Вы находитесь на странице: 1из 54

pdf version of the entry on

Action at a Distance in Quantum Mechanics


http://plato.stanford.edu/archives/spr2009/entries/qm-action-distance/ Action at a Distance in Quantum Mechanics
from the Spring 2009 Edition of the First published Fri Jan 26, 2007

Stanford Encyclopedia In the quantum realm, there are curious correlations between the
properties of distant systems. An example of such correlations is provided
of Philosophy by the famous Einstein-Podolsky-Rosen/Bohm experiment. The
correlations in the EPR/B experiment strongly suggest that there are non-
local influences between distant systems, i.e., systems between which no
light signal can travel, and indeed orthodox quantum mechanics and its
various interpretations postulate the existence of such non-locality. Yet,
the question of whether the EPR/B correlations imply non-locality and the
Edward N. Zalta Uri Nodelman Colin Allen John Perry
exact nature of this non-locality is a matter of ongoing controversy.
Principal Editor Senior Editor Associate Editor Faculty Sponsor
Focusing on EPR/B-type experiments, in this entry we consider the nature
Editorial Board
http://plato.stanford.edu/board.html of the various kinds of non-locality postulated by different interpretations
of quantum mechanics. Based on this consideration, we briefly discuss the
Library of Congress Catalog Data
ISSN: 1095-5054
compatibility of these interpretations with the special theory of relativity.

Notice: This PDF version was distributed by request to mem- 1. Introduction


bers of the Friends of the SEP Society and by courtesy to SEP 2. Bell's theorem and non-locality
content contributors. It is solely for their fair use. Unauthorized 3. The analysis of factorizability
distribution is prohibited. To learn how to join the Friends of the 4. Action at a distance, holism and non-separability
SEP Society and obtain authorized PDF versions of SEP entries, 4.1 Action at a distance
please visit https://leibniz.stanford.edu/friends/ . 4.2 Holism
4.3 Non-separability
Stanford Encyclopedia of Philosophy
Copyright c 2009 by the publisher 5. Holism, non-separability and action at a distance in quantum
The Metaphysics Research Lab mechanics
Center for the Study of Language and Information
5.1 Collapse theories
Stanford University, Stanford, CA 94305
5.2 Can action-at-a-distance co-exist with non-separability and
Action at a Distance in Quantum Mechanics
Copyright c 2009 by the author holism?
Joseph Berkovitz 5.3 No-collapse theories
All rights reserved. 6. Superluminal causation
Copyright policy: http://plato.stanford.edu/info.html#c.

1
Action at a Distance in Quantum Mechanics Joseph Berkovitz

7. Superluminal signaling other, they each encounter a measuring apparatus that can be set to
7.1 Necessary and sufficient conditions for superluminal measure their spin components along various directions. Although the
signaling measurement events are distant from each other, so that no slower-than-
7.2 No-collapse theories light or light signal can travel between them, the measurement outcomes
7.3 Collapse theories are curiously correlated.[1] That is, while the outcome of each of the
7.4 The prospects of controllable probabilistic dependence distant spin measurements seems to be a matter of pure chance, they are
7.5 Superluminal signaling and action-at-a-distance correlated with each other: The joint probability of the distant outcomes is
8. The analysis of factorizability: implications for quantum non- different from the product of their single probabilities. For example, the
locality probability that each of the particles will spin clockwise about the z-axis
8.1 Non-separability, holism and action at a distance in a z-spin measurement (i.e., a measurement of the spin component along
8.2 Superluminal signaling the z direction) appears to be ½. Yet, the outcomes of such measurements
8.3 Relativity are perfectly anti-correlated: If the left-hand-side (L-) particle happens to
8.4 Superluminal causation spin clockwise (anti-clockwise) about the z-axis, the right-hand-side (R-)
8.5 On the origin and nature of parameter dependence particle will spin anti-clockwise (clockwise) about that axis. And this is
9. Can there be ‘local’ quantum theories? true even if the measurements are made simultaneously.
10. Can quantum non-locality be reconciled with relativity?
10.1 Collapse theories
10.2 No-collapse theories
10.3 Quantum causal loops and relativity
Bibliography
Other Internet Resources
Related Entries Figure 1: A schematic illustration of the EPR/B experiment.
Particle pairs in the spin singlet state are emitted in opposite
directions and when they are distant from each other (i.e., space-
1. Introduction like separated), they encounter measurement apparatuses that can
be set to measure spin components along various directions.
The quantum realm involves curious correlations between distant events.
A well-known example is David Bohm's (1951) version of the famous The curious EPR/B correlations strongly suggest the existence of non-
thought experiment that Einstein, Podolsky and Rosen proposed in 1935 local influences between the two measurement events, and indeed
(henceforth, the EPR/B experiment). Pairs of particles are emitted from a orthodox ‘collapse’ quantum mechanics supports this suggestion.
source in the so-called spin singlet state and rush in opposite directions According to this theory, before the measurements the particles do not
(see Fig. 1 below). When the particles are widely separated from each have any definite spin. The particles come to possess a definite spin only
other, they each encounter a measuring apparatus that can be set to with the first spin measurement, and the outcome of this measurement is a

2 Stanford Encyclopedia of Philosophy Spring 2009 Edition 3


Action at a Distance in Quantum Mechanics Joseph Berkovitz

with the first spin measurement, and the outcome of this measurement is a and Grassi 1994 and Redhead and La Riviere 1997.) On this view,
matter of chance. If, for example, the first measurement is a z-spin quantum states of systems do not always reflect their complete state.
measurement on the L-particle, the L-particle will spin either clockwise Quantum states of systems generally provide information about some of
or anti-clockwise about the z-axis with equal chance. And the outcome of the properties that systems possess and information about the probabilities
the L-measurement causes an instantaneous change in the spin properties of outcomes of measurements on them, and this information does not
of the distant R-particle. If the L-particle spins clockwise (anti-clockwise) generally reflect the complete state of the systems. In particular, the
about the z-axis, the R-particle will instantly spin anti-clockwise information encoded in the spin singlet state is about the probabilities of
(clockwise) about the same axis. (It is common to call spins in opposite measurement outcomes of spin properties in various directions, about the
directions ‘spin up’ and ‘spin down,’ where by convention a clockwise conditional probabilities that the L- (R-) particle has a certain spin
spinning may be called ‘spin up’ and anti-clockwise spinning may be property given that the R- (L-) particle has another spin property, and
called ‘spin down.’) about the anti-correlation between the spins that the particles may have in
any given direction (for more details, see section 5.1). Thus, the outcome
It may be argued that orthodox quantum mechanics is false, and that the of a z-spin measurement on the L-particle and the spin singlet state
non-locality postulated by it does not reflect any non-locality in the (interpreted as a state of knowledge) jointly provide information about the
quantum realm. Alternatively, it may be argued that orthodox quantum z-spin property of the R-particle. For example, if the outcome of the L-
mechanics is a good instrument for predictions rather than a fundamental measurement is z-spin ‘up,’ we know that the R-particle has z-spin
theory of the physical nature of the universe. On this instrumental ‘down’; and if we assume, as EPR did, that there is no curious action at a
interpretation, the predictions of quantum mechanics are not an adequate distance between the distant wings (and that the change of the quantum-
basis for any conclusion about non-locality: This theory is just an mechanical state of the particle pair in the L-measurement is only a
incredible oracle (or a crystal ball), which provides a very successful change in state of knowledge), we could also conclude that the R-particle
algorithm for predicting measurement outcomes and their probabilities, had z-spin ‘down’ even before the L-measurement occurs.
but it offers little information about ontological matters, such as the nature
of objects, properties and causation in the quantum realm. How could the L-outcome change our knowledge/ignorance about the R-
outcome if it has no influence on it? The simplest and most
Einstein, Podolsky and Rosen (1935) thought that quantum mechanics is straightforward reply is that the L- and the R- outcome have a common
incomplete and that the curious correlations between distant systems do cause that causes them to be correlated, so that knowledge of one outcome
not amount to action at a distance between them. The apparent provides knowledge about the other. [2] Yet, the question is whether the
instantaneous change in the R-particle's properties during the L- predictions of orthodox quantum mechanics, which have been highly
measurement is not really a change of properties, but rather a change of confirmed by various experiments, are compatible with the quantum
knowledge. (For more about the EPR argument, see the entry on the EPR realm being local in the sense of involving no influences between systems
argument, Redhead 1987, chapter 3, and Albert 1992, chapter 3. For between which light and slower-than-light signals cannot travel (i.e.,
discussions of the EPR argument in the relativistic context, see Ghirardi space-like separated systems). More particularly, the question is whether
and Grassi 1994 and Redhead and La Riviere 1997.) On this view, it is possible to construct a local, common-cause model of the EPR/B
4 Stanford Encyclopedia of Philosophy Spring 2009 Edition 5
Action at a Distance in Quantum Mechanics Joseph Berkovitz

it is possible to construct a local, common-cause model of the EPR/B denote the pair's state before any measurement occurs. Let l denote the
experiment, i.e., a model that postulates no influence between setting of the L-measurement apparatus to measure spin along the l-axis
systems/events in the distant wings of the experiment, and that the (i.e., the l-spin of the L-particle), and let r denote the setting of the R-
correlation between them are due to the state of the particle pair at the measurement apparatus to measure spin along the r-axis (i.e., the r-spin
source. In 1935, Einstein, Podolsky and Rosen believed that this is of the R-particle). Let xl be the outcome of a l-spin measurement in the L-
possible. But, as John Bell demonstrated in 1964, this belief is difficult to wing, and let yr be the outcome of a r-spin measurement in the R-wing;
uphold. where xl is either the L-outcome l-spin ‘up’ or the L-outcome l-spin
‘down,’ and yr is either the R-outcome r-spin ‘up’ or the R-outcome r-
2. Bell's theorem and non-locality spin ‘down.’ Let P λ l r(xl & yr) be the joint probability of the L- and the
R-outcome, and P λ l(xl) and P λ r(yr) be the single probabilities of the L-
In a famous theorem, John Bell (1964) demonstrated that granted some and the R-outcome, respectively; where the subscripts λ, l and r denote
plausible assumptions, any local model of the EPR/B experiment is the factors that are relevant for the probabilities of the outcomes xl and yr.
committed to certain inequalities about the probabilities of measurement Then, for any λ, l, r, xl and yr:[3]
outcomes, ‘the Bell inequalities,’ which are incompatible with the
quantum-mechanical predictions. When Bell proved his theorem, the Factorizability
EPR/B experiment was only a thought experiment. But due to P λ l r(xl & yr) = P λ l(xl) · P λ r(yr).
technological advances, various versions of this experiment have been
(Here and henceforth, for simplicity's sake we shall denote events and
conducted since the 1970s, and their results have overwhelmingly
states, such as the measurement outcomes, and the propositions that they
supported the quantum-mechanical predictions (for brief reviews of these
occur by the same symbols.)
experiments and further references, see the entry on Bell's theorem and
Redhead 1987, chapter 4, section 4.3 and ‘Notes and References’). Thus, a The state λ is typically thought of as the pair's state at the emission time,
wide consensus has it that the quantum realm involves some type of non- and it is assumed that this state does not change in any relevant sense
locality. between the emission and the first measurement. It is (generally) a
different state from the quantum-mechanical pair's state ψ. ψ is assumed
The basic idea of Bell's theorem is as follows. A model of the EPR/B
to be an incomplete state of the pair, whereas λ is supposed to be a (more)
experiment postulates that the state of the particle pair together with the
complete state of the pair. Accordingly, pairs with the same state ψ may
apparatus settings to measure (or not to measure) certain spin properties
have different states λ which give rise to different probabilities of
determine the probabilities for single and joint spin-measurement
outcomes for the same type of measurements. Also, the states λ may be
outcomes. A local Bell model of this experiment also postulates that
unknown, hidden, inaccessible or uncontrollable.
probabilities of joint outcomes factorize into the single probabilities of the
L- and the R- outcomes: The probability of joint outcomes is equal to the Factorizability is commonly motivated as a locality condition. In non-
product of the probabilities of the single outcomes. More formally, let λ local models of the EPR/B experiment, the correlations between the
denote the pair's state before any measurement occurs. Let l denote the distant outcomes are accounted for by non-local influences between the
6 Stanford Encyclopedia of Philosophy Spring 2009 Edition 7
Action at a Distance in Quantum Mechanics Joseph Berkovitz

distant outcomes are accounted for by non-local influences between the


distant measurement events. For example, in orthodox quantum
mechanics the first spin measurement on, say, the L-particle causes an
immediate change in the spin properties of the R-particle and in the
probabilities of future outcomes of spin measurements on this particle. By
contrast, in local models of this experiment the correlations are supposed
Figure 3: A space-time diagram of a local model of the EPR/B
to be accounted for by a common cause—the pair's state λ (see Fig. 2
experiment. The circles represent the measurement events, and the
below): The pair's state and the L-setting determine the probability of the
cones represent their backward light cones, i.e., the boundaries of
L-outcome; the pair's state and the R-setting determine the probability of
all the subluminal and luminal influences on them. The dotted
the R-outcome; and the pair's state and the L- and the R-setting determine
lines denote the propagation of the influences of the pair's state at
the probability of joint outcomes, which (as mentioned above) is simply
the emission and of the settings of the measurement apparatuses
the product of these single probabilities. The idea is that the probability of
on the measurement outcomes.
each of the outcomes is determined by ‘local events,’ i.e., events that are
confined to its backward light-cone, and which can only exert subluminal A Bell model of the EPR/B experiment also postulates that for each
or luminal influences on it (see Figure 3 below); and the distant outcomes quantum-mechanical state ψ there is a distribution ρ over all the possible
are fundamentally independent of each other, and thus their joint pair states λ, which is independent of the settings of the apparatuses. That
probability factorizes. (For more about this reasoning, see sections 6 and is, the distribution of the (‘complete’) states λ depends on the
8-9.) (‘incomplete’) state ψ, and this distribution is independent of the
particular choice of measurements in the L- and R-wing (including the
choice not to measure any quantity). Or formally, for any quantum-
mechanical state ψ, L-settings l and l′, and R-settings r and r′:

λ-independence
Figure 2: A schematic common-cause model of the EPR/B ρ ψ l r(λ) = ρ ψ l′ r(λ) = ρ ψ l r′ (λ) = ρ ψ l′ r′ (λ) = ρ ψ (λ)
experiment. Arrows denote causal connections.
where the subscripts denote the factors that are potentially relevant for the
distribution of the states λ.

Although the model probabilities (i.e., the probabilities of outcomes


prescribed by the states λ) are different from the corresponding quantum-
mechanical probabilities of outcomes (i.e., the probabilities prescribed by
the quantum-mechanical states ψ), the quantum mechanical probabilities
(which have been systematically confirmed) are recovered by averaging

8 Stanford Encyclopedia of Philosophy Spring 2009 Edition 9


Action at a Distance in Quantum Mechanics Joseph Berkovitz

(which have been systematically confirmed) are recovered by averaging the distribution of the possible pairs' states λ is dependent upon the
over the model probabilities. That is, it is supposed that the quantum- settings. Since the settings can be made after the emission of the particle
mechanical probabilities P ψ l r(xl & yr), P ψ l(xl) and P ψ r(yr) are pair from the source, this kind of violation of λ-independence would
obtained by averaging over the model probabilities P λ l r(xl & yr), P λ l require backward causation. (For advocates of this way out of non-
(xl) and P λ r(yr), respectively: For any ψ, l, r, xl and yr, locality, see Costa de Beauregard 1977, 1979, 1985, Sutherland 1983,
1998, 2006 and Price 1984, 1994, 1996, chapters 3, 8 and 9.) On some
Empirical Adequacy readings of John Cramer's (1980, 1986) transactional interpretation of
P ψ l r(xl & yr) = ∫λ P λ l r(xl & yr) · ρ ψ l r(λ) quantum mechanics (see Maudlin 1994, pp. 197-199), such violation of λ-
P ψ l(xl) = ∫λ P λ l(xl ) · ρ ψ l(λ) independence is postulated. According to this interpretation, the source
P ψ r(yr) = ∫λ P λ r(yr) · ρ ψ r(λ). [4] sends ‘offer’ waves forward to the measurement apparatuses, and the
apparatuses send ‘confirmation’ waves (from the space-time regions of
The assumption of λ-independence is very plausible. It postulates that
the measurement events) backward to the source, thus affecting the states
(complete) pair states at the source are uncorrelated with the settings of
of emitted pairs according to the settings of the apparatuses. The question
the measurement apparatuses. And independently of one's philosophical
of whether such a theory can reproduce the predictions of quantum
view about free will, this assumption is strongly suggested by our
mechanics is a controversial matter (see Maudlin 1994, pp. 197-199,
experience, according to which it seems possible to prepare the state of
Berkovitz 2002, section 5, and Kastner 2006). It is noteworthy, however,
particle pairs at the source independently of the set up of the measurement
that while the violation of λ-independence is sufficient for circumventing
apparatuses.
Bell's theorem, the failure of this condition per se does not substantiate
There are two ways to try to explain a failure of λ-independence. One locality. The challenge of providing a local model of the EPR/B
possible explanation is that pairs' states and apparatus settings share a experiment also applies to models that violate λ-independence. (For more
common cause, which always correlates certain types of pairs' states λ about these issues, see sections 9 and 10.3.)
with certain types of L- and R-setting. Such a causal hypothesis will be
In any case, as Bell's theorem demonstrates, factorizability, λ-
difficult to reconcile with the common belief that apparatus settings are
independence and empirical adequacy jointly imply the Bell inequalities,
controllable at experimenters' will, and thus could be set independently of
which are violated by the predictions of orthodox quantum mechanics
the pair's state at the source. Furthermore, thinking of all the different
(Bell 1964, 1966, 1971, 1975a,b). Granted the systematic confirmation of
ways one can measure spin properties and the variety of ways in which
the predictions of orthodox quantum mechanics and the plausibility of λ-
apparatus settings can be chosen, the postulation of such common cause
independence, Bell inferred that factorizability fails in the EPR/B
explanation for settings and pairs' states would seem highly ad hoc and its
experiment. Thus, interpreting factorizability as a locality condition, he
existence conspiratorial.
concluded that the quantum realm is non-local. (For further discussions of
Another possible explanation for the failure of λ-independence is that the Bell's theorem, the Bell inequalities and non-locality, see Bell 1966, 1971,
apparatus settings influence the pair's state at the source, and accordingly 1975a,b, 1981, Clauser et al 1969, Clauser and Horne 1974, Shimony
the distribution of the possible pairs' states λ is dependent upon the 1993, chapter 8, Fine 1982a,b, Redhead 1987, chapter 4, Butterfield 1989,
10 Stanford Encyclopedia of Philosophy Spring 2009 Edition 11
Action at a Distance in Quantum Mechanics Joseph Berkovitz

1993, chapter 8, Fine 1982a,b, Redhead 1987, chapter 4, Butterfield 1989, P λ l r(xl / yr) = P λ l r(xl) and P λ l r(yr / xl) = P λ l r(yr)
1992a, Pitowsky 1989, Greenberger, Horne and Zeilinger 1989,
P λ l r(yr) > 0 P λ l r(xl) > 0,
Greenberger, Horne, Shimony and Zeilinger 1990, Mermin 1990, and the
entry on Bell's theorem.) or more generally,

3. The analysis of factorizability OI


P λ l r(xl & yr) = P λ l r(xl) · P λ l r(yr).
Following Bell's work, a broad consensus has it that the quantum realm
involves some type of non-locality (for examples, see Clauser and Horne Assuming λ-independence (see section 2), any empirically adequate
1974, Jarrett 1984,1989, Shimony 1984, Redhead 1987, Butterfield 1989, theory will have to violate OI or PI. A common view has it that violations
1992a,b, 1994, Howard 1989, Healey 1991, 1992, 1994, Teller 1989, of PI involve a different type of non-locality than violations of OI:
Clifton, Butterfield and Redhead 1990, Clifton 1991, Maudlin 1994, Violations of PI involve some type of action-at-a-distance that is
Berkovitz 1995a,b, 1998a,b, and references therein). [5] But there is an impossible to reconcile with relativity (Shimony 1984, Redhead 1987, p.
ongoing controversy as to its exact nature and its compatibility with 108), whereas violations of OI involve some type of holism, non-
relativity theory. One aspect of this controversy is over whether the separability and/or passion-at-a-distance that may be possible to reconcile
analysis of factorizability and the different ways it could be violated may with relativity (Shimony 1984, Readhead 1987, pp. 107, 168-169, Howard
shed light on these issues. Factorizability is equivalent to the conjunction 1989, Teller 1989).
of two conditions (Jarrett 1984, 1989, Shimony 1984): [6]
On the other hand, there is the view that the analysis above (as well as
Parameter independence. The probability of a distant other similar analyses of factorizability [7]) is immaterial for studying
measurement outcome in the EPR/B experiment is independent of quantum non-locality (Butterfield 1992a, pp. 63-64, Jones and Clifton
the setting of the nearby measurement apparatus. Or formally, for 1993, Maudlin 1994, pp. 96 and 149) and even misleading (Maudlin 1994,
any pair's state λ, L-setting l, R-setting r, L-outcome xl and R- pp. 94-95 and 97-98). On this alternative view, the way to examine the
outcome yr: nature of quantum non-locality is to study the ontology postulated by the
various interpretations of quantum mechanics and alternative quantum
PI theories. [8] In sections 4-7, we shall follow this methodology and discuss
P λ l r(xl) = P λ l(xl) and P λ l r(yr) = P λ r(yr). the nature of non-locality postulated by several quantum theories. The
discussion in these sections will furnish the ground for evaluating the
Outcome independence. The probability of a distant above controversy in section 8.
measurement outcome in the EPR/B experiment is independent of
the nearby measurement outcome. Or formally, for any pair's state 4. Action at a distance, holism and non-separability
λ, L-setting l, R-setting r, L-outcome xl and R-outcome yr:

12 Stanford Encyclopedia of Philosophy Spring 2009 Edition 13


Action at a Distance in Quantum Mechanics Joseph Berkovitz

4.1 Action at a distance In the literature, there are various characterizations of holism. Discussions
of quantum non-locality frequently focus on property holism, where
In orthodox quantum mechanics as well as in any other current quantum certain physical properties of objects are not determined by the physical
theory that postulates non-locality (i.e., influences between distant, space- properties of their parts. The intuitive idea is that some intrinsic properties
like separated systems), the influences between the distant measurement of wholes (e.g. physical systems) are not determined by the intrinsic
events in the EPR/B experiment do not propagate continuously in space- properties of their parts and the spatiotemporal relations that obtain
time. They seem to involve action at a distance. Yet, a common view has between these parts. This idea can be expressed in terms of supervenience
it that these influences are due to some type of holism and/or non- relations.
separability of states of composite systems, which are characteristic of
systems in entangled states (like the spin singlet state), and which exclude Property Holism. Some objects have intrinsic qualitative
the very possibility of action at a distance. The paradigm case of action at properties and/or relations that do not supervene upon the intrinsic
a distance is the Newtonian gravitational force. This force acts between qualitative properties and relations of their parts and the
distinct objects that are separated by some (non-vanishing) spatial spatiotemporal relations between these parts.
distance, its influence is symmetric (in that any two massive objects
influence each other), instantaneous and does not propagate continuously It is difficult to give a general precise specification of the terms ‘intrinsic
in space. And it is frequently claimed or presupposed that such action at a qualitative property’ and ‘supervenience.’ Intuitively, a property of an
distance could only exist between systems with separate states in non- object is intrinsic just in case that object has this property in and for itself
holistic universes (i.e., universes in which the states of composite systems and independently of the existence or the state of any other object. A
are determined by, or supervene upon the states of their subsystems and property is qualitative (as opposed to individual) if it does not depend on
the spacetime relations between them), which are commonly taken to the existence of any particular object. And the intrinsic qualitative
characterize the classical realm.[9] properties of an object O supervene upon the intrinsic qualitative
properties and relations of its parts and the spatiotemporal relations
In sections 4.2 and 4.3, we shall briefly review the relevant notions of between them just in case there is no change in the properties and
holism and non-separability (for a more comprehensive review, see the relations of O without a change in the properties and relations of its parts
entry on holism and nonseparability in physics and Healey 1991). In and/or the spatiotemporal relations between them. (For attempts to
section 5, we shall discuss the nature of holism and non-separability in the analyze the term ‘intrinsic property,’ see for example Langton and Lewis
quantum realm as depicted by various quantum theories. Based on this 1998 and the entry on intrinsic vs. extrinsic properties. For a review of
discussion, we shall consider whether the non-local influences in the different types of supervenience, see for example Kim 1978, McLaughlin
EPR/B experiment constitute action at a distance. 1994 and the entry on supervenience.)

4.2 Holism Paul Teller (1989, p. 213) proposes a related notion of holism, ‘relational
holism,’ which is characterized as the violation of the following
condition:

14 Stanford Encyclopedia of Philosophy Spring 2009 Edition 15


Action at a Distance in Quantum Mechanics Joseph Berkovitz

condition: Like holism, there are various notions of non-separability on offer. The
most common notion in the literature is state non-separability, i.e., the
Particularism. The world is composed of individuals. All violation of the following condition:
individuals have non-relational properties and all relations
supervene on the non-relational properties of the relata. State separability. Each system possesses a separate state that
determines its qualitative intrinsic properties, and the state of any
Here, by a non-relational property Teller means an intrinsic property composite system is wholly determined by the separate states of
(1986a, p. 72); and by ‘the supervenience of a relational property on the its subsystems.
non-relational properties of the relata,’ he means that ‘if two objects, 1
and 2, bear a relation R to each other, then, necessarily, if two further The term ‘wholly determined’ is vague. But, as before, one may spell it
objects, 1′ and 2′ have the same non-relational properties, then 1′ and 2′ out in terms of supervenience relations: State separability obtains just in
will also bear the same relation R to each other’ (1989, p. 213). Teller case each system possesses a separate state that determines its qualitative
(1986b, pp. 425-7) believes that spatiotemporal relations between objects intrinsic properties and relations, and the state of any composite system is
supervene upon the objects’ intrinsic physical properties. Thus, he does supervenient upon the separate states of its subsystems.
not include the spatiotemporal relations in the supervenience basis. This
view is controversial, however, as many believe that spatiotemporal Another notion of non-separability is spatiotemporal non-separability.
relations between objects are neither intrinsic nor supervene upon the Inspired by Einstein (1948), Howard (1989, pp. 225-6) characterizes
intrinsic qualitative properties of these objects. But, if such supervenience spatiotemporal non-separability as the violation of the following
does not obtain, particularism will also be violated in classical physics, separability condition:
and accordingly relational holism will fail to mark the essential
Spatiotemporal separability. The contents of any two regions of
distinction between the classical and the quantum realms. Yet, one may
space-time separated by a non-vanishing spatiotemporal interval
slightly revise Teller's definition of particularism as follows:
constitute two separate physical systems. Each separated space-
Particularism*. The world is composed of individuals. All time region possesses its own, distinct state and the joint state of
individuals have non-relational properties and all relations any two separated space-time regions is wholly determined by the
supervene upon the non-relational properties of the relata and the separated states of these regions.
spatiotemporal relations between them.
A different notion of spatiotemporal non-separability, proposed by Healey
In what follows in this entry, by relational holism we shall mean a (see the entry on holism and nonseparability in physics), is process non-
violation of particularism*. separability. It is the violation of the following condition:

4.3 Non-separability Process separability. Any physical process occupying a


spacetime region R supervenes upon an assignment of qualitative
intrinsic physical properties at spacetime points in R.

16 Stanford Encyclopedia of Philosophy Spring 2009 Edition 17


Action at a Distance in Quantum Mechanics Joseph Berkovitz

intrinsic physical properties at spacetime points in R. having z-spin ‘down’ (i.e., spinning ‘down’ about the z-axis) can be
represented by the orthogonal vector, |z-down>. Particle pairs may be in a
5. Holism, non-separability and action at a distance state in which the L-particle and the R-particle have opposite spins, for
in quantum mechanics instance either a state |ψ1> in which the L-particle has z-spin ‘up’ and the
R-particle has z-spin ‘down,’ or a state |ψ 2> in which the L-particle has
The quantum realm as depicted by all the quantum theories that postulate z-spin ‘down’ and the R-particle has z-spin ‘up.’ Each of these states is
non-locality, i.e., influences between distant (space-like separated) represented by a tensor product of vectors in the Hilbert space of the
systems, involves some type of non-separability or holism. In what particle pair: |ψ1> = |z-up> L |z-down>R and |ψ 2> = |z-down>L |z-up> R ;
follows in this section, we shall consider the nature of the non-separability where the subscripts L and R refer to the Hilbert spaces associated with
and holism manifested by various interpretations of quantum mechanics. the L- and the R-particle, respectively. But particle pairs may also be in a
On the basis of this consideration, we shall address the question of superposition of these states, i.e., a state that is a linear sum of the states |
whether these interpretations predicate the existence of action at a ψ1> and |ψ 2>, e.g. the state represented by
distance. We start with the so-called ‘collapse theories.’
|ψ 3> = 1/√2 (|ψ 1> − |ψ 2>)
5.1 Collapse theories = 1/√2 (|z-up> L |z-down>R − |z-down>L |z-up> R ).

5.1.1 Orthodox quantum mechanics In fact, this is exactly the case in the spin singlet state. In this state, the
particles are entangled in a non-separable state (i.e., a state that cannot be
In orthodox quantum mechanics, normalized vectors in Hilbert spaces decomposed into a product of separate states of the L- and the R-particle),
represent states of physical systems. When the Hilbert space is of infinite in which (according to the property-assignment rules of orthodox
dimension, state vectors can be represented by functions, the so-called quantum mechanics) the particles do not possess any definite z-spin (or
‘wave functions.’ In any given basis, there is a unique wave function that definite spin in any other direction). Thus, the condition of state
corresponds to the state vector in that basis. (For an entry level review of separability fails: The state of the particle pair (which determines its
the highlights of the mathematical formalism and the basic principles of intrinsic qualitative properties) is not wholly determined by the separate
quantum mechanics, see the entry on quantum mechanics, Albert 1992, states of the particles (which determine their intrinsic qualitative
Hughes 1989, Part I, and references therein; for more advanced reviews, properties). Or more precisely, the pair's state is not supervenient upon
see Bohm 1951 and Redhead 1987, chapters 1-2 and the mathematical the separable states of the particles. In particular, the superposition state
appendix.) of the particle pair assigns a ‘correlational’ property that dictates that the
outcomes of (ideal) z-spin measurements on both the L- and the R-
For example, the state of the L-particle having z-spin ‘up’ (i.e., spinning
particle will be anti-correlated, and this correlational property is not
‘up’ about the z-axis) can be represented by the vector |z-up> in the
supervenient upon properties assigned by any separable states of the
Hilbert space associated with the L-particle, and the state of the L-particle
particles (for more details, see Healey 1992, 1994). For similar reasons,
having z-spin ‘down’ (i.e., spinning ‘down’ about the z-axis) can be
the spin singlet state also involves property and relational holism; for the
18 Stanford Encyclopedia of Philosophy Spring 2009 Edition 19
Action at a Distance in Quantum Mechanics Joseph Berkovitz

the spin singlet state also involves property and relational holism; for the equation and the property assignment rule called ‘eigenstate-eigenvalue
above correlational property of the particle pair also fails to supervene link.’ According to the eigenstate-eigenvalue link, a physical observable,
upon the intrinsic qualitative properties of the particles and the i.e., a physical quantity, of a system has definite value (one of its
spatiotemporal relations between them. Furthermore, the process that eigenvalues) just in case the system is in the corresponding eigenstate of
leads to each of the measurement outcomes is also non-separable, i.e., that observable (see the entry on quantum mechanics, section 4).
process separability fails (see Healey 1994 and the entry on holism and Microscopic systems may be in a superposition state of spin components,
nonseparability in physics). energies, positions, momenta as well as other physical observables.
Accordingly, microscopic systems may be in a state of indefinite z-spin,
This correlational property is also ‘responsible’ for the action at a energy, position, momentum and various other quantities. The problem is
distance that the orthodox theory seems to postulate between the distant that given the linear and unitary Schrödinger dynamics, these indefinite
wings in the EPR/B experiment. Recall (section 1) that Einstein, Podolsky quantities are also endemic in the macroscopic realm. For example, in a z-
and Rosen thought that this curious action at a distance reflects the spin measurement on a particle in a superposition state of z-spin ‘up’ and
incompleteness of this theory rather than a state of nature. The EPR z-spin ‘down,’ the position of the apparatus’s pointer gets entangled with
argument for the incompleteness of the orthodox theory is controversial. the indefinite z-spin of the particle, thus transforming the pointer into a
But the orthodox theory seems to be incomplete for a different reason. state of indefinite position, i.e., a superposition of pointing ‘up’ and
This theory postulates that in non-measurement interactions, the evolution pointing ‘down’ (see Albert 1992, chapter 4, and the entry on collapse
of states obeys a linear and unitary equation of motion, the so-called theories, section 3). In particular, in the EPR/B experiment the L-
Schrödinger equation (see the entry on quantum mechanics), according to measurement causes the L-apparatus pointer to get entangled with the
which the particle pair in the EPR/B experiment remains in an entangled particle pair, transforming it into a state of indefinite position:
state. This equation of motion also dictates that in a spin measurement,
the pointers of the measurement apparatuses get entangled with the |ψ4> = 1/√2 (|z-up> L |z-down>R |up> LA − |z-down>L |z-up> R |down>LA)
particle pair in a non-separable state in which (according to the theory's
property assignment, see below) the indefiniteness of particles’ spins is where |up> LA and |down>LA are the states of the L-apparatus pointer
‘transmitted’ to the pointer's position: In this entangled state of the displaying the outcomes z-spin ‘up’ and z-spin ‘down,’ respectively.
particle pair and the pointer, the pointer lacks any definite position, in Since the above type of indefiniteness is generic in orthodox no-collapse
contradiction to our experience of perceiving it pointing to either ‘up’ or quantum mechanics, in this theory measurements typically have no
‘down.’ definite outcomes, in contradiction to our experience.

The above problem, commonly called ‘the measurement problem,’ arises In order to solve this problem, the orthodox theory postulates that in
in orthodox no-collapse quantum mechanics from two features that measurement interactions, entangled states of measured systems and the
account very successfully for the behavior of microscopic systems: The corresponding measurement apparatuses do not evolve according to the
linear dynamics of quantum states as described by the Schrödinger Schrödinger equation. Rather, they undergo a ‘collapse’ into product
equation and the property assignment rule called ‘eigenstate-eigenvalue (non-entangled) states, where the systems involved have the relevant

20 Stanford Encyclopedia of Philosophy Spring 2009 Edition 21


Action at a Distance in Quantum Mechanics Joseph Berkovitz

(non-entangled) states, where the systems involved have the relevant Rimini 1990, and Butterfield et al. 1993). Similarly to orthodox collapse
definite properties. For example, the entangled state of the particle pair quantum mechanics, in the GRW models the quantum-mechanical state of
and the L-apparatus in the EPR/B experiment may collapse into a product systems (whether it is expressed by a vector or a wave function) provides
state in which the L-particle comes to possess z-spin ‘up,’ the R-particle a complete specification of their intrinsic properties and relations. The
comes to possess z-spin ‘down’ and the L-apparatus pointer displaying state of systems follows the Schrödinger equation, except that it has a
the outcome z-spin ‘up’: probability for spontaneous collapse, independently of whether or not the
systems are measured. The chance of collapse depends on the ‘size’ of the
|ψ5> = |z-up> L |z-down>R |up> LA. entangled systems—in the earlier models the ‘size’ of systems is
predicated on the number of the elementary particles, whereas in later
The problem is that in the orthodox theory, the notions of measurement
models it is measured in terms of mass densities. In any case, in
and the time, duration and nature of state collapses remain completely
microscopic systems, such as the particle pairs in the EPR/B experiment,
unspecified. As John Bell (1987b, p. 205) remarks, the collapse postulate
the chance of collapse is very small and negligible—the chance of
in this theory, i.e., the postulate that dictates that in measurement
spontaneous state collapse in such systems is cooked up so that it will
interactions the entangled states of the relevant systems do not follow the
occur, on average, every hundred million years or so. This means that the
Schrödinger equation but rather undergo a collapse, is no more than
chance that the entangled state of the particle pair in the EPR/B
‘supplementary, imprecise, verbal, prescriptions.’
experiment will collapse to a product state between the emission from the
This problem of accounting for our experience of perceiving definite source and the first measurement is virtually zero. In an earlier L-
measurement outcomes in orthodox quantum mechanics, is an aspect of measurement, the state of the particle pair gets entangled with the state of
the more general problem of accounting for the classical-like behavior of the L-measurement apparatus. Thus, the state of the pointer of the L-
macroscopic systems in this theory. apparatus evolves from being ‘ready’ to measure a certain spin property
to an indefinite outcome. For instance, in a z-spin measurement the L-
5.1.2 Dynamical models for state vector reduction apparatus gets entangled with the particle pair in a superposition state of
pointing to ‘up’ and pointing to ‘down’ (corresponding to the states of the
The dynamical models for state-vector reduction were developed to L-particle having z-spin ‘up’ and having z-spin ‘down’), and the R-
account for state collapses as real physical processes (for a review of the apparatus remains un-entangled with these systems in the state of being
collapse models and a detailed reference list, see the entry on collapse ready to measure z-spin. Or formally:
theories). The origin of the collapse models may be dated to Bohm and
Bub's (1966) hidden variable theory and Pearle's (1976) spontaneous |ψ6> = 1/√2 (|z-up>L |up>AL |z-down> R − |z-down> L |down>AL |z-up>R) |ready>AR
localization approach, but the program has received its crucial impetus
where, as before, |up> AL and |down>AL denote the states of the L-
with the more sophisticated models developed by Ghirardi, Rimini and
apparatus displaying the outcomes z-spin ‘up’ and ‘down’ respectively,
Weber in 1986 (see also Bell 1987a and Albert 1992) and their
and |ready>AR denotes the state of the R-apparatus being ready to
consequent development by Pearle (1989) (see also Ghirardi, Pearle and
Rimini 1990, and Butterfield et al. 1993). Similarly to orthodox collapse
22 Stanford Encyclopedia of Philosophy Spring 2009 Edition 23
Action at a Distance in Quantum Mechanics Joseph Berkovitz

and |ready>AR denotes the state of the R-apparatus being ready to


measure z-spin. In this state, a gigantic number of particles of the L- the outcome ‘down,’ the GRW collapse models would successfully
apparatus pointer are entangled together in the superposition state of resolve the measurement problem. Technically speaking, a precise
being in the position (corresponding to pointing to) ‘up’ and the position localization is achieved by multiplying |ψ7> by a delta function centered
(corresponding to pointing to) ‘down.’ For assuming, for simplicity of on the position corresponding to either the outcome ‘up’ or the outcome
presentation, that the position of all particles of the L-apparatus pointer in ‘down’ (see the entry on collapse theories, section 5 and Albert 1992,
the state of pointing to ‘up’ (‘down’) is the same, the state |ψ6> can be chapter 5); where the probability of each of these mutually exhaustive
rewritten as: possibilities is ½. The problem is that it follows from the uncertainty
principle (see the entry on the uncertainty principle) that in such
|ψ7> = 1/√2 (|z-up>L |up>p1 |up>p2 |up>p3 … |z-down> R − localizations the momenta and the energies of the localized particles
|z-down> L |down>p1 |down>p2 |down>p3 … |z-up>R) |ready >AR would be totally uncertain, so that gases may spontaneously heat up and
electrons may be knocked out of their orbits, in contradiction to our
where pi denotes the i-particle of the L-apparatus pointer, and |up> pi
experience. To avoid this kind of problems, GRW postulated that
(|down> pi ) is the state of the i-particle being in the position corresponding
spontaneous localizations are characterized by multiplications by
to the outcome z-spin ‘up’ (‘down’). [10] The chance that at least one of
Gaussians that are centered around certain positions, e.g. the position
the vast number of the pointer's particles will endure a spontaneous
corresponding to either the outcome ‘up’ or the outcome ‘down’ in the
localization toward being in the position corresponding to either the
state |ψ7>. This may be problematic, because in either case the state of
outcome z-spin ‘up’ or the outcome z-spin ‘down’ within a very short
the L-apparatus pointer at (what we characteristically conceive as) the
time (a split of a micro second) is very high. And since all the particles of
end of the L-measurement would be a superposition of the positions ‘up’
the pointer and the particle pair are entangled with each other, such a
and ‘down.’ For although this superposition ‘concentrates’ on either the
collapse will carry with it a collapse of the entangled state of the pointer
outcome ‘up’ or the outcome ‘down’ (i.e., the peak of the wave function
of the L-apparatus and the particle pair toward either
that corresponds to this state concentrates on one of these positions), it
|z-up> L |up> p1 |up> p2 |up> p3 … |z-down>R also has ‘tails’ that go everywhere: The state of the L-apparatus is a
superposition of an infinite number of different positions. Thus, it follows
or from the eigenstate-eigenvalue link that the position observable of the L-
apparatus has no definite value at the end of the measurement. But if the
|z-down>L |down>p1 |down>p2 |down>p3 … |z-up> R .
position observable having a definite value is indeed required in order for
Thus, the pointer will very quickly move in the direction of pointing to the L-apparatus to have a definite location, then the pointer will point to
either the outcome z-spin ‘up’ or the outcome z-spin ‘down.’ neither ‘up’ nor ‘down,’ and the GRW collapse models will fail to
reproduce the classical-like behavior of such systems. [11]
If (as portrayed above) the spontaneous localization of particles were to a
precise position, i.e., to the position corresponding to the outcome ‘up’ or In later models, GRW proposed to interpret the quantum state as a density
the outcome ‘down,’ the GRW collapse models would successfully of mass and they postulated that if almost all the density of mass of a
system is concentrated in a certain region, then the system is located in
24 Stanford Encyclopedia of Philosophy Spring 2009 Edition 25
Action at a Distance in Quantum Mechanics Joseph Berkovitz

system is concentrated in a certain region, then the system is located in


5.2 Can action-at-a-distance co-exist with non-separability and
that region. Accordingly, pointers of measurement apparatuses do have holism?
definite positions at the end of measurement interactions. Yet, this
solution has also given rise to a debate (see Albert and Loewer 1995, The action at a distance in the GRW/Pearle models is different from the
Lewis 1997, 2003a, 2004, Ghirardi and Bassi 1999, Bassi and Ghirardi Newtonian action at a distance in various respects. First, in contrast to
1999, 2001, Clifton and Monton 1999, 2000, Frigg 2003, and Parker Newtonian action at a distance, this action is independent of the distance
2003). between the measurement events. Second, while Newtonian action is
symmetric, the action in the GRW/Pearle models is (generally)
The exact details of the collapse mechanism and its characteristics in the
asymmetric: Either the L-measurement influences the properties of the R-
GRW/Pearle models have no significant implications for the type of non-
particle or the R-measurement influences the properties of the L-particle,
separability and holism they postulate—all these models basically
depending on which measurement occurs first (the action will be
postulate the same kinds of non-separability and holism as orthodox
symmetric when both measurements occur simultaneously). Third (and
quantum mechanics (see section 5.1.1). And action at a distance between
more important to our consideration), in contrast to Newtonian action at a
the L- and the R-wing will occur if the L-measurement interaction, a
distance, before the end of the L-measurement the state of the L-
supposedly local event in the L-wing, causes some local events in the R-
apparatus and the R-particle is not separable and accordingly it is not
wing, such as the event of the pointer of the measurement apparatus
clear that the influence is between separate existences, as the case is
coming to possess a definite measurement outcome during the R-
supposed to be in Newtonian gravity.
measurement. That is, action at a distance will occur if the L-
measurement causes the R-particle to come to possess a definite z-spin This non-separability of the states of the particle pair and the L-
and this in turn causes the pointer of the R-apparatus to come to possess measurement apparatus, and more generally the fact that the non-locality
the corresponding measurement outcome in the R-measurement. in collapse theories is due to state non-separability, has led a number of
Furthermore, if the L-measurement causes the R-particle to come to philosophers and physicists to think that wave collapses do not involve
possess (momentarily) a definite position in the R-wing, then the action at action at a distance. Yet, the question of whether there is an action at a
a distance between the L- and the R-wing will occur independently of distance in the GRW/Pearle models (and various other quantum theories)
whether the R-particle undergoes a spin measurement. depends on how we interpret the term ‘action at a distance.’ And, as I will
suggest below, on a natural reading of Isaac Newton's and Samuel
The above discussion is based on an intuitive notion of action at a
Clarke's comments concerning action at a distance, there may be a
distance and it presupposes that action at a distance is compatible with
peaceful coexistence between action at a distance and non-separability
non-separability and holism. In the next section we shall provide more
and holism.
precise characterizations of action at a distance and in light of these
characterizations reconsider the question of the nature of action at a Newton famously struggled to find out the cause of gravity.[12] In a letter
distance in the GRW/Pearle collapse models. to Bentley, dated January 17 1692/3, he said:

26 Stanford Encyclopedia of Philosophy Spring 2009 Edition 27


Action at a Distance in Quantum Mechanics Joseph Berkovitz

That this phenomenon is not produced sans moyen, that is without


You sometimes speak of Gravity as essential and inherent to
a cause capable of producing such an effect, is undoubtedly true.
Matter. Pray do not ascribe that Notion to me, for the Cause of
Philosophers therefore can search after and discover that cause, if
Gravity is what I do not pretend to know, and therefore would take
they can; be it mechanical or not. But if they cannot discover the
more Time to consider it. (Cohen 1978, p. 298)
cause, is therefore the effect itself, the phenomenon, or the matter
In a subsequent letter to Bentley, dated February 25, 1692/3, he added: of fact discovered by experience … ever the less true?

