It's clear that DK's scheme won't work - nor will any scheme that is based on unitary linear orthodox quantum theory using orthogonal base states.
However, concerning Valentini's, Josephson, Weinberg, Stapp & my different & independent from from DK's approaches: while the trace operation to get expectation values of observables on quantum density matrices is invariant under unitary transformations of the base states which preserve orthogonality, that is not true for the transformation from an orthogonal Fock basis to the non-orthogonal Glauber coherent state basis, which is clearly a non-unitary transformation that is OUTSIDE the domain of validity of orthodox quantum theory. I think many Pundits have missed this point?
Hawking's former assistant Bernard Carr spells this out clearly in Can Psychical Research Bridge the Gulf Between Matter and Mind?" Bernard Carr Proceedings of the Society for Psychical Research, Vol 59 Part 221 June 2008
Begin forwarded message:
From: nick herbert <quanta@cruzio.com>
Subject: Re: AW: AW: More on the |0>|0> term
Date: June 14, 2013 11:14:57 AM PDT
To: Suda Martin <Martin.Suda.fl@ait.ac.at>
Thank you, Martin.
I finally get it.
My confusion lay in the attribution of the short calculation below.
I thought this calculation (which leads to rA) was due to Gerry.
Instead it is a calculation done by Gerry but attributed to DK.
It was not a calculation that DK ever carried out but
arose from Gerry taking Gerry's FULL CALCULATION,
applying the Kalamidas approximation
and getting an incorrect result.
The correct result is Zero
on which you and Gerry agree.
So if Kalamidas would have carried out the calculation this way
he would have gotten an incorrect answer.
I hope I have now understood the situation correctly.
But Kalamidas did not carry out the calculation that Gerry displays.
DK did not start out with the FULL CALCULATION and then approximate.
DK starts with an approximation and then calculates.
DK starts with an approximation and carries out a series of steps which all seem to be valid
but whose conclusion is preposterous. Furthermore the approximation (weak coherent states)
is an approximation used in dozens of laboratories by serious quantum opticians without
as far as I am aware leading to preposterous or impossible conclusions.
Therefore it seems to me that the calculation below is another nail in the Kalamidas coffin, BUT
THE BEAST IS STILL ALIVE.
1. No one yet has started with Kalamidas's (approximate) assumptions, and discovered a mistake in his chain of logic.
2. No one yet has started with Kalamidas's (approximate) assumptions, followed a correct chain of logic and shown that FTL signaling does not happen.
Martin Suda came the closest to carrying out problem #2. He started with the Kalamidas (approximation) assumptions and decisively proved that all FTL terms are zero. But Martin's proof contains an unphysical |0>|0> term that mars his triumph.
I am certain that the Kalamidas claim is wrong. The FULL CALCULATION refutations of Ghirardi, Howell and Gerry are pretty substantial coffin nails. But unless I am blind there seems still something missing from a clean and definitive refutation of the Kalamidas claim. See problems #1 and #2 above.
I do not think that Nick is being stubborn or petty in continuing to bring these problems to your attentions. I should think it would be a matter of professional pride to be able to bring this matter to a clean and unambiguous conclusion by refuting Kalamidas on his own terms.
Thank you all for participating in this adventure whatever your opinions.
Nick Herbert
On Jun 14, 2013, at 3:29 AM, Suda Martin wrote:
Nick,
Thank you for comments!
I would still like to explain my short considerations below a bit more precisely, anyway. I feel there was perhaps something unclear as regards my email (12th June), because you wrote "you were confused".
I only considered the following:
DK disclosed a calculation (see attachment) which is completely wrong because he made a mathematical limit (see first line, where he omitted the term ra^{+}_{a3}) which is absolutely not justifiable here (just as CG mentioned, see below) because both parts are equally important if you make the expectation value properly. If you take both parts you get exactly zero: alpha^{*}(tr^{*}+rt^{*})=0.
So one does not obtain a quantity like (r alpha)^{*}.
That’s all. There is absolutely no discrepancy between me and CG.
Nice regards,
Martin
-----Ursprüngliche Nachricht-----
Von: nick herbert [mailto:quanta@cruzio.com]
Gesendet: Mittwoch, 12. Juni 2013 23:33
Betreff: Re: AW: More on the |0>|0> term
"And again, the notion that an alleged approximate calculation (I say "alleged" because as with everything else there are correct and incorrect approximate calculations) based on a weak signal coherent state somehow trumps an exact computation valid for any value of the coherent state parameter, is, well, just insane. If you want to see where things go wrong just take more terms in the series expansions. Add up enough terms and, viola, no effect! One can't get much more specific than that." --Christopher Gerry
Actually, Chris, one can get much more specific than that by explicitly displaying the Correct Approximation Scheme (CAS) and showing term by term than Alice's interference vanishes (to the proper order of approximation).
