I have finally obtained and am reading Michael Behe's book DARWIN'S BLACK BOX. I thought it might be interesting, both for myself and for others, to post not a single overall review, but a series of messages on my reactions as I delve deeper into the book.

Up front, so everyone is aware of it, I'd like to say that I start out disagreeing strongly with Behe's basic idea that evolutionary theory can't explain biochemistry -- or to use his own metaphor, that the "black box" of the cell proves to be a gap that Darwin's process cannot leap over. However, I do try to be open-minded, and I genuinely want to see if Behe's arguments are stronger than the other material I've read makes them sound. With that caveat, herewith the first of my partial reviews . . .

I realize I am not the intended audience for this book, but if I were, Behe would be in trouble from the start. In the preface he gets off on entirely the wrong foot with a statement that:

"evolution means a process whereby life arose from nonliving matter and subsequently developed entirely by natural means. That is the sense that Darwin gave to the word, and the meaning that it holds in the scientific community. And that is the sense in which I use the word evolution throughout this book." -- DARWIN'S BLACK BOX, p. xi, emphasis in original

What's wrong with this? Well, for one thing, Darwin never used the word "evolution" in this sense. To Darwin, evolution was a process that explained the origin of new species from old ones. Darwin never claimed that his theory could explain the origin of life. To my knowledge, no modern biologist makes any such claim either. It appears that Behe is deliberately overstating the reach of evolutionary theory, to make it easier to knock it down later. In the parlance of logic, this is called 'erecting a strawman," and it's a rather serious no-no.

In Chapter 1, Behe lays out his basic thesis. He briefly reviews the history of Darwinism, points out that it does succeed in explaining small-scale evolution quite handily, and adds that he does not question the basic idea of common descent of organisms from more primitive ancestors.

I have no expertise to address the next few pages, on the history of biology in general and biochemistry in particular. What Behe says tallies with what I've read in other sources.

On p. 15, Behe states that while Darwin's theory can adequately explain the small steps of microevolution, it cannot explain macro-evolution, large-step evolution. It appears that Behe is aiming toward the idea that while Darwin's theory can explain adaptive improvements in existing structures, on any scale from whole-organism to intracellular, it cannot explain the origin of entirely new structures. OK, this seems a reasonable line to take; after all, creation of something new is always more difficult than modifying something that already exists. Behe also touches on one of my pet ideas by talking about how the complexity of life appears to be almost fractal: no matter how deep you go, there's always another layer of complexity below you.

Behe spends a couple of pages describing the complexity of vision, and wraps up with a declaration that this level of complexity is utterly beyond the explanatory ability of evolutionary theory. He follows this with a declaration that really raises my hackles: that all other evidence, all other branches of science, are now irrelevant. Fossil evidence doesn't matter. Anatomy doesn't matter. Biogeography doesn't matter. Field studies of selection in action don't matter. Population genetics doesn't matter. If Darwin's theory can't explain the processes of biochemistry, then Darwin's theory is dead. In effect, Behe is very carefully eliminating all evidence from other fields in favor of evolutionary theory, and demanding that it rigidly meet his standards alone, based on his field alone. Inductive arguments that "it explains all of these things, so maybe it can explain biochemistry too" will not be permitted. I find this extremely odd -- after all, if there are gaps in the theory's explanatory powers at any level, there should be gaps in its powers at all levels. It makes no logical sense for a theory to work so well at the macroscopic level and yet fail catastrophically at the microscopic level.

Behe ends Chapter 1 with a candid declaration that his purpose in this book is to demonstrate conclusively that the theory he calls "neo-Darwinism" or the "neo-Darwinian synthesis" cannot adequately account for the molecular level of biology, and therefore that the theory is not true. A challenging goal, indeed. Can he do it? I don't think so, but I've been wrong before.

Back to top

The second part of my running critique of Behe's book DARWIN'S BLACK BOX, concerning Chapter 2, "Nuts and Bolts."

