Question about discussions around quantum interpretations

[bhobba]

Thank you for the recommendation, I'll look into it!
Also, I checked Dr. Fatima- it was a great finding. I really like her videos on science, colonization and other social issues.

Oh yes. She is terrific, just so rational. I became a patron after watching her first video, and I am glad I did. She is living proof that it doesn't matter what you believe; if your arguments are rational, it is always of value. That applies to all areas, of course, not just science. Her Sokal video is IMHO exceptional - of course, I was being vexatious in my comments about the science wars - it is much more nuanced, but really belongs to your area than mine.

Thanks
Bill

 

[gentzen]

If you have not read it, read: Surely You're Joking, Mr Feynman.

Or maybe, rather don't read it. This book was not written by Richard Feynman, but by Ralph Leighton. This is hinted at by "Adventures of a Curious Character as told to Ralph Leighton" on the title page which is hidden inside of the book. (Yes, this does not just feels fishy, it really is!)
And don't make the mistake to confuse Ralph Leighton with Robert B. Leighton of Feynman lectures fame (together with Matthew Sands). He is his son, but I don't know anything about him except for what I learned from Angela Collier in that video.

 

[PeterDonis]

Or maybe, rather don't read it. This book was not written by Richard Feynman

Why would that be a reason not to read it? Even if it wasn't written by Feynman, strictly speaking, it's still about him, and has a lot of direct quotes from him.

 

[gentzen]

Why would that be a reason not to read it? Even if it wasn't written by Feynman, strictly speaking, it's still about him, and has a lot of direct quotes from him.

How can you trust Ralph that the quotes are directly from Feynman, given he is so dishonest when it comes to the actual author of the book? Angela Collier argues that the stories themselves are with almost absolute certainty not factual.

 

[Morbert]

Oh, that makes sense. Are there any probabilistic no-collapse interpretations?

If by no-collapse interpretation you mean interpretations that do not posit collapse as a physically real process, then consistent/decoherent histories is a no-collapse interpretation which permits stochastic/probabilistic dynamics.

 

[PeterDonis]

How can you trust Ralph that the quotes are directly from Feynman, given he is so dishonest when it comes to the actual author of the book?

Oh, for goodness' sake. I'm sorry to sound blunt here, but if you feel cheated because you didn't think to look at the inside cover of the book before buying it, to me that's on you, not Feynman or Ralph Leighton. Not to mention that it couldn't have been just Leighton: the publishers of the book had to know how it was written, and they listed Feynman as an author.

Angela Collier argues that the stories themselves are with almost absolute certainty not factual.

Angela Collier wasn't alive when those events took place. Other people who were and who knew Feynman personally, such as Freeman Dyson (hardly an unreliable source or one given to hyperbole), have given their own accounts of some of the same stories, which tally fairly closely with those in the book. Indeed, some of them, such as the safecracking stories in Los Alamos, AFAIK are now corroborated by declassified government documents.

Was Feynman given to exaggerating? Sure. Is that a reason not to read the book? IMO, not at all.

 

[PeterDonis]

Angela Collier

For a critique of her treatment of Feynman from someone who was in a much better position than she to know relevant facts, see this:

https://www.feynmanlectures.caltech...Baez_regarding_Angela_Colliers_sham_video.pdf

 

[Fra]

I wont add more than necessary here, but as the OP was a sociologist, and this came up... I'll just add this.

If experimentally quantum mechanics has yielded more and more evidence of probabilistic behavior, why is it that there are so many interpretations looking to reintroduce fundamental determinism?

IMO this is at least related to the type of explanation you want. The idea of deterministic chaos that needs fine tuning is conceptaully simpler as it is based on the traditional system dynamics. So one natural reason for what you ask my be that if possible, it would be an explanation within a conceptaully traditional type of model. Ie system dynamcs. Where we have initial or boundary conditions and differential equations.

The alternative is conceptually more challenging, and likely involved emergence in place of fine tuning (though the final results may well be consistent) but that is traditiional, so more "werid", like a new paradigm.

Thank you! I'll read this later tonight- I want to pay attention to it. Since I haven't read it yet, I can't say for sure what you mean, but I don't think that most sociologists consider reality a social construct in a strict sense.

As you are a sociologist, interesting one of Lee Smolings books and works of evolution of law (which is related to the alternative to system dynamics) was in cooperation with a social theorist, Roberto Unger. Ive read most of those books and talks, and it does not surprise me how some of hte ideas was "inspired" but social interactions and emergent social laws.

