Professional Documents
Culture Documents
Quantum Mechanics
Information
(and only
a little
more)
Christopher A.Fuchs
Computing Science Research Center
Abstract
In this
paper,
I
try
once
again to
cause some
ever stop
the taxpayer
burdening
with conferences
found
no end
to reduce
statements
quantum
of crisp
theory
physical
axiomatic) significance.
The suspicion is
a means is
or three
to two
(rather
than abstract,
In this regard,
no tool appears
better calibrated
for
a strained application of
a time-honored problem, this method
precisely because a large partbut not
information
information.
has always
paper, though
core, corrects one
This
its
observations
particular,
mechanics
in the
beyond
mistake
the
one
Iidentify
and
traditionally
is
ascribed to
the
offers
previous
version.
element
that I
would not label
theoryit
as
sev-eral
taking quant-ph/0106166
of
In
quantum
a subjective term
integer
parameter
a quantum system
via its
Hilbert-space dimension.
Introduction
Quantum
as a
with us for
theory
weather-sturdy
75
years now.
one can
a year has gone by in the last 30
when there was not a meeting or conference
devoted
to some
aspect of the quantum
check that not
foundations.
o, Quantum
is only
mystery?
for want of
the
it is, it cannot be
Whatever
a self-ordained
the
solution: Go to
any
,
,
with all
Consistent
Historians
[4],
the
[6]
the Einselectionists
Objectivists
and
[7]
many more
the Contextual
Everettics
[9, 10],
This
viewed
paper,
as a
though
present
paper,
substantially
continuation
and
can be found
Section 1
1
longer,
amendment
should
to
Ref.
in the Appendix
Substantial
further
be
[1].
to the
arguments
defending
transition
stance implicit
stance implicit
from
the
objective
can be found
Bayesian
Bayesian
inRef. [2].
1
A Fraction of the Quantum Foundations Meetings since 1972
1972
1973
1974
Quantum Mechanics,
1975
1976
International Symposium
1977
1978
Stanford Seminar
1979
1980
Quantum Theory and the Structures of Time and Space, Tutzing, Germany
1981
Marburg, Germany
a Half
on Fifty
Vienna, Austria
on the Foundations
on Quantum
Optics, Experimental
a Tribute to Louis
1982
1983
Tokyo, Japan
1984
1985
Symposium
on the Foundations
1986
New Techniques and Ideas in Quantum Measurement Theory, New York, USA
1987
Symposium
on the Foundations
1988
1989
1990
Symposium
Washington, DC,USA
on the Foundations
1991
1992
Symposia
onthe Foundations
1993
International Symposium
on Fundamental
Oviedo, Spain
1994
1995
The Dilemma of Einstein, Podolsky and Rosen, 60 Years Later, Haifa, Israel
1996
on Fundamental
Oviedo, Spain
on Conceptual
1997
Sixth UK Conference
1998
1999
2nd Workshop
2000
2001
on Fundamental
on Decoherence
o,Sweden
wrong
this! If
light,
any
with
there
simply
year-after-year
would
conference.
clear enough: If we
not
The
be
the
verdict
seems
care about
us as a
meetings
are
quantum foundations,
community
to ask
then it behooves
why
these
accusations
Despite
incompleteness,
of
nonsensicality,
sees one
religion making against the other, I
see little to
no difference in any of their canons. They all
irrelevance,
look
equally
quantum
to want
single
detached
practice to
a firm
God
one
and surreality
from
often
the
can
they
point
world
of
seems
theoryi.e., a
each
to and declare,
even when
there
worked
provocative
differences
statement
in what
the churches
proceed
from the
same
label
to be
they nonetheless
abstract
starting
all
point
every system,
space
1.
For
2.
3.
are observable
H.
4.
But what
nonsense
is this,
is this, and
you must
no one
has said it
more
clearly than
Where present-day
quantum-
so unlike
paper on special
relativity.
do,
rather
have the
name
than,
published
one
some
say
could
of them
that
most
the
empirical
before Einstein
of
community
so
striving
electromagnetic
a mystery to
particularly, this was a
mystery
that
More
of
be conquered.
sense
hard to make
phenomena
remained
heaping
ad
further
hoc
(mathematical)
solve.
