Why do we experience the Born Rule in Many Worlds?

[jbergman]

TL;DR Summary

In real QM experiments we see outcomes that align with the Born rule. How is such a phenomenon justified in a Many Worlds Interpretation?

I hate to bring up an old saw again, but I've been listening to Carroll and some others wax poetically about Many Worlds. And Jurek's work on decoherence and pointer states seem to address some of the problems with the MWI.

However, I haven't seen any compelling explanation as to why an observer in this universe would see experimental results consistent with the Born rule.

If you do a repeated spin up or down measurement 3 times you will have 8 different worlds where each observer will have seen some distinct combination of outcomes, {UUU, UUD, UDU, UDD, DDD, DDU, DUD, DUU} which seems to suggest that there is an equal probability of any outcome even if the preparation is such that the magnitude of the |U> coefficient has a magnitude of say .95.

I haven't seen any explanation that convincingly addresses this question and it seems like the Born Rule statistics are fundamentally at odds with what one would expect to experience if MWI were true.

 

[PeterDonis]

I haven't seen any compelling explanation as to why an observer in this universe would see experimental results consistent with the Born rule.

That's because there isn't one. There have been many attempts to derive the Born Rule in the context of the MWI, but none of them has gained a clear consensus that it is correct. Proponents of the MWI assert that sooner or later some convincing explanation will be found, but as it stands today that is just an assertion of their belief.

 

[vanhees71]

A nice summary of the efforts to derive Born's rule from the other standard postulates can be found in

S. Weinberg, Lectures on Quantum Mechanics

 

[LostInSpaces]

IIRC Sean Carroll, in his Something Deeply Hidden book, essentially concludes that some observers will see Born Rule violations, but most won't, and that's that. It seems most Everettians these days go along with this sentiment that some observers will have 'bad luck' and observe outcomes that seem to violate QM, but since most won't most of the time the appeal of MWI is worth the sacrifice of complete coherence across all branches. This stance has always annoyed me, particularly when you contrast it to the response superdeterminism always gets about rendering the scientific method 'impossible'. To me, the concept of postulating that "sometimes the universe violates QM for some observers" is no better than the apparent 'conspiracy' of superdeterminism.

Emily Adlam's work from 2014 is an excellent summary of this problem in Everett: https://philpapers.org/rec/ADLTPO-14

 

[vanhees71]

That's interesting! Is there a scientific paper, where one can find the arguments for, why MWI should predict different probabilities than what comes out of standard QT using Born's rule, and how these different probabilities are derived? Is all this then consistent with the experiments directly tesing Born's rule, as, e.g., in this three-slit interference experiment:

https://arxiv.org/pdf/1007.4193

 

[Demystifier]

essentially concludes that some observers will see Born Rule violations, but most won't

That's also true in standard QM, or in any probabilistic process for that matter. There is always a chance that the coin will land tails ten times in a row.

 

[Demystifier]

That's interesting!

Actually it's not, because it doesn't happen under conditions that can be controlled by an experimentalist.

 

[jbergman]

That's also true in standard QM, or in any probabilistic process for that matter. There is always a chance that the coin will land tails ten times in a row.

Yes. We have to be careful how we phrase things when talking about probabilities. Technically, the born rule could be wrong and we live in a universe where a highly unlikely series of events have occurred to make it appear true.

However, with enough measurements we render such situations less likely. For instance, if we measure the spin 1 million times instead of 3 times with a preparation that is mostly spin up, in the MWI if we assign equal weight to each world, you would be much more likely to end up in a world with an equal number of spin up and spin down measurements than in a world with mostly spin up measurements. Whereas in our actual universe with the born rule that isn't true.

 

[vanhees71]

Actually it's not, because it doesn't happen under conditions that can be controlled by an experimentalist.

Does this mean, it's something unobservable in principle or is it just difficult to realize? I always thought that at the end also "Many Worlds" doesn't give anything observable that's different from standard QT.

 

[jbergman]

Does this mean, it's something unobservable in principle or is it just difficult to realize? I always thought that at the end also "Many Worlds" doesn't give anything observable that's different from standard QT.

There are two parts to the born rule.
1. Probability of observing an observable is the magnitude of the coefficient in a suitable basis.

2. After measurement the state vector "collapses" to the measured stated.

Many Worlds is compatible with 2 by asserting that within a branch you are effectively in the measured state.

