A Crisis at the (Western) Edge of Physics

The New York Times had an op-ed piece this weekend by Adam Frank and Marcelo Gleiser, entitled A Crisis at the Edge of Physics. They make some of the usual criticisms of string theory and the multiverse, ending with

Are superstrings and the multiverse, painstakingly theorized by hundreds of brilliant scientists, anything more than modern-day epicycles?

I mostly agree, although I don’t think they make clear what the real problem is, that these theories predict nothing and explain nothing. In contrast, epicycles were a quite useful, well tested model that was highly predictive and approximately correct. If we had modern day epicycles, that would be a huge advance…

Last Friday in his concluding talk at a Nordita conference on particle physics and cosmology, Michael Turner gave his take on the multiverse:

Most important discovery since Copernicus?
Is it science? (not testable)
Many true believers (left coast) and not enough doubters.

He makes clear his opinion on these questions with this graphic:


and I think this expresses well the majority opinion of the physics community. A major question here is whether the problem of pseudo-scientific multiverse mania is one of the “edge” of physics (the “left edge”, as Turner notices and was discussed here), or whether it has infected the center. Some days I’m quite discouraged to see how widespread this is, other days it seems to me that we may finally be getting over this. There’s only so long you can get media attention for your empty but easy to understand new “Copernican revolution” before people lose interest and move on to something else. Perhaps we’re getting to that point. I think this was the first year that the World Science Festival here in New York didn’t have a program promoting the multiverse, and maybe that’s a sign of change.

For quite a few years now, there have been few scientific talks trying to use a multiverse to do calculations at serious string theory conferences (see for instance this week’s String Pheno 2015, or Strings 2015 later this month), with the multiverse mainly appearing in promotional talks to the public. Maybe the public is finally getting bored and starting to adopt the point of view that Turner’s graphic suggests (and that I think the physics community should get behind).

Update: Physics Today has an opinion piece entitled Could the evolution of theoretical physics harm public trust in science? This addresses an issue I don’t think some theorists realize the seriousness of. If you start arguing that conventional notions of testability don’t matter, this can be a very dangerous thing to do in an environment where public trust in science is an issue. Put differently, if physicists publicly promote the pursuit of speculative ideas in an ideological framework that can never be falsified, they create a real danger of a public perception that science is just one more ideology.

This entry was posted in Multiverse Mania. Bookmark the permalink.

23 Responses to A Crisis at the (Western) Edge of Physics

  1. Pingback: A Crisis at the (Western) Edge of Physics | &Ph...

  2. Jeff M says:

    Honestly, epicycles were not “predictive” in any meaningful way. They were just a fancy geometric way of building a table of planetary positions based on past positions. The geometry was there simply because it was clear there had to be some geometric explanation for the positions. But they made no predictions, in the sense that if someone found a new plant, epicycles would tell you nothing about where it should be. Of course, if you want you can consider epicycles to be the correct theory of the solar system, since it’s just a change of coordinates.

  3. gadfly says:

    I would say homeopathy is worse than multiverse mania, but at least the placebo effect and therapeutic “consultations” actually help people.

  4. Eric Habegger says:

    Seriously, is it really necessary to characterize the multiverse mania as a West Coast phenomena? I seem to recall that there is a prominent faculty member at Columbia who has written many books that popularized the whole multiverse idea. The idea of one coast being the fount of misinformation does a disservice to all interested parties and also promotes a “new” source of prejudice that will cause problems down the line. I’m from the West Coast and I’ve never entertained the multiverse as anything serious.

  5. Peter Woit says:

    Good point about Columbia. There’s also someone at MIT… Still though, I think there’s a statistically significant effect, was interested to see that Turner shares that perception.

  6. Jeff M says:

    Oh, and BTW, there’s a previously unprintable word missing between “I” and “believe” in that great sign 🙂

  7. @Jeff M: “Epicycles”. Half in jest, presumably the AIC (or its variants, including Watanabe’s WAIC) would penalize the epicycles in comparison to an ellipse for the extra parameters required. An interesting question is whether or not the ellipse would lose in favor of epicycles because of poor observational accuracy, assuming those observational errors were not, at the time, well understood or estimated.