2
Very
might
briefly,
as
be
by
captured
actual
universes
though,
in
Hamiltonian.
these
some
are
universal
state
vector
with
preconceived
tensor-product
Spontaneous
Collapsians
vectorthough
now
for
it
is
For
and
an
the
the
the persuasion,
supplemented
decompositions
respect
is
with
function
upon
struct
the
vector
sometimes
positions
reality
space.
wave
(Depending
two entities
the
Bohmians,
coordinate
universal
be.
of
the state
supplementing
it is the
of
For the
follows.
trajectory
Everettics,
a cartoon
ures
again
individual
to
of the
various
) For
the
the
state
systembut
Hamiltonian
dynamics
collapse
mechanism.
reality
is captured
state and
preferred
is supplemented
For
the
with respect to
Hamiltonian
with
of
measures
valued
objective
Historians
an initial quantum
by the addition
positive-operator
an
Consistent
a set
sets of historiesalong
of
(POVMs)
with
minor modifications
own way,
to the meanings of
a few
contribute
words in the
3
What
was
of
years
was an understanding
structuresome
abstract,
simple,
mathematical
crisp
physical
mathematics
Einstein supplied
would
that
be
indisputable.
and became
one
of the
structure
transformations
to
of
two
the
simple
Lorentz
statements
space is
of its source, 2)physics
appear
should
the
same
frames.
The deep significance
problem
should
up
anyone
to
overpoweringly
stand
and
who
speak
admires
these
principles.
Einsteins
debate
a handful
plicity
of Einsteins
ready
for the
imagine
that
next
any
with the
principles,
step.
mindeven
supreme
physics
sim-
became
Is it possible
to
Einsteinscould
from
the
original,
abstract
Lorentz transformations?
only
Indeed,
find
in
one can
A structure
adequate?
empirically
structure
of
the
that
was
Iwould
pursuing
the
same
strategy
say no.
we will
of
we are:
Symbolically, where
Where
Speed of light
vt
p1
v 2 /c 2
t vx/c
p1
we need to be:
is constant.
/c
sense of the
more structure,
more defini-tions, more science-fiction imagery on
top of them, but to throw them away wholesale
The
task
quantum
and start
asking
is not
axioms
to make
heaping
afresh. We should
ourselves:
principles
might
mathematical
should
by
From
we
derive
structure?
be crisp; they
be relentless
what
deep
this
Those
should
be
in
physical
exquisite
principles
compelling.
high school, I
sat down with Martin
book Relativity
Gardners
came
away
sustains
So, throw
pro-ceed?
the existing
away
mechanics
I
myself
contemplate
techniques,
deep
and
axioms
of quantum
see no
and
alternative
hard
the
the implications
reason
to
but
tasks,
the
of quantum
is simple, and I
about
information.
It is just
that
the
Quantum Mechanics:
operators
an information
reason if possible!
theoretic
an information
theoretic
Give
H.
measures
Give
reason if possible!
H is a complex vector
not
(POVMs)
}on H.
space,
Give
an information
theoretic
reason if possible!
quaternionic module.
spaces,
HAB
= HA
HB
an information
reason if possible!
theoretic
an information
reason if possible!
theoretic
an information
theoretic
Give
By
maps.
Give
evolve
Give
maps.
reason if possible!
an information
reason if possible!
Give
theoretic
This, I
see
as
pursue
with
quantum
system represents
independent
a collection
of
we
should
consistency:
The
relentless
something
real and
with
our
separate
5
the
But physicists
forever
trying
to
The task
subjective
before
us
is to
wheat
are, at bottom, a naive
come to terms
breed,
out
there by
methods
in
essence
the
same
element of contact
as a well-placed
5
the
from
chaff.
represents
much
subjective
its
of
the
If
same
support
character?
state
then
information,
mathematical
might be of that
quantum
how
structure
Some of it,
and
an
every
task should be to
go to
each
information
Only when
theoretic
we are
justification
finished
picking
if
we can.
can
will we
of terms) that
be
as subjective information
be
a position to make real progress in quantum
foundations.
The
raw distillate left
behindminuscule though it may be with respect
to the full-blown theorywill be our first glimpse
of what quantum mechanics is trying to tell us
interpreted
in
about
me try to
give
this by making
was
in his recognizing
I
would
say
it
field
one
coordinate
feels in an accelerating
a
case
elevator is
description
program
an
of
observer.
be
and
viewed
coordinate choices.
things that
are
the observer-free
come
about
motional
additional
numerically
situationi.e.,
purely
by
that
[15] to
bringing
agent, coordinate
(scientific
phenomena
as consequences of our
It was in identifying all the
the
observer
was a true
out
interpreted
all
as
behind finally
Riemannian
breakthrough.
