Tom Gilson at Thinking Christian has been hosting some interesting discussions about materialism and Intelligent Design (ID).
This post is taken from my response to Tom's defense of ID. Is ID's claim that it has predictions a true claim?
The first is that design has not been scientifically well-defined by ID advocates. Design in scientific context involves modeling of the physical world, planning to goals, and manufacturing. So design as we know it is intimately related to goals and utility.
You can see this in the SETI program. SETI looks for narrow-band transmissions because narrow band transmissions are 1) easy to find if the receiver is looking for them, 2) energy efficient and 3) useful for communications. (1) may say something about us, but (2) and (3) say something about ET, namely, that ET doesn't have an infinite resource of energy, and ET communicates using EM radiation because alternatives are cost-prohibitive. There's an implicit assumption that it is expensive for ET to travel through the universe. ET would hardly waste time beaming us prime numbers at Hydrogen-times-Pi if it could instantly transport directly to our planet and speak to us in English. We are making many naturalistic assumptions about ET before we begin to detect him/her/it.
ID is another one of those "third options" that theists love so much. Just as a naturalistic definition of determinism leaves only randomness as residue, a naturalistic definition of explanation leaves only unexplained things as residue. In science, you have to have naturalistic explanations.
Yet, the "tests" of ID proposed by ID advocates are that no natural explanation for life can be found. But this just means that life would be unexplained, or perhaps, improbable, if no probable natural explanations are found. There is no scientific meta-rule that says "ID is a default explanation unless you find an alternative." For ID to be a valid scientific theory, it has to make predictions of its own. It cannot just predict that everyone else's predictions won't pan out because that resorts to the illegal meta-rule.
I realize that ID advocates claim to make predictions, but true predictions have to follow from one's premises, and that's where things break down for ID. Without justified predictions, ID lacks explanatory power.
Let me give you an example. One so-called ID prediction is that there will be no junk DNA. Where does this prediction come from?
Evolutionary models don't say whether there will or will not be junk DNA. The models predict non-coding sequences, but don't leap to any conclusions about the regulatory function of non-coding sequences. However, it's likely that the cell environment is such that every non-coding sequence affects gene expression in at least some small way. Just like the non-functional ornaments in building architecture must subtly affect the people who work in and around those buildings.
But why does ID predict that junk has an effect? What is ID assuming about the designer that dictates the utility of junk DNA? Why does the designer care about the junk at all? Why can't the junk simply be a designed reserve of genetic material for future mutation? What, if anything, prevents the recurrence of junk since design occurred? Isn't it possible that junk DNA is a designed feature?
The ID web sites claim that their rationale is that a designer wouldn't create structures (our junk DNA) without their being functional. However, this is certainly not an inference from human design because human software applications generate plenty of junk and non-coding sequences, but remain functional due to their modularity. In contrast, DNA is non-modular spaghetti code. That is, not only does DNA not exhibit the hallmarks of human design, there's no reason why junk DNA could not have been designed to be junk (just like an aging defragmented software heap).
Of course, there are clear reasons why the perpetrators of the ID hoax would make this claim. Based on what we already know, it's a prediction that's certain to be true, even if only to some tiny extent.
The same goes for the other predictions of ID. They simply fail to follow from the premises because ID has no models with enough specificity to justify its predictions.