October 28, 2005

A counterexample to ID

A generic argument for Intelligent Design goes something like this:

  1. Organism O exhibits biological system S
  2. Biological system S has distinct parts X, Y, and Z
  3. Without any of X, Y, and Z, S would not function properly
  4. Hence, S exhibits irreducible complexity (IC)
  5. The best explanation for IC is the existence of a designer
  6. Hence, by inference to the best explanation, from S we conclude the existence of a supernatural designer

The most obviously contentious claim is (5), but more often opponents of ID attack the inference from (3) to (4) as begging the question against natural selection -- experiments reconstructing the evolution of, for example, the eye, show that 'irreducible complexity' is a bogus notion. My interest here is to provide counterexamples to (5). (Continued below the fold.)

Consider first an apple pie. While simple enough by our standards that one can be thrown together in just a couple hours, the pie is a rather sophisticated culinary construct.

First, there's the crust, which requires flour, a solid fat (usually butter or margarine), and knowledge of how to turn those into pastry dough -- knowledge which certainly never occurred spontaneously as someone idly watched a wild cow grazing in a field of undomesticated grass. Similarly, there are some surprising ingredients (cinnamon comes from the bark of a tree, for example) in the filling of an apple pie, and many are crucial to its final flavour and structure. And, of course, the crust and the filling have to be brought together and cooked properly, which requires the technology of an oven.

At the very least, an apple pie cannot 'function' as an apple pie wtihout all of the following: flour, solid fat, water, apples, brown sugar, spices including cinnamon, nutmeg, and allspice, salt, a metal pie plate, and an oven capable of surrounding the pie with a roughly regular temperature throughout the entire cooking time. Missing just one will lead to a defective and undesirable pseudo-pie.

Hence, an apple pie exhibits irreducible complexity.

But is it right to say the apple pie was designed? I argue that this is not the case. I'm not a culinary anthropologist, so I can't recount for you the history of pastry; but certainly the first apple pie recipe did not spring into existence ex nihilo. Rather, the inventor of the apple pie no doubt drew upon previous experience baking other sorts of pastries; and pastry itself probably developed gradually, as bakers experimented with the different agricultural products surrounding them, pursuing those developments which proved to be the tastiest. That is, pastry (and thus the apple pie) evolved gradually, in a fashion that is much more reminiscent of natural selection than intelligent design.

Another, example is beer. Successfully brewing beer requires the technology to make glass bottles, sterilize them, and hermetically seal them; as well as cultivate and exploit the metabolism of yeast, and raise, roast, and integrate hops to provide flavour. Beer exhibits irreducible complexity, but it is ridiculous to think the complete brewing process, from field to Braeuhaus, was first conceived in its entirety. Small innovations, tiny steps and experiments, were the key to the development and evolution of beer.

Jams and jellies also exhibit irreducible complexity, as cooking, the stabilizing effects of sugar, and sterilization techniques are all necessary for a successful jam (not to mention knowing which fruits can and can't be jammed). But, again, no designer looked at a blackberry bush and came up with the whole preserving process in one fell swoop.

Hence, not only is the conclusion of design not inference to the best explanation; there are actually empirical cases where it is simply false. There's an important difference here with the experiments done to reconstruct the history of the eye: those experiments merely show that the defender of natural selection is a viable alternative to ID, not that ID itself is flawed. (5) is not shown to be false by naturalistic accounts, merely cast into doubt. The examples I've given here, on the other hand, show (5) cannot be true. To preserve their theory, an IDer must qualify (5), and argue why it applies in the case of biological evolution, but not culinary evolution.

No comments: