Dembski Responds to Rosenhouse

Unsupported assertion, and Texas Sharpshooter fallacy besides.

That assumes the mutations must occur simultaneously, which is an assertion you have have been asked to substantiate.

This points out a particularly silly problem that underlies the entire premise of Behe’s argument in “The Edge”. He finds a particular observation with a certain probability, then he identifies (often wrongly, but that’s a different matter) another process that would produce a result with the same probability. He then assumes that the second accounts for the first.

To illustrate: Suppose I observe that one out of every six marbles I pull out of a bag is red.

I also know that if someone throws a six-sided die the odds of any specific number coming up is also one in six.

I therefore conclude that the reason 1/6 of the marbles in the bag are red is that the color of each marble is determined by a throw of a die.

That would be stupid, wouldn’t it?


@Rumraket , @Faizal_Ali , @Roy , @Mercer
Ya know, you guys might actually have to read the book to understand it. Most of the evolutionary analysis here is based on bad concepts of the arguments. Is that why you’re asking us to summarize it for you?

But watch out, you might find some things in the book interesting: A German Scientist Speaks Out about Intelligent Design - YouTube

This looks good until you realize antibodies can bind to any spot on an antigen that happens to have a sequence that can act as the complement. So the location is somewhat random and there are no further implications. For a protein complex, you need the complement at a “right” location so that the resulting quaternary structure has meaningful catalytic activity. Even if you get complementary binding sites, having an active protein complex is far from given. The requirements for protein complexes are many orders of magnitude more stringent than B cells.

I am basing my arguments on the extensive writing that Behe has himself made about his argument, as well as what @lee_merrill, who supposedly has read the book, is saying about it. So if those two are misrepresenting Behe’s book, there’s not much we can do about it, is it? But I feel pretty confident that Behe is not misrepresenting. Why would he?

That is objectively, empirically wrong:

1 Like

It looks to me like it’s you guys who read something you don’t really think about or assess critically, you then come here quoting it thinking it somehow settles the matter, and then act flabbergasted when it’s picked apart as the pile of rubbish, handwaving, and mere assertion that it is.

So tell me again, why are we supposed to believe it takes 5-6 amino acid substitutions to produce an efficient protein-protein binding site, and why are we supposed to believe only one out of 20 amino acids at each position confers binding? Are you aware that every single one of those assumptions is an outright demonstrable falsehood. I already cited the literature that shows this.

That’s true for any two proteins that bind each other. They really can in principle bind each other anywhere that is sufficiently complementary. And it usually just takes a single mutation for that to happen.

Sorry but now you’re just making additional hurdles up.

First of all you’re assuming that protein-protein binding is only meaningful if it’s a complex of enzymes.
Second, many proteins form complex multimers without being enzymes.
Third, usually enzymes are still enzymes that have substantial catalytic activity even without being part of a complex of multiple proteins, and only some of them are more efficient when they stick together into a larger multi-enzyme complexes.

For others these changes are entirely neutral and seem to have evolved simply as a consequence of chance occurrence of sticky patches on their surface:

[Citation needed]

I invite you to consider the case of the irreducibly complex, gated membrane channel T-URF13 which evolved from non-protein-coding RNA:

1 Like

I’m not asking you or anyone else to summarise Dembski’s book for me. Or Behe’s book. I’ve read enough of Dembski’s (and Behe’s) work to know for myself how bad their arguments are.

I’m not going to take Bechly’s opinion of a book I’ve read over my own, not least because Bechly’s own articles are just as bad.


Don’t you think that you might actually have to learn some biology to understand it, instead of reading books full of falsehoods written to support a particularly bad theology?

You appear to have confused a book with a video.

Can they bind to “spots” on an antigen that have the same shape as host molecules?

Really? Looks like the Texas Sharpshooter fallacy raising its ugly head.

  1. Why quaternary and not tertiary or secondary?
  2. Do all protein complexes have catalytic activity, much less “meaningful” catalytic activity?
  3. Please apply your hypothesis to a muscle myosin and its catalysis of ATP → ADP + Pi. What binding is required?

