Causality, local realism, and all that
john baez
publicly since they come up often. In fact, I'm thinking of compiling
some sort of coherent introduction to quantum quandaries in my copious
and ever-growing spare time. So if anyone has saved any posts of mine
on this subject that they found enlightening (e.g., my recent
divagations on pure vs mixed states), please send me a copy.
He writes:
>from what i understood, you're saying that causality is the restriction that
>no influence can travel at FTL, while local realism is the position that
>parts of the system can be considered independent of the whole system.
>i'm pretty sure i'm not expressing either of these rigorously enough.
>basically the EPR experiment + confirmations showed that one of these
>must go, and Bohr chose to lose local realism.
That's basically right. Let me repeat, perhaps a bit more clearly, just
so everyone in the world hears, the distinction I was trying to make
between causality and local realism.
Causality says *roughly* that what we decide to do at point A of
spacetime cannot affect what occurs at point B unless light or something
slower can get from A to B. This is really too rough to be of much use
in *mathematical* physics, and besides, if one doesn't interpret these
words correctly one might come away from the Aspect experiment thinking
that causality has been violated.
In physics generally the most basic concepts are "state" and "observable".
Roughly, states represent what we know about how things are,
while observables represent what we can measure about how things are. I
emphasize that states and observables are mathematical objects, part of
a model (theory) of the universe or some subsystem. In different
theories states and observables will look pretty different. All we need
to insist upon is that a given state S and a given observable O combine
to give a probability distribution on the real line, called the
probability distribution for O in the state S.
In relativity we can speak about the state of the whole universe, which
allows us to get the probability distribution of *any* observable. We
can also speak about the state in some region of spacetime, which only
allows us to get probability distributions for observables that can be
measured by experiments done in that region. Given the state of the
whole universe, we can "restrict" it to any region and get the state in
that region.
Okay, here's the definition of causality. Say we know the state S in
some region R of spacetime. And say that the observable O can be
measured by an experiment done in a region R' of spacetime. And say
that R' is in the "causal shadow" of the region R - in other words,
every path moving backwards in time starting from R' and moving slower
or equal to lightspeed hits R:
----------- R'
/
/<-- typical path
/
----------------------- R
Here time is the vertical axis as usual and diagonal lines correspond to
lightspeed.
This is the precise way of saying that no influence can travel at faster
than the speed of light.
Now for local realism. This says simply the following. Suppose we have
two regions R and R' of spacetime. The union R U R' is another region
containing both of these (and the smallest such). Suppose we know the
state in the region R and the state in the region R'. Local realism
says that this is enough to determine the state in the region R U R'.
>are you using local
>realism to be roughly synonymous with hidden variables?
No. There are actually a bunch of hidden variables theories, but
most of them refer to some sort of way of embedding a quantum theory
in a classical theory. Since classical theories act a lot differently from
quantum theories, one must play some sneaky tricks to achieve this,
hence the word "hidden." Roughly speaking, Bell showed that while there
may be hidden variable theories that mimic quantum theory, there can't
be local realistic ones.
>what i'm really interested in is one of the details of the experiment which
>shows that one or the other of these must be invalid -- not the interpretation
>of the experimental results. the gist of what aspect seemed to be getting
>at was that he was defining "non-local" to mean that the act of measurement
>would produce those wacky correlations at another point.
Well, let's not worry about what "produces" the correlations; since we
can't observe the correlations or lack of correlations without doing a
measurement it's not very meaningful to worry about whether they were
there beforehand. The non-locality just means the failure of local
realism as defined above.
>"separable" seemed
>to be the added contraint that those correlations were occuring at
>points which were outside of each others light cones... and what i'm
>really interested in, was if the aspect experiment was designed to isolate
>the acts of measurement from each other to insure that the points were
>space-like seperated.
Paul Budnik would be glad to provide you with references to a paper that
argues that the experiment was not done carefully enough to ensure this.
I haven't gone through this stuff myself, and one would need to be a
good experimentalist to make sure things were (or weren't) really
functioning as advertised.
Paul Budnik uunet!mtnmath!paul
>[...] Causality says *roughly* that what we decide to do at point A of
> spacetime cannot affect what occurs at point B unless light or something
> slower can get from A to B. This is really too rough to be of much use
> in *mathematical* physics, and besides, if one doesn't interpret these
> words correctly one might come away from the Aspect experiment thinking
> that causality has been violated.
There is no question of causality violation in Aspect's experiments because
the time delays were not measured. In the absence of such a measurement one
simply does not know if there were superluminal effects.
However quantum mechanics does predict that causality as you just loosely
defined it will be violated in these experiments. Bell and Eberhard proved
that manipulating an experimental apparatus at one location can superluminally
effect detections at a distant location if QM is true.
> [...] Okay, here's the definition of causality. Say we know the state S in
> some region R of spacetime. And say that the observable O can be
> measured by an experiment done in a region R' of spacetime. And say
> that R' is in the "causal shadow" of the region R - in other words,
> every path moving backwards in time starting from R' and moving slower
> or equal to lightspeed hits R:
>
> ----------- R'
> /
> /<-- typical path
> /
> ----------------------- R
>
> Here time is the vertical axis as usual and diagonal lines correspond to
> lightspeed.
>
> This is the precise way of saying that no influence can travel at faster
> than the speed of light.
I think you set up the groundwork but then forgot to state the definition.
I assume the definition you intended to state is that the *probability*
of observing some value for the observable O cannot be influenced by events
outside of R. The problem with this definition is that the *value* for
the observable O can be influenced without influencing the probability
that it will take on a particular value. The absolute probability of an
observation is not changed by distant manipulations but the *conditional*
probability (conditional on a second detection) is changed. This
counterintuitive kind of influence happens in quantum mechanics and can
be detected experimentally because two space like separated observations
are predicted to be correlated in a special way. This correlation requires
that manipulation of a distant experimental apparatus must be able to
influence the detection of a distant event or the correlation could not
be as great as it is. This is what Bell and Eberhard proved and I
consider it most unfortunate that so many competent physicists fail to
understand this simple and important result.
>
> Now for local realism. This says simply the following. Suppose we have
> two regions R and R' of spacetime. The union R U R' is another region
> containing both of these (and the smallest such). Suppose we know the
> state in the region R and the state in the region R'. Local realism
> says that this is enough to determine the state in the region R U R'.
>
> >are you using local
> >realism to be roughly synonymous with hidden variables?
>
> No. There are actually a bunch of hidden variables theories, but
> most of them refer to some sort of way of embedding a quantum theory
> in a classical theory. Since classical theories act a lot differently from
> quantum theories, one must play some sneaky tricks to achieve this,
> hence the word "hidden." Roughly speaking, Bell showed that while there
> may be hidden variable theories that mimic quantum theory, there can't
> be local realistic ones.
And Eberhard showed that this is because QM is not a local theory. Only
a theory that has superluminal causal influences can mimic QM because the
mathematics of QM has nonlocal operations that cannot be factored out.
The attempt to distinguish between local effects and local state is a
rationalization. It is the manipulation of an experimental
apparatus that changes the state. This has an experimentally detectable
effect. To deny that this is a causal effect is to redefine causality in
a way that contradicts our usual understanding of the term.
> >what i'm really interested in is one of the details of the experiment which
> >shows that one or the other of these must be invalid -- not the interpretation
> >of the experimental results. the gist of what aspect seemed to be getting
> >at was that he was defining "non-local" to mean that the act of measurement
> >would produce those wacky correlations at another point.
>
> Well, let's not worry about what "produces" the correlations; since we
> can't observe the correlations or lack of correlations without doing a
> measurement it's not very meaningful to worry about whether they were
> there beforehand.
That's a nonsequitur if I ever heard one. It is clear that the act of
measurement is an integral part of the process that creates the correlations.
Trying to understand what kind of physical mechanism might generate these
effects is a profoundly important question. Simply saying they come from the
measurement does not answer this question. We know the correlations occur.
There is a large body of experimental evidence on this. What we do not know
is if there is a space like separation between when the experimental
apparatus changes and when this affects the probability of joint detections.
No experiments have ruled out the possibility that there is physical local
mechanism that is responsible for the correlations. Experiments may ultimately
prove the delay between when the apparatus changes and the probability of joint
detections changes is long enough to be caused by a local mechanism.
If so understanding this mechanism will to say the least be an interresting
physics problem.
> [...] Paul Budnik would be glad to provide you with references to a paper that
> argues that the experiment was not done carefully enough to ensure this.
> I haven't gone through this stuff myself, and one would need to be a
> good experimentalist to make sure things were (or weren't) really
> functioning as advertised.
Even a mathematical physicist like yourself can readily understand that
you cannot prove space-like separations occurred without measuring both time
and distance. You can also readily understand that one cannot
base estimates of time on a classical model of a particle traversing a
polarizer and then being detected. You must directly measure the time
delay between when the apparatus changes and this has a physical effect.
The problem with Aspect's experiment is not some subtlety of experimental
method. In fact the error is so gross that had the result been in conflict
with QM I am sure it never would have got past the reviewers of his paper.
It is not just politicians that have a tendency to believe what they want
to believe. I am not saying his paper should not have been published but
it should have included a big disclaimer about what conclusions could be
drawn from it.
Paul Budnik
William Unruh
>Now for local realism. This says simply the following. Suppose we have
>two regions R and R' of spacetime. The union R U R' is another region
>containing both of these (and the smallest such). Suppose we know the
>state in the region R and the state in the region R'. Local realism
>says that this is enough to determine the state in the region R U R'.
Uh, you must mean by "know" Know completely and exactly-know in such a
way that there is absolutely no uncertainty left so that from out
knowledge we can predict everything about the system in R or R'. Becasue
even in Statistical classical thoeries (which are surely ralistic local
thoeries) knowledge of the state (in the statistical sense) in R and R' is not
sufficient to know the state in RUR'.
William Unruh
>However quantum mechanics does predict that causality as you just loosely
>defined it will be violated in these experiments. Bell and Eberhard proved
>that manipulating an experimental apparatus at one location can superluminally
>effect detections at a distant location if QM is true.
Oh dear, one spends time posting rebuttals to the net, and noone
listens. Eberhard did not prove superliminal anything. He may have
claimed to but you shouldn;t believe everyone's claims. Look at the
proof (or lack thereof).
john baez
>ba...@guitar.ucr.edu (john baez) writes:
>
>>Now for local realism. This says simply the following. Suppose we have
>>two regions R and R' of spacetime. The union R U R' is another region
>>containing both of these (and the smallest such). Suppose we know the
>>state in the region R and the state in the region R'. Local realism
>>says that this is enough to determine the state in the region R U R'.
>Uh, you must mean by "know" Know completely and exactly-know in such a
>way that there is absolutely no uncertainty left so that from out
>knowledge we can predict everything about the system in R or R'. Becasue
>even in Statistical classical theories (which are surely realistic local
>theories) knowledge of the state (in the statistical sense) in R and R' is not
>sufficient to know the state in RUR'.
You're right, I screwed up. I had originally written something that
began with a vague explanation of pure and mixed states, and then
for some reason thought I could avoid this, but one can't. I really
should have said:
Suppose we know
state in the region R and the state in the region R'. Local realism
says that this is enough to determine the state in the region R U R', if
that state is pure.
I repeat my request for anyone who might have saved my more detailed
explanation of pure and mixed states to email a copy back to me.
I want to compile an all-purpose quantum quandaries file.
john baez
>Oh dear, one spends time posting rebuttals to the net, and noone
>listens. Eberhard did not prove superliminal anything. He may have
>claimed to but you shouldn;t believe everyone's claims. Look at the
>proof (or lack thereof).
People listen, just not the people you are rebutting. :-) Paul will
never change his opinion about this; I tried my best, and you tried even
better.
Paul Budnik uunet!mtnmath!paul
>
> Oh dear, one spends time posting rebuttals to the net, and noone
> listens. Eberhard did not prove superliminal anything. He may have
> claimed to but you shouldn;t believe everyone's claims. Look at the
> proof (or lack thereof).
>
I do not believe everything I read and have demonstrated that by my postings.
I think I have made a small contribution to the general knowledge in
sci.physics by showing that wide spread and widely published beliefs
about Aspect's experiment are mistaken.
I have carefully read Eberhard's papers and find them convincing on this
point.
Posting to the net is not generally considered a way of establishing
scientific truth. If you think you have an effective rebuttal to Eberhard's
arguments you should publish your results. If your arguments are true
they are important. I do not have the time to rebut complex technical
arguments that do not have much credibility in my mind and that seem to
be awkwardly stated as yours were. Referee's on the other hand are compelled
to come up with at least some flaw or problem in a paper.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
John, When have you ever posted a rebuttal to Eberhard's arguments? I doubt
that you have read Eberhard. Did you carefully read Unruh's arguments
and do you think they are sound? (His writing is just as precise as his
typing and spelling.)
When are you going to answer Bell's objections to Everett that I have
posted several times.
I do not think you have good answers to Bell and I do not think
you understand Bell's results about locality and Eberhards refinement of
them. Until you understand this material I do not think you are in any
postion to argue these points.
I am more than willing to listen to convincing arguments. I have demonstrated
that I change my mind when I am proven wrong as happened with Keith Ramsay's
proof.
Pehaps you are the one with a closed mind.
Paul Budnik
Aephraim M. Steinberg
Paul Budnik uunet!mtnmath!paul <pa...@mtnmath.UUCP> wrote:
>In article <28due9$4...@nntp.ucs.ubc.ca>, un...@physics.ubc.ca (William Unruh) writes:
>>
>> listens. Eberhard did not prove superliminal anything. He may have
>> claimed to but you shouldn;t believe everyone's claims. Look at the
>> proof (or lack thereof).
>>
>point.
>
>Posting to the net is not generally considered a way of establishing
>scientific truth. If you think you have an effective rebuttal to Eberhard's
>arguments you should publish your results. If your arguments are true
>they are important. I do not have the time to rebut complex technical
(Given that I don't even have time to _read_ all of your postings, I find
this last point hard to believe, but we'll let it slide.)
I'm not in the mood to jump back into one of these discussions (and haven't
even been following this thread for the most part), but I would like to
defend Eberhard from slander. I've worked with him, and I _assure_ you
that he does not believe in any superluminal effects at all like your
description in an earlier post ("by adjusting parameters at one location,
one can superluminally affect distant experiments" or words to that effect--
please correct & forgive me if I'm simply misreading your prose). I
believe he worked on the question of whether or not Aspect's experiment
even showed that the "spooky actions at a distance" were instantaneous, and
may have come up with some number like 2.5c as a lower bound the experiment
had placed on their speed. But he certainly does not believe in the poss-
ibility of communicating superluminally.
In time-honored quantum-mechanical tradition, he believes in parameter-
independence but not outcome-independence: that is, your experiment's
result might depend nonlocally on the outcome of my own experiment, but
not on the settings of my analyzers. Since it is only the latter I can
control, I cannot send you any signal.
I'm trying to figure out how the discussion has reached to this point,
since I know that Eberhard & Ross's _disproof_ of the possibility of
communicating superluminally has been discussed here in the past.
--
Aephraim M. Steinberg | "If the human brain were simple
UCB Physics | enough for us to understand, we
aeph...@physics.berkeley.edu | would be too simple to understand
| it." -- anonymous
Paul Budnik uunet!mtnmath!paul
[...]
> I'm not in the mood to jump back into one of these discussions (and haven't
> even been following this thread for the most part), but I would like to
> defend Eberhard from slander. I've worked with him, and I _assure_ you
> that he does not believe in any superluminal effects at all like your
> description in an earlier post ("by adjusting parameters at one location,
> one can superluminally affect distant experiments" or words to that effect--
> please correct & forgive me if I'm simply misreading your prose).
Let me quote Eberhard:
Let us consider two measuring apparata located in two different places
A and B. There is a knob *a* on apparatus A and a knob *b* on apparatus B.
Since A and B are separated in space, it is natural to think what will
happen at A is independent of the setting of knob *b* and *vice versa*.
The principles of relativity seem to impose this point of view if the
time at which the knobs are set and the time of the measurements are
so close that, in the time laps, no light signal can travel from A to
B and *vice versa*. Then, no signal can inform a measurement apparatus
of what the knob setting on the other is. However, there are cases in
which the predictions of quantum theory make that independence assumption
impossible. If quantum theory is true, there are cases in which the
results of the measurements A will depend on the setting of the knob
*b* and/or the results of the measurements in B will depend on the
setting of the knob *a*.
-- Bell's Theorem without Hidden Variables, Il Nuvuo Cimento, V38 B, N. 1.
> I believe he worked on the question of whether or not Aspect's experiment
> even showed that the "spooky actions at a distance" were instantaneous, and
> may have come up with some number like 2.5c as a lower bound the experiment
> had placed on their speed. But he certainly does not believe in the poss-
> ibility of communicating superluminally.
Nor do I but that is not what I said.
The lower bound in the speed of these effects in Aspect's experiment
is less than C. See the following references:
J. D. Franson, Physical Review D, pgs. 2529-2532, Vol. 31, No. 10, May 1985.
A. Leggett, Foundations of Physics, V 17, p. 875. 1987
The distinction between superluminal effects and superluminal
communication is a point that many physicists are confused on
including obviously yourself. It is not possible to use this
affect for superluminal communication. Information about the setting of one
knob is transferred superluminally but it is in effect encrypted with
quantum uncertainty. One can prove that information was transferred
superluminally but one can only do so by comparing two distant measurements
using standard communication techniques. This result is only possible in
a theory like quantum mechanics that claims probabilities are irreducible.
In classical theories one cannot get superluminal effects without the
theoretical possibility of superluminal communication.
> In time-honored quantum-mechanical tradition, he believes in parameter-
> independence but not outcome-independence: that is, your experiment's
> result might depend nonlocally on the outcome of my own experiment, but
> not on the settings of my analyzers. Since it is only the latter I can
> control, I cannot send you any signal.
This contradicts well known, well understood and elementary results.
Perhaps you need some additional discussions with Eberhard on this point.
What you are thinking about I suspect is that the probability of detection
is independent of settings of distant parameters. This is correct. However
there are other experimentally detectable results. The joint
probability of detection is not independent of distant settings and the
predictions of quantum mechanics about this joint probability require the
superluminal transfer of information (as that term was defined by Shannon).
The argument that these effects do not support superluminal communication
is more subtle then saying that distant parameter settings have no effect
on distant results. That statement is false if quantum mechanics is true.
> I'm trying to figure out how the discussion has reached to this point,
> since I know that Eberhard & Ross's _disproof_ of the possibility of
> communicating superluminally has been discussed here in the past.
You have apparently read few of my postings. I have often said
superluminal communication is not possible. I use to religiously
add it every time I talked about superluminal effects so there would be
no misunderstanding. Perhaps its just as well that I did not this time
since it caught your attention. Maybe you will learn something from this
exchange.
Paul Budnik
lam...@cs.washington.edu
>you cannot prove space-like separations occurred without measuring both time
>and distance. You can also readily understand that one cannot
>base estimates of time on a classical model of a particle traversing a
>polarizer and then being detected. You must directly measure the time
>delay between when the apparatus changes and this has a physical effect.
>The problem with Aspect's experiment is not some subtlety of experimental
>method. In fact the error is so gross that had the result been in conflict
>with QM I am sure it never would have got past the reviewers of his paper.
>It is not just politicians that have a tendency to believe what they want
>to believe. I am not saying his paper should not have been published but
>it should have included a big disclaimer about what conclusions could be
>drawn from it.
so aspects experiment did not address seperability as advertised? have
experiments been done which have?
--
Lamont Granquist drugz: ftp.u.washington.edu:/pub/user-supported/alt.drugs
lam...@cs.washington.edu personal: !finger lam...@cs.washington.edu | more
"Conservative: n. One who admires radicals centuries after they're dead."
lam...@cs.washington.edu
>Suppose we know
>state in the region R and the state in the region R'. Local realism
>says that this is enough to determine the state in the region R U R', if
>that state is pure.
Okay, what i gathered from this lecture i went to was that bohr's position
was that the entire measurement apparatus must be considered holistically
and that if you started to divide it up you'd start reaching wrong
conclusions. And what i gathered was that Bell showed that the results
of the EPR experiment supported Bohr. Is this the practical application
of violating local realism is terms i can grasp?
lam...@cs.washington.edu
>even been following this thread for the most part), but I would like to
Did Eberhard attempt to prove that the results of an EPRish experiment
showed superluminal influence? *influence* here being that thing which
Einstien called "spooky action at a distance" and which is not synonymous
with (or useful for) communication. And the debate here is that some
believe that all is observed is a correlation with no communication or
even influence occuring?
And where was this published?
john baez
>ba...@guitar.ucr.edu (john baez) writes:
>>Suppose we know
>>state in the region R and the state in the region R'. Local realism
>>says that this is enough to determine the state in the region R U R', if
>>that state is pure.
>Okay, what i gathered from this lecture i went to was that bohr's position
>was that the entire measurement apparatus must be considered holistically
>and that if you started to divide it up you'd start reaching wrong
>conclusions. And what i gathered was that Bell showed that the results
>of the EPR experiment supported Bohr. Is this the practical application
>of violating local realism is terms i can grasp?
Sorry, the above is too fuzzy for me to grasp. The "practical
applications" of violations of local realism that might be easiest to
grasp are Bennett's work on quantum cryptography and quantum
teleportation (the last one was given the name it has only to make
people go "ooh" and "ah," but it's still pretty cool). I'm terribly
busy now but maybe some kind soul who saved our earlier discussions of
these phenomena can email 'em to you, or at least dig up a reference, or
post an explanation.
Paul Budnik uunet!mtnmath!paul
> aephraim@physics5 (Aephraim M. Steinberg) writes:
> >I'm not in the mood to jump back into one of these discussions (and haven't
> >even been following this thread for the most part), but I would like to
> >defend Eberhard from slander...
>
> Did Eberhard attempt to prove that the results of an EPRish experiment
> showed superluminal influence? *influence* here being that thing which
> Einstien called "spooky action at a distance" and which is not synonymous
> with (or useful for) communication. And the debate here is that some
> believe that all is observed is a correlation with no communication or
> even influence occuring?
>
> And where was this published?
Bell and Eberhard proved there is influence or action at a distance if QM
is true. The experimental jury has yet to decide the issue. That
influence cannot be used for superluminal communcation.
"On the Einstein Podolosky Rosen Paradox", John S. Bell, Physics, 1,
195-200 (1964).
"Bell's Theorem without Hidden Variables", P. H. Eberhard, Il Nuovo Cimento,
38 B 1, p 75, (1977).
"Bell's Theorem and the Different Concepts of Locality", P. H. Eberhard,
Il Nuovo Cimento 46 B, p 392, (1978).
Paul Budnik
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <28e2t2$e...@galaxy.ucr.edu>, ba...@guitar.ucr.edu (john baez) writes:
> > In article <28due9$4...@nntp.ucs.ubc.ca> un...@physics.ubc.ca (William Unruh) writes:
> >
> > >Oh dear, one spends time posting rebuttals to the net, and noone
> > >listens. Eberhard did not prove superliminal anything. He may have
> > >claimed to but you shouldn;t believe everyone's claims. Look at the
> > >proof (or lack thereof).
> >
> > People listen, just not the people you are rebutting. :-) Paul will
> > never change his opinion about this; I tried my best, and you tried even
> > better.
>
> John, When have you ever posted a rebuttal to Eberhard's arguments? I doubt
> that you have read Eberhard. Did you carefully read Unruh's arguments
> and do you think they are sound? (His writing is just as precise as his
> typing and spelling.)
>
What has been posted by several individuals, including me several
times, is an explanation of how in a certain sense QM is totally local
with the Everett interpertation. More specifically there is a perfectly
good interpertation of QM that has NO superluminal casual influences,
contrary to what you have said several times.
> When are you going to answer Bell's objections to Everett that I have
> posted several times.
>
What objections? Please post them and, even though I am not called
John, I will respond to them. Perhaps the fact that John knows that
your objections have been delt with in detail is part of the reason
that he does not want to spend a lot of time writing out responses.
> I do not think you have good answers to Bell and I do not think
> you understand Bell's results about locality and Eberhards refinement of
> them. Until you understand this material I do not think you are in any
> postion to argue these points.
>
You are mistaken about what has been proven since you are saying that
there have to be superluminal casual influences in ANY interpertation
of QM. The easy proof is by a counterexample. Specifically the Everett
interpertation. Now it is possible to not accept the Everett
interpertation as what is going on, but it certainly does show that
your statement is wrong.
Given that fact I think that you should be careful in trying to say
that others do not understand the material since you certainly have
some misconceptions about it.
> I am more than willing to listen to convincing arguments. I have demonstrated
> that I change my mind when I am proven wrong as happened with Keith Ramsay's
> proof.
>
Then could you respond to what I just said? THis point came up a while
ago, and I gave the same explanation to you, and you do not seem to
have understood it.
> Pehaps you are the one with a closed mind.
>
Anything is possible.
Ben Tilly
lam...@cs.washington.edu
>>Okay, what i gathered from this lecture i went to was that bohr's position
>>was that the entire measurement apparatus must be considered holistically
>>and that if you started to divide it up you'd start reaching wrong
>>conclusions. And what i gathered was that Bell showed that the results
>>of the EPR experiment supported Bohr. Is this the practical application
>>of violating local realism is terms i can grasp?
>
let me try to rephrase that.
what i'm looking for is a physical application of violating local
realism. and what i'm thinking of is the results of the EPR experiment.
my understanding of what EPR argued was that since we can predict with
certainty what will be measured in the 2nd observation after making the
1st observation that there must be some aspect of reality representing it(i.e.
hidden variable) which is zooming across the room. and that this was
subsequently shown to not be the case (not unless you start adding
higher dimentions or Everett or other weirdness). so, the idea would
be that even though we know that the measurement will result in a spin
up photon, before we measure it, it isn't. thus, the entire experiment,
including the measurement apparatus, would have to be considered
holistically.
john baez
>what i'm looking for is a physical application of violating local
>realism. and what i'm thinking of is the results of the EPR experiment.
>my understanding of what EPR argued was that since we can predict with
>certainty what will be measured in the 2nd observation after making the
>1st observation that there must be some aspect of reality representing it(i.e.
>hidden variable) which is zooming across the room. and that this was
>subsequently shown to not be the case (not unless you start adding
>higher dimentions or Everett or other weirdness). so, the idea would
>be that even though we know that the measurement will result in a spin
>up photon, before we measure it, it isn't. thus, the entire experiment,
>including the measurement apparatus, would have to be considered
>holistically.
Hmm. The phrase "even though we know the measurement will result
in a spin up electron, before we measure it, it isn't" is really the
sort of thing that means nothing to me, I'm afraid. Worrying about
whether this or that observable "has a value before we measure it"
seems like mere monkeying with language since I don't know how to test
such things. Maybe you mean something operational (testable) but I
can't tell what.
Similarly, "the entire experiment must be considered holistically"
doesn't mean anything to me. I don't know what it means, in concrete,
practical terms, to consider something holistically or not.
I have found it very necessary to concentrate on 1) rigorous mathematics
and 2) operational physics (i.e., statements along the lines of "if you
do thus and so, and measure thus and so in such a way, you will get a
reading of <5 volts with probability .3") in order to make any sense of
quantum mechanics. All sorts of things like "Everett's branches" may
seem very metaphysical, but for me at least they are merely a convenient
shorthand for statements of a mathematically/operationally precise
character.
I should also add, just in case you hadn't pondered it yet, that
it would be very naive to say that just because a
measurement at A tells us with certainty about what's going on at B,
something must zoom across the room when the measurement occurs. This
is something that goes on all the time in classical mechanics: take two
balls, one red and one blue, and randomly send one to Alpha Centauri and
the other to Sirius. The Sirian, upon seeing (say) a red ball,
immediately knows that a blue one will be seen on Centauri, many
light years. But of course no information is being transmitted faster
than light. (I say "of course," but really a definition of information
is needed to make this mean anything; I'm using the usual Shannon one.)
Anyway, if you are looking for a physical application of violating local
realism, probably Bennett's "quantum teleportation" paper is a nice one.
To quote myself,
Anyway, the paper is
Teleporting an unknown quantum state via dual classical and
Einstein-Podolsky-Rosen channels, Phys Rev Lett 70 (1993), 1895
by Bennett and a slew of other folks. For starters, let me note that I
think the language in Bennett's abstract is deliberately provocative.
The article itself is much more careful. For one, it emphasizes that
this new "quantum teleportation" trick can NOT occur over spacelike
intervals. From my few conversations with Bennett and reading his work
on quantum cryptography and logical depth, it's clear that he's a sharp
character. Anyone interested in the bizarre tricks one can REALLY do with
quanta (as opposed to superluminal ledgermain) should read his stuff.
Here's the rough idea of Bennett's paper - I'm too lazy to fill in the
details. Say Alice has a spin-1/2 particle whose state |phi> she
wants to communicate to Bob. Well, she could just give it to him. She
*cannot* make a copy of it and send that to him, since one cannot "clone
a quantum" - this beautifully simple result can be found in
Wooters and Zurek, Nature 299 (1982) 802.
Crudely speaking, classical information can be perfectly copied but not
quantum information. This is a crappy way of speaking, though, since
it's utterly unclear what "classical information" is as opposed to
"quantum information" - that's one of the things I dislike about the
abstract.
