See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321362644

Review and suggested resolution of the problem of Schrodinger’s cat

Article in Contemporary Physics · November 2017

DOI: 10.1080/00107514.2017.1401368

CITATIONS READS
23 907

1 author:

Art Hobson
University of Arkansas at Fayetteville
126 PUBLICATIONS 1,163 CITATIONS

SEE PROFILE

All content following this page was uploaded by Art Hobson on 25 May 2018.

The user has requested enhancement of the downloaded file.

Art Hobson Review and suggested resolution of Schrodinger's cat 1

Published in Contemporary Physics on 28 November 2017.

Link to this article: https://doi.org/10.1080/00107514.2017.1401368

Review and suggested resolution of the problem of

Schrodinger's cat
Art Hobson
Professor Emeritus of Physics, University of Arkansas, Fayetteville, AR 72701, USA

Abstract

This paper reviews and suggests a resolution of the problem of definite outcomes of
measurement. This problem, also known as "Schrodinger's cat," has long posed an
apparent paradox because the state resulting from a measurement appears to be a
quantum superposition in which the detector is in two macroscopically distinct states
(alive and dead in the case of the cat) simultaneously. Many alternative interpretations of
the quantum mathematical formalism, and several alternative modifications of the theory,
have been proposed to resolve this problem, but no consensus has formed supporting any
one of them. Applying standard quantum theory to the measurement state, together with
the analysis and results of decades of nonlocality experiments with pairs of entangled
systems, this paper shows the entangled measurement state is not a paradoxical
macroscopic superposition of states. It is instead a phase-dependent superposition of
correlations between states of the subsystems. Thus Schrodinger's cat is a non-paradoxical
"macroscopic correlation" in which one of the two correlated systems happens to be a
detector. This insight resolves the problem of definite outcomes but it does not entirely
resolve the measurement problem because the entangled state is still reversible.

Keywords: quantum measurement, problem of definite outcomes, Schrodinger's cat,

quantum entanglement, quantum nonlocality, quantum correlations.

1. The measurement problem

Quantum state collapse is a standard principle of quantum physics. Given a quantum (e.g. a
quantum object such as a photon, electron or atom) described by a superposition of
eigenstates of some observable operator O, the principle asserts that a "measurement" of O
must yield an eigenvalue of O and that the measurement causes the state of the quantum to
collapse, or jump, into the corresponding eigenstate. This raises a host of questions: What
exactly do we mean, physically and mathematically, by a "measurement" of a quantum?
Does the collapse occur all at one instant? Wouldn't an instantaneous collapse contradict
special relativity? If the collapse occurs during a time interval, then what equation
describes this time-evolution? Quantum states are presumed to follow the Schrodinger
equation, which prescribes a continuous time evolution; how can instantaneous state
collapse be reconciled with this smooth evolution? And how can we resolve the "problem
of outcomes" that appears to arise when a superposed quantum's (e.g. a photon or electron
Art Hobson Review and suggested resolution of Schrodinger's cat 2

or atom) state is measured by a "which-state" detector, creating a so-called entangled state

of the quantum and detector that appears to be an indefinite superposition of two
macroscopically distinct states of the composite system?
Such questions comprise the quantum measurement problem, surely quantum
physics' most enduring puzzle and, according to some, an unsolvable logical paradox.
Many alternative interpretations of the quantum physics mathematical formalism, and
several alternative modifications of the theory, have been proposed to resolve this
problem, with no consensus on a solution. It is remarkable that, despite the unparalleled
experimental success of quantum theory across a vast range of experiments, most of these
suggested solutions differ from standard quantum physics in one or more significant
respects. Most of them involve new interpretations of the standard mathematical
formalism, interpretations such as "human minds collapse the quantum state" or "all the
possible collapses occur but only one of them occurs in our particular universe" or even a
rejection of the physical reality of the quantum world and the assumption that quantum
probabilities (and hence changes in those probabilities, such as quantum state collapse) are
mere measures of personal degrees of belief. Other suggestions assume modifications of
the standard mathematical formalism, such as an additional mechanism that causes
quantum states to spontaneously collapse from time to time, or new "hidden" and hence
uncontrollable variables that create the illusion of quantum randomness. This paper
describes the measurement problem and suggests a resolution of the problem of definite
outcomes that lies entirely within standard quantum physics.
This paper does not entirely resolve the measurement problem. The measurement
problem comprises two more-or-less independent conundrums: the problem of definite
outcomes and the problem of irreversibility, i.e. the problem of making a macroscopic
record of one outcome. This paper suggests a resolution of the former but not the latter.
Let's begin with the question of definitions. What do we mean, physically, by a
quantum measurement? The great John Bell railed against the very use of this term. He
was the theorist who first gave us, in 1964, a practical mathematical condition ("Bell's
inequality") that a probabilistic theory must satisfy if it is to be considered "local," meaning
that the theory allows no unmediated or instantaneous physical action at a distance.
Furthermore, he showed that standard quantum physics fails this test [1]. This validated
the much earlier conclusion of Einstein and others that standard quantum physics makes
nonlocal predictions in certain specific physical situations [2]. Bell's last publication prior
to his untimely death in 1990 [3], provocatively titled "Against Measurement," urged that
this term "should now be banned altogether in quantum mechanics." Bell complained that
quantum physics concerns itself exclusively with "measurements" made in laboratories,
and that "to restrict quantum mechanics to be exclusively about piddling laboratory
operations is to betray the great enterprise. A serious formulation will not exclude the big
world outside the laboratory" [3].
I agree that we need a non-anthropomorphic definition of the concept known as
"measurement," but rather than ban this widely-used term, let's just broaden its physical
definition to include Bell's "big world" as follows: A "quantum measurement" means any
quantum process that results in a macroscopic effect, regardless of whether humans or
laboratories are involved. Thus not only is an electron striking a laboratory viewing screen
and creating a visible flash a measurement, a cosmic-ray muon striking and
Art Hobson Review and suggested resolution of Schrodinger's cat 3

macroscopically moving a sand grain on a planet in some other galaxy is also a

measurement.
To analyze measurement, let's look at a specific experiment: An electron beam
passes through a pair of double slits (two narrow closely-spaced parallel slits cut into a
partition) and then impacts a viewing screen. This might be physicists' most popular
experiment; it was described by Richard Feynman as "a phenomenon which is impossible ...
to explain in any classical way, and which has in it the heart of quantum mechanics" [4].
Just as in Thomas Young's similar double-slit experiment using light, performed in 1801,
the pattern formed on the viewing screen shows interference between the two portions of
the electron beam coming through the two slits: A broad dark-and-bright striped pattern
diffracts (spreads out) widely on the screen, much wider than the slits, indicating regions of
destructive (dark) and constructive (bright) interference [5]. On closer inspection, the
bright lines are formed by zillions (my word for a really large number) of tiny individual
electron impacts, each one making a small flash on the screen [6]. According to our above
definition, each flash is a measurement of the position of an electron as it hits the screen.
Experiments like this make us wonder whether electrons are tiny particles or waves
in a continuous spatially-extended field: We see particle-like behavior in the individual
flashes, and wave-like behavior in the interference pattern made by zillions of flashes. As I
have argued elsewhere [7], this and many other experiments are impossible to explain by
assuming electrons are small particles. Instead, each electron is a spatially extended bundle
of energy that comes through both slits and interferes with itself at the viewing screen.
Slowed-down experiments in which electrons come through the slits one at a time
demonstrate this [6]. Just prior to impact, each electron is extended over the entire extent
of the interference pattern. The interaction between each electron and the atoms of the
screen then collapses each electron to atomic dimensions. Although each individual
electron must ultimately be conceptualized as a wave in a universal "matter field" or "psi
field" (the official term is "electron-positron field"), the experiment displays both wave and
particle aspects (Chapter 5 of [8]). Similar effects occur in Young's interference experiment
with light, but with non-material photons replacing the material electrons.
Each electron's flash on the screen is a measurement. But for purposes of analysis,
it's better to consider a related example of a measurement, still based on the double-slit
experiment. Suppose an electron detector is installed at the slits, a detector that can detect
the electron's position as it passes through the slits while disturbing each electron only
minimally (in the precise sense described below).
Measurement, even by a minimally-disturbing "which-path detector," changes
everything. Exactly when the detector turns on, the pattern on the screen changes from the
striped interference pattern to a smoothly-spread-out sum of two single-slit patterns, each
showing diffraction but no interference. The interference pattern abruptly vanishes. So far
as I know, this experiment showing the jump from interference to non-interference has not
been performed with electrons but there is little doubt how it would turn out. The
analogous experiment has been done using light (photons) instead of electrons, and using
an interferometer rather than a double-slit interference setup. A which-path detector was
randomly switched on or off as each photon passed through this experiment; the photons
for which the detector was "off" formed an interference pattern while the photons for
which the detector was "on" formed the expected non-interference pattern [9]. In this so-
called "delayed-choice experiment," the collapses were instantaneous to within the
Art Hobson Review and suggested resolution of Schrodinger's cat 4

accuracy of the fast switching between the two states; each collapse was executed entirely
while the photon was inside the interferometer.
We can gain considerable insight by studying how quantum theory describes this
which-path measurement (pp. 63-5 of [10]). Note that this is in fact a measurement as
defined earlier, because the detector registers "slit 1" or "slit 2" macroscopically for each
electron. Using the formulation of quantum physics that describes states as vectors in a
mathematical Hilbert space, let's denote the state of one electron passing through slit 1 as
|ψ1> and the state of one electron passing through slit 2 as |ψ2>. John von Neumann, the
first to carefully analyze measurement in purely quantum-theoretical terms [11], insisted
on treating not only the measured quantum but also the macroscopic detector as a
quantum system because, after all, detectors are made of atoms and they perform a
quantum function by detecting individual quanta.
Accordingly, let's represent the "ready to detect" quantum state of the detector by
|ready>, and the state of the detector after detecting an electron by |1> if |ψ1> was
detected, and by |2> if |ψ2> was detected. A properly operating detector will surely
transition from |ready> to |1> upon measurement of an electron that has been prepared
(perhaps by simply shutting slit 2) in state |ψ1>. As a limiting idealization, let's assume,
with von Neumann, that measurement of an electron prepared in state |ψ1> leaves the
electron still in state |ψ1> after detection. Such a minimally-disturbing measurement
would cause the electron-plus-detector composite system, initially in the composite state
|ψ1>|ready>, to transition into the final state |ψ1>|1>. We can summarize this process as

|ψ1>|ready> è |ψ1>|1>. (1)

Similarly, the minimally-disturbing measurement of an electron initially prepared in |ψ2> is

described mathematically by

|ψ2>|ready> è |ψ2>|2>. (2)

Now suppose both slits are open so each electron can pass through either slit, and
suppose the preparation and the experiment (e.g. the slit widths) is symmetric with respect
to the two slits. Then the state of each electron as it approaches the slits prior to detection
must be described by the symmetric superposition

|ψ> = (|ψ1> + |ψ2>)/√2 (3)

where the 1/√2 factor is required for normalization. But quantum physics, including its
time dependence, is linear. Thus (1) and (2) imply that |ψ>|ready> evolves according to

|ψ> |ready> è (|ψ1>|1> + |ψ2>|2>)/√2, (4)

The final state

|Ψ> = (|ψ1>|1> + |ψ2>|2>)/√2 (5)

Art Hobson Review and suggested resolution of Schrodinger's cat 5

following the detection is said to be "entangled" because it cannot be factored into a simple
product of states of the two sub-systems. Quantum entanglement is a large complex topic
(Chapter 9 of [8]). As indicated schematically in Figure 1, when two independent quanta
pass near each other, interact, and subsequently separate, the interaction generally
entangles the two quanta and the entanglement then persists after the interaction
regardless of how far apart the two quanta might eventually travel, provided only that the
two quanta experience no further interactions. Despite their possibly wide spatial
separation, entangled quanta have a unity not possessed by non-entangled quanta. This
unity is the source of quantum non-locality, as I'll discuss later. Entanglement is ubiquitous
in nature; for example, the electrons in any many-electron atom or molecule are entangled
with each other [12]. Erwin Schrodinger has said that quantum entanglement is "the
characteristic trait of quantum mechanics" (the emphasis is Schrodinger's) [13].

Figure 1. When the green quantum on the left interacts with the orange quantum at the
bottom, their two spatially-extended quantum states entangle with each other to form a
single "bi-quantum" (the two green-and-orange quanta form a bi-quantum). This highly
unified composite state has non-local characteristics.

The entangled "measurement state" (5) at the heart of quantum measurement is

remarkably subtle. To grasp it, we first need to understand quantum superpositions. A key
quantum principle says that any linear combination of possible quantum states of a system,
as in (3) and (5) for example, is also a possible quantum state of that system. Figure 2
pictures an experiment that demonstrates such a superposition of states. It shows a layout
of optical paths called a "Mach-Zehnder interferometer." A light beam enters at the lower
left by passing through a "beam splitter" BS1; this is a small plate of glass (shown edge-on
in Figure 2) angled at 45 degrees so that the reflected beam makes a right angle with the
incoming direction while the transmitted beam passes straight through (with refraction at
the two surfaces). It's designed to reflect 50% and transmit 50% of the incident light. So
Art Hobson Review and suggested resolution of Schrodinger's cat 6

the beam splits and each half traverses one of the two paths; mirrors M bring the paths
back to a crossing point as shown. Devices called "phase shifters," denoted by φ1 and φ2,
are placed into each path. A phase shifter can add a short variable length to a path, perhaps
by using mirrors. A second beam splitter BS2 can be placed at the crossing point. Without
BS2, each half-beam moves straight ahead along one path to the detector on that path.
Things get more interesting with BS2 in place. Because 50% of each of the two
beams then goes to each detector, BS2 mixes the two beams together so they can show
interference. The interferometer is constructed so that, when the phase shifters are set to
zero, the two "optical paths" (the number of wavelengths, after accounting for phase
changes upon reflection and refraction) from the entry point to D1 are equal while the two
optical paths to D2 differ by half a wavelength. It is then found that the light interferes
constructively at D1 and destructively at D2, so all the light goes to D1. If we then use φ1 or
φ2 to add half a wavelength to either path, light then interferes constructively at D2 and
destructively at D1 so all the light goes to D2. And as we continuously vary the length of
one or the other path by varying one or the other phase shifter, we find the amount of light
arriving at D1 varies continuously from 100% down to 0%, while the amount arriving at D2
varies from 0% to 100%. The two paths are clearly interfering. This experiment is the
interferometer-based analog of Young's double-slit interference experiment demonstrating
the wave nature of light.

Figure 2. A Mach-Zehnder interferometer can demonstrate the interference of light when

both beam splitters are present and the phase shifters alter the length of either path. But
light is made of indivisible photons. What happens when only one photon is present?