To his credit, Behe makes it fairly clear that the dissent brewing in evolutionary biology through the 20th century was over gradualism, the so-called "neo-Darwinian synthesis," not over Darwin's basic theory. This is unusually perceptive - many people don't see the difference. Unfortunately, as I read further it isn't entirely clear whether Behe really sees the difference himself.

Behe mis-states punctuated equilibrium on p. 27. Specifically, he says it "postulates" stability and rapid evolution. There's no mention of the fact that this pattern is what we should expect to see from modern models of speciation and population genetics. The reader is given the impression that Gould and Eldredge simply made up an explanation for the lack of speciation events in the fossil record.

On p. 28, Behe makes a rather subtle mis-statement. He says that the apparent short period of the Cambrian explosion, perhaps as little as 10 million years, means a lot of trouble for Darwin's idea that complex life evolved gradually from simple life. Problem with this is that no mention is made of the history of multicellular life before the Cambrian, which is substantial though rather sketchy.

The next couple of pages form a "Parade of Quotes" from people allegedly disenchanted with neo-Darwinian theory. The broad implication is that if neo-Darwinian synthesis fails, then Darwin's theory fails. Typical anti-evolution tactic. Disappointing. Behe has apparently missed the point that punc. eq. and other suggested variants are still Darwinism. Stephen Jay Gould, as often, said it best: "It is gradualism that we must reject, not Darwinism."

Behe gives a detailed explanation of bombardier beetle defense mechanism. (Side note: I personally disagree with Behe's description of Dawkins as "the best modern popularizer of Darwinism around." I would apply that description to SJ Gould.) Behe reviews the standard creationist line about the beetle, using Francis Hitching's book THE NECK OF THE GIRAFFE as his source, then gives Dawkins's counterpoint which explains how the bombardier could have evolved from other types of beetle. However, Behe then points out that Dawkins's explanation of how the bombardier beetle's defense mechanism evolved really explains nothing, because it doesn't go into enough detail. "All we can conclude at this point," Behe says, "is that Darwinian evolution might have occurred. If we could analyze the structural details of the beetle down to the last protein and enzyme, and if we could account for all these details with a Darwinian explanation, then we could agree with Dawkins. For now, though, we cannot tell whether the step-by-step accretions of our hypothetical evolutionary stream are single-mutation 'hops' or helicopter rides between distant buttes."

This seems reasonable. To a large degree, he's right - to be sure the Darwinian explanation was plausible, we'd have to analyze it to that level. We haven't. Right now, we can't, due to lack of knowledge.

On the other hand, it's also true that if each stage in the bombardier's evolution can be shown to be no greater a step than that which is spanned by ordinary genetic changes every day, then one may form a strong inductive argument that yes, evolution could account for the bombardier. Unfortunately, since Behe ruled all non-biochemical evidence inadmissible in Chapter 1, this argument will never be heard in his court.

Next, Behe introduces the notion of irreducible complexity. Quote: "An irreducibly complex system cannot be produced directly (that is, by continuously improving the initial function, which continues to work by the same mechanism) by slight, successive modifications of a precursor system, because any precursor to an irreducibly complex system that is missing a part is by definition nonfunctional."

I think there's a logical fallacy hidden in this statement. On the surface, it seems eminently reasonable, but look deeper. Behe is saying something along the lines of "you can't get from the first floor to the second floor of a house by climbing a staircase, because if you take away a section of the second-floor floorboards, you'll fall right back to the first floor." It is possible that the evolutionary precursor to an allegedly irreducibly complex system could have simpler but still functional parts. A two-cylinder engine is still useful for generating power, even though it doesn't do it as well as a four-cylinder engine.

