Problems with Many Worlds Interpretation

[Hurkyl]

There is still a great deal of misunderstanding around the terms "empiricist" and "rationalist", and it is causing lots of problems in getting what I am saying. Here's what I mean and what I don't mean:

Let me describe another position. (actually, positions on at least two different topics are being combined)

Knowledge is a repository of propositions one has acquired over time and has ascribed levels of confidence to. The role of observation is as a method of inference -- it is a means of producing a piece of knowledge. The role of reason is also as a method of inference -- combining pieces of knowledge together to produce a new piece of knowledge. (Also, to create new pieces of knowledge in the form of tautologies)

Science ascribes particular importance to Bayesian inference (or similar) -- when we obtain a new piece of knowledge X, it adjusts our confidence in other pieces of knowledge Y based on our confidence in the knowledge "Y implies X".

Nature is said to obey laws in the practical sense that we have confidence in knowledge of predictions about what will be seen 'in nature'.

One major role of theory is to provide for the ideas and concepts we use to think about things -- especially[/i] on topics we can't expect to have confidence in our intuition.

Another major role of theory is to provide a means to translate our intuition into forms we can reason with.

 

[Ken G]

And even with very precise measurement, you won't be able to chose between 2 versions of a theory:

A. World = BeautifulFormula()

and

B. World = BeautifulFormula() + s*UglyAddition()

if small value of s makes the value of UglyAddition() beyond measurement.

We can all agree on what attributes we are looking for in theories-- even empiricists. But we don't always agree on how to interpret those theories-- an interpretation is already an "ugly addition" to any theory, but we need them because "shut up and calculate" really just doesn't cut it.

 

[Ken G]

Consistent with decoherence is not the same thing as being decoherence-based. On the topic of the quantum-classical interface:

• A collapse-based interpretation roughly says that quantum states collapse into a pure"eigenstate", which we see in the classical realm as a non-parm state.
• A decoherence-based interpretation roughly says that quantum states decohere into a mixture of "eigenstates", which we see in the classical realm as a parmstate.

• But it is purely your own claim that your approach is "decoherence based." I see no evidence for it-- I think CI is completely "decoherence based" because the entire purpose of an observation, from the CI perspective, is to achieve decoherence. Perception of the outcome comes later, but it is also an essential part of interacting with our reality (imagine the football game where the referee tosses the coin but no one looks at it-- "red team calls heads, it's a mixed state, game cancelled." The empiricist says, if you try to decohere something to gain information about it, and if every time you do that you also get a collapse, then the "decoherence based" interpretation is to conclude that collapse is a natural consequence of perceiving decoherence. But either way, collapse is certainly something separate from decoherence, because it only appears when you add the concept of perception to the concept of decoherence.
  
  I view the decoherence as already having happened by the time we perceive the outcome, and it is only that perception that requires an interpretation, so I don't see any interpretations as having anything in particular to do with decoherence-- including MWI.

  This classification is independent of other features, such as the collapse-based question^* of just information (a wave-function is just a new way to encode ignorance that can't be thought of as a set of possibilities!) versus something real (there's some new physics to discover that violates unitary evolution to produce non-parmness in the limit).

  I'm not sure what you are saying here, but it sounds like something I agree with-- I agree with that first characterization of the wave function, and I find it highly likely that there is indeed some new physics to discover that will reveal unitarity as only nearly unbroken. It's nearly unbroken status will explain why QM works, and breaking it will explain the collapse, but to get there, we need to understand what perception is. Or maybe, by getting there, we will finally figure out what perception is.
  

  Bohm, I think, is a "new physics" interpretation on the decoherence-based side (but one that actually exists rather than being hypothetical), with the interesting feature of retaining the parmness of the quantum state, but couching observations in terms of non-parm Bohmian particles.

  I view Bohm's approach as angels on a different pin. The problem is how to retain determinism (instead of just chucking it as CI does), and the price is those angels-- MWI's angels are those other worlds, which maintain the deterministic evolution of the unitary state (and ultimately saying that closed systems don't change at all, Parmenides would be proud) but are not observed (you claim they are, I say that shows you are not at all an empiricist). Bohm's angels are the unknowable elements of the "initial condition" that determine the outcomes which we, not having access to the unknowable information, see as probabilistic.
  
  The three interpretations are really all a question of what you are distributing over (I'm reminded of Bob Dylan's lyrics: "it may be the devil, or it may be the lord, but you got to serve somebody."). CI distributes over identical trials of the same experiment. MWI distributes over observers in the many worlds. Bohm distributes over unknowable information in the initial condition. Perhaps Bob would put it, "it may be the experiments, or it may be the observers, but you got to distribute over something." Three interpretations, three distributions, all the same physics-- the only thing that changes is the world view they inspire. Maybe we just shouldn't use physics theories to inspire world views in the first place!
  

  Hence "suspect". The motivation for my suspicion is that MWI will be falsified if observations are made that cannot be explained as unitary evolution of a larger system.

  All interpretations of QM will be falsified if QM is falsified, but we'll feel less silly if we didn't build a world view around it first.
  

  I'm thoroughly unconvinced that sharing your prior philosophical biases about "what we measure" is a requirement for being an empiricist or for adopting axioms like "reality is what we measure".
  
  Show me a derivation of "there are no parmstates" from the hypothesis "reality is what we measure".

  Simple: we don't measure parmstates, we measure eigenvalues that associate with eigenstates. You keep saying that doesn't prove there aren't parmstates, but that's not what I said.

 

[Ken G]

One major role of theory is to provide for the ideas and concepts we use to think about things -- especially[/i] on topics we can't expect to have confidence in our intuition.

Another major role of theory is to provide a means to translate our intuition into forms we can reason with.

I agree with every single thing you said in that post, and not a word of it adjudicates between MWI and CI. That is simply because not a word of it is controversial for either a rationalist or an empiricist. The key issue you didn't address is at the crux of the matter: how seriously do you take those concepts as you are using them in the ways you state. Tegmark, as complete a rationalist as can possibly exist, takes them so seriously as to flat out assert that the concepts spawn realities, rather than the other way around (the empiricist view). If you will take MWI as your world view, then Tegmark is your oracle, and how this cannot make crystal clear the difference between rationalism and empiricism I just don't get. I think you are a dyed-in-the-wool rationalist who for some reason hesitates to self-identify as such. It isn't a bad word, though I admit I have been critical-- I have no doubt that Tegmark's contributions to human thoughts about reality will leave a far more lasting mark than any of mine will, even if just as a warning of what can happen if one goes a bit overboard on rationalism!

 

[Hlafordlaes]

Ken G... But either way, collapse is certainly something separate from decoherence, because it only appears when you add the concept of perception to the concept of decoherence.

I view the decoherence as already having happened by the time we perceive the outcome, and it is only that perception that requires an interpretation, so I don't see any interpretations as having anything in particular to do with decoherence-- including MWI.

Hi, Ken G, if you get a chance to digress for a sec can you go on about these two paragraphs a bit? Not sure I get it. Actually, I am sure I do not!

If, as from http://en.wikipedia.org/wiki/Quantum_decoherence" : "Decoherence occurs when a system interacts with its environment in a thermodynamically irreversible way," what more is needed for collapse to be said to also occur? And if as you say decoherence has already happened prior to perception, then how is perception required to explain collapse?

I mean, what's going on here?

