Wheeler's delayed choice doesn't change the past

[microsansfil]

Classical mechanics is ultimately about a trajectory in phase space, given by the dynamical evolution (since the state of the system in classical mechanics is represented by a point in phase space).

Quantum mechanics is ultimately about the evolution of the probabilities (or probability distributions) given by the statistical operator and the eigenvectors of observables in any picture of time evolution.

In the context of Statistical mechanics it seem that Classical and Quantum formulations are highly analogous : "Classical and quantum dynamics of density matrices" http://www.scs.illinois.edu/mgweb/Course_Notes/chem544/notes/Ch9.pdf

best regards
Patrick

 

[Fra]

In the context of Statistical mechanics it seem that Classical and Quantum formulations are highly analogous : "Classical and quantum dynamics of density matrices" http://www.scs.illinois.edu/mgweb/Course_Notes/chem544/notes/Ch9.pdf

In particular they both fit well into what Smolin calles the Newtonian schema. Where you indeed have an underlying determinstic evolution in a timeless statespace, by means of timeless laws.

Its more on the causal part that thet differ. One simple way I like to think of the difference is to draw an analogy to economy, and how a palyer can determine market value of things. Here we can consider market value as the "observables". One traditional way is to value the substance value, corresponding to in some realist sense the "actual" hard physical values. The other way is to consider a pure expectation picture, where the actual values are no more and no less than the expected value of all other players in the market. And these expectations (wether rational or not) determine the actions fo the players on the market. The former is like classical mechanics and the latter is more like quantum mechanics, where the causality are based on expectations, rather than actualities.

If you analyze the substance value picture, and require that the substance values must be experimentally determinend, you really end up in a situation such as QM. Because not even traditional hard values have objective values. For example gold, and diamonds? The value of gold is also in principle subject to speculation and expectations.

I suspect its this that is strange to some with QM.

/Fredrik

 

[vanhees71]

As I see it, question or "choosing observables" imples a change in the internal structure of the observer (or measurement device, or information processing agent if you prefer).
Thus technically, asking totally different questions correspond to different observers, so i do not see a problem with this. This is why its "subjective". But its not subjective in a mystical way imo?
/Fredrik

It's not subjective. Why should it be subjective? If I measure observable $A$, I get something different than measureing observable $B$, supposed $B$ is not a unique function of $A$. There's nothing subjective in this, nor is it very specific to QT.

The main difference between classical and quantum physics is not the measurement of observables but the meaning of the state, i.e., what can be "prepared". In classical physics you can in principle determine the state of the system such that all possible observables have a determined value, while in QT this is only possible for sets of compatible observables, and the meaning of the state, even if it's one implying maximal possible knowledge about the system, i.e., the preparation in a pure state, is probabilistic. That's the true content of the Heisenberg-Robertson uncertainty principle. A great deal of confusion concerning the meaning of the quantum state is thus due to Heisenberg, who got one of his most famous discoveries (the uncertainty relation) wrong. He was immediately corrected by Bohr, but unfortunately the wrong first interpretation by Heisenberg stuck, and you get misleading statements about the measurability or non-measurability of observables still today. Indeed, you can measure any observable as precisely as you like (modulo practical technical problems with precise measurements) for any state the measured system is prepared in. What you can't do is to prepare precisely two incompatible observables, but still you can measure either of them with arbitrary precision, in whatever state the system is prepared in. Neither preparation nor measurement is subjective if the used procedures/protocols fulfill the constraints of reprodicibility making them to scientific empirical facts rather than subjective imaginations.

 

[Fra]

It's not subjective. Why should it be subjective? If I measure observable $A$, I get something different than measureing observable $B$, supposed $B$ is not a unique function of $A$. There's nothing subjective in this, nor is it very specific to QT.

I think you react on the word "subjective", with subjective i simply mean that the expectations and the probabilitis are conditional to (ie subjective) to the measurement device (including its choice of observables).

So I agree that in the first level of analysis, there is no actual subjectivity as in "ambigousness".

