Brian Green's Beam Splitter Experiments

[selfAdjoint]

Unless I'm missing a crucial point or the wikipedia article is misleading... if this is the point, please clarify.

Wikipedia of course can be updated by anybody and only if there is a protest do many of these get corrected. So although it is tremendously useful, there are times when Wiki is misleading or just plain wrong. In my personal opinion, the line you quote and speculations about time travel are cases of this.

 

[Rake]

"It would seem that the 'choice' to observe or erase the which-path information can change the position where the photon is recorded on the detector, even after it should have already been recorded."
- Wikipedia

If wiki is correct then this would mean that an experiment can be devised such that a future event can change the past.

So for example, let's say that the experiment is changed so that the paths to the idler detectors D1,D2,D3,D4 are stretched out by 1 light year.

Then would this would mean that if an experimenter is staring at the spot on D0 where the twin signal made a mark precisely 1 year earlier then they are going to see that mark disappear and reappear somewhere else depending on the twin idlers recording by D1-D4?

This does not sound right...

 

[moving finger]

Gosh, one might conclude from this that the future influences the past. That view would certainly make a mess of interpretations or theories that rely on non-local mechanisms, n'est pas?

Not necessarily - check out Cramer's Transactional Interpretation of QM - in this (non-local) interpretation the future "influences" the past just as much as the past "influences" the future. Nothing strange about that notion if one believes in strict causal determinism.

http://en.wikipedia.org/wiki/Transactional_interpretation

Best Regards

 

[moving finger]

If wiki is correct then this would mean that an experiment can be devised such that a future event can change the past.

So for example, let's say that the experiment is changed so that the paths to the idler detectors D1,D2,D3,D4 are stretched out by 1 light year.

Then would this would mean that if an experimenter is staring at the spot on D0 where the twin signal made a mark precisely 1 year earlier then they are going to see that mark disappear and reappear somewhere else depending on the twin idlers recording by D1-D4?

This does not sound right...

It's not just wiki that claims this. Check the link provided by SelfAdjoint :

The position of a photon at detector D0 has been registered and scanned. Yet the actual position of the photon arriving at D0 will be at one place if we later learn more information; and the actual position will be at another place if we do not.

Ho-hum. Another experimental proof of QM. This is the way it works, folks.

What is missing is an explanation of the ontology - an explanation of just what is going on. One way it can be understood is in terms of the "future influencing the past", in exactly the way that Cramer's Transactional Interpretation describes.

Can anyone come up with another explanation?

Best Regards

 

[selfAdjoint]

Finger, you keep plugging Cramer. I used to like Cramer too, but how does it work now that the "future absorber" it depends on, once identified with the supposed future big collapse of the universe, has apparently been done away with by the accelerated expansion?

 

[selfAdjoint]

Well then, here's a thought experiment / question:

What if we observe the which-path information, in which case we see no interference pattern, and then erase the which-path information. Would we then see the interference pattern? Probably not...

At this case "our knowledge" of the which-path information would affect the outcome of the experiment..?

No this fails as an explanation. It was our actions, not our knowledge, that caused the behavior. And it is easily possible to imagine our actions performed automatically, completely independent of our knowledge.

Let me add another thing that confuses people thinking about this. It is natural to think of the photons as lttle bullets that have a succession of well defined positions along the various paths as they "move" from source to target. But QM says no such thing! Whether you use the "wave function" view or the Feynman "sum over histories" view, there are no successive postions. Source, paths, target are all described by one extended "shape" in spacetime. And that shape constrains what happens, not "particle motion".

 

[kvantti]

Yes I realized that and therefore deleted my post... apoligies.

 

[Rake]

The position of a photon at detector D0 has been registered and scanned. Yet the actual position of the photon arriving at D0 will be at one place if we later learn more information; and the actual position will be at another place if we do not.

Ho-hum. Another experimental proof of QM. This is the way it works, folks.

The first sentence states that the position of the photon at detector D0 has been registered. And I think that this never changes.

What changes is the position of the photon ARRIVING at D0. MEANING that once the twin idler info has been gleaned, then the experimenter will be able to say something about the PATH that the twin signal took to get to the position on D0.

I think that the wording of the conclusion is poor and is what is misleading here. The past is not being changed at all. Only new information is being attained about a past event with the recording of the idler. That information has to do with the position of the photon as it arrives D0.

