Exploring Relational Quantum Mechanics: An Alternative Interpretation

[Farsight]

I came across this on another thread, regarding "Relational Quantum Mechanics".

http://plato.stanford.edu/entries/qm-relational/

"Relational quantum mechanics is an interpretation of quantum theory which discards the notions of absolute state of a system, absolute value of its physical quantities, or absolute event. The theory describes only the way systems affect each other in the course of physical interactions. State and physical quantities refer always to the interaction, or the relation, between two systems."

It struck me as something that offered a less mystical interpretation wherein the observer could be another system and reality could only be the interaction.

Vanesh, is this hogwash or does it offer possibilities?

 

[selfAdjoint]

I came across this on another thread, regarding "Relational Quantum Mechanics".

http://plato.stanford.edu/entries/qm-relational/

"Relational quantum mechanics is an interpretation of quantum theory which discards the notions of absolute state of a system, absolute value of its physical quantities, or absolute event. The theory describes only the way systems affect each other in the course of physical interactions. State and physical quantities refer always to the interaction, or the relation, between two systems."

It struck me as something that offered a less mystical interpretation wherein the observer could be another system and reality could only be the interaction.

Vanesh, is this hogwash or does it offer possibilities?

Check on the Beyond the Standard Model board, the thread "Everyone sees the same Elephant (or so says Rovelli)" It is about Relational QM and Vanesch has contributed to it. He is not a believer in it, but I am attracted to it for the same reasons you cite. The thread was triggered by a paper by Matteo Smerlak and Carlo Rovelli (one of the big names in Quantum Gravity research), entitled Relational EPR, quant-ph/0604064 on the arxiv.

In a nutshell, they regard interacting quantum systems as real, and the basic makeup of reality, and discard the classical observer, which seems to be the block in every quantum physicist's mind. For the authors, "observation" is just a specialized task for a prepared quantum system, just another interaction. Eigenvalues are not objective reality but only relative to the other interacting system, and they work out how all of this comes out right (that is, quantumly) in the EPR case.

 

[vanesch]

Check on the Beyond the Standard Model board, the thread "Everyone sees the same Elephant (or so says Rovelli)" It is about Relational QM and Vanesch has contributed to it. He is not a believer in it, but I am attracted to it for the same reasons you cite. The thread was triggered by a paper by Matteo Smerlak and Carlo Rovelli (one of the big names in Quantum Gravity research), entitled Relational EPR, quant-ph/0604064 on the arxiv.

Well, I think there have been misunderstandings about my claims in that thread: I DO like (a part of) RQM. What I don't like is some wordings about its possibilities. The way I understand RQM is as the "solipsist" or "single" observer version of MWI. By this, I don't mean that RQM cannot handle multiple observers, but RQM assigns a different "reality" to each observer (just as MWI does). The only difference I really see is that RQM denies the existence of an objective reality, while MWI does give that (although it sounds crazy).
However, in the articles discussed in the thread SA mentionned, it is my impression that there is a big verbal exercise to burry the fact that there's a denial of objective reality (nevertheless, that's about the contents of "hypothesis 2" in one of the two papers).

So, my idea is that RQM is fine from a single observer PoV, where the "subjective" reality of an observer is described (although this is called an "objective observer-dependent reality" or something of the kind).
What's real for A has "no element of reality" for B.
Now, this is exactly also the PoV of MWI. What MWI adds, is also an objective reality (in casu the overall wavefunction), which then explain the different subjective realities of different observers.

So, I'm not against RQM "as a first mental exercise in MWI" say :-)
However, RQM as a "complete vision" hurts deeply my Platonic concept of the world, with a denial of the existence of an objective reality.
Nevertheless, if it were clearly stated that THIS is the fundamental idea (there's no objective reality, only one's subjective reality counts), I wouldn't mind. But the word "objective" is sprawled all over the place in order to hide this subjective character - and this, I find regrettable, because it lures in people who might at first think that this denial isn't so deep, but that it is only a matter of "switching coordinate systems" or something.

