Non-Locality, Causality and Aether in Quantum Mechanics: J, H. Poincare

55
Astron. Nachr., 303 (1982)r, 55-80
Non-Locality, Causality and Aether in Quantum Mechanics

J, P. VIGIER,Paris
Laboratoire de Physique Theorique, Institut H. POINCARE
With 16 figures (Received 1981July 20)
Introduction
Since MICHELSON’Sexperiment, EINSTEIN has acquired, (in the general public) the reputation of being the
destroyer of the Aether concept. This is a misconception: he only buried the “absolute” classical aether models.
Indeed in a little known paper (EINSTEIN 1924,i.e. published after the discovery of general relativity) he describes
the curved spacetime as a changing real material field (i.e. as a covariant aether) which carries gravitational inter-
actions along the light cone.
This astonishing paper (entitled “Uber den Ather”) remained largely unnoticed due to EINSTEIN’S defeat in
the BOHR-EINSTEIN controversy and his subsequent isolation in the physics community. It anticipates by a few
years EINSTEIN’S well-known attempt to unify matter and gravitational field by considering particles as singularities
in the g,, field. This description, which implies the geodesic trajectories, constitutes (to my knowledge) the only
materialistic explanation of the objective nature of the laws of motion which now naturally result from the local
continuity of the gravitational field. This description of matter as part of extended fields evidently conflicts with
the irreducible unsurpassabIe statistical character attributed by BOHRto the waves of quantum theory. This is
natural. Already EINSTEIN’S description of probability as limit of distributions induced by subquantal stochastic
processes had led (after 1927)to the famous epistemological divorce between Relativity and the Copenhagen Inter-
pretation (CIQM) of Quantum Mechanics.
The aim of the present report is to revisit EINSTEIN’S 1924 ideas in the light:
a ) of recent developments in the Stochastic Interpretation of Quantum Mechanics (SIQM)
b) of the growing experimental evidence in favor of non-local interactions in recent experimental tests of the EIN-
STEIN-PODOLSKY-ROSEN (EPR) paradox.
In other terms we are going in successive sections
A) to add to the usual g, terms stochastic 8gpvterms and describe spacetime as a real subquantal covariant random
medium, which implies subquantal fluctuations at the velocity of light,
B) to show that the corresponding Brownian like motions of the classical extended model of spinning particles satisfy
the relativistic generalisation of NELSON’Sform of NEWTON’S law i.e. :
m(Dcvd- Dus) = F + (0-1)
where D,and D , represent drift and stochastic derivatives,
vd und usdrift and stochastic four velocities, F + the external force.
C) To demonstrate that the non-locality recently connected with quantum mechanics in the theoretical analysis
and experimental tests of the EPR paradoxon can be interpreted in this extended gravitationa1 model since one can
deduce from it the SIQM, the superluminal propagation of the quantum potential and the causality (in EINSTEIN’S
sense) of quantum phenomena.
Indeed this new possible unification of gravitation and quantum theory at the microscopic level
I ) interprets the non-local interactions predicted by wave mechanics and describes the EPR paradoxes as action at
a distance between spacelike separated particles.
2) implies possible new experimental tests which would help to choose between CIQM and SIQM.
1. EINSTEIN’S
Random Aether Model
A) The idea to use random fluctuations of the g,, field (considered by EINSTEIN himself (EINSTEIN 1924)to be
a possible representation of a real all pervading material field) as the origin of the real quantum forces which justify
the stochastic interpretation of quantum mechanics is not new in the literature. Indeed EINSTEIN’S ideas on the
question were reactivated by JANOSSY (1972)who considered the cosmological term AgPyin:
56 J. P. VIGIER:
Non-Locality, Causality and Aether in Quantum Mechanics
as a representation of the energy momentum tensor distribution of the background gravitational aether. Moreover
the introduction of a random g, component (which goes back t o A. MARCH(1934, 1g37), M. A. MARKOV(1938),
H. YUKAWA (1966) and D. I. BLOCHINTSEV (1975) has been recently revived by C. FREDERICK(1976), K. NAMSRAI
(1980) and G. F. CEROFOLINI(1980).
As suggested by FREDERICK (1976) stochastic ,g, can be defined as follows:
-
a) At each world point the tangent space-time is not defined by the MINKOWSKIS, but by d,, + &,,where
denotes stochastic tensor components.
b) All measurements of dynamical variables correspond t o controvariant components of tensors.
c) the probability P ( z , t ) t o find a particle can be written P ( z , t) = A (ac, t) I/Twhere g is the determinant of the
metric.
d) The 8gsvare small and practically add linearly, so that the rnetrics due t o different physical situations (I) and (2)
yield a metric:
g g = (f) (gS) g:”y’) * +
Assumptions (a), (b), (c), (d) imply the following evident consequences:
I) If one denotes the metric at a point (I) by&, we cannot determine i t precisely but only ask for P(&) which
is the probability of a particular &,. However we can define its average value
volume element, so that &, = &,, @ +
,, with (I&,,)
-
= (&) over a four dimensional
= (sag,,,) = 0. The fluctuating Sg,, field can thus be compared
t o a chaotic superposition of random wavelets on a curved regular sea. This is physically reasonable since gravita-
tional waves originating from all independent (practically uncorrelated) sources in the universe are expected t o
build a random background gravitational noise i.e. the “gravitational vacuum”.
2) The use of contravariant quantities in measurement theory is necessary if we represent test particles as massive
singularities of the g, field which thus necessarily follow geodesics. Indeed the random &, imply fluctuating geode-
sics (i.e. random gravitational forces) since we have on each such particle:
where c} are now fluctuating CHRISTOFFELsymbols and- the proper time derivative. Relation (2)evidently generates
a random set of motions around the usual mean geodesics defined from &”. These motions which can be neglected
for macroscopic bodies imply an irreducible stochastic character upon all motions of microscopic test particles : an
evident possible basis for SIQM. Since we must introduce measurable dx, in (2) we must be able t o define the distance
t o our singular test particles. Calling this distance i’(= F), the covariant equivalent or the radial coordinate 7 is
I
ti and 5, = g,,e = g,,[’ = (Y/(I - zGm/r)) which yields for the contravariant distance J d7 = 7 in contradistinction
1 0
with the covariant distance 5, = J d(7/(1 - 2Gmlr) = co so that only the contravariant distance is observable.
zPv
0
3) The stochastic random motions induced by the occur a t the velocity of light. This is the essential property of
our model. It can be demonstrated as follows. Following FREDERICK’S line of argument (1976) we start with a particle
a t rest (for an inertial observer) a t a space point 0,.
After a time dt the EULER-LAGRANGE equations yield a proba-
bility D,(x) of being in a region bounded by x and x +
dx. After a second time interval dt the new distribution is
m
D , +&). From probability theory (JAMMER 1966) this is the convolution D I + , ( x )= D,(y) D,(x - y ) dy. Assuming
-a
vacuum isotropy (so that the EULER-LAGRANGE equations give the same distribution independently of the starting
point) we have for our Brownian like motion D,(x) = D,(x). Since g,,(x) = (gvy(xl),gpy(x2)...} and thus D,(x) =
= {Ill(.), D,(x) ...} are identically disiributed random variables, the motion of any test particle is equal t o the
repeated convolution Drtz+..,(x)...which yields a normal distribution by the centrallimit theorem. The position spread
of the particle a t t >o is a Gaussian. To calculate the spreading velocity we remark that after N such convolutions
( N large) we obtain a normal distribution with variance 0 2 . Still according t o the central limit theorem (JAMMER
1966),this distribution is N times the variance of D,(x). If we represent this variance of Dl(x) by a, i.e.:
var (Ox)
=a
we see that the distribution Dl is obtained after d t so that we can write after N convolutions
dx = var ( D2
: I ) = Na
resulting from N time intervals dt. This yields:
So the initially localized test particle spreads with a constant velocity a which must be equal t o c since the
result is frame independent. The same result is obtained with a particle with a definite position and a momentum
distribution with the difference of having a different distribution D, due t o the uncertainty of the particle’s direction
nf propagation.
J. P. V ~ G I E R :
Non-Locality, Causality and Aether in Quantum Mechanics 57
Aether : BROWN-NELSON
2. Random Motions Associated with EINSTEIN’S Motions
We are now in a position t o examine the stochastic motions which result from our fluctuating zPy
field. To do
this we shall consider an ensemble of extended particles which combine average drift motions with fluctuations
induced by our background gravitational noise.
This classical relativistic model generalizes the non relativistic stochastic hydrodynamical model of QM of
BOHMand VIGIERin terms of a fluid with irregular fluctuations (BOHM,VIGIER1954).It contains three new physical
features.
I) The fluid elements which follow the lines of flow of the fluid with irregular fluctuations are built from extended
elements in the sense discussed by BOHMand VIGIER(1958)and HALBWACHS, SOURIAU and VIGIER(1961).They are
the classical counterpart of J = O,I or I/Zspinning particles (FENECH, MOLES and VIGIER1979).
2 ) As just shown the stochastic fluctuations occur a t the velocity of light.
3) The fluid is a mixture of extended particles (and antiparticles) : the latter being mathematically equivalent to
particles moving backward in time (TERLETZKI and VIGIER1961).
The existence of such fluctuations (which induce in the particle a MARKOVtype of Brownian motion) has been
shown (BOHMand VIGIER1954)to lead any initial distribution of the particles in the fluid into a limiting equili-
brium distribution const. e(x,(r)) proportional t o the fluid‘s average conserved drift density e(x,(r)).This means
that the fluctuations of our MADELUNGfluid induce on our particles stochastic jumps which can be decomposed
into the regular drift motion vd plus an apparent spacelike random part uswith
v d = dx,/dt (2-1)
t representing the proper time along the drift lines: so that v d - vd = -c2,
Indeed any velocity w represented by a point P (with w, . zu, = 0 ) of the light cone can be decomposed into
the sum of two four-velocities v d and us,i.e. w = v d + >
us with us- us 0 : Since the three independent Compo-
nents of w determine the four components of us.As a consequence if one considers a particle of the preceding type
it undergoes two independent types of motions:
a) regular motions along the fluid’s drift lines of flow with the fluids own velocity v d .
b) stochastic jumps in any direction with the velocity of light with a four velocity w satisfying w w = 0. -
To establish (I)let us first recall that a particle or a regular fluid element (which can be compared with the
stochastic particle and the thermostat’s elements in the usual Brownian motion) are now represented in four dimen-
sional space-time like hypertubes instead of time-like lines. These hypertubes can be naturally assumed t o have a
minimum spacelike radius 112 which yields the mimimum distance 7 which separates two continuous particles in
any spacelike section passing through their centre of mass. Independently of the stochastic jumps our drifting fluid
is thus comparable with a timelike set of extended fibers and the mimimum time needed t o pass from one of these
hypertubes t o the next is thus ?/c = d r since the jumps occur a t the velocity of light. This implies that the proper-
time variables which correspond t o adjacent events in our stochastic model have non zero minimum temporal se-
paration A t .
In this model all free extended elements are thus characterized by an internal angular momentum Saj added t o
their four momentum G, and unitary four velocity i,,= dx,/dt (i.e. i,,k@ = -19) which are no longer collinear.
The conservation equations yield:
(j,z= o (i.e. G,G~ = -mEc2)
and
with
S,,= G,i, - G,X,
SP V S ~=V +oz = const.
1 (2-2)
Since we must use extended (and not point) particle elements we must introduce internal angular momenta S@
which will act as the classical counterpart of spin. Indeed we know that the relativistic internal angular momenta
(classical spin) of any extended structure (HALBWACHS, SOURIAU and VIGIER1961)must map (in the particle’s
frame) two-dimensional plane surfaces onto themselves (FENECH, MOLES and VIGIER1979).The first of these planes
is space-like and remains orthogonal t o the axis of rotation in that space: it yields the condition Sap2 = o (with
Zfl = ax@/atrepresenting the four velocity) and corresponds t o spin I. The second contains this rotation axis and yields
spin 0. The third and fourth are both tangent t o the light cone and yield the spin 112 of particles and antiparticles.
On this basis one can introduce three classical extended models corresponding t o the three spin states (i.e.
-
J = 0 , J = 112 and J = I) by the three -
constraints S,,3cv = 0 , S,,vxv = o&,, and S,,viy= 0. One then obtains S,S. =
= const and s,sP = const with S , = S,,,Gv and s, = Sp$. I n particles, internal motions then include circular ro-
tations of a centre-of-matter density (denoted by (c.c) . x,(r)) around a centre-of-mass (denoted by (C.M) and
y,(6) moving parallel to G, with a constant radius Y,, (defined by mZc2Rp= S,,,Gr) and DE BROGLIE frequency
Aw = mc2 with respect t o its proper time. This corresponds t o a classical counterpart of the quantum - Zitter-
bewegung -. The (c.c)’sinternal DARBOUX-FRENET frames which now represent internal spin do not rotate ( J = o),
rotate twice ( J = 112)or once (J = I), while the (c.c) rotates once around the (C.M). Antiparticles have opposite
(mirror) internal motions. Along the (C.M) motion one thus recovers classically PLANCK’S formula A = mc6 (where
A represents the relativistic action and 8 the path interval) which can also be written as A = m c 6 = h x ; x denoting
Our new internal DE BROGLIE phase and It its conjugate momentum. The corresponding internalrotations (oscillations)
also yield 6 = (hlmc)x = 2%where 2 = h/mc is COMPTON’S classical radius, i.e. the (C.M) internal motion associated
with one (2n) internal rotation.
58 J. P. VIGIER:Non-Locality, Causality and Aether in Quantum Mechanics
The second step is just t o generalize t o our relativistic model the average velocities utilized by DE LA PEGA
and CETTO(1975) t o discuss the non relativistic stochastic theory of classical and quantum mechanical systems.
Let us start (fig. I) from a four dimensional volume limited on the side by the fluid’s regular lines of flow and
by two spacelike constant phase surfaces (VIGIER1979) S , and S,. If the domain is small enough such surface sare
separated by an interval 242: an interval A t separating Sland S,from a median section S,. Of course lAtl 2 A<
As a consequence of the assumed stochastic equilibrium we can treat on the same footing the fluid behaviour
and an ensemble of similarly prepared particles characterized by the density e(z,t) in configuration space where
x represents a point in four-dimensional space-time.
We shall now establish that the preceding model leads t o the correct quantummechanical statistics (governed
in our simplified case by the KLEIN-GORDON equation) in the simple case of a charged scalar particle. The simplifi-
cation is justified since the introduction of spin complicates, but does not modify significantly, the various steps of
our demonstration.
-
X
Fig. I.
We can describe the average local motions of the elements of the ensemble by the selection of all particles that
at proper time z = t,are contained in a small four-dimensional volume element around the point r = r, with co-
ordinates ( Y , ) ~ . This is necessary in our model, since if one starts from a particle in its local drift rest frame (i.e. the
frame in which the neighbouring fluid elements are practically a t rest) its stochastic jumps along the light cone can
bring it into any neighbouring line of flow: both in the forward and backward proper time direction. As a conse-
quence our general stochastic model implies the use of a four-dimensional stochastic space-time volume element to
recover all possible stochastic jumps of each drifting particle. We have thus made the new theoretical step of intro-
ducing along with the average space positions the new concept of an average time in a four-dimensional volume ele-
ment.
I n order t o describe the global motion of this element we select the particles that a t proper time t, are con-
tained in a small section (space volume element) of s, limited by the hypertubes boundary. According t o our model
i t is possible t o distinguish two different kinds of motion of this volume element during a short interval At. Besides
its motion as a whole in the hypertube (which preserves the fluid’s scalar density e) the element will suffer varia-
tions of p due t o the stochastic jumps which move matter from one line of flow t o another and will bring fluid across
the hypertubes’ boundary. Generalizing DE LA PEGA’Sand CETTO’S (1975)ideas we can obtain a simplified descrip-
tion in terms of two quasilocal statistical velocities. If we take any one of the particles of our volume element
+
and call rl and r2 its average mean position a t tl= t, - AT and z1= t, A t we can calculate the average of
r, - r, over the subensemble defined by the particles which belong t o our small volume element. We call these
average values the mean and denote them with ( ). We thus write
r, - r, = (r, - r z ) + 6,r
and r, - r1 = ( r , - rl) 6-r . +
(2-3)
Since one must assume (in our model) the homogeneity, isotropy and time independence of our stochastic mecha-
nism the change variable S g i must satisfy (6+ri) = (6-ri) so that we can omit the subscripts f from such ex-
pressions and write in general (&;) = 0.
We can now derive from (3) two different velocities i.e.
b + ( z )= ( ( r 3- r,)/At) and k ( 2 )= ((r, - r l ) / A t ) ,
whose mean values
~+= ( 4( b + ( z ) )= < ( ( r 3- r,)/At)) and ~ ( z=) ( k ( z ) >= ( ( ( r , - r l ) / A t ) )
are the relativistic generalization of the mean forward and backward velocities. From these one can derive the
regular fluid’s velocity vd and a stochastic velocity usthrough the relations
~ ~ (=-2 ( () ( r 3- r l ) / z A t ) )= f (v+ + R ) (2-4)
u,(z) = ( [ ( r 3- Y,) - (r, - r 1 ) ] / 2 A t )= f (v, - t7-) (2-5)
J. P. VIGIER:
Now the stochastic velocity uscan be determined in any spacelike direction by calculating the flow between zl
and z3of all elements which cross-a drift timelike plane passing through r2and orthogonal t o a spacelike direction x.
Indeed let us consider (see fig. 2) an ensemble of fluid elements (particles) which are a t z2 in the neighbourhood of
x. If e1L(eIR) then represents the scalar densities in the neighbourhood of xIL(xIR)a t z = zlwe see that these densi-
ties are related t o n+ and n- through
a+ = -
( X ~ R x2) e 3 =~ (x2 - eIL and a- = (X,R - eIR = (x2 - x 3 d e3~.
e3L
c.
c-
4-
Fig. 2. particles at t, = t 2- A t and x3R(x3L)is the average position of the same particle
x I L ( x I Ris) the average position of the $z+(n-)
at t g= t 2+ A t ; eL(eR)being the densities of particles t o the left (right) of x = x2.
This yields
x1 + + +
x3 - 2x2 = (I/(n+ a-)) [-e1L(x2 - x , ~ ) ~erR(%R - ~ 2 ) e3R(x3R
~ - 4' - e3dx2- xd2I +
which can be averaged over the ensemble. Since each of the parentheses then become ( ( 6 ~ ) we
~ ) can write t o the
first approximation (with n, +
k = 2e(z2)A x ) :
if we define as usual the diffusion coefficient as D = ( ( S r i ) 2 ) / 2 A z and neglect higher order terms in AT. D is always
> o since our quantum jumps are spacelike.
This is exactly the relativistic generalization of EINSTEIN'S definition of the stochastic velocity in Brownian
motion. We have further v+ = v d f uswhich connect our forward (particle) and backward (antiparticle) velocities
with the fluids regular drift velocity vd and its stochastic velocity us.
Moreover an expansion in TAYLOR series yields
To calculate D we now assume as done independently by CEROFOLINI(1980):

a) That in our stochastic jumps D describes (on the average) one jump only.
b) That in each individual jump the particles' (c. c) undergoes one rotation only (FENECH, MOLES and VIGIER1979)
around the (C. M). This is natural since jumps do not cancel rotation.
Assumptions a) and b) evidently represent the simplest possible Brownian-motion scheme which preserves both
the average drift motion and the local phase correlation and average ordering of the DE BROGLIE-C1OCk zitter-
bewegung. If no interaction occurs during one stochastic
~-
jump, we can consider it as the v -+c limit of a free-particle
motion described by the Lagrangian 62 = -mc d-ipxp dz, where dz represents the corresponding drift interval
(FENECH, MOLES and VIGIER1979). As stated above, during the corresponding free jump the (c. c) rotates of a
quantity 27c, so that the corresponding distance ((6x)2)1/2 must be equal t o the COMPTONwave-length A = (film).
When going t o the v = c limit we thus write ( ( 6 ~ ) =~ )c2 A z 2 , which yields (since D = ((6x)2)/zAz)the diffusion
constant D = h/zm, with:
(6ri * 8 r i / 2 A t ) = DGii ,
in the drift rest frame ;diffusion in time representing, as before, particle-antiparticle transition; 6ri and 61;. denoting
any pair of Cartesian components of 6*r which are assumed t o be statistically independent if i =/= j .
The second step is t o associate the two velocities needed t o describe our motion t o four accelerations required
t o describe the forward and backward changes of these velocities. To do this we require the existence of our mini-
mum proper time interval which allows us t o define the four accerlerations
b+(3)- b + ( z )= a: + 6+b+
i
*
b - ( 3 )- k ( z ) = a+ + B+b-,
b + ( 2 )-b+(I)= 0 ; + d-b+ t
(2-7)
b.(2) - b-(I) = a1 + 6-b- 8
which evidently lead t o systematic drift and stochastic derivative operators. Indeed if we define as Dd and D, the
following operations on a general function f ( r ) of the stochastic variable r, i.e.
Ddf(r2) = <[f(r3) - f(rl)l/2dtand Dsf(r2) = + f ( r )- z f ( r 2 ) l / 2 A z >
<[f(rl) J
which are evidently related with the forward (D+)and backward (D-) derivative operators through the relation
D* = Dd D,; we see they thus correspond to scalar (proper time type) derivatives in timelike and spacelike
directions and yield the drift and stochastic velocities through 6 d = Ddr and u, = Dsr, where the dummy in-
dex 2 has been omitted. This generalizes vp = dx,/dt and leads t o the preceding mean accelerations through the
expressions a$ = D+v* and a? = D-v*.
equation, in our model. To do
The third (essential) step is t o derive the covariant generalization of NELSON’S
that we recall that any detailed description must start from the general equation
=f d + fs (2-8)
wheref, represents the drift spacelike forces and f, the purely random effect, the dot denoting proper-time derivatives.
The corresponding statistical theory must, according t o our model, start from the ensemble of particles which a t
any proper time t2lie in the neighbourhood of r,. The mean of the preceding relation thus becomes
m ( r ) = Fd + F, = where Fd = (fd) F, = (fs) = 0 .
with
Since the mean value of r is taken over the same ensemble utilized t o define our average velocities and acce-
lerations in the preceding steps, it must be expressed as a linear combination of af. To determine these combinations,
we remark that ( r ) and ( f d ) can be split into two parts, i.e. a part ( r ) +(or ( f a ) + ) which is invariant under proper
time reversal i.e. z, - t, +tl - z2 and a part (#)- or (fa)-) that changes sign under this discrete symmetry which
changes vd but conserves us.Combining equation (9) with its counterpart obtained through a proper-time reversal
Qperation we obtain the new set of equations
m(i;)* = F*d . (2-10)
We now make the final step in our demonstration of NELSON’S

