Undergrad Anti-realist Interpretations of QM

[Lord Jestocost]

...that there is no mind-independent reality?

You have to distinguish between the "empirical reality" and the "ultimate reality". As Bernard d'Espagnat points out:

What quantum mechanics tells us, I believe, is surprising to say the least. It tells us that the basic components of objects – the particles, electrons, quarks etc. – cannot be thought of as "self-existent". The reality that they, and hence all objects, are components of is merely "empirical reality".

This reality is something that, while not a purely mind-made construct as radical idealism would have it, can be but the picture our mind forces us to form of ... Of what ? The only answer I am able to provide is that underlying this empirical reality is a mysterious, non-conceptualisable "ultimate reality", not embedded in space and (presumably) not in time either.

From: https://www.theguardian.com/science/blog/2009/mar/17/templeton-quantum-entanglement

 

[Lynch101]

I can only repeat what I've said. I believe you are conflating having a value with having a possible set of observable values. Things interact because they do and we have the formalism and data to prove it.

Apologies, I don't quite follow this point because I'm not sure that I am conflating those two.

If we go back to the contention that "it is meaningless to talk about the quantum system prior to measurement", this begs the question as to why it is meaningless. This usually stems from the instrumentalist notion that the mathematical formalism only gives probabilistic predictions for the outcomes of measurements. If the mathematics only describes the outcome of experiments i.e. only gives the probability that a particular measurement will be observed, then it doesn't appear to describe the properties of the system prior to measurement.

The istrumentalist position doesn't go any further than to say the mathematics only allows us to make probabilistic predictions. It is essentially the shut up and calculate position and isn't an interpretation of the mathematical formalism.

The anti-realist interpretation does go further and makes ontological deductions from the instrumentalist position. It says that properties don't have well defined values prior to measurement because there are no properties prior to measurement.

So, would it be accurate to say;

realism = systems have property values prior to measurement?

I would be more inclined to say that realism, at its most fundamental, is simply the position that systems have properties prior to measurement, not necessarily property values.

Since the realism view is born out of a classical macroscopic intuition we evolved with, I say that it's unsupported by the measurements and should be abandoned. There is no reason to expect it to hold for purely quantum objects.

If we take realism to be as outlined above, then anti-realism has, what I would think, is an insurmountable issue, namely, how something without any properties can interact with a measurement device in the first place.

 

[Lynch101]

Anti-realism re/ QM pertains to the anti-realism of properties considered by QM (i.e. those represented by regions of Hilbert space), as opposed to all properties entirely.

Thanks Morbert, this offers a different insight that I haven't come across before. I need a bit of help parsing what it means though.

I keep coming back to the shared position of instrumentalism and anti-realism, which says that the mathematical formalism is only a predictive tool which gives us the probability that a certain measurement outcome will be observed.

If the mathematical formalism only describes the probability that a certain measurement outcome will be observed, then it wouldn't seem to describe the properties of the system prior to measurement. The statement about the meaninglessness of talking about the quantum system prior to measurement seems to stem from this idea.

The instrumentalist position remains agnostic on the question of the properties of the system prior to measurement, while the anti-realist position appears to say that the mathematical formalism doesn't describe an ontic state and only gives us probabilistic measurement outcomes because there is no ontic state prior to measurement i.e. there are no properties whatsoever.If we take the interpretation you propose above that the anti-realist position of QM refers to anti-realism of properties considered by QM, does that mean that there are other properties which QM does not consider? Would these then qualify as hidden variables?

 

[Lynch101]

You have to distinguish between the "empirical reality" and the "ultimate reality". As Bernard d'Espagnat points out:

What quantum mechanics tells us, I believe, is surprising to say the least. It tells us that the basic components of objects – the particles, electrons, quarks etc. – cannot be thought of as "self-existent". The reality that they, and hence all objects, are components of is merely "empirical reality".

