Many Worlds Interpretation and act of measuring

[bhobba]

Again, almost all physicists are frequentists. From a mathematucal perspective Bayesianism is much more common than for the typical physicist. QBism will be much more diificult for physicists to accept than the MWI.

Yes. If you ask them they will probably say Copenhagen, but if you pin them down you will find its really Ballentine's Ensemble because they don't really understand Copenhagen has a subjective view of probabilities. That's because beginning texts like Griffiths don't explain the nuances of it and only some graduate texts like Ballentine delve deeper.

Thanks
Bill

 

[Rajkovic]

Its weird, and I don't ascribe to it - but it can't be debunked.

If it has aspects that appeal you may like Consistent Histories - its sometimes described as MW without the many worlds. I probably have given a link to it before, but here it is again:
http://quantum.phys.cmu.edu/CQT/index.html

Thanks
Bill

It can't be debunked to a certain point until some new experiment put an end to all of this (hope that day comes soon)

which theory fits the best to you Bill? do u think these doubts will be solved soon?

I will read it, thx

 

[atyy]

Yes. If you ask them they will probably say Copenhagen, but if you pin them down you will find its really Ballentine's Ensemble because they don't really understand Copenhagen has a subjective view of probabilities. That's because beginning texts like Griffiths don't explain the nuances of it and only some graduate texts like Ballentine delve deeper.

Copenhagen simply has many flavours, and a frequentist view of probability is also a flavour of Copenhagen. Ballentine's Ensemble is not new, and he should not get credit for work that is wrong where it is novel, and simply an orthodox flavour of Copenhagen where it is correct.

I should also note that Ballentine's Ensemble (1998, p46) does have a subjective notion of the state because he uses a conceptual ensemble that does not exist, except in his mind: "Thus, although the primary definition of a state is the abstract set of probabilities for the various observables, it is also possible to associate a state with an ensemble of similarly prepared systems. However, it is important to remember that this ensemble is the conceptual infinite set of all such systems that may potentially result from the state preparation procedure, and not a concrete set of systems that coexist in space."

 

[bhobba]

Copenhagen simply has many flavours, and a frequentist view of probability is also a flavour of Copenhagen. Ballentine's Ensemble is not new, and he should not get credit for work that is wrong where it is novel, and simply an orthodox flavour of Copenhagen where it is correct.

Yes there are many variants of Copenhagen, but when I mention it I mean something along the lines of the following:
http://motls.blogspot.com.au/2011/05/copenhagen-interpretation-of-quantum.htm

Note point 1.
A system is completely described by a wave function ψ, representing an observer's subjective knowledge of the system.

Frequentists have a differing view.

Thanks
Bill

 

[atyy]

Yes there are many variants of Copenhagen, but when I mention it I mean something along the lines of the following:
http://motls.blogspot.com.au/2011/05/copenhagen-interpretation-of-quantum.htm

Note point 1.
A system is completely described by a wave function ψ, representing an observer's subjective knowledge of the system.

Frequentists have a differing view.

Thanks
Bill

Even if one used a purely subjective definition of the state, that doesn't mean that one is using a Bayesian interpretation of probability. Frequentists and Bayesians differ about the interpretation of Kolmogorov's axioms, which has no concept of the quantum state. So even if one considers the state "subjective", the probability obtained from the Born rule can be Frequentist.

 

[julcab12]

... I'm always getting reminded to be very critical, skeptic and careful. MWI is no exemption. So i'll leave a note to these guys.

 

[atyy]

... I'm always getting reminded to be very critical, skeptic and careful. MWI is no exemption. So i'll leave a note to these guys.

Carroll does not claim MWI is without problems. He gives a list of serious issues in the MWI approach about which consensus has not been reached in http://www.preposterousuniverse.com...ion-of-quantum-mechanics-is-probably-correct/.

"The fierce austerity of EQM is attractive, but we still need to verify that its predictions map on to our empirical data. This raises questions that live squarely at the physics/philosophy boundary. Why does the quantum state branch into certain kinds of worlds (e.g., ones where cats are awake or ones where cats are asleep) and not others (where cats are in superpositions of both)? Why are the probabilities that we actually observe given by the Born Rule, which states that the probability equals the wave function squared? In what sense are there probabilities at all, if the theory is completely deterministic? These are the serious issues for EQM ..."