equation by examining the
(0-1) implications of
eq. (10).
The first implication is the importance ot the proper-time relation m(r)+= Fd+which evidently represents the
stochastic generalization of Newton’s law for our model. Indeed the usual four-dimensional acceleration
classical point x satisfies xx = o (since x x = -c2) and is invariant under proper-time reversal. The same holds
of a *
for our stochastic case since: a) the drift acceleration v d is orthogonal t o Vd; b) the stochastic spacelike velocity us
is locally orthogonal t o v d so that the corresponding stochastic accelerations (which vanish on the average since
( F , ) = 0) are thus always orthogonal t o V d .
The second implication is that ( r ) +must be expressed by just the linear combination of relation (7) which
+ +
are proper-time-inversion invariant i.e. (a? u T ) or (a$ ar) or a linear combination therefrom.
The third implication is that a mean acceleration (which corresponds mathematically t o second-order proper-
time derivatives) should be defined physically only by the motions of fluid elements surrounding r, i.e. enclosed
within the four-dimensional volume element limited by S, and S, utilized t o define mean quantities. We deduce
therefrom and from the explicit form of the a’s given in eq. (7), that the only quantity of this type invariant under
t --f -t is (a? + +
a+).Indeed the definition of (a$ a=)implies knowledge of the behaviour of fluid elements which
lie outside our volume since it contains four-velocities of elements which are crossing S, and S, in the backward and
forward directions i.e. are leaving this volume. Moreover one sees that the combination ( r ) = (a? aT) evidently +
represents the relativistic definition of the sum of the mean accelerations of antiparticles (a?)and particles (a;)
passing through r, a t t - z:,
As a consequence we must write relation (10) in the form
$ m ( d +a;) = F + , (2-11)
which is exactly the relativistic generalization of the form given by DE LA PEAAand CETTO (1975)t o NELSON’S
equation. Clearly eq. (11) contains particle - antiparticle symmetry.
The same argument applies t o the - part of (10).Indeed the only combinations of a$ that change sign under
proper-time reversal are (a$ - a:) and (a? - a+)and the second only is exclusively defingd by the motion of fluid
elements between S, and S,. We thus have f m(a? - a+) = F- which satisfies the continuity equation and is
compatible with the introduction of the LORENTZ force for charged fluid elements. Moreover these relations can be
rewritten with the help of the definitions of Dd and D, into the form
?n(Dd??d- n,u,) = P+ a
(2-12)
ant1
+
m(Ddu, Dsvd)= I?- . (2-12)b
I n eq- (12 b) both sides tend (as they should) t o zero in the nonstochastic limit.
From this analysis we derive three consequences:
(1980) on the
I) From the preceding calculation one immediately deduces (as was done independently by NAMSRAI
basis of MARKOVprocess analysis) the stochastic equations (0-1) :
(D.D- 6D. 6D) xP = o and Dp = o (2-13)
and RUGGIERO
(i.e. the form given by GUERRA (1978) to equation (0-1) for free particles) where D = a/az Pa, +
and 6D = Wa, - (h/zm) represent the total drift and stochastic derivatives; b F = Dx,, and 8 b P = BDxP denoting
J . P. \’ICIER: Non-Locality, Causality arid Aethcr iii Quantum Mechanics 61
the drift and stochastic velocities. Moreover, as we shall see the stochastic velocity is given by 6@ = -(h/nz)a. x
x log (ex/*)where e is a density which satisfies the continuity equation along the drift tubes of flow.
11) That (as one knows (HALBWACHS 1960))all three wave equations can be mapped into relativistic hydrodynami-
cal representations where particles move (on the average) from one line of flow to another and jump (at the velocity
of light) from one drift line to another. The quantum potential (which now also includes quantum torques) re-
presents stochastic interactions (action and reaction) between our “particles” and the waves, both carried on EIN-
STEIN’S .ilether. DE BROGLIE phase correlation clearly expresses the wellknown property ( HILLION1957)that oscil-
lators can only transfer energy by contact t o their immediate neighbours, provided they oscillate in phase. This
property is true for example for neighbouring closed vortex tubes in classical fluids - or for connected spring systems
(HILLION 1957).I t implies that the stochastic jumps do not disturb the regular ordering of DE BROGLIE clocks in DE
BROGLIE waves and justifies DE BROGLIE’S idea (HALBWACHS, SOURIAU and VIGIER1961)that the subquantum ther-
mostat is built with a superposition of independent J = 0 , J = 112 and J = I co-variant thermostats.
111) The three J = 0, J = I, J = 112 wave equations can be directly obtained by application of the results of
HALBWACHS, SOURIAU and VIGIER1961 to the three classical fluids of spin J = 0,I/Z and I. One obtains in this
way, the KLEIN-GORDON (VIGIER1979;GUERRA and RUGGIEI~O 1978;LEHRand PARK1977),FEYNMAN-GELL’MANN
CUFAKO-PETRONI and VIGIER1g81),and PROCA (CUFARO-PETRONI and VIGIER1979)equations.
3. Causality and Locality
We now come back t o the present experimental and theoretical discussion on causality a i d locality. As
everybody knows MICHELSON’Sexperiment exploded the classical aether models and EINSTEIN built relativity theory
on the idea that n? material particle (i.e. information) could move faster than light between source and observer.
He also assumed that all interactions had t o travel continuously between interaction objects. His scheme thus
eliminates NEWTON’S conceptual difficulties tied to gravitational instantaneous “action-at-a-distance” acting through
the classical empty vacuum which results from MICHELSON’S remarkable result.
As we shall now establish this suppresses the possibility of real free long distance travelling tachyons and will
be shown (by ZEEMAN’S theorem; ZEEMAN 1964)t o be the basis of EINSTEIN’S “locality principle” i.e. to the relati-
vistic causal and local description of the behaviour of matter.
Indeed if one accepts the idea that particles are dimensionless points, then EINSTEIN’S analysis of the objective
reality and validity of the LORENTZ and POINCAR& transformations (and tlieir subsequent confirmation by experi-
ment) established two fundamental new points.
a) All objective laws can be written in a covariant form independent of the observers’ frame.
b) No interaction and/or information can propagate faster than the velocity of light c, which implies the essential
concept that no information and energy can travel faster than c.
A 5 a consequence,
I) The possible causal origin of any situation a t any given world point P is limited within its past light-cone.
2) The possible causal consequences of any event a t a world point P are limited t o its forward light-cone.
In other terms if particles are points then causality (in ZEEMAN ’s sense) implies locality. Vice-versa EINSTEIN’S
locality yields causality if only point particles carry interaction and/or information. To follow EINSTEIN one can
state that causality is mathematically equivalent t o the POINCARB group plus dilatations. This is an essential
point. If one accepts:
a) the existence of a tangent MINKOWSKIspace at each space-time point, then
b) ZEEMAN(1964)has shown, on mathematical grounds, that if one represents causality by a partial ordering of
test particles on Minkowski space the group of all automorphisms which preserve this partial ordering is necessarily
generated by the inhomogenous LORENTZ group with dilatations and space reversal.
To summarize his demonstration, if one denotes tangent space-time by M (with M 3 X(x,, x,, x,, x3) and
introduces the characteristic quadratic form Q ( x ) = - x: - x t - ther cis a partial ordering on M given by x y <
if a n event at x can influence an event at y . More precisely, x < y when y - x is a time-like vector: Q(x - y ) > o
<
with xo yo. If a function f then represents a one t o one mapping M -+M , f is a causal automorphism if both f
(and f-1) preserve locally the partial ordering i.e. x < <
y -+fx fy with all x , y E M . By definition these causal
automorphisms f(f-l) form a group G which we call the causality group. Now ZEEMAN has demonstrated that f(f-l)
only preserve the partial ordering if, and only if, they preserve the relation associated with transitions at the velo-
city of light Q ( x - y ) = 0.
Since this implies (ZEEMAN 1964) that f(f-’)must map:
I) light rays t o light rays
2) parallel light rays t o parallel light rays
3 ) each light ray linearly
4) parallel equal intervals on light rays t o parallel equal intervals.
One sees that the only possible group is G = P (the POINCARB group) plus dilatations and space reversal.
As everybody knows, the causal group G implies the essential consequences of EINSTEIN’S
(local) restricted
relativity theory. We only mention two here. Causality implies :
I) that no particle can overcome the velocity of light c so that no particle can leave the light cone.
11) there are no particle (or object) endowed with negative energy which can propagate in the direction of positive
time: so that all physical signals have to be carried towards the future by objects endowed with positive energy.
This is the basis of the modern theory of information and G implies in fact the existence of antimatter: which is
mathematically equivalent t o a particle endowed with. negative energy travelling backward in time with (if extended)
space reversed (mirror) internal motions.
111) contrary t o widespread opinion (and as remarked by EINSTEIN and DIRAC(1938)) the causality group G does
not imply locality in the precise sense that, if extended test particles can propagate internal superluminal inter-
action, then one can construct in principle phase phenomena (like collective excitations in superconductivity) which
propagate, by contact, faster than light on a material “aether” filled with stochastic motions. Of course such mo-
tions cannot propagate signals but, as we shall see later, can explain in principle the superluminal interactions re-
cently observed in ASPECT’S experiment (ASPECT1975) on the E. P. R. Paradoxon.
As we shall see later, these limitations of the effective past (and future consequences) introduced by EINSTEIN play
an essential role in the discussion of the concept of causality introduced by quantum mechanics, because BOHR’S
measurement formalism (STAPP 1977)’ as well as experiment, now suggest (though not yet definitively) that there
exist space-like interactions (between space-like macroscopically separated measurement apparatus measuring
associated pairs of correlated particles) which correspond exactly t o the physical situation discussed in the B. P. R.
paradoxon (EINSTEIN1935).
We have emphasized here the locality-causality connection because many people still believe that they are
unseparable concepts. This is evidently not true and does not follow from ZEEMAN’S theorem.
Indeed if we drop the point-like aspect of relativistic particles and consider them as extended time-like hyper-
tub&) which can transmit superluminal interactions localized within their internal structure we see :
A) that though they preserve causality locally (in the sense that the sequence of causes and effects observed by
any observer along the time like paths followed by a characteristic internal point (center of mass or center of matter
density) does not change for any observer) they break locality on small distances.
B) that they can break locality macroscopically in the sense they can carry on local superluminal interactions which
are not transmitted by individual particles (limited t o timelike motions) but result from the superluminal propaga-
tion of a real physical collective excitation (i.e. a density wave) which propagates like a phase phenomenon (ana-
logous t o the successive lighting of electric bulbs on a Christmas tree) on the top of a continuous thermostat of
such extended elements which then corresponds t o the material “vacuum” which supports the real physical y
waves associated t o individual particles in the B. V. particular version (BOHM,VIGIER1954) of the causal inter-
pretation of quantum mechanics.
Point A is easily established even for individual isolated extended particles. The idea that extended particles are
non local in nature i.e. t h a t they can propagate in their interior superluminal interaction and/or information is not
new in the literature. DIRAC(1938) was the first t o notice that if one treats the classical extended electron as a point
charge imbedded in its own radiating e.m. field the equations obtained are of the same form as those already in
current use; but that in their physical interpretation the finite size of the electron reappears in a new sense: the
interior of the electron being a region of space through which signals can be transmitted faster than light. Physi-
cally this can be understood as follows. If we send out a pulse from a point A and a receiving apparatus for electro-
magnetic waves is set up a t a point B and if we suppose there is an extended electron on the straight line joining A
t o B then the disturbed electron will be radiading appreciably a t a time u-l before the pulse has reached its centre
so that this emitted radiation will be detectable a t B a t a time u-l earlier than when the initial pulse (which travels
from A t o B with the velocity of light) arrives. I n this way a signal can be sent from A t o B faster than light so
that it is possible for a signal t o be transmitted faster than light through the interior of an electron; the whole
theory being, of course perfectly LORENTZ invariant and causal in EINSTEIN’Ssense since no particle travels faster
than light.
The same result can be obtained from the relativistic generalization of the motion of extended rigid body. As
one knows there is no such thing as a perfectly classical extended rigid body in relativity since the distance between
l) This extension opens new paths t o interpret old and newly discovered quantum numbers ( T3,Y,C etc.) as describing internal
periodic stochastic particle motions: see for example GUERET et al., Lett. in Math. Phys. 3 (1979)47,N.CUFARO-PETRONIet al., J. Phys.
A 14 (1981)501 and Nuov. Cim. Lett. 29 (1980)565.
J. P. VIGIER:Non-Locality, Causality and Aether in Quantum Mechanics 63
two arbitrary subelements depends on the spacelike cross-section of the timelike hypertube defined by the observer.
However if we consider in any such arbitrary spacelike region a given chain of contiguous subelements and assume
that the corresponding strings cannot cross a t any time another chain built with different elements we arrive a t the
relativistic concept of an elastic solid i.e. t o the closest possible analogue of a rigid body in classical theory. This
has been utilized by GUTKOWSKY, MOLESand VIGIER(1977)t o construct the classical counterpart of DIRAC’S electron.
If we then add t o this concept the idea that we are dealing with a spherically shell of matter with a current J , =
= av, such that with v, -
v, = -c2:
I) a vanishes everywhere a t any fixed time except a t the surface of a sphere of centre z and radius r where CY $; o is
constant ;
a ) Y is a constant;
3 ) v, depends on time but a t any time is constant on the surface of the sphere;
4) conditions a), b), c) hold with respect t o every proper inertial frame of the charged sphere and with respect t o
a particular external inertial frame,
we fall exactly on the wellknown rigidity conditions introduced by BORN(1909)in relativity theory. In this sense
such a rigid body has only three degrees of freedom: the remaining three degrees of freedom of the spherical shell
being determined, as shown by POUNDER (1g54), by requiring that BORN’Srigidity conditions are satisfied on the
surface.
.
Fig. 4. In this model aparticle whose boundaries are denoted by “and centre motion z by - - moves along the drift average four velocity
- a
ud. It undergoes stochastic jumps a t the velocity of light w from point I to point 2, after which it reintegrates the drift flow. A and B
(as A’B‘ and A”B”) represent the opposite extremities of a particle’s diameter.
BORN’S “rigid” shell evidently implies transmission of superluminal interaction and/or information since such
a shell travelling a t (or every close to) thevelocity of light implies knowledge and interaction which crosses the sur-
face of the light cone. Indeed knowledge of the position A on the shell (or of its centre z ) implies, as shown in fig. I,
knowledge of B on the opposite end of a diameter, and a collision (or interaction) a t A’ which switches the hypertube
(and the path of 2) back into the forward light cone (into A” and B“) implies a deflection of the path of B’ beyond
the light cone’s surface, i.e. a nonlocal modification of the physical situation a t B’. Of course such nonlocal inter-
actions can be neglected in classical relativity theory since the actual size of the particles cross-sections have been
shown t o be very small (perhaps of the order of PLANCK’S length, i.e. -~ so that the introduction of a n
I O - ~cm
extended structure can be approximated by the statement that their centre-of-mass and centre-of-matter density
associated with extended particle models are restricted t o the forward light cone.
Fig. 5.
64 J. P. VIGIER: Non-Locality, Causality and Aether in Quantum Mechanics
The distribution of superluminal interactions inside such a rigid body can be easily described in a covariant way.
Assuming the shell described in its geometrical center’s rest frame, we can pass through this center o a distribution
of tachyons with a constant surface density on the pseudo mass shell @,,pfi = c2 which is a timelike surface. A Lo-
RENTZ transform rn aintains this distribution (fig. 5 ) which remains unchanged on any space-like section.
4. Action at a distance and causality in relativistic mechanics
We now revisit, for a more detailed analysis, the connection between causality and locality in relativistic
mechanics. This is necessary, as we shall presently see, t o clarify the discussion of quantum mechanics.
First let us recall the general result that unless one introduces specific constraints on nonlocal interactions the
corresponding action-at-a-distance breaks causality and implies retroactions in time. This is shown in the following
fig. (6).
I;ig. b.
Let us consider two relatively moving observers 0,, 0, with respective rest frame S, S’. At the event el, 0,
sends (in its relative future) a superluminal signal t o 0, which absorbs it a t E $ ; after some time, at the event E ~ 0,
,
sends (in its relative future) another superluminal signal t a 0, which absorbs it at e4. I t is easy t o verify (MBLLER
1962) that we can always arrange this experiment so that E~ precedes cl so that we can use superluminal signals
in order t o modify the absolute past of O,! In other terms, as can be checked easily, the criss-crossing of space-like
paths allows the transport of positive energy backward into time from el t o E ~ .
Moreover the order in time of events E, and E, (or &,*andc4) depends on the external observer. This raises the
specific problem of the existence of relativistic constraints which could preserve the sequence of events on two
timelike paths despite the existence of superluminal interactions. One deduces from ZEEMAN’S theorem that this
implies that superluminal interactions which preserve causality cannot allow the forbidden combination of fig. 6.
This problem (which is also not new in the literature (MARTIN1973) has recently received a satisfactory solution
by DROZ-VINCENT (1977) in the hamiltonian formalism. I shall summarize his argument for two particles since his
line of reasoning can be easily generalized for N particles. In this formulation, the hamiltonians are just the genera-
ting functions leading t o the equations of motion. These equations admit a POISSON bracket form which involves N
independent parameters (the proper times or a suitable generalization of them).
The masses are not taken as constant a priori, but rather considered as constants of the motion. Accordingly,
our two-particle phase-space has sixteen dimensions. We consider the canonical coordinates q, p , q’, p‘ where q and q’
are points in MINKOWSKIspace, while p and 9’ are four vectors. These canonical coordinates satisfy the standard
POISSON bracket relations. Beware that, due t o the so-called “No interaction theorem” (CURRIE1963, 1966; CURRIE,
JORDAN and SUDARSHAN 1963, the q, q‘ cannot be confused with the positions x, x‘ when interaction is present
(DROZ-VINCENT 1977, 1979). When possible, greek indices running from o t o 3 are omitted. For instance p stands
for PIC,etc. Scalar products are written in compact from :
p 2 =$ * p =pap., etc.
J , P. ViC;ihK : Non-Locality, Causality aiid Alethcr in Quaiitutii Atcclialiics 65
We separate external from internal variables by setting r = x - X I , z = q - q’, p = p $’, y = I/z($ - p‘). +
Application of the projector LIZ = SZ - PaPp/P2t o any object will be noted -. For instance P = IT;,@. The energy
is the time component of P : but t,he hamiltonians (who generate the motion in manifestly covariant form) are
phase-space scalar functions H and H’.
Interacting models are obtained when the free hamiltonians :
Ho = a
H‘ = i p ‘ 2 p 2 (4-1)
are completed with additional terms. They are constants of the motion and identified with I/Z of the squared
masses. Thus, on the orbits of the system, we have numerically:
H = H’ = r m p 2 (4-4
But, as phase-space functions, the Hamiltonians can be written, in a general way, as:
H = Ho V W + +
H’ = HL V-W + (4-3)
In contrast with the lion relativistic case, the nonlocal “relativistic potentials” V +
W , V - W , have the dimen-
sion of a squared mass and cannot be chosen arbitrarily.
In order t o insure the existence of world-lines (CUFARO-PETRONI, DKOZ-VINCENT and VICIER 1981 ; DROZ-
V i N C m - r 1975) it is geometrically necessary that :
{ H , H ’ } = const. (4-4)
tlie { } denoting POISSON brackets.
Moreover the physical interpretation of H and H‘, as well as, independently, the requirement of symmetry
under particle exchange, demands that in fact:
{H, H’} = 0 . (4-5)
In practice eq. (4-5) is a condition on V and W .
In order t o make easier the forthcoming calculations we shall use the combinations 4(H H ’ ) - p z and +
H - H‘. Assuming that V and W are Poincare invariant, {H, P 2 } and {H‘, P 2 )vanish. Thus we can replace eq.
(4-5) by the equivalent condition :
(4(H + H‘) - P2,H - H’} = 0 . (4-6)
Using the obvious identities:
+
4(Ho ff;) p 2 = 4Y2 (4-7)
Ho - H: = y.€’ (4-8)
we get:
4(H + H’) - pz = 4y2 8V + (4-9)
H - H’ =y.P ZW+ (4-10)
into (4-6), performing the calculation, we can finally write the condition :
Inserting (4-10)
{y2,W } + {V,y.P>+ Z{V, W } =0 (4-11)
wliicli is equivalent t o (4-5).
A large class of solutions can be obtained by setting W = 0. This particular situation will be referred t o as the
single-potential case. This case deserves special interest because of its simplicity: then the condition (4-11) is com-
pletely solved (DROZ-VINCENT 1977, 1979) by requiring that V depends on Z2, p2,j 2 ,2.5,y . P but does not depend
on z. P which is just, in the rest-frame of the system (DROZ-VINCENT 1977, I979), the relative coordinate-time, up to a
factor ]PI. In the particular case where m = m‘ the external coordinates represent the c center of mass and the supp-
lementary condition means that the relative four momentum y is in the particles rest frame since (with W = 0)
we get y . P = 0.As we shall later see this is satisfied, for two quantum scalar particles, by the superluminal quantum
potential.
The prec~~diiig condition5 (4-11) evidently iniply causality (which is thus disjointed from locality) siiicc :
’1) the scalar hamiltonians H , H’ are separately invariant under the causal group P,
B) everything goes as if the two particles moved causally (under each other’s influence of course) along their paths
as a consequence of {H, H ’ } = 0. One has thus reestablished, in relativity theory, tlie causal mechanism of
classical mechanics.
5. The BOHR-EINSTEIN
Controversy
We now come t o the present theoretical implications of recent experiments on the E. P. R. paradoxon. With
the hindsight of history we can subdivide the history of the BOHR-EINSTEIN controversy in two stages:
The first stage started with EINSTEIN’S attempt t o demonstrate the incomplete character of quantum mechanics
i.e. t o justify the possibility of subquantum hidden variables which would:
I) explain the statistical laws associated with wave mechanics
2) interpret corresponding distribution of point like quanta in measuring devices.
This attempt materialized in the famous 1935 paper of EINSTEIN,PODOLSKI and ROSEN(1935) went beyond
this particular objective. It raised indeed not only the completeness problem but also the local character of quantum
5 .Ltiuuou. Ndchr. Ud. 303,Il. I
mechanics. Wether EINSTEIN realized this a t the time (or not) is a question I shall leave t o the historians of science.
All I know is that the E. P. R. paradoxon, first proposed as a Gedanken experiment, has (since 1935) played the
part of a delayed time-bomb left behind b y EINSTEIN for the benefit of his opponents.
After EINSTEIN’S defeat a minority of physicists generally stuck t o EINSTEIN’Sprogram of developing local
hidden variable models t o account for quantum statistics. Well known attempts by DE BROGLIE(1972) (pilot-wave
then double solution) BOHM(1952)~BOHMand VIGIER(1954) were successively proposed with various connexions
with relativistic hydrodynamics (HALBWACHS, SOURIAU and VIGIER1961) and stochastic theory (VIGIER 1979 ;
CEROFOLINI 1980). It is not my purpose t o analyse them here. Many aspects of these attempts are still used and/or
revived up t o this day. Let us just say they acted as a stimulant t o the crucial theoretical controversy by the disco-
very that any local hidden variable theory implied measurable consequences in E.P.R. types of measurement (i.e.,
more precisely, in joint but separate measurements made on individual particles emitted by the same source in
correlated singlet states) which could be compared with the predictions of quantum theory. This led
A) t o a heated theoretical discussion
B) t o various experimental tests.
A) To cut a long story short let us recall
a ) that BELL’Sinequalities (which limit observed correlation in E.P.R pairs) are valid for all local theories (as further
shown by SELLERIand TAROZZI (1979); CLAUSERand SHIMONY (1978) so that their disproval, as shown precedently,
implies a t least the introduction of non-local variables in hidden variable theories
b) that here is now a general consensus that quantum mechanics implies non-locality (D’ESPAGNAT 1978) (i.e. super
luminal correlation between two measuring devices) in E.P.R. measurements.
This is a fundamental step because BELL’Sresult has thus opened a new deeper stage of the E.P.R. contro-
versy. Everything goes as if the two great theories on which physicists have fed side by side, since fifty years have
now reached an antagonistic stage. This is clearly summarized by SHIMONY (1980) in a striking sentence “If the
experiments of ASPECTand RAPISARDA confirm (as I believe) the predictions of quantum mechanics we shall be
obliged t o drop EINSTEIN’S criterion of reality or t o accept some form of action-at-a-distance”.
I n other terms the development of the discussion since 1965 has deeply modified the initial BOHR-EINSTEIN
controversy. Their initial starting point (wether quantum statistics are complete or not) is now superseded by the
locality-causality problem. The discovery of BELL’Sinequalities implies a development (or a modification) of BOHR’S
measurement theory and of hidden variable theories. If confirmed the existence of non local correlations, it
implies, if one wants t o remain within MINKOWSKI’S space-time, a modification of the two antagonistic lines of re-
search which started a t the Solvay Congress. Faced with the consequences of BELL’Sinequalities the disciples of
BOHRand EINSTEIN’S followers, cannot remain on the initial positions of the antagonists. They must create some-
thing new or drop their initial interpretations of quantum mechanics.
B) The two E.P.R. consequences a) and b) in A can (after 45 years) be checked independently. Naturally
enough people started with point a) i.e. the experimental tests of realistic local theories via BELL’Stheorem.
The principle is simple and starts from a variant of E.P.R. proposed by BOHM(1951) who suggested t o measure
the relative orientation of polarisations of two correlated photons emitted in the singlet state. The corresponding
experimental device is described, in principle, in the experimental device shown in (fig. 7).
X
de&tor
8
Fig. 7.
That this device allows a comparison of local hidden variable theory versus quantum mechanical measure-
ment prediction can be simply shown by their confiicting predictions. Indeed “if one emits two correlated photons
y A and y B emitted from a single source with opposite helicities from which one observes the coincidences after passing
through two linear polarizers PAand PB(fig. 7)”, one sees that the rates of coincidence which depend on the angle
0 = OB - O A of the two polarizers must satisfy (once the photons are separated) different relations“ (ASPECT,
GRANGIER and ROGER1981).
This can be illustrated by a simple space-time drawing which illustrates the preceding scheme (fig. 8).
The conflicting interpretations imply:
- For local hidden variable theories that the dices are cast in S so t h a t the phenomenon is determined by local
variables in S, P A and Pg. Any observed correlation depends on their respective values in any separate given ob-
servation the correlation distribution resulting from their “hidden” stochastic distribution.
- For the Copenhagen Interpretation that the dices are not cast in S. The photons have no “hidden” polarisation
in S but acquire polarisations when they are observed in PAand PB(at times t A and t B ) despite the fact that these
measurements are separated by a space-likeinterval. In EINSTEIN’Sterms for BOHR,“God plays dices” in A and B,
a n idea he vehemently rejected. In other terms when the E.P.R. paradoxon was first proposed nobody realized that
J. 1’. VIGIEK:Non-Locality, Causality and hether in Quantum Mechanics
the experimental rejection of Bell’s inequalities and the verification of quantum mechanics in this case implies not
only the rejection of local hidden variables but also a strong indication in favour of the existence of quantum non
local interactions between measuring devices.
A set of increasingly precise experiments started from BOHM’Sproposal (WILSON,LOWEand BUTT1976;
BRUNO, D’AGOSTINO and MARONI1977; FREEDMAN and CLAUSER1972; FRY1973); with two exceptions (FARACI,
GUTKOWSKI, NOTTARIGO and PENNISI 1974; HOLT1972) they confirmed Q.M. The last one, just finished by ASPECT,
GRANGIER and ROGER(1981) definitely kills BELL’Sinequalities. Using a device similar t o the CLAUSER-1- ’REEDMAN
experiment they have obtained results in excellent agreement with the quantum mechanical predictions (which
strongly violate generalized Bell’s inequalities) and thus rule out the whole class of realistic local theories. They
are summarized in the following figure 9.
One must add here an important remark: ASPECT’S experiment goes beyond former experiments because no
significant deviations from quantum mechanics was observed when source polarizer separations were increased up
to 6.5 m, i. e. when the wave packets of each element of the photon pair were clearly separated in space.
The next stage of ASPECT’S experiment (also performed on similar lines by Rapisarda) is t o complicate the
scheme of fig. 7, by two switches AL and AN which fluctuate a t random in a time interval <
L/zc so that the
emitted photons choose arbitrarily their polarizers. If experiment then reproduces the result of fig. g the success of
the quantum mechanical prediction implies the existence of non local (superluminal) correlations between L and N .
This raises a formidable problem. Everything goes as if the polarizers could not be correlated since they are
outside of each other’s light cone. If established, this only leaves the alternative antagonistic answer of fig. XI. What
is a t stake is the problem of causality in physics. Indeed if one destroys it a t the microscopic level one knows it
vanishes at the macroscopic level since one controls amplifying techniques (triggers) good enough t o explode an
atomic bomb with an individual trigger-photon. Up t o now only two alternative interpretations have been proposed
0 90 780 270 360

