Interpreting QM without Schrödinger's equation

[Jarvis323]

Rovelli, in his recent paper, writes:

In his celebrated 1926 paper [1], Erwin Schr ̈odinger introduced the wave function ψ and computed the spectrum of hydrogen from first principles. This spectrum, however, had already been computed from first principles by Pauli four month earlier [2], using the theory that emerged from Werner Heisenberg’s 1925 breakthrough [3], based on the equation $$qp − pq = iℏ$$
with no reference to ψ. The theory we call “quantum mechanics”, in fact, had already evolved into its current full set of equations in the series of articles by Born, Jordan and Heisenberg himself [4, 5]. Dirac, equally inspired by Heisenberg’s breakthrough, got to the same structure independently in 1925, the year before Schr ̈odinger’s work, in a paper titled “The fundamental equations of quantum mechanics” [6]. (See [7, 8] for a detailed historical account.) Properly, the only Nobel Prize with the motivation “for the creation of quantum mechanics” was assigned to Heisenberg.

So, what did Schr ̈odinger achieve in 1926? With hindsight, he took a technical and a conceptual step. The technical step was to translate the unfamiliar algebraic language of quantum theory into a familiar one: differential equations. This brought the novel ethereal quantum theory down to the level of the average theoretical physicist. The conceptual step was to introduce the notion of “wave function”, which soon evolved into the general notion of “quantum state”, ψ, endowing it with ontological weight.

The Relational Interpretation of Quantum Mechanics, or RQM, is based on the idea that this conceptual step, which doesn’t add anything to the predictive power of the theory, was misleading: we are paying the price for the confusion it generated.

The mistaken idea is that the quantum state ψ represents the “actual stuff” described by quantum mechanics. This idea has pervaded later thinking about the theory, fostered by the toxic habit of introducing students to quantum theory in the form of Schr ̈odinger’s “wave mechanics”, thus betraying history, logic, and reasonableness.

https://arxiv.org/pdf/2109.09170.pdf

Do you think Rovelli is correct in implying the wave function is redundant and misleading? If we disregard the wave function's significance, beyond convenience, say we forgot it even existed, how would that change (or should it change) how we interpret QM? Would there be any point, or basis, for MWI without the wave function? Which interpretations become meaningless without the wave function?

 

[PeterDonis]

Do you think Rovelli is correct in implying the wave function is redundant and misleading?

If it is, then so is Heisenberg's matrix mechanics, since the two are mathematically equivalent. The same issue with interpreting the wave function as "the actual stuff" applies equally well to the state vectors in the Heisenberg formulation. And just as you can choose to adopt a non-realistic interpretation of state vectors, you can choose to adopt a non-realistic interpretation of wave functions.

Which interpretations become meaningless without the wave function?

None of them. If you don't use wave functions, you have to use something else that's mathematically equivalent, and you get the same predictions for all experimental results, and you get the same interpretational issues that you have now.

 

[WernerQH]

Do you think Rovelli is correct in implying the wave function is redundant and misleading?

I certainly agree with Rovelli that the wave function is misleading. It has led many people to believe that there is a one-to-one correspondence between the wave function and an individual system, whereas it is (at least for me) obvious that wave function can only be interpreted statistically. The wave function is but one piece of a bigger machinery: to arrive at measurable quantities, a $|\text{ket}\rangle$ always has to be combined with a $\langle\text{bra}|$ and a trace be taken. The bras and kets have the opposite time-dependence, or none at all in the Heisenberg picture. The fundamental time-symmetry of QM at the microscopic level is lost when you separate unitary evolution and "measurements".

I also concur with Rovelli that QM is a theory that predicts the probabilities of (or correlations between) events (or "facts"), but I'm not an adherent of the Relational Interpretation.

 

[A. Neumaier]

The same issue with interpreting the wave function as "the actual stuff" applies equally well to the state vectors in the Heisenberg formulation.

