. He was illustrating that descent with modification should be obvious because we see it in the changing designs of automobiles. He was also illustrating that evolution should be obvious because of that same automobile example. The first illustration is explicit; the second is implicit, more on that later when I answer your question about the “blunder”.

    “Does anybody have the context?”

    Yes, I do. Mr. Berra’s book title offers a great starting point for context as does the blurb that you cited.

    “Is he saying that Corvettes evolved naturalistically, or is the context about evidence for common descent?”

    Here, I believe you have posited a false dilemma. None of us thinks Corvettes evolved naturalistically, but because of that are we to agree that Berra is writing about common descent? Not at all. One other (of many possible) option is that Berra is suggesting that the evolutionary process (for living things) that is evidenced, however unconvincingly from some points of view, by the fossil record is supported by the change in automobiles designs even though automobile design is a different process altogether.

    “I would say that Berra’s example does illustrate common ancestry. What’s the blunder if it illustrates what he claimed it does?”

    Although common ancestry is certainly in the mix, he is also writing about evolution and by implication natural selection and random mutation.

    Here is Berra:

    “The point is that the Corvette evolved through a selection process acting on variations that resulted in a series of transitional forms and an endpoint rather distinct from the starting point. A similar process shapes the evolution of organisms.” (my italics)

    Now, I think we can all agree that common ancestry is not a process, but an artifact. In the context of a pro-evolution book, and in a section using the fossil record (of all things) for support, it is when Mr. Berra jumps over to “process” that he blunders.

    Mr. Johnson rightly enjoins the argument here. How are these processes similar/dissimilar? That is the question! It is exactly at this point that Mr. Johnson notes that the process in automobile development is not naturalistic, but design.

    If I may say so, game over.

  43. H’mm:

    The point is that the Corvette evolved through a[N INTELLIGENTLY DRIVEN] selection process acting on variations that resulted in a series of transitional forms and an endpoint rather distinct from the starting point. A similar process shapes the evolution of organisms.” (

    $0.02

    GEM of TKI

  44. 44

    Tim [42], game not over, no. By “context” I didn’t mean “the title of the book” (we all knew that already) or “repeat the passage with added emphasis.” We all know he’s making the case for evolution in the book. But what is the context of the quoted discussion?

  45. 45

    David Kellog:

    I would say that Berra’s example does illustrate common ancestry. What’s the blunder if it illustrates what he claimed it does?

    Varying models of Corvettes do not illustrate common ancestry, as no car has ever had an ancestor except in the figurative sense. There’s no analogy between living things producing offspring over multiple generations and someone repeatedly producing new objects from raw materials.

  46. Testing to see if my comments are still being placed in the moderation queue, plus a reply to earlier comments by Clive Hayden and gpuccio.

    Clive and gpuccio,

    Both of you seem to be missing my point, so allow me to restate the argument more explicitly:

    1. Evolutionary theory predicts that species are formed by descent with modification from common ancestors. This produces a branching pattern of descent — an evolutionary tree.

    2. Evolutionary theory predicts that intergenerational genetic change is gradual. It predicts, for example, that you won’t see thirteen coordinated mutations arising in a single generation to produce a novel trait. The odds are just too steep. (Note that this says nothing about the pace of evolution over longer timescales, which is why there is no conflict between gradualism and the theory of punctuated equilibrium or the multi-million year Cambrian “explosion”).

    3. Because of #1 and #2, it is possible to reconstruct phylogenetic trees based on morphological and molecular evidence, subject to principles of parsimony and maximum likelihood.

    4. Because of #1 and #2, evolutionary theory predicts that phylogenetic trees reconstructed on the basis of multiple independent traits and molecules will be consistent.

    5. In fact, these methods have produced stunning results. As Douglas Theobald explains in his 29+ Evidences for Macroevolution:

    …the standard phylogenetic tree is known to 38 decimal places, which is a much greater precision than that of even the most well-determined physical constants. For comparison, the charge of the electron is known to only seven decimal places, the Planck constant is known to only eight decimal places, the mass of the neutron, proton, and electron are all known to only nine decimal places, and the universal gravitational constant has been determined to only three decimal places.

    6. Therefore, we conclude that the predictions of evolutionary theory — which does not invoke a designer — are wildly successful.

    7. Can the hypothesis of a designer be reconciled with these results? Yes — because literally any set of observations can be explained by invoking a designer. Why are things the way they are? Because the designer made them that way.

    8. This, obviously, is no longer science. It’s faith. Starting with the assumption of a designer and then interpreting the evidence to fit that assumption is not “following the evidence where it leads” — it’s leading the evidence where you want the conclusion to follow.

    9. Science, when practiced correctly, asks for the best, most economical explanation of the observational data. It does not demand that the data be reinterpreted to fit a lesser, uneconomical explanation.

    10. The designer hypothesis does not predict common descent. It does not predict gradualism. It does not predict the congruence of phylogenetic trees reconstructed on the basis of distinct sets of independent characters and molecules. These are possible under the designer hypothesis, but they are not predicted by it.

    11. They are, however, predicted by evolutionary theory.

    12. There are literally millions of design schemes that do not produce the appearance of common descent, gradual change, and consistent phylogenies.

    13. Of the millions of possible design schemes, only a minuscule fraction would produce the appearance of these things.

    14. Given these facts, which of these explanations is more economical (think Occam’s Razor) and fits the data best?

    A) There is a designer, and he happened to choose one method of design out of millions that happens to create the appearance that we would expect if he weren’t there at all. (Note that I have said nothing about the designer’s motives, abilities or constraints).

    B) There is no designer.

    15. If you choose (A), then you are choosing to believe in a designer despite the evidence, not because of it. That’s not science. It’s faith.

  47. Greetings again to Mr. Kellogg,

    Well, I’m not sure how to respond. You asked for context. I provided it by means of the title and refering readers to the very citation you gave. Your comment about added emphasis is spurious as my added emphasis had nothing to do with providing context, but the issue at hand.

    I chose to emphasize those words and sections to call readers’ attention to the blunder which in my opinion was Mr. Berra’s use of so much evolution-baggage-laiden terminology in the illustration and of course the term, process.

    Mr. Kellogg, what is your interpretation of Mr. Berra’s use of process?

  48. 48

    Skeech,

    I reject common descent. Therefore the Designer isn’t denied his design in light of it. And I’m really skeptical of what we should “expect” from the Designer, as far as knowing what he would or wouldn’t do, which is what you’re implying, even though you say you’re not implying it. I’m not sure what you mean about literally millions of possible design schemes that do not produce the appearance of common descent, gradual change, and consistent phylogenies. What are just a few of these literal millions? Do you mean a design appearance that doesn’t show common descent? I think that’s what we have. Do you mean a design that doesn’t show gradual evolution? That’s exactly what we have too. Again, we cannot get into the motives of the Designer, which is what you do, even though you say you don’t. And again, by your scheme, the conclusion that there is no Designer would be wrong, even though he tried to hide himself.

  49. Clive Hayden (#33): “To insist that there is a designer is to abandon science and resort to faith.”

    Error.

    Faith is based on facts—that’s why we have faith. Your use of the word “faith” assumes as true a pro-Atheism corruption of the concept.

    Ray

  50. 50

    Scott Andrews,

    There’s no analogy between living things producing offspring over multiple generations and someone repeatedly producing new objects from raw materials.

    You mean there’s no analogy between things produced in nature and the alleged activities of the designer?

  51. 51

    Ray,

    I didn’t write what you credit me with in your quote. I was quoting Skeech, he wrote it. I know faith is based on facts. I have facts that back up and support my faith in my friends and family. It’s no different with any person–what we deal with when we are dealing with a person or a personality is faith; we have faith in them based on facts. The more we know the person, the more faith we may have in our assurances in how they’ll deal with us, and whether we believe them in their actions.

  52. I hope this comment doesn’t go into the moderation queue like all of the others I’ve submitted today.

    Clive,

    If you read the entire thread, you’ll see that the designer hypothesis has a huge problem whether or not common descent is true.

    In any case, like it or not, the evidence points overwhelmingly to common descent. Take another look at Douglas Theobald’s point about the accuracy of the phylogenetic tree:


    …the standard phylogenetic tree is known to 38 decimal places, which is a much greater precision than that of even the most well-determined physical constants. For comparison, the charge of the electron is known to only seven decimal places, the Planck constant is known to only eight decimal places, the mass of the neutron, proton, and electron are all known to only nine decimal places, and the universal gravitational constant has been determined to only three decimal places.

    If phylogenetic trees based on independent traits converge with that kind of stunning accuracy, then we can be logically certain that either a) there is no designer, or b) the designer exists, but he chose a design method that makes common descent appear absolutely, overwhelmingly true.

    Contrary to your repeated assertions, I do not claim to know why a designer would choose such a strategy (indeed, I see it as a purely academic point since there is no reason to think that a designer exists in the first place). Is it because he wants to remain hidden? I don’t know. Because he wants to test our faith? Can’t say. Because he is constrained to design this way in order to achieve some other, unknown goal? Who knows?

    It doesn’t matter for our purposes. The question is not why a designer would design that way. Our task is to determine which explanation is most economical and best fits the data.

    You asked for an examples of design schemes that would not conform to neo-Darwinian expectations.

    That’s easy. I could design a set of cars so that if you tried to infer a “phylogenetic tree” based on the types of engines used, you’d get one tree, but if you did the same based on air conditioners or radios or fuel injectors you would get completely different trees that bore no resemblance at all to the first tree.

    We don’t see this in nature. Real phylogenetic trees are amazingly consistent. The evidence points overwhelmingly to common descent.

    And as I pointed out to gpuccio earlier in the thread, conceding the truth of common descent doesn’t solve the problem. See my earlier comments for details.

  53. skeech plus:

    I still can’t follow your reasoning. You point #1 is common descent, which I do accept. So please, at least with me, don’t bring again evidence of common descent: I agree.

    But where is your evidence of gradualism? Especially at the molecular level? Where are the pathways which bring from one protein to a different one with coordinated mutations of not more than two or three aminoacids (the only realistically achievable ones), and all the intermediate states selectable? Please, answer that, and give specific examples.

  54. 54

    David Kellog:

    You mean there’s no analogy between things produced in nature and the alleged activities of the designer?

    That’s not what I said, but it’s an accurate statement. You can’t draw an analogy between something and itself. There is no analogy between a Jetta and a car because a Jetta is a car.
    But that wasn’t the point.
    A good analogy sheds light on one or both of its subjects, helping others to understand them better.
    Comparing deliberate changes in a the design of a car to unintended variations in animals caused by undirected breeding doesn’t shed light on either. If someone accepted the premise, his understanding of at least one of its subjects would be diminished, not increased. All analogies break down, but they should impart something useful first.

  55. Scott:

    Correct: instantiation is not analogy.

    (Case in point — the cell’s DNA, RNA, ribosomes, enzymes etc instantiate a re-programmable, digitally stored data, coded, algorithm-implementing information processing system, i.e a computer. [And, what do we know from the observed causes responsible for computers whose origins we directly see?)

    Further to the original post, Berra’s blunder was to try to compare a claimed spontaneous process tracing to chance + necessity only, to a known designed process. He only succeeded in showing that common design can account for apparent evolutionary sequences.

    This has a lot of implications for many claimed evidences and icons of evolution, as we see here how they fail to distinguish themselves from common design.

    Thus also we see the censoring significance of the recently imposed rule of methodological naturalism, which in effect rules out a priori the ability to raise this point, and still be deemed — note the reference to an authority — “scientific.” (For, methodological naturalism changes science from the pursuit of the truth about our world based on evidence, to the best evolutionary materialistic account of our world. Big difference.)