I use the word subjective synonymous to "conditional to observer" which also means conditional to the choice of observables. But wether there "choice" of observables is "free" or not, is a different discussion. It also depends on if we are talking about the freedom of experimenter to tweak the detectors, or freedom of a nucleus to "choose" its observables. In the former case, there is a FAPP freedom, but in the later case i think the nucleus aligns its "choice of observables" in order to stabilize itself in its environment and get maximal predictive power. However in the latter case Bohrs idea of the requirement for a CLASSICAL measurement device also breaks down. So for this reason it tried to keep the discussion at the current QM level, in order to stay on topic.

/Fredrik

 

[vanhees71]

As I see it, question or "choosing observables" imples a change in the internal structure of the observer (or measurement device, or information processing agent if you prefer).

Thus technically, asking totally different questions correspond to different observers, so i do not see a problem with this. This is why its "subjective". But its not subjective in a mystical way imo?

/Fredrik

Well, of course, to measure the position of a particle you need a different device than to meausure its momentum. This is not specific to QT but also the case within classical physics. Indeed there's no problem with this, and it's in no way mystical at all.

I think you react on the word "subjective", with subjective i simply mean that the expectations and the probabilitis are conditional to (ie subjective) to the measurement device (including its choice of observables).

So I agree that in the first level of analysis, there is no actual subjectivity as in "ambigousness".

I use the word subjective synonymous to "conditional to observer" which also means conditional to the choice of observables. But wether there "choice" of observables is "free" or not, is a different discussion. It also depends on if we are talking about the freedom of experimenter to tweak the detectors, or freedom of a nucleus to "choose" its observables. In the former case, there is a FAPP freedom, but in the later case i think the nucleus aligns its "choice of observables" in order to stabilize itself in its environment and get maximal predictive power. However in the latter case Bohrs idea of the requirement for a CLASSICAL measurement device also breaks down. So for this reason it tried to keep the discussion at the current QM level, in order to stay on topic.

/Fredrik

But this is an abuse of the word "subjective". All you describe are objective properties of objective observations in nature. Of course the probabilities depend on which quantity is measured. That's not even surprising, let alone mystical or subjective. For me it doesn't make sense to say "a nucleus chooses its observables". A nucleus just is a welldefined entity of nature. What I observe at it (e.g., it's position or momentum) is my free choice, and QT helps me to tell the probabilities for the outcome of the corresponding measurement, provided I've given the state of the nucleus (which is a formal mathematical description of (an equivalence class of) a specific prepartion procedure.

I've also never understood Bohr's "classical measurement device". According to QT everything is quantum, including macroscopic systems making up measurement devices. The classical behavior of the relevant macroscopic observables (which are coarse-grained by averaging many microscopic degrees of freedom over microscopically large, macroscopically small space-time regions) is emergent.

 

[Demystifier]

Classical mechanics is ultimately about a trajectory in phase space,

Perhaps in the Hamilton formulation, but not in the Newton or Lagrange formulation. In the latter two formulations, what matters is the configuration space, not the phase space.

 

[Demystifier]

Classical mechanics is ultimately about a trajectory in phase space, given by the dynamical evolution (since the state of the system in classical mechanics is represented by a point in phase space).

Quantum mechanics is ultimately about the evolution of the probabilities (or probability distributions) given by the statistical operator and the eigenvectors of observables in any picture of time evolution.

Can classical mechanics be formulated without an explicit reference to measurement?
Can quantum mechanics be formulated without an explicit reference to measurement?
If the first answer is "yes" and the second "no", don't you feel that it is a problem?

 

[vanhees71]

True, but that doesn't make any principal difference to the distinction between classical and quantum physics. If you derive classical approximations from quantum theory it's clear that the Hamilton formulation is the only save starting point. E.g., in QFT you always have to start with the "Hamiltonian path integral" to be sure to get the correct "Lagrangian path integral". A naive application of the Lagrangian form already leads to wrong results in quite simple cases as, e.g., for the thermal-field theory treatment free charged Klein-Gordon field in the grand-canonical approach at finite chemical potential!

 

[vanhees71]

Can classical mechanics be formulated without an explicit reference to measurement?
Can quantum mechanics be formulated without an explicit reference to measurement?
If the first answer is "yes" and the second "no", don't you feel that it is a problem?

