Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Relational QM

  1. Aug 8, 2006 #1
    Hi,

    recently, I have read the paper of Marchildon, Why Should we Interpret Quantum Mechanics?, who raise a question about the Rovelli interpretation. This encourage me to finally read the Rovelli's paper, but I don't understand what really is the relationnal interpretation. I stopped in the begining of the paper, and I can't continue until I understand what some of his idea means.

    Rovelli argue that one observer make a measurement, while the other don't. It's like one is outside a box with the Schrödinger's cat inside. Rovelli say that the man outside and the cat inside described the same sequence of events. For the cat, there is a collapse, and for the man outside, there is no collapse. So they give different account of the same sequence of events. The way Rovelli explain this make me feel that he consider an objective reality, that is, an objective sequence of events independant of any observer, described be different observer.

    My own question is first: How can Rovelli say that "the man outside" and "the cat inside" describe the "same sequence of events"? The first event is the one where the cat and the quantum-gun with their respective states were put in the box. I aggree that this event is "seen" by the man and the cat. After, these two agree that the boxe is closed. But now, the man outside witness only a black box (a long sequence of events : "the man see a black box"). But in the box, the cat will see the quantum-gun fire, while the man outside will still observe the boring black box. How can we say that the cat and the man "describe the same sequence of events"?

    What is an event for Rovelli?

    The question of Marchildon is : "To someone who believes there is a state of affairs of some sort behind the description, the difference between O’s and P’s point of view means that one of them is mistaken. To put it differently, the problem with the argument is that expressions like “standard quantum mechanics” or “conventional quantum-mechanical description” are ambiguous. They can refer either (i) to strict unitary Schrodinger evolution, or (ii) in the manner of von Neumann, to Schrodinger evolution and collapse. Once a precise definition is agreed upon, either description (1) or description (2) (but not both) is correct." ("O" is what I called the cat, "P" is what I called the man; "(1)" describe the collapse while "(2)" describe the unitary evolution).

    I need to understand these points before to continue my lecture of Rovelli's paper.

    Thanks,

    Tipi
     
    Last edited: Aug 8, 2006
  2. jcsd
  3. Aug 21, 2006 #2

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    This has been discussed here:

    https://www.physicsforums.com/showthread.php?t=117286

    My personal view on the Rovelli interpretation is that it does 2 things:
    - it gives an "informational" account of quantum theory (in other words, quantum theory describes "information you have" and not "what's happening")
    - it denies a priori the existence of a unique ontology (it doesn't attempt to describe a "nature out there", but only what "an observer knows" = a list of observations made by that observer).

    As such, I call this vision "solipsist", in that the observer's observations are a purely subjective phenomenon, limited to its own subjective world - but proponents of the Rovelli interpretation didn't agree with that. Take your pick.

    I also made the statement, and tried to defend it, that the Rovelli vision is the "single observer" vision of a many-worlds interpretation (MWI). In a MWI, there IS an objective reality, but observers experience only part of it, and the parts they see do not need to be in agreement (however, the parts they see others see, are in agreement of course). The "equivalence class" of observers in agreement (and in interaction) are of course a branch in the MWI, and observers in different branches cannot interact, so their disagreement is not observable. I saw Rovelli's argument as the "single branch" version of MWI, and IMO, the "error" Rovelli makes is the following:

    He identifies: "everybody sees the same elephant" with "everybody sees everybody agreeing on seeing the same elephant".
    Observationally, only the last statement is verifiable (and is what happens within one branch in MWI). Nevertheless, the last statement doesn't exclude the possibility that there are classes of non-interaction observers which do NOT see the same elephant - as is the core idea in MWI.
    Nevertheless, Rovelli makes the jump from the second, to the first statement, and calls that "objective reality".
    Clearly, if we were to do that in MWI, we'd have a problem, because we wouldn't know WHICH branch to take as defining "objective reality". However, by using the words "objective reality" Rovelli (IMO) tricks us in thinking that suddenly he solved the parallel branch problem, but his definition of "objective reality" is erroneous (given that there are several, incompatible, versions of it).

    Formally, however, there's nothing wrong with Rovelli's viewpoint, as long as you tell yourself that it is happening from the viewpoint of a single observer (and hence in a single branch). But it is not innovative either: it is what was already understood in MWI since long. The hype seems to come from the apparent resolution of the difficulty of defining an objective reality which does not consist of parallel branches, and this is, IMO, erroneous.

    Nevertheless, I think the papers on the RI are an interesting read, and they stimulate reflection on the subject. However, I didn't see any magical solution of a problem in it:
    Or you have to accept the craziness of MWI and the parallel branches,
    or you have to confine yourself to the description of your purely subjective world denying any objectively existing world shared with other observers (you just have a list of "observational experiences" in your subjective world - something I call solipsism). The way Rovelli introduces "objectivity" is, IMO, totally erroneous.
     
