Is MWI Self-Contradictory and Does Time Travel Need a New Approach?

[Demystifier]

And I don't get why people are so obsessed with Born rule in MWI. It is so obviously correct It's part of the math of QM, it works equally well in all interpretations, it has been proven to be the only sensible measure that works. It is intuitively well understood (in thermodynamics sense as a ratio of the number of microstates leading to different macrostates). Yes, counting those microstates is a pain because it all comes from the measurement and measurement is a complicated multi-stage process, but the progress has been made in this direction. Actually this problem (going from micro to macro) applies equally to all interpretations, at least those that try to study the process of measurement rather just postulate it as in Copenhagen.

The question is not whether the Born rule can be incorporated into MWI. Of course it can, in many ways. The question is whether incorporation of the Born rule ruins the initial beauty and simplicity of MWI. The answer is that it does! Perhaps not more than in other interpretations, but not less either. So if MWI with the Born rule incorporated is no simpler or more beautiful than other interpretations, then why one should prefer it over other interpretations? I'm sure one can find many reasons to still prefer it, and that is fine, but the point is that the initial STRONGEST argument for preferring MWI is now lost. If there was no such a loss, then MWI could easily dominate on the world-interpretation map, and that's why people are so obsessed with the Born rule, in both MWI and anti-MWI camps.

 

[Fyzix]

First off: no Tim Maudlin's paper is not a "review" it's a direct response to the problem of ontology in MWI...

Also, you are completley wrong.
MWI isn't a theory, it's a idea, but that idea makes certain claims, which IT HAS TO PROVE.
It's not up to the skeptic.
If I state "There is a God" is it up to other people to disprove it? No, it's up to me who claim this God exists.
If I make my statement more detailed like "God moonwalks on the moon 3:50 am each night" we could test this claim.
Exact same way we can test Born Rule in QM and show it doesn't work in MWI means YES MWI as a "pure interpretation" is disproved.

Like Demystifier just said, SURE you can introduce ADDITIONAL POSTULATES to get Born Rule, such as particles, wavefunction collapse, many minds dualism of mind etc.
However would you still call this many worlds? It's no longer pure wavefunction and this is what is being discussed here.

The claim that pure wavefunction taken as ontologically real can somehow make worlds (which is questioned by many) and that these worlds correspond to Born Rule (which is downright disproved at this point).
So in this sense Bohm is also a "many worlds" interpretation in that it has the wavefunction, but also particles.
The particls allow the Bohmian interpretation to get both of the things MWI can't: a physical ontology that makes sense in 3-Space (local beables) and it can derive Born Rule.
However you wouldn't call this many worlds, it's not even inspired by Everett.
Hell deBroglie was thinking about this back in the 1920's...

 

[Hurkyl]

It's not up to the skeptic.

You're not expressing skepticism of MWI -- you're denying MWI. A denial is a claim which has a burden of proof.

Like Demystifier just said, SURE you can introduce ADDITIONAL POSTULATES to get Born Rule

MWI doesn't need to add postulates -- anything "core" MWI postulates were already there in ordinary quantum mechanics.

Besides, I've always found it baffling that people would criticize any interpretation for assuming "the probabilities computed by the mathematics match the probabilities observed by experiment".

 

[Fyzix]

You're not expressing skepticism of MWI -- you're denying MWI. A denial is a claim which has a burden of proof.

Which I've provided in sources...

MWI doesn't need to add postulates -- anything "core" MWI postulates were already there in ordinary quantum mechanics.

Besides, I've always found it baffling that people would criticize any interpretation for assuming "the probabilities computed by the mathematics match the probabilities observed by experiment".

read the papers please...

 

[Delta Kilo]

The question is whether incorporation of the Born rule ruins the initial beauty and simplicity of MWI. The answer is that it does!

Could you please elaborate on that? How does it ruin the beauty?

I agree that it is difficult to introduce Born rule in MWI without circularity. But there is a reason or it. Typically, in order to derive Born rule one would have to set up some sort of toy model, show that it works (eg. from symmetry arguments), then generalize to more complicated scenarios. But to do that in MWI one needs Born rule to set the stage (to pin down the slice of reality if you wish). In other words, one needs macroscopic objects (pointer states, observers) to operate on but these objects are emergent, not given. It's a bootstrap problem.