Absent a correct CAS and its refutation these general claims are little more than handwaving.
Produce a CAS.
Refute it.
Is anyone up to this new Kalamidas challenge?
Or does everyone on this list except me
consider deriving a CAS a waste of time?
Nick Herbert
On Jun 12, 2013, at 2:03 PM, CHRISTOPHER GERRY wrote:
We are both right: the two terms cancel each other out! That the
whole expectation value is zero is actually exactly what's in our
paper's Eq. 9. This happens because the reciprocity relations must
hold. That Kalamidas thought (or maybe even still thinks) his
calculation is correct, is at the heart of the matter, that is, that
he is either unable to do the calculations or that he can do them but
chooses not too because they don't get him where he wants to go.
The Kalamidas scheme will not work not work on the basis of general
principles as we showed in the first part of our paper (see also
Ghirardi's paper).
And again, the notion that an alleged approximate calculation (I say
"alleged" because as with everything else there are correct and
incorrect approximate calculations) based on a weak signal coherent
state somehow trumps an exact computation valid for any value of the
coherent state parameter, is, well, just insane. If you want to see
where things go wrong just take more terms in the series expansions.
Add up enough terms and, viola, no effect! One can't get much more
specific than that.
Christopher C. Gerry
Professor of Physics
Lehman College
The City University of New York
718-960-8444
christopher.gerry@lehman.cuny.edu
---- Original message ----
Date: Wed, 12 Jun 2013 12:28:16 -0700
From: nick herbert <quanta@cruzio.com>
Subject: Re: AW: More on the |0>|0> term
To: Suda Martin
All--
Excuse me for being confused.
Gerry refutes Kalamidas by showing that an omitted term is large.
Suda refutes Kalamidas by showing that the same term is identically
zero.
What am I missing here?
I wish to say that I accept the general proofs. Kalamidas's scheme
will not work as claimed.
That is the bottom line. So if the general proofs say FTL will fail
for full calculation, then it will certainly fail for approximations.
The "weak coherent state" is a common approximation made in quantum
optics. And dozens of experiments have been correctly described using
this approximation. So it should be a simple matter to show if one
uses Kalamidas's approximation, that FTL terms vanish to the
appropriate level of approximation. If this did not happen we would
not be able to trust the results of approximation schemes not
involving FTL claims.
Gerry's criticism is that Kalamidas's scheme is simply WRONG--that he
has thrown away terms DK regards as small.
But in fact they are large. Therefore the scheme is flawed from the
outset.
If Gerry is correct, then it seems appropriate to ask: Is there a
CORRECT WAY of formulating the Kalamidas scheme using the "weak
coherent state" approximation, where it can be explicitly shown that
this correct scheme utterly fails?
It seems to me that there are still some loose ends in this Kalamidas
affair, if not a thorn in the side, at least an unscratched itch.
It seems to me that closure might be obtained. And the Kalamidas
affair properly put to rest if everyone can agree that 1. DK has
improperly treated his approximations; 2. Using the CORRECT
APPROXIMATION SCHEME, the scheme abjectly fails just as the exact
calculation says it must.
Why should it be so difficult to construct a correct description of
the Kalamidas proposal, with CORRECT APPROXIMATIONS, and show that it
fails to work as claimed?
AS seen from the Ghirardi review, there are really not that many
serious FTL proposals in existence. And each one teaches us
something-- mostly about some simple mistakes one should not make when thinking
about quantum systems. Since these proposals are so few, it is really
not a waste of time to consider them in great detail, so we can learn
to avoid the mistakes that sloppy thinking about QM brings about.
When Ghirardi considers the Kalamidas scheme in his review, I would
consider it less than adequate if he did not include the following
information:
1. Kalamidas's scheme is WRONG because he treats approximations
incorrectly.
2. When we treat the approximations correctly, the scheme fails, just
as the general proofs say it must.
Gerry has provided the first part of this information. What is
seriously lacking here is some smart person providing the second
part.
Nick Herbert
On Jun 12, 2013, at 8:50 AM, Suda Martin wrote:
Dear all,
Yes, if one calculates precisely the Kalamidas - expression given in
the attachment of the email of CG one obtains exactly
alpha^{*}(tr^{*}+rt^{*})=0
due to the Stokes-relation of beam splitters. No approximations are
necessary. So, I am astonished about the sloppy calculations of
Demetrios.