So the experimenters do not know which of the photons make up an interference pattern vs. which ones do not until the arrival and subsequent correlation of the idlers into D1-D4 has been completed.

 

[MojaveJoe]

The experiment is set up to seem to demonstrate that ("Delayed Choice"), but consider; if automatic software were designed to (pseudo-)randomly scan or not scan with no human the wiser (double-blind) as to which "choices" had been made and only long after did anyone look at the results. Do you think the same patterns of behavior would result, or not? I believe they would. I think the "our knowledge" issue is a red herring, an artifact of the experimental design.

It does seem like a red herring doesn't it? How can nothing tangible (the "quantum information"), have a tangible effect on the experimental outcome? But that's exactly what this experiment implies and that's why I asked...swore I misunderstood the experiment. I assumed there was some hidden variable or something...

The problem is how the heck do you set up software to monitor something but we can't "know" the answer or that screws up the experiment!

I take that back, quantum information isn't nothing...to me, its the 800 lb gorilla in the room!

Cheers!

 

[eep]

The first sentence states that the position of the photon at detector D0 has been registered. And I think that this never changes.

What changes is the position of the photon ARRIVING at D0. MEANING that once the twin idler info has been gleaned, then the experimenter will be able to say something about the PATH that the twin signal took to get to the position on D0.

I think that the wording of the conclusion is poor and is what is misleading here. The past is not being changed at all. Only new information is being attained about a past event with the recording of the idler. That information has to do with the position of the photon as it arrives D0.

So the experimenters do not know which of the photons make up an interference pattern vs. which ones do not until the arrival and subsequent correlation of the idlers into D1-D4 has been completed.

I don't see how this fixes anything. But I also don't understand how a registered position could "change". I assume it's not possible that the photon that hits D0 already "knows" what it's entangled partner is going to do - that would imply so-called hidden variables, correct?

 

[selfAdjoint]

David Lindley has written a book called https://www.amazon.com/gp/product/0465067867/?tag=pfamazon01-20 about quantum theory. Here are a couple of excerpts from his discussion of the delayed choice experiment:

If we want to know which way the photon went, we can't see the interference pattern. If we want to see an interference pattern, we can't ask to know which way the photon went.
...
The problems only arise when we surrender to the urge to start imagining, once we know what the outcome of an experiment is,that we can deduce accurately what must have happened along the way. That is, if we see an interference pattern, we think the photon must have divided itself into separate pieces, whereas if we know the photon was detected along one pathway, that it must have actually gone down that pathway and not the other. But then when we do a delayed choice experiment, we find that insisting on one or the other of these two mutually exclusive interpretations leads us into trouble, because it appears that the photon must know in advance what is going to happen, so that it can choose the appropriate behavior to follow. But, and this is the fundamental point, those two possible behaviors are not actual behaviors that we know the photon must in fact have followed, but inferred, deduced, or (more accurately) speculated behaviors...

The lesson of the delayed choice experiment is that any hypothesis that forces the photon to adopt one or the other of two distinct and contradictory behaviors is, in fact, not reasonable.

And we should have known this in advance. What was the first thing we learned about quantum measurements from the Stern-Gerlach experiment? That the spin of an electron is strictly an undetermined quantity until an experiment yields a measurable value for it. And, moreover, that any attempt to say, having found the electron to be in an "up" orientation, that it must have been that way all along, is demonstrably incorrect.
..
And the delayed choice experiment brings that point forcefully into the open. Seeing an interfrence pattern, or detecting a photon in one path or the other, is at bottom a simple measurement made on a complicated system. And any attempt to think that we know, once the measurement is done, what really went on inside the system (whether the photon went one way or the other, or both ways at once) is precisely an attempt to pin down the prior state of a system after a measurement has been made. It causes trouble and we musr resist the temptation to do it.

The minimal interpretation of quantum mechanics embedded in the first two sentences of this {quotation} is like a wise but stern parental injunction against certain kinds of teenage behavior: limiting, to be sure, but it keeps people out of danger.

Quantum mechanics says the state, whether particle or wave, doesn't exist until an interaction produces a value for it; and furthermore that value is for that interaction only and gives no warrant for saying the value existed going into the interaction.


.

 

[MojaveJoe]

Quantum mechanics says the state, whether particle or wave, doesn't exist until an interaction produces a value for it; and furthermore that value is for that interaction only and gives no warrant for saying the value existed going into the interaction.