    Thus, much of what has been going on in recent years is plain for those willing to look.

    Sadly

    GEM of TKI

  56. 56

    This is an old issue, but I’ve got the book now. I asked for some context of Berra’s analogy. If he’s talking about naturalistic mechanisms, Johnson has a point. If he’s just talking about common descent, I would argue that Johnson is both wrong and misleading.

    Having read the chapter, it is clear that Berra here is simply talking about descent with modification. He is definitely not talking about a naturalistic mechanism at this point, nor is he making a “claim that there is no need for a creator” (as Johnson puts it). He is simply saying that the Corvette evolves “in the sense of ‘descent with modification’” (page 118, in a passage Johnson does not quote).

    Does Johnson say Corvettes didn’t evolve in that sense? I would say they did.

    Does Johnson say Berra’s analogy doesn’t work to illustrate the mechanism of naturalistic evolution? So what? Berra doesn’t use the analogy for that purpose.

  57. 57

    He is simply saying that the Corvette evolves “in the sense of ‘descent with modification’”

    But no Corvette has ever been known to descend from another Corvette, with the second Corvette having modifications (or not.)
    It works an illustration of what the effects of descent with modification would look like. But to illustrate descent with modification, he should have used something that descended and modified.

  58. 58

    ScottAndrews [57],

    But no Corvette has ever been known to descend from another Corvette, with the second Corvette having modifications (or not.)

    The Corvette people would disagree with you. From a General Motors press release:

    Equipped with a 377 kW LS7 V8 engine and built on a lightweight chassis, the Corvette Z06 is a road going descendant of the championship-winning Corvette racing programme that includes the Le Mans 24-Hour race in its annual schedule. With its racing-inspired powertrain and suspension features, the Corvette Z06 delivers exceptional levels of capability and technology.

    We even know its ancestor:

    The original 1953 Corvette is a direct descendant of the one-of-a-kind prototype LE SABRE, a multi-million dollar research project hailed in 1952 as “The Car of the Future.” Many elements of the LE SABRE were incorporated in the production version Corvette, including the wraparound windshield and subtle tail fins.

    Of course it’s an engineered descent, but Berra never denies that. His point holds. There’s no “blunder,” because he’s not arguing about the mechanism.

  59. 59

    Here’s another quote, this one about the engine:

    The basic Chevrolet small-block V-8 design has remained in continuous production since its debut in 1955, longer than any other mass-produced engine design in the world auto industry, though current versions share few if any parts interchangeable directly with the original. Descendants of the basic small-block OHV V-8 design platform in production today have been much modified with advances such as aluminum block and heads, electronic engine management, and sequential port fuel injection, to name just a few improvements over the 54-model-year design life of the engine concept to date.

    Link

    Of course, the word “design” is here too. But again, Berra is not arguing against design in the chapter where he uses the analogy. He’s arguing about evidence for descent with modification. That’s the point of the analogy, and it holds. There’s no blunder at all.

  60. 60

    gpuccio,

    I would very much appreciate your kind attention to my post number 41 dated 03/30/09, 1:28 PM.

    Yours in collegiality and scientific rigor,

    Adel

  61. 61

    The Corvette people would disagree with you.

    I would disagree with them as well, but I know that they were using the word “descendant” metaphorically.
    Obviously Berra doesn’t mean to say that one Corvette literally descended from another. But he says

    If you compare a 1953 and a 1954 Corvette, side by side, then a 1954 and a 1955 model, and so on, the descent with modification is overwhelmingly obvious. This is what paleontologists do with fossils.

    He writes of “descent” but means two entirely different things in relation to the car and to the fossils. If the two were really comparable, we would have to conclude that either the fossils were manufactured or that the cars reproduced. It’s a verbal shell game.
    The analogy is intended to make Berra’s point. But it raises the question, how does understanding why the Corvettes are similar tell us why the fossils are similar?
    I don’t like making analogies because they are so easy to pick apart. But in this case it’s the underlying premise that’s flawed, not the details.

  62. 62

    ScottAndrews,

    He writes of “descent” but means two entirely different things in relation to the car and to the fossils. If the two were really comparable, we would have to conclude that either the fossils were manufactured or that the cars reproduced.

    I frankly don’t see why you make such a claim. He’s merely saying that, if you have those examples, and if you know the time they were produced, you have enough to infer a relationship of descent. That’s all.

    In fact, I’d go further. Let a person examine twenty Corvettes of different dates over thirty years or so. Only reveal the dates of three cars: one toward the beginning, one toward the end, and one somewhere in the middle. I bet that alone is enough information for an intelligent examiner to put the 17 other cars in approximately correct order.

    That’s the kind of thing Berra is talking about in the chapter. If you read the whole chapter you’ll see that Johnson’s charge is about an argument that Berra does not use the example to make.

  63. Adel (#42 and 59):

    Thank you for mentioning your previous post: indeed I had completely mussed it, in the complex flows of information here at UD…

    So, I will try to answer as much in detail as I can.

    Indeed, detail is exactly what I like in discussions about ID and darwinian theory. Ans so I am happy that you raise so many pertinent and constructive point which could well give rise to a very constructive discussion.

    Let’s begin…

    You say:

    “In any case, how would you have dealt with the shortcomings you perceive in the McLennan review? Please help me understand what kinds of arguments McLennan should have given by providing an example of what you consider to be a more correct approach.”

    Well, that’s not really a very important point, so I will not spend much time on that. My reaction to that paper was probably due more to the context it was presented here, as evidence of a concept which I really dislike, than to its intrinsic merits or demerits. Let’s say that, where it suns up:

    “Molecular analysis indicates
    that ? crystallins are homologous with heat shock proteins”

    it could have given a few more details on the level of homology, and where it says:

    “Phylogenetic and functional analyses indicated that
    ? crystallins originated following two duplication events”

    it could give at least an elementary explanation of why these duplications are inferred, and where it says:

    “In the first event one of the copies of a heat shock
    protein (copy 1) retained its plesiomorphic function (protection
    against damage) and plesiomorphic location (throughout the body), while the other (copy 2) added a change in function (focusing light) to the old function of protection
    from degradation in the plesiomorphic location.”

    It could try to discuss what kind of molecular changes could have given the reported changes, especially the changes in location, which IMO remain probably arbitrary even in a more detailed discussion like we could find in the original papers. And so on.

    But again, the point is not the quality of the review. I think that what I find arbitrary in that review is probably arbitrary in all darwinian literature, and so I used that review just as an example, probably mixing up what was untold just because it was a summary review (which is bad but forgivable) and what was untold because it is always untold (which is bad and unforgivable). I will try to detail better this last concept in the following part of my post.

    You say:

    “I thought you understood that there is no “established truth” in science. Correct me if I’m wrong, but I thought you wrote that very thing on another thread. All scientific inferences are hypotheses, are they not? ”

    I understand that and I absolutely agree. That’s one of my most important convictions about science.

    “So, given the observation of sequence similarity between two genes, how would you go about disproving that they both derived from a common ancestor?”

    That’s a very important question. First of all, I don’t want to disprove that. I do believe that in many cases homology between proteins is strong evidence that they derive from a common ancestor. I have probably expressed badly my opinion about that. I don’t believe that homology, of any level, is evidence that one protein derived from another through darwinian mechanisms. But I do believe in common descent, and even if I have sometimes vague doubts about universal common descent of all living beings (it’s an issue about which I am really curious, and open to all possibilities), I have no doubt that very similar proteins have a common descent. And when two proteins are really similar, and the function is the same, the observed differences in species can really be due only to random variation, and reflect the natural history of those species.

    And yet, I would not accept a general rule that homology always implies direct derivation, and I would never accept that it implies derivation through darwinian mechanisms. First of all, homology can sometimes just imply similar function, even if direct derivation is not true (that is similar to the difference between the concept of common descent and the concept of common design). I do believe that homology can mean both things (similar function or derivation), and probably it is possible in principle to distinguish between the two.

    I stop a moment here, and I will go on in the next post (just a trick to avoid being accused of posts too long).

  64. Adel (continued):

    First of all a correction: your post was #41, and not #42.

    So, to go on, my point is that the confrontation between darwinian theory and ID is about causal explanation, and not about natural history.

    I want to be very clear about that: I don’t personally accept darwinian natural history as a whole: I think that darwinists are very often declaring a gradualism which they badly need for their theory, and which is not supported by facts. But still, this is not the real important point. ID is perfectly compatible with gradualism, even if darwinian theory is certainly not at ease with saltation and “explosions”.

    But the true point is causal explanation. Causal explanation means, as you well know, that if we have a theory about how something happens or happened, that theory must be first of all logically consistent, it must be as quantitative as possible, and finally it must explain known facts (and if possible future facts).

    Well, that’s exactly what darwinian theory does not do. All the supposed “evidences” for darwinian theory are evidences (if and when they are) for a natural history, but they lack any explanatory power of that history. Why? Because the only explanatory mechanisms on which all the theory is built (RV and NS) are completely unable to explain the natural history which is provided by the theory itself, if we go a little into details and if we require a quantitative assessment. In other words, when put to test as explanatory models, darwinian assumptions invariably become just so stories.

    That’s why we in ID insist on probability calculations: it’s not just to confound people with high numbers, but rather because darwinists strictly avoid any convincing quantitative assessment of their own theory. And why should darwinian theory be assessed through probability calculations? it’s very simple. Because that theory (and not ID) is based on random variation as the only engine of generation of new information (with all due respect to Allen MacNeill, who goes on counting engines which are not causal explanations of anything).

    IOW, a theory based on random events “must” count its probabilities, otherwise it’s not a theory at all!

    So, let’s see your two points:

    1) “So, given the observation of sequence similarity between two genes, how would you go about disproving that they both derived from a common ancestor” by a darwinian mechanism? (obviously, the last part is added by me).

    It’s simple: let’s imagine that we have two genes in two different species, and that the two genes have a satisfying level of homology, so that both you and me agree that the second is derived from the first, and we have reason to agree on which is the oldest, and the two proteins coded by the genes have different structure and function, and the functional difference can be shown to be due to a difference in primary sequence of at least 30 aminoacids, all ot them necessarily different so that the new function may substitute the old, and that we can experimentally show that no intermediate protein between the two has any improvement in the originary function, or any detectable level of the new function, so that it could have been selected by NS. Although those assumptions may appear a lot, it is perfectly reasonable that while our understanding of proteins and their function increases (and it is increasing!) we will be able to verify many of those things in real cases. So, let’s go on with our imaginary experiment.

    So, let’s say that we have all the reasons to infer that a coordinated mutation of at least 30 aminoacids is necessary to pass from protein A to protein B. And we know that “natural history” did indeed pass form protein A to protein B, when the second species appeared. and we can determine with enough precision how much time passed from A to B. And the maximum rate of mutations for those species in that period.

    And then, with all that knowledge available with enough detail and precision and credibility, we make our probability calculations and… the probability of the proposed coordinated mutation in that period of time, even taking into account all possible functional variants, is, say, 10^-200. (I am just giving numbers her, please don’t take them too seriously: it’s just to show the method).

    What would be your conclusion? Mine would be that something is wrong in the model.

    Well you can say that probably some of our assumptions is mysteriously wrong. But then we repeat our reasoning for other cases, with fresh new data, always more detailed, always more credible, and the result is the same: very unlikely events (indeed, completely unlikely events) seem to have happened all the time in our perceived “natural history”.

    What would that leave you with? Maybe with magic. Or, more reasonably, with the evidence that a causal theory is wrong. Our assumptions were not wrong. It is our causal model which is wrong.

    And what is wrong? Is NS wrong? No. We know that NS can work, given the right context, we have a few models of that, real models like antibiotic resistance.