 

[stevendaryl]

There is a slight difference between the problems of MWI and the problems of other interpretations. With MWI, the problem is purely about how to interpret the theory. The theory itself is unambiguous, it seems to me (just a state vector evolving continuously and unitarily). But in the case of other interpretations of QM, the problems are about additions to (or exceptions to) unitary evolution. Bohm's theory has additional "elements of reality", namely definite particle locations at all times. Von Neumann's "collapse" interpretation has an additional type of state change--collapse following an observation.

 

[rootone]

In the MWI the different worlds 'exist' in an abstract mathematical space, Hilbert Space. So don't take this visualisation too literally.

That is where the problem lies.
Not with the interpretation as a concept, but with the presentation of it as being literal, by some folks.
I'm not having a go at Carrol in particular, but I do think people like Greene, Tegmark, and Kaku could be deemed 'guilty' of encouraging misunderstanding among the general public.
Some of whom then go on to quote them on mystical woo type sites.

 

[atyy]

There is a slight difference between the problems of MWI and the problems of other interpretations. With MWI, the problem is purely about how to interpret the theory. The theory itself is unambiguous, it seems to me (just a state vector evolving continuously and unitarily). But in the case of other interpretations of QM, the problems are about additions to (or exceptions to) unitary evolution. Bohm's theory has additional "elements of reality", namely definite particle locations at all times. Von Neumann's "collapse" interpretation has an additional type of state change--collapse following an observation.

I'm not sure it is so different. From what I understand, Carroll's concern is whether MWI matches experiment.

 

[stevendaryl]

Even if one used a purely subjective definition of the state, that doesn't mean that one is using a Bayesian interpretation of probability. Frequentists and Bayesians differ about the interpretation of Kolmogorov's axioms, which has no concept of the quantum state. So even if one considers the state "subjective", the probability obtained from the Born rule can be Frequentist.

I'm always bothered by frequentist versus bayesian arguments. To me, the differences just amount to different ways of talking about the same things. There is really no substantive difference.

In my opinion, frequentism is just a collection of rules-of-thumb for dealing with probability. It isn't an interpretation of probability. To say that probability means frequency is nearly vacuous. Any real experiment is only conducted a finite number of times, so the concept of relative frequency in the limit of infinitely many trials seems irrelevant. Now, what you can do, with a finite number of trials, is to compute the probability that the observed relative frequency is different from the theoretically predicted probability, and you can argue that for a large enough number of flips of an unbiased coin, the difference between the relative frequency of heads and 1/2 will be vanishingly small, with a probability close to 1. But that latter part--"with a probability close to 1"--is using some theoretical notion of probability that is NOT given a frequentist meaning.

In a real experiment, you never do anything more than a finite number of times, so what happens in an infinite number of trials seems irrelevant. Of course, frequentists have ways of dealing with that situation, involving measures of the level of significance. But as I said, those techniques seem like ad hoc rules of thumb---they aren't actually justified by any frequentist understanding of probability.

 

[julcab12]

Carroll does not claim MWI is without problems. He gives a list of serious issues in the MWI approach about which consensus has not been reached in http://www.preposterousuniverse.com...ion-of-quantum-mechanics-is-probably-correct/.
.."

No objections. But i was wondering since he strongly stated. " The potential for multiple worlds is always there in the quantum state, whether you like it or not." Would we be able to get a different viewpoint on that manner? "Yes! It appears we are detecting seemingly series of different state + classical time. We should make a basic assumption from it -- Multiples". BUt wait! How about Time? Would it have any effect on that scale -- Like a gravitational time dilation/ distortion or/ lensing that causes things to appear decohered. Think of like visual slices of possible outcomes. It should have decoherance of time not just a decoherence of state.

