tmb@tardis.UUCP (01/27/86)
|||when one performs a measurement on a "mathematical object that represents |||the state of the universe", the outcome is only *probable*. That is, ||Nevertheless, abstractly, the theory only tells us how to calculate ||one future, one that incorporates both outcomes of the experiment. |Let me put it this way: how can a deterministic theory have it that the |very same initial conditions (particle, measuring device setup, etc.) lead |in different trials to the different outcomes experimenter-thinks-"E1!" |and experimenter-thinks-"E2!"? Let's talk about a variation on Schroedinger's cat, perhaps that will clarify the standpoints and end this silly discussion. Assume I put a cat into a box, together with a device that performs an experiment whose two outcomes are equally likely, according to quantum mechanics, and one outcome of which causes a cyanide capsule to break within the box and the cat to die. ALL THAT QUANTUM MECHANICS PREDICTS is that if you make an infinite number of such boxes and perform the experiment with all of them then, after the experiment, in one half of the boxes you will find dead cats, in the other half you will find live cats. Now, people are unhappy with this (I, for example, like cats). There are some who say that you didn't make the boxes perfectly identically in some parameter that isn't incorporated into our quantum mechanical theory of nature. They say that if you discovered this 'hidden parameter' and made the boxes identical in it as well, then the outcome of the experiment would be certain: either all cats die or all cats live. This is a valid position, although there is absolutely no evidence to support it. Other people suggest that the cat isn't really dead until someone observed that it is dead, i.e. that the decision about its death isn't made until someone observes it. In different terms, we cannot say what happens inside an isolated system until it isn't isolated anymore, because the outside observer has to 'open' it. This point has no relevance. A physical theory does not gain anything by assuming that the decision is made at the point of observation; in fact, it looses simplicity. It is like argueing about the existence of an ether with very special properties (namely such properties that relativity works) vs. the theory of relativity. One theory isn't any more powerful than the other, and we simply pick the one which is less verbose. Finally, some people suggest that instead of the cat dying or not dying, both happen, in 'branching universes'. This, again, is like postulating the existence of an ether: it doesn't add anything to our theory and just introduces verbiage. There is no experiment that I can perform after opening the box and finding a dead cat that shows that the cat is still alive in some 'other universe'. ALTOGETHER, quantum mechanics is a theory which makes predictions about the statistical outcome of a large number of identical experiments. It does not predict the outcome of individual experiments, and you may feel free, therefore, to refuse to call it a 'deterministic' theory. At the moment, quantum mechanics is simply the best theory we have. We might be able to find a theory which predicts the outcomes of individual experiments, or we might not. There is no way to decide that question by philosophising. Thomas.
dgary@ecsvax.UUCP (01/29/86)
A bit of clarification on the Copenhagen interpretation of quantum mechanics: In the referenced article Thomas Breuel describes the famous Schroedinger's cat gedanken experiment in which a cat in a box has its fate determined by a radioactive decay (which QM makes nondeterministic). Thomas talks about the hidden-variable and many-worlds views, and says also: >Other people suggest that the cat isn't really dead until someone observed >that it is dead, i.e. that the decision about its death isn't made until >someone observes it.... Yes, some hold this view, but a large number consider the cat's death, or even the decay itself, to be the "observation" that makes the wave equation collapse. Any irreversible quantum-mechanical event amounts to an observation in this view. I have it on good authority that this was Nils Bohr's own interpretation of quantum theory: I heard John Wheeler say this is what Bohr told him. >. . . This point has no relevance. A physical theory >does not gain anything by assuming that the decision is made at the point >of observation; in fact, it looses simplicity... It's certainly true that we can do physics without really understanding quantum mechanics on more fundamental level, but I don't think such philosophizing is pointless. The number of physicists who engage in it would suggest otherwise. It is entirely possible that speculations along these lines will ultimately lead to a better understanding of the universe (including ways of distinguishing among views by experiments we haven't thought of yet). Remember that Newton called physics "experimental philosophy." -- D Gary Grady Duke U Comp Center, Durham, NC 27706 (919) 684-3695 USENET: {seismo,decvax,ihnp4,akgua,etc.}!mcnc!ecsvax!dgary
torek@umich.UUCP (Paul V. Torek ) (01/30/86)
Had the person who posted the article to which I'm replying broken NETNEWS tradition and actually READ all the articles he was replying to, he might have known that we (Ken Rimey and I) were already aware of all the points he made. Mildly disgusted, --Paul Torek torek@umich
lambert@boring.uucp (Lambert Meertens) (01/30/86)
