Dave-Platt%LADC@CISL-SERVICE-MULTICS.ARPA (Dave Platt) (12/05/85)
Re Niket Patwardhan's question about FTL information transfer, and Robert Firth's example of instantaneous action... there was some discussion of this general topic in Scientific American a couple of years ago. The principal of instantaneous action (or, more precisely, correlation of particle pairs) shows up in the violation of something called the Bell inequality, under certain conditions; it's due to the fact that the particle pair can be defined as a single quantum-mechanical waveform, and perturbing one aspect (the photon in Pittsburg) naturally results in a "simultaneous" change in the state of the other aspect. Unfortunately, there is a very good theoretical/practical reason why you can't use this effect to transfer information faster than light from the POV of an observer (at least, you can't transfer any useful information in this way). Let's do a thought experiment to take a look at why this is so. Be warned... I'm flying strictly based on memory here, and will probably oversimplify and/or accidentally speak falsehoods in what's about to follow... but it's how I understand the situation. Imagine that you want to transfer information between Trantor and Pittsburg. To do this, you'll have to set up a station midway between the two endpoints which is emitting a matched pair of particle beams (let's use electrons, for the sake of convenience). Naturally, each individual "beam" will have to be in some known, "reference" spin state (let's say that Pittsburg receives a "spin up" beam, and Trantor receives "spin down"), so that each party can detect the fact that the other party is sending information (by comparing the beam it receives against some reference standard to see whether it has been perturbed). If each beam is not in a well-known state, then the only way to tell whether a particular particle has been perturbed or not is to phone the person at the other end of the hookup (via a light-speed hookup of some sort) and say "My detector went *ping*; did yours?"... which defeats the whole purpose of the arrangement. How to produce such nice beams? You can certainly use well-known reactions (e.g. spontaneous pair production from gammas) to create pairs of electrons and positrons. However... although you can be sure that each pair of emitted particles is correlated, you have no way of ensuring that different sets of pairs are also correlated. For example, your first pair might come out spin-up/spin-down, and the second and third might come out spin-down/spin-up. So... the output of your correlated-pair source is not, itself, correlated over time; instead, each beam of pair-particles is quite random, and you don't have a "reference" state. In short, you can't tell whether any individual particle has been perturbed or not, because you can't tell what its original, unperturbed state was. So what? Why not sort the particles coming out of your pair-generator, and eliminate the ones that aren't in the state that you want? You could eliminate all spin-down particles from the Pittsburg beam, and remove the spin-up particles from the Trantor beam, and be left with two nice clean reference beams that could be modulated to your heart's content; they'd only be half as strong as the unsorted beams, but you'd just have to soup up the black hole you're using as a power source to compensate. Here's where Heisenberg bites us. The very act of sorting the particles according to their spin is enough to scramble their spin state! The paradox is that by figuring out what the spin state of the particle WAS when it entered your detection device, you must apply sufficient energy to the particle (in the very act of detection) that you can't tell what the state IS when it leaves the detector. You're left with the unhappy knowledge that the particles that made it through your filter were all spin-up... but that does you no good, as roughly half of them will have flipped over while passing through your spin detector. You are once again left with a randomly-correlated beam (albeit only half as strong). The sender on Trantor can modulate his/her beam, but the receiver in Pittsburg has no way of telling whether the beam has been modulated... a signal-to-noise ratio of 0! Well.... that's my understanding of the situation. I've used electrons as the particle in question mostly because that's how I remember the original Sci.Am. article being phrased, and I'm not entirely sure how the spin-scrambling-detection issue would apply if you use photon polarization as your signal carrier, rather than electron/positron spin (I guess I don't grok just what happens to a photon when you run it through a polarizing filter... I can only assume that the act of doing so must perturb the photon enough to scramble its state in an interesting way). Perhaps someone who reads SPACE, and is also a net.physics person can correct any misinformation I've dealt out above, and/or make things clearer for us all?
space@ucbvax.UUCP (12/08/85)
Hi Dave, still at LADC? How's Alderaan? Wing Wong? John (the pirate) Flint? I'm up in Oregon chasing a PhD. Try to establish a mail path to me. I'd love to know what's going on with you guys. Doug Pase (author of version 1 of A., and past LADC employee)