Talk:Delayed choice quantum eraser/Archive 01

From Wikipedia, the free encyclopedia

Rich.lewis 22:14, 10 April 2006 (UTC) I disagree with part in the Discussion section about no usable information traveling backward in time is both unnecessary and wrong. As described on bottomlayer.com in the basic delayed choice(http://www.bottomlayer.com/bottom/basic_delayed_choice.htm) and delayed choice quantum eraser (the experiment performed by Kim, et al. http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm), the "delayed choice" does change the result of the observation on the primary detector. It changes the results of an observation made in the past, that is the whole point.

It doesn't change the past observations, just our ability to interpret them. See the section 'Why this is not an ansible". DMPalmer (talk) 19:28, 1 March 2008 (UTC)

I think that the most interesting result of this experiment, also as described on bottomlayer.com, is this shows the underlying flaws in the copenhagen explanation of the double slit experiment. As Wheeler observed, an experimenter may choose to know a property after the event should already have taken place. In other words, as described in the bottomlayer.com explanation, "the observer's choice would determine the outcome of the experiment – regardless of whether the outcome should logically have been determined long ago." Furthermore, Wheeler's thought experiment has been experimentally verified by the results from Kim, et al.

So, I think that comment about causality not being violated takes away from the point of this experiment. Wheeler was disturbed by the fact that causality seemed to be violated by the Copenhagen interpretation, and Kim, et al. have shown experimentally that this is true. It is a fundamental paradox that has never been resolved. Copenhagen said that he himself did not understand it and anyone who said they did was a liar.

Contents 1 Discussion section comments 2 My response 3 My reply 4 rlewis reply 5 Article Cleanup (Fact vs. Fiction) 6 Looks like Alain Aspect's Orsay team reachs the same conclusions 7 Clarification please 8 Revert 9 Something seems missing in the article 10 And something is really strange in the experiment 11 Problematical assertion 12 Polarization 13 Ungrounded assertions 14 Diagrams 15 Removed two paragraphs

[edit] Discussion section comments

A few points:

• The experiment at bottomlayer.com has not, to my knowledge, been carried out. So any expected results from the experiment is speculation and shouldn't be treated as facts in a Wikipedia article.

• Traveling faster than the speed of light and going back in time are, IMO, essentially the same thing.

• The delayed choice experiment that was carried out, as is all such experiments of this nature, requires a coincidence detector, that must match (using time coincidence) the detection of the reference particle to it's "twin" slit passing particle. You can, in principle, put the dectection (and choice) of the reference particle millions of light years away in a distance galaxy, but information (or confirmation) about exactly when the particle was detected (to match the corresponding slit particle), and thus which choice was made, can only travel at the speed of light (and will still take millions of years). Thus, DCQE cannot be used for faster then light communications, or to violate causation.

• Quantum mechanics predicted the strange effects of DCQE, long before the experiment was actually carried out. The Copenhagen interpretation gave an interpretation to exactly what QM was predicting. I don't quite see how the experimental results of DCQE, confirming QM, could contradict the Copenhagen interpretation.

[edit] My response

Rich.lewis 22:14, 10 April 2006 (UTC)

• The experimental setup I was refering to was the delayed choice quantum eraser, which was carried out by Kim et al., included in the references, and the bottom layer has a sort of layman's interpretation which points out some of the paradoxes illustrated by their results. (http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm)

• The detection of the "twin" particle can be delayed an arbitrary amount of time without putting the secondary detector in a far away galaxy, instead, use a mirror to reflect the twin particle back to a detector near the experimental setup. You could also slow down the twin particle in a number of ways, for example by using electrons instead of photons and slowing the twin by passing it through a decelerating electric field. This is tricky but in principle can be done, my point being that the detection of the twin particle does not necesrily take place at a distance separated relativisticly from the primary detector.

• Technically the copenhangen interpretation does predict this bizzare result, which while it is not actually time travel does allow classical information to be transmitted backwards in time. This information could be used to create a paradox. The fact that the copenhagen interpretation predicts a paradox seems to me to be a contradiction, and by predicting a contradiction this means it is logically false.

If you think about it, the two slit experiment itself boggles the mind. How can something be both a particle and a wave? How can a single "particle" pass through both slits in the experiment to create the interference pattern? Maybe, there is no particle, just a wave. The whole idea of wave particle duality is nonsense. The wave function is real, the particle is not. What we experience as reality is only one possible superposition of states of the wave function, but that is beacuse our experience of reality is limited to only experience one or the other superpositions. In fact, the superpositions always exist, and the wave function does not collapse. The wave function is real, our percepetion of the wave function as a single particle at any one time is an illusion, or more accurately a limitation of our perception. Otherwise, we would have information flowing backwards in time, paradoxes, and logical contradictions.

PS I changed some of my original comments that were not clear. The point I am trying to get at is that the delayed choice qunatum eraser experiment demonstrates a very spooky paradox in QM.

[edit] My reply

--Lostart 19:53, 11 April 2006 (UTC)

• I didn't read very much of the bottom layer article so I can't comment on it directly. I would only say that while there certainly seems to be a lot of paradoxes surrounding QM (and in particular, DCQE), it has nowhere near been settled as to whether the paradoxes as actual, or simply apparent, stemming from our human--thus biased--view of reality. Until the question is settled (if it ever can be), I think it's approprate to present this (in a Wiki article) as active speculation, but not as being an established fact.

• I do know it's quite posssible (in fact it's been done) to slow a particle, even a photon, even to the point of stopping it for an indefinite period of time. But I'm not at all sure that QM allows you to to use that with entangled particles in such a way that create a causality paradox.

Keep in mind that it is impossible to predict (or determine) the path of each individual particle (one slit or both) using the entangled pair. An individual particle can be detected anywhere on the screen whether you force the particle through one slit or both slits. You need a large collection of particles to determine the pattern. You can (in principle) collect all the particles from the screen, while "holding" all the reference particles until you make the decision (a minute or a year later) about how to read them. But the data you collected on the screen is worthless until you do the second part of the experiment, and use the coincidence detector to corrolate the particles, and let you see the outcome (ie the pattern). You still don't get to see the result until after you make the decision.

I don't have much formal experience in the mathematics of QM, but I'm not aware of anyone doing the mathematical formalism (ie formal proof) of the above experiment, or published a paper establishing that QM allows such causal parodox. If you know of such a published paper please reference it. My guess is doing the formalism would reveal just how QM makes it impossible, perhaps from the Hiesenberg uncertainly principle (ie keeping the particle so long makes it's location so uncertain as to make it impossible to match it with it's twin).

• The Copenhangen Interpretation (and QM) may predict an apparent paradox/contridiction; not necessarily a real one.