    It’s RV which cannot work. Or, at least, it cannot do those things. It’s the “engine” which is wrong.

    Another break.

  65. Adel (continued):

    So, your second point:

    2) “Or, alternatively, how would you construct a “credible causal pathway” leading to the opposite conclusion?”

    Here comes the positive part: we were at a point where, in our well documented natural history, apparently magical events happen all the time. But wait a moment: are they really magical? No, not at all. They are magical only if we insist that their cause was random variation. RV cannot do those things, so if we believe it did we are believing in some kind of magic. And that could be a little uncomfortable for a scientist who has not yet embraced Harry Potter as his Bible.

    But what if we shift to another kind of explanation? We can definitely try necessity: but, alas, notwithstanding our repeated efforts, all our theories based on necessity don’t work. Even worse is the fact that there are theorical reasons why it seems that no theory based on necessity will ever work: it’s not just a question of our ignorance, it’s the form of the problem which seems intractable to theories based on necessity.

    But… before we give up to utter despair, we are reminded of some vague memory: where did we see something like that really happen? Is there some empirical, real, observable context where improbable configurations of information emerge easily, and out of any known context of strict necessity?

    Yes, there is. We have always seen that happen. We have seen that happen while we were talking to our friends or writing a letter to our son. We have seen that happen when we tried our new program in C. We have seen it happen reading our favorite book, or seeing the last movie from our favorite director.

    We have seen that happen when our colleague from the near department of protein engineering has proudly told us of his new success, of his new artificial functional protein.

    And so on. Lots of information, Lots of magic. Everywhere. And what is the cause of all that magic? Conscious intelligent agents. Us.

    And the good thing is that it’s not magic at all: it’s nature. Consciousness and intelligence are part of nature. we are part of nature.

    Yes, that other colleague of ours from the AI department would object, but that’s not a problem: his discipline has been now obsolete for many decade, after repeated failures to prove anything worthwhile, and is kept in the curriculum only for historical reasons… :-)

    And so, everything is bright again. Our despair vanished, we start on our new glorious scientific path. The mystery is solved. The magic has vanished. We have a theory. OK, it’s just an hypothesis, we know that: after all, all scientific knowledge is made of hypotheses. Bu hey, this is a good one, at last! So, let’s enjoy this perfect moment of provisional, but deeply satisfying, scientific attainment: we “know” how all that mysterious information came about. It’s simple. Some conscious intelligent agent designed it! OK, we have now to understand who that agent was, how did he act, and many other details, but who cares? We are on the right path at last.

    In a sense, it was easy. We had just never thought of that. Well, it happens. That’s the mark of genius: thinking of something that nobody has ever thought before.

    Never… But what is this other vague memory which is there, somewhere, in our subconscious?

  66. Adel (continued):

    Well, let’s go back to serious discussion now.

    “Therefore, I ask you now to make the distinction clear by describing the alternative causal mechanisms established or envisioned by design theory. And it would be especially helpful if you would stipulate how one would go about deciding which explanation is the better one.”

    Let’s go back to our proteins A and B, of which we know all that is knowable. And still, it remains a mystery how B appeared out of A. Or, if you prefer, how A became B.

    But we know it happened, in a definite time, and it could not have happened by random variation, and it could not have happened by RV + NS. So, how did it happen? That’s a legitimate scientific question, I suppose.

    Well, the answer can be as follows: the pattern of modification required was wxactly the correct one needed to implement a new function, and the new function was exactly the right one needed in the new species, the right one which, in coordination with 1000 other similar changes, makes the difference between the two species. So, our legitimate hypothesis is that there is a plan, a purpose in all that.

    Is that the end of our scientific reasoning? Absolutely not. We are just making an assumption: intelligent, purposeful information was implemented in the passage from A top B, and in the passage from A’s species to B’s species. But a lot of questions, scientific questions, remain: how did that happen? was it gradual or sudden? which physical laws allow conscious intelligent beings to manipulate material configurations so that they can express intelligent information, without apparently violating known laws of necessity? is that a quantum effect? or some other level of reality we still can’t appreciate? who was the designer? was it one or many? can we infer anything about that from the design itself? can we infer something from other sources? what is the purpose of specific designs? was the implementation of design obtained by direct controlled variation, or by targeted random variation followed by intelligent selection, or by both? or by other strategies?

    And so on. And, as a collateral field of invewstigation: what is consciousness? how does it work? how does it interact with physical realities? and so on, and so on.

    You say:

    “And it would be especially helpful if you would stipulate how one would go about deciding which explanation is the better one.”

    IMO, design at present is not only the best explanation for biological information: it’s the only explanation available. That comes form my deep conviction that the actual alternative explanations (indeed, I know one only which has a minimum of structure: classical darwinism) are obviously false.

    But I can agree that not everyone seems to share my certaonty about that. well, I can live with that. I do believe in science. I am absolutely sure that, as our knowledge of facts increases, it will become increasingly obvious what is the best explanation. Indeed, it will become increasingly obvious that we have only one explanation, and that the others are daily losing any credibility.

    In that sense, while I do believe in science, I am not a fan of the definitions of science. All that stuff form Popper and similar, about falsifiability and predictions and method, is certainly interesting, but in no way a dogma, at least not for me. I am more a fan of Polanyi and Feyerabend: science is too rich a reality to limit it with our definitions.

    But science works: as knowledge about facts goes on, also reasoning about facts goes on. And intuition. And creativity.

    Physics has had its Einsteins, Bohrs, Diracs, Feynmans, and others. Biology is probably expecting its revolution.

  67. Adel (continued):

    You say:

    “I am pleased to agree that it is a goal of science to provide credible causal explanations. Therefore, I ask you to present the alternative credible causal explanation of the origin of alpha crystallins based on design theory or any alternative theory you may prefer. (If my estimation of your character is correct, I am confident that you would not require evolutionary theory to meet a standard that your own theory does not meet.)”

    It’s easy. The design of new species required that a protein with the characteristics of alpha crystallins were available in the eye, The designer used therefore the most easily available protein with a structure which implied the least modifications, duplicated the pertinent gene, resttricted it to the pertinent environment through a specific modification of the code which control the transcriptome in that kind of cells (whatever that code may be), and then operated a more or less gradual modification of that inactive code. When the functional result was ready, the new code was activated, and its regulation finely tuned to all the other changes that the designer had realized in the meantime in other structures. That is a model which makes specific previsions, which in time can be explored. But it is consistent, and it requires only the implementation of new information at specific points of natural history.

    And this is the end, for now. :-)

  68. Adel:

    As the thread is a little bit cold, I just mention here that I have answered your post. I hope you see the message.

  69. 69

    gpuccio,

    Thank you for the bump in your #66, but I can only access the Internet occasionally, hence this delay.

    You are amazing. In reply to my modest requests in my post #41, you have presented me with five posts totalling 3,000 words! And excellent words and sentences they are. Such generosity – you must have known that today is my birthday.

    Since this thread has already been pushed below the fold as you noted, we will have to search harder to find each other, so I will not try to drag things out in this location and I will not make any attempt to critique every one of your 3,000 words. (I don’t have your strength.)

    So I would like to touch on a point or two before I am exhausted, beginning with the following, in which you refer to the deficiencies you perceive in the McLennan review on co-option:

    It could try to discuss what kind of molecular changes could have given the reported changes, especially the changes in location, which IMO remain probably arbitrary even in a more detailed discussion like we could find in the original papers. And so on.

    Before I start citing and quoting from the pertinent literature in an attempted counter, would you clarify your meaning in the above? Do you mean that you question whether science has any evidence about the mechanism of tissue-specific crystalline gene expression?

  70. Adel:

    Yes, I mean that the mechanisms which regulate the specific transcriptomes in different cell types are at present very poorly understood, and while some final detail of transcritpion control may be known, the general procedures which determine the individual transcriptomes in cells are still mysterious. That’s what I mean. I don’t know anything specific about crystallins, and as I have said I would be glad to hear from you what is objectively known.

  71. 71

    gpuucio,

    Thanks, try this reference, which covers much of the issue, although it is a bit dated (Lens crytallins and their genes: diversity and tissue-specific expression, by J. Piatigorsky, 1989):

    http://www.fasebj.org/cgi/reprint/3/8/1933

    (I hope the link works for you.)

    Expression is on pages 1936-1937.

  72. Adel:

    I have read the paper (or at least, the expression part). I would like to make some quick comment, just to introduce further discussion, if you want. Please consider that the specific subject is completely new for me, and so I apologize if I am wrong in some of my impressions.

    First of all, I don’t think that the paper, which is indeed detailed and very clear, says anything very different from what I had hypothesized. The expression section details many facts about the expression of different crystallin proteins in various tissues, and the highly preferential expression of some of them in the lens, and by means of deletion experiments it identifies many different regulation sequences in 5′ DNA, acting both as promoters and enhancers, in different animal species. Well, that’s interesting, but I don’t think that it gives us a clear idea of how a “restriction” is working. It just tells us that some recognition sequences are implied in the process.

    I will be more clear. Take just as an instance the following paragraph:

    “Gel retardation and methylation experiments using oligonucleotides show evidence for selective binding of different chicken lens nuclear proteins to the proximal and distal regions of the mouse aA promoter (29). The proximity of the distal and proximal control elements
    suggests that the putative trans-acting factors physically interact when situated on their cognate sequences. Similar nuclear proteins (or protein complexes) appear
    to be present in chicken erythrocytes and HeLa cells
    (29),supporting the notion that the same or related proteins found in many tissues are necessary but not sufficient for expression of the mouse aA-crystallin gene.”

    That illustrates well my point. Specific promoters or enhancers may well be necessary to realize the selective expression of a protein in a cell type, but they are certainly not sufficient for that. That is obvious form the simple fact that those sequences are the same in all cell types, like the rest of the genome. Therefore, it’s not the recognized sequence in itself which accomplishes the “restriction”. It is only an appropriate gignal which is used when the appropriate recognition apparatus in working. And the recognition apparatus (that is, the right balance of specific transcription factors) is determined by the specific transcriptome of that cell type, which means the control and regulation of mant different complex proteins (at least the final transcription factors). Otherwise, the “simple” DNA recognition sequence is useless.

    That’s more or less what I had generically stated:

    “Yes, I mean that the mechanisms which regulate the specific transcriptomes in different cell types are at present very poorly understood, and while some final detail of transcritpion control may be known, the general procedures which determine the individual transcriptomes in cells are still mysterious.”

    So, a variation in the genome which restricts a gene to a cell type can be determined by a variation in the recognition sequence, provided that a well organized recognition apparatus is working in that cell type.

    So, to have a detailed model of how credible it is that a random variation can restrict the expression of a gene to a specific cell type, and particularly to the cell type where it will work functionally, we should assess at least the following:

    1) What genomic variations in promoter and enhancer sequences are necessary and sufficient to determine that step in a previous genome where the gene is not localized.

    2) What are the quantitative probabilities that such a variation be achieved by random mutation. As you can see in the paper, long sequences of nucleotides are often involved, often more than one sequence is necessary, and distant sequences interact in the process, and the sequences themselves are different in different species. So, if we are building a quantitative evolutionary model, we should restrict our analysis to a specific acquisition of the restriction in one species, and hypothesize which regulating sequences were acquired that restricted the expression in a way which was not present previously, and then calculate how likely that coordinated mutation is, by RV and NS only.

    3) Moreover, we should try to understand what is the specific transcriptome and set of transcription factors which allows the restriction in the lens, and if that transcriptome was already active in the previous genome where the restriction is supposed not to be present. And, if other specific changes were necessary to accomplish the restriction (for instance, the acquisition of a new transcription factor, or just a different regulation of the same in the lens), then calculate also the probabilities that such a necessary, but unrelated event could happen by those mechanisms.