The problem is with 1. What most many worlders do is add the Born Rule even if it is in some sense unexplained or even incompatible with the interpretation. Therefore I don't think you could perform an experiment to contradict it.

Instead you can only ask how the Born rule is compatible. To me it isn't. I think the closest attempts at really justifying it in the MWI are Zurek's with his envariance paper. His out is that he ends up creating more worlds proportional to the probability of each outcome. I don't see how this is justified based on the arguments I've seen for how entanglement works.

 

[LostInSpaces]

That's also true in standard QM, or in any probabilistic process for that matter. There is always a chance that the coin will land tails ten times in a row.

It's different in MWI though, as it guarantees that this happens to observers all the time, thus undermining these observers' ability to make any definite claim about probability, while forcing those who do observe the Born Rule to assume that they are simply the 'lucky ones', despite sharing ancestor observer state. It becomes circular quickly. One can, of course, still insist, as many MWIers do, that "this is just how it is", but to me that is not a satisfactory solution to MWI's probability problems

 

[Demystifier]

It's different in MWI though, as it guarantees that this happens to observers all the time

How MWI guarantees that?

 

[Demystifier]

I always thought that at the end also "Many Worlds" doesn't give anything observable that's different from standard QT.

Right.

 

[jbergman]

Right.

I think that depends on whether you think Many Worlds entails the Born Rule probabilities or not.

I personally don't believe it does because I think the most logical thing is to assign a probability measure baesed in the number of worlds that will be generated.

 

[Demystifier]

I think that depends on whether you think Many Worlds entails the Born Rule probabilities or not.

I personally don't believe it does because I think the most logical thing is to assign a probability measure baesed in the number of worlds that will be generated.

Agreed with both remarks.

MWI in its simplest and most natural form predicts probabilities that do not agree with observations. Hence those who think that MWI is on the right track must construct more complicated versions of MWI. There is no consensus within the MWI community how do it. For outsiders, it is a good reason to not take MWI seriously.

 

[vanhees71]

Right.

So the claim that there are deviations in the predictions of MWI vs. standard minimal-interpreted QT, is fake.

 

[PeterDonis]

MWI in its simplest and most natural form predicts probabilities that do not agree with observations.

I wouldn't state it that way. I would say that MWI in its simplest and most natural form does not predict probabilities at all; it says that every time a measurement happens, all possible results happen. Everything is completely deterministic and there are no probabilities anywhere.

 

[vanhees71]

What does then MWI predict at all? Now we have contradicting claims: (a) MWI predicts different probabilities than standard QT with Born's rule (without any reference in the literature, (b) this morning this claim was taken back and now (c) it's claimed, that MWI doesn't predict any probabilities.

It would be great to have a (review) paper about the theory, which says what it really says in the sense as a physical theory. I tried several times to understand this interpretation an failed. Also Everett's original RMP doesn't tell me, what's the content of it might be.

Now, looking at, how QT is used in the real world to confront its predictions as a physical theory with observations, one simply uses Born's rule to calculate probabilities for the outcome of measurements. It doesn't matter, whether the universe splits, only because I see an electron on a fluorscence screen. I just see this detection of an electron at a before unknown spot. All the knowledge of the quantum state of this electron implies is the probability for registering this electron at some point on the screen. Also there's always a unique outcome of the measurement, i.e., a single electron makees one point on the screen, and thus I need many electrons, always prepared in the same quantum state, to be able to make the statistics about the distribution of this ensemble of electrons. So with MWI there's nothing gained, and I'm again back at the minimal statistical interpretation ;-)).

 

[PeterDonis]

What does then MWI predict at all?

The MWI is an interpretation of QM, and like all interpretations, it makes the same predictions as "basic" QM without any interpretation.

Those who are (to say the least) skeptical of the MWI say that the MWI has to do a lot of handwaving (to say the least) in order to make those predictions.

Now we have contradicting claims: (a) MWI predicts different probabilities than standard QT with Born's rule (without any reference in the literature, (b) this morning this claim was taken back and now (c) it's claimed, that MWI doesn't predict any probabilities.

These are not claims about what the MWI predicts according to its proponents. I'm not sure where you're getting (a) or (b) from. (c) was my post, and it should be clear that I was talking, not about the MWI as its proponents claim, but about what a skeptic comes up with by just looking at the unitary math without any wave function collapse, which is the first thing the MWI tells you to do. There are also plenty of papers in the literature expressing similar skepticism: that's why the title question of this thread is an open question in the literature.

