You are on page 1of 85

Causality, Locality and Joint

Probabilities in Quantum
Mechanics
S. M. Roy, HBCSE, TIFR, Mumbai
HBCSE,TIFR, 10 June 2014
From paradoxes to useful quantum ef-
fects . The EPR paradox, Schrodingers
Cat paradox and Zeno paradox have long
focussed attention on counter-intuitive fea-
tures of quantum foundations: superposi-
tion principle ,entanglement and measure-
ment theory. These striking features are now
elevated from paradoxes to useful quan-
tum eects, and account for the striking
gains of quantum computation with respect
to classical computation. I survey quantum
foundational issues illuminated by the EPR
paradox and some applications of the as-
sociated non-classical features, e.g. quan-
tum violations of Bell inequalities following
from Einsteins local reality principle and of
the much more stringent inequalities follow-
ing from quantum separability or absence of
quantum entanglement. The Bell-CHSH in-
equalities and their multiparticle generaliza-
tions bring out violation of Einstein local re-
ality principle by a factor 2
(N1)/2
by certain
entangled states of N particles. For quan-
tum computation, the more relevant prop-
erty is entanglement or its opposite, viz.
separability. Quantum separability implies
inequalities exponentially stronger than Bell
inequalities for large N (S. M. Roy). They
are violated by a factor 2
N1
by some en-
tangled states; these states are natural can-
didates to exploit in seeking improvements
over classical computation. They also enable
exponentially enhanced accuracy in measur-
ing physical quantities such as time , and
interaction strengths (Roy-Braunstein).
Introduction: Quantum Versus Classical
World View The beautiful and the grotesque
in QM:
States Rays in Hilbert Space;
Measurement/Collapse Postulate.
Bells theorem [0] demonstrating that quan-
tum mechanics violates Einsteins principle
of local reality [0] is sometimes regarded as
the most fundamental discovery of the 20th
century [0].This theorem underlines the ba-
sic dierences between the classical and quan-
tum world views.In 1989, the year Bell de-
parted,he gave three lectures Against Mea-
surement at the Erice conference on Six-
tytwo Years of Uncertainty which will re-
main an inspiration for those who dare to
question the magnicent edice of QM.
In the classical, i.e. relativistic but pre-quantum
world view the instantaneous state of a sys-
tem is specied by listing the phase space co-
ordinates: the positions (q
i
[i = 1, 2, ..N) and
the momenta (p
i
[i = 1, 2, ..N) of the particles
and the amplitudes of waves (such as elec-
tromagnetic waves) and their time deriva-
tives at all space points.Here the dierent
indices i can specify dierent particles or dif-
ferent spatial components for the same parti-
cle.The time evolution is then specied uniquely
as trajectories in phase space in terms of
the laws of motion (e.g. relativistic Newtons
laws for the particles and relativistic wave
equations ). Classical mechanics has the fol-
lowing fundamental characteristics:
Objective Reality .
The phase space co-ordinates are objective
properties of the system independent of their
observation.
Causality/ Determinacy.
Dierent results (viz. phase space co-ordinates)
arise from dierent causes (Jauchs formula-
tion of Causality).
Locality .
The phase space co-ordinates are indepen-
dent of, i.e. not inuenced by, actions per-
formed at a spacelike separation.
In the quantum worldview a unied descrip-
tion of waves and particles is achieved . The
particular disposition of experimental appa-
ratus can bring out the particle (photon) na-
ture of electromagnetic waves (as in photo-
electric eect), or the wave nature of parti-
cles (as in neutron interference experiments).
The wave and particle attributes which can-
not be revealed simultaneously may be called
complementary (Bohrs complementarity prin-
ciple). This unication is achieved at an
enormous philosophical cost, e.g. loss of
causality or determinism, loss of a unied
description of the world due to its arbitrary
division into quantum object (or system) and
classical subject (or observer/apparatus), and
as shown by Einstein and Bell, a consequent
loss of local reality.
John Bell gave a seminar at TIFR in 1980
October which brings out the surprise of quan-
tum mechanics vis a vis classical mechanics
beautifully. To make it more accessible, I
quote the rst transparency of his seminar
entitled, What in the world is quantum me-
chanics about exactly ? :
QM is about: Wave functions , Opera-
tors H, Schr odinger Eqn,
i/t = H (1)
and how to solve it. But what in the world
? The wave function is not like the world:
= + (2)
would be unrecognizable. So: Statistical In-
terpretation. Statistics of what? measure-
ment results.
A quantum mechanics based on the statis-
tics of measurement results is fundamentally
ambiguous because it requires an unden-
able boundary between the quantum object
and the classical measuring apparatus. No-
body knows what quantum mechanics says
exactly about any situation. For nobody
knows where the boundary really is, between
wavy quantum system and the world of par-
ticular events. This is the problem of quan-
tum mechanics. It is no problem in practice-
because practice is not accurate enough-and
may be never will be. Bell considered this
fundamental drawback far more serious than
the loss of determinism.
Einstein in 1933 [0] expressed dissatisfaction
with quantum theory being merely a set of
rules about the statistics of measurement re-
sults :I still believe in the possibility of giv-
ing a model of reality which shall represent
events themselves and not merely the prob-
ability of their occurence, and more speci-
cally, I am, in fact, rather rmly convinced
that the essentially statistical character of
contemporary quantum theory is solely to
be ascribed to the fact that this (theory)
operates with an incomplete description of
physical systems .
Nonexistence of History in
Ordinary Quantum Mechanics. At the
1927 Solvay conference (the year of the un-
certainty principle), Einstein discussed the
example of a particle passing through a nar-
row hole on to a hemispherical uorescent
screen which records the arrival of the parti-
cle.
Solvay Conference 1927
Suppose that a scintillation is seen at a point
P at time t = T, and suppose that the hole is
so narrow that the wave packet correspond-
ing to the particle is uniformly spread all over
the screen at t slightly less than T. Was the
particle somewhere near P at t = T (
small)? Ordinary quantum mechanics says
that the probabilities at t = T for the par-
ticle being anywhere on the screen are uni-
form (and not particularly large in the vicinity
of P). Thus the naive history corresponding
to the idea of a particle with a trajectory
(any trajectory) is denied.
There have been attempts to dene con-
sistent histories in quantum mechanics of
open systems
1
. Apart from detailed features
found unattractive by some
2
, there is the
basic proposition by the authors themselves
that only very special sets of histories are
consistent, and only these can be assigned
probabilities. For example, in a double slit
interference experiment we cannot assign a
probability that the particle reached a region
of the screen having earlier passed through
slit 1 (except in the case of vanishing inter-
ference).
Lack of Causality in Ordinary
Quantum Mechanics. One of the deni-
tions of causality (e.g. that advocated by
Jauch) is that Dierent results should have
dierent causes. Consider a quantum su-
perposition [z+)+[z) for a spin-1/2 par-
ticle, where [z+) and [z) are eigenstates of

z
with eigenvalues +1 and -1 respectively.
When the same state is prepared repeatedly
and passed through a Stern-Gerlach appara-
tus to measure
z
,
[z+) +[z) STERN GERLACH
(3)
quantum mechanics cannot predict whether
the result + or the result will occur in
a given trial; it merely says that a fraction
[[
2
of the particles ends up at the detec-
tor corresponding to
z
= +1, and a frac-
tion [[
2
goes to the detector corresponding
to
z
= 1. Thus dierent results (going
to one detector or the other) arise from ex-
actly the same cause (the same initial state).
Of course this lack of causality might be re-
stored in a theory in which the wave function
is not a complete description of the state of
the system (ref. Einstein quote).
Schr odingers Cat Paradox. Due to the
linearity of the Schrodinger eqn, the joint
state of the particle and the detectors D
+
, D

is transformed as,
([z+) +[z))D
+
D


[z+)D

+
D

+[z)D
+
D

, (4)
which is a superposition of macroscopically
dierent states D

+
D

and D
+
D

with D

denoting excited states of the detectors.


(Schrodingers Cat Paradox: if a cat located
at D
+
is killed by the arriving particle, we
have a superposition of dead and alive cats).
Context Dependence of Quantum Real-
ity. In quantum mechanics,
[(x, t)[
2
dx
is the probability of observing position to
be in dx if position were measured. It is not
the probability of position being in dx in-
dependent of observation. In fact, the same
state vector also yields
[

