Talk:Wheeler's delayed choice experiment/Archive 1

There must be something wrong the article, or with its source(s). The photon that strikes the detector screen is either reflected back toward the emitter (but scattered unless the detector screen is a mirror), or it is absorbed -- which means that it disappears and an electron in orbit around the nucleus of some atom is bumped into higher orbit. Anything that goes through the detector screen and is observed by a telescope is a different photon. It's like two bullets going through the window of a car. One bullet hits a passenger and the other bullet goes through the car and is observed when it goes right down the barrel of somebody's telephoto lens on a very fast movie camera that is capable of photographing its own imminent destruction. Even in that case one would not know whether the bullet that hit the passenger came from the same assault rifle that hit the camera. If there were two shooters standing side by side the camera could be aimed at one and so get a direct hit from its bullet, but the passenger could have been in direct line with the shot fired by the adjacent shooter. P0M (talk) 09:29, 23 November 2007 (UTC)

First attempt at analysis
The article gets confused between one photon and many photons. The reason for the confusion may be that the article on which it depends is not much better. I have difficulty believing that a physicist as eminent as Wheeler could have made the elementary mistakes of reasoning displayed in this article. The cited article is not much better than the Wikipedia article.

Part of the problem may be that if one assumes the conclusion that one purports to prove the whole exercise is compromised. One of the questions that seems to be implied in the Wikipedia article and in the article behind it is whether light is a wave or light is a particle, and the writers seem to assume that light is really a particle.

The description of the large-scale "experiment" is absurd as it is presented in the article because it assumes that one could isolate one photon and determine which of two telescopes that one photon comes through. The reality, of course, is that there is a star somewhere far away that is sending off a high flux of photons from points all over its surface, and many of those photons are going to reach the observer simultaneously -- and even more are going to arrive within so short a time interval that it would be impossible to separate them in temporal sequence in our experience.

To cut to the quick, Imagine that one has a detector screen in place and observes the expected interference pattern. Now walk out in front of that screen with telescope in hand and aim at either slit (or either side of the distant black hole or galaxy. Or bore a couple of holes through the detector screen and point your telescope through those holes. Now dissolve the detector screen. Does anything change? P0M (talk) 18:18, 23 November 2007 (UTC)

The illustration given in the external article suggests that if 14 photons were emitted and passed through a double slit, then 14 photons would light up the detector screen, spreading across in a pattern consisting of 1, 3, 6, 3, and 1. But if the screen were removed and two telescopes placed where the groups of 3 protons were earlier detected, then in this run of the experiment 7 photons would be detected at the same place in space where previously only 3 photons had been detected. So half of the photons would go into each telescope. If that were the case, then simply cutting away the detector screen except at the two secondary maxima positons would cause the photons to suddenly grow much brighter as the 6 photons that previously went to the center and the two photons that went to the polar positions would all pop into focus over the two remaining parts of the screen.

The diagram also assumes that the two secondary maxima are located on a straight line between the source and the slits. If that were true, then wouldn't the interference pattern produced in the double slit experiment depend on the distance between the light source and the slit screen?

If such results can be supported by experience, and it should be dead easy to do, then a positive result would indicate that in the presence of a continuous detector screen "6" photons would hit directly opposite the emitter, but when somebody takes out the center of the screen those photons don't just head toward that spot and keep on going until they actually hit something, but instead they "know" that nothing is going to be there so they alter their trajectories so they will hit where there is something to hit.

If that were the case, any bit of space flotsam between the distant star and earth that happened to fall on a maxima ought to have redirected all the photons that pass by onto that one molecule of amino acid or speck of shattered comet, and that would prevent any light from getting to earth.

Going back to the basic double-slit experiment: Both slits are important. The wavelength of the radiation is important. The distance between the slits is important. The distance between the barrier wall and the detector wall is important. Whatever it may "really" be, something has to have free passage through both slits. We do not see anything until a light energy is delivered to a spot on a piece of photographic film, the retina of one of our eyes, etc. When we see a distant flash of light we infer the trajectory it has followed to get into our eye. We depend on the observation that is well attested in experience that light travels in a straight line. If we want to clarify to ourselves and to demonstrate to others where a light source may be found, e.g., we observe the cigarette glow of a distant sniper from time to time, we can use a longish tube affixed to a stable mount like a tripod. If we can see the light through the tube (or the gun sight) we are pretty sure that if we follow that line back (with a laser beam or a bullet) we can strike the source of that light. But as people who try to shoot arrows at fish have learned, there is an inference involved here and things may be more complicated than we imagine.

