Cleverly drawing comparisons more or less in agreement with the creationist and/or intelligent design definition of complexity and what appears to be designed, Dawkins sets the stage for natural selection as not an alternative but the answer to the illusion of design. Yet, though he boasts, he provides no evidence or supporting arguments for his claims. In chapter two entitled Good Design, he acquiesces with the eighteen century theologian Willaim Paley and is impressed by the complexity of the sonar system bats use to coordinate themselves in space (and their respective environment). But he merely claims that Paley was wrong with no sound argument or conclusive evidence to support the modern day hypothesis that the job was done in gradual evolutionary stages by natural selection.
Granted that soft tissue (like brain) does not fossilize, we are left with guess work. We must look at various existing species of bats and using their ancestral relationships determine which form of sonar system is preceeded the other. Granted that he cannot show the transitional stages of the sonar system, he still failed to provide even a minimal basis for his confident claims. He only asserts that natural selection is the correct mechanism with no reasoning or supporting evidence by which natural selection could accomplish such an “illusion of design.” Nor does he build a strong case for natural selection in chapter three. The computer program he uses to illustrate natural selection works with a pre-determined goal oriented process - the exact opposite of natural selection!– though to his credit he acknowledges this flaw.
It seems to me that one could never expect a computer program to demonstrate the power of small changes acquired by mutation. Mutations in a grandoise sense, serves as a raw material for natural selection to work on. Considering that each mutation (roughly defined as a change in the genetic code of the individual) is unique and varies in the impact that it makes on the individual’s phenotype (it may be silent, deterimental, or a hox gene mutation with interesting results), it appears to be a mathematical impossibility to calculate the probabilities of each mutation’s effect on phenotype. To go from one stage to the next in the evolutionary process, you must recast the dice and calculate the probability of acquiring a new trait which increases your fitness value.
In the end, we are left with a book that does not satisfactorily makes an argument for natural selection, and with the impression of an “intellectually fulfilled atheist” author who’s arguments fall much below par.
Check out Avida as a MUCH better life simulator.
http://www.carlzimmer.com/articles/2005/articles_2005_Avida.html
Also, please note from “what good is half an eye”
“The shortest equals program Ofria could write is 19 lines long. The chances that random mutations alone could produce it are about one in a thousand trillion trillion.”
That’s 1*10^27, not even the lower limit of *the somewhat arbitrary* UPB, but a large number for a ‘blind process’ non the less.
Comment by Rich — June 30, 2006 @ 10:19 pm
For those of us following along on the web, what is the next book?
Comment by Jonathan Bartlett — June 30, 2006 @ 11:59 pm
We will have one more evening of discussion of Dawkins’ The Blind Watchmaker next Tuesday, and then move on to Michael Behe’s Darwin’s Black Box. Those watching us through the blog window might find it interesting to know that a participant in the seminar is volunteering to lead the discussion of each chapter in each book, and that this will mean that ID folk will sometimes lead the discussion of evolution books and vice-versa, along with the converse. We think this makes for a more lively discussion and tends to facilitate comprehension by the chapter leaders. It’s worked so far, anyway…
The full lineup of texts and the dates on which we will cover them (along with the readings and suggested references for final research papers) can be found under the syllabus and readings links in the menu bar to the right —>
Comment by Allen MacNeill — July 1, 2006 @ 12:14 am
Perhaps I should make it clear that each participant is volunteering to lead the discussion of the chapters of the texts we are using. This means that each participant will lead the discussion of several chapters from several books, sometimes by an author in the evolution “camp” and sometimes one in the ID “camp”. Hope that clears it up…
Comment by Allen MacNeill — July 1, 2006 @ 12:17 am
As to the actual usefullness of Dawkins’ arguments, they are (as some of the participants have pointed out) arguments by analogy (i.e. transduction), and as such are very weak.
However, this is certainly not to say that all arguments used by scientists are of the same quality as those used by Dawkins. On the contrary, most of the concepts and principles that scientists have formulated over the past few centuries have been constructed using a combination of inductive, deductive, and abductive arguments, all of which include analogies (you can’t get around this), but which are much stronger than simple arguments by analogy. This includes the bulk of Darwin’s arguments for descent with modification by means of natural selection, as presented in the Origin of Species.
The alternatives to argument via transduction, induction, deduction, and abduction (i.e. the core of objective logic) are arguments via elimination (which encompasses most of ID theory) and via revelation (which encompasses virtually all of creationism). As Francis Bacon first formally pointed out, arguments via elimination are the last resort in science, as they generally are made only in the face of ignorance, which eventually yields to new research. And arguments via revelation aren’t arguments at all, they are merely assertions (and that way “there be monsters”).
Comment by Allen MacNeill — July 1, 2006 @ 3:14 pm
In anticipation of the remainder of Dawkins book, pages 303-304 have something of interest where the work of Motoo Kimura is discussed. It’s a shame that so little was covered here because Kimura’s work may have unwittingly overturned Darwin’s ideas using arguments from population genetics.
Kimura and the advocates of neutral evolution argue the majority of molecular evolution was not subject to natural selection. They justified their claims by pointing out populations would have to be giving birth and killing too many offspring to sustain the evolution of that many molecules! Think about it, the human genome alone has 4 billion nucleotide pairs. That’s a lot of molecules.
Neutral evolution mathematically demonstrated that population resources were inadequately available to natural selection. Natural selection essentially did not have enough “fuel” to carry the load of molecular evolution. The solution, then, was to say natural selection was not involved in the majority of molecular evolution. However, this “fix” could really only be asserted by requiring the majority of molecules under the purview of neutral evolution to be non-functional. And thus a clever “truce” was achieved between the neutralists and selectionists.
Many of the molecules subject to neutral evolution were then labeled junk DNA. Junk DNA was believed to occupy 98% of the human genome. Junk DNA could be comfortably compartmentalized away from the issues which concerned selectionists like Dawkins who focused on functional systems. Functional features were explained by natural selection, and evolution of junk by neutral evolution, and all was well until….recent developments in molecular biology.
Previously 98% of the human genome was viewed as junk. That number has steadily decreased to (by some estimates) 70%, and the number keeps still falling as new data arrive. And that means more of the formerly “junk DNA” is not really junk afterall.
Why is this significant? Answer: it shows natural selection was not the principal mechanism for the creation of biological complexity! Kimura after all proved these now-discovered-to-be-functional molecules were not subject to selection. We have therefore large scale complexity that evolved indpendent of natural selection.
The neturalists however, in winning one battle, may have lost the war because neutral evolution is a poor explanation for complex design (not that natural selection was either, but I hope one gets the point). Thus neutralist and selectionist evolution may be on the brink of collapsed. They have found fatal flaws in each other’s theories.
At least, that’s the way many IDers see it. We will see.
Salvador
[When Behe’s book is studied, see pages 225 through 226. Behe’s 1996 book (now 10 years old!) referred to the issue of junk. Recent developments since 1996 may favor the general ID leaning that there is a lot less junk than previously assumed. It would be instructive to see the criticism which Ken Miller directs at Michael Behe in light of recent findings. ]
Comment by Salvador T. Cordova, IDEA GMU — July 1, 2006 @ 4:41 pm
I suggest that Sal’s mumblings about readings that have yet to be read be ignored until the class actually has the opportunity to get to them.
Sal, Allen has a course plan and schedule, and has kindly published it on this web site. Don’t you think it’s a bit rude to try to jumpstart the conversation into areas not yet covered?
Comment by Don Baccus — July 1, 2006 @ 7:52 pm
Well, I’ve never yet had a professor scold me for working ahead :), and the rest of Blind Watchmaker is supposed to be read and prepared for by our next class (Thursday).
I don’t expect any of us will have time to discuss it this weekend, but there was nothing wrong with Salvador’s post.
Comment by Hannah — July 2, 2006 @ 8:31 am
A stylistic note for Rabia and future contributors: it would be good form to name the book you are reviewing. It is named in the comments, but not in your opening post.
The computer program he uses to illustrate natural selection works with a pre-determined goal oriented process - the exact opposite of natural selection!– though to his credit he acknowledges this flaw.
Hopefully if/when you get around to reading the Protein Science paper by Behe and Snokes, you will remember making this comment.
Comment by ivy privy — July 2, 2006 @ 1:32 pm
Oh, but there are many, many things wrong with Sal’s post. You might want to start with doing a little research on “junk DNA” …
Comment by Don Baccus — July 2, 2006 @ 1:58 pm
I haven’t analyzed Salvador’s post. The complaint you made originally was invalid. That’s all.