DK

 

[IllyaKuryakin]

I've never been able to accept the concept that the entire universe splits in two every time a leaf falls to the ground face up or face down. Even if the math can be made to work, the concept is silly. Considering the number of particles that have made quantum choices over the history of our universe, it would take a number of universes much greater than the number of particles in our universe, and of course, every time a leaf falls to the ground in one of those, it splits in two as well, ad infinitum. Occams razor slices this to ribbons. This reminds me of the solution to string theory, if you suppose an infinite number of universes; one will surely fit the theory. Since we can never observe anything outside our universe, this can never be confirmed by experiment. Surely if you grant infinite degrees of freedom, just about anything can work out. But nothing, other than crackpot theories, ever has infinite degrees of freedom. It falls into the category of religion in my opinion.

 

[stefanbanev]

I've never been able to accept the concept that the entire universe splits in two every time a leaf falls to the ground face up or face down.

Well, it was a bad popularization of MWI. "Quantum universe" remains in its superposition state and it is not the universe you perceive as such. The branching describes the evolution of observer in this quantum realm, actually the picture is more complicated since observer as you perceive it is just one of slices of this realm.

 

[Fyzix]

Stefan: ready to defend your view ?
Or are you just going to continue to "ignore" these technical difficulties and keep the faith?

 

[Delta Kilo]

I've never been able to accept the concept that the entire universe splits in two every time a leaf falls to the ground face up or face down. Even if the math can be made to work, the concept is silly. Considering the number of particles that have made quantum choices over the history of our universe, it would take a number of universes much greater than the number of particles in our universe, and of course, every time a leaf falls to the ground in one of those, it splits in two as well, ad infinitum.

This picture is misleading and that's why I don't like terms "branching" and "splitting". Instead, picture the multiverse as a turbulent flow. Close trajectories represent similar worlds. A single macroscopic state is represented by a compact parcel of gas or water where differences between trajectories are so small that they are unobservable on macroscopic level. As the parcel goes with the flow it gets split into smaller parcels which go different ways, and again and again. If the observer had perfect memory of past events then the size of it parcel (or the proportion of the flow cross-section) would shrink every time the observer learns new fact, in the limit shrinking into a single point which represent the exact configuration of the world with no fuzziness at all. However, real observers are not perfect and tend to forget things. Once a piece of information is forgotten, the corresponding "worlds"/"branches" join back together. However from the global point of view its just a flow of water under the bridge.

And please don't take this analogy too literally. In my opinion it is better than the everbranching tree of life but still it will only carry you that far before it breaks down.

Another way of looking at it: first consider Bohm model. In there you have your wavefunction which evolves all by itself and then you have a single point that dances around in configuration space, representing single definite solution. But in order to make this single point move, the wavefunction has to be computed everywhere. So you get all the motion of other points basically for free. So instead of a lone point moving around, you get a field that flows.

Occams razor slices this to ribbons. ... Surely if you grant infinite degrees of freedom, just about anything can work out.

But this is exactly the issue! Hilbert space is just too darn big! You get all these degrees of freedom out of the box whether you like it or not. Just check how much memory you need to simulate 32 bits. Now do that for 32 qubits. Feel the difference? What do you do with all these numbers, where do you stick them all? (officers! silence!).

 

[Varon]

I will take a look at the paper later today.
I've actually discussed MWI with the author through mail just some weeks ago.
He objects to MWI due to the probability problem and the fact that there is no way to say where a world starts/ends, which is really problematic.
However I'm not sure if his Many minds interpretation is the same as others, usually they require dualism of the mind.
Which I personally reject.
There is some brief information of it on wikipedia, if you google his name + many minds he also got a FAQ

Fyzix, pls give comment on the Michael Lockwood paper which is not based on dualism. Which part of it don't you agree? It is supposed to explain the born rule.

 

[Delta Kilo]

Adrian Kent
http://arxiv.org/PS_cache/arxiv/pdf/0905/0905.0624v2.pdf

The author considers several attempts at deriving Born rule in MWI by counting branches. The branching is taken as granted, no attempt is made to look at the branching process in details. No wonder it does not work. As I said the probabilities have to come out as a ratio of microstates per indistinguishable macrostates during conversion from quantum to classical. The ppaper is nowhere near that.

Jacques Mallah
http://onqm.blogspot.com/2009/09/dec...es-to-mwi.html
http://arxiv.org/abs/0808.2415

Same as above. I may like MWI and at the same time I don't have to like decision-theoretic approach.