[edit] rlewis reply

Rich.lewis 16:42, 14 April 2006 (UTC)

I certainly agree with your comments above, and I'd like to point out that what really shocked me was this experiment in fact has been done, by Kim, et al, who I referenced but probably not very clearly since I buried the link in the text of the article. (http://xxx.lanl.gov/pdf/quant-ph/9903047 Yoon-Ho Kim, R. Yu, S.P. Kulik, Y.H. Shih, and Marlon O. Scully Phys.Rev.Lett. 84 1-5 (2000)) It is true that a single particle could not be used to effectively communicate information over time, you need many observations to construct a pattern which either shows interference (wave like behavior) or no interference (particle like behavior). However, I propose you could construct many identicle experimental aparatuses (sp?) and conduct many observations simulataneously, then actually combine the observations into a single pattern to show particle or wave behavior.

While it is also true that it is fundamentally impossible to predict (or determine) the path of each individual particle (one slit or both), which would amount to knowing both the position and momentum of the particle. However, using exactly this setup to essentially transmit information back in time, on could in essence observe the behavior of a particle after the fact and then send the information back in time to before the experiment was performed, essentially predicting the behavior of each individual particle. This would be so abhorent to the laws of QM that it must in fact be impossible to do this. But, the outcome of the DCQE experiment seems to sugest this may be possible.

Kim et al make some very interesting comments in the introduction to their paper which I will quote here:

"This paper reports a “delayed choice quantum eraser” experiment proposed by Scully and Dr¨uhl in 1982. The experimental results demonstrated the possibility of simultaneously observing both particle-like and wave-like behavior of a quantum via quantum entanglement. The which-path or both-path information of a quantum can be erased or marked by its entangled twin even after the registration of the quantum."

and in the following section...

"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 [1]. Since then, complementarity is often super- ficially 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 �x�p ≥ ¯h 2 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 Dr¨uhl found a way around this position-momentum uncertainty obstacle and proposed a quantum eraser to obtain which-path or particle-like information without scattering or �Permanent Address: Department of Physics, Moscow State University, Moscow, Russia otherwise introducing large uncontrolled phase factors to disturb the interference. To be sure the interference pattern disappears when which-path information is obtained. But it reappears when we erase (quantum erasure) the which-path information [3,4]. Since 1982, quantum eraser behavior has been reported in several experiments [5]; however, the original scheme has not been fully demonstrated."

But I agree these speculations are not established fact and should not be included in the main article as such. My real objection was to the assertion that it would be impossible to transmit usable, classical information backwards using a DCQE device. I believe this may be possible and remains to be shown by experimental observation if this could in fact be done.

I once saw a program on QM where they were describing how photons can appear to move faster than the speed of light through qunatum tunneling. First, on physicist explained how of course no classical information could actually be TRANSMITED faster than the speed of light, since qunatum uncertainty randomizes the particles and destroys any signal they may carry. The next segment had another scientist who had infact set up a simple experiment to show a microwave signal going through an intervening object, by quantum tunneling, and he had connected the device to a CD player playing classical music. He could then play back the signal (which was the result of photons using quantum tunneling to travel faster than light) and there was distorted but recognizable music coming out the other end. The conclusion was that obviously information CAN be transmitted via quantum tunneling. Scientists often revert to these kinds of "arm waving" arguments to describe what they believe to be an obvious conclusion which may in fact be false.

[edit] Article Cleanup (Fact vs. Fiction)

The article as it stands presents wildly speculative information regarding the ramifications of the Delayed Choice Quantum Eraser Experiment to which there is neither empirical nor theoretical evidence.

Whether or not the Delayed “Choice” Quantum Eraser Letter in the Physical Review supports the idea that in this experiment an outcome "in the past" is changed can be debated no more than is possible in Wheeler's original delayed choice experiment. Theoretically the implications are no different for a particle to have entered both slits and then be consigned to one than the probability of a particle's state at a point (in this case a detector) being altered due to delayed observation based on the path of an entangled particle.

The discussion portion of this article should be removed entirely. Its tendency towards science fiction detracts from the overall credibility of the article and falls prey to an unscientific interpretation of QM that has lately become popular. The statement that "quantum mechanics does not seem to have much of a problem with time travel" is incorrect. The Bohm interpretation regarding entanglement was shown recently by Polkinghorne to act more as two singers who sound out of tune unless heard together than two people instantaneously transmitting information. Thus preventing the thin use of the Bohm interpretation to permit time travel. Moreover, QM prevents an outsider from interacting in such a transfer of information, luckily we need not worry as QM prevents the possibility of such a transfer in the first place.

Separating science fact from science fiction is an arduous task. The majority of the media today can rarely tell one from another and supposing too much from antiquated theories and media interpretation can be disastrous. The discussion section of the article page goes far beyond any information presented in the Physical Review Letter and has no proof for the suppositions put forward.

I propose that in order to its scientific reliability this article must present no more information than can be logically derived from the published Letter, unless sufficient proof can be brought forward.

Your proposal seems reasonable--go ahead and make the changes you consider necessary. I would recommend not removing the whole discussion section, but just take out the portions that are wildly speculative and/or scientifically inaccurate, and preferably replace them with a better way of saying it.--that's is just my recommendation, of course. --Lostart 16:54, 16 May 2006 (UTC)

I'd also like to point out that widely held speculation would not be inappropriate in this (or any Wiki article), as long as it's stated clearly as such, as it adds only adds to information about the subject to the reader. IMO --Lostart 17:03, 16 May 2006 (UTC)

I would not describe the discussion section as wildly speculative, or science fiction. I'm not sure what parts you object to, but I made an attempt simply to paraphrase the conclusions from Kim et al. (which I quoted above). It was not my intent to go far beyond the information presented in that article. If you object to the statement that QM does not have a problem with time travel, very similar arguments are made by Feynman in his book, QED. At one point he describes an antiphoton as a photon moving backwards in time, and states that in fact that is exactly what it is. I agree with this description. What I really meant is that QM does not produce an "arrow of time." You can not easily distinguish, from a simple qunatum interaction, the direction in which time is flowing. Higher level phenomena, for example entropy and the 2nd law of thermodynamics, only emerge when you consider the behavior of large numbers of particles. If you object to my description of the results of this experiment as "time travel", I meant that changing the outcome of an event, which is the detection of the "first" photon on the "primary" detector, by choosing to observe or not observe the "second" entangd photon (the delayed choice) was analagous to time travel. This is the central paradox. I chose to call it time travel and I think that IS an appropriate description of the experimental result. I certainly did NOT mean literally travelling back in time. I simply meant the outcome of the observation at the primary detector is aparently changed by the action of the "delayed choice" AFTER the photon has already been detected. That seems like time travel to me.

Frankly, I think the results of Scully and Druhl (1982) and Kim et al. (2000) are highly provocative. The paradox which they present, of "apparent" time travel, is analagous to the apparently impossible observation that the speed of light was contant and was NOT relative to the observer's frame of reference. The puzzling out of that paradox resulted in Eintstein's theory of relativity. I believe that a similar reconcilition of the delayed choice paradox may result in a revolutionary discovery in QM. This is why I think it is appropriate to describe this result as time travel. In a sense that is what is happening. Of course this is impossible, so some important piece of the puzzle must be missing. I just think the result of this experiment is important, and describing the paradoxical observations as a kind time travel is appropriate. Rich.lewis 21:33, 1 June 2006 (UTC)

[edit] Looks like Alain Aspect's Orsay team reachs the same conclusions

No need to argue about that : this reference mentions everything there is to know about their own experiment :

Just click here and read

81.64.199.60 13:17, 2 November 2006 (UTC)

[edit] Clarification please

To a novice, this section is confusing:

"First, generate a photon and pass it through a double slit apparatus. After the photon goes through slit A or B, a special crystal (one at each slit) uses ..."

I was confused by the significane of the first double slit. And, sure enough, on reading the commentary in http://www.bottomlayer.com/bottom/kim-scully/kim-scully-web.htm I find that:

"(This double slit is a bit of a red herring. It is only a method of randomizing emissions from region A or region B of the crystal, which are themselves the equivalent of the two slits of Young's experiment.)

I was going to try to change the page text to make that "red-herring"-ness clearer (it's what I suspected when I first read the page). However, I really don't know enough to make the change safely. But I think some clarification would help people like me (who are, presumably, the core audience for a non-technical exposition like this). —The preceding unsigned comment was added by 66.69.240.110 (talk) 10:30, 7 February 2007 (UTC).

I don't think it is necessary to burden the article with this detail. The original article said that they used a single crystal. If they had just shined a laser on the single crystal they wouldn't have gotten two neat beams through it. P0M (talk) 09:13, 12 December 2007 (UTC)

[edit] Revert

Calling legitimate edits vandalism is not appropriate, even if you happen to disagree, nor are wholesale reverts (which includes grammatical corrections) without explanation. My edits were honest attempts to improve the readability and focus of the article. If you disagree with the edits, justify each revert in the talk section, preferably editing the reverts youself, and sign your comments. --Lostart (talk) 04:30, 26 November 2007 (UTC)

[edit] Something seems missing in the article

I've found a good image of the original drawing, one that shows D₄, and have made an SVG image to go with this article. The article does not as yet explain the function of all the beam-splitters and how they "erase" the path knowledge. P0M (talk) 09:08, 12 December 2007 (UTC)

One function of the BBO, which makes it useful to divide the left-slit path from the right-slit path, is that it polarizes the light differently. The Glen-Thompson prism directs the differently polarized light in diverging directions. Some of the discussions I have found mention the polarization issue, but I haven't found anything very helpful yet. It would be good to know what kind of beam splitters are used. Some beam splitters are polarizing. P0M (talk) 16:32, 12 December 2007 (UTC)

[edit] And something is really strange in the experiment

Each photon then interferes only with itself. Interference between two different photons never occurs. —Paul Dirac

The laser pumps out one photon. On the other side of the dual slits this photon disappears and is replaced by two new, entangled, photons, which then go off on different trajectories. (It isn't clear from the published materials why they do so. The diagrams make it appear that the entangled photons are generated moving 90 degrees apart.) It is one new photon that then enters the lens and is directed to a detector screen where it either interferes with itself or doesn't interfere with itself. If BBO were replaced by an analogous photon emitter such as another laser or just an LED then it would not ever produce an interference fringe. For there to be an interference fringe, a "something" has to originate in the alternate spot on the BBO, and it has to be "the same photon," at least according to Dirac.

It looks as though the original photon and the derivative pair of photons have to be entangled themselves. In other words, the derivative photons each have "inherited" the characteristics of the original wave pattern that was simultaneously generated at the exit ports of the two slits. So, in the diagram, the red line going into the lens has to represent something "real," and the blue line going into the lens has to represent something that is equally "real." The successfully interfere despite the fact that their sources are in a photon in one case and the absent photon in the other case. (I just mean that "the photon" that has been so painstakingly contrived to be generated and labeled by polarity near the output port of the double slits would presumably not exist in two instances. And its own progressively expanding wave front ought to proceed from where it is created, not for there and from a second point that is historically related but physically discrete.)

If we follow the method of Huygens, we can trace through the process without reference to the supposed photon with a presumed trajectory. The wavefront moves out of the laser (at the left end of the diagram), encounters the barrier wall, is propagated simultaneously through the two slits, and radiates with diffraction from both slits. Depending on how near the BBO is to the barrier wall, the two wave fronts might or might not be superimposed within the BBO. Regardless, the wave front "collapses" and -- then what? The paper (Kim, et al.) says that a pair of entangled photons are the result.

If we work backward from D₀, probability waves have been superimposed and a photon has flashed at a point of high probability. Tracing the probability waves back, how do we account for the second wave? Do the both come from one output port of the BBO? In that case we could presumably block one output port of the BBO and still get interference fringes. Does one come from the second output port? If so, why? Presumably nothing is happening there. But the wave front that came out of the second slit propagated to that point until it collapsed.

If we follow the progress of dual wavefronts through the prism, then there is the logical possibility that they follow mirror courses. In that case if one wavefront propagates to D₄ then its mirror version will propagate to D₃. This dual wave could collapse in either detector. If the dual wavefronts go through the first two beam splitters, they could end up in D₁ or D₂. But they could also end up in the same detectors but by symmetrically being reflected or passed by the third beam splitter.

Symmetry seems to be very important in this experiment. If the two wavefronts do not behave symmetrically, one wavefront could end up in D₂ and its double could end up in D₂ as well. So it would be interesting to check those two detectors for interference fringes.

Was detector D₄ so easily just sketched in early on and later omitted entirely because the authors assumed that what was detected by D₄ could be symmetrical to what was detected by D₃

Are any of these considerations discussed anywhere? P0M (talk) 03:47, 14 December 2007 (UTC)

[edit] Problematical assertion

However, it should be noted that an interference pattern can only be observed after the idlers have been detected, and the experimenter plots only the subset of signal photons that are matched with idlers that went to a particular detector such as D₁.

I've removed the above assertion for discussion for the following reasons:

I don't think that the first half of this assertion is correct. The article by Kim et al. gives a specific time lag between the firing of D₀ and the arrival of the entangled twin photon at one of the detectors from D₁ to D₄. The coincidence counter lists out matches between the detector for the signal photon and one of the four detectors that pick up the idler photon. Their arrivals are not synonymous.

Moreover, it is not clear from anything that I have read elsewhere that a signal photon might be received but that the corresponding idler photon would yet be lost somehow. Presumably the entire apparatus is shielded from extraneous light sources too. The statement I have removed insinuates that there is a significant subset of measured/observed signal photons that are not matched by measured/observed idler photons.