Comment by Hannah — July 2, 2006 @ 2:05 pm
I would like to point out that Dawkins does not call the selection in his simulation natural selection or argue that is demonstrating the power of natural selection, but rather calls this form of selection “cumulative selection” and leaves it up to the reader to draw the conclusion that this is synonymous with natural selection or is how natural selection works.
I believe an additional flaw in his argument is that he never makes this argument explicit.
Is natural selection a form of cumulative selection? What argument would make the case that it is? Does “cumulative selection” ever take a step backwards, or is it ever towards the goal? What about natural selection?
Does cumulative selection have the “power” that Dawkins attributes to it absent a goal or telos, absent the teleology that is presumably being denied? Both of these, I think, need to be answered. Both need to be demonstrated “scientifically” before we can rationally accept the “mutation + selection” argument.
Comment by Scott — July 2, 2006 @ 3:16 pm
In the other topic I have annonced that I will not be posting further due to family committments, but I will make this quick posting before leaving.
In fact, Dawkins explicitly says that it is not natual selection.
Comment by Ian Musgrave — July 2, 2006 @ 6:34 pm
Thanks Hannah for exposing the Dawkinian fallacy of “cumulative selection”. As shown in Ian’s post, although Dawkins previously confirmed the necessity of a “target phrase”, he quickly backpeddles and says nature does not have a target phrase (i.e, a long term goal). Is there a target phrase or not? If not, evolution cannot work. It can’t be both ways, and this part of Dawkins writings is typical–only convincing to the already convinced.
Comment by Roddy Bullock — July 2, 2006 @ 9:21 pm
Sorry, in my last post I said “Hannah” when I believe I should have credited “Rabia”. Please forgive the oversight.
Comment by Roddy Bullock — July 2, 2006 @ 9:24 pm
I fail to see where Dawkins explicitly states that it is not natural selection in your provided quote. Does he say it elswhere? Too bad you won’t be participating further, or I would ask you to point it out.
So what is the point of Dawkins’ demonstration if it is explicitly not natural selection that is being demonstrated? If natural selection doesn’t work in this same manner, what on earth is the logical connection between the simulation and Dawkins’ “Blind Watchmaker” argument?
Thank you for reminding me of this additional deficiency in his argument.
Comment by Scott — July 3, 2006 @ 1:37 pm
That is where he says it’s not like natural selection.
Then what’s the point of ANY simplified analogy in helping one understand ANY principle in ANY area of science?
Comment by Don Baccus — July 3, 2006 @ 1:48 pm
Dawkins says no such thing. Ian posted a relatively short paragraph, is it really asking too much of you to ask that you understand it rather than shoot down a misrepresentation of what it says?
He says his weasel program has a “target phrase”. He never says a “target phrase” is a necessity. And of course there exist other evolutionary models that don’t have targets of this sort and work quite nicely.
Comment by Don Baccus — July 3, 2006 @ 1:52 pm
Is there a target phrase or not? If not, evolution cannot work. It can’t be both ways, and this part of Dawkins writings is typical–only convincing to the already convinced.
Evolution cannot work without a target phrase? Please show your work. One cannot construct examples either with or without a target phrase? Please show your work. It appears that some are determined to remain unconvinced, no matter how much they have to torture reason to accomplish that goal.
Comment by ivy privy — July 3, 2006 @ 2:14 pm
Dawkins probably should have stated the point of WEASEL more explicitly in order to prevent the proliferance of strawmen. To most people, it’s obvious that he’s simply demonstrating the advantages of cumulative selection over non-cumulative selection. But other people, for whatever reason, ascribe much more ambitious goals to his illustration.
Comment by secondclass — July 3, 2006 @ 2:56 pm
proliferanceproliferationComment by secondclass — July 3, 2006 @ 3:19 pm
Dawkins’ example when read it the book is quite clear and yet seems to cause much confusion. That’s too bad as it seems to have generated many strawmen arguments on this board.
No Dawkins’ Weasel has no relevance to Behe and Snoke’s work, no Dawkins’ weasel has no relevance to natural selection in evolution.
In nature, the ‘fitness landscape’ is determined by the available mechanisms of variation and the environment. Changes in the genome either significantly increase, or decrease the fitness and selection can act or no significant change takes place and the effect is (near)-neutral. There is no ‘ultimate’ target, which is why the Weasel example has little relevance to evolution other than to show how selection can affect the probabilities involved. In the case of the Weasel example, fitness landscape is determined by the variation and a target. In evolution, it’s a relative fitness which decides the winners and losers.
Certainly Dawkins’ Weasel shows how many ID proponents (and perhaps even evolutionists) are unfamiliar with even the basic concepts of evolutionary theory and instead rely on a ‘gut feeling’ evaluation of what they believe evolution is all about. As such, it is much easier to quickly reject evolution. Sadly enough, what is rejected may just be a strawman version of the scientific concept of evolution.
Comment by Pvm — July 3, 2006 @ 6:05 pm
Regarding Salvador’s comment about “junk” DNA:
“Junk” is not a technical term, and thus lacks an official, agreed upon definition. Some refer to junk DNA as DNA that has no useful function; others mean DNA that does not code for protein and is not involved in gene expression regulation; but most commonly, it simply means non-coding DNA. In the latter two meanings, “junk” does not equate with “useless”. Now, Salvador writes:
And, if you take “junk” to mean “non-coding”, it still remains the case that over 98% of human DNA is “junk”, some of which is extremely useful. Less than 2% percent of all human DNA codes for proteins.
On the other hand, if by “junk”, Salvador means “useless” - and the rest of the paragraph seems to imply that he does - then I challenge him to present some evidence that it was ever the commonly held opinion of mainstream biology that 98% of human DNA is useless. After all, by the time human genome sequencing was completed, scientists have known about regulatory functions of non-coding DNA for at least two decades. (Searching for “regulatory DNA” on pubmed produces relevant articles going all the way back to the late 1970’s.)
What appears to be going on here is that Salvador is relying on different possible meanings of “junk DNA” to convey the impression that them silly evolutionists used to believe that 98% percent of the human genome was useless. I am certain that he is aware of the type of fallacy this constitutes.
Comment by Leonid Meyerguz — July 3, 2006 @ 6:10 pm
You cannot be serious. Are you one of the students in the class?
The point of an analogy is that “this” is like “that.” So what is the point of Dawkins’ “analogy” if immediately afterwards he admits that “this is not like that?”
Is there anyone else here who thinks that Dawkins is arguing from analogy?
Comment by Scott — July 3, 2006 @ 6:38 pm
But it is a necessity for the purposes of Dawkins’ argument.
Why didn’t Dawkins use one of them?
Does everyone here grasp what it is that Dawkins is arguing with his program? He is attempting to show “the power of cumulative selection.” If he cannot show the power of cumulative selection, then his larger argument fails, for it is premised upon this “power of cumulative selection.”
Dawkins is trying to show the power of cumulative selection by comparison to “single-step selection.” Why is this important? Because Dawkins is arguing that improbable things do not come about “by chance” (i.e., single-step selection).
But that improbable things can easily come about if we use “cumulative selection.” That’s it’s supposed power. To make the improbable probable.
But how can it do this lacking a goal or, in this case a “target phrase?” It logically cannot. At least not that I can see. CAn anyone else explain how?
Comment by Scott — July 3, 2006 @ 6:48 pm
It’s obvious to me, at least, that he is claiming to be demonstrating the advantages of cumulative selection over non-cumulative selection. Or as I would put it, the power of cumulative selection over single-step selection.
But does he really accomplish this? And how? And what, if anything, does it mean in the context of his overall argument? Those are the important questions.
He sets a TARGET for the single-step selection to reach.
He sets a TARGET for the cumulative-selection to reach.
With cumulative selection, the TARGET is easily reached.
In the context of his program, it is inevitable that the target would be reached. So it’s not just more probable, it’s certain.
But what if there are is NO TARGET? Then what happens to “the power of cumulative selection?”
Comment by Scott — July 3, 2006 @ 7:02 pm
Oh, well, maybe I was off by 3%
There are many examples of this. However, it was evident, up until the 1990’s one risked ridicule for trying to find function in “junk”.
Apparently, it has been the perception of some that there had been a prevailing viewpoint prejudiced to see the non-coding DNA as evolutionary leftovers, hence, “junk” echoed exactly their feelings rather than functional (which suggests design).
Amazingly, the tune has changed. All these duplicate regions make perfect sense in evolutionary theory (for whatever reason)….
Finally if large function is discovered for “junk”, a major argument for common descent and the tree of life is weakened and common design with a Linnaeuan hierarchy is strengthened. We will see.
Salvador
Comment by Salvador T. Cordova, IDEA GMU — July 3, 2006 @ 8:05 pm
Why should the selection criteria matter in terms of the power of cumulative selection? If there’s NO TARGET of course we can’t predict what selection will lead to, but no evolutionary biologist claims that humans were a necessary outcome of the evolutionary history of this planet.
Comment by Don Baccus — July 3, 2006 @ 8:29 pm
Salvador wrote:
I am afraid you are off by more than that. You have posted an extract from a 7-year news article from JNCI. The extract stated that when it was discovered that only 5% of the human genome codes for protein (we now know this to be inaccurate, as the actual figure is below 2%), the function of the remaining 95% of the genome was unknown. It went on to say that Ohno’s evolutionary remnant hypothesis was one possible explanation for the large amounts of non-coding DNA. Nowhere does it say that it was a commonly held belief among scientists that all 95% of the DNA with unknown function constituted useless junk.
So, once again, I invite you to provide serious evidence for your assertion that scientists viewed 98% (or even 95%) percent of the genome as useless junk. To my knowledge, this this is simply inaccurate - again, regulatory functions of non-coding DNA have been known since the at least the 1970s. Nonetheless if you could provide citations to published scientific papers expressing the opinion that 95% percent of human DNA is useless, or even some full in-context quotations from scientists in the relevant fields expressing this view in the present tense - i.e. no retrospective “we used to think it was useless, but now that we know better” quotes - I would very much appreciate it. Otherwise, for the sake of accuracy, it would be best to either retract or revise that part of your comment.
Incidentally, I’m a little curious about the source of the 70% percent figure you quote as well. I am thinking you might be referring to the work of Rigoutsos et. al. (PNAS 2006, 103[17]:6605-6610), but I could be wrong. Regardless, at this point, it is simply too early to say how much of the human genome might be functional. To my knowledge, this has always been the most commonly held view in the scientific community.
Given the nearly 10,000 papers on pubmed between the year 1970 and 1990 matching the terms “transcription regulation” alone, I trust you will forgive me if I doubt the anecdotal, slightly tongue-in-cheek excerpt you cite to justify this claim. Can you perhaps give an example of a single scientist who was ridiculed for researching the possible functions of “junk” DNA?
As for why non-coding DNA received so much less attention before the 1990s, the explanation is simple. Prior to the introduction and subsequent spread of PCR in the mid to late 1980s, genome sequences ware extremely costly and difficult to obtain. With such limited resources, scientists primarily focused their energies on protein-coding genes, which had already been an active area of research for a long time. Once genomic information became much easier to come by with the help of PCR, many scientists started giving non-coding regions significantly more richly-deserved attention.
Finally, by the time Behe’s book came out, many functional or potentially functional features of non-coding DNA have been discovered, and many more were expected, as the very articles you cite point out. His vague suggestion in DBB that some more of the junk DNA might turn out to have some unknown function was hardly visionary: by that time, pretty much everyone else in the revelevant sciences knew that already.
Comment by Leonid Meyerguz — July 4, 2006 @ 2:54 am
Genius of Junk
However, what shall I do, get a poll of the scientific community 30 years ago to establish an obviously qualitative assessment on my part? If Mattick views this disregard of non-coding DNA as a mistake, I take that as meaning some people certainly felt there was something culturally pervasive.
But focusing on such qualitative judgments is a distraction, and for the sake of this discussion then, I’ll withdraw it rather than allow this distraction to be a major issue in this thread.
The real issue was my point that the majority of molecular evolution had to be independent of natural selection, and if we continue to find more and more functionality in sections formerly deemed beyond the reach of natural selection, this could invalidate Dawkins hypothesis.
And I point out something Allen wrote in another thread:
This will be interesting because if short term effect of “parasitic” DNA is demonstrated to be potentially weakening to immediate fitness, but is also demonstrated to be anticipatory to certain possible future environmental stresses, this suggests foresight in the design, which is exactly what non-teleological evolutionary theory would not predict.
Redundancy or contingency designs are weakly selectable (for reasons I hope are obvious), and the “arrival of redundancy” is not easily justified as the work of a blind watchmaker because such designs are indicative of foresight to environmental conditions rather than to immediate short term organismal needs.
Finally, Lewontin in a Santa Fe paper winter 2003 made the observation of the difficulty of describing functionality with respect to fitness.
Lewontin did not phrase it this way, and he would think my interpretation of his work represents his view, however, I think he unwittingly found a serious flaw. (See Santa Fe Winter 2003.)
The issue is subtle, but it significant: one can not mathematically describe our conceptions of functionality and complex design in terms of natural selection. Natural selection in regards to evolution of complex design is shown to be self-contradictory, because biological complexity has no inherent mathematical meaning in the framework of blind watchmaking. Complexity is a meaningless phrase in Darwinian evolution, thus it is a severe category error to try to explain functional complexity in the world of natural selection, when functional complexity has no meaning within natural selections axioms!
In other words, it might not even in principal be able to formally describe functional complexity in terms of selective advantage. The only way to do so would be through and waving an ultimately circular arguments. Which is exactly a flaw in Dawkins hypothesis.
Salvador
Comment by Salvador T. Cordova, IDEA GMU — July 4, 2006 @ 12:39 pm
I don’t understand your question. I’ve been talking about the logical necessity for a target in order for Dawkins’ argument to work. I never said anything about selection criteria.
But let’s look at the selection criteria. The program looks at the letter at each position in the string and evaluates how close the new string is to the target phrase, and retains those strings that are closer, no matter how slight. Not gonna quote the book but that’s the gist of it. So the selection criteria is looking at the letters and evaluating how close to the target the offstring is.
Change the selection criteria. Don’t compare to the target sequence, compare to some other sequence perhaps. Present your alternative selection criteria.
Let’s say that we compare the new string to the phrase “oh what a wonderful world.” How many generations do you think it would take to arrive at the target phrase “methinks it is like a weasel” given this new selection criteria?
Do you really think Dawkins would be able to make the case for “the power of cumulative selection” if he changed the selection criteria of his program? I don’t. Anyone here who does?
Missing the entire point. Deal with my argument. It’s not about being able to predict a given outcome. It’s about increasing the likelihood of a given outcome. That’s what Dawkins’ program demonstrates how to do.
If there is NO TARGET, there is no way to compare the probabilities associated with single-step selection as compared to cumulative selection nor the number of steps it requires to reach the target if we apply cumulative selection, and thus no basis upon which to make the assertion that cumulative selection is any more “powerful” than single-step selection.
That’s the whole point of my argument and it shows Dawkins’ argument to be completely misguided and misleading.
If you disagree, then show how “cumulative selection” is more powerful than single-step selection without introducing a target.
You cannot do it. It’s logically impossible. Go ahead, try.
Comment by Scott — July 4, 2006 @ 3:23 pm
Sal again shows how limited familiarity with the available research may quickly lead one to conclude that “junk DNA” meant ‘no function’ when in fact junk DNA covers a variety of possibilities such as disfunctional genes (pseudogenes), and many many other all under the much more accurate term of “non-coding DNA”.
Sal however has yet to explain why ID predicts function for such DNA when ID also claims that the genome is deteriorating… In fact, I argue that Sal’s assertion is based not on foundational principles of ID but rather on religous foundations.
I invite Sal to prove me wrong and encourage him to show how function for “Junk DNA” follows from the foundational principles of ID.
Comment by PvM — July 4, 2006 @ 3:55 pm
Sal: Redundancy or contingency designs are weakly selectable (for reasons I hope are obvious), and the “arrival of redundancy” is not easily justified as the work of a blind watchmaker because such designs are indicative of foresight to environmental conditions rather than to immediate short term organismal needs.
Which is why, rather than redundancy, we observe degeneracy, something which can be very well understood from an evolutionary perspective. It is important to understand how modularity, degeneracy, evolvability arise almost ‘for free’ or ‘naturally’ in evolving systems.
There is an extensive set of research on this topic ranging from Fontana, Schuster, Stadler, Toussaint and countless others. Until Sal can explain to us why their research should be ignored, I find his claims, especially since they are so strongly contradicted by science, of relatively little interest.
Comment by Anon — July 4, 2006 @ 4:20 pm
This sounds like you’re saying that the 95% of ‘non-coding’ DNA is ‘degenerate.’ Which almost sounds like you’re calling it ‘junk DNA’. Could you explain what you mean by ‘degenerate’, and how this degeneracy ‘evolves’?
Comment by Lino D'Ischia — July 4, 2006 @ 6:13 pm
Not that I want to get dragged in to this, but the two claims are not mutually exclusive. Elementary logic. In fact, a deteriorating genome presumes functionality, don’t you think? What sense would it make to say that non-functioning regions of the genome are deteriorating?
But a question. Was the source of the idea of “junk DNA” theoretical or observational? I thought it was the former.
