Talk:Delayed-choice quantum eraser/Archive 3

Modified diagram
I've modified the diagram according to criticism received on my talk page on Commons. The article mis-identifies "Glan-Thompson" as "Glen-Thompson," an easy mistake to make I guess. There was also an extra prism not pictured in the article's drawing (which is also missing several other features), so I had mis-labeled the prism that they did draw in. P0M (talk) 17:35, 7 November 2010 (UTC)

Removal of discussion section for alleged copyright violation
Somebody cut out this section, saying it was a copyright violation. I have restored it. The Discussion section has existed for nearly all the existence of this article. It has been brought into its present shape edit by edit. Anybody claiming a copyright violation must be able to show specifically what is wrong.P0M (talk) 03:33, 7 March 2011 (UTC)


 * Hi, that was me. No harm was intended. I removed the first two paragraphs of the "Discussion" section because they seem to have been substantially copied from the referenced paper with a little bit of rewording.


 * Here is the removed text, which was added in 2006 (emphasis mine):


 * In their paper, Kim, et al.[1] explain that the concept of complementarity is one of the most basic principles of quantum mechanics. According to the Heisenberg Uncertainty Principle, it is not possible to measure both precise position and momentum of a quantum particle at the same time. In other words, position and momentum are complementary. In 1927, Niels Bohr maintained that quantum particles have both "wave-like" behavior and "particle-like" behavior, but can exhibit one kind of behavior only under conditions that prevent exhibiting the complementary characteristics. This complementarity has come to be known as the wave-particle duality of quantum mechanics. Richard Feynman believed that the presence of these two aspects under conditions that prevent their simultaneous manifestation is the basic mystery of quantum mechanics.


 * The actual mechanisms that enforce complementarity vary from one experimental situation to another. In the double-slit experiment, the common wisdom is that the Heisenberg Uncertainty Principle makes it impossible to determine which slit the photon passes through without at the same time disturbing it enough to destroy the interference pattern. However, in 1982, Scully and Drühl found a way around the position-momentum uncertainty obstacle and proposed a quantum eraser to obtain which-path or particle-like information without introducing large uncontrolled phase factors to disturb the interference.[4]


 * And from the paper "A Delayed Choice Quantum Eraser published in 1999:


 * Complementarity, perhaps the most basic principle of quantum mechanics, distinguishes the world of quantum phenomena from the realm of classical physics. Quantum mechanically, one can never expect to measure both precise position and momentum of a quantum at the same time. It is prohibited. We say that the quantum observables “position” and “momentum” are “complementary” because the precise knowledge of the position (momentum) implies that all possible outcomes of measuring the momentum (position) are equally probable. In 1927, Niels Bohr illustrated complementarity with “wave-like” and “particle-like” attributes of a quantum mechanical object. Since then, complementarity is often superficially identiﬁed with “wave-particle duality of matter”. Over the years the two-slit interference experiment has been emphasized as a good example of the enforcement of complementarity. Feynman, discussing the two-slit experiment, noted that this wave-particle dual behavior contains the basic mystery of quantum mechanics [2]. The actual mechanisms that enforce complementarity vary from one experimental situation to another. In the two-slit experiment, the common “wisdom” is that the position-momentum uncertainty relation δxδp ≥ ¯h2 makes it impossible to determine which slit the photon (or electron) passes through without at the same time disturbing the photon (or electron) enough to destroy the interference pattern. However, it has been proven [3] that under certain circumstances this common interpretation may not be true. In 1982, Scully and Drühl found a way around this position-momentum uncertainty obstacle and proposed a quantum eraser to obtain which-path or particle-like information without scattering otherwise introducing large uncontrolled phase factors to disturb the interference. [...]


 * Clearly this is the original source of those two paragraphs; several sentences or phrases are copied verbatim, and others are only slightly paraphrased. I couldn't find any Wikipedia policy on what exactly is considered a "derivative work" but my guess is that this qualifies. 72.177.91.150 (talk) 22:06, 7 March 2011 (UTC)


 * Good catch, and here I was feeling great admiration for the level of acumen of whatever Wikipedian had written the second of those two copied sentences.


 * Let's get some other opinion on how far a recapitulation of somebody else's work has to depart from the original to avoid breaking their copyright. Sometimes if I run something entirely through my own mind and write without looking at the original I get challenged regarding whether the original paper or book really said/meant what I wrote that it did, so I can see some justification for somebody's sticking pretty close to the original in this case. As people who only try to write encyclopedia articles, many of us are not deeply enough into a subject to boldly recast a technical explanation into a new form for fear of getting it messed up. (I just had to go back and fix the discussion pertinent to the second diagram above because in reading the article I somehow skipped over the Glan-Thompson prism. I got the diagram wrong, and that caused me to get the discussion of how the apparatus works wrong. Somebody with greater depth would not have made that mistake.) And, to be fair to whoever produced these two paragraphs, most of the first paragraph and one sentence in the second paragraph are such widely accepted ideas that Scully and Drühl can hardly be said to have been entirely original in writing them. It's not quite as cut and dried as asking how many ways one can say, "Paris is the capital of France." On the other hand, Scully and Drühl have an excellent technique for technical writing (just basing my judgment on this one paraphrased and copied sample).


 * What we really need is for somebody with real depth in the subject to abstract the content of these two paragraphs, reorganize the bare bones of it, and then rewrite without looking at how the original authors put it.


 * For the present I've put the two quoted sentences in quotation marks as surely ought to have been done in the first place. That way my guess is that Wikipedia is at least legally sort of covered -- but I'm not a lawyer, so who knows. P0M (talk) 04:13, 8 March 2011 (UTC)


 * Sounds fine to me. I'm certainly not knowledgeable enough about this subject to be confident about rewriting the section from scratch. (I was just using the article to find the arXiv paper out of my own curiosity, and happened to notice that the introduction looked awfully familiar. This details are mostly way over my head!) 72.177.91.150 (talk) 06:21, 8 March 2011 (UTC)

Questions on slightly changed setup
Based on the diagram for the experiment in the main article:
 * 1) I understand the overall pattern at D0 is no interference.
 * If there were only that branch of the experimental apparatus, then there would be an interference pattern because a photon's wavefunction has gone through both slits, and the two wavefunctions overlap on the detection screen just as in a vanilla double-slit experiment. The question is, however, what will happen if the entangled twin of this photon has something done to it that will result in a physical change (absorption by one of the other detection screens), and in that process "which path" information is gained.
 * Let me clarify the paragraph above. Suppose that you took out the BBO, the crystal that is responsible for connverting single photons into entangled twin photons. You could leave the Glan-Thompson prism out too, since it is there basically to help separate pairs of entangled photons into more widely diverging paths, and now we don't have any entangled photons to worry about. So you would just have the laser (the blue part), the double slit (the black part), and then for D0 you might as well just use a standard detection screen. There is no longer any need for measuring very faint signals or for seeing what signals coordinate with what other signals. All those complications have gone away because you've reduced the complicated apparatus to a very basic Young experiment apparatus.
 * Next, let's rebuild a part of the more complicated apparatus. We put the BBO back in, so we get entangled photons, and we put the Glan-Thompson prism back in and also the PS prism so that we can more easily make certain which stream of entangled photons we are looking at. If we left all the rest of the apparatus in the bottom limb of the experimental apparatus back in storage, there would still be some question about what happens to the entangled photons. What happens if you put something in front of both exits from the PS prism that blots up the photons before there is any chance of their interfering with themselves? Since the photons in the lower limb (now amputated to a stump) could not interfere with themselves, it would seem that their twins in the upper limb could also not interfere with themselves. So in this case, it seems that virtually anything that would sop up the photons in the lower limb without their having the opportunity to interfere with themselves would make interference in the upper limb disappear.
 * The other thing we could do would be to put in a couple of mirrors that would direct the photons coming out of the PS prism out into interstellar space. Our telescopes pick up photons that have never been absorbed by anything from some point near the beginning of the universe until they hit our telescopes, so it is also possible that their status as self-interfering or not-self-interfering would remain indeterminate for a very long time. I suspect that most of them will eventually be absorbed without having interfered with themselves. The delayed choice experiments are said to exhibit retrocausality, which means, I think, that a spark of light showing up somewhere a gazillion miles from here and far in our future will be matched by what shows up here in our present. I am not sure that it would make anybody happier about this alleged state of affairs to assert that while the appearance of the two entangled photons may be separated by a long time interval according to someone in our frame of reference, the appearance of the photons will be simultaneous for someone in an appropriate frame of reference. Anyway, because most of the photons that travel a gazillion miles or years before being absorbed will not interfere with themselves, we can expect that any interference pattern that may appear in D0 will be very badly washed out by the photons that are matched by absorbed twins. On top of that, there will probably be a lot of photons reaching D0 that are not entangled, not desired to be there, and were the reason that the original version of the experiment required the Coincidence Counter to be sure that we look only at entangled photons and discard the rest.P0M (talk) 02:09, 24 July 2012 (UTC)


