Modern version of LET (was: Einstein vs. Poincare)
Klaus Kassner
Matthew P Wiener wrote:
> Klaus Kassner wrote:
>> Matthew P Wiener wrote:
> >> >> No, my statement is correct. Perhaps you'd like to bring phlogiston
> >> >> theory up to date, and tell us that the result _is_ phlogiston theory?
> >> Perhaps you would care to answer my question above?
> >Why should I?
> Since in attempting to do so and failing miserably, you would see the
> error of your ways. Much better than trying to force petty and silly
> "I know I'm right so there" games.
> > I hope you are intelligent enough to see
> >your mistake yourself.
> See what I mean?
:No. I will give you a little time to think.
:Tune in for the answer next week.
The week is over, so here is the answer.
First, I don't feel inclined to bring *any*
discarded theory up to date, neither phlogiston
theory, nor LET. I am sufficiently satisfied
with the current status of the theories of heat
transfer and of SR.
Second, *if* I saw an updated version of
phlogiston theory, I might be curious enough
to consider it.
Third, I do not *have* to bring LET up to date,
because this *has been done* by others.
A recent reference is: "Grenzgeschwindigkeiten
und ihre Paradoxa. Gitter, Aether, Relativitaet."
by H. Guenther, Teubner, Stuttgart, 1996.
The author is a tenured German professor, not
a crank.
In case, you don't read German (or are not
willing to look into the literature at all),
I give a short summary of the book (I have read
only half of it so far, and won't have time for
a while to read the rest).
The translation of the title is: "Limit speeds
and their paradoxa. Lattices, ether, relativity."
The author considers the dynamics of dislocations
in a crystal. He exhibits the derivation (given
earlier by Seeger) of the equation of motion of
these defects in the continuum limit. This is
the sine gordon equation, but with the speed of
light replaced by the speed of sound.
He then asks the question: given this dynamics,
what would be the appropriate way to measure lengths
and time in the "world of dislocations". Now, due
to the nonlinearity of the sine gordon equation
you can use it to define a length and a time unit
(which you can't, for example, with Maxwell's
equations which only give you a ratio of a length
and a time unit).
The length unit is defined by the width of the
region of change of the basic stationary kink
solution to the equation (a hyperbolic tangent),
the time unit by the periodicity of the
basic breather solution. These solutions exist
in moving frames, too. It turns out, not
surprisingly, that if you adopt these rules
for length and time measurement, the resulting
metric is the Minkowski metric, with the
limit speed given by the speed of sound.
So we have a model for special relativity,
valid in the world of dislocations, which is
essentially Lorentz ether theory.
What are the features of this model?
a) Ether is easily identifiable; it is the
original crystal.
b) The Lorentz symmetry is not postulated,
it results from the continuum limit of
a lattice possessing neither Galilean
nor Lorentzian symmetry (but a discrete
translation symmetry). I guess this is about
the closest one can get to the derivation
of the Lorentz transformations from
something "more fundamental" as desired
by Ed Green.
c) The entire dynamics is dynamics of
defects in this ether.
d) There is no "crippling" tie to EM
whatsoever.
e) The ether is essentially undetectable.
Essentially, because if you allow external
events that break the Lorentz symmetry,
you can detect it.
f) The limiting speed is the speed of sound.
Of course, we know that you can have things
go faster than at the speed of sound in the
crystal, e.g. by shooting some particle into
it that is fast enough. Would this not lead
to a violation of causality? No. For some
observers the fast particle would move into
the past; but this could not be used to
send things into their *own* past, since for
that you would have to be able to move fast
enough in the forward and backward directions.
But this is not possible. For moving frames,
sending a signal in the direction of motion
would be possible with arbitrary speed (even
into the past) while in the opposite direction,
there would be an upper speed limit c^2/v
(where v is the velocity of the frame w.r.t.
the crystal). That a causality violation
does not occur, follows immediately from the
description of events in the distinguished
frame, for which sending a signal forth and
back between two observers will always take
a positive time. Causality violation at
faster-than-light speeds in SR results from
assuming the Lorentz symmetry not to be
broken at these speeds. In our toy model,
it *is* broken for speeds in excess of the
speed of sound.
Clearly, the conclusion that there *must* be an
ether in our world as there is one in this
toy model is as *inadmissible* as the conclusion
that since we cannot imagine curved surfaces
except when embedded in flat three space
curved three-space must be embedded in flat
four (to six) space.
So what would be the advantage of such a LET
as a replacement of SR?
In principle, it would be able to accomodate
certain faster-than-light signal transmissions
which are not allowed by SR. But it could do
so only at the price of a breaking of the
Lorentz symmetry. Now is there evidence for
such events in our world? I would say, no.
For example, the oft-cited violation of
Bell's inequalities cannot (if there is a
need for its explanation at all) be explained
by LET. The reasons: a) we have good evidence
that the Schroedinger-typ QM holds in *any*
inertial frame at low velocities (we have
the spectra of distant galaxies); b) QM predicts
the instantaneous collapse of the wave function
in *any one* of these systems. If we assumed the
wave function to be some objective field and
its instantaneous collapse to signify some kind
of superluminal transport, this would have to
work for all inertial systems. In LET, it
doesn't; LET could explain the prediction of
QM only in a single inertial system.
Now, as regards the experimental situation, it
is much less clear. We have Aspect's and similar
experiments but these were all done on earth,
i.e. essentially in a single inertial system.
So, *if* we happened to live in an inertial
system with a velocity close to zero (w.r.t.
the absolute frame defined by ether) -- a
very unlikely scenario -- , then LET might explain
the experiment but would in addition predict
QM to fail in *different* inertial systems. An
experiment to rule this out would be to redo
the Aspect experiment on a space ship moving
fast enough w.r.t. earth. My suspicion would
be that the QM prediction will be verified and
the LET explanation ruled out.
So, the current status is that SR and LET are
equivalent for all practical purposes but that
SR makes more predictions (impossibility of
superluminal transport, nonobjectivity of the
quantum wave function), i.e. is more easily
falsifiable, in principle. Thus SR should be
favoured. (For Ilja: The same kind of
predictions would, by the way apply to GR in
comparison with PG, wherefore GR should be
preferred over PG).
By the way, I find the explicit construction of
an ether model based on a discrete symmetry group
quite fascinating. If a similar thing could be
done for GR (I don't know if that is possible),
this would definitely be worth publishing.
But this means not to *postulate*
the properties of the ether (as done in Ilja's
approach) but to derive them from a model
(of the underlying crystal).
Matthew P Wiener
In article <33D748...@physik.uni-magdeburg.de>, Klaus Kassner <klaus.kassner@physik writes:
>Matthew P Wiener wrote:
>> Klaus Kassner wrote:
>>> Matthew P Wiener wrote:
>> >> >> No, my statement is correct. Perhaps you'd like to bring
>> >> >> phlogiston theory up to date, and tell us that the result
>> >> >> _is_ phlogiston theory?
>> >> Perhaps you would care to answer my question above?
>> >Why should I?
>> Since in attempting to do so and failing miserably, you would see the
>> error of your ways. Much better than trying to force petty and silly
>> "I know I'm right so there" games.
>The week is over, so here is the answer.
You *still* did not answer my question.
>First, I don't feel inclined to bring *any* discarded theory up to
>date, neither phlogiston theory, nor LET. I am sufficiently satisfied
>with the current status of the theories of heat transfer and of SR.
I didn't *ask* you to bring any discarded theory up to date. Sheesh.
>Second, *if* I saw an updated version of phlogiston theory, I might
>be curious enough to consider it.
But do we speak of such a hypothesized revised theory as "phlogiston
theory"? The answer is no. That is all.
>Third, I do not *have* to bring LET up to date, because this *has
>been done* by others.
And is this Lorentz's ether theory?
>A recent reference is: "Grenzgeschwindigkeiten und ihre
>Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
>Stuttgart, 1996. The author is a tenured German professor, not a
>crank.
No, it is Guenther's revised ether theory.
>In case, you don't read German (or are not willing to look into the
>literature at all), I give a short summary of the book (I have read
>only half of it so far, and won't have time for a while to read the
>rest).
I'm willing to look in the U Penn library, and it's not there.
--
-Matthew P Wiener (wee...@sagi.wistar.upenn.edu)
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> Third, I do not *have* to bring LET up to date,
> because this *has been done* by others.
>
> A recent reference is: "Grenzgeschwindigkeiten
> und ihre Paradoxa. Gitter, Aether, Relativitaet."
> by H. Guenther, Teubner, Stuttgart, 1996.
> The author is a tenured German professor, not
> a crank.
Thank you for this reference. Have you more information
(e-mail etc.?)
> So we have a model for special relativity,
> valid in the world of dislocations, which is
> essentially Lorentz ether theory.
>
> What are the features of this model?
>
> a) Ether is easily identifiable; it is the
> original crystal.
> b) The Lorentz symmetry is not postulated,
> it results from the continuum limit of
> a lattice possessing neither Galilean
> nor Lorentzian symmetry (but a discrete
> translation symmetry). I guess this is about
> the closest one can get to the derivation
> of the Lorentz transformations from
> something "more fundamental" as desired
> by Ed Green.
> c) The entire dynamics is dynamics of
> defects in this ether.
> d) There is no "crippling" tie to EM
> whatsoever.
> e) The ether is essentially undetectable.
> Essentially, because if you allow external
> events that break the Lorentz symmetry,
> you can detect it.
> f) The limiting speed is the speed of sound.
What I have named "atomic ether theory".
> So what would be the advantage of such a LET
> as a replacement of SR?
> In principle, it would be able to accomodate
> certain faster-than-light signal transmissions
> which are not allowed by SR. But it could do
> so only at the price of a breaking of the
> Lorentz symmetry. Now is there evidence for
> such events in our world? I would say, no.
I'm not.
> For example, the oft-cited violation of
> Bell's inequalities cannot (if there is a
> need for its explanation at all) be explained
> by LET.
But they are at least not in conflict with LET.
> The reasons:
Fine, seems to be your old, wrong refutation of quantum PG.
Let's discuss it now in detail.
> a) we have good evidence
> that the Schroedinger-typ QM holds in *any*
> inertial frame at low velocities (we have
> the spectra of distant galaxies);
We have good agreement with SR-QFT. That's all.
> b) QM predicts
> the instantaneous collapse of the wave function
> in *any one* of these systems.
We can choose different frames, make QM computations and are in
agreement with observation. If some real wave function really
instantaneously collapses we cannot conclude from these observations.
That's all.
SR-QFT shows that we cannot detect any non-Lorentz-invariant effects
by observation of infinite scattering matrices. LET-QFT is (of
course, only in the continuous approximation, no longer in the crystal
model) after omitting all unobservable things, SR-QFT in one frame,
named the preferred frame.
If we use the wrong frame in LET-QFT, we do not describe the reality
correctly. But this incorrect description doesn't matter, because the
observable predictions at least for infinite scattering coinside.
Thus, we are not able to detect this error. Especially, there is no
difference between the predictions of LET-QFT in the correct (or a
wrong) frame, as far as they are observable from point of view of SR,
and the predictions of SR-QFT in any frame.
> If we assumed the
> wave function to be some objective field and
> its instantaneous collapse to signify some kind
> of superluminal transport, this would have to
> work for all inertial systems.
Why? LET is not Lorentz-invariant, even if their observable
predictions are. The unobservable wave function may be different and
incompatible between different Lorentz-frames, as well as the
unobservable absolute contemporaneity and the state of the ether are.
> In LET, it
> doesn't; LET could explain the prediction of
> QM only in a single inertial system.
What does this mean? We are free to use the coordinates we like, in
every theory. In LET, we have a simplest choice and prefer this
system. But we can use every other system too.
> Now, as regards the experimental situation, it
> is much less clear. We have Aspect's and similar
> experiments but these were all done on earth,
> i.e. essentially in a single inertial system.
> So, *if* we happened to live in an inertial
> system with a velocity close to zero (w.r.t.
> the absolute frame defined by ether) -- a
> very unlikely scenario -- , then LET might explain
> the experiment but would in addition predict
> QM to fail in *different* inertial systems. An
> experiment to rule this out would be to redo
> the Aspect experiment on a space ship moving
> fast enough w.r.t. earth. My suspicion would
> be that the QM prediction will be verified and
> the LET explanation ruled out.
Complete nonsense, sorry. The SR-QFT prediction (violation of Bell's
inequality) will be verified if LET is correct, because it agrees with
the LET prediction in the preferred frame, whatever we choose as the
preferred frame (even if we assume a wrong one to be the preferred for
the computation).
To avoid confusion: SR-QFT I name here the theory which allows to
compute some probabilities of the scattering matrix, with Bohr's
notion of causality (IMO the same as to say "without causality" ;-).
Experimentally falsified by Aspect is SR with EPR-like causality. LET
is compatible with EPR-causality and the violation of Bell's
inequality.
> So, the current status is that SR and LET are
> equivalent for all practical purposes but that
> SR makes more predictions (impossibility of
> superluminal transport,
More detailed explanation please. Which type of observation which is
in agreement with LET falsifies SR? (remark: an observable
superluminal phone line would be in contradiction with LET, because it
allows to built an absolute clock, but all clocks are time-dilated by
interaction with the ether.)
Moreover, I want to play the following game with you: Assume there is
a working superluminal phone line. But I doubt that this strange
device violates relativity. You have to prove that it really violated
relativity by interpreting the observations you can make. I criticize
your interpretation using standard relativistic arguments, used to
show that the violation of Bell's inequality does not falsify
relativity. (I want to see your rejection of these arguments ;-)
> nonobjectivity of the quantum wave function),
Interesting, which observation allows to falsify the "non-objectivity"
of the wave function.
> i.e. is more easily
> falsifiable, in principle. Thus SR should be
> favoured. (For Ilja: The same kind of
> predictions would, by the way apply to GR in
> comparison with PG, wherefore GR should be
> preferred over PG).
That's why I'm very interested. But let's remember that for PG vs GR
we have found some possibilities for falsification of PG: detect a
wormhole, fall behind the horizon of a black hole. Moreover, we have
now a lot of quantum gravity predictions, use gr-qc/9706055 to compute
what you like and try to falsify. Unfortunately, we have no quantum
GR predictions to compare with.
> By the way, I find the explicit construction of
> an ether model based on a discrete symmetry group
> quite fascinating. If a similar thing could be
> done for GR (I don't know if that is possible),
> this would definitely be worth publishing.
Means, to unify this sub-LET construction with my PG? Good idea for
future research.
> But this means not to *postulate*
> the properties of the ether (as done in Ilja's
> approach) but to derive them from a model
> (of the underlying crystal).
Means to postulate the properties of the crystal and to derive the
continuous approximation. Postulating is always involved. And the
continuous ether theory is interesting enough even without an
underlying atomic model.
But even if this will be done, you probably will reject publication
until we have successfully quantized this theory. And after this, you
will probably simply ignore this theory as not beautiful enough or so.
(as you ignore gr-qc/9706055).
Ilja
--
I. Schmelzer, D-10178 Berlin, Keibelstr. 38, <schm...@wias-berlin.de>
http://www.cyberpass.net/~ilja
Ilja Schmelzer
wee...@sagi.wistar.upenn.edu (Matthew P Wiener) writes:
> >Third, I do not *have* to bring LET up to date, because this *has
> >been done* by others.
>
>
> >Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
> >Stuttgart, 1996. The author is a tenured German professor, not a
> >crank.
>
Formally correct. But why increase confusion among other readers? If
we talk about Guenther's revised ether theory, nobody has an idea what
we are talking about.
I suggest to name LET the theory which coinsides with SR but adds one
preferred frame. This preferred frame is used to define absolute
time, absolute contemporaneity, absolute space, and the rest frame of
a stationary ether. Every material clock is time-dilated if in
movement against the ether by the same well-known formula.
This is the straightforward generalization of LET to the current
state, with strong and weak forces and QFT. To reject to name it LET
is artificial because of the triviality of this modification. The
ether dilates all clocks in the same way.
For the case of gravity, there it is not so obvious, a new name for
the generalization is really justified - that's why I have introduced
the name post-relativistic gravity (PG).
Guenther's revised ether theory is an crystal model which coinsides
with this LET in the large scale approximation, thus, really something
different.
But for SR with preferred frame the name LET is appropriate, as well
as we use the name SR today too.
Klaus Kassner
Matthew P Wiener wrote:
>
> In article <33D748...@physik.uni-magdeburg.de>, Klaus Kassner <klaus.kassner@physik writes:
>
> >> Klaus Kassner wrote:
> >>> Matthew P Wiener wrote:
>
> >> >> >> No, my statement is correct. Perhaps you'd like to bring
> >> >> >> phlogiston theory up to date, and tell us that the result
> >> >> >> _is_ phlogiston theory?
>
I think I did.
> >First, I don't feel inclined to bring *any* discarded theory up to
> >date, neither phlogiston theory, nor LET. I am sufficiently satisfied
> >with the current status of the theories of heat transfer and of SR.
> I didn't *ask* you to bring any discarded theory up to date. Sheesh.
It seemed to me you did.
> >Second, *if* I saw an updated version of phlogiston theory, I might
> >be curious enough to consider it.
> But do we speak of such a hypothesized revised theory as "phlogiston
> theory"? The answer is no. That is all.
No. The answer could be yes. Depending on how many changes
were introduced.
> >Third, I do not *have* to bring LET up to date, because this *has
> >been done* by others.
> And is this Lorentz's ether theory?
Yes, unless the changes made are sufficient to warrant a new name.
> >A recent reference is: "Grenzgeschwindigkeiten und ihre
> >Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
> >Stuttgart, 1996. The author is a tenured German professor, not a
> >crank.
> No, it is Guenther's revised ether theory.
Most definitely not. He is not the only one looking at
this type of theory. The class name is "Lorentz ether theory".
> I'm willing to look in the U Penn library, and it's not there.
Sure. It is from 1996, and in German. What do you expect?
My claim is there *are* modern versions of LET. Yours was
there aren't. Usually one counterexample suffices to bring
down a general claim.
Klaus Kassner
Ilja Schmelzer wrote:
>
> Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > A recent reference is: "Grenzgeschwindigkeiten
> > und ihre Paradoxa. Gitter, Aether, Relativitaet."
> > by H. Guenther, Teubner, Stuttgart, 1996.
> > The author is a tenured German professor, not
> > a crank.
>
> Thank you for this reference. Have you more information
> (e-mail etc.?)
No, I just have the book.
> Fine, seems to be your old, wrong refutation of quantum PG.
> Let's discuss it now in detail.
>
> > a) we have good evidence
> > that the Schroedinger-typ QM holds in *any*
> > inertial frame at low velocities (we have
> > the spectra of distant galaxies);
>
> We have good agreement with SR-QFT. That's all.
No, we have good agreement even
with the nonrelativistic Schroedinger theory.
> > b) QM predicts
> > the instantaneous collapse of the wave function
> > in *any one* of these systems.
> We can choose different frames, make QM computations and are in
> agreement with observation. If some real wave function really
> instantaneously collapses we cannot conclude from these observations.
Right. We can even *exclude* it from SR. But the point is the
"real". The wave function need not be real in any sense, it
need not be objective. Then of course it can collapse
instantaneously in any frame of reference.
> > If we assumed the
> > wave function to be some objective field and
> > its instantaneous collapse to signify some kind
> > of superluminal transport, this would have to
> > work for all inertial systems.
>
> Why?
Because of experimental evidence.
> LET is not Lorentz-invariant, even if their observable
> predictions are. The unobservable wave function may be different and
> incompatible between different Lorentz-frames, as well as the
> unobservable absolute contemporaneity and the state of the ether are.
Yes, but only if it is not an objective quantity. If it is just
a quantity like the gradient part of the vector potential, then
everything remains o.k.. But then Lorentz symmetry remains
intact for every observable.
> > In LET, it
> > doesn't; LET could explain the prediction of
> > QM only in a single inertial system.
>
> What does this mean? We are free to use the coordinates we like, in
> every theory. In LET, we have a simplest choice and prefer this
> system. But we can use every other system too.
This is not the point. In LET, the collapse of the wave function
is not instantaneous in, say a far-away galaxy moving with
respect to us at almost the speed of light. It is even asymmetric,
traveling from A to B at negative speeds but from B to A at
a maximum speed of c^2/v. This is not supported by current
Schroedinger quantum theory.
> > By the way, I find the explicit construction of
> > an ether model based on a discrete symmetry group
> > quite fascinating. If a similar thing could be
> > done for GR (I don't know if that is possible),
> > this would definitely be worth publishing.
> Means, to unify this sub-LET construction with my PG? Good idea for
> future research.
>
> > But this means not to *postulate*
> > the properties of the ether (as done in Ilja's
> > approach) but to derive them from a model
> > (of the underlying crystal).
> Means to postulate the properties of the crystal and to derive the
> continuous approximation. Postulating is always involved. And the
> continuous ether theory is interesting enough even without an
> underlying atomic model.
But the continuous ether theory has already been given by
Einstein... See my post on "Einstein's ether theory".
>> So, the current status is that SR and LET are
>> equivalent for all practical purposes but that
>> SR makes more predictions (impossibility of
>> superluminal transport,
>More detailed explanation please. Which type of observation which is
>in agreement with LET falsifies SR? (remark: an observable
>superluminal phone line would be in contradiction with LET, because it
>allows to built an absolute clock, but all clocks are time-dilated by
>interaction with the ether.)
No, a superluminal phone line would not necessarily be
in contradiction with LET. Just consider the model of Guenther.
Supersonic transmission of messages does not falsify or
violate the theory. It just falls outside its realm of
predictions. It would be a means to verify the existence of
the ether.
Of course, if you postulate in your theory the impossibility
of building absolute clocks, you immunize it against falsification
vis a vis SR. In that case, of course, the Aspect experiment
would have to falsify LET, too, if it falsifies SR. (It could
eventually be used to detect the absolute frame, and then you
could build an absolute clock: just take an ordinary clock
in the absolute frame.)
By the way, I do not quite see what you want. On the one hand,
you stated yourself that SR makes one more prediction than
LET which allows it to be falsified and LET to survive. On
the other hand, you stated that any falsification of GR would
have to falsify PG, too. Now, if you can falsify SR without
falsifying LET, you should also be able to falsify GR without
falsifying PG. But then your basic argument why to favour PG
over GR would fall (as long as there is no agreement on the
fact of falsification of SR/GR). Then both PG and GR would have
possibilities to be falsified without falsifying the other
theory. Hence, one has to look for a different criterion.
Now if you ask for *ease of falsification* in addition, obviously
GR wins, because you can falsify it *everywhere*, whereas
in PG you have to fall into a black hole or explore a (global)
topology.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > We have good agreement with SR-QFT. That's all.
> No, we have good agreement even
> with the nonrelativistic Schroedinger theory.
Nice.
> > > b) QM predicts
> > > the instantaneous collapse of the wave function
> > > in *any one* of these systems.
> > We can choose different frames, make QM computations and are in
> > agreement with observation. If some real wave function really
> > instantaneously collapses we cannot conclude from these observations.
> Right. We can even *exclude* it from SR. But the point is the
> "real". The wave function need not be real in any sense, it
> need not be objective. Then of course it can collapse
> instantaneously in any frame of reference.
Ok. But if it is real, it collapses in one frame. We are not able to
observe which frame it is, but this is not problematic.
> > > If we assumed the
> > > wave function to be some objective field and
> > > its instantaneous collapse to signify some kind
> > > of superluminal transport, this would have to
> > > work for all inertial systems.
> > Why?
> Because of experimental evidence.
Here seems to be the key of the misunderstanding: Experimental
evidence does not tell anything about this collapse.
You conclude from experimental evidence (Lorentz-invariance of
observables) that the relativity principle holds (Lorentz-invariance
of anything real). But this conclusion is under discussion here.
In LET this is explicitely wrong.
Of course, the observable conclusion - the violation of Bell's
inequality - is valid in all inertial frames. But the internal, hidden
explanation in LET is of course different for different frames. We
observe a violation for space-like separated events A, B. In a frame
with t_A<t_B, this will be explained by A->B, in the other case by
B->A.
> > LET is not Lorentz-invariant, even if their observable
> > predictions are. The unobservable wave function may be different and
> > incompatible between different Lorentz-frames, as well as the
> > unobservable absolute contemporaneity and the state of the ether are.
> Yes, but only if it is not an objective quantity. If it is just
> a quantity like the gradient part of the vector potential, then
> everything remains o.k.. But then Lorentz symmetry remains
> intact for every observable.
Sorry, let's not confuse notions. In LET we have to distinguish
observable and real objects, in SR we have an equality real =
observable = "objective". Of course, we can discuss also the
underlying philosophical concepts, positivism vs. Popper. But LET is
certainly not a positivistic theory, it accepts that existence and
observability are different things.
Here you have claimed to have found contradictions between QM based on
LET/PG and observation (at least, this was my impression). If this is
true, you should be able to derive the false prediction using LET,
thus, apply LET metaphysics.
> > What does this mean? We are free to use the coordinates we like, in
> > every theory. In LET, we have a simplest choice and prefer this
> > system. But we can use every other system too.
> This is not the point. In LET, the collapse of the wave function
> is not instantaneous in, say a far-away galaxy moving with
> respect to us at almost the speed of light. It is even asymmetric,
> traveling from A to B at negative speeds but from B to A at
> a maximum speed of c^2/v. This is not supported by current
> Schroedinger quantum theory.
?????????? Seems, you have again developed an own nonsense theory to
beat it. The collapse is instantaneous in absolute time. Using PG
cosmology, we can conclude from symmetry argumentation that in the
expanding flat universe this absolute contemporaneity ~ coinsides with
equal proper time after big bang. Point.
Please show which observation is not in agreement with Schrödinger
theory in this absolute time.
> > Means to postulate the properties of the crystal and to derive the
> > continuous approximation. Postulating is always involved. And the
> > continuous ether theory is interesting enough even without an
> > underlying atomic model.
> But the continuous ether theory has already been given by
> Einstein... See my post on "Einstein's ether theory".
Einstein's ether does not have some properties of my ether: density,
velocity, stress tensor, conservation laws, existing in absolute space
and time.
> >More detailed explanation please. Which type of observation which is
> >in agreement with LET falsifies SR? (remark: an observable
> >superluminal phone line would be in contradiction with LET, because it
> >allows to built an absolute clock, but all clocks are time-dilated by
> >interaction with the ether.)
> No, a superluminal phone line would not necessarily be
> in contradiction with LET. Just consider the model of Guenther.
> Supersonic transmission of messages does not falsify or
> violate the theory. It just falls outside its realm of
> predictions. It would be a means to verify the existence of
> the ether.
Observable supersonic transmission and supersonic transmission are
different things, if you have only a very restricted set of
observational devices.
> Of course, if you postulate in your theory the impossibility
> of building absolute clocks, you immunize it against falsification
> vis a vis SR. In that case, of course, the Aspect experiment
> would have to falsify LET, too, if it falsifies SR.
No, because LET can explain it by an unabservable causal influence,
A->B or B->A, depending on the ordering in absolute time. Which may
be described f.e. by Bohmian mechanics. The direction of influence is
unobservable, the statement (A->B or B->A) proven by observed
violation of Bell's inequality. SR does not allow (A->B or B->A).
> (It could
> eventually be used to detect the absolute frame, and then you
> could build an absolute clock: just take an ordinary clock
> in the absolute frame.)
No, we cannot use it to detect the absolute frame, because it does not
contradict LET.
> By the way, I do not quite see what you want. On the one hand,
> you stated yourself that SR makes one more prediction than
> LET which allows it to be falsified and LET to survive. On
> the other hand, you stated that any falsification of GR would
> have to falsify PG, too.
Seems I have already tried to explain this often enough. Of course, if
you use a version of GR which accepts the EPR argument, we have the
same situation as with SR - the theory is falsified by Aspect's
experiment. If you reject EPR, you have immunized it against this
falsification.
In PG you can accept EPR without becoming falsified by Aspect, as LET.
> Now, if you can falsify SR without
> falsifying LET, you should also be able to falsify GR without
> falsifying PG.
Yep.
> But then your basic argument why to favour PG
> over GR would fall (as long as there is no agreement on the
> fact of falsification of SR/GR). Then both PG and GR would have
> possibilities to be falsified without falsifying the other
> theory.
This argument was about the version of GR which does not accept EPR
and follows Bohr. As far as I understand, the current mainstream
position.
Of course, for a version of GR which accepts EPR, my basic argument
for preference no longer holds - we have found a difference in the
predictions between GR and PG. In this case, we have another reason to
prefer PG: that GR is falsified by Aspect.
> Hence, one has to look for a different criterion.
No. The non-immunized theory is dead, for the immunized theory my
argument still holds. Until somebody presents another new experiment
which allows to falsify GR without falsifying PG.
Klaus Kassner
Ilja Schmelzer wrote:
> > > > If we assumed the
> > > > wave function to be some objective field and
> > > > its instantaneous collapse to signify some kind
> > > > of superluminal transport, this would have to
> > > > work for all inertial systems.
> > > Why?
> > Because of experimental evidence.
> Here seems to be the key of the misunderstanding: Experimental
> evidence does not tell anything about this collapse.
No. My argument was a bit more subtle.
Experimental evidence supports QM.
QM tells us that the collapse may be considered
instantaneous. This is an indirect conclusion.
> Of course, the observable conclusion - the violation of Bell's
> inequality - is valid in all inertial frames. But the internal, hidden
> explanation in LET is of course different for different frames. We
> observe a violation for space-like separated events A, B. In a frame
> with t_A<t_B, this will be explained by A->B, in the other case by
> B->A.
Which is perfectly compatible with SR.
Providing we use your "special" notion of causality.
> > Yes, but only if it is not an objective quantity. If it is just
> > a quantity like the gradient part of the vector potential, then
> > everything remains o.k.. But then Lorentz symmetry remains
> > intact for every observable.
> Sorry, let's not confuse notions. In LET we have to distinguish
> observable and real objects, in SR we have an equality real =
> observable = "objective". Of course, we can discuss also the
> underlying philosophical concepts, positivism vs. Popper. But LET is
> certainly not a positivistic theory, it accepts that existence and
> observability are different things.
Right. Maybe we have to replace some misuses of "real" that I made
before by "objective". This means observer-invariant and hence
refers only to observables.
> Here you have claimed to have found contradictions between QM based on
> LET/PG and observation (at least, this was my impression). If this is
> true, you should be able to derive the false prediction using LET,
> thus, apply LET metaphysics.
But I did so already. If the QM prediction of instantaneous
collapse is right (and the wave function real *in the sense
of LET*) then there is a contradiction, because the collapse
cannot be instantaneous in every frame according to LET
(while it would be according to QM).
It can be instantaneous only in one frame.
There is no problem, of course, if you accept the wave function
as nonreal. There is also no problem, if you say, QM does
not really state that the wave function collapses instantaneously;
QM is ambiguous about this.
> > > What does this mean? We are free to use the coordinates we like, in
> > > every theory. In LET, we have a simplest choice and prefer this
> > > system. But we can use every other system too.
> > This is not the point. In LET, the collapse of the wave function
> > is not instantaneous in, say a far-away galaxy moving with
> > respect to us at almost the speed of light. It is even asymmetric,
> > traveling from A to B at negative speeds but from B to A at
> > a maximum speed of c^2/v. This is not supported by current
> > Schroedinger quantum theory.
> ?????????? Seems, you have again developed an own nonsense theory to
> beat it. The collapse is instantaneous in absolute time. Using PG
> cosmology, we can conclude from symmetry argumentation that in the
> expanding flat universe this absolute contemporaneity ~ coinsides with
> equal proper time after big bang. Point.
I knew this was coming :-).
There are *not only* galaxies *at rest* w.r.t. the cosmological
fluid. There are enough of them that have proper motion and thus don't
have equal proper time with the average. Your argument is valid
only for the average but does not hold for the fluctuations.
The universe just is not as homogeneous and isotropic as
cosmological models suggest...
However, Schroedinger theory seems to work in the
proper frame of all galaxies, for which a redshift has been measured.
> Please show which observation is not in agreement with Schrödinger
> theory in this absolute time.
I am not arguing against Schroedinger theory. I am arguing against
LET; well, not even this: only against the pretended different
predictions of LET and SR, regarding Bell.
> > >More detailed explanation please. Which type of observation which is
> > >in agreement with LET falsifies SR? (remark: an observable
> > >superluminal phone line would be in contradiction with LET, because it
> > >allows to built an absolute clock, but all clocks are time-dilated by
> > >interaction with the ether.)
> > No, a superluminal phone line would not necessarily be
> > in contradiction with LET. Just consider the model of Guenther.
> > Supersonic transmission of messages does not falsify or
> > violate the theory. It just falls outside its realm of
> > predictions. It would be a means to verify the existence of
> > the ether.
> Observable supersonic transmission and supersonic transmission are
> different things, if you have only a very restricted set of
> observational devices.
No. Transmission, sending signals, etc. *implies* observability.
It is part of the definition.
> No, because LET can explain it by an unabservable causal influence,
> A->B or B->A, depending on the ordering in absolute time.
Well, I think this is the crux of the whole thing.
SR can explain this as well. Assuming *your* definition of
causality, SR has no problem whatsoever with superluminal
"causal" connections... Because it only forbids *observable*
superluminal causal connections. Since you have partially moved
causality to the realm of the unobservable with your definition,
SR does not have a problem with it. You see, SR does not even
have a problem with a "real" but "nonobjective" wave function.
> Which may
> be described f.e. by Bohmian mechanics. The direction of influence is
> unobservable, the statement (A->B or B->A) proven by observed
> violation of Bell's inequality. SR does not allow (A->B or B->A).
Of course it does. You can choose any observer as the one
whose time is "absolute". Calculate the time in this observer's
frame to get the causal_(in Ilja's sense) ordering of two events.
Then you have exactly this structure. It does not have anything
to do with *real* causality, but this is rather a problem of
the definition than of SR.
> No, we cannot use it to detect the absolute frame, because it does not
> contradict LET.
You don't need a contradiction with LET to detect
the absolute frame. Unless you put in an *extra* axiom: the
absolute frame is unobservable. But the whole charm of ether
theories with such a frame is that it might be somehow detectable.
And in a model analogous to the one of Guenther it would in
principle be detectable. For example via tachyons. They
would move, depending on direction, at restricted velocities
in moving frames but could be arbitrarily fast in the absolute
frame.
On the other hand, you would have a contradiction with *SR*,
if you found *these* tachyons.
> > But then your basic argument why to favour PG
> > over GR would fall (as long as there is no agreement on the
> > fact of falsification of SR/GR). Then both PG and GR would have
> > possibilities to be falsified without falsifying the other
> > theory.
> This argument was about the version of GR which does not accept EPR
> and follows Bohr. As far as I understand, the current mainstream
> position.
Ah, so there are now different versions of GR?
That's new to me. And one of them you would still
accept as not falsified by your argument?
> Of course, for a version of GR which accepts EPR, my basic argument
> for preference no longer holds - we have found a difference in the
> predictions between GR and PG. In this case, we have another reason to
> prefer PG: that GR is falsified by Aspect.
No. Since GR has no problems with your statement about
Bell's inequality, given your definition of causality.
It just does not consider the latter meaningful in any sense.
> > Hence, one has to look for a different criterion.
> No. The non-immunized theory is dead, for the immunized theory my
> argument still holds. Until somebody presents another new experiment
> which allows to falsify GR without falsifying PG.
This is simple. Find a field that does not couple the same way
to the metric as the others (one example from your point of
view would be the wave function; if it can collapse instantaneously,
it couples differently, meaning that either PG is falsified or
you have to accept the possibility of different couplings, the
view you seem to take; but then there is no reason to restrict
this to quantum mechanical fields, *any* new field, unknown
hitherto, will do).
GR would be dead. PG, or a modified version thereof would
continue to live happily.
Matthew P Wiener
In article <33D764...@physik.uni-magdeburg.de>, Klaus Kassner <klaus.kassner@physik writes:
>Matthew P Wiener wrote:
>> In article <33D748...@physik.uni-magdeburg.de>, Klaus Kassner <klaus.kassner@physik writes:
>> >Matthew P Wiener wrote:
>> >> Klaus Kassner wrote:
>> >>> Matthew P Wiener wrote:
>> >> >> >> No, my statement is correct. Perhaps you'd like to bring
>> >> >> >> phlogiston theory up to date, and tell us that the result
>> >> >> >> _is_ phlogiston theory?
>> >The week is over, so here is the answer.
>> You *still* did not answer my question.
>I think I did.
Only in the sense of continually ducking it, and then brazenly claiming
to answer it after a week's delay.
>> >First, I don't feel inclined to bring *any* discarded theory up to
>> >date, neither phlogiston theory, nor LET. I am sufficiently satisfied
>> >with the current status of the theories of heat transfer and of SR.
>> I didn't *ask* you to bring any discarded theory up to date. Sheesh.
>It seemed to me you did.
I did not. Not even close. I was only very explicit.
>> >Second, *if* I saw an updated version of phlogiston theory, I might
>> >be curious enough to consider it.
>> But do we speak of such a hypothesized revised theory as "phlogiston
>> theory"? The answer is no. That is all.
>No. The answer could be yes. Depending on how many changes
>were introduced.
The amount of changes needed to bring LET up to date is rather large.
>> >Third, I do not *have* to bring LET up to date, because this *has
>> >been done* by others.
>> And is this Lorentz's ether theory?
>Yes, unless the changes made are sufficient to warrant a new name.
Which is what it is.
>> >A recent reference is: "Grenzgeschwindigkeiten und ihre
>> >Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
>> >Stuttgart, 1996. The author is a tenured German professor, not a
>> >crank.
>> No, it is Guenther's revised ether theory.
>Most definitely not. He is not the only one looking at
>this type of theory. The class name is "Lorentz ether theory".
Whoever's. It is not Lorentz's.
>> I'm willing to look in the U Penn library, and it's not there.
>Sure. It is from 1996, and in German. What do you expect?
Gosh. So why did you bother with the snide remarks about me being
too lazy to go look it up?
>My claim is there *are* modern versions of LET. Yours was
>there aren't.
No, that was not my claim. My claim is that modernized LET is not LET
itself.
> Usually one counterexample suffices to bring
>down a general claim.
