The Everettian, aka Many Worlds, interpretation of Quantum Mechanics states that the wave function of the universe never collapses and evolves according to the plain Schrodinger equation.

When a measurement of a quantum system occurs, instead of having wave function collapse, the system gets entangled with the measuring device and with the environment.

As a consequence, measurement always puts the state of the universe in a superposition of states each of which describes a "copy" of our macroscopic world, and these copies (the "branches") differ by the output of the measurement itself.

A physical phenomenon called decoherence implies that in MWI it is impossible to empirically establish the existence of the other branches (the ones "we don't live in" in a given moment).

In MWI every quantum measurement outcome is realized in a branch, and has 100% probability of happening in the universe (by "universe" it's meant the whole set of degrees of freedom, comprising all the branches). The usual probabilistic picture of QM is retrieved by the fact that each observer splits into copies of itself along with the rest of the macroscopic world it lives in. Probability is now translated into the indexical uncertainty that each future copy of the observer is subject to before peering at the measuring device and seeing the measurement result.

The role of probability in MWI is not universally accepted. But in this question I'm not asking about this.

Karl Popper is famous for his demarcation criterion, falsificationism, for distinguishing scientific theories from non-scientific ones.

Q. On the surface, it could seem that MWI fails Popper's criterion, because we can't empirically access the branches (worlds) we don't live in. Is this really true? What did Karl Popper think about the MWI of quantum mechanics? Did he write anything about that?

  • 1
    Carroll, a big defender of MWI, claims (without citation) that Popper "was very impressed with Many-Worlds" and called it "a completely objective discussion of quantum mechanics". However, Popper proposed his own interpretation of QM (based on propensities), see Del Santo, Karl Popper's Forgotten Role in the Quantum Debate, and he was an indeterminist, so it is unlikely to have been appealing to him.
    – Conifold
    Sep 12, 2022 at 10:24
  • @Conifold: What I'd be curious to know, most of all, is whether Popper would have ditched MWI based on his demarcation criterion.
    – Qfwfq
    Sep 12, 2022 at 10:45
  • Carroll says no:"Another objection is that the theory isn’t falsifiable, since we can’t observe the other worlds. But the worlds aren’t the theory; they are a prediction of the theory. To falsify a theory, we just have to do an experiment that is incompatible with one of its predictions. In the case of Everett, that’s simple; just find an example where a wave function doesn’t obey the Schrödinger equation even when it’s not interacting. In other formulations of quantum mechanics, that can happen, but not in Many-Worlds." But again, I do not know what his source on Popper is.
    – Conifold
    Sep 12, 2022 at 12:32
  • 1
    In any case, Carroll's assessment is contrary to the consensus, see Kragh, Testability and epistemic shifts in modern cosmology:"Multiverse physics or cosmology does not agree very well with the standard ‘‘definition’’ (or intuition) of science, which in this case was taken to include as a crucial element Popper’s falsifiability criterion... if anthropic multiverse physics is accepted as truly scientific, it will constitute an epistemic shift, a major methodological discontinuity in the history of modern physics."
    – Conifold
    Sep 12, 2022 at 13:25
  • @Conifold: is Kragh's opinion close to the consensus? (I haven't read the link yet)
    – Qfwfq
    Sep 12, 2022 at 13:52

2 Answers 2


Popper criticised the MWI in "Quantum Theory and the Schism in Physics" in section 5 of the introduction. He claimed that the MWI violates conservation laws. He also claimed that since quantum theory is time symmetric Everettians would have to say that worlds arise out of fusion of worlds as well as by splitting and that this is absurd because it would require that your memories can fuse with those of another version of you.

Advocates of the Everett interpretation, such as David Deutsch, would say that worlds can fuse if they haven't undergone decoherence and this is what explains interference, see for example "The Fabric of Reality" by David Deutsch Chapter 2. He also wrote a paper claiming that if we constructed a quantum computer AI different versions of it could fuse and this would allow a test of the Everett interpretation:


Deutsch would also disagree with the idea that the Everett interpretation is untestable. You say:

On the surface, it could seem that MWI fails Popper's criterion, because we can't empirically access the branches (worlds) we don't live in. Is this really true?

This argument doesn't make sense since, for example, it would imply that dinosaurs don't exist since nobody has ever seen a dinosaur, only fossils. Deutsch would say you should accept the existence of something if it is required by your current explanations, see "The Fabric of Reality" Chapter 4. And by that standard the existence of the multiverse as described by Everett is unavoidable because there is no other explanation of single particle interference experiments and many other more complicated experiments, such as the EPR experiment.

Deutsch explains more about the testability of the Everett interpretation in this paper:


  • 1
    "Advocates of the Everett interpretation, such as David Deutsch, would say that worlds can fuse" – only Deutsch says that, and it's because he uses the word "world" differently from everyone else. See the last paragraph of this answer. Nothing that Deutsch says on this subject is representative. Also, Popper's two objections in your first paragraph aren't philosophical; they're mathematical claims about the theory which are demonstrably false, so it really just shows that he didn't understand it. I suppose that still answers the question (or part of it).
    – benrg
    Sep 12, 2022 at 18:43
  • @benrg I agree that Popper's objections to the MWI are wrong, but they're not mathematical claims since momentum, energy, etc. are part of physics not mathematics. You can do maths without ever referring to energy or momentum or their conservation.
    – alanf
    Sep 12, 2022 at 20:47
  • 1
    @benrg You say that interference doesn't exploit other "worlds". there is a sense in which this is true cuz worlds are emergent large scale structure: a point that Deutsch has made arxiv.org/abs/quant-ph/0104033, pp. 19-20, see also philsci-archive.pitt.edu/8888/1/… However, single particle interference can only be explained by multiple versions of the particle. The interpretation controversy is about what is happening in reality to produce the results predicted by quantum mechanics, so the MWI is supposed to be ordinary quantum mechanics.
    – alanf
    Sep 12, 2022 at 21:10
  • I have to retract my claim that Deutsch uses the word "worlds" differently, since he doesn't seem to use it at all; but he talks about "universes" and describes the Everett/Wheeler ideas as "many-universes" so I think it amounts to the same thing. That aside, the fact remains that no one besides Deutsch says that "worlds [universes] can fuse", and to speak of "advocates of the Everett interpretation, such as David Deutsch" is to misrepresent the beliefs of Everett and most other advocates of RSI/MWI. I don't understand how your last comment changes that.
    – benrg
    Sep 13, 2022 at 2:09
  • @benrg Do you think DD is correct or wrong about the fact that multiple versions of a particle that can fuse are required for the explanation of single particle interference? If not, what is the correct explanation? And in any case why are you talking about whether he is representing other people correctly instead of talking about object level issues?
    – alanf
    Sep 13, 2022 at 6:44

I think you need to read Quantum Theory and the Schism in Physics, one of the three volumes of Popper’s 'Postscript to the Logic of Scientific Discovery'.

You should also look at Popper's Experiment.

The Measurement Problem, as addressed by the various interpretations of QM, is generally considered philosophy, not science. That is, as not regarding what can be distinguished by experiments, but about how we make sense of the world. That would make it not a topic where falsification is relevant (yet).

  • Popper held that there was no formal dividing line between philosophy and science. His demarcation was only for the border between science and pseudo science. Popper generally applied the falsification/modification process to his own philosophy, and to others.
    – Dcleve
    Sep 12, 2022 at 16:13
  • The measurement problem, in my view, pertains both to philosophy and to the foundation of physics. It's "philosophical" not only in the philosophers' sense, so to speak, but also in the "Einstein's sense".
    – Qfwfq
    Sep 12, 2022 at 18:58

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .