Decoherence relationship to the measurement problem

[Feeble Wonk]

Decoherence relationship to the "measurement problem"

I have heard the argument that "decoherence" in quantum states causes the quantum collapse to occur, and that this solves the "measurement problem". But I'm still left with a nagging question... Does decoherence only produce quantum collapse when there is an observation made of the quantum system with sufficient specificity to identify the decoherence. If so, how can that possibly be considered a solution to the "measurement problem"?Sent from my iPhone using Physics Forums

 

[atyy]

Decoherence does not solve the measurement problem. At the end of decoherence one gets a reduced density matrix that is an improper mixture, whereas to solve the measurement problem it has to be a proper mixture. The treatment of an improper mixture as a proper mixture is an additional postulate equivalent to introducing wave function collapse.

 

[Jilang]

I am a bit unclear as to how one distinguishes between a proper and an improper mixed state. Is there an experiment that differentiates them?

 

[StevieTNZ]

I am a bit unclear as to how one distinguishes between a proper and an improper mixed state. Is there an experiment that differentiates them?

See GianCarlo Ghirardi's thought experiment in his book "Sneaking a Look at God's Cards".

 

[Jilang]

Thanks Stevie, have you got a page number for that?

 

[Feeble Wonk]

If I can, I'd like to return to my previous question that I'm still confused by. Is observation/measurement required with sufficient specificity to identify quantum decoherence to cause the entangled quantum system to collapse? Does the entire entangled quantum system remain in superposition to the extent that observable detail allows, regardless of potential decoherent states "below" observable detail? Sent from my iPhone using Physics Forums

 

[atyy]

I am a bit unclear as to how one distinguishes between a proper and an improper mixed state. Is there an experiment that differentiates them?

If one measures observables confined to the subsystem, then no experiment can distinguish proper and improper mixed states. However, if one can measure the system and the environment then proper and improper mixed states can be distinguished. http://philsci-archive.pitt.edu/5439/1/Decoherence_Essay_arXiv_version.pdf (see section 1.2.3 on p10)

Regardless of what one measures, decoherence does not solve the measurement problem because the system as a whole is still in a superposition of measurement outcomes. It does not make sense to say there is an outcome on the subsytem, but no outcome on the environment.

If I can, I'd like to return to my previous question that I'm still confused by. Is observation/measurement required with sufficient specificity to identify quantum decoherence to cause the entangled quantum system to collapse? Does the entire entangled quantum system remain in superposition to the extent that observable detail allows, regardless of potential decoherent states "below" observable detail?

Decoherence does not solve the measurement problem. For decoherence to solve it for "practical purposes", we have to make the assumption that an improper mixed state is a proper mixed state.

http://arxiv.org/abs/quant-ph/0312059
As Pessoa (1998, p. 432) puts it, "taking a partial trace amounts to the statistical version of the projection postulate."

https://www.amazon.com/dp/0198509146/?tag=pfamazon01-20 (see "The quantum-classical boundary" on p80)

What does decoherence solve? A nice example is why some molecules are chiral. Also, it shows the emergence of a "pointer basis". For solutions to the measurement problem, one has to look to de Broglie-Bohm theory, or perhaps many-worlds (among other possibilities).

 

[Feeble Wonk]

So... If I understand what you are saying... the quantum state of any system exists in a superposition state to the degree allowed by observable measurement. And in that system there will exist decoherent potential states that will remain in superposition until such time as there is a observable event and/or measurement that dictates a collapse to one of those states. Is that correct? Sent from my iPhone using Physics Forums

 

[bhobba]

I am a bit unclear as to how one distinguishes between a proper and an improper mixed state. Is there an experiment that differentiates them?

That's the point - no experiment can.

That's why decoherence explains APPARENT collapse.

Thanks
Bill

 

[bhobba]

See GianCarlo Ghirardi's thought experiment in his book "Sneaking a Look at God's Cards".

It doesn't do what Stevie claims.

Its an experiment to determine the difference between a superposition and a mixed state.