the
becomes
manifold
can
be
left
that
we
call
spacetimea
us
in
things
coordinate
For
something
one can
is
do
with
gambling
commitments,
information,
Summary
I
say to the House
as
I
said to ministers who have joined
this government, I
have nothing to offer but blood, toil,
us an ordeal
of the most
us many, many
months of
war. War
can answer
us. You
ask, what is
our
aim? I
I
should point out to the reader that in opposition to
the picture
of general relativity,
coordinate
systemi.e.,
observerchanges
tells
us
what
nothing
kind
more
that
the
matter
is
there).
necessary part
things
once
mechanics
same
pure
a
spacetime, rather
(when
how
to
system.
matter
move
is
(when
agents,
scientific
Observers,
world
will be
coordinate
curve
matter
tells
[16]
to
will pick
the observer
it off with
how
(it only
manifold
matter into
gridding
tells spacetime
Matter
the
the
the
the observer
reintroducing
like introducing
than simply
about
of sensations
up), I
do not suspect
Here I
suspect
where reintroducing
reintroducing
they
are on
the
scene?
Yes. If quantum
truly
think it is this.
not something
It is
task for
productive
points
community
Philosophy
philosophers,
college.
The
say
to be left to the
I
am apt to
quantum
is too important
view.
of
community
information
Likewise,
forth from
a medieval
once
said.
foundations.
The structure of the remainder
is
as
In Section
follows.
paper
Why
Ireiterate
Information?,
of the
the
cleanest
argument I
know of that the quantum state is
solely
an
expression
informationthe
considering
subjective
of
one
information
no
The argument
the
has about
objective
reality in
is then refined by
phenomenon
quantum
of
teleportation[23].
In
Section
What?,
I
tackle
head-on.
The
consequences
of
into nature.
quantum
About
Information
that
very
question
answer is the
our experimental
Once freed
measurement
[24]
potential
interventions
to
be
about
revealing traces of
(or beable [26]),
as a
basic notion.
theory,
utility of generalized
Indeed quantum
measurements
or
on
the
positive
measures
operator-valued
suggests
one
[27]
(POVMs)
as
The productivity
the
of this
probability
of the quantum
derivation
rule recently
by Joseph Renes
spaces over
the
a strengthened argument
noncontextuality
theorem.
In
International
start the
des Poids
assumption
4.2,
Section
process
as well for
even for
and
In Section 4.1, I
give
for
spaces,
Hilbert
et Mesures
of defining
in this
Bureau
Le
two-dimensional
and
what
Paris,
it
viewto
means
accept
notion
of fixing
In Section 5 Wither
ask whether entanglement
as
far
as quantum
Entanglement?,
is all it is touted to be
foundations
are
concerned.
the
mechanics,
combat
one
the
departure
tensor-product
trait
that
from classical
this, Igive
as
really
characteristic
Schr odinger
quantum
of
its
enforces
entire
lines of thought?
simple
derivation
To
of the
spaces
Hilbert
as its starting
use of
Gleason-like
7
If
you want to
in
the
means,
ask
me over a
beer sometime.
8
There
have
been
other
soundings
and computation
theory
deep
the foundations
of quantum
about
the
of
can tell us
information
idea
that
something
mechanics.
See
In
the
previous
quant-ph/0106166,
information
emphasize
probability
nothing
so
and
states
ascribes
than
the
to
of
paper,
quantum
Bayesian
are
states
and
did
not
idea
that
the
phenomenon
to that
Bayesian, probabilities
this
knowledge
gambling
of
called
one
more
version
I
variously
amounts
commitments
to
one
is
phenomenon. To the
subjective
all the
way
to turn
my
somewhat
paper, I
start
earlier de-emphasis
dangerous
the long
around
process
of trying
(even though it is
is little
more
paper)
the
word
knowledge
than
a modification
i.e.,
knowledge
is either right
as
The conception
much
as
or wrong
no
such thing
beliefs
call
or
mathematical
gambling
quantum
information
more
opinions,
already
commitments.
states
beliefs,
piece
sense
as a
right
paper
is specifically
Thus I
now
states
than is becoming
a set
and gambling
commitments.
of
variously
of
belief,
common
community),
is
and true
for capturing
of
version.
of this
information
understand
of
(though, by this I
mean information
subjective
quantum
particular
a quantum state
symbol
completed
in the present
possible
that
and only
of
an already
that
in a
in the
judgments,
Believe
me,
of jaws that
will
a sense
fears that I
will become
crystal-toting
New
medicine[22]I
to be absolutely
Age
of butterflies
practitioner
in the
solipsist [21]
of
or a
homeopathic
of
some
of
theory is part of
fruitful
a larger program to
in a
way.