An explanation of what a mutation is and what it does is badly oversimplified. Behe uses the "genes = list of instructions for building an organism" metaphor. This is MUCH too simple. Genes tell how to assemble proteins. Those proteins can have any of a wide variety of functions. They might be used as building blocks for cellular machinery. They may catalyze a chemical reaction in the cell, or perhaps inhibit one. They might simply float around waiting for another component they need to do something. They might activate or deactivate other genes. Organisms are "built" out of the interactions of thousand upon thousands of proteins. This is clearly far more complex than the simplistic "genes = instructions" metaphor Behe is using. I have to wonder why Behe would use an analogy that is so badly flawed. Does he think the flaws are irrelevant? Does he not recognize the flaws himself? Or does he know and not care because the flawed analogy makes his point, where the truth would not? I think any of the three answers may be equally damning.

The famous Mousetrap Analogy. Well-presented and well-reasoned. I would agree that the mousetrap is an irreducibly complex device. I've seen arguments that no, it isn't, but they seem to me to be somewhat specious. Question is not "do irreducibly complex systems exist?" They do. Question is, "do any biological mechanisms fit the definition of an irreducibly complex system?"

The chapter closes with a paragraph that reads oddly like a sidelong sneer at Dawkins for simplifying the evolutionary process in his books. Be reasonable, Dr. Behe: if Dawkins didn't simplify it, how many people would read his books? His intent is to simplify and popularize, so anyone can understand, not to do rigorously accurate science. Maybe he oversimplifies sometimes, but for Behe to accuse him of that seems somewhat hypocritical. People who live in glass houses, et cetera.

Back to top

(Yes, I know this is the second one tonight. I said I'd post these as I read the book, and my reading doesn't follow a set schedule. :-) )

The next few chapters of the book describe various biochemical systems that Behe believes to be "irreducibly complex." Chapter 3 focuses on the cilium and bacterial flagellum. Each is described briefly and simply, tagged as "irreducibly complex," and added to the list of things that Darwinian evolution can't explain.

Behe is good at his chosen topic, make no mistake. His explanations are pretty well written, and each case is examined and summarized with the skill and style of an expert prosecutor. It's easy to see why a lot of people find him convincing. Heck, I'd probably find him convincing, if not for previous experience I've had with exactly the same kind of argument. I've watched lawyers at work, and I've seen other people build the same kind of one-sided, subtly slanted argument. Behe's good at it, as I said, but if you look hard and long enough you can see that he's committing a couple of errors.

One error is that he's only presenting the aspects of the case that support his thesis. This is what a prosecutor is supposed to do, but it still results in the reader getting only one side of the story. For example, a strawman "evolutionist explanation" of the cilium's origin is deftly defined and deftly demolished. It's a strawman because it's the old "tornado in a junkyard" analogy placed at a biochemical level. That's just as fallacious at the biochemical level as it is at the whole-organism level.

Another error is subtle and well-concealed, so much so that I would not be surprised to find that Behe doesn't see it himself. It comes back to the same thing noted previously: the evolutionary precursor to an allegedly irreducibly complex system could have simpler but still functional parts. Behe makes no attempt to show that this is not the case. He appears to assume that evolution works like a kid putting together a set of TinkerToys. It doesn't. In fact, if you think carefully about the way genes are reproduced, and how genes vary, evolution couldn't follow a TinkerToy model.

In 1996, there was a creationist hanging around Compuserve's Dinoforum and Religious Issues who loved to claim that a woodpecker as a whole organism was irreducibly complex, and therefore could not have evolved. (Note: Longtime readers of the newsgroup talk.origins will probably recognize these arguments too, and the person behind them.) His argument disintegrated when someone pointed out that every facet and feature that makes a woodpecker so well adapted for hammering trees can be found in simpler, less evolved versions in other birds that are closely related to woodpeckers. Behe's case for the evolutionary impossibility of the cilium appears to suffer from the same oversight: do relatives of ciliated organisms have "protocilia" organelles?