    K. P. Burnham, D. R. Anderson, Model Section and Multimodel Inference, 2nd edition, Springer, 2002.

    S. Konishi, G. Kitagawa, Information Criteria and Statistical Modeling, Springer, 2008.

  8. Aren’t epicycles really just a Fourier expansion of the apparent motion in the earth’s frame of reference? The issue is then a poor choice of coordinates making life hard 🙂

  9. Bee says:

    I’m not sure that this push towards phenomenology is useful at all, and I say that as someone who normally preaches the value of phenomenology. Look, on the one hand there’s people who couldn’t care less whether their theories are testable at all. That’s bad of course. But on the other hand (should I say other coast?) are people who cook up models just because they can be tested with the next experiment, and I can’t see there’s more value in that. Lisa Randall’s talk (at the same conference) was an explicit example for that. That’s searching under the lamp post. Does any one really believe a model becomes more plausible because it can soon be testable? No. So why do they do it? Because it will get published and when it doesn’t get found one twiddles some parameters and moves the “phenomenology” to next decade’s experiments. That’s equally sick if you ask me.

  10. chris says:

    Those of you who think epicycles were not predictive should read Keplers Astronomia nova and look at how he got to the Ellipse shape. Peter Woit is spot on with his analysis.

    Oh, and I don’t believe in science. I wonder who does. I simply accept proven statements about nature as being true. Pictures as the one above are a real disservice to science, putting it on the same footing as religion.

  11. Dom says:

    The word “believe” is natural to use but highly problematic, when challenged I once said I believed in facts and evidence, this was dismissed by the other person as simply “another belief system” as if the alternatives were all equally valid.

  12. Peter says:

    Epicycles are essentially a Fourier expansion.

  13. Peter Woit says:

    At least in particle theory, my impression is that the hiring prejudice in favor of phenomenology is now changing, as departments decide that the LHC is not going to provide a lot of new physics. The pressure on “string theorists” to produce something “testable” has now been there for a long time, being met by claims that AdS/CFT is going to solve non-HEP problems. Unclear to me where this is going…

  14. Peter Erwin says:

    Of course the Ptolemaic model was predictive — that was one of its main features, and one of the reasons it stayed popular for centuries: you could use it to predict the future positions of the planets. The idea that a model of the heavens should not just provide a handwaving, semi-metaphorical picture of what was going on, but should make numerical predictions for future observations, was a key achievement.

    These weren’t terribly accurate predictions, and for a long time people weren’t very concerned with their inaccuracy. And, of course, the model failed to explain a lot of things, and failed to predict things that the Copernican model did (e.g., the phases of Venus).

    (It may be worth remembering that Copernicus’s heliocentric model also had epicycles — indeed, there was one variant that had secondary epicycles attached to the primary epicycles.)

    The problem in thinking about Ptolemaic (and Copernican) epicycles as “Fourier expansions” and the like is that they were constrained to have uniform circular motion. If you allow the epicycles to be *elliptical*, with non-uniform motion, then, yes, you can think of them as approximations to actual planetary motion.

  15. al says:

    Copernic’s book was published with a preface by Andreas Osiander who took pains to tell everybody that actually heliocentrism si just another way of ‘saving appearances.’ It was presented as a mathematical hypothesis, a way of doing calculations, not as physics. But in medieval universities two courses were taught: one with the Aristotelian-physical model which was deemed ‘true’ but known to be descriptively inexact; the other ‘untrue’ but exact (by the way Ptolemy had the ‘equant’ ( punctum equans) , a point inside an orbit where non uniform movement appeared as uniform).

  16. phil fogle says:

    As a member of the ‘public’, I constantly see evidence of the undermining of the principle of falsifiability, and HEP seems to be playing a major role.

    Almost all my friends have heard about the Higgs field; not one has any idea of its significance. Their minds are full of black holes, holograms and multiverses that they believe are proven facts.

    …actually, I see homeopathy (sorry for being OT) as extremely insideous, as it hijacks the placebo function. Medics should be honest about placebos.

  17. Brian says:

    Funny thing is about the update section, I thought thats what the fuss was about. Science is useful because it allows us to make predictions and test them. It lets us either discover how the world works or at least create a narrative that works, but either way its consistent, rigorous, objective, and sometimes has practical applications. I’m not arguing against religion, superstitions, or odd beliefs, if thats physiologically helpful to you thats all fine, so long as we have an objective way of knowing how the world works that we can use to acquire knowledge free from the subjective views and interests of others. Thats why I found it so unsettling, about a year or two ago when I read a popular science article by someone related to the field. they where saying that it was time to move on to “post empirical science ” (what ever that is). Another talking about getting rid of Popper’s falsification, the bottom line is the creation of a “science” that does not depend on experiment or verification. It is troubling to hear the words you would expect from a new age woo mister or some incomprehensible post-modernist professor coming from scientists or philosophers of science.