 

[Ken G]

Hi, Ken G, if you get a chance to digress for a sec can you go on about these two paragraphs a bit? Not sure I get it. Actually, I am sure I do not!

If, as from http://en.wikipedia.org/wiki/Quantum_decoherence" : "Decoherence occurs when a system interacts with its environment in a thermodynamically irreversible way," what more is needed for collapse to be said to also occur? And if as you say decoherence has already happened prior to perception, then how is perception required to explain collapse?

I mean, what's going on here?

Irreversibility is a weird subject, to be sure. If you watch a movie in which 100 coins, variously heads and tails, are lying on a platform, and the platform jumps up suddenly and sends the coins to flipping, and they all settle on "heads", you are quite likely going to conclude you are watching a movie run backward. If there are 1000 coins, even more certain will you be. But why is this? The actual state you observed, before the "all heads" state, is no more or less likely than all heads, if you distinguish each of the coins individually. So you are watching a movie in which one incredibly unlikely state flips into another equally incredibly unlikely state, so how do you know the movie is being run backward?

All the same, I'll bet if you actually do see that, then indeed the movie will have been run backward. The reason has to do with how difficult it is to get a configuration of all heads in any way but manually, whereas any state of various combinations of heads and tails will serve as the final state of the flips, for making a movie like that. There is something "non-generic" about all heads, stemming from the smallness of the information we need to specify that state (two words!). So the irreversibility of the movie stems entirely from the information content, because there is a connection between the amount of information you need to describe the state, and the number of states belonging to the class that can be described with that much information (the former is the natural log of the latter).

So where does this leave collapse? Well, there's the rub-- the nature of information in quantum mechanics is very much what is being disputed in this thread. In MWI, any closed system is a stationary state (it matters little if it is treated as a mixture of stationary states, they're independently stationary because their relative phase is not defined), so there is no such thing as either reversibility or irreversibility. But that's because the information there is not extractable. In CI, or an empiricist approach, we do have access to information about subsystems with which we interact, so we can define a concept of irreversibility. Decoherence will appear to involve extracting information if we ignore the coherences that persist-- so in that interpretation, "decoherence" means "we no longer care about the coherences", more so than the "coherences are really gone." But in MWI, those coherences are still there-- the decoherence is an artifact of the projection onto the substate, the whole closed system is still stationary and has zero entropy at all times.

Bottom line: the Wiki entry is from the CI standpoint, and there, taking a subsystem in an initially pure state, and decohering it by attaching it to a measuring device, is irreversible-- the resulting mixed state has more entropy than the original pure state, whereas the entropy of the measuring device is not treated separately from what it measures. But perceiving the definite outcome, which is what I mean by "collapse", returns the state to one of zero entropy. The reduction in entropy can happen because the act of perception includes increased entropy in the observer.

 

[Dmitry67]

an interpretation is already an "ugly addition" to any theory, but we need them because "shut up and calculate" really just doesn't cut it.

This is true for any interpretation except MWI. MWI is NULL interpretation, all wordy stuff in it is just an explanation (for humans) that no extra stuff is needed and pure QM math is enough. This is the only interpretation compatible with MUH.

 

[Hurkyl]

But it is purely your own claim that your approach is "decoherence based."

To the best of my knowledge, "collapse-based interpretation" and "decoherence-based interpretation" are standard phrases in the taxonomy of interpretations. I'm 99% sure I didn't invent the terms.

I see no evidence for it-- I think CI is completely "decoherence based" because the entire purpose of an observation, from the CI perspective, is to achieve decoherence.

But the distinguishing point is that you don't interpret the resulting mixed state as referring to a parmstate, but instead interpret it as expressing ignorance.

I'm not sure what you are saying here,

I believe my intent was to preempt some response I had imagined was possible. But I'm glad I made it because of your response.

and I find it highly likely that there is indeed some new physics to discover that will reveal unitarity as only nearly unbroken. It's nearly unbroken status will explain why QM works, and breaking it will explain the collapse, but to get there, we need to understand what perception is. Or maybe, by getting there, we will finally figure out what perception is.

This touches on the underlying motivation that I think drives many CI devotees -- and is one of the aspects associated with CI that I have the most contempt for. I feel like people are effectively saying:

I don't like quantum mechanical view, so I will try my best to make do with pre-quantum views until the next scientific theory comes along, which I tacitly hope with be more palatable.​

For the record, I think "QM-style theories keep working all the way up", "new physics that produces collapse and works all the way up" and "new physics diverges even further from the classical intuition but works all the way" are all perfectly good outcomes.

As an aside, doesn't this position essentially boil down to "I think that QM is explained by a hidden-variable theory?" It's been a while since I thought about this matter -- isn't the known problems with hidden-variable theories a rather serious blow to the prospects of this position?

we'll feel less silly if we didn't build a world view around it first.

You keep saying this. But why should we feel silly?

Making mistakes is normal. And even if you put a lot of effort into something that ultimately turns out to be in error, you can still usually salvage lots of useful stuff.

And even if it turns out that a quantum mechanical world view would need to be discarded 25 years down the road, so what? You're almost surely better off having used a quantum mechanical world view for those 25 years than you would have been struggling along with a pre-quantum world view.

(And, of course, your pre-quantum world view might not be useful at all for the next theory after quantum mechanics!)

Simple: we don't measure parmstates, we measure eigenvalues that associate with eigenstates.

As far as I can tell, the observed eigenvalue is parm.


Remember, Einstein said "make the theory as simple as possible -- but no simpler!" Rejecting parmstates makes things too simple, because parmstates are needed to fully evaluate how QM connects to observation.

 

[Fyzix]

This is true for any interpretation except MWI. MWI is NULL interpretation, all wordy stuff in it is just an explanation (for humans) that no extra stuff is needed and pure QM math is enough. This is the only interpretation compatible with MUH.

I just can't let this blatant lie fly..
There is no way to get the Born rule in the math
There is no way to get a preferred basis in the math, the math does NOT give you a branching structure.
Decoherence is not enough to explain branching...

Also as for your MUH, not even David Deutsch supports it.

Also it blows my mind that you make the error of not even understanding that whatever the true theory underlying QM is, it will have the correct math, that's just logic 101.
As for MWI we know it does not have the right math.
So it is INCONSISTENT with MUH.

 

[Dmitry67]

Lets not start it over again. Born rule had been derived by Max Tegmark.
And please don't confuse people with your misconceptions. MWI does not have any math except the one QM and Decoherence provide. If you don't agree, provide any 'MWI-specific formula'

 

[Dmitry67]

I just have my opinion. Having an opinion does not mean 'trying to believe desperately in something'. You have a different opinion, which is normal, but please stop injecting into thread the emotional stuff, telling me what I should 'desperately' do.

As you did not reply to my question in your post above (good example of what Max Tegmark calls 'word baggage' - mostly it contains names you want to hide behind), I have to repeat it and to add another one.

1. Please provide any 'MWI-specific' formula which does not exist neigher in standard QM nor in the derivation of the Qunatum Decoherence. Formula does mean formula, not a wordy stuff.
2. Please name the 'additional postulates' of MWI (except the Born riule, which is derived by Max Tegmark). To be specific, 'postulate' is an axiom, which can't be derived from the theory. For example, Quantum Decoherence is not an axiom.

