I The typical and the exceptional in physics

[A. Neumaier]

"Quantum mechanical predictions consist of predicting properties (typically expectations or conditional probabilities) of the measures defined in Axiom A5"

The collapse selects one possibility,

The collapse means taking probabilities conditioned on the known observations. Thus it is included in my axioms. But it amounts to a change of the modeling assumptions rather than to a change in the system.

This is exactly the same what people handling stocks do - they use propbabilities based on the most recently available information to make predictions, hence collapse their model probability distributions each time new information comes in. But I have never heard of a financial analyst complain about the weirdness of classical stochastic modeling.

Weirdness appears only when one mistakenly ascribes the collapse to the system rather than to the change in the model.

 

[vanhees71]

In case 4. additional polarizer change polarization of passing photons from linear polarization $|V \rangle$ to $|45^0 \rangle$ linear polarization but information update should not cause physical change.

Of course it does. The interaction of the photons with the middle polarizer is a physical interaction of the photon with the charges in the polarizer material and thus a dynamical process changing the photons' polarization (if they get through, if they are aborbed there are no more photons but you have an excitation of the polarizer material manifesting itself as heat).

 

[zonde]

Of course it does. The interaction of the photons with the middle polarizer is a physical interaction of the photon with the charges in the polarizer material and thus a dynamical process changing the photons' polarization (if they get through, if they are aborbed there are no more photons but you have an excitation of the polarizer material manifesting itself as heat).

Yes of course. But then it contradicts Demystifier's definition.

Definition 2:
Wave-function update is a mental act by a person who understands QM. In this act, an old wave function is replaced by a new wave function, with intention to better represent the new knowledge acquired by new measurement results.

Result of inserting middle polarizer can not be viewed as acquisition of new knowledge and nothing more.

 

[Demystifier]

It doesn't work, because the update is nonlocal in the sense that a correct way of updating the wave function is to assign it to a hyperplane of simultaneity, and then updating it instantaneously by non-Schroedinger evolution across the hyperplane.

There is no collapse, only updating, but the updating can be considered nonlocal.

Objecting to a nonlocal update does depend on ontology, since the update does not conflict with the locality of relativistic QFT.

Another way of saying it is that ontology is a tool for performing wave function updating. In that ontology, wave function updating is nonlocal. And there is nothing about that nonlocality that conflicts with QFT - unless one is mistaking the locality of QFT to be ontological.

Well you may call it "non-local" and "ontological" if you want, but then you are using those words in a rather unusual way.

 

[Demystifier]

But there is a density operator describing the LHC.

If we accept the instrumental interpretation according to which there is no wave function of the universe, then, in the same interpretation, there is no density operator describing the whole LHC.

Of course, you are free to use a different interpretation, but then your interpretation is not the minimal instrumental interpretation advocated e.g. by Peres..

 

[Demystifier]

If this is fair representation of vanhees71 position then it is wrong at least in some cases.
1. Say we have beam of unpolarized photons that we register using detector. No information, no update.
2. We insert polarizer before detector. According to presented viewpoint we gain information about polarization but no sudden change of wave-function.
3. We insert two crossed polarizers in the way of beam. There are no photons so we get no information.
4. Now between two crossed polarizers we insert third polarizer at 45 deg. This operation according to point 2. should not cause sudden change of wave-function but it does as now we get 1/8 of photons instead of none.

First, you are using a too narrow meaning of the concept of "wave function". The wave function is nothing but a vector in the Hilbert space of all possible physical states. In this sense, even the vacuum |0> is a wave function (more precisely, a wave functional) in the Hilbert space of quantum electrodynamics. Therefore, your 3. is wrong; even though there are no photons, there is still wave function and associated information.

Second, the change in 4. is not sudden. It is a continuous transition (although very fast, due to fast decoherence) from a state with one number of photons to a state with another number of photons.

 

[atyy]

Well you may call it "non-local" and "ontological" if you want, but then you are using those words in a rather unusual way.

Well, I don't think it weakens my main point that it wrong to use the locality of relativistic QFT to object to any nonlocality associated with updating.

 

[Demystifier]

Result of inserting middle polarizer can not be viewed as acquisition of new knowledge and nothing more.

In the operational interpretation of QM, insertion of the middle polarizer (or any other polarizer) should be viewed as a choice of the effective time-dependent Hamiltonian. The Hamiltonian defines the corresponding Schrodinger equation, so any change induced by insertion of the polarizer should be viewed as evolution by Schrodinger equation.