But light is made of photons, and photons are indivisible. So how does nature solve
this problem when we dim the light to the point where only one photon at a time traverses
the interferometer? After all, the photon still traverses BS1, yet it cannot split in two
because a quantum is unified and can't be split. With BS2 removed, we find either D1 or D2
registers a single entire photon, randomly, i.e. with 50-50 probabilities, regardless of how
the phase shifters are set. Careful tests verify this randomness as absolute, i.e. more
random then any human macroscopic game, such as coin flips, that mimics randomness.
Nature invents quantum randomness in order to deal with obstacles such as beam splitters
while preserving the unity of the quantum (Chapter 6 of [8]). Detectors never register half
a photon. You get either a whole photon or no photon.
What happens in the single-photon experiment with BS2 present? Beginning from
path lengths yielding constructive interference at D1 and destructive interference at D2, as
the phase shifters vary the probabilities of detecting the photon at D1 and D2 vary as in
Art Hobson Review and suggested resolution of Schrodinger's cat 7

Figure 3, which shows the percentage of photons impacting D1. Importantly, these results
don't depend on which phase shifter the experimenter chooses to vary. Since each photon
responds to changes in either path length, each photon must follow both paths! This
verifies the superposition principle and shows that quanta can be in two places at the same
time. This is paradoxical if you assume photons are tiny particles, but if you assume
photons are waves in a universal field it's not paradoxical: Each photon simply spreads
along both paths, interfering with itself at D1 and D2 [7].
The interferometer experiment of Figure 2 can be performed with atoms and even
molecules, with the same results: Atoms and molecules can be superposed along two
paths, and can interfere with themselves just as photons can. So these objects are also
waves in fields, not tiny particles. Search on "atom interferometry" for more information.

Figure 3. Evidence of quantum superposition in the experiment of Figure 2 with BS2

inserted. Each photon must follow both path 1 and path 2 because these probabilities vary
no matter which phase shifter is varied.

Furthermore, we must conclude that each photon travels both paths even when BS2
is not present to directly verify this, because once a photon enters the interferometer it
must behave in the same manner regardless of whether BS2 is placed or not placed at the
far end. Jacque et al's delayed-choice experiment [9] referred to above provides further
evidence for this conclusion: Since these photons "do not know" whether BS2 will be
inserted, they must travel both paths on all the trials including those for which BS2 is not
inserted.
This is connected with entanglement. With BS2 removed, the situation is like the
double-slit experiment with a which-slit detector present: Each photon is entangled with
macroscopic detectors D1 or D2 as in the right side of (5). With BS2 present, the two paths
mix and the situation is like the double-slit experiment with no which-slit detector: each
photon follows two paths to each detector where it interferes with itself, and we detect the
interference state (3).
Art Hobson Review and suggested resolution of Schrodinger's cat 8

All of this suggests that measurements collapse superposed quantum states via
entanglement of the superposed quantum with a detector.

2. The apparent paradox of Schrodinger's cat

It's physicists' favorite tale. As Schrodinger told it [14]:

One can even set up quite ridiculous cases. A cat is penned up in a steel
chamber, along with the following device (which must be secured against
direct interference by the cat): In a Geiger counter there is a tiny bit of
radioactive substance, so small, that perhaps in the course of the hour one of
the atoms decays, but also, with equal probability, perhaps none; if it
happens, the counter tube discharges and through a relay releases a hammer
which shatters a small flask of hydrocyanic acid. If one has left this entire
system to itself for an hour, one would say that the cat still lives if meanwhile
no atom has decayed. The psi-function of the entire system would express
this by having in it the living and dead cat (pardon the expression) mixed or
smeared out in equal parts.
It is typical of these cases that an indeterminacy originally restricted
to the atomic domain becomes transformed into macroscopic indeterminacy,
which can then be resolved by direct observation. That prevents us from so
naively accepting as valid a "blurred model" for representing reality.

Mathematically, the nucleus and cat have become entangled in the measurement
state (5), with |ψ1> and |ψ2> representing the undecayed and decayed states of the nucleus
and |1> and |2> representing the alive and dead cat. According to Schrodinger's
understanding of the situation, the indeterminacy of the nuclear state "becomes
transformed into macroscopic indeterminacy" of the cat, and we cannot comfortably accept
this as a "blurred" state--a cat that is in a superposition of being both alive and dead. This
paper will show that, according to standard quantum physics, Schrodinger's 1937
understanding was incorrect: The composite system (cat-plus-nucleus) is not predicted to
be in a superposition of two states of the cat, or nucleus, or composite system. Instead, the
composite system is predicted to be in a superposition of two correlations between the cat
and nucleus, one in which a live cat is 100% correlated with an undecayed nucleus, and the
second in which a dead cat is 100% correlated with a decayed nucleus. Entanglement has
transformed a pure state superposition of nuclear states to a pure state superposition of
correlations between subsystem states. We will see that this is precisely what one expects,
and is not paradoxical.
The problem of definite outcomes applies of course to more than Schrodinger's
dramatized example. Regardless of whether the measuring instrument is a which-slit
detector, a Geiger counter, or a cat, the entangled state (5) applies. This state appears at
first glance to represent a quantum superposition in which the detector is in two
macroscopically different states simultaneously. If so, then there is an inconsistency within
quantum physics, because it obviously cannot be this easy to create a macroscopic
superposition.
Art Hobson Review and suggested resolution of Schrodinger's cat 9

Is it true that (5) really represents a macroscopic superposition? There is more to

this entangled state than meets the eye. If you assume the detector to be in a superposed
state a|1> + b|2> where a and b are complex constants, you soon find that (5) necessitates
either a=0 or b=0 [15], implying that the detector is not in an individually superposed state
within its own Hilbert space. The same applies to the detected quantum: It is not in a
superposed state a|ψ1> + b|ψ2> with both a≠0 and b≠0. The entanglement process leaves
neither sub-system superposed! So far as I know, this simple fact has long been ignored by
analysts of the measurement problem.
The density operator formalism for quantum physics provides a stronger version of
this conclusion. If you aren't familiar with density operators, you can find a
straightforward presentation in Section 2.4 of [10]. The density operator for a quantum
system whose state is |ψ> is simply the projection operator

ρ = |ψ> <ψ|. (6)

If a system is in a state whose density operator is ρ, then the standard quantum expectation
value <O> of an arbitrary observable O is found from

<O> = Tr (ρO) (7)

where "Tr" represents the trace operation (the sum of the diagonal elements). This
approach is especially useful if the quantum system is a composite of two subsystems A and
B. Define the density operator ρ A for subsystem A alone by

ρ A = TrB ρ (8)

where " TrB" means that the trace is taken only over the states of subsystem B. It is then
easy to show [10] that the standard quantum expectation values for subsystem A alone (the
values obtained by an observer of A) are

<OA> = Tr (ρ A OA ) (9)

where OA means any observable operating on system A alone (i.e. operating within A's
Hilbert space).
Applying this formulation to the measurement state (5), the reduced density
operators for the quantum system (call it A) and its detector (call it B), respectively, are

ρ A = (|ψ1> <ψ1| + |ψ2> <ψ2|)/2, (10)

ρ B = (|1> <1| + |2> <2|)/2. (11)

The plus signs in (10) and (11) make one think of superpositions such as (3), but these are
not superpositions. The density operator for (3) has cross-terms:

ρ = |ψ> <ψ| = (|ψ1> <ψ1| + |ψ1> <ψ2| + |ψ2> <ψ1| + |ψ2> <ψ2|)/2. (12)
Art Hobson Review and suggested resolution of Schrodinger's cat 10