A more clever way for Alice to pass on |phi> to Bob is called "spin
exchange." She can take an ancillary system (Latin for extra) in a
known state |a_0> and let it interact with |phi> such that
|phi> x |a_0> -> |phi_0> x |a>
where |phi_0> is a known standard state of the spin-1/2 particle.
Now the information about |phi> is in |a> and she can hand the ancillary
system to Bob, who can reverse the process to get |phi>. Note that
Alice has lost all information about |phi>.
"Teleportation" is a more clever trick which I will not fully describe.
Let's suppose |phi> is a spin-1/2 particle and call it |phi_1>.
Alice prepares two more spin-1/2 particles, 2 and 3, in an EPR-type state
(|up_2, down_3> - |down_2, up_3>)/sqrt(2)
(supressing normalization). Alice sends Bob particle 3 and keeps 2.
Then Alice does a measurement on particles 1 and 2 - a measurement of an
observable that has as eigenvectors (corresponding to distinct
eigenvalues) the orthonormal states
(|up_1, down_2> - |down_1, up_2>)/sqrt(2)
(|down_1, up_2> - |up_1, down_2>)/sqrt(2)
(|up_1, down_2> + |down_1, up_2>)/sqrt(2)
(|down_1, up_2> + |up_1, down_2>)/sqrt(2)
Now, after Alice announces this result to Bob, Bob can do something to
particle 3 which enables him to figure out the original state phi_1!
Spoooooky! But not against any of the usual laws of quantum mechanics.
Note: I only skimmed this so anyone seriously interested in getting the facts
right should either try figure out what Bob does, or read the paper.
Paul Budnik uunet!mtnmath!paul
> [...] I should also add, just in case you hadn't pondered it yet, that
> it would be very naive to say that just because a
> measurement at A tells us with certainty about what's going on at B,
> something must zoom across the room when the measurement occurs. This
> is something that goes on all the time in classical mechanics: take two
> balls, one red and one blue, and randomly send one to Alpha Centauri and
> the other to Sirius. The Sirian, upon seeing (say) a red ball,
> immediately knows that a blue one will be seen on Centauri, many
> light years. But of course no information is being transmitted faster
> than light. (I say "of course," but really a definition of information
> is needed to make this mean anything; I'm using the usual Shannon one.)
> [...]
The point of Bell and Eberhard's proofs is that the correlated predictions
of QM do require superluminal transfer of information in the Shannon sense.
If your carefully apply Shannon's definition to these experiments
this is the inescapable conclusion. We know what the mechanisms are in
the mathematics of QM that transfers this information in the model. We
do not know if there is a superluminal transfer of information
because we do not know if these predictions are correct.
This is a point on which most physicists who write on this subject
make false inaccurate claims. Shannon was clever in constructing a definition
that applys to any conceivable circumstances and it applys to this one. The
correlations do not just involve the particle state. They involve the
relationship of the state to experimentally controllable parameters.
Information about those parameters must be transferred in at least one
direction superluminally to reproduce these predictions.
It is only because QM claims that probabilities are irreducible that we
are able to have superluminal information transfer without superluminal
communication. In effect the information transferred is encrypted with
quantum uncertainty. We can prove it was transferred and verify when the
transfer occurred but we can only do so afterwards by comparing results
from both locations.
Your devotion to mathematical precision is apparently weaker then your refusal
to face the problems in QM.
Paul Budnik
Daryl McCullough
>I have found it very necessary to concentrate on 1) rigorous mathematics
>and 2) operational physics (i.e., statements along the lines of "if you
>do thus and so, and measure thus and so in such a way, you will get a
>reading of <5 volts with probability .3") in order to make any sense of
>quantum mechanics. All sorts of things like "Everett's branches" may
>seem very metaphysical, but for me at least they are merely a convenient
>shorthand for statements of a mathematically/operationally precise
>character.
How is "you will get a reading of <5 volts with probability .3"
operational? Operationally, you either get <5 volts, or you don't.
Probability is not observable, only relative frequency is. Of course,
the two are connected by the various laws of large numbers, which
imply that as an experiment is performed again and again, the relative
frequencies for outcomes approaches the probability...except for a set
of histories with probability zero!
So, in my opinion, probabilistic theories are never completely
operationalized---they have no predictive value unless you assume that
your history happens to be typical: in the limit, everything happens
according to its probability. So far, it seems to be a safe assumption;
we don't have many instances of a million sequential coin tosses that
all end up "heads".
>I should also add, just in case you hadn't pondered it yet, that
>it would be very naive to say that just because a
>measurement at A tells us with certainty about what's going on at B,
>something must zoom across the room when the measurement occurs. This
>is something that goes on all the time in classical mechanics: take two
>balls, one red and one blue, and randomly send one to Alpha Centauri and
>the other to Sirius. The Sirian, upon seeing (say) a red ball,
>immediately knows that a blue one will be seen on Centauri, many
>light years. But of course no information is being transmitted faster
>than light. (I say "of course," but really a definition of information
>is needed to make this mean anything; I'm using the usual Shannon one.)
I think that the original poster was simply being sloppy, what he
probably meant was that *either* the information zoomed from A to B,
*or* they both shared the information from the start. In the case you
give, the information as to whether a red ball is sent to Sirius is
determined in the intersection of the backwards light cones of the
correlated events of 1. arrival of one ball at Sirius, and 2. arrival
of another ball at Alpha Centauri.
Daryl McCullough
ORA Corp.
Ithaca, NY
john baez
>ba...@guitar.ucr.edu (john baez) writes:
>>I have found it very necessary to concentrate on 1) rigorous mathematics
>>and 2) operational physics (i.e., statements along the lines of "if you
>>do thus and so, and measure thus and so in such a way, you will get a
>>reading of <5 volts with probability .3") in order to make any sense of
>>quantum mechanics. All sorts of things like "Everett's branches" may
>>seem very metaphysical, but for me at least they are merely a convenient
>>shorthand for statements of a mathematically/operationally precise
>>character.
>How is "you will get a reading of <5 volts with probability .3"
>operational? Operationally, you either get <5 volts, or you don't.
>Probability is not observable, only relative frequency is. Of course,
>the two are connected by the various laws of large numbers, which
>imply that as an experiment is performed again and again, the relative
>frequencies for outcomes approaches the probability...except for a set
>of histories with probability zero!
True.
>So, in my opinion, probabilistic theories are never completely
>operationalized---they have no predictive value unless you assume that
>your history happens to be typical: in the limit, everything happens
>according to its probability.
I now think you are pushing your conclusion too far. Of course, this
depends on a choice of definition of "operational." I think the real
world is too messy to get anywhere with a notion of operational that
doesn't make kind of assumption you make, as well as other equally
outrageous assumptions. :-) For example, we can never do the same
experiment *in the same time and place* twice so unless we assume some
kind of translation-invariance even the notion of frequency - "do the
SAME experiment over and over and count how many times X happen" - is
dubious. Similarly and much more importantly, any real-world experiment
is subject to many weird external influences that one doesn't put in
ones model -- the tides, etc. -- so one tries to show that all the
effects one can think of are negligable, but one is never sure there
aren't other such effects.
So in other words I would go for a kind of "soft operationalism" that
fully admits there are many things that we cannot prove logically but
just seem to work:
>So far, it seems to be a safe assumption;
>we don't have many instances of a million sequential coin tosses that
>all end up "heads".
All I want is for my theorizing to give predictions in conjunction with
such "common-sense" assumptions.
Silver Omega
> It is only because QM claims that probabilities are irreducible that we
> are able to have superluminal information transfer without superluminal
> communication. In effect the information transferred is encrypted with
> quantum uncertainty. We can prove it was transferred and verify when the
> transfer occurred but we can only do so afterwards by comparing results
> from both locations.
By which time, by definition, the information has had time to get to
you subluminally. It is only because you insist that the state at
both positions has to be objective, even if we don't know what it is
yet, that you get a contradiction.
Until you look at the system over at B, no matter whether the people
over there have looked at it, you have no reason to believe that the
result has yet been determined.
Harry.
--
"A foolish consistency is the hobgoblin of little minds" - Emerson
Harry Johnston, uda...@bay.cc.kcl.ac.uk
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <290q2n$h...@galaxy.ucr.edu>, ba...@guitar.ucr.edu (john baez) writes:
>
> > [...] I should also add, just in case you hadn't pondered it yet, that
> > it would be very naive to say that just because a
> > measurement at A tells us with certainty about what's going on at B,
> > something must zoom across the room when the measurement occurs. This
> > is something that goes on all the time in classical mechanics: take two
> > balls, one red and one blue, and randomly send one to Alpha Centauri and
> > the other to Sirius. The Sirian, upon seeing (say) a red ball,
> > immediately knows that a blue one will be seen on Centauri, many
> > light years. But of course no information is being transmitted faster
> > than light. (I say "of course," but really a definition of information
> > is needed to make this mean anything; I'm using the usual Shannon one.)
> > [...]
>
> The point of Bell and Eberhard's proofs is that the correlated predictions
> of QM do require superluminal transfer of information in the Shannon sense.
> If your carefully apply Shannon's definition to these experiments
> this is the inescapable conclusion. We know what the mechanisms are in
> the mathematics of QM that transfers this information in the model. We
> do not know if there is a superluminal transfer of information
> because we do not know if these predictions are correct.
>
A point that has come up repeatedly is that your statements about the
consequences of their reasoning has been shown to be false several
times. Specifically your comments about what the consequences of their
reasoning is in the case where you believe in the Everett
interpertation. I challenge you to show that within the Everett
interpertation there is no transfer of information.
[...]
> Your devotion to mathematical precision is apparently weaker then your refusal
> to face the problems in QM.
Your understanding of the way that the Everett interpertation works
boggles my mind. Here is the fact. Within the Everett interpertation
there is no change in the information. The EPR experiments are
explained by a local process. So on and so forth. Got it?
Ben Tilly
Paul Budnik uunet!mtnmath!paul
> [...]
> A point that has come up repeatedly is that your statements about the
> consequences of their reasoning has been shown to be false several
> times. Specifically your comments about what the consequences of their
> reasoning is in the case where you believe in the Everett
> interpertation. I challenge you to show that within the Everett
> interpertation there is no transfer of information.
I have said many times that Bell and Eberhard assume only elementary
statistics and geometry. Within that context their proof is valid. You have
to deny some of these elementary assumption or you have to show their
mathematics is invalid to disprove their results.
Everett may well be in contradiction with these elementary assumptions. There
is overwhelming evidence that these assumptions hold and not a shred of
evidence that Everett's strange assumptions have any validity. Please stop
quoting Everett as if he were some accepted physical theory. His claims
are speculative and his results are not generally accepted and not accepted
at all by me. If you want to attack the Bell and Eberhard results explain
which of their assumptions are wrong and why. However do not expect this
to change my mind or change what I write. As I said there is overwhelming
evidence in support of the assumptions Bell and Eberhard use. You take
on a Herculean task if you expect to show some of these assumptions are
invalid.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> In article <1993Oct07....@mtnmath.UUCP>,
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> > It is only because QM claims that probabilities are irreducible that we
> > are able to have superluminal information transfer without superluminal
> > communication. In effect the information transferred is encrypted with
> > quantum uncertainty. We can prove it was transferred and verify when the
> > transfer occurred but we can only do so afterwards by comparing results
> > from both locations.
>
> By which time, by definition, the information has had time to get to
> you subluminally. It is only because you insist that the state at
> both positions has to be objective, even if we don't know what it is
> yet, that you get a contradiction.
>
> Until you look at the system over at B, no matter whether the people
> over there have looked at it, you have no reason to believe that the
> result has yet been determined.
Yes you have to believe that other people can make reliable observations
and that recording devices can make reliable observations to test superluminal
effects. Otherwise the only way to test them would be if *you* could travel
superluminally. All ordinary science assumes objective observations are
possible. If you want to deny that objective observations can be made by
other people and devices that is fine but you have left the realm of
conventional science for a never never land of solipsism where your
observations are the only ones that count.
Paul Budnik
Keith Ramsay
|All ordinary science assumes objective observations are possible.
Indeed. "Objective" is not, however, very easy to define precisely.
It is necessary that one be able to observe "objectively", in a
relatively broad sense of the term. It is not clear that this sense of
"objectivity" is sufficient to rule out Everett's interpretation.
Everett's interpretation is a simplification of QM as usually applied.
Simpler theories are, all else being equal, to be preferred. It
implies that, although one observes just one value of an observable
[*], that more than one outcome is liable to exist. It is the
uniqueness of the outcome, after it has occurred, which you are
assuming and which doesn't follow from objectivity alone. It amounts
to imposing classicality upon parts of the system. Until we can agree
upon there being evidence that the world-in-the-large is classical,
this is not compelling.
[*] Some portions of state space for the system, including observer,
represent an observer having retrieved a single value of the
observable. Lo, the wave function is supported on such portions of the
state space.
Keith Ramsay
Paul Budnik uunet!mtnmath!paul
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> |All ordinary science assumes objective observations are possible.
>
> Indeed. "Objective" is not, however, very easy to define precisely.
> It is necessary that one be able to observe "objectively", in a
> relatively broad sense of the term. It is not clear that this sense of
> "objectivity" is sufficient to rule out Everett's interpretation.
>
> Everett's interpretation is a simplification of QM as usually applied.
Quantum mechanics does not require any interpretation
to make predictions. *Adding* an interpretation to it is not a simplification.
Interpretations do not have objective scientific content unless they make
predictions different from uninterpreted QM. In other words they are excess
philosophical baggage. No interpretation could be considered a simplification.
> Simpler theories are, all else being equal, to be preferred. It
> implies that, although one observes just one value of an observable
> [*], that more than one outcome is liable to exist.
It claims there are *other* outcomes that have no observable effect. How is
this a simplification and how does it qualify as science?
> It is the
> uniqueness of the outcome, after it has occurred, which you are
> assuming and which doesn't follow from objectivity alone.
What I claim is that only *observable* effects qualify as being objective
and as being science. Observability in different `worlds' that we have no
contact with does not count.
> It amounts
> to imposing classicality upon parts of the system. Until we can agree
> upon there being evidence that the world-in-the-large is classical,
> this is not compelling.
Horse manure. Accusing someone of classical thinking has become an all
purpose put down to use when you have no plausible arguments to make.
Limiting objective to what is observable is science *period* not
just classical science.
> [*] Some portions of state space for the system, including observer,
> represent an observer having retrieved a single value of the
> observable. Lo, the wave function is supported on such portions of the
> state space.
How is retrieving a single value from the state space simpler then assuming
the observation yields *the* single value in the state space? I understand
the motivation for this thinking. It may seem simpler if one thinks
it is a god given law that the wave function only evolves linearly. Creating
this philosophical garbage is a poor excuse for avoiding the obvious
conclusion that the wave function undergoes objective nonlinear changes
that we do not yet understand.
Paul Budnik
Silver Omega
> In article <ramsay.7...@unixg.ubc.ca>, ram...@unixg.ubc.ca (Keith Ramsay) writes:
>> Everett's interpretation is a simplification of QM as usually applied.
>
> Quantum mechanics does not require any interpretation
> to make predictions. *Adding* an interpretation to it is not a simplification.
> Interpretations do not have objective scientific content unless they make
> predictions different from uninterpreted QM. In other words they are excess
> philosophical baggage. No interpretation could be considered a simplification.
True in a way. The Everett interpretation simplifies the _idea_ of
QM, but certainly not the practice. Standard QM is in contradiction
with SR - as you have pointed out. Everett is the simplest way of
removing this contradiction.
>> Simpler theories are, all else being equal, to be preferred. It
>> implies that, although one observes just one value of an observable
>> [*], that more than one outcome is liable to exist.
>
> It claims there are *other* outcomes that have no observable effect. How is
> this a simplification and how does it qualify as science?
It is a simplification because you require no mysterious process to
remove the other outcomes apparently at random. It simplifies the
mathematics of the theory, without changing the predicted results.
Daryl McCullough
>ram...@unixg.ubc.ca (Keith Ramsay) writes:
>>
>> Everett's interpretation is a simplification of QM as usually applied.
>
>Quantum mechanics does not require any interpretation to make predictions.
That's not true. Without at least the minimal interpretation of the
square of the wave function as a probability, QM doesn't predict
anything at all.
>*Adding* an interpretation to it is not a simplification.
I think that Keith means that Everett's interpretation, where there is
no collapse, only smooth evolution, is simpler than the interpretation
that requires wave function collapse, if only because there is only
one type of evolution.
>Interpretations do not have objective scientific content unless they make
>predictions different from uninterpreted QM. In other words they are excess
>philosophical baggage. No interpretation could be considered a simplification.
Once again, it is the tossing out of the collapse that is a
simplification. The "many worlds" part of Everett is simply a way of
envisioning the resulting theory, as far as I'm concerned.
>> Simpler theories are, all else being equal, to be preferred. It
>> implies that, although one observes just one value of an observable
>> [*], that more than one outcome is liable to exist.
>
>It claims there are *other* outcomes that have no observable effect.
>How is this a simplification and how does it qualify as science?
It is a simplification because it eliminates the need for two kinds of
evolution. The unobservable aspects of the Everett interpretation need
not cause any more concern than the fact that when I'm in New York, I
can't observe what is happening in China, and vice-versa. Why should
one expect to be able to observe everything that exists?
[stuff deleted]
>> [*] Some portions of state space for the system, including observer,
>> represent an observer having retrieved a single value of the
>> observable. Lo, the wave function is supported on such portions of the
>> state space.
>
>How is retrieving a single value from the state space simpler then assuming
>the observation yields *the* single value in the state space?
Look, Paul, we have a theory (QM with no collapse) that is in
*perfect* agreement with experience---that is, there is absolutely
nothing in our experience that is not accounted for by linear QM.
How, then can you not agree that it is simpler to assume that there is
only linear evolution? How can you think that it is simpler to assume
that there is a nonlinear collapse, when there is no experimental
evidence in favor of it?
>I understand
>the motivation for this thinking. It may seem simpler if one thinks
>it is a god given law that the wave function only evolves linearly.
It isn't that it is god-given; it is that it is in perfect (as far as
we can tell) agreement with experiment.
>Creating this philosophical garbage is a poor excuse for avoiding the
>obvious conclusion that the wave function undergoes objective
>nonlinear changes that we do not yet understand.
When linear quantum mechanics is in perfect agreement with experiment,
it seems strange that you claim that nonlinear changes are an obvious
conclusion. Until we have experimental reasons to think that linear QM
is wrong, it seems to me that it makes more sense to try to understand
it than it does to invent unnecessary modifications.
Daryl McCullough
>I have said many times that Bell and Eberhard assume only elementary
>statistics and geometry.
And either hidden variables (in the case of Bell) or contrafactual
definiteness (in the case of Eberhard).
>Within that context their proof is valid. You have
>to deny some of these elementary assumption or you have to show their
>mathematics is invalid to disprove their results.
>Everett may well be in contradiction with these elementary assumptions.
It certainly is. It denies hidden variables *and* contrafactual
definiteness.
>There is overwhelming evidence that these assumptions hold
No, there is *not*! What kind of evidence are you talking about?
>and not a shred of evidence that Everett's strange assumptions have any
>validity.
The important thing about Everett's interpretation as far as I'm
concerned is that it shows that Eberhard is wrong; one *can't* prove
that quantum mechanics violates locality, since Everett's
interpretation is a counterexample with perfectly local evolution and
the same predictions as orthodox QM.
>Please stop quoting Everett as if he were some accepted physical theory.
>His claims are speculative and his results are not generally accepted and
>not accepted at all by me.
When you attack Everett, it is on the basis of almost complete
ignorance, since you have never read him. The "results" of Everett are
all perfectly standard quantum mechanics, showing that the apparent
success of assuming wave collapse after an observation can be
completely accounted for by linear evolution if one takes into account
the wave functions of observers and measuring devices. This fact is a
purely mathematical result of quantum mechanics, and is not disputed
(as far as I know) by anyone who has actually read Everett. All the
dispute is over the significance of the result. The many-worlds
interpretation was really an invention of others, not Everett, to give
a way to interpret no-collapse quantum mechanics.
>If you want to attack the Bell and Eberhard results explain
>which of their assumptions are wrong and why.
Hidden variables and counterfactual definiteness. It isn't that these
assumptions are necessarily wrong, it is only that there is no good
reason to assume that they are true.
>However do not expect this to change my mind or change what I write.
>As I said there is overwhelming evidence in support of the assumptions
>Bell and Eberhard use.
No, there is not. There are *philosophical* reasons for making the
assumptions that Bell and Eberhard use, but not evidence.
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <CELxx...@dartvax.dartmouth.edu>, Benjamin...@dartmouth.edu (Benjamin J. Tilly) writes:
> > [...]
> > A point that has come up repeatedly is that your statements about the
> > consequences of their reasoning has been shown to be false several
> > times. Specifically your comments about what the consequences of their
> > reasoning is in the case where you believe in the Everett
> > interpertation. I challenge you to show that within the Everett
> > interpertation there is no transfer of information.
>
> I have said many times that Bell and Eberhard assume only elementary
> statistics and geometry. Within that context their proof is valid. You have
> to deny some of these elementary assumption or you have to show their
> mathematics is invalid to disprove their results.
>
__OR__ you have to construct a counterexample. I find the last easiest
to do.
> Everett may well be in contradiction with these elementary assumptions. There
> is overwhelming evidence that these assumptions hold and not a shred of
> evidence that Everett's strange assumptions have any validity.
If you are going to claim that there is "overwhelming evidence" then
give me some. Otherwise stop posting bull to the net. And when you go
over the *exact* assumptions that you need for Everett there are some
very good reasons to accept them.
> Please stop
> quoting Everett as if he were some accepted physical theory. His claims
> are speculative and his results are not generally accepted and not accepted
> at all by me. If you want to attack the Bell and Eberhard results explain
> which of their assumptions are wrong and why. However do not expect this
> to change my mind or change what I write. As I said there is overwhelming
> evidence in support of the assumptions Bell and Eberhard use. You take
> on a Herculean task if you expect to show some of these assumptions are
> invalid.
I *do not* quote Everett as an accepted physical theory. I *have*
explained how Everett follows logically from some assumptions that a
lot of people would agree with. In any case for the point at hand what
is important is not that you believe Everett or not, but rather it is
that the fact that the Everett interpertation exists shows that various
comments of yours about the consequences of the Bell and Eberhard
results are false as statements about all possible interpertations of
QM. For that point there is no reason to show that their assumptions
are invalid about the real world. All that you need to do is show that
they do not *have* to be valid. Do you see the difference?
Ben Tilly
john baez
>In article <1993Oct09....@mtnmath.UUCP>
>pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>> Everett may well be in contradiction with these elementary
>> assumptions. There is overwhelming evidence that these assumptions
>> hold and not a shred of evidence that Everett's strange assumptions
>> have any validity.
>If you are going to claim that there is "overwhelming evidence" then
>give me some. Otherwise stop posting bull to the net. And when you go
>over the *exact* assumptions that you need for Everett there are some
>very good reasons to accept them.
The problem seems to be that Budnik mistakes his philosophical
predilections, which are those of "common sense," for "overwhelming
evidence." He seems to think, for example, that Everett's work violates
our usual notions of "objective reality." Of course, speaking about
"objective reality" is one of the worst things one could do in physics,
because it is such a vague and emotional concept -- being against
"objective reality" is like being against Mom and apple pie, yet it's
not at all clear what being *for* it means! In Budnik's case, it winds
up meaning being against quantum mechanics. This is especially ironic
in that quantum mechanics was something that nature, or if you like
"objective reality," dragged us kicking and screaming into believing.
In other words, experiment forced people into a model of nature quite
different from that of everyday common sense. This is in retrospect not
surprising, since common sense was built upon experience with large
objects, and quantum effects are typically easiest to notice for small
objects.
With Everett's work, it's clear too that quantum mechanics
fits quite nicely what we observe in the macroscopic world as well as
the microscopic world. (Let me repeat here for the nth time that
Everett's work is not to be confused with the popularlizations that
attempt to make it seem far-out.) So while it might be nice to have a model of
nature that pleases everyday common sense, it is wrongheaded to think
that a simple appeal to common sense suffices to overthrow quantum
mechanics. After arguing with Budnik for some time (and having given up
quite some time ago), it became clear to me that this is what his
arguments boiled down to. Note that I do not dogmatically *rule out*
future physical theories that do the job of quantum mechanics with less
violence done to common sense. I do not, however, expect such theories,
and it is up to those who feel the need for such theories to actually
construct them.
Robert Firth
>being against
>"objective reality" is like being against Mom and apple pie, yet it's
>not at all clear what being *for* it means!
Wearing a tattoo that says "My mother drunk and/or sober"?
lam...@cs.washington.edu
The Everett interpretation is most definitely not solipsistic. In the
first place, the role of consciousness in the Everett interpretation is
debatible. In the second place, you can't come up with an Everett
universe that violates quantum mechanics.
lam...@cs.washington.edu
>Look, Paul, we have a theory (QM with no collapse) that is in
>*perfect* agreement with experience---that is, there is absolutely
>nothing in our experience that is not accounted for by linear QM.
>How, then can you not agree that it is simpler to assume that there is
>only linear evolution? How can you think that it is simpler to assume
>that there is a nonlinear collapse, when there is no experimental
>evidence in favor of it?
This is going to sound kind of odd, but i tend to be opposed to the
Everett interpretation on the grounds that it is too simple. I tend
to think that nature is more creative than a bunch of early 20th
century European physicists.
Ron Maimon
|>
|> Quantum mechanics does not require any interpretation
|> to make predictions. *Adding* an interpretation to it is not a simplification.
|> Interpretations do not have objective scientific content unless they make
|> predictions different from uninterpreted QM. In other words they are excess
|> philosophical baggage. No interpretation could be considered a simplification.
|>
So, Paul, since you seem to know so much standard quantum mechanics, why don't you
please tell us what the prediction of standard quantum mechanics is for the
following experiments:
I have a plutonium atom in a highly excited state coupled to the electromagnetic
field. It is going to emit a photon _really_ soon.
I have a tiny sphere of aluminum atoms surrounding the plutonium atom. The
amplitude that the photon will ionize one of the aluminum atoms on its way
out is very close to 1.
There is a much larger sphere of aluminum atoms surrounding this sphere. These
aluminum atoms are missing many electrons- so that the potential difference
between the two spheres is huge, but not enough to cause the aluminum atoms to
ionize.
Assume that the probability that the photon will ionize one of the aluminum atoms
in the small sphere is very close to 1. This electron will then travel to the
other sphere, where it will have so much energy that it will cause a tremendous
localized excitation at that spot.
what is the initial and final wavefunction of the atom-spheres system?
alternatively. I have a pigment molecule and a photon of just the right frequency
to excite the pigment molecule. Unfortunately, the photon is in a superposition
state, so that it will only pass near the pigment with a probability of 50%. Or,
if you like, it's wave packet has two peaks, one which is moving toward the
pigment and one which is moving away from the pigment. What is the final state of
the pigment? If the pigment is connected to a larger system, and causes some
change in the larger system, what is the resulting state of the final system?
Ron Maimon
john baez
>This is going to sound kind of odd, but i tend to be opposed to the
>Everett interpretation on the grounds that it is too simple. I tend
>to think that nature is more creative than a bunch of early 20th
>century European physicists.
Yes, that does sound odd. Nature is undoubtedly more creative than we
are, or, to avoid anthropomorphizing, nature is the way it is and we
clearly haven't figured that out yet. However, if this is a reason to
oppose the work of Everett, or quantum theory, it is also a reason to
oppose all theories of physics! All current theories of physics are at
best approximations to the truth, as is evident from the fact that we
have no overarching theory of physics that combines the existing ones
and is consistent (much less correct!). The strategy in physics,
however, is a conservative one: we throw out something fundamental only
when we are quite convinced it yields either a mathematical
contradiction or a false prediction, OR if we happen to come up with
something better. Also, we never throw out something because it is too
simple -- only if it is too simple to fit the known facts. ("As simple
as possible, but no simpler.")
Benjamin J. Tilly
> da...@oracorp.com (Daryl McCullough) writes:
> >Look, Paul, we have a theory (QM with no collapse) that is in
> >*perfect* agreement with experience---that is, there is absolutely
> >nothing in our experience that is not accounted for by linear QM.
> >How, then can you not agree that it is simpler to assume that there is
> >only linear evolution? How can you think that it is simpler to assume
> >that there is a nonlinear collapse, when there is no experimental
> >evidence in favor of it?
>
> This is going to sound kind of odd, but i tend to be opposed to the
> Everett interpretation on the grounds that it is too simple. I tend
> to think that nature is more creative than a bunch of early 20th
> century European physicists.
And would you have been opposed to Newtonian physics on the same basis?
If you had then you would have been right. :-)
That is why I have been so clear about explaining the underlying
assumptions. It is clear to me that QM in the future will have to at
least be modified to get gravity in there, and it may go through
several modifications. However the Everett interpertation is IMHO
probably going to remain nearly right through all of these changes.
Therefore the explanation *as I have given it* is not going to be
exactly right. But it will probably be basically right in the same way
as the explanations from Newtonian physics are basically right, even
though the assumptions that it makes are actually wrong.