It would be great to have a (review) paper about the theory, which says what it really says in the sense as a physical theory.

There are many, many papers about the MWI in the literature. AFAIK Everett's original thesis is not taken to be the "canonical" version of the MWI by its proponents; Bryce DeWitt's early 1970s (IIRC) paper is a better version. But there are also many more recent papers. I don't think your implied claim here that this interpretation has never been properly expounded in the literature is correct. You might not think the interpretation is correct, and you would find plenty of physicists to agree with you. But that's not the same as the interpretation not having any good description in the literature.

looking at, how QT is used in the real world to confront its predictions as a physical theory with observations, one simply uses Born's rule to calculate probabilities for the outcome of measurements

And doing this does not require adopting any QM interpretation at all (unless you want to call this the "shut up and calculate" interpretation--but as PF uses the term "interpretation", this is not an interpretation, it's just the basic math of QM). So if this is all you think QM is, then @Demystifier is right that you should not even be posting in this subforum, since this subforum is for the express purpose of not just discussing the basic math of QM, but particular interpretations. Threads whose questions can be resolved by just using the basic math of QM belong in the regular QM forum, not this one.

 

[vanhees71]

The MWI is an interpretation of QM, and like all interpretations, it makes the same predictions as "basic" QM without any interpretation.

Those who are (to say the least) skeptical of the MWI say that the MWI has to do a lot of handwaving (to say the least) in order to make those predictions.These are not claims about what the MWI predicts according to its proponents. I'm not sure where you're getting (a) or (b) from. (c) was my post, and it should be clear that I was talking, not about the MWI as its proponents claim, but about what a skeptic comes up with by just looking at the unitary math without any wave function collapse, which is the first thing the MWI tells you to do. There are also plenty of papers in the literature expressing similar skepticism: that's why the title question of this thread is an open question in the literature.

(a) was claimed in #15 by @Demystifier . Since the probabilties according to Born's rule have always been confirmed in all tests of QT, the claime that "MWI in its simplest and most natural form predicts probabilities that do not agree with observations." is equivalent to the statement that MWI predicts different probabilities than standard QT with Born's rule.

(b) was stated, again by @Demystifier in #13.

There are many, many papers about the MWI in the literature. AFAIK Everett's original thesis is not taken to be the "canonical" version of the MWI by its proponents; Bryce DeWitt's early 1970s (IIRC) paper is a better version. But there are also many more recent papers. I don't think your implied claim here that this interpretation has never been properly expounded in the literature is correct. You might not think the interpretation is correct, and you would find plenty of physicists to agree with you. But that's not the same as the interpretation not having any good description in the literature.

Well, maybe I'm simply not able to understand the interpretation. Do you have the complete reference of DeWitt? Usually I can understand his papers pretty well.

And doing this does not require adopting any QM interpretation at all (unless you want to call this the "shut up and calculate" interpretation--but as PF uses the term "interpretation", this is not an interpretation, it's just the basic math of QM). So if this is all you think QM is, then @Demystifier is right that you should not even be posting in this subforum, since this subforum is for the express purpose of not just discussing the basic math of QM, but particular interpretations. Threads whose questions can be resolved by just using the basic math of QM belong in the regular QM forum, not this one.

A mathematical scheme without physical interpretation is not a physical theory either. You need at list the minimal statistical interpretation. This minimal statistical interpretation is, how QT is used when it's compared with real-world experiments. It's no about the basic math of QM but about interpretation. Also you cannot properly discuss also interpretations without the "basic math of QM".

On the other hand, e.g., when I try to discuss basic mathematical and physical properties of relativistic QFT (locality), it's almost instantaneously shifted to the interpretations subforum, because obviously it's pretty much intermingled with interpretational issues, because also here the issues are indeed interpretational.

You can't have a physical theory without interpretation, but you can have a minimal interpretation, which avoids as much as possible philosophical implications, that do not add anything to the physics content, for which however you need this minimal interpretation.

 

[PeterDonis]

A mathematical scheme without physical interpretation is not a physical theory either. You need at list the minimal statistical interpretation. This minimal statistical interpretation is, how QT is used when it's compared with real-world experiments.

This "minimal interpretation" is what I was referring to as "the basic math of QM". Sorry for the confusion.

 

[PeterDonis]

Do you have the complete reference of DeWitt?

