|
Author
|
Topic: The Evolution of the Bacterial Flagellum
|
William A. Dembski
Member
Member # 7
|
posted 12. November 2003 01:39
On the Talk Reason website, Nicholas Matzke claims to provide a detailed Darwinian model for the origin of the bacterial flagellum -- go here. For my response to Matzke's article go here.
On October 11, 2003, the Talk Reason website posted an article by Nicholas Matzke titled "Evolution in (Brownian) Space: A Model for the Origin of the Bacterial Flagellum" (http://www.talkreason.org/articles/flagellum.cfm). Talk Reason advertises itself as a website that "presents a collection of articles which aim to defend genuine science from numerous attempts by the new crop of creationists to replace it with theistic pseudo-science under various disguises and names." The most obvious target here is intelligent design. Indeed, Matzke's article attempts to rebut one of the main challenges that intelligent design has raised against Darwinian evolution, namely, how to explain the emergence of irreducibly complex biochemical machines like the bacterial flagellum.
[ 12. November 2003, 08:25: Message edited by: Moderator ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 12. November 2003 04:57
There seem to be two parts to the response: various ad hominem-style comments (perhaps they are not quite outright attacks), and a serious consideration of the model that's being proposed.
While being frustrated with particularly vocal and occasionally pseudonymous critics is understandable, I'd suggest that less space be devoted to the personal side of things, especially since the original article was quite focused on the topic at hand and not the participants.
If personal comments were to be made, perhaps it would suffice to say, "He's a geographer, and I am a mathematician and philosopher, so by training, neither of us has qualifications to evaluate flagellum evolution (or structure or function)."
Or, maybe, leave background out of it. Biology is broad, but not terribly deep. An interested party without a Ph.D. can become one of the world's experts in any particular area they choose (although I wouldn't necessarily trust them to actually design experiments, as a good deal of practical experience helps there).
Okay, let's get to the substantive stuff.
I'm going to agree with both Dembski and Matzke.
Matzke's model is commendable for its depth and extensive literature support. Because of this, it is difficult to simply dismiss away many of the suggestions, as can be done with all too many evolutionary stories. For example, the references to similarities between Type III systems and the flagellum make it pretty clear that the two are functionally related. In a divide and conquer approach, this is most helpful, as it appears that something like a Type III system would be easier to build from other cellular components than the full flagellum. (Other, more detailed hypotheses are also well supported, but I won't mention those as they require too much background to understand in the space I want to use here.) Because of this, the piece's place on Talk Reason is well deserved.
However, I also agree with Dembski's main criticism that the model presented isn't detailed and testable. As Dembski says
quote:
We don't have the intermediates that Matzke posits nor the ancestral type III secretion system. We don't know what they look like. We don't have their precise biochemical specification. We don't know if the intermediate systems that Matzke's model hypothesizes would work. We have no way of determining how easy or hard it is for the Darwinian mechanism to bridge the steps in Matzke's model.
Also, the various models proposed don't seem to come with clear predictions made by each that would, at least in theory, allow for a test of the hypothesis given appropriate data. An interested reader could probably begin to fashion something, but it's too generous to say that the models are explicitly testable as they stand.
But are Dembski's demands upon a model reasonable? quote:
Matzke, throughout his article, invokes gene duplications and mutations at key points where the Darwinian mechanism is supposed to effect transitions that are reasonably probable. But what gene exactly is being duplicated? And what locus on a gene is being mutated?
Well, I ask: what good would it do us if we did know the exact locus? The "protein folding problem" is, indeed, a problem. Given an amino acid sequence--even given an existing protein and a point mutation change--we cannot predict the function of the new protein with much success. And the cross-reactivity of most proteins is a complete mystery (except to people who have pulled down way too much junk in a yeast two-hybrid screen).
A major reason why evolutionary hypotheses are not at this level of detail is because nobody knows what the biological consequences are of specified changes at this level!