 

[stevendaryl]

To me, there are legitimate criticisms of MWI, having to do with the issues of how to interpret probability for a deterministic system, and how to count worlds, and so forth. But the most common complaint about MWI is about the "many worlds" aspect--the idea that when something happens due to random chance--Schroedinger's cat either dies or doesn't, depending on a random quantum event--there is a world in which the cat lives and a world in which the cat dies. People complain that this is "multiplying entities" in violation of occam's razor. But it's hard for me to see how to get just one world without ADDING something to the quantum formalism (such as a special status for "measurements")

 

[atyy]

In my opinion, frequentism is just a collection of rules-of-thumb for dealing with probability. It isn't an interpretation of probability. To say that probability means frequency is nearly vacuous. Any real experiment is only conducted a finite number of times, so the concept of relative frequency in the limit of infinitely many trials seems irrelevant. Now, what you can do, with a finite number of trials, is to compute the probability that the observed relative frequency is different from the theoretically predicted probability, and you can argue that for a large enough number of flips of an unbiased coin, the difference between the relative frequency of heads and 1/2 will be vanishingly small, with a probability close to 1. But that latter part--"with a probability close to 1"--is using some theoretical notion of probability that is NOT given a frequentist meaning.

Is the latter part in the law of large numbers really not given a Frequentist meaning? It would seem consistent to apply the Frequentist meaning there too. It would be nearly vacuous as you say, or at least circular. But I think it is as vacuous as physics - what is a charge? It is a thing on which an electric field exerts a force. What is an electric field? It is a thing which exerts a force on a charge. It is saved because we somehow have conventions as to what operations in real life correspond to certain dynamics of the mathematical objects.

 

[stevendaryl]

Is the latter part in the law of large numbers really not given a Frequentist meaning? It would seem consistent to apply the Frequentist meaning there too. It would be nearly vacuous as you say, or at least circular. But I think it is as vacuous as physics - what is a charge? It is a thing on which an electric field exerts a force. What is an electric field? It is a thing which exerts a force on a charge. It is saved because we somehow have conventions as to what operations in real life correspond to certain dynamics of the mathematical objects.

The law of large numbers says (in a quantitative way) that the probability that relative frequency for many repeated events differs significantly from the probability of a single event goes to zero as the number of repetitions goes to infinity. So using the law of large numbers to justify a frequentist interpretation of probability seems circular, as you say. I don't have any problem with circular definitions, but if you're using a circular definition, then you're not really defining the concepts, you're just axiomatizing them. So you're basically treating "probability" as an undefined term. If that's what you're doing, then it's not really frequentist, it's just abstract probability, which doesn't distinguish between frequentism and Bayesianism.

The only real source of disagreement between frequentists and Bayesians, it seems to me, is over whether it is meaningful to talk about probability of something that only happens once (or a small number of times). But the blunt fact is that EVERYTHING only happens once or a small number of times (for some definition of "small"). We always have a limited collection of samples. So the frequentist disdain for talking about probabilities on a small sample space really to me makes it impossible to do anything with probabilities (except through ad hoc means, such as cut offs and levels of significance, etc., which are not justified by the frequentist interpretation, as I already said).

 

[atyy]

The law of large numbers says (in a quantitative way) that the probability that relative frequency for many repeated events differs significantly from the probability of a single event goes to zero as the number of repetitions goes to infinity. So using the law of large numbers to justify a frequentist interpretation of probability seems circular, as you say. I don't have any problem with circular definitions, but if you're using a circular definition, then you're not really defining the concepts, you're just axiomatizing them. So you're basically treating "probability" as an undefined term. If that's what you're doing, then it's not really frequentist, it's just abstract probability, which doesn't distinguish between frequentism and Bayesianism.

The reason the Frequentist interpretation does more than just rename probability is that it is an operational definition, ie. it's a way of converting the abstract Kolmogorov axioms into statements about real operations.

The only real source of disagreement between frequentists and Bayesians, it seems to me, is over whether it is meaningful to talk about probability of something that only happens once (or a small number of times). But the blunt fact is that EVERYTHING only happens once or a small number of times (for some definition of "small"). We always have a limited collection of samples. So the frequentist disdain for talking about probabilities on a small sample space really to me makes it impossible to do anything with probabilities (except through ad hoc means, such as cut offs and levels of significance, etc., which are not justified by the frequentist interpretation, as I already said).

