Delayed Choice Quantum Eraser & The Observer (2000)

[atyy]

Assume that a device with a pointer is located close to ONE of the paths. After propagating further over a really long distance and time, the wave packet reaches a detection zone. Just before this happens, we can decide to switch the experiment between a setting where the pointer reveals the path information, to one where it can't. Depending on this decision, we se or don't see interference.

Now, does it matter how "macroscopic" the pointer device is? If it is just an atom or something, then it's plausible that it could give back the energy or whatever that it exchanged with the photon, and no one any the wiser. On the other hand, let's say the device is a photomultiplier tube, and the cascade of electrons should have reached the point of no return long before the wavepacket could reach the detection zone. What then? Can you put the toothpaste back in the tube even at this point? Or should we say that the cascade would be triggered nonlocally - or not triggered at all - depending on the scenario that exists in the detection zone at the last femtosecond? Would it make a difference whether or not the coherence length spans the PMT and the detection zone?

bhobba is right that for deriving the pointer states form decoherence, one usually needs a macroscopic number of degrees of freedom, However, if we are only talking about some aspects of the measurement, we can restrict our attention to just the system, or the system and a small part of the measurement apparautus often called an "ancilla", which can be a small quantum system. What matters is whether you can "reverse the interaction" between the ancilla and the system. If you can, then there is no collapse. If you cannot or do not want to control the ancilla enough to "reverse the interaction", then you can treat it as having collapsed the wave function. In this case, you typically do not observe the atom, so you can always delay the collapse. Collapse is only necessary when an observation is made, and a successive observation is made. An observation is when you get a definite outcome. Whether you have a definite outcome or not is subjective.

An example of an atom being used to cause collapse, provided we decide that we are not going to control the atom after it has interacted with the system is given in Fig. 1 of Zurek's http://arxiv.org/abs/quant-ph/0306072.

A related idea is the principle of deferred measurement.
http://en.wikipedia.org/wiki/Deferred_Measurement_Principle
http://www-bcf.usc.edu/~tbrun/Course/lecture07.pdf (p20)

 

[Derek Potter]

And yet, the macro device can "put the toothpaste back in the tube" long after the wavepacket has passed by it?

Not really. The erasing device is macroscopic but the observation is still just one photon carrying "which path" information. Such an observation is relatively easy to erase. However if it is amplified to a macroscopic pointer state, it cannot, for all practical purposes, be erased. As to your specific question, once the wavepacket has gone past the second set of detectors, the opportunity to erase the record is lost for ever.

Or is it that the toothpaste comes out of the tube (or doesn't) according to a nonlocal signal from the detection zone?

Interpreting the DCQE experiment in directly realist terms is hard. A signal would have to travel backwards in time from the idler detectors to the signal detector. This may, possibly, strike you as even more bizarre than spooky action at a distance. However it is all easily resolved if the whole system remains in superposition until "long after the wavepacket has passed by it". Of course some folk are a bit squeamish about macroscopic superpositions too, but you can't please everyone.

 

[Swamp Thing]

atyy, bhobba, Derek Potter: Thanks for the replies (and the links).

I realize that asking questions around classical analogies could be a bit annoying, but I can't resist this analogy (with apology) for atyy's description of an "ancilla":
A detector could consist of Schoedinger's flea biting or not biting an elephant. The flea would perhaps be an ancilla in a superposition, in which case you can still delay your choice... until a certain time by which the flea would have a non-negligible probability of having interacted with the elephant...

Sorry. Really.

 

[Derek Potter]

atyy, bhobba, Derek Potter: Thanks for the replies (and the links).

I realize that asking questions around classical analogies could be a bit annoying, but I can't resist this analogy (with apology) for atyy's description of an "ancilla":
A detector could consist of Schoedinger's flea biting or not biting an elephant. The flea would perhaps be an ancilla in a superposition, in which case you can still delay your choice... until a certain time by which the flea would have a non-negligible probability of having interacted with the elephant...

Sorry. Really.

Far less than an attosecond. After which quantum erasure is impossible because the flea has had an irreversible interaction with the elephant even though the idea of biting it hasn't crossed its mind. Of course it's only an analogy and all analogies break down sooner or later. In this case I would suggest sooner.

 

[atyy]

Sorry. Really.

That's alright. It's John Wheeler's fault. In comparison, Kaku deserves the Nobel.

http://schroedingersrat.blogspot.com/2013/11/do-not-work-in-quantum-foundations.html
"Experimental demonstration that the pagan god Mawu does not manifest when the pentacle is open. Did Mawu know in advance that we were going to open the pentacle? To appear in Nature Communications."

 

[Derek Potter]

That's alright. It's John Wheeler's fault. In comparison, Kaku deserves the Nobel.

He certainly deserves one for saying "It is often stated that of all the theories proposed in this century, the silliest is quantum theory. In fact, some say that the only thing that quantum theory has going for it is that it is unquestionably correct." Pity about the rest.

