measuring measuring... fast?
Claude Chaunier
Quantum Physics and they look like offering something interesting
when we put them together. Does anyone here knows why it isn't so?
Here they are.
- From a source of correlated particles, measuring a quantum state
of one particle immediately determines the correlated state of the
other particle in the pair, however far the two particles are from
each other. In effect, one is measuring a system - the two
particles - at once.
- Measuring some quantum state of a particle limits the subsequent
quantum indeterminacy of its behaviour. It can be statistically
measured whether a flow of particles has gone through some earlier
drastic measuring apparatus.
As a result it seems we can measure faster than whatyouknow whether
someone is switching on or off some measuring apparatus miles away...
I guess I'm wrong somewhere. I'd love to know where.
Regards.
Matt McIrvin
<chau...@handy.univ-lyon1.fr> wrote:
>- From a source of correlated particles, measuring a quantum state
>of one particle immediately determines the correlated state of the
>other particle in the pair, however far the two particles are from
>each other. In effect, one is measuring a system - the two
>particles - at once.
>
>- Measuring some quantum state of a particle limits the subsequent
>quantum indeterminacy of its behaviour. It can be statistically
>measured whether a flow of particles has gone through some earlier
>drastic measuring apparatus.
>
>As a result it seems we can measure faster than whatyouknow whether
>someone is switching on or off some measuring apparatus miles away...
It turns out that in these sorts of experiments, you can only tell
that the measurements were correlated by comparing notes after the
fact, by ordinary means-- the actual "transmitted information" from
one measurement to the other is always completely random.
Avoiding this kind of faster-than-light measurement communication is
actually a major consideration in constructing quantum field theories.
In the QFTs that are used to describe the world, observables are
associated with finite regions of space-time, and it's always the
case that observables separated by spacelike intervals must commute.
That is, if two measurements are separated by a faster-than-light
interval, the results might have a quantum correlation, but it
can never matter to the results which order the measurements happened
in. This is necessary for consistency with relativity, since a
change of reference frame can vary the order.
--
Matt McIrvin http://world.std.com/~mmcirvin/
scott...@my-deja.com
mmci...@world.std.com (Matt McIrvin) wrote:
> It turns out that in these sorts of experiments, you can only tell
> that the measurements were correlated by comparing notes after the
> fact, by ordinary means-- the actual "transmitted information" from
> one measurement to the other is always completely random.
>
> Avoiding this kind of faster-than-light measurement communication is
> actually a major consideration in constructing quantum field theories.
> In the QFTs that are used to describe the world, observables are
> associated with finite regions of space-time, and it's always the
> case that observables separated by spacelike intervals must commute.
[...]
Given your response, why is the standard interpretation of the Aspect
experiment, which apparently confirms Bell's inequality, that either
locality or objectivity of phenomenon must be sacrificed to conform to
the experiment. Is there something more subtle in these experiments
that has not been adequately communicated?
Scott
Sent via Deja.com
http://www.deja.com/
[Moderator's note: Quoted text trimmed. -MM]
Cl.Massé
message : 94sti5$d40$1...@news4.isdnet.net...
> - From a source of correlated particles, measuring a quantum state
> of one particle immediately determines the correlated state of the
> other particle in the pair, however far the two particles are from
> each other. In effect, one is measuring a system - the two
> particles - at once.
>
> - Measuring some quantum state of a particle limits the subsequent
> quantum indeterminacy of its behaviour. It can be statistically
> measured whether a flow of particles has gone through some earlier
> drastic measuring apparatus.
>
> As a result it seems we can measure faster than whatyouknow whether
> someone is switching on or off some measuring apparatus miles away...
Quantum correlations like in Aspect's experiment don't entail FTL
(Faster Than Light) information travel. The reason is that the
outcomes of the two measurements have to be collected at one point
in order to observe the correlation, and for that, they have to be
transmitted not FTL to this point. It can be seen it is
impossible to use this experiment to communicate between the two
ends of the setup.
--
~~~~ %20cl...@free.fr%20 LPF
Matt McIrvin
>Given your response, why is the standard interpretation of the Aspect
>experiment, which apparently confirms Bell's inequality, that either
>locality or objectivity of phenomenon must be sacrificed to conform to
>the experiment. Is there something more subtle in these experiments
>that has not been adequately communicated?
The "either-or" part is the key thing. A more precise way of saying it is
that we must sacrifice either hidden variables, *or* locality. Most
physicists prefer to give up hidden variables and keep locality. A
dissenter was David Bohm, who preferred to make the opposite choice.
QM is local in the sense that you can't send faster-than-light messages
with it. But, in and of itself, it lacks hidden variables: there's
nothing hiding behind our inability to predict which value an observable
is going to have. I suppose that could be referred to as rejecting
"objectivity of phenomenon."
Bohm's objective theory is actually completely compatible with the Aspect
results-- it isn't rejected experimentally-- even though it has hidden
variables. But those hidden variables act faster than light. In Bohm's
model, the dynamics behind the scenes involves "pilot waves" that can
transmit information at infinite speed.
Claude Chaunier
> Quantum correlations like in Aspect's experiment don't entail FTL
> (Faster Than Light) information travel. The reason is that the
> outcomes of the two measurements have to be collected at one point
> in order to observe the correlation, and for that, they have to be
> transmitted not FTL to this point. It can be seen it is
> impossible to use this experiment to communicate between the two
> ends of the setup.
I understand it about usual measurement of quantum variables -
the values measured on (let's say) our side are perfectly random
when they are considered on their own without knowing the values
on the far side.
How about the measurement I was speaking about, namely measuring
whether some quantum variable has already been measured and its
indeterminacy has been consequently reduced? A usual measurement
would or would not take place on the far side, while we would try
to measure how much indeterminate the system is on our side. There
is no need to wait and be transmitted a result from the far side
to see whether our side shows some interference due to its quantum
indeterminacy or behaves classically.
For example two-slit interference experiments show fringe patterns
only if the particle's linear momentum keeps indeterminate until
the particle gets any chance to interfere with itself. So I was
thinking, if a source emits two particles in random opposite
directions so that their linear momentum is correlated and the
particle on our side goes through a two-slit interference
experiment, we should also keep the other particle free from any
linear momentum measure if we want to see our fringe pattern.
I am however getting an answer from reading Mark P. Silverman's
description of quantum interference with correlated particles
(in "More than One Mistery", Springer-Verlag 1995, Section 3.2).
Namely, there is never going to be any fringe pattern on our side
proper, whatever is done or not done on the other side with the
correlated particle! The particle on our side always looks like
behaving classically and never interfering with itself alone, once
it came from a correlated pair. Even though its initial linear
momentum and the path it then takes is really indeterminate - it
can be seen to yield some quantum interference when joint results
about the two particles are considered as a whole.
After all that is what Matt McIrvin and Cl.Massé were saying, only
I needed to be sure it held with "second-order" measurements too.
Claude Chaunier
Daryl McCullough
>QM is local in the sense that you can't send faster-than-light messages
>with it.
I'm sure most people already know this, but just for clarification,
there is another sense in which quantum field theory is nonlocal.
In classical field theory, you can divide the universe up into a bunch
of small localized regions. Then the state of the universe is uniquely
determined by the conditions (the values of fields) in each region.
In quantum field theory, however, the state of the universe is not
uniquely determined by conditions in each localized region. That's
because there are additional "entanglement" relationships that are
inherently nonlocal.
--
Daryl McCullough
CoGenTex, Inc.
Ithaca, NY
Gerry Quinn
>I'm sure most people already know this, but just for clarification,
>there is another sense in which quantum field theory is nonlocal.
