Thomas Dent
Jun25-04, 07:16 AM
<jabberwocky><div class="vbmenu_control"><a href="jabberwocky:;" onClick="newWindow=window.open('','usenetCode','toolbar=no, location=no,scrollbars=yes,resizable=yes,status=no ,width=650,height=400'); newWindow.document.write('<HTML><HEAD><TITLE>Usenet ASCII</TITLE></HEAD><BODY topmargin=0 leftmargin=0 BGCOLOR=#F1F1F1><table border=0 width=625><td bgcolor=midnightblue><font color=#F1F1F1>This Usenet message\'s original ASCII form: </font></td></tr><tr><td width=449><br><br><font face=courier><UL><PRE>"Charlie Stromeyer Jr." <cstromey@hotmail.com> wrote\n\n> The comments I made earlier in this newsgroup in the message below\n> about non-causality are still true for both theoretical and\n> experimetal reasons, i.e. various non-causal phenomena have already\n> been detected by experimentalists.\n\n> [3] "Nutational Stimulated Photon Echoes", Optics Letters, v27(iss 13) (2002), pp..\n\nAs a public service I tried to read the Optics Letters article. Now I\nam not an expert in optics, but it seems that Charlie has been\nconfused by the terminology "noncausal direction".\n\nI quote: "It is well known that a probe pulse shorter than tau_21 has\nto be in the k+ direction to generate a real echo, which should\npropagate only along k- because of phase matching and causality" (...)\n"Instead of a brief pulse, we consider a quasi-continuous probe field,\nincident along k+" (...) "Typically, a brief probe along k- should\nnot generate an echo. But a quasi-continuous probe field along k-\ngives a different result." (...) "The field in noncausal direction k+\ndoes not cancel as is would for the typical brief pulse SPE". (...)\n\nNow, I\'m getting to the point:\n\n"Eq.(7) also represents an echo field with the same delay as the echo\nin the causal direction. The existence of this echo does not violate\ncausality. The reason that an echo occurs in the noncausal direction\nis that the probe pulse is longer than the delay tau_21. The front\npart of the probe pulse interacts with the grating and creates a\ncoherence that does not generate a propagating field because of the\nphase-matching condition. However the probe field incident into the\nmedium anytime tau_21 later than the front part rephases this\ncoherence into the propagating direction k+ and emits an echo."\n\nSo, on closer examination, the use of the phrase "noncausal direction"\nis only a piece of jargon that does not mean that causality is\nactually violated. I suspect that other supposed examples of causality\nviolation will also turn out actually to be examples of\nmisinterpretation.\n\nIt is hardly surprising that QED respects causality since the theory\nwas explicitly set up to do so.\n\nHowever, this does not rule out "faster-than-light" signalling:\nScharnhorst photons can do the trick. But they only exist in a context\nwhich explicitly breaks Lorentz invariance, such as between a pair of\nconducting plates. Hence they do not threaten causality directly.\n\n> For instance, QED is the most overall accurate scientific theory ever.\n> In a paper about acausality in QED [quant-ph/9802056], the author\n> proves in section 5 that photons can exhibit acausal behavior in both\n> time-like and/or space-like (superluminal speed) directions.\n\nAs for this paper, I am not sure if their results are correct, i.e.\nwhether their mathematical expressions for superluminal transmission\nactually represent something measurable, or only the fact that QM is a\nnonlocal theory. For example, if the wavefunction of a particle is\nnonvanishing at two spacelike separated points, there would appear to\nbe a finite probability of finding the particle at both points, in\nwhich case the particle could be said to have travelled faster than\nlight. Nevertheless, we do not see such things happening, which\nindicates that this interpretation of QM as acausal is incorrect, or\nat least incomplete. Similarly the Einstein-Podolsky-Rosen nonlocal\ncorrelations do not directly threaten causality. There is no evidence\nthat the future can influence the past.\n\nOne should keep in mind that any such explicit causality violation\ncapable of creating causal loops would likely produce an internally\ninconsistent theory: it would set up logical paradoxes and possible\nlead to divergent quantum fluctuations. (I don\'t have a reference for\nquantum fluctuations diverging close to causal loops, perhaps someone\ncan find it for me.)\n\n> At first, one might suspect that such a result might rule out certain\n> ideas within string theory such as some ideas about large extra\n> dimensions.\n\nWhy?\n\n> However, I myself would not make such a hypothesis for\n> various reasons, e.g. I have not yet looked at any of the recent\n> papers about photon propagation in noncommutative QED which of course\n> is a different and perhaps more fundamental environment than the usual\n> context for QED.\n\nNoncommutativity is of course another way of having a nonlocal theory,\nbut that doesn\'t necessarily imply lack of causality either.\n\n> Albert Einstein once said which is that "No one really understands the\n> nature of the photon". Richard Feynman also said or implied multiple\n> times that no one really seems to understand QT.\n\nBut it doesn\'t follow from that that anything goes.\n\n> At first, I thought that the AJL model must be flawed for reasons\n> mentioned in post [1], however, this may no longer be the case because\n> there are two recent papers which argue for separate reasons that\n> superluminal signals may actually be compatible with GTR [3].\n\n> > [3] http://arxiv.org/abs/gr-qc/0304059\n\nThis is a different ref. [3]! - *one* paper by Graham Shore. I quote\nfrom p.25:\n\n"Serious conceptual issues do arise through the apparent prediction\n(...) of a superluminal speed of light. We have (...) discussed the\nsubtle questions of interpretation that arise. (...) It appears that\ndespite an anomalous refractive index featuring a superluminal\nlow-frequency phase (and signal) velocity, the critical high-frequency\nlimit which determines the characteristics or wavefront velocity\nremains equal to c. Nonetheless, we described how superluminal\nvelocities could be accommodated in the framework of GR while\npreserving the essential notion of stable causality."\n\nSo Shore doesn\'t support the thesis that QED in curved space can\nviolate causality.\n\n> 1) Various quantum phenomena have already been demonstrated to be\n> "noncausal" both theoretically and sometimes even experimentally. (...)\n> except for the one very important conundrum I ask about below, the\n> term "noncausal" means only that there is no discernable and meaningful\n> dependence upon causality which is different from the idea that there\n> is some kind of explicit violation of Einstein causality via\n> superluminal signals.\n\nThis is confused. Causality means exactly that one can define a past\n\'light-cone\' and it is distinct from the future \'light-cone\'. Either a\ngiven phenomenon is consistent with causality, in which case it makes\nno sense to call it "noncausal", or it violates casuality. Also,\nsuperluminal signals do *not* imply causality violation unless you\nalso have unbroken global Lorentz invariance. See section 5.1 in the\nShore paper http://arxiv.org/abs/gr-qc/0304059. The curved backgrounds\nand SEP-violating theories in which superluminal propagation appears\nto happen do not have the required global Lorentz invariance.\n\n> For instance, experiments with TmYAG crystals have shown that\n> stimulated photon ehoes (SPEs) can exist in the noncausal direction\n> [3]\n\nDebunked above.\n\n> and separate experiments have shown that the phase and energy of\n> a photon pulse can travel faster than c, the speed of light in a\n> vacuum, but there does not (yet, anyways) seem to be meaningful\n> information transmitted via superluminal signals due to cancellation\n> of such potential superluminal signals because of complicated\n> diffraction and diffusion effects [4].\n\n> [4] J.J. Carey et al., Phys Rev Lett, v84(no7) (2000), p.1431.\n\nI\'m not sure that the experiment really shows that energy can travel\nfaster than c, or whether they are just cutting off the back of a\npulse and then noting that the pulse centroid is superluminal. The\nShore paper says the energy-transfer velocity is always < c (p.16).\n\nNote that the interpretation of the Carey paper is disputed by Mochan\nand Brudny. Carey et al. reply that they never meant to imply that\ncausality was violated: "We believe that it should be possible to\nprove theoretically that casuality is not violated in FTIR". See Phys.\nRev. Lett. 87, 119102 (2001).\n\n> Suppose that one definition of the presence of "acausality" would be\n> the existence of uncertainty which is clearly non-statistical (or\n> non-probabilistic). Well, this is what happens in the photon\n> experiment described in this brief four page paper\n> http://www.arxiv.org/abs/quant-ph/0102109,\n> yet there are no superluminal communications necssarily entailed !\n\nNow you have changed the definition of acausality, hence it is not\nsurprising that the meaning of the word has also changed. Let us read\nthe abstract:\n\n"The photon\'s wave function affects only one out of the atoms,\nregardless of its position within the row, thereby manifesting not\nonly non-local but also non-sequential characteristics. It also turns\nout that, out of n atoms, each one has a probability which is higher\nthan the normal 1/n to be the single affected one."\n\nWhy should we read any further? We are to believe that there exist n\npositive numbers, each of which is greater than 1/n, but that still\nsum to 1. This is indeed non-statistical and non-probabilistic, but\nalso non-sensical.\n\n> Thomas, since you are from Scandinavia you should heed what the Prince\n> of Denmark once said and also do not forget about ghosts :-)\n\nPrecisely, ghosts are non-physical; and the Prince of Denmark never\nspoke those words, only an actor on a stage.\n\n> http://arxiv.org/abs/gr-qc/0403121\n\nNote that this last paper (which I am not endorsing!) *defends* the\nnotion of a casual arrow of time which distinguishes the future from\nthe past.\n\n</UL></PRE></font></td></tr></table></BODY><HTML>');"> <IMG SRC=/images/buttons/ip.gif BORDER=0 ALIGN=CENTER ALT="View this Usenet post in original ASCII form"> View this Usenet post in original ASCII form </a></div><P></jabberwocky>"Charlie Stromeyer Jr." <cstromey@hotmail.com> wrote
> The comments I made earlier in this newsgroup in the message below
> about non-causality are still true for both theoretical and
> experimetal reasons, i.e. various non-causal phenomena have already
> been detected by experimentalists.
> [3] "Nutational Stimulated Photon Echoes", Optics Letters, v27(iss 13) (2002), pp..
As a public service I tried to read the Optics Letters article. Now I
am not an expert in optics, but it seems that Charlie has been
confused by the terminology "noncausal direction".
I quote: "It is well known that a probe pulse shorter than \tau_21 has
to be in the k+ direction to generate a real echo, which should
propagate only along k- because of phase matching and causality" (...)
"Instead of a brief pulse, we consider a quasi-continuous probe field,
incident along k+" (...) "Typically, a brief probe along k- should
not generate an echo. But a quasi-continuous probe field along k-
gives a different result." (...) "The field in noncausal direction k+
does not cancel as is would for the typical brief pulse SPE". (...)
Now, I'm getting to the point:
"Eq.(7) also represents an echo field with the same delay as the echo
in the causal direction. The existence of this echo does not violate
causality. The reason that an echo occurs in the noncausal direction
is that the probe pulse is longer than the delay \tau_21. The front
part of the probe pulse interacts with the grating and creates a
coherence that does not generate a propagating field because of the
phase-matching condition. However the probe field incident into the
medium anytime \tau_21 later than the front part rephases this
coherence into the propagating direction k+ and emits an echo."
So, on closer examination, the use of the phrase "noncausal direction"
is only a piece of jargon that does not mean that causality is
actually violated. I suspect that other supposed examples of causality
violation will also turn out actually to be examples of
misinterpretation.
It is hardly surprising that QED respects causality since the theory
was explicitly set up to do so.
However, this does not rule out "faster-than-light" signalling:
Scharnhorst photons can do the trick. But they only exist in a context
which explicitly breaks Lorentz invariance, such as between a pair of
conducting plates. Hence they do not threaten causality directly.
> For instance, QED is the most overall accurate scientific theory ever.
> In a paper about acausality in QED [http://www.arxiv.org/abs/quant-ph/9802056], the author
> proves in section 5 that photons can exhibit acausal behavior in both
> time-like and/or space-like (superluminal speed) directions.
As for this paper, I am not sure if their results are correct, i.e.
whether their mathematical expressions for superluminal transmission
actually represent something measurable, or only the fact that QM is a
nonlocal theory. For example, if the wavefunction of a particle is
nonvanishing at two spacelike separated points, there would appear to
be a finite probability of finding the particle at both points, in
which case the particle could be said to have travelled faster than
light. Nevertheless, we do not see such things happening, which
indicates that this interpretation of QM as acausal is incorrect, or
at least incomplete. Similarly the Einstein-Podolsky-Rosen nonlocal
correlations do not directly threaten causality. There is no evidence
that the future can influence the past.
One should keep in mind that any such explicit causality violation
capable of creating causal loops would likely produce an internally
inconsistent theory: it would set up logical paradoxes and possible
lead to divergent quantum fluctuations. (I don't have a reference for
quantum fluctuations diverging close to causal loops, perhaps someone
can find it for me.)
> At first, one might suspect that such a result might rule out certain
> ideas within string theory such as some ideas about large extra
> dimensions.
Why?
> However, I myself would not make such a hypothesis for
> various reasons, e.g. I have not yet looked at any of the recent
> papers about photon propagation in noncommutative QED which of course
> is a different and perhaps more fundamental environment than the usual
> context for QED.
Noncommutativity is of course another way of having a nonlocal theory,
but that doesn't necessarily imply lack of causality either.
> Albert Einstein once said which is that "No one really understands the
> nature of the photon". Richard Feynman also said or implied multiple
> times that no one really seems to understand QT.
But it doesn't follow from that that anything goes.
> At first, I thought that the AJL model must be flawed for reasons
> mentioned in post [1], however, this may no longer be the case because
> there are two recent papers which argue for separate reasons that
> superluminal signals may actually be compatible with GTR [3].
> > [3] http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0304059
This is a different ref. [3]! - *one* paper by Graham Shore. I quote
from p.25:
"Serious conceptual issues do arise through the apparent prediction
(...) of a superluminal speed of light. We have (...) discussed the
subtle questions of interpretation that arise. (...) It appears that
despite an anomalous refractive index featuring a superluminal
low-frequency phase (and signal) velocity, the critical high-frequency
limit which determines the characteristics or wavefront velocity
remains equal to c. Nonetheless, we described how superluminal
velocities could be accommodated in the framework of GR while
preserving the essential notion of stable causality."
So Shore doesn't support the thesis that QED in curved space can
violate causality.
> 1) Various quantum phenomena have already been demonstrated to be
> "noncausal" both theoretically and sometimes even experimentally. (...)
> except for the one very important conundrum I ask about below, the
> term "noncausal" means only that there is no discernable and meaningful
> dependence upon causality which is different from the idea that there
> is some kind of explicit violation of Einstein causality via
> superluminal signals.
This is confused. Causality means exactly that one can define a past
'light-cone' and it is distinct from the future 'light-cone'. Either a
given phenomenon is consistent with causality, in which case it makes
no sense to call it "noncausal", or it violates casuality. Also,
superluminal signals do *not* imply causality violation unless you
also have unbroken global Lorentz invariance. See section 5.1 in the
Shore paper http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0304059. The curved backgrounds
and SEP-violating theories in which superluminal propagation appears
to happen do not have the required global Lorentz invariance.
> For instance, experiments with TmYAG crystals have shown that
> stimulated photon ehoes (SPEs) can exist in the noncausal direction
> [3]
Debunked above.
> and separate experiments have shown that the phase and energy of
> a photon pulse can travel faster than c, the speed of light in a
> vacuum, but there does not (yet, anyways) seem to be meaningful
> information transmitted via superluminal signals due to cancellation
> of such potential superluminal signals because of complicated
> diffraction and diffusion effects [4].
> [4] J.J. Carey et al., Phys Rev Lett, v84(no7) (2000), p.1431.
I'm not sure that the experiment really shows that energy can travel
faster than c, or whether they are just cutting off the back of a
pulse and then noting that the pulse centroid is superluminal. The
Shore paper says the energy-transfer velocity is always < c (p.16).
Note that the interpretation of the Carey paper is disputed by Mochan
and Brudny. Carey et al. reply that they never meant to imply that
causality was violated: "We believe that it should be possible to
prove theoretically that casuality is not violated in FTIR". See Phys.
Rev. Lett. 87, 119102 (2001).
> Suppose that one definition of the presence of "acausality" would be
> the existence of uncertainty which is clearly non-statistical (or
> non-probabilistic). Well, this is what happens in the photon
> experiment described in this brief four page paper
> http://www.arxiv.org/abs/http://www.arxiv.org/abs/quant-ph/0102109,
> yet there are no superluminal communications necssarily entailed !
Now you have changed the definition of acausality, hence it is not
surprising that the meaning of the word has also changed. Let us read
the abstract:
"The photon's wave function affects only one out of the atoms,
regardless of its position within the row, thereby manifesting not
only non-local but also non-sequential characteristics. It also turns
out that, out of n atoms, each one has a probability which is higher
than the normal 1/n to be the single affected one."
Why should we read any further? We are to believe that there exist n
positive numbers, each of which is greater than 1/n, but that still
sum to 1. This is indeed non-statistical and non-probabilistic, but
also non-sensical.
> Thomas, since you are from Scandinavia you should heed what the Prince
> of Denmark once said and also do not forget about ghosts :-)
Precisely, ghosts are non-physical; and the Prince of Denmark never
spoke those words, only an actor on a stage.
> http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0403121
Note that this last paper (which I am not endorsing!) *defends* the
notion of a casual arrow of time which distinguishes the future from
the past.
> The comments I made earlier in this newsgroup in the message below
> about non-causality are still true for both theoretical and
> experimetal reasons, i.e. various non-causal phenomena have already
> been detected by experimentalists.
> [3] "Nutational Stimulated Photon Echoes", Optics Letters, v27(iss 13) (2002), pp..
As a public service I tried to read the Optics Letters article. Now I
am not an expert in optics, but it seems that Charlie has been
confused by the terminology "noncausal direction".
I quote: "It is well known that a probe pulse shorter than \tau_21 has
to be in the k+ direction to generate a real echo, which should
propagate only along k- because of phase matching and causality" (...)
"Instead of a brief pulse, we consider a quasi-continuous probe field,
incident along k+" (...) "Typically, a brief probe along k- should
not generate an echo. But a quasi-continuous probe field along k-
gives a different result." (...) "The field in noncausal direction k+
does not cancel as is would for the typical brief pulse SPE". (...)
Now, I'm getting to the point:
"Eq.(7) also represents an echo field with the same delay as the echo
in the causal direction. The existence of this echo does not violate
causality. The reason that an echo occurs in the noncausal direction
is that the probe pulse is longer than the delay \tau_21. The front
part of the probe pulse interacts with the grating and creates a
coherence that does not generate a propagating field because of the
phase-matching condition. However the probe field incident into the
medium anytime \tau_21 later than the front part rephases this
coherence into the propagating direction k+ and emits an echo."
So, on closer examination, the use of the phrase "noncausal direction"
is only a piece of jargon that does not mean that causality is
actually violated. I suspect that other supposed examples of causality
violation will also turn out actually to be examples of
misinterpretation.
It is hardly surprising that QED respects causality since the theory
was explicitly set up to do so.
However, this does not rule out "faster-than-light" signalling:
Scharnhorst photons can do the trick. But they only exist in a context
which explicitly breaks Lorentz invariance, such as between a pair of
conducting plates. Hence they do not threaten causality directly.
> For instance, QED is the most overall accurate scientific theory ever.
> In a paper about acausality in QED [http://www.arxiv.org/abs/quant-ph/9802056], the author
> proves in section 5 that photons can exhibit acausal behavior in both
> time-like and/or space-like (superluminal speed) directions.
As for this paper, I am not sure if their results are correct, i.e.
whether their mathematical expressions for superluminal transmission
actually represent something measurable, or only the fact that QM is a
nonlocal theory. For example, if the wavefunction of a particle is
nonvanishing at two spacelike separated points, there would appear to
be a finite probability of finding the particle at both points, in
which case the particle could be said to have travelled faster than
light. Nevertheless, we do not see such things happening, which
indicates that this interpretation of QM as acausal is incorrect, or
at least incomplete. Similarly the Einstein-Podolsky-Rosen nonlocal
correlations do not directly threaten causality. There is no evidence
that the future can influence the past.
One should keep in mind that any such explicit causality violation
capable of creating causal loops would likely produce an internally
inconsistent theory: it would set up logical paradoxes and possible
lead to divergent quantum fluctuations. (I don't have a reference for
quantum fluctuations diverging close to causal loops, perhaps someone
can find it for me.)
> At first, one might suspect that such a result might rule out certain
> ideas within string theory such as some ideas about large extra
> dimensions.
Why?
> However, I myself would not make such a hypothesis for
> various reasons, e.g. I have not yet looked at any of the recent
> papers about photon propagation in noncommutative QED which of course
> is a different and perhaps more fundamental environment than the usual
> context for QED.
Noncommutativity is of course another way of having a nonlocal theory,
but that doesn't necessarily imply lack of causality either.