Fig. 9. Normalized coincidence rate as a function of the relative polarizers’ orientation. Indicated errors are: d standard deviation. The
solid curve is not a fit t o the data but the prediction by quantum mechanics.
5’
t o interpret the non local superluminal correlations between the pairs of polarizers which iiicasurc: tlie idioton pairs
emitted from the source S., ie.
a) the propagation backward in time from PAt o S (by an advanced wave) followed by a retarded wave from S to-
wards PB (COSTADE BEAUREGARD-STAPP, CAPRA,MATTUCK, JOSEPHSON et al.).
b) The direct action at-a-distance between PAand PBvia a superluminal propagation of the quantum potential
(BOHM,VIGIER,HILEY,CUFARO-PETRONI,GARUCCIO, POPPER, DROZ-VINCENT et al.).
They are represented in fig. 11 and will be discussed in the next sections.
Fig. 11.
6. A Copenhagen Interpretation of non-locality
A) We now analyse t o the so-called Gottingen-Copenhagen interpretation 6.e. the present WIGNEK(1967),
STAPP(1g77), JOSEPHSON (1g79), COSTA (1977) extension of BOHR’Sinitial ideas) t o interpret non-locality in the
measurement process. Recent discussions have clarified the problem. It can be summarized as follows. Starting
with a source S emitting correlated pairs of particles (i.e. pairs of fermions or photons measured in A and B) one
knows quantum mechanics predicts non local correlations between A and B so that the dices are cast in A and B
(and not in S) where and when a measurement is performed. Now COSTADE BEAUREGARD adds (1978) that if he uses
the FEYNMAN propagator D,for two fermions one knows that this implies that these measurements are correlated
via their common past in S. This is certainly not correct. Indeed if one uses the FEYNMAN propagator
D, = 2 ( 2 ~ )$$J$
- ~ d4k exp [ik%,J S+(k,,k” +ka)
J . P. VIGIER:
(i.e. D, = B - (I/z) ( D , +D - ) t o describe the evolution of particles (and antiparticles) only positive (negative)
energies can move forward (backward) in time. . . so that (if experiment confirms non-local interactions between
the quantum measurements in A and
B) one cannot correlate these measurements via their p a s t . . . unless one adds something t o D,.Indeed D,
univocally implies the arrival in S of negative energy coming from A(B) followed by positive energy moving for-
ward t o B(A). This is equivalent t o the creation of particle-antiparticle pair in S . . . so that the corresponding signal
t o A and B would cast the dices in S, and yield BELL’Sinequality, thus destroying the quantum mechanical predic-
tion. The propagator D, thus does not explain and/or interpret non local interactions between A and B . . . so that
COSTAmust (explicitly (1978)or implicitly) add Djp t o D, t o interpret measurement theory with
D(x) = -(~n)-~
JJJ d% exp [ikkvp]~ ( k )
and must assume (fig. 12)that positive energy can flow from A t o S and S t o B. ThiB anti-telephone mechanism is
certainly not contained in D,as indicated by BELLa t the Geneva conference.
S
Fig. DE BEAuREGARn’s picture of
12.COSTA the propagation of quantum influences from the space-time point A to the point B. An anti-
telephone mechanism.
Contrary t o a statement of COSTA DE BEAUREGARD (1980) the (ac, ct) and (k,r/c) wave pictures are tied in the
sense that t and E must have the s a h e sign in the FEYNMAN picture: despite the fact that they cannot be focussed
simultaneously as a consequence of HEISENBERG’S uncertainty relations. Indeed the orthochronous LORENTZ trans-
forms and their discrete automorphisms transform them simultaneously. . . so that they represent identical states
when one passes from configuration t o momentum space. Retrocausation thus implies positive energies moving
backward in time.
One must remark a t this point that the fact that D, implies non local interactions between A and B but does
not interpret them is a straightforward consequences of BOHR’Sinterpretation of Q.M. which forbides any detailed
causal description of the measuring process. The Gottingen-Copenhagen proponents thus now go beyhnd BOHR’S
own ideas and this should be statCd explicitly (SELLERIand VIGIER1980).
Contrary t o COSTADE BEAUREGARD’S belief (1980)the use of D, implies that the description of particles (anti-
particles) as always moving forward in time with positive energies and/or antiparticles (particles) as moving back-
ward with negative energies is not arbitrary even in the case of bosons. Indeed if one builds bosons with fermion-
antifermion pairs the particle-antiparticle transition of both components transforms a positive energy boson moving
forward in time into a negative energy boson (physically equivalent t o a boson) moving backward in time. I n this
sense the use of D, in boson theory implies that only positive (negative) energy particles can move in the forward
(backward) time directions. D, thus implies the second law of thermodynamics and EINSTEIN’S causality.
C) We have not wrongly accused COSTADE BEAUREGARD (1980)of dropping D, in the initial pair emission: but
only remarked (as a consequence of I) and 2)) that if one wants (following STAPP(1977)and COSTADE BEAUREGARD
(1978))to use the ASB (or BSA) time zigzag to connect A and B one must replace or add t o D, either the JORDAN-
FAULI(Djp = D , + +
D- = Dadv Dmr) or the “unphysical” (C. DE BEAUREGARD 1980) anti-FEYNMAN D,, =
D + +
(I/z) x (D+ D-)propagators which both allow positive energy popagation into the time-like past. This of
courqe
- breaks causality and the second principle of thermodynamics a t the microscopic level;
- breaks, in S, EINSTEIN’S energy-momentum conservation law and allows the “miraculous” appearance (i.e.
“creation”) of,new energy coming from the future;
- breaks the usual concept of experiments since they pave the way t o possible retropsychokinetic experiments and
various URI-GELLERtypes (C. DE BEAUREGARD 1980)of parapsychological phenomena.
- implies the possibility to interfere with one’s own past :
a passibility which certainly represents a miraculous causal anomaly. We can illustrate this with a simple drawing
in fig. I3 (HALBWACHS, SOURIAU and VIGIER1961)in which we have added COSTADE BEAUREGARD’S antitelephone
mechanisms,. . . which evidently allow the classical relativistic forbidden situation,of fig. 13.
Fig. 13. Causal anomaly
Antitelephone mechanisms acting in S, and Sh induce an action of on E~ To quote HEISENBERG (1964)this

