X-Message-Number: 7939 Date: Tue, 25 Mar 97 01:42:26 From: Mike Perry <> Subject: Many-worlds and FSMs [Original Date sent: Sat, 22 Mar 1997 17:30:27] Peter Merel writes (#7910): >Mike Perry writes, > >>Many-worlds upholds locality. [... only ...] apparently >>non-local correlations are seen. (It might be objected too >>that many-worlds, even if it does preserve locality, >>introduces an additional complication--the splitting iself. >>So we have what are called "non-deterministic finite >>state machines"--still finite-state machines however, and >>also still "deterministic" in the sense that we always >>know what is going to happen.) > >Ah, but the point of saying that a finite space has a finite >number of states is to suggest that, in an infinite universe, >this state must cycle. This would be the case over infinite time (i.e. some states must be revisited). However, the finite-state-machine model is really not intended to apply to infinite time, if we are talking about immortal persons. A person is an FSM over finite stretches of time, but could also be a "growing automaton"--not limited to any fixed size if you take eter- nity into account. >I think that, even with only "apparently" non-local >correlations, this does not hold. Instead we must take the >whole universe into account in order to describe the >process of any finite subspace of it; "apparently" or not, >the state depends on doings outside the subspace. So >FSMs, non-deterministic or otherwise, lose their >applicability. With that I disagree. "Apparently non-local correlatons" are still local, despite some appearances to the contrary. It's certainly true, that what happens in one bounded region will depend (ultimately) on what happens in other places far removed, and it may look as if it happened faster than it really did, but that doesn't require that "FSM's lose their applicability." Another consideration may be that, even though FSMs are predictable, given their inputs, the inputs themselves need not be predictable. (I'm wondering if this could salvage the FSM theory even if we had to accept non-locality.) For what it's worth, I'll quote something Tipler says on the human as FSM (*The Physics of Immortality* p. 31, emphasis in the original): "According to quantum mechanics, any physical system is *exhaustively* described by its quantum state. That is, a system *is* its quantum state. The physicist Jacob Beken- stein showed that quantum systems--and according to physics, everything in sight is a quantum system--have only a finite number of states. In particular, a human being can be in one of 10^(10^45) states at most, and can undergo at most 4 x 10^53 changes of state per second. These numbers are of course enormous, and as a matter of fact I'm sure the actual numbers of states and changes are much less than these upper bounds. But these bounds are nevertheless finite, and firmly based on the central laws of quantum mechanics. They thus prove that a human being is a finite state machine, and *nothing but* a finite state machine." >>I haven't seen this article yet [on the quantum computer], >>but as I've said, progress with the quantum computer >>tends to support many-worlds. > >Um, so far as I know, it only knocks Copenhagen on the >head. That still leaves several interpretations in the >running. > >For my money Cramer's transactional interpretation is far >and away the pick of the litter, catering to the quantum >computation and related issues without requiring the >entire universe to split into two with every subatomic >event. While universe-splitting might be fun for science >fiction writers, I can't see that it's appealing when simpler >explanations are readily available. Many-worlds doesn't really "require the entire universe to split in two with every subatomic event" because the split only propagates at finite speed and in some cases can "heal"--though in general it doesn't. As for the quantum computer, it seems to provide a way of getting results that are hard to explain unless you assume many, many compu- tations are being done in parallel--a suggestion that many- worlds is true, i.e. the extra "worlds" really exist. (I'm not saying that all other interpretations are ruled out, however, but clearly the others have some explaining to do.) As for the Cramer transactional model, and the position of many- worlds relative to other interpretations of quantum reality, here is what Michael C. Price had to say (17 Feb. 1995, "Many-worlds FAQ," http://www.airtime.co.uk/users/station/m-worlds.faq): "Q3 What are the alternatives to many-worlds? "There is no other quantum theory, besides many-worlds, that is scientific, in the sense of providing a reductionist model of reality, and free of internal inconsistencies, that I am aware of. Briefly here are the defects of the most popu- lar alternatives: [snip] "5) Transactional model [C]. Explicitly non-local. An imaginative theory, based on the Feynman-Wheeler ab- sorber-emitter model of EM, in which advanced and re- tarded probability amplitudes combine into an atemporal "transaction" to form the Born probability density. It re- quires that the input and output states, as defined by an observer, act as emitters and absorbers respectively, but not any internal states (inside the "black box"), and, conse- quently, suffers from the familiar measurement problem of the Copenhagen interpretation. "If the internal states *did* act as emitters/absorbers then the wavefunction would collapse, for example, around one of the double slits (an internal state) in the double slit experiment, destroying the observed interference fringes. In transaction terminology a transaction would form between the first single slit and one of the double slits and another transaction would form between the same double slit and the point on the screen where the photon lands. This [is] never observed. "[C] John G Cramer _The transactional interpretation of quantum mechanics_ Reviews of Modern Physics Vol 58 #3 647-687 (1986)" Mike Perry http://www.alcor.org Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=7939