Cheers,
Martin
________________________________________
Von: CHRISTOPHER GERRY [CHRISTOPHER.GERRY@lehman.cuny.edu]
Betreff: Re: More on the |0>|0> term
I probably shouldn't jump in on this again, but...
I can assure you that there's no thorn in the side of the quantum
optics community concerning the scheme of Kalamidas. There are only
people doing bad calculations. Despite claims to the contrary, our
paper, as with Ghirardi's, does specifically deal with the Kalamidas
proposal. It is quite clearly the case that EXACT calculations in
the Kalamidas proposal shows that the claimed effect disappears. To
suggest that it's there in the approximate result obtained by series
expansion, and therefore must be a real effect, is simply
preposterous. All it means is that the approximation is wrong; in
this case being due to the dropping important terms.
The whole business about the |00> and whatever (the beam splitter
transformations and all that) is not the issue. I'm astonished at
how the debate on this continues. The real problem, and I cannot
emphasize it enough, is this: Kalamidas cannot do quantum optical
calculations, even simple ones and therefore nothing he does should
be taken seriously. As I've said before, his calculation of our Eq.
(9), which I have attached here, is embarrassingly wrong. It's
obvious from the expression of the expectation value in the upper
left that there has to be two terms in the result both containing
the product of r and t. But Kalamidas throws away one of the terms
which is of the same order of magnitude as the one he retains. Or
maybe he thinks that term is zero via the quantum mechanical
calculation of its expectation value, which it most certainly is
not. His limits have been taken inconsistently. So, he not only
does not know how to do the quantum mechanical calculations, he
doesn't even know how or when the limits should be taken. There's
absolutely no point in debating the meaning of the results incorrect
calculations. Of course, by incorrectly doing these things he gets
the result he wants, and then thinks it's the duty of those of us
who can do these calculations to spend time showing him why his
calculations are wrong, which he then dismisses anyway.
My point in again bringing this specific calculation of his is not
to say anything about his proposal per se, but to demonstrate the
abject incompetence of Kalamidas in trying to do even the most
elementary calculations. And if anyone still wonders why I'm angry
about the whole affair, well, what should I feel if some guy unable
to do simple calculations tries to tell established quantum optics
researchers, like me and Mark Hillery, that our paper showing where
he's wrong dismisses ours as being "irrelevant?" He doesn't even
seem to know that what he said was an insult.
And finally, the continued claim that the specific proposal of
Kalamidas has not been addressed must simply stop. It has been
repeatedly. I suspect this claim is being made because people don't
like the results of the correct calculations. That's not the problem
of those of us can carry through quantum optical calculations.
CG
Christopher C. Gerry
Professor of Physics
Lehman College
The City University of New York
718-960-8444
christopher.gerry@lehman.cuny.edu
---- Original message ----
Date: Tue, 11 Jun 2013 14:12:19 -0700
From: nick herbert <quanta@cruzio.com>
Subject: Re: More on the |0>|0> term
To: "Demetrios Kalamidas" <dakalamidas@sci.ccny.cuny.edu>
yer right, demetrios--
the |00> term on the right is always accompanied in Suda's
calculation by a real photon on the left.
But this is entirely non-physical.
No real or virtual quantum event corresponds to this term.
Especially with the high amplitude required for
Suda-interference-destruction.
So your specific approximate FTL scheme despite many general
refutations still remains a puzzlement.
A thorn in the side
of the quantum optics community.
if any think otherwise
let them put on the table
one unambiguous refutation
OF YOUR SPECIFIC PROPOSAL--
not of their own
nor of somebody else's
totally different FTL signaling scheme,
Nick
On Jun 11, 2013, at 1:27 PM, Demetrios Kalamidas wrote:
Nick,
The EP and CSs do derive from the same laser pulse: part of the
pulse pumps the nonlinear crystal and the other part is split off
accordingly to create the CSs.
However, you are still misssing the point: If no EP pair is
created, then you will certainly get '00' on the right
sometimes.... BUT there will be no left photon in existence. The
problem with the Suda term is that when it appears, it appears
only accompanied by a left photon in a superposition state: ie it
always appears as (10+e01)(00+11).
Think of it this way: Suppose you just have an EP source that
creates pairs, with one photon going left and the other right.
Imagine that on the right there is a highly trasnparent BS with
say
|r|^2=0.001. That means that only one out of every thousand right
photons from the EP are reflected, and 999 are transmitted. So,
this means that for every 1000 counts ON THE LEFT, there will be
999 counts tranmitted on the right. Now introduce, at the other
input of that same BS, a CS so that it has a tiny reflected
portion of amplitude |ralpha>. Allegedly then, there will arise
cases where no photon is found in the transmitted channel with
probability equal to |ralpha|^2. Since alpha is arbitrary, we can
choose |
ralpha|=0.1. This means that the probabilty of getting no
photon in
the transmitted channel will be |ralpha|^2=0.01.....Which now
means that, for every 1000 EP pairs created, we will get 1000
counts on the left, but only 900 counts in the transmitted channel
on the right! Whereas, without the CS in the other channel, there
would be
999 counts on the right for that same 1000 counts on the left.
Demetrios
On Tue, 11 Jun 2013 09:44:42 -0700
nick herbert <quanta@cruzio.com> wrote:
Demetrios--
I don't know how the entangled pair (EP) and CSs are generated.
I supposed all three are created with a single PULSE in a non-
linear crystal.
Now one can imagine that this pulse fails to create an EP but
does create a CS
Then some of Bob's detectors will fire but no ES is formed.
So this kind of process could lead to lots of |0>|0> terms.
However what we need are not "lots of |0>|0> terms" but a precise
amplitude (rA) of |0>|0> term.
Given our freedom (in the thought experiment world) to
arbitrarily select
the efficiency of the non-linear crystal, it is hard to see why
the elusive |0>|0>
term would have exactly the right magnitude and phase to cancel
out the interference.
Your original FTL scheme still continues to puzzle me.
Nick
On Jun 11, 2013, at 6:54 AM, Demetrios Kalamidas wrote:
Nick,
The 'entire experimental arrangement' is indeed where the
problem (mystery) arises:
When both CSs are generated it is easy to understand that '00'
will arise, simply because each CS has a non-zero vacuum term.
However, the entire arrangement means inclusion of the
entangled photon pair:
Any time that pair is generated, you are guaranteed to get a
photon on the right, regardless of whether the CSs are there.
So, when entangled pair and CSs are present, there must be at
least one photon at the right. In fact, when only one photon
emerges at the right WE KNOW both CSs were empty.
On Mon, 10 Jun 2013 10:34:30 -0700
nick herbert <quanta@cruzio.com> wrote:
Demetrios--
Sarfatti sent around a nice review of quantum optics
by Ulf Leonhardt that discusses the structure of path-uncertain
photons.
Here is an excerpt:
The interference experiments with single photons mentioned in
Sec. 4.3 have been
performed with photon pairs generated in spontaneous
parametric downconversion
[127]. Here the quantum state (6.28) of light is essentially
|01> |02> + ζ |11>|12 >. (6.29)
In such experiments only those experimental runs count where
photons are counted,
the time when the detectors are not firing is ignored, which
reduces the quantum
state to the photon pair
|11> |12> .
Postselection disentangles the two-mode squeezed
vacuum.
We argued in Sec. 4.3 that the interference of the photon pair
|11> |12> at a 50:50 beam splitter generates the entangled
state (4.24). Without postselection,
however, this state is the disentangled product of two single-
mode squeezed vacua,
as we see from the factorization (6.6) of the S matrix. The
notion of entanglement
is to some extent relative.
this excerpt suggests a possible origin for Suda's |0>|0> term.
In the above process, it's just
the inefficiency of the down converter that generates a |0>|0>
term. That won't do the trick.
But in your more complicated situation--containing two properly
timed coherent states--
when Bohr's "entire experimental arrangement" is considered,
the
| 0>| 0> term may
arise naturally with the proper amplitude and phase. It would
correspond to events when
the coherent states were successfully generated but there were
no events in either upper or lower path.
If this conjecture can be shown to hold true, then the
original Kalamidas proposal would
be refuted by Suda's calculation.
The trick would be to examine--in a thought experiment way--
exactly how those two |A> beams
are created--looking for entanglement with |0>|0> states in
the part of the experiment considered in your proposal.
Nick
ref: Ulf Leonhardt's wonderful review of quantum optics,
starting with reflections from a window pane and concluding
with
Hawking radiation.
OK, here is a simple case - not same as Kalamidas mind you - that seems to be outside the rules of orthodox quantum theory.
Alice the receiver has an ordinary orthodox quantum bit with base states |0> & |1> for a given orientation of her apparatus which never changes in the experiment. Bob the sender has two distinguishable non-orthogonal Glauber coherent eigenstates |z> and |w> of the non-Hermitian observable boson destruction operator a, where z and w are complex numbers. Right at this point we have violated one of the axioms of orthodox quantum theory in a factual way since Glauber states are facts.