( p, t)[
2
d p
which is the probability of observing momen-
tum to be in the interval d p if momentum
were to be measured. Accurate simultane-
ous measurement of position and momen-
tum is not possible; if it were possible, it
would collapse the state vector into a si-
multaneous eigenstate of position and mo-
mentum which does not exist. Thus, quan-
tum mechanics does not give probabilities
for position and momentum in the same ex-
perimental situation or context. Moreover,
quantum mechanics cannot be embedded in
a more complete (hidden variable) theory in
which reality is context independent
3
.
John S. Bell
Einsteins Principle of Local Reality and
the EPR Paradox Bohr asserted that quan-
tum reality of a system is contingent upon
the disposition of the apparatus in the entire
world, not just in the nearby part of it. This
was stated most forcefully by Bohr in the
now famous Bohr-Einstein debates on quan-
tum mechanics [0]. In conict with Bohrs
assertion, Einstein proposed a principle, now
called the Principle of local reality. This prin-
ciple led to the EPR paradox, Bell inequal-
ities and experiments which proved that lo-
cal reality is violated in nature. It continues
however to provide insights into the nature
of quantum entanglement, and via Bell in-
equalities, a quantitative measure of entan-
glement.
It is convenient to consider Einsteins local
reality principle as a combination of the de-
nitions of locality and reality spelt out be-
low.
Einstein Locality. Suppose two systems S
1
and S
2
which have interacted in the past
have now separated and are experimented
upon by two observers in spatially separated
regions. Observed properties of the two sub-
systems can ofcourse be correlated due to
past interactions. Einstein insisted, But on
one supposition we should,in my opinion, ab-
solutely hold fast: the real factual situation
of the system S
2
is independent of what is
done with the system S
1
, which is spatially
separated from the former.
Physical Reality. The meaning of real fac-
tual situation in Einsteins formulation was
claried by the denition of physical reality
in the landmark paper of Einstein, Podol-
sky and Rosen [0]: If without in any way
disturbing a system, we can predict with cer-
tainty (i.e. with probability equal to unity)
the value of a physical quantity, then there
exists an element of physical reality corre-
sponding to this physical quantity.
Einstein, Podolsky and Rosen applied the
principle of local reality to argue that Quan-
tum Mechanics is incomplete. Consider a
two particle system. In quantum mechanics,
the position observable q
i
and the momen-
tum observable p
i
cannot be simultaneously
specied sharply; but the observables q
1
q
2
and p
1
+p
2
are commuting observables and
there is a quantum state
[q
1
q
2
= q
0
)[p
1
+p
2
= p
0
)
in which they are specied arbitrarily sharply,
and have values q
0
and p
0
. Ignore momen-
tarily the diculty that such states are not
normalizable, a diculty removed later by
Bohm and Aharonov [0] by considering spin
observables. Suppose observers A and B
are spacelike separated. In such a state,
if B chooses to measure q
2
she predicts q
1
with certainty (without disturbing that parti-
cle since it is spatially separated), and hence
q
1
must have physical reality; equally, if she
chooses to measure p
2
she predicts p(1) to
have reality. By the principle of local reality,
reality for particle 1 must be independent of
choices made by observer B. Hence, both q
1
and p
1
must have physical reality for particle
1. No quantum state allows simultaneously
sharp q
1
and p
1
. Thus, EPR conclude that
quantum theory must be incomplete. This
result, usually referred to as EPR paradox, is
more appropriately the EPR theorem follow-
ing from the local reality hypothesis. EPR
envisaged that an extension (or completion
) of quantum mechanics agreeing with local
reality would be possible. Bells theorem
3
shattered this hope by constructing a math-
ematical formulation of the EPR ideas and
demonstrating that quantum correlations vi-
olate the local reality principle.
Bohm-Aharonov EPR Experiment, Lo-
cal Hidden Variables and Bells Theorem
Bell formulated the EPR local reality idea us-
ing the Bohm-Aharonov [0] example of two
spin-half particles or two qubits (quantum
bits) prepared in a singlet state
[) = [ )/

