I The typical and the exceptional in physics

[atyy]

As I said before, in Landau Lifshitz I cannot find the word collapse (by searching the electronic copy I have ;-)), and Weinberg holds the view that the question in interpretational issues is undecided (after a brilliant analysis in an early chapter of his QM lectures book). Then he happily goes on using the standard representation.

I don't think that a Heisenberg cut is in any way justified by the formalism of QT nor is it in anyway justified by observations. It's just a matter of technological challenge how to sufficiently isolate macroscopic systems from perturbations to avoid decoherence to demonstrate quantum effects also on them. You have to decide from case to case at which point in an experimental setup you can treat things (semi-)classically. I think that Bohr was right saying that a measurement apparatus should be within the validity of the semi-classical description. It must be an open system such that you can store the information on the measurements made, which is an irreversible process.

The interaction between measured object and the macroscopic measurement device are part of the Hamiltonian and as such, according to the successful relativistic QFT, a local interaction. The assumption that such a local interaction can cause far-distant instantaneous responses is thus a contradiction in adjecto. According to the standard (minimal) interpretation there's also no need to explain it by far-distant correlations as described by entanglement. It's all standard QT (or functional analysis if you wish).

I replied to this above, but add another comment to show that it is very clear that Weinberg's standard interpretation has a Heisenberg cut p81, section 3.7:

"The discussion of probabilities in Section 3.1 was based on what is called the Copenhagen interpretation of quantum mechanics, formulated under the leadership of Niels Bohr. According to Bohr, “The essentially new feature of the analysis of quantum phenomena is ... the introduction of a fundamental distinction between the measuring apparatus and the objects under investigation."

So LL and Weinberg both have the Heisenberg cut. The Heisenberg cut is part of the standard or minimal interpretation.

 

[zonde]

The interaction between measured object and the macroscopic measurement device are part of the Hamiltonian and as such, according to the successful relativistic QFT, a local interaction. The assumption that such a local interaction can cause far-distant instantaneous responses is thus a contradiction in adjecto. According to the standard (minimal) interpretation there's also no need to explain it by far-distant correlations as described by entanglement. It's all standard QT (or functional analysis if you wish).

In order to speak about local measured object you first have to trace out the distant part and you trace it out in particular basis. Now let's say that you trace out the distant part and after that you measure local object in different basis than the one used for tracing out. And then do the same for distant part. Can you still come up with correct predictions for correlation?

 

[vanhees71]

I don't understand what you want to say here. It's very clear how the far-distant parts of an entangled system are to be "traced out" to get the statistical operator of the corresponding other part of interest. It's also clear, how A updates her knowledge about the entire system, including B's part after gaining information by a measurement. For B to make use of A's information he needs to get this information from A since he cannot instantaneously get it somehow from his system which is far distant from A (due to the locality of interactions in relativistic QFT, which I consider valid). The correlation is inherent in the system from the very beginning by preparing it in the entangled state; it's not due to the local measurments of either A or B.

 

[vanhees71]

I replied to this above, but add another comment to show that it is very clear that Weinberg's standard interpretation has a Heisenberg cut p81, section 3.7:

"The discussion of probabilities in Section 3.1 was based on what is called the Copenhagen interpretation of quantum mechanics, formulated under the leadership of Niels Bohr. According to Bohr, “The essentially new feature of the analysis of quantum phenomena is ... the introduction of a fundamental distinction between the measuring apparatus and the objects under investigation."

So LL and Weinberg both have the Heisenberg cut. The Heisenberg cut is part of the standard or minimal interpretation.

In this sense, of course the Heisenberg cut is there. The only thing I see no justification for is to claim that on a fundamental level there is a "quantum world" and a "classical world" governed by different dynamical laws. I think the classical behavior of macrocopic objects under usual conditions is a phenomenon that can be understood from QT, using the standard ("coarse-graining") techniques of many-body theory.

 

[zonde]

It's also clear, how A updates her knowledge about the entire system, including B's part after gaining information by a measurement. For B to make use of A's information he needs to get this information from A since he cannot instantaneously get it somehow from his system which is far distant from A (due to the locality of interactions in relativistic QFT, which I consider valid). The correlation is inherent in the system from the very beginning by preparing it in the entangled state; it's not due to the local measurments of either A or B.

Let's get it straight. There are two sides of the situation we try to model. One is physical situation that results in measurement records of A and B. The other side is how we analyze obtained measurement records by correlating them. This other side is of no concern for physics theory. Look at it this way: some hypothetical physical mechanism results in certain related measurement records for A and B. We test that this physical mechanism works as expected by looking at correlation between A and B records.

 

[stevendaryl]

In this sense, of course the Heisenberg cut is there. The only thing I see no justification for is to claim that on a fundamental level there is a "quantum world" and a "classical world" governed by different dynamical laws. I think the classical behavior of macrocopic objects under usual conditions is a phenomenon that can be understood from QT, using the standard ("coarse-graining") techniques of many-body theory.