You’re comparing two different things. Protein complexes are, well, proteins, and B cells are much, much more than proteins.

And how many orders of magnitude, exactly? To me as an actual biologist, it appears that you’re just making this up as you go.


The German scientist in question is Gunter Bechly, which should come as no surprise to anyone here.


Why can’t you discuss it?

Behe claims that the chances of this occurring are prohibitively high. Antibodies and your own post demonstrate that this isn’t true.

How many “right” combinations are there?

This is the question that ID proponents stumble over. They use target thinking and commit the Sharpshooter fallacy. They focus on the protein complexes that do evolve and then claim that these are the only right combinations. They then calculate the probability of that specific amino acid sequence and that specific protein-protein interaction and pretend that the high improbability of that specific feature evolving is too high. What they ignore is all of the other possibilities that didn’t evolve.


No, he quotes a study by Greg Winter where 10-100 million sites needed to be provided in order to give a match for a new site.

That’s partly the study I mentioned above, and partly reasoning about protein shape space. It’s several pages long.

Speaking of Winter’s work, Behe writes “In all of these experiments, mutations were deliberately confined to a coherent patch of amino acids that were close to each other on the surface of the protein, to make as many novel, binding-site-sized regions as possible. If the workers had not deliberately directed the changes to a coherent patch on the protein’s surface, most changes would be scattered, unable to effectively interact.” (p. 132)

Well, yes, we’re looking for the probability of a given binding site to develop, that is a probability of interest, and a reasonable objective.

Because “that would be an event of approximately the right frequency.”

And Behe acknowledges that: “between ten and a hundred million binding sites have to be searched in a shape space library to find one that will bind with a modest affinity to a second protein.” (p. 273)

“12.As discussed in Chapter 7, there are different kinds of mutations—deletions, duplications, and so on. But point mutation represents the conceptually simplest, most straightforward route. This calculation uses consensus values for important variables. One could certainly imagine other scenarios for making a new protein-binding site, for example by first invoking gene duplication and then point mutation. But those are either unlikely to help much (Behe, M. J., and Snoke, D. W. 2004. Simulating evolution by gene duplication of protein features that require multiple amino acid residues. Protein Sci. 13:2651–64) or likely to involve special circumstances that amount to a Just-So story. All alternative scenarios would have to confront the fact that no new binding sites have turned up in the best-studied evolutionary cases of malaria and HIV, as described later in the text.”


13.Even though protein-binding sites often involve a score of amino acids on each of the partners, experiments have shown that only a fraction of those are important for having the two proteins stick to each other. (For example, see Braden, B. C., and Poljak, R. J. 1995. Structural features of the reactions between antibodies and protein antigens. FASEB J. 9:9–16; Lo Conte, L., Chothia, C., and Janin, J. 1999. The atomic structure of protein-protein recognition sites. J. Mol. Biol. 285:2177–98; Ma, B., Elkayam, T., Wolfson, H., and Nussinov, R. 2003. Protein-protein interactions: structurally conserved residues distinguish between binding sites and exposed protein surfaces. Proc. Natl. Acad. Sci. USA 100:5772–77.) In terms of the swimming pool analogy, the five or six residues represent bumps and magnets that are aligned very nicely; if enough are aligned, then it doesn’t matter so much if other features aren’t aligned, as long as they don’t actively block the surfaces from coming together."

Three or four of the mutations are supposed to occur simultaneously, since they are concluded to be deleterious: “So let’s suppose that of the five or six changes that have to happen to a protein to make a new binding site, a third of them are neutral. They could occur before the other key mutations, as a separate step, without harm. Although finding the right neutral changes would itself be an improbable step, we’ll again err on the conservative side and discount the average number of neutral mutations from the average number of total necessary changes. That leaves three or four amino acid changes that might cause trouble if they occur singly. For the Darwinian step in question, they must occur together.” (Edge of Evolution, p. 134)

But protein structure is the right area to investigate when you’re trying to change a protein to have a new binding site.

No, he just made up a fictitious scenario in which that is the case. If there really was a trait that required 3 or 4 specific mutations occurring simultaneously in a single individual then of course that is unlikely. Jet packs on bacteria are also unlikely. So what has either of those obvious facts to do with evolution?