I recommend to look at it
https://pirsa.org/08100049
and there is also some books

https://www.amazon.com/Singular-Universe-Reality-Time-Philosophy/dp/1107074061?tag=pfamazon01-20
- cowritten by a social theorist.

To not misunderstand this before you get to the depth of this, the important keys in here is NOT to view the progress of human science, as sociology. Indeed Smoling has writteon on that topic as well in the critique on string theory in other books. But that is NOT what this is about. This is a more deep serious suggestion that suggests that physical reality and its LAWS (that in system dynamics are constraints) are EMERGENT, loosely analogous to how social laws emerge. As always, taking analogies too serious will miss the point.

/Fredrik

 

[gentzen]

Oh, for goodness' sake. I'm sorry to sound blunt here, but if you feel cheated because you didn't think to look at the inside cover of the book before buying it, to me that's on you, not Feynman or Ralph Leighton.

I didn't buy the book, I read the copy from a public library. I don't remember whether I noticed that actually Leighton was the author. But I certainly remember that I was convinced that Leighton was a physicist and coauthor of the Feynman lectures. Maybe Preface and Acknowledgment of "QED: The Strange Theory of Light and Matter" were responsible for this:

If you are planning to study physics (or are already doing so), there is nothing in this book that has to be “unlearned”: it is a complete description, accurate in every detail, of a framework onto which more advanced concepts can be attached without modification. For those of you who have already studied physics, it is a revelation of what you were really doing when you were making all those complicated calculations!

This book purports to be a record of the lectures on quantum electrodynamics I gave at UCLA, transcribed and edited by my good friend Ralph Leighton. Actually, the manuscript has undergone considerable modification. Mr. Leighton’s experience in teaching and in writing was of considerable value in this attempt at presenting this central part of physics to a wider audience.

When I read those words again, your comment about the publishers came to my mind:

Not to mention that it couldn't have been just Leighton: the publishers of the book had to know how it was written, and they listed Feynman as an author.

At least for QED, listing Feynman as one of the author was certainly mandatory. And here too, listing Feynman as one of the author was probably mandatory. In fact, Angela Collier says in the video that there are tape recordings (they are even on the internet), and that according to James Gleick the stories in the book roughly correspond to those tapes, but heavily filtered. Not listing Ralph Leighton as author is the fishy part. Both books appeared 1985, QED is probably the one which appeared first. Maybe not being listed as author on the stories book was Ralph's revenge for being denied authorship of QED?

For a critique of her treatment of Feynman from someone who was in a much better position than she to know relevant facts, see this:

https://www.feynmanlectures.caltech...Baez_regarding_Angela_Colliers_sham_video.pdf

Michael Gottlieb is a friend of Ralph Leighton, and also has other conflicts of interest. Hence, it is unfortunate that he wrote: "In closing I will mention that Angela is making money from publishing this poisonous trash." Overall, my impression is that he is simply bad at coping with that stuff, but not acting in bad faith. I'm not convinced by his:

She gives false and misleading information about other books too, claiming, for example, that all the stories in Feynman’s autographical books are lies, without giving any basis for that claim, other than her speculations.

Angela did give evidence, for example from James Gleick and Murray Gell-Man. And his

the exercises were originally published in the 1960s, by Feynman and his coauthors

doesn't fact check either: Feynman was not listed as author when the excercises were originally published in the 1960s. But he is listed as author in Michael Gottlieb's publication. And of course, attacking John Baez before even trying to contact Angela Collier was not wise either:
John Baez (1/2):

... where Angela Collier will ruthlessly dissect the mythology he built around himself. You probably won't agree with everything she says, and you may hate some of it, but it will still be thought-provoking.

John Baez (2/2):

I was not implicitly endorsing all of @acollierastro's claims in my first post. I merely said what I wanted to say.

Nor am I implicitly endorsing Gottlieb's claims here. I hope Gottlieb and Collier can discuss this without using me as an intermediary.

Still, Michael Gottlieb somehow managed to convince me that Angela Collier guesses at the motivations of Ralph Leighton, Michael Gottlieb, and other "self appointed coauthors" of Feynman are off. What drove this home for me was
Blake C. Stacey

I'll go ahead and disagree with Collier's take on the *Feynman's Lost Lecture* book. The Goodsteins give *more* and *more accurate* credit than Feynman did.