This reality is something that, while not a purely mind-made construct as radical idealism would have it, can be but the picture our mind forces us to form of ... Of what ? The only answer I am able to provide is that underlying this empirical reality is a mysterious, non-conceptualisable "ultimate reality", not embedded in space and (presumably) not in time either.

From: https://www.theguardian.com/science/blog/2009/mar/17/templeton-quantum-entanglement

Thanks Lord Jestocost, I would be inclined to ideas like that. I have only recently been introduced to d'Espagnat but I find myself nodding along in agreement with much of what I have read so far. I must read some more of his work.

This position would be a denial of the anti-realist position, wouldn't it?

 

[Lord Jestocost]

This position would be a denial of the anti-realist position, wouldn't it?

I don't think so. David J. Chalmers writes in "Ontological Anti-Realism":

"The basic question of ontology is “What exists?”. The basic question of metaontology is: are there objective answers to the basic question of ontology? Here ontological realists say yes, and ontological anti-realists say no."

 

[DrChinese]

So, would it be accurate to say;

realism = systems have property values prior to measurement?

Since the realism view is born out of a classical macroscopic intuition we evolved with, I say that it's unsupported by the measurements and should be abandoned. There is no reason to expect it to hold for purely quantum objects.

My thoughts exactly.

Einstein (1935, 1948) thought systems DID have property values prior to measurement. He defined the essential debate early on.

"According to [QM], the particle possesses neither a sharply defined momentum nor a sharply defined position. In which sense shall I imagine that this representation describes a real, individual state of affairs? Two possible points of view seem to me possible and obvious and we will weigh one against the other:

(a) The (free) particle really has a definite position and a definite momentum, even if they cannot both be ascertained by measurement in the same individual case. According to this point of view, the ψ-function represents an incomplete description of the real state of affairs. ...

(b) In reality the particle has neither a definite momentum nor a definite position; the description by ψ-function is in principle a complete description. The sharply-defined position of the particle, obtained by measuring the position, cannot be interpreted as the position of the particle prior to the measurement. ..."

Einstein believed (a). Bohr and most all other believed (b). Bell and Aspect had not yet come along to settle matters. So formulated as above, (a)=realist; (b)=anti-realist.

If you extend the "anti-realist" concept to a quantum interpretation, then you would reject Bohmian Mechanics and certain other interpretations. Those assert (a) explicitly. Of course, you can accept (b) - which is essentially a rejection of "local realism" - without picking a particular interpretation. But even then, you would essentially be adopting your perspective anyway, which translates to this:

anti-realism = systems don't have [well-defined] property values prior to measurement

 

[Paul Colby]

If we take realism to be as outlined above, then anti-realism has, what I would think, is an insurmountable issue, namely, how something without any properties can interact with a measurement device in the first place.

You keep repeating this and people here keep reinforcing this. It's a non-sensical statement, IMO.

Neglecting the point I was trying to make, let's take the DrChines's response to my Lotto number choice for example. His point was that classically he could take whatever picking device I chose and replace it with a prior device that would preselect the very same choice. This of course is nonsensical even from a classical perspective. Errors and noise would rapidly accumulate the further back in time one pushes the next choosing device. At some point, the residual quantum noise due to macroscopic objects really being composed of quantum ones would prevail. So even in the macroscopic case extrapolating a system to its infinite past is nonsense.

Now that said, Bell's theorem is a statement about the statistics of measurement. One uses measurements and statistics of those measurements showing that predefined values for a system can't exist. Okay, that not the same as the system doesn't exist or the system can't interact or the system isn't real.

 

[DrChinese]

Now that said, Bell's theorem is a statement about the statistics of measurement. One uses measurements and statistics of those measurements showing that predefined values for a system can't exist. Okay, that not the same as the system doesn't exist or the system can't interact or the system isn't real.

Agreeing with your essential comment: The quantum object is real and exists. Properties (of that object) don't seem to appear until a final measurement. The measurement requiring an interaction with a measuring device.

Obviously, my perspective gives weight to the concept of a "final measurement" - with all the problems associated with that. So again, that's where one's interpretation comes to the rescue (or not).