 

[zonde]

Second, the change in 4. is not sudden. It is a continuous transition (although very fast, due to fast decoherence) from a state with one number of photons to a state with another number of photons.

Yes, I used poor wording. I should have said "physical change of subensemble that is described by wave-function".

 

[Demystifier]

Well, I don't think it weakens my main point that it wrong to use the locality of relativistic QFT to object to any nonlocality associated with updating.

Of course, but @vanhees71 does not have any objections against updating. He only objects against collapse. To clarify his opinion, I have defined the notions of "update" and "collapse" in a manner consistent with his own understanding of those words.

 

[A. Neumaier]

If we accept the instrumental interpretation according to which there is no wave function of the universe, then, in the same interpretation, there is no density operator describing the whole LHC.

This does not follow. The only obstacle to interpreting the wave function of the universe in the Copenhagen interpretation is the lack of an outside observer.

But the LHC is actually observed from the outside, and consists of a finite number of massive particles only, hence even has a description in nonrelativistic quantum mechanics with relativistic correction terms. You forgot to demonstrate why in this description the LHC has no state!

Of course, you are free to use a different interpretation, but then your interpretation is not the minimal instrumental interpretation advocated e.g. by Peres..

Of course, I could use my thermal interpretation. But even in the minimal interpretation, there are enough observable copies of the LHC (namely the single LHC shifted in time) so that one can apply the minimal interpretation in the form of Peres, independent of Copenhagen.

 

[zonde]

In the operational interpretation of QM, insertion of the middle polarizer (or any other polarizer) should be viewed as a choice of the effective time-dependent Hamiltonian. The Hamiltonian defines the corresponding Schrodinger equation, so any change induced by insertion of the polarizer should be viewed as evolution by Schrodinger equation.

In other words this measurement is not just update of information about photon polarization, right?

 

[Demystifier]

But the LHC is actually observed from the outside, and consists of a finite number of massive particles only, hence even has a description in nonrelativistic quantum mechanics with relativistic correction terms. You forgot to demonstrate why in this description the LHC has no state!

What I wanted to say is that, in practice, most of LHC is described by classical physics. Therefore, in practice, one cannot associate a quantum state with it.

 

[Demystifier]

In other words this measurement is not just update of information about photon polarization, right?

Insertion of the polarizer is not an update at all. To update the information, you must also have a detector (polarizer is not a detector) and take a look at it.

The measurement consists of a polarizer and a detector, but measurement is not yet an observation. Observation needs a conscious being who will look at detector. Only the conscious being (with some knowledge of QM) can do the update.

 

[A. Neumaier]

What I wanted to say is that, in practice, most of LHC is described by classical physics. Therefore, in practice, one cannot associate a quantum state with it.

What I wanted to say, however, is that there is no theoretical obstacle to describing all of the LHC by quantum mechanics. Thus one can associate a quantum state with it, though in practice it is only poorly known. But one can translate every classical knowledge about it into a corresponding Wigner density operator, so that even in practice one could give a reasonable density operator for it.

In any case, for the foundational discussion here it doesn't matter whether one can specify the state in practice but only that it exists.

 

[A. Neumaier]

Only the conscious being (with some knowledge of QM) can do the update.

A computer program written by a conscious being with some knowledge of QM can do the update, too. Just as the classical models of the stock market are automatically updated by computer programs. Thus consciousness is in no way involved in the collapse.

 

[Demystifier]

Yes, I used poor wording. I should have said "physical change of subensemble that is described by wave-function".

In terminology of post #126, do you mean physical 1 or physical 2?

 

[zonde]

Insertion of the polarizer is not an update at all. To update the information, you must also have a detector (polarizer is not a detector) and take a look at it.

When we analyze experimental setup we imagine that we have a detector at the place where we describe wave-function. A suppose that it goes without saying.

 

[Demystifier]

A computer program written by a conscious being with some knowledge of QM can do the update, too. Just as the classical models of the stock market are automatically updated by computer programs. Thus consciousness is in no way involved in the collapse.

Now you are changing my terminology which I fixed in post #154. Without fixing the meaning of words it's very hard to have a meaningful discussion. In #154 I fixed the terminology in a way that would please @vanhees71, but it's impossible to fix terminology in a way that would please everybody.