implies “that the law of causality does not apply in quantum theory and the laws of conservation are not longer
valid in elementary particle theory”.
7. Non-local causal description of interacting particle pairs

in the stochastic interpretation of quantum mechanics
I n this section, following EINSTEIN’S line of thought, we are going t o develop an alternative causal realist
description of the statistics of quantum mechanics in terms of subquantum random phenomena.
In EINSTEIN’S world there is no room left for unpredictable (i,e. true chance) phenomena. The theory indeed
predicts that statistical phenomena and the laws which govern them are a special perfectly objective subset of the
world machine i.e. the subset of phenomena, governed by a set of complex uncorrelated causes. Despite the often
discussed dialectical unity of change and causality, LAPLACE’S and MAXWELL’S classical materialist outlook clearly
require non statistical processes t o underlie all phenomena in physical science. I n modern time and modern language
we must say, with VON MISES,that objective probabilities (which have nothing t o do with the observer’s ignorance
of the outcome of any particular event submitted t o stochastic causes) are just the limit of the objective frequencies
resulting from a complex set of uncorrelated (or loosely correlated) causal phenomena. In other terms given a suffi-
cient knowledge, statistical processes are a straight-forward consequence of causality; if one assumes (as one should)
that causality also contains the causal effe6t of an unlimited set of loosely correlated causes. As everybody knows,
this concept of probability was the root of EINSTEIN’S criticism of the Copenhagen School and the starting point of
all research and models of possible hidden variables in the causal interpretation of quantum mechanics.
we thus start with EINSTEIN’S&,field i.e. with a real subquanta1 vacuum alive with fluctuation and random-
ness and fill it with a material covariant vacuum (DIRAC’S aether (DIRAC1951,VIGIER1980))hence forward called
“DIRAC’S vacuum” on which will propagate the localized particles and DE BROGLIE waves of the SIQM.
Using EINSTEIN’S “Aether” as basic stage we can fill it with actors (extended localized particles) in two steps:
a) The first step is t o fill EINSTEIN’S
“Aether” with covariant isotropic stochastic distributions of the three types of
classical spinning extended particles discussed above, i.e. t o fill EINSTEIN’S“Aether” with DIRAC’S spinning, co-
variant, distribution (VIGIER 1980).This is necessary since we thus built real stochastic distributions which mate-
rialize the covariant counterpart of the quantum vacuum states. Being covariant DIRAC’S distributions are unobser-
vable i.e. devoid of spin and qantum numbers. More precisely we introduce within each microscopic space-time
volume a distribution of extended structure given by fig. 14.
b) In the second step we now describe observed matter (DE BROGLIE waves and concentrated particles) as represen-
Ping collective excitations travelling on the top of DIRAC’S“Aether” which can be approximated by real relativistic
waves carrying soliton like concentrated particles i.e. comparable t o moving oscillators surrounded b y their own
real sound wave. This is a relativistic generalization of the hydrodynamical model of QM developed by MADELUNG
(1926), TAKABAYASI (1952)and extended t o spinning particles by various authors (HALBWACHS 1960).
One can compare this process with a plane flying a t Mach I within its own sound wave: the waves themselves
(thus very different in nature from ordinary sound waves) can be best described as organized “spin waves” propa-
gating on a chaos of the spinning classical J = 0 , J = 112 and J = I tops which now constitute DIRAC’Svacuum
with spin. The wave equations are now deduced (and not postulated as usual) from our stochastic model, and shown
J. P. VIGIER: Non-Locality, Causality and Aether in Quantum Mechanics 71
invariants of the POINCARB

Fig. 14. The shell’s correspond to the two CASIMIR group, i.e. p,p = -macz and S,SP = constant >.o
to describe waves of the three t y p e of spinning tops propagating over DIRAC’S“vacuum” in EINSTEIN’S “Aether”
with random fluctuations a t the velocity of light. If we analyse their hydrodynamical representations, we see
(HALBWACHS 1960) that their corresponding (DE BROGLIE’S “pilot”) wavelike solutions all satisfy DE BROGLIE’S
“phase correlation principle” : a principle characterized b y the fact that all internal (zitterbewegung plus spin) ro-
tations of the associated extended particle (which follow the drift lines of these DE BROGLIE“pilot” waves) rotate
in phase with the fluid elements which surround them.
I will not come back here on the thermodynamical analysis of the DIRAC“vacuum” (Aether) and DE BROGLIE
waves (VIGIER1980) but just recall the essential property that the quantum potential (which now represents the ran-
dom vacuum’s action on the drift lines of the wave equation) propagates with superluminal velocities. To show this
we will limit ourselves t o the J = o case since the introduction of spin complicates but does not alter the result. We
proceed in two steps:
As already indicated (VIGIER1968, HALBWACHS 1960) the stochastic wave equation can be written on a RIE-
MANN surface iPv as:
which can be split (with y = R exp (iS/b))into:

ar(R2 I/--gp
a,s) = o
and
afis avs + ~ : C Z - h2(nR/R)= o
iflv
which can be interpreted with the help of the unitary four velocity up = a,S/Mo with the “variable mass” ME =
= (mz - (h2/c2)(@R/R)and the scalar density e = Jq. R2(Mo/mo) as the HAMILTON-JACOBIequation:
p a,s avs+ M2, = o
and the continuity equation = de/dt = a,(eu”) = o where d(M,u,)/dt = -ap(Moc2) now represents NEWTON’S
equation. The stochastic force (per unit of variable mass) is then given by
K,,= -9 a,(ln M,)
and one sees that:
I ) The KLEIN-GORDON fluid satisfies the first two laws of thermodynamics. This can be shown directly from the
decomposition of KLEIN-GORDON energy-momentum tensor explicitly performed by HALBWACHS (1960). Indeed
he has deduced from the Lagrangian 2 = app* ahp - (mzc2/hz)y*y the expression
+ +
tpv= e ~ , , ~ vq/1zlV qvm, @
,,+
where q, denotes a heat current (orthogonal t o the conserved current j , = em,,) with up@ = -c2. From avt,,, = o
+
one immediately obtains introducing the scalar wc4 = tasuauBand writing w = e(& mc2) (where E represents the
internal energy) exactly ECKART’S form (ECKART 1940)of the first law i.e.
mDg + +
(I/c)[a=qa qa6a +
w4 a a v q = o
where mu, = t&&$ with qflv= S, +(mpuv/cz).
As noted by DE BROGLIE the second law is a straightforward consequence of -slk = A/h, since the entropy S is thus
maximized along the fluid drift lines: if we write for the action A = S with S = -M,,c2.
2) The heat current qa takes the form qu = -GqaB as@, where G represents the local thermal conductivity of the
KLEIN-GORDON fluid and 0 the local (probabilistic) temperature density 0 = T . R2 = (mc2/k) R2. Indeed its-
calculated form (HALBWACHS 1960) :
qu = e(h2/M:) -a, In R .qPva v In R .
formula (ECKART
can be transformed into ECKART’S 1940):
-
qa = (kh2/zMo moc2)uRa, (In R) qpvavo= Gqpyavo.
We now come t o the two particle case and are going to discuss the E.P.R. paradox on the basis of the very
probable result that the forthcoming experiment of ASPECTand RAPISARDA et al. (1981) will confirm in the near
future the non locality predicted by quantum mechanics.
This is an important event, since many people still believe in the antagonistic character of non-locality and
causality. This belief is understandable since one knows (M~LLER1962) that unless one imposes particular restric-
tions on possible superluminal interactions one faces causal paradoxes tied with possible retroaction in time. Our
aim is t o discuss the relation of these two concepts in the particular case of two identical non interacting quantum
particles in order t o interpret causally the corresponding E.P.R. situation analyzed (in the non-relativistic limit)
by BOHMand HILEY (1975).
We first define what we mean with the word causality by three properties:
a) The system of our two particles can be solved in the forward (or backward) time direction in the sense of the
CAUCHYproblem ;
b) The paths of all material particles must be time-like;
c) The formalism must be invariant under the POINCARE group P = T @ 2 t .
As shown by DROZ-VINCENT (1979)~CUFARO-PETRONI, DROZ-VINCENT and VIGJER(1981) one can have action-
at-a-distance and preserve this causality in the following case: we start with the two free Hamiltonians H,,, = P32 =
= &/2, and Hm = PZl2 = mZl2, completed with additive interaction terms V , and V , which are non local potentials.
Note that:
I) We shall call V,, V, potentials for convenience, although they have not the dimensions of energy and must have
the dimensions of squared masses;
11) The Hamiltonians are not directly related with energy but rather with half the squared masses. We thus get :
i =
H Ho1+ V1, H2 == HOB +V2 (7-1)
now defined in the sixteen-dimensional phase space qf, q:, pf, 9:. One sees immediately that the potentials cannot be
chosen arbitrarly since the existence of world-lines requires, for identical particles, the vanishing of POISSON’S
brackets {H,, H , } (DROZ-VINCENT 1977). Phase space has 16 dimensions and the standard brackets are assumed
among &, & and unconstrained Pf, P:.
We now perform the following separation of internal and external variables :
so that, in the case that V, = V , = V and m, = m2 = m we have:

