Multiverse theory -- Why don't strange things happen here sometimes?

[vanhees71]

I like Vaidman's emphasis of locality ;-)).

 

[pinball1970]

New to the party, I have read the first 20 or so posts. Should different/weird physics ever be expected? In our universe? I thought the whole point of MWI was that a branch is in a world we are not connected with AND why would it be weird?
Also I thought "multi verse" was to do with Guth's inflation not a measurement in QM.
Yes- My sources are popsci - just easier to get it out of the way.

 

[PeterDonis]

Should different/weird physics ever be expected? In our universe?

Not according to MWI; the MWI assumes our current laws of physics.

 

[LostInSpaces]

Not according to MWI; the MWI assumes our current laws of physics.

I think what most people who learn MWI through pop-sci think about is rather weirdness as in "If MWI is true, then that means I tunnel through the wall in some branch if I walk into the wall right now" or "the neurons in every person's brain could and thus will reconfigure such that we elect a duck as president of the whole planet". Neither of these explicitly violate the known laws of physics and since "everything happens" is preached, this is where the weirdness and confusion takes over

 

[PeterDonis]

Neither of these explicitly violate the known laws of physics

That "explicitly" hides the fact that the burden of proof is on the person proposing the weird scenario to show that it doesn't violate the known laws of physics--not just "explicitly" but at all. But nobody ever actually does that.

 

[vanhees71]

After reading some of the quoted texts in this forum, my main problem with MWI is that it promises more than it holds. First of all all of these texts are full of philosophy hiding the math and physics behind a wall of gibberish. The Vaidman article was the best I've seen yet:

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

The main obstacle from my physics-point of view is that they don't proof Born's rule for the probabilities, which finally are all that's observable also in this interpretation, from the other postulates but just assume them also more or less explicitly to begin with. That then the repetition of many realizations of an experiment on equally prepared systems (ensembles) reproduces these probabilities is just the central-limit theorem of probability theory. The probabilities a la Born are simply also assumed as they are in minimally interpreted QT. The rest is also just standard theory of "open quantum systems" with decoherence.

What I like is the strict avoidance of the quantum-classical cut a la Heisenberg, which is in no way observed yet but the validity of QT for larger and larger systems could be established. Also the emphasis of locality is on the plus side of this interpretation.

 

[GarberMoisha]

The MWI is not a complete interpeation like the CI. It would be strange if any interpretation was able to recover the Born's rule without somehow implicating the one doing the experiment.
How would they accomplish such a feat?
Physical matter is neither classical nor quantum. It is sort of both but in reality - neither. These are 2 aspects of something else that is matter in reality.

There is no law that dictates that quanta must make up a chair and act classically. There is also no law of physics that dictates that chairs must act quantum mechanically under certain circumstances.
These endless conceptual issues arise due to the ingrained intuitive Newtonian picture of what physical matter must be.

 

[LostInSpaces]

That "explicitly" hides the fact that the burden of proof is on the person proposing the weird scenario to show that it doesn't violate the known laws of physics--not just "explicitly" but at all. But nobody ever actually does that.

I agree, but this nuance is rarely, if ever, elucidated in popular science accounts of Everettian QM. Usually, the only caveat highlighted goes along the lines of "as long as it doesn't violate known laws of physics, it will happen." This dispels the idea that there are parallel universes where gravity is different or some other Lewisian modal worlds where dragons exist but does nothing to explain why MWI does not permit people to walk through walls in extremely low-weighted worlds. Afterall quantum tunneling is a thing, and as MWI is presented in pop-sci even ridiculously negligible probabilities are realized in some branches.

 

[PeterDonis]

Usually, the only caveat highlighted goes along the lines of "as long as it doesn't violate known laws of physics, it will happen."

Yes, but we don't actually know that all the things described meet that condition. For example, we don't know that the known laws of physics actually permit people to tunnel through walls, even with very low probability. We know electrons can tunnel through barriers, but people are not electrons.

Also, even the criterion as you state it is too broad. The actual criterion is that only the things that are included with nonzero amplitude in the actual wave function of the universe will happen. There is no reason to believe that everything that is possible according to the laws of physics has a nonzero amplitude in the actual wave function of the universe. Of course we don't know the actual wave function of the universe so we don't know which things actually have nonzero amplitudes in it and which don't. But that still doesn't justify the claim that everything that is possible according to the laws of physics will actually happen in some branch in the MWI.