But your rules for working with quantum mechanics are different for macroscopic variables and microscopic variables. A macroscopic variable such as a pointer position has a definite value at all times, while that value may only predictable from past states probabilistically. In contrast, with a microscopic variable such as the z-component of an electron's spin, it doesn't make any sense to say that it has a definite value at all times, it only makes sense to say that when that variable is coupled to a macroscopic value (via a measurement), it acquires a definite value.

I think it is false to say that the properties of macroscopic variables are derivable from the properties of microscopic variables through coarse-graining. If it were true that the same physics applies to both, you should be able to formulate the laws of physics in a way that is independent of whether you are talking about micro or macro. In contrast, the minimalist interpretation involves both.

 

[vanhees71]

A pointer variable of a macrocopic measurement device has on a very coarse grained level a definite value at all times. The statistical (quantum and thermal) fluctuations are much smaller than the required accuracy of this pointer variable's value!

 

[stevendaryl]

To me, the issue about macroscopic versus microscopic is pretty straight-forwardly illustrated by an EPR experiment involving anticorrelated spin-1/2 particles.

Alice measures her particle's spin along the z-axis, and gets spin-up. The question is: Is that the result, "Alice measured spin-up", an objective, physical fact about the universe? In the minimal interpretation, it is treated as an objective physical fact. But in contrast, for an electron that is in a superposition of spin-down and spin-up, neither "The electron is spin-up" nor "The electron is spin-down" is considered an objective physical fact about the universe. People often say that it has no spin until it is measured. So this seems to be a big distinction between microscopic objects and macroscopic objects: macroscopic objects have definite, objective properties, while microscopic objects do not.

I understand that macroscopic objects can't (for long) be in superpositions, because decoherence will rapidly cause correlations between the macroscopic object and a larger chunk of the universe. But decoherence doesn't make properties such as spin more definite and objective, it just enlarges the size of the system that must be considered to be in an indefinite state.

It seems to me that there are only three possibilities:

1. There is some physics that governs macroscopic interactions (measurements) that doesn't apply to microscopic systems (objective collapse theories, for example), or
2. Contrary to what is commonly believed, macroscopic systems do not have definite properties, either (that's Many-Worlds), or
3. Contrary to what is commonly believed, microscopic systems do have definite properties (Bohmian mechanics, for example).

I can understand people who say that there is no need to resolve the question, since we have a recipe for using quantum mechanics that works well enough without answering the question. But I don't understand the people who claim that there is no issue to resolve in the first place.

 

[vanhees71]

You are mixing theories all the time. According to classical physics, applicable to macroscopic physics (in a sense to be understood below!), any observable of an object has a definite value.

Quantum theory extends the validity of our theory tremendously beynd the range of validity of classical physics. Thinking about how the restricted validity of classical physics can be understood from the more comprehensive (in fact today there's no limit of validity of QT known, except our inability to construct a satisfatcory quantum description of gravity), leads at least me to the conclusion that classical physics is valid in an average coarse-grained sense. You get classical behavior of observables averaged over many appropriate microscopic observables. Then very often these coarse grained macroscopic observables are sufficient to effectively describe the behavior of the macroscopic object in terms of classical dynamics. Within the accuracy of the validity of a coarse-grained description the statistical fluctuations of the corresponding macroscopic observables are negligible, and Ehrenfest's theorem leads to the validity of the classical dynamics in this sense.

 

[Ken G]

I don't see how paradigm building and shifts, or even just mathematical intuition, could be reduced to noticing sameness and differences... it's a variety of qualitatively different operations that come into play.

Here I just mean, when Copernicus makes a model that says orbits are circles with the Sun at the center, the samenesses are the circles, and the differences are the different scales of the orbits. Call it symmetries and breaks in symmetries, if you like.

Actually I don't see why, in this context, we should worry about the nature of thinking itself! All we need to know about epistemology is that it concerns our ways of knowing, which involves a plurality of factors. We refer to that knowing with respect to the operations we perform in the lab: we can leave it at intuition, it's even simpler than having an ontology to worry about. We don't need a theory of mind to do physics, why are you worrying about it?

I agree that trying to put in some kind of theory of thinking is premature, we just don't know enough about the mind. I think it will advance physics dramatically once we understand better how we do it, but that's not going to happen in this thread, and maybe not for a thousand years for all I know. Your point is that it is a detour from the basic questions of ontology vs. epistemology, and that's true. So getting back to epistemology, the core issue here is that we like to motivate our epistemologies with ontologies, and my only point is that we always run into trouble when we take the ontologies too seriously by framing them as the purpose of the endeavor. We should instead regard them as tools of understanding, effective pictures we use as we think, because this solves all the problems we have with ontology-- in particular, it justifies why ontologies are always nonunique, and it explains why they invariably end up getting replaced with others that are almost completely different. The epistemology of science is to both seek out and discard useful ontologies, but not because we are looking for the right one, any more than a hermit crab is looking for the right shell.