2 = [ _ _)/

2
(5)
at a source S and then ying apart, to be de-
tected by two spacelike separated observers,
each equipped independently, (e.g. with ro-
tatable Stern-Gerlach magnets) , to measure
any arbitrarily chosen component of the par-
ticle spin.
EPR-Bohm-Aharonov Experiment
Since the singlet state is rotationally sym-
metric, the spin components measured along
any direction by the two observers must be
opposite. Observer 1 can predict with cer-
tainty the result of measurement of any com-
ponent of
(2)
.a by observer 2 by previously
measuring the same component
(1)
.a for
particle 1.Since Einstein locality implies that
the choice of magnet orientation made by
the remote observer 1 does not aect the re-
sult obtained by observer 2, the result of any
such observation must be predetermined.The
initial quantum wave function does not spec-
ify the result of an individual measurement;therefore
the predetermination requires a more com-
plete specication of the state , say by adding
additional variables (hidden variables) .
EPR Experiment: Perfect Anticorrelation
Hidden Variables
Hidden variables achieving perfect anti-correlation
(in each individual shot) between + and
results along the same direction for two discs
shot o along opposite directions is easy to
visualize classically. We may even allow prob-
abilistic rather than deterministic hidden vari-
ables. Suppose the hidden variables , with
probability distribution () ,determine the
probability p
r
1
(, a) of observing the value r =
1 of the observable A(a) =
(1)
.a and the
probability p
s
2
(,

b) of observing the value s =


1 of the observable B(b) =
(2)
.

b. Einstein
locality implies that p
r
1
(, a) and p
s
2
(,

b) are
independent of the orientations of the re-
mote measuring apparatus.Then the prob-
ability p
r,s
(a, b) of observing A(a) = r and
B(b) = s is,
p
r,s
(a, b) =

d()p
r
1
(, a)p
s
2
(,

b) (6)
What would the predictions of such a lo-
cal hidden variable theory (LHV) be in
the context of Einstein-Podolsky-Rosen type
measurements of P
QM
(a, b) =<
1
a
2

b >
,when we consider four possible measure-
ments, with two orientations a, a
t
on one side
and two orientations b, b
t
on the other side ?
Wigners proof of Bells Theorem. No-
tice that Bells hypothesis actually allows the
construction of a joint probability
p
r,r
t
,s,s
t
(a, a
t
, b, b
t
) =

d()p
r
1
(, a)p
s
2
(,

b)
p
r
t
1
(, a
t
)p
s
t
2
(,

b
t
)(7)
for A(a), A(a
t
), B(b), B(b
t
) to have the values
r, r
t
, s, s
t
respectively, such that the result of
any of the four feasible experiments can be
obtained as a marginal. E.g.
p
r,s
(a, b) =
r
t
=1,s
t
=1
p
r,r
t
,s,s
t
(a, a
t
, b, b
t
).
(8)
The experimental correlation A(a)B(b)) is
then,
P(a, b) =
r=1,s=1
rsp
r,s
(a, b)
=
r=1,s=1,r
t
=1,s
t
=1
rsp
r,r
t
,s,s
t
(a, a
t
, b, b
t
),
with similar expressions for P(a
t
, b), P(a, b
t
), P(a
t
, b
t
).
Hence, using the positivity of probabilities,
[P(a, b) P(a, b
t
)[
= [
r,s,r
t
,s
tr(s s
t
)p
r,r
t
,s,s
t
(a, a
t
, b, b
t
)[

r,s,r
t
,s
t[r(s s
t
)[p
r,r
t
,s,s
t
(a, a
t
, b, b
t
),
and
[P(a, b) P(a, b
t
)[ +[P(a
t
, b) +P(a
t
, b
t
)[

r,s,r
t
,s
t

[r(s s
t
)[ +[r
t
(s +s
t
)[

p
r,r
t
,s,s
t
(a, a
t
, b, b
t
)
= 2 (9)
which is the Bell-CHSH (Clauser-Horne-Shimony-
Holt) local reality inequality. Here we used
[r(s s
t
)[ +[r
t
(s + s
t
)[ = 2 and the require-
ment that the joint probability summed over
all values of r, s, r
t
, s
t
must be unity.
In contrast, Quantum Mechanics gives,
P
QM
(a, b) = a.

b. (10)
The choice of coplanar vectors such that
a.

b = a.

b
t
=

a
t
.

b =

a
t
.

b
t
= 1/

2 yields the
value 2

2 for the left-hand side of the Bell-


CHSH inequality in violation of local reality!
It shows that even if the measurements of

1
a and
2

b are made at spacelike separa-


tion, statistical predictions of quantum me-
chanics are inconsistent with the assumption
that the measured value of
1
a has a reality
that is independent of whether
2

b or
2

b
t
is measured together with it. Experiments
of Alain Aspect and others support quantum
mechanics, and disprove the existence of a
single joint probability p
r,r
t
,s,s
t
(a, a
t
, b, b
t
) (im-
plied by the EPR local reality hypothesis),
which reproduces all four correlations.
EPR Correlations and Teleportation
A
B
Bellstate
Measurement
Classical Communication
U
Teleported
State
System
1
2
Entangled
EPR Source
Usual Bennett et al
5
Protocol For
Teleportation.
[