presence
of classical
notion
communication.
structure
tensor-product
of entanglement
With the
established,
follows
the
very
in step. This
probability
is
ask
one
why
Bayes Rule?,
expect
the
updating
completion of
it actually
for
upon
the
that
on average
measurement
that
does
a tautology!)
proof technique
always
information
this is not
rule
derivation
increases
should
used for
not
itself
appearance
discard
otherwise,
indicates
an
quantum
collapse
probability
strong analogy
extremely
and Bayes
overall
unitary
readjustment
the
readjustment
initial
state
takes
the
of
into
as
system
for the
description
an
Up to
theory:
between
rule in classical
interactionquantum
account
ones
as
well
ones
measurement
collapse
is
precisely
more
impetus
sections.
theorems
Gleason-like
the
behind
previous
6.1, Accepting
In Section
Mechanics,
of
I
complete
the
process
the
two
Quantum
started
in
any
I-know-not-what
that
mea-surement
to an
is
leads
about
the
outcome
potential
of
the
Section
Information?,
a quantum state
7,
describing
Else
is
I
argue that, to the extent that
is
be the assignment
for
What
quantum
state
upon
the
completion
some
measurementgenerally
of
POVMwhose
precisely
draw
an
the
same
rule must be
sake. To
for consistencys
theory,
distribution
P(h)
probability
distribution
the state-change
rule d
the
andstatistical
the
model
some
for
and
P(d| h) enacting
the
transition
Bayesian
probabilities
are
subjectiveeven
the
probabilities
P(d|h)
of
a statistical
Specializing to the
is gathered,
one
completely positive
time-evolution
case
that
no information
are
maps
themselves nothing
more
than
subjective judgments.
In Section 8 Intermission,
breather to sum
up what
I
give
a slight
we are headed.
where
In
Unknown
I
tackle
very
these
Section
States?,
words.
use
ubiquitous
Quantum
the conundrum
Despite
posed
the
in the quantum
information
literature, what
by
phrases
be?
informationmust
someone, if it exists at
hand, for many an application
known by
other
it
information,
imagine
that
background
is
there
be
quite
is always
describing
grounding
in quantum
contrived
someone
to
in the
system
being
we are
the
or manipulated,
measured
doing
would
always be
all. On the
the
phenomenon
with
case
is found
version
of
of quantum-state
through
de
Finettis
unknown probabilities.
classic
mechanical
theorem
on
[33]. Maybe
one
of the most
numbers,
tomography
a quantum
spaces
suggesting
over
that
perhaps
the
whole
might not be
I
flirt
Quantum,
tantalizing
Oyster
with
the
and
most
There is
that
to our touch.
8
It has
in
optimal
method
be
nothing
more
of reasoning
than
is
the
and processing
a fundamental
a concrete proposal
(wonderful)
for
Zing!,
sensitivity.
potential
I
consider
traditionally
As
mathematical
the integer
ascribed to
expression
parameter
a quantum system
of
by
way
of its Hilbert-space
Why Information?
Realists
can
dimension.
be tough customers
indeedbut
there
is
no
reason to be afraid
of them.
immaculately
In particular, he
absolutely
was
the first
unambiguous
person to say
in
same
thing,
beliefs and
otherwise).
His
argument
quantum-state
was
assignment
that
a
a system can
simply
for
be
go one way or
forced to
a part
with
causal connection
The paradigm
known
paper
with
course
the Einstein,
a long
the
one
Podolsky,
thought had
no
here is of
through
[34]
pre-history
well
Rosen
of the train of
himself.
The best
was
in
essence
this.
Take two
some
entangled
quantum
state
AB
i. By
of one or another
on system A alone, one can
her
immediately
of two
write down a new state for system
B. Either the state will be drawn from one set
Bi
of states {| the
i}state
or another
wil l
be{|
drawn
i}, depending
from one
i
upon which observable is measured. 11 The key
the measurement
performing
of two
observables
two systems
are
other
concluded
fine
that
details
of the
whatever
world.
these
any
of
Einstein
things
called
affairs
the
more
cannot
be construed
to be something
than
a reflection
ones
initial
of
some
tricky combination
and
information
the
the language
cannot
be
of Einstein,
complete
the
as more
of
knowledge
Expressed
the quantum
description
in
state
of
the
quantum system.
Here is the
way
Dash,
11
verve,
Generally
between
there
need
be
hardly
only
any
that
relation
when
the
states
are
together
weighted
by
their
probabilities,
B alone. For
precise statement
operator
they
mix
for system
see
of this freedom,
Ref.