HI + H , = 4 P a + y 2 + 2V (7-3)
Hl- H,=yP
The condition for the existence of world-lines then reduces t o the relation:
( y P , v>= 0 . (7-4)
Jf we define the projector Il: = Sf - PpPv/P2and 9 = n f z v , 7” = Ilfyv, relation (4) implies that V depends on
2, P2,y2, 27, y P but does not depend on Z P which is in the rest frame of the system, the relative time coordinate
up t o a factor IPI. Moreover one finds {P”, H l } = {P”, H,) = o so that the center of mass momentum P i s constant
and one can slice space-time with 3-planes orthogonal to P p and connect the two particles by space-like lines in
these hyperplanes.
We now come to the description of two quantum non interacting particles. For a system of two classical re-
lativistic particles interacting (DROZ-VINCENT 197g), is described by two parameters: zl,z,, i.e. the proper times
of the two particles. The movement is generated in the phase space T(M,) x T(M,) in a symplectic way by the CO-
variant Hamiltonians H , and H , analyzed in the first part of this section. Of course we can build the canonical trans-
formation theory in this covariant framework (HIRONDEL 1977). The transformation which solves the motion equa-
tion is generated by the JACOBI’S principal function S but it is simpler t o consider the covariant HAMILTON- JACOBI
+
characteristic function W = S - (m2/z)(zl z,)which is determined by the HAMILTON-JACOBI system (Hirondel1977)
4
One remarks here (in accordance with the well-known no-interaction theorem (CURRIE1963) that the canonical
variables &, qt are not coincident with the positions x:, x; except when the interaction vanishes.
By straightforward quantization of this multitemporal canonical formalism we obtain, for a system of two
free particles, the KLEIN-,GORT)ON system (for h = c = I)
J . P. VIGIER:Non-Locality, Causality and Aether in Quantum Mechanics 73
where y is a two points dependent function. Out of (6) we can extract the usual main equation:
(nl+ 02)y(xl, x2) = 2m2y(x,,x2) (7-7)
completed by the so-called “subsidiary” condition :
(01 - n2)Y = 0 (7-8)
We now introduce in a relativistic way the concept of quantum potential (BOHM, VIGIER1958, DE BROGLIE
+
1964). Following the original DE BROGLIE’S method we set y = exp (R i W ) where R, W are real functions. Se-
parating eq. (7) in real and imaginary part we get for the real part
where we have:
I n spite of a n obvious analogya) the system (9) cannot be immediately identified with eq. (5).To be more specific we
will consider the case of a y eigenstate of P, = i(af 3:): +
= exp i(K, (x: + x:/z)) p(z”); so, where K , is a constant timelike vector, we have:
(a: + at) R 0 , (a; + a:) w = ~ f i (7-4
Moreover the difference of the equations (9) gives:
so that R only depends on ZP = x; - xg and more precisely only through its spatial part with respect t o KP, namely
z p = zfl - (zvky)KP/K2. I n this case from (11)we have U, = U2 = U = f ( z T ) . But, as seen before, U has not a
suitable expression because it depends only on zI and it cannot satisfy the condition (YP,V } = 0. In fact this pro-
cess gives U as a function of z” and KF and not of z9, PP.
Making the substitution :
Z f i -+2 (7-13)
in U we get finally V = f(Z) which depends on Pfi in a correct way so that we can interpret it as a relativistic
potential.
Equations:
I
- (a,,w
2
a p v ) + v(q = --,
2
m2 1 (7-14)
are now coincident with (5) if qp = x, i.e.:
We remark here that the variable xf = qf:and x$ = qg are canonical for the free quantum system as well as for the
classical interaction system. Moreover they are also position variables for the quantum free system but )hey do not
represent the positions in the classical interacting system except in the particular rest frame system where we recover
the HAMILTON-JACOB] equations for a classical system in interaction through the potential v.
At this stage of our work, as was the case for the old DEBROGLIE’S derivation, we have only exhibited a themati-
cal analogy between a system of two quantum free particles and a system of “fictitious” but causally interacting
particles. We are going now t o recall and summarize the physical interpretation of this fictitious system (in theframe-
work of the stochastic interpretation of Quantum Mechanics) in two steps.
I) Our starting point is just the two-particle generalization in configuration space of our one-particle model
(GUERRA and RUGGIERO 1978, VIGIER1979, LEHRand PARK1977). Indeed let us assume two identical scalar particles
labelled I and z imbedded in DIRAC’S stochastic “aether” (DIRAC 1951; VIGIER1980). The pair’s motions along any
world line, in configuration space-time, build a fluid in this space-time. These motions are not independent (since
the presence of particle I disturbs the “aether” i.e. the motion of particle 2 and vice-versa) and one assumes that we
are dealing (as in the one-particle case) with stochastic jum ps a t the velocity of light (in physical space-time) which
pass the pair I, z from one drift line of flow (in configuration space) t o another. Physically this amounts (in the
hydrodynamical model of BOHM and VIGIER) t o the superposition space-time of two interacting fluids I and 2
which undergo light-like internal stochastic motions, particle-antiparticle transitions and possible number-preserving
transfers from one fluid t o another. . . so that we have a conserved scalar fluid particle density in configuration
space.
a) Of course eq. (5) are written in terms of qf, q$ while eq. (9)involves ~ $ 4
but; for the original free system the position varia-
bles are canonical so that we can write without problem q< = zy, q% = %; which makes manifest the analogy between (5) and (9).
Mathematically this model can thus be described (CUFARO-PETRONI and VIGIER 1979) in an eight-dimensio-
nal configuration space where a pair position is defined by an eight-component vector X i (i = I, ..., 8) where
{xi>;=,
,...8 {xf; xz”}p,v=o ,...3
with xf, x: four-vectors of the position of each body. The metric is defined by
1 0 0 0 0 0 0 0
0 - 1 0 0 0 0 0 0
0 0 - 1 0 0 0 o 0
0 0 0 - 1 0 0 0 0
0 0 0 0 1 0 0 0
0 0 0 0 0 - 1 0 0
0 0 0 0 0 0 - 1 0
0 0 0 0 0 0 0 - 1
so that
~2 = xixi= g&”xi = (x1)2 + (x2)2.
If x f ( t , ) , x!Jz,) are the trajectories for the two particles, the trajectory in configuration space will be an Xi(t,, t,).
As a consequence of NELSON’Sequation (0-1) we can now generalize the differential operators defined by GUERRA
and RUGGIERO (1978) for the single-particle case t o a system of two identical particles
a a h
I~ = - +at,- + +at,i a ~ , SD=Sbiai--

2m
0,
bi = OX,, Sb, = SDX, .

Now a direct extension of GUERRAand RUGGIERO(1978) formulae gives the following dependence of 6bi on a density
e(Xi, tl,t2):
h
dh, = --
m %
a. log e’lz , (7-16)
where for the density we have as continuity equations
with
I n our model, as a generalization of the assumption that e is independent of the proper time in the one-body case,
we make the physical hypotheses that the total number of particles (i.e. pair in the real space-time) is conserved, and
thus we write
-+-=
ae ae o,
at, at,
so that our continuity equations in configuration space are
ai(ebi) = o . (7-17)
11) We assume as before (GUERRA 1978, VIGIER 1979, LEHRand PARK
and RUGGIERO 1977) that our fluid motions
are irrational, so that
where +(Xi,
t,,t 2 )is a phase function, and, if we look for a steady state (i.e. proper time independent) equation,
Now it is clear that (as generally assumed and later demonstrated by CUFARO-PETRONI and VIGIER (1979)) NEW-
TON’S equations for’thetwo free particles can be written in the compact form
(DD - SDSD) Xi= O . (7-20)
Starting from I7,20 and using (16,-18, 19) we obtain a HAMILTON-JACOBI-type equation ( R = ex/*)for our two-body
svstem i.e. :
which yields for the continuity equation the form

2 a , ~ais + R a,ais = o (7-24
J. P. VIGIER
: Non-Locality, Causality and Aether in Quantum Mechanics 75
Finally if we consider (21)as the real part and (22)as the imaginary part the total equation for y =R exp [iS/h]is
(7-23)
which is equivalent t o relation (8). Moreover the model implies the causal condition (12) since in the rest frame of
particles I and 2 we have by the principle of action and reaction P, = - P 2 i.e.
(PI.
P, - P2.P2) = (P,+ P2)(P, - P2) = 0
-
These results lead t o an immediate causal, non-local, interpretation of the E.P.R. paradox which we shalI discuss
in the non relativistic limit.
From relation (23) one evidently deduces, in the nonrelativistic limit, the usual two-particle SCHRODINGEP
equation which (writing y(x,, x2,t ) = R(x,, x2,t ) exp [iS/h])splits into real and imaginary parts i.e. :
with P = R2 = y * y and
+
with Q = -(h2/zm)[(V:R/R) (VzR/R)].Clearly relation (24) represents the conservation of the probability
P = y * y in configuration space (x1x2) while relation (25) as discussed by BOHM and HILEY(1975) corresponds t o a
HAMILTON- JACOBIequation for two particles which interact through a nonlocal quantum potential Q with which
they have interpreted the first form of the E.P.R. paradox in its original position-momentum formulation. I n the
causal interpretation of this situation one adds of course that our particle momenta are described in real space by
V , = VIS/mon V2= V2Slm as the mapping of configuration space into real space suggests (CEROFOLINI1980).
Now as discovered by EINSTEIN and emphasized by BOHM and HILEY(1975) this is an entirely new unclassical
situation. Indeed a) the quantum potential Q(x,,x2,t ) does not in general produce a vanishing interaction between
the two particles I and 2 as x, - x2--f co so that space-like separated distant systems may still be in strong direct
interconnection b) The quantum potential Q cannot be expressed as an universally determined function of all the CO-
ordinates but depends on y(xl,x2,t ) as a whole including its real physical boundary conditions. This is classicaly
a fantastic situation since an unphysical probability wave now appears t o imply the exchange of energy (and in-
formation) between the two particles.
To show how this works let us return to the original Gedankenexperiment proposed by E.P.R. (31) (fig. 15)
In this proposal one considers two non interacting particles represented by the wave function
[. ‘z
y = e x p zP- x1 x2]6(x, - x2 -
which represents a state in which the total momentum is P and the two particles are separated by the macro-
scopic distance a. If one measures the position x1 then x2 = a +
x,. As remarked by E.P.R. this could perhaps be
explained by assuming that Pl and P2 were so correlated before the measurement but this contradicts the uncer-
tainty principle since one could have instead measured P, (and obtained the correlated result P2= P - P,) and we
know x1 and P, cannot be measured together. The only way out is t o say that somehow the measurement of p,
actually “put” this particle into a definite state of momentum PI2 while it “put” particle 2 into the correlated state
P - P,. This means that particle 2 seems t o “know”into which state it should go without any interactions that could
transmit this information which can evidently be made t o travel a t superluminal velocity by choosing suitable time
intervals t o measure P, and P2.This can be made clearer (though not less paradoxical) in the language of the causal
interpretation by substituting to the &function the “needle” function f sharply peaked a t x1 - x2 = a. Indeed the
corresponding p w a v e
y = exp [. ; Z x2] f(Xl - x2)
now interacts through the quantum potential Q = -(h2/zm) v2f(x1- x2)If(x1 - x2) so that any measurement (i.e.
momentum determination) of particle I must influence particle 2 through the quantum potential. One sees however
that nothing is gained in this way from the point of view of the contradiction with locality, since the distance a and
the measurement time of P, and P2 can be so chosen that this influence of Q would also have t o travel faster than
the velocity of light.
8. New experimental tests in the BOHR-EINSTEIN