 * Note that the experimenters have inserted a lens into the path of the wavefunctions in the upper branch. The reason they did that was to shorten the path. Shortening the path will make a photon show up at D0 before its twin shows up at D1, D2, D3, or D4. One would think that what happens first cannot be determined by what happens afterwards, and one would think that what is farther away from the laser must happen later than what happens at D0, and therefore what happens at D0 ought to determine what happens "later" in the bottom branch.
 * The problem with that picture is that what happens in the top branch and what happens in the bottom branch is all one event, and a single event happens when it happens.
 * Notice the Coincidence Counter. If something strikes D0 at t and something strikes, e.g., D2 at T = t + n, then the Coincidence Counter ought not to click -- because the events do not coincide in time.
 * To me, this part is extremely spooky. P0M (talk) 02:48, 15 April 2011 (UTC)


 * Hi, the part in your response about seeing an interference pattern in the the top half of the experiment if the bottom half does not exist has me confused. I would have thought that once the particles have passed the BBO and become entangled, there would always just be a blob detectable. Wouldn't the only way to get an interference pattern (or even just the slit pattern) out of the entangled particles be to measure the idler and correlate that information with the blob information to get the subset that forms a pattern? Do you have a reference to an experiment where the experimenter was able to directly observe either an interference pattern or a clear slit-pattern on entangled particles (i.e. without correlating the idler data with blob data to get a subset that makes one or the other pattern)? --89.253.76.71 (talk) 02:35, 21 July 2012 (UTC)
 * http://www.scribd.com/doc/58640003/Walborn may be what you want. However, I don't see any indication that they have actually looked at whether an interference pattern is delivered to Dp, their upper-limb detector. P0M (talk) 15:10, 24 July 2012 (UTC)


 * The problem may have been that I was not very precise about what not having the bottom half exist would entail. The only way to really get rid of the bottom half is to get rid of the BBO. Once you have entangled photons and their twins going in different direction, you have two halves, top and bottom. If you get rid of the BBO, you will have nothing to prevent an interference pattern from forming in the upper limb. If you trap all the photons in some carbon nano-tangle trap so that they must get "detected" without having interfered with themselves, then you will destroy the interference pattern topside. Fixing things in the bottom limb of the experiment so that the photon twins must interfere with themselves would mean that the bottom limb becomes a functional duplicate of the top limb, and so interference will show up in both limbs.
 * If you put back the entirety of the original lower limb, then you get a problem that the experimental design introduces by the way phase changes work out. On one detector you would get a pattern of amplitudes something like:
 * 10203030403030202 vs.
 * 20203030403030201
 * But you could flip the phase relationships by redesigning the experimental apparatus and get something like
 * 10203030403030202 and
 * 20203030403030201
 * and the interference pattern should be clear. (The interference patterns are not perfectly symmetrical left for right, but that is just the way things are.) The interference effects are there. There is no cheat or fudge factor involved. It's just a question of what strategy to use to get the results fixed so that they do not obscure each other.P0M (talk) 02:09, 24 July 2012 (UTC)


 * 2) What if I extended the path to D0, say to 1 light minute, and let the photons hit D1 to D4 first. I believe pattern at D0 is unchanged.
 * That's what they do experiments for. My guess is that you are right. If D0 goes into operation after D1, D2, D3, or D4 go into operation, and if what happens to D0 is determined by the event at the other detector, at least the time sequence is "cause before effect." So maybe there is less cognitive dissonance in this experimental apparatus.P0M (talk)
 * 3) What if I then change the setup so photons reaching at D3 and D4 to be like D1 and D2. That is no which path info. I believe pattern at D0 is unchanged. But I don't know why; in this experimental setup there doesn't seem to be any leak of which-path-info.
 * If there is which path information revealed in the bottom branch, then I believe the experimental results are that the photon that hits D0 will always have just the characteristic of a diffraction pattern, not the characteristic of an interference pattern. (They would need to collect and make an image for all the "hits" to D0 matched to hits at D1, an image for all the hits to D0 matched to hits at D2, and so on to sort these results out.) But if they do something to D3 and D4 as you suggest, and no which path information were available, then each of the four image "slices" should, over many runs, form an interference pattern.P0M (talk)
 * Thanks P0M, for taking the time to answer my questions.
 * In the scenario I give (ie: get D3/D4 to be like D1/D2), is there "which path information revealed in the bottom branch"? I think you are saying there is not, which means you are saying there will be an overall interference pattern at D0?
 * (I understand you are saying using info at D1 to D4, I can see 4 interference pattern within D0. But that is not I am interested in. I want to know the overall pattern at D0; don't care what additional info on D1 to D4 I can use.)110.175.53.141 (talk) 02:30, 16 April 2011 (UTC)
 * Because D3 and D4 were originally places where a photon could only reach by one route (imagine going from a two-lane interstate to a one-lane interstate) then for anything that showed up at the end of that route had to have come by a single path. If somebody makes it so now the same photon can reach it by two different paths (so it is getting whatever goes through each side of the double-slit apparatus), there is no longer any which-path information. So all photons and entangled twins that go through the apparatus will interfere with themselves. They will therefore do what would ordinarily form a single, very clear, interference pattern at D0. However, you asked about "overall pattern," and the article shows how by messing around with mirrors and beam splitters in the bottom branch the experimenters have created a situation in which one interference pattern looks something like:

.|.|.| vs.
 * So its a little like throwing two movies from two projectors on the same screen at the same time. The bright spots from one movie will fill in the dark spots from the other movie. When you next make your contribution to the situation by letting D3 and D4 photons contribute to interference patterns, I think you can only make the "overall pattern" more like a continuous band. Well, actually, with just D1 and D2 contributing you already were seeing an "overall pattern" that was basically a band of light, brighter in the middle and dimmer at the ends. So there is even less variation than you would get by aiming four movie projectors at the same screen. The movies are still there. But each tends to wash out the other.
 * It's really spooky, to me, to see that what happens in the lower branch is mimiced, somehow, in the upper branch. It even goes to displacing the interference patterns associated with the entangled twins that end up at different detectors in the lower branch. Why something that happens in the lower branch should affect what happens in the upper branch is not clear at first. However, it looks like it boils down to the fact that in the lower branch the experiment forces the photon to "make decisions" as it were. In other words, in the original version of the experiment not enough was done to determine where and how D1 and D2 photons showed up there to make them unable to interfere with themselves. So their entangled twins also showed up interfering with themselves. Nevertheless, taking either the paths to D1 or the paths to D2 did do things to the phase relationships involved. D1 photons and D2 photons were originally in phase, but going through the lower branch of the apparatus put them out of phase. So their entangled twins were also put out of phase. So looking at what we get at D0 can tell us whether the entangled twin went through a D1 path or a D2 path. That's something interesting, but it still does not tell us which side of the double-slit apparatus the photon went through.P0M (talk) 10:00, 16 April 2011 (UTC)
 * It's really spooky, to me, to see that what happens in the lower branch is mimiced, somehow, in the upper branch. It even goes to displacing the interference patterns associated with the entangled twins that end up at different detectors in the lower branch. Why something that happens in the lower branch should affect what happens in the upper branch is not clear at first. However, it looks like it boils down to the fact that in the lower branch the experiment forces the photon to "make decisions" as it were. In other words, in the original version of the experiment not enough was done to determine where and how D1 and D2 photons showed up there to make them unable to interfere with themselves. So their entangled twins also showed up interfering with themselves. Nevertheless, taking either the paths to D1 or the paths to D2 did do things to the phase relationships involved. D1 photons and D2 photons were originally in phase, but going through the lower branch of the apparatus put them out of phase. So their entangled twins were also put out of phase. So looking at what we get at D0 can tell us whether the entangled twin went through a D1 path or a D2 path. That's something interesting, but it still does not tell us which side of the double-slit apparatus the photon went through.P0M (talk) 10:00, 16 April 2011 (UTC)


 * 4a) What If I then drop a polarizer in front of D0, does that bring back the interference? I think not, but worth a try...
 * If there is no which-path information available for the bottom branch, then an interference pattern (or, actually, four interference patterns that may be out of phase with each other. I'm just guessing) will form. Putting a polarizer in the upper path will only have an effect (not passing a photon with the wrong polarity) if a significant number of photons in the lower branch get polarized. If my memory is correct, what passes through a beam splitter has one polarity, and what reflects off the first surface of a beam splitter has the opposite polarity. So how a photon travels through the lower branch may affect its polarity. Then some of the wavefunctions that travel in the upper branch might be expected to have one polarity, and some might have the opposite polarity. So if that guess is correct, then some wavefunctions should be unable to pass through the polarizer. The result would be that the interference pattern observed at D0 would be dimmer. But then the question would be whether any photon that went through the polarizer to D0 could be associated with a path in the lower branch because that branch had the correct polarity to let the wavefunctions pass through the polarizer in the upper apparatus. P0M (talk)
 * 4b) Is there anything I can drop just before D0 to bring back the interference pattern?
 * It's already been established that removing which-path information from the lower branch would mean that the upper branch would always show results consistent with an interference pattern. If there were any way that a single polarizer could have an effect in the upper branch, I think it would be that some incoming photons would be vertical in polarity while others would be horizontal, and that a polarizer would block one or the other. There might be some way that a circular polarizer could be used before the rectangular polarizer so as to undo in advance the blockage of wavefunctions by the rectangular polarizer. I'm too sleep at this point to try to see how that might be done.