 

[Morbert]

The instrumentalist position remains agnostic on the question of the properties of the system prior to measurement, while the anti-realist position appears to say that the mathematical formalism doesn't describe an ontic state and only gives us probabilistic measurement outcomes because there is no ontic state prior to measurement i.e. there are no properties whatsoever.

If we take the interpretation you propose above that the anti-realist position of QM refers to anti-realism of properties considered by QM, does that mean that there are other properties which QM does not consider? Would these then qualify as hidden variables?

We should distinguish between a system having no ontic state, and the ontology of the system.

When we we say there is no ontic state, what we mean is we do not model the physics of the system terms of an ontic state or ontic state space. But we can still discuss the ontology/metaphysics of the system apart from our physical models. E.g. We can say the system exists. We can say the system interacts with our measurement apparatus in a predictable way. We can discuss the system as perhaps possessing some lawmaker property that's responsible for its dynamics etc.

 

[Lynch101]

I don't think so. David J. Chalmers writes in "Ontological Anti-Realism":

"The basic question of ontology is “What exists?”. The basic question of metaontology is: are there objective answers to the basic question of ontology? Here ontological realists say yes, and ontological anti-realists say no."

I'm mindful of the possibility of deviating too far from a discussion of QM interpretations on this point but the above would seem to turn on the meaning of "objective answers".

There are probably more than two ways to interpret the idea of "objective answers", but two examples I can think of that relate to QM would be taking it to mean:
a) measurables or observables
b) objective reality, independent from the mind of each person.

If the ontological anti-realist says that we cannot measure/observe ultimate reality, then this would be more a statement of realism with the implication being that there is a fundamental limit on how deep we can probe the world around us. But if that is the anti-realist position with respect to QM, to my mind, it would be tantamount to saying that there are hidden-variables and, again, would essentially be a realist position. It might be contrasted with an alternative realist position that says it will be possible to measure fundamental reality in the future, when we further refine our instruments.

In short, both ontological realism and anti-realism would, in this sense of "objective answers", point to an incomplete theory and the existence of hidden-variables (or "beables").The interpretations of anti-realism in QM, which I have encountered, don't seem to point to hidden-variables rather they deny the presence of them.Apologies for continually returning to this point, but I find it the easiest to digest and use as the basis for explanation and inquiry; if we go back to the question of the quantum system prior to measurement and talk about the mathematical formalism:

If we say that the mathematical formalism only gives us probabilistic predictions for the outcomes of experiments then it would seem to imply that the mathematics doesn't describe the quantum system prior to measurement. Am I correct in saying that this is what lead to the EPR paper and the charge of incompleteness?

If the mathematical formalism doesn't describe the quantum system prior to measurement we might then ask the question, is there a quantum system prior to measurement? If there is a quantum system prior to measurement then I think it is reasonable to state that it must have some/any properties, even if we cannot ascribe definite values to these properties. It might be down to my misinterpretation, but this to me sounds like a statement that there are hidden-variables.

The alternative to this, as far as I have reasoned thus far, would be the position that either there is no quantum system prior to measurement or that it has absolutely no properties whatsoever.

 

[Lynch101]

You keep repeating this and people here keep reinforcing this. It's a non-sensical statement, IMO.

Apologies for the repetition. I was trying to delineate what I saw as a key difference but it is one that you allude to below.

Now that said, Bell's theorem is a statement about the statistics of measurement.

It might help if I outline, by way of a crude example, the point I am trying to make on this particular point. It might offer insight into where I am going wrong.

If we imagine the crudest of crude quantum experiments where we have a black box with two screens at either end, divided into two panels, the top (or up) panel and the bottom (or down) panel. There is a button that we can press and a little while after pressing the button a flash of light appears on each screen on the end. The flashes of light always appear in opposite panels - if the flash appears on the up panel on the left screen then the flash on the right screen will always appear in the down panel, and vice versa.