Frankly, I don't expect any answers, except 'In my private conversations with Deutsch, Wallace, Newton and Einstein they all had mentioned that they hated MWI' :)

 

[Fyzix]

I literally JUST TOLD YOU which postulates YOU have to come up with.
That's not my job...

No Born Rule is not derived by Tegmark, at best with his additional infinite universe(+1 postulate) he MIGHT have managed to get the Born Rule within that speculative picture.
So no and even if you want to keep insisting on it, it added a additional postulate to achieve it. (although I realize that you probably don't have anything against postulating infinity when u already accept MUH).

As for "private conversations", yes I have had them with both Deutsch and Wallace, but both are strong adherents of MWI, I have never said anything else.
If you want Deutsch's newest thoughts on the matter read his book and then read this critical review of his book: http://physicsworld.com/cws/article/indepth/47021 which shows how it reintroduces the preferred basis problem and still haven't solved the Born Rule.

As for Wallace he keeps changing axioms to get the Born Rule in his decision theoretic approach and he constantly changes the whole ontology of his MWI.
He has some vague "emergence" assumptions, but Maudlin has pointed out why this doesn't work.
Which you can read in the Many Worlds? book or Maudlin's new book.

Zeh openly says he postulate the Born Rule (read the papers on his site)

 

[Dmitry67]

Thank you for mentioning 'infinite universe' - yes, sure, it is an additional postulate, and it was my mistake for not mentioning it. Our universe is nearly flat and current observations can't tell us if it is finite or infinite. However, that postulate sounds very reasonable for me for the reasons of mathematical beauty, also described by Max Tegmark in his work.

So I think it is much easier to accept the 'Infinite Universe' and to derive the Born Rule postulate than postulate to the Born Rule explicitly.

Do you know if Wallace or Zeh have any motivation for not accepting the derivation by Max Tegmark?

 

[Fyzix]

Thank you for mentioning 'infinite universe' - yes, sure, it is an additional postulate, and it was my mistake for not mentioning it. Our universe is nearly flat and current observations can't tell us if it is finite or infinite. However, that postulate sounds very reasonable for me for the reasons of mathematical beauty, also described by Max Tegmark in his work.

So I think it is much easier to accept the 'Infinite Universe' and to derive the Born Rule postulate than postulate to the Born Rule explicitly.

Do you know if Wallace or Zeh have any motivation for not accepting the derivation by Max Tegmark?

Well then atleast you admit that it is a additional postulate and then we can agree, that maybe indeed Tegmark can succesfully derive the Born Rule within that model.

I'm not sure about Wallace, but seeing as he has been working on the decision-theoretic approach the last decade I think he is pretty much convinced that approach is the right one.
As for Zeh I don't think it would be right to publish his personal correspondence on a public forum like this, but mainly he objected to it due to the fact that it rests on the assumption of infinity.

 

[Hurkyl]

I have seen NOONE else endorse this approach and everyone I have asked (even prominent MWI'ers) say that the approach is flawed.

The specific section on the confusion operators and deriving the frequentist limit is a good one. When I figured it out on my own a few years back, I was given the impression that the argument was already old hat then. It's the rest of the paper that's 'weird', trying to recast QM as an ensemble over repetitions in an infinite universe.

 

[Ken G]

This is true for any interpretation except MWI. MWI is NULL interpretation, all wordy stuff in it is just an explanation (for humans) that no extra stuff is needed and pure QM math is enough.

MWI is only a null interpretation if you adopt a rationalist perspective, that is what I keep pointing out. The rationalists on this thread just don't seem to get emiricism, to the extent they don't even recognize the other possibility so make claims like "MWI is a null interpretation." I can only repeat: in empiricism, the purpose of an observation is not to check a theory, but to establish what is true. Thus, it is the purpose of a theory to make sense of the observation. If the theory has to postulate "other worlds" that are not observed to make sense of the observations, it has overstepped its authority and is now asserting what is true. Observations do that, only.

This is the only interpretation compatible with MUH.

In my view, MUH completely removes any importance in MWI. MUH asserts the existence of any world that is describable mathematically, so the "many worlds" of MWI are now completely superfluous, as they are "just more" worlds that can be described mathematically, and have no more connection to our observations than mathematical theories which clearly don't agree with our observations (but exist in "some other world"). MUH is where MWI goes to die.

 

[Hurkyl]

If the theory has to postulate "other worlds" that are not observed to make sense of the observations, it has overstepped its authority and is now asserting what is true. Observations do that, only.

Wonderful. I just flipped a coin, and my observation was parm -- I saw heads.

What? My observation was not parm? What did you observe to tell you that? Oh, there wasn't an observation that told you that?

If a theory has to postulate repeated collapses to make sense of observations, it has overstepped its authority and is now asserting what is true. Observations do that, only.

 

[Fyzix]

Hurkyl: do you really believe in MWI or are you playing "Devil's Advocate" ?

 

[Hurkyl]

By the way, have you yet encountered the paradox of grue and bleen? (the following story is mine)While out for a walk on the morning of December 22, 2013, you notice a large congregation of space aliens in the park. Curious, you go to investigate and ask one what's going on.

Alien: "Oh, your planet is about to experience an interesting phenomenon! The sky and ground are going to swap colors! Aren't you excited?"

You: "Swap colors? What do you mean?"

Alien: "It's the darnedest thing! Your sky is bleen and your grass is grue, but at noon today they'll swap -- your sky will be grue and your grass will be bleen! You have such wonders on your planet."

You: "Grue? Bleen?"

Alien: "Oh, that's right. I heard you have some weird colors... blue and green, I think. I heard your blue color is normally our bleen, but at until noon today, your blue color switches to grue. And conversely for green. Where did you ever get such a strange notion of color anyways? Weird coincidence that your colors switch the same time the sky and ground swap colors. Oh well, I've got to get a good spot to watch the change!"

 

[Hurkyl]

Hurkyl: do you really believe in MWI or are you playing "Devil's Advocate" ?

https://www.physicsforums.com/showpost.php?p=3529275&postcount=393

 

[Fyzix]

I think decoherence-based interpretations are the superior to collapse-based interpretations for understanding quantum mechanics. Among those, MWI appears to be the most promising approach for understanding the quantum-classical interface -- but that purpose does remain conjectural until quantum thermodynamics is sufficiently well-developed.

Also, I suspect that the MWI has more ways to fail than the CI. Since it hasn't failed, that would imply MWI enjoys a greater share of empirical support than CI.

Insomuch as one uses science to understand reality, MWI seems to be the understanding that QM tells us.

But what about the preferred basis problem ?
Also sure you thought of something similar as Tegmark when it came to Born Rule, but it still requires infinity.
Last but not least, ontology objection from Maudlin?

 

[Ken G]

To the best of my knowledge, "collapse-based interpretation" and "decoherence-based interpretation" are standard phrases in the taxonomy of interpretations. I'm 99% sure I didn't invent the terms.

It was never my claim that you invented the misnomer, I merely gave an argument, which you did not address, why it was a misnomer. If you hold that it is not a misnomer, please explain CI to me without decoherence. The irony is, people who criticize CI invariably describe it in a way that leaves out decoherence, and then they blame CI for not being informed by decoherence. The actual fact is, CI needs decoherence just as badly as MWI does, as anyone who understands CI knows. If there were times when Bohr didn't get decoherence, he would have been troubled by something missing from his perspective, and then when he did get decoherence, he would have said "ah, exactly, now what I've been saying all along finally becomes clear." Decoherence is what enables a subsystem in a pure state to become a mixed state when the closed system is projected onto that open subsystem, all prior to the perception of that subsystem. CI is how we interpret the resulting mixed state, prior to said perception (a perception that will return the substate to a pure state). Now, please tell me what is "not decoherence based" about that?

