Thursday, January 19, 2006


Well, I did it again. I went debating intelligent design proponents. I gotta quit doing that.

Still, I composed this interesting rant in the process. Consider it another lesson on explanation.

An explanation is a rule of implication. Cause A leads to effect B. If you see A, then B must happen.

Not "B can happen." "B must happen."

A scientific inference would be to observe A preceding B and theorize that A causes B.

A blind guess would be to observe only B, and claim A implies B.

Still, every explanation, even if guessed blindly, makes a prediction that cause A always implies B.

Having a lack of an explanation means that either 1) you cannot identify any such a rule, or 2) you postulate a rule such that every event has its own private, hidden cause (cause A is undetectable).

Hence, generic ID is NOT an explanation! It's not even a blind guess at an explanation. It proposes no testable rule of implication. Yes, an intelligent alien could have designed life on Earth! That's utterly trivial. The problem is that generic ID doesn't predict what we see. Generic ID doesn't imply the world must be as we see it. Generic ID is just compatible with what we see. As are lots of things.

A specific theory of ID has to predict the world we see, rather than some world we don't. A scientific theory of ID has to be experimentally testable, at least in principle. For example, say the aliens designed life for purpose X, using tools Y, at time T, leaving corresponding evidence E. You can have gaps in your ID theory, but it must make a prediction to be science.

ID always fails to do this because the designer is always God. God theories have an infinite number of free parameters, so fitting your God theory to an arbitrarily large number of observations still fails to make a prediction. It's like fitting a curve to points on a graph and somehow failing to predict any interpolated or extrapolated point. This is done by making your "curve" tautologically equal to the points to which you are fitting. It's not a theory or an explanation if it's just a restatement of your data.

ID tries to pull the wool over the eyes of the public by cooking up formulas that "detect design". How do they do this? They try to construct a function that returns true only on human artifacts and on biological mechanisms.

But this isn't science, my friends. If I cook up a function that returns true on typewriters and supernova remnants, have I proved that typewriters are capable of star formation? (Hint: No.) My function is just a paraphrasing of my prior observations of supernovas and typewriters.

Writing down a formula for kinetic energy (one half mass times velocity squared) doesn't make a theory. If I define Floopdedoo as mass-cubed times velocity squared, I haven't defined the law of Floopdedoo. It is no more than a mathematical transformation on my raw observations.

The theory is that the transformation can express a law of nature and make a prediction. Conservation of energy is a law. It is the statement that energy out must be equal to energy in. Must!

Similarly, a law of Floopdedoo, if there was one, would have to say that given Floopdedoo in, Floopdedoo out must be constrained (predicted).

So far ID has one or more formulas for "CSI". CSI is just a formula that returns true on things that look designed to humans. Well, that's just a formula, just a paraphrased observation. If you want a theory, the theory is that CSI is correlated with design in non-human artifacts (e.g., biological structures). So that's what ID has to substantiate. Of course, the only designed artifacts are human, so there's no way to test the formula except by seeing actual evidence that life was designed and built by aliens! CSI is a formula, it is not evidence in and of itself, no matter how well it correlates with human artifacts and biological structures. No matter how many times or how precisely you say life "looks designed", you still have no evidence.

In lieu of a scientific theory, ID is forced to make the claim that NDE is somehow broken. As if that lent any weight to ID's claims. NDE may be incomplete, but at least it's successfully predictive and explanatory. ID is neither.

1 comment:

rob said...

Personally, I think everythings too complicated to be designed but if I had to make a prediction it would be that you will get banned again.