It helps to notice what is being said.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > Of course, the observable conclusion - the violation of Bell's
> > inequality - is valid in all inertial frames. But the internal, hidden
> > explanation in LET is of course different for different frames. We
> > observe a violation for space-like separated events A, B. In a frame
> > with t_A<t_B, this will be explained by A->B, in the other case by
> > B->A.
>
> Which is perfectly compatible with SR.
> Providing we use your "special" notion of causality.
No. (A->B or B->A) for space-like separated events as compatible with
SR? No, Einstein causality clearly states "not (A->B or B->A)". Point.
You can use a weaker notion of causality so that Aspect does not prove
(A->B or B->A). For example, accepting Bohr's rejection of EPR. But
(A->B or B->A) itself is not compatible with SR.
> > > Yes, but only if it is not an objective quantity. If it is just
> > > a quantity like the gradient part of the vector potential, then
> > > everything remains o.k.. But then Lorentz symmetry remains
> > > intact for every observable.
> > Sorry, let's not confuse notions. In LET we have to distinguish
> > observable and real objects, in SR we have an equality real =
> > observable = "objective". Of course, we can discuss also the
> > underlying philosophical concepts, positivism vs. Popper. But LET is
> > certainly not a positivistic theory, it accepts that existence and
> > observability are different things.
> Right. Maybe we have to replace some misuses of "real" that I made
> before by "objective". This means observer-invariant and hence
> refers only to observables.
Let's see.
> > Here you have claimed to have found contradictions between QM based on
> > LET/PG and observation (at least, this was my impression). If this is
> > true, you should be able to derive the false prediction using LET,
> > thus, apply LET metaphysics.
> But I did so already. If the QM prediction of instantaneous
> collapse is right (and the wave function real *in the sense
> of LET*) then there is a contradiction, because the collapse
> cannot be instantaneous in every frame according to LET
> (while it would be according to QM).
> It can be instantaneous only in one frame.
Yes. It is instantaneous in the correct frame. But what is the
contradiction with observation?
> There is no problem, of course, if you accept the wave function
> as nonreal. There is also no problem, if you say, QM does
> not really state that the wave function collapses instantaneously;
> QM is ambiguous about this.
No, the wave function may be real. We can even consider Bohmian
mechanics, to make the things more certain.
Thus, if we use a wrong frame initially, we obtain a wrong result
about what really happens. But this does not lead to observable
differences as far.
> > ?????????? Seems, you have again developed an own nonsense theory to
> > beat it. The collapse is instantaneous in absolute time. Using PG
> > cosmology, we can conclude from symmetry argumentation that in the
> > expanding flat universe this absolute contemporaneity ~ coinsides with
> > equal proper time after big bang. Point.
> I knew this was coming :-).
Fine.
> There are *not only* galaxies *at rest* w.r.t. the cosmological
> fluid. There are enough of them that have proper motion and thus don't
> have equal proper time with the average.
Fine. Means, the true frame of LET is not the rest frame of this galaxy.
Not it becomes interesting.
As far as I understand, you have said that the observation agrees with
the Schrödinger equation if we use the wrong frame - the rest frame of
this galaxy.
> However, Schroedinger theory seems to work in the
> proper frame of all galaxies, for which a redshift has been measured.
Fine. I agree.
The internal description of what really happens is wrong, if we do
this. But that does not mean that we can observe this. Thus, it is
nonetheless in agreement with observation. It seems to work. But not
more.
> I am not arguing against Schroedinger theory. I am arguing against
> LET; well, not even this:
So what: is LET-QFT (what means SR-QFT with an assumed preferred
frame) in agreement with observation or not? If it is ok, we can
forget one old claim against PG. If not, you clearly argue against
LET.
> only against the pretended different
> predictions of LET and SR, regarding Bell.
This is a different question, and I want to discuss this question too.
> > Observable supersonic transmission and supersonic transmission are
> > different things, if you have only a very restricted set of
> > observational devices.
> No. Transmission, sending signals, etc. *implies* observability.
> It is part of the definition.
Do you want to talk about names or about arguments? Let's use a
different name for hidden but real causal influences connected with
hidden but real variables in LET.
> > No, because LET can explain it by an unobservable causal influence,
> > A->B or B->A, depending on the ordering in absolute time.
> Well, I think this is the crux of the whole thing.
Yep.
> SR can explain this as well. Assuming *your* definition of
> causality, SR has no problem whatsoever with superluminal
> "causal" connections... Because it only forbids *observable*
> superluminal causal connections.
That's a clear modification of the theory. All laws of nature have to
be Lorentz-invariant or not all?
Is it a law of nature that
1. there exists a partial order A->B in spacetime (no closed causal
loops);
2. Causal influence from A to B is possible only if A->B;
3. The relation A->B is Lorentz-invariant?
> Since you have partially moved
> causality to the realm of the unobservable with your definition,
> SR does not have a problem with it.
I do not see where I have moved causality to the realm of the
unobservable. Reversely, I prefer a notion of causality which allows
to observe causal relations by observation. Especially, I consider
the observation of a violation of Bell's inequality as the observation
of the fact that (A->B or B->A). Following Bohr, we do not have to
accept this as an observation of a causal relation. Using the same
argumentation, we can reject any observation of causal relations.
Who moves causality to the realm of the unobservable?
> > Which may
> > be described f.e. by Bohmian mechanics. The direction of influence is
> > unobservable, the statement (A->B or B->A) proven by observed
> > violation of Bell's inequality. SR does not allow (A->B or B->A).
> Of course it does. You can choose any observer as the one
> whose time is "absolute". Calculate the time in this observer's
> frame to get the causal_(in Ilja's sense) ordering of two events.
> Then you have exactly this structure. It does not have anything
> to do with *real* causality, but this is rather a problem of
> the definition than of SR.
So what are the laws of *real* causality?
> > No, we cannot use it to detect the absolute frame, because it does not
> > contradict LET.
> You don't need a contradiction with LET to detect
> the absolute frame. Unless you put in an *extra* axiom: the
> absolute frame is unobservable. But the whole charm of ether
> theories with such a frame is that it might be somehow detectable.
For example, by an underlying atomic ether theory which breaks the
symmetry of the continuous approximation.
> And in a model analogous to the one of Guenther it would in
> principle be detectable. For example via tachyons.
You do not need tachyons. You simply have a different symmetry
group. If you are accurate enough to observe the differences between
the atomic theory and the continuous approximation, you have it.
But atomic ether theory is a different theory. It falsifies continuous
LET for small distances. We are talking about continuous theories yet
- LET and PG.
> On the other hand, you would have a contradiction with *SR*,
> if you found *these* tachyons.
As well as with continuous LET and PG.
> > > But then your basic argument why to favour PG
> > > over GR would fall (as long as there is no agreement on the
> > > fact of falsification of SR/GR). Then both PG and GR would have
> > > possibilities to be falsified without falsifying the other
> > > theory.
> > This argument was about the version of GR which does not accept EPR
> > and follows Bohr. As far as I understand, the current mainstream
> > position.
> Ah, so there are now different versions of GR?
> That's new to me. And one of them you would still
> accept as not falsified by your argument?
Yes, the variant which does not allow to use EPR-like argumentation
and accepts Bohr's rejection.
I criticize this theory as an ad hoc immunization of relativity, but I
have to accept that it is not falisfied by observation.
Note that in PG I explicitly accept EPR-like argumentation. Note also
that EPR-like argumentation is widely used by common sense. A theory
which rejects it has much less predictive power.
> > No. The non-immunized theory is dead, for the immunized theory my
> > argument still holds. Until somebody presents another new experiment
> > which allows to falsify GR without falsifying PG.
> This is simple. Find a field that does not couple the same way
> to the metric as the others (one example from your point of
> view would be the wave function; if it can collapse instantaneously,
> it couples differently, meaning that either PG is falsified or
> you have to accept the possibility of different couplings, the
> view you seem to take; but then there is no reason to restrict
> this to quantum mechanical fields, *any* new field, unknown
> hitherto, will do).
(Looking into my papers about PG ;-) There is no such field in my PG.
Sorry, not found.
> GR would be dead. PG, or a modified version thereof would
> continue to live happily.
A modified version, but not PG. Let's not mingle claims about certain
well-defined theories (GR and PG as classical theories of gravity) and
guesses about possible meta-theoretical developments.
Klaus Kassner
Ilja Schmelzer wrote:
>
> Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > Of course, the observable conclusion - the violation of Bell's
> > > inequality - is valid in all inertial frames. But the internal, hidden
> > > explanation in LET is of course different for different frames. We
> > > observe a violation for space-like separated events A, B. In a frame
> > > with t_A<t_B, this will be explained by A->B, in the other case by
> > > B->A.
> >
> > Which is perfectly compatible with SR.
> > Providing we use your "special" notion of causality.
>
> No. (A->B or B->A) for space-like separated events as compatible with
> SR? No, Einstein causality clearly states "not (A->B or B->A)". Point.
causality, for which the statement is entirely compatible with SR.
Period.
I guess we should stop arguing. I think we agree that we don't
agree.
Matthew P Wiener
In article <i3gd8o8...@fermi.wias-berlin.de>, Ilja Schmelzer <schmelze@fermi writes:
>wee...@sagi.wistar.upenn.edu (Matthew P Wiener) writes:
>> >been done* by others.
>> And is this Lorentz's ether theory?
>> >A recent reference is: "Grenzgeschwindigkeiten und ihre
>> >Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
>> >Stuttgart, 1996. The author is a tenured German professor, not a
>> >crank.
>> No, it is Guenther's revised ether theory.
>Formally correct. But why increase confusion among other readers?
It decreases confusion.
>If we talk about Guenther's revised ether theory, nobody has an idea what
>we are talking about.
Too bad. Give a reference.
>This is the straightforward generalization of LET to the current
>state, with strong and weak forces and QFT. To reject to name it LET
>is artificial because of the triviality of this modification.
It's only trivial in retrospect. Partially because of the powerful
constraints imposed by SR, the correct form of the strong and weak
forces, and QFT itself, were discovered over the past several decades.
Recasting these backwards allows an ether theory to be read off, but
as alternative historical reconstruction, it is terribly implausible.
>The ether dilates all clocks in the same way.
It's the brain of physicists that I'm talking about.
>For the case of gravity, there it is not so obvious, a new name for
>the generalization is really justified - that's why I have introduced
>the name post-relativistic gravity (PG).
>Guenther's revised ether theory is an crystal model which coinsides
>with this LET in the large scale approximation, thus, really something
>different.
Indeed.
>But for SR with preferred frame the name LET is appropriate, as well
>as we use the name SR today too.
But that is not Guenther's theory. It fixes a particular model for
the ether, one that Lorentz did not have.
Ilja Schmelzer
wee...@sagi.wistar.upenn.edu (Matthew P Wiener) writes:
> >If we talk about Guenther's revised ether theory, nobody has an idea what
> >we are talking about.
> Too bad. Give a reference.
Here it is:
> >> >A recent reference is: "Grenzgeschwindigkeiten und ihre
> >> >Paradoxa. Gitter, Aether, Relativitaet." by H. Guenther, Teubner,
> >> >Stuttgart, 1996. The author is a tenured German professor, not a
> >> >crank.
> >This is the straightforward generalization of LET to the current
> >state, with strong and weak forces and QFT. To reject to name it LET
> >is artificial because of the triviality of this modification.
>
> It's only trivial in retrospect. Partially because of the powerful
> constraints imposed by SR, the correct form of the strong and weak
> forces, and QFT itself, were discovered over the past several decades.
> Recasting these backwards allows an ether theory to be read off, but
> as alternative historical reconstruction, it is terribly implausible.
Nonsense. The idea to make all physics compatible with Lorentz
transformation, as far as possible, is already present in Poincare's
1905 paper. That the time dilation caused by the ether is the same for
all clocks, even for clocks based on strong and weak forces, is the
simplest assumption about the behaviour of these clocks.
The question if there is a preferred frame or not is orthogonal to the
theory of weak and strong forces.
> >The ether dilates all clocks in the same way.
> It's the brain of physicists that I'm talking about.
???
> >For the case of gravity, there it is not so obvious, a new name for
> >the generalization is really justified - that's why I have introduced
> >the name post-relativistic gravity (PG).
>
> >Guenther's revised ether theory is an crystal model which coinsides
> >with this LET in the large scale approximation, thus, really something
> >different.
>
> Indeed.
>
> >But for SR with preferred frame the name LET is appropriate, as well
> >as we use the name SR today too.
>
> But that is not Guenther's theory. It fixes a particular model for
> the ether, one that Lorentz did not have.
Agreement. But, I hope, we nonetheless can talk about LET today - as
SR with preferred frame and ether interpretation.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > > Of course, the observable conclusion - the violation of Bell's
> > > > inequality - is valid in all inertial frames. But the internal, hidden
> > > > explanation in LET is of course different for different frames. We
> > > > observe a violation for space-like separated events A, B. In a frame
> > > > with t_A<t_B, this will be explained by A->B, in the other case by
> > > > B->A.
> > >
> > > Which is perfectly compatible with SR.