It is inconceivable that inanimate Matter should, without the Newton's and Clarke's comments suggest that for them gravity was a law-
Mediation of something else, which is not material, operate upon, governed phenomenon, i.e., a phenomenon in which objects influence
and affect other matter without mutual Contact…That Gravity each other at a distance according to the Newtonian law of gravity, and
should be innate, inherent and essential to Matter, so that one body that this influence is due to some means which may be invisible and
may act upon another at a distance thro’ a Vacuum, without the intangible and of a different nature from mechanism. On this conception
Mediation of any thing else, by and through which their Action of action at a distance, there seems to be no reason to exclude the
and Force may be conveyed from one to another, is to me so great possibility of action at a distance in the quantum realm even if that realm
an Absurdity that I believe no Man who has in philosophical is holistic or the state of the relevant systems is non-separable. That is,
Matters a competent Faculty of thinking can ever fall into it. action at a distance may be characterized as follows:
Gravity must be caused by an Agent acting constantly according to
Action at a distance is a phenomenon in which a change in
certain laws; but whether this Agent be material or immaterial, I
intrinsic properties of one system induces a change in the intrinsic
have left to the Consideration of my readers. (Cohen 1978, pp.
properties of a distant system, independently of the influence of
302-3)
any other systems on the distant system, and without there being a
Samuel Clarke, Newton's follower, similarly struggled with the question process that carries this influence contiguously in space and time.
of the cause of gravitational phenomenon. In his famous controversy with
We may alternatively characterize action at a distance in a more liberal
Leibniz, he said:[13]
way:
That one body attracts another without any intermediate means, is
Action* at a distance is a phenomenon in which a change in
indeed not a miracle but a contradiction; for 'tis supposing
intrinsic properties of one system induces a change in the intrinsic
something to act where it is not. But the means by which two
properties of a distant system without there being a process that
bodies attract each other, may be invisible and intangible and of a
carries this influence contiguously in space and time.
different nature from mechanism …
And while Newton and Clarke did not have an explanation for the action
And he added:
at a distance involved in Newtonian gravity, on the above
characterizations action at a distance in the quantum realm would be

28 Stanford Encyclopedia of Philosophy Spring 2009 Edition 29


Action at a Distance in Quantum Mechanics Joseph Berkovitz

characterizations action at a distance in the quantum realm would be and their wave function determine the outcomes of any measurements (so
explained by the holistic nature of the quantum realm and/or non- long as these outcomes are recorded in the positions of some physical
separability of the states of the systems involved. In particular, if in the systems, as in any practical measurements).
EPR/B experiment the L-apparatus pointer has a definite position before
the L-measurement and the R-particle temporarily comes to possess There are various versions of Bohm's theory. In the ‘minimal’ Bohm
definite position during the L-measurement, then the GRW/Pearle models theory, formulated by Bell (1982), [15] the wave function is interpreted as
involve action at a distance and thus also action* at a distance. On the a ‘guiding’ field (which has no source or any dependence on the particles)
other hand, if the R-particle never comes to possess a definite position that deterministically governs the trajectories of the particles according to
during the L-measurement, then the GRW/Pearle models only involve the so-called ‘guiding equation’ (which expresses the velocities of the
action* at a distance. particles in terms of the wave function). [16] The states of systems are
separable (the state of any composite system is completely determined by
5.3 No-collapse theories the state of its subsystems), and they are completely specified by the
particles’ positions. Spins, and any other properties which are not directly
5.3.1 Bohm's theory derived from positions, are not intrinsic properties of systems. Rather,
they are relational properties that are determined by the systems’ positions
In 1952, David Bohm proposed a deterministic, ‘hidden variables’ and the guiding field. In particular, each of the particles in the EPR/B
quantum theory that reproduces all the observable predictions of orthodox experiment has dispositions to ‘spin’ in various directions, and these
quantum mechanics (see Bohm 1952, Bohm, Schiller and Tiomno 1955, dispositions are relational properties of the particles— they are (generally)
Bell 1982, Dewdney, Holland and Kyprianidis 1987, Dürr, Goldstein and determined by the guiding field and the positions of the particles relative
Zanghì 1992a, 1997, Albert 1992, Valentini 1992, Bohm and Hiley 1993, to the measurement apparatuses and to each other.
Holland 1993, Cushing 1994, and Cushing, Fine and Goldstein 1996; for
an entry level review, see the entry on Bohmian mechanics and Albert
1992, chapter 5).

In contrast to orthodox quantum mechanics and the GRW/Pearle collapse


models, in Bohm's theory wave functions always evolve according to the
Schrödinger equation, and thus they never collapse. Wave functions do
not represent the states of systems. Rather, they are states of a ‘quantum
field (on configuration space)’ that influences the states of systems. [14]
Also, particles always have definite positions, and the positions of the
particles and their wave function at a certain time jointly determine the
trajectories of the particles at all future times. Thus, particles’ positions
and their wave function determine the outcomes of any measurements (so
Figure 4. The EPR/B experiment with Stern-Gerlach
30 Stanford Encyclopedia of Philosophy Spring 2009 Edition 31
Action at a Distance in Quantum Mechanics Joseph Berkovitz

Figure 4. The EPR/B experiment with Stern-Gerlach the particles that are emitted below this plane, like the L-particles 4-6,
measurement devices. Stern-Gerlach 1 is on, set up to measure the will be disposed to spin ‘down.’ Similarly, if the R-measurement occurs
z-spin of the L-particle, and Stern-Gerlach 2 is off. The horizontal before the L-measurement, the guiding field and the position of the R-
lines in the left-hand-side denote the trajectories of six L-particles particle at the emission time jointly determine the disposition of the R-
in the spin singlet state after an (impulsive) z-spin measurement particle to emerge either above the z-axis (i.e., to z-spin ‘up’) or below the
on the L-particle, and the horizontal lines in the right-hand-side z-axis (i.e., to z-spin ‘down’) according to whether it is above or below
denote the trajectories of the corresponding R-particles. The center the center plane, independently of the position of the L-particle along the
plane is aligned orthogonally to the z-axis, so that particles that z-axis.
emerge above this plane correspond to z-spin ‘up’ outcome and
particles that emerge below this plane correspond to z-spin ‘down’ But the z-spin disposition of the R-particle changes immediately after an
outcome. The little arrows denote the z-spin components of the (earlier) z-spin measurement on the L-particle: The R-particles 1-3 (see
particles in the ‘non-minimal’ Bohm theory (where spins are Fig. 4), which were previously disposed to z-spin ‘up,’ will now be
intrinsic properties of particles), and are irrelevant for the disposed to z-spin ‘down,’ i.e., to emerge below the center plane aligned
‘minimal’ Bohm theory (where spins are not intrinsic properties of orthogonally to the z-axis; and the R-particles 4-6, which were previously
particles). disposed to z-spin ‘down,’ will now be disposed to z-spin ‘up,’ i.e., to
emerge above this center plane. Yet, the L-measurement per se does not
To see the nature of non-locality postulated by the minimal Bohm theory, have any immediate influence on the state of the R-particle: The L-
consider again the EPR/B experiment and suppose that the measurement measurement does not influence the position of the R-particle or any
apparatuses are Stern-Gerlach (S-G) magnets which are prepared to other property that is directly derived from this position. It only changes
measure z-spin. In any run of the experiment, the measurement outcomes the guiding field, and thus grounds new spin dispositions for the R-
will depend on the initial positions of the particles and the order of the particle. But these dispositions are not intrinsic properties of the R-
measurements. Here is why. In the minimal Bohm theory, the spin singlet particle. Rather, they are relational properties of the R-particle, which are
state denotes the relevant state of the guiding field rather than the intrinsic grounded in the positions of both particles and the state of the guiding
properties of the particle pair. If the L-measurement occurs before the R- field.[17] (Note that in the particular case in which the L-particle is
measurement, the guiding field and the position of the L-particle at the emitted above the center plane aligned orthogonally to the z-axis and the
emission time jointly determine the disposition of the L-particle to R-particle is emitted below that plane, an earlier z-spin on the L-particle
emerge from the S-G device either above or below a plane aligned in the will have no influence on the outcome of a z-spin on the R-particle.)
z-direction; where emerging above (below) the plane means that the L-
particle z-spins ‘up’ (‘down’) about the z-axis and the L-apparatus While there is no contiguous process to carry the influence of the L-
‘pointer’ points to ‘up’ (‘down’) (see Fig. 4 above). All the L-particles measurement outcome on events in the R-wing, the question of whether
that are emitted above the center plane aligned orthogonally to the z- this influence amounts to action at a distance depends on the exact
direction, like the L-particles 1-3, will be disposed to spin ‘up’; and all characterization of this term. In contrast to the GRW/Pearle collapse
the particles that are emitted below this plane, like the L-particles 4-6, models, the influence of the L-measurement outcome on the intrinsic

32 Stanford Encyclopedia of Philosophy Spring 2009 Edition 33


Action at a Distance in Quantum Mechanics Joseph Berkovitz

models, the influence of the L-measurement outcome on the intrinsic does not involve state non-separability. For recall that in this theory the
properties of the R-particle is dependent on the R-measurement: Before state of a system does not consist in its wave function, but rather in the
this measurement occurs, there are no changes in the R-particle's intrinsic system's position, and the position of a composite system always
properties. Yet, the influence of the L-measurement on the R-particle is at factorizes into the positions of its subsystems. Here, the non-separability
a distance. Thus, the EPR/B experiment as depicted by the minimal Bohm of the wave function reflects the state of the guiding field. This state
theory involves action* at a distance but not action at a distance. propagates not in ordinary three-space but in configuration space, where
each point specifies the configuration of both particles. The guiding field
Bohm's theory portrays the quantum realm as deterministic. Thus, the of the particle pair cannot be factorized into the guiding field that governs
single-case objective probabilities, i.e., the chances, it assigns to the trajectory of the L-particle and the guiding field that governs the
individual spin-measurement outcomes in the EPR/B experiment are trajectory of the R-particle. The evolution of the particles’ trajectories,
different from the corresponding quantum-mechanical probabilities. In properties and dispositions is non-separable, and accordingly the
particular, while in quantum mechanics the chances of the outcomes ‘up’ particles’ trajectories, properties and dispositions are correlated even
and ‘down’ in an earlier L- (R-) spin measurement are both ½, in Bohm's when the particles are far away from each other and do not interact with
theory these chances are either one or zero. Yet, Bohm's theory postulates each other. Thus, process separability fails.
a certain distribution, the so-called ‘quantum-equilibrium distribution,’
over all the possible positions of pairs with the same guiding field. This In the non-minimal Bohm theory [18], the behavior of an N-particle system
distribution is computed from the quantum-mechanical wave function, is determined by its wave function and the intrinsic properties of the
and it is typically interpreted as ignorance over the actual position of the particles. But, in contrast to the minimal theory, in the non-minimal
pair; an ignorance that may be motivated by dynamical considerations and theory spins are intrinsic properties of particles. The wave function
statistical patterns exhibited by ensembles of pairs with the same wave always evolves according to the Schrödinger equation, and it is
function (for more details, see the entry on bohmian mechanics, section interpreted as a ‘quantum field’ (which has no sources or any dependence
9). And the sum-average (or more generally the integration) over this on the particles). The quantum field guides the particles via the ‘quantum
distribution reproduces all the quantum-mechanical observable potential,’ an entity which is determined from the quantum field, and the
predictions. evolution of properties is fully deterministic.[19]

What is the status of this probability postulate? Is it a law of nature or a Like in the minimal Bohm theory, the non-separability of the wave
contingent fact (if it is a fact at all)? The answers to these questions vary function in the EPR/B experiment dictates that the evolution of the
(see Section 7.2.1, Bohm 1953, Valentini 1991a,b, 1992, 1996, 2002, particles’ trajectories, properties and dispositions is non-separable, but the
Valentini and Westman 2004, Dürr, Goldstein and Zanghì 1992a,b, 1996, behavior of the particles is somewhat different. In the earlier z-spin
fn. 15, and Callender 2006). measurement on the L-particle, the quantum potential continuously
changes, and this change induces an immediate change in the z-spin of the
Turning to the question of non-separability, the minimal Bohm theory R-particle. If the L-particle starts to spin ‘up’ (‘down’) in the z-direction,
does not involve state non-separability. For recall that in this theory the the R-particle will start to spin ‘down’ (‘up’) in the same direction (see

34 Stanford Encyclopedia of Philosophy Spring 2009 Edition 35


Action at a Distance in Quantum Mechanics Joseph Berkovitz

the R-particle will start to spin ‘down’ (‘up’) in the same direction (see always evolve according to unitary and linear dynamical equations (the
the little arrows in Fig. 4).[20] Accordingly, the L-measurement induces Schrödinger equation in the non-relativistic case). And the orthodox
instantaneous action at a distance between the L- and the R-wing. Yet, quantum-mechanical state description of systems is supplemented by a set
similarly to the minimal Bohm theory, while the disposition of the R- of properties, which depends on the quantum-mechanical state and which
particle to emerge above or below the center plane aligned orthogonally to is supposed to be rich enough to account for the occurrence of definite
the z-direction in a z-spin measurement may change instantaneously, the macroscopic events and their classical-like behavior, but sufficiently
actual trajectory of the R-particle along the z-direction does not change restricted to escape all the known no-hidden-variables theorems. (For
before the measurement of the R-particle's z-spin occurs. Only during the modal interpretations, see van Fraassen 1973, 1981, 1991, chapter 9,
R-measurement, the spin and the position of the R-particle get correlated Kochen 1985, Krips 1987, Dieks 1988, 1989, Healey 1989, Bub 1992,
and the R-particle's trajectory along the z-direction is dictated by the 1994, 1997, Vermaas and Dieks 1995, Clifton 1995, Bacciagaluppi 1996,
value of its (intrinsic) z-spin. Bacciagaluppi and Hemmo 1996, Bub and Clifton 1996, Hemmo 1996b,
Bacciagaluppi and Dickson 1999, Clifton 2000, Spekkens and Sipe
Various objections have been raised against Bohm's theory (for a detailed 2001a,b, Bene and Dieks 2002, and Berkovitz and Hemmo 2006a,b. For
list and replies, see the entry on Bohmian mechanics, section 15). One an entry-level review, see the entry on modal interpretations of quantum
main objection is that in Bohmian mechanics, the guiding field influences theory. For comprehensive reviews and analyses of modal interpretations,
the particles, but the particles do not influence the guiding field. Another see Bacciagaluppi 1996, Hemmo 1996a, chapters 1-3, Dieks and Vermaas
common objection is that the theory is involved with a radical type of 1998, Vermaas 1999, and the entry on modal interpretations of quantum
non-locality, and that this type of non-locality is incompatible with theory. For the no-hidden-variables theorems, see Kochen and Specker
relativity. While it may be very difficult, or even impossible, to reconcile 1967, Greenberger, Horne and Zeilinger 1989, Mermin 1990 and the entry
Bohm's theory with relativity, as is not difficult to see from the above on the Kochen-Specker theorem.) [21]
discussion, the type of non-locality that the minimal Bohm theory
postulates in the EPR/B experiment does not seem more radical than the Modal interpretations vary in their property assignment. For simplicity,
non-locality postulated by the orthodox interpretation and the we shall focus on modal interpretations in which the property assignment
GRW/Pearle collapse models. is based on the so-called Schmidt biorthogonal-decomposition theorem
(see Kochen 1985, Dieks 1989, and Healey 1989). Let S1 and S2 be
5.3.2 Modal interpretations systems associated with the Hilbert spaces HS1 and HS2, respectively.
There exist bases {|αi>} and {|β i>} for HS1 and HS2 respectively such
Modal interpretations of quantum mechanics were designed to solve the that the state of S1+S2 can be expressed as a linear combination of the
measurement problem and to reconcile quantum mechanics with following form of vectors from these bases:
relativity. They are no-collapse, (typically) indeterministic hidden-
variables theories. Quantum-mechanical states of systems (which may be |ψ8 > S1+S2 = ∑ i ci |α i> S1 |β i> S2.
construed as denoting their states or information about these states)
always evolve according to unitary and linear dynamical equations (the

36 Stanford Encyclopedia of Philosophy Spring 2009 Edition 37


Action at a Distance in Quantum Mechanics Joseph Berkovitz

(similarly to the state |ψ9>) the L- and the R-particle have definite z-spin
When the absolute values of the coefficients ci are all unequal, the bases properties: Either the L-particle has z-spin ‘up’ and the R-particle has z-
{|αi>} and {|β i>} and the above decomposition of |ψ 8 > S1+S2 are unique. spin ‘down,’ or the L-particle has z-spin ‘down’ and the R-particle has z-
In that case, it is postulated that S1 has a determinate value for each spin ‘up,’ [24] where the probability of the realization of each of these
observable associated with HS1 with the basis {|α i>} and S2 has a possibilities is approximately 1/2. In the (earlier) z-spin measurement on
determinate value for each observable associated with HS2 with the basis the L-particle, the state of the particle pair and the apparatuses evolves to
{|βi>}, and |ci| 2 provide the (ignorance) probabilities of the possible the state:
values that these observables may have.[22] For example, suppose that the
state of the L- and the R-particle in the EPR/B experiment before the |ψ11> = ((1/√2+ε) |z-up>L|up>AL | z-down> R −

measurements is: (1/√2-ε′) |z-down> L|down>AL | z-up>R) |ready>AR

|ψ9> = (1/√2+ε) |z-up> L| z-down>R − (1/√2-ε′) |z-down>L| z-up> R where (as before) |up> AL and |down>AL denote the states of the L-
apparatus pointing to the outcomes z-spin ‘up’ and z-spin ‘down’,
where 1/√2 >> ε,ε′, (1/√2+ε) 2+(1/√2-ε′)2 = 1, and (as before) |z-up> L (|z- respectively. In this state, either the L-particle has a z-spin ‘up’ and the L-
up> R ) and | z-down>L (| z-down>R ) denote the states of the L- (R-) apparatus points to ‘up,’ or the L-particle has z-spin ‘down’ and the L-
particle having z-spin ‘up’ and z-spin ‘down’, respectively.[23] Then, apparatus points to ‘down.’ And, again, the probability of each of these
either the L-particle spins ‘up’ and the R-particle spins ‘down’ in the z- possibilities is approximately 1/2. The evolution of the properties from
direction, or the L-particle spins ‘down’ and the R-particle spins ‘up’ in the state |ψ10> to the state |ψ 11> depends on the dynamical laws. In
the z-direction. Thus, in contrast to the orthodox interpretation and the almost all modal interpretations, if the particles have definite z-spin
GRW/Pearle collapse models, in modal interpretations the particles in the properties before the measurements, the outcomes of z-spin measurements
EPR/B experiment may have definite spin properties even before any will reflect these properties. That is, the evolution of the properties of the
measurement occurs. particles and the measurement apparatuses will be deterministic, so that
the spin properties of the particles do not change in the L-measurement
To see how the modal interpretation accounts for the curious correlations and the pointer of the L-apparatus comes to display the outcome that
in EPR/B-type experiments, let us suppose that the state of the particle corresponds to the z-spin property that the L-particle had before the
pair and the measurement apparatuses at the emission time is: measurement. If, for example, before the measurements the L- and the R-
particle have respectively the properties z-spin ‘up’ and z-spin ‘down’, the
|ψ10> =
(earlier) z-spin measurement on the L-particle will yield the outcome ‘up’
((1/√2+ε) |z-up>L |z-down> R − (1/√2−ε′) |z-down> L |z-up>R) |ready>AL |ready>AR
and the spin properties of the particles will remain unchanged.
where |ready>AL (|ready> AR) denotes the state of the L-apparatus (R- Accordingly, a z-spin measurement on the R-particle will yield the
apparatus) being ready to measure z-spin. In this state, the L- and the R- outcome ‘down’. Thus, in this case the modal interpretation involves
apparatus are in the definite state of being ready to measure z-spin, and neither action at a distance nor action* at a distance.
(similarly to the state |ψ9>) the L- and the R-particle have definite z-spin

38 Stanford Encyclopedia of Philosophy Spring 2009 Edition 39


Action at a Distance in Quantum Mechanics Joseph Berkovitz

For nothing in the above property assignment implies that in |ψ11> the
However, if the measurement apparatuses are set up to measure x-spin spin properties that the L-particle has as a ‘separated’ system and the spin
rather than z-spin, the evolution of the properties of the L-particle and the properties that it has as a subsystem of the particle pair be the same: The
L-apparatus will be indeterministic. As before, the L-measurement will L-particle may have z-spin ‘up’ as a separated system and z-spin ‘down’
not cause any change in the actual spin properties of the R-particle. But as a subsystem of the particle pair.
the L-measurement outcome will cause an instant change in the spin
dispositions of the R-particle and the R-measurement apparatus. If, for Furthermore, the dynamics of the properties that the L-particle (R-
example, the L-measurement outcome is x-spin ‘up’ and the L-particle particle) has as a separated system and the dynamics of its properties as a
comes to posses x-spin ‘up,’ then the R-particle and the R-apparatus will subsystem of the particle pair are generally different.[25] Consider, again,
have respectively the dispositions to possess x-spin ‘down’ and to display the state |ψ10>. In the (earlier) z-spin measurement on the L-particle, the
the outcome x-spin ‘down’ on a x-spin measurement. Thus, like the spin properties that the L-particle has as a separated system follow a
minimal Bohm theory, the modal interpretation may involve action* at a deterministic evolution — the L-particle has either z-spin ‘up’ or z-spin
distance in the EPR/B experiment. But, unlike the minimal Bohm theory, ‘down’ before and after the L-measurement; whereas as a subsystem of
here spins are intrinsic properties of particles. the particle pair, the spin properties of the L-particle follow an
indeterministic evolution — the L-particle has no definite spin properties
In the above modal interpretation, property composition fails: The before the L-measurement and either z-spin ‘up’ (with approximately
properties of composite systems are not decomposable into the properties chance ½) or z-spin ‘down’ (with approximately chance ½) after the L-
of their subsystems. Consider, again, the state |ψ10>. As ‘separated’ measurement.
systems (i.e., in the decompositions of the composite system of the
particle pair+apparatuses into the L-particle and the R- The failure of property composition implies that the quantum realm as
particle+apparatuses and into the R-particle and the L- depicted by the above version of the modal interpretation involves state
particle+apparatuses) the L- and the R-particle have definite z-spin non-separability and property and relational holism. State separability
properties. But, as subsystems of the composite system of the particle pair fails because the state of the particle pair is not generally determined by
(e.g. in the decomposition of the composite system of the particle the separate states of the particles. Indeed, as is easily shown, the actual
pair+apparatuses into the particle pair and the apparatuses), they have no properties that the L- and the R-particle each has in the state |ψ9> are
definite z-spin properties. also compatible with product states in which the L- and the R-particle are
not entangled. Property and relational holism fail because in the state |
A failure of property composition occurs also in the state |ψ11>, where ψ9> the properties of the pair do not supervene upon the properties of its
the L- and the R-particle have definite z-spin properties both as subsystems and the spatiotemporal relations between them. Furthermore,
‘separated’ systems and as subsystems of the particle pair (though in process separability fails for similar reasons.
contrast with |ψ10>, in |ψ 11> the range of the possible properties of the
particles as separated systems and as subsystems of the pair is the same). The failure of property composition in the modal interpretation calls for
For nothing in the above property assignment implies that in |ψ11> the explanation. It may be tempting to postulate that the properties that a
system (e.g. the L-particle) has, as a separated system, are the same as the