Suppose we have the entangled state
|A,B> = (1/2)^1/2[|0>|z> + |1>|w>]
then using the orthodox Born probability rule in density matrix formulation gives
p(0) = p(1) = (1/2)[1 + |<z|w>|^2]
p(0) + p(1) = 1 + |<z|w>|^2 > 1
the entanglement signal at Alice's receiver is |<z|w>|^2
violating conservation of Born's rule for probability - because the observable is not hermitian and actually a closer examination shows a non-unitary time evolution. This is a larger theory that reduces to orthodox quantum theory in the appropriate limit.
note
http://en.wikipedia.org/wiki/Coherent_states
Now, we can squirm out of this by a-priori ad-hoc forcing of the non-universal normalization
|A,B>' = [1 + |<z|w>|^2]^-1/2|A,B>
giving
p'(0) = p'(1) = 1/2 with no signaling
Note, that Bob does not need to use that normalization at all because of Alice's <0|1> = 0.
That's why I use "non-universal" above.
However, it's not clear the Nature works this way without more testing.
On Jun 1, 2013, at 1:04 PM, Ghirardi Giancarlo <ghirardi@ictp.it> wrote:
Il giorno 01/giu/2013, alle ore 18:38, JACK SARFATTI <adastra1@me.com> ha scritto:
Ghirardi: I do not agree at all on this. The actual situation is that there has never been a clear cut indication that in Kalamidas serf-up something (probabilities, outcomes or whatever you want) actually changes something at left as a consequence of preparing one or the other state at right, so that it can be used to send faster than light signals. It is his duty and not ours to prove that the effect exist. I believe to have argued against its existence and I have also checked that for the most natural observables at left no difference occurs when you choose one or the other of the two initial states. The game is back to Kalamidas. And, sincerely, I am a little bit disturbed by all this enormous mess and many inadequate and unjustified statements that have been put forward during the debate. I am not keen to follow the matter any more.
On Jun 1, 2013, at 1:54 PM, Suda Martin <Martin.Suda.fl@ait.ac.at> wrote:
Dear all,
thanks to everybody for emails, papers, contributions to discussion and comments. I enjoyed very much the highly interesting dialogues. I can fully agree to the arguments of CG and GG, of course.
Only a comment with respect to the question of the approximation:
As regards the approximation done in the calculation of DK, I would like to point out again - and I sent a pdf called Interf_BS_50_50_Suda.pdf two days ago - that because of such an approach the normalization of the output wave function behind the 50/50 BS has been changed to (1+2|alpha|^2+|alpha|^4), see Eq.(7), instead of being exactly 1. The probabilities for the potential "interference part" (see Eq.(6)) are (|p_10|^2+|p_01|^2)/4=2|alpha|^2 and the other parts give all together 2(|q_10|^2+|q_01|^2)/4=1+|alpha|^4. One keeps therefore precisely the modified normalization of Eq.(7).
One can clearly see that the "interference part" and the other parts are outcomes from an incorrect normalization.
Nice regards,
Martin
Begin forwarded message:
From: CHRISTOPHER GERRY <CHRISTOPHER.GERRY@lehman.cuny.edu>
Subject: Re: The Kalamidas affair
Date: June 1, 2013 9:46:37 AM PDT
To: nick herbert <quanta@cruzio.com>
Cc: Ghirardi Giancarlo <ghirardi@ictp.it>, Demetrios Kalamidas <dakalamidas@sci.ccny.cuny.edu>, John Howell <howell@pas.rochester.edu>, Suda Martin <martin.suda.fl@ait.ac.at>, Ruth Kastner <rekastner@hotmail.com>, JACK SARFATTI <adastra1@me.com>, "Romano rromano@iastate.edu [MATH]" <rromano@iastate.edu>
Nick and everyone,
The specific failings of the Kalamidas proposal have, in fact, been pointed out in the papers you mentioned and elsewhere. I don't understand why anyone continues to say otherwise. To say that they have not been addressed does not make it so, and comes off merely an act of denial. This has been an interesting episode, but I think it's time to stop beating a dead horse. Chris
On Jun 1, 2013, at 9:13 AM, nick herbert <quanta@cruzio.com> wrote:
Kalamidas fans--
NH: I believe that everyone is in agreement that general considerations prove that the Kalamidas proposal must fail.
JS: Yes
In both Ghirardi's and Gerry's papers, they emphasize these general considerations and decline to engage in the specifics of Kalamidas's calculations. Whether one wishes to engage the specifics or not is a matter of taste. But Kalamidas is asking us to engage in specifics. As he puts it: Since you know that I am wrong, it should be "easy pickins" to
point out exactly where I am mistaken.
Gerry comes closest to meeting Kalamidas's challenge to move out of the safety of generalities and deal with specifics.
In the conclusion of Gerry's paper he states "Clearly, if the exact calculation shows no interference, but the approximate calculation does, there is something wrong with the approximate calculation. Looking at Eq 6, one notes that while some terms to order rA have been kept in going from 6a to 6c, the terms labeled "vanishing" in Eq 6b are also of this order and have been discarded. Thus the approximate calculation in {1} is inconsistent and wrong."
Gerry engages in specifics. He is meeting Kalamidas on his own terms. But he neglects to specify exactly which terms of order rA Kalamidas has mistakenly labeled as "vanishing". When Gerry displays these wrongly-neglected terms (perhaps in an informal note), he would have definitively "slain the beast in his own lair" and we can all get on with the non-Kalamidas aspects of our lives.
JS: Agreed, thanks Nick :-)
Nick
PS: There is still the fascinating Martin Suda Paradox which was discovered in the context of the Kalamidas refutation, but that is a separate issue altogether.
JS: What is that Nick? Please give details.
Begin forwarded message:
From: JACK SARFATTI <adastra1@me.com>
Subject: [ExoticPhysics] Fwd: The Kalamidas affair
Date: June 1, 2013 7:45:42 AM PDT
To: Exotic Physics <exoticphysics@mail.softcafe.net>
Reply-To: Jack Sarfatti's Workshop in Advanced Physics <exoticphysics@mail.softcafe.net>
Sent from my iPad
Subject: Re: The Kalamidas affair
yes I agree with this
any attempt at signaling within axioms of orthodox quantum theory will fail e.g. Adrian Kent's papers
however, antony valentini, myself and others (Stapp, Weinberg, Josephson) have all independently proposed several extensions giving a more general non-orthodox post quantum theory containing orthodox quantum theory as a limiting case. In particular, the non-hermitian boson destruction operator is a macroscopic observable with Glauber coherent eigenstates that are non-orthogonal distinguishable violating orthodox quantum theory. Furthermore, they obey a non-unitary dynamics given by the c-number landau-ginzburg equation for spontaneous broken symmetry ground/vacuum state emergent local order parameters. These order parameters entangle with others and also with orthodox qubits, so we have a new larger theory here analogous to general relativity in relation to special relativity.
Furthermore, there is no violation with the group structure of relativity because intervals are frame invariant and what matters is the interval between actual irreversible detections. What is violated is the retarded casuality axiom appended to relativity that is adhoc like Euclid's fifth axiom. Again the analogy to non-Euclidean geometry is appropriate.
Sent from my iPad
On Jun 1, 2013, at 6:40 AM, CHRISTOPHER GERRY <CHRISTOPHER.GERRY@lehman.cuny.edu> wrote:
Everyone,
I'm in total agreement with Prof. Ghirardi's assessment. The beam splitter transformations are not the essential point here, as even if the are done correctly, the claimed effect goes away. We addressed the beam splitter issue in our comment to demonstrate that sloppy calculations in general are contained in the Kalamidas paper. We then assumed that the one case of his t and r of parameters that would satisfy the reciprocity relations actually held, thus ensuring that his transformations did not violate unitarity (for that one case!) and from there showed via an exact calculation that the effect disappears. As I said, it will disappear even with totally correct, unitary beam splitter transformations, just as stated by Prof. Ghirardi. Chris
Christopher C. Gerry
Professor of Physics
Lehman College
The City University of New York
718-960-8444
christopher.gerry@lehman.cuny.edu
---- Original message ----
Date: Sat, 1 Jun 2013 14:57:07 +0200
From: Ghirardi Giancarlo <ghirardi@ictp.it> Subject: The Kalamidas affair To: CHRISTOPHER GERRY <christopher.gerry@lehman.cuny.edu>, Demetrios Kalamidas <dakalamidas@sci.ccny.cuny.edu>, John Howell <howell@pas.rochester.edu>, nick herbert <quanta@cruzio.com>, Suda Martin <martin.suda.fl@ait.ac.at>, Ruth Kastner <rekastner@hotmail.com>, JACK SARFATTI <adastra1@me.com>, "Romano rromano@iastate.edu [MATH]" <rromano@iastate.edu>
Dear all,
attached herewith you will find a letter (even though it looks like a paper for technical reasons) that I have decided to forward to you to make clear the conceptual status of the situation. I hope of having been clear and I wait for comments.