 * Really, the best thing is to understand the experiment in the diagram correctly. What it all comes down to is that when a photon is emitted by a laser and goes through a double-slit device, then half of its wavefunction goes by one path and half of it goes by another. If anything is done to prevent those two halves from landing on the same detection screen in proper registration, then interference cannot occur. If, however, something is done after the two halves have been created by the double slits and one of those two halves is deliberately "lost,"then one may endeavor to do something else to get the two parts back together again. This experiment uses these relations to explore the idea of time and causation. The rest of what we can learn from this experiment is how correct are our ideas about what to do to a photon (or a wavefunction) to create or to block interference.

P0M (talk) 02:48, 15 April 2011 (UTC)

WhoperJ12 (talk) 05:45, 13 April 2011 (UTC) some sort of answer
 * I hope it is at least approximately correct. ;-) P0M (talk) 10:00, 16 April 2011 (UTC)

FTL possibility in no-communication theorem
http://en.wikipedia.org/wiki/No-communication_theorem#Opposing_viewpoint


 * For example B. Dopfer, a graduate student of Anton Zeilinger, has indicated via experiment[1][2] that it is possible to cause or prohibit an ensemble of photons into making an interference pattern on a screen, by remotely manipulating their entangled twins.[clarification needed] Physicist John Cramer  is currently attempting to replicate Zeilinger's experiment for the purpose of communication. (The first experiment, attributed to A. Zeilinger, was actually done by Zeilinger's graduate student B. Dopfer).[3]


 * Of course Zeilinger and Dopfer's experiment does not prove superluminal communication, but neither does the no-communication prohibit all forms of communication. If superluminal communication is prohibited, it is not because of the no-communication theorem. Thus, the question of superluminal communication remains open.

A discrepancy here. --21:34, 28 April 2011 (UTC) —Preceding unsigned comment added by 146.96.35.52 (talk)
 * Is your point that a theoretical statement ("theorem") cannot trump a consistent set of empirical observations?P0M (talk) 01:49, 29 April 2011 (UTC)
 * No, my point is that this:
 * "The total pattern of signal photons at the primary detector never shows interference, so it is not possible to deduce what will happen to the idler photons by observing the signal photons alone, which would open up the possibility of gaining information faster-than-light (since one might deduce this information before there had been time for a message moving at the speed of light to travel from the idler detector to the signal photon detector) or even gaining information about the future (since as noted above, the signal photons may be detected at an earlier time than the idlers), both of which would qualify as violations of causality in physics."
 * suggests that there is NO way to make this a way of communication, while the earlier paragraph states that they're experimenting on the possibility. --68.160.195.48 (talk) 13:19, 29 December 2011 (UTC)
 * By some coincidence I just added a little to the article that should make the time sequence problem clearer. The facts about this limitation are, I am sure, perfectly clear to people like Dr. Cramer who are interested in the possibility of faster than light communication. I believe that despite these obvious difficulties there are also probably other people who consider the possibility that these limitations may be overcome. For instance, one of the factors that confuses the picture in the experiment that is the topic of this article is that a phase change in the lower (four detector) part of the experiment puts the two interference patterns out of phase with each other. It should be possible to introduce a further phase change in the lower part to put the two interference patterns in the same phase relationship. The center of the Detector 0 interference pattern would be overlaid with the diffraction pattern, but the fringes more distant from the center would be visible. The lower part of the apparatus might periodically be manipulated to prevent interference, and that expedient would create an on-off situation that could be the basis of communication. If I remember correctly, this kind of modified set-up was the basis of the experiment proposed by Dr. Cramer. Even if it is "theoretically impossible", that does not preclude people from trying to do something. I'll have another look at the article to see whether this conflict can be made to disappear or to become more easily apparent.P0M (talk) 15:57, 29 December 2011 (UTC)
 * It has not been easy to reformulate the section in question. I have made a start at it, but I need to see whether I will need to remove some old stuff to avoid repeating the new first paragraph of this section. I don't think that we have an article on Cramer's work on an "ansible." Perhaps there are other attempts going on that I do not know about.
 * Does the new paragraph make it clearer that the theoretical impossibility of something has not made everybody give up? There are still people trying to make eternal motion machines.P0M (talk) 02:20, 30 December 2011 (UTC)

Move subsection here
I moved the following part here because it seemed very out of place where it was in the article. It was as if the image-description text stopped and the following was inserted, and then after the following text, the image-description text started up again. I believe it requires an expert to replace it in a more appropriate place in the article. Also, the usage of bold emphasis is not within guidelines. (MOS:BOLD)

"Now it seems that, regardless of appearances, something may in all cases have traveled along both paths.

But what if the choice to "erase" the information is in fact delayed, until after the target phase?

Kim, et al.,  have shown that it is possible to delay the choice to "erase" the quantum information until after the photon has actually hit its target.

Under those conditions an interference pattern can be recovered, even if the information is erased after the photons have hit the detector. The experimental apparatus is considerably more elaborate than that shown and described above."

– Paine Ellsworth  (  C LIMAX   )  04:06, 25 October 2011 (UTC)

Reading through the rest of the article, the info above that I removed is detailed later on in the article. So it did not need to be where it was, as it disrupted the explanation of the easier-to-follow diagram in that section. If any editor wants to put this back where it was in the article, then it needs to be discussed here before that happens. – Paine Ellsworth  (  C LIMAX   )  05:17, 25 October 2011 (UTC)

Critical missing piece of information in summary.
The summary doesn't mention the dates the experiment was first run. It's possible to guess it was the 1980's sometime from the discussion section, but it would be expedient and helpful to include that piece of information unambiguously in the summary. I don't have time right now to chase the references to find out the correct info. 162.111.235.36 (talk) 15:38, 28 December 2011 (UTC)


 * The article describing the experiment was submitted in January 1999 and published in January of 2000. That's probably as close as I can get without buying the article to see whether they tell when they made the first run of the experiment. However, once they had their apparatus set up it shouldn't have taken very much time to run the experiment enough times to be sure they were getting reliable results. So I would guess that most of the work occurred during 1998.P0M (talk) 16:09, 28 December 2011 (UTC)
 * That's grand. Thank you, but shows how much guessing from context in the article does.  I think if I ever have the time, I'll fix that.  162.111.235.34 (talk) 16:49, 28 December 2011 (UTC)
 * I made mention of the date the article was submitted.P0M (talk) 23:07, 28 December 2011 (UTC)

Forward-propogating information too?
One statement "It is impossible to know which group a photon appearing at Detector 0 at time T1 may belong to until after its entangled partner is found at one of the other detectors and their coincidence is measured at some slightly later time T2." seems overly strong to me. The position of the photon at D0 landing on the camera at D0 is known, and because there is correlation of the D0 position with the D1/D2/D3/D4 detection, then by observing D0 we may already have some information about which of D1/D2/D3/D4 will be detected.

For instance, if I observe the photon at D0 landing in a node (zero) of the interference pattern associated with D1, then I know with certainty that D1 will not be detected, and there will be 50%/25%/25% chance of D2/D3/D4. Likewise for D2's interference pattern. If instead the photon lands halfway between node and antinode, then I can assign 25%/25%/25%/25% chances for D1/D2/D3/D4. --24.85.247.169 (talk) 19:06, 21 January 2012 (UTC)


 * The reason it is impossible to do what you propose is that the nodes are not in the same positions in all instances. One fringe pattern looks like the left-to-right mirror image of the other:
 * |_|_|_|_|_|
 * _|_|_|_|_|_
 * So the result is a more-or-less solid band. The swap is due to phase changes in the bottom part of the experimental apparatus.P0M (talk) 22:59, 21 January 2012 (UTC)


 * Yes, those offset stripes are part of the reason I would expect that partial information is obtained. To be specific, I'm going to refer to the figures as the end of this paper. I agree that if you add up the joint detection rates R01 R02 R03 R04, or even just R01+R02, a smooth bump is obtained, but that's beside the point.
 * Suppose that I am sitting at D0 and detect a photon at 1.2 mm position. The joint detection rate R02 is about three times larger than R01 for this position -- thus, I know that if a photon in the lower apparatus is observed, it is three times more likely to be in detector 1 than detector 2. In particular I calculate P(D1) = R01/(R01+R02+R03+R04) = 50/405 = 12%, P(D2) = ... = 31%, and P(D3) = P(D4) = 28%.
 * We could also examine at 1.33mm, halfway between node and antinode: 22%/20%/29%/29%. (I'm not sure why D4 and D3 are more probable than D1/D2. Losses in the path to D1/D2 I guess.) --24.85.247.169 (talk) 02:14, 22 January 2012 (UTC)


 * D3 and D4 receive photons by one path only, so they would be the ones that have the spread that is associated with diffraction. So almost all of the photons hit near the center. There is some spread, but far less than when an interference pattern is involved, as with D1 and D2. Or, to be more precise, the probability that a photon that has a known single path will appear anywhere very far from the center is vanishingly remote. (George Gamow wrote that it is possible that all the air molecules in a room might by random "choice" all head north at the same instant, but that one need not hold one's breath in preparation for this event.) It is not clear that the experiment is set up in such a way that it catches photons that diverge far from the center of the detection screens.