The difference is a bit as in relativity. In relativity, there is an "objective reality" which is described by the spacetime manifold (or the equivalence class of isomorphic manifolds for nitpickers) and there are of course the different coordinate representations (say, observers) which assign different numbers to each event.
Now, of course a coordinate description is dependent on an observer, but things like the statement "the explosion of the firecracker took place at coordinates (x,y,z,t) for observer O1" is an objectively true statement, which can be acknowledged by all observers. That is, observer O2 will have of course (X,Y,Z,T) as coordinates for that firecracker, but, knowing the coordinate transformation to O1, he'll be able to conceive that O1 has coordinates (x,y,z,t) for the event. So observer O2 has no difficulties taking the statement that O1 has coordinates (x,y,z,t) for the event as an objectively true (or a false) statement.

This is slightly as in MWI: there's a description of an objective reality (in casu the wavefunction, or better yet, the unitary structure), and then there are the different states corresponding to different observers.
There too, one can say that for observer state O1, who observed a pink elephant, and observer state P1, entangled with observer state O1, BOTH can be in agreement that observer state O1 saw a pink elephant.
They can also agree that observer state O2 has seen a blue elephant, although they'll never interact with it. But the objectively true statement that O1 observed a pink elephant is clear for everybody.

In RQM, however, things are different. Here, in O1's reality, O1 DID see a pink elephant, but in P1's reality, O1 DIDN'T SEE a pink elephant until P1 interacted with O1. P1's description of O1 is still that O1 is in a superposition of having seen a pink or a blue elephant, while this is NOT the case for O1 itself. This is not only because P1 has no information yet: P1 cannot assume that O1 did see one OR did see the other - he still needs to entertain O1 in a quantum superposition, while O1 did clearly see one of both. So the truth value of the statement "O1 saw a pink elephant" is dependent on who's saying so. There is no "objective truth value" to be assigned, EVEN to statements regarding the relationship between an observer and a system.
This is necessary in RQM in order to be able to deny Bell's theorem: O1 "didn't have a result yet" from P1's PoV (while from O1's PoV, O1 DID have a result of course).

This can only be the case if we're talking about two different realities: the one related to O1 and the one related to P1. "Realities" in which IDENTICAL STATEMENTS have different truth values.

So RQM is a web of subjective realities with no objective reality to link them, while MWI is an objective reality from which subjective realities can be deduced.

And RQM needs to cheat a bit to make the different subjective realities coincide (see my post about that in the other thread).

In a nutshell, they regard interacting quantum systems as real, and the basic makeup of reality, and discard the classical observer, which seems to be the block in every quantum physicist's mind. For the authors, "observation" is just a specialized task for a prepared quantum system, just another interaction. Eigenvalues are not objective reality but only relative to the other interacting system, and they work out how all of this comes out right (that is, quantumly) in the EPR case.

This is exactly what can be said about MWI in fact. Only, in RQM, the term "reality" has become observer-dependent, and hence subjective, with a denial of the existence of an objective, observer-independent reality (or at least, the possibility to describe it logically).

So, RQM as the "single observer/solipsist" viewpoint of MWI, I'm all in favor. RQM as a complete vision, I think it is even more bizarre than MWI.

In other words, I regret a bit the verbal marketing around RQM: there's no miracle, and it is a bit sold as if there were.

 

[Farsight]

Thanks selfAdjoint, Vanesh. I'll read up some more taking careful note of your points.

By the way, do you know if there's some reason why it's called

Relational Quantum Mechanics

rather than...

Quantum Relativity!



Sorry to hijack the thread Dyamios. I'll pipe down for a while.

 

[selfAdjoint]

I hate to harp on an OT subject, but this

However, in the articles discussed in the thread SA mentionned, it is my impression that there is a big verbal exercise to burry the fact that there's a denial of objective reality (nevertheless, that's about the contents of "hypothesis 2" in one of the two papers).
So, my idea is that RQM is fine from a single observer PoV, where the "subjective" reality of an observer is described (although this is called an "objective observer-dependent reality" or something of the kind).
What's real for A has "no element of reality" for B.

seems wrong to me.

