They should have asked Ethan Siegel. Because a few days ago he strayed from the path of awesome news about the universe to inform his readership that “String Theory is not Science.” Unlike Davide however, Ethan has not yet learned the fine art of not expressing opinions that marks the true science writer. And so Ethan dismayed Peter Woit, Lubos Motl, and me in one sweep. That’s a noteworthy achievement, Ethan!Upon my inquiry (essentially a polite version of “wtf?”) Ethan clarified that he meant string theory has no scientific evidence speaking for it and changed the title to “Why String Theory Is Not A Scientific Theory.” (See URL for original title.)
Now, Ethan is wrong with believing that string theory doesn’t have evidence speaking for it and I’ll come to this in a minute. But the main reason for his misleading title, even after the correction, is a self-induced problem of US science communicators. In reaction to an often raised Creationist’s claim that Darwinian natural selection is “just a theory,” they have bent over backwards trying to convince the public that scientists use the word “theory” to mean an explanation that has been confirmed by evidence to high accuracy. Unfortunately, that’s not how scientists actually use the word, have never used it, and will probably never use it.
Scientists don’t name their research programs following certain rules. Instead, which expression sticks is mostly coincidence. Brans-Dicke theory, Scalar-Tensor theory, terror management theory, or recapitulation theory, are but a few examples of “theories” that have little or no evidence speaking in their favor. Maybe that shouldn’t be so. Maybe “theory” should be a title reserved only for explanations widely accepted in the scientific community. But looking up definitions before assigning names isn’t how language works. Peanuts also aren’t nuts (they are legumes), and neither are Cashews (they are seeds). But, really, who gives a damn?
Speaking of nuts, the sensible reaction to the “just a theory” claim is not to conjure up rules according to which scientists allegedly use one word or the other, but to point out that any consistent explanation is better than a collection of 2000 years old fairy tales that are neither internally consistent nor consistent with observation, and thus an entirely useless waste of time.
And science really is all about finding useful explanations for observations, where “useful” means that they increase our understanding of the world around us and/or allow us to shape nature to our benefits. To find these useful explanations, scientists employ the often-quoted method of proposing hypotheses and subsequently testing them. The role of theory development in this is to identify the hypotheses which are most promising and thus deserve being put to test.
This pre-selection of hypotheses is a step often left out in the description of the scientific method, but it is highly relevant, and its relevance has only increased in the last decades. We cannot possibly test all randomly produced hypotheses – we neither have the time nor the resources. All fields of science therefore have tight quality controls for which hypotheses are worth paying attention to. The more costly experimental test of new hypotheses becomes, the more relevant is this hypotheses pre-selection. And it is in this step where non-empirical theory assessment enters.
Non-empirical theory assessment was topic of the workshop that Davide Castelvecchi’s SciAm article reported on. (For more information about the workshop, see also Natalie Wolchover’s summary in Quanta, and my summary on Starts with a Bang.) Non-empirical theory assessment is the use of criteria that scientists draw upon to judge on the promise of a theory before it can be put to experimental test.
This isn’t new. Theoretical physicists have always used non-empirical assessment. What is new is that in foundational physics it has remained the only assessment for decades, which hugely inflates the potential impact of even smallest mistakes. As long as we have frequent empirical assessment, faulty non-empirical assessment cannot lead theorists far astray. But take away the empirical test, and non-empirical assessment requires utmost objectivity in judgement or we will end up in a completely wrong place.
Richard Dawid, one of the organizers of the Munich workshop, has, in a recent book, summarized some non-empirical criteria that practitioners list in favor of string theory. It is an interesting book, but of little practical use because it doesn’t also assess other theories (so the scientist complains about the philosopher).
String theory arguably has empirical evidence speaking for it because it is compatible with the theories that we know, the standard model and general relativity. The problem is though that, for what the evidence is concerned, string theory so far isn’t any better than the existing theories. There isn’t a single piece of data that string theory explains which the standard model or general relativity doesn’t explain.
The reason many theoretical physicists prefer string theory over the existing theories are purely non-empirical. They consider it a better theory because it unifies all known interactions in a common framework and is believed to solve consistency problems in the existing theories, like the black hole information loss problem and the formation of singularities in general relativity. Whether it is actually correct as a unified theory of all interactions is still unknown. And short of a uniqueness proof, no non-empirical argument will change anything about this.
What is known however is that string theory is intimately related to quantum field theories and gravity, both of which are well-confirmed by evidence. This is why many physicists are convinced that string theory too has some use in the description of nature, even if this use eventually may not be to describe the quantum structure of space and time. And so, in the last decade string theory has become regarded less as a “final theory” and more as mathematical framework to address questions that are difficult or impossible to answer with quantum field theory or general relativity. It yet has to prove its use on these accounts.
Speculation in theory development is a necessary part of the scientific method. If a theory isn’t developed to explain already existing data, there is always a lag between the hypotheses and their tests. String theory is just another such speculation, and it is thereby a normal part of science. I have never met a physicist who claimed that string theory isn’t science. This is a statement I have only come across by people who are not familiar with the field – which is why Ethan’s recent blogpost puzzled me greatly.
No, the question that separates the community is not whether string theory is science. The controversial question is how long is too long to wait for data supporting a theory? Are 30 years too long? Does it make any sense to demand payoff after a certain time?
It doesn’t make any sense to me to force theorists to abandon a research project because experimental test is slow to come by. It seems natural that in the process of knowledge discovery it becomes increasingly harder to find evidence for new theories. What one should do in this case though is not admit defeat on the experimental front and focus solely on the theory, but instead increase efforts to find new evidence that could guide the development of the theory. That, and the non-empirical criteria should be regularly scrutinized to prevent scientists from discarding hypotheses for the wrong reasons.
I am not sure who is responsible for this needlessly provocative title of the SciAm piece, just that it’s most likely not the author, because the same article previously appeared in Nature News with the somewhat more reasonable title “Feuding physicists turn to philosophy for help.” There was, however, not much feud at the workshop, because it was mainly populated by string theory proponents and multiverse opponents, who nodded to each other’s talks. The main feud, as always, will be carried out in the blogosphere...
Tl;dr: Yes, string theory is science. No, this doesn’t mean we know it’s a correct description of nature.