So what is the point of coming up with flagellum evolution stories at all? It is to guide us when searching for new evidence (e.g. as bacterial genomes are completed, primitive Type III-like systems will be of significant interest!). It is to provide a sanity check, to make sure that an evolutionary hypothesis is not blatantly ridiculous at least compared to other evolutionary hypotheses. And it is to illustrate the pathways that absolutely must be considered when computing the improbability of the flagellum.
So Matzke's work, in effect, puts the ball back in the court of the ID camp, and Dembski's reply does nothing offset this. (Nor would one expect it to, given the length, content, and rapidity of reply.)
"Evolution in (Brownian) Space" provides a model that proposes plausible functional intermediates, and this model circumvents the reasoning used to estimate a low probability for the flagellum in all published ID work. Did the flagellum actually evolve this way? Almost certainly not. Humans aren't very good, in general, at predicting biology. However, in order to make a strong case that the flagellum is too improbable to have evolved, this case must be ruled out.
One might object--do we have to do a whole new set of probability calculations every time someone comes up with a fanciful new hypothesis?
Yes. You do. That is what it means to rule out all chance hypotheses and regularities.
Either rule out every vaguely plausible model explicitly, or make a general claim that can rule out every possible pathway, explicitly, mathematically, and in a way supported by experiment.
It's a tall order. The converse--demonstrating the actual evolutionary pathway of the flagellum--is a tall order, too. I don't expect the matter to be settled one way or the other right away. As a sociological matter, this is more troubling for ID as a competing hypothesis than evolution as the dominant hypothesis; ID needs the wins, while evolution doesn't.
[ 12. November 2003, 05:03: Message edited by: Rex Kerr ]
IP: Logged
|
|
Evan
Member
Member # 164
|
posted 12. November 2003 08:02
Rex, this is a good overall response to Dr. Dembski’s response, so much of what I say will be repetitive - still, I would like to add my comments.
Response to the personal remarks
1) I agree that the amount of space Dembski spends on the personal issues, and his tone, deflect from the substance of his response. The fact that Matzke has been posting on the internet about these issues under various pseudonyms for seven years or so is irrelevant.
First, Matzke is interested, and he likes to study the issues - this is a good thing. As an educator, I find calling Matzke a “PubMed junkie” a rather demeaning insult - I wish I had more students who were broadly curious outside of their specific field, knew how to use the internet to expand their knowledge, and were interested in synthesis and analysis of what they read.
Secondly, his use of various internet pseudonyms is not important. One of the most prominent and scientifically specific proponents of ID, with a particular interest in the flagellum, posts exclusively under the pseudonym “Mike Gene” at ARN and here at ISCID, has a website under the same name, and has credentials that are totally unknown. But, as “Mike” argues, it is his ideas and his work that should be taken be taken into account, not his credentials, and I think the same applies to Matzke.
2) In respect to biology, Matzke is less of an amateur than Dembski is. Enough said.
3) Dembski writes that Matzke wrote under the name “Tamzek” “until a year or two ago, when he blew his cover by publicly attacking Jonathan Wells at UCSD.”
I know nothing about this incident, but I am inclined to think that the phrases “blew his cover” and “attacking” are excessive.
Matzke has written a long article in response to Wells’ book that he [Matzke] entitled “Icons of Obfuscation,” which I found at www.talkdesign.org, another site about ID. The author is listed as Nic Tamzek, but that is endnoted to the remark that “[1] Nic Tamzek is the pen name of Nicholas Matzke. http://www.geog.ucsb.edu/~matzke/” It is unlikely that Matzke saw his use a pseudonym on the internet as a “cover” for anything - it is more likely that after he wrote and got support for such an extensive review he felt comfortable and justified in starting to attach his real name to his work. (Perhaps “Mike Gene” will also feel like doing this some time.)
And last, I wonder about the use of the phrase “attacked Wells.” In science, ideas are “attacked” and that is considered acceptable and expected - it is through the forceful challenging of new ideas that the fruitful are winnowed from the chaff. Personal attacks, on the other hand (and I know these have been aimed at Wells by others), are not acceptable even though they do regrettably occur at times.