>
>In classical field theory, you can divide the universe up into a bunch
>of small localized regions. Then the state of the universe is uniquely
>determined by the conditions (the values of fields) in each region.
>
>In quantum field theory, however, the state of the universe is not
>uniquely determined by conditions in each localized region. That's
>because there are additional "entanglement" relationships that are
>inherently nonlocal.
>
Philosophically, that's rather Bohmian, isn't it? It makes the
entanglement relationships look almost like a tangible part of the
world.
Of course QFT is not necessarily (or by all accounts, probably) the
ultimate description of the universe.
- Gerry Quinn
Cl.Massé
> How about the measurement I was speaking about, namely measuring
> whether some quantum variable has already been measured and its
> indeterminacy has been consequently reduced? A usual measurement
> would or would not take place on the far side, while we would
> try to measure how much indeterminate the system is on our side.
> There is no need to wait and be transmitted a result from the
> far side to see whether our side shows some interference due to
> its quantum indeterminacy or behaves classically.
The trick is that the indeterminacy has well been reduced, but we
can't know what the outcome of the measurement on the far side
actually is. In order to measure the indeterminacy on the near
side, we need to measure many particles which partner has a
probabilistic distribution of measured values on the far side if a
measure have been done there, or a quantum superposition of states
if not. In the first case, on our side, we measure a probability
density, and in the latter one, a quantum superposition, which
gives the same result according to Born's probabilistic
interpretation. There is no mean to know the indeterminacy of a
single particle since we don't measure the wave function, but a
given observable, so changing the wave function.
> For example two-slit interference experiments show fringe
> patterns only if the particle's linear momentum keeps
> indeterminate until the particle gets any chance to interfere
> with itself. So I was thinking, if a source emits two particles
> in random opposite directions so that their linear momentum is
> correlated and the particle on our side goes through a two-slit
> interference experiment, we should also keep the other particle
> free from any linear momentum measure if we want to see our
> fringe pattern.
>
> I am however getting an answer from reading Mark P. Silverman's
> description of quantum interference with correlated particles
> (in "More than One Mystery", Springer-Verlag 1995, Section 3.2).
>
> Namely, there is never going to be any fringe pattern on our
> side proper, whatever is done or not done on the other side with
> the correlated particle! The particle on our side always looks
> like behaving classically and never interfering with itself
> alone, once it came from a correlated pair. Even though its
> initial linear momentum and the path it then takes is really
> indeterminate - it can be seen to yield some quantum
> interference when joint results about the two particles are
> considered as a whole.
More exactly, the two-slit experiment shows a fringe pattern if
the incoming particle is in a state of well defined linear
momentum. If it is the partner of a momentum correlated pair, it
is not in such a state, but in a quantum superposition of states
of well defined momentum. Each of these states will give a
different pattern, and these will superpose and destroy each
other. As there is no resulting interference pattern, it makes no
difference if the other particle has been measured or not.
--
~~~~ %20cl...@free.fr%20 LPF
Liberty, Equality, Profitability.
Jim Carr
} Given your response, why is the standard interpretation of the Aspect
} experiment, which apparently confirms Bell's inequality, that either
} locality or objectivity of phenomenon must be sacrificed to conform to
} the experiment. Is there something more subtle in these experiments
} that has not been adequately communicated?
In article <2001013003...@world.std.com>
mmci...@world.std.com (Matt McIrvin) writes:
>The "either-or" part is the key thing. A more precise way of saying it
>is that we must sacrifice either hidden variables, *or* locality. Most
>physicists prefer to give up hidden variables and keep locality. A
>dissenter was David Bohm, who preferred to make the opposite choice.
It was not clear to me if the questioner understood that the
Aspect-type experiments demonstrate that Bell's Inequality is
*violated* and that, as a result, one of the assumptions used
to derive it must be false. Hence the "either-or" problem.
[ObNitPick self-correction: Aspect's experiment is not good enough
to really establish a violation, but more recent ones are. Further,
they actually test a different inequality but the consequences are
the same. In addition, there are several other assumptions you can
choose to discard, but most people like those even more than the
ones listed above.]
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
"The half of knowledge is knowing where to find knowledge" - Anon.
Motto over the entrance to Dodd Hall, former library at FSCW.
Charles Francis
scott...@my-deja.com:
>In article <952641$j0b$1...@news.state.mn.us>,
> mmci...@world.std.com (Matt McIrvin) wrote:
>> It turns out that in these sorts of experiments, you can only tell
>> that the measurements were correlated by comparing notes after the
>> fact, by ordinary means-- the actual "transmitted information" from
>> one measurement to the other is always completely random.
>>
>> Avoiding this kind of faster-than-light measurement communication is
>> actually a major consideration in constructing quantum field theories.
>> In the QFTs that are used to describe the world, observables are
>> associated with finite regions of space-time, and it's always the
>> case that observables separated by spacelike intervals must commute.
>Given your response, why is the standard interpretation of the Aspect
>experiment, which apparently confirms Bell's inequality, that either
>locality or objectivity of phenomenon must be sacrificed to conform to
>the experiment. Is there something more subtle in these experiments
>that has not been adequately communicated?
There is something more subtle. As Matt says the locality condition in
quantum field theory is that observables separated by spacelike
intervals commute. This condition applies to operators, and it is not
the same as the locality condition which is used in the derivation of
Bell's theorem. We should conclude that the locality condition used in
the derivation of Bell's theorem is incorrect, and that we do not have
to reject either locality or objectivity of phenomenon. The correct
expression of locality, as given in field theory does not allow you to
derive the Bell inequalities.
--
Regards
Charles Francis
cha...@clef.demon.co.uk
Paul P. Budnik Jr.
> mmci...@world.std.com (Matt McIrvin) writes:
> >The "either-or" part is the key thing. A more precise way of saying it
> >is that we must sacrifice either hidden variables, *or* locality. Most
> >physicists prefer to give up hidden variables and keep locality. A
> >dissenter was David Bohm, who preferred to make the opposite choice.
>
> It was not clear to me if the questioner understood that the
> Aspect-type experiments demonstrate that Bell's Inequality is
> *violated* and that, as a result, one of the assumptions used
> to derive it must be false. Hence the "either-or" problem.
It is important to understand that this is a matter of philosophical opinion
and not scientific fact. Quantum mechanics violates classical definitions of
locality. Bell proved it and Eberhard proved it without reference to
hidden variables. ("Bell's Theorem without Hidden Variables",
_Nuovo Cimento_, V 30 B, pp 75-89, 1977). Quantum mechanics is
able to obtain a limited consistency with special relativity because
the direction of the nonlocal effects are masked by quantum uncertainty.
In tests of Bell's inequality two experimental manipulations and two
detentions are involved One of the distant manipulation instantaneously
affects one of the detentions but one cannot tell the direction of the
effect and thus the predictions are the same in all frames of reference.
To call this a local model is to redefine the term in a very strange
way.
> [ObNitPick self-correction: Aspect's experiment is not good enough
> to really establish a violation, but more recent ones are. Further,
> they actually test a different inequality but the consequences are
> the same. In addition, there are several other assumptions you can
> choose to discard, but most people like those even more than the
> ones listed above.
Wrong again. The two most recent experiments independently addressed
the two primary loopholes, detection efficiency and timing measurement.
However both papers agree that their respective experiments
did not simultaneously address both
loopholes. That is clearly needed for a decisive experiment. Further the
timing measurement loophole is elusive because quantum mechanics
provides no objective definition of when a measurement is complete. One
needs a macroscopic measurement of the timing to be reasonably sure
that a measurement cannot be reversed.
It is interesting to note that even before Aspect attempted to measure
timing he claimed a conclusive experimental result.