(But her analysis of the stories in Leighton's book is not affected by this.)

 

[PeterDonis]

I was convinced that Leighton was a physicist and coauthor of the Feynman lectures.

Robert Leighton was a physicist and coauthor of the Feynman lectures. His son Ralph Leighton was not, nor did he ever claim to be. Again, this doesn't look to me like a case of anyone trying to deliberately mislead; it looks like a case of you not being very careful.

I'm not going to bother arguing any further about Angela Collier's claims. Both of us have given some references, and other readers can make up their own minds, and it's off topic for this thread anyway.

 

[RUTA]

Here is a 2023 paper relevant to this thread Many-Worlds: Why Is It Not the Consensus?

Abstract:​

In this paper, I argue that the many-worlds theory, even if it is arguably the mathematically most straightforward realist reading of quantum formalism, even if it is arguably local and deterministic, is not universally regarded as the best realist quantum theory because it provides a type of explanation that is not universally accepted. Since people disagree about what desiderata a satisfactory physical theory should possess, they also disagree about which explanatory schema one should look for in a theory, and this leads different people to different options.

 

[bhobba]

Just a quick question to Ruta.

First, I always enjoy your posts, and this is a nice one, especially for helping out a newbie from a different area, sociology.

My question is, do you think QFT with the field considered real a realist interpretation?

Thanks
Bill

 

[RUTA]

Just a quick question to Ruta.

First, I always enjoy your posts, and this is a nice one, especially for helping out a newbie from a different area, sociology.

And also to you :-)

My question is, do you think QFT with the field considered real a realist interpretation?

Thanks
Bill

I'm not a philosopher, my philosophy colleague and coauthor Michael Silberstein handles these kinds of questions, but naively my answer is "yes". Many (most?) in foundations believe quantum fields are the fundamental building blocks of reality. I don't see how you can get any more "real" than that, but I may not appreciate the philosophical nuance.

 

[ojitojuntos]

Thank you for the replies in this thread! I'll check the various sources you've recommended, although it will probably take me a while haha
I also wanted to clarify that I'm aware that the laws of physics are not emergent in the same sense as social structures and dynamics are. When I mentioned that as a sociologist accepting stochasticity as inherent to reality was easier, I meant from a epistemic point of view, but I'm not equating social dynamics to physics; I know that these are very different areas of knowledge.

Related to this, I'm having some trouble understanding a couple of concepts: I understand that the wavefunction evolves deterministically once you have the measurement; however, what we observe is that, before the measurement, reality looks inherently probabilistic, and that this represents the measurement problem, which quantum interpretations try to solve, right?

Now, at the effective scale of human experience, even if we assume a probabilistic interpretation of quantum, does this make a difference? Or am I wrongly assuming a barrier between the quantum and classical?

 

[PeterDonis]

the wavefunction evolves deterministically once you have the measurement; however, what we observe is that, before the measurement, reality looks inherently probabilistic, and that this represents the measurement problem, which quantum interpretations try to solve, right?

I think you have it somewhat backwards.

In a typical quantum experiment, we prepare a system, it goes through some kind of process, and then we measure it. Preparing the system determines the starting wave function; the wave function then undergoes unitary evolution (which is deterministic) through the process in the middle, and only when we measure at the end do any probabilities come into play.

As an example, take the Stern-Gerlach experiment (or at least an idealized version of it). We prepare a spin-1/2 particle in a definite state, say spin-z up. Then we pass it through a Stern-Gerlach magnet oriented in, say, the x direction. Doing that induces a unitary (i.e., deterministic) evolution of the state. Then we measure the particle with a detector screen downstream of the S-G magnet: here we have one of two possible results, corresponding to two different places where the particle could hit the screen: one place corresponds to a measurement result of spin-x up, the other corresponds to a measurement result of spin-x down. There is a 50% probability of each result; that's the only place where probability comes into play at all.

The measurement problem, in the context of the experiment just described, is this: the state of the particle after it goes through the S-G magnet is a superposition of spin-x up and spin-x down (actually it's an entangled superposition, with the particle's spin being entangled with the direction of its momentum--the two different momentum directions point at the two different spots on the detector screen). How is it that when we measure the particle, we don't measure any such superposition, but instead, we measure either spin-x up or spin-x down? Or, to put it another way, why do we measure only one spot on the detector screen where the particle hits, instead of two? What is it about the screen that makes the particle just have one measurement result?