  4. Aug 21, 2006 #3

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    Patrick, I think it's a little unfair of you to respond to a request for clarification of RQM with a comercial for your own position, including the "solipsist" canard which is, ahem, particular to you alone as far as I know.

    ALso your interpretation of the "everybody sees the same elephant" is based on your own assumptions about "reality" which lead in fact to your adoption of MWI; to caricature, I think not too unfairly, you want a real, don't-need-to-observe-it reality there in wave function land, but behaving quantally, hence MWI. In a truly realtional world view, if the observers agree, and speak truly qaccording to their own experiences, then you can't ask for more objectivity than that, and in relational-think, all observers are equal, the reverse of solipsism, as everybody but you defines it.
     
  5. Aug 21, 2006 #4
    Dear vanesch,

    thanks for your answer.

    So, if I understand well, Marchildon is right. As you said
    Wich is the point of Marchildon. It is like if the Schrödinger cat reality1 is "I'm alive" when the observer outside the box reality2 is "the cat is half-alive/half-death". How can we concile reality1 with reality2?

    I think your right when you say
    Tipi
     
  6. Aug 21, 2006 #5
    Dear selfAdjoint,

    thanks for your help.
    A little question : what we can do in the cat case? In other word, what we do when two observers dont agree? (One see the cat alive when other see it death-alive). What is the reality then?

    Thanks,

    Tipi
     
  7. Aug 21, 2006 #6
    I'm reading this :

    https://www.physicsforums.com/showthread.php?t=117286

    and I see some answer. Continue reading...

    Tipi
     
  8. Aug 21, 2006 #7

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Well, reading the article quoted by the OP, I'm seemingly not alone :smile:, and you have to admit that I *first* pointed to the long thread where we had this discussion and then explicitly stated *my own opinion on the matter*

    With all respect, I think you didn't understand my criticism of Rovelli then. Also, you don't seem to understand my position on MWI. As I said here several times, my position is the following: *given* standard quantum theory, MWI seems to me to be the interpretation that is closest to the formalism, avoids most of the formal difficulties, and allows one to postulate an ontology that goes with it. That doesn't mean that I think that it is ultimately true, and - contrary to what some might think - I'm absolutely not a religious MWI adept. But face it, when a theory gives you a wavefunction and an evolution equation, what's more natural than to say then that this theory (and the toy world it describes) takes this wavefunction as an element of description of (toy) nature, and the evolution equation as the "law of toy-nature" that goes with it ?

    Whether "real" nature is like this "toy nature" or not is, IMO, absolutely not established one way or another and second, is a totally unsolvable issue (unless empirical falsification of QM), in which we're back to square one.

    However, the critique you have about my elephant business is Rovelli's OWN argument, which YOU cited here (long ago). Rovelli himself uses the wavefunction of the two "observer versions" to show that each version will be in agreement with the observations made by another observer in the same term (branch).

    Here's your post with the quote:
    https://www.physicsforums.com/showpost.php?p=1004474&postcount=102

    So the only thing that is established there, and which serves as an adage to say that "everybody sees the same elephant", is that observers in the same branch see the same elephant, and see all the other observers they can see and talk to agree upon the same elephant.

    But that doesn't exclude the case where there are several disjunct classes of observers with different elephants in each class, and it is even the case that was illustrated there. The fact that there are many means then that these DIFFERENT consistent sets of observation cannot be used to define ONE SINGLE objective reality. THIS is the joke: there's a DIFFERENT objective reality for each class of observers if you take that seriously, or there's a DIFFERENT objective reality for each observer in all those branches. Now, observer-dependent objective reality is in my book an oxymoron: it is in fact *subjective reality*. When all that exists, is *subjective reality*, then that is then solipsism. When what exists are *multiple subjective realities*, then that's MWI.

    Of course things don't have to be like that in reality - only in quantum theory when pushed far enough and taken seriously enough, way beyond what is empirically established - that's what I mean with the toy world of quantum theory (like there is the toy world of Newtonian mechanics, and the toy world of general relativity).

    But given that Rovelli USES this quantum world with superpositions of observers (as in the part you quoted yourself) to show that observers only see eachother agree, he should take that seriously enough to see also that there are separate classes of "agreeing observers" and hence he cannot use that to define "objective reality". This is not due to my penchant for MWI, of which I explained for the n-th time that it is only relative, but rather due to the argument Rovelli uses himself: he cannot use an expression describing SEVERAL SETS of mutually agreeing observers to make his point, and then define "objective reality" to be what observers agree mutually upon.