Ben Tilly
Paul Budnik uunet!mtnmath!paul
> > [...]
> > Interpretations do not have objective scientific content unless they make
> > predictions different from uninterpreted QM. In other words they are excess
> > philosophical baggage. No interpretation could be considered a simplification.
>
> True in a way. The Everett interpretation simplifies the _idea_ of
> QM, but certainly not the practice. Standard QM is in contradiction
> with SR - as you have pointed out.
It contradicts SR + causality. Dropping causality is the simpler
way and the one most often used to resolve this problem.
> Everett is the simplest way of removing this contradiction.
Multiplying the number of universes by some indeterminate
huge number is not a simplifying assumption. This is a technique
that could be used to quickly `solve' any problem in science. It is
just as scientific in my mind and almost equivalent logically to saying
some things are simply God's will and beyond human understanding.
> > [...]
> > It claims there are *other* outcomes that have no observable effect. How is
> > this a simplification and how does it qualify as science?
>
> It is a simplification because you require no mysterious process to
> remove the other outcomes apparently at random. It simplifies the
> mathematics of the theory, without changing the predicted results.
Remove what *other* outcomes. There are no other outcomes
that we can observe. Standard QM says nothing about other outcomes.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> >> Everett's interpretation is a simplification of QM as usually applied.
> >
> >Quantum mechanics does not require any interpretation to make predictions.
>
> That's not true. Without at least the minimal interpretation of the
> square of the wave function as a probability, QM doesn't predict
> anything at all.
You of course need some way to relate the mathematics to the the world
of real experiments. Interpretations rationalize this relationship. They
do not help to do experiments. We understand operationally what the
probability of an observation is and how to set up experiments. This
knowledge is independent of interpretations.
If an interpretation could derive the macroscopic from the microscopic
it would have scientific content. As it stands they are all start
out with *undefined primitive* terms that refer to macroscopic observations.
Thus they are only philosophical rationalizations of what
we already understand operationally. None of them provide a
scientific explanation of the connection between
the QM wave function and macroscopic observations.
> >*Adding* an interpretation to it is not a simplification.
>
> I think that Keith means that Everett's interpretation, where there is
> no collapse, only smooth evolution, is simpler than the interpretation
> that requires wave function collapse, if only because there is only
> one type of evolution.
That's fine as far as it goes. However the real world we observe is not
a smooth linear evolution. It is extremely bumpy. To say that Everett
provides the simplist explanation to deal with the messy bumpy real world
is ridiculous. First he does not explain anything he uses instrument
readings as undefined primitive terms and second multiplying universes
is hardly simplifying.
>[...]
> Once again, it is the tossing out of the collapse that is a
> simplification. The "many worlds" part of Everett is simply a way of
> envisioning the resulting theory, as far as I'm concerned.
He does not just toss out collapse. He replaces it with something that
is more absurd, improbable and equally ill defined.
> >It claims there are *other* outcomes that have no observable effect.
> >How is this a simplification and how does it qualify as science?
>
> It is a simplification because it eliminates the need for two kinds of
> evolution. The unobservable aspects of the Everett interpretation need
> not cause any more concern than the fact that when I'm in New York, I
> can't observe what is happening in China, and vice-versa. Why should
> one expect to be able to observe everything that exists?
It does not eliminate the need for two kinds of evolution.
It explains nothing. It simply *assumes* there are such things as instrument
readings. Where did these bumpy nonlinear instrument readings come from?
> >How is retrieving a single value from the state space simpler then assuming
> >the observation yields *the* single value in the state space?
>
> Look, Paul, we have a theory (QM with no collapse) that is in
> *perfect* agreement with experience---that is, there is absolutely
> nothing in our experience that is not accounted for by linear QM.
Bullshit! You arbitrarily assume there are instrument readings as primitive
undefined terms and then you claim you have accounted for them scientifically.
What you can account for is the probability of making observations. Neither
Everett no anyone else has a coherent scientific explanation of what
observations are.
> How, then can you not agree that it is simpler to assume that there is
> only linear evolution? How can you think that it is simpler to assume
> that there is a nonlinear collapse, when there is no experimental
> evidence in favor of it?
I am amazed at the depths of your rationalization. The world is not linear.
How much evidence for this do you need? No theory or interpretation of
QM explains this nonlinearity. They all simply assume it exists as collapse
or observations or instrument readings or whatever.
> >[...] It may seem simpler if one thinks
> >it is a god given law that the wave function only evolves linearly.
>
> It isn't that it is god-given; it is that it is in perfect (as far as
> we can tell) agreement with experiment.
*Every* experiment demonstrates nonlinear changes. Your statement is absurd.
> >Creating this philosophical garbage is a poor excuse for avoiding the
> >obvious conclusion that the wave function undergoes objective
> >nonlinear changes that we do not yet understand.
>
> When linear quantum mechanics is in perfect agreement with experiment,
> it seems strange that you claim that nonlinear changes are an obvious
> conclusion. Until we have experimental reasons to think that linear QM
> is wrong, it seems to me that it makes more sense to try to understand
> it than it does to invent unnecessary modifications.
It is in perfect agreement with how probabilities evolve. It says nothing
about how particular events come into being. These are nonlinear.
Do you think there is one universal wave function that has
evolved in a continuous linear way since the big bang? Is not the
switching of a transistor, the conception of a child or the death of
a cat a nonlinear effect? Do you think there is universal wave function
that encodes all the possible ways that everything *might* have developed?
How do you *rigorously* define what all these possibilities are?
Do you see that as a simplification? Can you claim this is science when
there is not a clear definition of what an observation or instrument
reading is?
Paul Budnik
john baez
>I am amazed at the depths of your rationalization. The world is not linear.
>How much evidence for this do you need? No theory or interpretation of
>QM explains this nonlinearity. They all simply assume it exists as collapse
>or observations or instrument readings or whatever.
A bunch of people get confused by the fact that "linearity" in quantum
mechanics has an utterly different meaning than that in, say, Maxwell's
equations (where the superposition of two electromagnetic fields in a
vacuum, both satisfying Maxwell's equations, is again a solution).
A simple example:
The harmonic oscillator problem in classical mechanics
x''(t) = -x(t)
is linear in the sense that the sum of two solutions is again a
solution. It's a linear ODE. The anharmonic oscillator is not:
x''(t) = -x(t) - cx(t)^3
All the brouhaha about chaos etc. is due to the fact that problems
nonlinear in this sense display much more interesting phenomena than
linear ones.
Now suppose we quantize these systems. Thus instead of the position of
the particle x(t) we deal with a wavefunction psi(t,x) such that
|psi(t,x)|^2 is the probability density of finding the particle at x at
time t. For the harmonic oscillator this satisfies a version of
Schrodinger's equation, with quadratic potential:
i (d psi/d t) = -(1/2)((d^2 psi/d x^2) + x^2 psi)
This is a linear PDE. For the anharmonic oscillator:
i (d psi/d t) = -(1/2)((d^2 psi/d x^2) + x^2 psi + cx^4 psi/2)
Again, a linear PDE! What's going on?? Does this mean quantum
mechanics is "more linear than classical mechanics" and hence fatally
flawed in some manner as Budnik hints?
No way, Jose! The point is that if we write out quantum mechanics in
the Heisenberg representation we see that it is just as nonlinear as
classical mechanics -- nonlinear in terms of the position x(t). This
nonlinearity has zip, zilch, nada to do with linearity in terms of the
wavefunction psi! In fact, one can easily describe classical mechanics
in terms of a wavefunction just as one does quantum mechanics -- the
only reason this is rarely done is that it's overkill. (To get a rough
idea of how, see the stuff on "Koopmanism" in the first volume of Reed
and Simon's series on mathematical physics -- although there is more to
say than what they say there!) If one describes the anharmonic
oscillator this way one gets an equation for psi that's perfectly linear
in psi. But this linearity has only to do with how *amplitudes* add,
and does not in the leasty contradict the fact that *observables* like
x(t) satisfy nonlinear differential equations.
So Budnik's statement "The world is nonlinear. How much evidence for
this do you need?" is correct in one sense -- observables evolve
nonlinearly -- but utterly misleading in another, in that he seems to be
saying this contradicts the fact that quantum mechanics has the
wavefunction satisfying a linear differential equation. Nope. One need
only crack open a quantum field theory book to see that the basic laws
of physics are quantized versions of *nonlinear* equations like the
Yang-Mills equation, and that this nonlinearity is why quantum field
theory is able to describe our world with such stunning precision.
Okay, now for the Heisenberg form of the quantized anharmonic
oscillator. I'll write "x" instead of "q" for the position observable
to clarify the relation to the classical case.
dx/dt = -i[H,x] dp/dt = -i[H,p]
where the Hamiltonian H is
H = (p^2 + x^2)/2 + cx^4/4
so
dx/dt = p, dp/dt = -x - cx^3
or
d^2 x/dt^2 = -x - cx^3
which is the SAME as in the classical case, except that now x(t) is the
position *operator*.
I did not really expect the above to make sense to people who haven't
learned quantum mechanics yet, but there are still plenty of people
who've learned it to some extent and are confused about this "linearity"
business. I know I was; I just had to figure it out myself. Upon
talking to experts, it became clear they had all figured it out for
*them* selves too.
Paul Budnik uunet!mtnmath!paul
> >I have said many times that Bell and Eberhard assume only elementary
> >statistics and geometry.
>
> And either hidden variables (in the case of Bell) or contrafactual
> definiteness (in the case of Eberhard).
Contrafactual definiteness is a standard assumption in statistics.
I regard Eberhard's results as a clarification or extension of Bell's.
> >Within that context their proof is valid. You have
> >to deny some of these elementary assumption or you have to show their
> >mathematics is invalid to disprove their results.
>
> >Everett may well be in contradiction with these elementary assumptions.
>
> It certainly is. It denies hidden variables *and* contrafactual
> definiteness.
>
> >There is overwhelming evidence that these assumptions hold
>
> No, there is *not*! What kind of evidence are you talking about?
Let us stick to contrafactual definiteness since Eberhard showed we
do not need to make assumptions about hidden variables. This is simply
the assumption that we can argue about all the possible outcomes of an
experiment even though only some of these will be realized in a
given experiment. One argues this way all the time in analyzing the
statistics of experiments.
>[...]
> The important thing about Everett's interpretation as far as I'm
> concerned is that it shows that Eberhard is wrong; one *can't* prove
> that quantum mechanics violates locality, since Everett's
> interpretation is a counterexample with perfectly local evolution and
> the same predictions as orthodox QM.
In Everett's model there will appear to be nonlocal
effects. These nonlocal effects come from the *linear* evolution of
the wave function in configuration space as you have pointed out many
times. I do not understand in what sense (if any) Everett gets around
Eberhard's results.
>
> >Please stop quoting Everett as if he were some accepted physical theory.
> >His claims are speculative and his results are not generally accepted and
> >not accepted at all by me.
>
> When you attack Everett, it is on the basis of almost complete
> ignorance, since you have never read him.
I am attacking him based on Bell's criticism of Everett. That is not `almost
complete ignorance' and more than adequate basis for a posting. I would
not attempt to publish an article on the subject without reading him
directly.
> The "results" of Everett are
> all perfectly standard quantum mechanics, showing that the apparent
> success of assuming wave collapse after an observation can be
> completely accounted for by linear evolution if one takes into account
> the wave functions of observers and measuring devices.
This is garbage. All interpretations have macroscopic references as
primitive undefined terms. These are the functional equivalent of collapse.
They are simply assumed and are not even clearly defined.
> This fact is a
> purely mathematical result of quantum mechanics, and is not disputed
> (as far as I know) by anyone who has actually read Everett.
I dare say Bell has read Everett. He quotes him in some detail. He does
not agree with your assessment.
> >If you want to attack the Bell and Eberhard results explain
> >which of their assumptions are wrong and why.
>
> Hidden variables and counterfactual definiteness. It isn't that these
> assumptions are necessarily wrong, it is only that there is no good
> reason to assume that they are true.
Hidden variables are not needed and contrafactual definiteness is a
standard assumption for analyzing the statistics of an experiment. I
doubt that you would even consider questioning it had you not been made
uncomfortable by Eberhard's results.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> The problem seems to be that Budnik mistakes his philosophical
> predilections, which are those of "common sense," for "overwhelming
> evidence."
My ideas about QM are far from common sense. They even
conflict with what physicists like yourself think is a reasonable
basis for a physical theory. You notion of common sense is offended
by a discrete space-time model.
> He seems to think, for example, that Everett's work violates
> our usual notions of "objective reality."
My argument against Everett is that he produces an overly complicated
analysis that accomplishes nothing. The detailed argument is the
criticism that Bell makes of him. I am still waiting for your answer
to Bell's arguments.
> Of course, speaking about
> "objective reality" is one of the worst things one could do in physics,
> because it is such a vague and emotional concept -- being against
> "objective reality" is like being against Mom and apple pie, yet it's
> not at all clear what being *for* it means! In Budnik's case, it winds
> up meaning being against quantum mechanics.
Why must claiming that quantum mechanics does not appear to be a
complete theory mean I am `against' it. I think it is
a marvelous achievement. I also think it leaves some important things out.
One job of a scientist is to focus on possible deficiencies
of existing theories.
> This is especially ironic
> in that quantum mechanics was something that nature, or if you like
> "objective reality," dragged us kicking and screaming into believing.
> In other words, experiment forced people into a model of nature quite
> different from that of everyday common sense. This is in retrospect not
> surprising, since common sense was built upon experience with large
> objects, and quantum effects are typically easiest to notice for small
> objects.
We all know the story and I agree that it was a remarkable extraordinary
achievement.
> With Everett's work, it's clear too that quantum mechanics
> fits quite nicely what we observe in the macroscopic world as well as
> the microscopic world.
Here we part company. To do this you need to explain the
macroscopic from the microscopic. Everett does not do this. He
simply assumes the existence of instrument readings as undefined primitive
terms.
> (Let me repeat here for the nth time that
> Everett's work is not to be confused with the popularlizations that
> attempt to make it seem far-out.) So while it might be nice to have a model of
> nature that pleases everyday common sense, it is wrongheaded to think
> that a simple appeal to common sense suffices to overthrow quantum
> mechanics.
Of course not. First no one is going to overthrow quantum mechanics. They
may refine it and even show that it is is only approximately correct.
This can only be done by experiments. Arguments are important because
they can show what experiments may be most revealing.
> After arguing with Budnik for some time (and having given up
> quite some time ago), it became clear to me that this is what his
> arguments boiled down to.
That is wishful thinking on your part. Let me summarize my principle arguments:
1. Bell and Eberhard have shown that QM predicts superluminal experimentally
detectable macroscopic effects. A number of physicists (such as Shimony)
realize that this means either causality or special relativity is false
at least for these effects. Baez is not among them in part I
suspect because he has yet to read Eberhard and Bell.
I think this prediction is unlikely to be true. I have pointed out that
an alternative possibility is that there is a delay between when experimental
parameters change and correlations are observed. At greater distances the
correlations may disappear as if the singlet state correlated particles
spontaneously collapsed into independent particles.
2. EPR like correlations come from the combination of absolute conservation
laws and probabilistic laws of observation. It is as if there were a Cosmic
Accountant that is continually monitoring all experimental results to insure
that the conservation books balance. I think it likely that there is some
underlying mechanistic process that explains how this comes about.
I would expect this to be a local process and that is the basis of my
predictions about what will be observed in effective tests of Bell's
inequality. I have explained how the class of models that I think should
be considered *might* produce such a result.
3. QM assumes that probabilities are irreducible. There is no mathematical way
to define such a model.
I think one should always in science look for
a complete explanation. To say that laws are probabilistic is to say that
we should not try and find a more complete explanation. To me this is
a sin against science. It is the equivalent of saying that some things
are simply God's will and beyond human understanding. That may be true
but we will never know unless we look for a more complete explanation.
4. There is no satisfactory explanation in existing theory to link the
macroscopic and the microscopic. I think the most likely explanation is
that the wave function is an objectively real physical object that
embodies the energy of the particle and that undergoes linear and nonlinear
changes. The wave function we use in our calculations
is only an average of the detailed behavior. I think this function is
continually undergoing nonliner changes and at times these changes collectively
contribute to a statistically irreversible macroscopic change. This is
my `interpretation' of QM. If it is true it will not be a philosophical
interpretation. It will be a scientific extension of the existing theory.
This are a number of experimentally testable consequences.
This assumption limits the diffusion of the wave function for a single
particle to the time it would take for that particle to transfer its energy
to a detector.
> Note that I do not dogmatically *rule out*
> future physical theories that do the job of quantum mechanics with less
> violence done to common sense. I do not, however, expect such theories,
> and it is up to those who feel the need for such theories to actually
> construct them.
My goal is not to formulate a theory that is more compatible with common
sense whatever that may mean. My goal is to formulate a theory that is more
complete and coherent than QM, i. e. one that does not introduce the
macroscopic world through arbitrary assumptions and primitive undefined terms.
The macroscopic should be derivable from the microscopic.
It is a safe bet that if such a theory is possible it will be at least as
difficult as QM to develop and probably a good deal more difficult.
No one is likely to be able to create such a theory from whole cloth and
introduce to the world as a fait accompli. Like QM it will only come about
through experiments that establish the need for a new theory and
play a central role in its development. What I am trying to do is show
how another perspective and another class of models might be the basis
for such a theory. This is important as a potential stimulus to doing the
right experiments. It may also serve as one way to help make sense out
of those experiments if they point in the direction I think they will.
Paul Budnik
Keith Ramsay
|Everett's interpretation is a simplification of QM as usually applied.
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
|Quantum mechanics does not require any interpretation
|to make predictions. *Adding* an interpretation to it is not a simplification.
|Interpretations do not have objective scientific content unless they make
|predictions different from uninterpreted QM. In other words they are excess
|philosophical baggage. No interpretation could be considered a simplification.
In physics QM seems uniquely plagued by having "interpretations".
Physicists lump many models together under the general rubric of
"quantum mechanics", and call specific forms of models
"interpretations". What I mean when I refer to Everett's view as an
"interpretation" is simply that it is one of the various models which
falls under the rubric of "quantum mechanics". Any model, of course,
needs to be interpreted before it can be regarded as a theory of
physics, in the sense that its connection with natural phenomena has
to be given as well as a formal structure.
I have trouble making sense of your phrase "uninterpreted QM". Once
you are specific enough about which version of QM you are invoking,
i.e., once you have a specific model in mind from among the usual menu
of possibilities, and once you have attached it to real world
predictions, then you already have an "interpretation", in both
senses, even if you are not making a big philosophical commitment to
it. You have, first, choosen a specific form of QM, and this specific
form has to have physical meaning.
I see no way to lay claim to particular simplicity by failing to
specify your model precisely at all! I don't understand, then, how you
conclude then that Everett's version is "more complicated" or "carries
excess philosophical baggage". It clearly runs *counter* to
philosophical baggage of certain kinds.
|> Simpler theories are, all else being equal, to be preferred. It
|> implies that, although one observes just one value of an observable
|> [*], that more than one outcome is liable to exist.
|
|It claims there are *other* outcomes that have no observable effect. How is
|this a simplification and how does it qualify as science?
The complexity of a theory is not to be measured by how extensive its
*consequences* appear, but how complex its *premises* are. My usual
"model" of the world includes the existence of quite a few physical
bodies which don't have any detectible influence on human beings,
being too small or too far away, and so on. This is okay, so long as
one is not adding complexity to the theory in order to tack them on.
Worse also is adding ad hoc assumptions to postulate them away.
Suppose that we are going down a river, run into a fork, and follow
the left hand fork. Someone who says, "Okay, let's say that the right
hand fork must dwindle out further downstream" is not making our
*model* of the river system simpler; they're just arbitrarily
postulating some phenomenon which would prune away the complexity *of
the river network*, at the expense of ad hoc patches to hydrodynamics.
Similarly someone who asserts, "Okay, now suppose that some nonlinear
phenomenon has caused that other portion of the wave function to go
away" is not simplifying out *model* of the world. They are just
presuming away some detail of the world by postulating another
(unverified) phenomenon. Supposing that other branches survive is not
adding complexity to the model, not unless there is some positive
reason to believe that the other branches die out (which is hard to
get, when you are too far away to see them). When they are close at
hand, you *can* check that they don't capriciously disappear.
The reason this qualifies as science is that science not only attempts
to produce *one* model or account of the phenomena of the world, but
the *best* one. Economy is one of the virtues we seek. It is perfectly
reasonable as part of science to attempt to produce a theory which
explains the very same facts as have been explained before, only with
fewer ad hoc assumptions.
|> It is the
|> uniqueness of the outcome, after it has occurred, which you are
|> assuming and which doesn't follow from objectivity alone.
|
|What I claim is that only *observable* effects qualify as being objective
|and as being science. Observability in different `worlds' that we have no
|contact with does not count.
A theory which predicts undetectible effects is not necessarily
"unscientific". There are lots of things in the world which we cannot
observe, but which we believe to be there, because the most economical
theory fitting our observations of the world includes them. It would
require ad hoc assumptions to describe how the world manages to avoid
them being included.
|> It amounts
|> to imposing classicality upon parts of the system. Until we can agree
|> upon there being evidence that the world-in-the-large is classical,
|> this is not compelling.
|Horse manure. Accusing someone of classical thinking has become an all
|purpose put down to use when you have no plausible arguments to make.
Sorry, probably I was writing too hastily concerning what your
position is. The essential problem is this: I feel you still have the
burden of explaining to us what you mean by "objective". I still
cannot see how it differs, if at all, from assuming that nature in the
large is classical.
Suppose I observe a dead cat. You apparently want to assert that a
"superposed" observation of a live cat musn't also be included in a
(sound) model of such transactions, because of the "objectivity" of
the observation.
There *is* a contradiction between "observing a dead cat which is also
live". There is also a problem with supposing that a person can invent
whichever alternative observation he wants at will. I feel that these
considerations are standard ingredients in objectivity. I'm willing to
take these as givens.
By my understanding of "objective", as I have just described it,
however, there is nothing about "objectively observing the dead cat"
which excludes there also being an associated part of the
wave-function of the world, corresponding to "objectively observing a
live cat". There is not a contradiction between the two.
For this reason, I believe that you are assuming something more by
"objectivity" than I am. The main obvious feature of this
"objectivity" is that it apparently leads in your thinking to the idea
that there aren't superpositions among "objectively observed"
phenomena.
Fairly soon we will either run out of useful things to discuss about
this "objectivity", or someone will introduce some way to characterize
what your notion of "objectivity" is, different from just "that
macroscopic superpositions can't happen". So as not to be hasty again
in interpreting you, I'll turn it over to you to explain it.
|Limiting objective to what is observable is science *period* not
|just classical science.
I hope that after reading this post you will see that I'm not trying
to argue for changing the standards of science. The best theory for
explaining the *observed* evidence may indeed have as a consequence
that there are unobserved phenomena in addition to observed phenomena.
|> [*] Some portions of state space for the system, including observer,
|> represent an observer having retrieved a single value of the
|> observable. Lo, the wave function is supported on such portions of the
|> state space.
|How is retrieving a single value from the state space simpler then assuming
|the observation yields *the* single value in the state space?
In order for observation to yield "the" value, and for it to be "the"
value in all of reality, there has to be a phenomenon occurring which
we haven't observed yet, wiping out the other alternatives: one needs
a `God-given' phenomenon of "observation" (or some broader type of
phenomenon which includes it as a special case), which somehow is
different from "non-observation" in that it forces this uniqueness. In
the other view, all the phenomena behaves by the same principles, and
our distinction between "observation" and "not observation" is just
our way of describing the world in terms convenient for us. *Our*
experience we call "observation", not because it's an intrinsically
different process from all the other interactions in nature, but
because we're *us*.
If there is such a distinction intrinsic to nature, then one should be
able to determine experimentally what it is, but there aren't any
experiments doing that for us, at least not yet. We can be confident
that state reduction *doesn't* occur in some cases, but all the
arguments to the effect that it *does* occur appears to depend upon
human observation being supposed to have a privileged position in
nature. We can see easily that photons behave "weirdly", but are less
easy about conceding that *we* might possible be behaving "weirdly" as
well.
|I understand
|the motivation for this thinking. It may seem simpler if one thinks
|it is a god given law that the wave function only evolves linearly. Creating
|this philosophical garbage is a poor excuse for avoiding the obvious
|conclusion that the wave function undergoes objective nonlinear changes
|that we do not yet understand.
I'm sorry that you consider our thinking on this topic trashy.
Keith Ramsay
ram...@unixg.ubc.ca
john baez
>My ideas about QM are far from common sense. They even
>conflict with what physicists like yourself think is a reasonable
>basis for a physical theory. You notion of common sense is offended
>by a discrete space-time model.
What physicists think and "common sense" are quite different. I don't
know whether or not a discrete spacetime model of lattice type goes
against common sense or not; my objection has nothing to do with that;
rather, it's that it violates diffeomorphism-invariance and makes it
difficult to formulate field theories such general relativity (and even
makes special relativity lose a lot of its symmetry). On the other
hand, the discrete spacetime structures that arise naturally when one
combines general relativity and quantum mechanics a la loop variables
preserve diffeomorphism invariance.
>> He seems to think, for example, that Everett's work violates
>> our usual notions of "objective reality."
>
>My argument against Everett is that he produces an overly complicated
>analysis that accomplishes nothing. The detailed argument is the
>criticism that Bell makes of him. I am still waiting for your answer
>to Bell's arguments.
I do not intend to get into a point-for-point discussion of this with
you, since there are now plenty of people on sci.physics who are
well-informed concerning Everett's work (to name a few: Unruh,
McCullough, McIrvin, Ramsay, Tilly and Maimon - really quite a startling
number compared to a couple of years ago) and who share my views as
closely as one could hope for given the nature of the subject. I would
not expect my own arguments to meet with greater success than theirs,
especially because I don't have the time this sort of thing requires.
In particular, since you seem to have only read about Everett's work
second-hand from someone criticizing him, it's all rather
pointless except in that some people on sci.physics are getting to hear
about his work from people who understand it, which is a good thing.
Similarly, since I have not ever had time to read the Eberhard stuff, I
don't want to get into discussions of that. Indeed, it is not very good
of me to criticize you without being prepared to dig in and spend lots
of time making everything I say precise. But I don't want to spend my time
doing this. (I have 2 books to finish up by year's end, classes to
teach, 2 grad students to boss around, and several papers I'm halfway
through writing.) That is why I decided quite a while back to no longer
reply to what you write. Since what you say so often provokes me to
some response, the only way I am able to stick to this decision is to
not read what you write. But occaisionally I glance at something and
can't resist firing off a brief salvo in response. Sorry! I will try
harder to leave you alone. The capable crew listed above are doing a
better job than I would ever have the patience for, in any event.
Daryl McCullough
>pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>>If you want to deny that objective observations can be made by
>>other people and devices that is fine but you have left the realm of
>>conventional science for a never never land of solipsism where your
>>observations are the only ones that count.
>
>The Everett interpretation is most definitely not solipsistic. In the
>first place, the role of consciousness in the Everett interpretation is
>debatible. In the second place, you can't come up with an Everett
>universe that violates quantum mechanics.
I'm sort of inclined to agree with Paul on this point. It may be that
the physics of the Everett interpretation never treats consciousness
as something special, but nevertheless, a consequence of the
interpretation is that all of the seemingly objective particular
features of the world---that there is an Earth, that there was a World
War II, that Bill Clinton was elected President of the US in 1992,
etc.---turn out to be features, not of the universe, but of one
particular path through the abstract set of possible histories. The
"history of the world" seen by me is, according to the Everett
interpretation, really only *my* history; a different observer,
performing different experiments (measuring different observables)
will have a different history.
Another point to be made is that there is absolutely no observable
difference (even in principle) between the Everett interpretation and
the assumption that consciousness collapses the wave function, but
only *my* consciousness. The Everett interpretation is equivalent, in
a certain sense, to starting from this rather solipsistic theory and
restoring observer-independence by throwing in not only my version of
history (where my observations collapse the wave function), but also
every other observer's history. (The neat thing about the technical
work of Everett is that it shows that all these different histories
*can* consistently be combined; that each person can assume that *his*
observations collapse the wave function, and no contradiction arises.)
Paul Budnik uunet!mtnmath!paul
> In article <1993Oct12....@mtnmath.UUCP> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> >I am amazed at the depths of your rationalization. The world is not linear.
> >How much evidence for this do you need? No theory or interpretation of
> >QM explains this nonlinearity. They all simply assume it exists as collapse
> >or observations or instrument readings or whatever.
>
> A bunch of people get confused by the fact that "linearity" in quantum
> mechanics has an utterly different meaning than that in, say, Maxwell's
> equations (where the superposition of two electromagnetic fields in a
> vacuum, both satisfying Maxwell's equations, is again a solution).
>
> [...]
> Again, a linear PDE! What's going on?? Does this mean quantum
> mechanics is "more linear than classical mechanics" and hence fatally
> flawed in some manner as Budnik hints?
This is not what I mean by nonlinear and you ought to know that.
The nonlinearity I am referring to is specific events. One observes
a particle at a particular location and time although its wave function
prior to that moment was spread over a much larger region.
> [...]