 * D1 and D2 receive photons by two paths each, and the difference between them lies in the phase relations among the photons they will receive, which in turn means that we get the two sort of mirror image interference patterns. If you will do what I just did and superimpose the image for D3 and for D4 (even though the article only gives one, they are both present and the twins of the photons which reach each of them will all reach D0, and also D1 and D2, then you will be in a position to add the amplitudes for all four beams. The peak will be about three times higher than the highest of the interference pattern peaks, and the sides will be fairly smooth.


 * It is more likely that a photon observed at any point represented on the graphs of the experimenters will be associated with D3 and D4. There are two of them, and they overlap in uninterrupted coverage of the central region, which is all the experimenters appear to have been measuring. If you look at a position on D0 that corresponds to the highest peak of D1, I think it is right to say that it is more likely that any photon that is found there corresponds to one that showed up at D3 or D4, or with less probability it will be associated with D1, and with even less probability that it will be associated with one that shows up at D2. But knowing the probability that something is the case is not the same as knowing what actually happened.
 * Let's say that you picked one of these photons and made a bet as to which detector its twin was associated with. You could not know whether you had won or lost until you checked with the coincidence indicator.
 * If you were bound and determined to bet on photons that were parts of an interference pattern, then you would do better if you could find an experiment that took into account more than the first three or four fringes. Interference patterns resulting from Young experiments are extremely broad. (I don't know whether anybody has tried to experimentally verify the greatest angular spread of a detectible pattern. Just with the naked eye, the pattern produced by interferences is dozens of times wider than the pattern produced by a single slit. So if you had a typical setup with the laser 3 meters from a detection screen, anything that was more than 25 cm. from the center would almost certainly be part of an interference pattern. In other words, if you set up a single slit and looked for photons at any distance from the central spot of light, you would probably need very sensitive apparatus to detect anything, and even then you would have to wonder whether there was a light leak in your darkroom.


 * I think you are right about there being varying probabilities, across the width of a detection screen, for which detection screen a photon may be associated with. If you bet on either D1 or D2 you already have a fifty percent probability of being wrong because the photon was associated with D3 or D4 plus some additional probability of being wrong because the photon was associated with the less likely of D1 or D2.


 * I think the reason that some people are interested in the "retrocausal" aspect of these experiments is because they hope that the phenomenon can be used for communications over such great distances that the speed of light is not good enough for us. (If you send a radio message to earth from a colony 100 light years away, you will not receive an answer in your lifetime.) If somebody could send a continuous electromagnetic signal to earth and keep the quantum twins running in circles at home, then after the signals reached earth the quantum twins at home could be modified so that they conveyed information to the people on earth. John Cramer had a plan he intended at one time to test, but he has not published anything further about it. Anyway, just knowing that a single photon was more likely to have been associated with D1 than with D2 would not be very useful in this context. P0M (talk) 08:25, 22 January 2012 (UTC)


 * Thanks for pointing out the additional complication of single-slit diffraction pattern. I had completely forgotten about that. I wonder why they didn't bother plotting the side-lobes in the arxiv paper?
 * Yes, I definitely agree that D3 and D4 tend to dominate and make things uncertain. The information from D0 is poor and indefinite at best. I know I wouldn't make any bets on it unless the house margins were very slim :). Now, here's a fun thing to think about: Let's say the photon is detected at D0, and the photon on the lower apparatus is stuck in a super long delay line (okay, two matched delay lines, one for each path). After D0 detection, I randomly decide to remove the beam splitter in front of the D1/D2 detectors, so that I am guaranteed to measure which-path information (D1 acts like D4, and D2 acts like D3). Or, instead I could remove the beam-splitters in front of D3/D4, so that I can never measure which-path information. Would this change my D0 result? I think that the overall pattern of D0 intensity should be the same, no matter what devious manipulations I perform with the lower photon. If not, I would say it's a very good case for retrocausality (and ability to send messages back in time / faster than light).
 * By the way, Hyperphysics has a nice graphic showing the difference between single and double slit. The double-slit intensity there can rise up to at most 4 times higher than the single-slit intensity curve, due to constructive interference, but on average it will only be 2 times brighter. I know what you mean about interference patterns looking a lot wider than single-slit patterns (I've noticed the same things in a few texts), but as far as I know that's mathematically not supposed to happen... it might just be an artifact of the increased brightness. --24.85.247.169 (talk) 20:50, 22 January 2012 (UTC)


 * Even in the classical account of this phenomenon (see Sears, Optics, p. 214ff) the math is expressed as a continuing series of values, i.e., the equations involve a series of integer multiples and there is no reason to stop calculating except that generally speaking the intensities of the peaks get smaller as you calculate fringe positions farther from the center. Eventually you will get to the point where your eyes are too weak to see anything. Some charge-coupled devices can be fired by only one photon hitting them, so the width of the detectable fringe pattern ought to be much more than the easily visible width of more than 4 feet observable at 20 feet or so from the double slit apparatus.
 * Single_%26_double_slit_experiment.jpg
 * I made the experiment and the photograph I made is still on Commons. In the article on the Double-slit experiment it has been replaced by a better image made with nicer equipment. The number of fringes in both images are approximately the same. The image I photographed was only the central 8 inches or so. The remaining 40 inches or so was weaker, and I saw no point in trying to photograph it. (I had to get close enough to the image to make my digital camera fire its shutter. Any farther away and the camera would refuse to fire because the image was "too dark" for its settings. With my present camera I could have put it on open shutter so I would have gotten a much broader image that would have been impossible to see without printing it horizontally on 8 x 10 paper. The photo by Sears shows about 36 fringes (opposite p. 222). I think I have seen more than a couple of books that assert directly that Psi functions are not bounded. The central portion, i.e., the portion where experience tells us to look for a photon to show up, is the central portion because that is where the probability of a photon showing up is highest. But there are lesser probabilities away from the center, and while the probabilities may drop down so low that experimenters might have to wait for a lifetime to actually see a photon show up there, one eventually will. (Somebody wins the Sweepstakes.) Richard Feynman says at one point something to the effect that if you fire off a laser, any given photon has some probability of showing up anywhere in the universe. Anyway, it should be clear just from the photo above that the visible fringes are not restricted to the width of the visible diffraction pattern.


 * You would enjoy the article on retrocausal communications published by John Cramer in Analog a few years ago. I think you are describing a version of his experiment. You can google it up. P0M (talk) 08:44, 23 January 2012 (UTC)


 * Heh, alright, you got me. I'm not entirely convinced but I myself don't have any experimental pictures to back up what I want to say, nor have I gone through the math myself. I would make one suggestion though, related to this single vs. double-slit pattern: In the third and fourth figures of this wikipedia article, the coincidence results for D3 and D4 have diffraction bands on them (good), but the spacing of the diffraction bands is the same as the interference bands on D1 and D2 (confusing). You might want to make them look more like the Kim data (plus side-lobes if you prefer). Anyhow, it's been a pleasure talking to you. Cheers. --24.85.247.169 (talk) 17:40, 24 January 2012 (UTC)

why not one photon
Why not try this experiment with just one photon instead of stream of them? Then you wouldn't need the coincidence counter would you? That should then enable looking at the D0 detector before the other ones and see if the information is there. — Preceding unsigned comment added by 64.22.160.1 (talk) 17:03, 3 February 2012 (UTC) One photon would not provide the information needed to detect a pattern. However shooting a series of photons at intervals that allowed recording of each before the next is sent would allow us to view the patterns as they are assembled. The results would not change but it might provide some insight and would be an interesting way to share the data with the community. — Preceding unsigned comment added by 76.178.252.56 (talk) 01:30, 30 June 2012 (UTC)
 * I'm just working from memory here, but I think the reason is that in equipment currently being used is "noisy" in the sense that while it is designed to produce entangled photons it also produces lots of photons that do not get entangled. It would be difficult, probably impossible, to filter out the unentangled photons somehow, and easier to just deal with everything that is produced and look for the photons that show up together.
 * It should be possible, however, to do what the very early experimenters did to pick out individual photons going through a double-slit apparatus, just cut down the delivery rate so that, generally speaking, only one photo would be delivered within a certain time period. As long as the separation was clear, then a CCD or whatever detection device being used could detect the positions and times of all photons received, wash out the photon records that were not matched, and then you would have a record of individual photon pairs produced under virtually identical conditions (same apparatus, same power supply, etc.).
 * To me it would seem one way to deal with the claim of some that quantum mechanics only deals with ensembles of measurements, as though the interference effects would disappear if they did not occur amidst hundreds or more other measurements that would show interference.P0M (talk) 15:59, 1 July 2012 (UTC)

No citations in section: Problems with using retrocausality
The discussion in the subsection "Problems with using retrocausality" is very clear, however, it has no citations whatsoever. I would like to dig deeper into understanding that part, but couldn't find the source of the main explanation. 89.253.76.71 (talk) 12:46, 13 August 2012 (UTC)


 * I'll see whether Dr. John Cramer still has his speculations (published in Analog Science Fiction and Fact a number of years ago, and on his website) in an available form. That may lead to citations of people who have published the critiques of this idea. P0M (talk) 18:28, 14 August 2012 (UTC)


 * There is a little information here: http://nextbigfuture.com/2009/09/john-cramers-retrocausal-experiments.html

P0M (talk) 18:32, 14 August 2012 (UTC)

File:Beam_Split_and_fuse.svg
This diagram incorrectly labels which beams receive a k phase shift and which ones receive a lambda/2 shift.

