George Ellis and the multiverse

George Ellis

George Ellis is a South African cosmologist who rose to fame almost half a century ago when he wrote a book together with Stephen Hawking (The Large Scale Structure of Space-Time, 1973), today considered a classic.

In an earlier post in this blog, published in November 2014, I mentioned that there are six independent theories about the multiverse, almost all of them incompatible with each other. In a recent article titled Theory Confirmation and Multiverses published in the book Why Trust a Theory?, edited by Radin Dardashti, Richard David and Karim Thébault (Cambridge University Press, 2019), George Ellis updates the different multiverse theories. He does not mention six, as I did five years ago, but nine, although he has left out one of the six I mentioned in my post (Smolin’s), perhaps because this theory has been abandoned in the meantime. The nine theories are:

1. The multiverse in the universe: Undetectable parts of our universe, beyond our event horizon, could have physical properties different from ours.
2. The chaotic inflationary multiverse: If our universe went through an inflation phase and then slowed down, there could exist many other similar bubbles in a much larger universe.
3. The multiverse of the M theory: This is an extension of string theory that adds an additional dimension, which would separate the various universes in that multiverse, each of which is called a brane.
4. Cyclic universes: I did not mention this version, because the different universes are not separated in space, but in time (each one follows the previous one and gives rise to the next). Cyclical universes were fashionable in the 80s, but today they don’t seem to be in favor.
5. The string theory landscape: It is more or less the same as multiverse number 3, but without the branes. In my post I joined these two theories into one.
6. Ramifications of the quantum mechanics wave function: This theory, which was proposed by Hugh Everett III in 1957, is what I called the quantum multiverse in time. It is also called the many worlds interpretation of quantum mechanics. According to this theory, everything that can happen, must happen in one of those universes.
7. Holographic projections: This is a new proposal, which assumes that our universe is the holographic projection of a multi-dimensional universe over three dimensions. If there is a projection, there could be others, which would give us a multiverse.
8. We live in a simulation: Our universe is not real, but a computer simulation. I spoke about this in another of my posts. This theory has been disproved by recent research (2018).
9. Tegmark’s mathematical multiverse: Everything that can exist, exists somewhere.

Ellis adds a tenth alternative possibility:

10. There is only one universe. Therefore the multiverse does not exist, it’s just a figment of the imagination of some physicists. He immediately adds that the available evidence and Occam’s razor favor this possibility.

Ellis rightly points out that the existence of so many alternative theories about the multiverse, none of which can be shown to be false, while several are incompatible with the others, rather than making the existence of a multiverse more plausible, makes it less plausible:

…if they conflict with each other, why should any of them be true?

All those theories are impossible to prove. In his words:

…we cannot test them, because we don’t have observational or experimental access to any other universes that may exist… Multiverse theories claim to tell us what is true in regions for which no observation data whatever are available! These theories are not modest in their aim.

Faced with this situation, some supporters of multiverse theories claim that science should relax the criteria to decide whether a theory is scientific or not. Rather than demanding that the theory can be proved false by a (possibly mental) experiment, we should resort to arguments such as: if we don’t have anything better, this theory must be true.

Ellis proposes an interesting test that would distinguish whether a theory is really scientific or philosophical (ideological):

What data/observation would lead you to abandon the multiverse model? If the answer to this question is ‘none’ then the theory is dogma, not science. Because multiverses can explain anything, the answer is, indeed, likely to be ‘none’.

The final conclusion of Ellis’s article is devastating:

The kinds of arguments given… are not adequate for claiming the theory is a well-established scientific result. Having a well-developed mathematical theory is not sufficient.

The same post in Spanish
Thematic thread on Multiverse and Fine Tuning: Preceding Next

Manuel Alfonseca