> So Budnik's statement "The world is nonlinear. How much evidence for
> this do you need?" is correct in one sense -- observables evolve
> nonlinearly -- but utterly misleading in another, in that he seems to be
> saying this contradicts the fact that quantum mechanics has the
> wavefunction satisfying a linear differential equation. Nope. One need
> only crack open a quantum field theory book to see that the basic laws
> of physics are quantized versions of *nonlinear* equations like the
> Yang-Mills equation, and that this nonlinearity is why quantum field
> theory is able to describe our world with such stunning precision.
>[...]
The nonlinearity I am referring to is something outside the linear
evolution of the wave function. The wave function only describes how
probabilities evolve. It says nothing about how these probabilities are
actualized into events. Once we have made an observation this imposes
a *nonlinear* constraint on the future evolution of the wave function
*or* on how we deduce probabilities from it. You understand this perfectly
well and you ought to understand that this is what I am talking about.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> In article <1993Oct12....@mtnmath.UUCP> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> >My ideas about QM are far from common sense. They even
> >conflict with what physicists like yourself think is a reasonable
> >basis for a physical theory. You notion of common sense is offended
> >by a discrete space-time model.
>
> What physicists think and "common sense" are quite different.
This suggests that common sense is some mental deficiency that we
are cursed with unless we become physicists. I would say that common sense
is a way of looking at the world based ones experience talents and
inclinations. It varys as widely as human experience does.
> I don't
> know whether or not a discrete spacetime model of lattice type goes
> against common sense or not; my objection has nothing to do with that;
> rather, it's that it violates diffeomorphism-invariance and makes it
> difficult to formulate field theories such general relativity (and even
> makes special relativity lose a lot of its symmetry).
Yes of course it does. That is an excellent common sense (for a sophisticated
physicist) argument against it. That is also why it was an emotional trauma
for Einstein to realize that such models may be needed.
I consider it quite possible that physics cannot be based on the
field concept, i. e., on continuous structures. In that case
*nothing* remains of my entire castle in the air gravitation
theory included, [and of] the rest of modern physics.
-- Einstein in a 1954 letter to Besso, quoted from:
"Subtle is the Lord", Abraham Pais, page 467.
>[...] I am still waiting for your answer
> >to Bell's arguments.
>
> I do not intend to get into a point-for-point discussion of this with
> you, since there are now plenty of people on sci.physics who are
> well-informed concerning Everett's work (to name a few: Unruh,
> McCullough, McIrvin, Ramsay, Tilly and Maimon - really quite a startling
> number compared to a couple of years ago) and who share my views as
> closely as one could hope for given the nature of the subject.
I am waiting for *anyone* to answer these arguments.
>[...]
> Similarly, since I have not ever had time to read the Eberhard stuff, I
> don't want to get into discussions of that.
It would probably take you less time to read the shorter of Eberhard's two
papers then it took you to write these recent responses.
> Indeed, it is not very good
> of me to criticize you without being prepared to dig in and spend lots
> of time making everything I say precise. But I don't want to spend my time
> doing this. (I have 2 books to finish up by year's end, classes to
> teach, 2 grad students to boss around, and several papers I'm halfway
> through writing.) That is why I decided quite a while back to no longer
> reply to what you write. Since what you say so often provokes me to
> some response, the only way I am able to stick to this decision is to
> not read what you write.
You need more self discipline. You might even learn something if you read
more of what I post even if you do not have time to respond.
Paul Budnik
SCOTT I CHASE
>
>This suggests that common sense is some mental deficiency that we
>are cursed with unless we become physicists.
No, it is a mental deficiency with which even physicists are burdened.
When you stop relying so heavily on your intuition, you will be in
a better position to do good physics. Intuition is sometimes a useful
tool. But overdependence upon an uneducated intuition can easily
lead you to a dead end. (As an extreme case of uneducated intuition,
we have Ludwig Plutonium, who just *knows* that there is only one order
of infinity, therefore the reals *must* be countable.)
-Scott
-------------------- Physics is not a religion. If
Scott I. Chase it were, we'd have a much easier
SIC...@CSA2.LBL.GOV time raising money. -Leon Lederman
lam...@cs.washington.edu
>If you had then you would have been right. :-)
I also would have been opposed to the aether on roughly the same basis.
>However the Everett interpertation is IMHO
>probably going to remain nearly right through all of these changes.
Severe emphasis should probably be placed on that IMHO. The Everett
Interpretation, and all of quantum mechanics, may simply be a useful
mathematical model for predicting results on the quantum scale. Admittedly
this is the same kind of criticism that Einstein levelled against
QM and was proven wrong every time. However, i think that Einstein was
correct in spirit, even if he couldn't produce anything really concrete.
john baez
>It would probably take you less time to read the shorter of Eberhard's two
>papers then it took you to write these recent responses.
I bet you underestimate the speed at which I can type! Plus, I spend
much of my time reading math and physics papers in my work, while
sci.physics counts as "rest and relaxation."
>You need more self discipline.
I do indeed need more self discipline. For example, I shouldn't post
this article, but I will. :-)
lam...@cs.washington.edu
>Yes, that does sound odd. Nature is undoubtedly more creative than we
>are, or, to avoid anthropomorphizing, nature is the way it is and we
>clearly haven't figured that out yet. However, if this is a reason to
>oppose the work of Everett, or quantum theory, it is also a reason to
>oppose all theories of physics!
You've almost hit the nail on the head there, but you've phrased it in
terms of black and white thinking.
>The strategy in physics,
>however, is a conservative one: we throw out something fundamental only
>when we are quite convinced it yields either a mathematical
>contradiction or a false prediction, OR if we happen to come up with
>something better. Also, we never throw out something because it is too
>simple -- only if it is too simple to fit the known facts. ("As simple
>as possible, but no simpler.")
I don't suggest that we blindly throw out quantum mechanics. However, the
applicability of quantum mechanics to the "classical" scale is something
which does IMHO need to be proven, rather than simply extrapolated. I don't
believe that the work of the past (70-80 years?) is sufficient for us to
determine that quantum mechanics really does represent the underlying
mechanics of reality and not just a computational approximation. And
i think we need to hold in mind that our theories of reality should, to an
extent, be considered guilty until proven innocent.
john baez
>In article <1993Oct12....@mtnmath.UUCP>, pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes...
>>This suggests that common sense is some mental deficiency that we
>>are cursed with unless we become physicists.
>No, it is a mental deficiency with which even physicists are burdened.
>When you stop relying so heavily on your intuition, you will be in
>a better position to do good physics.
I would prefer to say that ones intuition needs to be developed and
brought into accord with more and more facts about the world. The
trained intuition is a wonderful, amazing tool. But the intuition must
constantly be trained! It really seems to be a case of "no pain, no
gain," which is why one must always be on the lookout for various
"paradoxes" and puzzles, "inexplicable" facts, complicated formalisms
that work "for no good reason", etc..
As you say,
lam...@cs.washington.edu
>evolved in a continuous linear way since the big bang?
According to the Everett Interpretation this would be the case.
>Is not the
>switching of a transistor, the conception of a child or the death of
>a cat a nonlinear effect?
According to the Everett interpretation it is just one part of the
linear wave-function that we can observe and therefore only appears to
be non-linear because of our perspective.
>Do you think there is universal wave function
>that encodes all the possible ways that everything *might* have developed?
That is the Everett Interpretation.
>How do you *rigorously* define what all these possibilities are?
By Quantum Mechanics. Anything which quantum mechanics would allow given
the initial conditions does in fact happen.
>Do you see that as a simplification?
Yes it is a simplification. It establishes that Quantum Mechanics
rules the whole damn universe at every level, and thats probably the
best criticism against it.
lam...@cs.washington.edu
>I'm sort of inclined to agree with Paul on this point...
Depends on what your definition of solipsism is. My interpretation of
the way that Paul was using the word solipsism is that he was objecting
to the Everett interpetation based on some of the "pop"-interpretations
of Everett which give consciousness a special and active place in the universe.
My understanding of solipsism is that it theorizes that matter arises out
of mind rather than the reverse. The Everett Interpretation can be
twisted into producing such a universe, but is not necessarily solipsistic
in that sense. The OED definition of solipsism is "the view or theory
that the self is the only object of real knowledge or the only thing
really existant." Since the Everett Interpretation requries the wave-function
to be a "material" object in real existance then it is clearly not
solipsistic by the dictionary defintion.
Silver Omega
>> True in a way. The Everett interpretation simplifies the _idea_ of
>> QM, but certainly not the practice. Standard QM is in contradiction
>> with SR - as you have pointed out.
>
> It contradicts SR + causality. Dropping causality is the simpler
> way and the one most often used to resolve this problem.
How does dropping causality solve the problem? I think I must be
missing something. I haven't seen any other resolution besides
Everett and your self-collapsing system.
>> > It claims there are *other* outcomes that have no observable effect. How is
>> > this a simplification and how does it qualify as science?
>>
>> It is a simplification because you require no mysterious process to
>> remove the other outcomes apparently at random. It simplifies the
>> mathematics of the theory, without changing the predicted results.
>
> Remove what *other* outcomes. There are no other outcomes
> that we can observe. Standard QM says nothing about other outcomes.
Of course it does. When you calculate something in QM you get a
number of possible outcomes. Standard QM says "Well, the Hand of
God comes along and picks one at random." Everett says, "All the
possibilities occur, and _this_ is how come we only see one."
In a sense, the first is simpler. But very unsatisfactory as a
physical model - I guess it's a matter of taste, but I never did
like random processes!
Harry.
--
"A foolish consistency is the hobgoblin of little minds" - Emerson
Harry Johnston, uda...@bay.cc.kcl.ac.uk
Bruce Scott
Since your version of nonlinearity seems to differ from the usual, I'd
like to see a mathematical, not a verbal, explanation of what it is
you mean.
For example, in a 1D collisionless electrostatic plasma, the particle
distribution function for each species satisfies
(d/dt) f + v (d/dx) f + (q E/m) (d/dv) f = 0,
and the electric field satisfies
(d/dx) E = 4 pi \int dv (sum over species) q f,
where f is f(x,v,t), E is E(x,t), and q and m are charge and mass of
each particle. All derivatives are partial ones.
The first equation appears linear in f but the total system is
nonlinear since E is determined by integrals of all the f's over v.
[This is the Vlasov-Poisson system, usually expressed in terms of the
potential phi, where E = - (d/dx) phi.]
Maxwell's equations, on the other hand, are linear in all the fields,
as is Schroedinger's equation in the wave-function, psi.
Now, how does it work in your system? What in mathematical terms does
"nonlinearity" mean to you?
--
Gruss,
Dr Bruce Scott The deadliest bullshit is
Max-Planck-Institut fuer Plasmaphysik odorless and transparent
b...@hagar.ph.utexas.edu (to 12 Oct) -- W Gibson
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <1993Oct10....@bay.cc.kcl.ac.uk>, uda...@bay.cc.kcl.ac.uk (Silver Omega) writes:
> > True in a way. The Everett interpretation simplifies the _idea_ of
> > QM, but certainly not the practice. Standard QM is in contradiction
> > with SR - as you have pointed out.
>
> It contradicts SR + causality. Dropping causality is the simpler
> way and the one most often used to resolve this problem.
>
Which introduces a number of problems itself. But in any case the
Everett interpertation is in fact simpler than standard QM in a number
of ways. (Which I am not about to list again right here.)
> > Everett is the simplest way of removing this contradiction.
>
> Multiplying the number of universes by some indeterminate
> huge number is not a simplifying assumption. This is a technique
> that could be used to quickly `solve' any problem in science. It is
> just as scientific in my mind and almost equivalent logically to saying
> some things are simply God's will and beyond human understanding.
>
I agree. However I say that assuming that QM w/o collapse describes
macroscopic systems is a simplification because it allows a simple
answer to the domain of validity of QM and of explantions that use QM
which has a reasonable justification. (It was to point out the
importance of this problem that a certain animal abuse case was
described... :-) If it turns out, as it does, that this assumption
leads to an explanation of why there is effectively such a splitting,
then I call the theory a reasonable explanation.
> > It is a simplification because you require no mysterious process to
> > remove the other outcomes apparently at random. It simplifies the
> > mathematics of the theory, without changing the predicted results.
>
> Remove what *other* outcomes. There are no other outcomes
> that we can observe. Standard QM says nothing about other outcomes.
>
Standard QM says that there are other outcomes possible. The question
at hand here is what happens to the outcomes that we did not observe.
(First posed as why does one thing happen and the rest not. The problem
of the collapse of the wave function. A slight change in the question
is needed in the Everett interpertation for obvious reasons.)
Ben Tilly
Mats Grahm
(text deleted)
> to the Everett interpetation based on some of the "pop"-interpretations
> of Everett which give consciousness a special and active place in the universe.
(more text deleted)
Excuse my ignorance, but isn't Everett's interpretation a reaction
AGAINST giving consciousness a special role?
Knowing Everett only through quotations and explanations in
other authors books, I can't say I know much about it. But I thought
one of his ideas whas that by splitting the world, the wave function
does not has to collapse. And therefore no consciousness is needed
to make THAT thing happen. Am I totally wrong, or....?
Mats Grahm
Ron Maimon
|>
|> Do you think there is one universal wave function that has
|> evolved in a continuous linear way since the big bang? Is not the
|> switching of a transistor, the conception of a child or the death of
|> a cat a nonlinear effect? Do you think there is universal wave function
|> that encodes all the possible ways that everything *might* have developed?
|> How do you *rigorously* define what all these possibilities are?
|> Do you see that as a simplification? Can you claim this is science when
|> there is not a clear definition of what an observation or instrument
|> reading is?
|>
well....
yes. no. yes. it's easy. yes. yes.
Ron Maimon
lam...@cs.washington.edu
>In article b...@news.u.washington.edu, lam...@cs.washington.edu writes:
>> to the Everett interpetation based on some of the "pop"-interpretations
>> of Everett which give consciousness a special and active place in the universe.
>AGAINST giving consciousness a special role?
I can't answer this definitely. From what i understand, the Everett
interpretation was in response to the Copenhagen interpretation where
collapse comes as a function of the irreversible act of amplification in
a measurement. Then Everett and Cophenhagen wound up in this perverted
union in the pop books where conscious beings can "select" their
wavefunction or where consciousness causes the collapse.
Daryl McCullough
>> >*Adding* an interpretation to it is not a simplification.
>>
>> I think that Keith means that Everett's interpretation, where there is
>> no collapse, only smooth evolution, is simpler than the interpretation
>> that requires wave function collapse, if only because there is only
>> one type of evolution.
>
>That's fine as far as it goes. However the real world we observe is not
>a smooth linear evolution. It is extremely bumpy. To say that Everett
>provides the simplist explanation to deal with the messy bumpy real world
>is ridiculous. First he does not explain anything he uses instrument
>readings as undefined primitive terms and second multiplying universes
>is hardly simplifying.
Paul, you are going to have to stop saying what Everett says until you
find out what Everett says. You are basing your complete understanding
of Everett's work on Bell's essays. I have read both Everett's
articles and Bell's essays on Everett, and, as much as I like and
admire Bell, I think he is *extremely* sloppy in describing and
criticising Everett.
As far as I know, Everett does not give any preferred status to
instrument readings. Bell's statements to the contrary are just
*wrong*. Everett's theory allows one to calculate probabilities for
*any* observables whatsoever, whether they correspond to instrument
readings, or whatever.
The only place where instrument readings come into play is in
Everett's showing that his theory of the universal wavefunction is
consistent with the observations of physical systems like us, with
memories. And he succeeds in showing this, more or less.
>>[...]
>> Once again, it is the tossing out of the collapse that is a
>> simplification. The "many worlds" part of Everett is simply a way of
>> envisioning the resulting theory, as far as I'm concerned.
>He does not just toss out collapse. He replaces it with something that
>is more absurd, improbable and equally ill defined.
No, he does not.
>> >It claims there are *other* outcomes that have no observable effect.
>> >How is this a simplification and how does it qualify as science?
>>
>> It is a simplification because it eliminates the need for two kinds of
>> evolution. The unobservable aspects of the Everett interpretation need
>> not cause any more concern than the fact that when I'm in New York, I
>> can't observe what is happening in China, and vice-versa. Why should
>> one expect to be able to observe everything that exists?
>
>It does not eliminate the need for two kinds of evolution.
Yes it does. It has only has one kind of evolution, and reproduces the
same predictions as orthodox quantum mechanics.
>It explains nothing.
Yes it does. It shows why there is no contradiction in experiments
involving several observers, such as the EPR experiment. In orthodox
quantum mechanics, there *seems* to be a contradiction in such
experiments (at least with special relativity). Everett's main results
show that, even without assuming that observation collapses the wave
function, the results of repeated experiments by an observer will be
*as* if the wave function collapsed after each observation. This was a
real mathematical result, that did not assume *anything* about
specialness of instrument readings; it just treated an observer as a
quantum mechanical system in its own right.
>It simply *assumes* there are such things as instrument readings.
Everett assumed that instruments were quantum mechanical systems, just
like atoms and molecules, only much bigger and more complicated.
>Where did these bumpy nonlinear instrument readings come from?
I have no idea what you mean by "nonlinear" in the above sentence, but
as I said, Everett doesn't treat instruments any differently than atoms
or electrons, or whatever (except in their complexity).
>> >How is retrieving a single value from the state space simpler then assuming
>> >the observation yields *the* single value in the state space?
>>
>> Look, Paul, we have a theory (QM with no collapse) that is in
>> *perfect* agreement with experience---that is, there is absolutely
>> nothing in our experience that is not accounted for by linear QM.
>
>Bullshit! You arbitrarily assume there are instrument readings as
>primitive undefined terms and then you claim you have accounted for
>them scientifically.
I am only claiming that one doesn't need nonlinear evolution to
account for experiment. That was a nontrivial result, and is not at
all obvious. The version of quantum mechanics that Von Neumann first
formalized, with the two different kinds of evolution---linear between
observations and nonlinear after observations---would seem to give
numerically different results than a theory that assumes that there is
only linear evolution. Everett showed that it does not (not for any
experimentally testable cases, anyway). In other words, the assumption
that the wave function was collapsed by past observations is
unnecessary for statistical agreement with experiment. I don't think
that it is obvious what conclusions should be drawn from this, but it
does seem that it is premature to be looking for mechanisms for wave
function collapse when there is no evidence that there is such a thing.
>What you can account for is the probability of making observations. Neither
>Everett no anyone else has a coherent scientific explanation of what
>observations are.
I guess I agree with that.
>> How, then can you not agree that it is simpler to assume that there is
>> only linear evolution? How can you think that it is simpler to assume
>> that there is a nonlinear collapse, when there is no experimental
>> evidence in favor of it?
>I am amazed at the depths of your rationalization. The world is not
>linear.
What do you mean by that, and what makes you believe that? What I mean
by linearity is this: if |Psi1(t)> is a solution to Schrodinger's
equation, and |Psi2(t)> is a solution, then |Psi1(t)> + |Psi2(t)> is
also a solution. Any evidence that the world is nonlinear, in this
sense, would be evidence that Schrodinger's equation is incorrect.
(Of course, it *is* incorrect, it is only valid for nonrelativistic
systems, but in the relativistic case, there is a similar linear
evolution equation for the state.)
>How much evidence for this do you need?
Any evidence at all for the incorrectness of quantum mechanics would
count. (Experimental evidence, that is---examples of predictions made
by quantum mechanics that are contradicted by experiment.)
>No theory or interpretation of QM explains this nonlinearity. They
>all simply assume it exists as collapse or observations or instrument
>readings or whatever.
I don't know what you are talking about. The Everett theory is
completely linear, so there is no assumption of the existence of
nonlinearity.
>> >[...] It may seem simpler if one thinks
>> >it is a god given law that the wave function only evolves linearly.
>>
>> It isn't that it is god-given; it is that it is in perfect (as far as
>> we can tell) agreement with experiment.
>*Every* experiment demonstrates nonlinear changes.
No it does not! There has never been a single experiment that has
shown that wave function evolution is nonlinear.
>Your statement is absurd.
Well, I will agree that my statements, together with yours are absurd,
since we are asserting contradictory things.
Let me make my statement more explicit. I claim that there has never
been an experiment that has shown that past evolution of the wave
function was ever nonlinear. Suppose that we know that wave function
|A(0)> at time 0 evolves into to wave function |A(t)> at time t, and
that wave function |B(0)> at time 0 evolves into wave function |B(t)>
at time t. Linearity means that the superposition |A(0)> + |B(0)>
would evolve into the superposition |A(t)> + |B(t)>. Where is there
any evidence of a failure of this principle of linearity? (This would
be an example of the failure of quantum mechanics, since it predicts
such linearity).
>> When linear quantum mechanics is in perfect agreement with experiment,
>> it seems strange that you claim that nonlinear changes are an obvious
>> conclusion. Until we have experimental reasons to think that linear QM
>> is wrong, it seems to me that it makes more sense to try to understand
>> it than it does to invent unnecessary modifications.
>It is in perfect agreement with how probabilities evolve. It says nothing
>about how particular events come into being. These are nonlinear.
Linearity (as I have been using it) only applies to the evolution of
the wave function. I don't know what you mean by saying that the
way particular events come into being is "nonlinear".
Be that as it may, it seems that we are in agreement, that the
evolution of probabilities (probability amplitudes, actually) are
perfectly described by a linear evolution equation, as far as any
experimental evidence shows. The extra hypothesis that the wave
function sometimes undergoes nonlinear evolution is not justified by
any experimental results.
>Do you think there is one universal wave function that has
>evolved in a continuous linear way since the big bang?
I don't know, but it seems to agree with our observations.
>Is not the switching of a transistor, the conception of a child or
>the death of a cat a nonlinear effect?
I don't know what you mean by calling these things "nonlinear".
There is no reason to think (as far as I know) that the probability
amplitudes for these things evolve nonlinearly.
>Do you think there is universal wave function that encodes all the
>possible ways that everything *might* have developed?
Sure.
>How do you *rigorously* define what all these possibilities are?
I don't know how to describe babies from first principles starting
with quantum mechanics, no. However, I think that the only hope is to
assume that transitions, and babies and cats are made up of elementary
particles describable by Quantum field theory. Then there are various
ways of characterizing histories of particles as paths through
configuration space.
>Do you see that as a simplification? Can you claim this is science
>when there is not a clear definition of what an observation or
>instrument reading is?
Quantum mechanics as advocated by Everett doesn't treat observation or
instrument readings as essentially different from any other kind of
quantum mechanical event. They are all described by probability
amplitudes. As to whether that makes things simple, no it doesn't.
There is no way to make the physics of 10^23 interacting particles
simple. But it is *simpler* in principle than an ad hoc theory that
throws in ill-defined "collapses" for which there is no evidence or
mechanism or need.
Paul Budnik uunet!mtnmath!paul
> In physics QM seems uniquely plagued by having "interpretations".
> Physicists lump many models together under the general rubric of
> "quantum mechanics", and call specific forms of models
> "interpretations". What I mean when I refer to Everett's view as an
> "interpretation" is simply that it is one of the various models which
> falls under the rubric of "quantum mechanics". Any model, of course,
> needs to be interpreted before it can be regarded as a theory of
> physics, in the sense that its connection with natural phenomena has
> to be given as well as a formal structure.
There is a distinct difference between interpretations in QM and
interpretations in other physical theories. QM as it is currently
formulated must treat two separate and irreconcilable realms: the macroscopic
and the microscopic. These are not linkable within the current theory
except through arbitrary metaphysical assumptions. All interpretations make
such assumptions.
Of course we need this link in some form to apply the theory.
But we only need an operational link that tells us how to set up
experiments and how to apply the mathematics to the experiments. None of
the interpretations help us to do this. They only provide a rationalization
for what we already understand how to do.
> I have trouble making sense of your phrase "uninterpreted QM". Once
> you are specific enough about which version of QM you are invoking,
> i.e., once you have a specific model in mind from among the usual menu
> of possibilities, and once you have attached it to real world
> predictions, then you already have an "interpretation", in both
> senses, even if you are not making a big philosophical commitment to
> it. You have, first, choosen a specific form of QM, and this specific
> form has to have physical meaning.
By uninterpreted QM I mean the mathematics that describes how the wave
function evolves and how observations of macroscopic states apply constraints
on future predictions. One can use this mathematics without assuming collapse,
Everett or anything beyond the mathematics and the methodology for applying
it to experiments. None of the interpretations add anything to this model.
This is what most physicists do when they use QM. They certainly do not
worry about what some interpretation implys about their experiment or
analysis.
> I see no way to lay claim to particular simplicity by failing to
> specify your model precisely at all! I don't understand, then, how you
> conclude then that Everett's version is "more complicated" or "carries
> excess philosophical baggage". It clearly runs *counter* to
> philosophical baggage of certain kinds.
It is simple. The model is fully specified in the way I just described.
What does Everett add to this model?
>[...]
> The complexity of a theory is not to be measured by how extensive its
> *consequences* appear, but how complex its *premises* are. My usual
> "model" of the world includes the existence of quite a few physical
> bodies which don't have any detectible influence on human beings,
> being too small or too far away, and so on. This is okay, so long as
> one is not adding complexity to the theory in order to tack them on.
> Worse also is adding ad hoc assumptions to postulate them away.
But this is what Everett did. The problem interpretations address is
relating the macroscopic to the microscopic. Everett simply assumes the
macroscopic exists in the arbitrary from of `instrument readings'.
This is no different then the Copenhagen interpretation that assumes the
macroscopic exists in the form of observations. We know that macroscopic
observations constrain the future evolution of the wave function or at
least the probabilities we can deduce from it. This is the minimal simple
assumption.
> Suppose that we are going down a river, run into a fork, and follow
> the left hand fork. Someone who says, "Okay, let's say that the right
> hand fork must dwindle out further downstream" is not making our
> *model* of the river system simpler; they're just arbitrarily
> postulating some phenomenon which would prune away the complexity *of
> the river network*, at the expense of ad hoc patches to hydrodynamics.
> Similarly someone who asserts, "Okay, now suppose that some nonlinear
> phenomenon has caused that other portion of the wave function to go
> away" is not simplifying out *model* of the world. They are just
> presuming away some detail of the world by postulating another
> (unverified) phenomenon. Supposing that other branches survive is not
> adding complexity to the model, not unless there is some positive
> reason to believe that the other branches die out (which is hard to
> get, when you are too far away to see them). When they are close at
> hand, you *can* check that they don't capriciously disappear.
I do not think this analogy applies. We know how hydrodynamic works and
there is a vast body of experiment that confirms this.
My biggest problem with Everett is the false claim that he
explains the relationship between the microscopic and the macroscopic. This
is not true. The nonliner effects exist in all the interpretations and are
explained in none of them. It does not matter if you expand the wave function
in terms of `instrument readings' or you collapse at at `observations'. In
either case you introducing the macroscopic as a undefined primitive concept.
It is not derivable or explicable from the physical model.
The simplest model is to say the wave function behaves as if it collapses
when a thermodynamically irreversible events occur. This is the the working
assumption of every experimental physicist.
This fully describes how QM is done and all models must support this
minimal model in some way. How can adding other philosophical baggage that
has no experimental consequences be a simplification?
Physicists feel compelled to invent interpretations because of the implications
of this minimal model. It suggests that there are objective nonlinear changes
of the wave function associated with thermodynamically irreversible events.
If this is true quantum mechanics is an incomplete theory. Horror of horrors
lets invent an interpretation quick so we do not have to consider this
possibility!
> The reason this qualifies as science is that science not only attempts
> to produce *one* model or account of the phenomena of the world, but
> the *best* one. Economy is one of the virtues we seek. It is perfectly
> reasonable as part of science to attempt to produce a theory which
> explains the very same facts as have been explained before, only with
> fewer ad hoc assumptions.
I just explained what the most economical model is. I do not believe that
simplicity is the motivating factor. I think it an antiscientific desire
to rationalize out of existence what we do not understand. The scientific
approach is to admit our ignorance and try to improve upon it.
>[...]
> A theory which predicts undetectible effects is not necessarily
> "unscientific". There are lots of things in the world which we cannot
> observe, but which we believe to be there, because the most economical
> theory fitting our observations of the world includes them. It would
> require ad hoc assumptions to describe how the world manages to avoid
> them being included.
What exactly are you thinking about. We assume there are galaxies we
cannot now observe because we can observe regular patterns
and we can extrapolate from them. We do not believe in the existence of
galaxies that are *in principle* unobservable. It is only a matter of having
a sensitive enough instrument and looking at the right place. This is
fundamentally different then Everett. Here we take a perfectly good theory
and add to it the assumption that there is a vast range of effects that
are *in principle* not detectable. That is metaphysics not science.
> [...]
> The essential problem is this: I feel you still have the
> burden of explaining to us what you mean by "objective". I still
> cannot see how it differs, if at all, from assuming that nature in the
> large is classical.
What I mean is what every experimental physicists assumes in his work.
Thermodynamically irreversible processes are objective events.
> Suppose I observe a dead cat. You apparently want to assert that a
> "superposed" observation of a live cat musn't also be included in a
> (sound) model of such transactions, because of the "objectivity" of
> the observation.