Another point that Behe never mentions: the vast difference between life at the macroscopic and microscopic levels. His analogies for cilia are all macroscopic: a swimming human, a steamboat paddle, a toy fish. The assumption is that cilia work the same way as these things do. But when one is dealing on a scale of individual atoms and molecules, the relevant properties of water change and those analogies break down. For example, no human could do what a waterstrider insect can do, literally walk on water, but the 'strider can do it because it's too light to break the water's surface tension. This may or may not be relevant to Behe's argument, but if it is I'd like to see it discussed, and if it isn't I'd like to know why.

Chapter 4 concerns blood clotting. In reading this, I have an unfair advantage: I've already read some of the critiques of Behe, so even as I read, I'm comparing what he says to what his critics say. This is especially true where the blood-clotting cascade is concerned - I've read several talk.origins items on this subject, including a talk.origins Post of the Month and Keith Robison's review of the book, so I already know a plausible way that the clotting process might have evolved from a simpler one.

Behe does discuss the fact that many of the proteins in the clotting cascade are similar in amino-acid sequence and in chemical activity - for example, several require vitamin K as a catalyst or activator. This means, he says, that they may have appeared through gene shuffling and gene duplication. But he seems to brush this aside as irrelevant:

"By itself, however, the hypothesis of gene duplication and shuffling says nothing about how any particular protein or protein system was first produced - whether slowly or suddenly, or whether by natural selection or some other mechanism. Remember, a mousetrap spring might in some ways resemble a clock spring, and a crowbar might resemble a mousetrap hammer, but the similarities say nothing about how a mousetrap is produced. In order to claim that a system developed gradually by a Darwinian mechanism a person must show that the function of the system could 'have been formed by numerous successive, slight modifications.'" - DARWIN'S BLACK BOX, p. 90

If you can't figure out the logic in that argument, join the club. I can't make any sense of it either. Gene-shuffling is a way of generating a whole new protein suddenly, all at once. So creating a single new protein is not a matter of "numerous slight, successive modifications." It can be done all in one jump, one mutation. And because proteins can do such a wide variety of things, including catalyzing entire sequences of other reactions, that single new protein could have major consequences all through the organism. Here Behe's inadequate understanding of the evolutionary process jumps out in a very obvious way: so far he's talked about "evolution" and "selection" as if they were interchangeable words, but they aren't. Selection can fine-tune existing proteins, but it takes a mutation to create a new protein.

The rest of the chapter on blood clotting is a sustained pummeling of this already-established strawman. Behe does do one thing very, very well: he convincingly shows that the current human blood-clotting cascade could not have developed by TinkerToy construction. However, again the same hole is evident: the evolutionary precursor to the clotting cascade could have had simpler but still functional parts, and Behe makes no attempt to show that this is not the case. The TinkerToy model does not match how we know evolutionary change must occur, if it is to occur at all.

Back to top

Behe next takes on the intracellular transport system. Not a bad description-by-analogy - it only took me three reads to get all of it, and usually it takes me more than that where biochem is concerned. Again, well-reasoned writing, but again I smell something wrong. Maybe it's just me. I don't know. But the analogies - the car, the carrier truck, etc. - don't feel right. They're too simple. I have that prickling sensation again that tells me something is being left out here. Behe is not telling me everything about the intracellular transport system and how it works.

After some thought, I figure out at least part of the problem. Behe's analogies are colorful and easy to follow, but they aren't accurate. We are, after all, talking about chemicals here. Chemicals don't behave like cars, and cars don't behave like chemicals.

Behe tells the story of a girl who is going to die, killed by a malfunction in intracellular transport. Again relating this to his theme, he says that if such a minor malfunction is lethal, then surely this is an irreducibly complex system! And it could not have evolved. But again I see the same old problem: simply examining the intracell transport system in humans is not enough. To make a convincing case, Behe should also be examining that system in other organisms. In particular, I'd like to see an analysis of intracellular transport in halogen or methanogen bacteria, which are arguably the most primitive forms of life on Earth.