But the distinguishing point is that you don't interpret the resulting mixed state as referring to a parmstate, but instead interpret it as expressing ignorance.

Precisely. But we must go a step farther-- we must understand what "ignorance" actually is. To understand ignorance, you must have a model for knowledge. That means you have to choose empiricism or rationalism. CI chooses empiricism. This is the crucial element, because if one asserts that observation is truth, then one must frame knowledge of truth in terms of what is observable. That means one is not building a rationalistic model of "ignorance" (which doesn't work very well from the rationalist god's eye view where ignorance is an odd concept), one is saying that ignorance means not knowing how an observation will come out. But not knowing that means there is something we don't know about perception, it all comes down to perception because that is what an empiricist means when they talk about an "outcome" of an observation. Empiricism requires that we include the act of perception in our ontological description of reality, rationalism does not require that. This is the crux of the distinction between MWI and CI, it is absolutely inescapably intertwined with the different knowledge models of rationalists and empiricists.

I feel like people are effectively saying:

I don't like quantum mechanical view, so I will try my best to make do with pre-quantum views until the next scientific theory comes along, which I tacitly hope with be more palatable.​

No, you are quite wrong here. Really couldn't be much wronger, to be honest. You think that empiricists don't like quantum mechanics. How could you possibly think that Bohr and Heisenberg didn't find quantum mechanics palatable? They loved quantum mechanics, as do I. I'm mostly an empiricist, and I think quantum mechanics is just splendid, fascinating, amazing. I also think it is probably not exactly right, and as I said, I think unitarity is probably only nearly unbroken. None of that has anything to do with "not liking" QM, it has to do with a respect for the history of science, and a basic vein of skepticism that runs through me. I agree that empiricists are generally more skeptical of theories, whereas rationalists tend to marry them after one date, but that has nothing to do with how much they like them-- it has to do with what they think truth is.

You think empiricists don't like QM, and that's why they take a skeptical stance toward it, but your mistake about that motivation is actually revealing your own motivation for not liking empiricism! Empiricists would have adopted skeptical stances around every theory, including classical mechanics, so it has nothing to do with liking classical thinking better. Instead, it has to do with a basic recognition that physics is what we can say about nature, so we are in the conversation, including the fact that we are "classical" systems (which does not mean we are systems that obey classical laws, because empiricists don't think we really obey laws at all, they think laws are how we understand and predict behavior). If there aren't laws that "govern" nature, then the act of seeking laws is something quite a bit different from how rationalists frame that process, and this has nothing to do with the details of any theory.

For the record, I think "QM-style theories keep working all the way up", "new physics that produces collapse and works all the way up" and "new physics diverges even further from the classical intuition but works all the way" are all perfectly good outcomes.

Yes, at the end of the day we all just want the truth. Still, my point is that it is always questionable to build a world view on any particular set of postulates, it is just taking those postulates too seriously and it has never been right yet. It might be a valid constraint to impose when seeking the next theory, I view the seeking of theories to be the valid application of rationalism. It is the building of world views that I classify as rationalism taken to a naive extreme.

As an aside, doesn't this position essentially boil down to "I think that QM is explained by a hidden-variable theory?"

But you are again coming from a rationalistic perspective! You think that for QM to be wrong, some other theory has to be right. The empiricist just doesn't think that way-- there is no right theory, physics is about what we can say about nature, and that is never going to be nature herself. We get observations to tell us what is true, and we get theory to help us predict what will happen, and we get a limited sense that we understand why it happened. That's it, that's science-- that's all we get.

And even if it turns out that a quantum mechanical world view would need to be discarded 25 years down the road, so what? You're almost surely better off having used a quantum mechanical world view for those 25 years than you would have been struggling along with a pre-quantum world view.

Again, you see the options as "quantum world view" or "pre-quantum world view." Those are just the two possibilities open to rationalists. Empiricists prefer a different approach: skepticism of all world views. They are all seen as valuable for what they are good at, yet limited in truthfulness, and certainly not to be taken seriously. When Newton's laws made the future seem to be nothing but an echo of the past, the empiricist asked "what observation proves that?" None. When MWI asserts that the universe as a whole does not change at all, the empiricist asks "what observation proves that?" None. So it goes with world views.

Remember, Einstein said "make the theory as simple as possible -- but no simpler!" Rejecting parmstates makes things too simple, because parmstates are needed to fully evaluate how QM connects to observation.

So? QM invokes wave functions, we know that, it is part of the theory. We are talking about interpreting the parmstates, not "rejecting" them from the theory altogether.

 

[Hurkyl]

But what about the preferred basis problem ?

There is no problem with the idea that talking about "splitting" relative to a basis depends on your choice of basis.

Or are you referring to the question of whether QM predicts rulers can exist? That question has to wait for quantum thermodynamics to be developed.

But I'm aware of no obstacle posed by the observed existence of rulers, nor by the operational definition that rulers are things that measure position.

Also sure you thought of something similar as Tegmark when it came to Born Rule, but it still requires infinity.

Of course it does.

Frequency probability is the interpretation of probability that defines an event's probability as the limit of its relative frequency in a large number of trials. (ref: wikipedia)​

Last but not least, ontology objection from Maudlin?

I don't even know what this is.

 

[Ken G]

Wonderful. I just flipped a coin, and my observation was parm -- I saw heads.

You saw heads, that's all you get. What do you mean it was parm? I don't have to use observations to prove to you what you didn't observe, I just want to know what you did observe. Did you observe that heads was parm? How did you observe that?

 

[Ken G]

Alien: "Oh, that's right. I heard you have some weird colors... blue and green, I think. I heard your blue color is normally our bleen, but at until noon today, your blue color switches to grue. And conversely for green. Where did you ever get such a strange notion of color anyways? Weird coincidence that your colors switch the same time the sky and ground swap colors.

This presents no paradox to the empiricist, actually. Observations are of consistencies, not properties. We don't need a rationalist description of what green "really is" to say that we perceived it, the consistency of the perception provides sufficient basis for counting it as part of what is true. Yes, this means perception remains fundamentally a mystery, and yes, it means perception is something our brains are doing not something "out there", but that's just how it is. We won't really understand the connections between the empiricist and rationalist perspectives until we understand better what perception is, and I think its role in "collapse" is going to be the crucial avenue for reaching that understanding. Ergo, rejecting collapse merely makes it impossible to ever understand what perception is, which isn't surprising because MWI sees no significance in perception anyway.

 

[Hurkyl]

You saw heads, that's all you get. What do you mean it was parm? I don't have to use observations to prove to you what you didn't observe, I just want to know what you did observe. Did you observe that heads was parm? How did you observe that?

I saw heads.

It was parm because experiences are parm. You keep telling me they aren't parm. What do you mean? Did you observe that heads wasn't parm? How did you observe that?
This last paragraph is devil's advocate, because I know full well that (classically, at least) parmness is not observable, because the observations would share any parmness present, rather than transcending it. And, while I'm not quite sure, I think you know that too.

My observations do not include the adjective "not parm". When you tell me my observations are not parm, you are making theoretical assertions about what is true -- the very thing you are trying to criticize MWI of doing. But I don't see why MWI is more deserving of such criticism than CI.