I have no idea whether Matzke attacked Wells’ ideas or Wells’ himself at UCSD, although I am going to assume, given the substance of Matzke’s article about Wells’ book, that it was Wells’ ideas that were attacked. If this is so, then I think that there is nothing wrong with this, and that perhaps Dembski’s phrase “publicly attacking Wells” is misleading.
I am sorry to spend this much time on the personal issues raised in Dembski’s reply, but unfortunately Dembski chose to start with these issues and so I think they need to be addressed. If “attacking” people themselves as opposed to their ideas is not a good thing, then perhaps Dr. Dembski should model this behavior and forgo such attacks himself.
Response to the substantive remarks
Dembski’s key statement here is that to refute the ID arguments about the flagellum, Matzke “ must make good on his claim to provide a detailed, testable, step-by-step Darwinian model of how the bacterial flagellum could have originated.”
This, however, is setting the bar way too high, both in terms of what can be realistically expected about our scientific knowledge at this time of a foundational event that happened 3.5 billions years ago; and, more importantly, this is setting the bar way too high in respect to what is need to pose a significant challenge to ID, and specifically to Dembski’s Explanatory Filter (EF).
Let me concentrate on this second issue. From the beginning of my participation at ARN, my main interest has been in how the EF might in fact be empirically employed. Central to its premise is that probability calculations must be calculated for an event to see if it falls below some universal lower bound and thus qualifies as design. In order to do this, methods for calculating the probability of a long series of biological and chemical events involving numerous organisms (or pro to-organisms) over multiple generations, need to be developed. Nothing like this has been offered yet, so at this point ID theory can do nothing but contemplate how such calculations might play out if such tools were developed.
What Matzke has done is show how current scientific knowledge leads us in a direction by which we might ascertain the steps leading to the development of the flagellum. This is good science - it leads to further questions and further hypotheses. To complain that his paper does not provide a challenge to ID because he does not yet provide a “detailed, testable, step-by-step Darwinian model of how the bacterial flagellum could have originated” seems to me to indicate a (pardon what may be an excessive expression here) profoundly unrealistic sense of what science can do.
The only alternative scenario offered by ID that I know of occurs in chapter 5 of Dembski’s book “No Free Lunch,” and that calculation assumes that the flagellum assembled in one fell swoop with no intermediate steps from more or less isolated components into a flagellum. This is certainly a less “testable” hypothesis than Matzke’s, I would think.
So of course, Matzke’s work is in the “subjunctive” mode - all hypotheses are. But it is empirically based on known knowledge, it leads to experimental questions by which further data might either strengthen it or weaken it, and it’s a pretty good product for a geographer who is an amateur biologist with a fondness for drawing on the tremendous informational resources of the internet.
Work like Matzke’s present a challenge to Dembski that I believe he is avoiding - rather than responding negatively to Matzke’s personal credentials and to unrealistic expectations to science, Dembski should be working to positively advance his own theory by thinking about the central problem: what empirical data and realistic methods of calculation can be employed to in fact show that a hypothesized series of steps is in fact improbable enough to conclude design?
As I have often stated, a first step in doing this is to figure out how to do such calculations in actual situations in which we do know the steps involved in order to show that we can reliably conclude that something is not designed. Until the ID movement in general, and Dembski as the leader of this aspect of it, takes this problem seriously, work like Matzke’s will present a serious challenge to ID. Dembski does not further the cause by the type of deflective response he has addressed to Matzke.
Conclusion
Returning to Dembski’s key statement concerning a “ detailed, testable, step-by-step Darwinian model of how the bacterial flagellum could have originated.”
I continue to be bothered by a statement Dembski made here at ISCID a number of months ago in a response to an article by Ken Miller.