40 Stanford Encyclopedia of Philosophy Spring 2009 Edition 41


Action at a Distance in Quantum Mechanics Joseph Berkovitz

system (e.g. the L-particle) has, as a separated system, are the same as the located in neither the L-wing nor any other subregion of the universe.
properties that it has as a subsystem of composite systems. But, as Yet, due to the dynamical laws, properties like the position of pointers of
Bacciagaluppi (1995) and Clifton (1996a) have shown, such property measurement apparatuses, which appear to us to be local, behave like
assignment will be inconsistent: It will be subject to a Kochen and local properties in any experimental circumstances, and accordingly this
Specker-type contradiction. Furthermore, as Vermaas (1997) radical type of non-locality is unobservable (for more details, see
demonstrates, the properties of composite systems and the properties of Berkovitz and Hemmo 2006b, sections 8.1 and 9).
their subsystems cannot be correlated (in ways compatible with the Born
rule). Another way to try to explain the failure of property composition is to
interpret the properties of composite systems as holistic, non-
For what follows in the rest of this subsection, the views of different decomposable properties. On this interpretation, the z-spin ‘up’ property
authors differ widely. Several variants of modal interpretations were that the L-particle has as a subsystem of the particle pair in the state |ψ9>
developed in order to fix the problem of the failure of property is completely different from the z-spin ‘up’ property that the L-particle
composition. The most natural explanation of the failure of property has as a separated system, and the use of the term ‘z-spin up’ in both
composition is that quantum states assign relational rather than intrinsic cases is misleading (for more details, see Berkovitz and Hemmo
properties to systems (see Kochen 1985, Bene and Dieks 2002, and 2006a).[27]
Berkovitz and Hemmo 2006a,b). For example, in the relational modal
interpretation proposed by Berkovitz and Hemmo (2006a,b), the main The relational and holistic interpretations of properties mark a radical
idea is that quantum states assign properties to systems only relative to shift from the standard interpretation of properties in orthodox quantum
other systems, and properties of a system that are related to different mechanics. Other advocates of the modal interpretation have chosen not
systems are generally different. In particular, in the state |ψ10> the L- to follow this interpretation, and opted for a modal interpretation that does
particle has a definite z-spin property relative to the R-particle, the not violate property composition. While the property assignment above
measurement apparatuses and the rest of the universe, but (as a subsystem does not assume any preferred partition of the universe (the partition of
of the particle pair) it has no definite z-spin relative to the measurement the universe into a particle pair and the rest of the universe is as good as
apparatuses and the rest of universe.[26] On this interpretation, the the partition of the universe into the L-particle and the rest of the
properties of systems are highly non-local by their very nature. Properties universe), proponents of property composition postulated that there is a
like pointing to ‘up’ and pointing to ‘down’ are not intrinsic to the preferred partition of the universe into ‘atomic’ systems and accordingly a
measurement apparatuses. Rather, they are relations between the preferred factorization of the Hilbert space of the universe. This preferred
apparatuses and other systems. For example, the property of the L- factorization is supposed to be the basis for the ‘core’ property
apparatus pointing to ‘up’ relative to the particle pair, the R-apparatus and assignment: Properties are prescribed to atomic systems according to a
the rest of the universe is not intrinsic to the L-apparatus; it is a relation property assignment that is a generalization of the bi-orthogonal
between the L-apparatus and the particle pair, the R-apparatus and the decomposition property assignment.[28] And the properties of complex
rest of the universe. As such, this property is highly non-local: It is systems are postulated to be compositions of the properties of their atomic
located in neither the L-wing nor any other subregion of the universe. systems (see the entry on modal interpretations of quantum theory,

42 Stanford Encyclopedia of Philosophy Spring 2009 Edition 43


Action at a Distance in Quantum Mechanics Joseph Berkovitz

systems (see the entry on modal interpretations of quantum theory, are fundamentally relative. Systems have relative states, which are
section 2, and Bacciagaluppi and Dickson 1999). The challenge for this derivable from the various branches of the entangled states. For example,
atomic modal interpretation is to justify the assumption that there is a consider again |ψ11>.
preferred partition of the universe, and to provide some idea about how
such factorization should look like. |ψ11> = (1/√2+ε) |z-up>L|up>AL | z-down> R |ready>AR −
(1/√2-ε′) |z-down> L|down>AL | z-up>R |ready>AR .
Finally, while the modal interpretation was designed to solve the
measurement problem and reconcile quantum mechanics with special In this quantum-mechanical state, the L-apparatus is in the state of
relativity, it faces challenges on both accounts. First, in certain imprefect pointing to the outcome z-spin ‘up’ relative to the L-particle being in the
measurements (where there are imprefections in the coupling between the state z-spin ‘up,’ the R-particle being in the state z-spin ‘down’ and the
measured system and the pointer of the measurement apparatus and/or the R-apparatus being ready to measure z-spin; and in the state of pointing to
pointer and the environment), modal interpretations that are based on the the outcome z-spin ‘down’ relative to the L-particle being in the state z-
Schmidt biorthogonal-decomposition theorem (and more generally the spin ‘down,’ the R-particle being in the state z-spin ‘up’ and the R-
spectral decomposition theorem) fail to account for definite measurement apparatus being ready to measure z-spin. Likewise, the L-particle is in the
outcomes, in contradiction to our experience (see Bacciagaluppi and state z-spin ‘up’ relative to the L-apparatus being in the state of pointing
Hemmo 1996 and Bacciagaluppi 2000). For versions of the modal to the outcome z-spin ‘up,’ the R-particle being in the state z-spin ‘down’
interpretations that seem to escape this problem, see Van Fraassen (1973, and the R-apparatus being ready to measure z-spin; and in the state z-spin
1991), Bub (1992, 1997), Bene and Dieks (2002) and Berkovitz and ‘down’ relative to the L-apparatus being in the state of pointing to the
Hemmo (2006a,b). Second, as we shall see in section 10.2, a number of outcome z-spin ‘down,’ the R-particle being in the state z-spin ‘up’ and
no-go theorems challenge the view that modal interpretations could be the R-apparatus being ready to measure z-spin. And similarly, mutatis
genuinely relativistic. mutandis, for the relative state of the R-particle and the R-apparatus.

Everett's original formulation left the exact meaning of these relative


5.3.3 Everett-like interpretations
states and their relations to observers’ experience and beliefs open, and
In 1957, Everett proposed a new no-collapse interpretation of orthodox there have been different Everett-like interpretations of these states.
quantum mechanics (see Everett 1957a,b, 1973, Barrett 1999, the entry on Probably the most popular reading of Everett is the splitting-worlds
Everett's relative-state formulation of quantum mechanics, the entry on interpretation (see DeWitt 1971, Everett's relative-state formulation of
the many-worlds interpretation of quantum mechanics, and references quantum mechanics, Barrett 1999, and references therein). In the
therein). The Everett interpretation is a no-collapse interpretation of splitting-worlds interpretation, each of the branches of the state |ψ11>
quantum mechanics, where the evolution of quantum states is always refers to a different class of worlds (all of which are real) where the states
according to unitary and linear dynamical equations (the Schrödinger of the L-apparatus, R-apparatus and the particles are all separable: Class-
equation in the non-relativistic case). In this interpretation, quantum states 1 worlds in which the L-particle is in the state z-spin ‘up,’ the R-particle
are fundamentally relative. Systems have relative states, which are is in the state z-spin ‘down,’ the L-apparatus is in the state of pointing to

44 Stanford Encyclopedia of Philosophy Spring 2009 Edition 45


Action at a Distance in Quantum Mechanics Joseph Berkovitz

is in the state z-spin ‘down,’ the L-apparatus is in the state of pointing to Fourth, there is the question of how the splitting-worlds interpretation
the outcome z-spin ‘up’ and the R-apparatus in the state of being ready to accounts for the statistical predictions of the orthodox theory. In the
measure z-spin; and class-2 worlds in which the L-particle is in the state Everett-like interpretations in general, and in the splitting-worlds
z-spin ‘down,’ the R-particle is in the state z-spin ‘up,’ the L-apparatus is interpretation in particular, all the possible measurement outcomes in the
in the state of pointing to the outcome z-spin ‘down’ and the R-apparatus EPR/B experiment are realized and may be observed. Thus, the question
is in the state of being ready to measure z-spin. More generally, each term arises as to the meaning of probabilities in this interpretation. For
in state of the universe, as represented in a certain preferred basis, reflects example, what is the meaning of the statement that in the state |ψ10> (see
the states of its systems in some class of worlds; where the range of the section 5.3.2) the probability of the L-measurement apparatus pointing to
different classes of worlds increases whenever the number of the terms in the outcome ‘up’ in an earlier z-spin measurement on the L-particle is
the quantum state (in the preferred basis) increases (this process is called (approximately) ½? In the splitting-worlds interpretation the probability
‘splitting’). of that outcome appears to be 1! Furthermore, setting aside the problem of
interpretation, there is also the question of whether the splitting-worlds
The splitting-worlds reading of Everett faces a number of challenges. interpretation, and more generally Everett-like interpretations, can
First, supporters of the Everett interpretation frequently motivate their account for the particular values of the quantum probabilities of
interpretation by arguing that it postulates the existence of neither a measurement outcomes. Everett claimed to derive the Born probabilities
controversial wave collapse nor hidden variables, and it leaves the simple in the context of his interpretation. But this derivation has been
and elegant mathematical structure of quantum mechanics intact. But, the controversial. (For discussions of the meaning of probabilities, or more
splitting-worlds interpretation adds extra structure to no-collapse orthodox precisely the meaning of the coefficients of the various terms in quantum
quantum mechanics. Further, this interpretation marks a radical shift from states, in Everett-like interpretations, see Butterfield 1996, Lockwood
orthodox quantum mechanics. A scientific theory is not constituted only 1996a,b, Saunders 1998, Vaidman 1998, Barnum et al. 2000,
by its mathematical formalism, but also by the ontology it postulates, the Bacciagaluppi 2002, Gill 2003, Hemmo and Pitowsky 2003, 2005,
way it depicts the physical realm and the way it accounts for our Wallace 2002, 2003, 2005a,b, Greaves 2004 and Saunders 2004, 2005.)
experience. The many parallel worlds ontology of the splitting-worlds
interpretation and its account of our experience are radically different Other readings of Everett include the many-minds interpretation (Albert
from the ontology of the intended interpretation of orthodox quantum and Loewer 1988, Barrett 1999, chapter 7), the consistent-histories
mechanics and its account for our experience. Second, relative states are approach (Gell-Mann and Hartle 1990), the Everett-like relational
well defined in any basis, and the question arises as to which basis should interpretation (Saunders 1995, Mermin 1998) and (what may be called)
be preferred and the motivation for selecting one particular basis over the many-structures interpretation (Wallace 2005c). While these readings
others. Third, in the splitting-worlds interpretation each of the worlds in address more or less successfully the problems of the preferred basis and
the universe may split into two or more worlds, and the problem is that splitting, except for the many-minds interpretation of Albert and Loewer
(similarly to the collapse in orthodox collapse quantum mechanics) there the question of whether there could be a satisfactory interpretation of
are no clear criteria for when a splitting occurs and how long it takes. probabilities in the context of these theories and the adequacy of the
Fourth, there is the question of how the splitting-worlds interpretation derivation of the Born probabilities are still a controversial issue (see
46 Stanford Encyclopedia of Philosophy Spring 2009 Edition 47
Action at a Distance in Quantum Mechanics Joseph Berkovitz

derivation of the Born probabilities are still a controversial issue (see of perceiving definite measurement outcomes. For example, consider
Deutsch 1999, Wallace 2002, 2003, Lewis 2003, Graves 2004, Saunders again, the state |ψ10>. While in a first z-spin L-measurement, this state
2004, Hemmo and Pitowsky 2005, and Price 2006). evolves deterministically into the state |ψ11>, minds of observers evolve
indeterministically into either the state of perceiving the outcome z-spin
What kind of non-locality do Everett-like interpretations involve? ‘up’ or the state of perceiving the outcome z-spin ‘down’ with the usual
Unfortunately, the answer to this question is not straightforward, as it Born-rule probabilities (approximately 50% chance for each of these
depends on one's particular reading of the Everett interpretation. Indeed, outcomes). Since in this state the L-particle has no definite spin properties
all the above readings of Everett seem to treat the no-collapse wave and the L-apparatus points to no definite measurement outcome, and since
function of the universe as a real physical entity that reflects the non- in the later z-spin measurement on the R-particle the R-particle does not
separable state of the universe, and accordingly they involve state non- come to possess any definite spin properties and the R-apparatus points to
separability. But, one may reasonably expect that different readings depict no definite spin outcome, the question of whether there is action at a
different pictures of physical reality and accordingly might postulate distance between the L-particle and the L-apparatus on the one hand and
different kinds of non-locality. Thus, any further analysis of the type of the R-particle and the R-apparatus on the other does not arise.
non-locality postulated by each of these readings requires a detailed study
of their ontology (which we plan to conduct in future updates of this 6. Superluminal causation
entry).
In all the above interpretations of quantum mechanics, the failure of
For example, the question of action at a distance in the EPR/B experiment
factorizability (i.e., the failure of the joint probability of the measurement
may arise in the context of the splitting-worlds interpretation, but not in
outcomes in the EPR/B experiment to factorize into their single
the context of Albert and Loewer's many-minds interpretation. Albert and
probabilities) involves non-separability, holism and/or some type of
Loewer's interpretation takes the bare no-collapse orthodox quantum
action at a distance. As we shall see below, non-factorizability also
mechanics to be the complete theory of the physical realm. Accordingly,
implies superluminal causal dependence according to certain accounts of
the L-apparatus in the state |ψ11> does not display any definite outcome.
causation.
Yet, in order to account for our experience of a classical-like world,
where at the end of measurements observers are typically in mental states First, as is not difficult to show, the failure of factorizability implies
of perceiving definite outcomes, the many-minds interpretation appeals to superluminal causation according to various probabilistic accounts of
a dualism of mind-body. Each observer is associated with a continuous causation that satisfy Reichenbach's (1956, section 19) principle of the
infinity of non-physical minds. And while the physical state of the world common cause (for a review of this principle, see the entry on
evolves in a completely deterministic manner according to the Reichenbach's principle of the common cause).
Schrödinger evolution, and the pointers of the measurement apparatuses
in the EPR/B experiment display no definite outcomes, states of minds Here is why. Reichenbach's principle may be formulated as follows:
evolve in a genuinely indeterministic fashion so as to yield an experience
of perceiving definite measurement outcomes. For example, consider

48 Stanford Encyclopedia of Philosophy Spring 2009 Edition 49


Action at a Distance in Quantum Mechanics Joseph Berkovitz

PCC (Principle of the Common Cause). For any correlation PC(x), PC(y) and CC(x,y):
between two (distinct) events which do not cause each other, there
is a common cause that screens them off from each other. Or FactorUCP
formally: If distinct events x and y are correlated, i.e., P PC(x) PC(y) CC(x,y) (x & y) = P PC(x) CC(x,y) (x) · P PC(y) CC(x,y) (y).

(Correlation) Like PCC, the basic idea of FactorUCP is that the objective probabilities
P(x & y) ≠ P(x) · P(y), of events that do not cause each other are determined by their causal
pasts, and given these causal pasts they are probabilistically independent
and they do not cause each other, then their common cause, of each other. As is not difficult to see, factorizability is a special case of
CC(x,y), screens them off from each other, i.e., FactorUCP. That is, to obtain factorizability from FactorUCP, substitute λ
for CC(x,y), l for PC(x) and r for PC(y). FactorUCP and the assumption
(Screening Off) that the probabilities of the measurement-outcomes in the EPR/B
P CC(x,y)(x/y) = P CC(x,y)(x) P CC(x,y)(y) ≠ 0 experiment are determined by the pair's state and the settings of the
measurement apparatuses jointly imply factorizability. Thus, given this
P CC(x,y)(y/x) = P CC(x,y)(y) P CC(x,y)(x) ≠ 0.[29]
later assumption, the failure of factorizability implies superluminal
causation between the distant outcomes in the EPR/B experiment
Accordingly, CC(x,y) renders x and y probabilistically
according to any account of causation that satisfies FactorUCP (for some
independent, and the joint probability of x and y factorizes upon
examples of such accounts, see Butterfield 1989 and Berkovitz 1995a,
CC(x,y):
1995b, section 6.7, 1998b).[30]
P CC(x,y)(x & y) = P CC(x,y)(x) · P CC(x,y)(y).
Superluminal causation between the distant outcomes also exists
The above formulation of PCC is mainly intended to cover cases in which according to various counterfactual accounts of causation, including
x and y have no partial, non-common causes. But PCC can be generalized accounts that do not satisfy FactorUCP. In particular, in Lewis's (1986)
as follows: influential account, counterfactual dependence between distinct events
implies causal dependence between them. And as Butterfield (1992b) and
PCC*. The joint probability of any distinct, correlated events, x Berkovitz (1998b) demonstrate, the violation of Factorizability involves a
and y, which are not causally connected to each other, factorizes counterfactual dependence between the distant measurement outcomes in
upon the union of their partial (separate) causes and their common the EPR/B experiment.
cause. That is, let CC(x,y) denote the common causes of x and y,
and PC(x) and PC(y) denote respectively their partial causes. But the violation of factorizability does not imply superluminal causation
Then, the joint probability of x and y factorizes upon the Union of according to some other accounts of causation. In particular, in process
their Causal Pasts (henceforth, FactorUCP), i.e., on the union of accounts of causation there is no superluminal causation in the EPR/B
PC(x), PC(y) and CC(x,y): experiment. In such accounts, causal dependence between events is
explicated in terms of continuous processes in space and time that
50 Stanford Encyclopedia of Philosophy Spring 2009 Edition 51
Action at a Distance in Quantum Mechanics Joseph Berkovitz

explicated in terms of continuous processes in space and time that and 118). It is thus frequently claimed with respect to EPR/B experiments
transmit ‘marks’ or conserved quantities from the cause to the effect (see that there is no such thing as a Bell telephone, namely a telephone that
Salmon 1998, chapters 1, 12, 16 and 18, Dowe 2000, the entry on causal could exploit the violation of the Bell inequalities for superluminal
processes, and references therein). Thus, recalling (sections 1, 2, 4 and 5) signaling of information.[31]
that none of the interpretations of quantum mechanics and alternative
quantum theories postulates any (direct) continuous process between the The no-signaling theorem demonstrates that orthodox quantum mechanics
distant measurement events in the EPR/B experiment, there is no excludes any possibility of superluminal signaling in the EPR/B
superluminal causation between them according to process accounts of experiment. According to this theory, no controllable physical factor in
causation. the L-wing, such as the setting of the L-measurement apparatus, can take
advantage of the entanglement between the systems in the L- and the R-
7. Superluminal signaling wing to influence the statistics of the measurement outcomes (or any
other observable) in the R-wing. As we have seen in section 5.1.1, the
Whether or not the non-locality predicted by quantum theories may be orthodox theory is at best incomplete. Thus, the fact that it excludes
classified as action at a distance or superluminal causation, the question superluminal signaling does not imply that other quantum theories or
arises as to whether this non-locality could be exploited to allow interpretations of the orthodox theory also exclude such signaling. Yet, if
superluminal (i.e., faster-than-light) signaling of information. This the orthodox theory is empirically adequate, as the consensus has it, its
question is of particular importance for those who interpret relativity as statistical predictions obtain, and accordingly superluminal signaling will
prohibiting any such superluminal signaling. (We shall return to discuss be excluded as a matter of fact; for if this theory is empirically adequate,
this interpretation in section 10.) any quantum theory will have to reproduce its statistics, including the
exclusion of any actual superluminal signaling.
Superluminal signaling would require that the state of nearby controllable
physical objects (say, a keyboard in my computer) superluminally But the no-signaling theorem does not demonstrate that superluminal
influence distant observable physical phenomena (e.g. a pattern on a signaling would be impossible if orthodox quantum mechanics were not
computer screen light years away). The influence may be deterministic or empirically adequate. Furthermore, this theorem does not show that
indeterministic, but in any case it should cause a detectable change in the superluminal signaling is in principle impossible in the quantum realm as
statistics of some distant physical quantities. depicted by other theories, which actually reproduce the statistics of
orthodox quantum mechanics but do not prohibit in theory the violation
It is commonly agreed that in quantum phenomena, superluminal of this statistics. In sections 7.2-7.3, we shall consider the in-principle
signaling is impossible in practice. Moreover, many believe that such possibility of superluminal signaling in certain collapse and no-collapse
signaling is excluded in principle by the so-called ‘no-signaling theorem’ interpretations of quantum mechanics. But, first, we need to consider the
(for proofs of this theorem, see Eberhard 1978, Ghirardi, Rimini and necessary and sufficient conditions for superluminal signaling.
Weber 1980, Jordan 1983, Shimony 1984, Redhead 1987, pp. 113-116
and 118). It is thus frequently claimed with respect to EPR/B experiments