In article <10137@tardis.UUCP> tmb@tardis.UUCP writes: > Assume I put a cat into a box, together with a device that performs an > experiment whose two outcomes are equally likely, according to quantum > mechanics, and one outcome of which causes a cyanide capsule to break > within the box and the cat to die. > ALL THAT QUANTUM MECHANICS PREDICTS is that if you make an infinite number > of such boxes and perform the experiment with all of them then, after > the experiment, in one half of the boxes you will find dead cats, > in the other half you will find live cats. [...] > Other people suggest that the cat isn't really dead until someone observed > that it is dead, i.e. that the decision about its death isn't made until > someone observes it. [...] This point has no relevance. Now do the following: make two slits A and B close to each other in the box that can be opened or closed from the inside. Install equipment that monitors if the cat is alive (as used in intensive care) that causes slit A (initially open) to close, and B (initially closed) to open when the cat dies. After an elapsed time such that the probabilities of life vs. death are equal, let an electron be fired to the slits. It will pass through A or B, with equal probabilities. To an outside observer, to whom this devilish contraption is a black box, this is the same as a box with two open slits to which an electron is fired; if measurements concerning the passing of the electron are made, it is equally likely to pass through A as through B. If no such measurement is made but a photographic plate is placed behind the slits, we get an interference pattern because of self-interference of the "probability wave" passing through the slits, i.e., the superposition of the two waves corresponding to the two outlets. Now it seems to me that if this Russian roulette played on a cat results in the superposition of a live and a dead cat, then the output of the box is likewise the superposition of an electron passing through A, and the *same* electron passing through B. So we should see an interference pattern here as well. Is this indeed what QM predicts? (I honestly don't know; it's not my field. I have a gut feeling it does not, in which case I would really like to know why not.) This is an experiment that can be performed. (Instead of cats, dogs can be used, which, although not so stupid, are much more obnoxious:-) If the outcome shows indeed interference, then I don't see how other positions than "Schroedinger's cat (or Meertens' dog) is in a sense both dead and alive" can be maintained. On the other hand, if it does not show interference, that ambivalence would seem quite untenable to me. -- Lambert Meertens ...!{seismo,okstate,garfield,decvax,philabs}!lambert@mcvax.UUCP CWI (Centre for Mathematics and Computer Science), Amsterdam
gwyn@brl-tgr.UUCP (02/01/86)
> Had the person who posted the article to which I'm replying broken NETNEWS > tradition and actually READ all the articles he was replying to, he might have > known that we (Ken Rimey and I) were already aware of all the points he made. > > Mildly disgusted, > --Paul Torek torek@umich Perhaps Torek and Rimey should arrive at a consensus, then post that result for the rest of us to read, rather than carrying on such an extended private dialogue in the news group.
rimey@ernie.berkeley.edu.BERKELEY.EDU (Ken &) (02/02/86)
In article <6746@boring.UUCP> lambert@boring.UUCP (Lambert Meertens) writes: >In article <10137@tardis.UUCP> tmb@tardis.UUCP writes: >> Assume I put a cat into a box, ... > >Now do the following: make two slits A and B close to each other in the box >that can be opened or closed from the inside. Install equipment that >monitors if the cat is alive (as used in intensive care) that causes slit >A (initially open) to close, and B (initially closed) to open when the cat >dies. After an elapsed time such that the probabilities of life vs. death >are equal, let an electron be fired to the slits. It will pass through A >or B, with equal probabilities. Yes. Nothing profound here. >[Now consider two slits through] which an electron is >fired; if measurements concerning the passing of the electron are made, it >is equally likely to pass through A as through B. If no such measurement >is made but a photographic plate is placed behind the slits, we get an >interference pattern because of self-interference of the "probability wave" >passing through the slits, i.e., the superposition of the two waves >corresponding to the two outlets. Yes, one of our favorite thought experiments. >Now it seems to me that if this Russian roulette played on a cat > [ where the electron is constrained to go through slit A if > the cat is alive, and slit B if dead ] >results in >the superposition of a live and a dead cat, then the output of the box is >likewise the superposition of an electron passing through A, and the *same* >electron passing through B. So we should see an interference pattern here >as well. Is this indeed what QM predicts? No, according to quantum mechanics you will NOT get an interference pattern, because the electron's choice of slit is correlated with the state of the cat. If there is so much as a neutrino that would be going this way if the electron went through the left slit, but that way if the right, then you will not get the double-slit interference pattern. (This observation suffices to kill most amateur hidden-variables theories.) If there are two paths to exactly the same final state, you add their complex amplitudes (interference). But if these paths really lead to two different final states, the probability of getting to either is the sum of the probabilities of each (no interference). Ken Rimey
rimey@ernie.berkeley.edu.BERKELEY.EDU (Ken &) (02/05/86)
In article <2113@brl-tgr.ARPA> gwyn@brl-tgr.ARPA (Doug Gwyn <gwyn>) writes: >>Had the person who posted the article to which I'm replying broken NETNEWS >>tradition and actually READ all the articles he was replying to, he might have >>known that we (Ken Rimey and I) were already aware of all the points he made. >> >> Mildly disgusted, >> --Paul Torek torek@umich > >Perhaps Torek and Rimey should arrive at a consensus, >then post that result for the rest of us to read, >rather than carrying on such an extended private >dialogue in the news group. I am offended by your implication that Torek and I have been abusing this news group. I know, for example, that my side of that dialogue consisted of three postings spread over a month. Furthermore, what we wrote was not "private". This is clear from the discussion it initiated among half a dozen people. Your suggestion that Torek and I arrive at a consensus privately is strange. I also fail to understand why, if you are addressing both Torek and me, you quote a note from Torek to who-knows-who. Perhaps this recent quantum philosophy discussion is winding down, or perhaps it is just getting started, but in any case it doesn't look to me like it is swamping net.physics with junk mail. If you meant to say that it was, I wish you had said it more plainly. Ken Rimey