Here is the link:
http://books.google.com.au/books?id...ng a look at god's cards appendix 15a&f=false

It's an experiment to determine the difference between |final> = 1/root 2 (|V, A+> + |H, A->) and an ensemble ie a mixed state. |final> is a pure state which is different from an ensemble which is a mixed state and is expressed as operators - not vectors.

I posted this in another thread but will repeat it here:

The issue is states are not elements of a vector space as some books, especially those at the intermediate level like Griffiths, will tell you. They are in fact positive operators of unit trace defined by the general form of the Born Rule. To really grasp it you need to see the two axioms of QM as detailed by Ballentine in his text.

1. To each observation there corresponds a Hermitian operator whose eigenvalues give the possible outcomes of the observation.
2. There exists a positive operator of unit trace P such that the expected outcome of the observation associated with the observable O is E(O) = Trace (PO) - this is the Born Rule in its most general form. By definition P is called the state of the system.

In fact the Born Rule is not entirely independent of the first axiom, as to a large extent it is implied from that via Gleason's Theroem - but that would take us too far afield - I simply mention it in passing.

Also note that the state, just like probabilities, is simply an aid in calculating expected outcomes. Its not real like say an electric field etc. In some interpretations its real - but the formalism of QM is quite clear - its simply, like probabilities, an aid in calculation.

By definition states of the form |x><x| are called pure. States that are a convex sum of pure states are called mixed ie are of the form ∑ pi |xi><xi| where the pi a positive and sum to one. It can be shown all states are either pure or mixed. Applying the Born rule to mixed states shows that if you have an observation whose eigenvectors are the |xi><xi| then outcome |xi><xi| will occur with probability pi. Physically one can interpret this as a system in state |xi><xi| randomly presented for observation with probability pi. In such a case no collapse occurs and an observation reveals what's there prior to observation - many issues with QM are removed. Such states are called proper mixed states.

Pure states, being defined by a single element of a vector space, can be associated with those elements and that's what's usually done. Of course when you do that they obey the vector space properties so the principle of superposition holds ie if |x1> and |x2> are any two pure states a linear combination is also a pure state. This is what is meant by a superposition. Note it deals with elements of a vector space not convex sums of pure states when considered operators - they are mixed states. This means the state 1/root 2 |x1> + 1/root 2 |x2> is a pure state and is totally different from the mixed state 1/2 |x1><x1| + 1/2 |x2><x2|.

Now what decoherence does is transform a superposition like |x> = 1/root 2 |x1> + 1/ root 2 |x2> into a mixed state like X = 1/2 |x1><x1| + 1/2 |x2><x2|. When that is done it can be interpreted as a proper mixed state which solves many of the issues with collapse etc.

The thought experiment Stevie refers to determines the difference between a pure state |x> represented by a vector and the mixed state X represented by an operator. Doing an observation on either for |x1> and |x2> gives exactly the same result. However if you observe it for different things you can tell the difference, as the thought experiment demonstrates. An improper mixed state and a proper mixed state are exactly the same mathematically and represented by the operator X - there is no way to tell the difference.

Thanks
Bill

 

[atyy]

So... If I understand what you are saying... the quantum state of any system exists in a superposition state to the degree allowed by observable measurement. And in that system there will exist decoherent potential states that will remain in superposition until such time as there is a observable event and/or measurement that dictates a collapse to one of those states. Is that correct?

I don't understand exactly what you are saying, but yes, that seems essentially correct.

 

[bhobba]

So... If I understand what you are saying... the quantum state of any system exists in a superposition state to the degree allowed by observable measurement. And in that system there will exist decoherent potential states that will remain in superposition until such time as there is a observable event and/or measurement that dictates a collapse to one of those states. Is that correct?

Cant say I understand what you are saying.

