On the scientificity of string theory

I must emphasize that although I have a background in physics, by no means do I have expertise in particle physics or string theory. I was a condensed matter physicist, and an experimentalist not a theorist. I do not have any deep understanding of string theory beyond a general background knowledge. What I bring to the table is a bit of awareness of how physics research operates in practice, plus the cynicism that comes with the territory of being an ex-physicist.

String theory was essentially a scientific fad. I’m not going to go into the history, because I have no expertise on that, and the Wikipedia article is frankly opaque. Dr. Collier recently made a more accessible retrospective–although I find the video game irritating, and Dr. Collier is liable to get some things wrong.

The relevant part is that string theory was a fad in scientific research and a fad in popular science. The physicists were overly excited about it, and so was the public. Then there was backlash, which again occurred both among physicists, and in the public. String theory was criticized for failing to make any testable predictions. Peter Woit described the theory as “not even wrong” (and published a book with that title), because a theory could only earn the status of being wrong by making a prediction that was found to be false.

Today, long after all that went down, what I encourage among non-physicist readers is moderation. String theory isn’t exactly a success story, but in the end it’s still legitimate scientific research. Frankly, you probably shouldn’t have any opinion on string theory at all, and it was a mistake for science communicators to have ever encouraged you to have one.

Public opinion on scientific questions

I think it’s important to understand a social reality about science–scientists are trained to take sides. Sure, the scientific ideal is to neutrally adjudicate the evidence, but if you actually want to get paid or get published, you take a side. Specifically, it is a scientist’s job to argue and persuade fellow scientists (and grant providers) of the value of their own novel ideas.

In my opinion, this is the #1 bias in science. When the general public talks about bias among scientists, they think about scientists’ political leanings, or their failures to understand e.g. women’s issues. I’m sorry to say that the biggest bias in science is a lot more petty and self-serving. Scientists’ careers depend on them being able to demonstrate the value of their own contributions. These incentives are the root of the replication crisis as well as incidents of scientific fraud. Jonathan Pruitt didn’t fabricate results because he had a political commitment to some theory of spider behavior, it’s because he had a commitment to his own career success.

So what are the implications on scientific communication to the public? Usually the goal of communicating to the public is to teach what has been well-established in science. However, scientists commonly go beyond that to talk about the “cutting edge”, and are specifically biased towards talking about their own cutting edge. There are a lot of reasons for this. Sometimes, they want to impress upon people that science is an ongoing project (and one that requires public funding). Sometimes raising questions is just a way of intellectually engaging people. Sometimes the audience is mixed, and they’re actually hoping that the relevant experts in the audience will learn about their work. Sometimes it’s just about social status.

Whatever the reason, it can be a bit dangerous. The general public can’t necessarily tell where established science ends and self-advocacy begins. When a scientist talks about a scientific controversy, and tries to pass it off as if they’re totally winning, the general public does not have the tools to be skeptical about it, or even recognize what is happening.

And that’s basically how I see string theory. It was a mistake to communicate it to the public like it was for sure the next big thing. And the public, not knowing better, ate it up. Then the critics came along, and I don’t blame them at all for going to the public, because string theory was already there in the minds of the public, demanding a response. But if you asked me what is the ideal state, I think that the general public should not have any opinion on string theory one way or another.

Let me make a comparison to my field. As a member of the general public, do you have any opinion on the theory of cuprate superconductivity? What do you think is the explanation for the pseudogap observed in such materials? You don’t have an opinion on the subject, because you probably don’t know what a cuprate is, and almost certainly do not understand what a pseudogap is. I published papers about the pseudogap and even I cannot assert an opinion. You shouldn’t have an opinion either. Of what benefit is it to form an opinion on the solutions to a scientific problem that you don’t even understand?

Science communicators have encouraged you to think about and form opinions about string theory. And maybe that was successful at engaging people or whatever, but. If you don’t have the expertise, you know you’re just play-acting, right? You can’t really form a meaningful opinion on string theory.


I think part of the reason lay people felt entitled to having opinions against string theory, is because the main criticism is on a rhetorical level that they can understand. You might be familiar with falsifiability, which was used by philosopher Karl Popper to address the demarcation problem, i.e. distinguishing science from non-science. The argument is that string theory is unfalsifiable and therefore unscientific, and that’s an argument simple enough for anyone to form an opinion about it.