 

[romsofia]

Is there really no review/textbook, which just explains the Everett approach in a clear physicists' way as is done for orthodox QT, which I find way better (as long as one leaves out the collapse postulate and other oddities of the Copenhagen-interpretation mixtures) in just postulating the complete theory, including the probabilistic meaning of the quantum state right from the beginning, instead of pretending to derive it by just introducing it somehow assuming it through the "backdoor"?

DeWitt attempts to do this in his book "The global approach to quantum field theory I" in chapter 8 (QUANTUM THEORY OF MEASUREMENT), chapter 9 (INTERPRETATION OF THE QUANTUM FORMALISM I) and chapter 12 (INTERPRETATION OF THE QUANTUM FORMALISM II).

DeWitt goes above my head at times, so I haven't really sat down and worked through what he puts down, but it was his intent in this chapters.

 

[LostInSpaces]

Yes, but we don't actually know that all the things described meet that condition. For example, we don't know that the known laws of physics actually permit people to tunnel through walls, even with very low probability. We know electrons can tunnel through barriers, but people are not electrons.

I agree; I am merely explaining my hypothesis of why MWI causes so many people to think that "Everything that can happen will happen". I don't blame them, as this is how it is often depicted. We know particles can tunnel through barriers, and people and barriers are made of particles, so it is natural to extrapolate that indeed people can also tunnel through walls by this logic.

 

[Nugatory]

does nothing to explain why MWI does not permit people to walk through walls in extremely low-weighted worlds. Afterall quantum tunneling is a thing, and as MWI is presented in pop-sci even ridiculously negligible probabilities are realized in some branches.

That’s true of all interpretations, not just MWI, and the answer is the same: There’s not a lot of operational difference between the statements “X cannot happen” and “The probability of X happening is less than 10^-10000 per trial”.

It is somewhat perplexing to me that the probabilistic nature of QM’s macroscopic predictions gets so much attention, yet no one cares about the probabilistic basis of statistical mechanics and all of thermodynamics.

 

[berkeman]

We know particles can tunnel through barriers, and people and barriers are made of particles, so it is natural to extrapolate that indeed people can also tunnel through walls by this logic.

This means nothing without a video...

 

[OCR]

This means nothing without a video...

OK, then. . . .

Mysterious Cube

.

 

[LostInSpaces]

That’s true of all interpretations, not just MWI, and the answer is the same: There’s not a lot of operational difference between the statements “X cannot happen” and “The probability of X happening is less than 10^-10000 per trial”.

It is somewhat perplexing to me that the probabilistic nature of QM’s macroscopic predictions gets so much attention, yet no one cares about the probabilistic basis of statistical mechanics and all of thermodynamics.

Sure, however, "Every branch is realized" is the crux here. In single-world interpretations, such low probabilities just do not happen at all; ever, there's not enough space-time in existence, so naturally, they can be entirely discarded. It's not only that it rarely happens; it does not happen. I believe people's grasp of probability vanishes when they are told that it does indeed happen, even if it is only to an infinitesimal portion of future observers, hence threads like this.

 

[GarberMoisha]

This is the result of the mode of thinking in which quantum mechanics is supposed to underly everything. You get quantum paradoxes all the way up to human scales. Cats, Moons, people walking through walls, Wigner's friends, etc.
You don't get these if you consider that both classical and quantum behavior are not fundamental but are distinct aspects of something else that is matter.

 

[gentzen]

Sure, however, "Every branch is realized" is the crux here. In single-world interpretations, such low probabilities just do not happen at all; ever, there's not enough space-time in existence, so naturally, they can be entirely discarded. It's not only that it rarely happens; it does not happen. I believe people's grasp of probability vanishes when they are told that it does indeed happen, even if it is only to an infinitesimal portion of future observers, hence threads like this.

Sean Carroll is good at explainig why this way of thinking about MWI is a mistake (i.e. he agrees with you, and tries to explain why MWI should not change your probability based decisions -- or expectations).

Max Tegmark manages to get this wrong in the worst possible way:

https://arxiv.org/pdf/quant-ph/9709032.pdf

Read the shortish paper. It's quite informative and addresses several misconceptions and criticisms.

This paper says more about its author (and his ideas about mathematics) than about the MWI. And what it actually says about MWI (IV.B) is worse than merely being wrong.

 

[Nugatory]

You don't get these if you consider that both classical and quantum behavior are not fundamental but are distinct aspects of something else that is matter.

