The Trouble With Physics: The Rise of String Theory, The Fall of a Science, and What Comes Next

by Lee Smolin

  It seems rational to deny this request and insist that we should not change the rules of science just to save a theory that has failed to fulfill the expectations we originally had for it. If string theory makes no unique predictions for experiments, and if it explains nothing about the standard model of particle physics which was previously mysterious—apart from the obvious statement that we must live in a universe where we can live—it does not seem to have turned out to be a very good theory. The history of science has seen a lot of initially promising theories fail. Why is this not another such case?

  We have regrettably reached the conclusion that string theory has made no new, precise, and falsifiable predictions. But still, string theory makes some startling assertions about the world. Could an experiment or an observation one day reveal evidence for any of these surprising features? Even if there are no definite up or down predictions—predictions of the kind that could kill or confirm the theory—might we see evidence of a feature that is central to the stringy view of nature?

  The most obvious novelty of string theory is the strings themselves. If we could probe the string scale, there would be no problem seeing abundant evidence for string theory, if it is true. We would see indications that the fundamental objects are one-dimensional rather than pointlike. But we are not able to do accelerator experiments at anywhere near the energies required. Is there another way we could make the strings reveal themselves? Might the strings somehow be induced to become bigger, so that we could see them?

  One such scenario was proposed recently by Edmund Copeland, Robert Myers, and Joseph Polchinski. Under certain very special assumptions about cosmology, it might be true that some very long strings were created in the early universe and continue to exist.6 The expansion of the universe has stretched them to the point that they now are millions of light-years long.

  This phenomenon is not limited to string theory. For some time, a popular theory about the formation of galaxies suggested that they were seeded by the presence of huge strings of electromagnetic flux left over from the Big Bang. These cosmic strings, as they were called, had nothing to do with string theory; they were a consequence of the structure of the gauge theories. They are analogous to the quantized lines of magnetic flux in superconductors, and they can form in the early universe as a consequence of the universe going through phase transitions as it cools down. We now have definitive evidence from cosmological observations that such strings were not the main ingredient in the formation of structure in the universe, but there could still be some cosmic strings left over from the Big Bang. Astronomers search for them by looking for their effect on light from distant galaxies. If a cosmic string were to come between our line of sight and a distant galaxy, the gravitational field of the string would act as a lens, duplicating the image of the galaxy in characteristic ways. Other objects, like dark matter or another galaxy, can have a similar effect, but astronomers know how to distinguish between the images they generate and those produced by a cosmic string. Recently there was a report that such a lens might have been detected. It was labeled, optimistically, CSL-1, but when it was viewed by the Hubble Space Telescope, it turned out to be two galaxies close to each other.7

  What Copeland and his colleagues found is that under certain special conditions, a fundamental string, stretched to enormous lengths by the expansion of the universe, would resemble a cosmic string. So it might be observable through its action as a lens. Such a fundamental cosmic string may also be a prodigious radiator of gravitational waves, which might make it observable by LIGO, the Laser Interferometer Gravitational-wave Observatory.

  Predictions of this kind give us some hope that string theory might someday be verified by observations. Yet the discovery of a cosmic string, by itself, cannot verify string theory, because several other theories also predict the existence of such strings. Nor can failure to find one lead to a falsification of string theory, because the conditions for such cosmic strings to exist are specially chosen, and there is no reason to think they might exist in our universe.

  Besides the existence of strings, there are three other generic features of the stringy world. All sensible string theories agree that there are extra dimensions, that all the forces are unified into one force, and that there is supersymmetry. So even if we have no detailed predictions, we can see whether experiment can test these hypotheses. Since they are independent of string theory, finding evidence for any one of them does not prove that string theory is true. But here the opposite is not the case: If we learn that there is no supersymmetry or no higher dimensions or no unification of all the forces, then string theory is false.

  Let’s start with extra dimensions. We may not be able to see them, but we can certainly look for their effects. One way to do this is to search for the extra forces that are predicted by all higher-dimensional theories. These forces are transmitted by the fields that comprise the geometry of the extra dimensions. Such fields must be there, because you cannot limit the extra dimensions to producing only the fields and forces we so far observe.

  The forces that come from such fields are expected to be roughly as strong as gravity, but they may differ from gravity in one or more ways: They may have a finite range, and they may not interact equally with all forms of energy. Some current experiments are extraordinarily sensitive to such hypothetical forces. About ten years ago, one experiment saw preliminary evidence for such a force, which was called the fifth force. Further experiments did not support the claim, and as of yet there is no evidence for such forces.