52 Stanford Encyclopedia of Philosophy Spring 2009 Edition 53


Action at a Distance in Quantum Mechanics Joseph Berkovitz

7.1 Necessary and sufficient conditions for superluminal sufficient condition for controllable probabilistic dependence. But, recall
signaling (footnote 3) that in some models of the EPR/B experiment, in addition to
the pair's state and the setting of the L- (R-) measurement apparatus there
To simplify things, in our discussion we shall focus on non-factorizable are other local physical quantities that may be relevant for the probability
models of the EPR/B experiment that satisfy λ-independence (i.e., the of the L- (R-) measurement outcome. In such models, parameter
assumption that the distribution of the states λ is independent of the dependence is not a necessary condition for controllable probabilistic
settings of the measurement apparatuses). Superluminal signaling in the dependence. Some other physical quantities in the nearby wing may be
EPR/B experiment would be possible in theory just in case the value of relevant for the probability of the distant measurement outcome. (That is,
some controllable physical quantity in the nearby wing could influence let α and β denote all the relevant local physical quantities, other than the
the statistics of measurement outcomes in the distant wing. And in non- settings of the measurement apparatuses, that may be relevant for the
factorizable models that satisfy λ-independence this could happen just in probability of the L- and the R-outcome, respectively. Then, controllable
case the following conditions obtained: probabilistic dependence would obtain if for some pairs' states λ, L-
setting l, R-setting r and local physical quantities α and β, P λ l r α β(yr) ≠
Controllable probabilistic dependence. The probabilities of
P λ l r β(yr) obtained.) For the relevance of such models to the question of
distant measurement outcomes depend on some nearby
the in-principle possibility of superluminal signalling in some current
controllable physical quantity.
interpretations of quantum mechanics, see sections 7.3 and 7.4.
λ-distribution. There can be in theory an ensemble of particle (iii) The quantum-equilibrium distribution will not be the same in all
pairs the states of which deviate from the quantum-equilibrium models of the EPR/B experiment; for in general the states λ will not be
distribution; where the quantum-equilibrium distribution of pairs' the same in different models.
states is the distribution that reproduces the predictions of (iv) In models that actually violate both controllable probabilistic
orthodox quantum mechanics. dependence and λ-distribution, the occurrence of controllable probabilistic
dependence would render the actual distribution of λ states as non-
Four comments: (i) In controllable probabilistic dependence, the term equilbrium distribution. Thus, if controllable probabilistic dependence
'probabilities of measurement outcomes' refers to the model probabilities, occurred in such models, the actual distribution of λ states would satisfy
i.e., the probabilities that the states λ prescribe for measurement λ-distribution.
outcomes.
(ii) Our discussion in this entry focuses on models of the EPR/B The argument for the necessity of controllable probabilistic dependence
experiment in which probabilities of measurement outcomes depend only and λ-distribution is straightforward. Granted λ-independence, if the
on the pair's state λ and the settings of the measurement apparatuses to probabilistic dependence of the distant outcome on a nearby physical
measure certain properties. In such models, parameter dependence (i.e., quantity is not controllable, there can be no way to manipulate the
the dependence of the probability of the distant measurement outcome on statistics of the distant outcome so as to deviate from the statistical
the setting of the nearby measurement apparatus) is a necessary and predictions of quantum mechanics. Accordingly, superluminal
sufficient condition for controllable probabilistic dependence. But, recall transmission of information will be impossible even in theory. And if λ-
54 Stanford Encyclopedia of Philosophy Spring 2009 Edition 55
Action at a Distance in Quantum Mechanics Joseph Berkovitz

transmission of information will be impossible even in theory. And if λ- implies the failure of controllable probabilistic dependence, yet the
distribution does not hold, i.e., if the quantum-equilibrium distribution violation of λ-independence would imply the possibility of superluminal
holds, controllable probabilistic dependence will be of no use for signaling: If λ-independence failed, a change in the setting of the nearby
superluminal transmission of information. For, averaging over the model measurement apparatus would cause a change in the distribution of the
probabilities according to the quantum-equilbrium distribution, the model states λ, and a change in this distribution would induce a change in the
will reproduce the statistics of orthodox quantum mechanics. That is, the statistics of the distant (space-like separated) measurement outcome.
distribution of the λ-states will be such that the probabilistic dependence
of the distant outcome on the nearby controllable factor will be washed Leaving aside models that violate λ-independence, we now turn to
out: In some states the nearby controllable factor will raise the probability consider the prospects of controllable probabilistic dependence and λ-
of the distant outcome and in others it will decrease this probability, so distribution, starting with no-collapse interpretations.
that on average the overall statistics of the distant outcome will be
7.2 No-collapse theories
independent of the nearby controllable factor (i.e., the same as the
statistics of orthodox quantum mechanics). Accordingly, superluminal 7.2.1 Bohm's Theory
signaling will be impossible.
Bohm's theory involves parameter dependence and thus controllable
The argument for the sufficiency of these conditions is also
probabilistic dependence: The probabilities of distant outcomes depend on
straightforward. If λ-distribution held, it would be possible in theory to
the setting of the nearby apparatus. In some pairs' states λ, i.e., in some
arrange ensembles of particle pairs in which controllable probabilistic
configurations of the positions of the particle pair, a change in the
dependence would not be washed out, and accordingly the statistics of
apparatus setting of the (earlier) say L-measurement will induce an
distant outcomes would depend on the nearby controllable factor. (For a
immediate change in the probability of the R-outcome: e.g. the probability
proof that these conditions are sufficient for superluminal signaling in
of R-outcome z-spin ‘up’ will be 1 if the L-apparatus is set to measure z-
certain deterministic hidden variables theories, see Valentini 2002.)
spin and 0 if the L-apparatus is switched off (see section 5.3.1). Thus, the
Note that the necessary and sufficient conditions for superluminal question of superluminal signaling turns on whether λ-distribution
signaling are different in models that do not exclude in theory the obtains.
violation of λ-independence. In such models controllable probabilistic
Now, recall (section 5.3.1) that Bohm's theory reproduces the quantum
dependence is not a necessary condition for superluminal signaling. The
statistics by postulating the quantum-equilibrium distribution over the
reasoning is as follows. Consider any empirically adequate model of the
positions of particles. If this distribution is not an accidental fact about
EPR/B experiment in which the pair's state and the settings of the
our universe, but rather obtains as a matter of law, superluminal signaling
measurement apparatuses are the only relevant factors for the
will be impossible in principle. Dürr, Goldstein and Zanghì (1992a,b,
probabilities of measurement outcomes, and the quantum-equilibrium
1996, fn. 15) argue that, while the quantum-equilibrium distribution is not
distribution is λ-independent. In such a model, parameter independence
a matter a law, other distributions will be possible but atypical. Thus, they
implies the failure of controllable probabilistic dependence, yet the
conclude that although superluminal signaling is not impossible in theory,
56 Stanford Encyclopedia of Philosophy Spring 2009 Edition 57
Action at a Distance in Quantum Mechanics Joseph Berkovitz

conclude that although superluminal signaling is not impossible in theory,


it may occur only in atypical worlds. On the other hand, Valentini 7.3 Collapse theories
(1991a,b, 1992, 1996, 2002) and Valentini and Westman 2004) argue that
7.3.1 Dynamical models for state-vector reduction
there are good reasons to think that our universe may well have started off
in a state of quantum non-equilibrium and is now approaching gradually a In the GRW/Pearle collapse models, wave functions represent the most
state of equilibrium, so that even today some residual non-equilibrium exhaustive, complete specification of states of individual systems. Thus,
must be present.[32] Yet, even if such residual non-equilbrium existed, the pairs prepared with the same wave function have always the same λ state
question is whether it would be possible to access any ensemble of — a state that represents their quantum-equilbrium distribution for the
systems in a non-equilbrium distribution. EPR/B experiment. Accordingly, λ-distribution fails. Do these models
involve controllable probabilistic dependence?
7.2.2 Modal interpretations
Recall (section 5.1.2) that there are several models of state reduction in
The presence or absence of parameter independence (and accordingly the the literature. One of these models is the so-called non-linear Continuous
presence or absence of controllable probabilistic dependence) in the Stochastic Localization (CSL) models (see Pearle 1989, Ghirardi, Pearle
modal interprtation is a matter of controversy, perhaps due in part to the and Rimini 1990, Butterfield et al. 1993, and Ghirardi et al. 1993).
multiplicity of versions of this interpretation. Whether or not modal Butterfield et al. (1993) argue that in these models there is a probabilistic
interpretations involve parameter dependence would probably depend on dependence of the outcome of the R-measurement on the process that
the dynamics of the possessed properties. At least some of the current leads to the (earlier) outcome of the L-measurement. In these models, the
modal interpretations seem to involve no parameter dependence. But, as process leading to the L-outcome (either z-spin ‘up’ or z-spin ‘down’)
the subject editor pointed out to the author, some think that the no-go depends on the interaction between the L-particle and the L-apparatus
theorem for relativistic modal interpretation due to Dickson and Clifton (which results in an entangled state), and the specific realization of the
(1998) implies the existence of parameter dependence in all the stochastic process that strives to collapse this macroscopic superposition
interpretations to which this theorem is applicable. Do modal into a product state in which the L-apparatus displays a definite outcome.
interpretations satisfy λ-distribution? The prospects of this condition And the probability of the R-outcome depends on this process. For
depend on whether the possessed properties that the modal interpretation example, if this process is one that gives rise to a z-spin ‘up’ (or renders
assigns in addition to the properties prescribed by the orthodox that outcome more likely), the probability of R-outcome z-spin ‘up’ is 0
interpretation, are controllable. If these properties were controllable at (more likely to be 0); and if this process is one that gives rise to a z-spin
least in theory, λ-distribution would be possible. For example, if the ‘down’ (or renders that outcome more likely), the probability of R-
possessed spin properties that the particles have at the emission from the outcome z-spin ‘down’ is 0 (more likely to be 0). The question is whether
source in the EPR/B experiment were controllable, then λ-distribution there are controllable factors that influence the probability of realizations
would be possible. The common view seems to be that these properties of stochastic processes that lead to a specific L-outcome, so that it would
are uncontrollable. be possible to increase or decrease the probability of the R-outcome. If

58 Stanford Encyclopedia of Philosophy Spring 2009 Edition 59


Action at a Distance in Quantum Mechanics Joseph Berkovitz

be possible to increase or decrease the probability of the R-outcome. If 7.5 Superluminal signaling and action-at-a-distance
such factors existed, controllable probabilistic dependence would be
possible at least in theory. And if this kind of controllable probabilistic If superluminal signaling were possible in the EPR/B experiment in any
dependence existed, λ-distribution would also obtain; for if such of the above theories, it would not require any continuous process in
dependence existed, the actual distribution of pairs' states (in which the spacetime to mediate the influences between the two distant wings.
pair always have the same state, the quantum-mechanical state) would Indeed, in all the current quantum theories in which the probability of the
cease to be the quantum-equilbrium distribution. R-outcome depends on some controllable physical variable in the L-wing,
this dependence is not due to a continuous process. Rather, it is due to
7.4 The prospects of controllable probabilistic dependence some type of ‘action’ or (to use Shimony's (1984) terminology) ‘passion’
at a distance, which is the ‘result’ of the holistic nature of the quantum
In section 7.3.1, we discussed the question of the in-principle realm, the non-separability of the state of entangled systems, or the non-
controllability of local measurement processes and in particular the separable nature of the evolution of the properties of systems.
probability of their outcome, and the implications of such controllability
for the in-principle possibility of superluminal signaling in the context of 8. The analysis of factorizability: implications for
the CSL models. But this question is not specific to the CSL model and
quantum non-locality
(more generally) the dynamical models for state-vector reduction. It
seems likely to arise also in other quantum theories that model In sections 5-7, we considered the nature of quantum non-locality as
measurements realistically. Here is why. Real measurements take time. depicted by theories that violate factorizability, i.e., the assumption that
And during that time, some physical variable, other than the state of the the probability of joint measurement outcomes factorizes into the single
measured system and the setting of the measurement apparatus, might probabilities of these outcomes. Recalling section 3, factorizability can be
influence the chance (i.e., the single-case objective probability) of the analyzed into a conjunction of two conditions: OI (outcome
measurement outcome. In particular, during the L-measurement in the independence)—the probability of a distant measurement outcome in the
EPR/B experiment, the chance of the L-outcome z-spin ‘up’ (‘down’) EPR/B experiment is independent of the nearby measurement outcome;
might depend on the value of some physical variable in the L-wing, other and PI (parameter independence)—the probability of a distant
than the state of the particle pair and the setting of the L-measurement measurement outcome in the EPR/B experiment is independent of the
apparatus. If so, it will follow from the familiar perfect anti-correlation of setting of the nearby measurement apparatus. Bohm's theory violates PI,
the singlet state that the chance of R-outcome z-spin ‘up’ (‘down’) will whereas other mainstream quantum theories satisfy this condition but
depend on the value of such variable (for details, see Kronz 1990a,b, violate OI. The question arises as to whether violations of PI involve a
Jones and Clifton 1993, pp. 304-305, and Berkovitz 1998a, section 4.3.4). different kind of non-locality than violations of OI. So far, our
Thus, if the value of such a variable were controllable, controllable methodology was to study the nature of quantum non-locality by
probabilistic dependence would obtain. analyzing the way various quantum theories account for the curious
correlations in the EPR/B experiment. In this section, we shall focus on
the question of whether quantum non-locality can be studied in a more
60 Stanford Encyclopedia of Philosophy Spring 2009 Edition 61
Action at a Distance in Quantum Mechanics Joseph Berkovitz

the question of whether quantum non-locality can be studied in a more


Fourth, Howard's spatiotemporal separability condition (see section 4.3)
general way, namely by analyzing the types of non-locality involved in
requires that states of composite systems be determined by the states of
violations of PI and in violations of OI, independently of how these
their subsystems. In particular, spatiotemporal separability requires that
violations are realized.
joint probabilities of outcomes be determined as some function of the
8.1 Non-separability, holism and action at a distance single probabilities of these outcomes. Winsberg and Fine (2003) object
that as a separability condition, OI arbitrarily restricts this function to be a
It is frequently argued or maintained that violations of OI involve state product function. And they argue that on a weakened formalization of
non-separability and/or some type of holism, whereas violations of PI separability, a violation of OI is compatible with separability. Fogel
involve action at a distance. For notable examples, Howard (1989) argues (2004) agrees that Winsberg and Fine's weakened formalization of
that spatiotemporal separability (see section 4.3) implies OI, and separability is correct, but argues that, when supplemented by a certain
accordingly a violation of it implies spatiotemporal non-separability; ‘isotropy’ condition, OI implies this weakened separability condition.
Teller (1989) argues that particularism (see section 4.3) implies OI, and Fogel believes that his suggested ‘isotropy’ condition is very plausible,
thus a violation of it implies relational holism; and Jarrett (1984, 1989) but, as he acknowledges, this condition involves a nontrivial measurement
argues that a violation of PI involves some type of action at a distance. context-independence.[33]
These views are controversial, however.
Fifth, as the analysis in section 5 demonstrates, violations of OI might
First, as we have seen in section 5, in quantum theories the violation of involve action at a distance. Also, while the minimal Bohm theory
either of these conditions involves some type of non-separability and/or violates PI and arguably some modal interpretations do not, the type of
holism. action at a distance they postulate, namely action* at a distance (see
section 5.2), is similar: In both cases, an earlier spin-measurement in
Second, the explicit attempts to derive OI from separability or (say) the L-wing does not induce any immediate change in the intrinsic
particularism seem to rely (implicitly) on some locality conditions. properties of the R-particle. The L-measurement only causes an
Maudlin (1998, p. 98) and Berkovitz (1998a, section 6.1) argue that immediate change in the dispositions of the R-particle—a change that
Howard's precise formulation of spatiotemporal separability embodies may influence the behavior of the R-particle in future spin-measurements
both separability and locality conditions, and Berkovitz (1998a, section in the R-wing. But, this change of dispositions does not involve any
6.2) argues that Teller's derivation of OI from particularism implicitly change of local properties in the R-wing, as these dispositions are
relies on locality conditions. Thus, the violation of OI per se does not relational (rather than intrinsic) properties of the R-particle. Furthermore,
imply non-separability or holism. the action at a distance predicated by the minimal Bohm theory is weaker
than the one predicated by orthodox collapse quantum mechanics and the
Third, a factorizable model, i.e., model that satisfies OI, may be non-
GRW/Pearle collapse models; for in contrast to the minimal Bohm
separable (Berkovitz 1995b, section 6.5). Thus, OI cannot be simply
theory, in these theories the measurement on the L-particle induces a
identified with OI.
change in the intrinsic properties of the R-particle, independently of

62 Stanford Encyclopedia of Philosophy Spring 2009 Edition 63


Action at a Distance in Quantum Mechanics Joseph Berkovitz

change in the intrinsic properties of the R-particle, independently of Jarrett (1984, 1989), Ballentine and Jarrett (1997) and Shimony (1984)
whether or not the R-particle undergoes a measurement. Thus, if the R- hold that superluminal signaling is incompatible with relativity theory.
particle comes to possess (momentarily) a definite position, the EPR/B Accordingly, they conclude that violations of PI are incompatible with
experiment as described by these theories involves action at a distance — relativity theory, whereas violations of OI may be compatible with this
a stronger kind of action than the action* at a distance predicated by the theory. Furthermore, Sutherland (1985, 1989) argues that deterministic,
minimal Bohm theory. relativistic parameter-dependent theories (i.e., relativistic, deterministic
theories that violate PI) would plausibly require retro-causal influences,
8.2 Superluminal signaling and in certain experimental circumstances this type of influences would
give rise to causal paradoxes, i.e., inconsistent closed causal loops (where
It was also argued, notably by Jarrett 1984 and 1989 and Shimony 1984,
effects undermine their very causes). And Arntzenius (1994) argues that
that in contrast to violations of OI, violations of PI may give rise (at least
all relativistic parameter-dependent theories are impossible on pain of
in principle) to superluminal signaling. Indeed, as is not difficult to see
causal paradoxes. That is, he argues that in certain experimental
from section 7.1, in theories that satisfy λ-independence there is an
circumstances any relativistic, parameter-dependent theory would give
asymmetry between failures of PI and failures of OI with respect to
rise to closed causal loops in which violations of PI could not obtain.
superluminal signaling: whereas λ-distribution and the failure of PI are
sufficient conditions for the in-principle possibility of superluminal It is noteworthy that the view that relativity per se is incompatible with
signaling, λ-distribution and the failure of OI are not. Thus, the prospects superluminal signaling is disputable (for more details, see section 10).
of superluminal signaling look better in parameter-dependent theories, Anyway, recalling (section 8.2), if λ-distribution is excluded as a matter
i.e., theories that violate PI. Yet, as we have seen in section 7.2.1, if the of law, it will be impossible even in theory to exploit the violation of PI
Bohmian quantum-equilbrium distribution obtains, then Bohm's theory, to give rise to superluminal signaling, in which case the possibility of
the paradigm of parameter dependent theories, prohibits superluminal relativistic parameter-dependent theories could not be discounted on the
signaling. And if this distribution is obtained as a matter of law, then basis of superluminal signaling.
Bohm's theory prohibits superluminal signaling even in theory.
Furthermore, as we remarked in section 7.1, if the in-principle possibility Furthermore, as mentioned in section 7.1 and 7.4, the in-principle
of violating λ-independence is not excluded, superluminal signaling may possibility of superluminal signaling in theories that satisfy PI and violate
exist in theories that satisfy PI and violate OI. In fact, as section 7.4 seems OI cannot be excluded a priori. Thus, if relativity theory excludes
to suggest, the possibility of superluminal signaling in theories that satisfy superluminal signaling, the argument from superluminal signaling may
PI but violate OI cannot be discounted even when λ-independence is also be applied to exclude the possibility of some relativistic outcome-
impossible. dependent theories.

8.3 Relativity Finally, Berkovitz (2002) argues that Arntzenius's argument for the
impossibility of relativistic theories that violate PI is based on
assumptions about probabilities that are common in linear causal

64 Stanford Encyclopedia of Philosophy Spring 2009 Edition 65


Action at a Distance in Quantum Mechanics Joseph Berkovitz

assumptions about probabilities that are common in linear causal Parameter dependence (PI) postulates that in the EPR/B experiment the
situations but are unwarranted in causal loops, and that the real challenge probability of the later, distant measurement outcome depends on the
for these theories is that in such loops their predictive power is setting of the apparatus of the nearby, earlier measurement. It may be
undermined (for more details, see section 10.3). tempting to assume that this dependence is due to a direct influence of the
nearby setting on the (probability of the) distant outcome. But a little
8.4 Superluminal causation
reflection on the failure of PI in Bohm's theory, which is the paradigm for
In various counterfactual and probabilistic accounts of causation parameter dependence, demonstrates that the setting of the nearby
violations of PI entail superluminal causation between the setting of the apparatus per se has no influence on the distant measurement outcome.
nearby measurement apparatus and the distant measurement outcome, Rather, it is because the setting of the nearby measurmenent apparatus
whereas violations of OI entail superluminal causation between the influences the nearby measurement outcome and the nearby outcome
distant measurement outcomes (see Butterfield 1992b, 1994, Berkovitz influences the distant outcome that the setting of the nearby apparatus can
1998b, section 2). Thus, it seems that theories that violate PI postulate a have an influence on the distant outcome. For, as is not difficult to see
different type of superluminal causation than theories that violate OI. Yet, from the analysis of the nature of non-locality in the minimal Bohm
as Berkovitz (1998b, section 2.4) argues, the violation of PI in Bohm's theory (see section 5.3.1), the setting of the apparatus of the nearby
theory does involve some type of outcome dependence, which may be (earlier) measurement in the EPR/B experiment influences the outcome
interpreted as a generalization of the violation of OI. In this theory, the the nearby measurement, and this outcome influences the guiding field of
specific R-measurement outcome in the EPR/B experiment depends on the distant particle and accordingly the outcome of a measurement on that
the specific L-measurement outcome: For any three different directions x, particle.
y, z, if the probabilities of x-spin ‘up’ and y-spin ‘up’ are non-zero, the
While the influence of the nearby setting on the nearby outcome is
probability of R-outcome z-spin ‘up’ will generally depend on whether
necessary for parameter dependence, it is not sufficient for it. In all the
the L-outcome is x-spin ‘up’ or y-spin ‘up’. Yet, due to the determinism
current quantum theories, the probabilities of joint outcomes in the EPR/B
that Bohm's theory postulates, OI trivially obtains. Put it another way, OI
experiment depend on the settings of both measurement apparatuses: The
does not reflect all the types of outcome independence that may exist
probability that the L-outcome is l-spin ‘up’ and the R-outcome is r-spin
between distant outcomes. Accordingly, the fact that a theory satisfies OI
‘up’ and the probability that the L-outcome is l-spin ‘up’ and the R-
does not entail that it does not involve some other type of outcome
outcome is r-spin ‘down’ both depend on (l − r), i.e., the distance
dependence. Indeed, in all the current quantum theories that violate
between the angles l and r. In theories in which the sum of these joint
factorizability there are correlations between distant specific measurement
probabilities is invariant with respect to the value of (l − r), parameter
outcomes — correlations that may well be interpreted as an indication of
independence obtains: for all pairs' states λ, L-setting l, and R-settings r
counterfactual superluminal causation between these outcomes.
and r′, L-outcome xl, and R-outcomes yr and yr′  :
8.5 On the origin and nature of parameter dependence