"Our results in excellent agreement with the quantum mechanical
predictions, strongly violate the generalized Bell's inequalities, and rule
out the whole class of realistic local theories" , _PRL_, V 47, p 460, 1981.
Most physicists are thoroughly convinced that quantum mechanic is correct in
these predictions. As a result they seem far to eager to make excessive
claims about these experiments. Recent reports have generally made it clear
what loopholes remain while still claiming that their result seems decisive.
We need to close all the loopholes simultaneously for a truly decisive
experiment.
Locality is the single most powerful simplifying assumption in science. To
argue that a pair of photons remain entangled after being separated by a
billion light years and that the entanglement collapses when one of them is
observed is a mind boggling claim that requires extreme measures to justify.
Given quantum mechanics' failure to give an objective definition to
measurement we need to be even more careful.
Paul Budnik
pa...@mtnmath.com
www.mtnmath.com
Squark
<3B17190C...@mtnmath.com>):
>To call this a local model is to redefine the term in a very strange
>way.
My definition of ("relativistically") local is not-allowing superliminal
transmittion of information. Quantum mechanics does just that. Imagine
the following experiment: a random number generator in a black-box
produces a number which is copied onto two notes. Afterwards, the box is
divided, the two notes being in the two parts, and carried away
(luminally!) to two different places. In each of them, the box is opened
and the note read. How's that for an EPR effect? Now you might say: but
that's not fair, the information about the outcome of the experiment
_existed_ all the time, rather than appearing at the time of the
experiment, like in QM. However, this seems to me an unphysical question,
because it can't be tested! Moreover, if we take the multi-world point of
view, nothing special happens at the moment of the experiment, and there's
no essential difference between the black-box experiment and the EPR one.
Best regards,
Squark.
--------------------------------------------------------------------------------
Write to me at:
[Note: the fourth letter of the English alphabet is used in the later
exclusively as anti-spam]
dSdqudarkd_...@excite.com
gregegan
Squark<dSdqudarkd_...@excite.com> wrote:
> On Fri, 01 Jun 2001 04:33:22 GMT, Paul P. Budnik Jr. wrote (in
> <3B17190C...@mtnmath.com>):
> >To call this a local model is to redefine the term in a very strange
> >way.
>
> My definition of ("relativistically") local is not-allowing superliminal
> transmittion of information. Quantum mechanics does just that. Imagine
> the following experiment: a random number generator in a black-box
> produces a number which is copied onto two notes. Afterwards, the box is
> divided, the two notes being in the two parts, and carried away
> (luminally!) to two different places. In each of them, the box is opened
> and the note read. How's that for an EPR effect? Now you might say: but
> that's not fair, the information about the outcome of the experiment
> _existed_ all the time, rather than appearing at the time of the
> experiment, like in QM. However, this seems to me an unphysical question,
> because it can't be tested! Moreover, if we take the multi-world point of
> view, nothing special happens at the moment of the experiment, and there's
> no essential difference between the black-box experiment and the EPR one.
There's a wonderful paper by David Deutsch and Patrick Hayden that makes
this clearer than anything I've ever read. Anyone interested in EPR,
whether they're sceptical about MWI or not, should read this.
Essentially, it shows that information in an entangled system is
thoroughly localised and well-behaved, even though it might only be
*accessible* via measurements on the whole system.
<http://xxx.lanl.gov/abs/quant-ph/9906007>
Information Flow in Entangled Quantum Systems
ABSTRACT: All information in quantum systems is, notwithstanding Bell's
theorem, localised. Measuring or otherwise interacting with a quantum
system S has no effect on distant systems from which S is dynamically
isolated, even if they are entangled with S. Using the Heisenberg picture
to analyse quantum information processing makes this locality explicit,
and reveals that under some circumstances (in particular, in
Einstein-Podolski-Rosen experiments and in quantum teleportation) quantum
information is transmitted through 'classical' (i.e. decoherent)
information channels.
Charles Francis
<pa...@mtnmath.com> writes
>Quantum mechanics is
>able to obtain a limited consistency with special relativity because
>the direction of the nonlocal effects are masked by quantum uncertainty.
>In tests of Bell's inequality two experimental manipulations and two
>detentions are involved One of the distant manipulation instantaneously
>affects one of the detentions but one cannot tell the direction of the
>effect and thus the predictions are the same in all frames of reference.
>To call this a local model is to redefine the term in a very strange
>way.
Yes it is strange. The point is that the property of spin does not exist
without the measurement. Quantum mechanics gives a poor description of
the subatomic world because it describes it in terms of the properties
of classical measuring apparatus. Classical measuring apparatus are
themselves non-local, and in fact define the entire reference frame. We
have an apparently non-local effect in Bell's theorem because the thing
we intend to measure, spin, does not have a value until there is a
framework of measurement in terms of which it can be described, and this
descriptive framework is non-local.
When one experimenter measures spin, he also determines spin of the
other particle, outside the light cone. But he does not alter the other
particle, only his description is altered, and the other experimenter
finds no alteration in the results of his measurement. He does a
different measurement of spin, and gets a different description. But
while the experimenters are outside the light cone of each others
measurement these descriptions only hold for themselves, not for each
other. Only when the experimenters get together can they produce a new
description in which the correlation is observed, and the existence of
this correlation says nothing more than that the events had a common
cause.
Relativistic quantum field theory gives a far more appropriate
definition of locality, that no observable effect may be transmitted
faster than the speed of light. And this definition is obeyed in Bell's
theorem.
>Locality is the single most powerful simplifying assumption in science. To
>argue that a pair of photons remain entangled after being separated by a
>billion light years and that the entanglement collapses when one of them is
>observed is a mind boggling claim that requires extreme measures to justify.
Only the description is entangled because of insufficient evidence.
>Given quantum mechanics' failure to give an objective definition to
>measurement we need to be even more careful.
Measurement is something carried out by an individual observer, and is
in that sense subjective. What we require is that all observers can
carry out identical measurements under identical circumstances and
obtain results obeying identical laws.
Regards
--
Charles Francis
Jim Carr
|
| mmci...@world.std.com (Matt McIrvin) writes:
| >The "either-or" part is the key thing. A more precise way of saying it
| >is that we must sacrifice either hidden variables, *or* locality. Most
| >physicists prefer to give up hidden variables and keep locality. A
| >dissenter was David Bohm, who preferred to make the opposite choice.
|
| It was not clear to me if the questioner understood that the
| Aspect-type experiments demonstrate that Bell's Inequality is
| *violated* and that, as a result, one of the assumptions used
| to derive it must be false. Hence the "either-or" problem.
In article <3B17190C...@mtnmath.com>
"Paul P. Budnik Jr." <pa...@mtnmath.com> writes:
>
>It is important to understand that this is a matter of philosophical opinion
>and not scientific fact.
IMO, the "opinion" that when (A and B) ==> C, (not)C ==>
(not)A or (not)B is the basis for any discussion of the
comparison of observations (facts) with theory in science.
That the "(not)C" part is qualified in terms of a confidence
level derived from the limitations of the experiment, and
that the axioms have to be expanded to include possible
loopholes that were not part of the original theorem is a given.
>Quantum mechanics violates classical definitions of
>locality. Bell proved it and Eberhard proved it without reference to
>hidden variables. ("Bell's Theorem without Hidden Variables",
>_Nuovo Cimento_, V 30 B, pp 75-89, 1977). Quantum mechanics is
>able to obtain a limited consistency with special relativity because
>the direction of the nonlocal effects are masked by quantum uncertainty.
I disagree. IMO, QED is a local theory that is relativistic.
No "masking" is needed. Further, Bell's theorem, like that
of CHSH, does not rest on only one assumption.