At the very least there should be an explanation of whether there are too many signal photons or too many idler photons, why this is believed to happen, and a citation provided for each of these assertions if they are to be reinserted into the article. P0M (talk) 08:31, 15 December 2007 (UTC)

The assertion you removed here, which I have restored, was not intended to imply that there a significant number of idlers for which the signal photon is not detected at all (I have no idea if this is true or not...I also did not intend to imply that the arrival of the signal and idler were 'synonymous', I don't actually understand what you mean by that). Rather, it's just pointing out that if you look at the total pattern of signal photons at D0, some of them belong to signal-idler pairs where the idler went to D1, some belong to pairs where the idler went to D2, some belong to pairs where the idler went to D3, and some belong to pairs where the idler went to D4. It is only when you look at the subset of signal photons whose corresponding idler went to D1, or the subset of signal photons whose corresponding idler went to D2, that you see an interference pattern. But if you add these two subsets together to get the subset of signal photos whose idlers went to either D1 or D2, then since the peaks of the D0/D1 interference pattern line up with the troughs of the D0/D2 interference pattern (as explained in a sentence shortly after the one you quoted), no interference will be observed in this larger subset. Likewise, the D0/D3 subset and the D0/D4 subset show no interference, so if you look at the total pattern of signal photons whose idlers went to any of the four detectors, this group does not show interference either. I provided a link to a book by Brian Greene where he states that the total pattern of signal photons won't show interference, and this is also illustrated in the description of the experiment here. Hypnosifl (talk) 04:38, 9 March 2008 (UTC)

[edit] Polarization

Take a look at this diagram. Where are issues of polarization discussed? It will not do simply to ignore the issue. It would appear that if the lower half of the apparatus were simply taken out then interference fringes could never be observed in the photons that go through the upper paths. The reason is that the BBO crystal gives opposite polarizations to photons that emerge from its two output ports. Other eraser experiments are explicitly built on using polarization to "mark" the paths of photons. (See the Scientific American "do it yourself" article, for instance.) [The article by Jacques et al. (see below) makes the use of opposing polarizations to prevent interference explicit. (Note added P0M (talk) 07:59, 19 December 2007 (UTC))]

The photons that emerge from the BBO are entangled, so changing the polarization of one "twin" will change the polarization in the other one. The apparatus succeeds in producing interference fringes in the upper part of the apparatus because of the tricky way the reflections (which change polarizations) are provided for in the bottom half. If a photon does not get recorded at either D₁ or D₄, it goes through a triple reflection vs. double reflection process depending on whether it is identified with slit A or with slit B. The net result is that in the end polarizations are not crossed. The changes in polarization work on the entangled photon too, so its polarizations are not crossed and it can interfere with itself.

This part of the experiment is so obvious and so important to the physical process involved that it must have been discussed in published materials by now. Where?

The business about reflections altering polarization is discussed in Sears Optics, pp.171ff. P0M (talk) 06:33, 17 December 2007 (UTC)

Explicit use of polarization is discussed in Quantum eraser experiment.P0M (talk) —Preceding comment was added at 08:45, 17 December 2007 (UTC)

Explicit use of polarization is also mentioned in Experimental realization of Wheeler's delayed-choice GedankenExperiment, by Jaques, et al. P0M (talk) 07:56, 19 December 2007 (UTC)

[edit] Ungrounded assertions

The last two paragraphs of the "Discussion" section may be correct, but as written they seems to stand as original research. The passages are not really clear enough to be of greatest benefit to the average well-informed reader. It would be most helpful if whoever wrote them would back them up with references and cite those references in the article. P0M (talk) 09:12, 19 December 2007 (UTC)

[edit] Diagrams

The diagrams were definitely needed for this article; I'm glad they got added. A couple of suggestions:

-The small diagrams in the intro; they are hard to read, could that be made bigger (I mean the ones embedding in the text)?

-The main diagram is good, but complicated. I think it would be hard for a reader to conceptualize the experiment from this diagram. I think adding a simplified diagram that shows the experiment on an idealized/conceptual level only would be good.--Lostart (talk) 03:47, 26 December 2007 (UTC)

I'm learning how to do SVG and just figured out how to crop images. I just re-uploaded the first diagram. When it works its way through the system it should look better. It's always possible to change the "px" to a higher value.

The main diagram has to be that complicated, I think. It's impossible to see how impossible the results are unless you can see what is going on. It might be helpful to add in some lines to show the tricky way the light is made to bounce around in the lower path.

If you will check out the article you may agree with me that it leaves many things unsaid. Supplying a simplified diagram could obscure the tricky parts that need to be understood to avoid having readers draw false conclusions. P0M (talk) 04:42, 26 December 2007 (UTC)

The main diagram has to show the BBO to establish the presence of the entangled photons. It has to show all of the detectors, and it has to show all of the mirrors and beam splitters because they are what makes it possible to mix the "red path" and the "blue path" stuff going into d-1 and d-2. The only things that could be taken out would be the prism and the coincidence counters, but they don't really make the diagram harder to follow.

The thing that is confusing are the two paths shown in the following diagram by the dotted lines.

Would this diagram be better for the article? P0M (talk) 05:38, 26 December 2007 (UTC)

Actually I was suggesting ADDING another diagram (with some explanation about it, perhaps in a different section), NOT replacing the existing one. I agree the existing detailed diagram is important to show the actual experiment. Having both a conceptual simple diagram and the detailed one would aid in in the understanding for an uninitiated reader, IMO.--Lostart (talk) 17:02, 26 December 2007 (UTC)

How about this diagram then. The experiment is tricky to begin with, and the article about the Kim experiment obfuscates the fact that if they just chopped off the lower half of the apparatus the upper part (to detector 0) would not form an interference pattern because the polarities are opposed. So what I have done is make a schematic by taking out the parts that explain how they actually detect interference and replacing that part with drawings of interference patterns in the detectors, showing the polarities graphically, and using purple arrows to show each time the polarities get switched. The first time I looked at the experiment it made no sense. It sounded like "the magic of human observation." But when I followed it out step by step it became clear that the apparatus is complicated because it takes unequal numbers of reflections to get the polarizations going the same way when they enter detectors one and two. The incomprehensible part, to me, is what that would fix the polarization of whatever goes into detector zero before the corrections can happen in the bottom half. But that's the quantum mystery factor.

Here is my diagram:

Now I see that I need one more "reflection changes polarity" label, and the text size in the lower left is wrong. And my arrow heads are gone now. Rats. (Inkscape is not perfect.)