The easiest way to understand it is as tracing over the environment:
http://physics.stackexchange.com/qu...ake-the-partial-trace-to-describe-a-subsystem

Whats going on is the system being observed and the environment become entangled, which via the math detailed in the link leads to the system being observed being in an effective (another word is improper) mixed state. It looks exactly the same as a proper mixed state. A proper mixed state is a state that is randomly presented for observation and if it was a proper mixed state collapse would have occurred. But since it only LOOKS like a proper mixed state it only LOOKS like collapse has occurred - it really hasn't - which is why the word APPARENT is used.

However there is nothing stopping assuming it actually occurred there, Von-Neumann showed it could be placed anywhere, and in doing so many issues are solved.

That said, and without going into the details - issues do remain - but of course research is ongoing.

Thanks
Bill

 

[Maui]

That's the point - no experiment can.

That's why decoherence explains APPARENT collapse.

Thanks
Bill

This setup was invented so that people would believe that it's the environement that does the collapse. It doesn't(experiments prove this).

The whole setup is flawed from the beginning as it assumes that improper mixed states are the same as proper mixed states. They are not!

Then you add contrived structures that you label 'environment' which are foreign to qm. Of course decoherence is guaranteed to take place once you introduce that which all interpretations have as their endgoal - classicality.

Ok it's not religion, but if someone said it was, i wouldn't argue.

 

[atyy]

With respect to the measurement problem, I think decoherence is interesting for the many-worlds interpretation, which is a potential solution to the measurement problem. One problem with many-worlds is that the branches are ambiguous because a state can be decomposed in many ways. Decoherence by choosing a "pointer" basis could help that. It still needs some finessing, since decoherence is never perfect. There are very interesting ways to do the finessing in David Wallace's "The Emergent Multiverse" https://www.amazon.com/dp/0199546967/?tag=pfamazon01-20

 

[Maui]

With respect to the measurement problem, I think decoherence is interesting for the many-worlds interpretation, which is a potential solution to the measurement problem. One problem with many-worlds is that the branches are ambiguous because a state can be decomposed in many ways. Decoherence by choosing a "pointer" basis could help that. It still needs some finessing, since decoherence is never perfect. There are very interesting ways to do the finessing in David Wallace's "The Emergent Multiverse" https://www.amazon.com/dp/0199546967/?tag=pfamazon01-20

You still have to assume that improper mixed states are the same as proper one so that there be interactions and world splitting. This injects philosophy so it cannot be science.

 

[bhobba]

You still have to assume that improper mixed states are the same as proper one so that there be interactions and world splitting. This injects philosophy so it cannot be science.

Its not philosophy that there is no experiment that can tell the difference.

Until there is a way to tell the difference between interpretations experimentally its purely a personal choice based on all sorts of thing of which philosophy is just one.

Thanks
Bill

 

[Maui]

Its not philosophy that there is no experiment that can tell the difference.

Experiments prove that the world is quantum and not classical. You should not pretend you are explaining the quantum to classical transition by stating the obvious - that you always observe a classical world.

Until there is a way to tell the difference between interpretations experimentally its purely a personal choice based on all sorts of thing of which philosophy is just one.

Thanks
Bill

Yes, that I agree with. As long as you label your decoherence pet theory as another philosophical interpretation, we are on the same page. But you are presenting it as a significant progress on the measurement problem and that is... er... yes, ridiculous.

 

[atyy]

You still have to assume that improper mixed states are the same as proper one so that there be interactions and world splitting.

Why do you say that? It's not obvious to me since the pointer basis is selected without having to assume that improper and proper mixed states are the same.

 

[Maui]

Why do you say that? It's not obvious to me since the pointer basis is selected without having to assume that improper and proper mixed states are the same.

How is the pointer basis selected? If you do not assume they are one and the same(for all practical purposes) you will get no decoherence but pure states.

 

[eloheim]

Experiments prove that the world is quantum and not classical. You should not pretend you are explaining the quantum to classical transition by stating the obvious - that you always observe a classical world.

This seems as relevant a place as any to ask a question I've had. What would it mean to "observe a quantum world?" I mean we see non-classicality in experiments all the time. I assume when people say the above they mean macroscopic non-classicality? Or how would the world appear exactly, so that we knew it was the case?