controversy
In this last section I want t o stress the remarkable fact that the (now almost certain) discovery of non-local
correlations instead of weakening fatally EINSTEIN’S position in the BOHR-EINSTEIN controversy on the contrary
paves the way towards new experimental tests. . . which will help to choose between their conflicting points of
view. Indeed:
A) As one has seen their successors have now built two antagonistic interpretations, since recent discqssions and
experiments on the EINSTEIN-PODOLSKY-ROSEN paradox have led t o renewed confrontation between the (acausal)
Copenhagen (CIQM) and (causal) stochastic (SIQM) interpretations of quantum mechanics. In the E.P.R. case they
lead t o different mechanisms t o interpret possible space-like correlations between quantum measuring apparatus
operating on correlated particle pairs emitted in the single state. I n CIQM one is led (COSTA DE BEAUREGARD 1979)
t o introduce real advanced potentials and in SIQM phase-like collective propagations within DIRAC’Sstochastic
subquantal “aether” model (GARUCCIO and VIGIER1980).Despite their formal similarity (both interpretations use the
same wave equations), CIQM and SIQM differ on essential points.
I n CTQM:
I. The quantum states are associated with individual systems and represent an ultimate statistical knowledge.
Microphenomena are particles or waves, never the two simultaneously.
2. A measurement on a system provokes a discontinuous, space-like collapse of the state (wave packet reduction).
3. The HEISENBERG uncertainty principle restricts the simultaneous measurability of noncommuting observables
on individual systems.
In SIQM:
1‘. The quantum states represent real physical fields (waves) associated both with individual particles and ensembles
of identically prepared systems. Micro-objects are thus particles and waves simultaneously.
2’. These states (particles plus waves) evolve causally in time. Then there is no wave packet reduction, but they
are modified b y interaction with real physical macroscopic apparatus (CINI,DE MARIA,MATTIOLI and NICOLO1979,
MARIC and ZIVANOVI~ 1976).
3’. The HEISENBERG uncertainty principle does not restrict simultaneous measurability of noncommuting obser-
vables on individual objects (since particles follow world lines in space-time), but represent dispersion relations re-
sulting from their dual (wave plus particle) characters and from their subquantal stochastic motions.
3%)CIQM and SIQM lead t o conflicting testable predictions in particular experimental situations.
Indeed starting from these conflicting viewpoints
I) FITCHARD(1979)has recently proposed an experiment t o choose between 3 and 3’.
2) SELLERI
and VIGIER(1981)have devised a test t o choose between I and 1’.
Since they have both been already published I will limit myself here t o a third more easily testable proposal.
Assuming that ASPECTand RAPISARDA definitely confirm the existence of non local correlations the CIQM and
SIQM can be compared in an experiment proposed by GARUCCIO and VIGIER(1980),further developed by GARUCCIO,
POPPER and VIGIER(rg8r).Its aim is t o present an experimental test of the physical reality of the DE BROGLIE waves
since any proof of their independent existence can be reasonably considered as a strong argument in favour of the
SIQM. To that effect one must find a real situation in which CIQM and SIQM predict different testable experimental
results.
The following experiment has thus been proposed. It leads to conflicting predictions in the particular case of
interfering laser beams. Its principle is summarized in fig. (16).
MANDELand PFLEEGOR (1968)(M. P.) have observed interference figures with light originating in two inde-
pendent laser beams under conditions where the light intensity was so low that the mean interval between photons
was great compared with their transit time through the apparatus.
In CIQM this means a) that the photons are not emitted in one laser oaly; b) that the interference pxocess
results from the detection process since the localisation of the photon a t the detector makes it intrinsically uncer-
tain from which of the two sources it came. The correlation (coherence) of the phases in the sources thus expresses
the fact (DIRAC1938)that both lasers constitute (in CIQM) a single quantum mechanical source. In the SIQM
(following ANDRADE RE SILVAand DE BROGLIE 1972)one argues that the electromagnetic waves emitted by both
lasers are present (and interfere) even when only one (guided) photon is present during the measurement.
The initial proposal (GARUCCIO and VIGIER1980)was t o modify the apparatus of M.P. in fig. 16 by adding a
device which tells us from which laser the photons which appear in these interferences are really emitted. To that
effect one introduces along the beam of the first laser a semireflecting mirror M , which cuts its intensity in two
equal parts. The reflected part is then observed in a photomultiplier C. We then cut down the intensity of both beams
so that there can exist only one photon in each beam during the measurement. We then count signals coming from
A or B only if there is simultaneous signal coming from C , i.e. if a, b, c denote the signals coming from the photo-
multipliers A, B, C we measure N (a A c) and N(b A c).
From EHRENFEST’S theorem we then predict that one should and will observe the M.P. interference pattern.
Indeed in the quantum theory of light classical MAXWELLelectromagnetic waves are just the ly functions of Q.M.
It is not exactly so in the CIQM since the detection of the photon in C collapses instantaneously part of the
wave of laser I so that weakened interferences should be observed in principle. In the SIQM on the contrary one
expects stronger interferences because the states (particles plus waves) evolve causally in time. This yields inter-
J. 1’. I‘IGIER:
Non-Locality, Causality and Acther in (juantuni Mechanics 77
PHO?iiMUlT/PfIERA
1’1s. 16. Two light beams are derived from two independeiit single-mode lasers and their polarizations are aligned. After passing tliiougli
two attenuators, they are superposed with the help of the mirrors M3 and M4 on the interference detector R. By means of the unsilvered
glass plates MI and Mz, a smallportiop of the unattenuated beams is split off and passes through a pinhole t o a photptube whose func-
tion is t o register thedifference frequency of the two lasers and t o activate a gate for the interference detector for a zo-ysec period
whenever the beat frequency falls below 50 kclsec. The interference pattern is received on a stack of thin glass plates (insert) each of which
has a thicknesscorresponding t o a half fringe width. The plates are cut and arranged so that any light falling on the odd plate is fed t o
one photomultiplier (A), while light falling on the even plate is fed t o the other (B). One has added t o MANDELand PFLEEMR’S apparatus
a semitransparent mirror M,, which splits the beam of the first laser into two parts so that a new part is now reflected on mirror M5 and
detected on a photomultiplier C with which we can measure coincidences with photons arriving in the A and B photomultipliers.
fereiice in K due t o the superposition of tlic (photonless) transniittccl part of the MAXWELLwave and tlie plioton
carrying wave of laser 11.
Of course the preceding device can only be theoretically significant if we can be sure (i.e. with say a y ~ ‘ ; / ~
probability) that the second photon comes from laser 11. To quote HILEY(1981): “This is just what you cannot
control. It could equally have come from laser I and not be reflected by mirror M, but transmitted”.
The same point has been raised by DE MUYNCK(1981).If we approximate (as has been checked by experiment)
tlic two laser beams by the colierent state:
a,
la) 2
= e--ia-izlz
n=o
~
$i!
112) . (8-1)
The average number of observed photons is N = laI2 (with N = I in our case) so that the probability P(n) that
there are n photons in the beam in At is given by the POISSON
distribution:
(8-2)
with P(I)= e-1 and P ( 2 ) = e-1/2. Thus “the Probability of having one photon in each beam is P ( I )= ~ e-zN N
e-2 for N = I ) : wliicli should bc compared with the probability P ( z ) of having two photons coniing from 1
(i.cs.
<
(i.e. P ( z ) L e-I/z) so that we do not have P ( z ) P ( I ) ~. .. which makes difficult any measurement of CIQM-
SIQM differences”. To make sure that the CIQM indeed kills interferences (when coincidences between C and R are
observed) we now replace M.P.’s attenuator A and B by attenuators built with a succession (stack) of separated semi
transparent mirrors (with transmission coefficient 112)which cut down the photon flux t o N one photon a t a time
in the apparatus.
Indeed we are going t o show that if a POISSON mixture of probabilities 2 P(n)is emitted by our lasers this
ta
stack will enhance the isolated photon probability P ( I )in the transmitted wave (TW) and reduce P(z) ... P(n)
t o unsignificant values.
Our argument is that despit the fact that the RW (reflected wave) and TW contain (on the average) half the
photon number (intensity) of the I W (incoming wave) we can show (and will test) t h a t the TW (and RW) contains
a different proportion of isolated photons, photon pairs etc . . . and that M enhaces P(I)t o the expense of P(z) ...
P(n).The point is that for each mirror the incoming POISSON distribution is thus transformed (split) into the sum
of two POISSON reflected and transmitted distributions RW and TW. Moreover each photon arriving a t a mirror
reacts independently in phase and direction so that the stack preserves the phase correlation of the incoming laser
beam in its reflected and transmitted waves RW and TW. If we characterize the incoming beam IW before each
mirror by N tlie corresponding outgoing transmitted wave TW will be defined by N / 2 ... so that the ratios
P ~ w ( n ) / P ~ w=
( n(112)”
) -
eNL2will drop sharply with n.
78 J. P. VIGiER: Non-Locality, Causality and Aether in Quantum Mechanics
Let us then assume that the incoming POISSON distribution of laser I has been weakened down to N = I. If
we introduce a supplementary mirror we know that each photon on the beam (comparable a t this stage t o an ordi-
nary MAXWELLwave) behaves independently and has a fifty percent transmission (reflection) probability. . . so
that the (POISSON) RW and TW are now constructed with N = I/Z. This yields immediately P ( z ) = (e-I/*/z).
(112)~and P(I)= (e-II2) (I/z) i.e. PTW(Z)/PTW(I) = ( P I w ( 2 ) / P ~ w ( 1 )(1/2).
)
The same holds of course for all Plw/(n +
r)/PIw(n) ... so that from then on the intervention of M successive
mirrors enhances as strongly as one wishes the probability of finding one photon only. This results from the fact:
a) That the initial incoming POISSON distribution (or iplw wave) can be described by a linear superposition of indc-
pendent single photon waves with the same phase but regrouped as a sum 2 ylw(i) of waves containing I, 2, ..., i
photons. i
b) That each photon reacts independently on each isolated mirror (i.e. M,) which behaves like an elastic scattering
device which splits y ~ w
into a transmitted and reflected parts ~ T W and ~ R according
W to the Hamiltonian:
HYJ(1) = vca:ai + vip;ai
which yields a 50y0 traiisitioii into WW and y H W .
c) That the changed ratios P(x)/P(n) just result from the independent separation of the multiplet elements (photons)
which are scattered a t random with equal probabilities into ~ T W or yKw.One finds indeed for the ratios of incoming
and transmitted waves (which are described by N + N / z in relation (2) the ratios ( P ( n - I)/P(fz))TW/(l’(% I)/
P(n))rw= z since P(n - ~ ) / P ( g z )= n / N .
This justifies the existence of different CIQM-SIQM predictions. Indeed if one photon appears in the RW of
M there remains a small probability t o observe z , 3 ... photons in the TW but their contribution t o the interference
pattern is practically negligible.
This argument of course rests on two assumptions which should, can and will be tested separatly.
A) The first is that if one sends a strongly weakened laser beam on a. semi transparent mirror (so that like in M.P.’s
initial experiment only one photon a t a time is present in fhe apparatus) one observes anticorrelations between the
reflected and transmitted waves. Such anticorrelations have been observed by MANDELand DAj E N A r s (1978) with
incoherent sources but not in weakened coherent sources.
B) The second is that the ratios P ( I ) / P ( are ~ ) indeed weakened (according t o the preceding calculations) in the
same situation. With this new device we see that the simultaneous transformation of both initial (IW) states by the
filtering equipment implies that CIQM would predict weakened interferences when photon I is observed in C, since
the standard CIQM interference theory (RICHTER, BRUNNER and PAUL 1964) shows that the interference term (see
formula (4,4) in RICHTER,BRUNNER, PAUL(1964)) in an initial state given by z,p,, >
In is proportional t o zn
in+czc,,., and the term ~/TC;C, vanishes from the TW’s contribution if one photon is observed in C.
In other terms the collapse of the term c,~I) predicted by CIQM produces in this situation an observable de-
viation from the prediction of SIQM which remains identical t o the prediction of MAXWELL’Stheory.
One last observation. Nothing in our proposal prevents a modification of the relative lengths M,C and M,A(B).
In other terms photon I can arrive (at time t,) in C: a) in the past light cone (t, ~ A , B of < ) photon 2’s arrival in
photon 2’s arrival in A(B).

