"The wavefunction never collapses"

[Roberto Pavani]

The wavefunction itself is not a local variable.

So MWI has a non-local variable (the wavefunction) that contains all outcomes from the start. Isn't that structurally a non-local hidden variable theory, similar to Bohmian mechanics?

 

[PeroK]

So MWI has a non-local variable (the wavefunction) that contains all outcomes from the start. Isn't that structurally a non-local hidden variable theory, similar to Bohmian mechanics?

I don't understand that comparison at all. Bohmian mechanics is the complete opposite of MWI!

 

[Roberto Pavani]

I don't understand that comparison at all. Bohmian mechanics is the complete opposite of MWI!

Both have a non-local universal wavefunction that contains all outcome information before measurement. Bohm selects one outcome, MWI keeps all. But the non-local predetermination is the same, the difference is only in how many outcomes you keep.
This is all following the answers to my questions on this thread

 

[PeterDonis]

does the branching occur at A's measurement, at B's measurement, or at both independently? And if both, in which order? (Since they are spacelike separated, the order is frame-dependent.)

If you're going to bring in relativity, you can't use non-relativistic QM; you have to use QFT. So you need to go find a reference that analyzes a situation like this using QFT. I'm not sure if MWI proponents have settled on a single version of how the MWI works with QFT, but my general sense is that they view "branching" as occurring at each measurement independently and spreading at the speed of light. So there is no "time order" required; where the future light cones of A's and B's measurements meet is the only relevant factor, and that's an invariant.

 

[PeterDonis]

MWI has a non-local variable (the wavefunction)

In QFT there is no wave function, there are only quantum fields. How QFT captures nonlocality is more complicated than it is for non-relativistic QM.

 

[PeterDonis]

Both have a non-local universal wavefunction that contains all outcome information before measurement.

No, this is not correct for Bohmian mechanics. In Bohmian mechanics, outcome information is contained in the unobservable particle positions. The wave function is part of the nonlocal equation of motion for the unobservable particle positions. Bohmian mechanics explains single outcomes by attributing them to the unobservable particle positions--each particle has just one position. Those positions are nonlocal hidden variables; the fact that they're nonlocal (because their equation of motion includes the nonlocal information in the wave function) is what allows them to produce correlations that violate the Bell inequalities.

 

[Roberto Pavani]

I'm not sure if MWI proponents have settled on a single version of how the MWI works with QFT

Thank you.
If there's no settled version, isn't the branching structure underdetermined?

In Bohmian mechanics, outcome information is contained in the unobservable particle positions.

Thank you for the correction on Bohm. But the structural point remains: in MWI it's which branch you're in.
Both are inaccessible to the observer, both contain the outcome information. The specific variable changes, but the epistemological structure is the same.

 

[PeterDonis]

the structural point remains: in MWI it's which branch you're in.

But in the MWI, all branches correspond to actually occurring outcomes. In Bohmian mechanics, only one does. That's a big difference.

 

[PeterDonis]

If there's no settled version, isn't the branching structure underdetermined?

I don't think so; AFAIK there isn't any question about which branches there are. The question is about how MWI explains the branching process in the context of QFT, as opposed to non-relativistic QM.

 

[Roberto Pavani]

But in the MWI, all branches correspond to actually occurring outcomes. In Bohmian mechanics, only one does. That's a big difference.

The question is about how MWI explains the branching process in the context of QFT, as opposed to non-relativistic QM.

On the first point: agreed, the ontological claim is different. But epistemologically, the observer's situation is the same, inaccessible information determines the experienced outcome.

On the second: if there's no question about which branches exist, then for my earlier example of the entangled singlet, are there 4 branches (all spin combinations) or 2 (only those allowed by angular momentum conservation)? If 2, what eliminated the other 2?

 

[PeroK]

On the first point: agreed, the ontological claim is different. But epistemologically, the observer's situation is the same, inaccessible information determines the experienced outcome.

On the second: if there's no question about which branches exist, then for my earlier example of the entangled singlet, are there 4 branches (all spin combinations) or 2 (only those allowed by angular momentum conservation)? If 2, what eliminated the other 2?

Where did you get the idea that MWI does not respect entangled states? And might allow impossible uncorrelated outcomes?

 

[Roberto Pavani]

Neve said so, I'm just embracing the MWI and asking questions based on the answers received here on the thread. The post you quoted was just asking a number: 4 vs 2

 

[bhobba]

The question is about how MWI explains the branching process in the context of QFT, as opposed to non-relativistic QM.

David Wallace's book, The Emergent Multiverse, discusses its extension to QFT.

Thanks
Bill

 

[bhobba]

It collapses all the way down!

With my mentor's hat on, please refrain from quoting long posts.

Thanks
Bill

 

[PeterDonis]

epistemologically, the observer's situation is the same, inaccessible information determines the experienced outcome.

Not really. The bolded phrase is correct for Bohmian mechanics, but it's not correct for the MWI. In the MWI, all outcomes are determined to happen. There is nothing that determines that you experience any one particular outcome; there is no additional information, beyond the fact that you observed a particular outcome, that determines that you observed that outcome. You say "which branch you are in", but that is not determined by any "inaccessible information" in the MWI--the only information that tells you which branch you're in is the outcome you observe, which is not inaccessible to you.