>In tests of Bell's inequality two experimental manipulations and two
>detentions are involved One of the distant manipulation instantaneously
>affects one of the detentions but one cannot tell the direction of the
>effect and thus the predictions are the same in all frames of reference.
>To call this a local model is to redefine the term in a very strange
>way.
What is strange about saying that QED is local? It *is* local.
I think your comments are a result of taking such things as the
Copenhagen interpretation much too literally. Look at Youssef's
Bayesian approach as a completely different alterantive.
| [ObNitPick self-correction: Aspect's experiment is not good enough
| to really establish a violation, but more recent ones are. Further,
| they actually test a different inequality but the consequences are
| the same. In addition, there are several other assumptions you can
| choose to discard, but most people like those even more than the
| ones listed above.
>Wrong again. The two most recent experiments independently addressed
>the two primary loopholes, detection efficiency and timing measurement.
Those are included in my list. Most people (since most
physicists are experimentalists) like those even more than
relativity or locality (since they are based on assumptions
that predate all of modern physics), which is why the former
only appears recently. You don't see it in EPR.
>However both papers agree that their respective experiments
>did not simultaneously address both
>loopholes. That is clearly needed for a decisive experiment.
Note, however, that the "either or" is rejected; what has
to be eliminated is an even-more-special case where they
both appear in a way that, say, the detection efficiency
varies in a highly bizarre fashion where the detector can
know at FTL rates what efficiency it needs to have.
Also note that most critics were sure that closing one of
those loopholes (like Caroline Thompson's view of background
subtraction) would fix things. The fact that each one that
gets closed only leads to a more significant deviation from
the inequality bodes badly for the standard assumptions.
>Further the
>timing measurement loophole is elusive because quantum mechanics
>provides no objective definition of when a measurement is complete. One
>needs a macroscopic measurement of the timing to be reasonably sure
>that a measurement cannot be reversed.
That was done in one experiment.
>It is interesting to note that even before Aspect attempted to measure
>timing he claimed a conclusive experimental result.
>"Our results in excellent agreement with the quantum mechanical
>predictions, strongly violate the generalized Bell's inequalities, and rule
>out the whole class of realistic local theories" , _PRL_, V 47, p 460, 1981.
>
>Most physicists are thoroughly convinced that quantum mechanic is correct in
>these predictions. As a result they seem far to eager to make excessive
>claims about these experiments.
I thought he was trying to make a case for a Nobel Prize. IMO he
would have said the same thing if it had come out the other way.
What would be your opinion of the defenders of QM if it had come
out the other way and they started coming up with various exotic
loopholes to patch up the disagreement as better and better work
kept getting further from the QM prediction?
>Locality is the single most powerful simplifying assumption in science. To
>argue that a pair of photons remain entangled after being separated by a
>billion light years and that the entanglement collapses when one of them is
>observed is a mind boggling claim ...
So don't make it mind boggling. It is not mind boggling as long
as you keep treating them as a _pair_ of photons rather than as
two independent photons when you know they cannot be independent
and still conserve angular momentum. Is it puzzling that the
opposite faces of a die are correlated? Why should it be puzzling
that two members of a pair are correlated?
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
spammer fodder: ab...@aol.com ab...@yahoo.com ab...@hotmail.com
ab...@msn.com ab...@earthlink.com
Jim Carr
Charles Francis <cha...@clef.demon.co.uk> writes:
>The point is that the property of spin does not exist
>without the measurement.
That is like saying that the property of "Charles Francis" does
not exist without the measurement. I consider that an extreme
view of the nature of nature.
>Quantum mechanics gives a poor description of
>the subatomic world because it describes it in terms of the properties
>of classical measuring apparatus.
"What if I don't treat the detector classically, what
will happen to me?"
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
spammer fodder: ab...@aol.com ab...@yahoo.com ab...@hotmail.com
ab...@msn.com ab...@earthlink.com
Charles Francis
<j...@dirac.csit.fsu.edu> writes
>I think your comments are a result of taking such things as the
> Copenhagen interpretation much too literally. Look at Youssef's
> Bayesian approach as a completely different alterantive.
But I don't think it is a completely different alternative, so much as a
variant of the Copenhagen interpretation taken in its strict form, by
which I mean excluding Bohr's complementarity - the Copenhagen
interpretation proper does not say that we are describing a 'matter
wave' with a definite property like spin, but that whatever kind
material entity we are attempting to describe, the description by means
of macroscopic concepts is imperfect and misleading. For example on p15
of Physics and Philosophy Heisenberg says "Also this deficiency in
knowledge is expressed in the probability function".
The notion of expressing a level of knowledge by an amplitude sounds
pretty Bayesian, though I would rather describe it in terms of the more
modern ideas of many valued logic, which can trace their roots to
Bayesian reasoning. Personally I would hold back from calling qm
probability amplitudes Bayesian, but that is because I am a frequentist.
I hold that qm gives an objective prediction of real frequencies
resulting from repeated experiments, and I maintain that, in general,
probability theory is about such objective predictions. If I were a
Bayesian with respect to probability theory, then I would be with
respect to qm also.
Regards
--
Charles Francis
Daryl McCullough
>IMO, QED is a local theory that is relativistic.
I think it depends on what you mean by a "local theory". I
think that there are two parts to a theory being local:
(1) states are local, and (2) evolution is local. QED without
wave function collapse is local in the second sense, but not
in the first sense.
What I mean by states being local is this: The state of
the whole is uniquely determined by the states of the parts
(plus perhaps information on how the parts fit together).
Classically, if we divide the universe up into little cubes
of size 1 meter^3, then the state of the universe is completely
determined by the values of fields in each cube, their time
derivatives, and positions, momenta, angular momenta and kinds
of paricles within each cube. If we know the conditions in each
cube, and we know how the cubes fit together, then we know
everything there is to know about the universe.
Quantum mechanics is different, because of entanglement.
The closest that we can come to giving a quantum state
for a small volume of space is to give a density matrix.
But the density matrix for the universe as a whole is
not uniquely determined by the density matrices of each
small volume. Local information is not enough.
Paul P. Budnik Jr.
"Jim Carr" <j...@dirac.csit.fsu.edu> wrote in message
news:9g6q5u$dih$1...@news.fsu.edu...
> Jim Carr wrote in article <2001013003...@world.std.com>:
> In article <3B17190C...@mtnmath.com>
> "Paul P. Budnik Jr." <pa...@mtnmath.com> writes:
> >It is important to understand that this is a matter of philosophical
> > opinion and not scientific fact.
> IMO, the "opinion" that when (A and B) ==> C, (not)C ==>
> (not)A or (not)B is the basis for any discussion of the
> comparison of observations (facts) with theory in science.
>
> That the "(not)C" part is qualified in terms of a confidence
> level derived from the limitations of the experiment, and
> that the axioms have to be expanded to include possible
> loopholes that were not part of the original theorem is a given.
The issue I was raising has nothing to do with experimental results.
It is the claim that the the two alternatices are local theory or realistic
theory. That is a matter of philosophical opinion. Consider the
following quote:
"Quantum theory is nonlocal. Indeed, quantum theory
predicts correlations among distant measurement outcomes
that cannot be explained by any theory which involves
only local variables." from "Violation of Bell Inequalities by
Photons More Than 10 km Apart" by
W. Tittel, J. Brendel, H. Zbinden, and N. Gisin, _PRL_,
V 81, N 17, p 3563, 1998.
You may have anaother definition of nonlocal but the creation
of a new definition of what is an essentially mathematical term
requires more than the opinion of a few or even the majority
of physicists. There is no general agreement that quantum mechanics
requires us to adopt a new definition of locality no matter how
frustrating that may be for some physicists.
> What is strange about saying that QED is local? It *is* local.