If the mathematical formalism only tells us the probability of each flash, where it appears, and their always being in opposite panels then it doesn't appear to tell us what is happening inside the box. We can reasonably ask the question, what is happening inside the box.

One uses measurements and statistics of those measurements showing that predefined values for a system can't exist.

OK, so we cannot assign predefined values to what is happening inside the box but surely there must be something happening inside the box and it must have some properties, in some sense of the word - even if they aren't predefined values.

If we say yes, there is something happening inside the box but the mathematical formalism only gives us information about the measurement outcomes, then this would seem to suggest that there are hidden-variables in the box that the mathematics doesn't describe.

Okay, that not the same as the system doesn't exist or the system can't interact or the system isn't real.

My understanding is that the anti-realist position must say that there is nothing whatsoever happening inside the box because the alternative would be that there are realistic hidden-variables.

I think the idea that these hidden variables must have predefined values is a sticking point. If we drop the idea of the properties inside the box having predefined values then we are left with unqauntified properties, or what I am inclined to think are hidden-variables.

 

[Lynch101]

The quantum object is real and exists.

Would objects which exist and are real not have properties by necessity? Would real and existing not be properties in and of themselves?

Properties (of that object) don't seem to appear until a final measurement. The measurement requiring an interaction with a measuring device.

This is the part that I can't seem to get my head around. With what does the measurement device interact to give rise to the final measurement, if not the properties of the system?

I have a bit of an understanding of the idea that the system doesn't have predefined values prior to measurement and how properties might appear upon measurement. The analogy I think of is that of a single molecule of water in a dark box. We dip a piece of paper into the box and it comes out a little "wet". Here, the property of wetness only manifests upon measurement, with the interaction between the measuring device and the molecule of water, but we cannot say that the molecule was "wet" prior to measurement.

To my mind, the system must have properties to first, be considered to exist and be real, and second to interact with the measurement device.

 

[PeterDonis]

Would objects which exist and are real not have properties by necessity?

Why would they have to?

Would real and existing not be properties in and of themselves?

No. Bertrand Russell got this one right a century ago. "Existence" is a quantifier, not a property. He didn't explicitly talk about "real", but it seems to me like it would work the same way.

With what does the measurement device interact to give rise to the final measurement, if not the properties of the system?

Um, with the system?

Even in classical physics, we don't say the properties of one system interact with the properties of another system. We just say the systems interact.

 

[Lynch101]

We should distinguish between a system having no ontic state, and the ontology of the system.

When we we say there is no ontic state, what we mean is we do not model the physics of the system terms of an ontic state or ontic state space. But we can still discuss the ontology/metaphysics of the system apart from our physical models. E.g. We can say the system exists. We can say the system interacts with our measurement apparatus in a predictable way. We can discuss the system as perhaps possessing some lawmaker property that's responsible for its dynamics etc.

Argh! That's what happens when I start trying to use certain terms. Apologies for retreating to more familiar ground on this one, I keep relating it back to the statement about how its meaningless to talk about the system prior to measurement.

I wrote a post in reply to Paul Colby with a crude analogy which outlines my thinking. I hope you don't mind me referencing that to save writing it out again.

My reasoning, which is often flawed (especially when it comes to physics), is that if the mathematical formalism of QM only makes probabilistic predictions about the flashes of light that will appear at either end of a black box, then it doesn't tell us what is happening inside the box.

There are at least two conclusions that we can draw:
a) something is happening inside the box which causes the flashes
b) there is absolutely nothing happening inside the box

If we accept a) and the proposition that the mathematical formalism doesn't describe what happens inside the box it would seem to point to realistic hidden variables, even if they don't have defined values.

My understanding is that the anti-realist position says b).

 

[PeterDonis]

There are at least two conclusions that we can draw

No, there are three. Here's the third:

c) Since we are not observing or measuring anything inside the box, we cannot make any meaningful statements about whether or not anything is happening inside the box.