    The argument that the wavefunction just describes *possibitilities* doesn't hold: what's (still) possible for one, is not possible anymore for another. If there is a single objective reality then, only one "possiblity" is "real" and all the rest is ignorance description, which should then be suited for a probabilistic or deterministic hidden variable description, and we're back to EPR and non-locality and all that, in which Rovelli doesn't bring any new light. That's nevertheless the real problem to solve.
     
  9. Aug 21, 2006 #8

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    That's what's called objective reality, right ? What's there, independent of whether it is "observed" because "observation" is "also there" and just part of it. From the moment you "need observation" you need also an *outside* observer of course, because who/what's otherwise observing the observer (in order to be there in the first place) ? Hence an entity outside of the scope of the theory is needed, namely "the observer" (as in the good old days).

    Well, RQM gives you a description of *your observations* and *your observations only*, including *your observations of things you call "other observers"*. But from the start, it's only a list of your observations. If nothing else exists but your observations, which do not correspond to anything else but being observed by you, then that's solipsism - the denial of objective reality existing outside of your observational impressions. Now CALLING your list of observations "objective reality" while emphasizing that they do not correspond to anything, is just a game of words. It still means that there is an entity, outside of the theory (see above), which experiences a list of observations, to which nothing corresponds outside of said observations.

    From the moment that there is an objective reality, it is of course independent of observation, said observation being simply part of it, and described by it. THIS is what goes wrong with RQM when claiming to describe objective reality.
     
  10. Aug 21, 2006 #9
    Hello guys,

    I have read a lot in the other thread. But I still have questions.

    First, the MWI interpretation need the other branches to explain interference. If Rovelli interpretation don't keep the other branches, but only the one of the observer, what appends with interference phenomena?

    Another thing with the cat. Suppose there is a camera inside the box with the cat, and observer O1 watching it on tv. There is also observer O2 watching the outside of the boring blackbox. Now Rovelli say that O1 and O2 "describe the same sequence of events". But, what is a sequence of events for Rovelli? Just try to find an answer to this question... For me, "the same sequence of events" means that there is an objective reality out there. No? If yes, O1 and O2 can not be all right, one must be wrong about THE reality.

    In fact (if I understand what is said in the other thread), O1 and O2 know different things, and MQ describe what they know, right? So MQ is about what we know, and what we know is our own reality, right?

    But, this night, I made some dreams. And I know some things about it. There were a phantom, and a flying car. This knowledge didnt seem to respect the rules of QM. Why? QM is about what I know, right? What is the difference between what I acquired from dreams and what I acquired from "reality"? If QM is about information, what is the difference between dreams and reality?

    The problem of defining QM as a theory about what I know is that it suppose something out there about what I have information. The theory postulate (implicitly) something it can not describe. So it is incomplete, no?

    Tipi
     
    Last edited: Aug 21, 2006
  11. Aug 21, 2006 #10

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    The interference is still the quantum effect. Relational QM doesn't seek to 'explain" quantum interactions in terms of something else, but just to give a coherent account of quantum physics happening in the universe. See those are really two different questions, but for most physicsists, while everything else might be up for grabs, quantumness of nature is not.

    If you have an observer in the box with the cat, you have a classical system. Schroedinger's point with the cat is, if quantumness is observer dependent, and you don't know what's going on with the cat, then it's in a superposition of alive and dead. Superposition doesn't mean "both alive and dead" nor does it mean "either alive or dead". It's a new quantum idea, that breaks th excluded middle of classical logic.

    The Rovelli idea of same sequence of events turns on a quantum interaction. One system (an "observer") interacts with another (an "observed") in a completely quantum way. And each system changes state as a result of this basic interaction.

    The observer's states are called pointer states (think of a needle on a dial) and the observed change of state is what is often called the collapse of the wave function. But note that the observer was in a superposition of pointer states before the interaction (before the "measurement" if you must, but I'd rather keep people out of it). And after the interaction each system retains a trace of what its resultant state was. Later on they can compare notes and if they do, their notes - the comparison of the traces they have retained - will agree. And the authors calculate out in the paper that even if you extend this to a case of entangled particles and even if a third party comes and observes the goings-on, still everybody's stories will agree (Vanesch makes what I think is a desparate attempt to undermine this by distinguishing "Their Stories" from "What really Happened". But this is just literature; in Relational physics the stories ARE what really happened). It's as if he refused to give up a preferred frame in relativity.

    And so Rovelli asserts, and I accept: That's the only kind of objective reality there is. Vanesch calls it language games and I call his MWI stuff fantasy land. Take your choice or come up with something better!

    Yes, but it's a shared reality if our stories agree. Do you have any other source of confidence than the fact that you and other people agree on what you all see?

    IMHO this is what you get when you try to build "consciousness" into "quantumness". The former is an ill-defined concept whose boundaries are all over the place, the latter is sharply defined (derivable from axioms, permitting exact computation of results and prediction of experiments). Category error.