There is of course no such problem since no physics can be formulated without reference to measurements/observations. Physics is about measurements and observations!

 

[Demystifier]

There is of course no such problem since no physics can be formulated without reference to measurements/observations. Physics is about measurements and observations!

http://www.informationphilosopher.com/solutions/scientists/bell/Against_Measurement.pdf

Theoretical physics is distilled from experiments, there are no doubts about it. However, ones the distillation process is over, one may want to formulate the theory without an explicit reference to measurement. For instance, Landau and Lifshitz have written a great book on classical mechanics without mentioning measurements. On the other hand, it seems that something similar cannot be done for quantum theory (in the standard form).

 

[vanhees71]

I know this, but how can a physicist write a pamphlet against measurements? Even a theoretical physicist gets unemployed if there are no measurements done anymore (except he can switch to pure mathematics or, horribile dictu, philosophy ;-)).

 

[Demystifier]

I know this, but how can a physicist write a pamphlet against measurements?

Bell, of course, is not against doing measurements, or against using the results of measurement to formulate the theory. He is against measurement as a part of formulation of the theory. In classical mechanics, measurement is not a part of the formulation of the theory. In standard quantum mechanics, it is.

 

[vanhees71]

In classical mechanics already writing down a position vector in terms of its Cartesian components implicitly assumes measurements.

 

[Demystifier]

In classical mechanics already writing down a position vector in terms of its Cartesian components implicitly assumes measurements.

I don't think so, just as I don't think that decomposing a quantum state in a specified basis for the Hilbert space implicitly assumes measurement.

 

[Fra]

Well, of course, to measure the position of a particle you need a different device than to meausure its momentum. This is not specific to QT but also the case within classical physics. Indeed there's no problem with this, and it's in no way mystical at all.

The difference is that in classical mechanics you an in principle have several measurement devices at once, measuring all things at once without distorting the system. Ie. you can have a "collection of observers" asking all kinds of questions at once, and then simlpy joing the answers.

This is not possible in quantum mehanics, and its Bohrs point with complementarity.

In this case the measurent is "trivial" or marginalized to a practical matters in classical matter. In quantum mechanics, the choice of measurements distorts the system, in a way that that dependes on what you choose to measure. This is the "subjectivity".

But there is no point in disagreeing on the word. I think we roughly agree at the basic level. I just wanted to express that even though i may not share Demystifiers view, i still follow the objections and point of subjectivity (give or take the choice of words).

I've also never understood Bohr's "classical measurement device". According to QT everything is quantum, including macroscopic systems making up measurement devices. The classical behavior of the relevant macroscopic observables (which are coarse-grained by averaging many microscopic degrees of freedom over microscopically large, macroscopically small space-time regions) is emergent.

You can not in Borhs view properly speak of what is quantum without a classical reference. Surely, you can view the apparatous + system as another "new quantum system" BUT, then you need ANOTHER classical backdrop. To think you can repeat this until you end up with a complete wavefunction of the universe is IMO a fallacy that makes no sense. I think this is Bohrs point. But if we get into the details here, atl east i will raise questions that i think is beyond the scope of the original Einstein Bohr dispute. Bohrs point is I think the most acccurate one if you consider quantum theory as it stands. Even field theory needs a backdrop. The detectors must be attached in a classical world.

To generalized things beyond that, then we are at BTSM discussions.

There is IMO also a connection between when the Newtonian schema works (as per Smolin) AND Bohrs requirement for aclassical backdrop. Without the classical backdrop, which also severs the purposes of information sink, there is no rigid reference for the Newtonian schema - and we need a new understanding of physical law.

/Fredrik

 

[vanhees71]

We don't need a revolution of physics. It already happened in 1926 with the discovery of quantum theory (in 3 equivalent forms) and Born's probabilistic interpretation of the quantum state. There's no need for any other revolution since QT works very well in describing all known phenomena.

 

[PeterDonis]

Thread closed for moderation.

 

[bhobba]

Thread permanently closed. Original question answered - now just a discussion of different views of physics.