Copenhagen - What qualifies as "measurement" and "observer"?

[Jesayntly]

Hey guys and gals, clearly I'm new here, not only new here but new to QM (relatively speaking). I've read a few threads here pertaining to a thought I've been pondering on but the answer doesn't seem clear.

In the Copenhagen interpretation, what qualifies as a "measurement", and/or what qualifies as an "observer"?

I ask this because of the alleged seemingly common misunderstanding that the one doing the measurement needn't be sentient because otherwise it would stand to reason that prior to being "measured" by an "observer" the most basic building blocks of matter wouldn't be able to combine and interact due to being in an unstable state of possibilities and never collapsing into a physical point in the universe and well, we wouldn't have a universe. In other words, we'd have no objects in the universe as their components would still be in an unstable state of flux from not being measured because of a lack of sentient beings to measure it.

Is it simply the act of an outside "agent", sentient or not, that can collapse a particle's wave function?

So pardon the "noob" question, just looking for some clarification.

Thanks!

 

[atyy]

Ultimately, the one doing the measurement must be sentient or at least have "common sense" in Copenhagen. One can delegate the measurement to a non-sentient device. However, the important point is that the device is not considered quantum. The device is "macroscopic" or "classical". It is this subjective division of the world into classical and quantum parts that requires the observer that stands apart from the quantum system. Only the classical part is "real", while the wave function describing the quantum part may be just a tool to calculate the probability of measurement outcomes.

This is of course problematic, and is known as the measurement problem.
http://www.informationphilosopher.com/solutions/scientists/bell/Against_Measurement.pdf
http://www.tau.ac.il/~tsirel/download/nonaxio.html
http://arxiv.org/abs/quant-ph/0209123

 

[Jesayntly]

Can we assume that bigger units of matter can not form from particles whose wave functions remain "uncollapsed"? If we can, how could the now observable celestial bodies to come into existence when there are no sentient beings actively observing or at least none available to produce classical measuring devices?

 

[atyy]

Can we assume that bigger units of matter can not form from particles whose wave functions remain "uncollapsed"? If we can, how could the now observable celestial bodies to come into existence when there are no sentient beings actively observing or at least none available to produce classical measuring devices?

It is possible to do so, but not within Copenhagen. One needs an interpretation like Bohmian Mechanics. BM seems to work for non-relativistic quantum mechanics and some relativistic quantum phenomena, but it is unclear if it can be describe all relativistic quantum phenomena.

Another possibility is the Many-Worlds approach. However, I am not sure whether this approach is correct (some versions are, but those versions are variants of Bohmian Mechanics).

 

[Jesayntly]

Can we assume that bigger units of matter can not form from particles whose wave functions remain "uncollapsed"? If we can, how could the now observable celestial bodies to come into existence when there are no sentient beings actively observing or at least none available to produce classical measuring devices?

Is there any discussion of the formation of larger units of matter when it's base building blocks are still in a state of superposition?

 

[Jesayntly]

It is possible to do so, but not within Copenhagen. One needs an interpretation like Bohmian Mechanics. BM seems to work for non-relativistic quantum mechanics and some relativistic quantum phenomena, but it is unclear if it can be describe all relativistic quantum phenomena.

Another possibility is the Many-Worlds approach. However, I am not sure whether this approach is correct (some versions are, but those versions are variants of Bohmian Mechanics).

Curious. I'm sure this conundrum has been presented before. I wonder how it was handled. I'll delve more into Bohmian Mechanics. Thanks for the input!

 

[atyy]

Is there any discussion of the formation of larger units of matter when it's base building blocks are still in a state of superposition?

In a sense this is true in Copenhagen, Bohmian Mechanics and Many-Worlds. However, in Copenhagen, without the observer, the larger units and the smaller units are all not necessarily real. Reality has to be recognized by the observer in Copenhagen, or instantiated with hidden variables as in BM, or quite radically rethought as in Many-Worlds.

 

[Jesayntly]

I mean, if one supposes CI to be true and that larger particles can't form from uncollapsed quantum building blocks, the only other possible explanation is, well, something divine in nature is observing and measuring from outside our universe or we're encased in some amazing simulation. This is a disturbing notion. One I know Einstein viewed in much the same light.

Off to do some reading.