 

[stevendaryl]

You are mixing theories all the time. According to classical physics, applicable to macroscopic physics (in a sense to be understood below!), any observable of an object has a definite value.

Quantum theory extends the validity of our theory tremendously beynd the range of validity of classical physics. Thinking about how the restricted validity of classical physics can be understood from the more comprehensive (in fact today there's no limit of validity of QT known, except our inability to construct a satisfatcory quantum description of gravity), leads at least me to the conclusion that classical physics is valid in an average coarse-grained sense. You get classical behavior of observables averaged over many appropriate microscopic observables. Then very often these coarse grained macroscopic observables are sufficient to effectively describe the behavior of the macroscopic object in terms of classical dynamics. Within the accuracy of the validity of a coarse-grained description the statistical fluctuations of the corresponding macroscopic observables are negligible, and Ehrenfest's theorem leads to the validity of the classical dynamics in this sense.

That's just not true. If an electron is in a superposition of spin-up and spin-down along the z-axis, and it interacts with a measuring device, then the measuring device will evolve into a superposition of "measuring spin-up" and "measuring spin-down". Decoherence then would propagate the indefiniteness to the rest of the universe---the universe would evolve into a superposition of one universe in which the measuring device measures spin-up and another universe in which the measuring device measures spin-down. Coarse graining is not going to change that. It's a complete red herring to bring it up.

If it is true that macroscopic objects obey the same physics as microscopic objects, then a many-worlds type ontology follows. You want to affirm one side of an implication and reject the conclusion, but it is incoherent to do so. Bringing up coarse-graining and decoherence doesn't change the conclusion.

 

[ddd123]

Here I just mean, when Copernicus makes a model that says orbits are circles with the Sun at the center, the samenesses are the circles, and the differences are the different scales of the orbits.

What about General Relativity then?

Call it symmetries and breaks in symmetries, if you like.

When you devise a new symmetry, you have to think about it. Whether it's a new 'sameness' or a new 'difference', you have to think about it with a mix of creativity, intuition and logical deduction. It's not like saying '0' or '1' and mixing them up. I mean, at least if you're not going into stuff like generating proofs of theorems using brute force within a Godel representation of maths or something like that (but even if that worked, it would work for maths, physics is more fuzzy and involves more intuitive concepts).

We should instead regard them as tools of understanding, effective pictures we use as we think, because this solves all the problems we have with ontology-- in particular, it justifies why ontologies are always nonunique, and it explains why they invariably end up getting replaced with others that are almost completely different. The epistemology of science is to both seek out and discard useful ontologies, but not because we are looking for the right one, any more than a hermit crab is looking for the right shell.

I think the final point is good, but it's not going to be convincing if you avoid the object of the process: understanding what? Thinking what? By avoiding that sort of question you end up putting it in the spotlight: instead of focusing on matter, you focus on mind, and then suggest that there's no separate reality being described etc (which implies that subject and object are interdependent, which is a philosophical theory, which is outside the scope of physics). In short, you're not agnostic enough. If you're saying that we're not describing an outside reality, and if there is an outside reality (which you're agnostic about), then you're making the ontological statement that we are NOT describing it. Half of your position is subtly ontological. Maybe your readers don't go this far into the nature of your arguments but they sense that you're still trying to superimpose a no-picture picture on reality and this ends up being unconvincing.

It's much smoother, IMHO, to go like this:

We start from intuitive concepts like 'measurement outcome' and 'experimental preparation' in the abstract; to that semantics we associate the structured and collected information which we model by devising a theory (which may also involve a picture as an intuition aid, but of course the picture is in our heads, not outside of our heads).

Here I made no claims whatsoever on the ontology and not even on the relationship between epistemology and ontology: it is pure epistemology in the proper sense, that I'm not even separating epistemology and ontology, which necessarily would be a partially ontological claim.

 

[Ken G]

It seems to me that there are only three possibilities:

1. There is some physics that governs macroscopic interactions (measurements) that doesn't apply to microscopic systems (objective collapse theories, for example), or
2. Contrary to what is commonly believed, macroscopic systems do not have definite properties, either (that's Many-Worlds), or
3. Contrary to what is commonly believed, microscopic systems do have definite properties (Bohmian mechanics, for example).

I can understand people who say that there is no need to resolve the question, since we have a recipe for using quantum mechanics that works well enough without answering the question. But I don't understand the people who claim that there is no issue to resolve in the first place.

The way to make the problem go away, more so than "resolve" it, is to look for a fourth possibility: that physics doesn't govern things, and that systems do not have properties. Or more correctly, that when we use anthropomorphic language about "laws" and "governances", we are finding associations between our experiences doing experiments and the basic structures with which we have everyday familiarity, and when we talk about "properties" we mean "contraints on how we can successfully think about systems." Using these more precise replacements, replacements that actually dovetail with what we can observe when we watch a scientist do science, we can still use the language to do all the same things for our science, but we can avoid the quagmire of taking our ontologies too seriously. None of this causes any trouble if we frame it all as modes of thought and approaches to manipulating information that involve pictures, cartoons really. The cartoons should have captions such as "I like to picture what is happening like this", but the problems only appear when that correct and well-tested caption gets simplified to the invariably incorrect "this is what is actually happening."