> > > Providing we use your "special" notion of causality.
> >
> > No. (A->B or B->A) for space-like separated events as compatible with
> > SR? No, Einstein causality clearly states "not (A->B or B->A)". Point.
>
> causality, for which the statement is entirely compatible with SR.
> Period.
Describe your understanding of Einstein causality, especially where it
is different from mine. Have in mind, what I really need is
acceptance of the EPR argumentation. Remembering what the E in EPR
means, it seems hard to claim that this acceptance is incompatible
with "Einstein causality".
> I guess we should stop arguing. I think we agree that we don't
> agree.
I don't think that's a reason to stop arguing. If we agree, it would
be a reason to stop. If we do not agree, it becomes interesting.
But, of course, if you stop arguing, I cannot change this.
Klaus Kassner
Ilja Schmelzer wrote:
>
> Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > > > Of course, the observable conclusion - the violation of Bell's
> > > > > inequality - is valid in all inertial frames. But the internal, hidden
> > > > > explanation in LET is of course different for different frames. We
> > > > > observe a violation for space-like separated events A, B. In a frame
> > > > > with t_A<t_B, this will be explained by A->B, in the other case by
> > > > > B->A.
> > > >
> > > > Which is perfectly compatible with SR.
> > > > Providing we use your "special" notion of causality.
> > >
> > > No. (A->B or B->A) for space-like separated events as compatible with
> > > SR? No, Einstein causality clearly states "not (A->B or B->A)". Point.
> >
> > causality, for which the statement is entirely compatible with SR.
> > Period.
> Describe your understanding of Einstein causality, especially where it
> is different from mine. Have in mind, what I really need is
> acceptance of the EPR argumentation. Remembering what the E in EPR
> means, it seems hard to claim that this acceptance is incompatible
> with "Einstein causality".
I won't try to give a definition of causality (just as Einstein
avoided giving a definition of reality in the EPR paper :-)).
But I can say what I think is an important *element* that should
be present in *any* definition of causality. This is: if you
remove the cause, the effect should also go away (allowing you
to conclude from the absence of the effect to the absence of the
cause).
Now you have defined causality to be just an ordering in absolute
time. That is, if A happens before B (in absolute time), then A->B.
This does not satisfy my criterion. You can remove any number of
events happening before B, without affecting its coming to pass,
if only you avoid removing those that *really* caused B.
Now why is your statement compatible with SR?
The formal statement was:
not(A->B) and not(B->A) implies Bell's inequality.
Hence you can conclude from a violation of Bell's inequality
that either A->B or B->A. (I am not discussing the *correctness*
of the original statement here.)
But what does your statement read when expressed in ordinary
language without any reference to the word "causal"?
It reads: If neither A happens before B (in absolute time)
nor B happens before A (in absolute time) then Bell's inequality
must hold. Hence, we can
conclude from a violation of Bell's inequality that either
A must happen before B or B before A in the absolute frame.
(Which is however unobservable).
Let me translate this to SR. Of course, there is no absolute
frame in SR. But we have a democracy of frames. You can just
choose one as the absolute frame.
Then the statement reads: If in some frame A does not
happen before B and B does not happen before A, then
Bell's inequality holds. (I don't see why this should not
be true in SR, since the condition implies that A and B are
spacelike.) Hence, we can conclude from a violation of
Bell's inequality that there is a frame, in which A happens
before B or B happens before A. (This seems also perfectly
compatible with SR to me albeit an almost empty statement.)
Of course there is a slight difference between the LET and
SR interpretations: in LET, the frame, in which the simultaneity
of A and B implies Bell's inequality *must*
be the same for all realizations of the experiment.
(Well, depending on some additional conditions.)
In SR, it can be a different frame for different realizations.
But this difference does not play any role, because it is not
possible to detect the frame, in which the simultaneity holds
-- otherwise you could build an absolute clock...
The trap, in which you seem to have fallen, is that you have
mistaken your definition of causality to *be* the true causality.
However, I am not sure that the proof of your relation will
go through with a good definition of causality.
In any case, it would be
important to state at some point what you think are the events
A and B in a concrete experiment. I don't see well what you
mean by these. For example, it is not possible that A and B are the
measurements of one of the spins. For you can choose the temporal
sequence of these arbitrarily, *including* their happening at
the same time. QM predicts a violation of Bell's inequality
in all these cases.
> > I guess we should stop arguing. I think we agree that we don't
> > agree.
>
> I don't think that's a reason to stop arguing. If we agree, it would
> be a reason to stop. If we do not agree, it becomes interesting.
I'll answer that one later.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > Describe your understanding of Einstein causality, especially where it
> > is different from mine.
> But I can say what I think is an important *element* that should
> be present in *any* definition of causality. This is: if you
> remove the cause, the effect should also go away (allowing you
> to conclude from the absence of the effect to the absence of the
> cause).
> Now you have defined causality to be just an ordering in absolute
> time. That is, if A happens before B (in absolute time), then A->B.
> This does not satisfy my criterion.
Accepted. A->B denotes now actual causal influence of something that
happens at A to something that happens at B.
> Now why is your statement compatible with SR?
> The formal statement was:
> not(A->B) and not(B->A) implies Bell's inequality.
> Hence you can conclude from a violation of Bell's inequality
> that either A->B or B->A. (I am not discussing the *correctness*
> of the original statement here.)
This statement remains valid for actual causality too.
Let me explain where I think "potential" causality is useful. The
actual causality is not a Lorentz-invariant relation, not even
transitive. But we can formulate the following law as a law of
causality:
1. There exists some relation A~>B (potential causality) which is
a partial order and should be Lorentz-invariant.
2. A->B (actual causal influence) is possible only if A~>B.
Only now we can derive from the relativity principle (all laws of
nature are Lorentz-invariant means A~>B should be Lorentz-invariant)
that A->B falsifies Einstein-causality.
> But what does your statement read when expressed in ordinary
> language without any reference to the word "causal"?
> It reads: If neither A happens before B (in absolute time)
> nor B happens before A (in absolute time) then Bell's inequality
> must hold. Hence, we can
> conclude from a violation of Bell's inequality that either
> A must happen before B or B before A in the absolute frame.
> (Which is however unobservable).
That's the explanation in LET. Or, more accurate, it is not an
explanation, but shows that LET is not in contradiction with this
claim.
> Let me translate this to SR. Of course, there is no absolute
> frame in SR. But we have a democracy of frames. You can just
> choose one as the absolute frame.
> Then the statement reads: If in some frame A does not
> happen before B and B does not happen before A, then
> Bell's inequality holds. (I don't see why this should not
> be true in SR, since the condition implies that A and B are
> spacelike.) Hence, we can conclude from a violation of
> Bell's inequality that there is a frame, in which A happens
> before B or B happens before A. (This seems also perfectly
> compatible with SR to me albeit an almost empty statement.)
A very funny type of argumentation against an experimental
falsification of a theory. You show that you can derive from a given
experimental fact also some other statements, which are or may be in
agreement with the theory. Of course, I can derive a lot of things,
some of them may be in agreement with SR too ;-) But this does not
show a failure in the original argumentation.
And the original argumentation is very simple.
If A->B relativity is falsified.
If B->A relativity is falsified.
--------------------------------
If (A->B or B->A) relativity is falsified.
Or, may be, you assume we live in a superpositional state of two
universes, so that in above universes relativity is wrong but in our
superpositional state not? ;-)
> The trap, in which you seem to have fallen, is that you have
> mistaken your definition of causality to *be* the true causality.
> However, I am not sure that the proof of your relation will
> go through with a good definition of causality.
Let's look on the whole argumentation again, simply to clarify which
parts you accept and which not. Please, mark in the following the
enumerated statements which you accept and which you do not accept. I
hope, this modular subdivion helps to find out the critical points in
my argumentation.
1. There are no "good" and "bad" definitions of causality. IMO, there
are different physical theories of causality. We can subdivide them
into two classes:
a. The first class accepts the EPR argument - de facto takes it as an axiom.
b. The second class rejects the EPR argument, f.e. following Bohr.
I suggest to name the first class EPR-causality, the second
Bohr-causality.
2. We can formalize EPR causality in a way that Bell's theorem in my
form (A->B or B->A) => Bell's inequality becomes a strong theorem.
(If you do not accept this, replace "EPR-causality" below with "a
theory of causality where (A->B or B->A) => Bell's inequality holds".)
3. In Bohr causality, the proof is false.
4. That means, the combination (SR + EPR-causality) is falsified by Aspect.
5. The combination (SR + Bohr causality) is not falsified.
6. EPR-causality has more predictive power than Bohr causality. Proof:
comparison of 4. and 5. Note that (A->B or B->A) => Bell's inequality
may be applied in a lot of other different situations to detect causal
influences.
7. The combination (LET + EPR-causality) is not falsified by Aspect.
8. The combination (LET + Bohr-causality) is also not falsified by
Aspect.
9. The combination (LET + EPR) has less predictive power compared with
(SR + EPR). The additional prediction is Bell's inequality.
Especially, before Aspect's experiment the preference for SR vs. LET
was justified by Popper's criterion of empirical content.
10. The combination (LET + EPR) has more predictive power compared
with (SR + Bohr). This is because (5) Bell's inequality is no longer
predicted by (SR+Bohr), but (6) EPR has in general more predictive
power than Bohr.
11. That's why I criticize the replacement of (SR + EPR) with (SR +
Bohr) as an ad hoc immunization of SR.
12. In all of these statements we can replace SR with GR, LET with PG,
making the small error to forget about the minor differences between
PG and GR in other domains (topology, black holes).
Klaus Kassner
> > be present in *any* definition of causality. This is: if you
> > remove the cause, the effect should also go away (allowing you
> > to conclude from the absence of the effect to the absence of the
> > cause).
> > Now you have defined causality to be just an ordering in absolute
> > time. That is, if A happens before B (in absolute time), then A->B.
> > This does not satisfy my criterion.
> Accepted. A->B denotes now actual causal influence of something that
> happens at A to something that happens at B.
> > Now why is your statement compatible with SR?
> > The formal statement was:
> > not(A->B) and not(B->A) implies Bell's inequality.
> > Hence you can conclude from a violation of Bell's inequality
> > that either A->B or B->A. (I am not discussing the *correctness*
> > of the original statement here.)
> This statement remains valid for actual causality too.
I think it becomes wrong. More precisely, there are some
hidden assumptions which are not fulfilled in the EPR arrangement,
leading to a failure of your argumentation.
> Let me explain where I think "potential" causality is useful. The
> actual causality is not a Lorentz-invariant relation, not even
> transitive.
Actual causality *is* Lorentz invariant. I don't know of any
counter-example. Why transitivity should be important, I do not
know.
> But we can formulate the following law as a law of
> causality:
> 1. There exists some relation A~>B (potential causality) which is
> a partial order and should be Lorentz-invariant.
> 2. A->B (actual causal influence) is possible only if A~>B.
> Only now we can derive from the relativity principle (all laws of
> nature are Lorentz-invariant means A~>B should be Lorentz-invariant)
> that A->B falsifies Einstein-causality.
All laws of nature referring to observables. Laws of nature
containing nonobservables need not be Lorentz-invariant in SR.
(Of course, we could discuss whether there is such a thing.
However, since in practice quantities containing unobservable
parts are introduced for convenience, we do have formulations
of "laws" that are not Lorentz-invariant.)
Now, unless you are specifying what A means and what B,
I can't see what kind of events you are talking of.
I simply don't see what events you are referring to in
the EPR experiment. Give me an example of observable
events A and B, for which you think that Bell's correlations
imply a causal relationship.
Of course, it does. Very strange that you accept my presentation
as correct for LET and incorrect for SR. The problem is that
your definition of causality need not be Lorentz invariant even
in SR.
Klaus Kassner
Ilja Schmelzer wrote:
> 1. There are no "good" and "bad" definitions of causality. IMO, there
> are different physical theories of causality. We can subdivide them
> into two classes:
Sure there are bad definitions.
> a. The first class accepts the EPR argument - de facto takes it as an axiom.
> b. The second class rejects the EPR argument, f.e. following Bohr.
I do not accept this antinomy. It was not the *argument* of
EPR that was wrong. There was an ambiguity in their statement
of the *problem*. There is no difference between Einstein's and
Bohr's causality.
> 2. We can formalize EPR causality in a way that Bell's theorem in my
> form (A->B or B->A) => Bell's inequality becomes a strong theorem.
If this is true, it is a strong theorem in Bohr's causality, too.
But this is not what is happening.
> (If you do not accept this, replace "EPR-causality" below with "a
> theory of causality where (A->B or B->A) => Bell's inequality holds".)
This is not useful, *if* any such theory leads to non-Lorentz invariant
notions of causality (such as yours).
But maybe, it does not...
> 3. In Bohr causality, the proof is false.
Not necessarily. That depends on your hidden assumptions.
I think the Bohr explanation is that there is no A or no B.
> 4. That means, the combination (SR + EPR-causality) is falsified by Aspect.