>
> I think your comments are a result of taking such things as the
> Copenhagen interpretation much too literally. Look at Youssef's
> Bayesian approach as a completely different alterantive.
The Copenhagen interpretation is a philisophical
interpretation that I think is very unlikely to be true. I do not
take it literally or any other way. What I do take seriously are
the experimental predictions of quantum mechanics.
> What would be your opinion of the defenders of QM if it had come
> out the other way and they started coming up with various exotic
> loopholes to patch up the disagreement as better and better work
> kept getting further from the QM prediction?
Those are hardly comparable cases.
The difficulty is that these experiments
are trying to rule out all possible local (by the classical definition)
theories.
This includes an infinite number of theories that no one has ever thought
about. The criteria for a conclusive experiment are different than
if one were trying to rule out a specific theory It may be hard for
you to imagine how a classical local
theory could explain the existing experimental results but nature is
quite capable of being for more clever than any of us. That is why
we have to do experiments.
By the way no scientific theory needs defenders. Its `defense' is
the experimental record. Quantum mecahanics is an extraordinary
achievement with predictions far surpassing in accuracy any
other theory. Will it eventually be seen as only an approximation
like every other theory? Given the historical record and the
conceptual problems of the theory this seems likely.
> >Locality is the single most powerful simplifying assumption in science.
> >To argue that a pair of photons remain entangled after being separated by a
> >billion light years and that the entanglement collapses when one of them
> >is observed is a mind boggling claim ...
> So don't make it mind boggling. It is not mind boggling as long
> as you keep treating them as a _pair_ of photons rather than as
> two independent photons when you know they cannot be independent
> and still conserve angular momentum. Is it puzzling that the
> opposite faces of a die are correlated? Why should it be puzzling
> that two members of a pair are correlated?
What is mind boggling is that the correlation is not between photon
states but between experimental manipulations and experimental
observations. To get the best possible
prediction about what happens to a local photon I have to know what
was done to its partner after it has been separated by a billion light
years.
--
Paul P. Budnik Jr.
pa...@mtnmath.com
Charles Francis
<j...@dirac.csit.fsu.edu> writes:
>In article <Z4uFqUAu...@clef.demon.co.uk>
>Charles Francis <cha...@clef.demon.co.uk> writes:
>>The point is that the property of spin does not exist
>>without the measurement.
> That is like saying that the property of "Charles Francis" does
> not exist without the measurement.
True. If Charles Francis were not able to describe his observation of
himself, "I think therefore I am" or equivalent, then he would have no
way to say he exists. But that is likely to take us away from physics
and in to mysticism. In this ng let's stick to exterior observation of
nature, not the interior observation of self.
> I consider that an extreme
> view of the nature of nature.
But it is an orthodox interpretation of qm, advocated by Dirac,
Heisenberg, Von Neumann etc. And there does not appear to be any other
satisfactory interpretation. If you take a more modern information
theoretic expression of the view then you would say that a superposition
of spin states is a property of the information we have about the state,
and that the probability amplitude is a truth value for the result of a
measurement, if one were done. You could not say that a precise spin
exists without a measurement, or equivalent physical process.
>>Quantum mechanics gives a poor description of
>>the subatomic world because it describes it in terms of the properties
>>of classical measuring apparatus.
> "What if I don't treat the detector classically, what
> will happen to me?"
You will tidy up your language, but ultimately you will be saying the
same thing. That a precise value of spin is something which you get from
an interaction between particle and apparatus, and does not exist until
the two interact. You would end up saying the same things about
classical measurement as well as quantum measurement - for example that
length is always a value produced in measurement not a prior property of
space. And you would also take into account that even in a classical
apparatus there is uncertainty on a quantum scale. I look at things this
way in
http://xxx.lanl.gov/abs/physics/9909047
A Model of Classical and Quantum Measurement
Regards
--
Charles Francis
Frank Wappler
> Quantum mechanics violates classical definitions of locality.
> Bell proved it
No: Bell proved that certain "quantum mechanical expectatation
values cannot be represented, either accurately or arbitrary
closely, in a form [which follows from the hypothesis of a
local hidden-variable description of those expectation values]"
only for some _highly specific and restrictive_ class
of local hidden-variable descriptions
("On the Einstein-Podolski-Rosen Paradox", _Physics_ V 1,
pp 195-200, 1964).
Initially, Bell defines an expectation value as satisfying
both a local hidden-variable description (for _some_ general
local hidden-variable description),
as well as a quantum mechanical description,
if for any set of hidden variables { lambda } there exist
two sets of real numbers { A_lambda } and { B_lambda }
whose elements all have values +1 or -1,
and if there exist two real numbers alpha and beta such that
cos( alpha - beta ) =
Integral_{ lambda }_( A_lambda B_lambda d_lambda ) /
Integral_{ lambda }_( d_lambda ).
(eqs. 2, 3, and 14).
However, following (14), Bell restricts/strengthens that definition
by requiring, for the set { lambda } and along with the sets
{ A_lambda }, { B_lambda } and the numbers alpha, and beta
the existence of _a third_ set { G_lambda } (whose elements
all have values +1 or -1) and the existence of _a third_
real number gamma, that are supposed to satisfy
cos( alpha - gamma ) =
Integral_{ lambda }_( A_lambda G_lambda d_lambda ) /
Integral_{ lambda }_( d_lambda ), and
cos( beta - gamma ) =
Integral_{ lambda }_( B_lambda G_lambda d_lambda ) /
Integral_{ lambda }_( d_lambda ).
Since for certain given real numbers alpha, beta, and gamma there
cannot exist any sets { A_lambda }, { B_lambda }, and { G_lambda }
that would satisfy the three supposed equations, Bell obtains a
contradiction only in this case of the restricted/strengthened
definition, for local hidden-variable descriptions that would
require the existence of all three (or more) of such sets,
for any particular set { lambda }.
But that proof does obviously not apply to local hidden-variable
descriptions which only require the existence of two sets,
{ A_lambda } and { B_lambda }, as used in Bell's
initial, general definition of conditions on expectation values
that would satisfy a quantum mechanical description
together with _some_ a local hidden-variable description.
> and Eberhard proved it without reference to hidden variables.
> [("Bell's Theorem without Hidden Variables", _Nuovo Cimento_,
> V 38 B, pp 75-89, 1977). -- Volume number corrected. FW]
No: Eberhard's analysis (as well as a related one by Clauser,
Horne, Shimony, and Holt's analysis, to which Eberhard refers)
is similarly limited:
For any set of indices { j }, Eberhard supposes the existence
of no less than _four_ sets of real numbers
{ A_j }, { F_j }, { B_j } and { G_j },
whose elements all equal either +1, or -1;
together with four real numbers alpha, beta, phi, and gamma.
Not surprisingly, for certain values of alpha, beta, phi, and gamma
there cannot exist sets { A_j }, { F_j }, { B_j } and { G_j }
to satisfy all of
cos( alpha - beta ) = Sum_{ j }_( A_j B_j ) / Sum_{ j }_( 1 ),
cos( alpha - gamma ) = Sum_{ j }_( A_j G_j ) / Sum_{ j }_( 1 ),
cos( phi - beta ) = Sum_{ j }_( F_j B_j ) / Sum_{ j }_( 1 ), and
cos( phi - gamma ) = Sum_{ j }_( F_j B_j ) / Sum_{ j }_( 1 ).
