Intelligent Design: The Illusive Step

I've had this one on the back burner for quite some time now, so I'm going to give it a shot...

Boiled down to its simplest form, it seems like the scientific arguments of the ID movement reduce to two basic arguments, the one relying upon the other.

Behe's theory of irreducible complexity holds that certain complex systems cannot be the product of gradual evolution because if any one part were missing, the whole would not work.

Dembski's theory of specified complexity takes Behe's theory and runs with it, pointing out that it is impossible for whole systems or subsystems to jump into being in a single step (which they would have to if irreducible complexity is correct, and they cannot assemble gradually), because the statistical likelihood of such a complex system resulting from a random mutation (or other random event) is so infinitesimal as to be virtually impossible.

Essentially, Behe puts the ball on the tee by saying that complex systems such as cells must evolve one whole system at a time, and Dembski then performs a 450 yard drive by proving that it is, in turn, impossible for whole subsystems to evolve either.

Although one may (and people who, unlike me, are masters of all the math involved have) dispute the methodology by which Dembski has sought to calculate the specified complexity of individual biological systems, the basic principle which he has identified is unquestionably true. Given that time, space and matter are finite, some things are so statistically unlikely to have happened by chance, that one may more or less dismiss them as possibilities. Thus, the likelihood that some huge evolutionary step would happen spontaneously in one generation is virtually nill (and in sexually reproducing organisms, unhelpful as well, since such a 'hopeful mutant' would have nothing biologically compatible to mate with.)

The real crux of the ID debate is Behe's theory of irreducible complexity. If irreducibly complex systems cannot form gradually, then Dembski's theory kicks in to prove that they cannot form at all without some sort of outside intervention.

I think it's illustrative to step through the debate between Kenneth Miller and Michael Behe over Behe's flagship example of irreducible complexity: the bacterial flagellum.

1. In Darwin's Black Box Behe argued that the bacterial flagellum was irreducibly complex.
2. In Finding Darwin's God, Miller argued that Behe was wrong because there were a number of other designs of the flagellum found in nature, some of them significantly less complex than the one Behe described in his book.
3. Behe responded that he hadn't meant that the design itself couldn't be different, but rather than none of the essential elements of the flagellum (microtubules/filaments, basal body, membrane) could be missing.
4. Miller in turn responded by pointing to the type III secretory system, a waste disposal organelle whose proteins are directly homologeous to those the flagellum's base. Miller pointed out that the TTSS functioned successfully as a waste disposal mechanism, which in turn can be used as a primitive propulsion system, and thus could be an indicator of what sort of organelle the flagellum evolved from.
5. Behe responded that the TTSS has a different function from the flagellum, and so the fact that it was a working subset of the flagellum had nothing to do with the flagellum's irreducible complexity qua propulsion mechanism.

What strikes me from reading this exchange over the last few years is that Behe's point is not so much that the flagellum itself could not be the product of gradual evolution, but that to the rotary propulsion system qua rotary propulsion system could not have gradually evolved. Behe doesn't question (at least not necessarily, though he may in particular) that a given configuration of flagellum as found in a given species may be the product of descent with modification from an ancestral flagellum with a different configuration but the same basic working elements.

Recall that one of the basic tenets of natural selection is that for a feature to be selected for, it must at every stage benefit the creature. So, for example, it is only possible to say that the human eye is the product of evolution if one can show that each step along the line of development was of more benefit to its possessor than the step before it. (Or as Dick Dawkins said in one of his non-raving moments: "How much good is 50% of an eye? Well, a little bit more than 40% of an eye.")

If I follow correctly, Behe is saying that the problem with the flagellum being the product of evolution is that the purpose of the flagellum is to provide rotary propulsion. Yet if any of the three basic elements of the flagellum were missing, it could not rotate, and thus would be no use for propulsion. So what use could 2/3 of a flagellum be? Thus, a designer would have to be involved, someone who knew that 2/3 of a flagellum was a step on the way to a working flagellum -- or who simply created the flagellum whole.

I wonder, however, if Behe's line of thinking is too locked into a teleological view of the system that he's examining -- if he's thinking too much about the fact that the flagellum is a rotary propulsion system. If the evolutionary explanation is correct, the flagellum didn't start out to be a rotary propulsion system. It doesn't know it's a rotary propulsion system. Rather, the flagellum would have originated as an organelle that provided some sort of propulsive function, which, through a series of advantageous incremental changes, became capable of rotary propulsion.

Now, if you're still reading, you're already shown much more tenacity when faced with words like "flagellum" and "organelle" than the average bear, so let's bring this out into the real world of things that we're familiar with. A while ago, in an online conversation about ID, someone commented: "If you don't believe in irreducible complexity, go pull a part out of your car and see if it still runs."

It happens that I recently had to do some hands-on car repair: replacing my radiator. Removing the radiator would be a perfect example of slamming hard against the wall of irreducible complexity. Once you take it out, you drain the remaining coolant and (if, like my Camry, your car has an oil cooler as well) most of the oil out of your car into a bucket. If you started your car at this point, very bad things would happen very quickly.

However, say you built a car from scratch, and simply did not include a radiator: all you had was a pipe that allowed you to circulate oil through the engine, then bring it around and pump it back in again -- a very basic oil circulatory system to keep everything lubricated. Now clearly, your car has some functionality issues. It'll run, but only for 5-20 minutes (depending on how fast you're revving the engine and how cold it is outside) before it overheats. Then you need to turn it off and let it cool before you run it again.

However, there is something you can do to improve this performance. (And if someone chimes in at this point that I'm talking about a situation with a designer: I'm aware of that. Let's see where we get with all this.) If that pipe that circulates the engine oil back into the engine is flattened out and/or made longer so it can be farther from the engine and radiate heat more easily, your engine will run a little longer before overheating. If you ran the pipe next to a convenient heat sink (say, a tank full of water) you could last even longer. If you then took your heat sink, and ran it through an exposed grillwork to help it cool down... you'd have a radiator.

Now, the point of this example is not that a radiator could evolve, clearly, the changes here are the product of a designer trying to make his car work a little better. But the question is: at what point did the car start to have a radiator? Did it have one when all it had was an exposed oil circulation pipe which just happened to help radiate heat? Was it when that pipe was extended and flattened out to radiate more heat? In a certain functional sense, it was always a radiator. But if you'd examined the car at stage one, you would have said: "This car has no radiator." And it didn't. It had an oil circulation system that had some radiatory properties.

According to evolutionary theory, this is how the addition of a new cell organelle, or a morphological change in a species of multi-cellular creatures would take place. It's not a matter of "I need a rotary propulsion system, I'll add a flagellum" but rather "I need to excrete waste. I'll just blow it out of the cell through this membrane tube. Hmmm. That helps me move. I'll excrete waste when I want to move. Hmmm. Having some filaments in the tube helps control the direction of movement a little better. Hmmm. If I can whip the filaments around, I can move even when I'm not excreting. Whoa! What's this? I've got a rotary propulsion mechanism. What's up with that?"

Now, none of this explains by what evolutionary pathway the flagellum evolved, nor does it show if it evolved. However, the purpose of this overly long exercise is to ask: Does the concept of irreducible complexity assume that the function of a given system was always the function that it has now. If so, irreducible complexity essentially assumes that evolution is false as a starting point, since one of the basic ideas behind evolution via natural selection is that natural selection works with what is already there: finding new uses for existing characteristics, and then modifying them by selecting for those variants of the characteristic most useful for the new purpose.