What are Physical States?

Stephen Boughn⋇

Department of Physics, Princeton University, Princeton NJ

Departments of Physics and Astronomy, Haverford College, Haverford PA

Abstract
The concept of the physical state of a system is ubiquitous in physics
but is usually presented in terms of specific cases. For example, the
state of a point particle of mass m is completely characterized by its
position and momentum. There is a tendency to consider such states as
“real”, i.e., as physical properties of a system. This rarely causes
problems in classical physics but the notion of real quantum states has
contributed mightily to the philosophical conundrums associated with
quantum mechanics. The Einstein-Rosen-Podolsky paradox is a prime
example. In fact, quantum states are not physical properties of a system
but rather subjective descriptions that depend on the information
available to a particular observer. This realization goes a long way
toward resolving such dilemmas as Schrödinger’s cat, wave function
collapse, quantum non-locality, and parallel universes.

Introduction
The notion of the physical state of a system is a crucial concept in physics as well
as in other disciplines. For macroscopic objects, quantities like temperature, pressure,
composition, etc., are sufficient to characterize a state. Of course, such descriptions leave
out a great deal of information about a system, in particular, the details of the
microscopic constituents of the system. However, if a system is comprised of a single or
of a few particles, both classical and quantum mechanics assume that a proper physical
state includes all possible information about the system. In the case of a quantum system,
Einstein famously concluded that the quantum state (quantum wave function) “does not
provide a complete description of the physical reality”. Bohr, and probably many other
physicists, would disagree. The following is my pragmatic take on the completeness of
such physical states.

† This essay should be considered a companion to “Against ‘Reality’ in Physics” (Boughn 2019)
⋇ sboughn@haverford.edu
 2

Physical States as Complete Descriptions

For a specific observer, the physical state of a system, whether it be a classical or
a quantum system, is the complete description of the system consistent with the
information about the system possessed by the observer. By “complete” I mean that from
such a description, the observer can predict the results of all possible observations of the
system, whether they be precise or statistical in nature. Furthermore, the dynamical
theory of the system can be used to compute the time evolution of the physical state and,
therefore, to predict the results of all possible future observations. Note that this
definition of physical state is specific to a particular observer. A different observer with
different knowledge of the system might well assign a different physical state to the
system and that state can also be considered complete provided that it can be used to
predict the results of all possible observations of the system by that observer. Therefore,
the physical state is not a physical property of a system but rather is an observer
dependent yet complete description of that system.
In classical mechanics, one might insist that there is a true physical state of a
system but some observers might simply lack the information necessary to identify
the true state and, therefore, can only specify an approximate state based on the
information available to them. Classical mechanics, even statistical mechanics, is
amenable to the notion of hidden variables, and so permits this interpretation. On
the other hand, such a declaration of “truth” is metaphysical in the sense that it adds
nothing to the utility of the concept of physical state delineated in the previous
paragraph. If the specified state enables the observer to make accurate predictions
(whether exact or statistical) of all future measurements, then that state satisfies the
criterion of being a complete description. On the other hand, an “approximate state”
might imply predicted results that are inconsistent with what is actually observed.
For example, if the approximate state indicates a given uncertainty in position, e.g., a
Gaussian distribution, but the observations of such states result in a precise value of
the position or a distribution of positions that is narrower that that implied by the
approximate state, then the specified state is simply not the complete description
required by a “physical state” of the system (as defined above).
 3

In quantum mechanics, the notion of a true physical state of a system runs

into trouble. The quintessential example is an entangled state. If two components
of an entangled system are spatially separated, then an observation of one of the
subsystems results in new information for that observer, in which case the observer
is entitled to update the quantum state describing the entangled pair. This is
euphemistically referred to as “state reduction” or “collapse of the wave function”. If
there were, indeed, a true physical state, one is tempted to consider state reduction as a
physical process, in which case that part of the state describing the remote entangled
subsystem will also be instantaneously reduced, a seeming violation of relativity. This is
the source for claims of quantum non-locality. On the other hand, as long as the remote
observer lacks knowledge of the results of the observed subsystem, the original quantum
state will provide a complete description of all future observations, by that observer, of
the remote subsystem. In other words, the notions of a true physical state and state
reduction are again metaphysical appendages that add nothing to the utility of the state
concept.