Fixed that part but now Wikipedia is misinterpreting the SVG image and putting in a big black block in the text -- something that isn't there in the original and isn't even in the SVG if I download it from Commons and run it through Adobe or Inkscape. Maybe I'll just remove the text block and see what happens. P0M (talk) 06:38, 27 December 2007 (UTC)

fixed P0M (talk) 07:21, 26 January 2008 (UTC)

The only thing that is not essential to understanding what is going on is the prism, but it doesn't require any thinking on the part of the reader and taking it out would mean redrawing the rest of the diagram to make it appear that the beams diverge magically to go where the apparatus demands that they go. I guess I could remove the first two beam splitters, but that might cause readers to lose their orientation to the more complete diagram, and you don't have to really think about them either. P0M (talk) 05:48, 27 December 2007 (UTC)

A diagrammatically much simpler experiment is described in Quantum_eraser_experiment. The tricks with polarity are there, but the time sequence (delayed choice) maneuvers are left out. P0M (talk) 07:48, 27 December 2007 (UTC)

[edit] Removed two paragraphs

I have removed the following two paragraphs:

It might initially seem that the "choice" to observe or erase the which-path information of the idler can change the position where the signal photon is recorded on the detector, even after it should have already been recorded. However, as noted above, the total pattern of signal photons never shows interference, and it is only when one looks at a subset of signal photons whose idlers were seen at a particular detector that an interference pattern can be recovered.

Now I think I see what this paragraph is trying to convey. If the coincidence counter were taken out of the apparatus and the detection of photons at the four different detectors in the idler path were permitted to continue through a large series of trials, then the experimenters would get a mixture of interfering and non-interfering photons, and there would not be any way to sort them out. But the intention behind provision of the coincidence counter is so that the position of each arriving photon can be mapped to one of four records depending on which detector its twin shows up in. The idea of the experiment is that the apparatus itself randomly delivers idler photons to the several detectors, they arrive wherever they arrive, and when their arrival point is recorded it is found to correspond to the arrival point of the photon in the upper limb of the apparatus -- even though there is a deliberate temporal disparity between the two events.

A viewer of a remote "signal photon" detector would randomly receive either "particle" hits or "wave" hits, but s/he wouldn't have any way to learn (except later) whether a hit corresponded to a hit on idler path detector 1, 2, 3, or 4. But the original apparatus was not designed to send a message. It is the equivalent of a geiger counter acting as a telegraph key. To send a message, the apparatus would need to be redesigned so that there would be a real telegraph key that would cause the incoming photons to merge their paths or constrain them from maintaining their paths, therefore arranging for either self-interfering detection of a photon or the delivery of a photon at detectors at the ends of well-separated paths.P0M (talk) 03:45, 14 February 2008 (UTC)

I was the one who originally wrote this paragraph, and my meaning was slightly different than what you suggest above. I could have been clearer, but I was referring to the notion that it is possible to control whether the idlers go to a detector that erases their which-path information (D1 or D2) or to a detector that preserves the which-path information (D3 or D4)--to make sure they go to D1 or D2 you can replace two of the beam-splitters with mirrors, and to make sure they go to D3 or D4 you can remove the beam-splitters entirely. Thus, after you have already observed the complete pattern of signal photons at D0, you can either make the choice to erase the which-path information of all the idlers, or preserve the which-path information of all the idlers. One might naively think that this would retroactively determine whether or not the total pattern of signal photons shows interference, allowing you to send information back in time, but it turns out that the total pattern of signal photons never shows interference, even in the case where every single idler has its which-path information erased by being sent to D1 or D2. Interference will be seen in the subset of signal photons whose idler goes to D1, and also in the subset of signal photons whose idler goes to D2, but as noted in the text I re-added to the last paragraph in the section "The experiment", the peaks of the D1/D0 interference pattern line up with the troughs of the D2/D0 interference pattern and vice versa, so that their sum shows no interference.

Anyway, if you think that paragraph was unclear I don't object to it being deleted, but I think the article does require some explanation of the fact that the total pattern of signal photons will never show interference and thus you can't use the experiment to send information back in time, since this seems to be a common confusion. Hopefully the section I re-added to the last paragraph in "The experiment" will be acceptable to the other editors of this article. Hypnosifl (talk) 22:46, 8 March 2008 (UTC)

Thus, the experiment would certainly not allow one to send a message back in time, and whether the experiment requires any sort of backwards causality to understand it would depend on one's interpretation of quantum mechanics. The transactional interpretation would interpret the results in terms of genuine backwards causality, but other interpretations such as the Copenhagen interpretation, the Bohm interpretation and the many-worlds interpretation would predict the same experimental results without the need for backwards causality.

The first part is a non sequiter as far as I can see. The rest of this paragraph is dogmatic. The experiment is intended to show, and if it has been reliably reproduced, does indeed show that there is a determined relationship between what happens in the two limbs of the experiment. The Copenhagen interpretation generally follows the "rule" that one should not say anything about what one cannot observe. One observes the correlation between "hits" in the two arms of the apparatus at different times. On several criteria it would seems that the two events ought to be separate events, so it is upsetting to our common sense ideas when it turns out that they are not separated after all even though there is no physical connection at the times when photons are detected. Logically, it makes as much sense to me to say that the way the signal photon shows up in the uppermost detector determines which of the four detectors the idler photon shows up in one of the four lower detectors. Our habits of thought suggest to us, however, that the several beam splitters in the idler paths "do something to the photon," and that in turn "does something" to the signal photon. But that idea is at war with our idea of temporal sequence in causation. Bohm would presumably like some hidden variable to be set the same way for both photons and therefore to determine the relatedness of the outcomes. The "many worlds" interpretation would suggest the creation of one universe for each possible different outcome of each run of the experiment, but it would not account at all for the relatedness of the observed outcomes. Or, to put it another way, there presumably would never be a universe created in which the signal photon arrived in a "particle way" and the idler photon did not, and the same for "wave way" arrivals. P0M (talk) 03:34, 14 February 2008 (UTC)

I don't think thismaterial is relevant, and much of it is very unclear. P0M (talk) 05:44, 26 December 2007 (UTC)

Unclear, perhaps, but I would still consider it relevant. I'm just reading about this phenomenon for the first time today. Sending messages back in time was one of the first things that came to mind. Suppose that the idler path was a light year long, similar to what was mentioned at the very top of the talk page. At the end of that light-year-long path, let's say there's an astronaut at an observation post who's waiting to tell us, back on Earth, whether he sees a rogue planet headed on a collision course or something like that. If he sees one on January 1st 2100, he removes the prism for the idler path that let enabled us to detect which course the photons are taking. As a result, as of January 1st 2099, the interference pattern observed at D0 (which is with us on Earth) will change so that it no longer show any hints of a delayed choice being made. Thus we would know that one year in the future, our astronaut friend sees a rogue planet.

Dr. John Cramer wrote an article, in Analog I think, in which he suggests a sort of cosmic telegraph operated by (in your terms) moving the prism in and out.