I’d like to complicate this narrative by talking about how science works in practice. You might be familiar with the scientific method, often taught to kids in flow chart form. You make an observation, form a hypothesis to explain it, test it with an experiment, and draw conclusions. If you think about this for a moment it’s obviously greatly oversimplified. In particular, I observe that science is a collaborative effort. You don’t necessarily go through the whole flow chart personally. You have lots of people who specialize in just one tiny piece of the flow chart, or do stuff that isn’t listed there.

One of the biggest divides in physics is between theory and experiment. Theorists try to come up with explanations for observations, and use math or simulations to understand the implications of a theory. Experimentalists design instruments, perform experiments, and analyze results.

By the way, it’s often explained in popular science that there is a difference between a “hypothesis” (an untested idea) vs a “theory” (a well-tested framework). Often the main purpose is to explain why evolution isn’t “just” a theory, but is actually an extremely well-established and foundational framework. And that’s true about evolution, but it’s just not an accurate description of how “theory” is used by scientists. Scientists use “theory” in multiple ways (including the colloquial “just a theory” sense), but in physics it most frequently refers to the distinction between theory and experiment. So when we talk about string theory or, say, superconductivity theory, that’s what we mean. It’s the distinction between math and concepts vs instruments and observations. It does not typically refer to the distinction between tested and untested–we would never hear the end of arguments about it if that’s what it meant.

The point I’m getting to, is untestable theories are rather common in physics. From the perspective of a theoretical physicist, it’s one of the central challenges. What prediction can you make that is within practical reach of the experimentalists? How do you persuade the experimentalists to actually devote resources to this? Once the result has been found, how do you persuade other scientists that you actually succeeded in confirming/disconfirming your theory?

My experience comes from being on the other side. I’d listen to theorists say, “all you need to do is perform this experiment to test my theory” and we’d tell them that we’d already done that and hadn’t seen anything like that, and they’d tell us we needed to look harder, and we’d explain that we could look a tiny bit harder but there are physical constraints. And then if we do see something or don’t see something, I’m not necessarily convinced, and it certainly doesn’t convince our scientific rivals who have other ideas. And it goes on and on. This is basically the steady state for unresolved problems in physics.

When people talk about a theory being unfalsifiable, it comes off as scandalous. It’s like, don’t these smart people know about falsification, like what we learned in high school? But really, in many scientific controversies, whether theories are testable or not is precisely the subject of the controversy. Each theorist has their own ideas about how to test their theories, and the situation is that other scientists just aren’t buying it.

So the whole thing about string theory being unfalsifiable… that’s never what the problem was. The problem was and is one of magnitude. Critics argue that this theory has been around for decades and decades with so many people in the field, without any testable predictions. I’d bet that every string theorist has an idea or ten about how to test the theory, because they know it’s important, it’s just unclear whether any of these ideas hold any promise whatsoever.

I mean, it sure doesn’t look good for string theory. But I think it’s a mistake to think of it as a Popperian demarcation problem. Contra Popper, untestable theories are a natural problem occurring in the normal course of science. I think particle physics is particularly vulnerable because of its reliance on huge experiments, and the sheer complexity of the theory. And maybe it really is untestable like the critics say.

Either way, string theory does not look like non-science. At worst, it’s science that performed very poorly.

Negative results

Scientific success is by its very nature not entirely up to the scientist. When a scientist has an idea, it is not the scientist who decides whether that idea works out or not; it’s the world that decides. But there are also some ideas that are unpromising from the start. This poses a problem, because a negative result could be a good idea that simply didn’t pan out, or it could reflect negatively on the scientific project. And so we have the file drawer effect, a bias where researchers are less likely to publish negative results than positive results. After all, scientists don’t want to publish something that may reflect poorly on how they spent their grant money, and peer reviewers may not want to see that either.

Although the file drawer effect is usually discussed in the context of statistics-heavy fields like psychology, I have a feeling that it’s much more severe in physics.  People just don’t talk about it, because calling it a file drawer effect implies something going wrong, and yet in physics it feels so justified to bury negative results. The space of unpromising physics ideas is so vast and unstructured, you can’t just test every idea and pretend that you did good work when every result comes back negative. A negative result could mean you’ve disproven a theory, but it’s a lot more likely that there was an experimental problem, or the experiment was not powerful enough, or the theory was never testable by that method.