 

[Ken G]

I saw heads.

It was parm because experiences are parm. You keep telling me they aren't parm.

No, I'm not telling you what experiences are, no empiricist can tell you that. Experiences are the ultimate mystery in empiricism. They define truth, that's it, that is the axiom. They do not have attributes like "parm" or "not parm", the words "parm" or "not parm" apply to rationalistic conceptualizations of what is going on, the experience just is. I saw heads, heads is what happened. Saying the experience is parm or not parm is a category error in empiricism.

MWI cannot stop there. It must try to say why heads happened, or heads isn't what happened, because to a rationalist, what happens is not what is experienced, but rather what is conceptually true, mathematically true. So the postulates dictate truth, and the experiences check the postulates, but have no other connection to the truth. The postulates are where "parm" comes from, not the experiences, but the experiences are consistent with the rationalist description of the parm state. I never said observations ruled out parmstates, I said that if the observations are regarded as what is fundamentally true, then parmstates are not what is fundamentally true (which is quite a bit different from saying they are false, because only observations can say what is false also). In empiricism, parmstates are neither true nor false, they are aspects of a theory. But since the parmstates are not regarded as true (or false), they cannot be regarded as real. They can be regarded as unreal on the basis of not being real, it does not require they be falsified.

This last paragraph is devil's advocate, because I know full well that (classically, at least) parmness is not observable, because the observations would share any parmness present, rather than transcending it. And, while I'm not quite sure, I think you know that too.

Yes, absolutely, parmness is not observable. One can neither observe its presence, nor its absence. That's why the rationalist view cannot be ruled out by the observations, but it is ruled out by the empiricist anyway, based on the axioms of empiricism-- that which cannot be observed is angels on a pin, and the elements of any theory are the map not the territory. Whether or not one regards the observations as the territory or just another kind of map depends on how much of a realist one is, but the fact remains, the theories change but the observed outcomes don't.

My observations do not include the adjective "not parm". When you tell me my observations are not parm, you are making theoretical assertions about what is true -- the very thing you are trying to criticize MWI of doing.

Now you are saying that empiricism does not have axioms. That is untrue, empiricism does have axioms.

 

[Hurkyl]

In empiricism, parmstates are neither true nor false, they are aspects of a theory. But not being true means they are not real.

Then where do you get off claiming that empiricism rejects MWI because it connects the theory to perception through parmstates? And how do you not manage to reject CI for the same reason -- that it connects theory to perception through collapse into non-parmness?

 

[Hurkyl]

We don't need a rationalist description of what green "really is" to say that we perceived it,

The paradox relates to why we perceive green instead of grue, why we think our experiences lead us to believe the sky is blue and not bleen.

and I think its role in "collapse" is going to be the crucial avenue for reaching that understanding. Ergo, rejecting collapse merely makes it impossible to ever understand what perception is, which isn't surprising because MWI sees no significance in perception anyway.

I am interested in understanding the appearance of collapse -- I see no reason to insist on anything stronger.

Failing to insist on something stronger does not equate to "see[ing] no significance in perception".

 

[Ken G]

Then where do you get off claiming that empiricism rejects MWI because it connects the theory to perception through parmstates?

That isn't the problem that empiricists have with MWI. The problem is treating MWI like it was real. Which MWI enthusiasts always do, by the way (witness "quantum suicide"). Empiricists who understand QM have no issue with treating MWI like a valid pedagogy for picturing how QM works, they object to taking that valid pedagogy of a particular theory and following the faulty logic "the theory agrees with observations, so not only should the theory be regarded as true in reality (yuck), but a highly rationalistic interpretation of the theory should also be regarded as true in reality (double yuck)." Not by empiricist axioms it shouldn't.

And how do you not manage to reject CI for the same reason -- that it connects theory to perception through collapse into non-parmness?

Empiricism has no problem at all with "connecting theory to perception", indeed it views that connection as the whole point of a theory. The issue is with associating the theory with what is either true, or real. In empiricism, nothing but observations are ever true or ever real, and the history of science is rather clear on the advantage of thinking that way.

 

[Ken G]

The paradox relates to why we perceive green instead of grue, why we think our experiences lead us to believe the sky is blue and not bleen.

Again, that is no paradox to empiricism, because it is not the label of the perception that matters, but the consistency of the perception. If we came out tomorrow and said "henceforth green will be called grue", no empiricist would bat an eyelash, because the consistency would be perfectly maintained under the transformation green-->grue. That is also why there is no need to establish that two observers perceive the same thing when they perceive green-- the truth of the perception of green is entirely in the consistency of the perception, which allows the observers to agree on a "green" perception. There doesn't even have to be any such thing as the "actual" perception of green, merely the consistency class.

I am interested in understanding the appearance of collapse -- I see no reason to insist on anything stronger.

Yes, empiricist are also interested in understanding that, they just drop the word "appearance" because there is no such thing as a difference between collapse and a consistent appearance of collapse. Consistent appearance = truth.

Now, to stave off any complaints about "mirages" or some such, we can well agree that a mirage is a consistent appearance. Hence it is true-- it is true that the eye is really seeing something that the brain can erroneously interpret as water, and the reason the interpretation of the observation is erroneous is because the consistency of that perception breaks down when you go and stick your toes into the mirage. This is generally not any kind of problem for physics experiments, though we certainly try to get as many consistency checks as possible, and who knows-- tomorrow we might find a consistency breakdown that suggests "the Matrix" was right all along.

Failing to insist on something stronger does not equate to "see[ing] no significance in perception".

Explain what perception is from an MWI standpoint. Does it have any importance in the theory of quantum mechanics? (Not in testing the theory, in the theory itself-- in the description of what is happening, according to quantum mechanics.)

 

[Fra]

I interpret Hurky's meaning of "apperance" so as to acknowledge the de facto collapse but denial of it at deeper level.

In terms of information update, this is the same as to acknowledge that there sort of is an information update, but that there always is a different (larger) perspective in which the information is more complete and the information update is just understood as apparent when you truncate the picture.

The problem with this picture is that "removing the collapse" requires changing perspective. And this choice is non-physical for a given observer, because it's stuck with the perspective it has.

MWIers always refer back to the master view in which there is no collapse, and from this view surely you can say that the collapse is apparent.

The problem is IMO that this master view does seem to map onto any physical views. It's more some kind of mathematical view whose value is unclear beyond maintainin a unitary picture, even though not one corresponding to a real observer.

But this seems to pretty much what the core of entire thread has been about.

/Fredrik

 

[Ken G]

In terms of information update, this is the same as to acknowledge that there sort of is an information update, but that there always is a different (larger) perspective in which the information is more complete and the information update is just understood as apparent when you truncate the picture.

That's how I'm getting him too, if he means something else he will need to clarify.

The problem with this picture is that "removing the collapse" requires changing perspective. And this choice is non-physical for a given observer, because it's stuck with the perspective it has.

That's just how I see the problem too.

MWIers always refer back to the master view in which there is no collapse, and from this view surely you can say that the collapse is apparent.

The problem is IMO that this master view does seem to map onto any physical views. It's more some kind of mathematical view whose value is unclear beyond maintainin a unitary picture, even though not one corresponding to a real observer.