 

[Ken G]

That's just not true. If an electron is in a superposition of spin-up and spin-down along the z-axis, and it interacts with a measuring device, then the measuring device will evolve into a superposition of "measuring spin-up" and "measuring spin-down".

That's not strictly correct, the measuring device itself is not in a superposition state, only the wavefunction that includes the measuring device (if there is any such thing "in reality", which is very much the question). If you project onto the degrees of freedom of the measuring device, you get a mixed state, not a superposition. However, that's not really the problem, the problem is in deciding what that mixed state means-- is it the actual state of the measuring device, on grounds that a measuring device needs an actual unique state in the reality (which I reject)? If so, then "collapse" hasn't happened yet, it only happens when we look, and get a single outcome, returning the measuring device to a unique state. In my view, the need to regard anything as having a state stems purely from our desire to be able to create correct expectations about that thing, and is in no way some kind of requirement of reality, if there's even any way to give that latter language physical meaning.

Decoherence then would propagate the indefiniteness to the rest of the universe---the universe would evolve into a superposition of one universe in which the measuring device measures spin-up and another universe in which the measuring device measures spin-down.

This is the pre-collapse state, if one takes a universal wavefunction seriously. But the problem hasn't appeared yet-- the problem is when we look at the outcome and only see one. Now we need an interpretation, because our description of this uber-superposition is no longer gibing with our perceived outcome. We can now say that our outcome only represents a small fraction of what is actually happening, in which case we are forced to conclude that what we care about (what happens to us) is a limited amount of the full information. But it's the information we have! So we start with ontology, and are led back to epistemology-- our information is all that matters to us. The ontology has become useless!

Or, we take the Copenhagen view, and say that our information, which is what matters, must be everything that happens. Here we have made ontology matter, but only by shoving epistemology down its throat-- so it really doesn't matter here either, the ontology is so subservient to the epistemology that all that remains is the epistemology anyway!

Or, we can take the Bohmian view, and say that we don't have all the information, so the uber-wavefunction you describe never happens. What happens to us is determined by information we have no access to. So we do achieve an ontology that goes beyond the epistemology, but we do it in the usual way-- by postulating the existence of essentially invisible and unknowable higher powers, here acting in the form of details of the preparation that we could never know. "Preparations work in mysterious ways," where have we seen that before?

So the bottom line is, either the ontology is no more than the epistemology, or anything more that it is becomes a matter of essentially religious interest only.

If it is true that macroscopic objects obey the same physics as microscopic objects, then a many-worlds type ontology follows.

Yet we must carefully track all the suppositions that are implicit in that hypothetical:
1) that there are such things as "objects" and they can be either micro or macroscopic, they are not just concepts we use to manipulate information
2) that objects "obey" laws, as in they are in some form of communication with these immutable laws, rather than "obeying laws" being a familiar language we can use to make sense of the behaviors we see
3) that our current version of those "laws", the Schroedinger equation, is not just the current approximation that is spectacularly accurate in isolated instances, it is the actual immutable law that the actual objects are actually in some form of communication with.

So yes, you do get a many-worlds ontology if you make all those assumptions, but identifying precisely what the assumptions actually are clarifies greatly why we should not be surprised they lead us to a bizarre ontology. Such is always the way.

 

[Ken G]

I was walking across the golf course the other day and then suddenly this golf ball hit me right between the eyes. My doctor told me not to worry about the huge lump on my forehead as it was only epistemic

Epistemic means how your mind interacts with its stimulus . If anyone thinks it doesn't matter how your mind interacts with its stimulus, I have to wonder what they have been doing with their life all this time! When someone has late-stage terminal cancer, and is told they have only a week to live, and it will be the most painful and awful week of their entire life, and they are given an option to take a drug that will interfere with their brain's ability to detect that their body is dying in awful ways, do they care more about the ontology of the death of their body, or the epistemology of how they get to feel during that process? We always care more about epistemology than ontology, epistemology is no less than everything that matters to us, even if the true believer in ontology must see a dose of self-deception in epistemology.

 

[Ken G]

What about General Relativity then?

It stems from the equivalence principle, a classic example of our mind's ability to detect "sameness." But still, I agree this is a detour.

I think the final point is good, but it's not going to be convincing if you avoid the object of the process: understanding what?

If it is less convincing, that may be simply because we are so set in our ways of thinking even when they become barriers. (And I wager we've alll seen that phenomenon!) I'm claiming that understanding is just understanding, period, and claiming that it is understanding "of something" is just the nature of how we understand. If someone understands the fundamental theorem of integral calculus, then what is the "something" they understand? The fundamental theorem of integral calculus, of course-- but that's just more information. We understand by manipulating information, and what we understand is more information-- how could it be any different? How can manipulating information transcend itself? The whole idea of a "something" that we understand is just that-- it is an idea. We can see this, we merely watch someone manipulate that idea, and say to ourselves "yup, they are clearly manipulating an idea there."