With my best regards
GianCarlo
________________
remarks.pdf (83k bytes)
________________
_______________________________________________
ExoticPhysics mailing list
ExoticPhysics@mail.softcafe.net
http://mail.softcafe.net/cgi-bin/mailman/listinfo/exoticphysics
Jack Sarfatti This is hot. If the effect works it's the basis for a new Intel, Microsoft & Apple combined for those smart venture capitalists, physicists & engineers who get into it. This is as close as we have ever come since I started the ball rolling at Brandeis in 1960-61 & then in mid-70's see MIT Physics Professor David Kaiser's "How the Hippies Save Physics". I first saw this as a dim possibility in 1960 at Brandeis grad school and got into an intellectual fight about it with Sylvan Schweber and Stanley Deser. Then the flawed thought experiment published in the early editions of Gary Zukav's Dancing Wu Li Masters in 1979 - pictured in Hippies book tried to do what DK may now have actually done. That is, control the fringe visibility at one end of an entangled system from the other end without the need of a coincidence counter correlator after the fact. Of course, like Nick Herbert's FLASH at the same time late 70's, it was too naive to work and the nonlinear optics technology was not yet developed enough. We were far ahead of the curve as to the conceptual possibility of nonlocal retrocausal entanglement signaling starting 53 years ago at Brandeis when I was a National Defense Fellow Title IV graduate student.
Jack Sarfatti
about an hour ago near San Francisco
On Feb 5, 2013, at 12:28 PM, JACK SARFATTI <sarfatti@pacbell.net> wrote:
Thanks Nick. Keep up the good work. I hope to catch up with you on this soon. This may be a historic event of the first magnitude if the Fat Lady really sings this time and shatters the crystal goblet. On the Dark Side this may open Pandora's Box into a P.K. Dick Robert Anton Wilson reality with controllable delayed choice precognition technology. ;-)
On Feb 5, 2013, at 10:38 AM, nick herbert <quanta@cruzio.com> wrote:
Demetrios--
Looking over your wonderful paper I have detected one
inconsistency but it is not fatal to your argument.
On page 3 you drop two r terms because "alpha", the complex
amplitude of the coherent state can be arbitrarily large in
magnitude.
But on page 4 you reduce the magnitude of "alpha" so that
at most one photon is reflected. So now alpha cannot be
arbitrarily large in magnitude.
But this is just minor quibble in an otherwise superb argument.
This move does not affect your conclusion--which seems
to directly follow from application of the Feynman Rule: For distinguishable
outcomes, add probabilities; for indistinguishable outcomes, add amplitudes.
To help my own understanding of how your scheme works,
I have simplified your KISS proposal by replacing your coherent states with
the much simpler state |U> = x|0> + y|1>. I call this variation of your proposal KISS(U)
When this state |U> is mixed with the entangled states at the beamsplitters,
the same conclusion ensues: there are two |1>|1> results on Bob's side of the source
that cannot be distinguished -- and hence must be amplitude added.
The state |U> would be more difficult to prepare in the lab than a weak coherent state
but anything goes in a thought experiment. The main advantage of using state |U>
instead of coherent states is that the argument is simplified to its essence and needs
no approximations. Also the KISS(U) version shows that your argument is independent
of special properties possessed by coherent states such as overcompleteness and non-
orthogonality. The state |U> is both complete and orthogonal -- and works just as well
to prove your preposterous conclusion. --- that there is at least one way of making photon
measurements that violates the No-Signaling Theorem.
Thanks for injecting some fresh excitement into the FTL signaling conversation.
warm regards
Nick Herbert
Like · · Share
David Fernando López Torres, Keith Kenemer and 2 others like this.
View 1 more comment
Jack Sarfatti On Feb 5, 2013, at 1:15 PM, Demetrios Kalamidas <dakalamidas@sci.ccny.cuny.edu> wrote:
Nope, no refutation I can think of so far....and I've tried hard.
Demetrios
...See More
33 minutes ago · Like
Joe Ganser Jack do you know a lot of people at CUNY? I take ph.d classes there.
26 minutes ago · Like
Joe Ganser I'm interested in who may do these sorts of topics in NYC
25 minutes ago · Like
Jack Sarfatti Daniel Greenberger!
9 minutes ago · Like · 1
a few seconds ago · Like
On Feb 5, 2013, at 1:15 PM, Demetrios Kalamidas <dakalamidas@sci.ccny.cuny.edu> wrote:
Nope, no refutation I can think of so far....and I've tried hard.
Demetrios
On Tue, 5 Feb 2013 13:09:28 -0800
nick herbert <quanta@cruzio.com> wrote:
Thanks, Demetrios. I understand now that alpha can be large
while alpha x r is made small. Also I notice that your FTL signaling scheme seems to work both ways. In your illustration the photons on the left side (Alice) are combined at a 50/50 beam splitter so they cannot be used for which-way information. However if the 50/50 beamsplitter is removed, which-way info is present and the two versions of |1>|1> on the right-hand side (Bob) are now distinguishable
and must be added incoherently, which presumably will give a different answer and observably different behavior by Bob's right-side detectors. So your scheme seems consistent -- FTL signals can be sent in either direction.