 

[PeterDonis]

what eliminated the other 2?

Nothing had to eliminate them, because they were never there in the wave function in the first place.

 

[Roberto Pavani]

Thanks for the clarifications.
If I have understood, the wavefunction already "knew" at preparation time that only two outcomes were possible.
That information was fixed before measurement and inaccessible until measurement. That is an hidden information aka hidden variable ?

 

[Roberto Pavani]

To clarify my earlier question: I was referring to the perspective of an observer inside one branch. From their view, the outcome was fixed at preparation and revealed at measurement, operationally indistinguishable from a hidden variable.

Conversely, from a hypothetical "outside" perspective that sees all branches simultaneously, all outcomes are realized, they are not mutually exclusive. But Kolmogorov's axioms require mutually exclusive outcomes for probability to be defined. If all branches are equally real, in what sense does the Born Rule assign probabilities?

 

[Matterwave]

There is no "hidden variable" in MWI, it is not a hidden (or extra) variable theory. All there is is Schroedinger evolution of the (universal) wave function / state vector. You could try to make an argument that MWI (and standard QM in general) is incomplete, a-la EPR, but that's a different discussion.

 

[Roberto Pavani]

Agreed from the global perspective (external observer with "potential outcomes"). But we are always observers inside one branch, and from that perspective, it is operationally indistinguishable from a hidden variable.
Note: "potential outcomes" allows Kolmogorov (mutually exclusive events → probability defined). "All outcomes realized" does not (no mutual exclusivity → probability undefined).

 

[Lord Jestocost]

There is no "hidden variable" in MWI, it is not a hidden (or extra) variable theory. All there is is Schroedinger evolution of the (universal) wave function / state vector. You could try to make an argument that MWI (and standard QM in general) is incomplete, a-la EPR, but that's a different discussion.

To my mind, proponents of the many-worlds interpretation (MWI) have merely problems with the probabilistic structure of standard quantum mechanics (QM), as it contradicts their simple deterministic ideas. Since MWI ‘rejects’ - so to speak - probabilities, probabilities are masqueraded as world splitting and all possibilities are imagined as to be realized in inaccessible alternative branches of the world.

I have doubts that there is any scientific value in discussing such things?

 

[Matterwave]

Agreed from the global perspective (external observer with "potential outcomes"). But we are always observers inside one branch, and from that perspective, it is operationally indistinguishable from a hidden variable.
Note: "potential outcomes" allows Kolmogorov (mutually exclusive events → probability defined). "All outcomes realized" does not (no mutual exclusivity → probability undefined).

As Peter has already said -- there are no variables that tell you which branch you go to. It's not something "hidden" from you. It just doesn't exist within the theory.

If by "operationally indistinguishable" you mean makes all the same predictions, then yeah all interpretations, as far as we know, make the same predictions. If we know an interpretation to not make the same predictions as standard QM we no longer call them interpretations but new theories.

 

[Roberto Pavani]

there are no variables that tell you which branch you go to

If nothing in the theory determines which branch you end up in, yet you end up in one, isn't that precisely the definition of "incomplete" in the EPR sense?

 

[PeroK]

If nothing in the theory determines which branch you end up in, yet you end up in one, isn't that precisely the definition of "incomplete" in the EPR sense?

No.

 

[Roberto Pavani]

Do you mean no from the perspective of an observer inside a branch or from an "ideal global observer" "outside" the branches ?

 

[PeroK]

Do you mean no from the perspective of an observer inside a branch or from an "ideal global observer" "outside" the branches ?

No from the perspective that I can see no similarly between the arguments put forward in the EPR paper and the MWI.

 

[olgerm]

If nothing in the theory determines which branch you end up in, yet you end up in one

You end up in all branches. Just you in 1 branch does not know anything about you in another branch. In all branches to understand things it is easier to ignore other branches, because you cannot interact with other nor get information from other branches if these are sufficently decohered/separated from this branch. To explain why you ignore other branches in your calculations/physics you can say either that wavefunction branched into totally separated/decohered branches or say that in persvective of this branch wavefunction collapsed to configutration of this branch .
Time-evolution by only schrödinger-equation predicts that and that explains experiments. Adding some new mechanism that removes some of branches would make model of laws of physics more complicated and would not add any new effects that would explain anything that time-evolution by only schrödinger-equation would not explain. So it makes more sense to belive simpler model.

 

[PeterDonis]

the wavefunction already "knew" at preparation time that only two outcomes were possible.

Yes, because that's how the state was prepared.

That information was fixed before measurement

The information that only two outcomes were possible, yes.

and inaccessible until measurement.

No. It's known because you know how the state was prepared. You don't need any information beyond the state preparation to know that only two outcomes are possible if the spins are both measured along the same axis.

That is an hidden information aka hidden variable ?

No. A hidden variable is something that is not in the wave function.

 

[PeterDonis]

Time-evolution by only schrödinger-equation predicts that and that explains experiments.

According to MWI proponents, yes. But not all physicists agree with them. So your claim here is interpretation dependent.

 

[PeterDonis]

we are always observers inside one branch, and from that perspective, it is operationally indistinguishable from a hidden variable.

No, it isn't. See my post #118 just now.