Design Arguments and Probability - Reply to Ocham

I certainly don't mind being challenged, and I can usually count on my regular commenter, Ocham, to do just that. He seems to take issue with nearly everything I say. My last post on Intelligent Design was no exception. Here's his latest:

Ocham: I don't see how the designist argument counts as an explanation. As Dawkins says, it's a cop out. How do you explain how the designer came to exist, in all its complexity and so forth.

Another problem: as Einstein said, all arguments from probability disappear in the face of the improbable event actually happening. In the case that any highly improbable event occurs, namely the event that there exist conscious beings capable of asking how probable their existence is, naturally they will come to the conclusion that their existence cannot be chance at all. But in that case they would be wrong.

My reply: I'm surprised you're bringing in Dawkins here, Ocham. His favorite "Who designed the designer?" line, which he clearly takes to be a knock-down refutation of the design argument, is hardly that. When the ID theorist proposes that a mind or designer might serve as a basic explanation, she does not automatically commit herself, as Dawkins supposes, to such claims as

1. Basic designers must be complex beings.
2. Basic designers must be contingent beings.

What Dawkins' question does show is that the ID theorist is implicitly committed to a denial of both (1) and (2) - a truly basic designer must be an undesigned, non-emergent designer. But a being like God fits that bill perfectly, since God is a metaphysically simple, noncontingent being. So, far from refuting design arguments for theism, Dawkins' line actually helps the theist connect the dots from premise to conclusion. That Dawkins fails to see this shows only that he doesn't understand theism and refuses to take it seriously as a worldview.

Secondly, the probabilistic considerations you raise ignore the important distinction between prior and posterior probabilities. A priori, for example, any specified string of 10 poker hands is equally likely to be dealt. Hence, getting a string of 10 spade suite royal flushes in a row, say, is just as likely as getting any other string of 10 poker hands. In short, where X and Y are two arbitrary sequences of 10 poker hands, Prob(X)=Prob(Y). But when we are evaluating explanations we're not interested in prior probabilities but posterior probabalities. According to Bayes' theorem, the probability of a given explanatory hypothesis H given evidence E is

Prob(H/E) = [Prob(H)*Prob(E/H)]/Prob(E)

The prior probability of an arbitrary sequence of 10 poker hands gives us only Prob(E). The other quantities on the RHS also need to be evaluated before we can draw any interesting conclusions regarding the respective merits of chance-type hypotheses and design-type hypotheses. Intuitively, if a person gets one royal flush, we'd say he probably just got lucky. If he gets two in a single game, that's awfully suspicious. If he gets 3 or more, someone's likely to get shot for cheating. The reason is that a royal flush is not just any hand in poker. It's a very special hand, one that intelligent poker agents take (or are likely to take, or should take) particular interest in. So as the number of royal flushes in a game goes up, the design hypothesis (it's rigged; someone's cheating) rapidly becomes more plausible than the chance hypothesis (it's sheer luck or coincidence).