I would equate that "master view" with the "rationalist view" about what constitutes knowledge of truth. The debate we are having right now dates 2500 years back to Parmenides, who was the ultimate rationalist for the ages-- he said that observations were illusions, because they give the "appearance" of change, when change is in fact logically impossible. When what we experience disagrees with what we can know to be true, our experience is not a reliable guide. 2500 years later, the MWIers have taken up his cause, and interestingly, in exactly the same way-- that change is impossible (if the entire closed system is included).

 

[Fra]

Yes we seem to agree in our characterization of MWI.

our experience is not a reliable guide

This is partly true, but the point is that this is the only choice we have. This IS no reliable guide. Fantasizing about it, doesn't help. We need to make decisions based upon incomplete and imperfect decision base. But so does all other systems, so it's still a fair game. At least this is my firm view.

To me the very basic quest IS a decision problem based upon incomplete information. It seems to me it should be clear why "changing perspective into one where the information is complete" is completely missing the point.

But to be consistent in the critique here, it's in place to acknowledge that there exists also platonic structures also in CI (or actually in QM itself). So it's a fair question to those (like myself) that critique MWI view why we then can accept other platonic structures. I acknowledge that point and I do take it seriously, but my take on this is not that CI solves it all (it doesn't) I rather think the problem is in QM framework itself.

/Fredrik

 

[Hurkyl]

We need to understand perception by its observable behavior. I asserted a serious of propositions earlier, things like

• Observations are outcome-valued
• If I observe a system and see outcome X and you observe the same system and see outcome Y, then X = Y. (assuming our observations are reliable)
• If I observed X and I observed Y (both observations of the same thing), then X = Y.
• If I saw X, my future actions can be affected by X

In the stronger form of MWI, it's not good enough for observers to be completely external to a quantum system. While it may be a decent approximation, we are not gods that get to see wave-functions and observable algebras and eigenvalues. Instead, we are part of the system, and we only get to interact with the system through quantum-mechanical processes.

In other words, the relevant physical properties of a system should not be things like "that particle is at position X", but instead "that observer saw that particle at position X".The challenge, now, is to consider if and how the physical properties of a quantum system can reflect the behaviors of perception.Do I have more precise details on perception? Not really -- but then I don't have precise details even from a classical viewpoint.

Doing a better job than classical mechanics is too ambitious of a goal to insist upon -- it's good enough to simply put the above on comparable footing with classical mechanics, since it means we've fully integrated into quantum mechanics what we already know and understand about perception from centuries of pre-quantum physics. That in of itself is an important advance.The central idea to get started is that we know from classical statistical mechanics that our perceptions can be reflected by probability distributions over the phase space -- by mixed states. Everett's idea was that mixtures appear naturally in relative states.

And that's the central idea behind approaches to answer the challenge -- to reinterpret observations not as the experiences of external observers as CI might do, but instead as the experiences of internal observers -- observers that are part of the quantum system -- through the appearance of mixed states (ala Everett) and a correspondence (approximately) mathematically identical to the math of statistical mechanics.

And the main thing this pushes on us is to adjust our interpretation of statistical mechanics to be of (probabilistic) parmness, rather than being of ignorance probabilities.

 

[Ken G]

In other words, the relevant physical properties of a system should not be things like "that particle is at position X", but instead "that observer saw that particle at position X".

I agree, I mentioned that earlier. Indeed, I think this is a classic duality of <observer|observation>.

Do I have more precise details on perception? Not really -- but then I don't have precise details even from a classical viewpoint.

No one does-- perception remains the core mystery. Indeed, that's why I like CI-- it starts with this recognition, so there is nothing to explain away.

And that's the central idea behind approaches to answer the challenge -- to reinterpret observations not as the experiences of external observers as CI might do, but instead as the experiences of internal observers -- observers that are part of the quantum system -- through the appearance of mixed states (ala Everett) and a correspondence (approximately) mathematically identical to the math of statistical mechanics.

Right. But CI says that observers cannot be internal to the system, because the very definition of what we mean by "a system" involves the crucial subject/object dichotomy that is the core of empirical science. The alternative is to step away from empiricism, and allow that the observer is a part of the observation, not just responsible for the observation.

And the main thing this pushes on us is to adjust our interpretation of statistical mechanics to be of (probabilistic) parmness, rather than being of ignorance probabilities.

I think you pretty much have to end up "distributing over the observers in the many worlds", rather than distributing over trials of an identical experiment. But the distribution is the same, this is not really a fundamental difference, scientifically speaking. My deepest suspicion is that there never really was any scientific difference between CI and MWI.

 

[Fra]

And that's the central idea behind approaches to answer the challenge -- to reinterpret observations not as the experiences of external observers as CI might do, but instead as the experiences of internal observers -- observers that are part of the quantum system -- through the appearance of mixed states (ala Everett) and a correspondence (approximately) mathematically identical to the math of statistical mechanics.

Right. But CI says that observers cannot be internal to the system, because the very definition of what we mean by "a system" involves the crucial subject/object dichotomy that is the core of empirical science. The alternative is to step away from empiricism, and allow that the observer is a part of the observation, not just responsible for the observation.

I agree the point of inside observers is indeed a central issue.

In CI the classical background is where the entire theory and all statistics "lives". And the classical background is never affected by the systems backreaction. Also there is no "limit" of how much information that can be stored in the classical environment. This obviously does not make sense, except in a limiting case where a large observer embraces and observes a small subsystem. This is exactly where QM is successful anyway. This also where we can say that the system under observation is closed.

When observing the entire environment or a subsystem in a large uncontrolled environment the closedness assumption fails, so we need a strategy that works for open systems. And it's exactly when we are forced to formulate a theory of sliding windows of history that the finiteness of the observer becomes important too, this leads to incomplete basis for decisions and actions and leads to decoupling of parts of the universe. Two parts interact more weakly simply because they are unable to "remember everything". I think this is even why we have problems of renormalisation with some interactions since the entire formalism is made for the "classical background" does not quite make sense when we try to consistently use it to scale it with energy.

MWI doesn't seem to address any of this. The above issues is my "problems" with CI, and they precisely have to do with lack of "inside views".

/Fredrik

 

[Fra]

[*] If I observe a system and see outcome X and you observe the same system and see outcome Y, then X = Y. (assuming our observations are reliable)

Do you by "=" informally symbolize some equivalence class? So that "X = Y" means something like "there exists a transformation depending on the observers, that transforms X into Y"?

Otherwise, then only way to comapre X and Y IMHO are to let the observers interact (here I think a little bit like Rovelli in his RQM) either with each other or with a third party than makes the comparasion, and even then it doesn't seem a priori obvious to me that the transformation that defines the eqvuivalence is known to the system making the comparasion.

/Fredrik

 

[Hurkyl]

Indeed, that's why I like CI-- it starts with this recognition, so there is nothing to explain away.

Right. But CI says that observers cannot be internal to the system, because the very definition of what we mean by "a system" involves the crucial subject/object dichotomy that is the core of empirical science.

I get the impression it was exactly the other way around.

For half it's life, QM resisted attempts to reconcile the formalism of quantum mechanics with the fact of classical observation, and there were even theoretical proofs to that effect.

The "cut" between the quantum and classical "realms" was put in place because, despite the failed attempts to reconcile them, we could at least still work with the two realms separately, and heuristically manage their interface via the Born rule (to get the classical side's statistics) and the collapse postulate (to get the quantum side's dynamics).


