In his response to my review, Matthew Kleban takes a quote out of context. I did not say that multiverse theory was a “radical departure from normal mechanistic science.” I said it might appear that way, but “in many respects it is better seen as a continuation of that theme.” That’s the whole point: multiverse theory is trying to look like conventional science.
The greatest successes of science have come from making predictions that are later confirmed by experiments. And the most effective justification for scientific research is that it leads to successful technologies. So scientists naturally try to align themselves with these aims. But what do you do for a highly speculative area of theory which has no track record of producing either technological innovation or accurate predictions? A common answer is to imitate such outcomes by playing what we might call the technology game, and the prediction game.
The technology game is to argue that a particular field of research promises useful spin-offs, or vicariously take credit for past successes, even where the connection is remote. “From where I sit,” writes Kleban, “modern science is working pretty well, producing such near-magical marvels as smart phones, nuclear power, and space travel.” So, “If it ain’t broke, don’t fix it.” The implicit message is that opposing the current approach towards fundamental research, of the type that brought us string theory and multiverse theory, is equivalent to being anti-progress. In fact, as I showed in Truth or Beauty, despite the popularity of this argument, commercial spin-offs from fundamental physics have been rare (don’t expect a Higgs boson ray gun). And our ability to design a smart phone, which relies mostly on solid-state technology, certainly says nothing about our ability to determine the number of universes (we can’t actually call them up), or about Unger and Smolin’s point that cosmology is in crisis.
The prediction game can be played a number of ways. One is to say that a theory predicts something which everyone knows about (see Ed Witten’s claim that “String theory has the remarkable property of predicting gravity”), although this suffers from the drawback that it isn’t what prediction means. Another way to play is to make a bold prediction of some new phenomenon but leave it vague enough that it cannot be falsified. So far for example string theorists have predicted that every particle has a supersymmetric twin – thus doubling the number of particle types in existence – while remaining flexible about their exact properties such as mass. None has yet to be observed, but that isn’t a problem because it could just be that we need a larger accelerator to see them.
Cosmology is an even safer ground for playing the prediction game because it isn’t possible to run controlled experiments, and signals are often weak and elusive, more suggestion than hard proof. As in a children’s playground, all the surfaces are soft and forgiving. Scientists can stake out some open-ended predictions about a future experiment, and if the results are reasonably consistent, it is taken as a thumbs up, even though the results could be similarly consistent with any number of theories (developed or not). But if the prediction is proved false, again that doesn’t irreparably break the theory, because it can always be adjusted. It’s a game where you can win, but you can (almost) never lose. An example was the “detection” of gravitational waves which last year was touted as convincing evidence of cosmic inflation (a key component of many cosmological theories) but which turned out to be a mirage produced by space dust (the theories of course survived intact).
Here we learn that “the [string landscape] theory can be used to predict the signatures of cosmic bubble collisions.” Indeed that would be quite a find, as would a portal to a parallel universe, but it should be mentioned that so far, no related experiments have shown any such signals. I would be interested to see if the detection of positive spatial curvature actually falsified the theory – wouldn’t it just adapt? And the statement that “the standard model of particle physics combined with Einstein’s theory of general relativity … predict a large landscape quite similar to that of string theory” is news to me, and another interesting use of the word “prediction.”
Of course, real predictions are central to science. If the experiments mentioned do produce “nearly irrefutable evidence” of multiverse theory or string theory, as promised, or for that matter of Unger and Smolin’s singular universe, then that will be something to look forward to. Another possibility is that the prediction game will be allowed to run indefinitely, with none of the competing theories (which largely reflect aesthetic choices) gaining a conclusive advantage.
To see how this can work, note that the most sophisticated players of the science games are mainstream economists, who similarly take credit for all economic progress. According to an article by Peter Foster, kudos for mobile communications technology is really due to Adam Smith: “The Invisible Hand gives you the iPhone.” Economists even invented a theory, the efficient market hypothesis, which “predicts” that markets are unpredictable, thus providing a permanent excuse for their inability to foresee cosmic explosions such as the 2008 financial crash.
It is true that “rejecting science because one is uncomfortable with its implications has a decidedly poor track-record.” But that’s not what is going on here. String theory and multiverse theory are being rejected by many scientists because they have not provided a useful explanatory model of the universe. And holding on to theories that repeatedly fail to deliver doesn’t have much of a track record either.