)
1
=

[z+) [z)

1
[

)
1
=

[z) [z+)

1
[

)
23
=

[ )

23
/

2
[

) =

[ )

2
[
+
)
1
[

)
23
=
1
2

[
+
)
12
[

)
3
+[

)
12
[
+
)
3
[
+
)
12
[

)
3
[

)
12
[
+
)
3

Published Online May 29 2014 Science Ex-


press Index
Unconditional quantum teleportation between
distant solid-state quantum bits
W. Pfa1,*, B. Hensen1,H. Bernien1,S. B.
van Dam1,M. S. Blok1, T. H. Taminiau1,M.
J. Tiggelman1, R. N. Schouten1,M. Markham2,D.
J. Twitchen2,R. Hanson1,
+ 1Kavli Institute of Nanoscience Delft, Delft
University of Technology, P.O. Box 5046,
2600 GA Delft, Netherlands. 2Element Six,
Ltd., Kings Ride Park, Ascot, Berkshire SL5
8BP, UK.
Abstract
Realizing robust quantum information trans-
fer between long-lived qubit registers is a key
challenge for quantum information science
and technology. Here , we demonstrate un-
conditional teleportation of arbitrary quan-
tum states between diamond spin qubits sep-
arated by 3 m. We prepare the teleporter
through photon-mediated heralded entangle-
ment between two distant electron spins and
subsequently encode the source qubit in a
single nuclear spin. By realizing a fully deter-
ministic Bell-state measurement combined with
real-time feed-forward quantum teleportation
is achieved upon each attempt with an av-
erage state delity exceeding the classical
limit. These results establish diamond spin
qubits as a prime candidate for the realiza-
tion of quantum networks for quantum com-
munication and network-based quantum com-
puting.
Causal Quantum Mechanics and Joint
Position-Momentum Probabilities. The
search for a causal quantum mechanics led
to joint position-momentum probabilities (De
Broglie-Bohm) much before the discovery of
Bell inequalities for the EPR-Bohm-Aharonov
Experiment.Subsequently phase space Bell
inequalities have led to the theorem that for
general states in 2N dimensional phase space
more than N + 1 marginals cannot agree
with corresponding quantum probability den-
sities (Auberson-Mahoux-Roy-Singh).(N+1-
marginal theorem).
De Broglie and Bohm
4
(dBB) proposed a
theory with position as a hidden variable
so that x, [), i.e., the state vector supple-
mented by the instantaneous position is the
complete description of the state of the sys-
tem. Here x = (x
1
, , x
N
) denotes the con-
guration space co-ordinate which evolves
according to
dx
i
dt
=
1
m
i

i
S(x(t), t), (11)
where m
i
denotes the mass of particle i, and
the Schrodinger wave function is given by,
x[(t)) RexpiS, (12)
with R and S real functions of (x, t). DBB
show that if we start at t = 0 with an ensem-
ble of particles whose position density coin-
cides with [(x, 0)[
2
at t = 0, then such time
evolution leads to a position density that co-
incides with [(x, t)[
2
at any arbitrary time t.
Thus, the phase space density is

dBB
(x, p, t) = [(x, t)[
2

p

S(x, t)

(13)
whose marginal at arbitrary time reproduces
the position probability density,


dBB
(x, p, t) d p = [(x, t)[
2
. (14)
Bell: The De Broglie-Bohm picture dis-
poses of the necessity to divide the world
somehow into system and apparatus
(because the position co-ordinates exist with-
out the necessity of measurement). As far
as the position variable is concerned the dBB
theory restores history and causality without
altering the statistical predictions of quan-
tum mechanics. The lack of Einstein locality
is an essential feature of quantum mechan-
ics. It is enshrined in the dBB velocity of the
ith particle depending on the instantaneous
position of all the particles however far they
may be.
The momentum and other variables besides
position do not have the same favoured sta-
tus as position however. As Takabayasi
5
pointed out the dBB phase space density
does not yield the correct quantum momen-
tum density, i.e.,