We don’t need to go that far. If we consider quantum behavior to be fundamental, classical behavior for large collections of particles appears (analogous to the way that the ideal gas law emerges for macroscopic collections of molecules when we take Newton’s laws as fundamental).

The difficulty (which bothers some more than others) with considering quantum mechanics to be fundamental is not that it predicts that “strange things happen here” - it doesn’t. The difficulty is closing the gap between a prediction of various outcomes and the experience of exactly one outcome.

 

[vanhees71]

But isn't this as fundamental an observational fact as it is in classical physics? If we do a double-slit experiment with electrons, we just prepare a beam of electrons (or many single electrons) hitting the double slit with a pretty well defined momentum and then register, on which point of a screen far enough away from the double slit. It's a basic observational fact that for any electron we get "one pixel" registering one electron (where "a pixel" is just a macroscopically small region on a pixel detector, photoplate, etc.). According to QT where each individual electron is registered is random, and the probability distribution function is given by Born's rule from the calculated wave function at the place where we register the electrons. These are the simple observational facts we describe probabilistically with QT.

Of course, you can always ask, whether the necessity for a probabilistic description is due to some ignorance about the state of the electron in this situation, i.e., that there are maybe "hidden variables", whose determination would also determinate precisely the spot, at which we'll register a specific single electron. What's ruled out by many Bell-test experiments is that such a hidden-local-variable model can be made deterministic (realistic = all observables always take determined values) and local (space-like separated events cannot be causally connected).

Whether there may be some future deterministic theory describing everything in Nature, nobody can know, but the given empirical evidence is completely described by Q(F)T, except that there is no satisfactory QT which takes into account the gravitational interaction (or, if one takes the geometrodynamical paradigm of GR literally a QT of spacetime itself). IMHO that's the big open problem of contemporary physics and not some interpretational issues of QT, which are all solved theoretically (with relativsitic local QFT describing everything except the gravitational interaction) and consolidated empirically (with the long overdue Nobel prize of last year).

 

[GarberMoisha]

We don’t need to go that far. If we consider quantum behavior to be fundamental, classical behavior for large collections of particles appears (analogous to the way that the ideal gas law emerges for macroscopic collections of molecules when we take Newton’s laws as fundamental).

The difficulty (which bothers some more than others) with considering quantum mechanics to be fundamental is not that it predicts that “strange things happen here” - it doesn’t. The difficulty is closing the gap between a prediction of various outcomes and the experience of exactly one outcome.

If QM as it is known today were fundamental, you'd expect to get random detections here and there.
Instead, you get chairs and tables which have never been part of QM.
It can't be reformulated to be about chairs. Best you can do is say "it is important what we can say about nature, not what nature is".
It's dubious if you'd ever move forward if you instist on getting chairs from probabilities.

 

[PeterDonis]

you get chairs and tables which have never been part of QM

Sure they are: chairs and tables are just bound states containing $10^{25}$ or so atoms.

getting chairs from probabilities

Nobody claims that QM can do this, and it's the wrong way to get chairs and tables from QM anyway.

 

[Vanadium 50]

Why don't we see events with probability 10^-1000? Because the probability is 10^-1000! Interpretations don't even enter into it.

 

[Quantum Waver]

But isn't this as fundamental an observational fact as it is in classical physics? If we do a double-slit experiment with electrons, we just prepare a beam of electrons (or many single electrons) hitting the double slit with a pretty well defined momentum and then register, on which point of a screen far enough away from the double slit. It's a basic observational fact that for any electron we get "one pixel" registering one electron (where "a pixel" is just a macroscopically small region on a pixel detector, photoplate, etc.). According to QT where each individual electron is registered is random, and the probability distribution function is given by Born's rule from the calculated wave function at the place where we register the electrons. These are the simple observational facts we describe probabilistically with QT.

Of course, you can always ask, whether the necessity for a probabilistic description is due to some ignorance about the state of the electron in this situation, i.e., that there are maybe "hidden variables", whose determination would also determinate precisely the spot, at which we'll register a specific single electron. What's ruled out by many Bell-test experiments is that such a hidden-local-variable model can be made deterministic (realistic = all observables always take determined values) and local (space-like separated events cannot be causally connected).

Whether there may be some future deterministic theory describing everything in Nature, nobody can know, but the given empirical evidence is completely described by Q(F)T, except that there is no satisfactory QT which takes into account the gravitational interaction (or, if one takes the geometrodynamical paradigm of GR literally a QT of spacetime itself). IMHO that's the big open problem of contemporary physics and not some interpretational issues of QT, which are all solved theoretically (with relativsitic local QFT describing everything except the gravitational interaction) and consolidated empirically (with the long overdue Nobel prize of last year).