This is looking pretty scary.
Do you happen to have a refutation up your sleeve
or are you just as baffled by this as the rest of us?
Nick
Therefore, Nick it is premature for you to claim that the full machinery of the Glauber coherent states, i.e. distinguishable over-complete non-orthogonality is not necessary for KISS to work. Let's not rush to judgement and proceed with caution. This technology, if it were to work is as momentous as the discovery of fire, the wheel, movable type, calculus, the steam engine, electricity, relativity, nuclear fission & fusion, Turing machine & Von Neumann's programmable computer concept, DNA, transistor, internet ...
On Feb 5, 2013, at 12:18 PM, Demetrios Kalamidas <dakalamidas@sci.ccny.cuny.edu> wrote:
Hi Nick,
And thanks much for your careful examination of my scheme....however, there appears to be a misunderstanding.
Let me explain:
"On page 3 you drop two r terms because "alpha", the complex amplitude of the coherent state can be arbitrarily large in magnitude."
I drop the two terms in eq.5b because they are proportional to 'r'....and 'r' approaches zero. However, the INITIAL INPUT amplitude, 'alpha', of each coherent state can be as large as we desire in order to get whatever SMALL BUT NONVANISHING AND SIGNIFICANT product 'r*alpha', which is related to the terms I retain.
In other words, for whatever 'r*alpha' we want, lets say 'r*alpha'=0.2, 'r' can be as close to zero as we want since we can always input a coherent state with large enough initial 'alpha' to give us the 0.2 amplitude that we want.
So, terms proportional to 'r' are vanishing, while terms proportional to 'r*alpha' are small but significant and observable.
You state:
"But on page 4 you reduce the magnitude of "alpha" so that at most one photon is reflected. So now alpha cannot be arbitrarily large in magnitude."
The magnitude of 'alpha' is for the INITIAL coherent states coming from a3 and b3, BEFORE they are split at BSa and BSb. It is this 'alpha' that is pre-adjusted, according to how small 'r' is, to give us an appropriately small reflected magnitude, i.e. 'r*alpha'=0.2, so that the "....weak coherent state containing at most one photon...." condition is reasonably valid.
Demetrios
On Feb 5, 2013, at 12:28 PM, JACK SARFATTI <sarfatti@pacbell.net> wrote:
Thanks Nick. Keep up the good work. I hope to catch up with you on this soon. This may be a historic event of the first magnitude if the Fat Lady really sings this time and shatters the crystal goblet. On the Dark Side this may open Pandora's Box into a P.K. Dick Robert Anton Wilson reality with controllable delayed choice precognition technology. ;-)
On Feb 5, 2013, at 10:38 AM, nick herbert <quanta@cruzio.com> wrote:
Demetrios--
Looking over your wonderful paper I have detected one
inconsistency but it is not fatal to your argument.
On page 3 you drop two r terms because "alpha", the complex
amplitude of the coherent state can be arbitrarily large in
magnitude.
But on page 4 you reduce the magnitude of "alpha" so that
at most one photon is reflected. So now alpha cannot be
arbitrarily large in magnitude.
But this is just minor quibble in an otherwise superb argument.
This move does not affect your conclusion--which seems
to directly follow from application of the Feynman Rule: For distinguishable
outcomes, add probabilities; for indistinguishable outcomes, add amplitudes.
To help my own understanding of how your scheme works,
I have simplified your KISS proposal by replacing your coherent states with
the much simpler state |U> = x|0> + y|1>. I call this variation of your proposal KISS(U)
When this state |U> is mixed with the entangled states at the beamsplitters,
the same conclusion ensues: there are two |1>|1> results on Bob's side of the source
that cannot be distinguished -- and hence must be amplitude added.
The state |U> would be more difficult to prepare in the lab than a weak coherent state
but anything goes in a thought experiment. The main advantage of using state |U>
instead of coherent states is that the argument is simplified to its essence and needs
no approximations. Also the KISS(U) version shows that your argument is independent
of special properties possessed by coherent states such as overcompleteness and non-
orthogonality. The state |U> is both complete and orthogonal -- and works just as well
to prove your preposterous conclusion. --- that there is at least one way of making photon
measurements that violates the No-Signaling Theorem.
Thanks for injecting some fresh excitement into the FTL signaling conversation.
warm regards
Nick Herbert