 

[Demystifier]

What I wanted to say, however, is that there is no theoretical obstacle to describing all of the LHC by quantum mechanics. Thus one can associate a quantum state with it, though in practice it is only poorly known. But one can translate every classical knowledge about it into a corresponding Wigner density operator, so that even in practice one could give a reasonable density operator for it.

In any case, for the foundational discussion here it doesn't matter whether one can specify the state in practice but only that it exists.

Yes, I agree with that. But instrumental interpretation of QM does not care much about what is possible in principle. In principle it is possible to eat with a hammer, but it is irrelevant for both a repairman and a waiter.

 

[zonde]

In terminology of post #126, do you mean physical 1 or physical 2?

physical 2, but with addition that physical change means that there are no subjective aspect of it (no observer will say "nothing happened")

 

[RockyMarciano]

The collapse means taking probabilities conditioned on the known observations. Thus it is included in my axioms. But it amounts to a change of the modeling assumptions rather than to a change in the system.

This is exactly the same what people handling stocks do - they use propbabilities based on the most recently available information to make predictions, hence collapse their model probability distributions each time new information comes in. But I have never heard of a financial analyst complain about the weirdness of classical stochastic modeling.

Weirdness appears only when one mistakenly ascribes the collapse to the system rather than to the change in the model.

I believe the issue here is precisely that this comparison with classical stochastic modelling undermines the distinction you keep between change in the system and change in the modelling assumptions. That would be valid in Newtonian physics where measurements don't affect the system that is measured, but this is a key feature in QM and also in stochastic modelling where information directly affects the stock market.
It seems to me that your axioms simply ignore this difference because they might be tied to the classical mindframe.

 

[Demystifier]

When we analyze experimental setup we imagine that we have a detector at the place where we describe wave-function. A suppose that it goes without saying.

But detector is usually put only on the final point in the path of the photon, while wave function is described during the whole path.

 

[atyy]

Of course, but @vanhees71 does not have any objections against updating. He only objects against collapse. To clarify his opinion, I have defined the notions of "update" and "collapse" in a manner consistent with his own understanding of those words.

As I understand, vanhees71 objects to collapse on grounds of locality. I think that is not a correct reason to object to collapse (one can object to collapse for other reasons, eg. on reasons of ontology, under the hypothesis that Bohmian mechanics is correct).

 

[vanhees71]

Of course, but @vanhees71 does not have any objections against updating. He only objects against collapse. To clarify his opinion, I have defined the notions of "update" and "collapse" in a manner consistent with his own understanding of those words.

Particularly I object against claiming that there is an interaction at a distance, where there is none, as in the case of measuring the polarizations of entangled photons at far distant places. The interaction of the photon with the measurement devices are, by construction, local. Thus, there's now instantaneous effect of A's measurement on B's photon and vice versa.

The other case, discussed today, is about having three polarization filters (first H then $\pi/4$ then V in direction of the photon beam). Of course, here you have indeed local interactions leading to a change in the photon's polarization if going through or to its absorption if not going through. These are all caused by local interactions of the photons with the polarization foil.

 

[Demystifier]

As I understand, vanhees71 objects to collapse on grounds of locality.

Locality is a word with many different meanings, and he does not insist on all possible kinds of locality. He only objects against non-locality which contradicts Schrodinger equation, unless it is an update.

 

[vanhees71]

As I understand, vanhees71 objects to collapse on grounds of locality. I think that is not a correct reason to object to collapse (one can object to collapse for other reasons, eg. on reasons of ontology, under the hypothesis that Bohmian mechanics is correct).

Again, you have to be more precise in what you are saying! I deny the existence of nonlocal interactions (or "actions at a distance") in standard relativistic QFTs, because they are explicitly formulated with local interactions. Of course, there is the possibility of correlations between far-distant parts of an extended system like the here discussed polarization-entangled biphoton states. Here I deny collapse, because a local interaction of A's photon with her measurement device does not instantaneously affect B's photon's properties measured a far distance away from A.

 

[vanhees71]

Locality is a word with many different meanings, and he does not insist on all possible kinds of locality. He only objects against non-locality which contradicts Schrodinger equation, unless it is an update.

The Schrodinger equation describes action at a distance. It's a non-relativistic theory!

 

[zonde]

But detector is usually put only on the final point in the path of the photon, while wave function is described during the whole path.

So? Experimentalists describe changes of wave-function along the path of photon beam and arrive at correct predictions. Where is the problem?