Meir Hemmo
http://edelstein.huji.ac.il/staff/pi...rs/Paper45.pdf

The problem is to derive Born’s rule from the many worlds theory without assuming directly or indirectly a notion of typicality or a measure over branches.
Our aim is to show that this is impossible.

Cannot argue with that. A sensible measure of branch "thickness" is required. But that by itself does not disprove MWI.

 

[ExcessRed]

I apologize in advance if this has already been said. I only read the first page of this (kind of lengthy) thread, and felt the need to comment on this:

Nono, ofcourse I don't think the experimentalist will see both.

he splits just like whole world including the experiment.
However,

Pre-experiment you got:

1 Experimenter
A experiment with 2 possible outcomes

After the experiment

You got 2 experimenters each observing one of the 2 outcomes.
They repeat the experiment and this occurs again.

However unless MWI ASSUMES that there is something very special about consciousness that makes the universe somehow put the observers in the universe which would then correspond to the correct probability, they will see 50/50.

It's that simple.

Just draw it on a piece of paper and you will understand exactly what I mean (this thought experiment is Putnam's not mine originally).

90% chance that a blue light occurs, 10% chance that a red light occurs.

Unless you know ahead of time which experimenter to choose from the theoretically infinite number of realities, there is a good chance (in our case, 90% x 90% = 81%) you'd randomly grab two blue-lighters - we can't just choose two at random.

In contrast, if you somehow knew ahead of time that the two experimenters you choose are going to express different results, then the argument is about as relevant as saying... look in a bag of marbles (90% blue, 10% red), remove one marble of each color. From that new set of two, randomly choose one marble and you would have a 50% chance of picking blue.

In essence... wouldn't a single event with a precisely 90% chance of success result in 10 branched realities - 9 successes and 1 failure - thus preserving the global probability?

EDIT: Moved my edit to a new post, since someone had replied by the time I edited. Sorry!

 

[Fyzix]

The author considers several attempts at deriving Born rule in MWI by counting branches. The branching is taken as granted, no attempt is made to look at the branching process in details. No wonder it does not work. As I said the probabilities have to come out as a ratio of microstates per indistinguishable macrostates during conversion from quantum to classical. The ppaper is nowhere near that.

Same as above. I may like MWI and at the same time I don't have to like decision-theoretic approach.

could you give us a paper showing the technical details of this approach you are talking about?

 

[Fyzix]

I apologize in advance if this has already been said. I only read the first page of this (kind of lengthy) thread, and felt the need to comment on this:



90% chance that a blue light occurs, 10% chance that a red light occurs.

Unless you know ahead of time which experimenter to choose from the theoretically infinite number of realities, there is a good chance (in our case, 90% x 90% = 81%) you'd randomly grab two blue-lighters - we can't just choose two at random.

In contrast, if you somehow knew ahead of time that the two experimenters you choose are going to express different results, then the argument is about as relevant as saying... look in a bag of marbles (90% blue, 10% red), remove one marble of each color. From that new set of two, randomly choose one marble and you would have a 50% chance of picking blue.

In essence... wouldn't a single event with a precisely 90% chance of success result in 10 branched realities - 9 successes and 1 failure - thus preserving the global probability?

No and that's the whole point.

At each single event, the universe (if we look at it from MWI) would branch into 2 realities.
1 where the red light goes on
1 where the blue light goes on

This would happen each time and would always add up to 50/50.
There is no escaping it.

Take a piece of paper and draw a line, now pretend that you conduct a experiment and that it branches into blue on one side, red on the other, now branch a blue and red from each of those again and continue doing it.
How the hell are you going to say that there is a 10 vs 90% chance of one of them being observed if the experimenter pre-experiment will end up with BOTH branches and continue to do so at each experiment?

 

[ExcessRed]

No and that's the whole point.

At each single event, the universe (if we look at it from MWI) would branch into 2 realities.

I don't think the idea is that a branch occurs when a human eye observes it - that would be silly. An "observation" occurs any time a particle interacts with another particle in a way that requires it to have a definite quantum state.

Perhaps a more kosher way of describing this phenomenon would be that non-binary probabilities are the result of n binary subdivisions wherein 2 to the n-th worlds would be created.

If an experiment has success rate of approximately 90%, it might result in 1,048,576 different realities. 943,719 of these realities we would observe the result of as "blue light", and the other 104,857 we observe as "red light" - even though the observations of one blue light world and another blue light world are functionally the same, there are subtle differences below the observed level. In this case there were 20 subdivisions resulting in 2^20 worlds.