A value of a quantum number of an observed system is only real for the observing system interacting to produce that value from the many that could have resulted. Another system that is not in interaction with the observed one does not see that value and only knows it happened by report. Well, duh! The point of the RQM approach is that it's reaching to say that something that someone else (spacelike to you) sees and you don't is "really real" for you. If I'm remembering correctly, some Native American languages have syntax that makes this kind of distinction explicit.

 

[vanesch]

A value of a quantum number of an observed system is only real for the observing system interacting to produce that value from the many that could have resulted. Another system that is not in interaction with the observed one does not see that value and only knows it happened by report. Well, duh! The point of the RQM approach is that it's reaching to say that something that someone else (spacelike to you) sees and you don't is "really real" for you.

This is the kind of "jumping back and forth" between objective and subjective in the RQM papers which I find misleading.

The point is that there are two different kinds of statements:

S1: "A happened"

A being a specific description of an event, such as: Joe read "5V" on the voltmeter.

If there is an objective reality, then there's a truth value to be assigned to S1. It's either true or false, and this is (because of "objective") unrelated to any "truth observer" say. Of course A is related to Joe, but we'll be talking about the same Joe.

S2: B KNOWS that "A happened"

If B is "Joe" then we can assume that B does know that A happened.
If B is "Jack" then it depends on whether Joe told Jack, or whether Jack could deduce it from other knowledge.

So we have two different statements:
S2a: Joe KNOWS that "A happened"
S2b: Jack KNOWS that "A happened"

In other words, there's a difference between KNOWING the truth value of a statement, and the statement HAVING a truth value.

S2a and S2b being different statements, there's no problem in assigning different truth values to them, but their truth value has nothing to do with the truth value of S1 a priori.

However, in the articles concerning RQM, there's a lot of verbal exercise to create a confusion between the 3 different statements S1, S2a and S2b.

At some point, it is said that S2a and S2b 's truth values are to be the S1 truth value: that's how I read that "Although for Joe, A happened, for Jack, there is no element of reality assigned to "A happened" " (this is a paraphrasing of a litteral quote in one of the articles, which I discussed in the other thread).

So it seems that in RQM, S2a and/or S2b 's truth value is identified with S1's truth value (under the idea "reality = information"). But then of course THERE IS NO OBJECTIVE REALITY. That's my whole point. There are only different subjective realities. So this is what I call (maybe the term is not entirely exact, but you get the gist of it) a "solipsist" view on things, because "reality" only has specific truth values according to one "information possessor's" information.

And if this would be stated bluntly, it would have much less success. If you would say that whether Joe is dead or alive depends on whether *I* know about it, RQM's weirdness would be more open to the light of the day (and probably attract just as much critique as MWI for its weirdness). For example, it would then be sufficient for not having to broadcast information about, say, famine in Africa, for there not to be any famine in Africa in most realities (because most observers wouldn't KNOW about it).

So there's a verbal exercise to make you believe that nevertheless there is a kind of objective reality "on what all observers who interact agree upon". But the point is of course that the real difficulty in QM is when the observers CANNOT interact, because THAT's where Bell's theorem matters.
So if two groups of observers cannot interact, they have DIFFERENT realities (meaning: assigning different truth values to identical statements concerning observable and observed phenomena) according to RQM, and these two DIFFERENT realities are "one and the same objective reality". This is of course running into deep logical problems, but the statement is strictly necessary because otherwise Bell's theorem becomes applicable.

Nevertheless, this situation is VERY CLOSE to the MWI situation: there too, there are TWO DIFFERENT subjective realities, and one objective reality: but indeed, the two subjective realities are simply *two different terms* in the objective wavefunction, and it is the non-interaction between the two terms that allows their coexistence without observational contradiction.
However, claiming that two different subjective realities constitute the SAME objective reality is logically flawed IMO.

 

[selfAdjoint]

If there is an objective reality, then there's a truth value to be assigned to S1. It's either true or false, and this is (because of "objective") unrelated to any "truth observer" say. Of course A is related to Joe, but we'll be talking about the same Joe.