  String theorists have usually assumed that the extra dimensions are tiny, but several adventurous physicists realized in the 1990s that this did not have to be the case—that the extra dimensions could be large or even infinite. This is possible in a brane-worlds scenario. In such a picture, our three-dimensional space is actually a brane—that is, something like a physical membrane but with three dimensions—suspended in a world with four or more dimensions of space. The particles and forces of the standard model—electrons, quarks, photons, and the forces they interact with—are restricted to the three-dimensional brane that makes up our world. So using only those forces, you cannot see evidence of the extra dimensions. The sole exception is the gravitational force. Gravity, being universal, extends through all the dimensions of space.

  This kind of scenario was first constructed in detail by three physicists working at SLAC, the Stanford Linear Accelerator Center: Nima Arkani-Hamed, Gia Dvali, and Savas Dimopoulos. Surprisingly, they found that the extra dimensions could be quite large without conflicting with known experiments. If there were two extra dimensions, they could be as wide as a millimeter across.8

  The main effect of adding such large extra dimensions is that the gravitational force in the four- or five-dimensional world turns out to be much stronger than it appears to be on the three-dimensional brane, so quantum gravitational effects happen at a much larger length scale than otherwise expected. In quantum theory, a larger length scale means a smaller energy. By making the extra dimensions as large as a millimeter, one can bring down the energy scale at which quantum-gravity effects should be seen—from the Planck energy, which is 1019 GeV, to only 1,000 GeV. This would resolve one of the most stubborn questions about the parameters of the standard model: that is, Why is the Planck energy so many orders of magnitude bigger than the mass of the proton? But what is really exciting is that it would bring quantum-gravity phenomena within range of being revealed by the Large Hadron Collider (LHC), coming online in 2007. Among these effects could be the production of quantum black holes in collisions of elementary particles. This would be a dramatic discovery.

  Another kind of brane-world scenario was developed by Lisa Randall, of Harvard, and Raman Sundrum, of Johns Hopkins University. They found that the extra dimensions could be infinite in size as long as there was a negative cosmological constant in the higher-dimensional world.9 Remarkably, this too agrees with all observations to date, and it even makes predictions for new ones.


  These are adventurous ideas and fun to think about, and I’m deeply admiring of their inventors. That said, I’m troubled by the brane-world scenarios. They are vulnerable to the same problems that doomed the original attempts at unification through higher dimensions. The brane-world scenarios work only if you make special assumptions about the geometry of the extra dimensions and the way the three-dimensional surface that is our world sits inside them. In addition to all the problems suffered by the old Kaluza-Klein theories, there are new problems. If there can be one brane floating in the higher-dimensional world, couldn’t there be many? And if there are others, how often do they collide? Indeed, there are proposals that the Big Bang arose from the collision of brane worlds. But if this can happen once, why hasn’t it happened since? Some 14 billion years have gone by. The answer might be that branes are scarce, in which case we are back to depending on very finely tuned conditions. Or it might be that branes are precisely parallel to one another and don’t move much, in which case we again have finely tuned conditions.

  Beyond these problems, I am skeptical because these scenarios depend on special choices of background geometries, and this contradicts Einstein’s principal discovery, as set out in his general theory of relativity, that the geometry of spacetime is dynamical and that physics must be expressed in a background-independent manner. Nevertheless, this is science as it should be: bold ideas that are testable by doable experiments. Let’s be clear, though. If any of the predictions of brane worlds comes true, it will not amount to a confirmation of string theory. Brane-world theories stand on their own; they do not need string theory. Nor is there a completely worked-out realization of a brane-world model within string theory. Conversely, if none of the predictions of brane worlds are seen, this does not falsify string theory. Brane worlds are just one of the ways that the extra dimensions of string theory could manifest themselves.

  The second generic prediction of string theory is that the world is supersymmetric. Here, too, there is no falsifiable prediction, because we know that supersymmetry, if it truly describes the world we see, must be broken. In chapter 5, we noted that supersymmetry may be seen in the LHC. This is possible but by no means guaranteed, even if supersymmetry is true.

  Fortunately, there are other ways to test for supersymmetry. One possibility involves the dark matter. In many supersymmetric extensions of the standard model, the lightest new particle is stable and uncharged. This new stable particle could be the dark matter. It would interact with ordinary matter but only through gravity and the weak nuclear force. Such particles are called WIMPs, for weakly interacting massive particles, and several experiments have been mounted to detect them. These detectors exploit the idea that dark-matter particles will interact with ordinary matter via the weak force. This makes them very much like heavy versions of neutrinos, which also interact with matter only through gravity and the weak force.

  Unfortunately, because the supersymmetric theories have so many free parameters, there is no specific prediction for what the mass of the WIMPs should be or exactly how strongly they should interact. But if they do indeed make up the dark matter, we can deduce what range is possible for their masses, assuming that they have played the role we think they have in the formation of galaxies. The range predicted is comfortably within the one that theory and experiment suggest for the lightest superpartner.