The beam that passes through the glass should receive the k phase shift and the beam reflected from the mirror should receive the lambda/2 shift.

In addition, when the top beam reflects from the back of the mirror of the second beam splitter (instead of passing through it) it should pick up a phase shift of 2k.

The final result is labeled correctly.

Regards. KC. — Preceding unsigned comment added by Kcerb (talk • contribs) 01:30, 30 December 2012 (UTC)


 * I'll try to fix the diagram. Thanks. P0M (talk) 21:29, 13 January 2014 (UTC)

Horrible infix subclauses
The article contains a lot of really obfuscated hard-to-parse sentences. For example: The largest problem here is the large infix subclause providing a long range dependency. The distance from the subject "a very narrow beam of coherent light from a source" to its verb "is directed" really taxes the short-term memory, and long range dependencies are established as troublesome to learn and understand. Subclauses are quite OK and makes the language flow nicely, for example the postfixed "pierced by two parallel slit apertures", but infix subclauses should preferrably be short. Just as a general advice: the alternative to large infix subclauses, is to express their information in a separate sentence before or after the sentence. I'm not sure what's the best solution in this case. Rursus dixit. ( m bork3 !) 07:56, 1 October 2013 (UTC)
 * In the basic double slit experiment, a very narrow beam of coherent light from a source, is directed perpendicularly towards a wall pierced by two parallel slit apertures.

A fresh start
The article has acquired multiple issues due to POV differences between DParlevliet, Patrick0Moran, 129.217.159.124 and myself. We have enough of an understanding of ourselves and our differences, however, that I think we can go forward productively from the last generally accepted revision of 07:41, 1 October 2013, to which version I have reverted the article.

Let's be friends here. Stigmatella aurantiaca (talk) 19:10, 5 February 2014 (UTC)
 * The revert includes a lot of edits which are not disputed, so it is not allowed to remove. Only one part is an issue. We don't need to agree with everything in Wikipedia, it is no religion here. There are also items I don't agree, but don't change. That "All valid parts of classical physics are subsumed by quantum mechanics" does not mean that the classical explantion is forbidden. As long as the double slit article has a classical chapter I see no reason why it is not allowed here. The only question is, if the written part has errors. If QM scientist abhor the classical and find it useless, then don't read it. The QM explanation is there too, and on top DParlevliet (talk) 19:49, 5 February 2014 (UTC)

The idea is that we start with the Oct 1, 2013 version that the majority of us here have no issue with, and move forward WITH CONSENSUS.
 * Consensus first, THEN change
 * For example, I propose to delete the section titled Yet there are those who persevere in attempting to communicate retroactively. The references to this highly speculative section consist of (1) an article in a science fiction magazine and (2) an unrefereed arxiv article. Stigmatella aurantiaca (talk) 23:24, 5 February 2014 (UTC)
 * Yes, I am quite aware that Paul J. Werbos has a refereed article on this subject, but even in that article, he admits that there are highly speculative elements to the theory. In my opinion, covering this topic in this article touches on WP:UNDUE &mdash; Stigmatella aurantiaca (talk) 23:36, 5 February 2014 (UTC)
 * Werbos's refereed article on this subject could be used as a reference for the section "Problems with using retrocausality" provided that we also provide a second article for balance.
 * As Werbos and Dolmatova state, "True backwards time communication channels (BTCC) are absolutely impossible in most formulations of quantum theory but only almost impossible in the BTI formulation." Appropriate weight should focus on "most formulations of quantum theory". Stigmatella aurantiaca (talk) 01:59, 6 February 2014 (UTC)


 * DP wrote, "As long as the double slit article has a classical chapter I see no reason why it is not allowed here." This is not the double slit article. This is an article on quantum effects that ultimately do not have a classical explanation. Stigmatella aurantiaca (talk) 23:58, 5 February 2014 (UTC)


 * Until we can learn to stop bickering, the need for consensus applies to everybody. That applies to you, too, Patrick0Moran. Even though I agree with your removal of what sounds like an original research speculation in Problems with using retrocausality, and even though the removal was unrelated to the major POV differences that exist between DP, you, 129.217.159.124 and myself, it should have been at least announced here first. Stigmatella aurantiaca (talk) 04:20, 6 February 2014 (UTC)
 * You are right. Sorry about that. thought that the issue had been raised and no objection given to the idea of removing the speculative parts.P0M (talk) 05:20, 6 February 2014 (UTC)


 * Non-controversial addition of references &mdash; I will be making non-controversial additions of references to the discussion section with the aim of removing the "Unreferenced section" template. Stigmatella aurantiaca (talk) 05:04, 6 February 2014 (UTC)
 * Don't delete but edit I did not start the edit war. Deleting is not about majority but according arguments and Wiki delete rules and none Wiki rule is used for all deletions. There has been several edits since 1 October, with arguments. None has been disputed, not when edited nor now. So according Wikirules it is not allowed to delete all those. If there is any change in the past which afterwards is not right, then change that part, again with arguments. But part by part, not everything at once. There is only one part which is heavily criticised, so focus on that. Keep to wikirules, which demand that one should edit, not delete. So if anything needs editing, edit. If you think that the experiment cannot be explained by classical rules, then edit. DParlevliet (talk) 08:23, 6 February 2014 (UTC)
 * DParlevliet: Please don't edit war You have made three reverts in the last 24 hours against consensus, two reverts of my edits, one of Patrick0Moran's edits. Your entire history of contributions since 07:41, 1 October 2013‎ has been of addition of material expressing your POV against the consensus of (1) Patrick0Moran, (2) an anonymous IP who has actual experience in the field (who argued extensively with you on the talk page but refrained from actually editing the article himself), and (3) myself. Please do not revert again, and please do not think that you can hide behind the 24 hour rule by waiting a few extra hours before making your next revert. The 24 hour rule is merely a bright line rule. See WP:EDITWAR. Stigmatella aurantiaca (talk) 10:04, 6 February 2014 (UTC)
 * Whatever gave you the idea that the administrators would take your side when you reported me on the Edit Warring Noticeboard? Stigmatella aurantiaca (talk) 13:15, 6 February 2014 (UTC)

Fully protected
There looks to be a long running edit war on the article so I've fully protected it for a week so that everyone needs to discuss rather than revert each other. Plus continue to discuss the issues on this talk page rather than continuing to revert. Thanks, Callanecc (talk • contribs • logs) 12:05, 6 February 2014 (UTC)


 * Thanks! A week off will provide an opportunity for heads to cool down. So long as we are acting like this [[File:PissedOff.gif]], there won't be much chance for progress. Stigmatella aurantiaca (talk) 12:15, 6 February 2014 (UTC)
 * I will repeat my objection here: Stigmatella aurantiaca has reverted 2 months of editing without referring to a Wiki deletion rule and without discussing this revert on the talk page. The revert has been supported by Patrick0Moran and 129.217.159.124 (which declared that he has no plans to edit himself). None of them has questioned the edits during the last 2 months. Also at the moments of revert none of them has given arguments what was wrong in all those edits to justice a complete revert in stead of editing. There is only one small part which has caused an extended discussion, but that does involve all other edits. Therefore reverting so many edits without referring to Wiki deletion rules is not acceptable.DParlevliet (talk) 12:56, 6 February 2014 (UTC)
 * Whatever gave you the idea that the administrators would take your side when you reported me on the Edit Warring Noticeboard? Stigmatella aurantiaca (talk) 13:15, 6 February 2014 (UTC)
 * Is it "Thanks!" or not? You warned me that you would use the administrators. That was a good idea to let a moderator check if Wiki rules are followed, so I proceeded with what you proposed.. As you mentioned, better to cool down first. DParlevliet (talk) 13:52, 6 February 2014 (UTC)
 * You risked getting yourself blocked for your behavior. That was not my intention. As you yourself had to admit in your report, my initial revert was supported by Patrick0Moran and the anonymous IP. Stigmatella aurantiaca (talk) 14:08, 6 February 2014 (UTC)

Response to DParlevliet
Your entire history of contributions since 07:41, 1 October 2013‎ has been of adding material expressing your POV against the consensus of (1) Patrick0Moran, (2) an anonymous IP who has actual experience in the field (who argued extensively with you on the talk page but refrained from actually editing the article himself), and (3) myself.