The cut-offs and levels of significance are additional ad hoc criteria, but the Bayesian interpretation also has ad hoc things, like how one chooses the prior. So it is largely a matter of taste where one puts the ad hoc stuff. The real difference is that Frequentists at the meta-level don't mind being incoherent, whereas Bayesians believe in coherence. So in practical situations, one can apply Bayesian statistics while violating at at the philosophical level by being incoherent (ie. by saying that Bayesian statistics are just a tool).

*When I say Bayesian, I actually mean de Finetti's beautiful theory of coherence. I don't accept Jaynes's ugly objective Bayes philosophy.

 

[Quantumental]

As I have pointed out to you before the issue has yet to be resolved one way or the other. Certain key mathematical theorems are lacking.

But then it is very misleading to suggest that MWI is "beauty incarnate" and "mathematical elegance". Copenhagen is also very simple, as long as you don't worry about explaining collapse. So it Bohm, minus the non-locality etc. etc. I feel that MWI is getting a lot of unearned praise simply for existing as an idea that hasn't been fleshed out, or rather every time it's attempted to be fleshed out it's fatal flaws shine through.

At this moment in time, due to it's severe flaws, we have absolutely no reason to assume that these will just "work themselves out" anymore than we have a good reason to think that Bohm will somehow do away with it's problems with relativity.

I have seen papers where for simple models it is shown the results do NOT depend on the factorisation. They however need to be extended, and that hasn't been done yet.

Could you link?

 

[atyy]

At this moment in time, due to it's severe flaws, we have absolutely no reason to assume that these will just "work themselves out" anymore than we have a good reason to think that Bohm will somehow do away with it's problems with relativity.

At present, the standard model does not need exact relativity. For example, QED needs an high energy cutoff. Some advocate a lattice regularization, which violates special relativity, eg. Capitani, http://arxiv.org/abs/hep-lat/0211036: "In principle all known perturbative results of continuum QED and QCD can also be reproduced using a lattice regularization instead of the more popular ones." So if one accepts relativity as only approximate, Bohmian Mechanics does not necessarily have problems with relativity. As I understand, the main problem with Bohmian Mechanics is chiral fermions, but there is no theorem ruling that out yet, and plenty of recent work on it: eg. http://arxiv.org/abs/0709.3658, http://arxiv.org/abs/0912.2560.

 

[atyy]

At this moment in time, due to it's severe flaws, we have absolutely no reason to assume that these will just "work themselves out" anymore than we have a good reason to think that Bohm will somehow do away with it's problems with relativity.

I have a question. Some MWI advocates described Bohmian Mechanics as "Everett in denial". If this is the case, and given that there is high confidence with Bohmian Mechanics at least for non-relativistic physics, couldn't Bohmian Mechanics solve the problems of MWI, eg. simply asserting that all BM worlds are real and exist?

http://arxiv.org/abs/0712.0149 "A potentially more serious flaw arises from the so-called “Everett-in-denial” objection to realism (Deutsch 1996; Zeh 1999; Brown and Wallace 2005). ... Advocates of the Everett interpretation claim that, (given functionalism) the decoherence-defined quasiclassical histories in the unitarily evolving physically real wavefunction describe — are — a multiplicity of almost-identical quasiclassical worlds; if that same unitarily-evolving physically real wavefunction is present in DBB (or any other hidden-variable theory) then so is that multiplicity of physically real worlds, and all the hidden variables do is point superfluously at one of them."

 

[Quantumental]

I have a question. Some MWI advocates described Bohmian Mechanics as "Everett in denial". If this is the case, and given that there is high confidence with Bohmian Mechanics at least for non-relativistic physics, couldn't Bohmian Mechanics solve the problems of MWI, eg. simply asserting that all BM worlds are real and exist?