I think the dead cat speaks for itself. Of course you can assume that it
will be resurrected and go to cat heaven or you can assume that it is
superimposed with a live cat. You can assume anything you want to but
my experience is that dead cats stay dead. Of course you can assume that
there are things that have no physical and observable effect. This is
not science. Science only deals with what can be observed at least in
principle.
>[...]
> For this reason, I believe that you are assuming something more by
> "objectivity" than I am. The main obvious feature of this
> "objectivity" is that it apparently leads in your thinking to the idea
> that there aren't superpositions among "objectively observed"
> phenomena.
Thermodynamically irreversible events are objective. That is how we
all live our lives and it is how all experimental physics is done.
> Fairly soon we will either run out of useful things to discuss about
> this "objectivity", or someone will introduce some way to characterize
> what your notion of "objectivity" is, different from just "that
> macroscopic superpositions can't happen". So as not to be hasty again
> in interpreting you, I'll turn it over to you to explain it.
Do you object to the definition I just gave?
>
> |Limiting objective to what is observable is science *period* not
> |just classical science.
>
> I hope that after reading this post you will see that I'm not trying
> to argue for changing the standards of science. The best theory for
> explaining the *observed* evidence may indeed have as a consequence
> that there are unobserved phenomena in addition to observed phenomena.
I certainly think you are. Of course a model may have implications that
are not observable. This makes those aspects of the model speculative.
We establish scientific theories by experiments. Aspects of a model
that are not testable cannot become established science. For example one
should always keep in mind the experimental envelope in which existing
theories has been validated and recognize that far outside that envelope
even the most firmly established theories must be considered speculation.
>[...]
> In order for observation to yield "the" value, and for it to be "the"
> value in all of reality, there has to be a phenomenon occurring which
> we haven't observed yet, wiping out the other alternatives: one needs
> a `God-given' phenomenon of "observation" (or some broader type of
> phenomenon which includes it as a special case), which somehow is
> different from "non-observation" in that it forces this uniqueness.
It exists and has been extensively tested and described. It is called
thermodynamic irreversibility. It is the basis of all experimental records
and all experimental physics. You know that as well as I do. Why do you
keep talking as if this was not well known and well understood?
[...]
> |Creating
> |this philosophical garbage is a poor excuse for avoiding the obvious
> |conclusion that the wave function undergoes objective nonlinear changes
> |that we do not yet understand.
>
> I'm sorry that you consider our thinking on this topic trashy.
I think it is trashy in a deeply serious way. We all understand that
thermodynamically irreversible processes correspond to objective events.
That is how we do experimental physics and it is how we live our lives.
All this philosophical garbage is to avoid the implications of this
well understood and universally accepted (*in practice*) fact. This
is an extremely serious business. Instead of facing the implications
of this physicists have invented a vast web of rationalization.
I do think history will come down harshly on this effort. You are using
the resources that society makes available in a way that does not advance
our understanding of nature. You are using them to avoid asking the
tough questions.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> >
> >This suggests that common sense is some mental deficiency that we
> >are cursed with unless we become physicists.
>
> No, it is a mental deficiency with which even physicists are burdened.
> When you stop relying so heavily on your intuition, you will be in
> a better position to do good physics. Intuition is sometimes a useful
> tool. But overdependence upon an uneducated intuition can easily
> lead you to a dead end.
Intuition and common sense are different things. However I am happy to
repond to the point you are making.
I do not think my intuition as `uneducated'. If I was not able to argue
intelligently on these issues people would eventually ignore me.
It cuts both ways. One can rely too much on analysis and one can rely too
much on intuition. Western culture in general has a one sided bias to
a purely intellectual approach to problem solving and physics is
one of the most if not the most one sided of disciplines.
The greatest of physicists were not the ones with the most powerful intellect.
They were men like Bohr and the Einstein whos greatest strength was there
powerful intuition.
Neither intuition nor intellect can get you far alone. It is the
creative functioning of both of them that is central to scientific advance.
Perhaps what you see as my excessive reliance on intuition is largely a
reflection of the one sided intellectual way physics has developed.
If there were a little more faith in and reliance on intuition among
physicists much of the metaphysical garbage that I keep bitching about
what have been purged from the subject a long time ago. The problems that
I think are central would have received more attention and perhaps all of
us and society in general would be better off.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
>
> How does dropping causality solve the problem? I think I must be
> missing something. I haven't seen any other resolution besides
> Everett and your self-collapsing system.
Quoting Abner Shimony in "Conceptual foundations of quantum mechanics"
(The New Physics, Edited by Paul Davies, Cambridge Univ. Press, 1989)
The quantum mechanical probabilities of outcomes for e_x(1) and e_x(2)
are not independent of each other (since then the Bell independence
of conditions would hold), but they are due to reciprocal influence,
without singling out one event as the cause and one as the effect.
This kind of causal connectedness between two events with space-like
separation has no classical analogue, and no classical analogue
should be expected, since quantum mechanical potentiality has
essentially broadened the concept of an event. (page 388)
The predictions are the same in any frame of reference. It is only the
information transfer or causal mechanism that produces the correlations
that occurs in different sequences in different frames of reference.
> > [...] Remove what *other* outcomes. There are no other outcomes
> > that we can observe. Standard QM says nothing about other outcomes.
>
> Of course it does. When you calculate something in QM you get a
> number of possible outcomes. Standard QM says "Well, the Hand of
> God comes along and picks one at random." Everett says, "All the
> possibilities occur, and _this_ is how come we only see one." [...]
Standard QM does not predicts any outcome. It predicts the probability of
an outcome. To get even a single outcome you have to add something like
an observation. To get multiple outcomes you have to add the machinery
that defines what these are. In the case of Everett they are something like
all possible instrument readings. That sounds like a shaky foundation on
which to base a physical theory.
Paul Budnik
Daryl McCullough
[about whether the Everett interpretation is solipsism]
>Depends on what your definition of solipsism is. My interpretation of
>the way that Paul was using the word solipsism is that he was objecting
>to the Everett interpetation based on some of the "pop"-interpretations
>of Everett which give consciousness a special and active place in the
>universe. My understanding of solipsism is that it theorizes that matter
>arises out of mind rather than the reverse. The Everett Interpretation can be
>twisted into producing such a universe, but is not necessarily solipsistic
>in that sense. The OED definition of solipsism is "the view or theory
>that the self is the only object of real knowledge or the only thing
>really existant." Since the Everett Interpretation requries the wave-function
>to be a "material" object in real existance then it is clearly not
>solipsistic by the dictionary defintion.
I disagree. It is true that the Everett interpretation has an
objective wave function, but this wave function is actually not very
interesting. You may be thinking: "Of *course* it is interesting, it
contains every possibility anyone could imagine." True, but that is
what makes it uninteresting. A large enough block of marble contains
within it the possibility for every statue ever carved, its only
necessary to chip off the bits you aren't interested in. There is
nothing interesting about the set of all possibilities, it is only
particulars that are interesting.
And all the particulars of the universe in the Everett interpretation
only exist in the subjective view of an observer. There is no "object
of real knowledge" in the Everett interpretation except the knowledge
of some particular observer's history.
Solipsism is certainly not the starting point of Everett's
interpretation; it is the conclusion.
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> >[...] I am still waiting for your answer
> > >to Bell's arguments.
> >
> > I do not intend to get into a point-for-point discussion of this with
> > you, since there are now plenty of people on sci.physics who are
> > well-informed concerning Everett's work (to name a few: Unruh,
> > McCullough, McIrvin, Ramsay, Tilly and Maimon - really quite a startling
> > number compared to a couple of years ago) and who share my views as
> > closely as one could hope for given the nature of the subject.
>
> I am waiting for *anyone* to answer these arguments.
I thought that I already *had* answered them. However post them again
and *read* what I write carefully.
Ben
Benjamin J. Tilly
lam...@cs.washington.edu writes:
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> >Do you think there is one universal wave function that has
> >evolved in a continuous linear way since the big bang?
>
> According to the Everett Interpretation this would be the case.
>
(stuff deleted)
> >Do you see that as a simplification?
>
> Yes it is a simplification. It establishes that Quantum Mechanics
> rules the whole damn universe at every level, and thats probably the
> best criticism against it.
Not quite. There are several different versions that are usually not
distinguished. The one which I have defended says only that QM is a
good approximation to reality at our scale, and then establishes the
Everett interpertation as an explanation of the mechanism which makes
QM being a good approximation to reality consistent with our
observations. In this form it does not assert anything about the
ultimate nature of reality. This is similar to using fluid mechanics to
explain things about airplane wings, w/o our having to assert that the
fluid mechanics provides us with an exact description.
One version which would be false (which is an impression that you could
get as to what people mean) is that existing QM is reality exactly.
This is actually wrong since we do not have a quantum theory of
gravity. However that does lead to the one which you seem to actually
calling the Everett interpertation, which is that there is some version
of QM which we do not have yet that *is* an exact description of the
universe, in which Everett is right. I would agree with you that this
is rather speculative and I would not accept it as a theory basically
for the reason that you give. However note that this does not pose any
problem for the version that I was talking about.
Ben Tilly
Benjamin J. Tilly
lam...@cs.washington.edu writes:
> ba...@guitar.ucr.edu (john baez) writes:
> >Yes, that does sound odd. Nature is undoubtedly more creative than we
> >are, or, to avoid anthropomorphizing, nature is the way it is and we
> >clearly haven't figured that out yet. However, if this is a reason to
> >oppose the work of Everett, or quantum theory, it is also a reason to
> >oppose all theories of physics!
>
> You've almost hit the nail on the head there, but you've phrased it in
> terms of black and white thinking.
How did he miss?
> >The strategy in physics,
> >however, is a conservative one: we throw out something fundamental only
> >when we are quite convinced it yields either a mathematical
> >contradiction or a false prediction, OR if we happen to come up with
> >something better. Also, we never throw out something because it is too
> >simple -- only if it is too simple to fit the known facts. ("As simple
> >as possible, but no simpler.")
>
> I don't suggest that we blindly throw out quantum mechanics. However, the
> applicability of quantum mechanics to the "classical" scale is something
> which does IMHO need to be proven, rather than simply extrapolated. I don't
> believe that the work of the past (70-80 years?) is sufficient for us to
> determine that quantum mechanics really does represent the underlying
> mechanics of reality and not just a computational approximation. And
> i think we need to hold in mind that our theories of reality should, to an
> extent, be considered guilty until proven innocent.
Look up some of the results on superconductivity or on He2. There you
have clearly quantum mechanical effects taking place on a macroscopic
scale.
And by the same reasoning as what you are saying, we need to prove that
gravity works at interstellar distances before using it in theorizing
about galactic structure. For that matter I have yet to prove that
everyone else is not a figment of my imagination. Until I do that I
should not accept any physics...
You see you have to make some sort of assumptions to get along in the
real world, or to do science. One of the standard ones is to assume
that the best existing theories will continue to hold in regions that
they have not been tested in. (Of course that is something that we
would like to test as soon as possible, but we will assume that it is
true until we have contrary evidence.) This is what allows us to
analyze what stars are made of, even though none of us has ever tested
the laws of physics inside of a star.
Similarly there are grounds for accepting the assumptions of the
Everett interpertation until there is evidence against them.
Ben Tilly
Paul Budnik uunet!mtnmath!paul
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> Paul, you are going to have to stop saying what Everett says until you
> find out what Everett says. You are basing your complete understanding
> of Everett's work on Bell's essays. I have read both Everett's
> articles and Bell's essays on Everett, and, as much as I like and
> admire Bell, I think he is *extremely* sloppy in describing and
> criticising Everett.
>
> As far as I know, Everett does not give any preferred status to
> instrument readings. Bell's statements to the contrary are just
> *wrong*. Everett's theory allows one to calculate probabilities for
> *any* observables whatsoever, whether they correspond to instrument
> readings, or whatever.
>
> The only place where instrument readings come into play is in
> Everett's showing that his theory of the universal wavefunction is
> consistent with the observations of physical systems like us, with
> memories. And he succeeds in showing this, more or less.
I suppose I am going to have to read Everett if I keep getting so much
flack from people that think he provides answers to the objection I raise
about QM. Frankly I have a lot more faith in Bell's opinion of Everett
then I do of yours in part because I think Bell has a much deeper intuitive
understanding of these issues than you do.
My suspicion is that Bell is essentially right in his analysis because
you cannot derive events from the linear of the evolution of the wave function.
You have to add something. This something must connect the macroscopic to
the microscopic. Everett may be an excellent mathematician but he cannot
create something in a mathematical model that is not there to start out
with.
> >[...]
> >It does not eliminate the need for two kinds of evolution.
>
> Yes it does. It has only has one kind of evolution, and reproduces the
> same predictions as orthodox quantum mechanics.
But orthodox quantum mechanics can reproduce all the standard predictions
with just linear evolution. Collapse in QM computations is a way to put
the minimal complexity into the mathematics. You need collapse or *something*
to explain what objective events are. You do not need it to predict
probabilities.
>[...]
> Yes it does. It shows why there is no contradiction in experiments
> involving several observers, such as the EPR experiment. In orthodox
> quantum mechanics, there *seems* to be a contradiction in such
> experiments (at least with special relativity). Everett's main results
> show that, even without assuming that observation collapses the wave
> function, the results of repeated experiments by an observer will be
> *as* if the wave function collapsed after each observation. This was a
> real mathematical result, that did not assume *anything* about
> specialness of instrument readings; it just treated an observer as a
> quantum mechanical system in its own right.
I think you are assuming that it is only collapse that implys a conflict
between special relativity and QM. This is not true. It is the linear evolution
in configuration space that is nonlocal. Any realization of probabilities
that evolve in that manner is inconsistent with either causality or
special relativity. You cannot escape this by abolishing collapse.
>[...]
> >Where did these bumpy nonlinear instrument readings come from?
>
> I have no idea what you mean by "nonlinear" in the above sentence, but
> as I said, Everett doesn't treat instruments any differently than atoms
> or electrons, or whatever (except in their complexity).
I am asking how specific (nonlinear) observations come into being from the
linear evolution of the wave function. You need some auxiliary assumptions
beyond standard QM to deduce specific observations. This is not a philosophical
problem. We know macroscopic observations come into being as thermodynamically
irreversible events. There ought to be some way to construct a scientific
theory that relates the microscopic to these observations. This is not
possible without adding something to linear state space evolution.
>[...]
> I am only claiming that one doesn't need nonlinear evolution to
> account for experiment. That was a nontrivial result, and is not at
> all obvious. The version of quantum mechanics that Von Neumann first
> formalized, with the two different kinds of evolution---linear between
> observations and nonlinear after observations---would seem to give
> numerically different results than a theory that assumes that there is
> only linear evolution. Everett showed that it does not (not for any
> experimentally testable cases, anyway). In other words, the assumption
> that the wave function was collapsed by past observations is
> unnecessary for statistical agreement with experiment.
That is fine but it does not allow you to do away with something that
talks about the specific events we all observe as opposed to the linear
evolution of the state space which we only have indirect knowledge of.
I am not sure it is such a great achievement. All it amounts to is saying
that we can replace collapse with multiple constraints on the linear evolution
of the wave function. Instead of collapsing the wave function we use a series of
observations to restrict its future evolution. Since there is no mathematical
content to collapse other then restricting the wave function to be consistent
with the observation one would expect this result. Are you sure
this was not known until Everett's analysis?
> I don't think
> that it is obvious what conclusions should be drawn from this, but it
> does seem that it is premature to be looking for mechanisms for wave
> function collapse when there is no evidence that there is such a thing.
The evidence is the *specific* world we live in. The reason is to connect
the microscopic to the macroscopic using physics instead of metaphysics.
The reason is to explain the *physical* fact that irreversible thermodynamic
processes produce *specific* macroscopic observations. No linear state
space evolution can describe specific events. It can only describe the
probability of observing *any* event or *any* sequence of events.
> >What you can account for is the probability of making observations. Neither
> >Everett no anyone else has a coherent scientific explanation of what
> >observations are.
>
> I guess I agree with that.
Aha we have some common ground!
> >[...]
> >I am amazed at the depths of your rationalization. The world is not
> >linear.
>
> What do you mean by that, and what makes you believe that? What I mean
> by linearity is this: if |Psi1(t)> is a solution to Schrodinger's
> equation, and |Psi2(t)> is a solution, then |Psi1(t)> + |Psi2(t)> is
> also a solution. Any evidence that the world is nonlinear, in this
> sense, would be evidence that Schrodinger's equation is incorrect.
> (Of course, it *is* incorrect, it is only valid for nonrelativistic
> systems, but in the relativistic case, there is a similar linear
> evolution equation for the state.)
I mean that the linear evolution of probabilities cannot account for the
nonliner evolution of events which the world we all know and love (or hate)
is made of. You need to add something to liner evolution to describe these
and that something ought to be physics and not metaphysics. This means
that one should be able to derive the evolution of specific thermodynamic
observations from the evolution of the wave function since that wave function
fully describes the microscopic. A linear theory cannot do this.
>
> >How much evidence for this do you need?
>
> Any evidence at all for the incorrectness of quantum mechanics would
> count. (Experimental evidence, that is---examples of predictions made
> by quantum mechanics that are contradicted by experiment.)
>
Of course that is the only thing that will really upset the apple cart
and force people to face the obvious. I think those experiments will come
in time and they would come a great deal faster if there was less
rationalization in this area.
I am arguing there ought to be a physical way to describe the relationship
between the evolution of the wave function and the evolution of specific
physical events. If you do not accept this then you can argue as so many people
do that QM fully accounts for all observations. I think this is ridiculous
and I think metaphysics in the Copenhagen style or the Everett style is
a poor excuse for the physics that this problem demands.
> >No theory or interpretation of QM explains this nonlinearity. They
> >all simply assume it exists as collapse or observations or instrument
> >readings or whatever.
>
> I don't know what you are talking about. The Everett theory is
> completely linear, so there is no assumption of the existence of
> nonlinearity.
I am talking about the distinction between the linear evolution of the
probability of making an observation and the nonlinear evolution of specific
events. Everett claims to account for specific events. They are not explicable
by a fully linear theory only their probabilities can be accounted for
in such a theory.
>[...]
> Let me make my statement more explicit. I claim that there has never
> been an experiment that has shown that past evolution of the wave
> function was ever nonlinear. Suppose that we know that wave function
> |A(0)> at time 0 evolves into to wave function |A(t)> at time t, and
> that wave function |B(0)> at time 0 evolves into wave function |B(t)>
> at time t. Linearity means that the superposition |A(0)> + |B(0)>
> would evolve into the superposition |A(t)> + |B(t)>. Where is there
> any evidence of a failure of this principle of linearity? (This would
> be an example of the failure of quantum mechanics, since it predicts
> such linearity).
I agree that up to this point no experiment has shown a violation of
this principle. I hope you now have a better understanding of what I
mean when I say the *specific* results of every experiment are nonlinear
and a linear theory can only account for the probability of these results
and not the specific observations.
>[...]
> Be that as it may, it seems that we are in agreement, that the
> evolution of probabilities (probability amplitudes, actually) are
> perfectly described by a linear evolution equation, as far as any
> experimental evidence shows. The extra hypothesis that the wave
> function sometimes undergoes nonlinear evolution is not justified by
> any experimental results.
I think it is the simplest and most logical solution to the problem of
relating the microscopic to the macroscopic. I would say there is
overwhelming evidence for such nonlinear changes because without them
you have to use *all* past observations from the moment of the big
bang to constrain the probabilities you deduce from purely linearly
evolution. This is an onerously complex theory. The assumption that
the wave function undergoes objective nonlinear changes is a dramatic
simplification. The arguments against this position come largely from
the ingrained prejudice that QM as it currently stands is a complete theory.
Beyond this I claim the existing theory is incomplete because it has
no *objective* definition of what an observation is even though observations
play a crucial role in the theory. Nonlinear reversible changes in the wave
function and their evolution to statistically irreversible macroscopic
changes is the obvious solution consistent with what we know of physics
to explain what observations are and how they come into being.
>[...]
> >Is not the switching of a transistor, the conception of a child or
> >the death of a cat a nonlinear effect?
>
> I don't know what you mean by calling these things "nonlinear".
> There is no reason to think (as far as I know) that the probability
> amplitudes for these things evolve nonlinearly.
No but the events do.
> >Do you think there is universal wave function that encodes all the
> >possible ways that everything *might* have developed?
>
> Sure.
Every possible evolution of our universe from the moment of the big bang!
Of course no one know what `every possible' means in this context since
it has no rigorous definition. That is believing an awful lot that is
both ill-defined and not detectable even in principle.
>
> >How do you *rigorously* define what all these possibilities are?
>
> I don't know how to describe babies from first principles starting
> with quantum mechanics, no. However, I think that the only hope is to
> assume that transitions, and babies and cats are made up of elementary
> particles describable by Quantum field theory. Then there are various
> ways of characterizing histories of particles as paths through
> configuration space.
We agree here too. However specific babies have specific histories and
specific transistors make specific state changes. Quantum field theory
only deals with the potential evolution of any possible baby.
> >Do you see that as a simplification? Can you claim this is science
> >when there is not a clear definition of what an observation or
> >instrument reading is?
>
> Quantum mechanics as advocated by Everett doesn't treat observation or
> instrument readings as essentially different from any other kind of
> quantum mechanical event.
What is a *quantum mechanical* *event*? There are *no* events in the linear
evolution of the state function.
> They are all described by probability
> amplitudes.
It is the probability of events that are described by such amplitudes.
The events *themselves* are not.
> As to whether that makes things simple, no it doesn't.
> There is no way to make the physics of 10^23 interacting particles
> simple. But it is *simpler* in principle than an ad hoc theory that
> throws in ill-defined "collapses" for which there is no evidence or
> mechanism or need.
The choice is between collapse in some form or all the possible histories
(whatever that means) from the moment of the big bang. Any way you count
those they are going to make 10^23 seem like a tiny number indeed.
Of course I do not *know* what the solution is. However the evidence for
nonlinear objective changes in the wave function seems overwhelming to me.
Everett's solution suffers both from not defining objectively what observations
are and from introducing enormous and ill-defined complexity to our physical
model.
Paul Budnik
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <ramsay.7...@unixg.ubc.ca>, ram...@unixg.ubc.ca (Keith Ramsay) writes:
> > In physics QM seems uniquely plagued by having "interpretations".
> > Physicists lump many models together under the general rubric of
> > "quantum mechanics", and call specific forms of models
> > "interpretations". What I mean when I refer to Everett's view as an
> > "interpretation" is simply that it is one of the various models which
> > falls under the rubric of "quantum mechanics". Any model, of course,
> > needs to be interpreted before it can be regarded as a theory of
> > physics, in the sense that its connection with natural phenomena has
> > to be given as well as a formal structure.
>
> There is a distinct difference between interpretations in QM and
> interpretations in other physical theories. QM as it is currently
> formulated must treat two separate and irreconcilable realms: the macroscopic
> and the microscopic. These are not linkable within the current theory
> except through arbitrary metaphysical assumptions. All interpretations make
> such assumptions.
>
The two regions need not be "irreconcilable regions", but they are not
reconciled in many interpertations.
> Of course we need this link in some form to apply the theory.
> But we only need an operational link that tells us how to set up
> experiments and how to apply the mathematics to the experiments. None of
> the interpretations help us to do this. They only provide a rationalization
> for what we already understand how to do.
>
That is true for doing experiments in particle physics. That is not
true if we want to use QM as an explanation for things like chemical
properties of substances, or for the electrical properties of
substances and so on.
> > I have trouble making sense of your phrase "uninterpreted QM". Once
> > you are specific enough about which version of QM you are invoking,
> > i.e., once you have a specific model in mind from among the usual menu
> > of possibilities, and once you have attached it to real world
> > predictions, then you already have an "interpretation", in both
> > senses, even if you are not making a big philosophical commitment to
> > it. You have, first, choosen a specific form of QM, and this specific
> > form has to have physical meaning.
>
> By uninterpreted QM I mean the mathematics that describes how the wave
> function evolves and how observations of macroscopic states apply constraints
> on future predictions. One can use this mathematics without assuming collapse,
> Everett or anything beyond the mathematics and the methodology for applying
> it to experiments. None of the interpretations add anything to this model.
> This is what most physicists do when they use QM. They certainly do not
> worry about what some interpretation implys about their experiment or
> analysis.
>
They may not worry about it, but for applications such as the ones that
I give above they certainly do make assumptions which need some
justification if you are going to do it right. Incidentally these are
things which many interpertations do *not* justify, although Everett
does.
> > I see no way to lay claim to particular simplicity by failing to
> > specify your model precisely at all! I don't understand, then, how you
> > conclude then that Everett's version is "more complicated" or "carries
> > excess philosophical baggage". It clearly runs *counter* to
> > philosophical baggage of certain kinds.
>
> It is simple. The model is fully specified in the way I just described.
> What does Everett add to this model?
>
A lot. Look at the above for theoretical things which Everett allows
you to do which would the "uninterperted QM" cannot justify.
> >[...]
> > The complexity of a theory is not to be measured by how extensive its
> > *consequences* appear, but how complex its *premises* are. My usual
> > "model" of the world includes the existence of quite a few physical
> > bodies which don't have any detectible influence on human beings,
> > being too small or too far away, and so on. This is okay, so long as
> > one is not adding complexity to the theory in order to tack them on.
> > Worse also is adding ad hoc assumptions to postulate them away.
>
> But this is what Everett did. The problem interpretations address is
> relating the macroscopic to the microscopic. Everett simply assumes the
> macroscopic exists in the arbitrary from of `instrument readings'.
> This is no different then the Copenhagen interpretation that assumes the
> macroscopic exists in the form of observations. We know that macroscopic
> observations constrain the future evolution of the wave function or at
> least the probabilities we can deduce from it. This is the minimal simple
> assumption.
>
Everett does not assume this. Instead the assumption is that the
physical systems under consideration is a system that can be
effectively described by QM. This is not just an ad hoc assumption, as
I have explained several times. Furthermore it allows us to explain
things about the macroscopic world using QM, which is something that
the Copenhagen interpertation definitely does not do. Therefore there
is a real difference in the extent of what the interpertations allow QM
to be used for, although where both can be used they do agree with each
other.
> > Suppose that we are going down a river, run into a fork, and follow
> > the left hand fork. Someone who says, "Okay, let's say that the right
> > hand fork must dwindle out further downstream" is not making our
> > *model* of the river system simpler; they're just arbitrarily
> > postulating some phenomenon which would prune away the complexity *of
> > the river network*, at the expense of ad hoc patches to hydrodynamics.
> > Similarly someone who asserts, "Okay, now suppose that some nonlinear
> > phenomenon has caused that other portion of the wave function to go
> > away" is not simplifying out *model* of the world. They are just
> > presuming away some detail of the world by postulating another
> > (unverified) phenomenon. Supposing that other branches survive is not
> > adding complexity to the model, not unless there is some positive
> > reason to believe that the other branches die out (which is hard to
> > get, when you are too far away to see them). When they are close at
> > hand, you *can* check that they don't capriciously disappear.
>
> I do not think this analogy applies. We know how hydrodynamic works and
> there is a vast body of experiment that confirms this.
>
The same could be said for QM. If the macroscopc world can indeed be
described by QM then the analogy does hold. There is some evidence for
this claim, at the very least if we assume that it is true and then try
to predict properties of materials there is an agreement with the
assumption that QM is the basic mechanism.
> My biggest problem with Everett is the false claim that he
> explains the relationship between the microscopic and the macroscopic. This
> is not true. The nonliner effects exist in all the interpretations and are
> explained in none of them. It does not matter if you expand the wave function
> in terms of `instrument readings' or you collapse at at `observations'. In
> either case you introducing the macroscopic as a undefined primitive concept.
> It is not derivable or explicable from the physical model.
>
Please stop calling things false unless you can justify what you are
saying. In the Everett interpertation the macroscopic world's
properties follow from what is going on at the microscopic level.
However there is no explicit boundary as to what is macroscopic and
what is microscopic. That is true.
> The simplest model is to say the wave function behaves as if it collapses
> when a thermodynamically irreversible events occur. This is the the working
> assumption of every experimental physicist.
>
What is the theoretical boundary between a reversible event and an
irreversible one? (This ties to the last point I just made.) However in
practice I would agree with you here __IN TERMS OF WHAT WE CAN
OBSERVE__, and is something which can be theoretically justified within
the Everett interpertation. On the other hand I am not sure that every
experimental physicist would say that this is their working assumption.
I am not certain that they would even be able to specify what their
assumption is. However they would agree that if they look at the data
then it is observed. Whether it was actually observed in some sense
before that does not matter to them.
> This fully describes how QM is done and all models must support this
> minimal model in some way. How can adding other philosophical baggage that
> has no experimental consequences be a simplification?
>
1) Not all models *do* support this. If they did then there would not
have been any interest in Schrodingers cat. (To borrow a line from
someone--I think that it was you in fact. :-)
2) Everett does have experimental consequences as explained above.
3) Everett is also a simplification because it manages to make more
direct assumptions than this. Therefore it has a more compact set of
assertions.
4) Everett also removes one piece of philosophical baggage. Above you
are making an assertion about what is true. I remove the assumption
that it is what is real and only assume that it is true as far as we
can observe, which is a logically weaker statement. (However you could
argue that this is taken care of by the "...in some way...". However
your next paragraph suggests that you did not think about this.)