Here Dembski wrote
quote:
** Even if the trajectories are continuous, it still remains to be shown that they are Darwinian. I'm entirely comfortable with the evolution of the immune system, for instance. In fact, I'm comfortable with full common descent. My argument, always, is that evolutionary pathways to irreducibly complex systems, even if they exist, are non-Darwinian.
This remark seems to indicate that even if a “detailed, testable, step-by-step Darwinian model” of anything were offered, there would be no amount of detail that would convince Dembski that Darwinian processes were exclusively involved.
This is a very high bar. If this is really what Dembski thinks, then his response to continued research on any purported “irreducibly complex” biological feature will be a repeat of what he has said to Matzke - this isn’t detailed enough.
And, in the absence of tools to actually measure and compute the probability of biological events, Dembski argument will always be in principle potentially true.
And yet, as should be obvious, things that are always potentially true are in fact not very useful, and certainly not scientific. Dembski seems to be saying in the quote from ISCID, and confirming in his response to Matzke, that in fact no amount of science would be sufficient to overthrow his arguments.
So, to repeat myself for the umpteenth time, until the ID movement starts to work on developing the tools to compute the probabilities upon which the EF is based, ID will remain in the realm of philosophical speculation with no empirical content.
But, of course, what do I know - I’m just a amateur and pseudonominous internet junkie. ![[Smile]](smile.gif)
[ 12. November 2003, 09:16: Message edited by: Evan ]
IP: Logged
|
|
Matthew J. Brauer
Member
Member # 819
|
posted 12. November 2003 09:46
Wow.
I'd have thought that a serious and well-reasoned attempt to answer a specific biological question would, whether successful or not, be met with something other than vitriol and sarcasm. The advocates of ID would preusmably like their ideas to be taken seriously by biologists. Nick Matzke is a such a biologist, and he has taken the flagellum story seriously.
Also: many prople have a substantial internet presence because they engage in discussion with people on bulletin boards. They put their ideas into the public sphere for comment and criticism. They stay engaged, and participate in the resulting dialog. This does not make them, as Dr. Dembski would imply, "trolls" or "junkies". Rather, it can be an indication that they have some respect for the opinions and views of the people they're communicating with.
[edit for spelling]
[ 12. November 2003, 14:00: Message edited by: Matthew J. Brauer ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 12. November 2003 10:55
I highly recommend that readers watch my thread Does intelligence imply "motive"? (which is somewhat mis-named), in the future, with respect to this issue. Also of special interest is Dr. Dembski's post to ARN, which I copied into that thread on about page 5. NOW I must warn that most significant mathematical presentation is coming along slowly as I don't have time for a few weeks to completely finish presentation and some minor details need work. So don't expect a lot of material right away, but it will appear slowly.
I have already commented on some of the basic conclusions of the mathematical analysis of the Explanatory Filter's "reliability".
The point of relevance here is that one could view the flagellum issue in two levels:
1) Are specific models of how the flagellum developed well supported?
2) Is intelligent design inferred by information gathered on the flagellum, with respect to the events of cause of specific flagellum structures?
Now my thread deals in the reliability of the explanatory filter, and of different inference procedures that may relate or vary from that. But one thing that is shows (and will show with detailed mathematics) is the low level of reliability of the explanatory filter--precisely in the kind of case that is exemplified by the flagellum. Thus my thread is addressing number "2)" question.
What Matzke's article does most clearly is show that there are alternative natural explanations of the flagellum that can in the worst be considered as "missing distributions" with respect to applying the explanatory filter. I understand Dr. Dembski's objection based on use of most current information (and I quoted a version of that objection in that thread, as mentioned).
But the supersensitivity of the explanatory filter to "missing distributions" is at heart in its usefulness in inferring "intelligent design" in the case of the flagellum. Behe's inference of ID is essentially a version of the EF. Without that, without the explanatory filter, all that is present is some criticism of models, and no inference that relates to ID.
Dr. Dembski claims:
quote:
But to do so he must make good on his claim to provide a detailed, testable, step-by-step Darwinian model of how the bacterial flagellum could have originated.