By avoiding that sort of question you end up putting it in the spotlight: instead of focusing on matter, you focus on mind, and then suggest that there's no separate reality being described etc (which implies that subject and object are interdependent, which is a philosophical theory, which is outside the scope of physics). In short, you're not agnostic enough.

That's what I don't agree with. If I have no need to regard what you mean by a "separate reality" as anything but your mode of interpreting and organizing the consistent information of your senses (and I think you can agree that is what you are doing there), then why do I need anything more than just that? It's not that I don't need a concept of a separate reality, that's a vastly useful epistemological tool-- it's that I don't need to regard it as anything more than an epistemological tool, because that's the only way it ever gets used. That's agnosticism-- we only take what we need, what we actually use.

If you're saying that we're not describing an outside reality, and if there is an outside reality (which you're agnostic about), then you're making the ontological statement that we are NOT describing it.

I don't say I'm not describing outside reality, I'm saying I never use that I'm describing outside reality as it actually is, I only use what I mean by outside reality-- which is not "outside" me because I'm the one meaning something by it. I'm saying that all I ever use is the concept of an outside reality, a concept I need not take literally or seriously. I can just say "picture matter as though it were made of atoms", or "picture the sensory input you are experiencing as if it were coming from some separate outside reality that is independent of your sensations," and I do just fine. I get everything the true believer in ontology gets, without all the problems that the true believer in ontology has to contend with.

Now, sometimes people hear this wrong, and think I'm saying that if these are just pictures, then I shoudn't care if I drop a rock on my toe. But of course that's wrong, I do care if I experience pain, and the way I make sense of that experience is to say that a rock fell on my toe. I can avoid future pain using that picture, so I am in no way saying "don't model and test", I am saying "do model and test, but there's no need to pretend you are doing anything else."

We start from intuitive concepts like 'measurement outcome' and 'experimental preparation' in the abstract; to that semantics we associate the structured and collected information which we model by devising a theory (which may also involve a picture as an intuition aid, but of course the picture is in our heads, not outside of our heads).

And this is where we agree, I see a completely epistemological approach in your words there, and I would have put it very similarly. So the point is, notice how easily we frame the interpretations of QT in this light-- we say that a wavefunction is an informational structure that we use in the context of the Schroedinger equation and the Born rule, also informational structures, to make sense of the informational structure we like to call "reality." The way we manipulate information is present at every step, along with the limitations of our mental capacities, and the ultimate arbiter of our success is the outcome of observations-- which we have no need to regard as anything but yet another informational structure. The desire to take ontology seriously in any of this is essentially a form of religion, and leads to difficulties in the interpretations. All we need to use ontology for is to generate a nonunique picture that helps guide our sense of understanding, a sense that is itself purely epistemological but at the same time very important to us.

And to clarify, the act of noticing that it is epistemological, meaning it is the way we think, does not give us the power to think any way we like-- our epistemology is constrained to fit the information structures we are actually manipulating. So we create the notion of an objective reality to come to terms with why we cannot control the information we are manipulating, but of course that notion is just another example of how we manipulate the information that there are constraints on the information we manipulate. If someone says "all I mean by objective reality is the observation that I do not have complete control over the inputs to my senses, instead there are constraints that I share with other beings like myself", then that is a perfectly epistemological approach-- and would be refreshing to see in the context of QT interpretations because it puts the focus right where it belongs: how the mind choosing the interpretation copes with the constraints present in the informational structures they are attempting to understand. Above all, the information they are grappling with is not external to themselves, because information is not just a sequence of 0s and 1s, it is how we manipulate a sequence of 0s and 1s, the meaning we give to it.

 

[ddd123]

That's what I don't agree with. If I have no need to regard what you mean by a "separate reality" as anything but your mode of interpreting and organizing the consistent information of your senses (and I think you can agree that is what you are doing there), then why do I need anything more than just that? It's not that I don't need a concept of a separate reality, that's a vastly useful epistemological tool-- it's that I don't need to regard it as anything more than an epistemological tool, because that's the only way it ever gets used. That's agnosticism-- we only take what we need, what we actually use.

That sounds more like pragmatism. But I simply think your approach is too convoluted: you have to say, okay I have this characterization of outside reality but I don't characterize the characterization as actually referring to an outside reality... It's so duplicated. It may simply be a matter of words, but look at how much simpler my version of epistemology is. What I renounce in it is simply talking about ontology at all, not even in the negative.

The desire to take ontology seriously in any of this is essentially a form of religion, and leads to difficulties in the interpretations.