dBB
(x, p, t)dx ,= [

( p, t)[
2
.
For position, the value observed can be the
same as the preexisting value; not so for mo-
mentum. Momentum therefore has not the
same reality as Position.
Causal Quantum Mechanics
Symmetric in Position and Momentum
. We [S. M.Roy and V. Singh]
6,7
asked the
question , is it possible to remove this asym-
metrical treatment of position and momen-
tum and build a new causal quantum me-
chanics in which momentum and position
can have simultaneous reality? We spelt out
a non-unique but armative answer in one
dimensional conguration space and later (with
G. Auberson and G. Mahoux)
8
generalised
it to arbitrary dimensions. We recall rst the
original one dimensional construction with a
monotonic dependence of momentum on po-
sition.
The point of departure is to seek a phase
space density of the form
(x, p, t) = [(x, t)[
2
(p p(x, t)), (15)
where p(x, t) is not given by the dBB formula.
Rather, p(x, t) is to be determined by the
requirement

(x, p, t)dx = [

(p, t)[
2
. (16)
If we assume that p(x, t) is a monotonic func-
tion of x (non-decreasing or non-increasing),
(p p(x, t)) =
(x x(p, t))

p(x,t)
x

(17)
and
(x, p, t) =
[(x, t)[
2

p(x,t)
x

(x x(p, t)) (18)


If we determine p(x, t) such that
[(x, t)[
2
=

p(x, t)
x

( p, t)[
2
, (19)
we obtain
(x, p, t) = [

(p, t)[
2
(x x(p, t)) (20)
which obeys the desired Eq. (6). Two ex-
plicit solutions to Eq. (9), corresponding to
non-decreasing ( = 1) and non-increasing
( = 1) functions p(x, t) are given by,

p(x,t)

dp
t
[

(p
t
, t)[
2
=

dx
t
[(x
t
, t)[
2
. (21)
Instead of Eq. (5) or Eq. (10), the phase
space density may now be written in the sym-
metric form,
(x, p, t) = [(x, t)[
2
[

(p, t)[
2

dp
t
[

(p
t
, t)[
2

dx
t
[(x
t
, t)[
2

. (22)
To compare with previous results , note that
the Wigner distribution
9
, in 2N dimensional
phase space,
W( q, p, t) =
1
(2)
N

dy exp(i p.y)
q y/2[(t))(t)[ q +y/2) (23)
also reproduces [( q, t)[
2
and [

( p, t)[
2
, as
marginals. However, unlike our phase space
density, that distribution cannot have a prob-
ability interpretation as it is not positive def-
inite.E.g. for two states , ,

d qd p W

( q, p, t)W

( q, p, t)
= [

[
2
/(2)
N
, (24)
the right-hand side vainishes when the states
are orthogonal, so W

and W

cannot both
be positive denite all over phase space.
8. Experimental Tests of q,p Symmet-
ric Causal Quantum Mechanics . I pro-
pose tests of quantum phase space densi-
ties given by q, p symmetric causal quantum
mechanics by means of the best permissi-
ble joint measurements of conjugate observ-
ables devised by Von Neumann, and Arthurs
and Kelly
13
. Their idea is that the system
interacts with an apparatus which has two
commuting observables x
1
, x
2
and approxi-
mate values of system position and momen-
tum are extracted from accurate observation
of x
1
, x
2
. The Von Neumann-Arthurs-Kelly
interaction during the time interval (0, T) is,
H = K(qp
1
+pp
2
) (25)
where K is a const, with KT = 1 and the
other symbols denote the respective opera-
tors. During interaction time, H is so strong
that the free Hamiltonians of the system and
apparatus are neglected. Arthurs and Kelly
start with the system-apparatus initial state,
(q, x
1
, x
2
, t = 0) = (q)
1
(x
1
)
2
(x
2
) (26)
where,

1
(x
1
) =
1/4
b
1/2
exp(x
2
1
/(2b
2
)) (27)

2
(x
2
) =
1/4
(2b)
1/2
exp(2b
2
x
2
2
), (28)
and b/

2 is the uncertainty of x
1
in the ini-
tial apparatus state. The uncertainty of x
2
is chosen as above for optimum results.They
solve the Schroedinger Eqn. exactly and ob-
tain the nal joint probability density of the
apparatus variables to be just the Husimi
function
13
,
P(x
1
, x
2
) =

b,x
1
,x
2
[)