 

[bhobba]

Or, to put it another way, why do we measure only one spot on the detector screen where the particle hits, instead of two? What is it about the screen that makes the particle just have one measurement result?

Nutting things out for yourself is a great way to learn.

I could give my answer; however, central to this whole thing is something called Gleason's Theorem:
https://arxiv.org/pdf/quant-ph/9909073

Please take a moment to read it, put your thinking cap on, and see what emerges on the other side.

Post any thoughts here.

Thanks
Bill

 

[RUTA]

The measurement problem, in the context of the experiment just described, is this: the state of the particle after it goes through the S-G magnet is a superposition of spin-x up and spin-x down (actually it's an entangled superposition, with the particle's spin being entangled with the direction of its momentum--the two different momentum directions point at the two different spots on the detector screen). How is it that when we measure the particle, we don't measure any such superposition, but instead, we measure either spin-x up or spin-x down? Or, to put it another way, why do we measure only one spot on the detector screen where the particle hits, instead of two? What is it about the screen that makes the particle just have one measurement result?

Here are options for answering questions like this from Allori's paper linked in post #103:

Some theories are what Einstein [3] called constructive theories. For one thing, these theories have a microscopic ontology, which constitute the building blocks of everything else. Constructive theories allow one to understand the phenomena compositionally and dynamically: macroscopic objects are composed of microscopic particles, and the macroscopic behavior is completely specified in terms of the microscopic dynamics. Therefore, the type of explanation these theories provide is bottom-up, rather than top-down. According to Einstein, there is another type of theory, which he dubbed principle theory. Theories of this type, also called kinematic theories, are formulated in terms of principles, which are used as constraints on physically possible processes: they exclude certain processes from physically happening. In this sense, principle theories are top-down: they explain the phenomena identifying constraints the phenomena need to obey to. They are ‘kinematic’ theories because the explanations they provide do not involve dynamical equations of motion and they do not depend on the interactions the system enters into. Instead, by definition, constructive theories involve dynamical reductions in macroscopic objects in terms of the motion and interactions of their microscopic three-dimensional constituents. Flores [4] argued that this distinction could be expanded in terms of framework theories, which deal with general constraints, and interaction theories, which explicitly invoke interactions. He thought that framework theories are principle theories while interaction theories include a larger set of theories than constructive theories. Furthermore, he connected framework theories with unification and interaction theories with mechanistic explanation (see also [5,6]).

 

[Fra]

The measurement problem, in the context of the experiment just described, is this: the state of the particle after it goes through the S-G magnet is a superposition of spin-x up and spin-x down (actually it's an entangled superposition, with the particle's spin being entangled with the direction of its momentum--the two different momentum directions point at the two different spots on the detector screen).

Or, to put it another way, why do we measure only one spot on the detector screen where the particle hits, instead of two? What is it about the screen that makes the particle just have one measurement result?

My interpretation is that it's because the screen is just like the rest of us (an agent). Ie. If if we know the possible answers our decisions and behaviour reflect the uncertainty, but once we get the answer, it's precisely one of the possibilities, and our decisions align. I think the screen is no different.

So that gives the follow up question, is this look like it's all "ignorance", then why is bell inequality violated?

I think it's because in bells theorem, one assumes that the ignorance is agreed upon by all agents (an objective beable), thus the interactions between the agents should be possible to be described as an "average" of the mechanisms for each hidden value.

But it seems this idea is wrong - it seems to me uncertainty and ignorance is itself contextual. And when such contexts interact (ie two PARTS, or two AGENTS), quantum inferenece happens, that can't be explain in "classical terms"

/Fredrik

 

[PeterDonis]

the screen is just like the rest of us (an agent)

I don't see how this is a viable claim, since the screen does not exhibit any of the behaviors we exhibit as agents.

I think it's because in bells theorem, one assumes that the ignorance is agreed upon by all agents (an objective beable), thus the interactions between the agents should be possible to be described as an "average" of the mechanisms for each hidden value.

But it seems this idea is wrong - it seems to me uncertainty and ignorance is itself contextual. And when such contexts interact (ie two PARTS, or two AGENTS), quantum inferenece happens

I know this is the QM interpretations subforum, where the rules are a little broader, but still personal speculation is off limits here. You still need some kind of reference as a basis for making claims. Is this viewpoint proposed anywhere in the literature?