 

[Nugatory]

There's something not to like in every interpretation of quantum mechanics, and you've identified what's not to like in Copenhagen. It's the problem that Schrodinger pointed out with his famous thought experiment about the cat in the box: there's nothing in the theory that tells us which interactions count as observations causing collapse to a definite outcome and which lead to an uncollapsed superposition of the possible outcomes.

 

[Jesayntly]

I'm hoping new, recent experiments pertaining to Pilot Wave Theory give it some traction - I'm a bigger fan of a deterministic reality. The fact that some behaviors are being exhibited in a classical environment that many thought were exclusive to the quantum realm are giving me some hope.

 

[atyy]

I'm hoping new, recent experiments pertaining to Pilot Wave Theory give it some traction - I'm a bigger fan of a deterministic reality. The fact that some behaviors are being exhibited in a classical environment that many thought were exclusive to the quantum realm are giving me some hope.

I'm a fan of BM and other hidden variable interpretations too - but assuming you are reading https://www.quantamagazine.org/20140624-fluid-tests-hint-at-concrete-quantum-reality/ - then no, these do not lend the pilot wave theory any support.

Again, I would like to stress that it is unknown if the pilot wave theory can describe the full range of relativistic quantum phenomena.

 

[Jesayntly]

Different source - same information. Sigh. Here's to hoping. Copenhagen is just too creepy in what it infers.

 

[atyy]

I mean, if one supposes CI to be true and that larger particles can't form from uncollapsed quantum building blocks, the only other possible explanation is, well, something divine in nature is observing and measuring from outside our universe or we're encased in some amazing simulation. This is a disturbing notion. One I know Einstein viewed in much the same light.

Let's say BM is true. If that were the case, the CI is considered true but incomplete. CI can be derived from BM.

Dirac was a famous proponent of the view that CI is true but incomplete.

 

[Jesayntly]

Can't comment further until I gain a better understanding of BM. I'll keep that in mind though as I study.

 

[atyy]

I should add BM and the pilot wave theory are essentially synonymous.

 

[Jesayntly]

I should add BM and the pilot wave theory are essentially synonymous.

It would appear so.

 

[Jesayntly]

I know that BM is lacking in certain regards but it just rubs me the right way. CI just doesn't work for instance with my example of a developing universe. Either CI or BM is missing something HUGE.

 

[bhobba]

I know that BM is lacking in certain regards but it just rubs me the right way. CI just doesn't work for instance with my example of a developing universe. Either CI or BM is missing something HUGE.

All interpretations have issues - every single one.

You often hear QM has a measurement problem etc etc. That's not its real issue. The issue is it doesn't matter what problem in QM worries you, the measurement problem or any other, you can find an interpretation that fixes it. But fixing all the issues at once - that's the trick that no one has achieved. Further progress IMHO will need to wait until its possible to figure out (if its possible at all) how to experimentally distinguish them.

Another big issue is what worries some, others couldn't care less. For me QM is a theory about observations. Observations are the primitive of the theory like points and lines are the primitives of geometry. In modern times decoherence has morphed the measurement problem to why we get any outcomes at all. For me its a non issue - that observations exist is a primitive of the theory - worrying about it is like worrying about why we have points and lines. But for others its a deep concern.

First though you need to understand what CI says:
http://motls.blogspot.com.au/2011/05/copenhagen-interpretation-of-quantum.html

The modern way to fix its blemish alluded to in the above is Consistent Histories:
http://quantum.phys.cmu.edu/CHS/histories.html

Its a nice interpretation, basically many worlds without the many worlds, and actually shares many of the key theorems with the modern version of many worlds. Its issue IMHO is it gets around the measurements problem by not having measurements - in that interpretation QM is a stochastic theory about histories. It looks like defining your way out of difficulties rather than face them head on.

Another interesting thing is many QM interpretations are simply an argument about what probability is:
http://math.ucr.edu/home/baez/bayes.html

Thanks
Bill

 

[stevendaryl]

For me QM is a theory about observations. Observations are the primitive of the theory like points and lines are the primitives of geometry

I don't understand how it makes sense for observations to be primitive, when you have to have a lot of machinery--measurement devices--to make an observation. The fact that such and such a procedure constitutes a measurement/observation of such and such a property seems like a nontrivial deduction, rather than something primitive. I guess in practice you can do the Copenhagen split between a classical measuring device and a quantum system being measured, then use classical reasoning to figure out what is being measured, and then treat the measurement as a primitive when it comes to applying quantum mechanics. But the whole thing seems very much nonprimitive to me.

