Behe's answer to that objection:
That's what often happens when people who are adamantly opposed to an idea publicize their own definitions of its key terms—the terms are manipulated to wage a PR battle. The evident purpose of Miller and others is to make the concept of IC so brittle that it easily crumbles. However, they are building a straw man.
I never wrote that individual parts of an IC system couldn't be used for any other purpose. (That would be silly—who would ever claim that a part of a mousetrap couldn't be used as a paperweight, or a decoration, or a blunt weapon?) Quite the opposite, I clearly wrote in Darwin's Black Box that even if the individual parts had their own functions, that still does not account for the irreducible complexity of the system. In fact, it would most likely exacerbate the problem, as I stated when considering whether parts lying around a garage could be used to make a mousetrap without intelligent intervention.
In order to catch a mouse, a mousetrap needs a platform, spring, hammer, holding bar, and catch. Now, suppose you wanted to make a mousetrap. In your garage you might have a piece of wood from an old Popsicle stick (for the platform), a spring from an old wind-up clock, a piece of metal (for the hammer) in the form of a crowbar, a darning needle for the holding bar, and a bottle cap that you fancy to use as a catch. But these pieces, even though they have some vague similarity to the pieces of a working mousetrap, in fact are not matched to each other and couldn't form a functioning mousetrap without extensive modification. All the while the modification was going on, they would be unable to work as a mousetrap. The fact that they were used in other roles (as a crowbar, in a clock, etc.) does not help them to be part of a mousetrap. As a matter of fact, their previous functions make them ill-suited for virtually any new role as part of a complex system.
The argument of co-option
1. Co-option in microbiology means borrowing parts of systems from different places to form a new system. When in this way a new system is generated it has a new function. This proves evolution.
2a. But in the evolutionary scheme not all the systems were available by co-option.
2b. In the simple example of the flagellum of E-coli, only 10 out of 40 components can be traced back as having been developed by co-option.
2c. The rest of the 30 components are unique and new. There are simply no known homologues to them.
These 30 components were not available for co-option in hypothetic ancestral lines leading from e.g. a bacterium with no flagellum.
3. This again proves the existence of a designer who is no one else then God.
A Response to Sharon Begley's Wall Street Journal Column
Michael J. Behe
Darwin's Black Box, page 66.
The reason why a separate function for the individual parts does not solve the problem of IC is because IC is concerned with the function of the system:
By irreducibly complex I mean a single system which is composed of several well-matched, interacting parts that contribute to the basic function, and where the removal of any one of the parts causes the system to effectively cease functioning.
Darwin's Black Box, page 39.
The system can have its own function, different from any of the parts. Any individual function of a part does not explain the separate function of the system.
Miller applies his crackerjack reasoning not only to the mousetrap, but also to the bacterial flagellum—the extremely sophisticated, ultra complex biological outboard motor that bacteria use to swim, which I had discussed in Darwin's Black Box and which has becoming something of a poster child for intelligent design. No wonder, since anyone looking at a drawing of the flagellum immediately apprehends the design. Since the flagellum is such an embarrassment to the Darwinian project, Miller tries to distract attention from its manifest design by pointing out that parts of the structure can have functions other than propulsion. In particular, some parts of the flagellum act as a protein pump, allowing the flagellum to aid in its own construction—a level of complexity that was unsuspected until relatively recently.
Miller's argument is that since a subset of the proteins of the flagellum can have a function of their own, then the flagellum is not irreducibly complex, and Darwinian evolution could produce it. That's it! He doesn't show how natural selection could do so; he doesn't cite experiments showing that such a thing is possible; he doesn't give a theoretical model. He just points to the greater-than-expected complexity of the flagellum (which Darwinists did not predict or expect) and declares that Darwinian processes could produce it. This is clearly not a fellow who wants to look into the topic too closely.
In fact, the function of a pump has essentially nothing to do with the function of the system to act as a rotary propulsion device, anymore than the ability of parts of a mousetrap to act as paperweights has to do with the trap function. And the existence of the ability to pump proteins tells us nil about how the rotary propulsion function might come to be in a Darwinian fashion. For example, suppose that the same parts of the flagellum that were unexpectedly discovered to act as a protein pump were instead unexpectedly discovered to be, say, a chemical factory for synthesizing membrane lipids. Would that alternative discovery affect Kenneth Miller's reasoning at all? Not in the least. His reasoning would still be simply that a part of the flagellum had a separate function. But how would a lipid-making factory explain rotary propulsion? In the same way that protein pumping explains it—it doesn't explain it at all.
The irreducible complexity of the flagellum remains unaltered and unexplained by any unintelligent process, despite Darwinian smoke-blowing and obscurantism.
I have pointed all this out to Ken Miller on several occasions, most recently at a debate in 2002 at the American Museum of Natural History. But he has not modified his story at all.
As much as some Darwinists might wish, there is no quick fix solution to the problem of irreducible complexity. If they want to show their theory can account for it (good luck!), then they'll have to do so by relevant experiments and detailed model building—not by wordplay and sleight-of-hand.