The experiment, at least as it is described in the external article, seems to involve the use of a couple of telescopes precisely because the designer of the experiment wants to eliminate the possibility that light could reach observer A from both slit a and slit b, and just by observing a lit up spot on a detector screen we can't tell whether it went by way of a or b. (That's assuming that it didn't go by way of both of them.) The experimenter believes, on analogy with the behavior of laser beams and bullet trajectories, that aiming a tube at something means that anything that gets down to our end of the tube must have come on a straight line path that goes down parallel to the cylindrical walls of the tube.

One thing that should make us cautious about this line of argument is that we know from lots of experience that a lens in the middle of things can create the impression that, e.g., the sniper who is a mile away is located at a position that our binoculars and our binocular vision tell us is only a few hundred feet away. Looking at a bird through binoculars it may look as though we could reach out and touch it. There is an inference here that may catch us unawares.

One of the things that we know about light that passes through a single slit is that it diverges beyond that point. That's why we can use diffraction devices as flat magnifiers.

If we go back to the classical ideas of the Huygens and Fresnel, we expect to see a wave of light reaching us if we stand anywhere on the detector screen when a single slit is open. If we direct a telescope (or the tube from inside a roll of paper toweling for that matter) toward the open slit we can still pick up the wave of light and we can identify the direction from which it is coming. If that slit is closed and the other one is opened, we will have to shift our telescope's orientation unless it is one with a very wide barrel. But we will see the light coming from one slit not as approaching us in a straight line from its true source but as light coming from not quite a point source at the point where the slit is, but something pretty close to that. In other words, our senses will tell us that there is a "fire" in the mouth of the slit. So if we stand on the far port side of the apparatus we will falsely infer (it will appear to us that) the light source must be at the far starboard side of the apparatus somewhere behind the slit wall.

We would get the same general impression regardless of which single slit is open. Note that light comes to us on a line between our telescope and a slit in the wall. With a telescope of a sufficiently narrow bore we could differentiate between the two slits, however.

Now, what happens when an observer is standing far to the port or to the starboard side of the double-slit apparatus? With one slit open one gets a simple diffraction pattern, and far out in the bleachers it is quite dark. Open a second slit and if the observer is at one of the maxima of the resultant interference pattern a light suddens shines into that dark corner. Where does that beam of light seem to originate?P0M (talk) 19:58, 23 November 2007 (UTC)

Unless I'm missing something, if you do the classical physics, the physics that has predicted the way light goes through telescopes, microscopes, etc., the accurate description of what things look like (even though later studies have suggested a new model for understanding what is observed), what you will discover is that to the observer should see light coming down the bore of his/her telescope when the scope is aimed at the center point between the two slits.

I just did a quick construction with a compass and a somewhat ungenerous ball point pen, and that is the result I got. I can try to find a clear chart from Wikipedia Commons or a good textbook. If that is indeed what an observer will see (which can be checked out in the laboratory just to be sure that physicists since the 1800s haven't gotten it wrong), then the Wheeler experiment as described in the Wikipedia article and the external artical would not work.

So for right now it is back to be books to try to find an account of Wheeler that is not flawed. P0M (talk) 04:49, 24 November 2007 (UTC)

The article at: http://www.sciencemag.org/cgi/content/full/315/5814/966 is much better. It states the essence of the experiment as follows: "Either the interferometer is closed (that is, the two paths are recombined) and the interference is observed, or the interferometer remains open and the path followed by the photon is measured." "There is one photon." But what does that mean? Under one interpretation all we know about the photon is that we whacked the emitter in some appropriate way and after a period of time consistent with the known speed of light a flash is spotted on the target stream. Where the photon "is" or "goes" between times in unknowable. Under another interpretation a photon is a discrete entity and must go through one slit or the other. Moreover, it presumably travels in a straight line. Therefore, the reasoning goes, it should be possible to tell whether a photon has come out of slit A or slit B by pointing one telescope at slit A and another telescope at slit B. Since if you stick one finger in front of the lens of the left side of a pair of binoculars you never see it through the right tube, we expect that light that starts in at the top of one telescope will travel to the other end of that telescope and be recorded there. So we can tell which slit it has come through by observing which telescope it ends up in. That all sounds reasonable until we ask what will happen if we put our two telescopes out in left field. Sometimes light goes from emitter to slits within the bounds of a narrow cylinder (which is a little problematical because one emitter is sending light out "in a straight line" and it arrives at and passes through two slits that are separated by some distance -- but neither a pinhole nor a laser is really a point source, and diffraction does occur when light emerges from a narrow confine) and then it bends in its path. Diffraction does happen, and a change of path sometimes must happens so that one photon "goes around the corner." One does not, however, watch light move out from one or the other slit the way one could watch a baseball come out of one or another pitching machine. One just gets hit by a baseball and one can then try to follow its trajectory back to a source.