 

[atyy]

Again, you have to be more precise in what you are saying! I deny the existence of nonlocal interactions (or "actions at a distance") in standard relativistic QFTs, because they are explicitly formulated with local interactions. Of course, there is the possibility of correlations between far-distant parts of an extended system like the here discussed polarization-entangled biphoton states. Here I deny collapse, because a local interaction of A's photon with her measurement device does not instantaneously affect B's photon's properties measured a far distance away from A.

But the local interactions in QFT do not include the measurement apparatus. The Hamiltonian in the standard Bell test only includes the photons, which is why the state can be written $|hh \rangle + |vv \rangle$.

So the Hamiltonian has nothing to say about collapse.

 

[Demystifier]

The Schrodinger equation describes action at a distance. It's a non-relativistic theory!

By Schrodinger equation, I mean any equation of the form
$$H|\psi(t)\rangle = i\hbar\frac{\partial |\psi(t)\rangle}{\partial t}$$
where $H$ is the Hamiltonian of the considered system. Even relativistic QFT has a Hamiltonian.

 

[Demystifier]

So? Experimentalists describe changes of wave-function along the path of photon beam and arrive at correct predictions. Where is the problem?

No problem.

 

[vanhees71]

But the local interactions in QFT do not include the measurement apparatus. The Hamiltonian in the standard Bell test only includes the photons, which is why the state can be written $|hh \rangle + |vv \rangle$.

So the Hamiltonian has nothing to say about collapse.

The interaction of photons with anything is described by QED, at least I have no example, where this is clearly disproven. If collapse is not a dynamical process, it's simply not relevant to the entire discussion about measurement processes and state preparations.

 

[zonde]

I deny the existence of nonlocal interactions (or "actions at a distance") in standard relativistic QFTs, because they are explicitly formulated with local interactions.

Non local interactions in QED are hidden under change of basis.

 

[atyy]

The interaction of photons with anything is described by QED, at least I have no example, where this is clearly disproven. If collapse is not a dynamical process, it's simply not relevant to the entire discussion about measurement processes and state preparations.

What I would like to understand is: at the LHC - what is governed by unitary evolution? If you say there is no wave function of the universe, and no wave function of the LHC, then what is the biggest thing you are willing to assign a wave function to?

 

[RockyMarciano]

As long as there are artificially two kinds of evolution in QM, one by measurement and one in between measurements there is no way out of these eternal debates.
Let's hope the new formalism that may eventually arrive has the flexibility to model what all the quantum phenomenology shows, just one type of quantum evolution, the one observed.

 

[Mentz114]

But the local interactions in QFT do not include the measurement apparatus. The Hamiltonian in the standard Bell test only includes the photons, which is why the state can be written $|hh \rangle + |vv \rangle$.

So the Hamiltonian has nothing to say about collapse.

I don't think so. The settings of the polarizers is relevant to the outcome and should be included in the Hamiltonian. It doesn't make much difference in any case, but it does not support collapse as a physical process ( in my opinion, anyway ).

 

[Demystifier]

Particularly I object against claiming that there is an interaction at a distance, where there is none, as in the case of measuring the polarizations of entangled photons at far distant places. The interaction of the photon with the measurement devices are, by construction, local. Thus, there's now instantaneous effect of A's measurement on B's photon and vice versa.

In this regard I disagree with you and agree with @atyy . But to explain why I disagree I would need to talk about ontology, which is something you don't really care about. So I can't explain my reasons for disagreement in a way you would care about.

 

[vanhees71]

Non local interactions in QED are hidden under change of basis.

No, by construction the interactions are local in QED.

 

[vanhees71]

What I would like to understand is: at the LHC - what is governed by unitary evolution? If you say there is no wave function of the universe, and no wave function of the LHC, then what is the biggest thing you are willing to assign a wave function to?

You cannot even describe the LHC by classical mechanics of each atom it consists of, and even if you could, it's a total overkill to do so. Nevertheless the classical theory of macroscopic observables is just an effective theory of QT for the relevant degrees of freedom to describe (even construct) the LHC.

 

[vanhees71]

In this regard I disagree with you and agree with @atyy . But to explain why I disagree I would need to talk about ontology, which is something you don't really care about. So I can't explain my reasons for disagreement in a way you would care about.