No. There are three possibilities (as you probably know):
1. Local realism is wrong.
2. SR is wrong.
3. QM is incomplete.
One of these three must be true. Now your choice would be 2.
The problem is that even if you discard 2, replacing SR
by LET would *not* save 1. It is not *realism* that is the
problem. You can have realism. See Bohm's QM. The problem
is *local* realism. Bohm's QM leads to a violation of
Bell's inequalities even if "not (A->B) and not (B->A)",
because it requires the two spins to become determined
at strictly the same moment. So you get nonlocality, even if you
replace SR by LET. You cannot avoid infinite speed, which implies
nonlocality.
> 5. The combination (SR + Bohr causality) is not falsified.
Well, it might be. If (SR + EPR-causality) is falsified, then
also (SR + Bohr causality).
> 6. EPR-causality has more predictive power than Bohr causality. Proof:
> comparison of 4. and 5. Note that (A->B or B->A) => Bell's inequality
> may be applied in a lot of other different situations to detect causal
> influences.
No proof.
> 7. The combination (LET + EPR-causality) is not falsified by Aspect.
> 8. The combination (LET + Bohr-causality) is also not falsified by
> Aspect.
> 9. The combination (LET + EPR) has less predictive power compared with
> (SR + EPR). The additional prediction is Bell's inequality.
> Especially, before Aspect's experiment the preference for SR vs. LET
> was justified by Popper's criterion of empirical content.
No. Scientist never cared much about philosophers telling them why to
prefer certain theories. LET never was a serious competitor
for SR, even though it is strictly equivalent (once you disallow
the possibility to detect the ether).
> 10. The combination (LET + EPR) has more predictive power compared
> with (SR + Bohr). This is because (5) Bell's inequality is no longer
> predicted by (SR+Bohr), but (6) EPR has in general more predictive
> power than Bohr.
> 11. That's why I criticize the replacement of (SR + EPR) with (SR +
> Bohr) as an ad hoc immunization of SR.
Which is a meaningless statement to me, since I don't think that
Bohr and Einstein disagreed about causality. Einstein causality
and Bohr causality are the same. Schmelzer causality is different.
It is potential causality in LET and not even potential causality
in SR, since it includes a lot of spacelike situations, for which
there is *no* potential causality in SR.
Klaus Kassner
Klaus Kassner wrote:
> Schmelzer causality is different.
> It is potential causality in LET and not even potential causality
> in SR, since it includes a lot of spacelike situations, for which
> there is *no* potential causality in SR.
Correction: If it is potential causality in LET
then it can also be potential causality in SR.
This is possible only, if you assume causality not to depend
on energy exchange or information exchange.
(I don't think there is a theorem in SR that requires
Lorentz invariance for "causal" relationships of
such a type.)
If you assume the contrary, i.e., causal relationships are
always connected with energy or information transfer,
then Schmelzer causality is neither potential causality in LET nor in
SR. It would then just be a temporal half-ordering in some (arbitrary)
frame for both LET and SR (with no relation to causality whatsoever).
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > not(A->B) and not(B->A) implies Bell's inequality.
> > > Hence you can conclude from a violation of Bell's inequality
> > > that either A->B or B->A. (I am not discussing the *correctness*
> > > of the original statement here.)
> > This statement remains valid for actual causality too.
> hidden assumptions which are not fulfilled in the EPR arrangement,
> leading to a failure of your argumentation.
> > actual causality is not a Lorentz-invariant relation, not even
> > transitive.
> counter-example. Why transitivity should be important, I do not
> know.
I consider this as a minor misunderstanding not worth discussion.
But the Lorentz transformation of the pair of events A,B will be A',B'.
Now, there may be an actual causal influence A->B but no actual influence
A'->B'. Even if above are time-like, thus, A~>B <=> A'~>B'.
That's what I have had in mind.
> All laws of nature referring to observables. Laws of nature
> containing nonobservables need not be Lorentz-invariant in SR.
> (Of course, we could discuss whether there is such a thing.
> However, since in practice quantities containing unobservable
> parts are introduced for convenience, we do have formulations
> of "laws" that are not Lorentz-invariant.)
Again, I do not question that you can immunize SR using a sufficiently
weak formulation. For example, a formulation which allows to talk
about A->B only if we have a direct observation of an information
transfer.
If you do not like to explain the EPR correlation, I cannot help you.
> Now, unless you are specifying what A means and what B,
> I can't see what kind of events you are talking of.
> I simply don't see what events you are referring to in
> the EPR experiment. Give me an example of observable
> events A and B, for which you think that Bell's correlations
> imply a causal relationship.
??? The events observed in ideal Bell-like experiments, of course.
> > A very funny type of argumentation against an experimental
> > falsification of a theory. You show that you can derive from a given
> > experimental fact also some other statements, which are or may be in
> > agreement with the theory. Of course, I can derive a lot of things,
> > some of them may be in agreement with SR too ;-) But this does not
> > show a failure in the original argumentation.
> Of course, it does. Very strange that you accept my presentation
> as correct for LET and incorrect for SR. The problem is that
> your definition of causality need not be Lorentz invariant even
> in SR.
I agree that a definition of causality is possible which is in
agreement with SR and Aspect. But there is also another, stronger,
which is not. Of course, a proponent of relativity prefers the weaker
one. And I do not want to show you that the stronger is the better
one.
But this stronger notion of causality exists. And it is stronger
because Aspect falsifies SR + this stronger causality, but not SR +
your preferred causality.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > Schmelzer causality is different.
> > It is potential causality in LET and not even potential causality
> > in SR, since it includes a lot of spacelike situations, for which
> > there is *no* potential causality in SR.
>
> Correction: If it is potential causality in LET
> then it can also be potential causality in SR.
> This is possible only, if you assume causality not to depend
> on energy exchange or information exchange.
I assume no energy exchange, but hidden (but real) information
exchange.
> (I don't think there is a theorem in SR that requires
> Lorentz invariance for "causal" relationships of
> such a type.)
Depends on the axioms you use for causality. The question does not
have an anwer outside a concrete theory of causality.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > 1. There are no "good" and "bad" definitions of causality. IMO, there
> > are different physical theories of causality. We can subdivide them
> > into two classes:
;-)
1. IMO, there
are different physical theories of causality. We can subdivide them
into two classes:
>> a. The first class accepts the EPR argument - de facto takes it as an axiom
>> b. The second class rejects the EPR argument, f.e. following Bohr.
> I do not accept this antinomy. It was not the *argument* of
> EPR that was wrong. There was an ambiguity in their statement
> of the *problem*. There is no difference between Einstein's and
> Bohr's causality.
It's hard to argue with such claims. Especially, they have had other
problems, because they have not known Bell's paper and Aspect's
experimental results. Ok, let's specify:
a. There exists a strong axiomatic theory of causality so that the following
holds:
From the observable 100% correlation in Bell's device if we measure
the same direction at space-like separated events A and B, and
Einstein causality (that means, the result of combining this theory of
causality with SR, which forbids A->B and B->A as explanations)
follows the existence of predefined "elements of reality". For these
elements of reality we can apply the rules of classical logic and
probability theory.
In this theory, we can derive
(violation of Bell's inequality)=>(A->B or B->A)
I have choosen the freedom to name this EPR-causality because it
remembers the EPR argumentation.
b. There exists another strong axiomatic theory of causality where
this doesn't hold. This is obviously different from the previous
theory of causality.
I have choosen the freedom to name this Bohr-causality.
> > 2. We can formalize EPR causality in a way that Bell's theorem in my
> > form (A->B or B->A) => Bell's inequality becomes a strong theorem.
> If this is true, it is a strong theorem in Bohr's causality, too.
> But this is not what is happening.
Let's not argue about names. Of course, we can name EPR-causality also
Schmelzer-causality and Bohr-causality Kassner-causality. But that is
not the key. Moreover, it has more letters ;-(
> > 4. That means, the combination (SR + EPR-causality) is falsified by Aspect.
> > 5. The combination (SR + Bohr causality) is not falsified.
> > 6. EPR-causality has more predictive power than Bohr causality.
> > 7. The combination (LET + EPR-causality) is not falsified by Aspect.
> > 8. The combination (LET + Bohr-causality) is also not falsified by Aspect
> > 9. The combination (LET + EPR) has less predictive power compared with
> > (SR + EPR). The additional prediction is Bell's inequality.
> > 10. The combination (LET + EPR) has more predictive power compared
> > with (SR + Bohr). This is because (5) Bell's inequality is no longer
> > predicted by (SR+Bohr), but (6) EPR has in general more predictive
> > power than Bohr.
> > 11. That's why I criticize the replacement of (SR + EPR) with (SR +
> > Bohr) as an ad hoc immunization of SR.
Your answer was based on the rejection that there is a difference between
these two notions of causality. I hope, the comments have clarified this.
Even if not, it would be useful to know if you accept these claims,
under assumtion that the modified point 1 is correct.
> > Especially, before Aspect's experiment the preference for SR vs. LET
> > was justified by Popper's criterion of empirical content.
> No. Scientist never cared much about philosophers telling them why to
> prefer certain theories. LET never was a serious competitor
> for SR, even though it is strictly equivalent (once you disallow
> the possibility to detect the ether).
I have not claimed that scientists have cared much, I have claimed
that they have made a correct decision if we apply Popper's
methodology a-posteriori.
> > (If you do not accept this, replace "EPR-causality" below with "a
> > theory of causality where (A->B or B->A) => Bell's inequality holds".)
> This is not useful, *if* any such theory leads to non-Lorentz invariant
> notions of causality (such as yours).
> But maybe, it does not...
Interesting. You reject a theory of causality simply because you don't
like the consequences?
> > 4. That means, the combination (SR + EPR-causality) is falsified by Aspect.
> No. There are three possibilities (as you probably know):
> 1. Local realism is wrong.
> 2. SR is wrong.
> 3. QM is incomplete.
> One of these three must be true.
Or two, or all three.
> Now your choice would be 2.
> The problem is that even if you discard 2, replacing SR
> by LET would *not* save 1.
Who tells you that I want to save 1 - "local" realism? The thing which
is in contradiction with SR _is_ the non-local character.
> It is not *realism* that is the problem. You can have realism.
Yes.
> See Bohm's QM. The problem is *local* realism.
Once the realism should be non-local, we have a real violation of SR.
> Bohm's QM leads to a violation of
> Bell's inequalities even if "not (A->B) and not (B->A)",
> because it requires the two spins to become determined
> at strictly the same moment. So you get nonlocality, even if you
> replace SR by LET. You cannot avoid infinite speed, which implies
> nonlocality.
So what? I have non-local but LET-causal effects. But non-local effect
cannot be SR-causal.
> Which is a meaningless statement to me, since I don't think that
> Bohr and Einstein disagreed about causality. Einstein causality
> and Bohr causality are the same. Schmelzer causality is different.
> It is potential causality in LET and not even potential causality
> in SR, since it includes a lot of spacelike situations, for which
> there is *no* potential causality in SR.
I don't want to request the priority for a simple definition of
causality, which is IMO the most natural and simplest notion of
causality which is possible, and which IMO is the same as Einstein's.
But, if you think there is really something new in my definition of
causality, feel free to introduce the notion "Schmelzer causality" ;-)
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > > Let me explain where I think "potential" causality is useful. The
> > > > actual causality is not a Lorentz-invariant relation, not even
> > > > transitive.
> > > counter-example. Why transitivity should be important, I do not
> > > know.
> > I consider this as a minor misunderstanding not worth discussion.
> > But the Lorentz transformation of the pair of events A,B will be A',B'.
> > Now, there may be an actual causal influence A->B but no actual influence
> > A'->B'.
>
> same. They are only described by different observers.
Ok, this seems to be the misunderstanding between active and passive
transformations. Let's forget it.
> > If you do not like to explain the EPR correlation, I cannot help you.
>
> I believed, I was helping *you* :-).
>
> I think that QM *does* explain it. It gives you a mental image,
> the appropriate language and predictive power. What else do
> you expect from an explanation?
That it describes/defines the "element of reality" which causes a
correlation. This may be a hidden variable as well as a causal
influence. But it should be something real. That's what I expect
from a realistic explanation of an observable correlation.
It seems, to explain my position based on realism may be more
appropriate. Replace EPR-causality by EPR-realism (or
Schmelzer-realism, if you prefer ;-)
> On the other hand, the assumption
> of superluminal transport per se does not explain anything, as
> long as you can't show *how* it happens. You can always "explain"
> things that do not seem to fit in the current theoretical
> framework by postulate or by assumption.
If you require a realistic explanation, you receive at least some
"element of reality". An element with some problem-related
properties, but also with a lot of general properties - all rules of
classical logic and probability theory may be applied for elements
which "really" exist.
Thus, a realistic explanation in this sense gives something - a base
for other considerations which may lead to falsification of the
theory.
> > > Now, unless you are specifying what A means and what B,
> > > I can't see what kind of events you are talking of.