The anonymous IP (who has recently opened up an account as Cthugha82) summarized the situation in the following talk page diff. In this diff, he recommended that we revert to the 1 Oct 2013 revision of the article. Patrick0Moran and I were in agreement on this issue.


 * 16:57, Talk page: 5 February 2014

You claim that "None of them [i.e. us] has questioned the edits during the last 2 months." This is a completely false claim, as can be seen from even a cursory perusal of the talk page. (I'm a relatively recent addition to the debate so of course was not involved in most of these debates.) The pattern, repeated over and over, was that Patrick0Moran and 129.217.159.124 would argue with you, but you would completely ignore our recommendations and would proceed to edit the article the way you wanted.

Stigmatella aurantiaca (talk) 13:22, 6 February 2014 (UTC)
 * All discussions you mention concerns the small part regarding classical waves which I mentioned above. That is a technical discussion which I did keep outside the claim. I object to the revert of all other edits which are not related to this subject and were not disputed.DParlevliet (talk) 13:52, 6 February 2014 (UTC)


 * Selective reversion only of the disputed sections was not a possible option. The idea was to take the article to a non-controversial base point, and then add back in those contributions to which we could agree. The accepted edits would have included a number that you had made. Stigmatella aurantiaca (talk) 14:02, 6 February 2014 (UTC)


 * Yes that is possible, because the disputed subject is only one paragraph. DParlevliet (talk) 14:37, 6 February 2014 (UTC)


 * The dispute has covered more than the one paragraph. We have, on multiple occasions, both here and in the Double-slit experiment article, needed to decipher your English to figure out your intended meaning for edits outside of the primary dispute. Although I could, in principle, have performed a manual delete of the paragraph that is the primary topic of this dispute, I could not perform a targeted revert of this paragraph using the facility built into the Wikipedia history page because of these side-edits, which also need to be validated. Your side-edits, even when not disputed, have not generally improved the article. Stigmatella aurantiaca (talk) 17:55, 6 February 2014 (UTC)


 * Can you show in Deletion_policy the rule which allows deleting when improvement in language is needed? The last version before your revert includes all language edits. I don't think Wiki procedures force you to delete 2 months edits in stead of one paragraph. Regarding the paragraph I remind of the wiki advise "If editing can improve the page, this should be done rather than deleting the page". So don't delete, but edit and improve. Unless of course you can argument that the part written is wrong or violates wiki rules. DParlevliet (talk) 19:13, 6 February 2014 (UTC)

You are confused. Deletion_policy concerns itself with PAGE deletion policy. Page deletion policy is not relevant to my actions, which have been to revert to a version of the page that by the mutual consensus of Patrick0Moran, 129.217.159.124 and myself is an appropriate starting point for improving the article. We do not intend to exclude you from the improvement process, but we do insist that you abide by consensus rather than persistently ignoring our input and editing the article to reflect your POV. Stigmatella aurantiaca (talk) 20:21, 6 February 2014 (UTC)


 * Wikipedia has no police for late editors to go back three months and start all over again, but now with their permission. If it was no improvement, you should have acted earlier. Could you show were I multiple ignored your input, outside the disputed paragraph? Wiki is not only about majorities, but also its rules.DParlevliet (talk) 14:24, 7 February 2014 (UTC)

What is the consensus?
I don't want to get involved in the details of this discussion (which is very long-winded), but I would like to remind all of you that you'll get nowhere with recriminations. Before this latest spat, you appeared to be converging on some issues. Now would be a good time to identify clearly what the consensus is so far. Can you agree on some additions to the article? RockMagnetist (talk) 03:16, 7 February 2014 (UTC)
 * Patrick0Moran, Cthugha82 (129.217.159.124), and I would agree that all attempts to explain the delayed choice quantum eraser experiment with the aid of classical analogies are misguided. DParlevliet has argued to the contrary. We have had no promise from DParlevliet that he will work with us towards consensus. Stigmatella aurantiaca (talk) 08:17, 7 February 2014 (UTC)


 * Patrick0Moran, Cthugha82 (129.217.159.124), and I would agree that the 07:41, 1 October 2013 version of this article is an appropriate one to use as a basis for improvement. DParlevliet has argued to the contrary. We have had no indication from DParlevliet that he has changed his opinion. Stigmatella aurantiaca (talk) 08:22, 7 February 2014 (UTC)


 * From my point of view it seems most reasonable to restart from the 07:41, 1 October 2013 version and continue the "encyclopedic" way. Fortunately this is a topic about science, so the current state of the art is documented well in terms of peer-reviewed publications in respected journals. So my suggestion is to base further edits (if there is no immediate consensus) on whether the suggested edit is directly (without lots of interpretation) verifiable by comparing it to the literature in the field and citing the source. Cthugha82 (talk) 10:19, 7 February 2014 (UTC)


 * I agree that the 07:41, 1 October 2013 version is an acceptable starting point. The document for quite some time before then had undergone only minor changes aimed at clarity, and I recognize many of the names of contributors who made those changes, which indicates to me that quite a few knowledgeable people checked over it and found it basically sound. I think one important principle to keep in mind is that nature is the ultimate guide. Clearly we can report on what researchers report that they see in the laboratory. The equations of quantum mechanics are almost better than the observations in any one lab run because they have been verified so perfectly over nearly a century and it has been found that refinements to lab work bring observations closer to theory than the other way around. They can inform our work, but they would be unhelpful to the average well-informed reader. Interpretations of the math and of the lab results that try to say what photons really do and why things happen as they do can be problematical. Saying what nature ought to be doing on the basis of human models of nature is not a reliable procedure. I believe that in the present case the papers by Aspect. et al, Walborn et al, amd Kim et al should form the basis on which we rely, and that ancillary material ought to be added sparingly. (Perhaps there are times when people not in the field need to be warned away from certain false conclusions that those in the field would be unlikely to make.) P0M (talk) 12:08, 7 February 2014 (UTC)


 * Going back 3 months was a recent proposal from Cthugha82, I think. Until then is was not discussed nor mentioned to be needed. Both Stigmatella and Cthugha82 entered the dicussion recently, but long before POM did all editing, which in most cases I accepted, after discussing. It has improved the article (indeed also language) and I really value that, also because POM spends so much time improving Wiki. The discussion has been lengthen by a lot of misunderstandings between POM and me. I am not a QM scientist and POM wants exact scientific descriptions. Of course part is me to blame, for which my excuses. There is one exception: the paragraph about classical waves which I have added. POM wants it to be removed, I want it to be there because it is scientific right. When you search in google for images of "double slit", most of them will be classical, also from University sites. Also the Wiki "double slit" has a classical wave paragraph. So I don't understand why this is not allowed in the present article (also a double slit). If the experiment shows that classical cannot explain the experiment, then describe it. I did the same with the Kim experiment. I also thought that after clearing the last misunderstanding with POM, there was a mutual understanding. I don't mind improving the paragraph, adding or deleting if it is wrong. I hate edit wars, but two months editing was a lot of work for me and you can imagine that I don't accept that new editors delete all and start again. Therefore I asked an independent administrator if this is according Wiki rules. So I prefer to start with the version which POM and I discussed about, and then everyone can edit as intended by Wiki. But first I think the disputed paragraph has to be resolved. DParlevliet (talk) 15:11, 7 February 2014 (UTC)


 * 1) "I am not a QM scientist" &mdash; But Cthugha82 is a QM scientist working directly in this field, and he has repeatedly argued against your contributions, as has Patrick0Moran. I am a relative latecomer to these debates, but I agree entirely with Cthugha82 and Patrick0Moran.
 * 2) " I want it [a classical treatment] to be there because it is scientific right." &mdash; Patrick0Moran, Cthugha82 and I disagree with you. It is not "scientific right".
 * 3) "Also the Wiki "double slit" has a classical wave paragraph." &mdash; The delayed choice quantum eraser experiment is not the double slit experiment, and has no classical explanation.
 * 4) "If the experiment shows that classical cannot explain the experiment, then describe it." &mdash; The explanation is currently in the article, but the language of the explanation does need to be improved. Patrick0Moran and I are currently working on improving it, as can be seen elsewhere on this talk page.
 * 5) "I don't mind improving the paragraph, adding or deleting if it is wrong." &mdash; It is wrong.
 * 6) "I don't accept that new editors delete all and start again" &mdash;  Your previous contributions are not lost. They are in the page history, and it is not difficult to retrieve them for reinsertion, if they are found to be appropriate. We are not seeking to block you from editing. We do ask that you cooperate with us in the spirit of consensus.
 * 7) "Therefore I asked an independent administrator if this is according Wiki rules." &mdash; You were attempting to use the Edit Warring Noticeboard mechanism to get me blocked or banned.
 * 8) Do you promise in the future to cooperate with Patrick0Moran, Cthugha82 and myself in the spirit of consensus?
 * Stigmatella aurantiaca (talk) 16:41, 7 February 2014 (UTC)