I’d like to bring this into the context of theoretical physicist. Just as an experimentalist doesn’t know the result until they perform the experiment, a theorist can’t know the predictions until they do the math. Sometimes, an experiment doesn’t play out and the result is negative. And sometimes, a theory doesn’t play out, and no testable predictions pop out of it. An untestable theory is analogous to a negative experimental result. And that can be hard to publish because people have no concept of a file drawer effect in theoretical physics. People think if a theory is untestable it’s just bad.

So the question is, were string theorists unlucky? Or was it just never good to begin with? I really don’t know, I don’t have an opinion.

Personally, I think of scientific research as a bunch of lottery tickets. Very little research does anything valuable, until something does. I’m definitely coming at this from a cynical personal angle, because I came away from my scientific career feeling that I had done nothing of value. I published some novel ideas and made legitimate arguments, but the opposition would not have been persuaded by my work, and it did not go anywhere. Also, the entire field of cuprate superconductivity has virtually no applications, with actual applications using low-temperature superconductors. And that doesn’t mean it was bad to do the work, it just means we didn’t win.

So my take is that string theory is probably not valuable, but that doesn’t make it stand out. To be honest, I do not think string theory would be very valuable even if it were confirmed. Theories like that, they mainly fill the human need for a cosmology, an explanation of the universe. That makes it exciting, and the subject of a great deal of popular science, for better and for worse. But you do not need such a thing to live, and you do not need to have an opinion about it.


  1. Dunc says

    Also, it seems that quite a lot of people (on the internet, anyway) feel the need to have strong opinions about everything these days.

    Who knows, maybe in a few years, or decades, or whenever, somebody will come back to it with some new ideas and show that not only was it testable all along, but it actually explains a bunch of othewise problematic results that nobody had seen any connection between. Or maybe we’ll come up with some new paradigm that compeltely upends our understanding of fundamental physics in a way that makes all of our current ideas (including GR, QM, and the Standard Model) look like classical elementalism…

  2. Rob Grigjanis says

    A refreshing read.

    I came away from my scientific career feeling that I had done nothing of value.

    I think my attitude is a bit different from yours. My own work didn’t amount to much (if anything!) in the grand scheme of things, but I don’t care. I just loved doing it, and learning the background I needed to do it. I suspect I’d feel the same even if I’d worked in string theory.

  3. says

    As a layman, I do have an opinion on string theory: I find it interesting, but not compelling. I’ve read several popular works on the subject, and each time have had a similar experience. It starts out seeming like a possible breakthrough in understanding the universe, but by the end begins to strike me as an elaborate mathematical edifice that may or may not have any relationship to physical reality.
    Which may not be what most people would consider an “opinion.”

  4. flex says

    Of course there is the possibility that the vast majority of people didn’t express an opinion on the validity of string theory because they knew they didn’t know enough to have an opinion. The media published the views of people who had strong opinions, one way or the other, the remainder, the undecided, the rest of us, our opinions went right into the file drawer. 😉

    I don’t often comment on your posts because I know I have nothing to say, but….

    I came away from my scientific career feeling that I had done nothing of value.

    I spent the last 30 years selling my time/expertise to someone else so they could make more money. Oh, I did necessary work on meeting customer expectations and making cars safer, and I even had times when I enjoyed myself. I’m not upset or ashamed that I didn’t pursue a more scientific career, I took the money. Yet, while I don’t know that there is really a hierarchy of value, but adding to the sum of human knowledge even negatively has to rank higher than selling your talent to capitalism.

  5. Lassi Hippeläinen says

    IMHO it all boils down to the difference of Platonian vs. Aristotelian philosophy. Math is about ‘abstract’ ideas, and phys about ‘real’ life. Mathematicians will accept as a theory things that physicists won’t accept even as hypotheses. String theory is one of those.

    I don’t mean that is being in bad company. There have been mathematical conjectures that have produced truckloads of good science, even when the connection to real life is spurious at best. Fermat, Goldbach, Riemann…

  6. says

    I don’t think that characterization is correct. String theory isn’t pure math, and you’ve been misled if you think it is. String theorists are not refusing on principle to generate any real-world predictions. They agree that making predictions is important! It’s just that making predictions is difficult.

    I didn’t get into this, because I’m not confident on my knowledge of the science–but the reason string theory is difficult to test is because of renormalization theory, the theory of how physics transforms as we “zoom out”. Different physics on the micro scale will look the same on the macro scale. (Renormalization theory is also used extensively in condensed matter physics.) In principle you could get a powerful enough probe and just see what’s going on in the micro scale. But you can’t do that in practice, so the question is if there’s some clever workaround.