> > > I simply don't see what events you are referring to in
> > > the EPR experiment. Give me an example of observable
> > > events A and B, for which you think that Bell's correlations
> > > imply a causal relationship.
> >
> > ??? The events observed in ideal Bell-like experiments, of course.
>
> Which events? More precisely, please. Your whole argumentation
> is marred by imprecision. I think you can at most define one
> of your "events" (i.e. a local change) per two measurements in the
> EPR experiment, not two.
I don't understand your point here. A and B are space-like separated
events. Near A I measure - by random choice - the spin in direction 0,
60 or 120 degrees of particle 1. Near B I measure by another random
choice the spin of particle 2 in direction 180, 240, 300 degrees.
Then I'm looking at the results and wondering how it can happen that
1. always if I have measured the same direction, I have observed the
same result. Something I can explain by:
a) causal influence of the measurement decision at B on the result of A
b) causal influence of the measurement decision at A on the result of B
c) that the value of each measurement exists independent of the measurement
decision.
Believing into relativity, I have to reject a) and b) and to accept c).
Thus, we have three predefined values x,y,z in (0,1).
2. In this case, the truth value of the statements (x=y) (y=z) (z!=x)
is well-defined, and one of the statements is wrong. By choice, one is
tested. With probability 1/3 I have to find a false one. But I observe
a false one only in 1/4.
That's why I reject c), thus, a) or b) has to be correct. In above
cases, Einstein causality is false.
> It is not causality what you are
> talking about but reality.
Fine. May be that it is indeed better to talk here about reality.
> Causality refers to the relation "->",
> not to "A" or "B". You want to conclude to the existence of
> "A" in a relation "A -> B". But there is always the possibility
> that B happens either for a different reason or for no reason at
> all. This is not a question of causality. A theory of causality
> would deal with the relationship, i.e. *given* two events A and
> B, it might help you decide whether there is a causal relation
> between them. But given only B, it does not give you a unique
> A and sometimes no A at all.
I don't think this is the point. I want to accept EPR as it is, making
it even slightly stronger. If we name this "EPR-causality" or
"EPR-realism" I don't bother, I would even prefer realism. (I doubt it
is possible to make a clear cut between causality and realism.)
> > one. And I do not want to show you that the stronger is the better
> > one.
>
> > But this stronger notion of causality exists. And it is stronger
> > because Aspect falsifies SR + this stronger causality, but not SR +
> > your preferred causality.
>
> I don't think it is stronger. You are not free to change the
> meanings of words at will. Any reasonable use of "stronger" means *more*
> restrictive, not *less*. SR has a stronger notion of causality than
> Newtonian mechanics. But your notion of potential causality
> is closely related to the latter.
No. I think the definition has a general part - how can we conclude
from observation that a causal relation exists - and another part which
claims what causal relations are allowed.
Now, if the first part accepts EPR as a proof of hidden variables if
there is no causal relation A->B or B->A, than this first part is
stronger than a first part which does not accept EPR.
For the second part, SR causality is indeed stronger than LET causality.
Now, combine them and you observe that EPR+SR is too strong. We have
EPR+LET or no-EPR+SR as weaker alternatives.
> Moreover, neither for
> this potential causality nor for the "actual" causality
> that you see in the Bell correlations does SR require that
> it be a relation between non-spacelike events only.
I don't see that "actual" and "potential" causality is a point here.
The point is that I can formalize the theory in a way that the
EPR-argument is valid, i.e. I can simply take it as an axiom (of
causality or of realism is a question of denotation).
> So it does *not* lead to a falsification of SR.
No, if I use an appropriate connection between SR and the theory of
causality/reality which formalizes EPR.
> Find a different word than stronger. "Weaker" would be fine with
> me, but you can also call it "stronger in Ilja's sense", if you like.
Funny, if I combine the "weaker" notion of Ilja-causality with SR, I
obtain a falsification, if I combine it with the "stronger" version of
Klaus-causality, not?
Klaus Kassner
Ilja Schmelzer wrote:
> I assume no energy exchange, but hidden (but real) information
> exchange.
as a consequence of the second law of thermodynamics.
Klaus Kassner
Ilja Schmelzer wrote:
>
> > > Let me explain where I think "potential" causality is useful. The
> > > actual causality is not a Lorentz-invariant relation, not even
> > > transitive.
> > counter-example. Why transitivity should be important, I do not
> > know.
>
> I consider this as a minor misunderstanding not worth discussion.
> But the Lorentz transformation of the pair of events A,B will be A',B'.
> Now, there may be an actual causal influence A->B but no actual influence
> A'->B'.
This I consider to be impossible. Because the two events are the
same. They are only described by different observers.
> If you do not like to explain the EPR correlation, I cannot help you.
I believed, I was helping *you* :-).
I think that QM *does* explain it. It gives you a mental image,
the appropriate language and predictive power. What else do
you expect from an explanation? On the other hand, the assumption
of superluminal transport per se does not explain anything, as
long as you can't show *how* it happens. You can always "explain"
things that do not seem to fit in the current theoretical
framework by postulate or by assumption.
> > Now, unless you are specifying what A means and what B,
> > I can't see what kind of events you are talking of.
> > I simply don't see what events you are referring to in
> > the EPR experiment. Give me an example of observable
> > events A and B, for which you think that Bell's correlations
> > imply a causal relationship.
>
> ??? The events observed in ideal Bell-like experiments, of course.
Which events? More precisely, please. Your whole argumentation
is marred by imprecision. I think you can at most define one
of your "events" (i.e. a local change) per two measurements in the
EPR experiment, not two.
> I agree that a definition of causality is possible which is in
> agreement with SR and Aspect. But there is also another, stronger,
> which is not. Of course, a proponent of relativity prefers the weaker
It is not causality what you are
talking about but reality. Causality refers to the relation "->",
not to "A" or "B". You want to conclude to the existence of
"A" in a relation "A -> B". But there is always the possibility
that B happens either for a different reason or for no reason at
all. This is not a question of causality. A theory of causality
would deal with the relationship, i.e. *given* two events A and
B, it might help you decide whether there is a causal relation
between them. But given only B, it does not give you a unique
A and sometimes no A at all.
> one. And I do not want to show you that the stronger is the better
> one.
> But this stronger notion of causality exists. And it is stronger
> because Aspect falsifies SR + this stronger causality, but not SR +
> your preferred causality.
I don't think it is stronger. You are not free to change the
meanings of words at will. Any reasonable use of "stronger" means *more*
restrictive, not *less*. SR has a stronger notion of causality than
Newtonian mechanics. But your notion of potential causality
is closely related to the latter. Moreover, neither for
this potential causality nor for the "actual" causality
that you see in the Bell correlations does SR require that
it be a relation between non-spacelike events only.
So it does *not* lead to a falsification of SR.
Klaus Kassner
Ilja Schmelzer wrote:
> > I think that QM *does* explain it. It gives you a mental image,
> > the appropriate language and predictive power. What else do
> > you expect from an explanation?
>
> That it describes/defines the "element of reality" which causes a
> correlation. This may be a hidden variable as well as a causal
> influence. But it should be something real. That's what I expect
> from a realistic explanation of an observable correlation.
You have avoided defining "real".
Quantum mechanics gives a clear reason for the correlation.
It also tells you -- according to Bohr -- that naive realism
does not work. You should reconsider your position about what
to call real and what not.
> It seems, to explain my position based on realism may be more
> appropriate. Replace EPR-causality by EPR-realism (or
> Schmelzer-realism, if you prefer ;-)
No, EPR-realism is all right. But nothing new. Even Einstein
accepted that Bohr's position was logically correct.
> > > > Now, unless you are specifying what A means and what B,
> > > > I can't see what kind of events you are talking of.
> > > > I simply don't see what events you are referring to in
> > > > the EPR experiment. Give me an example of observable
> > > > events A and B, for which you think that Bell's correlations
> > > > imply a causal relationship.
> > >
> > > ??? The events observed in ideal Bell-like experiments, of course.
> >
> > Which events? More precisely, please. Your whole argumentation
> > is marred by imprecision. I think you can at most define one
> > of your "events" (i.e. a local change) per two measurements in the
> > EPR experiment, not two.
>
> I don't understand your point here. A and B are space-like separated
> events. Near A I measure - by random choice - the spin in direction 0,
> 60 or 120 degrees of particle 1. Near B I measure by another random
> choice the spin of particle 2 in direction 180, 240, 300 degrees.
>
> Then I'm looking at the results and wondering how it can happen that
What is the event "causing"? The measurement? That's impossible,
because you have no restrictions on the time of your measurement
or on the choice to do it or leave it.
And surely the measurement at A cannot cause the one at B, and
vice versa. Is it the "decision to make the measurement"?
(Are you kidding?)
Is it the outcome of the measurement? The outcome that spin A
is "up" "causes" spin "B" to be down? With or without measurement at
B? You still have not described what A and B are.
> 1. always if I have measured the same direction, I have observed the
> same result. Something I can explain by:
> a) causal influence of the measurement decision at B on the result of A
> b) causal influence of the measurement decision at A on the result of B
> c) that the value of each measurement exists independent of the measurement
> decision.
This is not an exclusive list. (It is based on asking the wrong
questions. Questions like: Are you still beating up your wife
regularly? Only allowed answers yes or no.)
Other possibilities:
d) You do not measure "at" A or "at" B. You measure
a single object, an entangled two-particle wave-function.
You are *not* measuring one or two particles. But your
measurement is of the disentangling type, so after it you
have two particles. The wave function has a *real* property
that leads to the correlation. (So A and B *do* have a
common cause but neither does A influence B nor does B
influence A.)
e) You measure A. The information travels back in time to
tell B to be in the right direction. So even if you measured
B before A, you get the result right.
f) When you measure A, the universe splits. Only the possibility
with spin B opposite to A remains in our universe, the others,
equally real, go to other universes.
> Believing into relativity, I have to reject a) and b) and to accept c).
No, you don't have to as I have shown in my other post.
SR does not contradict *all* types of causal influence that
are superluminal. SR is happy with superluminal
"Schmelzer-causality".
> > It is not causality what you are
> > talking about but reality.
> Fine. May be that it is indeed better to talk here about reality.
But then we can stop dicussing. We have different ideas about
reality. I don't think that you will convince me and I don't think
that I will convince you.
Which brings me to the answer of another question that I promised
you before. It is true that disagreement is *not* a reason to
stop a discussion (and I did not stop it). But if you can be
pretty sure that there will be no result of the discussion,
*that* is a good reason. We have spent much time until clarifying
that we were not discussing causality but reality. Now, I do not
wish to spend an equal amount of time trying to convey my view
of "the fabric of reality" (which still is developing) and
arguing against someone who seems to take reality as a granted concept
(given by the gods).
Nor do I consider these problems as urgent or extremely interesting.
For a minor engagement, they are o.k., but not for long exchanges.
The term is over now, I currently have no teaching obligations,
so I would like to use this time to do some research work
(and I have got to keep my Ph.D. students busy:-)). These discussions
are almost as time-consuming as preparing lectures...
I am quite willing to discuss every once in a while, and I certainly
am interested in discussing general relativity questions, but I really
consider this topic a bit sterile. It is just about questions of
interpretation, not really about physics.
Klaus Kassner
> > > exchange.
> > There is no information exchange without energy exchange,
> > as a consequence of the second law of thermodynamics.
>
No, I don't:-).
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > I think that QM *does* explain it. It gives you a mental image,
> > > the appropriate language and predictive power. What else do
> > > you expect from an explanation?
> >
> > That it describes/defines the "element of reality" which causes a
> > correlation. This may be a hidden variable as well as a causal
> > influence. But it should be something real. That's what I expect
> > from a realistic explanation of an observable correlation.
>
> You have avoided defining "real".
> Quantum mechanics gives a clear reason for the correlation.
> It also tells you -- according to Bohr -- that naive realism
> does not work. You should reconsider your position about what
> to call real and what not.
The existence of Bohmian mechanics shows clearly that naive realism
has no problem with classical quantum theory.
Naive realism becomes problematic only if we try to combine it with
Einstein locality.
Thus, I can easily define "real" following naive realism. I'm not
afraid it the result is SR is incompatible with such a formalized
naive realism ;-)
> > It seems, to explain my position based on realism may be more
> > appropriate. Replace EPR-causality by EPR-realism (or
> > Schmelzer-realism, if you prefer ;-)
> No, EPR-realism is all right. But nothing new. Even Einstein
> accepted that Bohr's position was logically correct.
To accept something as logically correct is a very weak statement.
Solipcism is logically correct.
I also do accept the rejection of EPR-realism (in my understanding) as
logically correct. It is a different theory.
> > > > > Now, unless you are specifying what A means and what B,
> > > > > I can't see what kind of events you are talking of.
> > > > > I simply don't see what events you are referring to in
> > > > > the EPR experiment. Give me an example of observable
> > > > > events A and B, for which you think that Bell's correlations
> > > > > imply a causal relationship.