The EPR Paradox

The 1935 paper by Einstein, Podolsky, and Rosen (EPR) considered the case of
an entangled pair of particles. The states of the two particles were presented not as
descriptions by observers but rather as “descriptions of reality”, i.e., true physical states.
EPR then find a contradiction in these assumptions and conclude: “the description of
reality as given by a wave function is not complete.”
Let’s investigate the entangled state scenario in more detail using the Bohm spin
singlet system, the same system employed by John Bell in his seminal paper on the EPR
paradox (Bell 1964). The system consists of the emission of two oppositely moving spin
½ particles in a singlet state, that is, the spins of the two particles are precisely opposite
each other but in an undetermined direction. According to quantum mechanics, their
combined wave function (quantum state) is given by
!
Ψ 1,2 = !
|1, ↑ ! |2, ↓ ! − |1, ↓ ! |2, ↑ ! (1)

where ↑ and ↓ indicate the up and down z components of the spins of particles 1 and 2.
Now suppose that the spin of particle 1 is measured with a Stern-Gerlach apparatus
 4

oriented in the 𝑧 direction and is determined to be ↑. Then we “know” that the z

component of the spin of particle 2 will be ↓ and the “true” state of particle 2 will be
instantaneously reduced to Ψ 2 = |2, ↓ ! . On the other hand, the original wave function
can also be expressed in terms of the x components (or in any direction for that matter) of
the spin,
!
Ψ 1,2 = !
|1, ↑ ! |2, ↓ ! − |1, ↓ ! |2, ↑ ! .

Then, if the spin of particle 1 is measured with a Stern-Gerlach apparatus oriented in the
𝑥 direction and is determined to be ↑, the spin of particle 2 will be ↓, and the “true” state
of particle 2 becomes Ψ 2 = |2, ↓ ! , a state fundamentally distinct from |2, ↓ ! . The
immediate transition of particle 2 to either the state |2, ↓ ! or |2, ↓ ! , depending on the
measurement performed on the distant particle 1, is the basis for claims of “action at a
distance”. Both Einstein and Bell accepted this picture but while Einstein was abhorred
by such “spooky action at a distance”, Bell embraced it as a demonstration of the “gross
non-locality of nature” (Bell 1975).
Now let’s look at this scenario from the perspective of a subjective description of
“state”. For the entangled singlet state considered above, Einstein would conclude
particle 2 has no unique (real) state. Let me push back on this conclusion. After their
emission, the polarization of neither particle is known. Therefore, the two particles can
be considered as two unpolarized particle beams. For the observer of particle 2, this can
be represented by a 50% mixture of spin up, |2, ↑ , and spin down, |2, ↓ , states (in any
direction), a so-called mixed state. A mixed state cannot be described by a pure quantum
state (e.g., Schrödinger wave function or vector in Hilbert space) but is well described by
its associated density matrix 𝜚,
! !
𝜌 2 = ! |2, ↑ 2,↑ + ! |2, ↓ 2,↓ (2)
with a similar expression for particle 1. This density matrix completely characterizes the
spin measurements of particle 2 with no mention of what measurements are made on the
other particle. For example, Eq. (2) implies that a measurement of the spin component in
any direction will yield ↑ for 50% of the measurements and ↓ for 50% of the
measurements subject to the usual statistical fluctuations. The measurement made on
 5

particle 2 is completely independent of any measurement carried out on particle 1.1

Okay, so far so good. Nevertheless, you may object that these single particle
mixed states are mute on the correlations of measurements made on the two states. Fair
enough, but now you are asking a different question. That is, after making these
independent measurements, what are the correlations between them? To answer this
question, to be sure, one needs the pure entangled state of Eq. 1 but this expression in no
way implies that the measurements made on particle 1 have any effect whatsoever on the
measurements made on particle 2 or that individually these measurements are other than
those predicted by the density matrix of Eq. 2.
So after the measurement of particle 1, what is the quantum state of particle 2? Is
it still the density matrix 𝜌 2 of Eq. 2 or is it the wave function |2, ↓ ! ? The premise of
this question is wrong. The quantum state is not an “it” in this sense. The only “it” is the
quantum system, the spin ½ particle. The quantum state is not a physical property but
rather a description of a physical object, and that description depends on the information
available to a particular observer. For Observer 1 who has just observed particle 1 to be
in the ↑ state, the available information leads to Ψ 2 = |2, ↓ ! while for Observer 2 the
appropriate state is still given by the density matrix 𝜌 2 . For a probabilistic theory like
quantum mechanics, confirmation can only come through the measurements of an
ensemble of identically prepared systems. For observer 2, the results of such
measurements will be completely described by the density matrix of Eq. 2. Precisely the
same is the case for Observer 1 prior to the observation of particle 1. After the
measurement, Observer 1 is justified in specifying the state of particle 2 as either |2, ↑ or
|2, ↓ , depending on the result that measurement. Finally, suppose that the results
obtained by Observer 1 are communicated (subluminally, of course) to Observer 2. With
this additional information in hand, Observer 2 will naturally specify Ψ 2 = |2, ↓ ! or
|2, ↑ to be the quantum state of particle 2 depending on this information. The point is, a
quantum state is a subjective description that depends on the information available to a
specific observer.

