A pervasive idea in fundamental physics and cosmology that should be retired: the notion that we live in a multiverse in which the laws of physics and the properties of the cosmos vary randomly from one patch of space to another. According to this view, the laws and properties within our observable universe cannot be explained or predicted because they are set by chance. Different regions of space too distant to ever be observed have different laws and properties, according to this picture. Over the entire multiverse, there are infinitely many distinct patches. Among these patches, in the words of Alan Guth, "anything that can happen will happen—and it will happen infinitely many times". Hence, I refer to this concept as a Theory of Anything.
Any observation or combination of observations is consistent with a Theory of Anything. No observation or combination of observations can disprove it. Proponents seem to revel in the fact that the Theory cannot be falsified. The rest of the scientific community should be up in arms since an unfalsifiable idea lies beyond the bounds of normal science. Yet, except for a few voices, there has been surprising complacency and, in some cases, grudging acceptance of a Theory of Anything as a logical possibility. The scientific journals are full of papers treating the Theory of Anything seriously. What is going on?
Have experiments revealed that our observable universe and the fundamental laws are too complicated to be explained by normal science? Absolutely not! Quite the opposite! On the macroscopic scale, the latest measurements show our observable universe to be remarkably simple, described by very few parameters, obeying the same physical laws throughout and exhibiting remarkably uniform structure in all directions. On the microscopic scale, the Large Hadron Collider at CERN (European Center for Nuclear Research) has revealed the existence of the Higgs, in accord with what theorists had predicted nearly 50 years ago based on sound scientific reasoning.
A simple outcome calls for a simple explanation for why it had to be so. Why, then, consider a Theory of Anything, that allows any possibility, including complicated ones? The motivation is the failure of two favorite theoretical ideas—inflationary cosmology and string theory. Both were thought to produce a unique outcome. Inflationary cosmology was invented to transform the entire cosmos into a smooth universe populated by a scale-invariant distribution of hot spots and cold spots, just as we observe it to be. String theory was supposed to explain why elementary particles could only have the precise masses and forces that they do. After more than 30 years investment in each of these ideas, theorists have found that they are not able to achieve these ambitious goals. Inflation, once started, runs eternally and produces a multiverse of pockets whose properties vary over every conceivable possibility—flat and non-flat; smooth and non-smooth; scale-invariant and not scale-invariant; etc. Despite laudable efforts by many theorists to save the theory, there is no solid reason known today why inflation should cause our observable universe to be in a pocket with the smoothness and other very simple properties we observe. A continuum of other conditions is equally possible. In string theory, a similar explosion of possibilities has occurred, driven by attempts to explain the 1998 discovery of the accelerated expansion of the universe. The acceleration is thought to be due to positive vacuum energy, an energy associated with empty space. Instead of predicting a unique possibility for the vacuum state of the universe and particles and fields that inhabit it, our current understanding of string theory is that there is a complex landscape of vacuum states corresponding to exponentially different kinds of particles and different physical laws. The set of vacuum space contains so many possibilities that, surely, it is claimed, one will include the right amount of vacuum energy and the right kinds of particles and fields. Mix the inflation and string theory, and the unpredictability multiplies. Now every combination of macrophysical and microphysical possibilities can occur.
I suspect that the theories would never have gained the acceptance they have if these problems had been broadly recognized at the outset. Historically, if a theory failed to achieve its goals, it was improved or retired. In this case, though, the commitment to the theories has become so strong that some prominent proponents have seriously advocated moving the goalposts. They say that we should be prepared to abandon the old-fashioned idea that scientific theories should give definite predictions and to accept a Theory of Anything as the best that can ever be achieved.
I draw the line there. Science is useful insofar as it explains and predicts why things are the way they are and not some other way. The worth of a scientific theory is gauged by the number of do-or-die experimental tests it passes. A Theory of Anything is useless because it does not rule out any possibility and worthless because it submits to no do-or-die tests. (Many papers discuss potential observable consequences, but these are only possibilities, not certainties, so the Theory is never really put at risk.)
I think a priority for theorists today is to determine if inflation and string theory can be saved from devolving into a Theory of Anything and, if not, seek new ideas to replace them. Because an unfalsifiable Theory of Anything creates unfair competition for real scientific theories, leaders in the field can play an important role by speaking out—making it clear that Anything is not acceptable—to encourage talented young scientists to rise up and meet the challenge. The sooner we can retire the Theory of Anything, the sooner this important science can progress.