[36].
9
What
relation
is there between
) ?
Does
we
situation (that
the
the state
function and
quantum
state
characterize
cannot
One
respond
to
unambiguously
represents
uncontrollable
the
(Heisenberg).
something
that
experience,
thetical.
intervention
The
real
is
in
state
immediately
(Comparable
therefore
to
accessible
always
rests hypo-
of motion.) Therefore
system
not
is
suppositions
a priori
can be taken
in
the law
this
a real
are,
one
an examination and
consequences
obliges us to admit that
only after
rigid
system separated
connection
between
parts of
space
in
the
an
arbitrary
way
(instantaneous
action at
distance,
, ,, ,
with
S12
function
which
12
and S2
is
which
are very far from each other at the instant t.If one
makes a complete measurement
on S1 which can
be done in different ways (according to whether one
measures, for example, the momenta or the
coordinates),
depending
on the result of the
measurement and the function 12 one can determine
of the second
different
measurement
according
forms,
applied to S1
the
excludes
measurement
S2
at
action
on S1
and therefore
effect
with (1) if
one
assuming
(1)
pass to
considers
quantum theory
amounts
as
the
a system
one
on
method
of
the
complete
(2)
is only
present
to renouncing
the
the supposition
If
the
Therefore
on
no
of
distance.
no
has
can assume
procedure
that
description
re-nunciation
of
if one
assumes
their description
that would
would
mean:
only to what
that
be useless. Otherwise
said,
observation
teaches
us
about
them
partial
knowledge
are
however
entirely
deterministic.)
Now,
statistical
accept
Icant
character
conditioned
by
that.
Ithink
of the present
the
choice
of
theory
an
that
the
is simply
incomplete
description.
admitting
that
systems has
any
the
a complete
very
meaning
notion
of
separate
at allone is led to
specification of
be
be
of
Einsteins
unwillingness
as a necessary
thought:
to accept such
an
namely,
2)
an
incompleteness
It is quite important
reason
...
step
more.
was, I
reason
a simple
psychological
For if the
unavoidable
12
on the other
hand,
even glorify.
So, I
suspect these words will fall on deaf
ears
ears
every event
says
no further
explanation is
most
progress,
new
phenomena
10
to look elsewhere for
complete description
so it would
of the
be clear from
very
the
beginning
description
are not
that
the elements
contained
within
of such
the conceptual
one
not serve
as the basis
of theoretical physics.
after confirmation
several variations
of it)
the
are
indeed violated
by
no-go
clarified to the
pictures
(and
can be
seems, is
that no
the
outcomes.
In
quantum
information
is simply
[40]
neither
But
Wolfgang Pauli
theory,
not complete
can
it be
maximal
information
completed.
As
Fierz [41],
The
well-known
mechanics
be
removed
by
an
existent
but certainly
cannot
is certainly
fact somehow-somewhere,
quantum
of
incompleteness
(Einstein)
reverting
to
classical
field
physics. Nor, I
would add, will the mystery of
that existent
an
ontological status.
The
complete
disconnectedness
the
of
spacetime
something
in the
intolerable.
The
is
telling
quantum
state
Subjec-tive, incomplete
information.
Put
considerations
deep:
right
It is
mindset,
a statement
this
us
is
information.
is
about
so
not
our
world.
us
from
ever
be captured
getting
more
through
infor-mation than
can
of
us to
the incomplete
perhaps the
But
things
first
question for
one
first.
anyone
more immediate
come this farand
The
who has
to be answered
that deserves
forthrightis
actually about? I
have hinted that I
would not
dare
say
our
some
that it is about
kind of hidden
might)
or even
about
(as
way to
build
up to an
no
quantum
further
to the
than
teleportation
[23]
another
part
of
phenomenon
Not
only
a system
can a
be forced
the
world
of
no
causal
significance,
can
that
make
anything
one wants
of
quantum
it to be.
state
one
assignment
Such
Bob sharing
AB
(1)
(|0i|0i
Bob then
goes to any
|1i|1i)
laboratory
prepares
any state |i that she
ultimately wants qubit
to impart
wit hany
ontostate
Bobs
| i
system.
t hat
She performs a Bell-basis measurement on the
two qubits in her possession. In the same vein as
wishes
Alice
another
qubit
thought
Einsteins
13
in her
with
longer
now,
entirely
pleased
refer to
for instance,
maximally
rigid
my
with this
gambling
warnings,
terminology.