So far, all Behe has managed is to show that there is a distinct lack of the kind of information required to decide if these systems could have evolved.

Discussion of antibodies: interesting. I have no way of knowing if it's accurate. Assuming that it is, I find it well-written but not convincing. Again, data from other fields comes into play - Behe has ruled it out, I know, but I can't. A recent issue of NATURAL HISTORY contained a detailed article about shark immunological systems. If I recall correctly, sharks don't have antibodies. They use other means of defending against diseases.

Yes, the antibody cascade is mind-bogglingly complex. Question: if Behe is indeed headed toward a hypothesis of intelligent design, how does *he* explain this nightmarish complexity? Surely an intelligent Designer would have more sense than to design a system dozens of steps long, where any breakdown in any step results in the whole process being badly weakened or failing completely. What Behe is describing sounds more than ever like "spaghetti code," a computer programmer's epithet for a program that accumulated bit by bit as new code was added to perform new functions, without any coherent overall design. If a program is designed de novo, and designed properly, then it should be clean and efficient with no trace of spaghetti code. Spaghetti code is an almost unmistakable indication that intelligent design was not involved in the development of that program.

Again Behe says that we can search the scientific literature in vain for any papers that discuss the evolution of these systems. Again, he strongly implies that because it hasn't been done, the reader should conclude that there's some insoluble problem here. This doesn't necessarily follow, for two reasons. One is that biochemistry generally doesn't fossilize, so it's a lot harder to study evolution in biochemistry than it is to study evolution in, say, bones or teeth, which do fossilize. Second reason is that researchers tend to follow their grant money (sad but true), and while research into the evolution of complex biochemical systems is scientifically interesting, it doesn't have the immediate pay-off that some other lines of research can produce.

Back to top

The fifth part of my ongoing thoughts on Michael Behe's book DARWIN'S BLACK BOX.

I'm going to skip Chapter 7, on the AMP sequence, because it offers nothing new, just another biochemical system that Behe thinks is irreducibly complex and therefore couldn't have evolved.

In a chapter called "Publish or Perish," Behe surveys the current state of research into abiogenesis. This appears to be pretty accurate based on what I've read elsewhere, although I think Behe makes it sound perhaps a bit too negative. Later in the chapter, Behe talks about protein- and gene-sequencing, and how both techniques can be used to build trees of evolutionary descent. Behe then repeats something he said in the first chapter: that protein sequences are good evidence for Darwinian common descent. In fact, at this point Behe says that he accepts Darwinian evolution to explain micro-evolution, and considers gene-sequencing strong evidence for larger-scale evolution. He just doesn't think it can explain the formation of irreducibly complex biochemical systems.

Again, Behe complains about the lack of any papers in the scientific literature that concern possible evolutionary paths for complex biochemical systems. Again, he seems to want the reader to conclude that because it hasn't been done, it can't be done. That was illogical before; it's still illogical here.

Next up: a chapter titled "Intelligent Design," in which Behe gets down to the matter of proposing his own countertheory. First, though, comes a discourse on mitochondria and the Margulis theory that mitochondria are degenerate cells that invaded and became symbiotic with larger cells. Behe complains that this is another pseudo-explanation, because it "explicitly assumes" that the invading cell already had the ability to manufacture foodstuffs and the host cell already was an acceptable habitat for the invader. This is a specious objection, because it has a very simple and obvious answer. Before mitochondria, all cells must have had the ability to generate their own energy, because obviously they didn't have mitochondria to do it for them. And any cell that wasn't a good habitat for the invader would not have been successfully colonized. All mitochondritic (eukaryotic) cells are descended from the few species, perhaps only one species -- perhaps, ultimately, a single individual - that was an acceptable habitat for the invader. Classic Darwinian evolution: a small variation yields a major adaptive advantage which the organism turns into tremendous evolutionary success.