And that's your definition of "objective reality" and it involves appealing to a hypothetical privileged classical observer to support it; an observer who sees all, knows all, and isn't bound by either relativiy or quantum mechanics. But I deny that observer can be appealed to even theoretically; Einstein banished her from kinematics, and we should get rid of her in synamics too.

There isn't any classical world; there are only quantum interactions. Quantum interactions ("wave reductions") only provide local information. Comparison of spacelike related quantum interactions is invalid from the git-go.

 

[Careful]

**And that's your definition of "objective reality" and it involves appealing to a hypothetical privileged classical observer to support it; an observer who sees all, knows all, and isn't bound by either relativiy or quantum mechanics. **

That is pure nonsense. Objective reality simply means that any interaction in spacetime leads to a unique, *determined* yes or no answer on any (local) physical question you want to ask. Too bad that this does not agree with QM (which is a failure of the latter IMO), but it does so with relativity.

**But I deny that observer can be appealed to even theoretically; Einstein banished her from kinematics, and we should get rid of her in synamics too. **

Huh, the dynamics of GR is clearly observer independent.

**There isn't any classical world; there are only quantum interactions. Quantum interactions ("wave reductions") only provide local information. Comparison of spacelike related quantum interactions is invalid from the git-go.**


Quantum interactions lead to no predictions unless you involve this consciousness mumbo jumbo (which is classical).

 

[vanesch]

And that's your definition of "objective reality" and it involves appealing to a hypothetical privileged classical observer to support it; an observer who sees all, knows all, and isn't bound by either relativiy or quantum mechanics.

No, not at all. I do not need ANY observer to talk about "objective reality" because that's the very definition of objective reality: it is the assignment of true and false values to statements (statements, which can, or cannot, involve observers).
You are right of course that this is hypothetical: the existence of an objective reality is indeed (and will always remain) a hypothesis ; this is exactly the unfalsifiability of solipsism.
What remains, when you deny it, is indeed just your subjective reality (= consciousness if you want to) and then we have a form of solipsism.

As I said in the other thread, what is "classical" is not the hypothetical existence of an objective reality, but rather the existence of definite results of measurement, simply because to a quantum system, the superposition principle applies, and hence also the superposition of measurement outcomes. I said here before that not arriving at a specific result is not a nasty aspect of the quantum FORMALISM, it is PUT IN FROM THE VERY START by imposing the superposition principle.
So the only CLASSICAL idea is that a measurement has a definite result: objective or subjective. If you *really* apply quantum theory all the way, there ARE NO MEASUREMENT RESULTS (as Careful points out). You always need some element of classicity to even be able to TALK about a definite measurement outcome because the superposition principle is then denied.

There isn't any classical world; there are only quantum interactions. Quantum interactions ("wave reductions") only provide local information. Comparison of spacelike related quantum interactions is invalid from the git-go.

Quantum interactions provide no "wave reductions", as per the superposition principle. Quantum interactions only provide entanglements, that is, "lumps of states that more or less belong together", and if the environmental interaction is helping, that will STAY together (that's decoherence). At no point, there is a wave reduction between the interaction of two QUANTUM systems: wave reduction occurs when the quantum-classical transition is involved.

I agree with you however that we could object to comparing spacelike QUANTUM interactions. However, denying the comparison of (by definition classical) MEASUREMENT OUTCOMES is begging the question. If B HAD a measurement result and KEEPS that result until he meets A, and A the same, and at the meeting point this is declared to be "real" then it is a stretch to say that the reality of this result wasn't "real enough" to be valid for Bell's theorem, no ? This is a game of words.

 

[vanesch]

This is a thread I split off from the original one (what decides the quantum state...), where an off-topic discussion on Relational Quantum Mechanics got mixed in, and which is in fact a continuation of a discussion on the "beyond the standard model" forum (everybody sees the same elephant).

I'll close this thread and invite further discussion, as SA points out, to that thread.