At the end of the DeWitt article on page 35, he suggests that the initial coherence of the universal wave function in the Big Bang may have empirical implications for cosmology. Sean Carrol has stated that MWI is popular among cosmologists. Carroll has mentioned working on showing how spacetime and gravity could emerge from Hilbert space (maybe in this paper: https://arxiv.org/pdf/2103.09780.pdf). So it's possible an interpretation of QM will have results for future physics. If you listen to people working in the foundation of QM, they do see it as making progress in physics, not simply philosophizing.

Aside from the above, the motivation is an attempt to understand what nature's doing when not being measured. The wave function works as a predictive model, but how is that so? If it's not describing anything real, then what makes it predictive? The beam of electrons are doing something in between being emitted and being registered at the wall. Unless you think they only exists as measurements, which raises the question of what does exist in the interim. A probability wave? What is that exactly? A potential value? What triggers the stochastic collapse from potential to a real, determinate value? Why does measurement make the difference? How is the measuring device not also a potential? DeWitt mentions on page 31 how the final state vector in his fifth equation does not have a unique value, because the apparatus has gone into superposition.

 

[PeterDonis]

What law of physics postulates that quanta must form bound states

You must be joking. The existence of bound states, for systems where they exist, follows from the Hamiltonian and Schrodinger's Equation. I think you need to learn some basic QM.

and act classically as chairs?

The laws of physics don't "postulate" chairs. Chairs are just one of a zillion possible bound states that are allowed by the laws of physics. Expecting the laws of physics to specifically tell you anything about chairs is foolish.

There is no adequate way to get chairs and tables from QM.

If you're going to be strict about "adequate" and require an explicit derivation, then classical physics is no more "adequate" than QM, since it doesn't say anything specific about chairs either. (In fact, strictly speaking, it's less adequate, since classical physics can't even explain atoms.)

 

[Quantum Waver]

Sure they are: chairs and tables are just bound states containing $10^{25}$ or so atoms.Nobody claims that QM can do this, and it's the wrong way to get chairs and tables from QM anyway.

Under MWI, the bound states of those atoms behave quantum mechanically, so it should be possible to calculate the likelihood that all the atoms would tunnel at the same time through the wall the table is next to. In the DeWitt article you linked to, he calls these "Maverick worlds" on page 34. However, he does admit that if Hilbert space is small enough, they may be excluded by the universal wave equation. Also that if there are Maverick worlds where the very low probability events happen regularly, then life might not be able to evolve and it would be devoid of observers.

 

[Quantum Waver]

Why don't we see events with probability 10^-1000? Because the probability is 10^-1000! Interpretations don't even enter into it.

True, it's overwhelmingly likely that we're normal observers and don't get to see those extremely rare events. But even without MWI, if the universe is infinite with a normal distribution of matter throughout, then it follows there will be regions with observers like us who do witness extremely low probability events. Given the measured topology is near flat, that's a possibility in cosmology.

 

[PeterDonis]

Under MWI, the bound states of those atoms behave quantum mechanically, so it should be possible to calculate the likelihood that all the atoms would tunnel at once through the wall the table is next to.

Such a calculation assumes that the potential and Hamiltonian for the case "macroscopic object next to wall" are sufficiently similar to the potential and Hamiltonian for the astronomically simpler case of "single quantum particle next to microscopic potential barrier" to make the known solution of the former an acceptable proxy for the unknown (since we have no way of actually writing down the applicable potential and Hamiltonian) solution of the latter. But this assumption seems to me to be handwaving to support an already determined conclusion, rather than anything that can reasonably be extracted from QM.

 

[Quantum Waver]

Such a calculation assumes that the potential and Hamiltonian for the case "macroscopic object next to wall" are sufficiently similar to the potential and Hamiltonian for the astronomically simpler case of "single quantum particle next to microscopic potential barrier" to make the known solution of the former an acceptable proxy for the unknown (since we have no way of actually writing down the applicable potential and Hamiltonian) solution of the latter. But this assumption seems to me to be handwaving to support an already determined conclusion, rather than anything that can reasonably be extracted from QM.

But under MWI, all the particles are in a superposition for all possible states, which includes the particles being on the other side the wall. The table and wall aren't fundamental to this. They're just emergent configurations of particles in decohered branches, but quantum mechanically speaking, they would be in a superposition. Unless there's something about the bound states prohibiting this.

