Science and nonscience again

I just can't escape that damned Demarcation Principle...

A fellow emailed me the other day, asking what I thought about String Theory. Was it science? He was trying to argue with Intelligent Design folk, and they brought String Theory up as a case of science that doesn't have any testable evidence yet. He responded "science is what scientists do", and ask my opinion about that claim...

I responded thus [names removed to protect the innocent]

[Name], you are stepping in deep, very cold, and very dank waters.

In public, when trying to deal with soundbite science, it is worthwhile saying something like "science is testable", but what that means is not easy to describe, let alone define. There is no such thing as a universal scientific method or procedure. Some science is testable, and some isn't. Some science is empirical and some (theoretical cosmology) isn't. Some science is quantifiable, some isn't.

In philosophy there's a thing known as a "family resemblance predicate" in which something is a member of a family of things if it exhibits most of the properties of that group, but not necessarily all. Science is like that. ID fails to be science for many reasons, not least being the lack of an active research program and a failure to discriminate its explanation from one without ID involving only selectionist explanations. But there's no simple knockdown criterion for excluding it, and in fact, at one time ID was an active research program. Granted, that was in the 17th and 18th centuries, but that only goes to show that what counts as science at a particular moment is relative to the past history and present activities of science.

So, you are right in that "science is what scientists do", but there's more to the story than that. Science is (FRPishly) empirical, explanatory, theoretical, active, historically related cognitive activities. It is partly political, for it is done by humans, and they are political animals. It is partly social, for humans are social animals. But it is also engagement with nature, the parts of the world that aren't us, and that is where ID fails - it doesn't engage with nature, it doesn't do research. It just takes an apriori view, and trawls for ways to make that view seem scientific.

The motto of the Royal Society - the first such society in the world - is "Nullius in verba" - nothing in words. The motto of the ID crowd is "Omnia in verba" - it's all in the words.

String Theory sits uncomfortably at the edge. It is clearly a progressive research program in Lakatos' sense, but it has no empirical consequences at present, so it is at best a conceptual program only. It is within science, but if it doesn't make any progress, eventually it will be abandoned. So it is the best example of "science is what scientists do" because if scientists stop working on it, it will cease to be part of science just in virtue of that fact.

By the way, I have a short post on demarcation criteria here, and here. The latter got included in the volume of best science blogging of 2006.

If anyone wants to argue this, I'm happy. But if you correct my Latin, I'll sulk. I decline to argue that.