    See above. The consciousness angle is in my opinion not just incomplete but incoherent.
     
    Last edited: Aug 21, 2006
  12. Aug 22, 2006 #11

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    What you write here is exactly my sentiment too:

    Yup.


    Exactly, that's why I call that vision "solipsist" in the sense that it only talks about a subjective set of observations by a subject, and refuses to consider that there is anything definite behind it (what's usually called objective reality) apart from its very observation by a subject, and which can be different according to different observers. Indeed, as you say, if objective reality is different for different observers, then that's the end of it, because objective reality is by definition what's independent of one's knowledge and observation. Because if all there is, is observation, then there are not even "other observers", but only you, having observations of other observers.

    Now, I think Rovelli wants us to consider the following situation:
    There are "islands" of genuine observers which interact by exchanging data. What they have exchanged and agreed upon is then called "objective reality". However, non-interacting islands of observers can then have different "objective realities" and one island of observers is not supposed to consider the possible alternatives of other islands of observers (because that would then lead to a hidden-variable theory, and hence run into troubles with Bell). So the only way to get out of that is by saying that what is objectively true for one island of observers is non-existant for the other, until they meet. This is some form of group-solipsism IMO, because there's no real use for having now "groups" of observers. If things your group hasn't seen, do not exist, until you see them, then that's just the same as doing that just on your own, and we're back to the same solipsist vision I think this is about, and for which I'm criticised by SA.

    Indeed, and the "completion" would simply be to add the other branches, in which case we're back to MWI.

    But, again, I'll get SelfAdjoint on my back for this, because this is *my* personal way of reading this stuff, and he sees this totally differently.

    Again, the only point I attack in Rovelli's viewpoint is his funny definition of "objective reality". All the rest is all right (but IMO not any different than what would your average MWIer do). But the whole sales argument for RQM is exactly this resolution of the contrast between "objective reality", subjective experience, and unitarity: namely that it does what cannot be done: postulate an objective reality which corresponds to our subjective experiences (naive realism, as in classical physics), while keeping the Schroedinger equation as a universal law.
     
  13. Aug 22, 2006 #12

    vanesch

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm not really on purpose distinguishing "what really happened" and "their stories". The problem is that, when there are two groups of observers having DIFFERENT stories, it is difficult to assign them to a single objective reality. This is what doesn't fit in Rovelli's story, and I haven't yet seen any sound explanation of it.

    In an EPR situation, we have the "Alice" group of observers, and the "Bob" group of observers. Now, because of their spacelike separation, they cannot (yet) interact (but of course the Alice observers interact amongst them, and the Bob observers do the same).
    So now we have an "objective reality according to the Alice group", and an objective reality according to the Bob group". So far, so good.
    However, this would mean that "objective reality" is the union of both, right ?
    And this is where things IMO go wrong. Rovelli says that you can't say that, that "objective reality for the Bob group" doesn't make any sense for the Alice group and vice versa. That's IMO an error, because it makes objective reality observer(group?) dependent, which it shouldn't, by definition.

    Now, Rovelli uses some rethoric to make one think that this is like coordinate transformations in relativity: of course different (groups of) observers have different descriptions of events. So the Alice group has one description (its own "objective reality by mutual agreement" plus Bob's group in a description of superposition), and Bob's group has another.
    The point is, however, that there is no transformation rule between them!
    In relativity, observer O works with (x,y,z,t) and O' works with (x',y',z',t'), which are of course different descriptions, but O KNOWS HOW TO TRANSFORM his (x,y,z,t) in O''s (x',y',z',t') and vice versa. It are different, but EQUIVALENT descriptions. However, this is not true for the quantum case. Alice's description of Bob's situation as a superposition is NOT a different but equivalent description of Bob's own view on his reality. Alice cannot find out what was Bob's situation (as O could calculate x',y',z',t'). You can now say that this is because Alice is LACKING INFORMATION. True, but that doesn't work. Because if Alice were only lacking information, then she could give a probability distribution of Bob's different *possible* situations, with the assumption that one of those IS Bob's situation. But that's nothing else but a hidden-variable description, which runs into troubles. So no, Alice's description of Bob's situation is NOT a different but equivalent description as the one seen by Bob, and is NOT a different but equivalent description with lacking information either.

    So no, I do not agree with this analogy with coordinate changes in relativity. Alice's account of Bob's reality is fundamentally different from Bob's account of his reality, contrary to any change of reference frame. I'd say, nice try.
     
  14. Aug 22, 2006 #13

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    Correct me if I'm wrong, but isn't what you're talking about here what others call "The collapse of the wave function" and I call an interaction? And if so, isn't this unobservable? Only the resulting states are observable. Now these states are the real-world traces of the interactions that produced them, and they, not the interactions, are what "everybody sees the same". Is it not so?
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Relational QM
Loading...