    Now, I understand that point 3 is very difficult to deal with, and that’s why I maintain that it’s hopeless at present to make assumptions about events which are not understood, such as the emergence of specific transcriptomes. But I would be happy, for a start, of a quantitative evaluation of points 1 and two in a specific evolutionary model.

    And that’s only a brief discussion about one single step of the evolutionary model suggested for the crystallins. All the other suggested steps should be quantitatively evaluated in a similar way, before they are suggested as credible by the proposed causal mechanisms.

  73. 73

    gpuccio [72]:

    “Yes, I mean that the mechanisms which regulate the specific transcriptomes in different cell types are at present very poorly understood, and while some final detail of transcritpion control may be known, the general procedures which determine the individual transcriptomes in cells are still mysterious.”

    I’m sorry, I gave you a reference that focused on cis-acting (promoter) elements, when the word “transcriptome” was staring me in the face. But your extended, and patient, further explanation may lead me on a more responsive path.

    First, a couple of clarifying questions:

    1. You speak of “mechanisms which regulate the specific transcriptomes,” “final detail” and “general procedures.” Do you mean that if I cite some papers on what is currently known about lens-specific transcription factors such information would constitute only “final detail” and will not adquately address your concerns?

    2. Please define “poorly understood” and “mysterious.” Do you mean that complete understanding has not yet been attained (which should be no surprise to anyone) or that further understanding (employing naturalistic methods) is impossible? Or do you mean something else?

  74. Adel:

    what I mean is IMO rather simple.

    1) The genome is one in all cells of one organism (with the partial exception of the immune system).

    2) The transctiptome (and therefore proteome) is different in each cell type, and in the same cell type in different situations.

    3) Therefore, IMO, there must be somewhere in the genome the information which guides the implementation of specific transcriptomes in specific cell types and situations.

    4) The idea that everything goes on because of feedbacks from the outer environment is, always IMO, rather silly.

    5) So, as I have many times argued here, at present we understand the effectors in the genome (the protein genes), but we have practically no idea of where the procedures are: IOW, where is the information which tells a lymphocyte what genes should be transcribed to be a lymphocyte, and so on. That is a lot of information, if you consider how many different cell types and states exist, and how many different transcriptomes are possible, even form a pool of “only” 20000 protein genes. And another factor is the relative quantitative expression of each gene, the time sequence, and so on.

    6) Even if we know that the final control network (which however is poorly known too) is the set of transcription factors (operating on specific DNA sequences), still that is not the real answer to the problem. Again, transcription factors may determine the transcriptome, but the problem remains of how the transcription factors are regulated, being those proteins too the same in the genome of all cells. So, we have gone back of one step, but we still don’t know where the information about specific transcriptomes of transcriprion factors is written. And if we don’t know where the information is, it becomes difficult to hypothesize if it could have arisen as a product of RV and NS. I would definitely say no, because regulating information (procedures) is probably even more complex and less “selectable” than effector information (protein sequences). But if we don’t know, we don’t know, and all discussions remain rather vague. That’s why I prefer to restrict the quantitative arguments to protein coding genes, whose information is much better known. But let’s remember that, in the human genome, protein coding genes are only 1.5%.

    7) You ask if complete understanding of all that is possible by naturalistic methods. I don’t see why not, but still that is only an assumption: if we don’t know, we don’t know. And, as I don’t love very much the words “nature”, “natural”, “naturalism” and similar, let’s say that if you ask if complete understanding of all that will be possible by application of what we already know (both in biology and physics), I will answer that it is possible, but IMO not likely. I personally believe that new approaches and paradigms will be necessary to completely understand this crucial point (so dear to our evo-devo colleagues), and that may include even new paradigms in biophysics. But, obviously, that’s only a personal view.

  75. 75

    gpuccio [74]:

    I gather from the above that no amount of information about the regulatory factors in crystallin gene expression will meet your stringent requirements. Nevertheless, I will point you to some relevant literature:

    http://www.ijdb.ehu.es/web/con.....;issue=8-9

    is the contents page of the Vol. 48 Nos. 8/9 (2004) issue of The International Journal of Developmental Biology, which is loaded with free-access papers on Eye Development. So far, I have read only the one by Cvekl et al “Regulation of gene expression by Pax6 in ocular cells: a case of tissue-preferred expression of crystallins in lens.” A quote from the abstract gives an indication of the complexities that had (up to 2004) been uncovered.:

    Here, we review mechanisms of lens-preferred expression of crystallin genes by employing synergism between developmentally regulated DNA-binding transcription factors: Pax6, c-Maf, MafA/L-Maf, MafB, NRL, Sox2, Sox1, RAR?/RXR?, ROR?, Prox1, Six3, ?FBP-B and HSF2. These factors are differentially expressed in lens precursor cells, lens epithelium and primary and secondary lens fibers. They exert their function in combination with ubiquitously expressed factors (e.g. AP-1, CREB, pRb, TFIID and USF) and co-activators/chromatin remodeling proteins (e.g. ASC-2 and BP/p300).

    I believe that the discussion in that paper addresses several of the points in your post above. Personally, I was deeply impressed by how much progress has been made in uncovering tissue-specific regulatory networks in only a couple of decades, and I see no reason why progress should not continue indefinitely using “naturalistic” methods. (Why don’t you like such terms?)

  76. 76

    I trust anyone reading the above understands that Greek letters, which abound in the quoted excerpt, are represented by question marks. I have not taken the trouble to insert the English words that correspond to those letters.

  77. Adel:

    I don’t like the term “nature” because it means different things to different people. We have debated that many times here. I can give you, for instance, the following different definitions of “nature”:

    1) All that exists. According to that defiition, even a God, if He exists, is part of nature. But I would definitely prefer the word “reality” for that meaning

    2) All that exists and has phenomenic manifestations. That’s more subtle, but requires some definite philosophical assumptions. But even noumenic realities have phenomenic manifestations, if they exists, so the problem remains.

    3) All phenomena. But if noumena exists, and cause phenomena, the problem remains.

    4) All that is material. That is very troublesome. What does “material” mean? Is consciusness material? Is consciousness part of nature? Are with, as conscious beings, part of nature? Is energy material? Is dark energy (whatever it may be) material?

    5) All that can be explaine according to present laws of physics. That is at least clear, and probably the meaning many “scientists” give to the word. But you will concede that it is unnecessarily restrictive. How can we be sure that the present laws of physics can explain everything? Is consciousness explained by the present laws of physics? (it isn’t, unless you are a dogmatic fan of strong AI).

    I think that “naturalism” is used at present as a form of “authority censorship” to deny existence to positions which are not “politically correct” according to the current ideology of science. Like ID. I don’t appreciate that usage, so I don’t appreciate the word.

    By the way, thank you for the link. I will do my homework. I do believe that progress will go on indefinitely using all methods available, however they are called,and I have great expectations from the studies about the network of transcription factors. The more we understand, the more it becomes easy to define what we still don’t understand. False knowledge thrives only in ignorance.

  78. 78

    gpuccio,

    Concerning “naturalism,” you are of course entitled to you aversions. I was using the term “naturalistic methods” to encompass empirical science as currently practiced.

    Please suggest an alternative term that you prefer.

  79. Adel:

    Empirical science is enough for me. That means any serious attempt at understanding the nature of reality through objectively sharable methods and approaches.

    Empirical science as it is currently practiced is OK, but empirical methods and their conceptions could change in the future. Why shouldn’t they? We are not certainly at the climax of empirical knowledge or of empirical philosophy.

    Moreover, there is a tendency today to deny the empirical status of important observable realities like consciousness and all its activities. I completely disagree with that. Future empirical science will have to consider consciousness as an empirical fact, and to investigate it appropriately.

    Finally, while empirical investigation is today very satisfying, epistemology, methodology and philosophy of science are IMO in great crisis, and very badly understood and practiced especially by scientists. That is a problem which must be faced, if we want empirical science to be able to meet its most important challenges. Contemporary science is gathering a lot of interesting data, but is often interpreting them in a very biased way. That “is” a problem, and we have to remember that true scientific culture is something which goes far beyond mere technology, and has deeper cognitive implications.

    Please, note that your phrase:

    “I see no reason why progress should not continue indefinitely using “naturalistic” methods.”

    interpreted in the light of your clarification, becomes:

    “I see no reason why progress should not continue indefinitely using the methods of science as currently practiced”.

    IMO, that is not true. Science has never been that way, and never will be. Science changes: in its methods, in its vision, in its perspective, in its philosophy, in its culture, in its morality, even in its errors. There is no reason that science will be, in 200 years, any more similar to current science than current science is to science of 200 years ago. Science is not the final truth, not the final culture, not the final model of everything. And yet it is precious, even more so for its weaknesses and limits. Science idolatry is IMO the greatest enemy of science itself.

  80. 80

    gpuccio [79]:

    Empirical science is enough for me.

    Agreed.

    Now, if you are willing, I would like to go back to the original topic of this thread: co-option.

    You will remember that in your post #18, you asked:

    For instance, how was the “change in location” achieved? By a random mutation in the rehulatory, and vastly unknown, procedures which determine the emergence of specific transcriptomes in different cells?

    In the two references I have given so far, Piatigorsky (1989), and Cvekl et al. (2004), there is information on the cis- and trans- acting factors, respectively, that are responsible for preferential expression of crystallin genes in vertebrate lenses. So I have addressed the “change in location” issue.

    Turning to the evidence for the origination of crystallin genes by duplication and functional modification, I offer this paper by de Jong et al. (1993), “Evolution of the [alpha]-crystallin / small heat-shock protein family:”

    http://mbe.oxfordjournals.org/cgi/reprint/10/1/103

    I should like to call your attention especially to the sequence alignments in Fig. 1, which provide compelling evidence for genetic relationships within this gene “family;” the phylogenetic tree that was deduced from the foregoing data (Fig. 2); Table 1, which compares the known (as of 1993) functional and structural properties of the alpha-crystallins and the small heat-shock proteins; and the statement on page 112:

    The 5’ flanking regions of all analyzed small-hsp genes and of the [alpha]B-crystallin gene contain one or more heat-shock elements. As a consequence, these genes-but not the [alpha]A gene-can be induced by elevated temperatures and various other types of stress…

    (My emphasis)

  81. Adel (#80):

    “Now, if you are willing, I would like to go back to the original topic of this thread: co-option.”

    Happy to do that.

    “In the two references I have given so far, Piatigorsky (1989), and Cvekl et al. (2004), there is information on the cis- and trans- acting factors, respectively, that are responsible for preferential expression of crystallin genes in vertebrate lenses. So I have addressed the “change in location” issue.”

    I don’t think so. You have shown, with satisfactory details, some of the tools acting to (partially) localize crystallin expression to the lens. Now you should give a model of how those tools were acquired by RV and NS form an ancestor where they were not present. For instance, ho dd the cis regulatory elements emerge? And what was the pobability ot their emergence by random mechanisms, before they could be selected? Ans so on.

    What you have done up to now is to show the existence of mechanisms which restrict the expression of those protein (partially)to the eye. Well, we know those mechanisms exist (after all, the expression of those proteins “is” partially restrticted to the eye. But your task was to give an explanatory mechanism, or at least amodel, of how those mechanisms were acquired.

    But let’s go to the more interesting part: the proteins themselves.

    You say:

    “Turning to the evidence for the origination of crystallin genes by duplication and functional modification, I offer this paper by de Jong et al.”

    Well, I have read with great interest that paper (a very good paper, by the way), and have trued to deepen my understanding of the subject by other sources too. I will therefore make some comments, apologizing in advance if I have misunderstood something, and always open to further discussion.