The two cross-terms, involving both |ψ1> and |ψ2>, are missing from (10). So (10) does not
describe a system in a superposition of two quantum states. However, (10) is precisely the
density operator you should use if you know the quantum system is either in state |ψ1> or
in state |ψ2> but you didn't know which and so, due your own to lack of information, you
simply assign a probability of 1/2 to each of these two possibilities (Section 2.4 of [10]).
The same goes for (11). (10) and (11) are "classical" probabilistic states--analogous to the
"states of knowledge" you would assign to a coin flip when you know the outcome to be
either heads or tails with equal probability but you don't know which. The situation
described by a density operator such as (10) is known as a "mixture" of the states |ψ1> and
|ψ2>, as distinct from the "superposition" of states observed in the experiment of Figure 2
and represented by (3).
Equation (9) tells us that all the correct statistics for subsystem A alone can be
found from the standard formula (7) applied to subsystem A alone. But we have just seen
that (10) is the density operator one should use if one knows A to be in either |ψ1> or |ψ2>
without knowing which. The same goes for subsystem B and (11). In the case of
Schrodinger's cat, it follows that an observer of the cat alone sees outcomes appropriate to a
cat that is either alive or dead, not both. For subsystems, the interference terms are
missing, and an "ensemble" of repeated trials must exhibit a nonsuperposed mixture rather
than a superposion. This is the clear prediction of quantum physics for the entangled state
(5). Others have long come to the same conclusion (see pp. 183-185 of [16], also [17]).
But we must be careful, because (10) and (11) are not complete descriptions of the
quantum states of the nucleus or the cat. In fact, (10) and (11) are not quantum states at all
but merely "reduced states" arising from the actual state (5) of the composite system. In
the case of Schrodinger's cat, (10) and (11) give the correct predictions for observations of
either the nucleus alone or the cat alone, but they do not represent the state of either
subsystem because this is given by (5). In fact, when two quanta are entangled, neither one
has a quantum state of its own (Figure 1)!
Physicists, philosophers and mathematicians who specialize in quantum
foundations have in the past objected to the argument that reduced density operators can
be adduced in this manner to clarify the measurement problem. They offer two key
objections (Section 2.4 of [10]): The first, "basis ambiguity," charges that the "basis set"
(the set of orthogonal eigenvectors) for the operator (11) (for example) is entirely
ambiguous, so (11) cannot represent a true quantum state. It's true that (11) doesn't
represent the true state of a subsystem, because (11) is actually just the identity operator
|1> <1| + |2> <2| in B's subspace, divided by 2, so that any other orthogonal basis set could
be used instead. For example, given only the description (11), subsystem B could just as
well be described by any other pair of orthonormal vectors in B's subspace, for example
(|1> ± |2>)/√2. But B's state of affairs is certainly not entirely described by (11). Rather, it
is described by the composite state (5). Equation (11) merely tells us the following: If the
cat and nucleus are in the state (5) then, when one looks at the cat, one is going to see a cat
that is either alive or dead. There is no claim that (11) represents the complete quantum
state of the cat. That is, there is no claim that the cat is really in either the state |1> or the
state |2>, because the state it's really in is admittedly (5). Thus the basis ambiguity
Art Hobson Review and suggested resolution of Schrodinger's cat 11

objection to our conclusion (namely, that Schrodinger's cat is either alive or dead, not both)
fails.
The second key objection is that (10) and (11) are "improper density operators"
because they arise not from insufficient knowledge (as classical probabilities arise) but
from reductions of the full density operator (5) to the Hilbert subspaces of each subsystem.
It's true that these reduced density operators do not arise from insufficient knowledge
about the actual state. In fact we have complete knowledge of the state of both A and B,
namely the measurement state (5). So the objection fails not because it is false but because
it is irrelevant: The reduced operators admittedly do not represent the state of the
composite system. They tell us what we will observe at the nucleus and at the cat, but they
tell us nothing about the correlations between these observations, so these density
operators do not tell us the real state of the system.

3. The unity of the quantum

And so the plot thickens. The entangled state (5) properly describes both individual
subsystems. However, the plus sign in (5) signifies a superposition of the two terms, yet
we know that neither subsystem A nor subsystem B is superposed. What is the meaning of
this plus sign? This superposition arose from the superposition represented by (3). We
cannot logically ignore it--a strategy known as the "shut up and calculate" approach to
quantum measurement. Instead, we must ask: Exactly what is superposed when two
subsystems are in this entangled state?
Superpositions preserve the all-important unity of the quantum. When Max Planck
proposed in 1900 that electromagnetic radiation occurs in energy steps of magnitude E =
hf, he tacitly implied the central quantum principle: The unity of an individual quantum.
Energy (electromagnetic energy in the case of radiation) comes in spatially extended
bundles, each having a definite and identical quantity of energy. You can't have half a
quantum, or 2.7 quanta. You must have either 0 or 1 or 2 etc. quanta. In its own way, this
is a fairly natural notion--apparently nature prefers to sub-divide the universe into a
countable or even a finite set of entities as opposed to an uncountable continuum. The
spatial extension of these bundles then entails nonlocality: If you have one quantum and
you destroy it (by transforming it to something else), you must destroy all of it everywhere
simultaneously, because you can't at any time have just part of a quantum. Louis de Broglie
put it perfectly in 1924, regarding another kind of quantum namely the electron:

The energy of an electron is spread over all space with a strong

concentration in a very small region. ...That which makes an electron an
atom of energy is not its small volume that it occupies in space--I repeat it
occupies all space--but the fact that it is undividable, that it constitutes a unit.
[18]

When you transform the state of a quantum, you've got to transform the entire extended
quantum all at once. Hence there are quantum jumps. Furthermore, composite entangled
systems such as atoms also behave in a unified fashion. This unity is also the source of the
nonlocality seen in experiments involving entangled pairs of photons. Nonlocality is
Art Hobson Review and suggested resolution of Schrodinger's cat 12

exactly what one would expect, given the unity and spatial extension of the quantum and
the unitary (i.e. unity-preserving) nature of the entanglement process.
Standard nonrelativistic quantum theory prescribes two kinds of time evolution:
collapse upon measurement, and the Schrodinger equation between measurements. A key
feature of the Schrodinger equation is that it prescribes a so-called "unitary" time
evolution, meaning time evolution that preserves pure states, i.e. transforms unit Hilbert
space vectors into other unit vectors. Again, this is required physically by the unity of the
quantum: If a quantum is described by a pure quantum state at t=0, it should remain pure
at later times. This notion prompts us to ask whether the measurement process also
preserves pure states. At least in the case of the idealized process described in (4), the
answer is "yes" because both the "before" and "after" states are pure.
The measurement state (5), since it is pure, represents a highly unified state of
affairs, even though one of its subsystems is a macroscopic detector. Thus we suspect that
this state, like its progenitor (3), is truly a superposition in which the superposed terms
represent two situations or states of the same object. But precisely what is that object, i.e.
what is superposed? We have seen that states of subsystem A are not superposed, nor are
states of subsystem B. The conventional interpretation (which, as we will see, is subtly
incorrect) of a product state such as |ψ1>|1> is that it represents a state of a composite
system AB in which subsystem A is in state |ψ1> while B is in state |1>. In this case, (5)
would represent a superposition in which AB is simultaneously in the state |ψ1>|1> and
also in the state |ψ2>|2>. The situation of Schrodinger's cat would be: a live cat and
undecayed nucleus superposed with a dead cat and decayed nucleus. This is at least as
physically outrageous as a live cat superposed with a dead cat, and it contradicts the
physical implications (a cat that is either alive or dead) of the reduced states (10) and (11)
as described in Section 2. Something is wrong.
The remainder of this paper will demonstrate that, according to standard quantum
theory (and of course according to experiment), the measurement state (5) represents
none of these paradoxical situations.