Ok, but this is explicitly contradicting the very foundations of the model (QFT) you want to interpret. So there's an explicit contradiction between the mathematical structure of the model and the (ontological?) interpretation. What should this be good for?

 

[Demystifier]

Ok, but this is explicitly contradicting the very foundations of the model (QFT) you want to interpret. So there's an explicit contradiction between the mathematical structure of the model and the (ontological?) interpretation.

There is no contradiction at all. For instance, Bohmian formulation of QFT retains all the mathematical structure of standard QFT. However, this formulation does not stop there but makes one further step by adding one additional equation. It is this additional equation which makes the theory non-local. This additional equation does not contradict any of the previous local equations.

 

[zonde]

No, by construction the interactions are local in QED.

I do not deny that known interactions are local in QED. I say that there is unknown (hidden) interaction that does not appear explicitly in QED.

 

[stevendaryl]

Before QM came along I suspect that very few scientists would have held that it is the job of science just to predict stuff, and not to say anything about 'reality'. Of course after QM, when it became awkward (to say the least) to ascribe some 'reality' to the state, I get the impression it was as if there was some collective decision to 'redefine' what science is about.

Right. That's why I suggest that it is revisionism to say that science is not about ontology, but only about epistemology. Or maybe it's sour grapes---when you don't know how to get something, you pretend that you never wanted it in the first place.

 

[vanhees71]

There is no contradiction at all. For instance, Bohmian formulation of QFT retains all the mathematical structure of standard QFT. However, this formulation does not stop there but makes one further step by adding one additional equation. It is this additional equation which makes the theory non-local. This additional equation does not contradict any of the previous local equations.

I don't know enough about Bohmian QFT, but if provides nonlocal interactions it obviously contradicts QFT. So it's a new theory rather than just an interpretation of standard QFT. In other words, are there observable consequences of such nonlocal interactions and if so, have they been tested in real experiments?

 

[vanhees71]

I do not deny that known interactions are local in QED. I say that there is unknown (hidden) interaction that does not appear explicitly in QED.

Then it's a new theory and not QED. To discuss it (I'm not sure that this is allowed in the forum) you'd need to specify it clearly. I can't get what your theory looks like ;-)).

 

[zonde]

Then it's a new theory and not QED.

No, I am not talking about new theory. I talk about mathematical elements and operations in standard QED model.

 

[stevendaryl]

The big problem with ignoring ontology is that people who claim to ignore it actually promote it: eg. Ballentine, Peres and vanhees71.

If they truly did not care about ontology, they would have no problems with collapse.

There are well respected positions that ignore ontology, eg. Bohr, Landau and Lifshitz, Copenhagen as described by Weinberg - but all of these have things which are disavowed by vanhees71, eg. collapse and the Heisenberg cut.

It seems to me that the minimal interpretation--that measurements produce eigenvalues of the observable being measured, with probabilities given by the Born rule--does have an ontology. What's considered real is the observation or measurement. My issue with it is that if you go on to say that measurements/observations are explainable in terms of microscopic objects and processes, then you're explaining real things in terms of things that are not real. I suppose this applies to both modern adherents to the minimalist interpretation but also the original adherents of the Copenhagen interpretation.

So it seems to me that there is a conceptual circularity involved in the minimalist interpretation. You explain quantum mechanics in terms of macroscopic measurements and explain measurements in terms of quantum mechanics. I suppose that the circularity doesn't ruin the theory, though, because circularity can be handled in an iterative fashion: Start with a classical description of measurements and measurement devices, and then make quantum corrections to this description to get an improved description, and iterate until you reach some kind of fixed point.

 

[stevendaryl]

Then it's a new theory and not QED. To discuss it (I'm not sure that this is allowed in the forum) you'd need to specify it clearly. I can't get what your theory looks like ;-)).

I'm thinking that it has two kinds of processes: (1) smooth Hamiltonian evolution of microscopic systems, and (2) measurements always return an eigenvalue of the observable being measured, with probabilities given by the Born rule.

Since (2) is not derivable from (1), you need both types of processes. QED only describes (1).

 

[stevendaryl]

I'm thinking that it has two kinds of processes: (1) smooth Hamiltonian evolution of microscopic systems, and (2) measurements always return an eigenvalue of the observable being measured, with probabilities given by the Born rule.

Since (2) is not derivable from (1), you need both types of processes. QED only describes (1).

Many-worlds advocates claim that (2) is derivable from (1), but others disagree.