This is not correct. The formalisms are equivalent, but their interpretation is not.

In the Heisenberg picture the state is independent of time, hence cannot be interpreted as being "the actual stuff" at a given time.

 

[PeterDonis]

In the Heisenberg picture the state is independent of time

I wasn't talking about the Heisenberg picture vs. the Schrodinger picture. I was talking about the wave function/differential operator representation (Schrodinger equation) vs. the state vector/matrix representation (matrix mechanics). That is what Rovelli appears to be talking about in the paper. You can do the Schrodinger or Heisenberg picture in either representation.

 

[A. Neumaier]

It has led many people to believe that there is a one-to-one correspondence between the wave function and an individual system, whereas it is (at least for me) obvious that wave function can only be interpreted statistically.

Hw do you interpret the fact that the analysis of the high precision measurements of the anomalous magnetic moment of electrons with a Penning trap (discussed, d.g., here) describes the single electron on which the measurements are made via a density matrix?

 

[Fra]

The mistaken idea is that the quantum state ψ represents the “actual stuff” described by quantum mechanics. This idea has pervaded later thinking about the theory, fostered by the toxic habit of introducing students to quantum theory in the form of Schr ̈odinger’s “wave mechanics”, thus betraying history, logic, and reasonableness.
...
Do you think Rovelli is correct in implying the wave function is redundant and misleading?

I think Rovelli is totally right in that the quantum state does not represents the “actual stuff”, but instead some sort of "relation",

But I do not think Rovellis specific idea as presented in his RQM paper is satisfactory. I think his relational idea between observer starts out right but in my eyes at least he misses some important points. I think that just as much as it's not the "actual stuff", it can also not be a pure relation(without context) between stuff, because I think the relation itself must be encoded and inferred from a perspective. This is the part where I find Rovellis solution (to a problem I agree with) to be not satisfactory.

/Fredrik

 

[WernerQH]

Hw do you interpret the fact that the analysis of the high precision measurements of the anomalous magnetic moment of electrons with a Penning trap (discussed, d.g., here) describes the single electron on which the measurements are made via a density matrix?

What is the problem? The classical is remarkably similar to the quantum picture: the spin direction of an electron is strongly correlated with its direction at a later time. Do you have reason to believe that $\langle \sigma(t) \sigma(0) \rangle$ does not correctly describe this?

 

[A. Neumaier]

What is the problem? The classical is remarkably similar to the quantum picture: the spin direction of an electron is strongly correlated with its direction at a later time. Do you have reason to believe that $\langle \sigma(t) \sigma(0) \rangle$ does not correctly describe this?

There are convincing reasons that the equations of motion derived from it correctly describe the experiment.

The point is, however, that to make sense of this quantum expectation you need to ascribe to the single electron a time-dependent state .Thus the quantum expectation cannot be given a statistical interpretation! This conflicts with your belief expressed in post #3:

the wave function is misleading. It has led many people to believe that there is a one-to-one correspondence between the wave function and an individual system, whereas it is (at least for me) obvious that wave function can only be interpreted statistically.

 

[A. Neumaier]

I wasn't talking about the Heisenberg picture vs. the Schrodinger picture. I was talking about the wave function/differential operator representation (Schrodinger equation) vs. the state vector/matrix representation (matrix mechanics). That is what Rovelli appears to be talking about in the paper. You can do the Schrodinger or Heisenberg picture in either representation.

Matrix mechanics (dynamics of operators) is the Heisenberg picture - the name for the latter derives from Heisenberg's original matrix mechanics. Whereas wave mechanics (dynamics of states) is the Schrödinger picture - the name for the latter derives from Schrödinger 's original wave mechanics.

 

[WernerQH]

The point is, however, that to make sense of this quantum expectation you need to ascribe to the single electron a time-dependent state .Thus the quantum expectation cannot be given a statistical interpretation!