4. Experimental nonlocality and entanglement

The unity of the quantum suggests that the measurement state (5) represents a unified,
hence superposed and pure, quantum state of the composite system. But precisely what is
superposed? We studied the simple (i.e. non-composite) superposition (3) via the
interference exhibited in the experiment of Figure 2. Varying the length of either path 1 or
path 2 created varying interference effects in the detectors, demonstrating each photon
really must travel both paths to its detector. Quantum theory agrees entirely with these
conclusions, as can be shown by using photon wavelengths to show that the path
differences correctly predict the interferences observed at each detector.
This suggests that, to understand the measurement state, we need to find and
analyze entanglement experiments that demonstrate interference. As it happens, this has
been done for several decades in connection with quantum nonlocality. The key theoretical
analysis was done by John Bell [1]. Many nonlocal interference experiments have been
done beginning with Clauser and Freedman [19], continuing with the definitive experiment
of Aspect et al [20] and other experiments such as the two described below, culminating in
Art Hobson Review and suggested resolution of Schrodinger's cat 13

experiments demonstrating nonlocality across great distances [21] and that

simultaneously closed all possible loopholes in all the previous experiments [22] [23] [24].
By now, it is well known that the entangled state (5) predicts nonlocal effects between its
two subsystems, and that phase variations of either subsystem cause instantaneous, i.e.
non-local, readjustments of the possibly-distant other subsystem.
But it's not easy to vary the phase of a cat, and as we saw in the experiment of Figure
2, one cannot understand a superposition without varying the phases of its superposed
parts. These nonlocality experiments are carried out with pairs of simpler quanta such as
photons. The nonlocal entanglement experiments most appropriate for investigating
measurement were conducted nearly simultaneously by Rarity and Tapster [25] and Ou,
Zou, Wang, and Mandel [26]. Figure 4 shows the layout for these "RTO" (for Rarity,
Tapster, and Ou) experiments. The "source" in Figure 4 creates entangled photon pairs by
"parametric down- conversion," a process which needn't concern us here.

Figure 4. The RTO experiments provide an ideal portrait of entanglement. In each trial, the
source emits two entangled photons A and B into a superposition of the solid and dashed
paths to create an entangled state.

Compare this layout with Figure 2. The RTO experiment is two back-to-back
interferometer experiments but with the first beam splitter for each photon located inside
the source of entangled photons. Without entanglement, each single photon (either A or B)
would interfere with itself at its own detectors according to its own phase shift φA or φB.
The two entangled photons are emitted into a superposition of the solid path connecting
detectors A1 and B1, and the dashed path connecting detectors A2 and B2. Note that the
two photons are already entangled when they are emitted.
The entanglement changes everything. No longer does either photon interfere with
itself at its own detectors. Instead, the photons are entangled in the measurement state (5)
with |ψ1> and |ψ2> representing (say) the solid-line and dashed-line states of A and |1> and
|2> representing the solid-line and dashed-line states of B, although in the RTO
experiments neither subsystem is macroscopic. Each photon now acts like a which-path
detector for the other photon. Recall the double-slit experiment: When a which-slit
detector is switched on, the pattern on the screen switches abruptly from the striped
interference pattern indicating the pure state nature of each electron across both slits, to a
phase-independent sum of two non-interfering single-slit patterns. The entanglement
between the electron and the which-slit detector breaks the pure state into two single-slit
Art Hobson Review and suggested resolution of Schrodinger's cat 14

parts, so that the measured electron comes through either slit 1 or slit 2. This suggests
that in the RTO experiment, the entanglement should break the pure-state superposition
(3) into two non-interfering parts.
This is exactly what is observed. Both photons impact their detectors as random 50-
50 mixtures, just like a flipped coin. The entanglement breaks the single-photon pure state
(3) observed in the experiment of Figure 2, causing each photon to behave "incoherently"
with no dependence on its phase setting.
But (5) is a pure state. Where has the phase dependence gone? The answer lies in
the phase-dependent but nonlocal relationship observed between Figure 4's solid and
dashed branches. This phase dependence is observed experimentally in coincidence (or
correlation) measurements comparing detections of entangled pairs. The flipped coins
mentioned above turn out to be correlated with each other. This phase dependence across
the two separated subsystems is essential to preserve the unity of the (now entangled)
quantum.
This is not an easy experiment to perform: The source creates a stream of photon
pairs, and one must compare the impact of a single photon A at detectors A1, A2 with the
impact of its corresponding entangled photon B at detectors B1, B2. RTO figured out how to
do this, with the result shown in Figure 5.

Figure 5. Nonlocal interference in the RTO experiments. As the nonlocal phase difference
φB - φA varies, the "degree of correlation" (see text for precise definition) between A and B
shows phase-dependent interference.

Figure 5 graphs the degree of correlation between A and B. This is a measure of the
agreement between the outcomes at A's detectors and B's detectors. A correlation of +1
means perfect, or 100%, agreement: Either both sets of detectors register outcome 1 (i.e.
A1 and B1 click) or both register outcome 2, as in a measurement: You want the which-slit
detector to register "slit 1" if the electron is in state |ψ1>, you want it to register "slit 2" if
the electron is in state |ψ2>, and you want this agreement on every trial. The opposite
extreme is a correlation of -1, meaning 100% disagreement: If one detector registers 1, the
other registers 2. Either correlation, +1 or -1, implies that either photon's outcome is
Art Hobson Review and suggested resolution of Schrodinger's cat 15

predictable from the other photon's outcome. A correlation of zero means one photon's
outcome does not at all determine the other's outcome: Each photon has a random 50-50
chance of either outcome regardless of the other photon. Correlations between 0 and +1
mean the outcomes are more likely to agree than to disagree, with larger correlations
denoting a higher probability of agreement; for example, a correlation of +0.5 means a 75%
probability of agreement. Similarly, correlations between 0 and -1 mean the outcomes are
more likely to disagree than to agree; a correlation of -0.5 means a 75% probability of
disagreement.
The RTO experiment agrees entirely with the predictions of standard quantum
physics. When an accounting is made of the optical paths for both photons, we obtain the
following result [27]:

P(correlated) = P(A1 and B1) + P(A2 and B2) = 1/2[1 + cos(φB - φA)] (13)

P(anticorrelated) = P(A1 and B2) + P(A2 and B1) = 1/2[1 - cos(φB - φA)] (14)

where P(correlated) is the single-trial probability that A's and B's detectors will agree, and
P(anticorrelated) is the single-trial probability that A's and B's detectors will disagree. The
degree of correlation, defined as P(correlated) - P(anticorrelated), is then simply cos(φB -
φA), as graphed in Figure 5.
In 1964, John Bell published a ground-breaking article stating a sufficient condition
for a statistical theory such as quantum physics to meet the condition known as "locality."
He defined locality to mean "that the result of a measurement on one system be unaffected
by operations on a distant system with which it has interacted in the past" [1]. Bell
expressed this sufficient condition in the form of an inequality (Eq. (15) of [1]) that any
local theory must obey. He also demonstrated that certain statistical predictions of
quantum physics violate Bell's inequality, i.e. quantum physics makes nonlocal predictions.
The results in Figure 5 turn out to be a case in point: Figure 5 violates Bell's inequality at
all phase differences φB - φA other than 0, π, and 2π. Let me underline the meaning of this:
The violation of Bell's inequality means that the statistics of measurements on photon A--
photon A's "statistical behavior"--is necessarily affected by the setting of photon B's phase
shifter.
In fact, even without Bell's condition, the nonlocality of this experiment is intuitively
obvious. Here's why: Suppose we set the phase shifters to zero and that all four optical
paths (two solid, two dashed) in Figure 4 are then equal; thus φB - φA is zero. Without the
two beam splitters BS, the two photons emitted into the solid pair and dashed pairs of
paths would impact either detectors A1 and B1 or A2 and B2 because of the symmetry of
the experiment and conservation of momentum. This is neither surprising nor nonlocal,
and would happen even if the photons were not entangled. But a beam splitter is a
randomizing device that mixes the solid and dashed paths; any photon passing through it
has a 50-50 chance of reflection or transmission. With non-entangled photons and both
beams splitters in place, there would then be no correlation between photon A's outcome
and B's outcome because the two photons are independent of each other. With
entanglement, the correlation is perfect. How does one photon know which path the other
photon took at the other photon's beam splitter? Each photon is now "detecting" the
quantum state of the other photon, from a distance that could be large. The perfect
Art Hobson Review and suggested resolution of Schrodinger's cat 16