That seems to answer the delayed choice past/present question - or at least state this is acceptable wierdness. However, it still doesn't seem to address the fact that the final data depends on whether or not we know the which path (a real empirical effect)...except that it is "dangerous" to draw this conclusion. That is unsatisfactory to me. Granted, I assumed I was wrong when I posted the initial question and I am content to simply know that I didn't misunderstand the experiment, but...

What exactly is the author afraid of?

Cheers!

Edit: I guess what I mean is I'm not asking whether there is an interference pattern or not or whether the photon went one way or both ways...what I am asking is why a particular type of knowledge has ANY effect at all. It shouldn't! Heck, it can't! Granted it could be, like you said, an experimental artifact...but to me that means we should dismiss an awful lot of experiments.

 

[MeJennifer]

Of course there is a completely different interpretation of all these "misteries".

So we have a photon, two slits, a screen, and a measuring device that is on or off and an experiment to follow.

What if the result is already determined classically before the experiment starts?
What if it is just us, experimenters, who don't know it yet.
And in our "infinite wisdom" we think that we are the ones who can influence the outcome of the experiment while in fact our actions are determined by the outcome.
Are we the owners taking the dog for a walk or is it perhaps the other way around?

As somebody once wrote "what if we are the hidden variables"?

Is anyone with me on this thought?

 

[Farsight]

Me.

http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm

If you look at page 16 and count the dots on figs 3 4 and 5, you see there's twice as many joint detections for R₀₁ and R₀₂ as for R₀₃₍₀₄₎.

In total there's four times as many, but I doubled the number of detections on fig 5 because they occur on both R3 and R4.

Does this mean there's something amiss with the 50:50 beamsplitters? Is there some simple geometric explanation for all this? Like the photon only goes through the beamsplitter when it's sideways on and therefore only goes through one of the slits rather than both.

 

[selfAdjoint]

What exactly is the author afraid of?

Mojave Joe, what he is afraid of is the kind of non-starter speculations that MeJennifer's post #44 expresses. If you purely imagine the photon moving somehow classically inside the equipment (which QM gives you NO reason to do!), you get into paradoxes, and in attempting to resolve those paradoxes in your mind you get into speculations like "We are the hidden variables" that are not only unwarranted, but don't work in experiments. The Bell theorem and the experiments like Aspect's that have confirmed it have ruled out hidden variables, in us or any other way.

In order to comprehend these experiments, EPR, double slit, Stern-Gerlach, delayed choice, and whatever the brilliant experimenters come up with next, you have to abandon thinking classically ("paths") about quantum systems.

And no, "our knowledge" isn't necessary for these things to happen. An unthinking quantum interaction can, and in fact does, drive the experiment.

 

[Farsight]

Stern Gerlach experiment:

http://plato.stanford.edu/entries/physics-experiment/app5.html

 

[Farsight]

SelfAdjoint: I was looking up EPR and came across this website:

http://freespace.virgin.net/ch.thompson1/

It looks reasonable to me. In your expert opinion is it trustworthy?

If not, please delete this post.

 

[selfAdjoint]

It is not VERY trustworthy. The late Carolyn Thompson was a fierce critic of the Aspect type experiments. Some of her initial critiques had some basis, and corrections were made in the areas she and others noted during later experiments. But she refused to give up criticizing, as she supported a classical view of physics that the EPR experiments contradicted. She posted here at PF quite a bit, and toward the end I think it is fair to say she was regarded as pretty much a crank.

Nobody I know of with enough training in the area to have an informed opinion now doubts that the EPR experiments do correctly eliminate classical and hidden variable explanations of microphysical interactions.

Rather than delete your post, I am leaving it up with this reply, since Ms. Thompson's views are out there and a warning reference on them is useful.

 

[moving finger]

Me.

http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm

If you look at page 16 and count the dots on figs 3 4 and 5, you see there's twice as many joint detections for R₀₁ and R₀₂ as for R₀₃₍₀₄₎.

In total there's four times as many, but I doubled the number of detections on fig 5 because they occur on both R3 and R4.

Does this mean there's something amiss with the 50:50 beamsplitters? Is there some simple geometric explanation for all this? Like the photon only goes through the beamsplitter when it's sideways on and therefore only goes through one of the slits rather than both.

The "number of dots" simply represents the measurement precision on the x-axis. For the D1 and D2 idlers (showing interference) the data are displayed at 0.1mm resolution (one data point each 0.1mm along the x-axis), whereas for the D3 and D4 idlers (which show no interference) the data are displayed at 0.2mm resolution.