I
am
no
I
would
or some
such
thing. See Ref. [2], pages 4950 and 5354. However, after
trying to reconstruct
conformity
with my
a more accurate
the section
this paragraph
new
representation
even more
several times to be in
terminology, I
finally decided that
would
break
the flow of
11
experiment,
system immediately
Bobs
the character of
one
of the states
takes
on
x |i,
|i,
y
on
|i,
the
or
characte
z |i. r
But
of on
that
e ofisthonly
e stat
insofar
es |i,
asxAlice
|i
is
y |i,
concerned.
14
Since there is
no
(reasonable)
for
If
now
Alice broadcasts
measurement
may
complete
protocol by performing
conditioning
is that Bobs
,,
the teleportation
system,
it
as
(I, x
one
of
z ) on his
on the information he
as Alice is concerned,
far
the state
system
final
|i.ly resi des predi ctably in the sta te |i
1516
How
can Alice
YES-NO measurement:
a measurement on a
causally disconnected
of causal
i.e.,
action
one
on
sharpen her
can
to complete certainty
predictability
has
a state
property
argues
Penrose
in his
book
The
a system
to be some
Emperors
|i
there
ought
corresponds
to its |iness.
the
corr system
espon ds
betoprepared to reveal a YES in the
case that Alice actually checks it? Asking this
rhetorical question with a sufficient amount of
command is enough to make many a would-be
informationist weak in knees. But there is a
crucial oversight
we
have already
implicit
in its confidence,
no one
ultimate
More importantly,
and
to anyone else in
else who
revelation
can predict
with certainty.
power to stand
up
and
say
with
17
consequence
4
revelation.
in Alices
confidence
interact
no
over to it and
There
mind
it, she
interact with
is only the
that, should
could
predict
she
the
of that interaction.
I
think
student
that
gets
the
sickliest
it, is that
notion
of
it is the
physics,
science
even
of
if
masses,
notion,
even
if
student
of the
ways
W.S.Franklin, 1903
untold lessons
for her, there were also a
sprinkling for me. For instance, I
started to see
her come to grips with the idea that there is a
world
independent
14
As
far
as
Bob
is
concerned,
nothing
whatsoever
mixed state
completely
As
far
as
Bob
12
way.
I
and remains that
is concerned,
nothing
whatsoever
the completely
16
The
repetition
typographical
17
I
adopt
Savages
the
when
I
and remains that
these
this
the
and
to be
terminology
person,
the
decisions,
acts
measurement.
in
world,
via
Inthis
Savage
the
paper,
to L. J.
similar
2, where
the
consequence
to the person,
he
way.
not
is
footnotes
error.
decision theory. A
happen
12
terms
acts,
in
he discusses
consequences,
context
of
is anything
writes,
capacity
of
where
rational
that
we
may
add
quantum
I
call what Savage calls the
or observer
instead.
12
of her desires. What struck
me was
the contrast
saw
grow
in her
that
on the
I
also
are aspects
The two go hand
there
of
in
it pushes back in a
way
(and
foreseen
could
may never
manipulate
to the complete
the world
desires
of her willI
would
be little
became
difference
convincedthere
between
wake
and
dream.
learned
anything
can she
think that
no
theory of measurement.
completely
How
undefined.
can
a
a theory
she have
measurement
once
told
problem refers to
a bit
of the truth.
community making
use
part
the
this
or
experimenter;
that experiment
to
see
it
contemplation
up
the
of
is
not
We set
how nature
and their
our data
consequences
books.
that
we
18
something
new
when
of
can
further
reality
distill
example
model
a description we
of a free-standing
of
I
have
description,
but it
can
us a compact
more if we
It gives
give much
want it to.
The important
that there is
no
that such
If the world is
we can never
such that
experimental
prepared
realitya
interventions,
then
our
of
we must
be
in its most
theory
identify
realityindependent
free-standing
minimal
and conceptually
18
But I
must stress that I
am not
think
that
primitive
of
the
physics
Vienna
an
No
Heisenberg
are by
one got
(in
closer
a quote
he
positivistic
as to
grounded
on a
be
as
the philosophers
interventions
experimenter
to be thoroughly
but
behind.
did. The
Circle
that
years
so
somehow
consequences
option
should
and their
records,
theory-laden.
necessity
attributed
one
at least
to the salient
have
no
It is just
theory
point
to Einstein
than
many
It
is quite
wrong
to try
founding
can observe.
very
theory
on
opposite
we
is
a very
process.