But of course one actually only needs to show that the probability of natural pathway for the event is not below that extremely low probability cutoff "alpha" of the GCE procedure. Dr. Dembski could claim that such probability is only found in what might be classified as "missing distributions" and is not directly calculable. My thread shows and will show how these missing distributions are amplified by the "ProbRes" factors M and N, and coincidence needs to be controlled for missing distributions just as it needs to be controlled for explicitly calculated distributions. That is why the cutoff "alpha" is usually chosen so small, such as the UPB of 10^-150.
With all the calls for detailed, testable, step by step models, such as:
quote:
Question: In what sense is this model detailed, testable, and step-by-step?
Where is there any detail on the intelligent design model?
What Matzke provides is the start on (if not considerable detail of) theory of how the evolutionary model would work. (Even if not all the details are worked out yet.) Because theory is present, that method has greater controls for coincidence. It is the control for coincidence that is lacking (though attempted) in the “design inference”. Please supply some detail of ID theory of the events leading to the flagellum.
By the way, if any ammunition is needed about what "league" I'm in as a "mathematician", I only have a master's degree in EE, and a BA in physics. Just thought I'd throw that out, since I use a pen name. Dr. Dembski might want to challenge my expertise as only that of an "amateur".
[ 12. November 2003, 12:31: Message edited by: gedanken ]
IP: Logged
|
|
Grape Ape
Member
Member # 399
|
posted 12. November 2003 12:29
Evan writes:
quote:
3) Dembski writes that Matzke wrote under the name “Tamzek” “until a year or two ago, when he blew his cover by publicly attacking Jonathan Wells at UCSD.”
I know nothing about this incident, but I am inclined to think that the phrases “blew his cover” and “attacking” are excessive.
Matzke has written a long article in response to Wells’ book that he [Matzke] entitled “Icons of Obfuscation,” which I found at www.talkdesign.org, another site about ID. The author is listed as Nic Tamzek, but that is endnoted to the remark that “[1] Nic Tamzek is the pen name of Nicholas Matzke. http://www.geog.ucsb.edu/~matzke/” It is unlikely that Matzke saw his use a pseudonym on the internet as a “cover” for anything - it is more likely that after he wrote and got support for such an extensive review he felt comfortable and justified in starting to attach his real name to his work. (Perhaps “Mike Gene” will also feel like doing this some time.)
It was actually Casey Luskin who outted Nic. He did so in an article that was a response to a short piece that Nic wrote on Icons, which was passed out at a Wells' talk at UCSB. (The episode is achived here on ARN.) Dembski is simply wrong to imply that this piece is directly responsible for Nic's outting. The party responsible is Casey Luskin, and the "attack" on Wells is only relevant insofar as it was this piece that Luskin was responding to. Luskin's method of discovering Nic's identity had to do with personal emails between to two, and had absolutely no connection to Wells' UCSB talk.
Furthermore, the phrase "blew his cover" implies that Nic had a vested interest in keeping his identity secret (which implies some sort of "cloak and dagger" mentality on Nic's part.) Had he been seriously concerned about this, he certainly would have picked a different pseudonym. You can see from the above link that Nic was disappointed at being outted, but moreso due to a breach of trust between himself and Luskin rather than having his identity revealed. The fact that the uses his real name on subsequent documents indicates that he's comfortable with having his identity known, yet wants to keep the connection to the pseudonym under which he's already written a lot of relevant material.
Finally, I agree with everyone who's pointed out the total irrelevancy of Nic's pseudonym(s), and the complete waste of space and distraction that writing about them entails. But if Dembski wishes to bring them up anyway, he should at least get the facts straight.
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 12. November 2003 12:40
I can confirm Grape Ape's points about the "outing" of Niiicholas. I was knowledgeable of the details and the negotiations and the apologies by Luskin, etc., as they happened. This is most certainly correct. Dr. Dembski should amend his presentation in that regard.