Calling it a religion is kind of aggressive (both towards religion and non-religion), and I think it doesn't hold in the sense of dogma. You say ontology is not provable... At some point nothing is provable. It's just easier but not more justified to say that ontology is not provable: if you doubt it characterizes external reality by doubting it possible to characterize external reality at all, the case is closed. Epistemology has the testing ground of it being effective in knowing: but at this point you can doubt your criteria of effectiveness, maybe you're predicting results that are completely skewed from an experimental perspective, because you put in confirmational biases and so on... You can never be fully sure that your epistemology makes sense, or even that epistemology at all makes sense (how would you know, you'd need to apply epistemology to itself...).

It's much more honest to say: having an ontology is a philosophical position, having an epistemology is a philosophical position, I think this is stronger, more justified than that, for these reasons...

 

[A. Neumaier]

A macroscopic variable such as a pointer position has a definite value at all times

Only at nearly all times. During the (positive) time a measurement is in progress, the pointer reading is not well-defined. Even for classical measurements one has to wait till a sufficiently stationary situation has been achieved, before one can take a reliable reading.

 

[A. Neumaier]

you should be able to formulate the laws of physics in a way that is independent of whether you are talking about micro or macro.

Such a formulation is indeed given by my thermal interpretation.

 

[Ken G]

That sounds more like pragmatism. But I simply think your approach is too convoluted: you have to say, okay I have this characterization of outside reality but I don't characterize the characterization as actually referring to an outside reality... It's so duplicated. It may simply be a matter of words, but look at how much simpler my version of epistemology is. What I renounce in it is simply talking about ontology at all, not even in the negative.

But we have to talk about ontology, because it's everywhere. We have to be able to converse with someone who says "I believe my life is embedded in a vast superposition of lives of a spectacular number of people very similar to me, on planets very similar to Earth, in universes very similar to mine, because that is what the Schroedinger equation says must be true." So it is to those people I aim my words-- I see their ontology as doing more than motivating how they solve the Schroedinger equation, I see it as a world view that could affect the decisions they make-- in potentially frightening ways, quite frankly (consider, for example, Tegmark's "quantum suicide" paradigm!). So if those people regard themselves as scientists, I want to say to them, but notice that science never uses your world view, it only ever uses the things that matter to you, which are the things you perceive as actually happening. Hence, what you perceive as actually happening is more important than what you regard as actually happening in some more "absolute" sense. The recognition of the difference between what we care about and what is "True" is the path from ontology to epistemology, so we must start in the one place to get to the other.

Calling it a religion is kind of aggressive (both towards religion and non-religion), and I think it doesn't hold in the sense of dogma. You say ontology is not provable...

It's not even testable, which is more to the point. All we can test is that information A is close to information B, that's it. Of course we wish to picture what is going on in that test, and we regard there to be lessons to be learned, but we don't need to take the ontology seriously in science. I am not badmouthing religion, I have no issue with ontology in religion-- I am saying that religion is the proper sphere for ontology whenever the ontology is taken seriously, rather than purely as a cartoon that supports the epistemology (and often in a nonunique way).

It's much more honest to say: having an ontology is a philosophical position, having an epistemology is a philosophical position, I think this is stronger, more justified than that, for these reasons...

Then let me put it this way: having an epistemology is the philosophy of doing science, but having an ontology, in the sense of taking it seriously, is never consistent with the philosophy of doing science-- even though this is not widely recognized.

 

[ddd123]

But we have to talk about ontology, because it's everywhere. etc

Well, quantum suicide isn't science because you can't publish your results!

If you want to talk people out of suicide it's great but it isn't science, it's social aid. Or, more to the point, what you're talking about is philosophy. In that context I start talking about ontology and discussing this topic in much more detail than this. But concerning hard science I stop at what I've said above.

I am saying that religion is the proper sphere for ontology whenever the ontology is taken seriously

No, philosophy is.

 

[Ken G]

No, philosophy is.

Fair enough, but it is not important to me to parse between religion and philosophy, it is only important to parse between science and non-science. The crux of my argument is that science is always pure epistemology, and taking an ontology seriously is always a brand of philosophy that is not science. I mentioned religion purely as an example that most people would agree is not science, but if that example raises flags, ignore it-- the important point is that we leave science when we do ontology (in the serious sense, not the sense of "picture this to help you understand"). What is so ironic about this point is that it is popular to characterize science as a path to arrive at ontology! That's wrong, and I don't think the history of science could possibly be much clearer on that.

 

[ddd123]

Fair enough, but it is not important to me to parse between religion and philosophy, it is only important to parse between science and non-science. The crux of my argument is that science is always pure epistemology, and taking an ontology seriously is always a brand of philosophy that is not science. I mentioned religion purely as an example that most people would agree is not science, but if that example raises flags, ignore it-- the important point is that we leave science when we do ontology (in the serious sense, not the sense of "picture this to help you understand"). What is so ironic about this point is that it is popular to characterize science as a path to arrive at ontology! That's wrong, and I don't think the history of science could possibly be much clearer on that.

Well, this I think as well, but I don't want to impose my view on science, if anything because some ideas may come to you only if you really believe in them.