gwyn@brl-smoke.ARPA (Doug Gwyn ) (02/06/86)
>>> Assume I put a cat into a box, ... ... >If there are two paths to exactly the same final state, you add their >complex amplitudes (interference). But if these paths really lead to >two different final states, the probability of getting to either is >the sum of the probabilities of each (no interference). My opinion is that the one real mystery (non-intuitive fact) of quantum theory is how the "probability amplitude" business can be understood. If one can really comprehend how the "two-slit experiment" works on an internalized intuitive basis, then Schr"odinger's cat and the other usual paradoxes would be straightforward extensions of the same idea. (EPR seems to raise a different issue.) I have never found a good explanation (as opposed to description) of how the superposition of probability amplitudes work. The best arguments that I have heard amount to "it's self-consistent, and it works", which I might agree with and still not think I understand what's going on. I will say that most people giving such arguments have not shown that they understand probability theory very well, so one wonders whether a satisfying explanation might not be possible after all. This issue is at the top of my list of spare-time (ha!) projects.
gwyn@brl-smoke.ARPA (Doug Gwyn ) (02/07/86)
>>>Had the person who posted the article to which I'm replying broken NETNEWS >>>tradition and actually READ all the articles he was replying to, he might have >>>known that we (Ken Rimey and I) were already aware of all the points he made. >>> >>> Mildly disgusted, >>> --Paul Torek torek@umich >> >>Perhaps Torek and Rimey should arrive at a consensus, >>then post that result for the rest of us to read, >>rather than carrying on such an extended private >>dialogue in the news group. Ken Rimey was offended by this suggestion. My perception was that the discussion was going nowhere, and that it would be more useful to post conclusive points of agreement rather than all the false leads. If people actually find the discussion useful, then I withdraw the suggestion. (It does remind me a lot of the sort of discussion you get among a few college undergrads, though. I haven't heard anything new yet.) Of course, THIS discussion is not useful, either, and should be continued off-line if at all.
apak@oddjob.UUCP (Adrian Kent) (02/10/86)
In article <6746@boring.UUCP> lambert@boring.UUCP (Lambert Meertens) writes: [in context of Schrodinger cat thought-experiment] >Now do the following: make two slits A and B close to each other in the box >that can be opened or closed from the inside. Install equipment that >monitors if the cat is alive (as used in intensive care) that causes slit >A (initially open) to close, and B (initially closed) to open when the cat >dies. After an elapsed time such that the probabilities of life vs. death >are equal, let an electron be fired to the slits. It will pass through A >or B, with equal probabilities. > >To an outside observer, to whom this devilish contraption is a black box, >this is the same as a box with two open slits to which an electron is fired; No, it's not - and this is a crucial point. A state describing two open slits is not the same as a superposition of two states each describing one open slit. >if measurements concerning the passing of the electron are made, it >is equally likely to pass through A as through B. If no such measurement >is made but a photographic plate is placed behind the slits, we get an >interference pattern because of self-interference of the "probability wave" >passing through the slits, i.e., the superposition of the two waves >corresponding to the two outlets. > >Now it seems to me that if this Russian roulette played on a cat results in >the superposition of a live and a dead cat, then the output of the box is >likewise the superposition of an electron passing through A, and the *same* >electron passing through B. So we should see an interference pattern here >as well. Is this indeed what QM predicts? (I honestly don't know; it's >not my field. I have a gut feeling it does not, in which case I would >really like to know why not.) You're right. QM predicts no interference. The wavefunctions |dead, A open> and |alive, B open> are orthogonal. The electron wavefunction becomes correlated with them, giving a state "|dead, A open, electron went through A> + |alive, B open, electron went through B>", but these don't interfere. (This is phrased loosely but can be made precise.) > >If the outcome shows indeed interference, then I don't see how other positions >than "Schroedinger's cat is in a sense both dead and >alive" can be maintained. On the other hand, if it does not show >interference, that ambivalence would seem quite untenable to me. I think that's too strong. The standard conclusion about this type of experiment is that the many-worlds interpretation (cat 'both dead and alive') and the collapsing-wave-packet picture (cat either dead or alive) make (for all practical purposes) the same experimental predictions. The problem (as I see it) is that a collapsing wave-packet picture requires a theory underlying QM (which hasn't been found), while the many-worlds picture isn't remotely plausible as a fundamental description of nature. (Lots of physicists disagree.) Adrian Kent > Lambert Meertens