The argument of the Wheeler camp seems to be that if "the interferometer is closed," i.e., if there is a full detection screen (they seem to talk in terms of a long strip of photographic film), then "interference" will take place. The implication would seem to be that if the detection screen is not there then interference will not take place. But that would be analogous to a water wave experiment in which two gaps in a sea wall would permit waves to be directed against one side of the wall and then pass through the two gaps after which there would be two separate but closely related series of waves spreading out from their respective slits and yet there would be no "interference" until the opposite shore is reached. But an observed will clearly see the waves and how they form a form a special kind of pattern as they either reinforce each other or cancel each other (or do something in between) at various times and places on the surface of the water between the sea wall and that remote shore. If there were an observer somewhere on the remote beach s/he would observe the oncoming waves coming toward him from a particular direction. But if the sea shore were suddenly replaced by more water and he then stood surrounded by water what he saw of the direction the waves were coming from would not be changed.

The alternative the aforesaid quotation provides is as follows: "If the interferometer remains open [then] the path followed by the photon is measured." We can't very well talk about this assertion by suddenly introducing an idea of a particle into the water wave analogy just given. So let's back up and try to describe the whole experiment in terms of a bullet that is shot and that eventually passes down one or the other of two pipes and ends up making a physical change in the target at the end of one of the pipes. We have a particle emitter, or, in our analogy, a gun at the far end of the experimental apparatus. We trigger the gun. We know from repeated experiment that when we load and trigger the gun we will very soon notice a new hole in the remote detector screen. But this time there is a barrier wall across the middle of the apparatus, and the barrier wall has two holes in it. Despite the fact that we have done our best to aim our gun dead center between the two slits, the bullet does not end up hitting the barrier wall. Perhaps in most instances the bullet fired does hit the barrier wall, but we generally do not notice because we are not interested in the bullets who do not make it through. And perhaps the gun shoots such a wide pattern that many of its bullets hit on the outside of the slits. But observers out at the line where the detector wall goes will only become aware of bullets that travel down one or the other of their telescope barrels. Suppose there is somebody with two telescopes directly opposite the emitter, the gun that fires the bullets, and s/he aims one telescope at the known location of one slit, and the other telescope at the remaining slit. If he were at the dead center position and real bullets were involved, the bullets would not hit him dead center. In fact, if he were thin or he were standing far away from the gun, the bullets would have to go to one side or the other of him. Opening and closing slits would not alter what happened in the course of many bullets being shot, only the sequence in which their impacts painted a band to his left and to his right. And if he felt the bullets were getting too close he might take advantage of a lull in firing to move far to the left or far to the right -- to places where he would find no bullet holes because real bullets go in more-or-less straight lines. With real bullets he is absolutely safe once he gets beyond the wedge delineated by drawing straight lines from the gun barrel to the two sides of slit A and to the two sides of slit B + some small amount to account for the fact that real bullets can drift a bit due to wind and other such factors. Does it matter to the guy dodging bullets whether there is a detector wall behind him? It does nothing to influence the trajectory of the bullets. Buy according to the Wheeler people the presence of a wall prevents protons from acting like bullets, and its absence causes protons to act like bullets.

So what happens if we put the guy far to the left side of the apparatus or far to its right side? With real bullets he is outside the war zone. With photons passing through a single slit there is a little fan-out as they diffract because of passing through a narrow slit, but he doesn't have to get very far away from the center, relatively speaking, to be safe. However, if a second slit is opened, then at some points in a regular distribution along the detector wall our observer will be met with a hail of bullets. Will anything be changed if the detector wall is removed? I think not. In fact, I have done a simple experiment that seems to indicate that the interference pattern is there regardless of whether there is a wall for it to interact with.