 

[Fra]

I don't use the "agent" label as specific claim of the screen, I used it as a change of conceptual modelling perspective, as my opinon is that it is at the heart of the problem. Agent/observer or observed/matter is to me mainly a matter of perspective of inference (and not a ontological claim in any way). And as different agents can observer it each, and agent is just a normal physical system - see from an external perspective.

Normally the agent concept is understood from it's internal perspective/drive (such as decision making; though these could in principle be stochastic self-organisation, as in Baranders view, not it does not imply consciousness). Thus the agent perspective can used without making specific assumptions or speculations of its evolution or structure. It's a modelling perspective.

In contrast matter(screens) is described from an external perspective(external agent beeing a macroscopic laboratory with human scientists), in terms of it's state in a state space with dynamical laws. We can similarly think of this without making specific assumptions or knowledge of the exact full dynamics law. It's a modelling perspective.

As the problem here isn't that QM doesnt describe this, it's that even with the model explicity under our nose, we have trouble to understand it, intuitively. So when I interpret the screen to be just an agent, I meant I try imagine how a screen might perceive and responds to an incoming particle, when only the preparation procedure is known, and seek some intuition from that view.

Ie try to reflect of the measurement problem from this perspective. Just like we can say that an agent is just a physical system; then what does a "physical system" composed of interacting agents look like? This is not an excplicit claim, it is a change of perspective, that is not palatable to all, but I think if offers many insights, that is lacking from the system dynamics view. In particular when you try to think about the difference between a classical uncertainty and an uncertainty constructed from information that forces the agent to maintain non-commutative structure at a single time.

/Fredrik

 

[PeterDonis]

I don't use the "agent" label as specific claim of the screen, I used it as a change of conceptual modelling perspective,

Yes, but you aren't giving any references at all to any literature where this "perspective" is discussed. As I said, even in this subforum, we are discussing interpretations that are given in the literature, not people's own personal home-brewed interpretations. You seem to be describing the latter, and that's off topic here.

 

[Fra]

Yes, but you aren't giving any references at all to any literature where this "perspective" is discussed. As I said, even in this subforum, we are discussing interpretations that are given in the literature, not people's own personal home-brewed interpretations. You seem to be describing the latter, and that's off topic here.

As to the agent part, its partly originated from qbism. But the the perspective change I refer to sits at the mathematical modelling perspective, and is general, and not something I brewed, and is not itself an interpretation i think. I tried to add perspectives as it helped me at least, everyone can make up their own interpretations in their heads.

Some papers relating perspectives

Hydrodynamic Limits of non-Markovian Interacting Particle Systems on Sparse Graphs
https://arxiv.org/abs/2205.01587

An elementary proof of convergence to the mean-field equations for an epidemic model
https://arxiv.org/abs/1501.03250

System Dynamics versus Agent-Based Modeling: A Review of Complexity Simulation in Construction Waste Management
https://www.mdpi.com/2071-1050/10/7/2484

The general idea is that markovian agent based models converge in the mean field limit a timeless system dynamics, as the number of agents -> inf. But Non-markovian agent based models, would converge to time dependent laws, in the same limit. Interesting as they give rise to different kinds of "time" even. And the idea is also that agent based models are "larger" in that they can model things "system dynamics can't", because system dynamics represent a limiting case of agent based models.

But both models have pros and cons and can be used together. System dynamics is a top-down approch, that is constraint based. Agent based modelling is more computationally driven be decentralized rules (which correspondes to the causal mechanisms), and the constrains of SD have a correspondence of limits of agent based models. I think as is seem in QM in particular, the system dynamics level really seem to encrypt the causal mechanisms. It seems hard to understand what is going on. This I think is likely a general feature of the paradigm.

Agent based models would be comptutational intense, so it will not replace system dynamics. Its in a way easier to describe continous field than dense population of "parts" interaction. But if one wants to understand the interactions, between the parts (which is really what I think we are talking about in the experiments) the "system view" will hide this. So no wonder evolution in hilbert space seems hard to make sense of.

/Fredrik

 

[PeterDonis]

Some papers relating perspectives

Thanks, these are helpful.

 

[gentzen]

My interpretation is that it's because the screen is just like the rest of us (an agent). Ie. If if we know the possible answers our decisions and behaviour reflect the uncertainty, but once we get the answer, it's precisely one of the possibilities, and our decisions align. I think the screen is no different.