It's referenced and briefly described in the Stanford Encyclopedia of Philosophy article (which also has a lot of other literature references):

https://plato.stanford.edu/entries/qm-manyworlds/

The 1970 DeWitt paper itself appears to be online here, but the scan is not very good:

https://pubs.aip.org/physicstoday/a...um-mechanics-and-realityCould-the-solution-to

 

[Demystifier]

You can't have a physical theory without interpretation, but you can have a minimal interpretation, which avoids as much as possible philosophical implications, that do not add anything to the physics content, for which however you need this minimal interpretation.

So, do you agree that even the minimal interpretation does not remove the philosophical implications completely?

 

[vanhees71]

Since it's minimal there are no additional mrtaphysical assumptions than those inferred from observations. The necessity for the probabilistic interpretation of the quantum state was enforced by the discrepancy between Schrödingers original interpretation of his wave function in the sense of classical field, in close analogy to light being described by the classical em. field. This didn't withstand the observational facts because of the observed particle properties. This of course was also the other problem with, what's called "wave-particle dualism" in the old QT, and also this is solved by the probabilistic reinterpretation of the wavd function and finally the representation independent formalismen a la Dirac, which delivered the common framework for both then known formulations, wave and matrics mechanics.

Unfortunately then came all the philosophical confusion into the game, and that's why we have all these debates, although the theory in its minimal interpretation is all there is needed for its use as aphysical theory.

But back to MWI: What's gained, frm a philosophical point of view, from the metaphysical conjecture of these splittings into parallel worlds, although all physicists agree on the outcome and storage of measurement results? Apparently there's only "one world" according to our experience.

 

[PeterDonis]

the theory in its minimal interpretation is all there is needed for its use as aphysical theory

Which, while true, is off topic in this subforum, since this subforum is explicitly for discussing interpretations that go beyond the "minimal interpretation", not because it makes any difference to the physical predictions, but because, to proponents of the various interpretations, they shed light on "what is really going on". You don't have to agree with that, but that is what this subforum is for.

What's gained, frm a philosophical point of view, from the metaphysical conjecture of these splittings into parallel worlds, although all physicists agree on the outcome and storage of measurement results?

MWI proponents believe that the MWI solves the problem of how the universe picks which one of the possible measurement results becomes actual--by saying that all of the possible measurement results become actual. Or, to put it another way, it solves the problem of how the universe decides when to evolve according to unitary dynamics and when to suspend that and do something else--by saying that the universe always evolves according to unitary dynamics and never does anything else.

Apparently there's only "one world" according to our experience.

The MWI explains this by saying that "our experience" corresponds to one particular branch of the wave function.

 

[vanhees71]

Then you have the same claimed problem as with the minimally interpreted theory, i.e., it doesn't answer the question, why we are in this specific branch of the wave function. Further from the article by DeWitt I again find that in fact Born's rule is indeed just conjectured as in the minimal interpretation and not derived as claimed. That's also nicely discussed in Weinberg's QM textbook. So isn't if much simpler to just accept that Nature behaves probablistically, and that the probabilities of the outcome of measurements are given by the quantum-theoretical formalism, including Born's rule? That there are definite outcomes in measurements is just due to empirical evidence. Why is this epistemic view still not acceptable for many philosophers?

 

[jbergman]

I wouldn't state it that way. I would say that MWI in its simplest and most natural form does not predict probabilities at all; it says that every time a measurement happens, all possible results happen. Everything is completely deterministic and there are no probabilities anywhere.

Although true, probabilities come into MWI with the assumption that an observer can only experience a single branch of the wave function.

When I conduct a spin measurement experience, in MWI, I will split into 2 copies of myself that live on different isolated branches. Afterwards, from the point of view of the observer on one of the branches it will have been a random process as to which branch they ended up on.

So before an experiment, I can ask the question what is the probability I will end up on a branch having observed X.

 

[PeterDonis]

Then you have the same claimed problem as with the minimally interpreted theory, i.e., it doesn't answer the question, why we are in this specific branch of the wave function.

According to the MWI, the question itself doesn't make sense: "we" are in all branches of the wave function (at least if we are considering branches resulting from some measurement we have made), not just one.

in fact Born's rule is indeed just conjectured

Yes, as I already said in post #2 of this thread.

isn't if much simpler to just accept that Nature behaves probablistically, and that the probabilities of the outcome of measurements are given by the quantum-theoretical formalism, including Born's rule?

Apparently MWI proponents don't think so.

That there are definite outcomes in measurements is just due to empirical evidence. Why is this epistemic view still not acceptable for many philosophers?