2
/(2) (29)
where

b,x
1
,x
2
(q) = (2)
1/4
b
1/2
exp(iqx
2
(x
1
q)
2
/(4b
2
)) (30)
is a minimum uncertainty system state cen-
tred at q = x
1
, p = x
2
. Although x
1
) = q),
and x
2
) = p), the dispersions in x
1
, x
2
are
larger than for the system,
(x
1
)
2
= (q)
2
+b
2
,
(x
2
)
2
= (p)
2
+
1
4b
2
. (31)
By varying b we obtain the measurement or
noise uncertainty relation , (units = 1 ),
x
1
x
2
1 (32)
where the minimum uncertainty is twice the
preparation uncertainty. Arthurs and Good-
man
13
have given a beautiful proof of this
fundamental uncertainty relation, which is
independent of any particular choice of the
Hamiltonian. Further,for x
1
distribution to
approximate q distribution closely, we need
b _ q; for x
2
distribution to approximate
p distribution closely, we need b (p)
1
. Hence, for good approximation of both q
and p distributions, we also need,
qp 1. (33)
Testing a phase space density against the
Arthurs Kelly (AK) distribution is meaning-
ful only in this region. Even when the phase
space density reproduces the quantum po-
sition and momentum probabilities exactly,
the test is not trivial, because the position-
momentum correlations in the two distribu-
tions may not agree. E.g. for the free parti-
cle Gaussian wave function,

(p, t) = ()
1/4
exp[
(p )
2
2
it
p
2
2m
],
(p)
2
=

2
, (q)
2
=
1 +(t/m)
2
2
, (34)
the Roy-Singh q,p symmetric causal quan-
tum mechanics gives,
p
RS
(q, t) = +
p
q
(q t/m). (35)
qp+pq)
RS
= (t/m)[2
2
+

1 +(m/(t))
2
].
(36)
This agrees with the Arthurs-Kelly result,
2x
1
x
2
)
AK
= (t/m)[2
2
+], (37)
when (m/(t)) _1, i.e. qp 1. In con-
trast, another phase space density which re-
produces quantum q, p probability densities,
viz. [(q, t)[
2
[

(p, t)[
2
gives 2(t/m)
2
for the
above correlation in disagreement with the
Arthurs-Kelly result. We know that for
(p)
1
_b _q,
the separate q and p probability densities given
by AK and RS agree. Therefore the crit-
ical test of the RS distribution against the
A K distribution will be to compare the
positions of the momentum peaks p
RS
(q)
and p
AK
(q). This has been performed by
Arunabha Roy
14
.
p
AK
(q) = (p
RS
(q) )[(1 +(b/q)
2
)
1

(1 (1/4)(qp)
2
)] (38)
The close agreement of the q, p correlations
in the two distributions in the relevant re-
gion (p)
1
_b _q is very striking.It in-
dicates the possibility of a causal description
of single slit diraction exactly in the region
qp 1 highlighted by Einstein.
The non-uniqueness of such a description
even for a xed choice of the canonical pair
q, p opens up the question of optimisation of
phase space probability densities. Recently I
noticed (S. M. Roy, Physics Letters A, 377
(34-36). pp. 2011-2015 (2013)) that a con-
vex combination of the two R-S phase space
densities, = 1
(q, p)
C
=
+