Wells at most may have had some responsibility for further promotion, of re-using the mistakes already apologized for.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 12. November 2003 18:52
I, for one, really liked Nic’s article. From my preliminary reading, I am struck by the way his style is vastly improved. He doesn’t seem to overstate his case and acknowledges points that are weak. It is a more detailed expansion of the EFM hypothesis that I address on my web page and has given me a few more things to think about. I’m not going to comment until I give it a better read and think more about some of the interesting points that Nic brings up. And that may take some time. Besides, RBH mentioned it was only a rough draft. Perhaps Nic is planning on submitting this to a scientific journal? In that case, it might be more appropriate to move a response into that venue (if such things are possible).
Since my name came up, I suppose I should address this:
quote:
Secondly, his use of various internet pseudonyms is not important. One of the most prominent and scientifically specific proponents of ID, with a particular interest in the flagellum, posts exclusively under the pseudonym “Mike Gene” at ARN and here at ISCID, has a website under the same name, and has credentials that are totally unknown. But, as “Mike” argues, it is his ideas and his work that should be taken be taken into account, not his credentials, and I think the same applies to Matzke.
Yes, I have argued from the beginning that what matters are the arguments, not the credentials. And we can extend this whole argument as it relates to the scientific literature, as pointing to something published in the literature is in the same realm as appealing to credentials. Thus, what ultimately matters is not who came up with the argument, or where the argument is presented, but the argument itself. For example, it’s interesting how we have been having this debate, over the last few years, in cyberspace and not in the literature.
Finally, there are big issues here that need to be discussed. Is the objective of Nic’s article to come up with a proposal for how the flagellum could have evolved? Or is it a hypothesis for how the flagellum did in fact evolve?
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 12. November 2003 19:07
Let's move this thread beyond pseudonymns, character assasinations, etc. I let the conversation go on for a few posts because it wouldn't be fair to give Dembski the only shot at rhetoric.
However, I think this thread would be a lot more productive if we focused on the *substantial* issues, especially the new ideas that Nic has presented in his paper, and the places where Dembski addresses these ideas.
So, let us move on. Regardless of what we think of the "preamble."
[ 12. November 2003, 19:09: Message edited by: Moderator ]
IP: Logged
|
|
Matthew J. Brauer
Member
Member # 819
|
posted 13. November 2003 11:28
A minor quibble on terminology in Dembski's response. He asks: "what locus on a gene is being mutated?"
This is a non-sequitur. The term "locus" is commonly spoken of as if it were synonymous with gene. However, a "locus" is a physical location on a chromosome that corresponds to a gene or some other genetic feature. Loci are not "on" or "in" genes, though, by any usage of the term.
I suspect that Dembski was intending to ask: "what nucleotide or amino acid is being mutated?" But this is not a meaningful question in the context of Matzke's model and it makes me think that Dembski is not yet engaging in "population thinking" (sensu Mayr) when he's analyzing these models. What's relevant in population genetics models is not mutation events per se, but rather substitution or fixation events.
This distinction may appear subtle, but it is of great importance. Discussing protein evolution in terms of "mutations" reinforces an image of single changes being made in stepwise, deterministic fashion. It focuses attention on the single molecule, rather than on the population, which is where the evolution actually happens. Finally, such terminology implicitly argues that a stepwise, deterministic model of arbitrary precision is theoretically achievable.
Such precision in stochastic population models is neither possible nor desirable. In discussing diffusion, a time-ordered list of the position vectors that a molecule takes as it moves across the room would not be useful, even if it were obtainable. Instead what is needed is a general law of diffusion.
Dembski's asking Matzke to account for "individual mutations" is akin to asking for the precise path of a molecule as it diffuses through space. Matzke's intention is to show that the variation arising in populations via well-known molecular genetic mechanisms is sufficient to encompass the intermediates in a potential evolutionary trajectory. That should be sufficient.