I think you have too much faith in the wave function and the special role of measurements. What you call "quantum expectations" are used a lot in condensed matter physics, and people there have no qualms calling them correlation functions, because they feel that there is something physically real that's correlated. Real even while the systems are not subjected to measurements in the precise sense as you define them.

Yes, I consider myself a realist (even an utterly naive realist), but not a strong realist in the sense used by Rovelli:

properties do not exist at all times: they are properties of events and the events happen at interactions.

I think Rovelli's relational interpretation is not radical enough. There is no need to assume that a "single electron" exists at every moment of time. The world line of an electron need not be continuous; it could be a series of events, of short-lived current pulses repeating on a time scale of zeptoseconds ($\hbar/mc^2$). This is consistent with all experiments that have been performed so far. What we call an electron is then just a special pattern of events in spacetime, and of course it's possible to apply statistics to such patterns (also if they represent just a single electron). QED (QFT) should be seen as a theory describing the correlations between microscopic events.

 

[PeterDonis]

Matrix mechanics (dynamics of operators) is the Heisenberg picture - the name for the latter derives from Heisenberg's original matrix mechanics. Whereas wave mechanics (dynamics of states) is the Schrödinger picture - the name for the latter derives from Schrödinger 's original wave mechanics.

This is not my understanding of the meaning of those two terms. My understanding is that "Schrodinger picture" means the time dependence is in the states and the operators stay the same, and "Heisenberg picture" means the time dependence is in the operators and the states say the same. (This is what you appeared to be saying in post #4.) Either of these pictures can be represented in both representations (differential operator vs. matrix). Rovelli's claim appears to be about the Schrodinger representation (differential operators), not the Schrodinger picture (time dependence in the states).

 

[A. Neumaier]

I think you have too much faith in the wave function and the special role of measurements. What you call "quantum expectations" are used a lot in condensed matter physics, and people there have no qualms calling them correlation functions, because they feel that there is something physically real that's correlated. Real even while the systems are not subjected to measurements in the precise sense as you define them.

Yes, I consider myself a realist (even an utterly naive realist), but not a strong realist in the sense used by Rovelli:

I think Rovelli's relational interpretation is not radical enough. There is no need to assume that a "single electron" exists at every moment of time. The world line of an electron need not be continuous;

But the theoretical explanation of the Penning trap assumes that the single electron has a continuous wave function. Wihout the latter no justification of the result...

 

[A. Neumaier]

Either of these pictures can be represented in both representations (differential operator vs. matrix).

But wave mechańics is the mechanics of the Schrödinger equation, hence assumes states and the Schrödinger picture. On the other hand, matrix mechanics is the mechanics of operators satisfying the canonical commutation relations, independent of whether the operators are thought of s infinite-dimensional matrices or differential operators. Indeed, Heisenberg's book on matrix mechanics contains Jordan's transformation theory (i.e., discrete basis leading to infinite-diemnsional matrices) and continuous bases (leading to differential operators) for the same operators, But everything is matrix mechanics!
From https://www.mathpages.com/home/kmath698/kmath698.htm

Heisenberg’s original matrix mechanics didn’t include the concept of a “state vector”, nor did it describe the results of a “measurement”. The time-dependence of the quantum variables in the Heisenberg picture represents purely unitary evolution.

Rovelli's claim appears to be about the Schrodinger representation (differential operators), not the Schrodinger picture (time dependence in the states).

Where does it appear so? Rovelli uses the terms in my sense:

Heisenberg lost the political battle against wave mechanics for a number of reasons: Differential equations are easier to work with than non-commutative algebras.
[...]
Heisenberg’s breakthrough is the idea of keeping the same equations as in the classical theory, but replacing commuting variables with non commuting ones

 

[PeterDonis]

But wave mechańics is the mechanics of the Schrödinger equation

Yes, the Schrodinger representation.

hence assumes states and the Schrödinger picture.