 

[PeroK]

True, it's overwhelmingly likely that we're normal observers and don't get to see those extremely rare events. But even without MWI, if the universe is infinite with a normal distribution of matter throughout, then it follows there will be regions with observers like us who do witness extremely low probability events. Given the measured topology is near flat, that's a possibility in cosmology

There was a thread somewhat recently about Bayesian methods being used to calculate the "probability" that the universe is infinite.

However, if you stick to purely frequentist probability methods, there is no way to calculate such a probability.

It's not that the Bayesians are wrong, but that statements like "it's 90% likely the universe is infinite" may never have a fully objective meaning.

That's why, IMO, statements about hypothetical planets where the laws of physics continue indefinitely to appear not to hold are potentially meaningless. The existence of such worlds is always predicated on some aspect of the theory that can never be observed.

This is very different from mathematical infinities, such as an infinite sequence, which can be proven to exist abstractly based on definite axioms.

 

[PeterDonis]

under MWI, all the particles are in a superposition for all possible states

No, they're not, although this is a common hand-waving claim made in discussions of the MWI. The universal wave function only includes terms that arise by unitary evolution from the universal wave function at the Big Bang and the applicable Hamiltonian. There is no guarantee that "all possible states" as we would naively interpret that term are included.

 

[PeterDonis]

this is a common hand-waving claim made in discussions of the MWI

Btw, I've often wondered why MWI proponents make such claims, since it seems like they're sawing off their own branch. It would seem to make much more sense under the MWI to explain the fact that we don't observe things like macroscopic objects tunneling through walls by invoking the fact that such outlandish things never had a nonzero amplitude in the universal wave function to begin with.

 

[Quantum Waver]

No, they're not, although this is a common hand-waving claim made in discussions of the MWI. The universal wave function only includes terms that arise by unitary evolution from the universal wave function at the Big Bang and the applicable Hamiltonian. There is no guarantee that "all possible states" as we would naively interpret that term are included.

Yeah, but the argument would be that the position of the particles in both the wall and table had become entangled at some point since the Big Bang because of some interaction like radioactive decay. They're not isolated systems.

Btw, I've often wondered why MWI proponents make such claims, since it seems like they're sawing off their own branch. It would seem to make much more sense under the MWI to explain the fact that we don't observe things like macroscopic objects tunneling through walls by invoking the fact that such outlandish things never had a nonzero amplitude in the universal wave function to begin with.

Because MWI isn't classical and QM allows for such possibilities. It's only outlandish to us normal branchers.

 

[PeterDonis]

the argument would be that the position of the particles in both the wall and table had become entangled

According to the MWI pretty much everything in the observable universe would have some entanglement with pretty much everything else. But that's not enough by itself to show that the table tunneling through the wall has a nonzero amplitude.

However, to get back to the original point of the thread, it's not even necessary to postulate events like "table tunneling through wall" to see that the MWI implies that there are "outlandish" branches in the wave function. A simple series of a sufficient number of Stern-Gerlach experiments is enough. So the specific question about tables tunneling through walls is really moot as far as this thread is concerned.

 

[Quantum Waver]

There was a thread somewhat recently about Bayesian methods being used to calculate the "probability" that the universe is infinite.

However, if you stick to purely frequentist probability methods, there is no way to calculate such a probability.

It's not that the Bayesians are wrong, but that statements like "it's 90% likely the universe is infinite" may never have a fully objective meaning.

That's why, IMO, statements about hypothetical planets where the laws of physics continue indefinitely to appear not to hold are potentially meaningless. The existence of such worlds is always predicated on some aspect of the theory that can never be observed.

This is very different from mathematical infinities, such as an infinite sequence, which can be proven to exist abstractly based on definite axioms.

I don't know whether statistics is necessary. It's really a question of whether the universe has zero curvature, in which case space extends infinitely. Then it becomes a question of whether matter/energy are uniformly distributed, and the observable universe is typical of that. It's known the actual universe is larger than the observable one, otherwise what's observable wouldn't increase over time as light has a chance to reach us from more distant regions.

The point is that a large enough (wouldn't have to be infinite) universe will have every physically possible event happen, including the extremely low probability ones. So it's not just an oddity of MWI.

 

[PeterDonis]

The point is that a large enough (wouldn't have to be infinite) universe will have every physically possible event happen

As long as "physically possible" means "consistent with the initial conditions". Again, we don't know that the initial conditions included everything that is possible according to the laws of physics.