+
(q, p) +

(q, p),

=
1
2

1
2

1 (2qp)
2
(39)
reproduces the quantum correlation < qp +
pq > 2 < q >< p > exactly.The possibility
of experimental tests via Arthurs-Kelly mea-
surements in quantum optics is an incentive
to look for this miracle for more general wave
functions.
Remote tomography and entanglement
swapping via von NeumannArthursKelly
interaction S. M. Roy, Abhinav Deshpande,
and Nitica Sakharwade, Phys. Rev. A 89,
052107 (2014).
INTERACTION
TRACKER 1
TRACKER 2
SYSTEM
PUMP PHOTONS STRONG
DISTANT
STATION
TOMOGRAPHY
STATION A
B
PHOTON
APPARATUS PHOTON 1
APPARATUS PHOTON 2
SYSTEM PHOTON P
(ENTANGLED WITH P)
REGION
P
99
J. S. Bell, Physics 1 (1964) 195; J. F. Clauser,
M. A. Horne, A. Shimony, and R. A. Holt,
Phys. Rev. Lett. 26 (1969) 880.
A. Einstein, B. Podolsky and N. Rosen, Phys.
Rev. 47 (1935) 777.
Stapp Phys. Rev. D3 (1971) 1303; Nuovo
Cim. B29 (1975) 270; Found. of Phys. 9
(1979) 1 .
A. Einstein, Phiolosopher Scientist, P.A. Schilp
Ed.. Library of Living Philosophers, Evanston,
Ill. (1949). See esp. Einsteins autobio-
graphical notes and replies by N. Bohr here
and in N. Bohr, Phys. Rev. 48 (1935)
696. See also, reprints of related articles in
,Quantum Theory and Measurement, J. A.
Wheeler and W. H. Zurek Eds., Princeton
(1983).
D. Bohm and Y. Aharonov, Phys. Rev. 108
(1957) 1070.
N. D. Mermin, Phys. Rev. Lett. 65 (1990)
1838.
S. M. Roy and V. Singh, Phys. Rev. Lett.
67 (1991) 2761.
M. Ardehalli, Phys. Rev. A 46 (1992) 5375;
A. V. Belinskii and D. N. Klyshko, Phys.
Usp. 36 (1993) 653; N. Gisin and H. Bechmann-
Pasquinucci,Phys. Lett. A 246 (1998) 1.
References
1. R.B. Griths, J. Stat. Phys. 36, 219
(1984); Phys. Rev. Lett. 70, 2201
(1993); M. Gell-Mann and J.B. Hartle,
in Proc. 25th Int. Conf. on High En-
ergy Physics, Singapore 1990, eds. K.K.
Phua and Y. Yamaguchi (World Scien-
tic, 1991); R. Omnes, Rev. Mod. Phys.
64, 339 (1992) and
The Interpretation of Quantum Mechanics
(Princeton Univ. Press 1994).
2. F. Dowker and A. Kent, J. Stat. Phys.
82, 1575 (1996); Phys. Rev. Lett. 75,
3038 (1995).
3. A.M. Gleason, J. Math. & Mech. 6, 885
(1957); S. Kochen and E.P. Specker, J.
Math. & Mech. 17, 59 (1967); J.S.
Bell, Physics 1, 195 (1964); A. Martin
and S.M. Roy, Phys. Lett. B350, 66
(1995) ; S. M. Roy, Int. J. Mod. Phys.
14,2075 (2000).
4. L. de Broglie, Nonlinear Wave Mechan-
ics, A Causal Interpretation, (Elsevier
1960); D. Bohm, Phys. Rev. 85, 166;
180 (1952); D. Bohm and J.P. Vigier,
Phys. Rev. 96, 208 (1954).
5. T. Takabayasi, Prog. Theor. Phys. 8,
143 (1952).
6. S.M. Roy and V. Singh, Deterministic
Quantum Mechanics in One Dimension,
p. 434, Proceedings of International Con-
ference on Non-accelerator Particle Physics,
2-9 January, 1994, Bangalore, Ed. R.
Cowsik (World Scientic, 1995).
7. S.M. Roy and V. Singh Mod. Phys. Lett.
A10, 709 (1995).
8. G. Auberson, G. Mahoux, S. M. Roy and
V. Singh, Phys. Lett. A300, 327 (2002);
Journ. Math. Phys. 44, 2729-2747
(2003), and 45,4832-4854 (2004).
9. E. Wigner, Phys. Rev. 40, 749 (1932).
10. S.M. Roy and V. Singh, preprint TIFR/TH/98-
42 and Phys. Lett. A255, 201 (1999).
11. A. Martin and S.M. Roy, Phys. Lett.
B350, 66 (1995).
12. S. M. Roy, in preparation (2012).
13. J. Von Neumann, Math. Foundations of
Quantum Mechanics, Princeton Univer-
sity Press (1955); E. Arthurs and J. L.
Kelly,Jr., Bell System Tech. J. 44,725
(1965); E. Arthurs and M. S. Goodman,
Phys. Rev. Lett. 60,2447 (1988); K.
Husimi, Proc. Phys. Math. Soc.Japan,
22,264 (1940), S. L. Braunstein, C. M.
Caves and G. J. Milburn, Phys. Rev.
A43,1153 (1991); S. Stenholm, Ann. Phys.
218,233 (1992); P. Busch, T. Heinonen
and P. Lahti, Phys. Reports 452,155
(2007).
14. A. S. Roy and S. M. Roy, in preparation
(2012)

You might also like