At least, that's what this experiment would seem to indicate as a possibility. It doesn't seem like it ought to be possible, but I don't know how to reject that possibility without rejecting the findings of this experiment. That's why I think the material in the 2 paragraphs you mentioned is still important, though a rewrite's definitely needed - I really need the explanation of why sending information to the past is impossible, or my brain's likely to explode. 205.175.225.22 (talk) 23:10, 7 February 2008 (UTC)

"Sending information to the past is impossible" is a deduction from theory. To me that stands as a dogmatic statement. The most one should say is that such an event is inconsistent with theories that have shown great utility and have been very well substantiated.

There are discussions in Greene's two books on this subject that you might find helpful. Or, put it this way, if it turned out to be possible to send information to the past then lots of complications would ensue. "I'm my own grandpa" might even turn out to work. People don't like to deal with "impossibilities."

The basic idea of what Greene says is that "temporal sequence" is intimately connected with the total probability of a sequence of events. It is possible that one day I throw my eraser at the blackboard at a certain angle, it strikes the board and is deflected into the eraser tray and from there it slides all the way down to the far end of the tray where it grinds to a stop. But so far it has only happened once in my lifetime, so I think it is not very likely. It is very likely that if I open a bottle of pills and sharply strike the bottom of the bottle with a rubber mallet then many of the pills will come flying out of the bottle more-or-less together and go flying all over the room. It is highly unlikely that even one pill thrown back along the original trajectory (even intending to get it into the bottle) would actually arrive there, and getting 100 pills to "unscatter" themselves and pile neatly into the bottle because of the very way they jostled each other on the way toward the bottle would have an extremely low probability. Just as pills have a higher probability of random dispersion, so does energy. If a ball lands in sand its momentum is dispersed into the pile. It is highly unlikely that random movements in the sand would send a surge of energy toward the rock and fling it back the way it came.

It is possible, according to George Gamow, that all the air molecules in a room should at the same time all line up pointing due north, and suddenly one part of the room would be in vacuum. (That wouldn't last long, of course.) The probability of that happening would be the product of the probabilities of each molecule to head north. There being a very large number of molecules in an ordinary room under earth normal conditions, the probability of a sudden vacuum developing is low enough that none of us should wear a pressure suit to bed.

If we look at very simple systems we may get an idea of what it would be like to "go back in time." What we would need to do would be to arrange a sequence of events that returned upon itself. We have a model for such sequences of events in biological clocks. They can consist of a sequence of chemical reactions that proceed in a circular way. Another model is a kind of circuit used in computers to continually repeat the sending of certain signals to various components on the motherboard. Turning on the computer causes a single chatter-free pulse to be originated. It is delivered to computer chip one, which sends out several signals, one of which is delivered to computer chip two, two does the same thing, delivering a signal to chip three, and (usually after a couple more chips are involved) the last chip in the chain sends its signal back to chip one -- which wouldn't do anything else if it didn't get this signal because the only reason it did anything to begin with was it got the start-up pulse that was triggered by the powering up of the computer. (One good but mysterious way for a computer to get sick is to have a weak chip in this circle that does not reliably send its signal around the merry-go-round. It may pass as being o.k. until the repairman decides on desperate measures and replaces chips in the crucial circuit one by one.)

We have to get energy to power the sequence from somewhere, but maybe we could count on "quantum flux" or some other mysterious source. We could "keep time" by watching the little LEDs attached to each chip, and we would see them light up one by one. If we had 12 chips we could have a 12 "hour" day in this little world. But there would only be one day as time is defined in this hypothetical world. To model a world more like our own, we would have to add components.

There doesn't seem to be an incipient paradox in this little universe. We would seem to have the kind of universe envisaged in much of early philosophical speculation. In India the idea of kalpas, eternal cycles of time turning back on itself and becoming its own beginning, is particularly prominent.

Greene suggests that any sequence of events that led around in a circle, via a wormhole perhaps, would have similarities to this simple circuit. There is no objection that I can see to things turning out differently. Suppose that somebody invents a time machine, hates himself for doing it, and goes back and causes his fetus to be aborted. From there on, the Universe takes a different course. If time is nothing other than the sequence of events (perhaps measured by sequences of events favored by various groups of observers -- I favor the atomic clock), there is nothing inherently paradoxical in this circularity as far as I can see. But it looks like the "trip back" (so to speak) would have to be done in isolation from one's original physical context -- perhaps by going to another inertial frame that does not strongly interact with one's native inertial frame. One's own biological clock would continue to cycle through its changes, cells would age and die, and when one "returned to one's own time-line" it would be as an older individual. '

It seems easy to imagine how circular time would work in the mini-Universe imagined here -- assuming that time is nothing more that the operational definition of time would imply. It is harder to imagine how circular time might apply to our Universe. But if the Big Bang is eventually followed by the Big Crunch that returns everything to the original state "before the beginning of space and time," that sequence would fairly closely follow the model of circular causation outlined above. Would it be then the same "pre Big Bang" whatever you call it, or would it be another "pre Big Bang" -- and does that question have any meaning in the absence of there being any characteristics by which to identify any "thing" at that point?

If, in opposition to what the relativity theory seems to be telling us, there is an actual, absolute "flow of time," then circularity would imply different events occupying the same space and time. It looks like there would be a kind of "superposition" in any case of time travel under that interpretation. There might be an Oswald in the book repository and a time traveler with his own weapon in the same room. But the time traveler would co-exist with a volume of air or perhaps co-exist with a volume of hard-bound books. Would a dead Oswald soon co-exist with a living Oswald and then shortly later a living Kennedy co-exist with a dead one? Maybe co-existing with a volume of air would not be too painful. Under those circumstances, meeting oneself on the street would not seem to pose any paradoxes. Killing oneself might be analogous to the Korean War era fighter pilots that riddled their own canopies with machine gun bullets by firing off their machine guns and then going into a power dive whose path intersected the trajectories of those bullets.

If causation is deterministic the way Laplace thought it was, then would a span of time be like a span of roadway that can consecutively have different vehicles traveling over it? Would there be a loop in time? The time traveler kills himself, so time goes on without him, so there is nobody to come back in time to kill himself, so... One of the interesting things about quantum mechanics is that it suggests that there is no certainty in causation, so after some number of iterations maybe the time traveler decides not to go back to kill himself.

One of the reasons for Kantian philosophy was the argument of what space and time are, an argument that started at least as early as St. Augustine. One view that eventually came to prominence was that space is only a relationship, and that time is only a relationship. But physics seems to be telling us that empty though it may be space has characteristics and existence. Perhaps time may in some way tell us that it too is not merely a relationship. P0M (talk) 05:32, 14 February 2008 (UTC)

I don't remember the details, but faster than light communication, were it possible, would enable some sentient entities to send a message from inertial frame one to inertial frame two to inertial frame three, and then back to the first inertial frame where, if the bounces were arranged properly, it could get there before it was sent. (I think it takes three inertial frames to create this paradox.)

