Last month marked the 50th anniversary of the launching of the "Many Worlds" Interpretation of quantum mechanics, in which parallel universes are constantly branching off from the one we experience, with different events taking place in them. The Many Worlds Interpretation competes with the Copenhagen Interpretation, championed by Niels Bohr, in which the quantum state of a system often abruptly changes when it is observed ("the collapse of the wave function"). One outcome is seen to happen and according to the Copenhagen Interpretation the parts of the quantum state predicting other possibilities simply vanish. Many Worlds says those other parts still exist, just not in our branch.

In a future issue we will have an article by journalist Peter Byrne about the author of the original Many Worlds paper—Hugh Everett III—including some little-known history of events around its publication in 1957 and afterward. A couple of weeks ago Peter was at a conference in Oxford on the Many Worlds theory and he sent me the following report to post here. (I wish I had thought to invite him to do this earlier. No, wait—I wish this were the branch of the universe in which I did think to do so earlier!)

— Graham P. Collins


Many Worlds in Oxford

By Peter Byrne

Meeting for three days in the rain-swept United Kingdom in mid-July, a roomful of philosophers and physicists heatedly discussed a theory that multiple universes exist. The occasion for the debate was the "Everett@Fifty" conference sponsored by the Foundational Questions Institute and hosted by the Philosophy Faculty of the University of Oxford in a building near Logic Lane in Oxford. Most of the participants took it for granted that the Copenhagen Interpretation of quantum mechanics, which dominated the philosophy of quantum physics from the 1930s until quite recently, does not adequately describe the continuous evolution of atomic reality because it arbitrarily separates the quantum and "classical" realms from each other.

Eager to improve upon the Copenhagen Interpretation, the thirty-odd academics differed on the best ways to talk about a completely quantum mechanical (non-Copenhagen) universe. Some think of it as "Everettian," i.e. as a "multiverse" composed of a large but indeterminate number of branching autonomous "worlds" in the sum of which everything that is physically possible happens. Others, including the "Bohmians," say there is only one universe and that the very concept of probability only allows that some things, not all things, happen. Be that as it may, fifty years ago, it would have been difficult to imagine this debate taking place at all—except, perhaps, inside the insane asylum of Bedlam.

In July 1957, Reviews of Modern Physics published an abridgement of the doctoral dissertation of Hugh Everett III, a student at Princeton University. Taking the mathematics of quantum mechanics at face value, Everett claimed to have solved the "measurement problem," i.e. the problem of accounting for how a definite classical reality emerges from quantum uncertainty. But his solution to the problem was challenging to say the least. The sticking point was not his logic or his mathematics, but the implication that in some sense (hence the name "Many Worlds" theory) it pointed to the actualization of all possibilities. It posited that every individual is constantly "splitting" into non-communicating copies of herself along with her immediate environment. It has taken a half century for Everett's peculiar theory to become a topic of serious scientific and philosophical concern to mainstream academics.

At the conference, the strongest "Everettian" was physicist David Deutsch of the Center for Quantum Computation at University of Oxford. Deutsch asserted that people have been wasting time for decades debating whether or not the Copenhagen Interpretation is meaningful because, "Only Everett's theory consists of taking quantum mechanics seriously." Deutsch was joined in that analysis by conference organizers Simon Saunders and David Wallace of Oxford; both have written extensively on how to deal with uncertainty in an Everettian multiverse wherein all outcomes are realized.

Preserving probability (in the form of the Born Rule) is critical for the acceptance of the Everett method. The Born Rule defines how to determine the probability that we will see various outcomes when we observe a system that is in a particular quantum state. For example, it may indicate we have a 99% chance of seeing an electron at location A and only a 1% chance of seeing it at location B. But what does it mean for location A to be 99 times more likely than location B when each outcome has a branch of the universe in which it occurs? Our hold on the mathematics of quantum mechanics depends upon being able to assign probability measures to events occurring inside the universe we experience, and as well as to events inside the universes we do not see! (And exactly who "we" are in this scheme that contains vast but indeterminate numbers of fractured selves is an ontological question related to the probability problem.)

Wayne C. Myrvold of the University of Western Ontario and Hilary Greaves of Rutgers University made a presentation backing up previous work by Deutsch, Saunders, and Wallace delineating a "decision-theoretic" statistical method that purports to preserve the Born Rule, i.e. probability, throughout Everett's multiverse. David Z. Albert of Columbia University objected to this argument: "Talk about the probability of this or that future event would seem to make no sense unless there is something about the future of which we are uncertain, and there seems to be no room for any such uncertainty in the context of anything along the lines of an Everettian picture." Albert critiqued the Everettians for trying to prove that their theory allows for the Born Rule to function by assuming the belief that the Everett theory is true (they should rather show how a non-believer could come to believe it to be true). Albert, who is sympathetic to the "pilot wave" or "non-local hidden variables" interpretations of quantum mechanics originally made by Louis de Broglie and David Bohm, described the Everettian's explanation of probability as at best explaining our betting behavior, which as an explanation of physical evidence is "sheer madness."

Tim Maudlin of Rutgers University is also an admirer of the Bohmian approach. He critiqued the Everettians for not being able to tie their abstract theory to the world of experience and for failing to provide evidence of probability. Jeffrey Bub of the University of Maryland and Itamar Pitowsky of The Hebrew University of Jerusalem attacked both the Everettians and the Bohmians as "dogmatic" for assuming that there is a solution to the measurement problem. And James Hartle of the University of California, Santa Barbara, succeeded in carving out an ideological space somewhere between the competing factions. He said that modern cosmologists owe a debt to Everett for his insight that a closed universe can be described by a single wave function.

In a very animated talk, metaphysician John Hawthorne of University of Oxford strongly critiqued the Everett Interpretation as "vague" and "not explanatory." In an emotional response, Michel Janssen of the University of Minnesota, an Everett partisan, exclaimed, "Science is not metaphysics!"

The debate, although fifty years in the making, is just starting to get hot. It continues from September 21 to 24 at the Perimeter Institute in Waterloo, Canada. Whereas at Oxford, the philosophers outnumbered the physicists, that ratio will reverse at the Perimeter meeting. Many sparks will fly.


Posted for Peter Byrne