The next thing to do is to imagine the photon as a particle and see what happens to it. P0M (talk) 08:58, 24 November 2007 (UTC)

In order for light to "bend" around a very massive object, space has to change. Light always travels by the most direct route, but now the most direct route has changed. (Suppose that in ancient times Earth was really flat. The shortest path between Japan and Los Angeles went through the central Pacific. Later, Earth transformed into a sphere, the Age of Magic died, and the shortest route went up near the North Pole.)A massive object can make space bend, and light's path around that object will change. (The classic experiment when Einstein's ideas about space and gravity were tested during an eclipse of the sun. As light rays from a distant star came nearer and nearer the eclipsed sun (since Earth was moving our point of view), the position of the star relative to other stars appeared to change. i.e, the distant star appeared to move nearer to the sun.) If a particle follows a straight line that changes into a line with two tails, the particle may go by either fork, but since it is a classical particle it cannot go both ways at once. It then appears to come toward the distant observer from around one edge or the other edge of the intervening massive object. The distant observer sees one particle approaching, and there is no problem to trap it in one tube or the other.

The Wheeler experiment does not deny that there is some wavelike "presence" that goes around both sides of the intervening massive object. It maintains, however, that it is possible to trap the discrete particle. Trapping the discrete particle will make the path known, and making the path known will eliminate the interference phenomenon. The detection of the path occurs late in the experiment, just before the photon was due to be discriminated.

If the particle is not detected then it could manifest at any of the maxima on the probability curve. If it is detected, then it can manifest at only one or at most two points. So the trajectory of the photon is determined by the eventual detection of the end of the trajectory -- even if it has taken hundreds or even thousands of years for the photon to come around the massive object in its path and proceed on to Earth. P0M (talk) 02:31, 27 November 2007 (UTC)

Further research
I found a quotation from one of Wheeler's interviews (?) in Scientific American that shows that he was clearly aware that the wave/particle dichotomy has everything to do with what is happening at the time that the photon is "observed" (or put into a state where it could be observed).

If we start thinking about rifle bullets or other such "particles," the whole two-slit experiment doesn't make sense. In the beginning, when people thought of light as a continuous disturbance in the aether, it was easy to see light waves moving out from the source like water waves moving out from a bobber jiggling on the surface of a pond. When a wave front encounters a barrier wall it is reflected back unless there is a hole or two in the wall, and in that case a portion of the wave goes through each hole. The two holes behave much like two small bobbers replicating the motion of the original bobber. Ripples spread out from each hole onto the water surface beyond. Their interference (reinforcement and cancellation) is manifest on the surface of the pond. Eventually they reach a barrier and manifest their energies in another way by sloshing water onto the barrier wall where there are maxima.

Wave motion spreads out from a "point" (actually a small volume) so it is easy to see how it can go through two holes in a barrier wall. Rifle bullets will, in reality, reach different points even when fired from rifles fixed firmly in vices. The reasons have to do with the machining of the rifle barrel, the differences in individual loads of power, the differences in individual slugs, etc. One can work toward the perfect rifle, the perfect rifle cartridges, and also arrange to shoot the rifle through a vacuum to avoid differences in air pressure, unlucky sparrows, etc. One can arrange for a tighter and tighter spread in the pattern of bullet holes in a target. If a well designed rifle were shot at a double-slit device constructed to scale, it seems likely that many of the shots would hit the part of the wall between slits, many of the shots would hit the parts of the wall outside the slits, and many of them would go through one or the other slit. The bullets that happen to go through the slits would not lose any energy in the process. The bullets that hit the barrier wall would lose energy to the wall. With water waves, a huge storm outside a barier reef will produce small waves in the protected cove.

Now go back to imagine (on the basis of what we have learned by observations) what happens when a single photon is emitted from some emitter, e.g., a laser. We can't even do the equivalent of bouncing bullets off flying bullets to try to "see" bullets. We have no way of observing whatever is going on between the time the photon is presumably emitted and the time when its presence is manifested by a flash on the detection screen. But it is clear that putting a wall up in the middle of things makes a big difference. My understanding of the work of the major physicists working on quantum mechanics is that there is some, very low, probability that the photon will "tunnel" right through the wall. There are some spooky things about how long it takes to transverse distance in such cases. But in the vast majority of cases the "spook" hits the wall and is either reflected or is absorbed. That is certainly what we expect from looking what happens when flashlights are shone on black walls, white walls, mirrors, etc.