With no hope in sight for merging the two realms, people took the cut seriously, and thus the myriad of collapse-based interpretations.

(Okay, I guess CI can be said to "start with the cut" in the same sense that MWI can be said to "start with unitary evolution")



But half a century ago, all of the strong reasons for the necessity of the cut vanished -- Bohm resurrected de Brogile's abandoned approach, and Everett started applying relative states to the problem, and both proved to be viable interpretational approaches lying outside the scope of the old impossibility proofs didn't apply.

As an outsider looking in, it really looks more like CI is operating more through inertia than by compelling arguments, and that these alternative approaches are rapidly gaining credibility.

very definition of what we mean by "a system" involves the crucial subject/object dichotomy that is the core of empirical science. The alternative is to step away from empiricism, and allow that the observer is a part of the observation, not just responsible for the observation.

I'm not familiar with the terms "subject/object" used in this fashion. But on the following sentence, I can remark that I can reflect upon my experiences. I think think about the fact that I observed things, I can observe and think about the fact that you observe things. I can reason with these facts, and make predictions involving them.

Formally, there doesn't appear to be anything special about "I observed X".

As an aside:

a classic duality of <observer|observation>

My first thought is "two different views on the same thing" -- e.g. the duality between algebra and geometry -- and your use of bra-ket reinforces that thought.

But would I be correct that you really meant "dual" in the sense of dualism? Two fundamentally different kinds of things?

 

[Dmitry67]

No one does-- perception remains the core mystery. Indeed, that's why I like CI-- it starts with this recognition, so there is nothing to explain away.

No. CI starts from a 'measurement' which is much weaker thing - it does not involve any living and conscious beings at all.

If for you perception is axiomatic and irreductable to the behavior of big systems of QM 'particles', then you prefer a different flavor of CI (it is even not called CI):

consciousness causes the collapse

http://en.wikipedia.org/wiki/Interp...rpretation:_consciousness_causes_the_collapse

let me ask you a question, can, in pronciple, perceptor be described as QM system?

 

[Fra]

I know you asked Ken but I'll add my answer:

let me ask you a question, can, in pronciple, perceptor be described as QM system?

Yes of course, but only relative to another perceptor! Without perceptors/observers there is no place to encode/define any descriptions.

A perceptor never describes herself in terms of a QM system. The observers own state itself acts IMO like a generalization of a prior in classical probability, and there is no way for her to question herself. Just like the prior is just the "starting point" for the placement of bets and learning. If the prior is "uncertain" with respect to another observer is irrelevant to the decision process of the prior itself.

This is on par with how you can act rationally on wrong information. A corrupt prior leads to corrupted bets. But I think sometimes "corrupted bets" are exactly what explains interactions. The fact that two systems are differently informed, leads them to contradicting expectations and interactions that serves to shift the "priors". Thus by on purpose "misinforming" two parties you can predict an interaction (conflict).

But I think the big difference is that "misinformation" is really happening due to physics. There simply is not notion of "correct information", no more than there exists a measure of right and wrong that humanity can agree upon because everyone always over-weights their own benefits. That's nothing wrong, it's just how things de facto work.

/Fredrik

 

[Dmitry67]

Fra, thank you.
Here is another question (I am still interested to get an answer from Ken).

What is a difference between perception and measurement?
Is perception a measurement performed by a conscious being?
And what collapses wavefunction in CI (as Ken prefers CI) – perception of measurement?

I am trying to lead Ken into a logical trap :)

 

[Hurkyl]

Aah! I forgot the main thing I wanted to say.

We already have wonderful examples of the benefit of 'observer in the system'. The symmetries of the laws of mechanics put severe constraints on what sorts of quantities are physically meaningful... but only on the hypothesis that observers obey the laws of mechanics as well.

For example, the special relativistic notion that absolute simultaneity is not meaningful. Einstein's thought experiments -- even if we try to interpret them as involving external observers (but I don't see how to do so) -- only imply the issue is subtle. We don't get the final conclusion that it's not physically meaningful until we work out the physical laws and see the symmetry.

But if we try to say that the observers aren't part of the system and aren't described by the laws of SR, we no longer have a compelling argument -- there's no longer any reason why simultaneity must be relative for observers. If we try to recover the notion, we need a whole new set of physical postulates about the interface between observers and special relativity.

 

[Fra]

What is a difference between perception and measurement?

I think it was Ken who started to use the word perception, so I'm not sure what distinction he likes to make.

But to speak for myself, above I used perceptor as synonymous to observer.

About "measurement" someones by measurement we include the whole package of preparations etc. But I see perception more like the "detection of outcome" of an measurement.

Is perception a measurement performed by a conscious being?

Not in the way I use the term. Conscioussness has nothing to do with this. Ianimate objects also have perceptions in MY view. (Note however that my view is not the "classical CI", I'm coming from CI but have a more radical view)

And what collapses wavefunction in CI (as Ken prefers CI) – perception of measurement?

The collapse of the wavefunction is nothing but an update of the observers state (which encodes information about it's environemnt - the "system" that he observes). The key is however that there exists no "neutral description" of the observers state. It can only be described incompletely from the perspecive of other observers, or for the observer itself it is like a prior.

So the collapse, and the breaking of unitary evolution, is in my view nothing else but "updating your expectations".

The information the observer has in it's "prior" we call wavefunction (+Hamiltonian(*)), IMPLIES a change, and defines a flow of time. Because the information contains a mix of information about initial states as well as information about "motion" and momenta, together the logical implication is evolution. In my view, the SE simply is the expression for this.

This information formes a "closed system" as long as you don't get new information and is subject to "self-evolution" described by SE. When you get and information update, it would be irrational to not update the expected evolution as well.

(*) In my personal extended view, the information contained in the state vector as well as the hamiltonian are part of one uinfied structure, and the hamiltonian flow are simply defined as an entropic kind of self imposed flow. I just wanted to mention this for consistentcy, but elaborating it gets out off topic.

/Fredrik

 

[Dmitry67]

Fra, now let's wait and see how Ken G answers these questions.

 

[Hurkyl]

Do you by "=" informally symbolize some equivalence class? So that "X = Y" means something like "there exists a transformation depending on the observers, that transforms X into Y"?

No, I meant equals. X and Y outcome-valued, and the expression "X=Y" is truth-valued. The assertion is that if X and Y actually reasonably behave like their intended interpretation, then X=Y has to be (very nearly) definitely true.

OTOH, if you set up the variables X and Y in a different fashion so that their respective outcome spaces aren't intended to be related by identity but instead some other function, then I would write "f(X) = Y" for the corresponding proposition. (or some other similar thing as appropriate)

Otherwise, then only way to comapre X and Y IMHO are to let the observers interact

I agree with you in two senses, and disagree in a third.

I disagree in the sense that "X=Y" is simply a mathematical object whose behavior can be treated by the theory.

I agree in the sense that "X=Y" is only meaningful for systems that include both me and you.

I also agree in the sense that a direct empirical test of "X=Y" requires interaction (direct or indirect) between the two observers.

(here I think a little bit like Rovelli in his RQM)

Rovelli's paper gave me the final key to my understanding of interpretations. Ironically, I feel like he missed the point! (although admittedly I have not fully digested his words)

I think he made the progress of realizing that Wigner's friend using a collapsed wave-function to analyze the system after he measures it and Wigner using the uncollapsed wave-function to analyze his friend doing measurement are both just as good descriptions of 'reality'.