All this means is that the authors possibly "threw away" (or merged) half of their data points from the D3 D4 idlers (becasue they did not need such high resolution data to show the overall shape of the distribution).

I have an interesting question - what does one get if one "adds together" the data from D1 and D2? (ie merge the two curves in figures 3 and 4?)

Best Regards

 

[moving finger]

Finger, you keep plugging Cramer. I used to like Cramer too, but how does it work now that the "future absorber" it depends on, once identified with the supposed future big collapse of the universe, has apparently been done away with by the accelerated expansion?

Cramer has attempted to address that issue in his 1983 paper here :

http://mist.npl.washington.edu/npl/int_rep/dtime/dtime.html

He postulates a boundary condition at T=0 (the Big Bang) which would in effect "reflect" advanced waves in time. Sounds weird I know.

One of the other interesting conclusions from his 1983 paper is that for a "Big Crunch" scenario :

such models are intrinsically time symmetric and cannot explain the dominance of retarded over advanced radiation (in the absence of additional special postulates).

In other words, an ever-expanding universe would be one way to explain the fundamental time asymmetry.

It's interesting to note that this paper was published when there was still some support for a Big Crunch scenario.

I accept that there are some cosmological issues to be addressed in the Transactional Interpretation. But right now I don't see many other candidates for explaining the ontology behind the weird results in delayed choice / quantum eraser experiments. Of course one can simply ignore the ontology and just accept the maths (the "shut up and calculate" approach), but for someone like me that's not a very satisfying philosophy

Best Regards

 

[selfAdjoint]

Hi mf, Yes I see the asymmetry argument but I still don't see the how Cramerite satz works in the asymmetric case. Originally the retarded quantum waves went into the future, and the advanced waves went into the past, and both were reflected by the supposed abosrbers, handwavingly (as far as I ever saw) identified with the big bang and the big crunch. Then the relected waves interfered, and here he did calculations in the simple cases at least to show that the famous quantum behavior resulted. But if there is no reflection for the retarded waves how does this work?

As far as "explaining reality" goes, I am of two minds currently. One is that quantum mechanics was never built for that purpose, it's complete in itself but it is a machine that gives accounts of behavior, not explanations, and to attempt to use it for that incorrect purpose results in science-fiction - Cramer's time travel or Everett's multiple universes.

My other mind leads me to look into the combination of decoherence and relational quantum mechanics, to see how much can be retrieved there. Decoherence can explain how the classical world automatically results from the quantum world, but it doesn't address the measurement problem. RQM is one way of addressing the measurement problem, but it seems to me very much a work in progress at this point. I don't really like Consistent Histories because it seems to be just SUAC in disguise.

 

[Farsight]

Thanks for that rather sad feedback, selfAdjoint. Could you recommend any papers/links on Aspect or EPR explaining why classical and hidden variables must be eliminated?

moving finger: I tried combining D1 and D2 via manual dotting and got myself a mess - there's a displacement that I missed. Maybe I'll try printing two copies and sticking through the top one with a pin. I have to say I don't like the sound of throwing away data points.

 

[selfAdjoint]

Thanks for that rather sad feedback, selfAdjoint. Could you recommend any papers/links on Aspect or EPR explaining why classical and hidden variables must be eliminated?

I don't think you can go wrong with https://www.amazon.com/gp/product/0521523389/?tag=pfamazon01-20 by Bell and Aspect themselves. It's a collection of papers at different levels of technical difficulty, including some that require no math at all (though they all require you to keep your wits carefully about you as if your were solving a difficult puzzle!).

Currently my favorite hard-nosed popular account of the issues with quantum mechanics is Where Does the Weirdness Go? by David Lindley, which I excerpted a biut earlier in this thread.

 

[physicsmasta]

I too read the book in question. But the answer to that question has evaded me also. I would love to if you find out anything surrounding the book.

 

[physicsmasta]

I just read, "Beyond the Quantum" and in it it suggest that light particles have mass and all we have to do measure speed, position and force simutamelously is to measure the mass of light in the presence of gravity on earth. Do you believe this would work please reply!

 

[Reagle]

I have read through this thread and I would respectfully like to verify a few conclusions I have reached regarding the whole

"delayed choice quantum eraser"

phenomenon.