Behe considers and rejects Kauffman's complexity theory as a possible 'explanation' for biochemical complexity. I have to admit this confuses me - I've heard "self-organizing complexity" offered as a hypothesis for abiogenesis, but never as one for individual biochemical processes. I have to wonder if this is another strawman. On reflection, though, I conclude that it isn't. Behe is doing what any good scientist with a novel hypothesis should do: carefully considering each possible counter-argument to his claims and showing (or trying to show) that it isn't a valid counter at all.

Behe declares flatly that you don't need to know anything about the designer in order to talk about whether or not the system was designed. I think this is a bit of artful question-dodging, but that's just my opinion. Yours may vary.

Behe next plunges into the question of "how do you detect intelligent design?" I can't find anything particularly wrong with his reasoning on this point: basically, he concludes that anything that can't reasonably be a product of natural processes must be a result of intelligent activity. This is pretty much the same argument that paleoanthropologists use to identify early, primitive stone tools and other artifacts. I accept it from them, so I can't very well reject it coming from Behe.

Behe then revisits the various "irreducibly complex" biochemical systems he talked about earlier in the book. Not surprisingly, he concludes that they can't be the result of natural processes, so they must be the result of intelligent design. To his credit, Behe does say that the fact that some biochemical systems are designed doesn't mean that all are designed. Some processes may indeed have evolved.

Behe next goes into the history of the argument from design, a large section of material that I don't fully follow and have no particular desire to - as usual in such discussions, it seems to wander off into "how many angels can dance on the head of a pin?" type questions, which I've always thought amounts to aimless mental wheel-spinning. About all I can extract from it is that Behe thinks Paley's "Watchmaker Argument" - that is, a watch (intelligent design) implies a watchmaker (a designer) - has never been adequately refuted. Indeed, it hasn't. In fact, it can't, because it's correct. The issue is not "does a watch imply a watchmaker." The issue is "is life analogous to a watch?" or more directly "is life the result of intelligent design?" Behe takes Dawkins and others to task for having handwaved the Watchmaker Argument aside, but in my understanding they did not do this. Dawkins doesn't argue that the Watchmaker Argument is flawed. He argues that the Watchmaker Argument doesn't apply to life, because life is not analogous to a watch. Life is not clearly the result of intelligent design. However, IF Behe is correct that life shows the hallmarks of intelligent design, THEN Paley's argument is applicable. Design implies a designer.

Next up is a discussion of one of the major arguments against design: that a designer would have done a better job of design on many aspects of life, such as the vertebrate eye. Behe claims this argument is flawed because it expects good design when there may have been many reasons to design a system poorly. I accept the argument and find his answer somewhat specious, because my opinion is that "an advanced intelligence would have to be an advanced stupidity" (to steal a phrase from Isaac Asimov) to create some of the systems we see in modern terrestrial life. But I suppose others could see it the other way around. Judgement call. No clear winner here.

Which is not to say that all of Behe's arguments on this point are equally sound. On p. 224, Behe offers the following "syllogism" (his word) to summarize this argument against design:

"1. A designer would have made the vertebrate eye without a blind spot.
"2. The vertebrate eye has a blind spot.
"3. Therefore Darwinian evolution produced the vertebrate eye."

"It is for reasoning such as this," Behe says, "that the term non sequitur was invented." In fact, this argument is a non sequitur (although it is not, strictly speaking, a syllogism), but it's also a strawman. The imperfection argument is properly used only to answer the claim of an intelligent designer, and the proper conclusion is "therefore, the vertebrate eye was not designed." I don't recall seeing anyone use this argument to defend evolution, only as an argument against a designer.

Behe next takes on the claim that vestigial organs and nonfunctional genes argue against a designer. This particular section had me scratching my head in confusion. One way that Behe answers this argument is by saying that "because we have not yet discovered a use for a structure does not mean that no use exists." So he's calling this an argument from ignorance. I find it extremely odd that he can recognize when someone else is using an argument from ignorance, and recognize that it's a fallacy, but he can't recognize the same thing in his own reasoning.