 

[Derek Potter]

Now you are shifting your claim.
First you need to define what you mean by relative state.

I do not see why I need to do anything of the sort. It is a term in common use. If you either don't understand or have issues with it, you should refer to Everett.

 

[bhobba]

I do not see why I need to do anything of the sort. It is a term in common use. If you either don't understand or have issues with it, you should refer to Everett.

All I am wanting is you to detail what you mean by these terms you bandy about. Its not hard.

Entanglement is a common term that's bandied about a lot and I have found peoples understanding of it often leaves a lot to be desired.

Its not hard - all you have to do is detail what you mean by it. I can state it in a few lines - that's all it takes.

Thanks
Bill

 

[Derek Potter]

Assume that a device with a pointer is located close to ONE of the paths. After propagating further over a really long distance and time, the wave packet reaches a detection zone. Just before this happens, we can decide to switch the experiment between a setting where the pointer reveals the path information, to one where it can't. Depending on this decision, we se or don't see interference.

Now, does it matter how "macroscopic" the pointer device is? If it is just an atom or something, then it's plausible that it could give back the energy or whatever that it exchanged with the photon, and no one any the wiser. On the other hand, let's say the device is a photomultiplier tube, and the cascade of electrons should have reached the point of no return long before the wavepacket could reach the detection zone. What then? Can you put the toothpaste back in the tube even at this point? Or should we say that the cascade would be triggered nonlocally - or not triggered at all - depending on the scenario that exists in the detection zone at the last femtosecond? Would it make a difference whether or not the coherence length spans the PMT and the detection zone?

This needed an answer before you got sidetracked into fleas and elephants.
The idea of a photodetector holding its breath until a remote photon makes up its mind whether to be detected at A, B, C or D is precisely what is implied by conventional QM. That is to say, the predictions are wrong if you assume the first detectors register their results immediately. Highly problematical for direct realism, no problem at all for Many Worlds, no problem for "shut up and calculate" (of course) and fertile ground for woo-mongers ...
A long coherence length might make interpretation less difficult, but in fact the wavepacket is very short indeed.

 

[bhobba]

Why do say "bandy about"? I am using the term correctly.

Dear oh dear its obvious you won't define it. If you did you may see the error you are making wrt MW.

Here is the definition. Suppose two systems are entangled eg 1/root 2 |a>|b> + 1/root 2 |b>|a>. The systems are entangled hence neither has an actual state. If system 1 is observed in state |a> then system 2 will be in state |b>, and conversely, if system 1 is observed in state |b>, then system 2 will be in state |a>. State |a> is relative to state |b> and conversely.

Now your claim was 'MW is a principle which follows tautologically from the assumption of linear (quantum) evolution.'. In fact in Everett's original interpretation, without the modern use of decoherence, used the concept of relative state, and INTERPRETED the relative states as separate worlds ie |a>|b> was one world and |b>|a> another. Its a key interpretive assumption - it does not follow tautologically because it requires that assumption - tautologies can't use extra things put into it - by definition a tautology is a circular statement.

Thanks
Bill

 

[Derek Potter]

Dear oh dear its obvious you won't define it. If you did you may see the error you are making wrt MW.

Here is the definition. Suppose two systems are entangled eg 1/root 2 |a>|b> + 1/root 2 |b>|a>. The systems are entangled hence neither has an actual state. If system 1 is observed in state |a> then system 2 will be in state |b>, and conversely, if system 1 is observed in state |b>, then system 2 will be in state |a>. State |a> is relative to state |b> and conversely.

Now your claim was 'MW is a principle which follows tautologically from the assumption of linear (quantum) evolution.'. In fact in Everett's original interpretation, without the modern use of decoherence, used the concept of relative state, and INTERPRETED the relative states as separate worlds ie |a>|b> was one world and |b>|a> another.

Its a key interpretive assumption - it does not follow tautologically because it requires that assumption - tautologies can't use extra things put into it - by definition a tautology is a circular statement.

Thanks
Bill

There is no assumption involved. The term "world" is fanciful and Everett did not like it, but to identify |a>|b> as a world is not an assumption, it is just a definition.

 

[bhobba]

There is no assumption involved. The term "world" is fanciful and Everett did not like it, but to identify |a>|b> as a world is not an assumption, it is just a definition.

Derek -- you are a bit confused between definition and assumption - world in MW is much more than a definition.

Thanks
Bill

 

[Derek Potter]

Derek -- you are a bit confused between definition and assumption - world in MW is much more than a definition.

Thanks
Bill

No confusion here, Bill. Rather than repeatedly insulting my intelligence I suggest you stick to the point and say what you think is the assumption.

 

[Derek Potter]

No confusion here, Bill. Rather than repeatedly insulting my intelligence I suggest you stick to the point and say what you think is the assumption.

... as I have asked you several times already.

 

[Dale]

And on that note, thread closed.