If the "spook" does advance forward, whatever it is seems to act in some ways analogous to the behavior of the water wave. When it reaches a wall with slits in it, something happens -- especially if it is the part of the "spook" that is in a direct line from the emitter. We do not know what happens to the "ghost," but we know from long experience that we can very reliably make predictions by following the methods of Huygens (drawing a series of circles along the gap, joining them in a smooth curve, drawing another series of circles along that curve, and so on). From a single slit we will see a diffraction pattern develop, and from a double slit we will see an interference pattern develop. These patterns move out from the second surface of the barrier wall. What happens next seems to depend on what is in the way of the expanding "ghost". It has a certain probability of appearing as a "flash of light" at different places -- but it is not only the place that is important but what is located at that place. A vacuum will rarely if ever produce anything. (Maybe part of the explanation for the spontaneous appearance of matter-energy in pure vacuums has something to do with the occasional light photon passing by?) If there is a half-reflective mirror in the path of its expanding wave front then it may pass through or it may be reflected. If it encounters a solid wall, e.g. a white plaster surface, it will almost surely be manifested as a flash of light at one of its maxima. Walls of different composition do not seem to alter the interference patterns manifested on them extect to the extent that black patches absorb light better than do white patches. The amount of energy, i.e., the number of photons, that gets manifested seems to be invariant.

The operational definition of the word "to observe" seems always to boil down to our doing something to a "ghost" that will make it manifest in a highly localized thing that we call a photon. If we have one slit, we have a slightly different kind of "localization" -- a localization by elimination. If we have two slits and attempt to determine which slit a given photon has gone through, most if not all attempts have involved detectors that cause the "ghost" to manifest as a photon in the detector so that we can observe it. But at that point the trajectory of the photon emitted by the laser has terminated, and if any photon goes on from that point it is a photon that is emitted because of the delivery of the first photon. Moving that detector farther and farther away changes the size of the mouth of the detector that is needed to ensure trapping the "ghost," but if the "ghost" is trapped it is no long a "ghost" but has manifested as a photon. Even if the detector is moved to just in front of the target wall, it will still cause decoherence because it will change "ghost A" into a detected photon and then into "ghost B" and "ghost B" will be what hits the detection screen.P0M (talk) 20:29, 27 November 2007 (UTC)

One of the seeming implications is that, e.g. in the case of the interstellar experiment, an interference pattern would appear on a detector screen if the two images of a distant star were brought together optically, but if the detector screen were taken away and replaced with only two telescopes then the interference pattern would disappear. What is the event that has observed, that has caused a photon to betray its presence? It is not just the telescopes, but also the photographic plate in the telescope or the retina of the eye of somebody looking through the telescope. The photon is located at that point. The photon causes a chemical change in the photographic emulsion or in the cells of the retina. What that tells us is that the wave function, the "ghost" that was emitted from the distant light source has traveled on and on through space and has "taken a fancy to" one or the other telescope tubes. If the telescopes had been placed where they would intercept minima in the expanding interference pattern, then nothing would have happened. But does the fact that there are telescopes pointed at both referred images of the distant star mean that every "ghost" that approaches the observatory is somehow channeled into one or the other of those telescopes?

What would be the practical difference in the following two situations:


 * The apparatus is same as usual. A detection screen detects the usual interference pattern.


 * The apparatus is the same as before except that in place of the original detection screen there is a set of tightly packed pairs of telescopes, each member of each pair directed at one or the other slit. Does it matter that the expanding wave fronts coming out of the two slits are interfering as before but meet a wall of telescope lenses instead of a wall of plaster? The probability waves say, essentially, that there is an X percent chance of the photon manifesting at such and such a point. Does "manifesting" include going down own telescope tube rather than another? P0M (talk) 23:37, 27 November 2007 (UTC)

Ungrounded assertions
The association of Wheeler's thought experiment with distant sources of light sent on two paths by gravitational lensing seems not to be in the article by Wheeler that has been cited. What is the source, if any, for the binocular telescope stuff in Walker's own writing? P0M (talk) 19:48, 27 December 2007 (UTC)