Behe points out in a rather ferocious tone (which is probably more or less necessary :-) ) that intelligent design theory is NOT necessarily the same thing as Young-Earth Creationism (YEC). This is true. One cannot refute the design theory by refuting YEC. (OTOH, one can refute YEC by refuting design, since a Designer is a necessary part of YEC.)

The next section again had me scratching my head. On first reading, it almost sounded like Behe was using intelligent design as a deus ex machina: if evolution can explain something, well and good, but if it can't, then that something must have been designed. On reflection, though, I have to admit that this isn't as bad as it first sounds. It's really no more than an application of Occam's razor. You apply the simplest explanation required by the facts.

Behe's final chapter, "Science, Philosophy, Religion," is a discourse on the relationship between those three topics, and why science (in his opinion) doesn't want to recognize this momentous discovery that Earthly life was designed. The first couple of pages probably do more to turn readers off Behe than anything else in the whole book, because they exude a distinct aura of self-congratulation. I can't say if Behe meant it to read this way, but it certainly sounds like he is patting himself on the back for being the one with the intelligence to recognize the evidence for design and the courage to declare it openly. Behe seems to think that most scientists are either stupid or craven cowards, either unable to see the evidence for design or too scared of peer pressure to admit it.

Back to top

This is the sixth and final part of my running critique of Michael Behe's book DARWIN'S BLACK BOX. This part is a summary of my feelings on the book and why it fails to convince me that Behe's got a valid case.

What's my overall opinion of Behe's book? It's well-written and well-argued, and I can see why many people find it convincing. However, I don't find it convincing. As I stated several weeks ago when I started this series of posts, that may be a function of my bias going in. I don't think so, though. I think there are at least three distinct reasons why Behe's argument, well-argued as it is, falls apart.

Reason #1 is that Behe doesn't appear to truly understand the evolutionary process. He thinks evolution should work like a kid with an Erector set: wherever an irreducibly complex system appears, it must have been assembled whole and complete. The idea that it might have developed by a circuitous, indirect route doesn't appear to enter into his thinking. To see why this is a fallacy, I suggest you read the talk.origins FAQ titled "Darwin's Black Box: Irreducible Complexity or Irreproducible Irreducibility?"

Reason #2 is that Behe's argument is basically a gigantic argument from ignorance: because we don't understand now how these systems might have evolved, we never will. Besides being conceptually flawed, this is factually flawed because some of the research Behe claims doesn't exist actually does exist. For details, refer to the same talk.origins FAQ cited above.

Reason #3 is a fundamental error in the root argument which Behe is offering. Basically, Behe's argument goes like this:

Premise 1: Irreducibly complex systems cannot evolve by Darwinian methods.
Premise 2: Biochemical systems exist which are irreducibly complex.
Conclusion A: These biochemical systems could not have evolved by Darwinian methods.
Premise 3: If a system wasn't evolved, it must have been designed.
Conclusion B: These biochemical systems must have been designed.

This is a perfect illustration of the difference between a valid argument and a true one. To be valid, an argument needs only to follow the rules of logic: that is, if the premises are true, then the conclusion(s) must be true. To be true, an argument has to be valid and the premises have to be true. In other words, validity is a matter of internal consistency, but truth is a matter of both internal and external consistency.

Behe's argument is valid. If the premises are true, the conclusions must be true. However, Behe has failed to show that the premises are true. In particular, he's failed to show that irreducibly complex systems cannot have evolved by Darwinian means. This makes the entire book little more than an exercise in abstract thinking. Behe may be right, but he hasn't proved it. A large probability exists that he's wrong. I think he is wrong, and I think future research will demonstrate this. DARWIN'S BLACK BOX is a good book, well-written. However, it's a well-written exploration of an inherently flawed concept. And it definitely is not a killer argument against the theory of evolution.