Some of the stuff that I cut is either flat out irrelevant or else it is relevant in some way that the writer understood but failed to convey (at least to me). Since I've read quite a little about it before, and I've even worked with the numbers enough to start getting familiar with how everything works, I somehow doubt that the average reader coming to this stuff for the first time would be able to make anything of it at all. As far as I can tell, it is all dogmatic assertion.

So what should the article say that can be backed up with solid citations?

Note that sending a message instantaneously does not in itself violate anything but our sense that we had finally figured out how the universe works and now there is an exception -- the non-locality that Einstein saw and saw as a denial of the possibility that quantum mechanics could be right.

Ah, I think I see. There is a problem that Wheeler saw and expressed well in an interview reported in Scientific American. People start out with the idea of there being one photon in one place. They are bound by their lifetime of habits to "see" the photon as a little ball. That being the case, the appearance of a photon at the exit port of side one or side two of some apparatus implies to them that the photon has stayed a discrete entity and has traveled one path to get to the one exit. So if an experimenter can do something a few light years away from the point where the photon either took path A or path B that determines whether the photon took one path or another, then that decisive action must reach back hundreds or thousands of years to the photon that was "deciding" which way to go around the intervening extremely massive body that formed a gravitational lens and give it some kind of shove toward one or the other path. But Wheeler says that it is fundamentally wrong to make such assumptions about the photon. The "photon in flight" responds by appearing as a wave phenomenon or a particle phenomenon depending on how we address it as it reaches where it finally discloses itself. If we permit it to interfere with itself we will get a reading consistent with the formation of an interference fringe. (One photon does not make a fringe, but photons show up in positions along a detector screen where they would not show up if they were not interfering with themselves.) If we segregate the two or more paths it could be traveling along, then it will show up at the end of one of the segregated paths and the most "spread" we will get out of it will be consistent with a diffraction pattern (if even that).

You can see what is, IMHO, a very much better depiction of what the photon can be understood to do in the article on Mach-Zehnder_interferometer.

So far, nobody has made an "ansible," even just one that can only send morse code. It shouldn't be very difficult to fabricate. If somebody makes one and can send an instantaneous message, e.g., by bouncing a beam off a reflector on the moon that originates in LA and is picked up in NYC, then such a device would excite great interest. As far as sending messages back in time, however, we would need the cooperation of sentient being of some kind traveling in different inertial systems. Without worrying about any special requirements for the velocities involved, until SETI pans out we probably will not have anybody to bounce messages off of unless we send crews out on long star treks. But being able to communicate instantaneously even with relatively near neighbors such as Mars would have great utility. Imagine being able to drive a Mars rover in real time. P0M (talk) 06:28, 8 February 2008 (UTC)

So what you're saying, in a nutshell, is that I'm not necessarily misinterpreting this experiment by believing it opens up a possiblity of backwards causality; that's a reasonable conclusion that, at this point in time, still conflicts with many other things we "know" about quantum physics. There's no resolution yet.

Backwards causality is in one sense a common thing, at least on the atomic level. I don't remember who pointed it out before, but some atomic-level interactions are sort of symmetrical with respect to time, i.e., you could look at one interaction as a "going forward in time the ordinary way" event, but there are other events that look just like it except that they go the other way, and you could think of them as being the same thing happening in reverse.

Some of the quantum eraser experiments involve entangled photons which are sent through two different experimental set-ups. The two have different optical path lengths, so that the photon that goes through set-up A will hit a detection screen sooner than the entangled photon that goes through set-up B. Nevertheless, something that happens to the entangled photon in B will determine how the one in A behaves. So if the photon in B is coerced into manifesting as a particle, then the "twin" in A will manifest as a particle, but if it is coerced into manifesting as a particle in B then its twin will do the same thing in A. It doesn't sound too bad until you get somebody who wants to make the optical path in B very long so that (in the extreme case) maybe years flow by before anything happens at a detection screen in B, but within minutes something has happened in A.

Nevertheless, somebody at B couldn't win bets by communicating with somebody at A to find out where the photon was going to appear at B. The reason is that the guy at B would still have to send out a message at a rate governed by c, which would get there too late to permit cheating.

To get to the point that you could cheat you would have to set up the sort of double-bank shot or maybe it is a triple-bank shot. Our everyday habits of thought get in the way of understanding how it works because we have a naive idea that things can occur "simultaneously." P0M (talk) 00:20, 12 February 2008 (UTC)

The Wheeler interview you mentioned makes sense. I'd gotten used to thinking of particles as fuzzy little clouds of possiblity that are sort of here, sort of there, sort of everywhere and nowhere, and they only collapse to a sensible particle-like state when the outside world interacts with them. The new part to me is the idea that something's "location" in time is as fuzzy as its location in space.

Wheeler was just saying that it is a matter of habit and prejudice to say that "the photon" goes one way or the other way around the high mass body that is doing gravitational lensing. So we imagine that if we can switch how things show up on earth (as particle or as wave) then we have switched how the photon "decided" to go around the high mass body (black hole or whatever). To the extent that the word "goes" still means anything, the photon goes both ways (or Feynman would say that it goes all ways). But the quantum eraser experiment seems to me too to indicate that our idea of location in time as well as in space is not what our macro experience would make us think.P0M (talk) 00:20, 12 February 2008 (UTC)

Here's one more idea, more plausible than keeping an instrument perfectly aligned when one part's on the earth and the other part's on Mars. What if we used mirrors to keep the photon bouncing back and forth, giving it a nice long path like a tenth of a light-second, before it reaches the optics in the idler path? Then the device can stay in 1 reasonably sized lab. I'm not suggesting this as a useful device, but as something to investigate causality here. It seems to me that if you were to remove the prism in the idler path (or just block it off by putting something opaque in the path), then the data at D0 would respond to the change 0.1s before the change occurred. I wonder how that would play out in the lab? Would the experimenter attempt to "trick" the device, reaching to block the idler path but pulling his hand away at the last moment, trying to get D0 to exhibit a change but then not taking the action that caused that change? Would his intent to "trick" the device ensure that D0 never exhibited a change at all, except when the experimenter really did block the idler path?

There have already been experiments done with entanglement kinds of lab apparatuses with very long paths. They've used fiber optics, typically. Have you checked out John Cramer's article? He's actually working on the kind of experiment you want to see.

It seems to me that the useful thing to do would be to have two streams of entangled photons involved. Let's say that you had an interstellar space ship that went to Wolf 359. Halfway there it established a mirror that will automatically orient itself so that it is pointing back toward our sun. It continues on toward Wolf 359. After it gets there and establishes itself, it picks up a laser beam from earth that started out roughly four years earlier. Earth is at the same time picking up the bounced twin that was send to the mirror. (The mirror can't be exactly on-line between Sol and Wolf 359, or the two beams would both get reflected by the same mirror.) So Earth has the twins of photons that are received by the space travelers orbiting Wolf 359. Getting a spate of photons that show up with particle-like behavior would be a "dash" and getting a spate that show up with wave-like behavior would be a "dot."