Nevertheless, for any values of alpha, beta, phi, and gamma
and any four sets if indices { j }, { k }, { l } and { n }
there exist sets { A_j } and { B_j }, { F_k } and { G_k },
{ R_l } and { S_l }, and { U_n } and { V_n } of
of real numbers +1 or -1 for which
cos( alpha - beta ) = Sum_{ j }_( A_j B_j ) / Sum_{ j }_( 1 ),
cos( alpha - gamma ) = Sum_{ k }_( F_k G_k ) / Sum_{ k }_( 1 ),
cos( phi - beta ) = Sum_{ l }_( R_l S_l ) / Sum_{ l }_( 1 ), and
cos( phi - gamma ) = Sum_{ n }_( U_n V_n ) / Sum_{ n }_( 1 ),
consistent with Bell's initial general definition of conditions on
expectation values that would satisfy a quantum mechanical description
together with _some_ a local hidden-variable description
(identifying the set of indices with the set of hidden variables).
Regards, Frank W ~@) R
Paul P. Budnik Jr.
"Daryl McCullough" <da...@cogentex.com> wrote in message
news:9gnvl...@drn.newsguy.com...
> j...@dirac.csit.fsu.edu (Jim Carr) says...
> >IMO, QED is a local theory that is relativistic.
> I think it depends on what you mean by a "local theory". I
> think that there are two parts to a theory being local:
> (1) states are local, and (2) evolution is local. QED without
> wave function collapse is local in the second sense, but not
> in the first sense.
....
I find this confusing since states evolve. If they did not they
would be local. All static models are local. What you may
mean is that fields in QED evolve locally and states do not.
I do not see how collapse is relevant. States evolve in
some version of configuration space which is inherently
nonlocal and nonrelativistic. If you project the resulting
probability distributions back onto physical space you
get an absolutely nonlocal evolution that is in direct contradiction
with special relativity. The actualization of those probabilities
avoids an absolute contradiction with relativity only because
it is impossible to know in which direction the nonlocal effect
traveled. The direction is masked by quantum uncertainty.
As a result the predictions are the same in all frames
of reference. This is not so with the evolution of probability densities.
In setting up configuration space you are forced to select a
frame of reference. In addition state evolution is deterministic.
There is no quantum uncertainty to mask the direction of
the nonlocal effect which is determined by how you
set up configuration space. Hence you can an absolute
contradiction with special relativity.
--
Paul P. Budnik Jr.
Mountain Math Software
pa...@mtnmath.com
www.mtnmath.com
Daryl McCullough
>"Daryl McCullough" <da...@cogentex.com> wrote:
>> I think that there are two parts to a theory being local:
>> (1) states are local, and (2) evolution is local. QED without
>> wave function collapse is local in the second sense, but not
>> in the first sense.
>
>....
>
>I find this confusing since states evolve. If they did not they
>would be local. All static models are local. What you may
>mean is that fields in QED evolve locally and states do not.
>
>I do not see how collapse is relevant.
Because if collapse were a physical phenomenon, then
it would necessarily be nonlocal evolution.
>States evolve in some version of configuration space
>which is inherently nonlocal and nonrelativistic.
In the Heisenberg "picture", you can put all the evolution
into the field operators, while states themselves are static.
The field operators evolve according to a deterministic,
local differential equations.
>If you project the resulting probability distributions back
>onto physical space you get an absolutely nonlocal evolution
>that is in direct contradiction with special relativity.
I disagree. There is no consensus that quantum field theory
in any way contradicts special relativity. There may even
be a consensus that it does not.
Anyway, don't argue with me---I was agreeing with you that
there is a sense in which quantum theory is nonlocal.
Paul P. Budnik Jr.
> Paul Budnik says...
> >"Daryl McCullough" <da...@cogentex.com> wrote:
> In the Heisenberg "picture", you can put all the evolution
> into the field operators, while states themselves are static.
> The field operators evolve according to a deterministic,
> local differential equations.
So you have static states and local evolution and yet locality
is violated. It seems like magic! I would guess that the state
involves a connection between singlet state particles that
adds a constraint to computing probability densities. Thus
everything remains local until you use the model to attempt to
compute probability densities in physical space. Is this correct?
Those probabilities do evolve in a decidedly nonlocal manner
and you have to be able to compute them as they change over
time from QED.
> >If you project the resulting probability distributions back
> >onto physical space you get an absolutely nonlocal evolution
> >that is in direct contradiction with special relativity.
>
> I disagree. There is no consensus that quantum field theory
> in any way contradicts special relativity. There may even
> be a consensus that it does not.
The predictions certainly do not but the mechanisms that
generate those predictions have to evolve in a way that
does.
> Anyway, don't argue with me---I was agreeing with you that
> there is a sense in which quantum theory is nonlocal.
My intuition suggest that locality violation is the modern
version of the black body radiation anomaly of classical physics.
It points the way to a new and deeper theory. Thus I
think its important to focus on how nonlocality enters into
the mathematics of QM. When you say there is local state
evolution and local field evolution you may be technically
correct but it is confusing if not misleading. For there has to
be a mechanism within the mathematics that produces locality
violation. You can avoid invoking that mechanism until
you need to compute probability densities but you cannot
eliminate it.
--
Paul P. Budnik Jr.
Jim Carr
Jim Carr <j...@dirac.csit.fsu.edu> writes:
|
| In article <Z4uFqUAu...@clef.demon.co.uk>
| Charles Francis <cha...@clef.demon.co.uk> writes:
| >The point is that the property of spin does not exist
| >without the measurement.
|
| That is like saying that the property of "Charles Francis" does
| not exist without the measurement.
In article <tfv4zgAa...@clef.demon.co.uk>
Charles Francis <cha...@clef.demon.co.uk> writes:
>
>True. If Charles Francis were not able to describe his observation of
>himself, "I think therefore I am" or equivalent, then he would have no
>way to say he exists. But that is likely to take us away from physics
>and in to mysticism. In this ng let's stick to exterior observation of
>nature, not the interior observation of self.
But IMO it is the analogy to an interior observation that you
were talking about. The exterior observation of the atom
itself is not subject to questions of existence.
| I consider that an extreme
| view of the nature of nature.
>But it is an orthodox interpretation of qm, advocated by Dirac,
>Heisenberg, Von Neumann etc. And there does not appear to be any other
>satisfactory interpretation. If you take a more modern information
>theoretic expression of the view then you would say that a superposition
>of spin states is a property of the information we have about the state,
>and that the probability amplitude is a truth value for the result of a
>measurement, if one were done. You could not say that a precise spin
>exists without a measurement, or equivalent physical process.
But you could say that the atom exists, just as you could say
that spin itself exists (the total, not the projection), in all
of those interpretations.
| >Quantum mechanics gives a poor description of
| >the subatomic world because it describes it in terms of the properties
| >of classical measuring apparatus.
|
| "What if I don't treat the detector classically, what
| will happen to me?"
>You will tidy up your language, but ultimately you will be saying the
>same thing. <... snip ...>
I don't think Wigner or I would agree that it is the same thing,
because there would be no classical apparatus. I would agree that
(IMO) the result of a full calculation of that type would have the
same "observational" results (meters count, etc).
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
e-mail info: new...@fbi.gov pyr...@ftc.gov enfor...@sec.gov
ab...@aol.com ab...@yahoo.com ab...@hotmail.com
Harry Johnston
> So you have static states and local evolution and yet locality
> is violated. It seems like magic! I would guess that the state
> involves a connection between singlet state particles that
> adds a constraint to computing probability densities. Thus
> everything remains local until you use the model to attempt to
> compute probability densities in physical space. Is this correct?
> Those probabilities do evolve in a decidedly nonlocal manner
> and you have to be able to compute them as they change over
> time from QED.
What do you mean by this? To the best of my knowledge, the
probabilities do NOT evolve nonlocally, except during wavefunction
collapse (which isn't essential anyway).
Can you provide an example of what you consider nonlocal evolution of
the probabilities?