 

[atyy]

How is the pointer basis selected? If you do not assume they are one and the same(for all practical purposes) you will get no decoherence but pure states.

As I understand it, in many-worlds one gets decoherence (selection of a pointer basis) and a pure state. Since there is no collapse, there is no transition from improper to proper mixture in many-worlds. The pointer basis is selected because if one has environment, apparatus, and system, perfect decoherence picks out a unique pointer basis.

There are still problems, but I don't think it has to do with decoherence leaving the system in a pure state. One problem is that decoherence is never perfect, so there may not be true uniqueness. Another issue is whether there is a unique division of the system into environment, apparatus and system.

 

[Feeble Wonk]

"This seems as relevant a place as any to ask a question I've had. What would it mean to "observe a quantum world?" I mean we see non-classicality in experiments all the time. I assume when people say the above they mean macroscopic non-classicality? Or how would the world appear exactly, so that we knew it was the case?"

Finally! This is the point I've been hoping to get to. I've been trying to follow your discussion with great interest and gratitude. Being a biologist rather than a physicist, some of the concepts are new to me, and I'm trying to acquire some of the sources you referred to. But it's the issue of "observation" that I think is at the core of your argument. It seems to me that macro world (classical) observations are simply course grain averages of all potential quantum states in superposition, suspended in that condition by the observational limitation in the specificity of the state being observed. Observing the "quantum world" (identifying the precise quantum state of any portion of the entangled system) would essentially produce a measurement to that degree of "fine grain" specificity. That would produce decoherence in the potential quantum states, and trigger a collapse. But WHAT constitutes the observation? Does the simple presence of the particle detector in Schrodinger's box make the observation, or does someone have to observe the detector? That's my real question. Does "decoherence" in potential quantum states, in and of itself, produce the observation and/or measurement?

 

[atyy]

"This seems as relevant a place as any to ask a question I've had. What would it mean to "observe a quantum world?" I mean we see non-classicality in experiments all the time. I assume when people say the above they mean macroscopic non-classicality? Or how would the world appear exactly, so that we knew it was the case?"

Finally! This is the point I've been hoping to get to. I've been trying to follow your discussion with great interest and gratitude. Being a biologist rather than a physicist, some of the concepts are new to me, and I'm trying to acquire some of the sources you referred to. But it's the issue of "observation" that I think is at the core of your argument. It seems to me that macro world (classical) observations are simply course grain averages of all potential quantum states in superposition, suspended in that condition by the observational limitation in the specificity of the state being observed. Observing the "quantum world" (identifying the precise quantum state of any portion of the entangled system) would essentially produce a measurement to that degree of "fine grain" specificity. That would produce decoherence in the potential quantum states, and trigger a collapse. But WHAT constitutes the observation? Does the simple presence of the particle detector in Schrodinger's box make the observation, or does someone have to observe the detector? That's my real question. Does "decoherence" in potential quantum states, in and of itself, produce the observation and/or measurement?

In the shut-up-and-calculate Copenhagenish interpretation, measurement is indeed fuzzy and quantum mechanics is fuzzy. There is a common sense macroscopic reality which includes our experimental results, and there is a quantum realm. To use quantum mechanics we always make this classical/quantum cut, and and we say that a measurement is when the macroscopic apparatus registers a macroscopic reading by probing the quantum realm. It works, but of course it is not satisfactory, which is the measurement problem.

Decoherence does not solve the measurement problem, although it can be part of the solution, for example, in the context of many-worlds. But in many-worlds all outcomes occur.

If we stick to each measurement producing one outcome, can we treat the wave function as "real", but providing incomplete information because it is a "coarse graining" of unseen variables? Yes! That is what the de Broglie-Bohm theory does for non-relativistic quantum mechanics. This is why the de Broglie-Bohm theory is a solution of the measurement problem for non-relativisitic quantum mechanics.