 

[bhobba]

I don't understand how it makes sense for observations to be primitive,

Its an intuitive idea like a point. Try and tie down a point exactly and you have exactly the same problem - we must rely on our intuition. The same for event in probability theory - in fact that seems quite similar to observation.

Thanks
Bill

 

[Feeble Wonk]

Ultimately, the one doing the measurement must be sentient or at least have "common sense" in Copenhagen. One can delegate the measurement to a non-sentient device. However, the important point is that the device is not considered quantum. The device is "macroscopic" or "classical". It is this subjective division of the world into classical and quantum parts that requires the observer that stands apart from the quantum system. Only the classical part is "real", while the wave function describing the quantum part may be just a tool to calculate the probability of measurement outcomes.

I'm confused again. Are you saying that Copenhagen DOES still posit the necessity for a conscious observer? I've asked this question multiple times before on PF, but the answers have been something less than consistent in response. Most recently I was told...
"You have to understand observer in QM is not necessarily an organic conscious observer. In modern times its anything that causes decoherence ie is an interaction. With that in mind in Copenhagen observer is absolutely fundamental."

You noted that sentience could be delegated to a macroscopic/classical "non-sentient device", but that sounds like a de facto acquiescence to decoherence as the "measurement" trigger. Can anyone tell me with clarity whether standard traditional Copenhagen still posits the need for a conscious observer?

I'm also confused by your functional delineation between the "classical" and the "quantum" in this context.

 

[bhobba]

I'm confused again. Are you saying that Copenhagen DOES still posit the necessity for a conscious observer?

It doesn't.

What Atty is alluding to is in the interpretation you need to be able to divide the total system into what is being observed and what does the observing. Even classical mechanics requires dividing a system into reasonable parts to solve problems. That must be done by some kind of intelligence. Its the factorisation issue. Most think, including me, it doesn't really matter how you divide it you should get the same results. But we don't have theorems to prove it one way or the other.

In my view its part of what an observation is - its a primitive of the theory. Its like event in probability. You require intuition to apply it - pinning it down exactly isn't easy. Even using the definition from decoherence, which is what I do, requires factoring the system.

But, and let's be clear about this, its the same issue you have in any area of applied math - you need a bit of common-sense intuition to apply it - that doesn't mean the theory requires conciousness.

Thanks
Bill

 

[atyy]

It doesn't.

What Atty is alluding to is in the interpretation you need to be able to divide the total system into what is being observed and what does the observing. Even classical mechanics requires dividing a system into reasonable parts to solve problems. That must be done by some kind of intelligence. Its the factorisation issue. Most think, including me, it doesn't really matter how you divide it you should get the same results. But we don't have theorems to prove it one way or the other.

In my view its part of what an observation is - its a primitive of the theory. Its like event in probability. You require intuition to apply it - pinning it down exactly isn't easy. Even using the definition from decoherence, which is what I do, requires factoring the system.

But, and let's be clear about this, its the same issue you have in any area of applied math - you need a bit of common-sense intuition to apply it - that doesn't mean the theory requires conciousness.

I think the observer is needed in Copenhagen in a way that is different from classical physics unless one can show that the factorization is objective, ie. any factorization gives the same result. However, it is unclear if this is true. Bohmian Mechanics solves the factorization problem in a way that is objective by supplying hidden variables that provide the objective factorization. But I think it also suggests that without hidden variables, the factorization is not objective - for example, can momentum be the hidden variable instead of position?

 

[stevendaryl]

I think the observer is needed in Copenhagen in a way that is different from classical physics unless one can show that the factorization is objective, ie. any factorization gives the same result. However, it is unclear if this is true. Bohmian Mechanics solves the factorization problem in a way that is objective by supplying hidden variables that provide the objective factorization. But I think it also suggests that without hidden variables, the factorization is not objective - for example, can momentum be the hidden variable instead of position?

I can't find the thread now, but some PF regular, maybe Demystifier, explained to me once why the Bohmian idea of giving privileged status to position doesn't work as well for other observables such as momentum. I think it might just be that interactions tend to be local in terms of position, but nonlocal in terms of momentum.