Comment by Scott — July 4, 2006 @ 6:57 pm
Lino: This sounds like you’re saying that the 95% of ‘non-coding’ DNA is ‘degenerate.’ Which almost sounds like you’re calling it ‘junk DNA’. Could you explain what you mean by ‘degenerate’, and how this degeneracy ‘evolves’?
Aha, I see the confusion. Degeneracy means something different. So let’s look at the terms redundancy and degeneracy.
Redundancy: Identical systems perform similar functions (backup systems)
Degeneracy: Dissimular systems perform similar functions.
While in technology we often see redundant systems, biology involves mostly degeneracy where different genes perform similar functions, increasing robustness.
Before I continue to explain how evolution explains degeneracy, I want to make sure that we accept and understand these definitions.
For instance the genetic code is degenerate since different triples code for the same aminoacid.
Scott: Not that I want to get dragged in to this, but the two claims are not mutually exclusive. Elementary logic. In fact, a deteriorating genome presumes functionality, don’t you think? What sense would it make to say that non-functioning regions of the genome are deteriorating?
But a question. Was the source of the idea of “junk DNA” theoretical or observational? I thought it was the former.
The argument is that Sal and ID claims that Junk DNA should have function after all and that this ‘prediction’ follows from ID.
Junk DNA is a term which describes ‘non coding’ DNA. That is DNA that does not gets coded into proteins. As such Junk DNA describes parts of the genome for which no function has (yet) been found. Since we 1) find conserved ‘junk DNA’ there is a good chance that there is function involved 2) On the other hand there are known examples of once functioning genes which now are pseudogenes and functionless or at least they do not perform their original function.
If ID predicts the deterioration of the genome then ID would predict functionless DNA, contrary to Sal’s prediction.
Hope this clarifies.
Comment by Pimothy — July 4, 2006 @ 7:38 pm
So let’s continue to explore the concept of redundancy and degeneracy (also referred to as distributed redundancy). As Wagner explains in Distributed robustness versus redundancy as causes of mutational robustness.
Abstract A biological system is robust to mutations if it continues to function after genetic changes in its parts. Such robustness is pervasive on different levels of biological organization, from macromolecules to genetic networks and whole organisms. I here ask which of two possible causes of such robustness are more important on a genome-wide scale, for systems whose parts are genes, such as metabolic and genetic networks. The first of the two causes is redundancy of a system’s parts: A gene may be dispensable if the genome contains redundant, back-up copies of the gene. The second cause, distributed robustness, is more poorly understood. It emerges from the distributed nature of many biological systems, where many (and different) parts contribute to system functions. I will here discuss evidence suggesting that distributed robustness is equally or more important for mutational robustness than gene redundancy. This evidence comes from large-scale gene deletion studies, as well as from the functional divergence of redundant genes. I also ask whether one can quantify the extent to which redundancy or distributed robustness contributes to mutational robustness.
So in other words, contrary to what we find in technological systems, robustness is not as much caused by duplicate systems as much as distributed redundancy.
Distributed Redundancy: In distributed robustness, many parts of a system contribute to system function, but all of these parts have different roles. When one part fails or is changed through mutations, the system can compensate for this failure, but not because a “back-up” redundant part takes over the failed part’s role.
So if we were to use the approach from analogy then we notice that in case of biology, nature does not really mimick technology but rather finds robustness in distributed redundancy aka degeneracy.
Seems that it’s time to return to the drawing boards for ID. Or perhaps some ID proponent can explain why ID would expect degenerate systems…
Once again the hard work of science is unraveling intricate details about biology which can be explained in terms of regularity and chance. Somehow the findings also seem to undermine the expectations from an ID perspective which would predict redundant systems to play a more prominent role (remember Sal’s magnetos…?)
Comment by Pimothy (PVM) — July 4, 2006 @ 11:24 pm
Would one of you folks opining on the blindness of the biological community on the junk DNA issue please explain to us:
(1) Why pufferfish get by with virtually no repetitive sequences in their DNA, resulting in a genome about 1/10th our size (IIRC) but with about the same number of genes as in the human genome?
(2) Why some plants (like certain ferns), salamanders, and amoebas can have 100-200 times as much DNA as the human genome (IIRC), mostly through vastly increased repetitive sequences?
(3) Why no ID advocates ever acknowledge these basic, textbook-level facts? They constitute strong evidence, known for 50 years or more, that most of the DNA in the human genome (and many other genomes), has at best a marginal, nonspecific, and rather dispensable function.
I have pointed out these facts to creationists many times, but it seems to go in one ear and out the other.
Comment by Nick (Matzke) — July 5, 2006 @ 12:59 am
I don’t know, and no one knows, but that is a reason to investigate it. Further, it AMOUNT of so called “junk-DNA” in a species has created what is known as the c-value paradox. That is, each speicies tends to have amount, and this suggests some sort of organizing prinicple.
What is perhaps of note is the ratio of non-coding DNA to total genomic DNA, not the raw numbers in isolation. Mattick (in an article available at www.noncodingdna.com) points out this ratio is correlated with increasing biological complexity. It thus appears an cybernetic organizing principle is at work, and that has teleology written all over it.
I just acknowldedge it and then some. Your objection is now negated (that’s not to say it ever was valid to begin with).
Dispensable? How does one define displensable with respect to fitness? Does one define it with respect to stress free environment, or a stressful environment. Lewontin (in the article I linked to above) points out why such definitions with respect to fitness are absolutely meaningless even when there is clearly biological function!
That’s exactly what is wrong with the fitness view of life. Organism with redundant backup systems will allow multiple knockouts of systems with little or no effect on fitness, overlooking the obvious fact function exists!
See the reason why at:
Airplane Magnetos, contigency designs, and reasons ID will prevail
Salvador
Comment by Salvador T. Cordova, IDEA GMU — July 5, 2006 @ 9:17 am
Sal counters simple facts with wild rhetoric and desperate assertion.
So Sal, since you know so much about junk DNA, please tell me what c-value (DNA amount) actually correlates with? I know the answer, do you?
It’s not “organism complexity”, because ferns and salamanders have 200 times as much DNA as humans, and within these groups organisms with the same amount . Mattick’s claims, further spun by Sal:
…are ludicrous, because his “data” on the ratio is based only on organisms with full genome sequences. Humans have the largest genome of any critter sequenced so far, and because the difference in genome size is mostly due to noncoding repetitive elements, humans have the most noncoding DNA of any organism sequenced so far. So Mattick’s data is an artifact. If someone would spend the bucks to sequence ferns and salamanders, his “pattern” would evaporate.
Comment by Nick (Matzke) — July 5, 2006 @ 12:02 pm
Sal continues his flawed analogy, comparing biology with redundant system like magnetos.
I assume that once he realizes how inappropriate these analogies are, he will abandon them and accept a flawed prediction of ID?
Sal: That’s exactly what is wrong with the fitness view of life. Organism with redundant backup systems will allow multiple knockouts of systems with little or no effect on fitness, overlooking the obvious fact function exists!
But organisms have no redundant backup systems but rather distributed redundancy aka degeneracy.
Another ID hopeful bites the dust.
Comment by Pimothy (PVM) — July 5, 2006 @ 12:39 pm
If you disagree, then show how “cumulative selection” is more powerful than single-step selection without introducing a target.
You cannot do it. It’s logically impossible. Go ahead, try.
Great timing, Scott! Dave Thomas has done exactly that in a post at the Panda’s Thumb, complete with working example. Why don’t you head over there and remind him that what he accomplished is logically impossible?
Comment by ivy privy — July 5, 2006 @ 4:46 pm
Pimothy:
This sounds like a mathematical definition of ‘degeneracy’, where you have more than one non-unique solution to a set of equations.
I’m afraid that ID doesn’t predict it; but the Law of Entropy does.
Since ID supposes that we’re dealing with a superior intellect, ID would ‘expect’ an informational system that at the very least would be equal to, but almost without doubt, would be beyond anything that man has already conceived. (N.B. Everyday we read about scientists stealing design ideas from ‘nature.’) This is what we seem to be seeing. Bill Gates has already said as much. And let’s remember that engineers have been building redundancy into systems for decades now.
Yes, and we’ve seen millions of conserved nucleotide bases being excised from normal mice ova, with the incredibly unexpected result that mice born from these ova are born perfectly normal! I suspect this is the exact reason that the Darwinists are now talking about ‘degeneracy’; afterall, when a result is in conflict with the ‘theory’, all you have to do is invent a new ‘word,’ and, to many, it sounds like you actually know what you’re talking about.
Comment by Lino D'Ischia — July 5, 2006 @ 5:36 pm
There were numerous typos in my last post (my apologies), and my wording was confusing, so I will attempt to make amends…
C-value is roughly the amount of DNA in a cell’s nucleus (the more exact definition in terms of haploid nucleus is given here).
Regarding the c-value paradox, here is a description:
Given a particular species, there tends to be constancy of c-values among the same members of that species, but what was surprising was that the c-values from species to species varied widely and bore no resemblance to the number of genes. This was viewed as paradoxical.
I suggested that c-value constancy within a species hint of an organizing principle, a cybernetic organizing principle. (A cybernetic organizing principle is a something of a general design goal, such as energy efficiency.)
This leads to Nick’s question:
The c-values correlate with cell volume, metabolic rate, cell division rate, and other things…
Thus c-value seems tied then to whatever organizing principle is driving metabolic rate and the other factors it correlates to…
But, if the amount of “junk DNA” seems governed by an organizing principle, why should we be so quick to view it as junk, since the amount of “junk DNA” seems to have some significance?
200 times? Salamanders have about 26 times the amount of DNA. Ferns about 47 times, not 200. (See PLU NSCI.)
Nick’s “refutation” assumes that the number of genes in fern and salamanders will not be governed by the pattern Mattick asserts.
In any case, Nick is trying to refute Mattick’s claims by data that hasn’t even been discovered. :-) Isn’t that a bit pre-mature?
Comment by Salvador T. Cordova, IDEA GMU — July 5, 2006 @ 5:47 pm
Nick Matzke:
Is this a prediction?
The way you’re framing this, it sure sounds like you’re arguing that ‘non-coding’ DNA is ‘junk-DNA’. Is that right? I thought Darwinists have known for 30 years or more that ‘junk-DNA’ isn’t really junk.
Comment by Lino D'Ischia — July 5, 2006 @ 5:50 pm
The definitions:
Biology has both, not either or. In the very article PvM quoted by Wagner:
In Wagner’s book, Robustness and Evolvability in Living Systems
Wagner’s point is redundancy is part of the overall robustness (tolerance to injury) of the organism.
I opt not to respond to the rest of PvM posts. They are filled with similar flaws. Taking time to clean up all the inaccuracies would only distract from the issues of this thread.
I used the concept of “contingency designs” (which include redundancy, degeneracy, etc.) to highlight the issue of functional complexities not easily visible to natural selection.
I suggested the neutral theory and these contingency designs (which are weakly selectable, and selectable only in certain stressful environments) pose a problem for natural selection because of their lack of visibility to natural selection.
Think about the ability to knock out genes without any phenotypic effect. How do such capabilities arrive in an organism since they are not immediately visible to natural selection? This is the kind of complexity invisible to Dawkins blind watchmaker, so how could it possibly evolve? This suggests that natural selection had little to do with the origination of these capabilities.
Comment by Salvador T. Cordova, IDEA GMU — July 5, 2006 @ 6:34 pm
Lino
I thought Darwinists have known for 30 years or more that ‘junk-DNA’ isn’t really junk.
To the extent the nucleus or nucleoid bears little resemblance to a trashcan, yeah, scientists have known for some time that “junk-DNA” isn’t “really” “junk.”
Comment by Susan Percell — July 5, 2006 @ 6:50 pm
Avida is badly flawed and it’s simulation only illustrates tautologies, and unrealistic simulation parameters, not real science. It thus has similar flaws to Dawkins Weasal.
[ example of a tautology: “all tables are tables”]
There are two critiques at ISCID.
Bits, Bytes and Biology: What Evolutionary Algorithms (Don’t) Teach Us About Biology
and Evaluation of Neo-Darwin theory by Avida by Royal Truman. A discussion in much greater and tedious detail is at ISCID: Evaluation of neo-Darwinian Theory with Avida Simulations
Salvador
Avida is a buggy program, and I can also claim I helped correct one of their source files in version 1.6 :-)
Comment by Salvador T. Cordova, IDEA GMU — July 5, 2006 @ 7:13 pm
Well, let’s see. I am an ID advocate and I acknowledge these basic, textbook-level (whatever that means)) facts.
So I’m in a bit of a quandry how I can possibly answer your question. It’s obviously based upon a false premise.
Comment by Scott — July 5, 2006 @ 7:57 pm
Because I am here and not there and don’t care to go slogging about int he morass that is the PT. Perhaps you can explain the basics to us here so that we all don’t have to go over there.
Comment by Scott — July 5, 2006 @ 7:59 pm
You’ve seemed to have missed the whole point. Nick Matzke seems almost to suggest that all this ‘excess’ DNA is of no utility, (you see ‘gobs’ of it, then you don’t see ‘any’ of it–so it must just be useless) and is, thusly, . . . . ‘junk’. So, which way is it?
Comment by Lino D\Ischia — July 5, 2006 @ 8:11 pm
ivy privy says:
Has he? The only references I see in that thread to cumulative vs single-step selection is Dave Thomas quoting Dawkins about WEASEL, and you quoting Scott from this thread.
Doesn’t it require a little more discussion to indicate how he did that, and what it means?
I think it is time to explore the principles involved in a little more detail.
That post will follow shortly.
Comment by Roger Rabbitt — July 5, 2006 @ 8:13 pm
ivy privy
Sorry to disappoint you, ivy privy, but it would appear that Dave Thomas has accomplished something of trivial importance. He selects as a kind of ‘target’ the solving of Steiner’s Problem. Then he gives diagrams of what they look like. However, THEN….. he gives photos of the various solutions to this class of problems: they’re soap bubbles between various post positions. What this means, regretably for the Darwinist darlings, is that he’s simulating something that is entirely a function of natural laws and forces; that is, all the ‘information’ necessary can, and is, already found in nature. In other words, if you write down the proper constraints, you’ll end up with the desired end-product. There’s just no parallel between this kind of modeling and biological reality (which is a function of chance and necessity). I suspect it’s no more than solving the ‘physics’ of the problem using a kind of perturbation theory. Unless there’s more to it than that, I don’t need to do anything beyond looking at the soap bubbles to know that we’re dealing with irrelevance.
Comment by Lino D\\Ischia — July 5, 2006 @ 8:28 pm
In Dawkins WEASEL program, we start out with the “target phrase” METHINKS IT IS LIKE A WEASEL. The point of the exercise is to supposedly demonstrate that “cumulative selection” can easily master problems that are daunting to “single-step selection”. Fair enough. Given the nature of the exercise, we need the target phrase to evaluate success, and compare how the two approaches work.
The problem with this approach is not just that we have the target phrase, but that we use the target phrase as part of the fitness function. As Dawkins acknowledges, that isn’t how evolution, or more specifically, natural selection, works. Hence Scott’s challenge, whence the “cumulative selection” advantage if there is no access to the target phrase as part of the fitness function? Or maybe more generally, how do we define a “cumulative selection” consistent with natural selection.
Now, Dave Thomas comes along with a GA that doesn’t have a “target phrase”. First, we need to understand that it isn’t a program designed to produce “phrases” (although one could label the “DNA” as a bit string “phrase”, that misses the significance of what the program produces and how the selection operates), so at a very basic level, a “target phrase” isn’t really appropriate. It doesn’t operate, like the WEASEL program, to produce, with a certainty approaching 1, a specific phrase as an output. It instead produces, depending on several variables, a collection of logical diagrams skewed toward certain characteristics, sometimes producing the “Steiner’s Solution”.
Now, given the odds, the program would seem to produce that “Steiner’s Solution” more often than one would expect from a “random” “single step selection” program, i.e. selecting only when that solution has been found.
So we seem to have a “cumulative selection” program outdoing a “single-step selection” program and we don’t have a “target phrase”. But is that correct?
Well, we still do have a “target”. And the fitness function is still using that target. From Dave’s posting:
And since he knows the “Steiner’s Solution” “is the most compact network of straight-line segments that connects the points”, selecting for short total-length is gonna greatly increase his odds. But of course, that is still forbiddedn to natural selection. It has no “goal”, hence can’t use any information about the “goal” to find it.
The problem in WEASEL, which still remains in Dave’s algorithm, is that we are using information about a goal to tweak the fitness function, and Natural Selection, especially as Dawkins defines it, can’t do that.
So, Scott’s question, more generically stated, still remains: How can we have cumulative selection in NS, if NS can’t use any information about a goal to get there?
Comment by Roger Rabbitt — July 5, 2006 @ 8:53 pm
In other words, life contains a spiritual, supernatural element and isn’t merely a function of the natural laws of physics, chemistry, etc?
Do you have empirical evidence to support this assertion?
Neither the program nor natural selection have a “goal”. Both select according to a fitness function. You seem to be arguing that no fitness function is allowed at all, which is silly.
Where is “there”? Seriously … thinks about it.
