Many Worlds or the multiverse was conceived by Hugh Everett as an interpretation of quantum mechanics. He proposed it as a way to make quantum mechanics deterministic and to avoid the collapse of the wave function. He argued that when the wave function shows that several different events are possible, it could be argued that all the possibilities happen, each in their own separate state. This interpretation would preserve the unitary development of the wave function in a smooth and continuous way so that quantum mechanics does not imply random processes are fundamental. What remains as random is which particular state of the universe we happen to find ourselves in. The original idea which he called the relative states formulation was mathematically incomplete. The major problem was to show how a probability distribution can be converted into a distribution of possible alternate worlds, since in practice the splitting is rarely so straight forward as a quanta choosing either one or other of two options, and the choices are weighted probabilities with a varying density distribution of possible events. Each of these possibilities then generates more universes in proportion to how likely each possibility is to occur. The idea was taken up by Bryce DeWitt who was the first to call it the Many Worlds interpretation and subsequently by a number of other physicists. The original concept, with the addition of some extra mathematical rigour, remains the same as when Everett first proposed it.
Branching leads to an exponential number of new universes continuously being created, including multiple copies of myself all writing slightly different versions of this article. Max Tegmark dismisses this as being a criticism. He argues that the number of universes is finite and that the parallel worlds are mostly very similar and frequently merge on the microscopic scale. Once they have become significantly different however they can no longer merge but still he considers the number of universes not a problem. Viewed from a timeless point of view, as Julian Barbour would, all possible configurations of the universe exist eternally so there is no growth at all. David Deutsch, who pioneered the early theoretical research in quantum computing, is also a strong supporter of the multiverse. He argues that the development of quantum computers is evidence in its favour as the power of the qubit depends upon its ability to interact with itself in those other universes.
Supporters of Many Worlds dismiss criticisms based on Ockham's razor. Tegmark sites the historical precedent that when Bruno proposed that the stars were distant suns, it was attacked as being not consistent with Ockham's razor. He argues that the theory is more economical with assumptions that form the basis on which quantum mechanics is built. While strictly true that we may may draw an analogy between stars being suns and the existence of other universes, this is a distraction from the very real criticisms that can be raised. I would point out that Many Worlds massively violates conservation of energy and every other conservation law besides. Also without adding on the extra assumption that gravity must be quantised and hence localised within each universe, the force of gravity would permeate all the universes and the multiverse would rapidly collapse. Deutsch's argument based on the power of the quantum computer could also be turned around to express how little we understand about the quantum nature of reality. The quantum computer is closely modeled on the Turing machine. Both the Turing machine and the quantum computer use the concept of computability and that all problems are in principle computable. Roger Penrose strongly disagrees with this and puts forward a powerful case in The Emperor's New Mind that human intelligence is not any form of Turing machine. Of all the arguments in favour of Many Worlds, Barbour's timeless vision is most persuasive but leaves itself open to arguments about how we can understand timelessness itself. What becomes of ideas such 'development' or 'process' or 'causality' if every possible state of the universe already preexists?
When I first started thinking about time travel I expected to find that we live in a multiverse. The alternative appeared to be that if time travel is possible in a singe universe, it must be perfectly deterministic but this has been ruled out by quantum mechanics which is intrinsically non deterministic. Today I am not so sure. The multiverse appears to be necessary but is itself problematic and in conflict with the experimental proposal under consideration for reasons I shall return to shortly. First let us see how it can help us understand how time travel might not lead to paradox.
Let us consider information arriving from the future concerned with science and technology. We can imagine that in this situation, many workers will seek to reproduce methods and results already known about in the future. This introduces us to one variant on the paradoxes of time travel. Who, it will be asked, originated an idea if the first to work on it learnt about the concept in papers not yet published. Will they find their own names on those papers or other? It looks like cheating, but it is not. Knowledge discovered in a parallel universe is arriving in ours. By creating pathways between universes we access the combined resources of all those linked worlds. If for, example, we are presented with a future map of the world showing how ocean levels have risen, this is not our future, but another future. Along with the map will come information regarding future technologies originated in those other worlds. They may be more sustainable and perhaps we can create a better future by putting them into practice sooner. There is no contradiction, we are choosing a different pathway along the many branches of the multiverse based on our greater knowledge. We only receive messages from other futures, our own future remains as mysterious and unknown as ever.
There is however a conflict between the Many Worlds theory and the proposal for transporting information back through time. The specific mechanism proposed depends on the existence of non locality. Non locality, the idea that a quantum event in one place or time can influence an entangled event far from it is an intrinsic 'weird' feature of standard quantum mechanics and an implicit part of Wheeler's Delayed Choice Paradox. In Many Worlds, non locality does not exist. The apparent 'spooky actions at a distance' can all be accounted for, so it is claimed, by straightforward means. If the research project discussed here is successful, information will be sent back through time, and that signal will be arriving in a different parallel world, but the Many Worlds theory forbids non locality so it will predict that it is impossible to send the signal. We will have a result which no current theory can explain.
I conjecture that there maybe a variant of the Many Worlds theory that can allow non locality and hence our signals to be sent. We can test to see if Many Worlds exist as they will act as a filter blocking certain types of information from being sent back from the future. We would receive messages from numerous different versions of the future and any variations that are purely random, perhaps for example the future state of the stock market as described in Implications, would become blurred by the overlapping of the different messages. Our instruments would produce ambiguous readings. In contrast a process that is presently beyond human understanding but for which there is almost certainly is a perfectly good deterministic theory, say being able to predict earthquakes, the warning should come through very clearly as in any version of the World that is substantially similar to this one, earthquakes will still happen at the same moments.
In general I am predicting that if Many Worlds exists, the specific way information can be sent will determine what information we can receive and what gets scrambled in its journey from the future. If on the hand, we find that there is no scrambling of the signal and that even intrinsically random messages arrive completely clearly, we can conclude that Many Worlds do not exist.