Bret Underwood, a friend of mine from my time in Madison, WI, saw my post on String Theory, and took issue with my statement that it wasn’t testable. I’m still standing behind what I said, but let’s address what Bret has to say.

I don’t understand your argument above for why string theory is untestable. In fact, it seems to me you just outlined the best possible case for string theory! What you said above is that if I have a string theory construction of a phenomenon (say, the Standard Model or Inflation), which uses a set of parameters X, and makes some predictions, then I can find another set of parameters Y that gives a different set of predictions. Wonderful! This means I can determine the parameters (either X or Y) by making measurements, and rule out models and parameter sets!

I don’t think that’s wonderful at all; I think that’s a very dangerous analogy. Why? Because string theory has too much “wiggle room” to be scientific. String theory isn’t at all predictive in this sense. How many possible values for the string vacua are there at this point? The typical estimate that people cite is about 10^{500}, according to the string landscape. For comparison, the number of subatomic particles in the entire Universe is somewhat less than 10^{91}. There are far fewer than 10^{500} parameters describing our entire Universe, so people argue that one of those models is likely to match with the Universe. And when we find out which one it is, then we can figure out everything.

But that isn’t a scientific theory to me, not as I understand science. A scientific theory makes definitive predictions *that are unique*. But string theory doesn’t do that; at least, not for any prediction that I know. Most models of string theory predict a negative cosmological constant, for instance. But we don’t say that this means string theory is wrong, we say that this rules out those models, because we see a positive one. String theory gives you a 10-dimensional Brans-Dicke theory of gravity; we observe a 4-dimensional one with a Brans-Dicke parameter of infinity. But somehow, we don’t say string theory is wrong, we say that we just need to get rid of 6 of those dimensions and make that parameter be infinite. We don’t say how, we just say somehow. Most models of string theory predict small tensor modes, but most models of inflation do, too. What does string theory predict **that’s unique to string theory**? Other than “everything is made of strings,” I don’t know of one. Until there is one, and one that can be tested, I can’t be comfortable calling it a scientific theory.

The nightmare scenario for me is the following: suppose I describe a phenomenon (Standard Model or Inflation) in string theory with a set of parameters X, which gives some predictions. Then I consider another different set of parameters Y that give the same predictions. Thus, predictions are not unique, so I cannot distinguish models!

But what’s even the point of distinguishing models if there’s no observation I can make or experiment I can do that *validates* string theory? Maybe I can find one (if you’re lucky) or more than one (if your nightmare comes true) set of parameters that agree with all the laws of physics in our Universe. But what can you tell me that I don’t already know? That’s what I mean by untestable.

I’m not saying that people who are interested in this shouldn’t work on it. But I’m saying that if string theory is going to bill itself as being science as opposed to mathematics, it needs to address the issue of “what does it predict that nothing else predicts?” I don’t know of any test that’s ever been devised, even in principle, that can test it. Do you?

**FYI:** I downloaded a talk by Michael R. Douglas at Rutgers, a string theorist who’s actually optimistic about string theory being a predictive theory. He asks the question: “Are there testable predictions of string theory?” He says yes initially, but then admits the following:

…none of the ideas which have been suggested so far are guaranteed signatures of string theory.

Yikes. Sorry, Bret, but that means I place it into the category of *an untestable hypothesis* at the present time. If you can’t validate it and you can’t falsify it, it isn’t yet a good scientific theory.