correlation certainly "feels" nonlocal even though (as mentioned above) this perfect
correlation at φB - φA = 0 does not violate Bell's inequality. Note that such a violation is a
sufficient but not necessary condition for nonlocality.
Non-locality is written all over the RTO experiment. Each photon "knows" which
direction the other photon takes at its beam splitter and adjusts its selection accordingly.
The key nonlocal feature of Figure 5 is that the graph, which is simply a cosine
function, has (φB - φA) as its independent variable. Thus any desired shift in correlations
can be made by an observer at either of the possibly-widely-separated phase shifters. Bell
suspected that this situation meant that observer A (call her Alice) could use her phase
shifter to alter the outcomes that would have occurred at both her own and observer B's
(call him Bob) detector and, following up on this hypothesis, derived his inequality
involving the probabilities at both Alice's and Bob's detectors which, if violated, implied
that both photons must have readjusted their states. Such readjustment is just what we
expect, given the unity of the quantum and thus the unity of atoms and other entangled
systems such as our two photons. The two photons form a single "bi-quantum," an "atom
of light," in the pure state (5). When Alice varies her phase shifter, both photons "know"
both path lengths and readjust their behavior accordingly to produce the proper
correlations. Analogously, a single photon "knows" both path lengths in the single-photon
interferometer experiment of Figure 2.
Finally, the central question of this paper: What is actually superposed in the
entangled superposition (5)? The experiment of Figure 2 tests the simple superposition
(3), while the RTO experiment tests the entangled superposition (5). We know what is
superposed in Figure 2, namely the quantum states |ψ1> (path 1) and |ψ2> (path 2). This is
deduced from the effect that either phase shifter has on both states. Now consider the RTO
experiment. What is the effect of shifting either phase shifter? One thing that does not
change is the state ("local state" would be a better term, as discussed in Section 5) of either
photon A or photon B: As we know, both photons remain in 50-50 mixtures regardless of
either phase setting. What does change with variations in either phase shifter is the
correlations between A and B. With φA = φB = 0 we have perfect correlation: Either A1 and
B1 (which we will denote (11)) or A2 and B2 (denoted (22)). As we vary either φA or φB we
obtain non-zero probabilities of anti-correlated individual trials, denoted (12) (outcomes
A1 and B2) and (21) (A2 and B1). When the non-local phase angle difference (φB - φA)
reaches π/2, we have zero correlation, and when it reaches π we have perfect anti-
correlation.
Table 1 summarizes this crucial point in more detail. The column titled "simple
superposition" shows how the superposition state of a single photon (Figure 2) varies from
"100% state 1" to "100% state 2" as the phase angle between the two states varies. The
column titled "entangled superposition of two sub-systems" shows that the state of each
photon remains unchanged throughout the entire range of both phase settings, while the
nonlocal correlation between the states of the two photons varies from "100% correlated" to
"zero correlation" and then to "100% anticorrelated" as either of the two local phase angles
varies.
What is superposed in the RTO experiment? The hallmark of a superposition is its
dependence on the phase difference between the objects that are superposed. But Table 1
exhibits no such phase dependence of the states of the two photons. Each photon remains
in an unchanging 50-50 mixture of its own "path 1" and "path 2" states, a situation that is
Art Hobson Review and suggested resolution of Schrodinger's cat 17

radically at odds with the true superposition of path 1 and path 2 exhibited by the
experiment of Figure 2. Thus in the entangled RTO state, neither photon is superposed.
We see here the source of the "classical" or non-superposed nature of the reduced density
operators (Eqs. (10) and (11)), not to mention the non-superposed and hence non-
paradoxical nature of Schrodinger's cat. Our examination of the phase-dependence of the
measurement state (5), as demonstrated by nonlocality experiments such as the RTO
experiment, reveals the true nature of Schrodinger's cat. The last column of Table 1 shows
us what actually is superposed when two subsystems are entangled in the measurement
state (5). Since the correlations between the two photons vary sinusoidally as the non-
local phase angle between the two photons varies, it is clearly these correlations between
the states of the two photons, and not the states themselves, that are interfering. The
entanglement has shifted the superposition, from the states of one photon A (Eq. (3), Figure
2) to the correlations between photon A and photon B (Eq. (5), Figure 4).

Simple superposition: Entangled superposition of two sub-systems:

φ State of photon φB -φA State of each photon Correlation between
the two photons
0 100% “1”, 0% “2” 0 50-50 “1” or “2” 100% corr, 0% anti
π/4 71% “1”, 29% “2” π/4 50-50 “1” or “2” 71% corr, 29% anti
π/2 50% “1”, 50% “2” π/2 50-50 “1” or “2” 50% corr, 50% anti
3π/4 29% “1”, 71% “2” 3π/4 50-50 “1” or “2” 29% corr, 71% anti
π 0% “1”, 100% “2” π 50-50 “1” or “2” 0% corr, 100% anti
Table 1. In a simple superposition, the photon's state varies with phase angle. In an
entangled superposition, the relationship between states of the two photons varies, while
individual states of both photons are phase-independent (or "mixed").

5. Summary and discussion

In order to resolve the problem of definite outcomes of measurements, aka Schrodinger's

cat, this paper analyzes the entangled state (5) of a microscopic quantum and its
macroscopic measuring apparatus. This state is a superposition of the two composite
entities |ψ1>|1> and |ψ2>|2>, with a phase angle between these entities that can range
over 2π radians. In a measurement, this phase angle is fixed at zero because we design the
detector so that the two basis states of the measured quantum are 100% positively
correlated with the basis states of the measurement apparatus.
To resolve the problem of definite outcomes we must ask: Precisely what does the
composite superposition (5) actually superpose, physically? In order to understand the
simple non-composite superposition (3), we looked at the effect of varying the phase angle
between the superposed entities |ψ1> and |ψ2> in an experimental setting such as the
interferometer of Figure 2. The theoretically predicted and experimentally observed
results then made it obvious that the quantum whose state is (3) flows simultaneously
along two separate paths described by |ψ1> and |ψ2>.
To understand the superposition (5), we should proceed similarly by studying
situations in which the phase angle between the superposed entities |ψ1>|1> and |ψ2>|2>
Art Hobson Review and suggested resolution of Schrodinger's cat 18