Finally, Dembski seems to think that "design" should be the null hyothesis for the origin of the flagellum. If this is the case, he should (like the good Fisherian he claims to be) set, a priori , the level of significance that he would accept as rejecting that null. What is Dembski's alpha for the design of the flagellum? In asking for arbitrary precision in Matzke's model, he is suggesting that it is 0.0, that is, that there is no amount of evidence that could reject this null hypothesis.
If that's the case, then Matzke and others are wasting their valuable time by engaging in this discussion. We all deserve the favor of a reply, and an (approximate!) value for Dembski's flagellar alpha, so we can have a better idea of whether or not his model can ever be rejected.
IP: Logged
|
|
Ryan Huxley
Member
Member # 366
|
posted 13. November 2003 19:09
I won't be able to add much, but just wanted to point out or question a few things:
1. Nic "attacked" Wells UCSD (note: D, not B, as in UCSB) lecture. I was present at this event and saw Nic's abridged version of his article on ICONS. Therefore, Dembski is correct to note UCSD (and I would agree with "attack" based on some of the ad-homenim attacks made in the article). Apologies to moderator for this one.
2. Being generally uninformed from a biological standpoint, the biggest problem I see for the Type III secretory system being the precursor to the flagellum is that the Type III is used for pumping toxins into eukaryotes, which weren't in existence for quite sometime after bacteria appeared (at least according to the fossil record). So, what must have happened for the basic premise Nic proposes (I've not read it all) is that this pump system evolved to pump into nothing (which therefore makes it essentially invisible to NS - there's nothing to pump into to further the bacteria colony), and then the flagellum evolved after that. Doesn't this seem a bit odd? Maybe I'm missing something?
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 13. November 2003 20:39
Ryan Huxley wrote quote:
2. Being generally uninformed from a biological standpoint, the biggest problem I see for the Type III secretory system being the precursor to the flagellum is that the Type III is used for pumping toxins into eukaryotes, which weren't in existence for quite sometime after bacteria appeared (at least according to the fossil record). So, what must have happened for the basic premise Nic proposes (I've not read it all) is that this pump system evolved to pump into nothing (which therefore makes it essentially invisible to NS - there's nothing to pump into to further the bacteria colony), and then the flagellum evolved after that. Doesn't this seem a bit odd? Maybe I'm missing something?
Read section 3.2 for a discussion of this question.
Brauer wrote quote:
Finally, Dembski seems to think that "design" should be the null hyothesis for the origin of the flagellum.
I agree that's the impression Desmbski very clearly gives. Let it be recorded that's a reversal from the logic of the Explanatory Filter, where chance+regularity is the null and design the alternative in a Fisherian hypothesis-testing framework.
RBH
[ 13. November 2003, 20:45: Message edited by: RBH ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 13. November 2003 20:41
Ryan,
You are missing something. The type III secretory system (TTSS) most likely evolved from a pre-existing flagellum. Nic does not envision the putative ancestral TTSS injecting toxins into other cells. He proposes what I call the EFM Hypothesis – Export, then Filament, then Motility. The ancestral function Nic envisions is simply something involved with the use of extracellular filaments, such as attachment.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 13. November 2003 21:25
I'd like to make a few remarks on "testability." In his response to Matzke, Dembski wrote quote:
That leaves testability. What are we to make of the repeated claim throughout Matzke's article that his model is testable? In claiming that his model is testable, Matzke confuses a precondition for his model with actual evidence for it. Fully two thirds of the 20 pages Matzke devotes to his model are concerned with establishing homologues for structures embedded in the bacterial flagellum. In several cases the homologues are completely absent. Thus, according to Matzke, his model is testable because if the homologues are found, they would confirm his model. But this argument evinces a deep confusion. If the homologues never existed in nature, then the systems that needed to be co-opted for the evolution of the flagellum never existed either, and so the flagellum could not have evolved in Darwinian fashion. In other words, if the homologues never existed, Matzke's model is dead in the water. Matzke's model therefore presupposes the existence of these homologues. They are a precondition for the model, and any evaluation of the model's adequacy proceeds on the assumption of those homologues. That's why I was willing earlier to grant their existence. To actually test Matzke's model requires being able to evaluate the likelihood of transitioning from one step in the model to the next. Because the systems at these various steps are in fact unspecified (they are hypothetical; they do not, as far as we know, currently exist in nature; they are not available in any laboratory; and researchers for now have no experimental procedures for generating them in the laboratory), there is no way to carry out this evaluation. It follows that Matzke's model is in fact untestable.