Actually this is very similar to the ideas that have been presented by several different researchers in the last years, all independent of each other.
Charles Sebens, who did the paper with Sean Carroll on their Born Rule in MWI proposal posted his main theory of many interacting worlds on Seans blog late last year: http://www.preposterousuniverse.com...racting-worlds-approach-to-quantum-mechanics/

Then almost at the same time another team of quantum foundation researchers posted a very similar proposal as can be seen here:
Quantum Phenomena Modeled by Interactions between Many Classical Worlds
http://journals.aps.org/prx/abstract/10.1103/PhysRevX.4.041013

Finally Kim Joris Boström's Combining Bohm and Everett: Axiomatics for a Standalone Quantum Mechanics
http://arxiv.org/abs/1208.5632 Is touching on the same.

 

[Quantumental]

Btw here is Kim Boström's most recent paper (just a few days old):
Quantum mechanics as a deterministic theory of a continuum of worlds
http://arxiv.org/pdf/1410.5653v4.pdf

 

[bhobba]

But then it is very misleading to suggest that MWI is "beauty incarnate" and "mathematical elegance".

Precisely why? For example what exactly is wrong with the decision theory analysis of Wallace that proves Born? I have gone through it in Wallace's book - The Emergent Multiverse and on page 475 has the Noncontextuality Theorem that he believes follows from any rational 'reward' function defined on the states. If its non-contextual then Gleason applies. Its a complex mathematical argument but I would like to know exactly what its flaw is. On page 189 he discusses why other possible ways of defining a rational 'probability' of outcomes is wrong. I believe its somewhat 'circular' in that if you go through them basically they are saying you can't reasonably define a decision type function on the space unless its basis independent - which is the essence of non-contextuality. But it's a rigorous elaboration of the idea.

Could you link?

Not off the top of my head - but this topic has been discussed many times before - I likely have posted it previously.

Thanks
Bill

 

[julcab12]

... Just a grain of doubt. Multiplicity is a predicted/recent phenomenon in cosmology. It is cause by the distortion of field or by gravitational lensing. A 'single object' will appear 'multiple' due to distortions and 'time' delay. Are there any similar effect in QM's vacuum or am i misunderstanding something on the vacuum effect?

http://newscenter.berkeley.edu/2015/03/05/distant-supernova-split-four-ways-by-gravitational-lens/

 

[rootone]

Gravitational lensing was predicted by GR, although I believe Einstein thought it would not be detectable.
It since turned out to be that it's very much detectable and highly useful.
The observation referred to in the link is a spectacular example, but it's not the first to be found.

I have not heard of any similar phenomena being predicted as a consequence of QM.

 

[bhobba]

Are there any similar effect in QM's vacuum or am i misunderstanding something on the vacuum effect?

What do you mean by he 'vacuum effect'?

The QFT vacuum is simply an artefact of the perturbation methods used and isn't really a teeming with particles popping in and out of existence as popularisations tell you.

A word of warning - just about anything you read outside a QFT textbook on QFT is likely wrong. Or to be more precise is a lot more nuanced than a literal reading.

Thanks
Bill

 

[atyy]

Precisely why? For example what exactly is wrong with the decision theory analysis of Wallace that proves Born? I have gone through it in Wallace's book - The Emergent Multiverse and on page 475 has the Noncontextuality Theorem that he believes follows from any rational 'reward' function defined on the states. If its non-contextual then Gleason applies. Its a complex mathematical argument but I would like to know exactly what its flaw is. On page 189 he discusses why other possible ways of defining a rational 'probability' of outcomes is wrong. I believe its somewhat 'circular' in that if you go through them basically they are saying you can't reasonably define a decision type function on the space unless its basis independent - which is the essence of non-contextuality. But it's a rigorous elaboration of the idea.

Wallace divides the probability problem in MWI into the "incoherence" and the "quantitative" problems. The derivation of the Born rule falls under the quantitative problem. Even if one were to grant the quantitative problem solved, it isn't clear that the incoherence problem is solved. Wallace favours an approach he calls "subjective uncertainty", but he indicates it is controversial (section 4.5 of http://arxiv.org/abs/0712.0149).

 

[bhobba]

but he indicates it is controversial

No argument that MW is controversial. No argument there is a factorisation problem that hasn't been settled. Other issues have been pointed out as well such as the assumption the environment is random.