Before we go to all of that trouble, using one fiber optic cable from NY to SF and another from NY to Omaha and back and then playing with that system should be a good proof of concept. Maybe the easy way to get photons capable of interfering with themselves would be to use two pairs, with a beam splitter at both ends of the telegraph operator's circular cable. The terminal beam splitter would be the telegraph key. The straight pair of cables would just merge their beams and watch to see what happens. (Actually, they could take turns, and the guy in SF could have an interruptible beam splitter to send with so that the guy in NY could monitor the now-unterminated cable pair to see whether there was any sign of interference.)

If that would work, then non-local communication would be possible. But would there be any causal paradoxes? What might work would be to make the Earth-return loop of the interstellar experiment short. Then the Earth-return message could be determined by what would happen about four years in the future. If the same thing were done in reverse, then the Wolf 359-return message with Earth news on it could be determined by what would happen about four years in the future. Or would it? I need to sit down somewhere with paper and pencil and go through this stuff step by step.P0M (talk) 00:20, 12 February 2008 (UTC)

Last thing, could you explain how different inertial frames are important? I don't understand how those come into play. Is redshift/blueshift important? I'll do my best to follow along. I studied quantum physics a little bit as an undergrad and spent a summer in a laser optics lab working on a new variety of interferometer (JILA lab at CU-Boulder, pretty cool facility), so I'm no layman but I'm still far from an expert. Thanks a lot for your time. 205.175.225.22 (talk) 17:25, 11 February 2008 (UTC)

Our idea of "simultaneity" is naive. Our idea of time is likewise naive. As long as everybody is moving along together (or in tech-talk, "in the same inertial frame") then "the same time" has a clear meaning. The faster two observers are moving with respect to each other, the more distortions are introduced into our ordinary-world ideas of time. If events are not really simultaneous then the alternative is that there is a temporal difference of some sort between them. I'll try to look up the reference in Greene where this stuff is explained. I certainly can't remember it or derive it now (if ever). P0M (talk) 00:20, 12 February 2008 (UTC)

See section 4.1 of Time travel. P0M (talk) 17:12, 12 February 2008 (UTC)

Thanks for pointing me towards John Cramer. I looked him up after my last post here. His site's like an encyclopedia of mind-blowing ideas. What's even more surprising than the ideas themselves is the fact that he's not a quack, he's a respected published researcher. Generally, scientific ideas this strange come from pseudoscientists selling home time travel kits and ranting about the government conspiracy to silence them.

I did have one other thought come to mind. The device in this experiment deals with only 1 photon (well, 2 entangled photons) at a time, correct? Does it have to? If the idler path is 0.1 light seconds long, are we limited to putting out at most 10 photons a second? If having a second photon in the idler path messes things up, then we face a huge restriction in data rate. If that's the case then a 1-light-year device would only handle 1 photon a year, and it takes dozens of photons to determine what kind of pattern we're seeing at D0, so it would take decades to establish 1 bit of the message. The only way I can think of to compensate for that would be to use not just 1 device, but an array of hundreds or thousands so you could get some reasonable amount of statistical certainty. Or am I going off on a pointless tangent, and having a million photons in the idler path at once would work just as well as only one? 205.175.225.22 (talk) 17:52, 12 February 2008 (UTC)

No problem with multiple photons. It's easier to talk about what actually seems to happen (and, experimentally, it makes certain thing clearer) when one photon at a time is run through an apparatus. But for practical purposes one would need many photons.

I fear that I just have not seen the gotchas yet, but if the apparatus described for this article could be made to have a relatively short path on the upper leg, and an interstellar distance on the bottom leg, then it looks to me as though messages could be received "before their time" and two sets of apparatus could make it possible to conduct conversations over long distances. A kind of "time travel" could result if one leg was very small in comparison to the other, and useful virtually instantaneous conversations could take place if the the pathways were approximately equal.

Here's what the schedule for a fictional conversation about an explorer who has suffered a life-threatening injury.

P0M (talk) 03:10, 13 February 2008 (UTC)

You may have to click all the way through to the actual svg image on this one. For some reason it is not getting handled right by Wikipedia. The svg is o.k., but when the system tries to produce a png image from it things seem to be going wrong.

Anyway, Earth station generates entangled photons, sends one through a double slit and in oppositely polarized beams on to Wolf 359. When it arrives, the explorers either let it alone or change the polarity so that the two beams can interfere. Back on earth, nearly four years before that happens, experimenters are watching the stream of twin photons to see whether they show up as interfering or not interfering.

The explorers near Wolf 359 have set up the same kind of apparatus, send one stream of entangled photons back toward Sol, and watch their twins to see what the experimenters on the Sol station have done with the received photons. It takes four years for the twins to reach their remote target. If there is a pair of mirrors at the halfway point, then the round-trip photons will come back to Sol at the same time their twins reach Wolf 359, and meddling with them from the Wolf 359 end will influence the behavior of their twins that are just then returning to Earth.

Is that the kind of communication device you were thinking of? I can't remember what Cramer said, but I think he has basically the same idea. Obviously he published it several years ago.

There would be huge collimation problems I believe. But maybe between Earth and Mars it would be useful for running Waldos. P0M (talk) 03:25, 13 February 2008 (UTC)

That's not exactly what I was picturing, but the concept is the same. I was imagining one device starting on Earth with its idler arm reaching into the station in deep space, and another device starting on the station in deep space with its idler arm reaching back to Earth. Keeping things perfectly aligned would be practically impossible, but hey, it's a hypothetical, not a patent application. The problem (or advantage) of this arrangement is that instead of being instant, every signal going either direction "arrives" long before it's "sent," which is very handy for some things but makes actual conversation impossible. Your suggestion would work better for real-time communication. My suggestion looks much more paradox-prone, but I'm of the opinion that a paradox could never happen.

I see the universe as a stable 4-dimensional entity where every object from the past, present and future has a fixed, stable line of existence. The future conditions of all those lines may be absolutely unknowable from inside the universe, but the lines exist nonetheless. A time-travel paradox would be impossible because only stable loops could be part of a stable 4-d universe, and as a result you can only interact with the past in ways that do not re-write the present. In sci-fi it shows up as the story of a time traveler finding out that the history he knew was all about his own actions in the past. But, like I said, this is just an opinion - and probably a completely untestable opinion at that. 205.175.225.22 (talk) 15:49, 13 February 2008 (UTC)

I think the only way you could test it would be to go back in time and change something. ;-) P0M (talk) 01:35, 14 February 2008 (UTC)