I don't use the "agent" label as specific claim of the screen, I used it as a change of conceptual modelling perspective, as my opinon is that it is at the heart of the problem. Agent/observer or observed/matter is to me mainly a matter of perspective of inference (and not a ontological claim in any way). And as different agents can observer it each, and agent is just a normal physical system - see from an external perspective.

So you label the screen as agent, because it plays the role of observer in your modeling of the S-G experiment? I don‘t like this way of dropping the distinction between observer and agent. An observer suggest something passive, like the screen. An agent suggest something more active, like an information gathering and using system (IGUS).
You should at least clarify how the screen is utilizing the information, if you want to label it as an agent. Is it using the information to store it for later retrieval, like a hard disk? Or on the other side, for transforming itself, like in a nanofabrication process?

 

[Fra]

So you label the screen as agent, because it plays the role of observer in your modeling of the S-G experiment? I don‘t like this way of dropping the distinction between observer and agent. An observer suggest something passive, like the screen.

I agree. But the passive nature of the screen is an approximation that is valid just because its huge.

Fundamentally passive observer is a fiction to me. In practice passive observer is a limiting case.

But once you consider the actual limit. I loose track of explanations and how dynamical law emerge.

An agent suggest something more active, like an information gathering and using system (IGUS).
You should at least clarify how the screen is utilizing the information, if you want to label it as an agent. Is it using the information to store it for later retrieval, like a hard disk? Or on the other side, for transforming itself, like in a nanofabrication process?

the microstate of a macroscopic screen is itself able to encode information ~ memory.

The internal physical processes in the screen and its interaction with the environment is the only "information processing" we need.

So to understand this "agent" is of coursr indistinguishable from understanding the in depth microstate and physical interactions of matter. Its only the perspective that differs.

also the agent is not not unique, just ad one screen can be thought of as beeinh made out of atoms with relations. An agent can been seen as a group of microagents.

/Fredrrik

 

[gentzen]

also the agent is not not unique, just ad one screen can be thought of as beeinh made out of atoms with relations. An agent can been seen as a group of microagents.

That is beside the point. After one has setup a model for a specific physical situation, the roles are fixed. Procrastinating over the fact that one could also have setup the roles and the model differently is not helpful. It even risks to confuse object level facts with meta-level stuff like:

The action to “acquire” devices on the other hand is on a kind of meta-level, which is of limited help for discussions of how to provide physical meanings and their connection to the formalism.

 

[Fra]

That is beside the point. After one has setup a model for a specific physical situation, the roles are fixed. Procrastinating over the fact that one could also have setup the roles and the model differently is not helpful. It even risks to confuse object level facts with meta-level stuff like:

I get your point, it does raise problems!

Also in the normal paradigm, we shouldnt confuse them.

But here we try to probe deeper I think this confusion is real and not only meta stuff. So my perspective is to accept the problems(including confusion) of what are object level facts and how those are contextual howto find an objective context in some limit. (Ie macroscopic reality or "classical world").

We might disagree on this.

/Fredrik

 

[ojitojuntos]

Hello guys. OP again. I appreciate the discussion and thorough explanations for a layman. I have one more question about the measurement problem:
If randomness arises at measurement, and we can’t pinpoint how the collapse occurs, why is it that most physicists (according to polls I’ve seen online) consider that reality is fundamentally probabilistic, instead of deterministic, but with epistemic uncertainty?
Is this correct? Or, in a simpler sense, do experimental results and the math seem to lean more towards fundamental probabilities?

 

[Fra]

If randomness arises at measurement, and we can’t pinpoint how the collapse occurs, why is it that most physicists (according to polls I’ve seen online) consider that reality is fundamentally probabilistic, instead of deterministic, but with epistemic uncertainty?
Is this correct? Or, in a simpler sense, do experimental results and the math seem to lean more towards fundamental probabilities?

Yes, this is the basic message of when experiment and QM violates Bells inqeuality.

Given some assumptions (that follow from a basic classical picture), if the results are only epistemic (in the sense of beeing due to physicists ignorance) then the inequality must hold - but it doesn't! That is the problem with the idea "deterministic, but with epistemic uncertainty".

/Fredrik

 

[PeterDonis]

Given some assumptions (that follow from a basic classical picture), if the results are only epistemic (in the sense of beeing due to physicists ignorance) then the inequality must hold

But there are QM interpretations where the results are epistemic and the assumptions you refer to are violated. For example, the Bohmian interpretation, in which the probabilities are purely due to our ignorance of the actual particle positions.