Behe and Irreducible Complexity

My response to Behe and his argument from Irreducible Complexity is that it appears to have nothing, except in the most superficial sense, with biology. Yes, he presents it as an argument about evolution and natural selection, but as presented, at least with his mouse trap example, it is almost totally divorced from actual biology. Let me explain why I think this is so.

He presents the mouse trap as a means of illustrating this idea of an object that is irreducibly complex and therefore cannot be produced by natural selection. He argues that you cannot take the several pieces of the mouse trap and assemble those pieces one at a time into the final mouse trap and have a functional mouse trap at each step. Many have pointed out that this argument leaves out change in function, which is a valid point, but even without that problem I think his example has nothing to do with natural selection, or even biology. When he talks about assembling the mouse trap by successively adding parts, he is clearly talking about the parts of the final mousetrap. So, if you were to try and assemble the trap starting with the striker bar, you must use the bar from the final trap. You could not start with something that on its own serves to kill mice and later, as more parts are added, is modified into the final striker bar. He makes no effort whatsoever to argue that you cannot get to the final mouse trap by such a process of addition and modification. He fails to show that you cannot assemble the final mouse trap by adding parts which are modified between additions. The parts in his analogy, in other words, do not change. It is this which is at odds with actual biology.

The problem is even more striking if we look at biology a little deeper. If we are to consider some system in an organism, analogous Behe's mouse trap, whether irreducible or not, then in each generation the system is modified slightly by whatever mutations might occur. So the "addition of a part" which is the only process that Behe discusses occurring is something that takes place from one generation to another. That is some offspring have one 'part' more than the parents had. But we know of the kinds of changes that take place between generations. The types of mutations include point mutations, insertions, deletions, gene duplications and cooption (the process whereby a part which exsists in one system serving one function is taken over by another system to serve another function). Most of these changes involve the modification of parts in exactly the way that is excluded in Behe's analogy. Furthermore, the only processes that might be accurately described as 'adding parts' are
1) cooption and
2) gene duplication followed by modification

The problem here then is that Behe and Dembski specifically deny that cooption is a force in evolution which eliminates option 1 and option 2 includes gene modification which is nonexistent in Behe's analogy.

In short the problem is that if Behe recognizes that the 'parts' of biological systems are modified by evolution, then his argument with the mouse trap says nothing about whether you can produce such a thing by the gradual process of natural selection. If he denies that biological 'parts' are modified in each generation he is at odds with most of what is actually observed about living things.