    Now to me that seems like giving up what science is in order to save a few careers. The whole argument depends on a lack of imagination, “string theory is right because there is no other game in town” (dubious). I have to ask what makes string theory the only game in town other than it looks fancy and some scientists/mathematicians like it? If science is simply a series of narratives that work, made more plausible by experiments, than any narrative would do in string theories place, since no experiment can be done to confirm it. Even fantastic ones like how the Greeks used to use mythological figures to explain the way the world worked, after all they have the same amount of experimental evidence. In the history of science these mythologies became unsustainable do to the scientific revolution, we could do experiments to come up with better stories, scientific theories. As far as the public is concerned, this is where the power of science comes from, the fact that it explains how the world works and can be demonstrated. Scientists would shot themselves in the foot by arguing that testability needs to go because its not their pet theory that benefits, its the snake oil salesman.

    anyway, I say its better to have no theory and science than have a theory and no science.

  18. al says:

    ‘Postempirical science’ is an attempt to shift the burden of proof. Standard science was and still is positivistic: if you make claims you have to point to some evidence. Positivism did not admit any ontological ‘known unknowns’. Postempiricists try to argue that you should give them credit as long as they have not been proved wrong. There is an evident shifting of power here as outsiders cannot beat experts at their game.
    (actually it was mathematicians in the 19 c. who first claimed the right to consider whatever they pleased as long as it was not obviously self-contradictory; physicists had to point to some external support, but ‘mathematical’ or ‘theoretical’ physics tended to erase the distinction)

  19. Nobody says:


    “Epicycles are essentially a Fourier expansion.”

    I think I originated that idea on the Compuserve SciMath forum about 20 years ago and I also believe that it is not true.

    But I’m not that smart, so maybe you or someone else can show me how you can get a Keplerian ellipse from a harmonic series of constant speed circles.

  20. Peter Woit says:

    Enough about epicycles please. One problem is that I really know nothing much about them, so can’t do a proper moderator’s job (e.g. delete misinformation…).

  21. Nathalie says:

    Dear Bee,

    You are being unfair to Lisa Randall.
    As the great Wolfgang Pauli used to say, when you don’t have a theory (as is currently the case for beyond the standard model) you should make “simple” models that can be tested and thus may provide some insight. And that is what Lisa Randall is doing.

    Esoteric movements, by not being testable, tend to lead nowhere.

    Once you have a theory which starts looking promising it is worth to work on polishing it, in order to understand better its structure.

  22. John Baez says:

    Jeff M wrote:

    Honestly, epicycles were not “predictive” in any meaningful way.

    Sure they were! It’s true they couldn’t predict what would happen to the Solar System if you added a new planet, but Newton’s laws couldn’t predict anything about chemistry and people don’t complain that he didn’t invent quantum mechanics.

    Anyone who mocks the epicycles Ptolemy’s Almagest probably hasn’t thought about how much better it was than previous models, and how long it took for heliocentric models to exceed its predictive power. Check out Ptolemy’s Almagest: Fact and Fiction by Richard Fitzpatrick of the physics department of U.T. Austin. Some quotes:

    The standard popular modern criticisms of Ptolemy’s model of the solar system are as follows. First, it is generally thought that Ptolemy’s thinking was shackled by accepted truths in ancient Greek philosophy (mostly due to Aristotle), which held, amongst other things, that the Earth was stationary, and that celestial bodies were constrained to move uniformly around circular orbits. Second, it is supposed that these mental shackles directly lead Ptolemy to introduce the concept of an epicycle as a sort of kludge to explain the observed retrograde motion of the superior planets without having to admit that this phenomenon was caused by the Earth’s motion. Third, it is generally held that Ptolemy’s model of the solar system was not particularly accurate, leading later Arabic and medieval European astronomers to add more and more epicycles in order to get better agreement with observations. The final version of the model is alleged to have contained an absurd number of epicycles, and to have essentially collapsed under its own weight, leaving the field clear for Copernicus and his, supposedly, much simpler, and much more accurate, heliocentric model of the solar system.

    Needless to say, the popular criticisms of the Almagest that I have just outlined are almost entirely wrong. What I want to do in this talk is to describe what Ptolemy actually did in the Almagest, and to contrast this with the mistaken popular view of what he did.


    The Almagest model of the geocentric solar orbit is surprisingly accurate. It can predict the position of the Sun, relative to the stars, to an accuracy of about 1 arc minute. That is, about 1/60 th of a degree.

  23. Kevin Henderson says:

    The general public in America favors ideologies of all sorts. In particular, the majority are in favor of things which scientists either believe or know to be magic (i.e., false). It is unfavorable, economically, at least, at this time to condemn all fantastic ideas based on unrealistic physics. People like to think science fiction is closer than it is.

Comments are closed.