EDIT: Basically, this:

As I said the probabilities have to come out as a ratio of microstates per indistinguishable macrostates during conversion from quantum to classical.

 

[Hurkyl]

In essence... wouldn't a single event with a precisely 90% chance of success result in 10 branched realities - 9 successes and 1 failure - thus preserving the global probability?

Taking the the decoherence-based philosophy that we should accept the statistical mixture as being reality rather than assuming collapse, the math tells us this is a weighted mixture of two states, with the weights being 90% and 10%.

While we could invent hidden degrees of freedom for the sole purpose of splitting the mixture into a uniform distribution with roughly 90% of the micro-states containing a blue event and the remaining 10% containing a red event, I think it's a bad idea for two reasons:

1. This moves us away from the fundamental idea of MWI -- to understand wave-functions undergoing unitary evolution. (This is the most important of the two reasons)
2. There is absolutely nothing to suggest probability is quantized, and to the best of my knowledge this approach doesn't solve any real problems.

(I don't count an attack on MWI by person who steadfastly refuses to even conceive the idea of a weighted mixture a "real problem")

 

[Lost in Space]

If the MWI theory is wrong, does this mean that consciousness does collapse wave functions? If so, how could the universal wave function collapse at the Big Bang if there were no observers around to collapse it?

 

[Hurkyl]

If the MWI theory is wrong, does this mean that consciousness does collapse wave functions?

If MWI is wrong, it means one of:

• Some aspect of quantum mechanics is wrong
• Experimentally observed statistics in subsystems are not the result of unitary evolution in larger systems
• There are other components to physical state beyond the wave-function which contribute to observations
• Some other mode of failure I didn't think of

The difference between physical state collapsing to a single outcome versus remaining a statistical mixture is not a physical difference -- no observation can distinguish between the two possibilities.

Consciousness has pretty much nothing to do with any of the above. Anything that talks about consciousness is a separate thing layered on top of the basics.

 

[rkastner]

If MWI is wrong (which I think it is), that does not imply that "some aspect of quantum mechanics is wrong". It simply implies that our interpretational approach has been limited by an unrecognized artificial barrier: we haven't gotten outside the nonrelativistic box. Keep in mind that nonrelativistic qm cannot be the whole story anyway. A big problem for most of the 'mainstream' interpretations is that they cannot define what an 'observation' is in the sense of ExcessRed's statement that "An "observation" [i.e., measurement] occurs any time a particle interacts with another particle in a way that requires it to have a definite quantum state." Well, what is this 'way' without resorting to an appeal to 'consciousness,' or an implicit human choice of what to observe as in decoherenc-based approaches? It is explained unambiguously in PTI with reference to the *relativistic* domain: a 'measurement' occurs when an annihilation operator acts on a quantum state -- i.e., when a quantum state is destroyed. The point is that the measurement problem cannot be solved within the nonrelativistic limit -- you have to get outside the nonrelativistic box to solve it. This will all be explained in detail in my forthcoming book on PTI.

 

[Fyzix]

If MWI is wrong, it means one of:

• Some aspect of quantum mechanics is wrong
• Experimentally observed statistics in subsystems are not the result of unitary evolution in larger systems
• There are other components to physical state beyond the wave-function which contribute to observations
• Some other mode of failure I didn't think of

Hurkyl you know this is ********, so why do you keep saying it?

Even wavefunction taken as all there is doesn't yield MWI, so, yeah...

Also what were your solution to the probability problem? The failed Desicion Theoretic approach?

 

[Hurkyl]

Even wavefunction taken as all there is doesn't yield MWI, so, yeah...

Of course not; there is also the hypothesis that the wavefunction only evolves according to Schrödinger's equation.^*, as opposed to the combination of Schrödinger's equation and collapse.

I suppose MWI doesn't become interesting, though, unless you consider the additional hypothesis that quantum mechanics remains valid on larger scales -- e.g. that the decoherence seen in measurement is a result of quantum thermodynamics.


If you want to keep insisting that MWI is equivalent to trying to insist that all probability distributions are uniform, that's your problem not mine.


* or whatever the appropriate notion of state and unitary evolution is needed for the particular variant of QM being considered

 

[StevieTNZ]

"Quantum universe" remains in its superposition state and it is not the universe you perceive as such.

Yup!