You'd have to ask them. This is not the kind of question that is answerable in a PF discussion.

 

[PeterDonis]

probabilities come into MWI with the assumption that an observer can only experience a single branch of the wave function.

No, this is misstated. If I make a measurement, I end up in all of the resulting branches of the wave function. The "I" in each branch experiences a different measurement result, but in all othe respects the "I"'s in all the branches are identical, and they all have an equal claim to be "I".

When I conduct a spin measurement experience, in MWI, I will split into 2 copies of myself that live on different isolated branches.

This is also misstated, or at least misleading. Time evolution in the MWI is always unitary, and unitary evolution can't make "copies" of anything. The "copies" you refer to are decoherent branches of the wave function, but no "copying" process is required to create such branches. All that is required is entanglement between the measured system, measuring apparatus, and observer, due to unitary interactions between them, and decoherence of the resulting wave function branches.

Afterwards, from the point of view of the observer on one of the branches it will have been a random process as to which branch they ended up on.

Sort of. See below.

So before an experiment, I can ask the question what is the probability I will end up on a branch having observed X.

You can of course ask that question, but under the MWI it's not clear that the question makes sense, since there are no probabilities in the MWI; all possible measurement results are realized, and the dynamics are entirely deterministic (since unitary evolution is deterministic).

Trying to make sense of this in the context of the MWI is at least a big part of, if not the primary part of, the problem pointed at in the title of this thread.

 

[vanhees71]

So MWI doesn't solve a problem but poses one? The greatest obstacle, however, seems to be that we experience a single world, i.e., there are fixed measurement results, and "I" at least have the impression to live in one and only one branch. There's no hint of the existence of the other branches.

 

[PeterDonis]

So MWI doesn't solve a problem but poses one?

According to its proponents, it solves one. If you don't agree with its solution, then of course you're not an MWI proponent. But, as I have already said, disputes about which intepretation is correct are not going to be solved by PF discussion and, according to the guidelines for this forum, are off topic here. The best we can do is to figure out, according to the most current literature, what each interpretation's proponents say it says, their arguments for it, and skeptical arguments in the literature against it.

 

[gentzen]

This doesn't make sense. If you have a forum on "interpretation of QT", it must be allowed to discuss about, which interpretation is considered correct or not for which reasons. Which sense should otherwise proposals for ever new interpretations have?

That is indeed tricky, especially with respect to MWI. In a certain sense, both Bohmian mechanics and "Everett" are still only math, which still needs to be interpreted, despite appearances to the opposite. For BM, at least there are only very few different schools, essentially just the Valentini school, the typicality school, and the rest. For MWI, there are nearly as many schools as there are variants of Copenhagen, and some of those schools are somewhat problematic in their behavior and claims:

... I still don't see the huge contradictions with the video that you notice, but maybe this is not important, see next point.

> And the comment about Many Worlds at the end of the video capped it off.

I initially brushed off those remarks about Many Worlds as irrelevant. That was a mistake, they should have triggered me to check whether Harvey R Brown is a well known MWI proponent, and if yes which "school". "Faculty of Philosophy and Wolfson College, Oxford" indicates that he indeed might be a well known MWI proponent of "Simon Saunders' decoherence/emergence" school. Indeed, he even cowrote a paper with David Wallace. And he seems to have a track record of not getting even Tim Maudlin's clearest points (like "Note that Brown and Wallace see fit to avail themselves of the terminology ‘configuration space’ while simultaneously denying the very existence of the particles that might be in any configuration and, presumably, the low-dimensional space that the multiplicity of particles might commonly inhabit.").

Therefore, I should have interpreted Brown's reference to Tim Maudlin as a way for Brown to express his disagreement. Instead I interpreted it as an acknowledgment of Maudlin's point. Sorry. So I was definitively wrong when I wrote:

> For me, this proves beyond reasonable doubt that you are accusing Harvey Brown of commiting a mistake that he did not commit!

I hope you understand that I don't want to publicly embarrass myself even more. So I will stop this "public discussion" here, especially since ...

How should you handle this state of affairs, in non-academic, public dicussions?

 

[vanhees71]

I'd tell the public precisely these state of affairs. It's symptomatic that people misunderstand each other due to the uncertainty in their statements. Everybody can then come to an own opinion about the issue.

 

[PeterDonis]

Thread closed for moderation.

 

[PeterDonis]

After moderator review, this thread will remain closed. Thanks to all who participated.