Comment by Don Baccus — July 5, 2006 @ 9:24 pm
Roger Rabbitt wrote
Thomas’s GA demonstrates that evolutionary mechanisms can ‘find’ solutions — note the plural — whilst “knowing” nothing but what’s better within a given population in a given selective environment. The fitness calculation knows nothing about Steiner’s Solution”. It “knows” only that “shorter is better”, just as in a predator-prey arms race, the prey population may “know” that “faster is better” with no specification at all of how to be faster. Roger apparently missed the fact that half a dozen ’solutions’ — MacGyvers — evolved: the system did not inevitably evolved the shortest path. It evolved short paths. Evolution is not at all “forbidden” to use relative fitness on some criterion defined by an environment, be the environment “natural” or determined by humans. Humans — scientists — do research by manipulating variables thought to be relevant in ‘natural’ contexts.Addressing Roger , Don Baccus wrote
Don has put his finger on the creationist misapprehension of GAs. They imagine that because humans use them as research platforms and manipulate variables like the selective environment to do research, humans are somehow injecting “information” into the algorithms. That’s reasoning on the same level as claiming that because Petri dishes and culture media are intelligently designed, research on E. coli that manipulates the composition of culture media is somehow injecting the results of the research into the media. Platforms like Thomas’s or Avida allow manipulating variables thought to be relevant to the evolution of stuff and test hypotheses about them. In contrast to Dawkins’ WEASEL, they are research platforms, not tutorial demonstrations.RBH
Comment by Richard B. Hoppe — July 5, 2006 @ 9:43 pm
Lino
if you write down the proper constraints, you’ll end up with the desired end-product. There’s just no parallel between this kind of modeling and biological reality (which is a function of chance and necessity).
“Just no parallel?” Here we go again. I guess it depends on how you define “just no parallel.” And that just depends on whether you are willing to deny obvious facts about nature and selection in an effort to deny evolutionary biology as science until, literally, no one except creationists pays any attention to you.
Selection works to change the genotype and phenotype of individuals in a population. Selection is “biological reality”, whether that selection is the result of humans putting selective pressure on organisms, or predators putting selective pressure on organisms, or the environment putting selective pressure on organisms (and the natural environment can include, of course, predators and/or humans). Natural selection operates on some level wherever there is a living organism that is capable of reproducing.
This is basic biology. 8th grade level or lower, I’d guess. I’m not sure what your problem is, Lino.
Also, how are Thomas’ results NOT a function of “chance and necessity”? The individuals need to have certain features to “reproduce” and the individuals have those features by chance. This seems parallel enough to mark your claim as clearly false.
As for this
if you write down the proper constraints, you’ll end up with the desired end-product
If only it were that easy! As any scientist who has performed genetic screens will tell you, determining the “proper constraints” which will enable you to achieve the “desired result” is often very complicated due to the intrinsic variability in populations and the unpredictability of the different ways in which living organisms can continue to live under conditions of selection.
But of course, that is still forbiddedn to natural selection. It has no “goal”, hence can’t use any information about the “goal” to find it.
Short version: “you can’t simulate evolution or any aspect of evolution to my satisfaction unless the result shows that something can not evolve without the input of intelligence.”
Right? Has there ever been a simulation of an aspect of evolutionary biology which satisfied you?
Returning to your strange claim that natural selection … has no “goal”, hence can’t use any information about the “goal” to find it I am struggling to understand what you could possibly mean by “natural selection” “using information” about a “goal” to achieve a “goal.”
Individual organisms either survive long enough to reproduce themselves or they do not reproduce themselves. As a species, E. coli never had a “goal” of becoming better able to survive in the human digestive tract. Tigers never had a “goal” of having sharp teeth. Chimpanzees never had a “goal” of looking a lot like human beings except with more hair.
Selective pressure on an organisms’ traits comes the environment exactly as the selection in Dave Thomas’ experiment comes from the environment.
Again: what is the problem? How can you say there is “just no parallel” with the way that selection “works” in nature? It’s perfectly parallel. The difference is not in the selection that is being put on the “organisms” but on the “organisms” in Thomas’ experiment. They aren’t alive. But that’s sort of the way that these simulations work so complaining about
that is just complaining about simulations generally which is … well, sort of hypocritical whining when it comes from an ID promoter who contributes zilcho original research to the discussion.
Comment by Susan Percell — July 5, 2006 @ 10:39 pm
if you write down the proper constraints, you’ll end up with the desired end-product. There’s just no parallel between this kind of modeling and biological reality (which is a function of chance and necessity).The desired end-product is survivability in the environment. In a simulation the environment is, by necessity, artificial. In nature and in the simulation the fitness function selects for organisms that are adapted to the environment. The fact that the author of the program knows the optimal solution does not mean anything if the software doesn’t know the optimal solution. The software just applies “shorter is better” and “disconnection is death” pressures to the population and occasionally reaches the optimal solution. More often it reaches a “good enough” solution. Even when it does achieve the optimal solution, it doesn’t know it. It will continue to produce more generations of organisms, some less than optimal, but all generally suited for the environment defined by the fitness function. This is exactly what evolutionary theory predicts and what we see in nature.
Comment by Matt Peterson — July 5, 2006 @ 11:17 pm
Sadly enough Sal ignores the issue and focuses on a strawman namely that nature has no redundancy. Of course Sal should have known that the issue of duplicate genes is a hot topic in evolution and a major source of innovation but Sal has yet to explain why nature shows far more prevalent degeneracy in the genome.
So when Sal writes
it seems clear that Sal is not only running away from the argument but also making additional assertions which he cannot support.
Why is it that ID has to hide itself in the shadows of ignorance anytime science exposes it for what it is?
When Sal is up for the task to address the real issue, he is certainly invited to respond and hopefully this time to the real issue.
So let’s explore further Sal’s ‘response’
But are they truely invisible to natural selection and is natural selection the only relevant mechanism in evolutionary theory? Of course not and compare this with how ID ‘explains’ degeneracy and redundancy… Poof?
And your argument is what Sal? That natural selection is not all there is? Do you even understand how science carefully explains how degeneracy arises? Do you understand how science explains the existence of redundancy?
The scientific vacuity of Sal’s ‘response’ indicates that ID has not only no explanations of its own but that the existence of degenerate systems is a major falsification of the ‘prediction’ by ID. Similarly, if duplications are not under selection, then is Sal’s argument that these are maintained by the designer(s) over the hundreds of millions of years?
Gene duplication is no friend of intelligent design but come to think of it, anytime science provides explanations, it is not very friendly to ID which relies on our ignorance to promote its scientifically vacuous ideas.
Degeneracy and redundancy are exquisitely well explained by natural processes which sort of means that ID has no role to play here and which may help understand why Sal did not address these arguments.
Comment by Pimothy--(PVM) — July 5, 2006 @ 11:20 pm
Roger And since he knows the “Steiner’s Solution” “is the most compact network of straight-line segments that connects the points”, selecting for short total-length is gonna greatly increase his odds. But of course, that is still forbiddedn to natural selection. It has no “goal”, hence can’t use any information about the “goal” to find it.
The problem in WEASEL, which still remains in Dave’s algorithm, is that we are using information about a goal to tweak the fitness function, and Natural Selection, especially as Dawkins defines it, can’t do that.
No, no no… There is no goal beyond a simple rule like the fitter the better. There is no global rule and in some cases, the solution gets stuck in local optima. Now I understand that the concept of a fitness function may be non-trivial to understand but if you can define what the global goal is and how this relates to the outcome then please explain.
Optimize fitness is not a global goal, reduce lenght is not a global goal… These are simple fitness functions which in no way prescribe the solution.
For instance a similar example exists in the Traveling Salesman problem, as Wesley Elsberry has so eloquently shown and which has forced Dembski to accept the existence of actual and apparent CSI, admitting that CSI is not a reliable indicator of design. Why do ID proponents not see the major concession by Dembski here?
Or does it matter that Dembski accepted that CSI is irrelevant?
There is no distance from ‘Steiner solution to actual solution’ in the fitness function… Remember… Unlike Dawkin’s distance which is a direct distance to the goal. All that is needed is a simple rule of minimizing length (or in case of evolution: optimizing fitness) without enforcing the solution.
Comment by Pimothy--(PVM) — July 5, 2006 @ 11:27 pm
Thanks…
Congrats! You looked it up! Now find me some ID webpages where ID advocates (DI, Dembski, IDEA clubs, etc.) where they talk about “junk” DNA and the blindness of mainstream biologists, and where the ID advocates actually discuss the c-value data and its implications. In my experience such pages don’t exist. Kind of a big hole, don’t you think?