 

[atyy]

I can't find the thread now, but some PF regular, maybe Demystifier, explained to me once why the Bohmian idea of giving privileged status to position doesn't work as well for other observables such as momentum. I think it might just be that interactions tend to be local in terms of position, but nonlocal in terms of momentum.

Yes, that is the informal argument. It is the same more or less the same argument as to why decoherence localizes objects. However, I don't know a real proof, not even at the physics level that momentum cannot be the hidden variable.

 

[DrChinese]

I know that BM is lacking in certain regards but it just rubs me the right way. CI just doesn't work for instance with my example of a developing universe. Either CI or BM is missing something HUGE.

There are other classes of interpretations. I am sure you are aware of Many Worlds. Are you familiar with the Time Symmetric class? (This includes Cramer's, Retrocausal and Relational BlockWorld.) In these, the future is part of the measurement context. There are state of the art experiments in which a future decision clearly affects the past. I should say that all interpretations still predict the same results. Also, just as FTL signalling is not predicted by any interpretation, Time Symmetric interpretations do not predict an ability to create a time paradox.

Interpretations such as these are very appealing once you get past their baggage. (as previously mentioned, every interpretation has baggage.)

 

[Feeble Wonk]

It doesn't.

What Atty is alluding to is in the interpretation you need to be able to divide the total system into what is being observed and what does the observing. Even classical mechanics requires dividing a system into reasonable parts to solve problems. That must be done by some kind of intelligence. Its the factorisation issue. Most think, including me, it doesn't really matter how you divide it you should get the same results. But we don't have theorems to prove it one way or the other.

In my view its part of what an observation is - its a primitive of the theory. Its like event in probability. You require intuition to apply it - pinning it down exactly isn't easy. Even using the definition from decoherence, which is what I do, requires factoring the system.

But, and let's be clear about this, its the same issue you have in any area of applied math - you need a bit of common-sense intuition to apply it - that doesn't mean the theory requires conciousness.

Thanks
Bill

I think the observer is needed in Copenhagen in a way that is different from classical physics unless one can show that the factorization is objective, ie. any factorization gives the same result. However, it is unclear if this is true. Bohmian Mechanics solves the factorization problem in a way that is objective by supplying hidden variables that provide the objective factorization. But I think it also suggests that without hidden variables, the factorization is not objective - for example, can momentum be the hidden variable instead of position?

There seems to be a difference of opinion here. Please help me understand what's in dispute. Is the factorization that both of you referred to the the delineation between the "observer" and the "observed"?

 

[Feeble Wonk]

Ultimately, the one doing the measurement must be sentient or at least have "common sense" in Copenhagen. One can delegate the measurement to a non-sentient device. However, the important point is that the device is not considered quantum. The device is "macroscopic" or "classical". It is this subjective division of the world into classical and quantum parts that requires the observer that stands apart from the quantum system. Only the classical part is "real", while the wave function describing the quantum part may be just a tool to calculate the probability of measurement outcomes.

This is of course problematic, and is known as the measurement problem.
http://www.informationphilosopher.com/solutions/scientists/bell/Against_Measurement.pdf
http://www.tau.ac.il/~tsirel/download/nonaxio.html
http://arxiv.org/abs/quant-ph/0209123

Is the factorization you referred to also related to the quantum/classical (real/potentially real?) division you noted here?

 

[bhobba]

There seems to be a difference of opinion here. Please help me understand what's in dispute. Is the factorization that both of you referred to the the delineation between the "observer" and the "observed"?

What's in dispute is the role of conciousness. I say its exactly the same as in any theory - theories require a reasonably intelligent conscious entity to apply them - Atty thinks for QM it goes deeper than that for reasons I don't quite follow.

Thanks
Bill

 

[atyy]

Is the factorization you referred to also related to the quantum/classical (real/potentially real?) division you noted here?

I believe it is. In Copenhagen, the observer does 3 things:
1) choose a factorization, which is usually classical/quantum
2) choose a preferred basis
3) determine when an observation has occurred.

In Copenhagen, these are all subjective. bhobba and I do seem to have a disagreement as to how the measurement problem is stated. To me, the measurement problem is how do you get rid of the observer, but bhobba usually says it is "why is there a definite outcome"?