My concern is precisely they are controversial. There seems to be this idea that because of that it must be wrong - it can't be beautiful mathematically (which it is) etc. They are controversial precisely because the issues haven't been settled and research is ongoing.

It seems to be rehashed over and over again eg:
https://www.physicsforums.com/threads/new-many-worlds-objections-right-wrong.781797/

As Kith said:
'Similar problems are also present in classical mechanics. The inital entropy problem is the question why was the initial entropy of the universe so low and the observer energy problem comes into play if we want measurements to be truly irreversible processes. For this, we need dissipation which contradicts the closed system assumption.'

I even recall one thread where it was shown the solution to a problem in classical mechanics crucially depended on a certain decomposition.

These are issues not confined to MW but crop up all over the place. I think far too much weight is given to them personally - they are issues - but not the big deal some want to make out they are.

Thanks
Bill

 

[atyy]

These are issues not confined to MW but crop up all over the place. I think far too much weight is given to them personally - they are issues - but not the big deal some want to make out they are.

What's your take (and anyone else who wants to comment) on Zurek's comment in http://arxiv.org/abs/1412.5206: "Quantum Darwinism shows why only such redundantly recorded pointer states are accessible to observers|it can account for perception of `quantum jumps'. However, full account of collapse involves 'consciousness', and may have go beyond just mathematics or physics."

I assume he is talking about it within his Existential Interpretation, which I have never understood, but it seems to be some variant of MWI.

 

[julcab12]

What do you mean by he 'vacuum effect'?

The QFT vacuum is simply an artefact of the perturbation methods used and isn't really a teeming with particles popping in and out of existence as popularisations tell you.

A word of warning - just about anything you read outside a QFT textbook on QFT is likely wrong. Or to be more precise is a lot more nuanced than a literal reading.

Thanks
Bill

..? A yes, maybe or no is fine with me. I'm referring to the vacuum where particle is observed in a state of superposition. But anyways i did found some insights; http://arnold-neumaier.at/physfaq/topics/vacfluc.

"... They are not changes in time. Instead, they describe uncertainties about what one gets when one tries to measure something. They describe the fluctuations in the measurement results when one repeats them under identical conditions - not fluctuations in what is measured... Fluctuations have a much better ontological status than virtual particles. Their properties are indeed computable nonperturbatively, hence are properties of the system under study and (unlike virtual particles) not only of the approximation method used. They are properties of the system, whether or not somebody measures it. In this
sense they exist independent of measurement, like a tree exists no matter whether someone looks at it."

...There much be a reason/evaluation on why they're taking this very literal -- Multiples as series of individual system/world = MWI. Non relativistic QM-- time and space are treated differently, with position being an operator and not time while QFT treats space and time on equal footing. GR- cosmology on the other hand has a natural intuitive solution to an OBSERVATION of 'multiplicity' -- distortion on the field causing it to appear multiple despite of being only 1 object. I might be making a weak comparison here but I want to understand also; why it can't happen to the very small?

"Time in QM is a label wrt to which things change.. Time is not a property of Quantum system so it makes no sense to promote it to an operator..Position of a particle however is a particular property of the system so it is an observable & hence an operator.. Time on the other hand is a universal property independent of any system..In fact this the beginning of Quantum Field theory where you demote position to a label to bring time & position on equal footing to get a Lorentz invariant theory."

 

[bhobba]

What's your take (and anyone else who wants to comment) on Zurek's comment in http://arxiv.org/abs/1412.5206: "Quantum Darwinism shows why only such redundantly recorded pointer states are accessible to observers|it can account for perception of `quantum jumps'

I would have to reacquaint myself with the detail of Quantum Darwinism to comment. But overall my take on Quantum Darwinism is it starts at the wrong end of the problem. It assumes states, Schroedinger's equation, unitary evolution etc to derive observations and the Born rule. Without delving into if it succeeds, I prefer the other way of attacking it - starting with observations as the primitive and deriving the other stuff.

However, full account of collapse involves 'consciousness', and may have go beyond just mathematics or physics."

Taking the modern interpretation of collapse as why do we get any outcomes at all we have interpretations like BM that do not require conciousness to explain it so obviously it's an incorrect statement.

Thanks
Bill