    1) If I understand well, we are discussing alha crystallin, and its relation to the family of small heat shock proteins. Other crystallins are not at stake here.

    2) If I understand well, alpha crystallins “are” heat shock proteins. This is the concept which I find everywhere. They are members of the family, they are even named as heat shock proteins, and their function “is” to work as molecular chaperones, like all other small hsps.

    3) Alpha crystallin is made up of two chains, A and B. While chain A is more (but not absolutely) restricted to the lens, chain B is well expressed also in other tissues (heart, muscle, kidney).

    4) In addition to its function as a chaperone and heat shock protein, “alpha-Crystallin has a structural function in the lens, contributing to the
    extremely high protein concentration, up to 50%, in the lens fiber cells, which warrants
    proper refractive properties and transparency”

    My conclusions:

    a) Where is the cooption?

    b) Alpha crystallin is and has always been a heat shock protein.

    c) It is higly expressed in (but not absolutely restricted to) the lens, because it is necessary there for its function.

    d) Having to contribute to the structural function of the lens, it is finely tuned to give, together with the other proteins in that tissue, the necessary “proper refractive properties and transparency”. This is an additional function, which coopts nothing, but which is added to the original function.

    e) I quote from the conclusions in the paper:

    “The recruitment of a-crystallin as a major lens protein now becomes
    understandable. Not only does its intrinsic structural stability makes it suitable for residing life long, without turnover, in the ens; but, by preventing undesirable protein interactions and restoring unfolded proteins, a-crystallin would contribute to the maintenance of lens transparency and integrity. In fact, the constitutively high level of crystallin in the lens cells would make them permanently stress tolerant.”

    And:

    “Logically, the functioning of a-crystallin as a stress protein in various tissues must have preceded its recruitment as an abundant lens protein.”

    And:

    “Whether the dual function of a-crystallins-as ( 1) stress proteins in and outside the lens and (2) structural lens proteins-imposes
    additional evolutionary constraints is yet uncertain.”

    f) Finally, let’s remember that the common sequence in the family of hsps (the C terminal 100 AAs sequence) is only part of the molecule: the N terminal part is very different in different molecules, and that part could well contribute different specificities and functions, while the C terminal part would be responsible of the hsp chaperone function (probably, more is known today about this: thatpaper is not very recent after all).

    g) So, to sum up: I an the designer, and I am planning a new organ, the eye, and in particular the lens. I have special constraints there: I need many functionl proteins to protect the tissue form light stress. Among thm, I decide to use a type of small heat shock protein. But I also need the lens to be trnspsrent, and to have a specific refractory power. So, I slight adapt one of the existing small hsps, so that it can serve as structural proteins, while preserving its more complex biochemical function as a chaperone.

    That is not cooption, not even guided cooption. It is rather direct utilization of an existing funtion in a different environment, with appropriate adaptation to serve also a different, simpler function (structural).

    And again, nothing in the quote literature shows how that could have happened ny RV + NS.

  82. Adel et al:

    Sorry for the many typos above (old age, I presume). I rewrite the last part here:

    g) So, to sum up: I am the designer, and I am planning a new organ, the eye, and in particular the lens. I have special constraints there: I need many functional proteins to protect the tissue from light stress. Among them, I decide to use a type of small heat shock protein. But I also need the lens to be transparent, and to have a specific refractory power. So, I slightly adapt one of the existing small hsps, so that it can serve as a structural protein, while preserving its more complex biochemical function as a chaperone.

    That is not cooption, not even guided cooption. It is rather direct utilization of an existing function in a different environment, with appropriate adaptation to serve also in a different, simpler function (structural).

    And again, nothing in the quoted literature shows how that could have happened by RV + NS.

  83. 83

    gpuccio [82]:

    I don’t think so. You have shown, with satisfactory details, some of the tools acting to (partially) localize crystallin expression to the lens. Now you should give a model of how those tools were acquired by RV and NS form an ancestor where they were not present. For instance, ho dd the cis regulatory elements emerge? And what was the pobability ot their emergence by random mechanisms, before they could be selected? Ans so on.

    What you have done up to now is to show the existence of mechanisms which restrict the expression of those protein (partially)to the eye. Well, we know those mechanisms exist (after all, the expression of those proteins “is” partially restrticted to the eye. But your task was to give an explanatory mechanism, or at least amodel, of how those mechanisms were acquired.

    The task that I set myself was to provide links to the scientific literature bearing on the issue of co-option with respect to lens crystallins. Literature that you seemed to be unaware of when you derided the idea of co-option back in post #18. So I am satisfied that I have made a good effort to reduce your ignorance of the empirical basis for some examples of the concept of co-option.

    And again, nothing in the quoted literature shows how that could have happened by RV + NS.

    It’s not clear what you find lacking. For example, the sequence alignments in de Jong’s Fig. 1 indicate that many mutational events have occurred between members of the small heat-shock protein family. There is no reason to think that those mutational steps were non-random with respect to fitness.

    Perhaps you’re asking for a mutation-by-mutation account of those changes, and how they affected the function of each organism in whose genome they occurred. I doubt that much relevant data are available yet. We are shocked, shocked (heat-shocked?) to observe that there are gaps in the evidence. If and when gaps are filled by data, I expect that they will be published. Without additional data, further hypothesizing seems unproductive.

    However, you have generously filled in those gaps with your preferred alternative hypothesis (though with rather scanty detail):

    I am the designer, and I am planning a new organ, the eye, and in particular the lens. I have special constraints there: I need many functional proteins to protect the tissue from light stress. Among them, I decide to use a type of small heat shock protein. But I also need the lens to be transparent, and to have a specific refractory power. So, I slightly adapt one of the existing small hsps, so that it can serve as a structural protein, while preserving its more complex biochemical function as a chaperone.

    An excellent and entertaining hypothesis, but you are not the designer (as far as I know – do you claim otherwise?) and there is no evidence for the existence of a designer. Nor do you have a way of testing your hypothesis. Your scenario is fanciful. A fable. A “Just-so Story.”

  84. 84

    gpuccio,

    In your post #63 you presented an argument for the existence of a designer of living beings. Some excerpts:

    Is there some empirical, real, observable context where improbable configurations of information emerge easily, and out of any known context of strict necessity?

    Lots of information, Lots of magic. Everywhere. And what is the cause of all that magic? Conscious intelligent agents. Us.

    OK, it’s just an hypothesis, we know that: after all, all scientific knowledge is made of hypotheses. Bu hey, this is a good one, at last! So, let’s enjoy this perfect moment of provisional, but deeply satisfying, scientific attainment: we “know” how all that mysterious information came about. It’s simple. Some conscious intelligent agent designed it! OK, we have now to understand who that agent was, how did he act, and many other details, but who cares? We are on the right path at last.

    This is an argument of the form:

    Some A have property B
    C is an example of A
    Therefore, C has property B

    The conclusion does not follow. There is a hidden premise: All A have property B

    Your argument:

    Some improbable configurations of information are/were designed and created by intelligent beings.
    Living beings are improbable configurations of information.
    Therefore, living beings are/were designed and created by an intelligent being.

    The conclusion does not follow; there is a hidden premise: All improbable configurations of information are designed and created by intelligent beings.

    In any case, if your conclusion is only an hypothesis, how can it be tested?

  85. Adel (#83):

    Well, now the discussion is becoming hotter! Good.

    The task that I set myself was to provide links to the scientific literature bearing on the issue of co-option with respect to lens crystallins.

    And you have not answered my basic objection that the literature you cited does not bear any support to the concept of cooption, as the main concept seems to be that alpha crystallin is a small heat shock proteins which has retained its main function while acquiring a supplementary structural function, which probably depends on other (non conserved) parts of the molecule. Again, that has nothing to do with cooption.

    Literature that you seemed to be unaware of when you derided the idea of co-option back in post #18.

    True. But I did not deride anything, I “criticized” the idea, and I still do. And I stated that there was not evidence in the original paper in support of the idea. Which was true. Just like now I state that there is no such evidence in the further literature which you graciously let me know, and for that I am really grateful.

    So I am satisfied that I have made a good effort to reduce your ignorance of the empirical basis for some examples of the concept of co-option.

    I am satisfied too, although not for the same reasons. You have certainly reduced my ignorance, and I am happy of that. But I can’t see how you reduce my ignorance “of the empirical basis for some examples of the concept of co-option”. For that empirical basis, I am still waiting.

    It’s not clear what you find lacking. For example, the sequence alignments in de Jong’s Fig. 1 indicate that many mutational events have occurred between members of the small heat-shock protein family. There is no reason to think that those mutational steps were non-random with respect to fitness.

    And so? Who is denying that random mutations happen? They do happen, and they are usually neutral (the really bad ones being usually eliminated by negative selection). I don’t see what you are inferring from that alignment. We have different member of a same family. The function is the same. The function remains the same. There are neutral mutations which do not affect the function. And so?

    What you had to show was how “positive” mutation could have produced the function form something which did not have it. How that 100 aminoacids sequence, in any of its functional implementation. could have been “found” in the search space of 20^100 sequences, starting from… what? A different protein? A differnet function? A coopted function? You choose…

    And again, what has that to do with cooption? Alpha crystallin “is” a small hsp. The function of those 100 aminoacids remains the same, even with some random modifications of the sequence.

    Perhaps you’re asking for a mutation-by-mutation account of those changes, and how they affected the function of each organism in whose genome they occurred.

    What I am asking for is an explanatory model which deserves the name of scientific hypothesis. Something that darwinists never want to give. It needs not have all the details. But at least some detail would be appreciated.

    I doubt that much relevant data are available yet.

    I doubt too.

    We are shocked, shocked (heat-shocked?) to observe that there are gaps in the evidence.

    “You” are shocked. I am rather cool about that. In my view of reality, it’s very normal that there are gaps, huge gaps of evidence in a wrong theory. Indeed, let’s say no evidence at all.

    If and when gaps are filled by data, I expect that they will be published. Without additional data, further hypothesizing seems unproductive.

    I would emphasize the “if”. And we are not hypothesizing, we are only discussing the existing evidence for existing hypotheses. At least, that was my impression.

    However, you have generously filled in those gaps with your preferred alternative hypothesis (though with rather scanty detail)

    My generous nature is one of my weaknesses! :-)
    And scanty details are better than no detail at all.

    An excellent and entertaining hypothesis, but you are not the designer (as far as I know – do you claim otherwise?)

    No, it was obviously a narrative tool. I am sure you had understood that.

    and there is no evidence for the existence of a designer.

    I suppose you mean “no evidence for the existence of a designer of biological information”. I must have missed something. Wasn’t that exactly what we were debating? Are you jumping to conclusions? Is our debate over? In case that was not clear, my point is that biological information itself is evidence of the existence of a designer.

    Nor do you have a way of testing your hypothesis.

    Why not? Great part of my previous discussion was about that. Do you remember the imaginary experiment about observing with certainty that something happened which could not have happened unless it was designed? I don’t remember any comment from you about that. Must I repost it?

    Your scenario is fanciful

    I take that as a compliment.

    A fable.

    I do like fables.

    A “Just-so Story.”

    That is probably less of a compliment!

    Let’s see: my scenario lacks some detail, that’s true. But is it a “just so story”? Has it really no explanatory power? (that is my idea of a “just so story”).

    What did I say? Ah, yes:

    “I am the designer”

    Well, I think we agree that is nopt literally true.

    “and I am planning a new organ, the eye, and in particular the lens.”

    Well, the eye did emerge in natural history, or not? My hypothesis is that it was designed, so it is perfectly natural that I hypothesize that the designer at some point planned it. You know, that’s what “design” means.