 

[Lost in Space]

I just wonder if those who consider MWI to be wrong and that consciousness plays an integral part in determining a result, whether a belief or expectational bias on behalf of an observer can also have a bearing on the probabilistic outcome of a collapse in wave function?

 

[Fyzix]

I just wonder if those who consider MWI to be wrong and that consciousness plays an integral part in determining a result, whether a belief or expectational bias on behalf of an observer can also have a bearing on the probabilistic outcome of a collapse in wave function?

Very very few of those who reject mwi (majority of physicists) believe in such nonsense that consciousness has ANYTHING to do with anything.

 

[Varon]

If MWI is wrong, it means one of:

• Some aspect of quantum mechanics is wrong
• Experimentally observed statistics in subsystems are not the result of unitary evolution in larger systems
• There are other components to physical state beyond the wave-function which contribute to observations
• Some other mode of failure I didn't think of

The difference between physical state collapsing to a single outcome versus remaining a statistical mixture is not a physical difference -- no observation can distinguish between the two possibilities.

Consciousness has pretty much nothing to do with any of the above. Anything that talks about consciousness is a separate thing layered on top of the basics.

Hurkyl, from 1 to 10, how true do you think is MWI? Is it 10? or 5?

 

[Fyzix]

Hurkyl doesn't even understand basic QM.
He thinks that if MWI is wrong it means QM is wrong...
He always claimed there are no Born Rule problem with MWI, yet he hasn't given any source to this claim.

 

[Lost in Space]

Very very few of those who reject mwi (majority of physicists) believe in such nonsense that consciousness has ANYTHING to do with anything.

Well, perhaps with the exception of the observation itself?

 

[Hurkyl]

Hurkyl, from 1 to 10, how true do you think is MWI? Is it 10? or 5?

You mean as compared to other interpretations? I don't think the question is meaningful.

Are you referring instead to the hypothesis that quantum thermodynamics can accurately treat larger systems? My impression is that physicists are fairly confident in that.

Are you referring instead to the hypothesis that quantum thermodynamics can accurately treat truly macroscopic systems in which gravity is negligible (such as a cat in a poison booth controlled by the decay of an atom)? It seems reasonable to be optimistic, although it seems a lot of theoretical developments have to be made to work out the mathematics. More importantly, all of the reasons I have ever heard to be severely pessimistic about it have been discredited.


The real question about interpretation is utility. As quantum mechanics is used to treat larger and larger systems, I don't see how one could possibly avoid thinking of things in a MWI-like fashion without deliberately avoiding to try and form any mental concept of the behavior a wave-function evolving unitarily.

 

[stefanbanev]

Hurkyl, from 1 to 10, how true do you think is MWI? Is it 10? or 5?

Excuse me for interfering, but from my perspective it's the key question to address MWI appeal:

First of all, in my opinion the theory cannot be "correct" or "right"; it may possesses more or less predictive power, from this perspective MWI does no prove definitively its superiority (so far). The Wrong/Right paradigm can be only applied if we assume that there is the Grand Design which science can reveal - it's apparently a religious assumption. The major appeal of MWI, in my opinion - MWI just makes more sense for some who is prone to look for more general explanations.

Stefan

 

[agentredlum]

Suppose MWI is correct, does that mean ALL possible realities exist physically? If the answer is yes then that means there exists a physical reality in which there is a law that prohibits the existence of ALL OTHER REALITIES! So you see MWI contradicts ITSELF. Guess what, we exist in THAT reality (the one that prohibits ALL OTHERS) MWI is incorrect. To give you an example of how absurd MWI is...are you seriously suggesting that there exists a physical reality in which i have 23 heads, 7 arms, 19 legs, a spinal column made of fractals, and 53 oragutans are having sex with the girl next door 24/7? The ABSURDITIES i can list here are finite, but the absurdities made possible by MWI must NECESSARILLY be infinite (play on a quote by Wittgenstien)

And now for The Granfather Paradox. This one is AGAINST time travel. Suppose you travel back in time and kill your grandfather BEFORE your father is CONCIEVED. Where does that leave you? If your father was never here how can you be here able to go back etc...

Now my idea...Time travel should be possible so that your grandfather can travel FORWARD in time and kill you just BEFORE you go back in time to kill him. This one is in favor of time travel. I call this MY GRANDFATHERS REVENGE THEORY. This theory is in reply to Steven Hawkins theory that the past must be made safe for historians. My theory makes the future safe for grandfathers.