My understanding is that the anti-realist position says b)

I think an anti-realist would be more likely to say c). Saying that "nothing is happening" inside the box is a realist statement: it says there is a real "reality" inside the box, and nothing is happening in that real "reality". In other words, it's a positive claim about "the way things are".

Anti-realism is about not making positive claims about the way things are, unless you absolutely have to.

 

[Paul Colby]

We can reasonably ask the question, what is happening inside the box.

Well, not all questions have a meaning. For example asking classically motivated questions about quantum systems can and often have no meaning. Are electrons red? Which trajectory did the photon take? Things like that. Valid scientific questions one might ask have to be consistent and well framed relative to the current theoretical understanding.

OK, so we cannot assign predefined values

Where this actually mean assign a specific eigenvalue (or hidden variable for that matter) of the observation device prior to measurement. There are edge cases but this really isn't a point anyone would make.

If we say yes, there is something happening inside the box but the mathematical formalism only gives us information about the measurement outcomes, then this would seem to suggest that there are hidden-variables in the box that the mathematics doesn't describe.

Yeah, no.

This is where I lose the bubble. There is a system. If it's been prepared, like in an experiment, it has a known quantum mechanical state or density matrix. This is a far cry from not having anything. In all likelihood it's as much as one can know. If you perform the Schrodinger cat experiment and never open the box you will still have a dead cat in the box after some mean time. To think otherwise is counter-factual.

My understanding is that the anti-realist position must say that there is nothing whatsoever happening inside the box because the alternative would be that there are realistic hidden-variables.

Yes, if they say this they are wrong. The system is evolving in time by well understood rules. Evolving in time as in something is happening.

 

[DrChinese]

This is where I lose the bubble. There is a system. If it's been prepared, like in an experiment, it has a known quantum mechanical state or density matrix. This is a far cry from not having anything. In all likelihood it's as much as one can know.

If it is as much as we can know, then it is complete in the Einstein sense (he thought QM to be incomplete). There is no more "complete" specification of the system. This is generally accepted in pretty much all interpretations.

OK, it has a density matrix. You can call this whatever you want. But whatever you want to describe it as, it must consider Bell. The Bell requirement was that it have a specific value prior to final measurement (and a measurement choice here should not affect an outcome there). So the question to you (or anyone) is: does it? When is the value assigned?

a) If it is preassigned: you are forced to conclude that the measurement itself forced a result that must be carried to a distant entangled partner and changes it as well. That so the statistical results match the predictions of QM. So you have an FTL influence to deal with.

b) If it is not preassigned, you are fine. Of course a new problem appears, which is how the result of the distant system will meet the statistical requirements. So you now need a mechanism for that. The non-realistic interpretations attempt to supply that.

 

[Paul Colby]

does it? When is the value assigned?

Question: what is it?

An observable is assigned an operator acting on a space of state vectors. An observable + the measurement device produces one eigenvalue. So the value produced by the observable + the measuring device is assigned when the measurement is made. The measurement can be non-local as in correlations between distant devices.

b) If it is not preassigned, you are fine. Of course a new problem appears, which is how the result of the distant system will meet the statistical requirements. So you now need a mechanism for that. The non-realistic interpretations attempt to supply that.

Why isn't the standard quantum mechanics just this mechanism?

 

[Morbert]

There are at least two conclusions that we can draw:
a) something is happening inside the box which causes the flashes
b) there is absolutely nothing happening inside the box

My understanding is that the anti-realist position says b).

Anti-realism is a claim about how our quantum model relates to the system, rather than a claim about the system alone. Our quantum model consists of:

a quantum state and state space
dynamics
a sample space
event algebras

These ingredients let us assign probabilities to different possible measurement outcomes. Anti-realism is the position that the properties in our sample space do not correspond to actually existing properties of the system we are modelling, and are instead part of the procedure for computing probabilities for outcomes of measurements on the systems.

Note that this is different from your position b). A physicist can adopt an anti-realist position regarding her model of the system in the black box, while still rejecting the claim that "there is absolutely nothing happening inside the box".