complicated
The phenomenon
measuring
apparatus. As
a result,
further
paths
produce
processes
and
sense
impressions
inour consciousness.
the phenomenon
to
must be able to
we can claim
underlying
impressions.
we
When
us to deduce
phenomena
claim
that
from
we
our
can
sense
observe
something
new, we
ought
not
assume
with
the
old
ones, we
from
the
upon them
to
phenomenon
consciousnessfunction
in such
a way
and
nevertheless
our
that
we can
rely
13
believe,
with words
that
rhyme
[47]. It
not
observables,
19
dingables.
is
about
We tap
theory
not
beables,
about
but
bell with
our
about
gentle
are
So what
the world?
the
What
are
measurable
operators.
modern
Or
Xi
(2)
to
set
operators {i } over
that form
to
correspond
perhaps
language,
corresponds
The
reaction?
a complete
of
to
each
say
it
in
observable
orthogonal
complex
Hermitian
Hilbert
more
there
projection
space
HD
=I
The index i
labels the potential outcomes of the
measurement
(or intervention
the
language
possesses
observer
most
promoted
generally
the
probability
P(i)
tr(i )
(3)
density
dictates
that he
outcomes
various
.
.
captured
mixed-state
mechanics
by
way
noncontextuality
outcomes [48]
measuring
two distinct
then
independent
the
PD
or
{i } and
projector
outcome
is
is this. Let
associated
space
it is associated
the statement
complex) Hilbert
contemplates
a single
of which observable
(real
one
of
simple
measurement
observables
probability
a very
for
In particular, if
rule
theorem [30]
kind
with
comes
an
When
the information
by
operatorquantum
can expect
above)
HD
and let f :
PD [0,1] be such that
with
for D 3,
Xi
=1
f(i )
(4)
a set
whenever
observable.
exists
f()
of
projectors
The theorem
{i } forms
concludes
that
an
there
(5)
In fact, in
theorem
convex set
of density operators).
In itself this is
no
Note that
an
no restrictions
is
no
differentiable,
theorem
prove
it [49]
assumption
nor that
amazing
required to
ones
upon
mentioned above.
that
it
need
be
too,
comes
from the
that they
are
complete
sets of orthogonal
linear vector
space.
one
Nonetheless,
should
projectors
ask:
onto
Does
this
a clearer
vision of where the probability rule comes from?
Astounding feats of mathematics are one thing;
insight into physics is another. The two are often
theorem
really
give
the
physicist
turn,
can
unifying strand
quantum
foundations
be drawn by viewing
in the light
of quantum
information.
The place to start is to drop the fixation that
the basic set of observables in quantum mechanics
are
complete
be
extremely
convenient
to expand
the
notion of measurement
positive
measures
(POVMs) [39,
in place
of the
50]
In other
textbook
19
14
operator-valued
words,
notion
Pronounced ding-ables.
usual
any set
of measurement,
positive-semidefinite
a resolution
(6)
on HD
that forms
|i 0,
h|Ed
{Ed } of
operators
for all |i HD
and
Xd
=I
Ed
(7)
measurement
P(d)
= tr(Ed )
(8)
The
set
{Ed } is called
POVM,
and
the
operators
the
while the
individual
represent
Ed
the potential
consequences
nature, while the
of
that
standard measure- ments,
there is no limitation
the number
the index
of values
Moreover, the Ed
is
no
intervention.)
may
requirement
be of
that
any
Unlike
can
on
take.
they
be
mutually
orthogonal.
way
The
measurement
POVM
this expansion
is usually
of the notion of
justified
is that
any
measurement
interacted
interest. Indeed,
are initially
and A
suppose
= S
(9)
some
USA U
SA
(10)
Now, imagine
via
on the total
on the
space
Hilbert
a set
d }. An outcome
standard
P(d)
measurement
standard
ancilla. It is described
d will be
found,
by
the
= tr
)U
(I
U(S
(11)
The
number
of
outcomes
in this
indirect
notion of measurement
by the
dimensionality
spacein
principle,
seemingly
is limited only
of the ancillas
there
can
be
Hilbert
arbitrarily
many.
As
advertised,
probability
it
turns
formula above
terms of operators
on the
can
out
that
systems Hilbert
the
be expressed
in
space
we let
orthonormal
basis
for the
respectively,
then
|s
i|ac
will be
we
get
P(d)
=
(s
Xc
hs |hac |
=
Xc
|s
)U
(I
d )U
i|ac i
hs
!|s
(I
hac |
|ac
A )U
(I
d )U
(12)
Letting trA and trS denote partial traces
over
the
15
where
= trS (S
Ed
(I
Ed
trA
A )U
(I
( I
d )U
(14)
an operator
original
acting
system.