> Physicists feel compelled to invent interpretations because of the implications
> of this minimal model. It suggests that there are objective nonlinear changes
> of the wave function associated with thermodynamically irreversible events.
> If this is true quantum mechanics is an incomplete theory. Horror of horrors
> lets invent an interpretation quick so we do not have to consider this
> possibility!
>
It definitely does not suggest that there is anything nonlinear when
there is a linear model that does the trick!
> > The reason this qualifies as science is that science not only attempts
> > to produce *one* model or account of the phenomena of the world, but
> > the *best* one. Economy is one of the virtues we seek. It is perfectly
> > reasonable as part of science to attempt to produce a theory which
> > explains the very same facts as have been explained before, only with
> > fewer ad hoc assumptions.
>
> I just explained what the most economical model is. I do not believe that
> simplicity is the motivating factor. I think it an antiscientific desire
> to rationalize out of existence what we do not understand. The scientific
> approach is to admit our ignorance and try to improve upon it.
>
In your explanation you made an extra assumption. Furthermore there
were other problems with your analysis as I noted.
> >[...]
> > A theory which predicts undetectible effects is not necessarily
> > "unscientific". There are lots of things in the world which we cannot
> > observe, but which we believe to be there, because the most economical
> > theory fitting our observations of the world includes them. It would
> > require ad hoc assumptions to describe how the world manages to avoid
> > them being included.
>
> What exactly are you thinking about. We assume there are galaxies we
> cannot now observe because we can observe regular patterns
> and we can extrapolate from them. We do not believe in the existence of
> galaxies that are *in principle* unobservable. It is only a matter of having
> a sensitive enough instrument and looking at the right place. This is
> fundamentally different then Everett. Here we take a perfectly good theory
> and add to it the assumption that there is a vast range of effects that
> are *in principle* not detectable. That is metaphysics not science.
>
*In principle* it is impossible for us to simeultaneously observe that
every organ in your body is made up of atoms. However we assume that it
is all of the time. Note however that you are right that there is some
metaphysics there. However there is also a lot of metaphysics there
when I assume that the world is real. (A rather basic assumption...)
> > [...]
> > The essential problem is this: I feel you still have the
> > burden of explaining to us what you mean by "objective". I still
> > cannot see how it differs, if at all, from assuming that nature in the
> > large is classical.
>
> What I mean is what every experimental physicists assumes in his work.
> Thermodynamically irreversible processes are objective events.
>
1) Not all experimental physicists are male. :-)
2) As pointed out before, the rest of this statement is somewhat
dubious. Depending on what "objective" means to you it is actually
explicitly false.
> > Suppose I observe a dead cat. You apparently want to assert that a
^^^^^^^
> > "superposed" observation of a live cat musn't also be included in a
> > (sound) model of such transactions, because of the "objectivity" of
> > the observation.
>
> I think the dead cat speaks for itself. Of course you can assume that it
> will be resurrected and go to cat heaven or you can assume that it is
> superimposed with a live cat. You can assume anything you want to but
> my experience is that dead cats stay dead. Of course you can assume that
> there are things that have no physical and observable effect. This is
> not science. Science only deals with what can be observed at least in
> principle. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^
Oh? What about the theory of ideal gasses. That is considered a part of
science, yet ideal gasses do not exist. What about the atomic theory?
Mach argued against it on the exact same grounds that you are arguing
against Everett, namely that you cannot observe atoms directly. What
science is about is coming up with the best theories (best by certain
standards that is) that explain the observations. However those
theories themselves may have things in them which cannot be observed,
but which do lead to a theory that is better in other ways. Also look
at what you are responding to. I do not think that you answered that
question at all. It does appear that you want to assert that if we
observe a dead cat then a scientific theory should __a priori__ not
include any superpositions with a live cat that we cannot observe.
There are good reasons to disagree with you on this.
> >[...]
> > For this reason, I believe that you are assuming something more by
> > "objectivity" than I am. The main obvious feature of this
> > "objectivity" is that it apparently leads in your thinking to the idea
> > that there aren't superpositions among "objectively observed"
> > phenomena.
>
> Thermodynamically irreversible events are objective. That is how we
> all live our lives and it is how all experimental physics is done.
>
Not true. In living our lives we assume a good deal less than that.
Furthermore irrelevant. In living our lives most of us assume that
Newtonian physics is right. In doing experiments we may be forced to
assume SR, but almost all of us can get by perfectly well w/o assuming
that GR is true. Furthermore we assume that various conservation laws
that are not always true in GR are in fact true. (ie Conservation of
energy.) That does not mean that GR is not a plausible scientific
theory however.
> > Fairly soon we will either run out of useful things to discuss about
> > this "objectivity", or someone will introduce some way to characterize
> > what your notion of "objectivity" is, different from just "that
> > macroscopic superpositions can't happen". So as not to be hasty again
> > in interpreting you, I'll turn it over to you to explain it.
>
> Do you object to the definition I just gave?
I do not disagree with the definition, but from that definition the
properties of objective are different than the properties that you
want. For example there is no contradiction between something
objectively happening one way and your theory saying that it also
happened a different way objectively, even though you would never
observe both things happening. If you understand that then there is no
disagreement between us. However the sort of statement that I just made
is in disagreement with how you want to use the idea of objective
events.
> >
> > |Limiting objective to what is observable is science *period* not
> > |just classical science.
> >
> > I hope that after reading this post you will see that I'm not trying
> > to argue for changing the standards of science. The best theory for
> > explaining the *observed* evidence may indeed have as a consequence
> > that there are unobserved phenomena in addition to observed phenomena.
>
> I certainly think you are. Of course a model may have implications that
> are not observable. This makes those aspects of the model speculative.
> We establish scientific theories by experiments. Aspects of a model
> that are not testable cannot become established science. For example one
> should always keep in mind the experimental envelope in which existing
> theories has been validated and recognize that far outside that envelope
> even the most firmly established theories must be considered speculation.
>
Could you explain to me the evidence that atoms really exist? The
existence of atoms has been hotly debated in the past, and is
definitely not observable. All that we can observe are the consequences
of the model. Then could you contrast it to the situation with
Everett's model?
> >[...]
> > In order for observation to yield "the" value, and for it to be "the"
> > value in all of reality, there has to be a phenomenon occurring which
> > we haven't observed yet, wiping out the other alternatives: one needs
> > a `God-given' phenomenon of "observation" (or some broader type of
> > phenomenon which includes it as a special case), which somehow is
> > different from "non-observation" in that it forces this uniqueness.
>
> It exists and has been extensively tested and described. It is called
> thermodynamic irreversibility. It is the basis of all experimental records
> and all experimental physics. You know that as well as I do. Why do you
> keep talking as if this was not well known and well understood?
> [...]
>
1) Does that wipe out the other alternatives?
2) Not everyone agrees that it has been tested.
3) It is not the basis of all experimental physics.
4) Could you give a precise definition of what is and what is not
irreversible? Please define the boundary _exactly_.
> > |Creating
> > |this philosophical garbage is a poor excuse for avoiding the obvious
> > |conclusion that the wave function undergoes objective nonlinear changes
> > |that we do not yet understand.
> >
> > I'm sorry that you consider our thinking on this topic trashy.
>
> I think it is trashy in a deeply serious way. We all understand that
^^^
> thermodynamically irreversible processes correspond to objective events.
NOT!
> That is how we do experimental physics and it is how we live our lives.
> All this philosophical garbage is to avoid the implications of this
> well understood and universally accepted (*in practice*) fact.
Actually we do not avoid the consequences of this "fact". However it
appears that you think that this "fact" means a lot more than it really
does.
> This
> is an extremely serious business. Instead of facing the implications
> of this physicists have invented a vast web of rationalization.
You call it rationalization, but have not really addressed what makes
it rationalization to you as far as I can see.
> I do think history will come down harshly on this effort. You are using
> the resources that society makes available in a way that does not advance
> our understanding of nature. You are using them to avoid asking the
> tough questions.
Your opinion is noted. My opinion is that you want to reject a
perfectly good answer with no really good reason. I think that instead
of trying to invent (nonexistent) problems, you should try to solve
real problems.
Ben Tilly
Fred Rice
> The important thing about Everett's interpretation as far as I'm
> concerned is that it shows that Eberhard is wrong; one *can't* prove
> that quantum mechanics violates locality, since Everett's
> interpretation is a counterexample with perfectly local evolution and
> the same predictions as orthodox QM.
I understand conceptually how this is true, however, has there
been a paper published on locality in the Everett interpretation
to anyone's knowledge? I'd be interested in reading such a paper.
Thanks,
Fred Rice
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <12OCT199...@csa2.lbl.gov>, sic...@csa2.lbl.gov (SCOTT I CHASE) writes:
> > >
> > >This suggests that common sense is some mental deficiency that we
> > >are cursed with unless we become physicists.
> >
> > No, it is a mental deficiency with which even physicists are burdened.
> > When you stop relying so heavily on your intuition, you will be in
> > a better position to do good physics. Intuition is sometimes a useful
> > tool. But overdependence upon an uneducated intuition can easily
> > lead you to a dead end.
>
> Intuition and common sense are different things. However I am happy to
> repond to the point you are making.
>
> I do not think my intuition as `uneducated'. If I was not able to argue
> intelligently on these issues people would eventually ignore me.
>
As long as Ludwig is getting responses on sci.math I would not rely on
this argument. :-)
> It cuts both ways. One can rely too much on analysis and one can rely too
> much on intuition. Western culture in general has a one sided bias to
> a purely intellectual approach to problem solving and physics is
> one of the most if not the most one sided of disciplines.
>
Have you met some new agers? They agree with you completely on this.
IMO you would be a lot safer to recognize that there is a balance that
has to be reached, and then to put a lot of thought into what that
balance should be.
> The greatest of physicists were not the ones with the most powerful intellect.
> They were men like Bohr and the Einstein whos greatest strength was there
> powerful intuition.
>
Those who were noted for their intuitions were likely also the ones who
were willing to put a lot of thought into training their intuition.
Furthermore they also came up with things that often violated
_everyones_ intuition. At that point they would use their intellect to
back up their intuitions. Think about SR for an example of this. The
final trained intuition was far from the original.
> Neither intuition nor intellect can get you far alone. It is the
> creative functioning of both of them that is central to scientific advance.
Agreed. However there is more. For example there are a set of
philosophical underpinnings that have to be specified to be able to do
science. A good deal of thought has gone into (and will continue to do
go into) deciding what these underpinnings should be. These are crucial
since they have to point the way in deciding between theories which we
cannot tell apart experimentally. Incidentally you can see some of the
things that I think are most important when I put forward my opinions
on why I think that Everett is a better theory.
> Perhaps what you see as my excessive reliance on intuition is largely a
> reflection of the one sided intellectual way physics has developed.
> If there were a little more faith in and reliance on intuition among
> physicists much of the metaphysical garbage that I keep bitching about
> what have been purged from the subject a long time ago. The problems that
> I think are central would have received more attention and perhaps all of
> us and society in general would be better off.
Much of that "metaphysical garbage" is the stuff that I just mentioned.
In your trying to reject it you are actually taking some very strong
positions and then refusing to examine them. For example you will not
accept a theory that has superpositions that we cannot observe. You
reject it out of hand. Why? What are your grounds for doing so? What
assumptions are you making? What justification do you have? By contrast
when I argue for that theory I state my reasons for accepting it. I say
why I make the assumptions that I do, and I justify it with examples of
how similar asssumptions are made all of the time. I do not reject an
assumption as "absurd" w/o giving a good reason why I think that it is
absurd. Now what do you think motivates my underlying reasoning? MY
INTUITION!!!
Ben Tilly
Ron Maimon
Well, Paul seems to have missed this. he didn't respond to it, and I don't know
if he ever will, since it seems I have joined the honored ranks of those killfiled
by Paul. But, it's worth a shot.
Here goes:
Why Bell's analysis is Wrong:
----------------------------
Bell claims that Everett is introducing a new and arbitrary assumption into
quantum mechanics in order to establish collapse, namely the "pointer basis". His
claim is that it is highly arbitrary in what way you split up the universe into
a macroscopic superposition and the way to do it is in no way determined by
quantum mechanics. For example, If I have an electron in a spin eigenstate, say
|+> then I measure it with a device which has a pointer, the pointer should
(if it is a good device) be put into an eigenstate of its position operator.
This means that if we have a pointer which swings left when the electron has
spin up, it should be put into the state "pointer on the left" if the electron
was in the state |+>. If it similarly swings right when the electron is in
the state |-> then if the electron is in the state |-> the pointer should end
up in the state "pointer on the right"
Now, says Bell, if we have the state 1/sqrt(2) ( |+> + |-> ) then the pointer
should end up in the state 1/sqrt(2) ( |right> + |left> ). According to Bell,
Everett says that this is to be interpreted as two universes, distinct and
noninteracting, one in which the pointer is in the state "right" and one in
which the pointer is in the state "left".
But aha! Says Bell, this is where that sneaky Everett devil gets in an extra
hypothesis! We don't have to consider the state 1/sqrt(2) ( |right> + |left> )
as a superposition- I mean it is a state in its own right. Why not say that there
has been no split at all, or that the split is into two universes, one in which
the pointer is in the state
A1 |right> + A2 |left>
and one where it is in the state
B1 |right> + B2 |left>
so long as A1+B1= A2+b2 = 1/sqrt(2) this is allowed. Then if we split the
universe along these lines we again get those eerie macroscopic superpositions.
In other words, Everett's unnatural assumption is that the splitting of the
universes occurs along the eigenstates of the pointer position operator. Different
eigenstates of the pointer correspond to different universes, and this is
arbitrary, unnatural, and just plain ugly.
hence Everett is just as bad as anyone else.
Well this is WRONG.
The reason is that (as many people have mentioned) there is no split of the
universe in the Everett interpretation. The state
1/sqrt(2) ( |right> + |left> )
is no more of a pair of universes then the state 1/sqrt(2) ( |+> + |-> ) of
spin for the electron.
then how come we never see eerie superpositions of position eigenstates?
Why is it that the "pointer basis" just happens to coincide with the basis along
which the universe seems to split?
this is because every human being has a preffered basis associated with him or
her self. This is the "states of mind" basis. The different states of this
basis are the different brain configurations that correspond to different states
of mind, or configurations of thoughts.
Any human being, when thrown into a superposition of states of mind will split
into several people, each of which has a different thought. Where before there
was only one path of mind, after there are several paths. These paths all have
the same memories up until the time of the experiment, and these all believe
different events have occured. This is the basis along which the universe
_subjectively_seems_to_split.
There is a problem with this however- what guarantees that eigenstate of my
state of mind are the same as eigenstates of the pointer position. If this wasn't
the case, then a definite state of my mind would correspond to an eerie neither
here nor there configuration of the pointer.
The answer is, NOTHING. It is perfectly possible to construct a computer with
sensors that respond to certain configurations by changing the internal state,
and these configurations are not necessarily eigenstates of position of a needle.
They might be closer to eigenstates of momentum of the needle. Such a computer
wouldn't see weird neither-here-nor-there needles, it would just "sense" momenta,
and won't be able to say to a very high accuracy where the needle is.
( As John Baez pointed out, this is only true to some approximation, the "states
of mind" aren't really eigenstates of some "state of mind" operator. The way to
think about it is to say that I have a bunch of questions I can ask the person,
and the answer to these questions is quite definite. The states which give a
certain set of answers are those which are "question eigenstates". These are only
approximately orthogonal. But this is of no consequence in the real world. Anyway,
what I have said is true to a _very_ good approximation, since macroscopically
distinguishable states have vanishing overlaps in Hilbert space.)
so why are the eigenstates of our thoughts the same as the position eigenstates
of the needle?
they aren't!
they are only very approximately position eigenstates of the needle.
this can be seen by the fact that when we look at a needle it doesn't start to
jump around erratically, it sort of moves on a smooth trajectory. This means
that when we look at a needle, we don't "collapse" it into a position eigenstate,
we only "collapse" it into an approximate position eigenstate. In Everett's
language, we are becoming correlated with a state that is neither an eigenstate
of the pointer's position, or its momentum, but approximately an eigenstate
to both, constrained by the uncertainty principle. This means that we don't have
such absurdly accurate eyes that can see the location of a pointer with superhigh
accuracy.
If we were determining the _exact_ position of the needle, we would have gamma
ray sensors for eyes and these gamma rays would have enough energy to visibly
jolt the needle whenever we looked at it.
In order to determine exactly what state we are correlated with, or if you like,
the world (subjectively) collapses to. You have to understand the mechanism of
our vision.
A light photon bouncing off a needle in a superposition
1/sqrt(2) ( |right> + |left> )
will bounce into a superpositon of the states |1> or |2> corresponding to
the direction it will get from either state. This same photon may then interact
with our eyes. The way it does this is to impinge upon a certain place in our
retina, and this place is highly sensitive to the direction of the photons
propagation. The response of the pigments in our eyes is both highly localized
in position (within the radius of a cell) and in momentum ( the width
of the aparature of our pupil determines the maximal resolution of our eyes). So
it is not surprising that our pigment excitation states become correlated with
approximate position and approximate momentum eigenstates of the needle. Hence
we see what we see.
If we had a good enough mathematical understanding of our eye we could say
in the Everett interpretation _exacly_ what state we seem to collapse the needle
into. Even lacking such information it is easy to see that we will put it in
a state resembling such states where Newton's laws are seen to hold, and
macroscopic reality emerges.
I realise, of course, that I am speaking as if to stone, because, Paul, you don't
even know enough quantum mechanics to understand why it is that the needle is
put in a superposition in the first place. You then cite lofty sources that say
that it isn't in a superposition when its obvious to any sophomore studying
quantum mechanics that this is the prediction of quantum mechanics. I don't expect
you to understand what I wrote, and I don't expect you to change your mind. I just
don't want you to use Bell's (well intentioned) analysis anymore.
because it is just plain wrong.
with my regards,
Ron Maimon
John C. Baez
Benjamin...@dartmouth.edu (Benjamin J. Tilly) writes:
>One version which would be false (which is an impression that you could
>get as to what people mean) is that existing QM is reality exactly.
>This is actually wrong since we do not have a quantum theory of
>gravity.
Let me take this chance to switch the subject to one that has been
discussed quite as much. :-)
It is interesting to note that it's precisely in the subject of quantum
gravity that "real physicists" are doing the most work on the
interpretational/foundational aspects of quantum theory. It's no
coincidence that Hartle and Gell-Mann put their papers on the preprint
server gr-qc, which stands for general relativity/quantum cosmology.
It's also no coincidence that Penrose has suggested many rather
surprising ideas concerning modifications of quantum mechanics due to
gravitational effects. That's because general relativity poses special
problems for quantum theory.
Why? Let me briefly count the ways, and say a bit more if there is
interest.
1) The wavefunction of the universe. General relativity is closely
connected with cosmology. Quantizing gravity would be necessary to
understand the super-early history of the universe in the big bang
model. For this, one presumably needs to understand the concept of the
"wavefunction of the universe" - or devise an acceptable substitute. If
one takes the notion of the wavefunction of the universe at face value,
it means the way we see the galaxies arranged is only one component of a
superposition!
2) The problem of time. In quantum gravity it appears that the only
way to describe the state of the universe is as a solution to the
diffeomorphism and Hamiltonian constraints (aka Wheeler-DeWitt
equation). If one takes this at face value, it means that the only way
to define time is relative to a choice of a particular clock, where now
a clock can no longer be idealized as a curve through spacetime, but is
a choice of observable like the position of the hand of an actual clock!
How can we describe dynamics this way?
3) The problem of observables. The only observables in quantum gravity
appear to be those invariant under all diffeomorphisms of space and
time. (This is another way of saying what I said at the beginning of
2.) I.e., the only way we can talk about "where" or "when" a field has
some value is by locating it relative to where some other field has a
certain value. We don't know how to describe such observables very
well.
4) The inner product problem. Due to 2) and 3), the usual rules for
choosing the correct inner product on the space of states in quantum
mechanics do not apply. the inner product problem is the problem of
finding the right inner product.
[I can't resist mentioning a raither radical proposal for tackling 4).
One might argue: if the universe is in a single given state,
the whole notion of a Hilbert space of states of the universe is
somewhat suspect. And after all, when we typically apply quantum
mechanics, we apply it not to the universe as a whole, but to some part
of it. So it is possible that what we really need is a Hilbert space of
"partial states" for any given part of the universe. (These are akin to
Everett's "relative states".) Remarkably, the relationship between knot
theory and quantum gravity gives an interesting way to implement these
notions mathematically, that Louis Crane and I have been pursuing from
rather different standpoints.]
5) The "classical domain" problem. Since the universe is quantum
mechanical, it is actually rather remarkable that - at least in certain
ways - it is approximately classical. Given a mathematical description
of a quantum universe, how could we tell which observables seemed to the
"approximately classical" ones?
There is a lot of work going on in all of these areas, and it is rather
a pity that the discussion in sci.physics focuses almost entirely on the
same old Bell's inequality/EPR paradox/Schroedinger cat stuff instead of
this. Of course, everyone needs to think about that stuff until they
feel they understand it well enough. But the people on gr-qc are
working on other things now.
John C. Baez
>One version which would be false (which is an impression that you could
>get as to what people mean) is that existing QM is reality exactly.
>This is actually wrong since we do not have a quantum theory of
>gravity.
Let me take this chance to switch the subject to one that has not been
Keith Ramsay
(john baez) writes:
|(Let me repeat here for the nth time that Everett's work is not to be
|confused with the popularlizations that attempt to make it seem
|far-out.) So while it might be nice to have a model of nature that
|pleases everyday common sense, it is wrongheaded to think that a
|simple appeal to common sense suffices to overthrow quantum mechanics.
...
|After arguing with Budnik for some time (and having given up
|quite some time ago), it became clear to me that this is what his
|arguments boiled down to.
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
|That is wishful thinking on your part. Let me summarize my principle
arguments:
I'm afraid I don't see very much more than appeals to common sense.
For example:
|1. Bell and Eberhard have shown that QM predicts superluminal
|experimentally detectable macroscopic effects. A number of physicists
|(such as Shimony) realize that this means either causality or special
|relativity is false at least for these effects. Baez is not among them
|in part I suspect because he has yet to read Eberhard and Bell.
|I think this prediction is unlikely to be true.
So your common-sense is to believe that "causality", which includes
counterfactual definiteness, and locality, should be true, rather than
expect the Aspect experiment to continue to work the same with the
detectors placed farther apart.
|2. EPR like correlations come from the combination of absolute conservation
|laws and probabilistic laws of observation. It is as if there were a Cosmic
|Accountant that is continually monitoring all experimental results to insure
|that the conservation books balance. I think it likely that there is some
|underlying mechanistic process that explains how this comes about.
I think most of us think of this as a "mechanistic" process.
|I would expect this to be a local process and that is the basis of my
|predictions about what will be observed in effective tests of Bell's
|inequality. I have explained how the class of models that I think should
|be considered *might* produce such a result.
Your explanation of how your "class of models" might produce this
effect strikes me as very tenuous indeed. Of course there ought to be
some way to patch up a model to make it work, but there are assorted
problems. For one, you appear not to be keeping consistency with
relativity: your discrete space-times have preferred directions in
them. Discretizing a continuous model can destroy conservation laws;
it can introduce irrelevant and nearly indetectible new ones; I don't
see why one should expect it to produce helpful new ones, particularly
of the kind which would cause particles to inform quantum correlated
partners what angular momentum to have.
Your "common sense" is to disbelieve in the existence of continuous
quantities. Aside from that, is it plausible to expect that a discrete
model will work better than a continuous one?
|3. QM assumes that probabilities are irreducible. There is no
|mathematical way to define such a model.
Your conception of what it means to "define" a model, is that it not
be a "merely" probabilistic model! Bringing "mathematical" into the
picture is a red herring.
|I think one should always in science look for
|a complete explanation. To say that laws are probabilistic is to say that
|we should not try and find a more complete explanation.
No, I disagree. What it means is that we don't insist that nature
*must* be behaving deterministically. We should be prepared to play
along with nature, should it happen to be playing dice. I don't really
understand the venom with which you approach this possibility:
|To me this is
|a sin against science. It is the equivalent of saying that some things
|are simply God's will and beyond human understanding.
But what is it exactly that you think we *can't* understand about a
probabilistic model?
|That may be true
|but we will never know unless we look for a more complete explanation.
I have no problem with looking for "more complete explanations", but
it should have some solid starting point, to be something more than
grinding air. You dislike the idea that radioactive nuclei decay
randomly. Okay, so you like the idea that they decay due to some
internal structure which we don't know about. I would enjoy that idea
as well. Unfortunately, we don't have any evidence of it. It's not got
anything to do with God at all.
Your common sense appears again to be weighing in a fairly arbitrary
way against a particular kind of model, because you have this sense
that nature has to be determinstic.
|4. There is no satisfactory explanation in existing theory to link the
|macroscopic and the microscopic.
I'll see if I can think of something more to say about your
common-sense intuitions about "the macroscopic" later. For now, the
connection is allegedly the fact that macroscopic bodies are made up
of microscopic parts!
...
|This are a number of experimentally testable consequences.
|This assumption limits the diffusion of the wave function for a single
|particle to the time it would take for that particle to transfer its energy
|to a detector.
I'd certainly be happy to see this kind of experiment taken far enough
further to check this.
Keith Ramsay
ram...@unixg.ubc.ca
Keith Ramsay
|I'm sort of inclined to agree with Paul on this point. It may be that
|the physics of the Everett interpretation never treats consciousness
|as something special, but nevertheless, a consequence of the
|interpretation is that all of the seemingly objective particular
|features of the world---that there is an Earth, that there was a World
|War II, that Bill Clinton was elected President of the US in 1992,
|etc.---turn out to be features, not of the universe, but of one
|particular path through the abstract set of possible histories. The
|"history of the world" seen by me is, according to the Everett
|interpretation, really only *my* history; a different observer,
|performing different experiments (measuring different observables)
|will have a different history.
There are cases in which various kinds of "relativity" of this sort
are embraced, and they don't necessarily require "solipsism".
For example, suppose that you weren't aware that there was more than
one cat in the world, and were given to talking about "the state of
the cat". Then, you discover that other people have other cats, and
that one can have "the cat is asleep" and "the cat is awake",
depending upon "point of view" (i.e., which cat you are talking
about). Likewise with such things as "is moving" and "is stationary".
It is a mistake in relativity to suppose that motion has become
"subjective" because our usual terms like "at rest" have been
relativized.
|Another point to be made is that there is absolutely no observable
|difference (even in principle) between the Everett interpretation and
|the assumption that consciousness collapses the wave function, but
|only *my* consciousness.
This is only true if your consciousness is 100% immune to interference
effects. I think this is liable to turn out to be only a good
approximation.
|The Everett interpretation is equivalent, in
|a certain sense, to starting from this rather solipsistic theory and
|restoring observer-independence by throwing in not only my version of
|history (where my observations collapse the wave function), but also
|every other observer's history. (The neat thing about the technical
|work of Everett is that it shows that all these different histories
|*can* consistently be combined; that each person can assume that *his*
|observations collapse the wave function, and no contradiction arises.)
I prefer to look at it from the other direction (even if not the
historical order of development). The solipsistic view is in some
sense what you get from Everett if you assume that yours is the unique
and only consciousness in the world (and your brain the only one which
produces "consciouness").
Keith Ramsay
ram...@unixg.ubc.ca
Lawrence R. Mead
: In article <CEuEE...@dartvax.dartmouth.edu>
Thank you for the very interesting summary. In order to follow the recent
work/developments/arguements in quantum gravity one must know something
about GR. When i was a graduate student (10-15 yrs ago) GR was not a
required course for anyone except those few (no one in my class of 20
at Washington University, ST. Louis) who intended to work in that area.
My only exposure was one summer when i schlogged through Adler, Bazin
and Schiffer on my own. Perhaps this is why there is less discussion here
in quantum gravity-- the lack of background in GR and lack of familiartiy
with recent developments. I would like to know two things. 1. Are graduate
schools still ignoring 1/4 of the laws of physics by not teaching regular
GR courses? and 2. Could anyone suggest text(s)/readings for someone with
a minimal background (i know quantum/field theory very well) in GR that
will bring me up to date regarding QG, etc. Thanks in advance.
--
Lawrence R. Mead (lrm...@whale.st.usm.edu) | ESCHEW OBFUSCATION !
Associate Professor of Physics
Fred Rice
>>Look, Paul, we have a theory (QM with no collapse) that is in
>>*perfect* agreement with experience---that is, there is absolutely
>>nothing in our experience that is not accounted for by linear QM.
I remember seeing something about not being able to get
chaos out of (linear) QM, and therefore this author arguing that this
violates the correspondence principle. (I only read a part
of that paper though.)
john baez
>Perhaps this is why there is less discussion here
>in quantum gravity-- the lack of background in GR and lack of familiarity
>with recent developments. I would like to know two things. 1. Are graduate
>schools still ignoring 1/4 of the laws of physics by not teaching regular
>GR courses? and 2. Could anyone suggest text(s)/readings for someone with
>a minimal background (i know quantum/field theory very well) in GR that
>will bring me up to date regarding QG, etc. Thanks in advance.