If we accept a) and the proposition that the mathematical formalism doesn't describe what happens inside the box it would seem to point to realistic hidden variables, even if they don't have defined values.

Variables are a feature of models a physicist constructs, and if a physicist does not model the inner happenings of the box, then there would be no need to use hidden variables in the model.

 

[Lynch101]

Why would they have to?

That's a good question! It might just be an unquestioned assumption on my behalf. I have been giving it some thought so it would be great to try to tease it out, although I am wary of the potential for it to turn into an exercise in navel gazing . Prior to that though, would my initial summation be accurate then, that the anti-realist position says that the quantum system has no properties whatsoever prior to measurement?

Thinking about the question, I was thinking in terms of a materialist paradigm, as opposed to say the paradigm of Cartesian Dualism. Again, I might have a hidden assumption in here but I was thinking that the materialist paradigm says that all things which exist are physical i.e. they are made of material substances which have physical properties.

As the "resident realist" , I would be interested to hear if @Demystifier has any thought on the idea that things can exist without any properties. He might be able to articulate it better than myself maybe.

Um, with the system?

How can a system without properties interact with something else? What part of the system is interacting with what part of the measuring device? When preparing the system, how can something with properties pepare a system that has absolutely no properties?

I know this might seem like a re-statement of the "measurement problem" but how can the system being measured contribute to the measurement outcome, if its properties don't play a part?

If the mathematical formalism only gives us the probability of a measurement outcome then it can't really be said to describe what goes on inside "the box", can it?

We might deduce from the mathematics that the properties cannot have predefined values, but that wouldn't be the same as saying there are no properties whatsoever, would it?

Even in classical physics, we don't say the properties of one system interact with the properties of another system. We just say the systems interact.

If we think about a baseball bat striking a baseball I feel like they interact the way they do because of their respective properties, in contrast to how neutrinos don't interact in the same way because of their properties. But again, that might be down to an assumption or a heuristic on my part.

 

[Lynch101]

I've just re-arranged the order of the points you make. I hope this doesn't misrepresent your post.

I think an anti-realist would be more likely to say c). Saying that "nothing is happening" inside the box is a realist statement: it says there is a real "reality" inside the box, and nothing is happening in that real "reality". In other words, it's a positive claim about "the way things are".

Anti-realism is about not making positive claims about the way things are, unless you absolutely have to.

This would be how I interpret the instrumentalist position and would distinguish it as being "non-realist". Instrumentalism doesn't make ontological claims while anti-realism does. I would take anti-realism to be the denial of realism.

No, there are three. Here's the third:

c) Since we are not observing or measuring anything inside the box, we cannot make any meaningful statements about whether or not anything is happening inside the box.

Again to my mind, that would be the instrumentalist position. It says that we cannot choose between the two options but I would still think there can only be two options a) something happening b) nothing happening.

I'll try and clarify my thinking in terms of Schroedinger's cat. We might say that we cannot say whether the cat is alive or dead until we open the box, and we might even say that the cat is in a superposition until we measure it, but I would say that there is still something inside the box or something happening inside the box.

 

[Demystifier]

As the "resident realist" , I would be interested to hear if @Demystifier has any thought on the idea that things can exist without any properties. He might be able to articulate it better than myself maybe.

I think they can't, but I don't have a particularly enlightening way to articulate it.

 

[PeterDonis]

Instrumentalism doesn't make ontological claims while anti-realism does.

You could be right, but I would want to see some actual references to anti-realist papers.

How can a system without properties interact with something else?

I don't know, but I also don't know that it isn't possible. Questions like this have a very poor track record of getting at real issues in quantum interpretations. Pretty much any a priori assumption you make about how things "must" be is likely to be wrong.

I would still think there can only be two options a) something happening b) nothing happening.

Same comment as above.

 

[Morbert]

You could be right, but I would want to see some actual references to anti-realist papers.