66 Stanford Encyclopedia of Philosophy Spring 2009 Edition 67


Action at a Distance in Quantum Mechanics Joseph Berkovitz

distance or superluminal causation between the distant wings of the


(PI)
EPR/B experiment. Yet, as quantum-mechanical states of systems are
P λ l r(xl & yr) + P λ l r(xl & ¬yr) = P λ l r′  (xl & yr′  ) + P λ l r′  (xl & ¬yr′  ).
assumed to reflect their physical states, the many-minds theory does
Parameter dependence is a violation of this invariance condition. postulate some type of non-locality, namely state non-separability and
property and relational holism.
9. Can there be ‘local’ quantum theories? Another way to get around Bell's argument for non-locality in the EPR/B
experiment is to construct a model of this experiment that satisfies
The focus of this entry has been on exploring the nature of the non-local
factorizability but violates λ-independence (i.e., the assumption that the
influences in the quantum realm as depicted by quantum theories that
distribution of all the possible pairs' states in the EPR/B experiment is
violate factorizability, i.e., theories in which the joint probability of the
independent of the measured quantities). In section 2, we mentioned two
distant outcomes in the EPR/B experiment do not factorize into the
possible causal explanations for the failure of λ-independence. The first is
product of the single probabilities of these outcomes. The motivation for
to postulate that pairs' states and apparatus settings share a common
this focus was that, granted plausible assumptions, factorizability must
cause, which correlates certain types of pairs' states with certain types of
fail (see section 2), and its failure implies some type of non-locality (see
settings (e.g. states of type λ1 are correlated with settings of type l and r,
sections 2-8). But if any of these plausible assumptions failed, it may be
whereas states of type λ2 are correlated with settings of type l′ and r′,
possible to account for the EPR/B experiment (and more generally for all
etc.). As we noted, thinking about all the various ways one can measure
other quantum phenomena) without postulating any non-local influences.
properties, this explanation seems conspiratorial. Furthermore, it runs
Let us then consider the main arguments for the view that quantum
counter to one of the most fundamental presuppositions of empirical
phenomena need not involve non-locality.
science, namely that in experiments preparations of sources and settings
In arguments for the failure of factorizability, it is presupposed that the of measurement apparatuses are typically independent of each other. The
distant measurement outcomes in the EPR/B experiment are real physical second possible explanation is to postulate causation from the
events. Recall (section 5.3.3) that in Albert and Loewer's (1988) many- measurement events backward to the source at the emission time. (For
minds interpretation this is not the case. In this interpretation, definite advocates of this way out of non-locality, see Costa de Beauregard 1977,
measurement outcomes are (typically) not physical events. In particular, 1979, 1985, Sutherland 1983, 1998, 2006 and Price 1984, 1994, 1996,
the pointers of the measurement apparatuses in the EPR/B experiment do chapters 3, 8 and 9.) Maudlin (1994, p. 197-201) argues that theories that
not display any definite outcomes. Measurement outcomes in the EPR/B postulate such causal mechanism are inconsistent. Berkovitz (2002,
experiment exist only as (non-physical) mental states in observers' minds section 5) argues that Maudlin's line of reasoning is based on unwarranted
(which are postulated to be non-physical entities). So sacrificing some of premises. Yet, as we shall see in section 10.3, this way out of non-locality
our most fundamental presuppositions about the physical reality and faces some challenges. Furthermore, while a violation of λ-independence
assuming a controversial mind-body dualism, the many-minds provides a way out of Bell's theorem, it does not necessarily imply
interpretation of quantum mechanics does not postulate any action at a locality; for the violation of λ-independence is compatible with the failure
distance or superluminal causation between the distant wings of the of factorizability.
68 Stanford Encyclopedia of Philosophy Spring 2009 Edition 69
Action at a Distance in Quantum Mechanics Joseph Berkovitz

of factorizability. models of the EPR/B experiment (i.e., viable models which do not
postulate any non-locality) are possible. Indeed, so far none of the
A third way around non-locality is to ‘exploit’ the inefficiency of attempts to construct local, non-factorisable models for EPR/B
measurement devices or (more generally) measurement set-ups. In any experiments has been successful.
actual EPR/B experiment, many of the particle pairs emitted from the
source fail to be detected, so that only a sample of the particle pairs is 10. Can quantum non-locality be reconciled with
observed. Assuming that the observed samples are not biased, it is now
relativity?
generally agreed that the statistical predictions of orthodox quantum
mechanics have been vindicated (for a review of these experiments, see The question of the compatibility of quantum mechanics with the special
Redhead 1987, section 4.5). But if this assumption is abandoned, there are theory of relativity is very difficult to resolve. (The question of the
perfectly local causal explanations for the actual experimental results compatibility of quantum mechanics with the general theory of relativity
(Clauser and Horne 1974, Fine 1982b, 1989a). Many believe that this way is even more involved.) The answer to this question depends on the
out of non-locality is ad hoc, at least in light of our current knowledge. interpretation of special relativity and the nature of the exact constraints it
Moreover, this strategy would fail if the efficiency of measurement imposes on influences between events.
devices exceeded a certain threshold (for more details, see Fine 1989a,
Maudlin 1994, chapter 6, Larsson and Semitecolos 2000 and Larsson A popular view has it that special relativity prohibits any superluminal
2002). influences, whereas theories that violate factorizability seem to involve
such influences. Accordingly, it is held that quantum mechanics is
Finally, there are those who question the assumption that factorizability is incompatible with relativity. Another common view has it that special
a locality condition (Fine 1981, 1986, pp. 59-60, 1989b, Cartwright 1989, relativity prohibits only certain types of superluminal influence. Many
chaps. 3 and 6, Chang and Cartwright 1993). Accordingly, they deny that believe that relativity prohibits superluminal signaling of information.
non-factorizability implies non-locality. The main thrust of this line of Some also believe that this theory prohibits superluminal transport of
reasoning is that the principle of the common cause is not generally valid. matter-energy and/or action-at-a-distance. On the other hand, there is the
Some, notably Cartwright (1989) and Chang and Cartwright (1993), view that relativity per se prohibits only superluminal influences that are
challenge the assumption that common causes always screen off the incompatible with the special-relativistic space-time, the so-called
correlation between their effects, and accordingly they question the idea ‘Minkowski space-time,’ and that this prohibition is compatible with
that non-factorizability implies non-locality. Others, notably Fine, deny certain types of superluminal influences and superluminal signaling (for a
that correlations must have causal explanation. comprehensive discussion of this issue, see Maudlin, 1994, 1996, section
2).[34]
While these arguments challenge the view that the quantum realm as
depicted by non-factorizable models for the EPR/B experiment must It is commonly agreed that relativity requires that the descriptions of
involve non-locality, they do not show that viable local, non-factorizable physical reality (i.e., the states of systems, their properties, dynamical
models of the EPR/B experiment (i.e., viable models which do not laws, etc.) in different coordinate systems should be compatible with each

70 Stanford Encyclopedia of Philosophy Spring 2009 Edition 71


Action at a Distance in Quantum Mechanics Joseph Berkovitz

laws, etc.) in different coordinate systems should be compatible with each Assuming (for the sake of simplicity) that measurement events occur
other. In particular, descriptions of the state of systems in different instantaneously, state collapse occurs along a single spacelike hyperplane
foliations of spacetime into parallel spacelike hyperplanes, which that intersects the spacetime region of the L-measurement event—the
correspond to different inertial reference frames, are to be related to each hyperplane that represents the (absolute) time of the collapse. But this
other by the Lorentz transformations. If this requirement is to reflect the type of collapse dynamics would involve a preferred foliation of
structure of the Minkowski spacetime, these transformations must hold at spacetime, in violation of the spirit, if not the letter of the Minkowski
the level of individual processes, and not only at the level of ensembles of spacetime.
processes (i.e., at the statistical level) or observed phenomena. Indeed,
Bohm's theory, which is manifestly non-relativistic, satisfies the The current dynamical collapse models are not genuinely relativistic, and
requirement that the Lorentz transformations obtain at the level of the attempts to generalize them to the special relativistic domain have
observed phenomena. encountered difficulties (see, for example, the entry on collapse theories,
Ghirardi 1996, Pearle 1996, and references therein). A more recent
However, satisfying the Lorentz transformations at the level of individual attempt to address these difficulties due to Tumulka (2004) seems more
processes is not sufficient for compatibility with Minkowski spacetime; promising.
for the Lorentz transformations may also be satisfied at the level of
individual processes in theories that postulate a preferred inertial In an attempt to reconcile state collapse with special relativity, Fleming
reference frame (Bell 1976). Maudlin (1996, section 2) suggests that a (1989, 1992, 1996) and Fleming and Bennett (1989) suggested radical
theory is genuinely relativistic (both in spirit and letter) if it can be hyperplane dependence. In their theory, state collapse occurs along an
formulated without ascribing to spacetime any more, or different intrinsic infinite number of spacelike hyperplanes that intersect the spacetime
structure than the relativistic metrics.[35] The question of the region of the measurements. That is, in the EPR/B experiment a collapse
compatibility of relativity with quantum mechanics may be presented as occurs along all the hyperplanes of simultaneity that intersect the
follows: Could a quantum theory that does not encounter the measurement spacetime region of the L-measurement. Similarly, a collapse occurs
problem be relativistic in that sense? along all the hyperplanes of simultaneity that intersect the distant (space-
like separated) spacetime region of the R-measurement. Accordingly, the
10.1 Collapse theories hyperplane-dependent theory does not pick out any reference frame as
preferred, and the dynamics of the quantum states of systems and their
The main problem in reconciling collapse theories with special relativity properties can be reconciled with the Minkowski spacetime. Further, since
is that it seems very difficult to make state collapse (modeled as a real all the multiple collapses are supposed to be real (Fleming 1992, p. 109),
physical process) compatible with the structure of the Minkowski the predictions of orthodox quantum mechanics are reproduced in each
spacetime. In non-relativistic quantum mechanics, the earlier L- reference frame.
measurement in the EPR/B experiment induces a collapse of the
entangled state of the particle pair and the L-measurement apparatus. The hyperplane-dependent theory is genuinely relativistic. But the theory
Assuming (for the sake of simplicity) that measurement events occur does not offer any mechanism for state collapses, and it does not explain

72 Stanford Encyclopedia of Philosophy Spring 2009 Edition 73


Action at a Distance in Quantum Mechanics Joseph Berkovitz

does not offer any mechanism for state collapses, and it does not explain
10.2 No-collapse theories
how the multiple collapses are related to each other and how our
experience is accounted for in light of this multiplicity. Recall (section 5.3) that in no-collapse theories, quantum-mechanical
states always evolve according to a unitary and linear equation of motion
Myrvold (2002b) argues that state collapses can be reconciled with
(the Schrödinger equation in the non-relativistic case), and accordingly
Minkowski spacetime even without postulating multiple different
they never collapse. Since the wave function has a covariant dynamics,
collapses corresponding to different reference frames. That is, he argues
the question of the compatibility with relativity turns on the dynamics of
with respect to the EPR/B experiment that the collapses induced by the L-
the additional properties —the so-called ‘hidden variables’— that no-
and the R-measurement are local events in the L- and the R-wing
collapse theories typically postulate. In Albert and Loewer's many-minds
respectively, and that the supposedly different collapses (corresponding to
theory (see section 5.3.3), the wave function has covariant dynamics, and
different reference frames) postulated by the hyperplane-dependent theory
no additional physical properties are postulated. Accordingly, the theory
are only different descriptions of the same local collapse events. Focusing
is genuinely relativistic. Yet, as the compatibility with relativity is
on the state of the particle pair, the main idea is that the collapse event in
achieved at the cost of postulating that outcomes of measurements (and,
the L-wing is modeled by a (one parameter) family of operators (the
typically, any other perceived properties) are mental rather than physical
identity operator before the L-measurement and a projection to the
properties, many find this way of reconciling quantum mechanics with
collapsed state after the L-measurement), and it is local in the sense that it
relativity unsatisfactory.
is a projection on the Hilbert space of the L-particle; and similarly,
mutatis mutandis, for the R-particle. Yet, if the quantum state of the Other Everett-like interpretations attempt to reconcile quantum mechanics
particle pair represents their complete state (as the case is in the orthodox with the special theory of relativity without postulating such a
theory and the GRW/Pearle collapse models), these collapse events seem controversial mind-body dualism. Similarly to the many-minds
non-local. While the collapse in the L-wing may be said to be local in the interpretation of Albert and Loewer, and contrary to Bohm's theory and
above technical sense, it is by definition a change of local as well as modal interpretations, on the face of it these interpretations do not
distant (spacelike) properties. The operator that models the collapse in the postulate the existence of ‘hidden variables.’ But (recalling section 5.3.3)
L-wing transforms the entangled state of the particle pair—a state in these Everett-like interpretations face the challenge of making sense of
which the particles have no definite spins—into a product of non- our experience and the probabilities of outcomes, and critics of these
entangled states in which both particles have definite spins, and interpretations argue that this challenge cannot be met without adding
accordingly it causes a change of intrinsic properties in both the L- and some extra structure to the Everett interpretation (see Albert and Loewer
the R-wing. 1988, Albert 1992, pp. 114-5, Albert and Loewer 1996, Price 1996, pp.
226-227, and Barrett 1999, pp. 163-173); a structure that may render these
In any case, Myrvold's proposal demonstrates that even if state collapses
interpretations incompatible with relativity. Supporters of the Everett
are not hyperplane dependent, they need not be incompatible with
interpretation disagree. Recently, Deutsch (1999), Wallace (2002, 2003,
relativity theory.
2005a,b) and Greaves (2004) have suggested that Everettians can make
sense of the quantum-mechanical probabilities by appealing to decision-
74 Stanford Encyclopedia of Philosophy Spring 2009 Edition 75
Action at a Distance in Quantum Mechanics Joseph Berkovitz

sense of the quantum-mechanical probabilities by appealing to decision- be the same in EPR/B experiments in which the measurements are not
theoretical considerations. But this line of reasoning has been disputed distant from each other (i.e., when the measurements are timelike
(see Barnum et al. 2000, Lewis 2003b, Hemmo and Pitowsky 2005 and separated). If so, relativistic parameter-dependent theories would involve
Price 2006). backward causal influences. But, he argues, in deterministic, relativistic
parameter-dependent theories these influences would give rise to causal
Modal interpretations constitute another class of no-collapse paradoxes, i.e., inconsistent closed causal loops.
interpretations of quantum mechanics that were developed to reconcile
quantum mechanics with relativity (and to solve the measurement Furthermore, Arntzenius (1994) argues that all relativistic parameter-
problem). Yet, as the no-go theorems by Dickson and Clifton (1998), dependent theories are impossible on pain of causal paradoxes. In his
Arntzenius (1998) and Myrvold (2002) demonstrate, the earlier versions argument, he considers the probabilities of measurement outcomes in a
of the modal interpretation are not genuinely compatible with relativity setup in which two EPR/B experiments are causally connected to each
theory. Further, Earman and Ruetsche (2005) argue that a quantum-field other, so that the L-measurement outcome of the first EPR/B experiment
version of the modal interpretation (which is set in the context of determines the setting of the L-apparatus of the second EPR/B experiment
relativistic quantum-field theory), like the one proposed by Clifton and the R-measurement outcome of the second EPR/B experiment
(2000), would be subject to serious challenges. Berkovitz and Hemmo determines the setting of the R-apparatus of the first EPR/B experiment.
(2006a,b) develop a relational modal interpretation that escapes all the And he argues that in this experiment, relativistic parameter-dependent
above no-go theorems and to that extent seems to provide better prospects theories (deterministic or indeterministic) would give rise to closed causal
for reconciling quantum mechanics with special relativity. loops in which parameter dependence would be impossible. Thus, he
concludes that relativistic, parameter-dependent theories are impossible.
10.3 Quantum causal loops and relativity (Stairs (1989) anticipates the argument that the above experimental setup
may give rise to causal paradoxes in relativistic, parameter-dependent
Recall (section 8) that many believe that parameter-dependent theories
theories, but he stops short of arguing that such theories are impossible.)
(i.e., theories that violate parameter independence) are more difficult or
even impossible to reconcile with relativity. Recall also that one of the Berkovitz (1998b, section 3.2, 2002, section 4) argues that Arntzenius's
lines of argument for the impossibility of relativistic parameter-dependent line of reasoning fails because it is based on untenable assumptions about
theories is that such theories would give rise to causal paradoxes. In our the nature of probabilities in closed causal loops—assumptions that are
discussion, we focused on EPR/B experiments in which the measurements very natural in linear causal situations (where effects do not cause their
are distant (spacelike separated). In a relativistic parameter-dependent causes), but untenable in causal loops. (For an analysis of the nature of
theory, the setting of the nearby measurement apparatus in the EPR/B probabilities in causal loops, see Berkovitz 2001 and 2002, section 2.)
experiment would influence the probability of the distant (spacelike Thus, he concludes that the consistency of relativistic parameter-
separated) measurement outcome. Sutherland (1985, 1989) argues that it dependent theories cannot be excluded on the grounds of causal
is plausible to suppose that the realization of parameter dependence would paradoxes. He also argues that the real challenge for relativistic
be the same in EPR/B experiments in which the measurements are not parameter-dependent theories is concerned with their predictive power. In

76 Stanford Encyclopedia of Philosophy Spring 2009 Edition 77


Action at a Distance in Quantum Mechanics Joseph Berkovitz

parameter-dependent theories is concerned with their predictive power. In that if such retro-causal theories were true, they would involve closed
the causal loops predicted by relativistic parameter-dependent theories in causal loops in which the probabilities of outcomes that these theories
Arntzenius's suggested experiment, there is no known way to compute the assign will certainly deviate from the statistics of these outcomes. And,
frequency of events from the probabilities that the theories prescribe. similarly to Arntzenius's argument, Maudlin's argument also rests on
Accordingly, such theories would fail to predict any definite statistics of untenable assumptions about the nature of probabilities in causal loops
measurement outcomes for that experiment. This lack of predictability (for a further discussion of Maudlin's and Berkovitz's arguments and,
may also present some new opportunities. Due to this unpredictability, more generally, the prospects of Cramer's theory, see Kastner 2004).
there may be an empirical way for arbitrating between these theories and Furthermore, Berkovitz (2002, sections 2 and 5.4) argues that, similarly to
quantum theories that do not predicate the existence such causal loops in relativistic parameter-dependent theories, the main challenge for theories
Arntzenius's experiment. that postulate retro-causality is not causal paradoxes, but rather the fact
that their predictive power may be undermined. That is, the probabilities
Another attempt to demonstrate the impossibility of certain relativistic assigned by such theories may fail to predict the frequency of events in
quantum theories on the grounds of causal paradoxes is advanced by the loops they predicate. In particular, the local retro-causal theories that
Maudlin (1994, pp. 195-201). (Maudlin does not present his argument in Maudlin considers fail to assign any definite predictions for the frequency
these terms, but the argument is in effect based on such grounds.) Recall of measurement outcomes in certain experiments. Yet, some other
(sections 2 and 9) that a way to try to reconcile quantum mechanics with theories that predicate the existence of causal loops, such as Sutherland's
relativity is to account for the curious correlations between distant (2006) local time-symmetric Bohmian interpretation of quantum
systems by local backward influences rather than non-local influences. In mechanics, seem not to suffer from this problem.
particular, one may postulate that the correlations between the distant
measurement outcomes in the EPR/B experiment are due to local Bibliography
influences from the measurement events backward to the state of the
particle pair at the source. In such models of the EPR/B experiment, Albert, D. (1992), Quantum Mechanics and Experience, Cambridge
influences on events are always confined to events that occur in their past Mass.: Harvard University Press.
or future light cones, and no non-locality is postulated. Maudlin argues Albert, D. and Loewer, B. (1988), “Interpreting the many-worlds
that theories that postulate such backward causation will be inconsistent. interpretation”, Synthese 77, 195-213.
More particularly, he argues that a plausible reading of Cramer's (1980, ––– (1995), “Tails of Schrödinger's cat”, in R. Clifton (1996b), pp.
1986) transactional interpretation, and any other theory that similarly 81-92.
attempts to account for the EPR/B correlations by postulating causation Alexander, H. G. (1956) (ed), The Leibniz-Clarke Correspondence,
from the measurement events backward to the source, will be Manchester: Manchester University Press (new reprint, 1998).
inconsistent. Arntzenius, F. (1992), “Apparatus independence in proofs of non-
locality”, Foundations of Physics Letters 5, 517-525.
Berkovitz (2002, section 5) argues that Maudlin's argument is, in effect,
––– (1994), “Spacelike connections”, British Journal for the
that if such retro-causal theories were true, they would involve closed
Philosophy of Science 45, 201-17.
78 Stanford Encyclopedia of Philosophy Spring 2009 Edition 79
Action at a Distance in Quantum Mechanics Joseph Berkovitz

Philosophy of Science 45, 201-17. British Journal for the Philosophy of Science 52, 125-130.
––– (1998), “Curiouser and curiouser: problems for modal Bell, J. (1964), “On the Einstein-Podolsky-Rosen paradox”, Physics
interpretations of quantum mechanics”, in Dieks and Vermaas 1998, 1, 195-200. Reprinted in Bell (1987), pp. 14-21.
pp. 337-377. ––– (1966), “On the problem of hidden variables in quantum
Bacciagaluppi, G. (1995), “A Kochen-Specker theorem in the modal mechanics”, Reviews of Modern Physics 38, 447-452. Reprinted in
interpretation of quantum mechanics”, International Journal of Bell (1987), pp. 1-13.
Theoretical Physics 34, 1205-1216. ––– (1971), “Introduction to the hidden-variable question”, in B.
––– (1996), “Topics in the modal interpretation of quantum Espagnat (ed), Foundations of Quantum Mechanics. Proceedings of
mechanics”, Ph.D. Thesis, University of Cambridge. the International School of Physics ‘Enrico Fermi’, Academic Press,
––– (2000), “Delocalized properties in the modal interpretation of a New York-London, pp. 171-181. Reprinted in Bell (1987), pp. 29-
continuous model of decoherence”, Foundations of Physics 30, 39.
1431-1444. ––– (1975a), “The theory of local beables”, TH-2053-Cern.
––– (2002), “Remarks on space-time and Locality in Everett's Reprinted in Epistemological Letters March 1976 and Bell (1987),
interpretation”, in Placek and Butterfield (2002), pp. 105-122. pp. 52-62.
Bacciagaluppi, G. and Hemmo, M. (1996), “Modal interpretations, ––– (1975b), “Locality in quantum mechanics: reply to critics”,
decoherence and measurements”, Studies in History and Philosophy Epistemological Letters, Nov. 1975, 2-6. Reprinted in Bell (1987),
of Modern Physics 27, 239-277. pp. 63-66.
Bacciagaluppi, G. and Dickson, M. (1999), “Dynamics for modal ––– (1976), “How to teach special relativity”, Progress in Scientific
interpretations”, Foundations of Physics 29, 1165-1201. Culture, Vol. 1, No 2. Reprinted in Bell 1987b, pp. 67-80.
Ballentine, L. E. and Jarrett, J. (1997), “Bell's theorem: does ––– (1982), “On the impossible pilot wave”, Foundations of Physics
quantum mechanics contradict relativity?”, American Journal of 12, 989-999. Reprinted in Bell 1987, pp. 159-168.
Physics 55, 785-792. ––– (1987a), “Are there quantum jumps?”, In C. W. Kilmister (ed.)
Barnum, H., Caves, C. M., Finkelstein J., Fuchs, C. A. and Schack Schrödinger: Centenary of a Polymath, Cambridge: Cambridge
R. (2000), “Quantum probability from decision theory?”, University Press, pp. 41-52. Reprinted in Bell (1987), pp. 201-212.
Proceedings of the Royal Society of London A 456, 1175-1182. ––– (1987b), Speakable and Unspeakable in Quantum Mechanics,
Barrett, J. (1999), The Quantum Mechanics of Minds and Worlds, Cambridge: Cambridge University Press.
Oxford: Oxford University Press. Bene, G. and Dieks, D. (2002), “A perspectival version of the modal
Bassi, A. and Ghirardi, G. C. (1999), “More about dynamical interpretation of quantum mechanics and the origin of macroscopic
reduction and the enumeration principle”, British Journal for the behavior”, Foundations of Physics 32, 645-671.
Philosophy of Science 50, 719-734. Berkovitz, J. (1995a), “What econometrics cannot teach quantum
––– (2001), “Counting marbles: reply to Clifton and Monton”, mechanics?”, Studies in History and Philosophy of Modern Physics
British Journal for the Philosophy of Science 52, 125-130. 26, 163-200.