    “I have special constraints there: I need many functional proteins to protect the tissue from light stress.”

    I though we agreed on those points. It’s also in your literature.

    “Among them, I decide to use a type of small heat shock protein.”

    Well, a type of small heat shock protein is in the lens. My hypothesis that the designer decided to use it becasue it was useful (and you do agree that it was useful, don’t you?) ia a very logical inference, if we are in a design scenario.

    “But I also need the lens to be transparent, and to have a specific refractory power.”

    I think you should agree on that too.

    “So, I slightly adapt one of the existing small hsps, so that it can serve as a structural protein, while preserving its more complex biochemical function as a chaperone.”

    Here I am just describing your natural history, and pointing out that it in no way is a history of “cooption”, but rather of added function. And I am pointing out a very trivial concept, that adding a new function, which is needed in the context, to an existing fucntion, which is also needed in the same context, is a rather trivial example of how a designed sequence usually appears.

    Just so story? I don’t think so. Just a simple example of how the design hypothesis “has” explanatory power of the observed natural history. A fanciful scenario and a beautiful fable which is also a very good explanation.

  86. Adel (#84):

    Here you really disappoint me. But you have probably not read the many past detailed discussions about that “argument” which I have entertained with others here at UD.

    But it’s late, so just a few tips:

    1) The design argument is not a syllogism, and not a logical deduction. It is an empirical inference, a search for the “best explanation”, like all empirical science. How is it that when darwinists are short of empirical arguments, they invarialble resort to ill inspired logic?

    2) You say:

    Some improbable configurations of information are/were designed and created by intelligent beings.
    Living beings are improbable configurations of information.
    Therefore, living beings are/were designed and created by an intelligent being.

    The conclusion does not follow; there is a hidden premise: All improbable configurations of information are designed and created by intelligent beings.

    Wrong. The right reasoning is as follows:

    a) All improbable configurations of information which we know of, with the only exception of biological information (which, I believe, is the object of our hypotheses)are/were designed by intelligent beings.

    b) Biological information shows the same type of improbable configurations of information, with the same formal properties.

    c) Therefore, and in the absence of any other credible hypothesis, at present our best explanation for biological information is that it was designed by one or more intelligent beings.

    It’s not exactly the same, do you agree? And c) does follow from a) and b): not like a mathemathical theorem follows from its premises (again, that is not a deductive reasoning), but exactly in the same way as all empirical inferences follow from existing data, and are considered good, or best explanations in empirical science.

    So, I have repeated that for the nth time. Who will be the next?

  87. 87

    gpuccio,

    My time is currently limited, and will be for a couple of days, so I will be brief and ask for your patience until I can reply at greater length.

    Your #86: I am delighted at your agreement that the argument for design is not deductive. This is an important clarification for me. But I would remind you that if it is inductive, it must provide a fruitful heuristic for further advances in understanding. It must make testable predictions. It must offer a better paradigm for working scientists than the existing evolutionary theory. (Thank Kuhn for that.)

    In sum, I don’t see that it’s worth anything scientifically. I would be pleased if you will provide empirical evidence (not argument) to the contrary.

    Therefore, and in the absence of any other credible hypothesis, at present our best explanation for biological information is that it was designed by one or more intelligent beings.

    Of course you are not seeing the elephant in the room. I hope to be your ophthalmologist.

  88. Adel:

    Please take all the time you need.

    If your elephant is evidence which makes non design theories of biological information credible, I will be happy to see what you can provide: I have never found anyone up to now who has given me (or others) something like that. So I am waiting for the fruitful heuristic of darwinian theory (or of any other similar theory).

    And obviously ID offers a better paradigm for working scientists. We will discuss that after you explain what is the utility of the darwinian paradigm, and which are its testable predictions. And please, one important disclaimer: evidence for common descent (with which I do agree) will not be considered by me evidence for darwinian mechanism: we are discussing causal explanations here, and not simple natural history.

  89. 89

    gpuccio:

    Just had a moment to catch this:

    And obviously ID offers a better paradigm for working scientists. We will discuss that after you explain what is the utility of the darwinian paradigm, and which are its testable predictions.

    Please don’t play games.

    You have already dismissed modern evolutionary theory as credible (your post #86, etc.), so it is clear that nothing I can say will satisfy you, which places the ball logically in your court.

  90. 90

    Ah, we are lucky. I have another few moments free. So, a brief question:

    I had written:

    Nor do you have a way of testing your hypothesis.

    You responded in #85:

    Why not? Great part of my previous discussion was about that. Do you remember the imaginary experiment about observing with certainty that something happened which could not have happened unless it was designed? I don’t remember any comment from you about that. Must I repost it?

    Yes, please. And while you’re at it, please explain how an imaginary experiment is an empirical test.

  91. Adel:

    What do you mean? I am playing no game. I have given you a lot of arguments. You have given me quotations which did not prove your points. I have commented about them in detail.

    You have promised many time to comment on my comments, but I am still waiting. You have promised to show me an elephant and heal my eyesight. I am still waiting. You have changed arguments many times, passing from defense of unguided cooption to logical errors to generic requests of fruitful heuristic, whatever that means, to sudden repositioning of the ball. So, who is playing games?

    If you are just tired of the discussion, just say it. You have no obligation to go on. I hope you have enjoyed it as much as I have.

    But if you want to go on, please say something, and be my ophthalmologist. I am here, ready to consider what you say. But at present, this looks ever more as a monologue.

  92. 92

    gpuccio [85]

    And you have not answered my basic objection that the literature you cited does not bear any support to the concept of cooption, as the main concept seems to be that alpha crystallin is a small heat shock proteins which has retained its main function while acquiring a supplementary structural function, which probably depends on other (non conserved) parts of the molecule. Again, that has nothing to do with cooption.

    You are disagreeing with the views of the scientists who think that the literature I presented does support the concept of co-option.

    Pardon me if I do not bow to your opinion, especially in light of your unfamiliarity with the literature and your evident determination to deny that the evidence indicates that co-option has occurred. (If you admitted that, you would be admitting to your original mistake in post #18. Horrors!)

    Your rationale above has it backwards. The concept in question is that the main function of the alpha crystallins is to provide a refractive matrix for the vertebrate lens. None of the other members of the small heat-shock protein family have anything to do with lenses. Their job is to protect cells from heat shock.

  93. Adel:

    Just read your #90. The following is from Wikipedia:

    Thought experimentation in science

    Scientists tend to use thought experiments in the form of imaginary, “proxy” experiments which they conduct prior to a real, “physical” experiment (Ernst Mach always argued that these gedankenexperiments were “a necessary precondition for physical experiment”). In these cases, the result of the “proxy” experiment will often be so clear that there will be no need to conduct a physical experiment at all.
    Scientists also use thought experiments when particular physical experiments are impossible to conduct (Carl Gustav Hempel labeled these sorts of experiment “theoretical experiments-in-imagination”), such as Einstein’s thought experiment of chasing a light beam, leading to Special Relativity. This is a unique use of a scientific thought experiment, in that it was never carried out, but led to a successful theory, proven by other empirical means.

    Relation to real experiments

    The relation to real experiments can be quite complex, as can be seen again from an example going back to Albert Einstein. In 1935, with two coworkers, he published a famous paper on a newly-created subject called later the EPR effect (EPR paradox). In this paper, starting from certain philosophical assumptions, on the basis of a rigorous analysis of a certain, complicated, but in the meantime realizable model, he came to the conclusion that quantum mechanics should be rejected as “incomplete”. There were of course influential people, e.g. Niels Bohr, who refuted Einstein’s analysis immediately. However it took decades, until the error in Einstein’s paper was located by a real experiment. This means that finally, after decades, there was a decision. The outcome: there was no error, neither in Einstein’s analysis nor in his model, but the above-mentioned philosophical starting assumptions were falsified (e.g. by the optical real experiments of Alain Aspect). This was only possible after the Bell inequalities had been published in 1964, another rigorous theoretical paper.

    Very simply, my imaginary experiment was intended go show a way in which an empirical test “can” be achieved, as our knowledge of the facts improves.

    I would just the same be curious to know how darwinists have tested their theory that RV + NS can generate biological information. But perhaps only you decide where the ball goes…

  94. 94

    gpuccio,

    We seem to be posting at the same time, so I missed your #91.

    I’m sorry to disappoint you, but I disagree that I have not responded to your arguments. If that were the case, why have there been so many replies from you?

    If it seems to be a monologue, maybe it’s because you say so much?

    I’ll give it a rest for now and let you catch up.

  95. Adel:

    Who has decided that the main function of the alpha crystallins is to provide a refractive matrix? Where is that written? Where is written that the chaperone fucntion has become secondary? Or has been lost?

    Eye. 1999 Jun;13 ( Pt 3b):403-8.Links
    Lens alpha-crystallin: function and structure.

    Horwitz J, Bova MP, Ding LL, Haley DA, Stewart PL.
    Jules Stein Eye Institute, UCLA School of Medicine 90095, USA. [email protected]
    alpha-Crystallin is a major lens protein, comprising up to 40% of total lens proteins, where its structural function is to assist in maintaining the proper refractive index in the lens. In addition to its structural role, it has been shown to function in a chaperone-like manner. The chaperone-like function of alpha-crystallin will help prevent the formation of large light-scattering aggregates and possibly cataract. In the lens, alpha-crystallin is a polydisperse molecule consisting of a 3:1 ratio of alpha A to alpha B subunits. In this study, we expressed recombinant alpha A- and alpha B-crystallin in E. coli and compared the polydispersity, structure and aggregation state between each other and native bovine lens alpha-crystallin. Using gel permeation chromatography to assay for polydispersity, we found native alpha-crystallin to be significantly more polydisperse than either recombinant alpha A- or alpha B-crystallin, with alpha B-crystallin having the most homogeneous structure of the three. Reconstructed images of alpha B-crystallin obtained with cryo-electron microscopy support the concept that alpha B-crystallin is an extremely dynamic molecule and demonstrated that it has a hollow interior. Interestingly, we present evidence that native alpha-crystallin is significantly more thermally stable than either alpha A- or alpha B-crystallin alone. In fact, our experiments suggest that a 3:1 ratio of alpha A to alpha B subunit composition in an alpha-crystallin molecule is optimal in terms of thermal stability. This fascinating result explains the stoichiometric ratios of alpha A- and alpha B-crystallin subunits in the mammalian lens.

  96. 96

    Very simply, my imaginary experiment was intended go show a way in which an empirical test “can” be achieved, as our knowledge of the facts improves.

    Fine. Please post it again.

    And now, good night.

  97. Adel:

    You say:

    “Pardon me if I do not bow to your opinion,”

    Maybe you could just discuss my opinion with arguments. Who asked that you bow?

    “You are disagreeing with the views of the scientists who think that the literature I presented does support the concept of co-option.”

    and:

    “especially in light of your unfamiliarity with the literature and your evident determination to deny that the evidence indicates that co-option has occurred. (If you admitted that, you would be admitting to your original mistake in post #18. Horrors!)”

    Yes, I am disagreeing. I thought that was still allowed. And maybe if you admitted that you are only using arguments from authority, you would be admitting that you are short of other arguments.

    And I beg your pardon, but is your only purpose in this long discussion to have me admit that I have made a mistake in post #18? A mistake I don’t think I have made?

    Strange, I may be naive, but I thought we were discussing in order to have a sincere, passionate and constructive intellectual confrontation about our scientific views. But you see, sometimes I do make mistakes…

  98. Adel (#96):

    Please, read it again at #64.

    And now, good night.

  99. 99

    gpuccio,

    At last I have some free time. Thank you for your patience and your reference to your #64. I see that your #93, concerning “thought experiments” intervened. Let’s look at that first:

    Very simply, my imaginary experiment was intended go show a way in which an empirical test “can” be achieved, as our knowledge of the facts improves.