> > > >
> > > > ??? The events observed in ideal Bell-like experiments, of course.
> > >
> > > Which events? More precisely, please. Your whole argumentation
> > > is marred by imprecision. I think you can at most define one
> > > of your "events" (i.e. a local change) per two measurements in the
> > > EPR experiment, not two.
> >
> > I don't understand your point here. A and B are space-like separated
> > events. Near A I measure - by random choice - the spin in direction 0,
> > 60 or 120 degrees of particle 1. Near B I measure by another random
> > choice the spin of particle 2 in direction 180, 240, 300 degrees.
> >
> > Then I'm looking at the results and wondering how it can happen that
>
> What is the event "causing"? The measurement? That's impossible,
> because you have no restrictions on the time of your measurement
> or on the choice to do it or leave it.
The choice of the measurement (which of the three) which has to be
done near A (means, a small, restricted region of spacetime around A,
so that for any point of this region and any point of another region
around B the distance between them is space-like).
How the choice is made is left open, free will, classical or quantum
randomness. I'm not talking about Aspect, but an ideal Bell-type
experiment, assuming that QM is correct for all of them.
> And surely the measurement at A cannot cause the one at B, and
> vice versa. Is it the "decision to make the measurement"?
> (Are you kidding?)
Yes. The decision which measurement has to be done.
> Is it the outcome of the measurement? The outcome that spin A
> is "up" "causes" spin "B" to be down?
Indirectly. Because the outcome is not predefined by a hidden variable
(because of the violation of Bell's inequality we have to exclude this
explanation), the related information is created at this moment, after
the choice to make the measurement in this direction.
> With or without measurement at B?
Depends on the time ordering between A and B.
> You still have not described what A and B are.
Small space-like separated regions is spacetime.
> > 1. always if I have measured the same direction, I have observed the
> > same result. Something I can explain by:
>
> > a) causal influence of the measurement decision at B on the result of A
> > b) causal influence of the measurement decision at A on the result of B
> > c) that the value of each measurement exists independent of the measurement
> > decision.
>
> This is not an exclusive list.
According to EPR it is an exclusive list. (Or a reformulation of the
EPR criterion in other words.) Name it the Schmelzer axiom that this
is an exclusive list.
> Other possibilities:
>
> d) You do not measure "at" A or "at" B. You measure
> a single object, an entangled two-particle wave-function.
The information which type of measurement I choose is created locally.
If I cannot create information locally, I can never prove the
existence of causal influences with superluminal velocity, even if
they exist, even if I have immediate phone contact to Andromeda.
Thus, explanation (d) makes the notion of Einstein causality
unfalsifiable.
> e) You measure A. The information travels back in time to
> tell B to be in the right direction. So even if you measured
> B before A, you get the result right.
Explanation (e) makes the notion of Einstein causality
unfalsifiable.
> f) When you measure A, the universe splits. Only the possibility
> with spin B opposite to A remains in our universe, the others,
> equally real, go to other universes.
Explanation (f) makes the notion of Einstein causality
unfalsifiable.
> > > It is not causality what you are
> > > talking about but reality.
> > Fine. May be that it is indeed better to talk here about reality.
> But then we can stop dicussing. We have different ideas about
> reality. I don't think that you will convince me and I don't think
> that I will convince you.
What about the following agreement:
1. naive realism may be formalized as Ilja-realism.
2. your realism is something different, I don't bother about the details.
3. SR is incompatible with Ilja-realism and experiment.
4. SR is compatible with Klaus-realism and experiment.
5. Ilja-realism is stronger than Klaus-realism, because Ilja-realism
forbids the explanations (d)-(f) for observable correlations.
6. Ilja-realism is compatible with LET and experiment.
> Which brings me to the answer of another question that I promised
> you before. It is true that disagreement is *not* a reason to
> stop a discussion (and I did not stop it). But if you can be
> pretty sure that there will be no result of the discussion,
> *that* is a good reason. We have spent much time until clarifying
> that we were not discussing causality but reality.
But we have clarified this. I have also learned that my usage of the
notion "event" has caused misunderstandings.
> Now, I do not
> wish to spend an equal amount of time trying to convey my view
> of "the fabric of reality" (which still is developing) and
> arguing against someone who seems to take reality as a granted concept
> (given by the gods).
In my response, you see that I accept that we have a different
understanding (different definition) of realism. I do not want to
convert you. But some simple consequences of my definition of realism
(or causality) are simply true, theorems.
You have claimed that there is something wrong with these statements.
If you make such claims, you have to show it.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > The existence of Bohmian mechanics shows clearly that naive realism
> > has no problem with classical quantum theory.
> degenerate ones: A -> B does not imply A before B in time, they
> can be at equal times; so A -> B -> A is possible; don't tell
> me it is not measurable; that is not an issue in naive realism).
Ok, we use a variant of Bohmian mechanics which explicitly forbids
different measurements at the same time. This variant has no such
causal loops. Nonetheless, the predictions coinside with QM as well
as for unmodified Bohmian mechanics.
Klaus Kassner
Ilja Schmelzer wrote:
>
> Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > > I think that QM *does* explain it. It gives you a mental image,
> > > > the appropriate language and predictive power. What else do
> > > > you expect from an explanation?
> > >
> > > That it describes/defines the "element of reality" which causes a
> > > correlation. This may be a hidden variable as well as a causal
> > > influence. But it should be something real. That's what I expect
> > > from a realistic explanation of an observable correlation.
> >
> > You have avoided defining "real".
> > Quantum mechanics gives a clear reason for the correlation.
> > It also tells you -- according to Bohr -- that naive realism
> > does not work. You should reconsider your position about what
> > to call real and what not.
>
> The existence of Bohmian mechanics shows clearly that naive realism
> has no problem with classical quantum theory.
David L Evens
Ilja Schmelzer (schm...@fermi.wias-berlin.de) wrote:
: Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
: > > exchange.
: > There is no information exchange without energy exchange,
: > as a consequence of the second law of thermodynamics.
: You apply thermodynamics outside it's range of applicability.
There are no limitations to the rules governing energy trasference.
: Of course, it may be, that a more fundamental theory (atomic ether
: theory) nonetheless describes this information exchange as combined
: with some energy transfer which is not observable in the continuous
: approximation. In this case, you can apply the second law to this
: theory. In an approximation, the energy exchange may be neglected.
When you deliberately make mistakes, as you do in all approximations,
mant important things will be neglected.
Of course, gicen the abscence of any validity in any ether theory, you
can be safely recognised as insignificant.
--
---------------------------+--------------------------------------------------
Ring around the neutron, | "OK, so he's not terribly fearsome.
A pocket full of positrons,| But he certainly took us by surprise!"
A fission, a fusion, +--------------------------------------------------
We all fall down! | "Was anybody in the Maquis working for me?"
---------------------------+--------------------------------------------------
"I'd cut down ever Law in England to get at the Devil!"
"And what man could stand up in the wind that would blow once you'd cut
down all the laws?"
------------------------------------------------------------------------------
This message may not be carried on any server which places restrictions
on content.
------------------------------------------------------------------------------
e-mail will be posted as I see fit.
------------------------------------------------------------------------------
Klaus Kassner
> > > has no problem with classical quantum theory.
> > degenerate ones: A -> B does not imply A before B in time, they
> > can be at equal times; so A -> B -> A is possible; don't tell
> > me it is not measurable; that is not an issue in naive realism).
>
> different measurements at the same time. This variant has no such
> causal loops. Nonetheless, the predictions coinside with QM as well
> as for unmodified Bohmian mechanics.
about the supposed real background structure. For this structure,
it is inevitable in Bohmian mechanics to have instantaneous
interactions and thus causal loops. If you say, only observables
are real, then you don't need Bohmian mechanics; but if you insist
on real concepts besides observables then you should also worry
about causality violations that are not observable.
(We were not talking about measurements but about realism.
Realism means taking the existence of properties for granted
even when they are not measured.)
Moreover, there are other problems in Bohmian mechanics which
you may not realize. For example, observables like momentum
or spin are as indeterminate ("nonreal") in Bohmian mechanics
as in ordinary QM. Only *positions* get "real". Thus, the EPR
"paradox" in its original form (a wave function where the
sum of momenta and the difference of positions are both
elements of reality, i.e. the wave function is an eigenstate
to these two operators, and you measure positions or momenta,
or both) might be "resolvable" by referring to a Bohmian argument.
But in its newer form, using spins (actually I think, Bohm himself
cooked it up), Bohmian mechanics does not help, since there is
no more an "agent" determining the far-away spin than in ordinary
QM. Quite an irony, because Bohm did not realize this, at least
initially. (Nor did Bell.)
There are papers by Detlev Duerr from Munich, a modern proponent of
Bohmian mechanics, in which it is clearly stated that neither
momentum nor spin are real properties of particles in Bohmian
mechanics.
Ilja Schmelzer
Klaus Kassner <klaus....@physik.uni-magdeburg.de> writes:
> > > > The existence of Bohmian mechanics shows clearly that naive realism
> > > > has no problem with classical quantum theory.
> > > degenerate ones: A -> B does not imply A before B in time, they
> > > can be at equal times; so A -> B -> A is possible; don't tell
> > > me it is not measurable; that is not an issue in naive realism).
> >
> > Ok, we use a variant of Bohmian mechanics which explicitly forbids
> > different measurements at the same time. This variant has no such
> > causal loops. Nonetheless, the predictions coinside with QM as well
> > as for unmodified Bohmian mechanics.
>
> That would not help. I was not talking about measurements, but
> about the supposed real background structure. For this structure,
> it is inevitable in Bohmian mechanics to have instantaneous
> interactions and thus causal loops. If you say, only observables
> are real, then you don't need Bohmian mechanics; but if you insist
> on real concepts besides observables then you should also worry
> about causality violations that are not observable.
> (We were not talking about measurements but about realism.
> Realism means taking the existence of properties for granted
> even when they are not measured.)
This artificial problem about instantaneous interactions is nothing I
really have to bother about. I can simply introduce another minor
modification: I assume not instantaneous interaction but real
influence with some, yet unknown, very high velocity, say 100000000
c. This theory cannot be distinguished from Bohmian mechanics using
current experimental data. It is in principle distinguishable if we
have accurate enough time measurement, because it predicts Bell's
inequality, thus, a violation of QM, for far away contemporary
measurements.
> Moreover, there are other problems in Bohmian mechanics which
> you may not realize. For example, observables like momentum
> or spin are as indeterminate ("nonreal") in Bohmian mechanics
> as in ordinary QM. Only *positions* get "real". Thus, the EPR
> "paradox" in its original form (a wave function where the
> sum of momenta and the difference of positions are both
> elements of reality, i.e. the wave function is an eigenstate
> to these two operators, and you measure positions or momenta,
> or both) might be "resolvable" by referring to a Bohmian argument.
> But in its newer form, using spins (actually I think, Bohm himself
> cooked it up), Bohmian mechanics does not help, since there is
> no more an "agent" determining the far-away spin than in ordinary
> QM. Quite an irony, because Bohm did not realize this, at least
> initially. (Nor did Bell.)
>
> There are papers by Detlev Duerr from Munich, a modern proponent of
> Bohmian mechanics, in which it is clearly stated that neither
> momentum nor spin are real properties of particles in Bohmian
> mechanics.
It does not have a great influence on the math if we use another set
of common observables which we name "real" in some variant of Bohmian
mechanics.
Of course, if you really want to propose Bohmian mechanics as the true
theory of everything, it looks strange if you define one spin
component (in the preferred direction of our universe ;-) to be real
and the other as not real. But I need this theory only to show that
there is no contradiction between realism and QM in principle. For
this compatibility proof even such a strange variant is sufficient.
Klaus Kassner
Ilja Schmelzer wrote:
> This artificial problem about instantaneous interactions is nothing I
> really have to bother about. I can simply introduce another minor
> modification: I assume not instantaneous interaction but real
> influence with some, yet unknown, very high velocity, say 100000000
> c.
What complete nonsense. Nobody knowing about Bohmian
mechanics would say anything like that. There is no "speed of
interaction"! The instantaneous influence of one particle on another
comes from the fact that the position of one particle depends on
that of all the others at the same time, no matter where they
are. You can't change this in any simple way without destroying
the theory. Instantaneous "interaction" is a *structural* property
of these theories.
> This theory cannot be distinguished from Bohmian mechanics using
> current experimental data.
There is no such theory.
> It does not have a great influence on the math if we use another set
> of common observables which we name "real" in some variant of Bohmian
> mechanics.
Again, a statement born out of ignorance. Only those variables that
are present already in classical mechanics can be made "real"
this way. Observables that would not exist classically, like
"spin", can't.
> and the other as not real. But I need this theory only to show that
> there is no contradiction between realism and QM in principle. For
> this compatibility proof even such a strange variant is sufficient.
Bohmian mechanics as a proof of the possibility of a naively
realistic theory equivalent to QM has *failed*.
If you state that there is such a theory, you have to provide
one and show its internal consistency.