But I cannot argue on this because I'm not a proponent of ontology, and it would be awkward to take that side. A pro-ontology user should answer you on this.

 

[Simon Phoenix]

Epistemic means how your mind interacts with its stimulus

That's not how I understand the onto/epi distinction within the context of physics. An ontic state is one that has some objective existence in the world - as Matt Leifer puts it if all the conscious beings in the world suddenly vanished an ontic state would still exist. An epistemic state is one that describes our current knowledge of a system; it is not something that exists in the external physical world.

So when we talk of a phase space point in classical physics this is an ontic state - it's a description that assumes a given particle has a specified position and momentum. If we suddenly ceased to exist those particles would still have a position and a momentum in this perspective.

In QM the ontic/epi argument is really about whether the wavefunction is one of these ontic states - or put more crudely "is the wavefunction real?" or is it just some mathematical device that encodes our knowledge? If we ceased to exist would objects still be 'in' quantum states?

If we hadn't needed QM to actually describe physical processes (in an ontic or epistemic fashion, whichever takes your fancy) then we wouldn't even be having this discussion since it's clear that classical physics deals with ontic states - or assigns probabilities to those states.

If you're going to suggest that the phase space point (p,q) of classical physics is an example of an epistemic state, then we'll just have to agree to disagree :-)

 

[Ken G]

That's not how I understand the onto/epi distinction within the context of physics. An ontic state is one that has some objective existence in the world - as Matt Leifer puts it if all the conscious beings in the world suddenly vanished an ontic state would still exist. An epistemic state is one that describes our current knowledge of a system; it is not something that exists in the external physical world.

What I mean by how the mind reacts to stimulus is the internal experience of having knowledge of that stimulus. I don't mean how someone else would invoke ontological pictures to be able to get an epistemic understanding of how my mind reacts to stimulus, I mean how my mind experiences stimulus. This is the whole point of the "collapse" in QT-- you never have any collapse until you look at how the mind experiences the sensation of a given experimental outcome. The origin of collapse is someone saying "hey, how come I got a single outcome from a theory that talks about a mixed state of outcomes?" You never get that at all until you include the experience of a single outcome.

So when we talk of a phase space point in classical physics this is an ontic state - it's a description that assumes a given particle has a specified position and momentum.

Not necessarily. A phase space can easily be turned into an epistemic tool, you simply say "let us treat this system as though it were a point in an abstract space with the following intellectual properties". The properties are then properties of the epistemology, not properties of the system. That subtle transformation completely escapes all the ontological quagmires of various interpretations of both classical and quantum physics, in one fell swoop.

If we suddenly ceased to exist those particles would still have a position and a momentum in this perspective.

Not in the way I just framed it-- yet I lost nothing in my power to do physics, I merely escaped making an assumption I never needed and never used (that the abstract space corresponded to something that would continue to exist if no mind conceived of it).

In QM the ontic/epi argument is really about whether the wavefunction is one of these ontic states - or put more crudely "is the wavefunction real?" or is it just some mathematical device that encodes our knowledge? If we ceased to exist would objects still be 'in' quantum states?

Yes, precisely. Notice how silly the entire question becomes if you simply stir in a dose of skepticism that quantum mechanics won't be replaced in a few hundred years by something almost completely different, as is the typical pattern in science.

If you're going to suggest that the phase space point (p,q) of classical physics is an example of an epistemic state, then we'll just have to agree to disagree :-)

Saying that it either is or isn't is ontological thinking. Saying that we can regard it either way is epistemic thinking. The latter is clearly correct though-- we clearly can regard a phase space as either an ontological or an epistemic entity, just like with a wavefunction! I'm saying you get all the problems one way, that you don't get the other way.

 

[secur]

In QM the ontic/epi argument is really about whether the wavefunction is one of these ontic states - or put more crudely "is the wavefunction real?" or is it just some mathematical device that encodes our knowledge? If we ceased to exist would objects still be 'in' quantum states?

To be careful, instead of "wavefunction" we should say "something corresponding to the wavefunction, which is basically equivalent to it". Otherwise hair-splitters will deny wavefunction and talk instead of Hermitian operators, or Bohmian beables, or whatever. The point (I believe) you want to make is: the ontic/epi argument asks whether "something basically equivalent to" the wavefunction must be ontic.

By the way,

I was walking across the golf course the other day and then suddenly this golf ball hit me right between the eyes. My doctor told me not to worry about the huge lump on my forehead as it was only epistemic

No doubt the doctor actually said it was only epidermic, i.e. didn't break the skin. But when you replied "Of course, nothing is real. No lump, no golf ball, no doctor. Only quantitative measurement!" he realized the brain damage was a lot worse than he thought :-)

 

[Simon Phoenix]

Notice how silly the entire question becomes if you simply stir in a dose of skepticism that quantum mechanics won't be replaced in a few hundred years by something almost completely different, as is the typical pattern in science.