As for 1., I can only report about what's up here at UCR. GR had been
regularly taught, but apparently no longer, since many of the physics
faculty have taken advantage of the good retirement deals the UC system
is offering to cut costs. (For while it seemed I'd be teaching it this
year, but I will instead be teaching general-purpose mathematical
methods in physics.)
As for becoming up-to-date on QG: first, go through the sections in
Wald's "General Relativity" and Misner Thorne and Wheeler's
"Gravitation" on quantum gravity. Then try the following review
articles:
Mathematical problems of non-perturbative quantum general
relativity (lectures delivered at the 1992 Les Houches summer school on
Gravitation and Quantization), by Abhay Ashtekar, 87 pp, Plain TeX,
available as gr-qc/9302024.
Paper: gr-qc/9304012
From: kuc...@mail.physics.utah.edu (karel kuchar)
CANONICAL QUANTUM GRAVITY, Karel Kuchar, Latex, 35pages, UU-REL-92/12/10
Paper: gr-qc/9211019
What can we learn from the study of non-perturbative quantum general
relativity? by Lee Smolin, LATEX 37 pages, no figures.
Then you might check out this very thorough book on one approach (my
favorite) to quantum gravity:
Lectures on Non-perturbative Canonical Gravity, by Abhay Ashtekar,
World Scientific Press, 1991. (ISBN 981-02-0573-2 or for paperback, ISBN
981-02-0574-0. This can be ordered by calling World Scientific at
1-800-227-7562.)
This is a not at all a complete answer to your question, because none of
these talk too much about the information loss problem for black holes,
quantum cosmology, wormholes, etc.. Perhaps someone else can more
easily dig up good references on those subjects.
Ron Maimon
|>
|> Let me take this chance to switch the subject to one that has been
|> discussed quite as much. :-)
|>
|> It is interesting to note that it's precisely in the subject of quantum
|> gravity that "real physicists" are doing the most work on the
|> interpretational/foundational aspects of quantum theory. It's no
|> coincidence that Hartle and Gell-Mann put their papers on the preprint
|> server gr-qc, which stands for general relativity/quantum cosmology.
|> It's also no coincidence that Penrose has suggested many rather
|> surprising ideas concerning modifications of quantum mechanics due to
|> gravitational effects. That's because general relativity poses special
|> problems for quantum theory.
|>
|> Why? Let me briefly count the ways, and say a bit more if there is
|> interest.
|>
|> 1) The wavefunction of the universe. General relativity is closely
|> connected with cosmology. Quantizing gravity would be necessary to
|> understand the super-early history of the universe in the big bang
|> model. For this, one presumably needs to understand the concept of the
|> "wavefunction of the universe" - or devise an acceptable substitute. If
|> one takes the notion of the wavefunction of the universe at face value,
|> it means the way we see the galaxies arranged is only one component of a
|> superposition!
|>
groovy
|> 2) The problem of time. In quantum gravity it appears that the only
|> way to describe the state of the universe is as a solution to the
|> diffeomorphism and Hamiltonian constraints (aka Wheeler-DeWitt
|> equation). If one takes this at face value, it means that the only way
|> to define time is relative to a choice of a particular clock, where now
|> a clock can no longer be idealized as a curve through spacetime, but is
|> a choice of observable like the position of the hand of an actual clock!
|> How can we describe dynamics this way?
|>
I don't understand this. There is no problem in the semi-classical theory, why
should there be a problem in the full quantum theory?
As far as I can see, you can (semiclassical gravity) label every point with four
numbers. This is a definite association. A clock, on the other hand, may not
indicate coordinate time, but may be in a superposition of different hand states,
and therefore indicate a different "clock time" for each superposed observer
looking at the clock.
If you are quatizing g, the metric, you can see that a superposition of (visibly
different) g's will lead to a superposition of visibly different clock-hand
positions, and this now gives a relative measurement of proper time, but time,
the coordinate label, is preserved as a number.
So I have no motivation to consider time as an operator, whatever that might mean.
But then again, perhaps I should ask a better question:
What in the world is the Wheeler-DeWitt equation?
|> 3) The problem of observables. The only observables in quantum gravity
|> appear to be those invariant under all diffeomorphisms of space and
|> time. (This is another way of saying what I said at the beginning of
|> 2.) I.e., the only way we can talk about "where" or "when" a field has
|> some value is by locating it relative to where some other field has a
|> certain value. We don't know how to describe such observables very
|> well.
|>
why can't you talk about where or when a field has a value relative to the
coordinates, and then establish coordinate invariance?
|> 4) The inner product problem. Due to 2) and 3), the usual rules for
|> choosing the correct inner product on the space of states in quantum
|> mechanics do not apply. the inner product problem is the problem of
|> finding the right inner product.
|>
|>
|> [I can't resist mentioning a raither radical proposal for tackling 4).
|> One might argue: if the universe is in a single given state,
|> the whole notion of a Hilbert space of states of the universe is
|> somewhat suspect. And after all, when we typically apply quantum
|> mechanics, we apply it not to the universe as a whole, but to some part
|> of it. So it is possible that what we really need is a Hilbert space of
|> "partial states" for any given part of the universe. (These are akin to
|> Everett's "relative states".) Remarkably, the relationship between knot
|> theory and quantum gravity gives an interesting way to implement these
|> notions mathematically, that Louis Crane and I have been pursuing from
|> rather different standpoints.]
|>
This doesn't sound so radical- it sounds correct and reasonable. Is there a
reason to think it might not be true?
BTW, what is the precise problem?
|>
|> 5) The "classical domain" problem. Since the universe is quantum
|> mechanical, it is actually rather remarkable that - at least in certain
|> ways - it is approximately classical. Given a mathematical description
|> of a quantum universe, how could we tell which observables seemed to the
|> "approximately classical" ones?
|>
|>
|> There is a lot of work going on in all of these areas, and it is rather
|> a pity that the discussion in sci.physics focuses almost entirely on the
|> same old Bell's inequality/EPR paradox/Schroedinger cat stuff instead of
|> this. Of course, everyone needs to think about that stuff until they
|> feel they understand it well enough. But the people on gr-qc are
|> working on other things now.
|>
agreed!
:)
Ron Maimon
Ron Maimon
|>
|> |I think one should always in science look for
|> |a complete explanation. To say that laws are probabilistic is to say that
|> |we should not try and find a more complete explanation.
|>
|> No, I disagree. What it means is that we don't insist that nature
|> *must* be behaving deterministically. We should be prepared to play
|> along with nature, should it happen to be playing dice. I don't really
|> understand the venom with which you approach this possibility:
|>
Well, I hate to agree with Paul on this one, but I don't see any way in which
a probabilistic model can be fundamental.
Every probabilistic model admits an "Everettization" in which it is a
deterministic model with "universe branching". However, for most theories,
this is a stupidly trivial operation, which does indeed lead to actual
universe splitting and time irreversibility.
for example, a _truly_random_ Everett coin toss which is observed leads to two
observers in two universes, one who observed heads and one which observed tails,
and they are truly and completely invisible to one another, forever, even in
principle.
The beautiful thing about quantum mechanics, is that it's Everettization is not
a stupidly trivial operation, and it doesn't lead to actual splitting of the
universe, but only to an apparent splitting, that is time reversible in essence.
A man observing a quantum coin toss will be put into a superposition of different
people that is for all practical purposes two uninteracting people, but that's
just an approximation that's true for short times. If I wait for the poincare
recurrence time of the system, I will find that the two people (long since
dead) will find their wavefunctions recombining into one "world" and reformng
the untossed coin.
This is a deterministic model, just weirdly deterministic.
If we were to find out that the laws of physics are not deterministic, I would
just say that we have discovered that the laws of physics are deterministic in
an Everett sense, but that the branching of the world is actual, and not just
apparent- and the process is time irreversible.
Then the choice between the "Everett style" interpretation or "collapse style"
interpretation would be a matter of philosophy. Unlike quantum mechanics, they
would be identical, not only in practice, but in consistency and in principle.
I would still use the Everett style description, but that's for philosophical,
not physical reasons.
|> |To me this is
|> |a sin against science. It is the equivalent of saying that some things
|> |are simply God's will and beyond human understanding.
|>
and these are the philosophical reasons, as Paul so (shudder) correctly
expressed.
|> But what is it exactly that you think we *can't* understand about a
|> probabilistic model?
|>
What is it that makes one possibility any more "actual" than another, when they
are both mathematical objects with equivalent mathematical existence. It would
say that there is more to the word "existence" than can be described with
the idea of mathematical existence, and would rule out the possibilty of finding
a mathematical model which _completely_ models the universe, either "exactly" or
(equivalently) to as high an accuracy as desired, since such a model would
not distinguish between say, the observation of a heads coin toss or a tails
coin toss and so it cannot say which of the two is "actual". This leaves the
world as a mystery forever, and that's extremely unneccesary, and anappatizing.
Paul's only problem is that he doesn't realize that quantum mechanics doesn't
do this horrible thing. In Quantum Mechanics, probabilities are not irreducible
objects- they arise from the dynamics and the inner product structure, and
they are only approximate descriptions of the world.
A complete model is simply one in which every "thing" in the universe has a
mathematical counterpart, and we can tell which "things" in the model have a
physical counterpart.
I would say the easy part is finding the model, it is just the set of all
(finite) mathematical objects. The hard part is finding which "things" in the
model have a physical counterpart.
But, unfortunately, I am level headed enough to know that this is not physics.
So I stop.
Ron Maimon
john baez
>In article <1993Oct13.2...@galois.mit.edu>, jb...@riesz.mit.edu (John C. Baez) writes:
>|> 2) The problem of time. In quantum gravity it appears that the only
>|> way to describe the state of the universe is as a solution to the
>|> diffeomorphism and Hamiltonian constraints (aka Wheeler-DeWitt
>|> equation). If one takes this at face value, it means that the only way
>|> to define time is relative to a choice of a particular clock, where now
>|> a clock can no longer be idealized as a curve through spacetime, but is
>|> a choice of observable like the position of the hand of an actual clock!
>|> How can we describe dynamics this way?
>I don't understand this. There is no problem in the semi-classical theory, why
>should there be a problem in the full quantum theory?
I'm not quite sure what you mean by the semiclassical theory. Anyway,
there are many approaches to QG, and while this problem can be avoided
in some, they have comparable problems - it's a bit like a lumpy rug.
I'll say a bit more on this after your next question, but mainly
let me suggest you read Chris Isham's big fat juicy review article,
Canonical Quantum Gravity and the Problem of Time. This is
gr-qc/9210011, which should still be available in compressed form
somehow. He goes through it all much better than I could. Anyway, I
think you are right that a better question is:
>What in the world is the Wheeler-DeWitt equation?
This is the basic equation of quantum gravity in the canonical approach.
Let me quote myself. This is from "week11" of my series of This Week's
Finds; sorry if it starts off with some stuff you probably know:
Both the technical problems of "canonical" quantum gravity and one of the
main conceptual problems - the problem of time - stem from the fact that
general relativity is a system in which the initial data have
constraints. So improving our understanding of quantizing constrained
classical systems is important in understanding quantum gravity.
Let me say a few words about these constraints and what I mean by
"canonical" quantum gravity.
First consider the wave equation in 2 dimensions. This is
an equation for a function from R^2 to R, say phi(t,x), where t is a
timelike and x is a spacelike coordinate. The equation is simply
d^2 phi/dt^2 - d^2phi/dx^2 = 0.
Now this equation can be rewritten as an evolutionary equation for
initial data as follows. We consider pairs of functions (Q,P) on R -
which we think of phi and d phi/dt on "space", that is, on a surface t =
constant. And we rewrite the second-order equation above as a
first-order equation:
d/dt (Q,P) = (P, d^2Q/dx^2). 1)
This is a standard trick. We call the space of pairs (Q,P) the "phase
space" of the theory. In canonical quantization, we treat this a lot
like the space R^2 of pairs (q,p) describing the initial position and
momentum of a particle. Note that for a harmonic oscillator we have an
equation a whole lot like 1):
d/dt (q,p) = (p, -q).
This is why when we quantize the wave equation it's a whole lot like the
harmonic oscillator.
Now in general relativity things are similar but more complicated.
The analog of the pairs (phi, d phi/dt) are pairs (Q,P) where Q is the
metric on spacetime restricted to a spacelike hypersurface - that is,
the "metric on space at a given time" - and P is concocted from the
extrinsic curvature of that hypersurface as it sits in spacetime.
Now the name of the game is to turn Einstein's equation for the metric
into a first-order equation sort of like 1). The problem is, in general
relativity there is no god-given notion of time. So we need to *pick* a
"lapse function" on our hypersurface, and a "shift vector field" on our
hypersurface, which say how we want to push our hypersurface forwards in
time. The lapse function says at each point how much we push it in the
normal direction, while the shift vector field says at each point how
much we push it in some tangential direction. These are utterly
arbitrary and give us complete flexibility in how we want to push the
hypersurface forwards. Even if spacetime was flat, we could push the
hypersurface forwards in a dull way like:
-------------------- new
____________________ old
or in a screwy way like
----
/ \ /\
/ --- \
------- new
____________________ old
Of course, in general relativity spacetime is usually not flat, which
makes it ultimately impossible to decide what counts as a "dull way" and
what counts as a "screwy way," which is why we simply allow all possible
ways.
Anyway, having *chosen* a lapse function and shift vector field, we can
rewrite Einstein's equations as an evolutionary equation. This is a bit
of a mess, and it's called the ADM (Arnowitt-Deser-Misner) formalism.
Schematically, it goes like
d/dt (Q,P) = (stuff, stuff'). 2)
where both "stuff" and "stuff'" depend on both Q and P in a pretty
complex way.
But there is a catch. While the evolutionary equations are equivalent
to 6 of Einstein's equations (Einstein's equation for general relativity
is really 10 scalar equations packed into one tensor equation), there
are 4 more of Einstein's equations which turn into *constraints* on Q
and P. 1 of these constraints is called the Hamiltonian constraint and
is closely related to the lapse function; the other 3 are called the
momentum or diffeomorphism constraints and are closely related to the
shift vector field.
For those of you who know Hamiltonian mechanics, the reason why the
Hamiltonian constraint is called what it is is that we can write it as
H(Q,P) = 0
for some combination of Q and P, and this H(Q,P) acts a lot like a
Hamiltonian for general relativity in that we can rewrite 2) using the
Poisson brackets on the "phase space" of all (Q,P) pairs as
d/dt Q = {P,H(Q,P)}
d/dt P = {Q,H(Q,P)}.
The funny thing is that H is not zero on the space of all (Q,P) pairs,
so the equations above are nontrivial, but it does vanish on the submanifold
of pairs satisfying the constraints, so that, in a sense, "the
Hamiltonian of general relativity is zero". But one must be careful in
saying this because it can be confusing! It has confused lots of people
worrying about the problem of time in quantum gravity, where they
naively think "What - the Hamiltonian is zero? That means there's no
dynamics at all!"
The problem in quantizing general relativity in the "canonical" approach
is largely figuring out what to do with the constraints. It was Dirac
who first seriously tackled such problems, but the constraints in
general relativity always seemed intractible (when quantizing) until
Ashtekar invented his "new variables" for quantum gravity, that all of a
sudden make the constraints look a lot simpler. Ashtekar also has
certain generalizations of Dirac's general approach to quantizing
systems with constraints, and part of what Tate (who was a student of
Ashtekar) is doing is to study a number of toy models to see how
Ashtekar's ideas work.
[This was a "review" of:
An algebraic approach to the quantization of constrained systems:
finite dimensional examples, by Ranjeet S. Tate, Syracuse University
physics department PhD dissertation, August 1992, SU-GP-92/8-1. (Tate
is now at rst...@cosmic.physics.ucsb.edu, but please don't ask him for
copies unless you're pretty serious, because it's big.)]
---------------------------------------------------------------------------
Okay. The above was all really classical GR. When we quantize we have
to deal with these constraints and there are two basic avenues: quantize
before constraining, or quantize after constraining. If we quantize
before constraining the Hamiltonian constraint H(Q,P) gets promoted to
a constraint H(Q,P)Psi = 0 on physical states, where now Q and P hence
H(Q,P) are *operators* and Psi is a function on the configuration space
of the classical theory (e.g., the space of all metrics on a 3-manifold,
"space".) This is the Wheeler-DeWitt equation. It plays a role in
quantum gravity comparable to the equation i (d/dt) Psi = H Psi for
plain old quantum theories. It is also known as the Hamiltonian
constraint. There is another constraint, the diffeomorphism constraint.
>why can't you talk about where or when a field has a value relative to the
>coordinates, and then establish coordinate invariance?
Good question. This is of course really the same question as you
raised before. Maybe the following will help - even though it's really
the same answer that I gave before! If you try to do what you want,
note that it is impossible to predict the value of the fields at time t
= T from their values at t = 0. This is because there are
diffeomorphisms mapping the surface t = T to any other surface t = T'
(T,T' > 0) and leaving a neighborhood of t = 0 alone, and the theory is
diffeomorphism-invariant. So while one can
imagine succeeding in doing what you want, you will still not have
equations of the usual "tell me what's going on now, I'll tell you
what's going on at t = T" sort. So the problem of time is still there
in a different guise.
>|> [I can't resist mentioning a raither radical proposal for tackling 4).
>|> One might argue: if the universe is in a single given state,
>|> the whole notion of a Hilbert space of states of the universe is
>|> somewhat suspect. And after all, when we typically apply quantum
>|> mechanics, we apply it not to the universe as a whole, but to some part
>|> of it. So it is possible that what we really need is a Hilbert space of
>|> "partial states" for any given part of the universe. (These are akin to
>|> Everett's "relative states".) Remarkably, the relationship between knot
>|> theory and quantum gravity gives an interesting way to implement these
>|> notions mathematically, that Louis Crane and I have been pursuing from
>|> rather different standpoints.]
>This doesn't sound so radical- it sounds correct and reasonable. Is there a
>reason to think it might not be true?
Again this is a big argument. *I* think it's correct, reasonable and
perhaps even true, so I am not the best source of objections.
>BTW, what is the precise problem?
Sorry, I'm spending too much time on this post already - maybe later?
In response to another post on this topic I listed some good places to
read about this stuff. The review article by Smolin might be a good one
for this particular issue.
Paul Budnik uunet!mtnmath!paul
> > I am waiting for *anyone* to answer these arguments.
>
> I thought that I already *had* answered them. However post them again
> and *read* what I write carefully.
Many of your answers have been confused and repeated statements. Repeating
the same thing over and over does not make it true. You do not understand
these issues nearly as well as you think. Admittedly I repeat many things
also but I try to limit my repetition in reponding to *each* individual.
In the forum of the net repetition when communicating with different people
is inevitable.
I did post these arguments in response to one of your posts and you did
not as far as I can recall respond to them specifically. You are
welcome to do so now.
Bell compares Everett's model unfavorably with De Brogiles pilot wave
model (the precursor to Bell's model). I am paraphrasing Bell since I
do not want to type in his full comments see `Speakable and unspeakable
in quantum mechanics', Cambidge, pg 96.)
1. Everett's special variables are anthroprocentric whereas de Broigle's
are objective.
2. Everett assumes all configurations are realized at any time (the
source of many World's I imagine). de Broigle has a *particular*
configuration.
3. Everett makes at best a half hearted attempt to link configuration
space to continuous trajectories in physical space (my beloved events).
De Broigle defines this link.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> Well, Paul seems to have missed this. he didn't respond to it, and I don't know
> if he ever will, since it seems I have joined the honored ranks of those killfiled
No Ron I did not miss it. I do not have a kill file.
> by Paul. But, it's worth a shot. [...]
There is little point in responding to someone who claims such a deep and clear
understanding of the problems that have been plaguing the best minds in
physics for decades. My first reaction to some of your comments was:
Sigh! I cannot even remember when I was that young and naive. Then I realized
I never was *that* naive.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> > [...]
> > Of course we need this link [macroscopic an microscopic]
> > in some form to apply the theory. [...]
> >
> That is true for doing experiments in particle physics. That is not
> true if we want to use QM as an explanation for things like chemical
> properties of substances, or for the electrical properties of
> substances and so on.
Gee I did not realize Everett was needed to apply QM to chemistry or
condensed matter physics. I doubt that the physicists who work in those areas
realize it either.
>[...]
> What is the theoretical boundary between a reversible event and an
> irreversible one? [...]
There is no *absolute* boundary. It is a statistical boundary. Everything
can be reversed but many events such as the demise of Schrodinger's cat
are so unlikely to be reversed that they can be considered absolutely
irreversible.
> > Science only deals with what can be observed at least in
> > principle. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> ^^^^^^^^^
> Oh? What about the theory of ideal gasses. That is considered a part of
> science, yet ideal gasses do not exist. [...]
You should think more carefully before you write. Physics uses many
mathematical models that are known to be only approximate. Such models
are used for a variety of practical and tutorial reasons. These models
are *not* established scientifc theory although they are part of the
mathematical machinery of science and may play an important role in
formulating established theory.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> There is a lot of work going on in all of these areas, and it is rather
> a pity that the discussion in sci.physics focuses almost entirely on the
> same old Bell's inequality/EPR paradox/Schroedinger cat stuff instead of
> this. Of course, everyone needs to think about that stuff until they
> feel they understand it well enough. But the people on gr-qc are
> working on other things now.
Of course if they are mistaken that physics has an acceptable solution to
the problems presented by EPR and Bell's inequality all this work may turn
out to be excellent mathematics but it is unlikely to have much physical
content. This is because the basic assumptions under which they are
working are physically false. Maybe its a shame that so much effort is
being put into projects that are doomed from the start.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> For example:
>
> |1. Bell and Eberhard have shown that QM predicts superluminal
> |experimentally detectable macroscopic effects. A number of physicists
> |(such as Shimony) realize that this means either causality or special
> |relativity is false at least for these effects. Baez is not among them
> |in part I suspect because he has yet to read Eberhard and Bell.
>
> |I think this prediction is unlikely to be true.
>
> So your common-sense is to believe that "causality", which includes
> counterfactual definiteness, and locality, should be true, rather than
> expect the Aspect experiment to continue to work the same with the
> detectors placed farther apart.
We have fundamental principles in conflict. One is locality. The other
is the existing formulation of QM. It is a matter of judgement as to
which you have more faith in. However this is a conflict between physical
principles *both* of which are supported by an enormous body of
experimental evidence. There is no experimental evidence that conflicts
with *either* principle. At the least doing these experiments properly
should command a high priority.
>
> |2. EPR like correlations come from the combination of absolute conservation
> |laws and probabilistic laws of observation. It is as if there were a Cosmic
> |Accountant that is continually monitoring all experimental results to insure
> |that the conservation books balance. I think it likely that there is some
> |underlying mechanistic process that explains how this comes about.
>
> I think most of us think of this as a "mechanistic" process.
It cannot be a local mechanistic process and there is no mechanism for
it in the existing theory. There is a mechanism to compute probabilities
that preserves conservation laws through nonlocal effects. There is no
mechanism for generating the specific results or observations.
> |I would expect this to be a local process and that is the basis of my
> |predictions about what will be observed in effective tests of Bell's
> |inequality. I have explained how the class of models that I think should
> |be considered *might* produce such a result.
>
> Your explanation of how your "class of models" might produce this
> effect strikes me as very tenuous indeed. Of course there ought to be
> some way to patch up a model to make it work, but there are assorted
> problems. For one, you appear not to be keeping consistency with
> relativity: your discrete space-times have preferred directions in
> them.
This is true and it is what bothered me the most when I first started
thinking about this class of models. I was extremely pleased when I first
ran across that quote from Einstein that I love to repeat because he seemed
to reach the same general conclusions as I did but even more so because
he recognized and accepted that this meant relativity could only be
approximately true.
I consider it quite possible that physics cannot be based on the
field concept, i. e., on continuous structures. In that case
*nothing* remains of my entire castle in the air gravitation
theory included, [and of] the rest of modern physics.
-- Einstein in a 1954 letter to Besso, quoted from:
"Subtle is the Lord", Abraham Pais, page 467.
> Discretizing a continuous model can destroy conservation laws;
In *must* destroy conservation laws if you discretize the field function.
However they are reintroduced in a global form by the repetition of states
effect I described. The destruction of the conservation laws at the local
level is a possible source of quantum uncertainty.
> it can introduce irrelevant and nearly indetectible new ones; I don't
> see why one should expect it to produce helpful new ones, particularly
> of the kind which would cause particles to inform quantum correlated
> partners what angular momentum to have.
I do not think the singlet state is something that exists in nature.
I think it only exists in human theory. I think it is an artifact of a
complex quantum collapse process extended in time and space. It is the
structure of this process that produces correlations that correspond to
the singlet state model.
I agree that it is a big step from my model
to generate these predictions. I do not have the kind of argument for
my model that will impress many physicists. It is the multitude of intuitive
connections between aspects of the model and known physical results that makes
me think this class of models has a good chance to work out in practice.
>
> Your "common sense" is to disbelieve in the existence of continuous
> quantities. Aside from that, is it plausible to expect that a discrete
> model will work better than a continuous one?
I think there are excellent philosophical and mathematical arguments to
doubt the existence of any *completed* infinite totalities or infinitesimal
structures. This is something I have thought a great deal about and I have a
done some considerable research into generalized recursion theory motivated
by these thoughts. It is much more than common sense. Virtually all of
mathematics can be formulated in terms of properties of Turing Machines
in a potentially infinite but nondeterministic (in the sense of following
multiple paths not of being probabilistic) universe. In such a universe
we can ask: will a species have an infinite number of descendant species.
The general formulation of this (if species are treated as Turing Machines)
requires quantification over the reals. The only mathematics that cannot
be cast in this light seems to be questions like the continuum hypothesis.
Those are question which I do not think are objectively either true or false.
I think such questions refer to different equally valid ways of extending
a mathematical model.
> |3. QM assumes that probabilities are irreducible. There is no
> |mathematical way to define such a model.
>
> Your conception of what it means to "define" a model, is that it not
> be a "merely" probabilistic model! Bringing "mathematical" into the
> picture is a red herring.
Not so. There is no mathematical way to formulate irreducible probabilities.
This is not my prejudice it is a fact of our current understanding of
mathematics.
> |I think one should always in science look for
> |a complete explanation. To say that laws are probabilistic is to say that
> |we should not try and find a more complete explanation.
>
> No, I disagree. What it means is that we don't insist that nature
> *must* be behaving deterministically. We should be prepared to play
> along with nature, should it happen to be playing dice. I don't really
> understand the venom with which you approach this possibility:
Of course we should but how do we *know* nature is playing dice. Maybe
she is just more clever then we are and she only appears to be playing
dice. If we give up too easily we will never find out. Since we have
no way of knowing how clever nature may be in fooling us I do not see
how we can determine *any* objective criteria for saying that are now
convinced nature is playing dice. We may suffer from having too much
faith in our own cleverness and too little comprehension of the
cleverness in nature.
>
> |To me this is
> |a sin against science. It is the equivalent of saying that some things
> |are simply God's will and beyond human understanding.
>
> But what is it exactly that you think we *can't* understand about a
> probabilistic model?
Nothing. What we can't understand is why a *specific* realization of
those probabilities happens in a *specific* experiment.
>
> |That may be true
> |but we will never know unless we look for a more complete explanation.
>
> I have no problem with looking for "more complete explanations", but
> it should have some solid starting point, to be something more than
> grinding air. You dislike the idea that radioactive nuclei decay
> randomly. Okay, so you like the idea that they decay due to some
> internal structure which we don't know about. I would enjoy that idea
> as well. Unfortunately, we don't have any evidence of it. It's not got
> anything to do with God at all.
We will be unlikely to find the evidence if we do not look for it. I have
no objections to a scientist saying I do not have any idea how to attack
that problem so I will attack another problem. I strenuously object to
a scientist saying I have no idea how to attack that that problem and
none of my eminent colleagues have any idea how to attack therefore
there must not exist any way to attack it.
>
> Your common sense appears again to be weighing in a fairly arbitrary
> way against a particular kind of model, because you have this sense
> that nature has to be deterministic.
Not at all. I am weighing in against arbitrarily asserting that a model
is complete because one cannot conceive of how a more complete model
might be constructed.
>
> |4. There is no satisfactory explanation in existing theory to link the
> |macroscopic and the microscopic.
>
> I'll see if I can think of something more to say about your
> common-sense intuitions about "the macroscopic" later. For now, the
> connection is allegedly the fact that macroscopic bodies are made up
> of microscopic parts!
Ah! that is the problem. The connection between them in existing theory
is not that a macroscopic object is made up of microscopic parts. Rather it
is that a wave function allows us to compute the probability of a macroscopic
observation. The two domains, that of the wave function and that of macroscopic
observations, are both assumed to exist prior to each other. Physics only
says how they relate. It does not say how to construct one from the other.
> ...
> |This are a number of experimentally testable consequences.
> |This assumption limits the diffusion of the wave function for a single
> |particle to the time it would take for that particle to transfer its energy
> |to a detector.
>
> I'd certainly be happy to see this kind of experiment taken far enough
> further to check this.
>
Wonderful. I hope I can convince more physicists that it worth taking such
experiments seriously.