Harry.
---
Harry Johnston, om...@ihug.co.nz
One Earth, One World, One People
Mike York
> In article <9ged7f$sj5$1...@news.fsu.edu>,
> Jim Carr <j...@dirac.csit.fsu.edu> writes:
> | In article <Z4uFqUAu...@clef.demon.co.uk>
> | Charles Francis <cha...@clef.demon.co.uk> writes:
> | >The point is that the property of spin does not exist
> | >without the measurement.
> >But it is an orthodox interpretation of qm, advocated by Dirac,
> >Heisenberg, Von Neumann etc. And there does not appear to be any other
> >satisfactory interpretation. If you take a more modern information
> >theoretic expression of the view then you would say that a superposition
> >of spin states is a property of the information we have about the state,
> >and that the probability amplitude is a truth value for the result of a
> >measurement, if one were done. You could not say that a precise spin
> >exists without a measurement, or equivalent physical process.
> But you could say that the atom exists, just as you could say
> that spin itself exists (the total, not the projection), in all
> of those interpretations.
Hmm.. The spin of what? How do you know what is there until you detect it's
properties, such as its (total) spin?
Classically there is an implicit assumption of continued existence. Hence we
are able to talk about observing the trajectory of a definite object, for
instance. But quantum mechanically we can only make discrete observations. So
typically we again project the continued existence of an object between
observations on grounds of continuity. But there are two difficulties with
this:
1. If there is an interaction and the object is changed in some way, perhaps
even its identity (i.e. type of object) at what point, since it was last
observed, did the original object cease to exist? And what other objects
existed between observations?
2. Even if we observe the same *type* of object, how do we know it is the
same one as in our last observation and that it didn't suffer some sequence
of transforming interactions, or even change places with a similar object
(identical particle exchange) in between observations?
I think we must recognize that classical assumptions of continued existence
do not hold in the quantum realm. And hence any prior knowledge of the
(total) spin in advance of a measurement is conditional on an assumption that
such a value would result if/when a measurement is made. I think this would
be a reasonable assumption if, and only if, the experiment was set up to
exclude any other result.
Mike
Tom Trotter
> "Daryl McCullough" <da...@cogentex.com> wrote in message
> news:9hq2d...@drn.newsguy.com...
> > Paul Budnik says...
> > >"Daryl McCullough" <da...@cogentex.com> wrote:
> > In the Heisenberg "picture", you can put all the evolution
> > into the field operators, while states themselves are static.
> > The field operators evolve according to a deterministic,
> > local differential equations.
> So you have static states and local evolution and yet locality
> is violated. It seems like magic! I would guess that the state
> involves a connection between singlet state particles that
> adds a constraint to computing probability densities. Thus
> everything remains local until you use the model to attempt to
> compute probability densities in physical space. Is this correct?
> Those probabilities do evolve in a decidedly nonlocal manner
> and you have to be able to compute them as they change over
> time from QED.
'Locality', in this context, seems to refer to relationships,
properties, events, etc., presumably occuring on a scale that we don't
directly (and sometimes not even indirectly) experience.
'Nonlocality', in this context, refers to our observations of certain
experimental events. The 'tests' of Bell's Inequalities have taught
us that it isn't a good idea to assume anything (necessarily) about a
scale of interactions that we can't, even in principle, experience.
They've also taught us, I think, that using the terms 'locality' and
'nonlocality' is probably not the best approach to clarify what's
going on.
Why isn't Bell's (or a Bell-like) formulation a good way of talking
about what's happening in some of these experiments? It specifies a
linear relationship between the relative orientations of the detectors
and the coincidence of detection. We knew from previous experiments
in optics that the relationship probably isn't linear. It's not
necessary to talk about what anybody assumed or didn't assume.
Jim Carr
|
<... some earlier snippage by York noted ...>
|
| In article <9ged7f$sj5$1...@news.fsu.edu>,
| Jim Carr <j...@dirac.csit.fsu.edu> writes:
| | In article <Z4uFqUAu...@clef.demon.co.uk>
| | Charles Francis <cha...@clef.demon.co.uk> writes:
| | >The point is that the property of spin does not exist
| | >without the measurement.
|
| >Heisenberg, Von Neumann etc. And there does not appear to be any other
| >satisfactory interpretation. If you take a more modern information
| >theoretic expression of the view then you would say that a superposition
| >of spin states is a property of the information we have about the state,
| >and that the probability amplitude is a truth value for the result of a
| >measurement, if one were done. You could not say that a precise spin
| >exists without a measurement, or equivalent physical process.
|
| But you could say that the atom exists, just as you could say
| that spin itself exists (the total, not the projection), in all
| of those interpretations.
In article <3B489D8B...@home.com>
Mike York <mike...@home.com> writes:
>
>Hmm.. The spin of what?
The atom in the (say) Stern-Gerlach apparatus.
>How do you know what is there until you detect it's
>properties, such as its (total) spin?
My point was that every QM interpretation agrees that the J of
atom in its ground state is not unknown when an atom is present.
>Classically there is an implicit assumption of continued existence. Hence we
>are able to talk about observing the trajectory of a definite object, for
>instance. But quantum mechanically we can only make discrete observations. So
>typically we again project the continued existence of an object between
>observations on grounds of continuity.
We do in QM, but not in a QFT. The above was in the context of
quantum mechanics. In that context, empirically confirmed
conservation laws also lead one to conclude that the baryons
and leptons are always there.
>But there are two difficulties with this:
>
>1. If there is an interaction and the object is changed in some way, perhaps
>even its identity (i.e. type of object) at what point, since it was last
>observed, did the original object cease to exist? And what other objects
>existed between observations?
Could be a real issue within a QFT or working with a collection
of particles, but I don't see it for an atom.
>2. Even if we observe the same *type* of object, how do we know it is the
>same one as in our last observation and that it didn't suffer some sequence
>of transforming interactions, or even change places with a similar object
>(identical particle exchange) in between observations?
We don't. We also do not have to, since the wavefunctions
have the appropriate bosonic or fermionic symmetry.
>I think we must recognize that classical assumptions of continued existence
>do not hold in the quantum realm. And hence any prior knowledge of the
>(total) spin in advance of a measurement is conditional on an assumption that
>such a value would result if/when a measurement is made. I think this would
>be a reasonable assumption if, and only if, the experiment was set up to
>exclude any other result.
Hmmm. I would say it is a reasonable assumption if the experiments
do not explicitly exclude it but show that the atom stays in a
particular energy state with only spin projections changing.
Charles Francis
<j...@dirac.csit.fsu.edu> writes
>In article <9ged7f$sj5$1...@news.fsu.edu>,
>Jim Carr <j...@dirac.csit.fsu.edu> writes:
>|
>| In article <Z4uFqUAu...@clef.demon.co.uk>
>| Charles Francis <cha...@clef.demon.co.uk> writes:
>| >The point is that the property of spin does not exist
>| >without the measurement.
>|
>| That is like saying that the property of "Charles Francis" does
>| not exist without the measurement.
>
> The exterior observation of the atom
> itself is not subject to questions of existence.
>
> But you could say that the atom exists, just as you could say
> that spin itself exists (the total, not the projection), in all
> of those interpretations.
>
I distinguish between the existence of the atom, and the existence of
the property spin. The atom exists without the measurement; in general a
crisp value of spin does not.
>| >Quantum mechanics gives a poor description of
>| >the subatomic world because it describes it in terms of the properties
>| >of classical measuring apparatus.
>|
>| "What if I don't treat the detector classically, what
>| will happen to me?"
>
>>You will tidy up your language, but ultimately you will be saying the
>>same thing. <... snip ...>
>
> I don't think Wigner or I would agree that it is the same thing,
> because there would be no classical apparatus. .