This is vague, mostly meaningless technobabble that you seem to have thought up over breakfast. If energy efficiency was the goal, then ferns wouldn’t be wasting billions of base pairs of DNA per cell, when the more complex Arabidopsis mustard plant can get by just fine with orders of magnitude less DNA.
Ding! This is correct. The fundamental correlate in the literature seems to be cell volume/nuclear volume. The correlation holds across many eukaryote groups and about five orders of magnitude in C-value.
Like I said before, it might have some kind of minimal, nonspecific function — perhaps big cells need “spacers” between their genes, and any old sequence will do, so repetitive elements are preserved. On the other hand, perhaps bigger cells can simply tolerate more junk because the cost of DNA replication is relatively tiny for a large cell.
These hypotheses, and related ones, are the ones being seriously discussed by the actual scientists who actually work in this area. This is why I am mystified when creationists and the press, even some science journalists, hyperventilate because some function (sometimes a “function”) is found in a pseudogene or something. Pseudogenes are a tiny fraction of the DNA. Repetitive elements are dominant, and it is differences in the amount of repetitive elements that accounts for the vast differences in c-value between various eukaryote species. The two serious families of explanations for the c-value differences and their correlation with cell volume are roughly (1) junk DNA and (2) non-specific structural function based on bulk amount of DNA. The idea that this repetitive DNA has some kind of amazing software function is extremely unlikely based on the data we have, and therefore the idea that the magical “design” explanation should be invoked to explain this imaginary “software” is even more ludicrous.
Anyone who is bashing biologists about “junk DNA”, and at the same time ignoring these basic facts well-known in the actual field, isn’t in the ballpark of making a serious argument. Unfortunately the whole ID movement falls into this category. It’s about as bad as when the young-earth creationists yammer about the Second Law of Thermodynamics and ignore the fact that the earth is an open system with massive amounts of free energy flowing through it.
200 times? Salamanders have about 26 times the amount of DNA. Ferns about 47 times, not 200. (See PLU NSCI.)
Eh, I said “IIRC” (if I recall correctly) at one point. I was going from memory. This webpage says that one fern species has 250 billion base pairs, compared to only 3 billion for humans. Amphibians also can get into the >100 billion base pair category.
Regardless, the basic argument remains the same. Are you seriously going to argue that ferns and salamanders have super-complex software in their genome requiring 100 billion extra base pairs of DNA above and beyond that found in the human genome?
Well, even without spending millions of dollars to sequence hundreds of billions of base pairs of boring repetitive sequence (Try putting in the grant for that! Except for humans and human relatives, labs and funding agencies prefer sequencing short genomes), even with more generic techniques we know (1) huge c-values; (2) relatively low organismal complexity (at least if we assume humans are near the top, which Mattick and creationists both do); (3) the large differences in c-values are always due primarily to variations in the amount of repetitive elements. The idea that ferns and salamanders will turn out to have hundreds of times more genes than humans is not credible, given these facts. But, if you like, you can call it a prediction. Evolution is good for those. What’s your prediction, Sal?
Comment by Nick (Matzke) — July 5, 2006 @ 11:45 pm
Apologies, this should have been in quotes:
Comment by Nick (Matzke) — July 5, 2006 @ 11:48 pm
Lino: What this means, regretably for the Darwinist darlings, is that he’s simulating something that is entirely a function of natural laws and forces; that is, all the ‘information’ necessary can, and is, already found in nature.
Excellent so natural forces can after all explain why evolutionary mechanisms/algorithms can generate complex specified information. Excellent progress we are making here. Similarly, all the information for the genome is found in the environment and is effectively ‘transferred’ to the genome via selection. Somehow I believe we are getting close to some resolution of ID’s claims.
Comment by Pimothy--(PVM) — July 5, 2006 @ 11:56 pm
Well, it seems to me that ferns owe a their large DNA size to massive duplications. Not only are they polyploid, some are (gasp) 1000-ploid (the haploid is composed of 1000 monoploid sets). Thus it may not be a matter of complexity, so much as repetition. Thus the idea that ferns may have hundreds of times more genes than humans is eminently reasonable given the fact of possibly 1000 of chromosome duplicates in the haploid, not just duplications of repetitive elements.
Salamanders are known to also be polyploid, so that may account for some of it too. And if there have been genome wide duplications, that will increase gene numbers as well.
Thus, your claim,
May not be sustainable, and Mattick may be right.
Salvador
Comment by Salvador T. Cordova, IDEA GMU — July 6, 2006 @ 12:48 am
So, are you telling me that duplicating the chromosomes makes an organism twice as complex? I guess increasing complexity is easy, then! You better tell Dembski et al. that you have sunk their entire argument.
Comment by Nick (Matzke) — July 6, 2006 @ 12:57 am
Susan wrote:
This isn’t basic biology, this is something right out of a Darwinian Bible Camps.
As an exercise, let’s replace the word ’selection’/'natural selection’ with the word ‘death’—since, after all, Dawkins tells us that selection is nothing more than the Grim Reaper.
Here’s how it sounds:
“Death works to change the genotype and phenotype of individuals in a population. Death is “biological reality”, whether that death is the result of humans putting deadly pressure on organisms, or predators putting deadly pressure on organisms, or the environment putting deadly pressure on organisms (and the natural environment can include, of course, predators and/or humans). Death operates on some level wherever there is a living organism that is capable of reproducing.”
Do you really want to go out on a limb and call this ’science’?
I don’t see any individuals over there at PT; all I see are soap films. You might need to take a look at it. Like I said earlier, one look at what was going on convinced me not to bother following the argument. Surprise me if you think my assessment is wrong.
[And, btw, let’s distinguish, please, between Roger Rabbit’s remarks and my own.]
Don Baccus wrote:
Fitness functions are to evolutionary algorithms what actuarial tables are to insurance agents–sources of important information. (This is what I meant when I said that EA’s sneak in information through a side door.)
Matt Peterson:
Sorry, but we don’t see ‘nature’ making these kinds of evaluations: “shorter is better”, and “disconnection is death”. “Shorter” is short-hand for what will get you the right answer. This is just indirectly sneaking in the goal we’re striving for. The same is true, but to a lesser degree, with “disconnection is death.” As I say, this is an uninteresting example that basically puts constraints on an essentially physico/chemical system. We’re not dealing with any life-form here.
Not so quick. You’ll have noticed that I had information bracketed with apostrophes. I was referring to information such as length, heighth, and any physical properties of the fluid they might use to model it–all of this ‘information’ being inputted by an intelligent agent. As I wrote: :In other words, if you write down the proper constraints, you’ll end up with the desired end-product.”Pimothy says:
Now the world is full of, let us say, quantum information. But the question is, where did this information come from? Was it through random mutation and natural selection? Or, to put it another way, can physicists explain to us why electrical charge is what it is, or any of the other physcial characteristics of the world?
Comment by Lino D'Ischia — July 6, 2006 @ 1:04 am
And anyway, apparently ploidy isn’t going to explain the huge c-value for the fern Psilotum, because its ploidy level is not particularly high compared to other ferns. Instead, it has very large chromosomes:
2002 paper in Annals of Botany
Comment by Nick (Matzke) — July 6, 2006 @ 1:11 am
Lino gets even closer to the solution of ID’s fallacy namely that if you write down the proper constraints you will end up with the desired end product.
What Lino misses however is that evolution via the concept of fitness and combined with the many constraints, delivers exactly what Lino is looking for, a functional solution.
While Lino is still confused about the concept of evolutionary algorithms and ’smuggling in’ information, he must have come to realize that the ’smuggling in’ is done by selection which ‘transfers’ information from the environment to the genome.
In addition, Lino still fails to explain how the example ’smuggles in’ the solution since it surely does not present a fitness function of distance to the final solution. In fact, evolutionary algorithms ‘find’ the best solution given the constraints.
Certainly there is no target in the sense of ‘Weasel’… All we have is local information… In fact the solution is not a single one but one of several ‘close matches’and some of them were not even Steiner solutions.
So how is information ’smuggled’ in ? I know ID proponents like to make this claim and if so, how does one distinguish between the information smuggled in by an algorithm such as natural selection and information smuggled in by a designer?
So many unanswered questions but at least PT has caused ID to move the goalposts once again.
Comment by PvM — July 6, 2006 @ 1:20 am
What about a genetic program that, when given a fitness function to produce a desired result, takes advantage of environmental factors the researchers were not even aware of to comes up with a completely novel and unexpected solution?
Comment by Matt Peterson — July 6, 2006 @ 1:38 am
What surprises me is how much was made of the explicit fitness function used by Weasel and when PT shows examples of fitness functions which do not rely on explicit distance, the tune suddenly changes.
Why is it so hard to accept that ID was wrong about its claims about evolutionary algorithms?