Not at all. You can write differential equations in the Heisenberg picture (time dependence in the operators). And you can write matrix equations in the Schrodinger picture (time dependence in the states).

Rovelli uses the terms in my sense:

No, he doesn't. He explicitly says differential operators vs. matrices. That's talking about which representation won, not which assumption about where the time dependence is won. If he were making a claim that the Schrodinger picture in the sense of always putting the time dependence in the states won, the claim would be false, since there has always been much theoretical work done in both pictures, and indeed there has also been much theoretical work done in the interaction picture, which is different from both the Schrodinger and the Heisenberg pictures. Are you claiming that Rovelli is so ignorant of the history of QM that he doesn't realize that people have continued to use the Heisenberg picture, and invented the interaction picture, after the "political battle" he refers to was long since finished?

 

[Fra]

The point is, however, that to make sense of this quantum expectation you need to ascribe to the single electron a time-dependent state .Thus the quantum expectation cannot be given a statistical interpretation! This conflicts with your belief expressed in post #3:

From the perspective of the "interacting and inferring agent" interpretation, there appears no such conflict as it is obvious that there is no such thing as to "repeat" a sequence from the past, so ANY sort of "statistics" implicity in the agents inferences necessarily take place at different times, but if it's possible to via cyclic process to encounter - from the agent - indistinguishable scenarios, then that seems to fully quality for a form of acquistion of statistics. But I agree that this is not really the samee as the idealized "ensemble", which is indeed fictional.

So an agent should be able to build a generalized statistics, from an interacting history, which conceptually seem similar to the tomographic sequenece Neumaier highlights. This is partly in line with my preferred view but with an inmportant difference, and that is that any agent has a saturation limit, where one is forced to compress and discard/radiate away information, which by construction is useless from the perspective of the agent, but not necessarily to other agents.

The "inference" implicit in a learning model (bayesian or other) is to be the sort of statistics. But the limiting information capacityy limits, implies that the concept of "frequentist statistics" of raw daata, is often not the most efficient one when it comes to storing as much as possible of USEFUL data.

But to describe this properly and clearly in terms of mathematics seems to require a grand reconstruction.

/Fredrik

 

[A. Neumaier]

But wave mechańics is the mechanics of the Schrödinger equation, hence assumes states and the Schrödinger picture. On the other hand, matrix mechanics is the mechanics of operators satisfying the canonical commutation relations, independent of whether the operators are thought of s infinite-dimensional matrices or differential operators.

Not at all. You can write differential equations in the Heisenberg picture (time dependence in the operators). And you can write matrix equations in the Schrodinger picture (time dependence in the states).

Rovelli uses the terms in my sense:

No, he doesn't. He explicitly says differential operators vs. matrices. That's talking about which representation won, not which assumption about where the time dependence is won.

Where does he say explicitly 'differential operators vs. matrices'? He says explicitly on p.1:

Differential equations are easier to work with than non-commutative algebras.

This is not 'differential operators vs. matrices' but Schrödinger picture (differential equation for the quantum state) vs. the Heisenberg picture (non-commutative algebras to describe transition probabilities),

Rovelli explicitly explains his goal on p.2:

A way to get clarity about quantum mechanics is to undo the conceptual mess raised by Schrödinger’s introduction of the “quantum state”. This is what the Relational Interpretation does.

This is geared against the Schrödinger picture in terms of a quantum state, Schrödinger's contribution to the setting. For Born and Heisenberg, a state was (before Schrödinger's work) always a state of fixed energy, not a state vector.

In the first paragraph of his Section 2, Rovelli refers to the linear Schrödinger evolution equation in abstract terms, thus painting the Schrödinger picture in full generality, without even mentioning a particular representation.

Are you claiming that Rovelli is so ignorant of the history of QM that he doesn't realize that people have continued to use the Heisenberg picture, and invented the interaction picture, after the "political battle" he refers to was long since finished?

I claimed neither. Don't try to put words into my mouth that I didn't use.