Paul Budnik
Keith Ramsay
|Of course if they are mistaken that physics has an acceptable solution to
|the problems presented by EPR and Bell's inequality all this work may turn
|out to be excellent mathematics but it is unlikely to have much physical
|content. This is because the basic assumptions under which they are
|working are physically false. Maybe its a shame that so much effort is
|being put into projects that are doomed from the start.
I've concluded that it is almost impossible to avoid "waste". If one
tries to explain to someone why their ideas are a waste of time,
generally they just get upset at you.
Keith Ramsay
ram...@unixg.ubc.ca
Benjamin J. Tilly
rma...@husc9.Harvard.EDU (Ron Maimon) writes:
> The beautiful thing about quantum mechanics, is that it's Everettization is not
> a stupidly trivial operation, and it doesn't lead to actual splitting of the
> universe, but only to an apparent splitting, that is time reversible in essence.
>
> A man observing a quantum coin toss will be put into a superposition of different
> people that is for all practical purposes two uninteracting people, but that's
> just an approximation that's true for short times. If I wait for the poincare
> recurrence time of the system, I will find that the two people (long since
> dead) will find their wavefunctions recombining into one "world" and reformng
> the untossed coin.
>
Looooooong since dead. The recurrence time of this system is of the
same order of magnitude as the time to come to equilibrium. That seems
to be taking a while...:-)
> This is a deterministic model, just weirdly deterministic.
Exactly.
Ben Tilly
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
The comments that you started with are always a possible sort of
problem in science. However historical examples suggest that a
suprising amount of the ideas that people worked with before the
theories changed actually can be salvaged. (Look at what happened to
classical thermodynamics when QM came in.)
But I take objection to the second statement. Why is it doomed? You
might think that it is, but can you really justify that statement? If
you cannot, then you should not say it out in the open where you will
be asked to do so. Not unless you want people thinking that you do not
know what you are talking about.
Ben Tilly
Dan S.
>We have fundamental principles in conflict. One is locality. The other
>is the existing formulation of QM. It is a matter of judgement as to
>which you have more faith in. However this is a conflict between physical
>principles *both* of which are supported by an enormous body of
>experimental evidence. There is no experimental evidence that conflicts
>with *either* principle. At the least doing these experiments properly
>should command a high priority.
There is no experimental evidence for counterfactual definiteness in
the microscopic domain. In fact, I would say the entire concept is
quite foreign to quantum formalism.
>>
>> |2. EPR like correlations come from the combination of absolute conservation
>> |laws and probabilistic laws of observation. It is as if there were a Cosmic
>> |Accountant that is continually monitoring all experimental results to insure
>> |that the conservation books balance. I think it likely that there is some
>> |underlying mechanistic process that explains how this comes about.
>>
>> I think most of us think of this as a "mechanistic" process.
>It cannot be a local mechanistic process and there is no mechanism for
>it in the existing theory. There is a mechanism to compute probabilities
>that preserves conservation laws through nonlocal effects. There is no
>mechanism for generating the specific results or observations.
That is true. There is no mechanism within quantum theory to account for
the fact that events happen at all, local or not. Even Everett fails to
accomplish this. This is called the quantum measurement problem, and
until we understand better what happens in a measurement I feel we are
not going to get anywhere on understanding the behavior of an EPR state.
>I do not think the singlet state is something that exists in nature.
>I think it only exists in human theory. I think it is an artifact of a
>complex quantum collapse process extended in time and space. It is the
>structure of this process that produces correlations that correspond to
>the singlet state model.
Neither did Einstein. But, in 1957 Aharonov and Bohm did an experiment to
see if the singlet state would indeed break down if the photons were
sufficiently separated. They're conclusion was that the singlet state
remained a singlet state regardless what the separation between the
two particles is.
>> |I think one should always in science look for
>> |a complete explanation. To say that laws are probabilistic is to say that
>> |we should not try and find a more complete explanation.
>>
>> No, I disagree. What it means is that we don't insist that nature
>> *must* be behaving deterministically. We should be prepared to play
>> along with nature, should it happen to be playing dice. I don't really
>> understand the venom with which you approach this possibility:
>Of course we should but how do we *know* nature is playing dice. Maybe
>she is just more clever then we are and she only appears to be playing
>dice. If we give up too easily we will never find out. Since we have
>no way of knowing how clever nature may be in fooling us I do not see
>how we can determine *any* objective criteria for saying that are now
>convinced nature is playing dice. We may suffer from having too much
>faith in our own cleverness and too little comprehension of the
>cleverness in nature.
I don't think anyone is saying we shouldn't investigate the possibility
of a non-probabilistic model. What most physicists do say is that it
appears that nature is inherently probabilistic.
>>
>> |To me this is
>> |a sin against science. It is the equivalent of saying that some things
>> |are simply God's will and beyond human understanding.
>>
>> But what is it exactly that you think we *can't* understand about a
>> probabilistic model?
>Nothing. What we can't understand is why a *specific* realization of
>those probabilities happens in a *specific* experiment.
Exactly. The quantum measurement problem again.
>> |4. There is no satisfactory explanation in existing theory to link the
>> |macroscopic and the microscopic.
>>
>> I'll see if I can think of something more to say about your
>> common-sense intuitions about "the macroscopic" later. For now, the
>> connection is allegedly the fact that macroscopic bodies are made up
>> of microscopic parts!
>Ah! that is the problem. The connection between them in existing theory
>is not that a macroscopic object is made up of microscopic parts. Rather it
>is that a wave function allows us to compute the probability of a macroscopic
>observation. The two domains, that of the wave function and that of macroscopic
>observations, are both assumed to exist prior to each other. Physics only
>says how they relate. It does not say how to construct one from the other.
It is quite easy to show that large collections of interacting microscopic
entities behave in a grossly macroscopic manner. Still, the theory only
allows you to predict probabilities for each event to happen. The mechanism
that connects the probabilities to each actual measurement is what is missing.
Dan S.
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <CEuBG...@dartvax.dartmouth.edu>, Benjamin...@dartmouth.edu (Benjamin J. Tilly) writes:
> > > I am waiting for *anyone* to answer these arguments.
> >
> > I thought that I already *had* answered them. However post them again
> > and *read* what I write carefully.
>
> Many of your answers have been confused and repeated statements. Repeating
> the same thing over and over does not make it true. You do not understand
> these issues nearly as well as you think. Admittedly I repeat many things
> also but I try to limit my repetition in reponding to *each* individual.
> In the forum of the net repetition when communicating with different people
> is inevitable.
>
Much the same thing could be said about you also. And to judge by the
responses that I have seen on the net people have more to criticize in
what you say than in what I say. Furthermore I repeat things because
you often seem to get them the first time and I want reasoned
responses.
> I did post these arguments in response to one of your posts and you did
> not as far as I can recall respond to them specifically. You are
> welcome to do so now.
>
> Bell compares Everett's model unfavorably with De Brogiles pilot wave
> model (the precursor to Bell's model). I am paraphrasing Bell since I
> do not want to type in his full comments see `Speakable and unspeakable
> in quantum mechanics', Cambidge, pg 96.)
>
This is the first thing. You quote Bell as if he were God. Why don't
you listen to what people say and respond to their points more.
> 1. Everett's special variables are anthroprocentric whereas de Broigle's
> are objective.
>
Actually Everett's variables are not at all anthroprocentric. However
in his model if you want to extract information about what people will
observe, then you need to analyze the process of observation. Therefore
there is an analysis of peoples experience. However that does not
reflect that there is anything physically important about us, instead
it reflects that if you want to know what we experience then you have
to factor us into it in some way. So this is not a disadvantage to me.
> 2. Everett assumes all configurations are realized at any time (the
> source of many World's I imagine). de Broigle has a *particular*
> configuration.
>
Why is this an advantage? It does not seem to me to be a very natural
description of what happens in QM in experiments.
> 3. Everett makes at best a half hearted attempt to link configuration
> space to continuous trajectories in physical space (my beloved events).
> De Broigle defines this link.
Same as before. Just because continuous trajectories in physical space
are what happens classically, the double slit experiment is good enough
reason for me to accept that it need not be a natural idea in QM.
Furthermore a disadvantage to me of the pilot wave model is that even
if it does give a description of what the world looks like in QM in
which there is only one reality, it still does not explain why it is
that we, being a quantum system, *percieve* reality in the way that we
do. For example why is it that a small quantum system shows
interference effects, but we do not observe interference effects with
classical objects which (by existing theories) are built up from the
quantum level? That is the question that Everett answered, which to me
is the most important question in this business. I could not care about
what the ultimate nature of the universe really is since I cannot know
if that is right, instead I want to understand how the properties of
the universe that we know follow from what we believe to be
(approximately) right. Which is, after all, what science is really
concerned with.
Ben Tilly
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <CEuIA...@dartvax.dartmouth.edu>, Benjamin...@dartmouth.edu (Benjamin J. Tilly) writes:
> > That is true for doing experiments in particle physics. That is not
> > true if we want to use QM as an explanation for things like chemical
> > properties of substances, or for the electrical properties of
> > substances and so on.
>
> Gee I did not realize Everett was needed to apply QM to chemistry or
> condensed matter physics. I doubt that the physicists who work in those areas
> realize it either.
>
The point is that if we want to use explanations from QM to describe
the macroscopic world then we need some justification. Everett provides
that by making the region of validity of the description used in QM
include macroscopic things. This is more problematical for theories
such as the one which you like which predicts a breakdown of the usual
QM so the descriptions used in QM, like the wave function, are not
applicable to macroscopic things. In those theories you need some
explanation of why it is that despite QM breaking down, (ie by some
sort of wave collapse) these explanations still work. Thus while
Everett per se is not needed, it does fill this gap which other
theories have problems with.
> >[...]
> > What is the theoretical boundary between a reversible event and an
> > irreversible one? [...]
>
> There is no *absolute* boundary. It is a statistical boundary. Everything
> can be reversed but many events such as the demise of Schrodinger's cat
> are so unlikely to be reversed that they can be considered absolutely
> irreversible.
>
Exactly. I actually said this several times with different points. One
was that there is no absolute definition of an observation in Everett's
interpertation for the exact same reason. Another was that an exact
definition is not always needed in scientific theories.
> > > Science only deals with what can be observed at least in
> > > principle. ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > ^^^^^^^^^
> > Oh? What about the theory of ideal gases. That is considered a part of
> > science, yet ideal gases do not exist. [...]
>
> You should think more carefully before you write. Physics uses many
> mathematical models that are known to be only approximate. Such models
> are used for a variety of practical and tutorial reasons. These models
> are *not* established scientifc theory although they are part of the
> mathematical machinery of science and may play an important role in
> formulating established theory.
There are different things that you could be talking about if you try
to talk about scientific theory. I tend to use the weaker version of
the body of theory used in science for explanatory and predictive
purposes. In that case the ideal theory of gases qualifies. However if
you want to use the restrictive definition of what we, to the best of
our knowledge consider to be true, then tell me how to _observe in
principle_ that an electron exists. For an easier question tell me how
to _observe in principle_ that the world has an actual existence. Yet
most physicists have little problem with these as statements of what is
believed to be true about the world. (Although some of them will say
that the electron is an approximation and the wave function is what is
real. :-)
Ben Tilly
Benjamin J. Tilly
pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
> In article <ramsay.7...@unixg.ubc.ca>, ram...@unixg.ubc.ca (Keith Ramsay) writes:
> > So your common-sense is to believe that "causality", which includes
> > counterfactual definiteness, and locality, should be true, rather than
> > expect the Aspect experiment to continue to work the same with the
> > detectors placed farther apart.
>
> We have fundamental principles in conflict. One is locality. The other
> is the existing formulation of QM. It is a matter of judgement as to
> which you have more faith in. However this is a conflict between physical
> principles *both* of which are supported by an enormous body of
> experimental evidence. There is no experimental evidence that conflicts
> with *either* principle. At the least doing these experiments properly
> should command a high priority.
Actually they are not in conflict. Proof, the Everett interpertation
combines w/o conflict.
(snip)
> It cannot be a local mechanistic process and there is no mechanism for
> it in the existing theory. There is a mechanism to compute probabilities
> that preserves conservation laws through nonlocal effects. There is no
> mechanism for generating the specific results or observations.
It is in the Everett interpertation.
> > Your explanation of how your "class of models" might produce this
> > effect strikes me as very tenuous indeed. Of course there ought to be
> > some way to patch up a model to make it work, but there are assorted
> > problems. For one, you appear not to be keeping consistency with
> > relativity: your discrete space-times have preferred directions in
> > them.
>
> This is true and it is what bothered me the most when I first started
> thinking about this class of models. I was extremely pleased when I first
> ran across that quote from Einstein that I love to repeat because he seemed
> to reach the same general conclusions as I did but even more so because
> he recognized and accepted that this meant relativity could only be
> approximately true.
>
> I consider it quite possible that physics cannot be based on the
> field concept, i. e., on continuous structures. In that case
> *nothing* remains of my entire castle in the air gravitation
> theory included, [and of] the rest of modern physics.
> -- Einstein in a 1954 letter to Besso, quoted from:
> "Subtle is the Lord", Abraham Pais, page 467.
>
Possible is not the same as believing that it is true. Until you patch
up a number of major problems in your model it seems to me to be very
dubious.
> > Discretizing a continuous model can destroy conservation laws;
>
> In *must* destroy conservation laws if you discretize the field function.
> However they are reintroduced in a global form by the repetition of states
> effect I described. The destruction of the conservation laws at the local
> level is a possible source of quantum uncertainty.
>
Please describe the effect again. Also note that possible here does not
even seem to mean that you have a solid reason to give as to why it
should be likely.
> > it can introduce irrelevant and nearly indetectible new ones; I don't
> > see why one should expect it to produce helpful new ones, particularly
> > of the kind which would cause particles to inform quantum correlated
> > partners what angular momentum to have.
>
> I do not think the singlet state is something that exists in nature.
> I think it only exists in human theory. I think it is an artifact of a
> complex quantum collapse process extended in time and space. It is the
> structure of this process that produces correlations that correspond to
> the singlet state model.
>
> I agree that it is a big step from my model
> to generate these predictions. I do not have the kind of argument for
> my model that will impress many physicists. It is the multitude of intuitive
> connections between aspects of the model and known physical results that makes
> me think this class of models has a good chance to work out in practice.
And a willingless to ignore a lack of any details. Think of the
problems which seemingly reasonable attempts to quantize the theory of
gravity run into. And there people were dealing with mathematics which
they had worked out in some detail. The problems in getting reasonable
approaches to work there should stand as a warning that what seems like
details before you try it may not be just details.
> >
> > Your "common sense" is to disbelieve in the existence of continuous
> > quantities. Aside from that, is it plausible to expect that a discrete
> > model will work better than a continuous one?
>
> I think there are excellent philosophical and mathematical arguments to
> doubt the existence of any *completed* infinite totalities or infinitesimal
> structures.
This is an old debate in math, and has very little relevence that I can
see.
> This is something I have thought a great deal about and I have a
> done some considerable research into generalized recursion theory motivated
> by these thoughts. It is much more than common sense. Virtually all of
> mathematics can be formulated in terms of properties of Turing Machines
> in a potentially infinite but nondeterministic (in the sense of following
> multiple paths not of being probabilistic) universe.
Depends on what you mean by virtually all. It is true that you can work
with the math that we need in practice in this setting. However if you
try to formalize things in this setting in a fairly natural way then
you can get wierd things like an uncountable set which is a subset of a
countable set.
> In such a universe
> we can ask: will a species have an infinite number of descendant species.
> The general formulation of this (if species are treated as Turing Machines)
> requires quantification over the reals. The only mathematics that cannot
> be cast in this light seems to be questions like the continuum hypothesis.
> Those are question which I do not think are objectively either true or false.
> I think such questions refer to different equally valid ways of extending
> a mathematical model.
Depends on how you do it. However I should point out that the
axiomatization of math has very little relevance to the way that it is
applied to physics, although there is the fact that it is impossible to
show that our math is consistent. But if you are willing to accept
that, which most of us are, then there is no problem.
> > |3. QM assumes that probabilities are irreducible. There is no
> > |mathematical way to define such a model.
> >
> > Your conception of what it means to "define" a model, is that it not
> > be a "merely" probabilistic model! Bringing "mathematical" into the
> > picture is a red herring.
>
> Not so. There is no mathematical way to formulate irreducible probabilities.
> This is not my prejudice it is a fact of our current understanding of
> mathematics.
>
What do you mean by mathematical in this setting? There are, in fact,
mathematical models which model QM exactly as it exists today, and
which give the Everett interpertation exactly. Now what is it that you
are asserting is impossible?
> > |I think one should always in science look for
> > |a complete explanation. To say that laws are probabilistic is to say that
> > |we should not try and find a more complete explanation.
> >
> > No, I disagree. What it means is that we don't insist that nature
> > *must* be behaving deterministically. We should be prepared to play
> > along with nature, should it happen to be playing dice. I don't really
> > understand the venom with which you approach this possibility:
>
> Of course we should but how do we *know* nature is playing dice. Maybe
> she is just more clever then we are and she only appears to be playing
> dice. If we give up too easily we will never find out. Since we have
> no way of knowing how clever nature may be in fooling us I do not see
> how we can determine *any* objective criteria for saying that are now
> convinced nature is playing dice. We may suffer from having too much
> faith in our own cleverness and too little comprehension of the
> cleverness in nature.
True. However in the Everett interpertation nature is *not* playing
dice.
> > I have no problem with looking for "more complete explanations", but
> > it should have some solid starting point, to be something more than
> > grinding air. You dislike the idea that radioactive nuclei decay
> > randomly. Okay, so you like the idea that they decay due to some
> > internal structure which we don't know about. I would enjoy that idea
> > as well. Unfortunately, we don't have any evidence of it. It's not got
> > anything to do with God at all.
>
> We will be unlikely to find the evidence if we do not look for it. I have
> no objections to a scientist saying I do not have any idea how to attack
> that problem so I will attack another problem. I strenuously object to
> a scientist saying I have no idea how to attack that that problem and
> none of my eminent colleagues have any idea how to attack therefore
> there must not exist any way to attack it.
I would agree with you here.
> >
> > Your common sense appears again to be weighing in a fairly arbitrary
> > way against a particular kind of model, because you have this sense
> > that nature has to be deterministic.
>
> Not at all. I am weighing in against arbitrarily asserting that a model
> is complete because one cannot conceive of how a more complete model
> might be constructed.
Actually the Everett interpertation is complete in a certain sense
which I would be happy to discuss if you wanted me to.
> >
> > |4. There is no satisfactory explanation in existing theory to link the
> > |macroscopic and the microscopic.
> >
> > I'll see if I can think of something more to say about your
> > common-sense intuitions about "the macroscopic" later. For now, the
> > connection is allegedly the fact that macroscopic bodies are made up
> > of microscopic parts!
>
> Ah! that is the problem. The connection between them in existing theory
> is not that a macroscopic object is made up of microscopic parts. Rather it
> is that a wave function allows us to compute the probability of a macroscopic
> observation. The two domains, that of the wave function and that of macroscopic
> observations, are both assumed to exist prior to each other. Physics only
> says how they relate. It does not say how to construct one from the other.
The connection in the Everett interpertation is *exactly* that the
macroscopic is made up from the microscopic. That this is not true in
some other interpertations is one of my objections to them.
Ben Tilly
Benjamin J. Tilly
da...@uxa.cso.uiuc.edu (Dan S.) writes:
> >It cannot be a local mechanistic process and there is no mechanism for
> >it in the existing theory. There is a mechanism to compute probabilities
> >that preserves conservation laws through nonlocal effects. There is no
> >mechanism for generating the specific results or observations.
>
> That is true. There is no mechanism within quantum theory to account for
> the fact that events happen at all, local or not. Even Everett fails to
> accomplish this. This is called the quantum measurement problem, and
> until we understand better what happens in a measurement I feel we are
> not going to get anywhere on understanding the behavior of an EPR state.
>
But is it a fact? It depends on what you mean by the word "events". If
you weaken the word events enough then they happen trivially. If you
strengthen it to the sort of level that Paul wants, then there is no
direct evidence that they happen at all.
> >> No, I disagree. What it means is that we don't insist that nature
> >> *must* be behaving deterministically. We should be prepared to play
> >> along with nature, should it happen to be playing dice. I don't really
> >> understand the venom with which you approach this possibility:
>
> >Of course we should but how do we *know* nature is playing dice. Maybe
> >she is just more clever then we are and she only appears to be playing
> >dice. If we give up too easily we will never find out. Since we have
> >no way of knowing how clever nature may be in fooling us I do not see
> >how we can determine *any* objective criteria for saying that are now
> >convinced nature is playing dice. We may suffer from having too much
> >faith in our own cleverness and too little comprehension of the
> >cleverness in nature.
>
> I don't think anyone is saying we shouldn't investigate the possibility
> of a non-probabilistic model. What most physicists do say is that it
> appears that nature is inherently probabilistic.
>
But it is possible with the Everett interpertation to give a
nonprobabilistic explanation of nature which explains why we observe
what look to be probabilistic processes.
(snip)
> >Ah! that is the problem. The connection between them in existing theory
> >is not that a macroscopic object is made up of microscopic parts. Rather it
> >is that a wave function allows us to compute the probability of a macroscopic
> >observation. The two domains, that of the wave function and that of macroscopic
> >observations, are both assumed to exist prior to each other. Physics only
> >says how they relate. It does not say how to construct one from the other.
>
> It is quite easy to show that large collections of interacting microscopic
> entities behave in a grossly macroscopic manner. Still, the theory only
> allows you to predict probabilities for each event to happen. The mechanism
> that connects the probabilities to each actual measurement is what is missing.
Could you clarify exactly what you mean here? I agree that it is not
universally accepted by all physicists (although I have hopes that this
will change), but the Everett interpertation provides a perfectly good
explanation IMHO.
Ben Tilly
Daryl McCullough
Well, in the case of special relativity, while some notions become
relativized, other notions become absolute, and these absolute notions
are agreed upon by all observers. Special relativity can be
reformulated in terms of invariant (or covariant) quantities that are
meaningful to all observers. The problem with the Everett
interpretation is that there are *no* observable absolutes. The only
absolute is the universal wave function, which is completely
unobservable.
>>Another point to be made is that there is absolutely no observable
>>difference (even in principle) between the Everett interpretation and
>>the assumption that consciousness collapses the wave function, but
>>only *my* consciousness.
>
>This is only true if your consciousness is 100% immune to interference
>effects. I think this is liable to turn out to be only a good
>approximation.
I believe that it is practically impossible for a person to observe
interference effects from different states of one's own consciousness.
I may be wrong, but here is my reasoning: To observe interference
between two brain states A1 and A2, there must be a later brain state
B such that there are sizable amplitudes for the transitions A1 --> B
and A2 --> B. This in turn implies that in state B, it is impossible
to say with certainty whether A1 or A2 occurred previously. This in
turn implies that there is no record of whether A1 or A2 occurred, which,
by my informal definition of observation, implies that there was no
observation made as to whether the brain was in state A1 or A2.
I know that this is pretty fuzzy reasoning, but I think that the
requirements of making an observation and of observing interference
are contradictory. To whatever extent you can be certain of having
made an observation, you can also be certain of there being no
interference effects.
>>The Everett interpretation is equivalent, in
>>a certain sense, to starting from this rather solipsistic theory and
>>restoring observer-independence by throwing in not only my version of
>>history (where my observations collapse the wave function), but also
>>every other observer's history. (The neat thing about the technical
>>work of Everett is that it shows that all these different histories
>>*can* consistently be combined; that each person can assume that *his*
>>observations collapse the wave function, and no contradiction arises.)
>
>I prefer to look at it from the other direction (even if not the
>historical order of development). The solipsistic view is in some
>sense what you get from Everett if you assume that yours is the unique
>and only consciousness in the world (and your brain the only one which
>produces "consciousness").
It depends on what you mean by "world". To me, the world is not a
universal wave function in which everything happens, but an arena in
which *particular* things happen. Clinton getting elected and Kennedy
being assasinated and the Allies winning World War II are facts about
the world. But in the Everett interpretation, none of these facts are
facts about the world, they are all facts about my particular path
through the world. And these facts don't tell me much at all about the
universal wave function (except that it nonzero amplitudes for certain
states; there is no way by observation that I can further quantify the
amplitudes for the universal wave function).
There is a similar philosophical problem with an infinite universe
with an infinite number of stars in classical mechanics. In that case,
it is to be expected that every finite configuration of matter that is
possible actually occurs somewhere in the universe. That being the
case, no observations that I ever make would tell me anything that I
didn't already know about the universe as a whole. The only knowledge
that is possible about such an infinite universe is knowledge about
the situation that one finds oneself in. Knowledge about the universe
as a whole in that case would be both useless and impossible.
Objective knowledge of the universe is only possible if observations
constrain in some way the set of possibilities for the universe, which
is not the case in the Everett interpretation.
Daryl McCullough
ORA Corp.
Ithaca, NY
Paul Budnik uunet!mtnmath!paul
> In article <1993Oct15....@mtnmath.UUCP>
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> > the problems presented by EPR and Bell's inequality all this work may turn
> > out to be excellent mathematics but it is unlikely to have much physical
> > content. This is because the basic assumptions under which they are
> > working are physically false. Maybe its a shame that so much effort is
> > being put into projects that are doomed from the start.
>
> might think that it is, but can you really justify that statement? If
> you cannot, then you should not say it out in the open where you will
> be asked to do so. Not unless you want people thinking that you do not
> know what you are talking about.
I said why. The basic assumptions they are working with are false. Of course
it depends on what ways they are false to determine what degree if any the
work could be salvaged for physics.
If Bell's inequality is not violated in nature there must be objective
nonlinear changes in the wave function. Attemts to unify quantum mechanics
and gravity that do not take this into account are unlikely to be productive.
If as Einstein suspected a discrete space time model is needed
to unify QM and gravity then they are barking up the wrong mathematical
tree. That does not guarantee that they will not do useful physics. One
cannot give precise arguments about such hypothetical questions.
In my judgment the odds would be badly against them.
Paul Budnik
Paul Budnik uunet!mtnmath!paul
> pa...@mtnmath.UUCP (Paul Budnik uunet!mtnmath!paul) writes:
>
> >We have fundamental principles in conflict. One is locality. The other
> >is the existing formulation of QM. It is a matter of judgement as to
> >which you have more faith in. However this is a conflict between physical
> >principles *both* of which are supported by an enormous body of
> >experimental evidence. There is no experimental evidence that conflicts
> >with *either* principle. At the least doing these experiments properly
> >should command a high priority.
>
> There is no experimental evidence for counterfactual definiteness in
> the microscopic domain. In fact, I would say the entire concept is
> quite foreign to quantum formalism.
I doubt that anyone has done a careful analysis to see what the implications
are of denying this. It is standard assumption that is used universally
in statistical analysis of experiments. One bases an analysis on a consideration
of all the possible outcomes even though only some are expected to occur.
It is not foreign to macroscopic observations that arise in QM experiments.
It is only macroscopic observations that are at issue in tests of Bell's
inequality.
No one would have ever questioned this except that they are made uncomfortable
by Eberhard's results.
> >It cannot be a local mechanistic process and there is no mechanism for
> >it in the existing theory. There is a mechanism to compute probabilities
> >that preserves conservation laws through nonlocal effects. There is no
> >mechanism for generating the specific results or observations.
>
> That is true. There is no mechanism within quantum theory to account for
> the fact that events happen at all, local or not. Even Everett fails to
> accomplish this. This is called the quantum measurement problem, and
> until we understand better what happens in a measurement I feel we are
> not going to get anywhere on understanding the behavior of an EPR state.
Thanks this is a point I have been making it repeatedly in discussions with
several people. Everett does not explain everything that happens because
he does not explain observations, measurements or events at all.
> >I do not think the singlet state is something that exists in nature.
> >I think it only exists in human theory. I think it is an artifact of a
> >complex quantum collapse process extended in time and space. It is the
> >structure of this process that produces correlations that correspond to
> >the singlet state model.
>
> Neither did Einstein. But, in 1957 Aharonov and Bohm did an experiment to
> see if the singlet state would indeed break down if the photons were
> sufficiently separated. They're conclusion was that the singlet state
> remained a singlet state regardless what the separation between the
> two particles is.
When the singlet state was first understood some expected the correlations
to disappear when the wave functions of the two particle no longer had
significant spatial overlap. Experiments showed this was not true. However I
think the particle decay that generates singlet state correlations has
a completely unknown structure. The only effective test of the singlet state
would be an effective test of Bell's inequality. You cannot base assumptions
about such issues on the standard wave function model for correlated state
particles. It is precisely this wave function not just the singlet state
correlations that these experiments are testing.
>[...]
> I don't think anyone is saying we shouldn't investigate the possibility
> of a non-probabilistic model. What most physicists do say is that it
> appears that nature is inherently probabilistic.
Many physicists are convinced that this is a dead issue. You would have
absolutely no chance of getting research funding for a project along these
lines. You would be far better off doing crazy useless speculation about the
wave function of the universe.
> It is quite easy to show that large collections of interacting microscopic
> entities behave in a grossly macroscopic manner. Still, the theory only
> allows you to predict probabilities for each event to happen. The mechanism
> that connects the probabilities to each actual measurement is what is missing.
Amen!
Paul Budnik