I don't follow you. A measuring apparatus just extracts a number from a
physical process involving both particle and apparatus. Perhaps that is
not a classical description of an apparatus unless you also try to fit
the set of all possible results of measurement into a classical
description of space time.
>
Regards
--
Charles Francis
Mike York
Jim Carr wrote:
> Jim Carr wrote in <9i5n6r$3o4$1...@news.fsu.edu>:
> |
> <... some earlier snippage by York noted ...>
> |
> |< sorry, I had to snip some more, to make this post readable>
>
> | But you could say that the atom exists, just as you could say
> | that spin itself exists (the total, not the projection), in all
> | of those interpretations.
>
> In article <3B489D8B...@home.com>
> Mike York <mike...@home.com> writes:
> >
> >Hmm.. The spin of what?
>
> The atom in the (say) Stern-Gerlach apparatus.
>
> >How do you know what is there until you detect it's
> >properties, such as its (total) spin?
>
> My point was that every QM interpretation agrees that the J of
> atom in its ground state is not unknown when an atom is present.
>
I disagree with this. I will agree however that *for an isolated state*, whatever
its (undetermined and, between measurements, indeterminable) internal composition
(one atom, two atoms, an ion and an electron, a planet + an anti-planet + an atom,
etc, etc), the total J will be the same for each subsequent measurement as any
previous measurement.
>
> >Classically there is an implicit assumption of continued existence. Hence we
> >are able to talk about observing the trajectory of a definite object, for
> >instance. But quantum mechanically we can only make discrete observations. So
> >typically we again project the continued existence of an object between
> >observations on grounds of continuity.
>
> We do in QM, but not in a QFT.
Ok..
> The above was in the context of
> quantum mechanics. In that context, empirically confirmed
> conservation laws also lead one to conclude that the baryons
> and leptons are always there.
>
I disagree. They tell us only that the total baryonic and leptonic number will be
the same for each succeeding measurement. I hope this doesn't sound like
nit-picking. I really am trying to make a point that I think is worth making.
>
> >But there are two difficulties with this:
> >
> >1. If there is an interaction and the object is changed in some way, perhaps
> >even its identity (i.e. type of object) at what point, since it was last
> >observed, did the original object cease to exist? And what other objects
> >existed between observations?
>
> Could be a real issue within a QFT or working with a collection
> of particles, but I don't see it for an atom.
>
I presume this is because you think of an atom as stable (i.e. below threshold for
any alternative system with the same conserved quantum numbers). But one of the
strange things about quantum mechanics is the possibility of "virtual" particles
(or virtual atoms?) which may violate such conditions between measurements. In
essence, this is another statement that the state of the system between
measurements is indeterminable.
Even if this were not the case, why do you exclude the possibility of an incoming
photon putting the atom in an excited state or even ionizing it? Possibly because
your apparatus is able to detect such an event and exclude it from the experiment.
But this is the sort of case I was referring to below ("...the experiment was set
up to
exclude any other result").
>
> >2. Even if we observe the same *type* of object, how do we know it is the
> >same one as in our last observation and that it didn't suffer some sequence
> >of transforming interactions, or even change places with a similar object
> >(identical particle exchange) in between observations?
>
> We don't. We also do not have to, since the wavefunctions
> have the appropriate bosonic or fermionic symmetry.
>
> >I think we must recognize that classical assumptions of continued existence
> >do not hold in the quantum realm. And hence any prior knowledge of the
> >(total) spin in advance of a measurement is conditional on an assumption that
> >such a value would result if/when a measurement is made. I think this would
> >be a reasonable assumption if, and only if, the experiment was set up to
> >exclude any other result.
>
> Hmmm. I would say it is a reasonable assumption if the experiments
> do not explicitly exclude it but show that the atom stays in a
> particular energy state with only spin projections changing.
>
Ok. I think this is saying the same as I intended. The essential point being that
the notion that all the other properties other than the spin projection remaining
fixed has to be a condition of the experimental setup (including, in particular,
an event filter designed to exclude events that violate this condition -- such as
an interaction with a stray incoming particle) in order for the statement that
they are all fixed to be guaranteed true.
Jim Carr
| >But it is an orthodox interpretation of qm, advocated by Dirac,
| >Heisenberg, Von Neumann etc. And there does not appear to be any other
| >satisfactory interpretation. If you take a more modern information
| >theoretic expression of the view then you would say that a superposition
| >of spin states is a property of the information we have about the state,
| >and that the probability amplitude is a truth value for the result of a
| >measurement, if one were done. You could not say that a precise spin
| >exists without a measurement, or equivalent physical process.
|
| that spin itself exists (the total, not the projection), in all
| of those interpretations.
In article <9jkouo$opb03$1...@ID-28113.news.dfncis.de>
Charles Francis <cha...@clef.demon.co.uk> writes:
>
>I distinguish between the existence of the atom, and the existence of
>the property spin. The atom exists without the measurement; in general a
>crisp value of spin does not.
But it can have a crisp value of J and not have a crisp value
of J_z, so you *can* say that the spin exists.
<... more snipped ...>
| | "What if I don't treat the detector classically, what
| | will happen to me?"
|
| >You will tidy up your language, but ultimately you will be saying the
| >same thing. <... snip ...>
|
| I don't think Wigner or I would agree that it is the same thing,
| because there would be no classical apparatus. .
>I don't follow you. A measuring apparatus just extracts a number from a
>physical process involving both particle and apparatus.
That would be a classical detector.
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
SirCam Warning: read http://www.cert.org/advisories/CA-2001-22.html
Jim Carr
|
| In article <3B489D8B...@home.com>
| Mike York <mike...@home.com> writes:
| >Hmm.. The spin of what?
|
| The atom in the (say) Stern-Gerlach apparatus.
|
| >How do you know what is there until you detect it's
| >properties, such as its (total) spin?
|
| My point was that every QM interpretation agrees that the J of
| atom in its ground state is not unknown when an atom is present.
In article <9jnfl8$ae78$1...@ID-28113.news.dfncis.de>
Mike York <mike...@home.com> writes (with long line lengths):
>
>I disagree with this. I will agree however that *for an isolated state*, whatever
>its (undetermined and, between measurements, indeterminable) internal composition
>(one atom, two atoms, an ion and an electron, a planet + an anti-planet + an atom,
>etc, etc), the total J will be the same for each subsequent measurement as any
>previous measurement.
I trust that you also agree that once a wavefunction is written down
for the atom, that every *interpretation* agrees on whether it is,
say, an eigenstate of J. That was my point.
<... snip point on baryon conservation that I will agree to disagree on ...>
| >But there are two difficulties with this:
| >1. If there is an interaction and the object is changed in some way, perhaps
| >even its identity (i.e. type of object) at what point, since it was last
| >observed, did the original object cease to exist? And what other objects
| >existed between observations?
|
| Could be a real issue within a QFT or working with a collection
| of particles, but I don't see it for an atom.
>I presume this is because you think of an atom as stable (i.e. below threshold for
>any alternative system with the same conserved quantum numbers). But one of the
>strange things about quantum mechanics is the possibility of "virtual" particles
>(or virtual atoms?) which may violate such conditions between measurements. In
>essence, this is another statement that the state of the system between
>measurements is indeterminable.
I was not discussing this in terms of a QFT. I make my statement
because the energetics at room temperature are such that the
probability of finding it in other states is negligible.
--
James Carr <j...@scri.fsu.edu> http://www.scri.fsu.edu/~jac/
SirCam Warning: read http://www.cert.org/advisories/CA-2001-22.html
e-mail info: new...@fbi.gov pyr...@ftc.gov enfor...@sec.gov