There appears to be a slightly different class of solutions to the measurement problem, also involving hidden variables. In de Broglie-Bohm theory, the wave function is a coarse grained variable in the sense that if we know the true state of the underlying variables, then that specifies the wave function. In this alternative class of theories, a given true underlying state can correspond to more than one wave function, so the wave function is not a traditional "real" coarse grained variable, and bit more like a "state of knowledge".

Incidentally, the question of fine/coarse grained is also important in David Wallace's version of the many-worlds approach, in which decoherence is an important element.

 

[Feeble Wonk]

Perhaps a better way if framing my question would be this... Does the concept of "decoherence" imply any mechanism of action what so ever. Or rather, is it simply a reference to the general idea that any sequence of quantum events will eventually result in potential quantum states that cannot logically coexist... that the logically coherent evolution of any quantum system will result in an either/or status and only one of the potential states can be manifestly observed.

 

[atyy]

Perhaps a better way if framing my question would be this... Does the concept of "decoherence" imply any mechanism of action what so ever. Or rather, is it simply a reference to the general idea that any sequence of quantum events will eventually result in potential quantum states that cannot logically coexist... that the logically coherent evolution of any quantum system will result in an either/or status and only one of the potential states can be manifestly observed.

Decoherence doesn't solve the measurement problem. In Schroedinger's cat, the cat evolves into a superposition of "dead" and "alive". However, that is not the only way to see it, the cat also evolves into a superposition of "dead-plus-alive" and "dead-minus-alive". So what are the possible outcomes? As a matter of fact, we only get a cat that is "dead" or a cat that is "alive". Why don't we get a cat that is "dead-plus-alive" or "dead-minus-alive", even though that is a valid superposition? What decoherence tells us is that if we include the quantum state of the environment, apparatus and cat, then the system is in a superposition in which the cat is "dead" and "alive". So decoherence gets rid of the superposition of "dead-plus-alive" and "dead-minus-alive". It doesn't solve the measurement problem because the system is still in a superposition of the cat being "dead" and "alive". But at least we got rid of the possibilities of "dead-plus-alive" and "dead-minus-alive".

Decoherence is just normal evolution of the wave function without collapse, except that we include the environment, apparatus and cat in the wave function. The system of environment, apparatus and cat as a whole does not decohere. Although the system as a whole does not decohere, if we only look at the cat, then "dead" or "alive" has been chosen over "dead-plus-alive" or "dead-minus-alive".

 

[Feeble Wonk]

So, if I understand you properly, the concept of decoherence is really the logical/mathematical process that addresses the entire quantum system ("environment") and rules out the event outcomes that cannot logically exist AT ALL, but still leaves all potential quantum states that CAN logically exist. It then requires something like Everett's "multiverse" theory to whittle down the remaining states to one that is actually manifest. Right?

 

[atyy]

So, if I understand you properly, the concept of decoherence is really the logical/mathematical process that addresses the entire quantum system ("environment") and rules out the event outcomes that cannot logically exist AT ALL, but still leaves all potential quantum states that CAN logically exist. It then requires something like Everett's "multiverse" theory to whittle down the remaining states to one that is actually manifest. Right?

Technically, what it says is that if we have control over the environment, then "dead-plus-alive" and "dead-minus-alive" cats are still logically possible. However, we never have enough control over the environment so the possibilities left to us are "dead" or "alive". And yes, something else like collapse or many-worlds is needed to pick one outcome, be it "dead" or "alive".

 

[Feeble Wonk]

Fantastic. Thanks so much for your patience. As I said... I'm a biologist and not a physicist. I'm struggling with the whole mind/body thing, and I need a better grasp of physics to wrap my mind (no pun intended) around it. It all seems to have more to do with information processing than physical mechanism. I have so many other questions, but it would take us far afield from this thread, and I don't want to waste your time. Thanks again.

 

[stevendaryl]

This seems as relevant a place as any to ask a question I've had. What would it mean to "observe a quantum world?" I mean we see non-classicality in experiments all the time. I assume when people say the above they mean macroscopic non-classicality? Or how would the world appear exactly, so that we knew it was the case?