80 Stanford Encyclopedia of Philosophy Spring 2009 Edition 81


Action at a Distance in Quantum Mechanics Joseph Berkovitz

26, 163-200. Bohm, D., and Hiley, B. J. (1993), The Undivided Universe: An
––– (1995b), “Quantum Nonlocality: an analysis of the implications Ontological Interpretation of Quantum Theory, London: Routledge
of Bell's theorem and quantum correlations for nonlocality”, Ph.D. & Kegan Paul.
Thesis, University of Cambridge. Bohm, D., Schiller, R. and Tiomno, J. (1955), “A causal
––– (1998a), “Aspects of quantum non-locality I: superluminal interpretation of the Pauli equation”, Supplemento al Nuovo Cimento
signalling, action-at-a-distance, non-separability and holism”,Studies 1, 48-66.
in History and Philosophy of Modern Physics 29, 183-222. Bub, J. (1992), “Quantum mechanics without the projection
––– (1998b), “Aspects of quantum non-Locality II: superluminal postulate”, Foundations of Physics 22, 737-754.
causation and relativity”, Studies in History and Philosophy of ––– (1994), “How to interpret quantum mechanics”, Erkenntnis 41,
Modern Physics 29, 509-545. 253-273.
––– (2001), “On chance in causal loops”, Mind 110, 1-23. ––– (1997), Interpreting the Quantum World, Cambridge: Cambridge
––– (2002), “On causal loops in the quantum realm”, in T. Placek University Press.
and J. Butterfield (2002), pp. 235-257. Bub, J. and Clifton, R (1996), “A uniqueness theorem for ‘no
Berkovitz, J. and Hemmo, M. (2006a), “How to reconcile modal collapse’ interpretations of quantum mechanics”, Studies in History
interpretations of quantum mechanics with relativity”, forthcoming and Philosophy of Modern Physics 27, 181-217.
in Philosophy of Science. [Earlier version available online.] Butterfield, J. N. (1989), “A space-time approach to the Bell
––– (2006b), “A modal interpretation in terms of relational inequality”, in Cushing and McMullin (1989), pp. 114-144.
properties”, in W. Demopoulos and I. Pitowsky (eds.), Physical ––– (1992a), “Bell's theorem: what it takes”, British Journal for the
Theory and Its Interpretation: Essays in Honor of Jeffrey Bub, Philosophy of Science 42, 41-83.
Western Ontario Series in Philosophy of Science, New York: ––– (1992b), “David Lewis meets John Bell”, Philosophy of Science
Springer, pp. 1-28. 59, 26-43.
Bohm, D. (1951), Quantum Theory, Englewood Cliffs, New Jersey: ––– (1994), “Outcome dependence and stochastic Einstein locality”,
Prentice-Hall. in. D. Prawitz and D. Westerstähl (1994) (eds), Logic and
––– (1952), “A suggested interpretation of the quantum theory in Philosophy of Science in Uppsala, Dordrecht: Kluwer, pp. 385-424.
terms of hidden variables. I & II”, Physical Review 85, 166-179 and ––– (1995), “Words, minds and quanta”, Aristotelian Society
180-193. Supplementary 69, 113-158.
––– (1953), “Proof that probability density approaches |ψ| 2 in causal ––– (1996), “Whither the minds?”, British Journal for the
interpretation of the quantum theory”, Physical Review 89(2), 15. Philosophy of Science 47, 200-221.
Bohm, D., and Bub, J. (1966), “A proposed solution of the Butterfield, J. N., Fleming, G. N., Ghirardi, G. C. and Grassi, R.
measurement problem in quantum mechanics by a hidden variable (1993), “Parameter dependence in dynamical models for state
theory”, Reviews of Modern Physics 38, 453-469. reductions”, International Journal of Theoretical Physics 32, 2287-
2303.

82 Stanford Encyclopedia of Philosophy Spring 2009 Edition 83


Action at a Distance in Quantum Mechanics Joseph Berkovitz

2303. Science 51, 155-164.


Cartwright, N. (1989), Nature's Capacities and Their Measurements, Cohen, I. B. (1978), Isaac Newton's Papers & Letters on Natural
Oxford: Clarendon Press. Philosophy, 2nd edition, Cambridge, Mass. and London.
Chang, H. and Cartwright, N. (1993), “Causality and realism in the Costa de Beauregard, O. (1977), “Time symmetry and the Einstein
EPR experiment”, Erkenntnis 38, 169-190. paradox”, Il Nuovo Cimento 42B, 41-64.
Clauser, J. F. and Horne, M. A. (1974), “Experimental consequences ––– (1979), “Time symmetry and the Einstein paradox – II”, Il
of objective local theories”, Physical Review D 10, 526-535. Nuovo Cimento 51B, 267-279.
Clauser, J. F., Horne, M. A., Shimony, A., and Holt, R. A. (1969), ––– (1985), “On some frequent but controversial statements
“Proposed experiment to test local hidden-variable theories”, concerning the Einstein-Podolsky-Rosen correlations”, Foundations
Physical Review Letters 23, 880-884. of Physics 15, 871-887.
Clifton, R. K. (1991),“Nonlocality in quantum mechanics: signalling, Cramer, J. (1980), “Generalised absorber theory and the Einstein-
counterfactuals, probability and causation”, Ph.D. thesis, Cambridge Podolsky-Rosen paradox”, Physical Review D 22, 362-376.
University. ––– (1986), “The transactional interpretation of quantum
––– (1995), “Independently motivating the Kochen-Dieks modal mechanics”, Reviews of Modern Physics 58, 647-687.
interpretation of quantum mechanics”, British Journal for Cushing, J. (1994), Quantum Mechanics: Historical Contingency
Philosophy of Science 46, 33-57. and the Copenhagen Hegemony, Chicago: Chicago University Press.
––– (1996a): “The properties of modal interpretations of quantum Cushing, J. and McMullin E., (eds) (1989), Philosophical
mechanics”, British Journal for Philosophy of Science 47 (1996): Consequences of Quantum Theories: Reflections on Bell's Theorem,
371-398. Notre Dame: University of Notre Dame Press.
––– (ed.) (1996b), Perspectives on Quantum Reality: Non- Cushing, J., Fine, A. and Goldstein, S. (eds) (1996), Bohmian
Relativistic, Relativistic and Field-Theoretic, Dordrecht: Kluwer. Mechanics and Quantum Theory: An Appraisal, Dordrecht: Kluwer.
––– (2000), “The modal interpretation of algebraic quantum field Deutsch, D. (1999), “Quantum theory of probability and decisions”,
theory”, Physics Letters A 271, 167-177. Proceedings of the Royal Society of London A 455, 3129-3137.
Clifton, R. K., Butterfield, J. N. and Redhead, M. L. G. (1990), Dewdney, C. Holland, P. R. and Kyprianidis, A. (1987), “A causal
“Nonlocal influences and possible worlds—a Stapp in the wrong account of non-local Einstein-Podolsky-Rosen spin correlations”,
direction”, British Journal for the Philosophy of Science 41, 5-58. Journal of Physics A 20(14), 4717-4732.
Clifton, R. K. and Monton, B. (1999), “Losing your marbles in De Witt, B. S. (1970), “Quantum Mechanics and Reality”, Physics
wavefunction collapse theories”, British Journal for Philosophy of Today 23, No. 9, 30-35. Reprinted in De Witt and Graham (1973),
Science 50, 697-717. pp. 155-165.
––– (2000), “Counting marbles with ‘accessible’ mass density: a ––– (1971), “The many-universes interpretation of quantum
reply to Bassi and Ghirardi”, British Journal for Philosophy of mechanics”, in B. d'Espagnat (ed), Foundations of Quantum
Science 51, 155-164. Mechanics: Proceedings of the International School of Physics

84 Stanford Encyclopedia of Philosophy Spring 2009 Edition 85


Action at a Distance in Quantum Mechanics Joseph Berkovitz

Mechanics: Proceedings of the International School of Physics Dürr, D., Goldstein, S. and Zanghì, N. (1992a), “Quantum
‘Enrico Fermi’, Course IL, New York: Benjamin, pp. 211-62. equilibrium and the origin of absolute uncertainty”, J. of Statistical
Reprinted in DeWitt and Graham (1973), pp. 167-218. Physics 67, 843-907.
DeWitt, B. S., and Graham, N. (eds) (1973), The Many-Worlds ––– (1992b), “Quantum mechanics, randomness, and deterministic
Interpretation of Quantum Mechanics, Princeton: Princeton reality”, Physics Letters A 172, 6-12.
University Press. ––– (1996), “Bohmian mechanics as the foundations of quantum
Dickson, M (1993), “Stapp's theorem without counterfactual mechanics”, in Cushing, Fine and Goldstein (1996), pp. 21-44.
committments: Why it fails nonetheless”, Studies in History and ––– (1997), “Bohmian mechanics and the meaning of the wave
Philosophy of Science 24, 791-814. function”, in Cohen, R. S., Horne, M., and Stachel, J., (eds.),
––– (1995), “Reply to H. Stapp's Comment”, Studies in History and Experimental Metaphysics—Quantum Mechanical Studies for Abner
Philosophy of Science 25, 965-966. Shimony, Volume One,Boston Studies in the Philosophy of Science
––– (1996a), “Is Bohm's theory local?”, in Clifton (1996b), pp. 321- 193, Boston: Kluwer Academic Publishers.
330. Earman, J. and Ruetsche, L. (2005), “Relativistic invariance and
––– (1996b), “Determinism and locality in quantum systems”, modal interpretations”, Philosophy of Science 72, 557-583.
Synthese 107, 52-82. Eberhard, P. (1978), “Bell's theorem and different concepts of
––– (1998),Quantum Chance and Nonlocality, Cambridge: locality”, Nuovo Cimento 46B, 392-419.
Cambridge University Press. Einstein, A. (1948), “Quantenmechanik und wirklichkeit”, Dialectica
Dickson, M. and Clifton, R. (1998), “Lorentz invariance in modal 2, 320-324.
interpretations”, in Dieks and Vermaas 1998, pp. 9-47. Einstein, A., Podolsky, R. and Rosen, N. (1935), “Can quantum-
Dieks, D. (1988) “The formalism of quantum mechanics: an mechanical description of physical reality be considered complete?”,
objective description of reality?”, Annalen der Physik 7, 174-190. Physics Review 47, 777-780
––– (1989), “Resolution of the measurement problem through Esfeld, M. (2001), Holism in Philosophy of Mind and Philosophy of
decoherence of the quantum state”, Physics Letters A 142, 439-446. Physics, Dordrecht: Kluwer Academic Publishers.
Dieks, D. and Vermaas, P. (eds) (1998), The Modal Interpretation of Everett, H. (1957a), “On the foundations of quantum mechanics”,
Quantum Mechanics, Dordrecht: Kluwer. Ph.D. thesis, Princeton University.
Donald, M. (1990) “Quantum theory and the brain”, Philosophical ––– (1957b), “‘Relative state’ formulation of quantum mechanics”,
Transactions of the Royal Society of London A 427, 43-93. Reviews of Modern Physics 29, 454-62. Reprinted in DeWitt and
––– (1995), “A mathematical characterization of the physical Graham (1973), pp. 141-150.
structure of observables”, Foundations of Physics 25, 529-571. ––– (1973), “The theory of universal wave function”, in DeWitt and
Dowe, P. (2000), Physical Causation, Cambridge: Cambridge Graham (1973), pp. 3-140.
University Press. Fine, A. (1981), “Correlations and physical locality”, in P. Asquith
and R. Giere (eds), PSA 1980, vol. 2, East Lansing, Michigan:

86 Stanford Encyclopedia of Philosophy Spring 2009 Edition 87


Action at a Distance in Quantum Mechanics Joseph Berkovitz

and R. Giere (eds), PSA 1980, vol. 2, East Lansing, Michigan: Gell-Mann, M., and Hartle, J. B. (1990), “Quantum mechanics in the
Philosophy of Science Association, pp. 535-562. light of quantum cosmology”, in W. H. Zurek (ed), Complexity,
––– (1982a), “Hidden variables, joint probability, and the Bell Entropy, and the Physics of Information, Proceedings of the Santa Fe
inequalities”, Physical Review Letters 48, 291-295. Institute Studies in the Sciences of Complexity, vol. VIII, Redwood
––– (1982b), “Joint distributions, quantum correlations and City, CA: Addison-Wesley, pp. 425-458.
commuting observables”, Journal of Mathematical Physics 23, Ghirardi, G. C. (1996), “Properties and events in a relativistic
1306-1310. context: revisiting the dynamical reduction program”, Foundations
––– (1986), The Shaky Game, Chicago: The University of Chicago of Physics Letters 9, 313-355.
Press. Ghirardi, G. C., Rimini, A. and Weber, T. (1980), “A general
––– (1989a), “Correlations and efficiency: testing the Bell argument against superluminal transmission through the quantum
inequalities”, Foundations of Physics 19, 453-478. mechanical measurement process”, Lettere Al Nuovo Cimento 27,
––– (1989b), “Do correlations need to be explained?”, in Cushing 293-298.
and McMullin (1989), pp. 175-194. ––– (1986), “Unified dynamics for microscopic and macroscopic
Fleming, G. N. (1989), “Lorentz invariant state reduction, and systems”, Physical Review D 34, 470-491.
localization”, in A. Fine and M. Forbes (eds), PSA 1988, vol. 2, East Ghirardi, G. C., Pearle, P. and Rimini, A. (1990), “Markov processes
Lansing, MI: Philosophy of Science Association, pp. 112-126. in Hilbert space and continuous spontaneous localisation of systems
––– (1992), “The objectivity and invariance of quantum predictions”, of identical particles”, Physical Review A 42, 78-89.
in D. M. Hull, M. Forbes and Q. Okruhlik (eds), PSA 1992 Vol. 1, Ghirardi, G. C., Grassi, R., Butterfield, J., and Fleming, G. N.
pp. 104-13, East Lansing: Philosophy of Science Association. (1993), “Parameter dependence and outcome dependence in dynamic
––– (1996) “Just how radical is hyperplane dependence?”, in Clifton models for state-vector reduction”, Foundations of Physics, 23, 341-
(1996b), pp. 11-28. 364.
Fleming, G. N. and Bennett, H. (1989), “Hyperplane dependence in Ghirardi, G. C., and Bassi, A. (1999), “Do dynamical reduction
relativistic quantum mechanics”, Foundations of Physics 19, 231- models imply that arithmetic does not apply to ordinary macroscopic
267. objects”, British Journal for the Philosophy of Science 50, 49-64.
Friedman, M. (1983), Foundations of Spacetime Theories: Ghirardi, G. C. and Grassi, R. (1994), “Outcome predictions and
Relativistic Physics and Philosophy of Science. Princeton: Princeton property attribution: the EPR argument reconsidered”, Studies in
University Press. History and Philosophy of Modern Physics 25, 397-423.
Frigg, R. (2003), “On the property structure of realist collapse Gill, R. D. (2003), “On an argument of David Deutsch”, in M.
interpretations of quantum mechanics and the so-called ‘counting Schürmann and U. Franz (eds), Quantum Probability and Infinite
anomaly’”, International Studies in the Philosophy of Science 17, 43- Dimensional Analysis, World Scientific, pp. 277-292. [Preprint
57. available online.]

88 Stanford Encyclopedia of Philosophy Spring 2009 Edition 89


Action at a Distance in Quantum Mechanics Joseph Berkovitz

Greaves, H. (2004), “Understanding Deutsch's probability in a Henry, J. (1994), “Pray do not ascribe that notion to me…”, in J. E.
deterministic multiverse”, Studies in History and Philosophy of Force and R. H. Popkin (eds), The Books of Nature and Scripture,
Modern Physics 34, 423-456. Dordrecht: Kluwer, pp. 123-147.
Greenberger, D. M., Horne, M., and Zeilinger, A. (1989), “Going Hesse, M. B. (1969), Forces and Fields: The Concept of Action At A
beyond Bell's theorem”, in M. Kafatos (ed), Bell's Theorem, Distance in the History of Physics, London: Thomas Nelson and
Quantum Theory, and Conception of the Universe, Dordrecht: Sons.
Kluwer, pp. 69-72. Holland, P. R. (1993), The Quantum Theory of Motion, Cambridge:
Greenberger, D.M., Horne, M.A., Shimony, A., Zeilinger, A. Cambridge University Press.
(1990),“Bell's theorem without inequalities”, American Journal of Howard, D. (1985). “Einstein on locality and separability”, Studies in
Physics 58, 1131-1143. History and Philosophy of Science 16, pp. 171-201.
Healey, R. (1989), The Philosophy of Quantum Mechanics, ––– (1989), “Holism, separability and the metaphysical implications
Cambridge: Cambridge University Press. of the Bell experiments”, in Cushing and McMullin (1989), pp. 224-
––– (1991) “Holism and nonseparability”, Journal of Philosophy 88, 253.
393-421. Hughes, R. I .G (1989), The Structure and Interpretation of
––– (1992), “Chasing quantum causes: how wild is the goose?”, Quantum Mechanics, Cambridge Mass.: Harvard University Press.
Philosophical Topics 20,181-205. Jarrett, J. (1984), “On the physical significance of the locality
––– (1994), “Nonseparable processes and causal explanation”, conditions in the Bell arguments”, Noûs 18, 569-589.
Studies in History and Philosophy of Modern Physics 25, 337-374. ––– (1989), “Bell's theorem: a guide to the implications”, in Cushing
Hellman, G. (1982), “Stochastic Einstein locality and the Bell and McMullin (1989), pp. 60-79.
theorems”, Synthese 53, 461-503. Jones, M. and Clifton, R. (1993), “Against experimental
Hemmo, M. (1996a), “Quantum mechanics without collapse: modal metaphysics”, in P. A. French et al. (eds), Midwest Studies in
interpretations, histories and many worlds”, Ph.D. Thesis, Cambridge Philosophy XVIII, Notre Dame, Indiana: University of Notre Dame
University. Press, pp. 297-316.
––– (1996b), “Possible worlds in the modal interpretation”, Jordan, T. F. (1983), “Quantum correlations do not transmit signals”,
Philosophy of Science, Supplement to Vol. 63, S330-S337. Physics Letters A 94, 264.
Hemmo, M. and Pitowsky, I. (2003), “Probability and nonlocality in Kastner, R. (2006), “Cramer's transactional interpretation and causal
many minds interpretations of quantum mechanics”, British Journal loop problems”, Synthese 150 no. 1, 1-14. [Earlier version available
for the Philosophy of Science 54, 225-243. [Preprint available online.]
online.] Kim, J. (1978), “Supervenience and nomological
––– (2005), “Quantum probability and many worlds”, forthcoming in incommensurables”, American Philosophical Quarterly 15, 149-156.
Studies in History and Philosophy of Modern Physics. Kochen, S. (1985), “A new interpretation of quantum mechanics”, in
P. Lathi and P. Mittelstaedt (eds), Symposium on the Foundations of

90 Stanford Encyclopedia of Philosophy Spring 2009 Edition 91


Action at a Distance in Quantum Mechanics Joseph Berkovitz

P. Lathi and P. Mittelstaedt (eds), Symposium on the Foundations of Malament, D. (1996), “In defense of dogma: Why there cannot be a
Modern Physics, Singapore: World Scientific, pp. 151-169. quantum mechanical theory of (localizable) particles?”, in Clifton
Kochen, S. and Specker, E. (1967), “The problem of hidden 1996b, pp. 1-10.
variables in quantum mechanics”, Journal of Mathematics and Maudlin, T. (1994), Quantum Nonlocality and Relativity, Oxford:
Mechanics 17, 59-87. Reprinted in Hooker (1975), pp. 293-328. Blackwell.
Krips, H. (1987), The Metaphysics of Quantum Theory, Oxford: ––– (1996), “Space-time in the quantum world” in Cushing, Fine and
Clarendon Press. Goldstein (1996), pp. 285-307.
Kronz, F. (1990a), “Jarrett completeness and superluminal signals”, McLaughlin, B. P. (1994), “Varieties of supervenience”, in E.
in A. Fine, M. Forbes and L. Wessels (eds), PSA 1990, East Lansing: Savellos and O. Yalchin (eds), Supervenience: New Essays,
Philosophy of Science Association, pp. 227-239. Cambridge: Cambridge University Press, pp. 16-59.
––– (1990b), “Hidden locality, conspiracy and superluminal signals”, McMullin, E. (1989), “The explanation of distant action: historical
Philosophy of Science 57, 420-444. notes”, in Cushing and McMullin (1989), pp. 272-302.
Langton, R. and Lewis, D. (1998), “Defining ‘intrinsic’”, Philosophy Mermin, N. D. (1990), “Quantum mysteries revisited”, American
and Phenomenological Research 58, 333-45. Reprinted in D. Lewis, Journal of Physics 58, 731-734.
Papers in Metaphysics and Epistemology, Cambridge: Cambridge -––– (1998), Mermin, N.D., “What is quantum mechanics trying to
University Press, pp. 116-132. tell us?”, American Journal of Physics 66, 753-767.
Laudisa, F. (1995), “Einstein, Bell and nonseparable realism”, The Myrvold, W. (2002a), “Modal interpretations and relativity”,
British Journal for the Philosophy of Science 46, 309-329. Foundations of Physics 32, 1773-1784.
Lewis, D. (1986), Philosophical Papers, Vol. 2, Oxford: Oxford ––– (2002b), “On peaceful coexistence: is the collapse postulate
University Press. incompatible with relativity?’ Studies in History and Philosophy of
Lewis, P. J. (1997), “Quantum mechanics, orthogonality and Modern Physics 33, 435-466. [Preprint available from the author.]
counting”, British Journal for Philosophy of Science 48, 313-328. Parker, D. (2003), “Finding your Marbles in wavefunction collapse
––– (2003a), “Four strategies for dealing with the counting anomaly theories”, Studies in History and Philosophy of Modern Physics 34,
in spontaneous collapse theories of quantum mechanics”, 607-620.
International Studies in the Philosophy of Science 17 no. 2, 137-142. Pearle, P. (1976), “Reduction of statevector by a nonlinear
Lockwood, M. (1996a), “Many minds interpretation of quantum Schrödinger equation”, Physical Review D 13, 857-868.
mechanics”, British Journal for the Philosophy of Science 47, 159- ––– (1989), “Combining stochastic dynamical state-vector reduction
188. with spontaneous localization”, Physical Review A 39, 2277-2289.
––– (1996b), “Many minds interpretations of quantum mechanics: ––– (1996), “Wavefunction collapse models with nonwhite noise”, in
replies to replies”, British Journal for the Philosophy of Science 47, Clifton (1996), pp. 93-109.
445-461. Pitowsky, I. (1989), Quantum Probability—Quantum Logic, Lecture
Notes in Physics 321, New York: Springer-Verlag.