    Maybe you showed a way, but I don’t see it. Could you point it out?

    I would just the same be curious to know how darwinists have tested their theory that RV + NS can generate biological information.

    As I understand it, random mutation changes the genetic information content of organisms. Take the crystallin example out of many. Changes in the regulatory regions and the protein-coding region occurred. The evidence also indicates that a gene-duplication event (which is an increase in genetic information) was involved. All empirical. Anybody can check it out. If this kind of evidence is not an adequate test in your mind, please explain.

    Turning to #64,

    So, let’s go on with our imaginary experiment.

    So, let’s say that we have all the reasons to infer that a coordinated mutation of at least 30 aminoacids is necessary to pass from protein A to protein B. And we know that “natural history” did indeed pass form protein A to protein B, when the second species appeared. and we can determine with enough precision how much time passed from A to B. And the maximum rate of mutations for those species in that period.

    And then, with all that knowledge available with enough detail and precision and credibility, we make our probability calculations and… the probability of the proposed coordinated mutation in that period of time, even taking into account all possible functional variants, is, say, 10^-200. (I am just giving numbers her, please don’t take them too seriously: it’s just to show the method).

    I have bolded the critical assumption here. It looks arbitrary to me. The only justification I can see is that it allows you to perform an elementary probability calculation: the number of ways 20 amino acids can be combined randomly, given 30 places, is 0.05 to the 30th power. The number I obtain is 9.3^-40, a very large number indeed, and adequate for your argument, but smaller than your number. (I am asking only for clarification: I wonder whether I have missed something that you included in your calculation.)

    But why assume that all of those changes must have occurred simultaneously? Why could they not have occured incrementally, as Darwin proposed?

    What would be your conclusion? Mine would be that something is wrong in the model.

    My conclusion is that the underlying assumption of the thought experiment is wrong.

  100. 100

    Correction:

    I said above,

    the number of ways 20 amino acids can be combined randomly, given 30 places, is 0.05 to the 30th power.

    I should have said “the probablity of a single outcome of 20 amino acids combining randomly.”

  101. Adel:

    I am happy we are discussing again. So, here are the answers to yopur questions, which point to fundamental issues.

    I will leave alone for the moment the thought experiment because it probably requires a longer discussion. So I will start with the other points.

    As I understand it, random mutation changes the genetic information content of organisms. Take the crystallin example out of many. Changes in the regulatory regions and the protein-coding region occurred. The evidence also indicates that a gene-duplication event (which is an increase in genetic information) was involved. All empirical. Anybody can check it out. If this kind of evidence is not an adequate test in your mind, please explain.

    And I explain. Again. Because, you see, the problem is always the same: you pass natural history for causal explanation. It’s not you fault, after all. You just repeat the fundamental error of all darwinian theory.

    I will try to be more clear. If alpha crystallin is in the beginning only a small hsp expressed in all the organisms, and them in a more recent species it becomes (partially) confined to the lens, where it acquires specific characteristics which allow it to be also a structural protein of the lens and contribute to its refractory properties, while retaining its original function, and if we assume common descent (as I do), then it is obvious and trivial that, as you say: “Changes in the regulatory regions and the protein-coding region occurred.”, passing from one species to another (from one without the lens to one with the lens, to be more clear). We agree on that.

    But why are you assuming that those changes can be explained by random mutations? That’s exactly the controversial point. That’s where you, and all darwinists, have no evidence at all.

    IOW, your evidence is just evidence that certain things happened. Not of why they happened. Is that clear?

    To be even more clear, I am not denying that random mutations happen. They do happen. So, my point is: random mutations happen, and certain functional changes do happen in natural history, but the functional changes are not explained by random mutations, and not even by any form of RV + NS. They can be explained only if we assume design, and therefore some source of guiding intelligent information.

    Is that clear? I will go into more detail, just to exemplify. Let’s consider our beloved alpha crystallin. I argued (as a hypothesis, but the point could be certainly an object of inquiry) that the chaperone function can be attributed to the C terminal sequence of about 100 AAs. That’s reasonable, because that is the sequence responsible for the homology with other shsps, IOW the conserved sequence. Well, that sequence is homologous in the family of proteins, but it’s not the same. There are differences. As the function is conserved, I hypothesize that most of those differences are due to neutral mutations (some of them, obviously, could have some functional meaning). So, what can we say about that? We can say that the chaperone function has been expressed in that family form the beginning and that, nothwistanding many neutral mutations, and assuming the elimination of negative mutations by NS, the function is conserve up to alpha crystallin, where it is well present and expressed.

    Let’s go to the new, structural function. What evidemce have you that it is expalined by the same sequence and structure which gives the chaperone function? Let’s remember that the N terminal sequence varies abundantly in the family, and that quaternary structure is also important to determine the physical properties of a protein. IOW, the physical and refractory properties of alpha crystallin in the lens could well be explained by adjustments in both primary and quaternary structure of the protein which have been realized explicitly for its local utilization in the lens.

    So, while the neutral mutations in the C terminal sequence contributing nothing to the existing function (if not in not reducing or cancelling it), the supposed changes in primary sequence and in regulatory regions which did create the new function of transparency and refractory power had to be specifically restricted to a target space. So, to show that they could happen as a result of random mutations and NS, you have to make an explicit model of at least some of those changes, of their complexity, and of the probabilistic resources necessary to achieve that result randomly. So, for example, if you assume that a specific regulatory region of, say 100 nucleotides was created by random mutation, and that that change is responsible for restriction to the lens, even if we admit that isolated restriction to the lens of alpha crystallin is sufficient to be a selectable step (which personally I don’t believe), still you should show that the assume modification of the genome could have taken palce by random events. IOW, shoe the starting genome of species A (where the protein is not restricted), show which changes in the genome in species B are responsible for the new function (restriction to the lens), and then let’s compute the probabilistic resources, the available time and population size and reproductive time and mutation rate, and let’s see if the observed result is even distantly compatible with our random model. That’s what I call an explanatory model.

    Let’s go to the other very important point. The coordinated mutations.

    I have bolded the critical assumption here. It looks arbitrary to me. The only justification I can see is that it allows you to perform an elementary probability calculation: the number of ways 20 amino acids can be combined randomly, given 30 places, is 0.05 to the 30th power. The number I obtain is 9.3^-40, a very large number indeed, and adequate for your argument, but smaller than your number. (I am asking only for clarification: I wonder whether I have missed something that you included in your calculation.)

    There are two different answers due here.

    The first is easier. Your calculations are right, I believe. But if you look at my original post, you will find, in the same part you quote, the following phrase:

    “(I am just giving numbers her, please don’t take them too seriously: it’s just to show the method).”

    Indeed, I was not trying to make any real calculation: I was just showing how a calculation can be made. I apologize if that was not clear enough. Your calculation is right and, as you say, adequate for my argument.

    But let’s consider now the problem of coordinated mutations. You say:

    But why assume that all of those changes must have occurred simultaneously? Why could they not have occured incrementally, as Darwin proposed?

    I use the term “coordinated” in a very specific sense (which is generally the sense all IDists use it), but that sense is probably not so intuitive, so I will explain it in detail.

    “Coordinated mutations” does not mean that “all of those changes must have occurred simultaneously”. The changes may well occur at different times. The meaning is that all those changes must be present at the same time in a single individual of the species to bear functional change.

    So, let’s go to my hypothetical model. We have protein A and protein B. Protein B expresses a new function (or some specific change in function) which we assume to be the selectable trait. Studying the two proteins, we observe that at least 30 AAs must change so that, starting from A, you can get the function in B. Please note that neutral mutations are not included in that computation. The 30 changes are all necessary, and must all be present at the same time, otherwise the B function is not observable. OK?

    Now, the point is: you camnot have NS occurring on the intermediates. And the 30 changes must all be present at the same time in one individual.

    To achieve that, it is not necessary that all the changes occur in the same individual. They can occur gradually. But the changes remain confined in the clone deriving form one individual. And, if selection does not happen, and the population is not expanding, the clone deriving form that individual will represent always approximately the same percentage of the total population. And even if some of those assumptions are not true, there is no reason that the clone with one, or two, of the required changes must expand in any significant way. Indeed, it could well become extinct.

    So, we are left with 30 successive changes which have to happen gradually practically in one individual clone (even if through many generations of that clone), and while the new changes accumulate any of the others already acquired can be lost. IOW, at any given moment any of the possible combinations of those 30 AAs can be present in the single clone, and any successive change is a random change, which can retain or lose any present aminoacid. IOW, if you have followed another busy thread here, there is no latching, neither explicit nor implicit, becasue there is no selection. Here, pure randomness reigns.

    Therefore, the probability of having the 30 aminoacids at the same time in one individual remains extremely low, approximately of the same order of having them at the same time in one generation (obviously, you can compute the probabilistic resources provided by having many generations and longer time: that is part of the computation).

    So, I mean with “coordinated mutations” a set of mutations which confers a functional advantage, presumably selectable, which cannot be achieved by a simpler mutation set, and for whom NS cannot be invoked as an intermediate causal factor.

    My conclusion is that the underlying assumption of the thought experiment is wrong.

    I believe the opposite. I don’t know if you will accept my arguments, but I have tried. And I am always ready to discuss.

  102. 102

    I am happy we are discussing again.

    I, also. And I apologize for my recent testiness.

    And I explain. Again. Because, you see, the problem is always the same: you pass natural history for causal explanation. It’s not you fault, after all. You just repeat the fundamental error of all darwinian theory.

    Your patience is exemplary, and I wish I could be as patient. Yes, the problem is always the same, because my error is the same as the error you perceive in evolutionary theory. I am grateful that you let me off the hook by placing me in the august company of those theorists.

    But it puzzles me that, despite that putative fundamental error, evolutionary theory should have been so successful as a participant in the spectacular advance of the biological sciences during the past century. You may remember Kuhn’s argument that a scientific paradigm is successful as long as it provides a framework for progress in a scientific discipline. That has been the case for evolutionary theory. We can can expect, if the future follows the patterns of the past, that evolutionary theory will be replaced by a more fruitful theory when it no longer meets the needs of the laborers in the field. That time has not arrived, as best I can tell.

    But why are you assuming that those changes can be explained by random mutations? That’s exactly the controversial point. That’s where you, and all darwinists, have no evidence at all.

    IOW, your evidence is just evidence that certain things happened. Not of why they happened. Is that clear?

    No, it’s not clear, because what you mean by the term ‘explanation’ is not clear. You go on to cite the crystallins.

    Let’s go to the new, structural function. What evidemce have you that it is expalined by the same sequence and structure which gives the chaperone function?

    I don’t have any evidence regarding that question, but I don’t understand why it should be an issue. Why must the same sequence and structure explain both properties? Protein molecules have domains, regions of structure that may perform different roles. One domain may provide the transparency function and another domain may provide the chaperone function. Or there may be overlap between domains and functions. I suspect that the literature may have information on such points, but that is not necessary to question the logic of your objection. And, please remember, if we don’t know everything it doesn’t mean we know nothing,

    So, to show that they could happen as a result of random mutations and NS, you have to make an explicit model of at least some of those changes, of their complexity, and of the probabilistic resources necessary to achieve that result randomly.

    Not at all. Neither I nor anyone working in the field has to meet those requirements. What you disparage as ‘natural history’ has apparently been enough for empirical scientific purposes in biology and will remain so until workers in the field encounter a need for potentially more useful explanations.

  103. 103

    Continuing,

    But let’s consider now the problem of coordinated mutations. You say:

    “But why assume that all of those changes must have occurred simultaneously? Why could they not have occured incrementally, as Darwin proposed?”