The interaction picture also uses quantum states, not noncommutative algebras. Noncommutative algebras are only used in the Heisenberg picture. That 'Heisenberg lost the political battle' does not mean that noncommuative algebras disappeared from quantum mechanics but only that they receded into the background - which can be seen in any modern textbook on quantum mechanics. The algebraic view is that of a minority.

The Heisenberg picture is essential only in quantum field theory, though even there it is often often dominated by functional integral techniques derived from the Schrödinger picture. But quantum field theory is not the topic of Rovelli's paper!

 

[PeterDonis]

This is not 'differential operators vs. matrices' but Schrödinger picture (differential equation for the quantum state) vs. the Heisenberg picture (non-commutative algebras to describe transition probabilities),

You and I appear to be interpreting this differently. But that's actually not the main point. See below.

Rovelli explicitly explains his goal on p.2...
This is geared against the Schrödinger picture in terms of a quantum state

But his goal only makes sense if everybody uses the Schrodinger picture and the Heisenberg picture (and the interaction picture) are neglected. But they're not.

I claimed neither.

I know you didn't claim it explicitly. But you appear to agree with Rovelli's point, which, as above, only makes sense if pictures other than the Schrodinger picture are neglected. So I asked you if you really believe that. Of course I don't think you do, and I don't think Rovelli does either; but I do think Rovelli is failing to realize the implications of what he is claiming.

 

[A. Neumaier]

But his goal only makes sense if everybody uses the Schrodinger picture and the Heisenberg picture (and the interaction picture) are neglected. But they're not.

As far as interpretation questions are concerned (which is Rovelli's topic), essentially everybody uses the Schrodinger picture and the Heisenberg picture (and the interaction picture) are neglected.

Or can you point to serious work on the interpretation of quantum mechanics framed in the Heisenberg picture (or the interaction picture) ?

 

[A. Neumaier]

I do think Rovelli is failing to realize the implications of what he is claiming.

Are you claiming that Rovelli is so ignorant of the implications of what he is claiming?

 

[PeterDonis]

As far as interpretation questions are concerned (which is Rovelli's topic), essentially everybody uses the Schrodinger picture

If Rovelli's purpose is to fill a gap in the literature on interpretation, then I would think this...

quantum field theory is not the topic of Rovelli's paper

...would be a better issue to tackle. After all, QFT is the most fundamental version of QM we have, correct? So interpreting that would seem to be the important gap to fill.

 

[PeterDonis]

Are you claiming that Rovelli is so ignorant of the implications of what he is claiming?

I was mystified by the fact that he seemed to be saying nobody uses any picture other than the Schrodinger picture. If he's only saying that no interpretation uses any picture other than the Schrodinger picture, that's a much less mystifying claim. I'm still not sure it's entirely true (for example, it doesn't seem to fit the Copenhagen interpretation), but it's much less mystifying.

 

[A. Neumaier]

If he's only saying that no interpretation uses any picture other than the Schrodinger picture, that's a much less mystifying claim.

It is point 3 in The 7 Basic Rules of Quantum Mechanics!

I'm still not sure it's entirely true (for example, it doesn't seem to fit the Copenhagen interpretation),

The Copenhagen interpretation was created in the Schrödinger picture. Born used the Schrödinger picture already in the paper

• Max Born, Zur Quantenmechanik der Stoßvorgänge, Zeitschrift für Physik 37 (1926), 863–867.

where he introduced what we call today Born's rule. Scattering in the Heisenberg picture is much more difficult to motivate, and had to wait until much later

• W. Heisenberg, Die „beobachtbaren Größen“ in der Theorie der Elementarteilchen, Zeitschrift für Physik 120 , (1943), 513–538.

when Heisenberg introduced the S-matrix.

 

[A. Neumaier]

Rovelli's purpose is to fill a gap in the literature on interpretation

Rovelli's purpose was different:

A way to get clarity about quantum mechanics is to undo the conceptual mess raised by Schrödinger’s introduction of the “quantum state”. This is what the Relational Interpretation does.