-
A(B); b) outside the past and future cones (t, ~ A ( B )of ) b’s arrival in A(B); c) in the future light cone (tc >
t A ( B ) ) of
In case a) (discussed above) SIQM predicts interferences CIQM does not. I n case b) the arrival of photon a in C
destroys in a non-local way the CIQM MANDEL-PFLEEGORinterferences which exist when C is not excited i.e. the
non excitation of C excites a positive interference phenomena. In case c) the vanishing of interferences predicted
by CIQM raises (in the author’s opinion) a new paradox. Indeed an event in C (photon 1’sabsorption) which occurs
after the arrival of photon z in A(B) modifies its behaviour. Moreover the non arrival of photon I in C re-establishes
the M P situation (i.e. observed interferences) while its arrival destroys it despite the fact that a t tA(B)there is then no
way t o know from which source the observed photon has originated.
Conclusion
We conclude with the remark that the causality implied in SIQM is absolute in the sense that the measuring
processes themselves (and the observers) satisfy the same causal laws and are real physical processes with antece-
dents in time. The measuring process (observer plus apparatus plus observed particles) is a set of particles which are
part of an overall causal process. I n this scheme the intervention of a measuring process contains no supranatural
“free will” or “observer consciousness” since quantum measuring devices act as spectral analysers which split into
subpackets the real DE BROGLIE’S waves associated with particles (which behave as planes flying a t Mach I within
their own sound waves): the particle entering into one of them according t o its random causal motion (CINI, DE
MARIA,MATTIOLIand NICOLO,1979 ; M A R I and ~ ZIVANOVIC, 1976) that scheme there are no “free will” signal produc-
tion and thus no possible causal paradoxes (MoLLER, 1962) : nothing exists beyond the motion and interactions of
material particles in a random stochastic aether.
This position evidently controdicts the prevailing opinion in theoretical Physics.
Indeed SIQM holds with EINSTEIN
A) that the flow of time is a real, irreversible and unidimensional phenomenon, which result sfrom the stochastic
character of the subquantum aether.
B) that only positive energies move in the forward time direction
J. P. VIGIER: Non-Locality, Causality and Aether in Quantum Mechanics 79
C) that the apparent microscopic time reversibility of the quantum mechanical wave equations only reflects the
particle antiparticle mixture of EINSTEIN and FEYNMAN which leads to correct perturbation theory.
The issues of the preceding discussion now lie in the hands of experimental physicists. The author of this report
believes that experiment will vindicat EINSTEIN’S views in the BOHR-EINSTEIN controversy.
References
Xsrac.r, A. : 1975, Pliys. Lett. 54, I I 7.

ASPECT,A.: 1976. Phys. Rev. D14, 1944.
ASPECT,A., GRANGIER, PH.and ROGER,Q. : 1981, Experimental tests of realistic local theories via BELL’Stheorem, Univ. Orsay,
Preprint.
COSTA DE BEAUREGARD, 0.: 1977. Nuovo. Cirn., B 42. 41.
COSTA UE UEAUREGARD, 0.: 1978, Phys. Lctt., 67 A, 171.
COSTADE BEAUREGARD, 0. : 1979, B 51, 267.
COSTA DE BEAUREGARD, 0.: 1979. Nuovo. CiIIl., 51 u, 267.
COSTADE BEAUREGARD, 0.: 1980, Nuovo. Cim. Lett., 17, 551.
COSTADE BEAUREGARD, 0.: 1980, Epist. Lett. 26, 11.
BEL, L. and MARTIN, J.: 1973, Phys. Rev., 8, 4347,
BELL,J. S.: 1964, Physics I, 195.
BLOCHINTSEV, D. : 1975, Fk. Eleni. Chastits. At. y&d.5, 606.
ROHM,D.: 1951, Quantum Theory, Prentice Hall.
BOHM,D.: 1952, Phys. Rev., 85,166, 180.
BOHM,D. and VIGIER,J. P.: 1954, Phys. Rev., 96,208.
BOHM,D. and VIGIER,J. P.: 1958, Phys. Rev., 109,882.
BOHM,D. and HILEY,B.: 1975, “Quantum Mechanics a half century latcr” edited by S. LEITE-LOPEZ and M. PATY.
BORN,M.: 1909. Ann. der Physik, 30, I .
DE BROGLIE, L. : 1964, Thermodynamique cachBe des particles, Gauthier-Villars. Paris.
DE BROGLIE, I,. : 1972, Une interpretation causale et non-lin6aire de la Mecanique Quantique, Gauthier-Villars. Paris.
BRUNO, M., D’AGOSTINO, M. and MARONI,C.: 1977, Nuovo. Cim. B 40. 143.
CEROFOLINI, G. F.: 1980, Nuovo. Cirn., B58, 286.
CEROFOLINI, G. F.: 1980, Nuovo. Gm. Lett. 29. 305.
CINI,M., MARIA,M., MATTIOLI, G. and NICOLO,F.: 1979. Found. of Phys. 9,579.
CLAUSER, J. F. and SHIMONY, A,: 1978, Rep. Prog. Phys. 41, 1881.
CURRIE,D. J . : 1963, J. Math. Phys. 4, 1470.
CURRIE,D. J., HORDAN, T. F. and SUDARSHAN, E. C. G.: 1963, Rev. Mod. Phys. 35,350.
CURRIE,D. J.: 1966, Phys. Rev. 142,817.
CUFARO-PETRONI, N. and VIGIER,J. P. : 1979, Physics Lett. A 73,289.
CUFARO-PETRONI, N. and VIGIER,J. P.: 1979, Nuovo. Cim. Lett. 26, 149.
CUFARO-PETRONI, N. and VIGIER,J. P.: 1981, Physics. Lett. A 81,12.
CUFARO-PETROKI, N., DROZ-VINCENT, PH. and VIGIER,J. I?.: 1981. Action - at-a-distance and causality in the SIQM, I.M.P.,
Preprint.
D~RAC P., A. M. : 1938, The Principles of Quantum Mechanics, 4th ed. Clarendon Press.
DIRAC,P. A. M.: 1938, Proc. Roy. SOC.,A 167. 448.
DIRAC,P. A.M.: 1951. Nature, 168,906.
DROZ-VINCENT, PH.: 1975, Reports on Math. Phys., 8, 79.
DROZ-VINCENT, PH.: 1977. Ann. Inst. H. Poincar6 27. 407.
DROZ-VINCENT, PH.: 1979. Phys. Rev. D 19. 702.
ECKART, C.: 1940. Phys. Rev. 58,919.
EINSTEIN, A. : 1924, Schweizerische Naturforschende Gesellschaft, Verhandlungen 85, 85.
EINSTEIN, A,, PODOLSKY, B. and ROSEN,N.: 1935. Phys. Rev., 47,7.
DIESPAGNAT, B.: 1978, Epistemological Letters (Lausanne), 19, 19.
FARACI, G.,GUTKOWSKI, D., NOTTARIGO, S. and PENNISI,A. R.: 1974, Nuovo. Cim. Lett. 9. 607.
FENECH,C., MOLES,M. and VIGIER,J. P.: 1979, Nuovo. Cim. Lett. 24, 56.
FITCHARD, E. E.: 1979, Found. of Phys. 9, 525.
FREEDMAN, S.J. and CLAUSER, J. F.: 1972, Phys. Rev. Lett. 28, 938.
I:RI‘.UERICK, C.: 1976, l’hys. Rev. L) 13,3183.
I~RY Ji., S.: 1973, l’liys. Rev. A8, 1219.
GARUCCIO, A. and VIGIER,J. 1’. : 1980, Found. of Phys. 10,797.
GARUCCIO, A., POPPER,K. and VIGIER,J. P.: 1981. Inst. H. Poincard, Preprint.
GUERRA, F. and RUGGIERO, P.: 1978, Nuovo. Cirn. Lett. 23. 529.
GUTKOWSKI, D., MOLES,M. and VIGIER,J. P.: 1977. Nuovo. Cim. B 39. 193.
HALBWACHS, F.: 1960, Theorie Relativiste des fluides i spin, Gauthier-Villars, Paris.
HALBWACHS, F., SOURIAU,’ J. M. and VIGIER,J. P.:’1961, Journ. Phys. Radium 22, 22.
HEISENBERG, W. : 1964,Physics andPhilosophy., Benjamin for a discussion see N. CUFARO-PETRONI, A. GARUCCIO, F. SELLERIand
J. P. VIGIER:1980, C. R. Acad. Sc. Paris 290, I I I .
HILEY,B. : 1981, Private communication.
HILLION,P. : 1957. Thesis, University of Paris.
HIRONDEL, D. : 1977, Thesis, Paris.
HOLT,R. A.: 1972, Ph. D. Thesis-Harvard.
JAMMER, M.: 1966. The Conceptional Development of Quantum Mechanics, McGraw Hill, New York.
JANOSSY, L.: 1972. Found. of Phys. 2, 9.
JOSEPHSON: 1979, in Science e t consuence, Colloque de Cordoue.
LEHR,W. and PARK,J.: 1977, J. Math. Phys. (N.Y.) 18, 1235.
MADELUNG, E.:1926, Zeitschrift f. Physik 40. 332.
MANDEL, L., DAJENAIS, K.: 1978, Phys. Rev. Lett. A 18, 2217.
MARCH,A.: 1934. Zeitschrift f. Physik 104,93 and 1937, 105, 620.
M A R I ~Z., and ZIVANOVI~, DJ. : 1976, in Quantum Mechanics determinism causality and particles, Reidel, Holland.
MARKOV, M. A. : 1938, Hyperons and k-Mesons, Fizmatgiz, Moskow.
MBLLER,C.: 1962, The Theory of Kelativity, Claretidon Press, Oxford, 52.
DE MUYNCK, B.: (1981). Private communication.
80 J. 1’. VIGIEH:Nori-Locality. Causality and Actlicr in Quaiituni Mechaiiics
NAMSRAI, K . : 1980. Found. of Phys. 10, 353 and 731.

DE LAPEEA. L. and CETTO,A. M.: 1975. Found. Phys. 5. 3 5 5
PFLEEGOR, R. L. and MANDEL, L.: 1967, Phys. Rev. 159, 1084; 1968, J. Opt. SOC.America, 1968, 58, 946.
POUNDER, R. J.: 1954. Comm. of Dublin Inst. ‘for adv. studies, A 11, I.
RICHTER,G., BRUNNER W. and PAUL, H.: 1964. Ann. der Physik 14, 239.
SELLERI,F. and TAROZZI, G. : 1979, Extension of the domain of validity of BELL’Sinequality, Bari. Univ. Prepriiit.
SELLERI,F. and TAROZZI, G.: 1979, Nuovo. Cim. Lett. 25. 151.
SELLERI,F. and VIGIER,J. P.: 1980, Nuovo. Cim. Lett. 29. 7.
SELLERI.F. and VIGIER.J. P.: 1981, YOURGRAU-MemOrial (in press).
SHIMONY, A. : 1980,Geneva Meeting on E. P. R. ’
ANDRADE DE SILVA,J. and DE BROGLIE,L.: 1972, Phys. Rev. 172.
STAPP,H. P.: 1977. Nuovo. Cim. B 40. 151.
TAKABAYASI. T.: 1952, Prog. Theor. Phys. (Japan) 8, 143 and 9, 187.
TERLETSKI,YA. P. and VIGIER,J. P.: 1961, Zurn. Eksp. Teor. Fiz. 13, 356.
VIGIER,J. P.: 1968, C. R. Acad. Sc. Paris 266, 266.
VIGIER. J. P.: 1979, Nuovo. Chi. Lett. 24. 265.
VIGIER,J. P.: 1980, Nuovo. Cim. Lett. 29, 467.
WIGNER,E. P.: 1967, Symmetries and Reflections, M. I. T. Press, 171.
WILSON,A. R., LOWE,J. and BUTT,D. K.: 1976, J. Phys. G 2, 613.
YUKAWA, H. : 1966, Res. Inst. Fund. Phys. Kyoto Univ. PIEB-55.
ZEEMAN,E, C.: 1964, Journ. Math. Phys. 5, $90.
Address of thc author:
JEAN-PIERRE VIGIER
Laboratoire de Physique Th6orique
E.R.A. NO 533 associ6e au CNRS
Institut H. Poincark, 11, rue P. & M. Curie
75231 Paris Cedex 05 (France)

Non-Locality, Causality and Aether in Quantum Mechanics: J, H. Poincare

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Non-Locality, Causality and Aether in Quantum Mechanics: J, H. Poincare

Uploaded by

Copyright:

Available Formats

55

Astron. Nachr., 303 (1982)r, 55-80

Non-Locality, Causality and Aether in Quantum Mechanics

Laboratoire de Physique Theorique, Institut H. POINCARE

With 16 figures (Received 1981July 20)

To calculate D we now assume as done independently by CEROFOLINI(1980):

We now make the final step in our demonstration of NELSON’S

3. Causality and Locality

4. Action at a distance and causality in relativistic mechanics

0 90 780 270 360

6. A Copenhagen Interpretation of non-locality

Fig. 13. Causal anomaly

Antitelephone mechanisms acting in S, and Sh induce an action of on E~ To quote HEISENBERG (1964)this

7. Non-local causal description of interacting particle pairs

invariants of the POINCARB

which can be split (with y = R exp (iS/b))into:

so that, in the case that V, = V , = V and m, = m2 = m we have:

are now coincident with (5) if qp = x, i.e.:

I~ = - +at,- + +at,i a ~ , SD=Sbiai--

bi = OX,, Sb, = SDX, .

where for the density we have as continuity equations

which yields for the continuity equation the form

8. New experimental tests in the BOHR-EINSTEIN

photon 2’s arrival in A(B).

Xsrac.r, A. : 1975, Pliys. Lett. 54, I I 7.

NAMSRAI, K . : 1980. Found. of Phys. 10, 353 and 731.

Address of thc author:

You might also like