But he erred in retaining the correspondence between observers and wave-functions -- he failed to make the final step to detaching observers from wave-functions. There's no particular reason why Wigner's friend must switch to the collapsed wave-function to continue his calculations, and there's no particular reason why Wigner can't consider the collapses of his wave-function. Or why they couldn't calculate using a third, completely different wave-function that makes the same predictions!

A classical analog would be something like

This is Wigner's coordinate chart; it's centered on him, and he never uses any other coordinates for calculation. That is Wigner's friend's coordinate chart; it's centered on him, and he never uses any other coordinates for calculation. But happily, we can transform between them​

instead of the more sensible alternative of just saying coordinate charts are coordinate charts, and observers are observers, and they have no prior connection to one another. Anyone can use any coordinate chart they want. Coordinate charts don't have to have to be linked to observers, and so forth.

 

[Fra]

But if we try to say that the observers aren't part of the system and aren't described by the laws of SR, we no longer have a compelling argument -- there's no longer any reason why simultaneity must be relative for observers.

I think I see your point and I think it's a great example indeed. This is the connection between observer invariance or covariance and symmetries in nature that effectively encode the laws of nature.

But I think I've expressed my view on this before (here I think I am far more radical then Ken though).

There are two ways to understand this, and my hunch is that you adopt the first, I adopt the second. Both solve the problem you address but in different way. Way(1) is the realist way, way(2) is the solipsist way.

(1) Either as a constraint on the set of observers if you think the laws are timeless platonic structures.
=> the laws of nature MUST be the same as inferred by all observers.

This is in fact how it works in SR and GR. One does not talke about arbitrary observers in relativity, the theory only works for observers that belong to the same equivalence class defined by the generators of the symmetry.

The question is: what is the most general set of observers? certainly poincare transformations are hardly the most general.

(2) Or as a democracy condition if you think of the laws as emergent.
=> there is observer democracy in the sense that each observer might infer different laws, but they all contribute in negotiating the consensus and the observer invariants. In this view laws are not timeless, they are evolving. Similarly the symmetries of the laws of physics are observer dependent. As we know the poincaare symmetry which defines notoins of particles etc are completely screwed up if we consider more general observers (outside the equivalence class defined by SR itself).

/Fredrik

 

[Ken G]

The "cut" between the quantum and classical "realms" was put in place because, despite the failed attempts to reconcile them, we could at least still work with the two realms separately, and heuristically manage their interface via the Born rule (to get the classical side's statistics) and the collapse postulate (to get the quantum side's dynamics).

I think Bohr's points on this matter are clear-- there is no quantum realm, and physics is what we can say about nature. The "cut" was never a cut between two worlds, it was always the cut between the observer and the observed, it is the subject/object dichotomy that is central to all scientific models because it is the beating heart of the scientific method.

With no hope in sight for merging the two realms, people took the cut seriously, and thus the myriad of collapse-based interpretations.

Again, I find your distaste for CI to be largely around not understanding it. I wouldn't like CI either if I think it was what you think it is! CI is nothing but the interpretation that is required by empiricism, with no added bells or whistles to make it seem like a more rationalistic world. Bohm and many worlds say that reality must make sense, but CI says all you'll ever have to make sense of is your perceptions of reality, so the rest will always be angels on a pin. That's empiricism.

(Okay, I guess CI can be said to "start with the cut" in the same sense that MWI can be said to "start with unitary evolution")

Precisely.

But half a century ago, all of the strong reasons for the necessity of the cut vanished -- Bohm resurrected de Brogile's abandoned approach, and Everett started applying relative states to the problem, and both proved to be viable interpretational approaches lying outside the scope of the old impossibility proofs didn't apply.

None of those are removed empiricism. Indeed, they were both throwbacks to old ways-- Bohm is a throwback to Newtonian determinism, where initial conditions control everything that happens later, and Everett is a throwback to ancient Greek rationalism, in which the truth is the mathematics. Then when the mathematical postulates change, the truth magically follows the changes.

As an outsider looking in, it really looks more like CI is operating more through inertia than by compelling arguments, and that these alternative approaches are rapidly gaining credibility.

Empiricism isn't inertia-- is the bedrock of science. Yes, there is a lot of inertia in bedrock.

Formally, there doesn't appear to be anything special about "I observed X".

The difference is between "I observed X" and "X was observed." Science always treats those sentences as identical, hence the problem. It's not that there's anything special about me, it's just the basic issue of whether anything like me needs to be included in the description.

My first thought is "two different views on the same thing" -- e.g. the duality between algebra and geometry -- and your use of bra-ket reinforces that thought.

But would I be correct that you really meant "dual" in the sense of dualism? Two fundamentally different kinds of things?

I mean dual as in the sense that the two can be interchanged, they are two halves of the same coin, neither existing without the other.

 

[Ken G]

No. CI starts from a 'measurement' which is much weaker thing - it does not involve any living and conscious beings at all.

In empiricism, the way we know anything is because we perceive it. The way we understand measurement is how we perceive measurements. So no, measurement is not a weaker thing than perception, it is a particular type of perception, one that involves scientific constraints. I think you are confusing empiricism with the whole "tree falling in the woods" business. Empiricism does not require that someone see the tree, it merely recognizes that everything we can say about a falling tree is given meaning via our perceptions. We don't need to be there for that particular tree to recognize this fact.

If for you perception is axiomatic and irreductable to the behavior of big systems of QM 'particles', then you prefer a different flavor of CI (it is even not called CI):

consciousness causes the collapse

Empiricism views "causation" as a rationalistic notion (see Hume). Hence, the concept belongs in the toolkit of how we try to understand our perceptions. But collapse is what we perceive-- it does not require a cause. Those who imagine a physical causal link between consciousness and collapse are imagining a kind of "god's eye view" of the process. CI, and empiricism, rejects that such a perspective is physically meaningful, it is more a form of imagination.

let me ask you a question, can, in pronciple, perceptor be described as QM system?

That depends on what you mean by "be described." The empiricist translates that your question into "can we gain understanding and predictive power over the perceptor/environment relationship by treating the perceptor as a QM system? And the empiricist answer is the skeptical one: I don't know, do you have any evidence that such power is obtainable that way? But of course, that's quite a different question from what I think you mean-- is the perceptor really a QM system! The empiricist answer to that is clear: no, nor is the perceptor really anything but what it is-- a perceptor. (Here even CI may dabble in rationalism, some of the descriptions it makes tend to overstress classical thinking rather than just staying rigidly attached to the observations.)

 

[Ken G]

But I think sometimes "corrupted bets" are exactly what explains interactions. The fact that two systems are differently informed, leads them to contradicting expectations and interactions that serves to shift the "priors". Thus by on purpose "misinforming" two parties you can predict an interaction (conflict).

But I think the big difference is that "misinformation" is really happening due to physics. There simply is not notion of "correct information", no more than there exists a measure of right and wrong that humanity can agree upon because everyone always over-weights their own benefits. That's nothing wrong, it's just how things de facto work.

I see a lot of value in this view. To me, the key is the recognition that "our fingerprints" are all over the physics we do. That has in the past been viewed as some kind of a problem to be minimized, like we would minimize uncertainty in observations. But QM brings us into contact with what you get after you have minimized uncertainty, so it's time to stop seeing uncertainty (or "corrupted information") as something bad, and start seeing it as something fundamental.