I think that's a good point. Let me expand on it a little:

The obvious answer to the question "what does it mean that we observe a classical world" is that we don't see superpositions of macroscopic objects. I'm either in New York, or I'm in Georgia, but there is a never a time when I'm in a superposition of being halfway in one location, and halfway in another. But if you push farther you could ask: Well, what would it "look like" if you were in a macroscopic superposition? Then there really is no good answer to that question. If I resolve to record my observations in a notebook, and write down either the sentence "I'm in Georgia" or "I'm in New York", then in the case of a superposition, I won't write down "Help! I'm in a macroscopic superposition!" Instead, my superposition will just "infect" my notebook, so that it's in a superposition. And it quickly spreads to the whole universe. That's the Many-Worlds view: being in a macroscopic superposition is not observable (or at least, not easily), and so we can't really use our observations to rule out the possibility.

 

[Maui]

As I understand it, in many-worlds one gets decoherence (selection of a pointer basis) and a pure state. Since there is no collapse, there is no transition from improper to proper mixture in many-worlds. The pointer basis is selected because if one has environment, apparatus, and system, perfect decoherence picks out a unique pointer basis.

There are still problems, but I don't think it has to do with decoherence leaving the system in a pure state. One problem is that decoherence is never perfect, so there may not be true uniqueness. Another issue is whether there is a unique division of the system into environment, apparatus and system.

Well, yes, that's how I think it's usually presented. But I do not see how the bolded part is warranted so the whole idea is mostly philosophy. You either accept it or don't.

 

[Feeble Wonk]

I shouldn't ask this, but I can't help myself... stevendaryl raises a very pertinent question. What determines the limit of "perceived" quantum superposition? I can't be in Georgia and New York at the same time, because it wouldn't be a logically coherent quantum state. But, I can perceive the interference bars in the double slit experiment. They form a macroscopically observable structure that represents a confluence of superpositional outcomes. What does this say about the state of the entire quantum system that includes the brain of the observer?

 

[atyy]

I shouldn't ask this, but I can't help myself... stevendaryl raises a very pertinent question. What determines the limit of "perceived" quantum superposition? I can't be in Georgia and New York at the same time, because it wouldn't be a logically coherent quantum state. But, I can perceive the interference bars in the double slit experiment. They form a macroscopically observable structure that represents a confluence of superpositional outcomes. What does this say about the state of the entire quantum system that includes the brain of the observer?

In a double slit experiment you still perceive definite outcomes. The interference bars are made of macroscopic spots.

 

[atyy]

Well, yes, that's how I think it's usually presented. But I do not see how the bolded part is warranted so the whole idea is mostly philosophy. You either accept it or don't.

If there is only one decomposition into environment/system/apparatus in which decoherence is perfect, the problem is solved. Of course, decoherence is not perfect, so the problem is open. But in a sense, solving the imperfection of decoherence could solve the environment/system/apparatus decomposition problem. I don't know if it really works, but I'd say try David Wallace's approach for a serious attempt at addressing these issues.

 

[Feeble Wonk]

In a double slit experiment you still perceive definite outcomes. The interference bars are made of macroscopic spots.

Well, yes. That's my point. They ARE macroscopic spots. Spots that are what they are, and are WHERE they are because of the unresolved quantum state of the fundamental particles' passage through the slits. They are a macroscopic representation of a quantum state superposition.

 

[atyy]

Well, yes. That's my point. They ARE macroscopic spots. Spots that are what they are, and are WHERE they are because of the unresolved quantum state of the fundamental particles' passage through the slits. They are a macroscopic representation of a quantum state superposition.

Yes, but even a dead cat is a macroscopic representation of a quantum superposition - it is a superposition of "dead-plus-alive" and "dead-minus-alive". I know I just seem to have contradicted myself here, but here unlike above I'm assuming there's no environment that has decohered the cat, because in the double slit experiment there is enough experimental control so that the environment has not decohered the quantum system.