Comment by PvM — July 6, 2006 @ 1:59 am
No, I didn’t suggest that, and that doesn’t represent my position.
You’re taking this discussion down to distracting issues, Nick, and you’re trying to controvert a well-received peer reviewed paper by invoking speculations about unmeasured data.
Back and forth of these side issues with me trying to refute your unsubstantiated speculations about why Mattick’s paper is wrong is taking focus away from the main points of this discussion.
If you insist the large sections of DNA are non functional, that Mattick is wrong, fine, give a percentage of how much you junk DNA your arguments can live with and maybe then this discussion can move forward. Endless speculation of fern c-value and fern complexity ratios are only jamming this thread.
This thread was not about defending ID, but the problems associated with the efficacy of natural selection in making complex designs.
I raised the issue of:
1. the problem of large sections of functional molecules effectively out of the reach of natural selection (based on Kimura)
2. the issue that functionality is essentially meaningless in the Darwinian view due to lack of rigor and essentially tautological definition of fitness (Lewontin’s paper, and others)
3. the problem posed by contingency designs which make them weakly selectable if not invisible to natural selection.
You’re invited to address those issues. (preferably not by obfuscating them to oblivion as PvM is doing, I hardly read any thing he posts as they are mostly mis-directions of my points rather than an attempt at a substantive refutation).
Sal
PS
Mattick’s complexity ratio would not be affected going from monoploidy to polyploidy in a haploid. If large chromosomes are due to large duplication events, Mattick’s ratios would also still hold.
Comment by Salvador T. Cordova, IDEA GMU — July 6, 2006 @ 2:10 am
This is an essentially tautologous statement (i.e. like “all tables are tables”). A trait may give advantage in one environment and disadvantage in another. This fluid valuing of traits says nothing of:
1. what is the likelihood of the evolution of complex design
2. how visible is a complex design to natural selection
In all of the computer simulations purporting to demonstrate the adequacy of natural selection (including Avida, Weasal, and the rest) either #1 or #2 are assumed to be sufficient in such quantity in the simulation to guarantee the outcome one wishes to prove.
In essense, one assumes the very thing one set out to prove, and this is circuitious reasoning.
Salvador
Comment by Salvador T. Cordova, IDEA GMU — July 6, 2006 @ 2:21 am
What is the threshold at which a design becomes “complex”?
Comment by Matt Peterson — July 6, 2006 @ 2:37 am
Or, to put it another way, can physicists explain to us why electrical charge is what it is, or any of the other physcial characteristics of the world?
Hustle now, move those goal posts!
Sorry, but we don’t see ‘nature’ making these kinds of evaluations: “shorter is better”, and “disconnection is death”. “Shorter” is short-hand for what will get you the right answer. This is just indirectly sneaking in the goal we’re striving for. The same is true, but to a lesser degree, with “disconnection is death.” As I say, this is an uninteresting example that basically puts constraints on an essentially physico/chemical system. We’re not dealing with any life-form here.
Yuor objections have gone well beyond the point of ridiculosity. And please explain how a life-form is not a physico/chemical system, without sneaking in any unproven supernatural concepts.
Comment by ivy privy — July 6, 2006 @ 11:52 am
Sal seems to be describing Intelligent Design here while misunderstanding the concepts of Avida etc which show how at least in principle evolutionary mechanisms can explain information and complexity.
ID, confronted with these facts, simply retreats to arguments that evolution is improbable, failing to provide much of any supporting evidence. After all for many IDers, the conclusion precedes the evidence. Add to the mix YECism and one may understand how some IDers maintain their faith despite the vaste amount of contradicting evidence.
Comment by PvM — July 6, 2006 @ 12:17 pm
Seems Sal is having a hard time dealing with my contributions :-) Let’s see how I can add to his despair by addressing some of his other poorly supported claims
Sal: This thread was not about defending ID, but the problems associated with the efficacy of natural selection in making complex designs.
But the two are the same. After all ID is nothing more than pointing out problems with Darwinian evolution to show ‘evidence’ of design.
So the question really is: Can Darwinian processes explain, at least in principle, complexity in nature and the answer is a resounding yes. As Adami et al and Schneider have shown, the simple processes of variation and selection are sufficient to explain increase in information/complexity in the genome.
When faced with this major problem, ID has to move its goalposts once again and point out that: Ok, these processes can at least in principle explain information/complexity so show us that they did… While a perfectly valid question, it should be clear by now that the lack of any ID-relevant hypothesis combined with the existence of plausible natural processes, reduces ID’s claims to dust.
So what’s there to do for ID but to ignore the obvious
I raised the issue of:
1. the problem of large sections of functional molecules effectively out of the reach of natural selection (based on Kimura)
Irrelevant. Neutrality and Selection are both mechanisms of evolution and need to be addressed by ID. Pointing out that natural selection does not explain it all is just a red herring.
Neutrality has been shown to be a selectable and essential trait. In other words, under the processes of selection, neutrality can be selected for and neutrality itself becomes the foundation for the efficiency of evolution.
2. the issue that functionality is essentially meaningless in the Darwinian view due to lack of rigor and essentially tautological definition of fitness (Lewontin’s paper, and others)
ah, the natural selection is a tautology argument again, laid to rest many many times and yet creationists still repeat these silly claims.
3. the problem posed by contingency designs which make them weakly selectable if not invisible to natural selection.
In fact, contingent design, or better stated degeneracy is very well understood by concepts of evolution such as gene duplication, preferential attachment, and selection. It’s remarkable how science has shown how simple and basic, and observed, processes can explain the existence of redundant and degenerate systems which both confer robustness as well as evolvability to the system.
In other words, once again science has triumphed by showing that natural mechanisms exist which can explain the observed data.
ID’s contribution? Poof… Non-existent really.
You’re invited to address those issues. (preferably not by obfuscating them to oblivion as PvM is doing, I hardly read any thing he posts as they are mostly mis-directions of my points rather than an attempt at a substantive refutation).
Shame on you Sal… I understand why you have a hard time dealing with degeneracy in nature since it disproves your redundancy arguments but to call this misdirection is just plainly silly.
I am merely exposing the scientific vacuity of Intelligent Design.
Comment by PvM — July 6, 2006 @ 12:57 pm
Matt asks
What is the threshold at which a design becomes “complex”?
This is a great question, Matt. I’ve never seen an ID promoter answer the question in a meaningful way (i.e., in a way that doesn’t lead you down a rabbit hole of other ill-defined terms).
The reason for the failure is obvious to most scientists: ID is nothing but vacuous hand-waving and something is “complex” if every aspect of its history can’t be proven in sufficient detail to convince an ID promoter. Real scientific, huh?
Comment by Susan Percell — July 6, 2006 @ 1:19 pm
Degeneracy
Gerald M. Edelman and Joseph A. Gally Degeneracy and complexity in biological systems PNAS November 20, 2001 vol. 98 no. 24 13763-13768
Edelman et al continue to observe that
structural level is sharpened by comparing design and selection in engineering and evolution, respectively. In engineering systems, logic prevails, and, for fail-safe operation, redundancy is built into design. This is not the case for biological systems.
In other words, biological systems are in many ways very different from ‘designed’ systems.
Indeed, not the least of Dar win’s achievements was to lay the argument by design to rest. But, for obvious economic reasons, design is by far the major component of most technical efforts in modern society. In general, an engineer assumes that interacting components should be as simple as possible, that there are no ‘‘unnecessary’’ or unplanned interactions, that there is an explicit assignment of function or causal efficacy to each part of a working mechanism, and that error correction is met by feedback, modeling, or other paradigms of control theory.
Again, redundancy but not degeneracy is the hallmark of design.
Protection can be af forded by planned redundancy, but adventitious compensation for error is neither expected nor usual. Irrelevancy is avoided from the outset. By contrast, in evolutionary systems, where there is no design, the term ‘‘irrelevant’’ has no a priori meaning. It is possible for any change in a part to contribute to overall function, mutations can prompt compensation, stochastic interactions with the environment can lead to strong selection, often there is no fixed assignment of exclusive responsibility for a given function, and, unlike the engineering case, interactions become increasingly complex. A theoretical analysis (1) suggests that this increase in complexity results not only from selection in rich environments
(which include other species) but also from the prevalence of degeneracy.
In other words, selection and degeneracy all explain information in the genome.
Now I understand that there is much relevant scientific evidence and articles out there for the average ID proponent to digest but these are essential papers for anyone interested in the issue of redundancy and degeneracy, especially if one were to make the argument from analogy.
In the past I have pointed to these papers and many others and still ID activists still seem to be ignoring the evidence.
Sal calls it obfuscation, I call it hard cold evidence… Either way, ID performs pretty poorly in this