 

[PeterDonis]

It is point 3 in The 7 Basic Rules of Quantum Mechanics!

That's hardly definitive since those rules were written to help moderate discussions here at PF, not as any kind of comprehensive claim about how QM must be done or interpreted.

The Copenhagen interpretation was created in the Schrödinger picture.

That doesn't mean it has to always use the Schrodinger picture.

Born used the Schrödinger picture already in the paper

• Max Born, Zur Quantenmechanik der Stoßvorgänge, Zeitschrift für Physik 37 (1926), 863–867.

where he introduced what we call today Born's rule. Scattering in the Heisenberg picture is much more difficult to motivate, and had to wait until much later

• W. Heisenberg, Die „beobachtbaren Größen“ in der Theorie der Elementarteilchen, Zeitschrift für Physik 120 , (1943), 513–538.

when Heisenberg introduced the S-matrix.

These don't seem to be about any particular interpretation, they are just part of the development of the basic mathematical machinery of QM.

Rovelli's purpose was different

He's proposing a new interpretation. That implies that he thinks there is something missing from all existing interpretations.

 

[A. Neumaier]

These don't seem to be about any particular interpretation, they are just part of the development of the basic mathematical machinery of QM.

The point is that before 1943, scattering (which was the main reason for Born's rule) could be done only in the Schrödinger picture. In the Heisenberg picture, one could do only stationary problems and semiclassical approximations. All the probabilistic stuff needed the Schrödinger picture.

He's proposing a new interpretation. That implies that he thinks there is something missing from all existing interpretations.

Not missing, but misguided! These are his words. And he spells out what he thinks is at fault: The quantum state, more precisely (as becomes apparent from the later parts) the interpretation of the dynamics of the quantum state, i.e., of the Schrödinger picture.

 

[A. Neumaier]

The point is that before 1943, scattering (which was the main reason for Born's rule) could be done only in the Schrödinger picture. In the Heisenberg picture, one could do only stationary problems and semiclassical approximations. All the probabilistic stuff needed the Schrödinger picture.

A nice, recent historical paper about this is

• A.S. Blum, The state is not abolished, it withers away: How quantum field theory became a theory of scattering, Manuscript (2020). arXiv:2011.05908.

 

[PeterDonis]

scattering (which was the main reason for Born's rule)

Would the Stern-Gerlach experiment be considered a "scattering" experiment?

 

[A. Neumaier]

Would the Stern-Gerlach experiment be considered a "scattering" experiment?

In a scattering experiment one starts with energy-momentum eigenstates of a free in-Hamiltonian and measures energy-momentum eigenstates of a free out-Hamiltonian, while the dynamics at intermediate times is given by an interacting Hamiltonian.

In case of the Stern-Gerlach experiment, the interaction is given by a classical magnetic field, the in-momentum is fixed in z-direction, and the out-momentum in the traditional planar idealization is two-valued, corresponding to opposite perturbations in the x-direction for the two possible values of the spin.

The hitting probabilities when intercepting at time $t$ (after the magnetic field was paased) a Stern-Gerlach particle (entering the magnetic field at time $0$) are found by squaring suitable matrix elements of the state vector $\psi(t)$ initialized at time $0$ by a plane wave.

 

[gentzen]

Would the Stern-Gerlach experiment be considered a "scattering" experiment?

Yes, the observed results in actual Stern-Gerlach experiments can be computed based on the interpretation as a "scattering" experiment. At least that is how I would interpret A. Neumaier's answer.

Whether conclusions in typical idealized Stern-Gerlach thought experiments allow an interpretation in terms of a "scattering" experiment probably depends on the details of the experiment and the conclusion. Often the intention of those thought experiments is to analyse properties of the quantum state and improve its intuitive understanding. The interpretation as a "scattering" experiment would at least defeat those intentions, so I would say that they cannot be considered as "scattering" experiments.