Lynch101 is using a nonstandard definition of anti-realism, one that does not square with its use in literature. E.g. Asher Peres is a contemporary anti-realist and in his book "Quantum Theory: Methods and Concepts" he says

"The role of physics is to study relationships between these [objective experimental] records. Some people prefer to use the word intersubjectivity, which means all observers agree on the outcome of any particular experiment. Whether or not there exists an objective reality beyond the intersubjective reality may be an interesting philosophical problem, but this is not the business of quantum theory. Quantum theory in a strict sense is nothing more than the set of rules whereby physicists compute probabilities of the outcomes of macroscopic tests."

Anti-realism is the claim that , whether or not there is an objective reality beyond the intersubjective reality, the variables that manifest in quantum mechanics are not to be interpreted as elements of that objective reality.

What Lynch is describing is more akin to property nihilism than anti-realism.

 

[Lynch101]

Well, not all questions have a meaning. For example asking classically motivated questions about quantum systems can and often have no meaning. Are electrons red? Which trajectory did the photon take? Things like that. Valid scientific questions one might ask have to be consistent and well framed relative to the current theoretical understanding.

That's very true. There may also be questions which have meaning but which we cannot answer due to limitations on our ability to probe the world around us. I'm inclined to think that the question of what happens prior to measurement is of the second variety.

Where this actually mean assign a specific eigenvalue (or hidden variable for that matter) of the observation device prior to measurement. There are edge cases but this really isn't a point anyone would make.

I might be misunderstanding this point, but I'm interpreting "observation device" to mean measurement device. So, instead of saying that we can't assign a value to the measurement device prior to measurement, I'm talking about the quantum system prior to measurement. The idea I'm trying to get at might be termed an unquantified property. Would that be what the term "beable" refers to?

This is where I lose the bubble. There is a system. If it's been prepared, like in an experiment, it has a known quantum mechanical state or density matrix. This is a far cry from not having anything. In all likelihood it's as much as one can know. If you perform the Schrodinger cat experiment and never open the box you will still have a dead cat in the box after some mean time. To think otherwise is counter-factual.

My misunderstanding might lie in what the mathematical formalism describes.

So there is a system and, not to be pedantic, but there is a system prior to measurement. As you say, it has been prepared. My first question would be how a system without properties can be prepared; what is it that is being prepared? To my mind, a system would have to be prepared with at least some properties, any properties, even if we cannot ascribe definitive values to the properties of the system.

You mention the density matrix. Does it describe the system prior to measurement or does it - please forgive my loose description here - simply form part of the probabilistic predictions for the outcome of the experiment? My understanding is that the anti-realist position says that it only forms part of the prediction.

 

[Lynch101]

tl;dr: I might be missing something somewhere, but it reads to me that what you and others are describing as "anti-realism" is what I understand as instrumentalism or Shut up and Calculate (SUAC).

Anti-realism is a claim about how our quantum model relates to the system, rather than a claim about the system alone. Our quantum model consists of:

a quantum state and state space
dynamics
a sample space
event algebras

These ingredients let us assign probabilities to different possible measurement outcomes. Anti-realism is the position that the properties in our sample space do not correspond to actually existing properties of the system we are modelling, and are instead part of the procedure for computing probabilities for outcomes of measurements on the systems.

This is what I understand to be the instrumentalist position or the SUAC position. It says that the mathematics is just a predictive tool. If it is necessary to make a distinction, I would interpret this more as "non-realism" than anti-realism.

The SUAC position stops short of making ontological inferences from the mathematics while the anti-realist position, as I understand it, does make ontological inferences from the mathematics. It goes further than simply saying the position that the properties in our sample space do not correspond to actually existing properties of the system we are modelling, it says that they do not correspond to actually existing properties because there are no existing properties. It denies the reality of such properties, which is what makes it an anti-realist interpretation, as opposed to the non-realist, SUAC position which remains silent on such questions of the state of the system prior to measurement.

Note that this is different from your position b). A physicist can adopt an anti-realist position regarding her model of the system in the black box, while still rejecting the claim that "there is absolutely nothing happening inside the box".