  Experimentalists have looked for WIMPs by using detectors similar to those used to detect neutrinos coming from the sun and distant supernovas. Extensive searches have been carried out, but so far no WIMPs have been found. This is of course not definitive—it means only that if they exist, they interact too weakly to have triggered a response from a detector. What can be said is that if they interacted as strongly with matter as neutrinos do, they would have been seen by now. Still, the discovery of supersymmetry by any means would be a spectacular triumph for physics.

  The main thing to keep in mind is that even if string theory requires that the world be supersymmetric on some scale, it makes no prediction about what that scale is. Thus, if supersymmetry is not seen in the LHC, that does not falsify string theory, because the scale at which it might be seen is completely adjustable. On the other hand, if supersymmetry is seen, that does not confirm string theory. There are ordinary theories that require supersymmetry, such as the minimal supersymmetric extension of the standard model. Even among quantum theories of gravity, supersymmetry is not unique to string theory; for example, the alternative approach called loop quantum gravity is completely compatible with supersymmetry.

  We now come to the third generic prediction of string theory: that all the fundamental forces become unified at some scale. As in the other cases, this idea is broader than string theory, so a confirmation of it would not prove that string theory is right; indeed, string theory allows several possible forms of unification. But there is one form that most theorists believe represents grand unification. As we discussed in chapter 3, grand unification makes a generic prediction, so far unverified, that protons should be unstable and decay at some time scale. Experiments have looked for proton decay and failed to find it. These results (or lack of them) killed certain grand unified theories but not the general idea. However, the failure to find proton decay remains a constraint on possible theories, including supersymmetric theories.

  A large number of theorists believe all three of these generic predictions will be confirmed. Consequently, experimentalists have put enormous effort into searching for evidence that would support them. It is not an exaggeration to say that hundreds of careers and hundreds of millions of dollars have been spent in the last thirty years in the search for signs of grand unification, supersymmetry, and higher dimensions. Despite these efforts, no evidence for any of these hypotheses has turned up. A confirmation of any of these ideas, even if it could not be taken as a direct confirmation of string theory, would be the first indication that at least some part of the package deal that string theory requires has taken us closer to, rather than further from, reality.

  12

  What String Theory Explains

  WHAT ARE WE to make of the strange story of string theory thus far? More than two decades have now passed since the first superstring revolution. During this time, string theory has dominated the attention and resources of theoretical physics worldwide—more than a thousand of the world’s most talented and highly trained scientists have worked on it. While there has been room for honest disagreement about the theory’s prospects, sooner or later science is supposed to accumulate evidence that allows us to reach a consensus about the truth of a theory. Mindful that the future is always open, I would like to close this section by offering an assessment of string theory as a proposal for a scientific theory.

  Let me be clear. First, I am not assessing the quality of the work; many string theorists are brilliant and well trained and their work is of the highest quality. Second, I want to separate the question of whether string theory is a convincing candidate for a physical theory from the question of whether or not research in the theory has led to useful insights for mathematics or other problems in physics. No one disputes that a lot of good mathematics has come out of string theory and that our understanding of some gauge theories has been deepened. But the usefulness of spin-offs for mathematics or other areas of physics is not evidence either for or against the correctness of string theory as a scientific theory.

  What I want to assess is the extent to which string theory has fulfilled its original promise as a theory that unites quantum theory, gravity, and elementary-particle physics. String theory either is or is not the culmination of the scientific revolution that Einstein began in 1905. This kind of assessment cannot be based on unrealized hypotheses or unproved conjectures, or on the hopes of the theory’s adherents. This is science, and the truth of a theory can be assessed based only on results that have been published in the scientific literature; thus we must be careful to distinguish between conjecture, evidence, and proof.

&
nbsp; One might ask whether it is too early to make such an assessment. But string theory has been under continuous development for more than thirty-five years, and for more than twenty it has captured the attention of many of the brightest scientists in the world. As I emphasized earlier, there is no precedent in the history of science, since at least the late eighteenth century, for a proposed major theory going more than a decade before either failing or accumulating impressive experimental and theoretical support. Nor is it convincing to point to the experimental difficulties, for two reasons: First, much of the data that string theory was invented to explain already exists, in the values of the constants in the standard models of particle physics and cosmology. Second, while it is true that strings are too small to observe directly, previous theories have almost always quickly led to the invention of new experiments—experiments that no one would have thought of doing otherwise.

  In addition, we have a lot of evidence to consider in making our evaluation. The many people working on string theory have given us a great deal to work with. Equally informative are the conjectures and hypotheses that have remained open despite intensive investigation. Most of the key conjectures that are unresolved are at least ten years old, and there is no sign that they will be resolved soon.

  Finally, string theory is, as a result of the discovery of the vast landscape of theories described in chapter 10, in a crisis that is leading many scientists to reconsider its promise. So, while we must remember that new developments could change the picture, this appears to be a good time to attempt an assessment of string theory as a scientific theory.