And your amps go to 11. OK, and…?

This one:

No. We’re looking for the probability of binding that elicits an adaptive response. Not binding through some pre-specified target sequence, because that assumes only that one particular sequence produces effective binding. Which is just false.

That doesn’t answer my questions at all. All you’re saying is that Behe is making those assumptions to support a third assumption, which was made in the first place by target thinking about a particular binding site.
So he’s assuming that in order for that binding site to have evolved, the ancestor would have to have begun 5-6 residues away from the descendant, and only by evolving one particular residue out of 19 others possible, at each of those 5-6 positions, simultaneously, could that binding site have evolved. And all intermediate stages would either be neutral or deleterious, because no other residues at those 5-6 positions would have produced a level of binding visible to selection.

It’s just imaginary hurdles all the way down. Target thinking with imaginary restrictive barriers.

No, that statement does not constitute acknowledgement that binding isn’t all or nothing, which is directly implied by his assumption that only one out of 20 possible amino acids at 5-6 positions produces effective binding. Behe there assumes that literally all other possible combinations of amino acids at those 5-6 sites produces no degree of binding that is visible to selection. That is an outright assumption that all intermediate stages have no effective binding affinity.

Please allow yourself to think about what he’s saying and what follows from his assumptions.


It’s not a reasonable objective, since reality doesn’t work that way and so any conclusion Behe reaches will be meaningless.

That’s why I wrote this:

Why did you delete it when responding?


No, Behe is reinforcing that false claim. He’s treating binding as binary. It isn’t.

What exactly does Behe’s citing of malaria have to do with that? Responding by merely copying Behe’s words constitutes an admission that you don’t know.

Hard to know if that is what Behe thinks he is saying there. Maybe he is saying that, yes, binding is not all or nothing. But even configurations that have a small degree of non-specific binding occur once in every 10-100 million.

I believe even some of Behe’s best informed critics make the error of presuming that Behe actually has a coherent argument against which to argue. He doesn’t. So one is left to disputing each error of fact and logical fallacy one at a time.

This works to his favour, as he can always pull out some other random assertion that must be refuted. The typical member of his intended audience then interprets this as an ongoing debate, rather than as one guy getting repeatedly called out on the many ridiculous errors he keeps making.


More to the point, what does that have to do with my comment to which this was intended as a response? Absolutely nothing, as far as I can see. But those amps definitely go to 11.

No, I believe it is these papers he is referring to (I couldn’t find the second one via Google): (Nissim, A., Hoogenboom, H. R., Tomlinson, I. M., Flynn, G., Midgley, C., Lane, D., and Winter, G. 1994. Antibody fragments from a “single pot” phage display library as immunochemical reagents. EMBO J. 13:692–98).

Dissociation constant on the order of nanomolar (Griffiths, A. D., Williams, S. C., Hartley, O., Tomlinson, I. M., Waterhouse, P., Crosby, W. L., Kontermann, R. E., Jones, P. T., Low, N. M., Allison, T. J., and Winter, G. 1994. Isolation of high affinity human antibodies directly from large synthetic repertoires. EMBO J. 13:3245–60).

That would be the probability of any new binding site, whereas we’re looking for the probability of some new binding site, a site, one site.

Well, about 1 in 10 to 100 million sequences produce a modest bond, according to Behe’s references.

Rather, no other residues at those positions would have produced a modest binding.

I agree, if “all other possible combinations” means “all other combinations in the pool of 10 to 100 million”…

It’s that in all the chances for malaria to develop a new binding site, none that we know of, have developed.

“This points out a particularly silly problem that underlies the entire premise of Behe’s argument in ‘The Edge’. He finds a particular observation with a certain probability, then he identifies (often wrongly, but that’s a different matter) another process that would produce a result with the same probability. He then assumes that the second accounts for the first.”

So which probabilities did you mean? I seem to have misunderstood your point.

Of course you have.

Read again the part of my comment that you didn’t quote:

Is the conclusion drawn in that example valid? Why or why not? Just stick with this for now, and forget about trying to figure out how it applies to Behe.