What I read this to mean is that Matzke identifies some homologues that have not yet been observed that are necessary to his model. That identification of not-yet-observed homologues constitutes a prediction, namely that the identified homologues will be found. If they are in fact found, that would constitute some degree of corroboration for the model; it would increase the model's plausibility. Failure to find such homologues after some reasonable search over appropriate representatives of the array of bacterial structures would reduce the plausibility of the model.
Dembski objects to this constituting a test of the model on the grounds that even if the homologues exist, Matzke still hasn't established some sort of likelihood of the transitions from step to step in the model.
Now I may be naive, but I've pretty much operated on the assumption that at rock bottom, "testability" means predicting observations not yet made, where the source of the prediction (hypothesis, model, theory) can accommodate some yet-to-be-made observations and not others. Now, Matzke's model makes some pretty firm predictions of that sort. It can accommodate - indeed requires - that certain homologues be found. It cannot accommodate the failure to find them. So if after some search over an appropriately wide array of candidates, those homologues continue to be assumed rather than observed, Matzke's model becomes less and less plausible and in the end should be abandoned as a guide for research. In the end, no news is bad news for Matzke's model.
Dembski's focus is on a quite different issue, namely his fixation on calculating probabilities of transitions (though given that here he uses the term "likelihood" one wonders if he's abandoning Fisher!). Clearly as a first step one needs to be assured that the various transitions are not implausible given the assumptions of the model. Matzke has done that. What is also clear is that calculating the probabilities of each and every mutational step in sequence is both impossible and irrelevant, as Brauer has pointed out. One can do that no more than one can watch the latter stages of an automobile that apparently bounced down a hill and subsequently reconstruct the exact moment-by-moment path of the automobile from the top to the bottom. One could make the hypothesis that the car did indeed bounce all the way down the hill and can appeal to various physical laws to explain it. That model would imply predictions of the form "If the rock bounced all the way down the hill, then I should find some marks it left up towards the top of the hill," and go look for those marks.
Failure to find the predicted marks would reduce the plausibility of one's hypothesis that the car bounced all the way down the hill; finding the predicted marks would strengthen one's confidence in the plausibility of the hypothesis. One can look for marks on the hill scattered here and there between the top and the bottom and one can reconstruct a rough path from them, but even 10 seconds after the fact one couldn't precisely reconstruct the exact path the car took. I've routinely done this sort of hypothesis testing in running on heavy rescue calls for 30 years. Given a 4-car accident with substantial secondary impacts off guardrails, trees, and other fixed objects, it's an interesting exercise in hypothesis testing to reconstruct a causal history of the paths and impacts of the several automobiles. Each hypothesized path implies predictions of observations at the accident scene in the light of knowledge of the physics of moving masses, and one goes and looks to see if the predictions are borne out or not by the observed evidence. There's a sort of dialogue between model formation, amendment, or replacement, and the observations of the physical traces left at the scene. Law enforcement is interested in that historical reconstruction for obvious reasons. Rescue workers are interested (and under time pressure) because the reconstruction can inform diagnosis and treatment of victims on the scene.
So in my view, Dembski's take on testability is a little strange.
RBH
[ 13. November 2003, 22:00: Message edited by: RBH ]
IP: Logged
|
|
Claire
Member
Member # 725
|
posted 13. November 2003 21:39
Evan,
Your post is one of the best I have read here in a long time.
Claire
IP: Logged
|
|
|
Printer-friendly view of this topic