    The most basic workaround is to predict what particles should exist. But there’s the issue of the model having too many free parameters, so it’s unsurprising that it can retrodict known particles–and predictions of unknown particles are ambiguous. Another route to test the theory is through astronomical observations, since the early universe effectively functions as a very powerful (but uncontrolled) probe.

  7. says

    Not a physicist, but as an engineer I’ve formed many a hypothesis and tried to confirm them with experiments when troubleshooting production problems. The times this yielded a clear and definitive answer are few and far between. Intuition and experience often play a larger role than many care to admit.

    In practice the world is messy and variable, and trying to isolate what you are interested in is often impossible or impractical.

    Engineers have been taught that Design Of Experiments is a good way to try and find causes of unstability/variation in production processes. The dirty little secret is that a stable process is required for the results of the experiment to yield meaningful data.

  8. Rob Grigjanis says

    Lassi @6: It seems to be a common misconception that string theory is just a fanciful mathematical exercise, but its history is motivated by observation, going back to the 1960s.

    Back then, more and more seemingly fundamental particles were being discovered, and it was an exercise to classify them. The quark model helped; the plethora of particles (the ‘particle zoo’) could be classified partly by their quark content. But there were still multiple particles with the same quark content, but different spin (intrinsic angular momentum) and mass. One tantalizing idea was that mesons were a quark-antiquark pair joined by a string. Different masses and spin were explained by the pair rotating around each other at different speeds, resulting in greater tension is the string.

    That picture ultimately failed, and was replaced by quantum chromodynamics (QCD). But studying the mathematics of this proto-string theory raised some intriguing possibilities. For example, you could find a vibrational state of a string corresponding to a spin 2, zero mass particle. A candidate for the graviton?

    Anyway, that pretty much exhausts my knowledge of the subject.

  9. Lassi Hippeläinen says

    @7 and 9: I don’t mean to claim that string theory is just a mathematical excercise. It is still work-in-progress. It may lead to something yet, but that remains to be seen. An insider can see goals of making connections to ‘real world’ physics, but we outsiders haven’t seen any results. So thus far it has been seen as pure math, and its scientificity judged as such.

  10. Ichthyic says

    I can look back on my research and definitely say the papers I published ended up not answering ANY of the questions I actually wanted to have answers to when I started.

    but the journey was bloody brilliant. took me to places I had no right to be, and sometimes to some rather scary adventures as well. I just wish I could have made it pay somehow.

  11. says

    @Lassi #10,
    That’s the viewpoint I was arguing against in the OP.

    @Ichthyic #11 & @Rob Grigjanis #2,
    I’m happy to hear accounts where people felt they enjoyed their work in science. Personally, it got me down how much of a gap there was between the alleged social value of scientific research, and the apparent low value of what we were actually doing. There was a lot of urgency to produce a positive result, which I found demoralizing. I do think it was an interesting problem though, and I’m still proud of my work. I just don’t think it did anything for anyone (but me, since I got a degree out of it).

  12. billseymour says

    I was never a scientist.  I think it’s interesting and I’ve read a few popular works—most recently Mano’s The Great Paradox of Science—but my main takeaway has always been how difficult it is to actually know something, and how easy it is for all of us to mislead ourselves.

    Like you, I retired without thinking that I had contributed much, and I won’t be remembered for any of it.

    In my first vocation as an electronics technician, I wired up a control room for an educational FM station in Pittsburgh, designed and built a box that would allow a commercial radio station in Greensburg, PA to control its (automated) FM station at the transmitter site from its AM control room downtown, and designed and built an electrocardiograph with an early microprocessor inside for a hospital in Pittsburgh.  Everything I did was already well understood; I just figured out how to put the pieces together.

    I later switched to being a computer programmer and recently retired after 31 years with the U. S. Postal Service.  I was a small fry on a team that was responsible for a system that connected desired movements of mail with available transportation.  If our system broke, the mail stopped moving; and the mail didn’t stop moving.  You’re welcome…I guess. 😎

    But I always enjoyed my work, always took great pleasure in doing it well, and towards the end, was pretty well paid for it; and I think that I’ve done my small bit to help society just a little even though I certainly won’t be celebrated by posterity.  That’s good enough. 😎

Leave a Reply

Your email address will not be published. Required fields are marked *