 * When I look at your material on classical waves, the thing that leaps out is the lack of citations. So in answer to your question, there is indeed a Wiki rule that justifies removing it: verifiability. If you add material that is likely to be challenged (or actually is challenged), you must support it with reliable sources. The burden is on you. I know that it is frustrating to do a lot of work and then have it removed, but the rules are clear. If you want to avoid further frustration, read the Core content policies and abide by them. RockMagnetist (talk) 16:51, 7 February 2014 (UTC)


 * When you look to double slit you will see that the main part has no citations either. That is because is it very old and familiar "classic" with simple high-school formula. But I do have a citation to the explanation of the quarter wave plate when polization is parrallel to the axis, which I think POM agreed with. The statement that setup 2 is the sum of 3 and 4 was mentioned in Waldborns article, and I did refer to that. Perhaps there are more items which needs citations, but that's where editing is for. If items are new, without citations, I will remove it. In an earlier stage I have proposed to get a classical optical scientist, to have a second view of the part. I have tried myself, but did not find yet (although one option running). It seems they don't excist anymore.... Please take into account that the "lot of work" does not concern the wave part, but all other changes which were not disputed since 2 months ago. That has no relation to the waves. DParlevliet (talk) 20:11, 7 February 2014 (UTC)


 * Everybody has access to the article. Please point me to a statement to the effect that "setup 2 is the sum of 3 and 4." Just so everyone is aware, the end of the article has a kind of graphical summary with 9 experimental configurations. Their Fig. 1 is not covered by DParlevliet's list, so their setup 2 is Walborn's 3 and so forth.P0M (talk) 23:27, 7 February 2014 (UTC)


 * Page 5, paragraph V, "The average sum of these two interference patterns gives a pattern roughly equal to that of figure 3". DParlevliet (talk) 08:03, 8 February 2014 (UTC)


 * WP:OSE is not a valid justification for adding material that does not adhere to Wikipedia standards of verifiability. Stigmatella aurantiaca (talk) 00:51, 8 February 2014 (UTC)


 * If the material is "classic" and not likely to be challenged, WP:V does not require that it be supported by citations (although it's still a good practice). Your material does need citations. RockMagnetist (talk) 01:13, 8 February 2014 (UTC)


 * See the "Test of..." section added above. P0M (talk) 01:48, 8 February 2014 (UTC)


 * Patrick0Moran, Cthugha82 and you should read the talk page guidelines, particularly maintain Wikipedia policy and be concise. Lengthy reasoning on a talk page is futile. It takes too much patience to read, and it's trumped by verifiability every time. RockMagnetist (talk) 17:02, 7 February 2014 (UTC)


 * In this case it is indeed futile. P0M (talk) 17:58, 8 February 2014 (UTC)

Citation needed of photon detection
I have never seen a published measurement where the photons are detector at the slit. DParlevliet (talk) 22:15, 6 January 2014 (UTC)


 * Haven't you read the Scientific American article on the quantum eraser yet? They have such a set-up in that article. Wikipedia is blocking the link, but use Google to search for "Rachel Hillmer" "Paul Kwiat" and look down the list for a PDF file on the quantum eraser experiment.


 * Photons are "detected" or labeled as to which path they came through, by polarizing whatever came through the left slit with one polarizer in vertical orientation and the other polarizer in horizontal orientation.P0M (talk) 22:48, 7 January 2014 (UTC)


 * They are marked, not detected. Those articles claim that because of the marking/labeling you can in principle later on detect through which slit the photon went, so not at the slit. DParlevliet (talk) 11:42, 11 January 2014 (UTC)


 * As long as photons are what you call "marked" they will not show up on the screen as components of an interference pattern. They will show up as part of a diffraction pattern, basically a somewhat blurry main dot and a couple of comparatively much dimmer side dots.


 * The language used in these experiments is often confusing. What physicists are calling a "detector" can be part of an apparatus that can in principle be used to determine what path a a given photon has followed. For instance, with the polarizers in place in the Scientific American apparatus it is possible to put a second polarizer in front of the detection screen. Any photon that shows up must correspond to a path that, at the slit, had the same polarization as the polarization of the polarizer nearer the detection screen.


 * The whole point is that anything that allows determination of a path, anything that would allow determination of a path if you bothered to do something about it, is enough to prevent interference. What that boils down to, if you follow through the experiments and see what the physicists have really done, is that if a physical change is made anywhere close enough to a slit to localize its effect to a photon that could find a path through that slit, then it will prevent interference at the detection screen. Polarization is one of the more subtle ways that one can interfere with the wavefunction that is associated with one slit because all it does, in effect, is to rotate the wavefunction so that it has one orientation and not the other. The wavefunction itself does not get changes as to its plot. It is just as though you took the wavefunction as plotted on graph paper and rotated the graph paper. So at the screen if you plotted the two wave functions corresponding to the two slits, one would be going along an x-axis and the other would be going along a y-axis. "Erasing" the marking of the photon amounts to re-rotating it (and its "twin") so that wavefunctions are aligned again in such a way that they can interfere. You must familiarize yourself with the Kim et al. article. It is not possible to go by ordinary macro-world intuition in this case. That's the whole reason that the double-slit experiment is so important and so revealing.P0M (talk) 18:19, 12 January 2014 (UTC)


 * Note that there has been no response. P0M (talk) 18:57, 8 February 2014 (UTC)

"The Experiment"
About POM's explanation: "The laser (blue box at the left) emits a photon, which encounters a diaphragm with two slits (the heavy black vertical line). From there, two pathways for the photon emerge, a blue path and a red path. The BBO crystal sends two entangled photons forward on two pairs of red and blue paths.....". If the laser photon goes through the red slit, in the BBO two red entangled photons are emitted, on going up and one down (in the diagram). There is no blue photon(s) at the time (it went though the red slit). So if the up-going red photon reaches Ds, where does its wave interfere with? DParlevliet (talk) 13:45, 11 January 2014 (UTC)


 * What possible evidence do you have for saying that there are two red entangled photons and no blue entangled photons? There is no possible evidence for your claim since the situation with photons is different from the situation with bullets or the like. We can photograph bullets in flight. Some people can see baseballs in flight. Nobody can see photons in flight.


 * The double-slit experiment works when photons are emitted one at a time. If, as you claim, a photon goes through one slit and therefore nothing goes through the other slit at the same time, then how do you explain interference? The clear experimental evidence show that if a photon has the possibility of going by two paths it will show interference. If it can only go by one path then it will not show interference. Are you trying to claim that this is not true?P0M


 * I don't know what to do in this situation. You should read the original experiment http://arxiv.org/abs/quant-ph/9903047 and depend on it and other legitimate sources, not make up rationales on your own.P0M (talk) 08:32, 12 January 2014 (UTC)


 * I just asked if somebody knew the explanation. Of course I did read the original and that is not clear while the "Explanation by physical optics" mentioned that is was wrong, but that explanation I understand neither. DParlevliet (talk) 12:01, 12 January 2014 (UTC)


 * Again, I cannot understand your syntax. If you read the Kim article you appear not to have understood it. You appear not to have understood the basic physics of the double-slit experiment. You should be able to follow Greenstein's Quantum Challenge and you will probably find it the best source against which to check your understanding.P0M (talk) 19:10, 12 January 2014 (UTC)


 * One other thought, is it possible that you do not understand that the BBO does pass the red and blue path wavefunctions on through. It is difficult, using ordinary logic, to see why this would happen since, as you point out, the photons presumably originate in different regions of the BBO. Nevertheless, the experiment shows interference fringes and that couldn't happen unless the red and blue paths are linked in such a way that whatever got set up at the original double-slit part of the experiment is continued all the way through on both the top part of the diagram and the bottom part of the diagram.P0M (talk) 20:00, 12 January 2014 (UTC)


 * You are right, that is what I mean. If the red and blue area would generate a double photon at the same time, that would be a very remarkable discovery. I have not read that anywhere. Take also into account the remark in the article that the lens is focussed on infinity and not to the red/blue area, as it should be. So the red/blue waves will not be focused on the detector. Interference can be caused by other effects, even errors. But I have no other explanation or reference, so have nothing to change in the article. DParlevliet (talk) 20:51, 12 January 2014 (UTC)


 * Citation to the "focused at infinity" remark, please.
 * Note that there has been no response. P0M (talk) 18:57, 8 February 2014 (UTC)


 * Have you studied entanglement yet? Nobody I know of has small enough eyes to watch what goes on in the BBO, but the experimental results clearly show that there will be interference between red path and blue path whenever they meet in a properly arranged detector, D1 and D2. You can see what you get when you pick out the hits on D0 with D1 (figure 3) and D2 (figure 4) in the Kim et al. article, 4th page.