 

[atyy]

What's in dispute is the role of conciousness. I say its exactly the same as in any theory - theories require a reasonably intelligent conscious entity to apply them - Atty thinks for QM it goes deeper than that for reasons I don't quite follow.

Well, the part I don't follow is why you say that the measurement problem is "why is there a definite outcome?". If that is the problem, why not just postulate that there is a definite outcome and the measurement problem will be solved?

 

[bhobba]

Well, the part I don't follow is why you say that the measurement problem is "why is there a definite outcome?". If that is the problem, why not just postulate that there is a definite outcome and the measurement problem will be solved?

That's what I do - with the twist of decoherence that its legitimate to view it as being these prior to observation because its a mixed state.

Thanks
Bill

 

[stevendaryl]

I believe it is. In Copenhagen, the observer does 3 things:
1) choose a factorization, which is usually classical/quantum
2) choose a preferred basis
3) determine when an observation has occurred.

In Copenhagen, these are all subjective. bhobba and I do seem to have a disagreement as to how the measurement problem is stated. To me, the measurement problem is how do you get rid of the observer, but bhobba usually says it is "why is there a definite outcome"?

I see the measurement problem differently. Here's the way I would say it:

Suppose you have a system described by a state [itex]|\psi\rangle[/itex] and you have some variable [itex]A[/itex]. Then you can rewrite [itex]|\psi\rangle[/itex] in terms of states with definition values for [itex]A[/itex] as follows:

[itex]|\psi\rangle = \sum_\alpha C_\alpha |\psi_{A \alpha}\rangle[/itex]

where [itex]|\psi_{A \alpha}\rangle[/itex] is the normalized projection of [itex]|\psi\rangle[/itex] onto the subspace in which [itex]A[/itex] has eigenvalue [itex]\alpha[/itex].

Then having written [itex]|\psi\rangle[/itex] this way, we would like to say that variable [itex]A[/itex] for that system has value [itex]\alpha[/itex] with probability [itex]|C_\alpha|^2[/itex]. That's the Born rule, essentially.

However, we can also choose a different observable [itex]B[/itex] and write

[itex]|\psi\rangle = \sum_\beta D_\beta |\psi_{B \beta}\rangle[/itex]

where [itex]\psi_{B \beta}[/itex] is a normalized state in which [itex]B[/itex] has value [itex]\beta[/itex].

However, various no-go results such as theorems by Kochen-Specker and Bell show that it is not consistent to suppose that EVERY physical variable has a value simultaneously. So, the Born probability rule can't be taken to simultaneously give the probabilities of all possible variables. It only applies to one variable at a time (or a collection of commuting observables). So somehow some variable is singled out.

According to the Bohm theory, what is singled out is location in configuration space. According to Copenhagen, what is singled out is whichever variable we chose to measure. I don't particularly like that way of putting it, because to me, a measurement is just an interaction like any other, and the measuring device is a physical component like any other.

Rather than separating the physical situation into observed system and measuring device, and saying that the measuring device measures some observable of the observed system, it seems to me that you get the same effect if you just say that

There is a composite system (which includes the observed system and the measuring device). This composite system is described by a composite state. The variable that is singled out is the "pointer values" of the measuring device (if the measuring device has a pointer--otherwise, it's whatever macroscopic quantity the device records).

So to me, there is a quantum mystery, which is: Why are some physical variables singled out to have definite values? But I don't see why observers and measurements and so forth need to be special in QM.

 

[atyy]

That's what I do - with the twist of decoherence that its legitimate to view it as being these prior to observation because its a mixed state.

But it that's what you do, then your interpretation would solve the measurement problem. Yet in other posts you say it doesn't solve the measurement problem.

Which is it?

 

[atyy]

However, various no-go results such as theorems by Kochen-Specker and Bell show that it is not consistent to suppose that EVERY physical variable has a value simultaneously. So, the Born probability rule can't be taken to simultaneously give the probabilities of all possible variables. It only applies to one variable at a time (or a collection of commuting observables). So somehow some variable is singled out.

I think that is only part (2) of the measurement problem - the choice of preferred basis. Before that, one has to choose the Hilbert space, which is equivalent to choosing a classical/quantum cut. After one choose the preferred basis, one has to say when the measurement occurred.