varies. As it happens, theorists and experimentalists studying the phenomenon of

nonlocality have been doing this for decades, but quantum foundations specialists have not
particularly noticed this work in connection with the measurement problem. In fact,
nonlocal aspects of the state (5) have been studied since Bell's 1964 theoretical paper [1]
and Clauser and Freedman's 1972 experiment [19]. The 1990 experiments of Rarity and
Tapster [25], and of Ou, Zou, Wang, and Mandel [26], furnish the ideal vehicle for such an
analysis and are the central feature of this paper.
One lesson of this analysis is that, in order to understand the measurement problem,
one must understand the significance of nonlocality. This is because the key measurement
state (5) that caused Schrodinger and decades of experts so much concern has nonlocal
characteristics. It must be understood as a superposition of correlations, rather than a
superposition of states, but this cannot become apparent until one considers the effect of
variations in the phase angle between its superposed terms. Such variations are not part
of the measurement process itself because measurements are designed to take place at
zero phase angle. Experimental or theoretical studies of such phase variations will have
nonlocal ramifications, because such variations are inherently nonlocal. This situation
would have prevented Schrodinger in 1935, or indeed anyone prior to Bell's 1964 paper
and the experimental confirmations of the reality of nonlocality beginning in 1972, from
understanding the entangled superposition (5).
It's worth emphasizing that, when two subsystems are entangled in the
measurement state (5), neither subsystem is superposed. Only correlations between the
subsystems are superposed. In the RTO experiments, the two correlations in question are
represented by the solid and dashed paths connecting pairs of outcomes. A pair of photons
entangled in the state (5) follows both of these paths simultaneously. The subsystems
themselves, however, are not in superpositions but are instead in indeterminate mixtures
of definite states. Thus observers of either subsystem will observe only definite outcomes,
as predicted by the local mixtures (10) and (11).
The RTO experiments are the entangled analog of the interferometer experiment of
Figure 2: a pair of back-to-back interferometer experiments, with an entangled pair of
quanta of which one quantum passes through each interferometer. The experiment and its
theoretical analysis shows that, when a superposed photon A becomes entangled with a
second photon B to form the state (5), the nonlocal aspect of A's superposition (Figure 2) is
transferred to the correlations between A and B (Figure 4). Thus an entangled state such
as (5) is neither a superposition of states of A nor of states of B, but instead a superposition
of the correlations between the states of A and the states of B.
To see this most clearly, let's compare the simple superposition (3) with the
entangled superposition (5). In the simple superposition, the state observed by a "which-
state" detector varies smoothly from 100% |ψ1>, through 50% |ψ1> and 50% |ψ2>, and
finally to 100% |ψ2> as the phase angle φ between |ψ1> and |ψ2> varies from 0 to π. In the
entangled superposition, neither the state of A nor the state of B varies as φA or φB varies;
both A and B remain in 50-50 mixtures throughout. What does vary is the correlation
between A and B. A non-local "correlation detector" (i.e. an RTO-type of experiment!)
would find the relation between the two subsystems varies from 100% positively
correlated (either the pair state 11 or 22, pictured by the solid and dashed paths in Figure
4), to 50% positively correlated and 50% anti-correlated, and finally to 100% anti-
Art Hobson Review and suggested resolution of Schrodinger's cat 19

correlated (12 or 21), as the nonlocal phase difference φB -φA varies from 0 to π. This is a
superposition of correlations, not a superposition of composite states or of non-composite
(single-system) states.
This conclusion implies that our standard physical description of a composite non-
entangled (i.e. factorable) product state such as |ψ1>|1> has been long mistaken. We
usually regard |ψ1>|1> as a state of the composite system AB, one in which subsystem A is
in state |ψ1> and subsystem B is in state |1>. But this leads us into the paradox of
Schrodinger's cat, where (|ψ1>|1> + |ψ2>|2>)/√2 represents a state in which two
macroscopically different composite states exist simultaneously as a superposition.
According to the present study, quantum theory and quantum experiments imply this
entangled state to be a superposition of correlations between states rather than a
superposition of composite states. Thus |ψ1>|1> is not a state of the composite system, but
instead a correlation between the two subsystems. That is, |ψ1>|1> means "subsystem A is
in the state |ψ1> if and only if subsystem B is in the state |1>," an important departure from
the usual description.
Even if one of the two subsystems happens to be a macroscopic detector, the
entangled state (5) is simply a non-paradoxical superposition of correlations. It says
merely that the state |ψ1> of A is correlated with the state |1> of B, and the state |ψ2> of A is
correlated with the state |2> of B, with the non-local phase angle φB -φA determining the
degree of each correlation. In the case of measurement, this phase angle is fixed at zero.
Regardless of phase angle, neither subsystem is in a superposition. The entangled
measurement state (5) is best described as a "macroscopic correlation": a pair of
superposed (i.e. phase-dependent) quantum correlations in which one subsystem happens
to be macroscopic. It is technically very difficult to create a macroscopic superposition,
but macroscopic which-path detectors routinely achieve the state (5). It's not paradoxical,
even though many analyses have puzzled over it.
At least in our idealized case of a minimally-disturbing von Neumann measurement,
the initial stage of the measurement process (through the formation of the measurement
state (5)) can be described as follows: A quantum in a simple superposition such as (3)
entangles with a macroscopic which-path detector. At the instant of entanglement, the
local states of the both the quantum and the detector undergo a radical change, a quantum
jump. Locally, the detector and the quantum jump into mixtures (10) and (11).
Simultaneously, the global state (5) continues evolving smoothly according to the
Schrodinger equation. Entanglement causes the superposed single quantum to be
instantly transformed into superposed correlations between the quantum and the detector.
This stage of the measurement process is entirely describable in terms of pure
global states following the Schrodinger equation. The collapse from a local superposition
to local mixtures occurs because of the formation of the entangled state (5) and the
resulting formation of subsystems whose local states (Eqs. (10) and (11)) have definite
outcomes. Note that the phenomenon of nonlocality is essential to preserving the pure-
state nature (the unity) of the composite system. To put this more intuitively, a re-
organization throughout the entire extent of the composite entangled system is required in
order to preserve the unity of the (now entangled) quantum.
According to Table 1, when two systems entangle to form the state (5), both collapse
into phase-independent local mixtures. Relativity requires this phase independence: If any
Art Hobson Review and suggested resolution of Schrodinger's cat 20

phase-dependent aspect of the entangled state were locally observable, instant

information-containing messages could be sent, violating special relativity. Local states of
entangled subsystems must be invariant to phase changes. Thus, only the relationship--the
correlations--between A and B, but not A or B themselves, can vary with phase angle. Since
local observers cannot detect these correlations, the entangled state cannot be used to send
superluminal signals. This is, ultimately, the reason Schrodinger's cat must be either alive
or dead rather than a superposition of both. A phase-dependent superposition involving
both local states would permit nonlocal signaling, violating relativity.
In entanglement, nature employs an ingenious tactic. She must not violate
relativistic causality, yet she must be nonlocal in order to maintain the pure-state nature of
the original single-quantum superposition over composite objects such as bi-photons.
Thus she accomplishes nonlocality entirely via the superposition of correlations, because
correlations cannot be locally detected and thus their superposition cannot violate
relativity. This tactic lies behind the nonlocal spread of phase-dependence over large
spatial distances. By means of the superposition of correlations--entanglement-- nature
creates a phase-dependent pure-state quantum structure across extended quantum
systems such as bi-photons.
I've frequently used the term "local" as contrasted with "global." For composite
systems, and especially the entangled measurement state, it's a crucial distinction.
Entangled states such as (5) have distinct local and global (nonlocal) aspects. The local
description means the situation observed by two (or N for an entangled N-body system)
observers, each observing only one subsystem. In the case of (5), this "local description" is
fully captured by the reduced density operators (10) and (11)--each local observer detects
a mixture, not a superposition, of one subsystem. The "global description" means the
evolving pure state of the entire composite system, in our case Eq. (5). It is a superposition
of nonlocal correlations that can only be detected by observing both subsystems and, via an
ensemble of trials that individually record corresponding outcomes at both subsystems,
determining the state of the correlations between them. Although the global state implies
the local description, the local description cannot hint at the global correlations because
any such hint would violate Einstein causality. Thus, when an electron shows up in your
lab, neither an examination of the electron nor an examination of an ensemble of
identically-created electrons can give you the least hint of whether or how this electron is
entangled with other quanta elsewhere in the universe.
This clarification of entanglement resolves the problem of definite outcomes, aka
Schrodinger's cat. An ideal measurement of a superposed microscopic system A by a
macroscopic detector B establishes the measurement state (5) at 100% positive
correlation. This state is equivalent to the logical conjunction "A is in local state |ψ1> if and
only if B is in local state |1>, AND A is in local state |ψ2> if and only if B is in local state |2>,"
where AND indicates the superposition. This conjunction is precisely what we want
following a measurement. Schrodinger's cat is not in the least paradoxical.
This analysis does not entirely resolve the quantum measurement problem. It
resolves the problem of definite outcomes associated with the measurement state (5), but
this state continues to obey Schrodinger's equation and is hence reversible. In fact, the
entangled state between a quantum and its which-path detector can actually be reversed in
the Stern-Gerlach experiment (Figure 11.1 of [8]). A quantum measurement must result in
Art Hobson Review and suggested resolution of Schrodinger's cat 21