    I use the term “coordinated” in a very specific sense (which is generally the sense all IDists use it), but that sense is probably not so intuitive, so I will explain it in detail.

    “Coordinated mutations” does not mean that “all of those changes must have occurred simultaneously”. The changes may well occur at different times. The meaning is that all those changes must be present at the same time in a single individual of the species to bear functional change.

    (My emphasis)

    The 30 changes are all necessary, and must all be present at the same time, otherwise the B function is not observable. OK?

    Now, the point is: you camnot have NS occurring on the intermediates. And the 30 changes must all be present at the same time in one individual.

    Not OK, because it’s not clear why intermediate levels of B function could not have existed and have provided selective advantages short of the current level of function.

    To achieve that, it is not necessary that all the changes occur in the same individual. They can occur gradually. But the changes remain confined in the clone deriving form one individual.

    Since populations of the eukaryotic organisms in question are most definitely not clonal, I don’t follow you.

    So, I mean with “coordinated mutations” a set of mutations which confers a functional advantage, presumably selectable, which cannot be achieved by a simpler mutation set, and for whom NS cannot be invoked as an intermediate causal factor.

    But your argument founders on the issue of clonality.

    I don’t know if you will accept my arguments, but I have tried. And I am always ready to discuss.

    I have given my objections, which you may rebut.

    I value your willingness to continue this discussion, which I find challenging and interesting.

  104. Adel:

    Thank you for your answers and for the constructive tone of the discussion. I will give you my counter arguments, but please consider that, if we find that we are repeating ourselves, we can stop any moment with full reciprocal respect. But as long as new interesting points come out, I am willing to go on.

    But it puzzles me that, despite that putative fundamental error, evolutionary theory should have been so successful as a participant in the spectacular advance of the biological sciences during the past century.

    Well, I will not spend much on this point. It is the usual “majority” argument. I have already stated that I am a minority guy. But I have to believe that my minority is on the right side, and in the case of ID I definitely do.

    Poper, Kuhn, Feyerabend and others have really enriched our culture in philosophy of science, a subject too often overlooked and forgotten. But none of them has given a final standard for reality (and probably none of them would have wanted to do that). I can agree with you that evolutionary biology has been the majority option for a long time. That would be a variant form of the “majority argument”, let’s call it the “long lasting majority argument”. But still I don’t buy it. The long status of evolutionary biology may help convince you that the theory is good. My only conclusion from that is that it is a shame for modern culture and science. And an anomaly, certainly. Not the only one, I would say. Strong AI is a good companion. IMO, the only evidence these two theories give us id the evidence of how ideology and dogma have deformed and permeated the scientific thought in the last decades.

    Finally, you think that the great advancements in biology are due to evolutionary theory. That’s a point on which we have to disagree. While I recognize with enthusiasm the great advancements of modern biology, and especially of molecular biology, I can’t see how they are the product of darwinist theory (please remember that I recognize the value of the theory of common descent). The great advancements in biology, the discovery of the genetic code, of the role of DNA, of the mechanisms of transcription and translation, the studies about protein structure and function, the sequencing of genomes, the reevaluation of non coding DNA, the interest for epigenetics, and so on, and all the related technologies which have been developed, are a demonstration of the work and ingenuity of a lot of good scientists, but IMO they are in no way dependent on darwinian theory, although they have certainly been developed in the “light” (or shadow, according to points of view) of that theory.

    But let’s go to more specific (and interesting) points. You say:

    No, it’s not clear, because what you mean by the term ‘explanation’ is not clear.

    in reference to my point about the difference between natural history (onserving that something has happened) and an explanatory model (building a causal model which explain why that thing happened, according to known laws and to a cause and effect model).

    I really don’t understand what is not clear in my use of the word “explanation”. I use it in a very trivial sense, the same sense it is always used in science. No deep philosophical assumption here.

    And it is not my invention that darwinian theory pretends to be an explanatory theory. It does. It is. A bad explanatory theory, IMO, one which does not work, but an explanatory theory just the same.

    Why? Because darwinian theory is not limited to the idea of “evolution” (common descent). It tries to “explain” why evolution happened. And the explanation is simple: ramdom variation (be it mutation, duplication, or anything else) and NS. That “is” an explanation. A model. And therefore, we can try to asses if it works or not. My point is simply that it does not work. First of all, the RV part and the NS part are never detailed. When does RV act, and at what point does NS act? The lack of any detail (even vague) makes it impossible to evaluate the two mechanisms quantitatively.

    a) RV must be evaluated quantitatively in relation to computable probabilities. Why? Because it is a mechanisms which invokes random events, and random events are governed by probability, and analyzed by probability theories.

    b) NS must be evaluated quantitatively in relation to the function which is supposed to be selected. Does that function exist? Is it capable of giving a reproductive advantage, and therefore to be selected?

    So, I mean with “explanation” exactly the same thing as darwinian theory does. The only problem is that darwinian theory gives an explanation which is easily falsified as soon as you try to detail and analyze it.

    Why must the same sequence and structure explain both properties?

    It must not. Indeed, I believe that it doesn’t, as I have tried to explain. My point is that if different sequences and structures are responsible for the two functions, and if only the C terminal sequence is homologous to the other proteins in the family, and if that sequence is responsible only of the chaperone function, which is perfectly conserved, and other parts of the molecule (which are not homologous to the family of hsp) are responsible for the structural function, well then I don’t see any cooption. Cooption would be present if the same structure which was responsible for a previous function were “coopted” for a new function, independent from the previous one. That is very simply my point.

    What you disparage as ‘natural history’ has apparently been enough for empirical scientific purposes in biology and will remain so

    I don’t disparage natural history. It is an important part of knowledge. And it has certainly been useful for empirical research. But it does nopt explain why things happen, without an explanatory theory.

    Darwinian theory “is” an explicatory theory. And a very bad one. That’s what I disparage. Natural history has all my respect, provided that it remains what it is: natural history.

    Not OK, because it’s not clear why intermediate levels of B function could not have existed and have provided selective advantages short of the current level of function.

    I am not saying that intermediate levels cannot exist. I am saying that I define coordinated mutations thgose which are necessary to pass from one level to another. You just show where the selectable trait appears, and I will do the computation form A to the new B. Maube it’s 30 coordinated mutations. Maybe it’s 15. Maybe it’s 100. We have to see for each case. Unfortunately, darwinisn has not provided a single case, up to now (always excluding microevolution, of 1 – 2 coordinated mutations).

    The alternative is that any transformation in proteins and genomes is deconstructable in steps of 1 – 2 mutations, all of them selectable for reproductive advantage. Do you really believe that? And if you believe that, have you any evidence of that?

    Since populations of the eukaryotic organisms in question are most definitely not clonal, I don’t follow you.

    The issue of clonality is not important. I was thinking in terms of a bacterial model. In ither models, the result is the same: the mutation which is not selected remain more or less confines to one allele and its propagation in the descendants, without any expansion. You may say that sexual reproduction allows for genetic drift. True, but genetic drift is a random mechanism. It selects nothing. A “useful” mutation has exactly the same probabilities to be “expanded” as any other mutation. So, genetic drift does not explain anything.

    But your argument founders on the issue of clonality.

    No, as I have tried to show. My argument founders on the lack of expansion of the single mutations. The coordinated mutations have therefore to accumulate in the same isolated gene, wherever it is.

    I value your willingness to continue this discussion, which I find challenging and interesting.

    Me too!

  105. 105

    I will give you my counter arguments, but please consider that, if we find that we are repeating ourselves, we can stop any moment with full reciprocal respect. But as long as new interesting points come out, I am willing to go on.

    Yes, repeating oneself is a pitfall of most discussions I have witnessed, even before the advent of the Internet! So, we should both feel free to call a halt at any time. Until then, I will try to keep my discourse interesting for you. And you are the judge of that.

    Well, I will not spend much on this point. It is the usual “majority” argument. I have already stated that I am a minority guy. But I have to believe that my minority is on the right side, and in the case of ID I definitely do.

    I was not saying that we must accept evolutionary explanations in biology because it is a majority view. My point was that it is the majority view because it has been fruitful.

    And more power to you for being a minority guy. All views are welcome in the marketplace of ideas.

    Finally, you think that the great advancements in biology are due to evolutionary theory.

    That was not my point and I don’t recall making it. I said it was a ‘participant’ in those advancements. At the very least, it has done no harm. But more than that, I don’t think the other contributions you mentioned would have the same value or would have been as strongly motivated in the absence of Darwinian hypotheses. (As Dobzhansky put it so nicely.)

    I really don’t understand what is not clear in my use of the word “explanation”. I use it in a very trivial sense, the same sense it is always used in science. No deep philosophical assumption here.

    And it is not my invention that darwinian theory pretends to be an explanatory theory. It does. It is. A bad explanatory theory, IMO, one which does not work, but an explanatory theory just the same.

    Why? Because darwinian theory is not limited to the idea of “evolution” (common descent). It tries to “explain” why evolution happened. And the explanation is simple: ramdom variation (be it mutation, duplication, or anything else) and NS. That “is” an explanation. A model. And therefore, we can try to asses if it works or not. My point is simply that it does not work. First of all, the RV part and the NS part are never detailed. When does RV act, and at what point does NS act? The lack of any detail (even vague) makes it impossible to evaluate the two mechanisms quantitatively.

    How can I respond without repeating myself? Perhaps I can’t. So I will reiterate that the level of detail you are asking for may not be available, but its absence does not currently seem to be a hindrance to progress in the field. You are not satisfied, I think, because your agenda (“My only conclusion from that is that it is a shame for modern culture and science”) is not the same as that of the workers in the field. This gets back to ‘majority views.’ The available levels of detail are evidently adquate to motivate continuing work among the members of that majority.

    Cooption would be present if the same structure which was responsible for a previous function were “coopted” for a new function, independent from the previous one. That is very simply my point.

    However, the idea of co-option is not restricted to a single structure. The term has been applied more broadly, as expressed by True and Carroll (2002) in the opening sentences of their abstract:

    “Co-option occurs when natural selection finds new uses for existing traits, including genes, organs, and other body structures. Genes can be co-opted to generate developmental and physiological novelties by changing their patterns of regulation, by changing the functions of the proteins they encode, or both.”

    gpuccio:

    The alternative is that any transformation in proteins and genomes is deconstructable in steps of 1 – 2 mutations, all of them selectable for reproductive advantage. Do you really believe that? And if you believe that, have you any evidence of that?

    Not every change need be advantageous. Not every change need be a point mutation. Some neutral mutations might become advantageous when recombined with other neutral mutations. The function of a gene will also depend on the context in which it is expressed in the mileu of other genes that are expressed. It may be very complicated…

    In ither models, the result is the same: the mutation which is not selected remain more or less confines to one allele and its propagation in the descendants, without any expansion.

    But we are dealing with populations, so all kinds of genetic experiments can take place, especially by recombination in sexually reproducing species.

    Enjoy your Easter/Passover/Vernal Equinox holiday!

  106. Adel:

    Happy Easter (or else) to you too!

    I must say that I find your last post very reasonable, and I don’t think I have much to add. Our positions are clear and well expressed, I hope. But if I realize of something which is worthwhile to add, I’ll come back after Easter. Otherwise, I am sure we will soon meet on some other thread. Again, a very happy Easter for you.

  107. 107

    gpuccio,

    I can see that you have had enough discussion with me for now, and though I have more to say, I am content to rest my humble case as well.

    I am most grateful to you for pushing my brain so hard it seemed at times to threaten explosion.

    But it’s still intact, and that is how we learn.

    Ciao, hasta la vista, bis später,

    Adel

Leave a Reply