Oh, the irony

Casey Luskin of the Discovery Institute wonders:

Question: What do you do when a theory logically predicts both (a) and not (a)?
Answer: Apparently you heavily promote it.

Luskin is referring to a couple of evolutionary papers that appear to do just that. It would seem that a theory that can predict both one thing and it's absolute opposite would be absolutely useless as far as science goes. And he would be right - to an extent. There are some problems with Luskins argument, however. (1) Predicting both (a) and not (a) is NOT necessarily bad. To see why, consider this: if (x) then (a); if (y) then not (a). Both (a) and not (a) are predicted but the conclusion depends on the premises of the argument. So, predicting two opposite outcomes is not necessarily a bad thing at all. (2) Evolutionary theory on it's own doesn't make any predictions. It is the auxilliary propositions added to them that do. So in a scenario where hyposesis (x) predicts (a) and hypothesis (y) predicts not (a) and (a) is the outcome, hypothesis (x) makes the successful prediction. Fine, you might say, but if evolution can make contradictory hypothesis it would still be able to predict anything. What is important to note, however, is that there are restrictions as to which auxilliary propositions can be used - they must be independently supported. For a starter, we're not allowed to assume that the theory we are testing is true, for then we would be indulging in circular reasoning. Second, we can't use the observations that we based our auxilliary proposition on to test our theory, for if we did, we could simply have our proposition predict whatever is observed. Lets use this knowledge to explain why Luskin is wrong when he claims that evolution predicts both (a) and not (a). Writes Luskin:

The first article about the evolution of Waterfowl genitalia contends, “Scientists had speculated that male waterfowl evolved longer phalluses to give them a competitive edge over those not as well-endowed when it came to successfully fertilizing females.” That makes sense, I suppose. But the article makes one admission that strikingly contradicts that little just-so hypothesis: “Most birds lack phalluses, organs like human penises. Waterfowl are among the just 3 percent of all living bird species that retain the grooved phallus…” If long phalluses are so advantageous for reproduction, why did so many birds supposedly lose them? Darwinists will look back retroactively and tell us that, in the environmental conditions for most bird species, long phalluses weren’t advantageous. The problem in so doing is that they now have a theory which can explain both (a) long phalluses, and also not (a).

The auxilliary proposition here (male waterfowl evolved longer phalluses to give them a competitive edge) can be tested by performing a "simple" experiment. One can take two male populations of fowl; one has short phalluses, one has long ones. Allowing for mating with female fowl, one would observe which population was the most reproductively successful. If the population with the long phalluses was more successful then the proposition is strengthened since it predictied that this would happen. Notice that the proposition does not require you to assume that evolution is true what-so-ever. Neither are you recycling old "known" observations.

Let us now have a look at one of Luskin's favorite ID predictions, namely that ID predicts that the should be no junk DNA. Luskin has made this claim so many times that my eardrums (or should that be retinas?) are starting are going numb. Let's examine how ID would fare under the above conditions. First of all, can we avoid circular reasoning for reaching this conclusion? As it turns out, no, we can't. ID makes no assumptions what-so-ever about the supposed designer. (The following is taken virtually verbatim from an article I wrote previously) In fact, ID does not and CAN NOT tell us anything about the designer; It cannot tell us anything about the designer’s intentions or purposes any more than it can tell us anything about whether the designer prefers brown over blue. The "junk-DNA claim" seems to stem from the observation (by IDists) that when human designers design things, they don’t tend to put junk into their designs. We should then reasonably expect that the “intelligent designer” would not have put junk into the design of our DNA. So, the claim comes down to is that since we can predict, based on human behaviour that there should be little junk DNA, it also follows that ID would predict the very same thing. An observant reader might notice something iffy here – namely that ID does not say anything about the designer. More precisely, ID says nothing about the designer thinking and acting like a human and even if it did, it says nothing about the designer’s desires about junk DNA. In order for ID to predict something regarding the existence of junk in DNA or anything for that matter, it will have to add some form of assumption about the designer’s intents or purposes. That assumption would include something about the designer's desires to put junk into DNA. I.e. the conclusion would also be found in the assumptions - pure circularity. So, ID is just as happy to predict the non-existence of junk in DNA as it is happy to predict it's ample existence.

Let's now, as opposed to Luskin's circular argument, go full circle and return the reason for this post, namely Luskin's initial assertion:

Question: What do you do when a theory logically predicts both (a) and not (a)?
Answer: Apparently you heavily promote it.

Which theory does that sound like?

Edited to add: EVOLUTIONBLOG has also commented on Luskin's post.