I’d like to preface this by saying that this post will be a bit more controversial than usual. I have somewhat unconventional opinions about the nature and purpose of science, and what I say below shouldn’t be taken as representative of the field in general.

A bit more than a week ago, Not Even Wrong had a post on the Fundamental Physics Prize. Peter Woit is often…I’m going to say annoying…and this post was no exception.

The Fundamental Physics Prize, for those not in the know, is a fairly recently established prize for physicists, mostly theoretical physicists.  Clocking in at three million dollars, the prize is larger than the Nobel, and is currently the largest prize of its sort. Woit has several objections to the current choice of award recipient (Alexander Polyakov). I sympathize with some of these objections, in particular the snarky observation that a large number of the awardees are from Princeton’s Institute for Advanced Study. But there is one objection in particular that I feel the need to rebut, if only due to its wording: the gripe that “Viewers of the part I saw would have no idea that string theory is not tested, settled science.”

There are two problems with this statement. The first is something that Woit is likely aware of, but it probably isn’t obvious to everyone reading this. To be clear, the fact that a certain theory is not experimentally tested is not a barrier to its consideration for the Fundamental Physics Prize. Far from it, the purpose of the Fundamental Physics Prize is precisely to honor powerful insights in theoretical physics that have not yet been experimentally verified. The Fundamental Physics Prize was created, in part, to remedy what was perceived as unfairness in the awarding of the Nobel Prize, as the Nobel is only awarded to theorists after their theories have received experimental confirmation. Since the whole purpose of this prize is to honor theories that have not been experimentally tested, griping that the prizes are being awarded to untested theories is a bit like griping that Oscars aren’t awarded to scientists, or objecting that viewers of the Oscars would have no idea that the winners haven’t done anything especially amazing for humanity. If you’re watching the ceremony, you probably know what it’s for.

Has this been experimentally verified?

The other problem is a difference of philosophy. When Woit says that string theory is not “tested, settled science” he is implying that in order to be “settled science”, a theory must be tested, and while I can’t be sure of his intent I’m guessing he means tested experimentally. It is this latter implication I want to address: whether or not Woit is making it here, it serves to underscore an important point about the structure of physics as an institution.

Past readers will be aware that a theory can be valuable even if it doesn’t correspond to the real world because of what it can teach us about theories that do correspond to the real world. And while that is an important point, the point I’d like to make here is a bit more controversial. I would like to argue that pure theory, theory unconnected with experiment, can be important and valuable and “settled science” in and of itself.

First off, let’s talk about how such a theory can be science, and in particular how it can be physics. Plenty of people do work that doesn’t correspond to the experimentally accessible real world.  Mathematicians are the clearest example, but the point also arguably applies to fields like literary analysis. Physics is ostensibly supposed to be special, though: as part of science, we expect it to concern itself with the real world, otherwise one would argue that it is simply mathematics. However, as I have argued before, the difference between mathematics and physics is not one of subject matter, but of methods. This makes sense, provided you think of physics not as some sort of fixed school of thought, but as an institution. Physicists train new physicists, and as such physicists learn methods common to other physicists. That which physicists like to do, then, is physics, which means that physics is defined much more by the methods used to do it than by its object of study.

How can such a theory be settled, then? After all, if reality is out, what possible criteria could there be for deciding what is or is not a “good” theory?

The thing about physics as an institution is that physics is done by physicists, and physicists have careers. Over the course of those careers, those physicists need to publish papers, which need to catch the attention and approval of other physicists. They also need to have projects for grad students to do, so as to produce more physicists. Because of this, a “good” theory cannot be worked on alone. It has to be a theory with many implications, a theory that can be worked on and understood consistently by different people. It also needs to constrain further progress, to make sure that not just anyone can create novel results: this is what allows papers to catch the attention of other physicists! If you have all that, you have all of the relevant advantages of reality.

String theory has not been experimentally tested, but it meets all of these criteria. String theory has been a major force in theoretical physics for the past thirty years because it can fuel careers and lead to discussion in a way that nothing else on the table can. It has been tested mathematically in numerous ways, ways which demonstrate its robustness as a theory of quantum gravity. In this sense, string theory is a prime example of tested, settled science.