 

[*now*]

It is point 3 in The 7 Basic Rules of Quantum Mechanics!

The Copenhagen interpretation was created in the Schrödinger picture. Born used the Schrödinger picture already in the paper

• Max Born, Zur Quantenmechanik der Stoßvorgänge, Zeitschrift für Physik 37 (1926), 863–867.

where he introduced what we call today Born's rule. Scattering in the Heisenberg picture is much more difficult to motivate, and had to wait until much later

• W. Heisenberg, Die „beobachtbaren Größen“ in der Theorie der Elementarteilchen, Zeitschrift für Physik 120 , (1943), 513–538.

when Heisenberg introduced the S-matrix.

On this, I just wonder what is meant by this, because hadn’t Born already named “Quantum Mechanics” for the discontinuous nature and written about probabilities and such, all before Schrödinger came up with a different easier approach, e.g. Born, Jordan, Z. Phys.33, 479-505, 1925; rec. 11-06-1925 ?

 

[DrChinese]

Rovelli, in his recent paper, writes:

"The mistaken idea is that the quantum state ψ represents the “actual stuff” described by quantum mechanics.

...

In RQM, it is a bookkeeping of known facts, and a tool for predicting the probability of unknown facts, on the basis of the available knowledge."

I am not sure I understand Rovelli here. Prior to "collapse", the quantum state - as described by a probability wave existing simultaneously at a variety of place in spacetime - can be manipulated at those places. How can that be termed as mere "bookkeeping" as Rovelli believes?

If we have a double slit setup, or anything where interference occurs between 2 possible paths: Obviously the mutual influence of both possible paths is responsible for the observed outcome pattern. There cannot be independence (product states) for the results.

Or: You can take the output ports of a polarizing beam splitter (PBS), and those ports obviously BOTH contain something that is "real". Because you can later recombine them, and you will return them to their original input state. In other words: what comes out of the PBS is NOT one the following:

a) A V beam from the transmitted port, and nothing from the reflected port.
b) An H beam from the reflected port, and nothing from the transmitted port.

Because those beams could not be recombined to restore the original input beam. What am I missing?

 

[Jarvis323]

I am not sure I understand Rovelli here. Prior to "collapse", the quantum state - as described by a probability wave existing simultaneously at a variety of place in spacetime - can be manipulated at those places. How can that be termed as mere "bookkeeping" as Rovelli believes?

Is Rovelli's "bookkeeping" view about the wave function different from other $\psi$-epistemic interpretations?

 

[Fra]

Rovelli's view shares parts of other epistemic views, but it's still different. The similarity is also why I share a part of Rovellis reasoning, but not his final steps.

"For relative facts, every interaction can be seen as a “Copenhagen measurement”, but only for the systems
involved. Any physical system can play the role of the “Copenhagen observer”, but only for the facts defined
with respect to itself. From this perspective, RQM isnothing else than a minimal extension of the textbook Copenhagen interpretation, based on the realisation that any physical system can play the role of the “observer” and any interaction can play the role of a “measurement”: this is not in contradiction with the permanence of interference through interactions because the “measured” values are only relative to the interacting systems themselves and do not affect other physical systems.
"
-- Rovelli, p3, https://arxiv.org/abs/2109.09170

My problem with this is that, my view is taking what Rovellis says seriously (ie that any systems qualifies as an observer) AND trying to understand the origin of actual interactions in terms of inferences (unification quest), between such arbitrary observers will force us to deform QM and the reason for this is that the real observers informationa capacity, puts a constraint to the "effective theory", and I think this is not bookkeeping, I think it has deep implications for the hierarchy of physical interactions. This makes things more complicate that what I perceive that Rovellis thinks. Rovells somehow seems to look for an interpretation that realizes a good idea, buth without breaking the QM formalism. I just do not see how that is possible.

/Fredrik