92 Stanford Encyclopedia of Philosophy Spring 2009 Edition 93


Action at a Distance in Quantum Mechanics Joseph Berkovitz

Notes in Physics 321, New York: Springer-Verlag. ––– (2005), preprint of “What is probability?”, forthcoming in A.
Placek, T. and J. Butterfield (eds) (2002), Non-locality and Modality, Elitzur, S. Dolev and N. Kolenda (eds), Quo Vadis Quantum
Proceedings of the NATO Advanced Research Workshop on Mechanics, New York: Springer. [Preprint available online.]
Modality, Probability and Bell's Theorems, Kluwer. Shimony, A. (1984), “Controllable and uncontrollable non-locality”,
Price, H. (1984), “The philosophy and physics of affecting the past”, in S. Kamefuchi et al. (eds), Foundations of Quantum Mechanics in
Synthese 16, 299-323. Light of the New Technology, Tokyo: Physical Society of Japan, pp.
––– (1994), “A neglected route to realism about quantum 225-230. Reprinted in Shimony (1993), pp. 130-139.
mechanics”, Mind 103, 303-336. ––– (1993), Search for Naturalistic Worldview, Vol. 2, Cambridge:
––– (1996), Time's Arrow and Archimedes’ Point. Oxford: Oxford Cambridge University Press.
University Press. Spekkens, R. W. and Sipe, J. E. (2001a), “Non-orthogonal core
––– (2006), “Decision-based probabilities in the Everett projectors for modal interpretations of quantum mechanics”,
Interpretation: Comments on Wallace and Greaves”, a typescript. Foundations of Physics 31, 1403-1430.
Redhead, M. L. G. (1987), Incompleteness, Nonlocality, and ––– (2001b), “A modal interpretation of quantum mechanics based
Realism, Oxford: Clarendon Press. on a principle of entropy minimization”, Foundations of Physics 31,
––– (1989), “Nonfactorizability, stochastic causality, and passion-at- 1431-1464.
a-distance”, in Cushing and McMullin (1989), pp. 145-153. Stairs, A. (1989), “Jarrett's locality Condition and Causal Paradox”,
Redhead, M. L. G. and La Rivière, P. (1997), “ The relativistic EPR in A. Fine and M. Forbes (eds), PSA 1988, East Lansing, MI:
argument”, in R. S. Cohen, M. Horne and J. Stachel (eds.), Philosophy of Science Association, pp. 318-325.
Potentiality, Entanglement and Passion-at-a-Distance (Kluwer, Stapp, H. (1989), “Quantum nonlocality and the description of
Dordrecht), pp. 207-215. nature”, in Cushing and McMullin 1989, pp. 154-174.
Reichenbach, H. (1956), The Direction of Time, Berkley, Calif.: ––– (1990), “Comments on nonlocal influences and possible
University of California Press. worlds”, British Journal for the Philosophy of Science 41, 59-72.
Salmon, W. (1998), Causality and Explanation, Oxford: Oxford ––– (1994), “Comments on "Stapp's theorem without
University Press. counterfactuals"”, Studies in History and Philosophy of Science 25,
Saunders, S. (1995), “Time, quantum mechanics and decoherence”, 929-934.
Synthese 102, 235-266. Suppes, P. and M. Zanotti (1976), “On the determinism of hidden
––– (1998), “Time, quantum mechanics, and probability”, Synthese, variable theories with strict correlation and conditional statistical
114, 405-444. [Preprint available from the author.] independence of observables”, in P. Suppes (1976) (ed), Logic and
––– (2004), “Derivation of the Born rule from operational Probability in Quantum Mechanics, Dordrecht: Reidel, pp. 445-455.
assumptions”, Proceedings of the Royal Society of London A, 460, 1- Sutherland, R. I. (1983), “Bell's theorem and backwards-in-time
18. [Preprint available online.] causality”, International Journal of Theoretical Physics 22, 377-384.

94 Stanford Encyclopedia of Philosophy Spring 2009 Edition 95


Action at a Distance in Quantum Mechanics Joseph Berkovitz

––– (1998), “Density formalism for quantum theory”, Foundations Oxford University Press.
of Physics 28 no. 7, 1158-1190. Vermaas, P. E. (1997), “A no-go theorem for joint property
Teller, P. (1986a), “Relational holism and quantum mechanics”, ascriptions in modal interpretations of quantum mechanics”, Physical
British Journal for Philosophy of Science 37, 71-81. Review Letters 78, 2033-2037.
––– (1986b) “Space-time as a physical quantity”, in R. Kargon and ––– (1999), A Philosopher's Understanding of Quantum Mechanics:
P. Achinstein (eds), Kelvin's Baltimore Lectures and Modern Possibilities and Impossibilities of a Modal Interpretation,
Theoretical Physics, MIT Press, pp. 425-448. Cambridge: Cambridge University Press.
––– (1989), “Relativity, relational holism and quantum mechanics”, Vermaas, P. E. and Dieks, D. (1995) “The modal interpretation of
in Cushing and McMullin (1989), pp. 208-223. quantum mechanics and its generalisation to density Operators”,
Vaidman, L., (1998), “On schizophrenic experiences of the neutron Foundations of Physics 25, 145-158.
or why we should believe in the many-worlds interpretation of Vink, J. (1993), “Quantum mechanics in terms of discrete beables”,
quantum theory”, International Studies in the Philosophy of Science Physical Review A 48, 1808-1818.
12, 245-261. [Preprint available online.] Wallace, D. (2003), “Everettian Rationality: defending Deutsch's
Valentini, A. (1991a), “Signal locality, uncertainty and the approach to probability in the Everett interpretation”, Studies in the
subquantum H-Theorem. I”, Physics Letters A 156, 5-11. History and Philosophy of Modern Physics 34, 415-438. [Preprint
––– (1991b), “Signal locality, uncertainty and the subquantum H- available online.]
Theorem. II”, Physics Letters A 158, 1-8. Winsberg, E. and Fine, A. (2003), “Quantum life: interaction,
––– (1992), On the Pilot-Wave Theory of Classical, Quantum and entanglement and separation”, Journal of Philosophy 100, 80-97.
Subquantum Physics, Ph.D. Thesis, International School of
Advanced Studies, Trieste, Italy. Other Internet Resources
––– (1996), “Pilot-wave theory of fields, gravitation and
cosmology”, in Cushing, Fine and Goldstein (1996), pp. 45-66. Donald, M. (1999), Progress in a many-minds interpretation of
van Fraassen, B. C. (1973), “Semantic analysis of quantum logic”, in quantum theory.
C. A. Hooker (ed), Contemporary Research in the Foundations and Fogel, B. (2004), The Structure of Quantum Holism. (PDF)
Philosophy of Quantum Theory, Dordrecht: Reidel, pp. 180-213. Larsson, J-A and Semitecolos, J. (2000), Strict detector-efficiency
––– (1981), “A modal interpretation of quantum mechanics”, in E. bounds for n-site Clauser-Horne inequalities.
Beltrametti and B. C. van Fraassen (eds), Current Issues in Quantum Larsson, J. A. (2002), A Kochen-Specker inequality.
Mechanics, New York: Plenium, pp. 229-258. Lewis, P. J. (2003b), Deutsch on Quantum Decision Theory.
––– (1982), “The charybids of realism”, Synthese 52, 25-38. ––– (2004), Interpreting spontaneous collapse theories (DOC file).
Reprinted in Cushing and McMullin (1989), pp. 97-113. Sutherland, R. I. (2006), Causally symmetric Bohm model (PDF).
––– (1991), Quantum Mechanics: An Empiricist View, Oxford: Tumulka, R. (2004), A relativistic version of the Ghirardi-Rimini-
Oxford University Press. Weber model (PDF).

96 Stanford Encyclopedia of Philosophy Spring 2009 Edition 97


Action at a Distance in Quantum Mechanics Joseph Berkovitz

Valentini, A. (2002), Signal-locality in hidden-variables theories


1. Intuitively, the spin-component of a particle in a certain direction can
(PDF).
be thought of as its intrinsic angular momentum along that direction. But,
Valentini, A. and Westman, H. (2004), Dynamical origin of quantum
as we shall see in section 5, the nature of spin properties depends on the
probabilities (PDF).
interpretation of quantum mechanics. In any case, the exact nature of this
Wallace, D. (2002), Quantum probability and decision theory,
quantity will not be essential for what follows in sections 1-4. The
revisited, also available from PhilSci Archive.
important thing is that in various quantum states the properties of distant
––– (2005a), Epistemology quantised: circumstances in which we
physical systems may be curiously correlated.
should come to believe in the Everett interpretation (PDF).
––– (2005b), Quantum probability from subjective likelihood: 2. Recall Bell's (1981) example of Bertlmann's socks. ‘Dr. Bertlmann
improving on Deutsch's proof of the probability rule, also available likes to wear two socks of different colours. Which colour he will have on
from PhilSci Archive. a given foot on a given day is quite unpredictable. But when you see that
––– (2005c), Everett and structure. the first sock is pink you can already be sure that the second sock will not
be pink. Observation of the first, and experience of Bertlmann, gives
Related Entries immediate information about the second.’

Bell's Theorem | causation: causal processes | intrinsic vs. extrinsic 3. Two comments:
properties | physics: holism and nonseparability | physics: Reichenbach's
common cause principle | quantum mechanics | quantum mechanics: (i) In some Bell-type models of the EPR/B experiment, it is assumed that
Bohmian mechanics | quantum mechanics: collapse theories | quantum in addition to the pair's state and the settings of the measurement
mechanics: Everett's relative-state formulation of | quantum mechanics: apparatuses, there are other factors that may be relevant for the
Kochen-Specker theorem | quantum mechanics: many-worlds probabilities of measurement outcomes. In particular, in his presentation
interpretation of | quantum mechanics: modal interpretations of | quantum of stochastic, local models of the EPR/B experiment Bell (1971, p. 37)
theory: the Einstein-Podolsky-Rosen argument in | supervenience | assumes that the setting of the apparatuses need not specify their entire
Uncertainty Principle relevant states. The outcomes may also be influenced by some other
aspects of the apparatus microstates, which may be different for the same
Acknowledgments settings (see also Jarrett 1984). More generally, in addition to the state of
the L- (R-) particle and the setting of the L- (R-) measurement apparatus,
For comments on earlier versions of this entry, I am very grateful to
there may be some other (local) physical quantities that are relevant for
Guido Bacciagaluppi.
the probability of the L- (R-) measurement outcome. That is, letting α and
β denote all the relevant local physical quantities (other than the settings)
Notes to Action at a Distance in Quantum that are relevant for the probability of the L- and the R-outcome
Mechanics respectively, in such models the single and joint probabilities of outcomes
will be: P λ l α (xl), P λ r β (xl) and P λ l r α β (xl & yr). We shall refer to this
98 Stanford Encyclopedia of Philosophy Spring 2009 Edition 99
Action at a Distance in Quantum Mechanics Joseph Berkovitz

probabilities of the big-space approach, e.g. P(xl / λ & r), have no


will be: P λ l α (xl), P λ r β (xl) and P λ l r α β (xl & yr). We shall refer to this correspondence in the many-spaces approach. Third, as we shall see in
type of models in section 7. But, for the sake of simplicity, in the rest of the next section, factorizability can be analyzed into two conditions:
this entry we shall focus on the simpler models above. parameter independence and outcome independence. Berkovitz (2002)
argues that the meaning of parameter independence need not be the same
(ii) There are two different approaches to modeling the probabilities in in the two different approaches. That is, in some circumstances the
Bell-type models of the EPR/B experiment: The many-spaces and the parameter independence of the big-space approach expresses different
big-space approaches (see Butterfield 1989, 1992a). In the many-space properties than the parameter independence of the many-spaces approach.
approach, which we use in this review, each triple of pair's state, L- and Indeed, in these circumstances the parameter independence of the many-
R-setting labels a different probability space of measurement outcomes. spaces approach fails, whereas the parameter independence of the big-
For example, letting l and l′ be different L-apparatus settings, the space approach holds. For arguments for the superiority of the many-
probability P λ l r(xl & yr) belongs to one probability space, whereas the spaces approach over the big-space approach, see Butterfield (1989, p.
probability P λ l′  r(xl′ & yr) belongs to another. By contrast, in the big- 118), and Berkovitz (2002, section 4.2).
space approach, all the probabilities of a Bell-type model belong to one
big probability space. In this approach the probabilities of outcomes are 4. Or when the range of the values of λ is discrete,
expressed in terms of conditional probabilities. For example, the
probabilities P(xl & yr / λ & l & r) and P(xl′ & yr / λ & l′ & r) correspond P ψ l r(xl & yr) = ∑ λ P λ l r(xl & yr) · ρ ψ l r(λ),
to P λ l r (xl & yr) and P λ l′ r (xl′ & yr), respectively. (Note that in contrast to P ψ l (xl) = ∑ λ P λ l (xl) · ρ ψ l(λ), and
the above notation, in the literature probabilities of spin-measurement P ψ r(yr) = ∑ λ P λ r(yr) · ρ ψ r(λ).
outcomes in the big-space approach are frequently expressed as
5. For a dissenting view, see Fine (1981, 1982a), Cartwright (1989) and
conditional probabilities of non-specific spin-outcomes, i.e. the non-
Chang and Cartwright (1993). We shall discuss this view in section 9.
specific outcomes ‘up’ or ‘down’, given certain settings: P(x & y / λ & l
& r), where ‘x’ and ‘y’ denote non-specific outcomes.) Mathematically, 6. While the fullest analysis of factorizability is due to Jarrett, precursors
the two approaches can easily be related to each other. In particular, one are Suppes and Zanotti (1976) and van Fraassen (1982).
can construct a big-probability space in which the conditional
probabilities of outcomes, given a pair's state and apparatus settings, are 7. See van Fraassen (1982) and Jarrett (1984, 1989).
equal to the corresponding unconditional probabilities in the many-spaces
8. As we shall see below, in the literature the term ‘interpretation’ is
approaches: P(xl & yr / λ & l & r) = P λ l r (xl & yr), P(xl′ & yr / λ & l′ &
frequently used to refer to alternative quantum theories. The question of
r) = P λ l′ r (xl′ & yr), etc. But, conceptually the two approaches are
whether this use is justified and the criteria for distinguishing between an
different. First, in contrast to the many-spaces approach, in the big-space
interpretation of orthodox quantum mechanics and an alternative quantum
approach it is presupposed that settings always have definite probabilities
theory will be insubstantial for the considerations below.
(Butterfield 1989, p. 118, 1992a, section 2). Secondly, some of the
probabilities of the big-space approach, e.g. P(xl / λ & r), have no

100 Stanford Encyclopedia of Philosophy Spring 2009 Edition 101


Action at a Distance in Quantum Mechanics Joseph Berkovitz

9. For a history of the notion of action at a distance, see Hesse (1969). ‘up’ and z-spin ‘down,’ respectively). Suppose that we perform a z-spin
measurement on particle 1, and switch off the R-apparatus. Then
10. In reality, the position of different particles will be different: |up i> pi (suppressing for simplicity's sake the free time evolution of the two
(|down i> pi ). But this is immaterial for the analysis below. wavepackets as they move towards their respective Stern-Gerlach devices
and the states of the Stern-Gerlach devices), the state of the guiding field
11. This is a variant of the so-called ‘tails problem’ (see the entry on
during the L-measurement will be:
collapse theories, section 12, and Albert 1992, chapter 5).
1/√2 f2(z2) ( f1(z1 + g1T) |z-up>1 |z-down> 2 − f1(z1 − g1T) |z-down> 1 |z-up>2)
12. For a recent interesting discussion of Newton's view of action at a
distance, see Henry (1994), and references therein. where g1 is the coupling constant for the spin measurement on particle 1
(coupling the position and the spin degrees of freedoms that are related to
13. For the Clarke-Leibniz correspondence, see Alexander (1956).
the guidance of particle 1); and T is the duration of the measurement.
14. Of course, here ‘field’ is not intended to mean a field in the sense of Since the guiding field of the particle pair factorizes into f2(z2) and ( f1(z1
quantum field theory. + g1T) |z-up> 1 |z-down>2 − f1(z1 − g1T) |z-down>1 |z-up> 2), it follows
from the guiding equation that particle 2's velocity along the z-axis does
15. For discussions of this version of Bohm's theory, see for example not depend on particle 1's position.
Dürr, Goldstein and Zanghì (1992a), Albert (1992) and Cushing (1994).
18. See Bohm, Schiller and Tiomno (1955), Dewdney, Holland and
16. The wave function propagates according to the Schrödinger equation Kyprianidis (1987), Bohm and Hiley (1993, chapter 10), and Holland
in the ‘configuration’ space of the particles, which for an N-particle (1993).
system is a 3N-dimensional space, coordinatized by the 3N position
coordinates of the particles. For more details, see the entry on Bohmian 19. While in the minimal and the non-minimal Bohm theories, the wave
mechanics. function is interpreted as a field, Dürr, Goldstein and Zanghì (1997,
section 12) propose that the wave function should be interpreted as a
17. Here follows a more technical account of the above experiment parameter of a physical law. This, they argue, may explain why there is
according to the minimal Bohm theory. Let the wave function, i.e. the no action of configurations of particles on wave functions.
state of the guiding field, before any measurement occurs be:
20. For discussions of the above experiment in the non-minimal theory,
ψ = 1/√2f1(z1) f2(z2) ( |z-up> 1 |z-down>2 − |z-down>1 |z-up> 2), see Dewdney, Holland and Kyprianidis (1987), Bohm and Hiley (1993,
section 10.6), and Holland (1993, section 11.3).
where f1(z1) and f2(z2) are non-overlapping Gaussian wavepackets; z1
and z2 are respectively the positions of particle 1 and particle 2 along the 21. For discussions of the prospects of relativistic modal interpretations,
z-direction; and |z-up> and |z-down> are z-spin eigenstates (i.e. z-spin see Dickson and Clifton (1998), Arntzenius (1998), Myrvold (2002a),
‘up’ and z-spin ‘down,’ respectively). Suppose that we perform a z-spin Earman and Ruetsche (2005) and Berkovitz and Hemmo (2005, 2006a,b).

102 Stanford Encyclopedia of Philosophy Spring 2009 Edition 103


Action at a Distance in Quantum Mechanics Joseph Berkovitz

Earman and Ruetsche (2005) and Berkovitz and Hemmo (2005, 2006a,b). 27. The challenge is to explicate the nature of such holistic properties and
We shall discuss this issue at the end of this section and in section 10.2. to relate them to our experience.

22. If some of the ci are degenerate, the Schmidt biorthogonal 28. That is, the property of a system is given by the spectral
decomposition is not unique, and the properties assigned by the above decomposition of its so-called ‘reduced state’ (a statistical operator
rule are projections onto multi-dimensional subspaces. obtained by a partial tracing). For example, the reduced state of the L-
particle in the state |ψ9> is obtained by a partial tracing of |ψ 9> over the
23. Note the difference between |ψ9> and the singlet state |ψ 3>. In |ψ 9>, Hilbert space of the R-particle.
the coefficients of the two branches of the superposition are unequal. And
while EPR/B-like experiments can be prepared with both the state |ψ3> 30. The above formulation of screening off is motivated by the fact that
and the state |ψ9>, the difference between these states is significant for we work in the framework of the many-spaces approach to the
the above interpretation. For unlike |ψ3>, |ψ 9> has a unique factorization. probabilities of outcomes in Bell-type models of the EPR/B experiment.
Accordingly, the L- and the R-particle each have definite spin properties In the literature, the formulation of screening off is slightly different:
in the state |ψ9> but not in |ψ 3>.
P(x/y & CC(x,y)) = P(x / CC(x,y)) P(y / CC(x,y)) ≠ 0
24. In fact, as we shall see later in this section (in the discussion of
‘property composition’), this claim needs some qualification. P(y/x & CC(x,y)) = P(y / CC(x,y)) P(x / CC(x,y)) ≠ 0

25. In the original modal interpretations, the question of the relation 30. In his celebrated theorem, Bell did not mention Reichenbach's
between the dynamics of properties of systems and the dynamics of the principle or FactorUCP. But it is reasonable to assume that he had in
properties of their subsystems has been largely overlooked. For a mind some similar principles.
discussion of this issue, see Vermaas (1997, 1999), Berkovitz and Hemmo
(2005, 2006a,b). 31. There are some obvious candidates for superluminal signaling. First,
the potentials in the Schrödinger equation are Newtonian. Therefore, if
26. Similarly to any other physical object, the brain of a human observer one is allowed to vary the potential somewhere, this will be felt
has many different sets of relational properties, i.e. sets of properties that instantaneously throughout space. But, in the context of this entry such
are related to different systems. Brain properties that are defined relative superluminal signaling is less interesting because it will be due to
to different systems are generally different. Thus, the question arises as to Newtonian effects rather than quantum effects. Second, wave functions
which of these different brain properties are correlated to our beliefs can spread instantaneously: If you have a particle confined to a box (so
about the properties of physical systems that figure in our experience. For that its wave function is zero outside the box) and open the box, the wave
a discussion of this question, see Berkovitz and Hemmo (2005, section 6, function will instantaneously be non-zero everywhere, and superluminal
2006b, section 6). signaling will be possible. It is noteworthy, however, that the preparation
of such state requires the existence of an infinite potential barrier—a state
that is impossible. In any case, in what follows we shall focus on the
104 Stanford Encyclopedia of Philosophy Spring 2009 Edition 105
Action at a Distance in Quantum Mechanics Joseph Berkovitz

that is impossible. In any case, in what follows we shall focus on the particles. But a little reflection on the grave difficulties involved with that
question of whether the non-locality in the EPR/B experiment, as depicted task suggests that the physical feasibility and plausibility of any such
by various interpretations of quantum mechanics, can be exploited to give separable model of the EPR/B experiment will be highly questionable.
rise to superluminal signaling.
34. Friedman (1983, sections 4.6-4.7) holds that special relativity per se
32. Note that according to this suggestion, the statistical predictions of does not prohibit superluminal signaling, but that such signaling will lead
Bohm's theory slightly deviate from the statistical predictions of orthodox to paradoxes of time travel. Maudlin (1994, pp. 112-116) argues that
quantum mechanics. superluminal signaling need not imply such paradoxes, as the conditions
for them are much more complex than merely the existence of
33. It is noteworthy, however, that while separability does not imply OI, superluminal signaling.
the prospects of separable models that violate OI are dim. To see why, let
us consider Maudlin's (1994, p. 98) criticism of Howard's claim that OI 35. As Maudlin (1996, pp. 292-293) notes, it is not clear that a general
follows from spatiotemporal separability. Maudlin invites us to consider criterion for identifying a structure of spacetime as intrinsic could be
the following model for the EPR/B experiment. Suppose that each particle found.
had some means of superluminal communication, which may be realized
by a tachyon. Suppose also that each of the particles carries the same Copyright © 2009 by the author

instructions: If it arrives at a measurement apparatus without having Joseph Berkovitz

received a message from its partner, and the measurement apparatus is in


a state of being ready to measure z-spin, its state of not having definite z-
spin evolves with equal chance to either the state of having z-spin ‘up’ or
the state of having z-spin ‘down.’ It then communicates to its partner the
setting of its measurement apparatus and the outcome of the
measurement. If it receives a message from its partner, its state is
modified accordingly, so that the new chance of spin outcomes agrees
with the predictions of orthodox quantum mechanics. Such a model will
involve a violation of OI, but by construction it is separable: The particles
and the tachyons have separable states at all times, and the joint state of
any two systems is just the product of their individual states. Yet, the
model will be separable in the intended sense only if the above set of
communication instructions could be encoded into the qualitative,
intrinsic properties of each of the particles, and each of the particles could
keep an open line of communication with its partner and no other
particles. But a little reflection on the grave difficulties involved with that

106 Stanford Encyclopedia of Philosophy Spring 2009 Edition 107