a macroscopic indication such as a recorded mark, and a mark is irreversible. The above
analysis shows the entangled state (5) describes a mixture of definite, not superposed,
outcomes of measurements, but these outcomes remain indeterminate and the global state
remains reversible.
The irreversibility problem is the question of how this nonlocal superposition of
correlations then further collapses irreversibly to just one of its possible outcomes, a
collapse that occurs in the RTO experiment only when one photon impacts a detector. The
present analysis does not claim to resolve this problem. In the case of the RTO experiment,
however, it seems fairly clear that the nonlocal superposition described by Eq. (5) must
irreversibly decohere [10] when either of its subsystems A or B interacts with a detector.
The RTO experiment furnishes a particularly good setting for this question, because the
two photons remain in the reversible entangled state (5) throughout their flights from the
source to detectors, and thus the two key questions of the measurement problem (the
problem of definite outcomes and the problem of irreversibility) can be analyzed
individually.

Acknowledgements

I thank University of Arkansas Physics Professor Surendra Singh and Philosophy Professor
Barry Ward for many enlightening discussions, the Department of Physics for providing
this retired professor with a congenial office, James Malley of the Center for Information
Technology at the National Institutes of Health as well as Mario Bunge of the McGill
University Philosophy Faculty for advice, retired physics professor Ulrich Harms of
Reutlingen, Germany, for discussions and support, and numerous referees for helpful
criticism. I especially thank my friend Peter Milonni, Professor of Physics, Department of
Physics and Astronomy, University of Rochester, for his help and advice.

Notes on contributor

Art Hobson, since retiring in 1999 from teaching physics at the

University of Arkansas, Fayetteville, has been able to pursue his long-time passion of trying
to sort out the widespread confusion about what quantum physics actually means. This
paper, along with his book Tales of the Quantum (Oxford University Press, 2017) and other
publications listed in this paper's references, reflect what he learned during this fascinating
and continuing adventure.

References
Art Hobson Review and suggested resolution of Schrodinger's cat 22

[1] J.S. Bell, On the Einstein-Podolsky-Rosen paradox, Physics 1 (1964), pp. 195-200.
[2] A. Einstein, B. Podolsky, and N. Rosen, Can quantum mechanical description of physical
reality be considered complete?" Phys. Rev. 47 (1935), pp. 777-780.
[3] J. S. Bell, Against 'measurement, Physics World 3 (1990), pp. 33-41.
[4] R. Feynman, The Character of Physical Law, MIT Press, Cambridge, MA, 1965.
[5] C. Johnson, Electron Diffraction at Multiple Slits, Am. J. Phys. 42 (1974), pp. 4-11;
translated from the original 1961 German publication.
[6] A. Tonomura, J. Endo, T. Matsuda, T. Kawasaki, and H. Exawa, Demonstration of single-
electron buildup of an interference pattern, Am. J. Phys. 57 (1989), pp. 117-120.
[7] A. Hobson, There are no particles, there are only fields, Am. J. Phys. 81 (2013), pp. 211-
223.
[8] A. Hobson, Tales of the Quantum: Understanding Physics' Most Fundamental Theory,
Oxford University Press, New York, 2017.
[9] V. Jacques, E. Wu, F. Grosshans, F. Treusart, P. Grangier, A. Aspect, J.-F. Roch,
Experimental realization of Wheeler's delayed choice gedanken experiment, Science 315
(2007), 966-968.
[10] M. Schlosshauer, Decoherence and the Quantum-to Classical Transition, Springer-
Verlag, Berlin, 2007.
[11] J. von Neumann, Mathematical Foundations of Quantum Mechanics, Princeton University
Press, Princeton, NJ, (1955); translated by R. T. Beyer from the original 1932 German
publication.
[12] J. Slater, Quantum theory of matter, McGraw-Hill Book Co., New York, 1951.
[13] E. Schrodinger, Discussion of probability relations between separated systems,
Mathematical Proc. of the Cambridge Phil. Soc. 31 (1935), 555-563.
[14] E. Schrodinger, The present situation in quantum mechanics: a translation of Schrodinger's
"cat paradox" paper, by J. Trimmer, Proc. Am. Phil. Soc., 124 (1937), pp. 323-338.
[15] A. Hobson, Two-photon interferometry and quantum state collapse, Phys. Rev. A 88
(2013), 022105.
[16] J.M. Jauch, Foundations of Quantum Mechanics, Addison-Wesley, Reading, MA, 1968.
[17] S. Rinner, and E. Werner, On the role of entanglement in Schrodinger's cat paradox,
Central European J. of Phys. 6 (2008), pp. 178-183.
[18] L. de Broglie, Recherches sur la Theorie des Quanta, PhD dissertation, Paris University,
1924; English translation quoted in J. Baggott, The Quantum Story, Oxford University Press,
Oxford, 2100, p. 38.
[19] J. F. Clauser, and S. J. Freedman, Experimental test of local hidden-variables theories,
Phys. Rev. Letts. 26, (1972) pp. 938-941.
[20] A. Aspect, J. Dalibard, and G. Roger, (1982), Experimental test of Bell's inequalities using
time-varying analyzers, Phys. Rev. Letts. 49 (1982), 1804-1807.
[21] J.-W. Pan et al, Satellite-based entanglement distribution over 1200 kilometers, Science 356
(2017), pp. 1140-1144.
[22] B. Hensen et al., Loophole-free Bell inequality violation using electron spins separated by
1.3 kilometers, Nature 526 (2015), pp. 682-686.
[23] M. Giustina et al., Significant-loophole-free test of Bell's theorem with entangled photons,
Phys. Rev. Lett. 115 (2015), 250401.
 Art Hobson Review and suggested resolution of Schrodinger's cat 23

[24] L.K. Shalm et al., Strong loophole-free test of local realism, Phys. Rev. Letts. 115 (2015),
250402.
[25] J.G. Rarity and P.R. Tapster, Experimental violation of Bell's inequality based on phase and
momentum, Phys. Rev. Letts. 64 (1990), pp. 2495-2498.
[26] Z.Y. Ou, X.Y. Zou, L.J. Wang, and L. Mandel, (1990), Observation of nonlocal interference
in separated photon channels, Phys. Rev. Letts. 66 (1990), pp. 321-324.
[27] M.A. Horne, A. Shimony, and A. Zeilinger, Introduction to two-particle interferometry, in
Sixty-Two Years of Uncertainty, A. I. Miller, ed. Plenum Press, New York, 1990, pp. 113-
119.

View publication stats