I’ve spent a bit of time recently being drawn back into discussions with creationists. It is always soul destroying coming up against yet another earnest believer channelling the same creationist lies from the same few sources. It is so frustrating trying to have a conversation about biology with folks who’s knowledge is so superficial that they have no real clue what scientists actually say, or why. Frustrating because, inevitably, such people are so convinced they are right and scientists are fundamentally stupid or ideologically blinded, that they refuse to actually learn any science to have the discussion. Invicible Ignorance. Bleagh. It is no wonder that so many Christians find creationism so embarrassing.
One of the recent tactics of the institutions that peddle creationism is the idea that evolutionary science is no more than a different interpretation of the same data. If you start with the assumption there is no God, you can read evolution into the evidence. Whereas if you start with the true knowledge of God, then the evidence clearly points to a young earth (or an old earth, depending on which creationist mill you get your information from — they disagree with one another about the details).
Science, and empiricism generally, works a bit differently.
Empiricism deals with “differentiable hypotheses”. These are simply explanations which can be told apart. All explanations have consequences, and differentiable hypotheses will have some consequences that are different from one another. Given two possible explanations, we work through until we find consequences that differ. Then we can go and check which one was right.
Often one of the hypotheses will be the “null hypothesis”, which is simply a way of saying that the other hypothesis is wrong. So if the hypothesis is that “Bill burgled the house.”, then the null hypothesis will be “Bill did not burgle the house.” We ask what the consequences of each would be, and find some situations in which they would be different. Those situations are evidence.
Note that this is predictive. You have to predict what the consequences will be. All empiricism is predictive. Not predictive of what will happen in the future, necessarily (we can do science on things in the past). But predictive of the consequences of things being true.
Explanations that lose in this process don’t tend to stay still. They evolve to explain the confounded expectation. Even good science makes duff predictions sometimes, and needs to adapt to bring confounded expectations into the theory. So having unexpected results doesn’t make something wrong per se. But the way those adaptations occur is a good way to tell between good and bad explanations.
As I pointed out previously, a hypothesis that keeps getting it wrong ends up accumulating explanations that get more and more far fetched, and which themselves have consequences that aren’t or can’t be verified. As per my game about contradictions, it is always possible to explain away any failure.
A good theory, however, moves forwards with its adaptations, explaining more of the data and providing more opportunities to check.
The classic creationist example is gradualism. The original expectation was that evolution happens at a fairly steady rate. We’d see gradual changes in phenotype over time. This prediction was wrong, as creationists love to point out. So this needed explanation. It was explained, through a series of careful studies in the 70s, and verified in mathematical models of evolution in the 80s and 90s. The new explanation had huge consequences, which we could check, and which drove evolutionary theory forwards considerably: feeding into work on neutral networks, regulation, and evodevo.
On the other hand, I’ve not seen anything in the creationist works I’ve read where one of their adaptive explanations is then analysed for its consequences and those consequences are then pursued and verified. The nearest you could get is the creationist prediction that there is no junk DNA, which then claim the the recent ENCODE papers as justification. But this only works at the level of quote-mining ENCODE, and relies on a serious misunderstanding of what ENCODE actually found (and even then, the results ENCODE reporting are very likely to be unreliable). Other predictions, such as irreducible complexity, have proven to be wrong (so in the manner of these things, the definitions have got more complex and the consequences remain hypothetical and unchecked). That’s the definition of pseudoscience, in my mind.
The purpose of this post is to have somewhere to refer to when this comes up. I’d love any thoughts on it.
 It isn’t clear how wrong this prediction was. Creationists love to oversell the lack of gradualism. And I’ve read some reaction pieces that seem to oversell the gradualism observed, or play down the expectation that evolution should be gradual. The actual patterns of expectation and reality are a lot more fuzzy than I’m suggesting here.