Sunday, April 16, 2006

ID, lugnuts, and the lugnuts who love ID

The DI's Casey Lumenskin Luskin writes a response to claims that there are no irreducibly complex systems known, and does so by asking "Do Car Engines run on Lugnuts?"

Setting aside the fact that Irreducible Complexity is incoherently defined (at least if we are interested in understanding evolution), and passing by the biological illiteracy that Luskin demonstrates, I want to focus on the analogy from his title.

He explains:

Car engines use various kinds of bolts, and a bolt could be seen as a small “sub-part” or “sub-system” of a car engine. Under Miller's logic, if a vital bolt in my car's engine might also to perform some other function—perhaps as a lugnut--then it follows that my car's whole engine system is not irreducibly complex. Such an argument is obviously fallacious.
Let's pass by the fact that a bolt is not a nut, and could not, therefore, be a lugnut. Ignore the lowhanging fruit.

Luskin saying that Behe claims "a system is irreducibly complex if the system stops functioning upon the removal of one part. This is the appropriate test of Darwin’s theory because it asks the question, 'Is there a minimal level of complexity which is required for functionality of this system?'"

Only it doesn't. In his own writings, Behe has acknowledged that IC as defined above is not a good test of evolution, because it's backwards looking, not forwards looking. In other words, it asks whether a system can be broken down, not whether it can be built up. This may seem trivial, but as we shall see, it isn't.

I think we can all agree that an engine would not function, could not even be constructed, without nuts and bolts. But Henry Ford didn't invent the nut and bolt. That work was done earlier, for a different purpose. Car manufacturers took advantage of an existing system for supplying that particular fastener. In the absence of the nut and bolt industry, Ford would have had to turn elsewhere for his fastener needs. Maybe he'd have invented a new fastening system. But it's a perfect example of exaptation, the alternative to IC that Luskin is trying to dismiss.

The internal combustion engine wasn't designed for a car, nor was the Model T chassis that different from the buggy chassis that came before. The various subsystems of a car were invented separately and brought together piecemeal. By the definition of IC presented, the car is IC, but in real terms, it's easy to see how subsystems in the car could be added separately. And that's what's wrong with the concept of IC. Nuts and bolts were developed in other contexts and were added to the car after they worked well already. So did its other parts.

Each of those parts evolved in their own way, becoming increasingly efficient on their own. Then, once integrated into the car, they became more specialized for that purpose, until the fit between components would seem irreducible. What good is a car that has a transmission but no engine? None, but why should we expect it to work? A car without a transmission could still be built, and taking the engine out of that car gives you something that would still operate. If you didn't have bolts you'd weld, or rivet, or nail wooden walls together.

Luskin insists that Behe acknowledged that point and dismissed it, but I don't buy it. To the extent IC has any coherent meaning, it is an attempt to operationalize the notion of unevolvability. If the concept doesn't capture viable evolutionary pathways, albeit circuitous ones, it doesn't do its job. You're left being able to say "it couldn't evolve this way." Which is scientifically interesting, but doesn't even approach the claim that "it couldn't evolve." In Dover, IC was offered as scientific proof that evolution couldn't explain some systems. Since IC isn't quite defined well enough to capture that idea, it's appropriate to dismiss it as evidence.