Yes - but the question is really whether there is anything about the structure of QM that forces us to ascribe an ontic character to the state (and that's independent of whether QM is replaced some way down the line) - but if you don't accept that the phase space point (p,q) of classical physics is an ontic state - then the whole discussion is somewhat irrelevant, because that's an example of what the quantum foundation people mean when they use the term 'ontic state'. If you think that the whole concept of an 'ontic state' can be transformed using your terminology into an 'epistemic state' then you're probably wondering what on Earth all these foundation people are wasting their time on :-)

What we're looking for is some kind of theorem that can settle the issue (or at least bound things) with the same kind of power and clarity that Bell used in his assassination of locally realistic hidden variable theories. And in the context of this discussion, Bell's treatment using hidden variables would definitely be described as an ontological model.

 

[ddd123]

but if you don't accept that the phase space point (p,q) of classical physics is an ontic state - then the whole discussion is somewhat irrelevant, because that's an example of what the quantum foundation people mean when they use the term 'ontic state'. If you think that the whole concept of an 'ontic state' can be transformed using your terminology into an 'epistemic state' then you're probably wondering what on Earth all these foundation people are wasting their time on :-)

I have explained this earlier:

When, for example, we talk about falsification, or even more radically, like in this thread, about ontic vs. epistemic interpretations, we are using concepts directly borrowed from philosophy which come with a baggage that isn't necessarily carried along with them properly across the border.

Ontic and epistemic are concepts borrowed from philosophy, possibly without understanding philosophy: this can be an issue. They lose their originary meaning.

Ken G is using "ontic" in the philosophical sense, not in the foundations of physics sense. If he is onto something, it means that the assumption behind all these talks, that there is a distinction between epistemic and ontic states in physics, is void, and thus also the discussions and theorems are based upon a meaningless premise.

The basic philosophical idea is that our senses are turned by our brain into intelligible experience as a representation, which not necessarily (and most likely not) corresponds to "what is out there". A point in space is an idea within our minds: space itself is an idea within our minds. What is out there, if there is such a thing at all, is outside the scope of our mind's objectification, it's intangible. So, for Ken G, devising this strange distinction within our minds of "what's really there" vs "a mental picture of what's really there" is kind of absurd. My answer to him is that, subtly, he's still trying to rein in science in a way that's not proved is justified, if the scope is advancing in science, developing new and successful theories: maybe such an absurdity works in that sense. Let me post an illustration from the official site of the 2df Galaxy Redshift Survey experiment, which shows that also scientists can be aware of this issue:

 

[secur]

@Ken G's point is, indeed, philosophical. But if you're going to get just one idea from philosophy, this is the one to get. For my taste Bishop Berkeley said it best in "A Treatise Concerning the Principles of Human Knowledge", whence his famous aphorism "Esse est Percipi". His take on it was called "Subjective Idealism". However moderns might find him incomprehensible (like Shakespeare). The 2dFGRS cartoon above might be more appropriate. Also movie "The Matrix". After all it's simple enough.

Since all our information comes very indirectly through the senses, we don't know what's out there (if anything) and never will. That's it. Something so obvious is impossible to "prove", although @Ken G's persistent defense of it has done some good.

Modern concepts like atoms, statistical thermodynamics, EM radiation, relativity and especially QM have made it very relevant to science. Once you get it, simply do the following. There are a lot of things we think of as ontically real, like people, rocks and wavefunctions. Continue calling them "real" but translate, in your thoughts, to "appears real to me, based on the persistent and convincing data of my senses, but of course there's no way to tell if it's really real." No further adjustment is needed. This paradigm shift can never reveal any new scientific facts, but it automatically ends a common type of endless debate.

If everybody accepts this fact you never need mention it again. We can happily continue to use words like "real", "ontic", and "epistemic", mutatis mutandis. However if you, or anyone, won't accept it, that's alright. We'll just translate your concept of "real" as described above, and move on. Unfortunately you'll continue to hunt down blind alleys, vainly seeking a through passage. Have fun.

 

[Simon Phoenix]

Continue calling them "real" but translate, in your thoughts, to "appears real to me, based on the persistent and convincing data of my senses, but of course there's no way to tell if it's really real."

Eh? This is supposed to be the sharp pointy thing that cuts through the Gordian knot?

All we have to do here is to play with words to get back to the issue at hand - "I must interpret the quantum state as if it were describing something real, even though I can never prove that anything is real" would then be something like the ontic position and "I must interpret the quantum state as being descriptive of my knowledge about some reality that only appears real" would be something like the epistemic position. OK I could probably phrase those better, but I fail to see how the rather obvious and trivial assertion that we can never prove anything to be real solves anything as far as the epi/onto problem in foundations physics is concerned.

So given the assumption that there is an external reality independent of our senses and consciousness do we have to take an ontic or epistemic view of the quantum state, according to the structure of QM? That's the issue - and it isn't in any sense resolved by merely saying the initial assumption is unprovable or false.