 * The BBO does produce two photons, two entangled protons that is, for each photon that enters from the double-slit diaphragm. Moreover, each entangled photon that emerges from the BBO has a red path origin and a blue path origin (or, to use more technical vocabulary, there is no path information available that could say that whether a photon came out of the red path or the blue path area). If this were not true, then there could not be a photon that ends up at the ends of both red and blue paths on D1 or D2.


 * There are experiments that show when two different lasers are placed close together side by side and are controlled so that they jointly can only emit one photon at a time, then an interference pattern will emerge on the detection screen. The argument is that it is impossible to determine which laser actually emitted the photon, so that there is no path information, and therefore there must be interference. If you don't believe me, look it up in Greenstein's Quantum Challenge.


 * Also, it is not clear that a BBO absorbs a photon by having an electron boosted in orbit and having it emit two photons in a two-hop trip back to its equilibrium state. All that can really be said is that one kind of thing goes in and two things of another kind go out. As far as I know, the assertion that different physical parts of the BBO are responsible for the two photons that emerge and go in different directions is just a matter of interpretation. The things that go out behave just as if they were emerging from a double-slit apparatus. In other words they seem to "inherit" the history of the earlier passage through the double-slit apparatus. This is a black box situation, pretty much. We know what goes into the box. We know what comes out of the box. We don't know what the "machinery" in the black box is. (Actually, I guess people know something about the crystalline structure of the BBO, but I don't know that they can explain what really happens to a photon in transit through the BBO.)P0M (talk) 21:37, 12 January 2014 (UTC)


 * Note that there has been no response. P0M (talk) 18:57, 8 February 2014 (UTC)

Point by point
I am trying to understand and make your section more understandable.

You say:

"There seems to be a misconception, regarding where this pattern originates. One might think that the signal photons shape the pattern on the way to detector D0 by interfering with each other. But at detector D0 only a blurred image of the double slit is projected by the converging lens, effectively smearing any direct interference patterns from the slits."

Whose misconception is this supposed to be? Are you challenging Kim et al.?

How can you assert that there is only a blurred image of the double slit at Detector 0? Give me a quotation from the Kim article in support of that.

You seem to attribute the blurring to the converging lens, but that assertion is not correct. All the lens does is to shorten the time it takes for photons to reach Detector 0 so that Detector 0 will do whatever it is going to do before any of the other detectors go into action.

Please answer these points with citations.P0M (talk) 19:35, 12 January 2014 (UTC)
 * Note that there has been no response. P0M (talk) 18:57, 8 February 2014 (UTC)

Not seeing any reply, I'll just have to challenge the next part that I find dubious:

"To understand the source of the derived interference pattern, one has to focus on the third beam splitter BSC, where the photon paths from both slits merge. At this point a phase difference exists between the merging paths, which is dependent on the different path lengths from slit A or B respectively to the splitter. Furthermore path length and phase difference depend in part on the deviation angle of the idler photon leaving the Glan-Thompson prism." If I am reading this correctly, I think this part may be o.k. I guess you are saying that if you observe the paths going vertically (on the diagram) from BSc then you will understand that both a blue path and a red path approach D2 therefrom, and that if you examine the paths going from the same beam splitter diagonally down and toward the left, you will also find both a blue and a red path. "At this point" is confusing to the reader since according to the rules of syntax "this point" should refer to some point in or on BSc, but you surely must mean instead the surfaces of D1 and D2. Any phase difference that exists will not ordinarily be present only at some point along any of the paths. The wavefunctions will be out of phase or in phase just depending on how far along their paths and what the exact geometry of the situation may be. (If you've ever tried to do this general kind of experiment you will realize that the slightest misalignment will throw everything out of whack and it is a painstaking business to get all the mirrors beam splitters and other components to line up properly.) The only point where phase differences make a difference to what is actually observed is along the surfaces of the detectors. With suitable tampering the two wavefunctions could be made to slide closer or farther apart horizontally just as two sheets of paper on a desk can be made so that one exactly covers the other or pushed sidewise far enough that no part of one is covered by or covering a part of the other. But all of this stuff is just details about the touchy adjustments that have to be made in the lab.

"While a fixed position on detector D0 is correlated to events at detector D1 or D2, only events at D1 or D2 are inspected, which share the same phase differences."

I don't know what you are trying to say. It is not true that there is one "fixed position" on D0 to which all signal protons are brought. Is this part of the unsubstantiated business about the lens focusing everything in the signal path at infinity? You have thus-far ignored my correction on that point. I have re-read the article and there is such assertion.

Furthermore, according to the precise syntax of your next sentence you are saying that researchers examined events at D1 and D2. In that case they would have ignored events at the other two idler detectors. That is not true. You then further assert that "events at D1 or D" happen to "share the same phase differences." I doubt that what you have said is what you meant, but you can't burden readers to try to figure out what you intended to say.

Are you trying to say that events at D1 and at D2 share the same phase dfferences, and so those are the events that are preferentially observed? Or are you trying to say something else?

"After leaving the beam splitter BSC, each phase difference may lead to constructive or destructive interference on the paths from the splitter to the detectors D1 and D2, while also allowing any arbitrary intermediate values. But one has to note, that the combined probabilities for both paths behind BSC always add up to 1, as the difference of the phase shifts between that paths amounts to π or 180°."

Interference is not manifested "on the paths." Interference is manifested only when wavefunctions are superimposed on some surface (object) and also the wavefunctions "collapse."

The business about the "combined probabilities" is only significant in the context of explaining what is seen at D0 (or if for some reason someone decided to sum the results for D1 and D2.

Then, out of the blue, you bring up the business about the phase shifts pertaining to the two detectors just mentioned. The difference occurs because of the different sequences of reflections and transmissions leading up to those two detectors. For the first it is: blue: Tr Rf Rf red: Tr Rf Tr For the secnd it is: blue: Tr Rf Tr  red: Tr Rf Rf What happens to blue in one situation happens to red in the other situation, and vice-versa. There are real consequences to this fact.

That is only significant because it explains why the curves in diagrams 3 and 4 are flipped, and their being flipped explains why, if you projected one pattern on top of the other the maxima of one would fill in the minima of the other and vice-versa.

These are all experimental details. What the experiment does, basically, is to run four interlaced experiments at random.

Every so often an idler photon lands on the first detector and will start to fill out one kind of interference pattern. It is interesting to see what the behavior its entangled twin manifests on D0. It turns out that it must contribute to an interference pattern of type one on that detector also.

Every so often an idler photon lands on the second detector and interference patterns of type 2 start to fill out on that second detector but also on the detector for signal photons.

Every so often an idler photon lands on the third detector and no interference can occur (only one path goes to that detector), and on the upper detector (O) a photon contributes to a diffraction pattern but not to an interference pattern.

Every so often an idler photon lands on the fourth detector and no interference can occur there either. As before, the upper detector sees only a photon being added to a diffraction pattern.

This experiment would be useless if we could not keep the four interlaced experiments separate, which is what the coincidence counter is used for. When there is a coincidence between D0 and one of the four detectors below, the pattern being added to (and needing to be sorted out) on D0 should be graphed on the sheet that corresponds to inputs that correlate to whichever of the four detectors lit up at the same time.

So what about erasure? The experiments are saying, as it were, if a photon shows up on the third detector of the fourth detector we are in the position of being able to stand with our eyeball where that detector is and see the bottom slit "light up." Or if a photon shows up on the fourth detector it is like being able to stand there and look directly into light blinking at us from the top slit. So we would say that we have found "which path" information. However, we can take that information away by standing at either D1 or D2, in which case we will not be able to tell whether the blip came from the top slit or the bottom slit. We cannot have "which path" information from those positions because both slits are superimposed in our visual field.

What this all boils down to is saying that we can direct light from a double-slit apparatus along diverging paths and can reliably expect one photon to show up for one photon emitted by the laser, and we can never see "half a photon" show up at the end of the two divergent paths, or else we can direct light from a double-slit apparatus along converging paths and exactly line them up, in which case there would be no interference because maxima and minima in the two copies of the wave function would be in exactly the same place, or else we can arrange for light from a double-slit apparatus to be slightly out of phase in which case we get superimposed wavefunctions, maxima filling in for minima at some points, maxima doubling the amplitude at some points (and therefore quadrupling the intensity at those points), and so forth.

It would be much better if you would discuss changes and gain a consensus in discussion before making changes. Making changes that have to be substantially changed creates turbulence that is bad for those who come to this article for information.P0M (talk) 05:40, 13 January 2014 (UTC)


 * Please read my starting post. I did not write that section. I don't understand it so asked if someone knows. That seems not to be so. DParlevliet (talk) 08:51, 13 January 2014 (UTC)


 * Sorry. I got confused by what you wrote in your starting point. I can't well understand what it's supposed to communicate. I'm just going to delete it. Thanks. P0M (talk) 14:19, 13 January 2014 (UTC)