1 After all, the measurement made on particle 2 could be made in advance of even selecting the type
of measurement to be made on particle 1.
 6

Entangled Classical State

There is a simple classical analog of an entangled two-particle state. Suppose
there are two balls, a white ball and a black ball. The two are randomly placed in either
of two boxes. The boxes are closed and sent off in opposite directions where they
encounter observers who open the boxes. In this case, the entangled state is a 50/50
mixture of a white ball in box 1 – black ball in box 2 and a black ball in box 1 – white
ball in box 2. On the other hand, it’s clear that each observer has a 50/50 chance of
finding a white ball and a 50/50 chance of finding a black ball and so is justified in
describing their subsystem as a 50/50 mixtures of white ball in the box and black ball in
the box. This can be confirmed by compiling the data from multiple (random) trials of
the experiment. There is no set of observations performed on an ensemble of random
trials that disagrees with this prediction, i.e., the 50/50 mixture of white and black balls
constitutes a complete description of boxes 1 and 2. If Observer 1 finds a white ball, that
observer is justified in “collapsing” the entangled state to a white ball in box 1 and a
black ball in box 2. However, until this information is transmitted to Observer 2, the
complete description of that observer is still give by a 50/50 mixture of white and black
balls in the box and this describes the predictions of future observations of box 2.
As in Bohm’s spin singlet quantum system, it is also clear that the findings of the
two observers will be completely correlated; if the observer of box 1 finds a white ball,
the observer of box 2 will necessarily find a black ball and vice versa. So for an observer
privy to the results of the observations of both boxes, the appropriate description is the
mixed entangled state. As in the quantum case, we would never claim that the act of
observing a white ball in box 1 causes a black ball to appear in box 2. Einstein would not
be bothered by this situation because, in his view, there is a true physical state of each
subsystem and so there is no necessity for positing the physical process of state reduction.
We’ve already noted that such an assertion is completely compatible with classical
physics. However, as was pointed out above, these considerations are metaphysical and
don’t add to the utility of the notion of a physical state. 2 (Of course, quantum

2 This toy model of an entangled classical state may seem a bit contrived; however, there is a well-
defined probabilistic formulation of classical mechanics based on the Hamilton-Jacobi formalism on
configuration space wherein the notion of entangled states and mixtures is completely natural. (Hall
& Reginatto 2016)
 7

entanglement is a much richer phenomenon than classical entanglement and exhibits all
the aspects of quantum interference with which we are familiar.)

Final Remarks
The notion of a physical state as a “complete description” need not encompass all
the information about the system. For example, one might characterize the position and
momentum of an electron with a spatial wave function that neglects any mention of the
spin state of the electron. In this case, the wave function can still be considered to
provide a complete description so long as observations are limited to position and
momentum and there are no magnetic fields that might couple spin with the spatial
trajectory. Similarly, the white ball/black ball classically entangled state presented above
can still be considered complete so long as the plethora of other properties of the balls
(e.g., mass, density, temperature, composition, etc.) are irrelevant to subsequent
observations of the system.
In the abstract I indicated that characterizing physical states as subjective
descriptions “goes a long way toward resolving such dilemmas as Schrödinger’s cat,
wave function collapse, quantum non-locality, and parallel universes”3. Wave function
collapse and quantum non-locality were addressed directly in the paper. It should be
clear how to apply the notion of subjective states in the resolution of other conundrums in
physics, and particularly in quantum mechanics, but I’ll leave this task to the reader.

Acknowledgement
Thanks to Marcel Reginatto who introduced me to the wonders of ensembles on
configuration space and the associated probabilistic formulation of classical mechanics.

References
Bell, J., 1964, “On the Einstein Podolsky Rosen Paradox”, Physics 1, 195–200.

Bell, J., 1975, “The Theory of Local Beables”, in Speakable and Unspeakable in
Quantum Theory, 2nd edn., by J. Bell (Cambridge Univ. Press, Cambridge 2004).

3 I’ve discussed Hugh Everett’s “many worlds interpretation” in detail elsewhere. (Boughn 2018)
 8

Boughn, S., 2018, “Making Sense of the Many Worlds Interpretation”,

arXiv:1801.08587 [quant-ph].

Boughn, Stephen, 2019, “Against ‘Reality’ in Physics”, arXiv:1903.08673

[physics.hist-ph] [quant-ph].

Einstein, A., Podolsky, B., and Rosen, N., 1935, “Can quantum–mechanical description
of reality be considered complete?”, Phys. Rev. 47, 777–780.

Hall, M., and Reginatto, M., Ensembles on Configuration Space: Classical, Quantum,
and Beyond (Springer, Berlin 2016).