My questions relate to Yoon-Ho Kim, R. Yu, S.P. Kulik, and Y.H. Shih fairly famous paper calledhttp://xxx.lanl.gov/PS_cache/quant-ph/pdf/9903/9903047.pdf"

As noted in this thread this paper is the subject of comments made by Ross Rhoades in his http://www.bottomlayer.com/bottom/ki...scully-web.htm" at Wikipeadia.

Here are my conclusions about this paper:

1.

It makes no difference what is at D3.

If the idler photon was sent to D3, and D3 had been a coffee cup (instead of a detector), then the interference pattern would be still destroyed.

2. If conclusion 1 is true then the interference pattern at DO is dependent on the existence of which path information, even if which path information is impossible to recover (such as a photon hitting a coffee cup at D3).

3. If conclusion 1 and 2 are true then a lot of people like to make overreaching assumptions about this paper, leading to unsupportable conclusions. Especially people like the producers of the "[MEDIA=youtube[/URL]. These people are interpreting the interference effect as being determined by "what an observer knows", when it really has nothing to do with an observer at all and only has to do with the existence, or non existence, of which path information.

Am I right or wrong regarding these 3 observations?

 

[selfAdjoint]

Here are my conclusions about this paper:

1.
Quote:
It makes no difference what is at D3.

If the idler photon was sent to D3, and D3 had been a coffee cup (instead of a detector), then the interference pattern would be still destroyed.

2. If conclusion 1 is true then the interference pattern at DO is dependent on the existence of which path information, even if which path information is impossible to recover (such as a photon hitting a coffee cup at D3).

3. If conclusion 1 and 2 are true then a lot of people like to make overreaching assumptions about this paper, leading to unsupportable conclusions. Especially people like the producers of the "What the bleep clip here", and even Ross Rhoades in this video. These people are interpreting the interference effect as being determined by "what an observer knows", when it really has nothing to do with an observer at all and only has to do with the existence, or non existence, of which path information.

Am I right or wrong regarding these 3 observations?

Sure looks right to me. "What the bleep.." is a notoriouus new age commercial that has been complained about by just about every physicst with a blog. Would that newspaper journalists would pick up on that.

And yes a whole lot of people have egg on their faces, not just mystigogical cranks.

In my mind (and here I am treading on sacred professional ground) is that the whole personalizing thing in QM (as in using "observation" instead of "interaction" to describe whatever it is that generates real number observable values) should be abandoned. It was introduced by Bohr who was a famous mutterer and "guru" to three generations of physicsists including some of the great ones, but it's time to see it for what it is, a crutch, and a broken one that is droppiing us on our keisters at that.

 

[Hurkyl]

Maybe you've figured this out already, but this is something that wasn't clear to me when I was trying to figure out the delayed choice eraser experiment.

If you're just looking at D0, there is no interference pattern. The "interference pattern" only surfaces when split all the photons striking D0 into two groups, each group corresponding to having its pair strike D3 or D4.

 

[moving finger]

Thanks for that rather sad feedback, selfAdjoint. Could you recommend any papers/links on Aspect or EPR explaining why classical and hidden variables must be eliminated?

moving finger: I tried combining D1 and D2 via manual dotting and got myself a mess - there's a displacement that I missed. Maybe I'll try printing two copies and sticking through the top one with a pin. I have to say I don't like the sound of throwing away data points.

isn't that what destructive interference is all about? take two waves 180 degrees out of phase, and you end up with no wave.

the whole experiment is about throwing away (or deleting) information - delete the "which way" information (by randomising with a beamsplitter) and you see the interference - but retain the information and you don't see the interference. It just seems to me that the phase shift between the D1 and D2 data is just enough so that when combined these two signals destructively interfere. Is that significant, or irrelevant?

Best Regards

 

[Reagle]

Thank you selfAdjoint, for your feedback. I had an unresolved question in my mind regarding this paper. So let me respectfully ask one more question:

Q) The whole point of Yoon-Ho Kim, R. Yu, S.P. Kulik, and Y.H. Shih's fairly famous paper calledhttp://xxx.lanl.gov/PS_cache/quant-ph/pdf/9903/9903047.pdf"
is that the position of the signal photon impact on detector D0 is belatedly dependent on if the idler photon goes to D3 (which establishes which path information) or goes to either D0 or D1 (which erases which path information).
This is a pretty standard delayed quantum eraser setup with entangled pairs.

But in my simple mind this is totally impossible, since once the location of the D0 signal photon impact has been recorded, it can not change.

What am I missing here?