I might not be picking up on some nuance, but it reads to me like saying a physicist can adopt an instrumentalist (SUAC) position while rejecting claim b). That I would completely agree with it. But that would just bring us back to the question about the description of the system prior to measurement.

Variables are a feature of models a physicist constructs, and if a physicist does not model the inner happenings of the box, then there would be no need to use hidden variables in the model.

This would then relate back to the issue Einstein had about QM not being a complete description of "the world", wouldn't it? The question about what happens prior to measurement, or what the description is, of the quantum system prior to measurement, is a question about the inner workings of the box. If the model doesn't account for the inner workings of the box then it wouldn't be considered a complete description of the system, would it?

Again, this is just my interpretation of the descriptions and arguments that I've read, and I might be missing something along the way.

 

[Lynch101]

If it is as much as we can know, then it is complete in the Einstein sense (he thought QM to be incomplete). There is no more "complete" specification of the system. This is generally accepted in pretty much all interpretations.

I think this is where I am falling between the cracks, so to speak.

I think we can ask a further question here: why is it as much as we can know? I think there are at least two answers to that question:
1) because there is a limit to our ability to probe nature.
2) because there is nothing more to know i.e. there is nothing inside the box.

Answer 1) would imply that our model is not a complete description of nature but it also says that there is no more complete specification of the system. Am I correct in saying that those parts of the system which cannot be modeled would be referred to as "beables"? I have been referring to them as properties of the system prior to measurement, but beables might be the more accurate term. I have been thinking of beables as hidden-variables, but I might be conflating two terms there.

Answer 2) is what I have come to understand as the "anti-realist" position.

OK, it has a density matrix. You can call this whatever you want. But whatever you want to describe it as, it must consider Bell. The Bell requirement was that it have a specific value prior to final measurement (and a measurement choice here should not affect an outcome there). So the question to you (or anyone) is: does it? When is the value assigned?

a) If it is preassigned: you are forced to conclude that the measurement itself forced a result that must be carried to a distant entangled partner and changes it as well. That so the statistical results match the predictions of QM. So you have an FTL influence to deal with.

b) If it is not preassigned, you are fine. Of course a new problem appears, which is how the result of the distant system will meet the statistical requirements. So you now need a mechanism for that. The non-realistic interpretations attempt to supply that.

Is it possible to talk about unquantified properties of the system. I'm thinking of properties at a fundamental level without preassigned values. I'm not exactly sure how that would work, but to my mind it seems like a conclusion that can be drawn.

 

[Lynch101]

I think they can't, but I don't have a particularly enlightening way to articulate it.

It's infuriating isn't it?! It's a bit like that quote about time " If no one asks me, I know what it is. If I wish to explain it to him who asks, I do not know."

 

[Lynch101]

You could be right, but I would want to see some actual references to anti-realist papers.

I don't know, but I also don't know that it isn't possible. Questions like this have a very poor track record of getting at real issues in quantum interpretations. Pretty much any a priori assumption you make about how things "must" be is likely to be wrong.

Same comment as above.

I guess that's part of what makes it so fascinating.

 

[Paul Colby]

I'm inclined to think that the question of what happens prior to measurement is of the second variety.

Please define what you mean by happen.

My definition is the system state vector evolves in time according to QM.

So, instead of saying that we can't assign a value to the measurement device prior to measurement, I'm talking about the quantum system prior to measurement.

And hence the eternal problem I see with many such discussions. No macroscopic measurements device, no data. These arguments are like the sound of one hand clapping.

System preparation, where to begin? For Schrodinger's cat experiment you've selected some cats to kill and put them in boxes and filled the poison vials. For a particle experiment you've turned on and tuned the accelerator and selected a target etc.

My first question would be how a system without properties can be prepared;

Again, define what you mean by system property. I have no problem finding cats.

In QM properties involves operators and state vectors plus an implied ideal measurement device which is always macroscopic AFAICT.