Universe or multiverse?

In the posts in this blog I have often said that theories about the multiverse (there are many) are not science, but speculations, because it is impossible to design an experiment that demonstrates the existence or non-existence of these multiverses.

In an article published in May 2023 in the journal Springer Nature, entitled Is Everyone Probably Elsewhere?, the authors claim that it would at least be possible to distinguish between the following two hypotheses:

1. Our universe is unique, it does not belong to any multiverse.
2. Our universe belongs to some multiverse. Of course, we would have no idea what type of multiverse it would belong to.

When it comes to choosing between two alternative hypotheses, several statistical tools can be used:

•         Bayesian inference: Requires assigning a priori probabilities and/or the cost of error types (false positives and false negatives). In this case, that is not possible, so this method is eliminated.

•         Maximum likelihood: This method would be applicable, since it would allow us to distinguish between two symmetrical hypotheses based on an estimate of the parameters that make them different, which in this case would be the fundamental constants of the universe.

•         Hypothesis testing: A basic hypothesis and an alternative hypothesis are tested. According to the authors of the article, this method would also be applicable.

I am not going to go into detail about the development of the article, which carries a considerable statistical load. I will simply summarize the conclusions:

The strong anthropic principle requires that the fundamental constants of our universe be compatible with life (which leads to the observation that there is fine tuning). The question raised by the article is this: To what extent are they compatible?

•         If our universe is unique, it is most likely that its fundamental constants, even if they comply with the anthropic principle, are not optimal in relation to the existence of intelligent life. They could be marginally compatible.

•         If our universe belongs to a multiverse with a huge number of universes, its fundamental constants are most likely optimal for the existence of intelligent life. Why? Because if there are many universes compatible with such life, they will be clustered around the optimal values ​​of the constants (see Figure 2), while there would be far fewer in compatible but less optimal areas closer to the areas where universes would be incompatible with the existence of intelligent life.

Figure 1 shows a toy analysis using the maximum likelihood method, which attempts to distinguish an isolated universe from a multiverse with two universes, both with a single fundamental constant α. If α<0.4, the analysis concludes that it is more likely that there is only one universe; if, on the other hand, α>0.4, it would be more likely that there are two universes.

Figure 1

Figure 2 shows another toy analysis performed using the hypothesis contrast method that would distinguish an isolated universe from a multiverse with an infinite number of universes, also with a single fundamental constant C_i. The values ​​of C_i compatible with life would correspond to the interval [0.3; 0.7]. If 0.3<C_i<0.4 or 0.6<C_i<0.7, this analysis concludes that it is more probable that there is a single universe; if, on the contrary, 0.4<C_i<0.6, it would be more probable that there are infinite universes.

Figure 2

The real case, of course, is much more complex, since the number of fundamental constants that we know is close to 40. For this reason, the authors recognize that, in the current state of science, it is not possible to distinguish between the two alternative hypotheses (there is only one universe or there are many). But they say that perhaps in a more or less distant future we could come to distinguish them.

I wonder if the fact that the masses of the top quark and the Higgs boson seem to match their values ​​in such a way that the universe is marginally stable would be enough to claim that our universe is unique, as this article claims. But, as I pointed out in another blog post, that is not confirmed either, so the conclusion in the previous paragraph is still valid.

My conclusion is that, despite this study, multiverse theories are not science, and even if they become science someday, in a fairly long time we can forget about it.

The same post in Spanish

Thematic Thread on Multiverse and Fine Tuning: Previous Next

Manuel Alfonseca