> Albert Einstein once said which is that "No one really understands the
> nature of the photon". Richard Feynman also said or implied multiple
> times that no one really seems to understand QT.
But it doesn't follow from that that anything goes.
> At first, I thought that the AJL model must be flawed for reasons
> mentioned in post [1], however, this may no longer be the case because
> there are two recent papers which argue for separate reasons that
> superluminal signals may actually be compatible with GTR [3].
> > [3] http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0304059
This is a different ref. [3]! - *one* paper by Graham Shore. I quote
from p.25:
"Serious conceptual issues do arise through the apparent prediction
(...) of a superluminal speed of light. We have (...) discussed the
subtle questions of interpretation that arise. (...) It appears that
despite an anomalous refractive index featuring a superluminal
low-frequency phase (and signal) velocity, the critical high-frequency
limit which determines the characteristics or wavefront velocity
remains equal to c. Nonetheless, we described how superluminal
velocities could be accommodated in the framework of GR while
preserving the essential notion of stable causality."
So Shore doesn't support the thesis that QED in curved space can
violate causality.
> 1) Various quantum phenomena have already been demonstrated to be
> "noncausal" both theoretically and sometimes even experimentally. (...)
> except for the one very important conundrum I ask about below, the
> term "noncausal" means only that there is no discernable and meaningful
> dependence upon causality which is different from the idea that there
> is some kind of explicit violation of Einstein causality via
> superluminal signals.
This is confused. Causality means exactly that one can define a past
'light-cone' and it is distinct from the future 'light-cone'. Either a
given phenomenon is consistent with causality, in which case it makes
no sense to call it "noncausal", or it violates casuality. Also,
superluminal signals do *not* imply causality violation unless you
also have unbroken global Lorentz invariance. See section 5.1 in the
Shore paper http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0304059. The curved backgrounds
and SEP-violating theories in which superluminal propagation appears
to happen do not have the required global Lorentz invariance.
> For instance, experiments with TmYAG crystals have shown that
> stimulated photon ehoes (SPEs) can exist in the noncausal direction
> [3]
Debunked above.
> and separate experiments have shown that the phase and energy of
> a photon pulse can travel faster than c, the speed of light in a
> vacuum, but there does not (yet, anyways) seem to be meaningful
> information transmitted via superluminal signals due to cancellation
> of such potential superluminal signals because of complicated
> diffraction and diffusion effects [4].
> [4] J.J. Carey et al., Phys Rev Lett, v84(no7) (2000), p.1431.
I'm not sure that the experiment really shows that energy can travel
faster than c, or whether they are just cutting off the back of a
pulse and then noting that the pulse centroid is superluminal. The
Shore paper says the energy-transfer velocity is always < c (p.16).
Note that the interpretation of the Carey paper is disputed by Mochan
and Brudny. Carey et al. reply that they never meant to imply that
causality was violated: "We believe that it should be possible to
prove theoretically that casuality is not violated in FTIR". See Phys.
Rev. Lett. 87, 119102 (2001).
> Suppose that one definition of the presence of "acausality" would be
> the existence of uncertainty which is clearly non-statistical (or
> non-probabilistic). Well, this is what happens in the photon
> experiment described in this brief four page paper
> http://www.arxiv.org/abs/http://www.arxiv.org/abs/quant-ph/0102109,
> yet there are no superluminal communications necssarily entailed !
Now you have changed the definition of acausality, hence it is not
surprising that the meaning of the word has also changed. Let us read
the abstract:
"The photon's wave function affects only one out of the atoms,
regardless of its position within the row, thereby manifesting not
only non-local but also non-sequential characteristics. It also turns
out that, out of n atoms, each one has a probability which is higher
than the normal 1/n to be the single affected one."
Why should we read any further? We are to believe that there exist n
positive numbers, each of which is greater than 1/n, but that still
sum to 1. This is indeed non-statistical and non-probabilistic, but
also non-sensical.
> Thomas, since you are from Scandinavia you should heed what the Prince
> of Denmark once said and also do not forget about ghosts :-)
Precisely, ghosts are non-physical; and the Prince of Denmark never
spoke those words, only an actor on a stage.
> http://arxiv.org/abs/http://www.arxiv.org/abs/gr-qc/0403121
Note that this last paper (which I am not endorsing!) *defends* the
notion of a casual arrow of time which distinguishes the future from
the past.