This
true.
which
everything
can
that
be said
a way to
think
measurements
standard
a sense
quantum
about
measure- ment,
standard
something
Indirect
about
though
one
the system
it
may seem,
in
turn
this
can
technique,
information
one
of
learns
of real interest.
powerful
otherwise [51]. A
in
capture
sometimes
be
revealing
re-vealed
very simple example is where a
a single qubit available for the
three potential messages. She
message in one
of three preparations of the system, even though
the system is a two-state system. To recover as
much information as possible, the receiver might
(just intuitively) like to perform a measurement
therefore has
a need to encode
the
outcomes.
were
to
limited
measurement,
If, however,
he
quantum
standard
outcomes. Thisperhaps
surprisinglygenerally
way
this standard
of justifying
up
the POVM is
any
Might
of
the
mysteries
is whether
quantum
mechanics
notion
basic
POVMs utility
of
measurement?
portend
a larger
Does
as
the
Standard
Generalized
Measurements
Measurements
{i }
{Ed }
h|i |i 0,|i
Pi
P(i)
=ij
|i 0,|i
Pd
=I
=tr(i
i j
h|Ed
Ed
P(d)
=I
= tr(Ed
I
try to make this point dramatic
lectures
by exhibiting
table above.
list
of
notion
right-hand
of
a transparency
in
my
of the
various
of
properties
for
the
standard
a quantum measurement. On
side, there is an almost identical
properties
for
the
POVMs.
The
the
list
only
one
is missing
condition required
of
the orthonormality
standard
The question I
ask the audience
measurement
addition
make
of that
the
mysterious?
Indeed, I
imagine
myself
teaching
a vote
with the
best
audience
of
all, the
measurement
would
you prefer?
I
am quite
sure
16
is the point! It would make
them, and it should make
no
no
difference
to
to
us
difference
measurement
will allow
us to see more
deeply
let
us pose
the
question
that
interventions.
Concerning
those probabilities,
let
us
(in analogy
f :ED [0,1]
(15)
where nE
ED
:
0 h|E|i
1 |i HD
(16)
a POVM,
=1
f(Ed )
(17)
(Ingeneral,
f(E) 0
(18)
and
satisfying
= constant
f(Ed )
a frame
over
HD .)
as no surprise,
of course, that
must exist
f(E)
a density operator
= tr(E)
for which
(19)
This
was
independently,
consider
us
now.
the
case
where
HD
and
the
operators
First, con sider
on itth
are
ecase
defined
whe re
only
H Dover the field
(complex) rational numbers.
see
that
f is linear
combinations
of operators
It is
no
with respect
of
problem to
to
positive
,,
thatpositiv
never ecom
go outside
bin ation
ED
For
consider ta
three-element
of operators
hat
nevergo outPOVM
side ED{E1
By assumption f(E1 )PO
+ f(E2
VM)+
{E1
f(E3 )
E2
E3
}.
= 1.
}.However,
By ass
we
mption
can f(
also
E1 group the
POVM to obtain a new
have f(E1
function
= 1. In other
) +f(E3 )
f must
be
additive
E2 )
f(E1 )
f(E2 )
(20)
Similarly for
in this
words, the
respect
to
= mf
f(E)
with
operation:
fine-graining
f(E1
E2
= nf
1
E
(21)
Suppose
n
m
1. Then if we write E
statement becomes:
n
f(G)
= nG, this
(22)
17
Thus
on ED
Thus
we immediately
we imme diately
on One
ED
have
a kind
of limited linearity
way.
Clearly
cap
off
the full
vector
space OD
of
Hermitian
on
HD is
the
setitia
ED n
since
that
v ector
spspanned
ace O D by
of H
erm
operator
set
among other things,
ED
all the projection
on contains,
HD
operators. Thus, we can write any operator E ED as
a linear combination
D
operators
2
D
Xi=1
Xi=1
i Ei
Xi=1
(23)
for
some
fixed operator-basis
f(E)
Xi=1
{Ei
i=1
Linearity of
f(Ei )
Xi=1
(24)
So, if we define
by solving
= f(Ei )
tr(Ei )
(25)
the D
linear equations
we would
f(E)
have
Ei
Xi
tr
Xi
= tr
= tr(E)
i Ei
(26)
and
essentially
normalization
density
be
done.
(Positivity
of f would require to be
operator.)
But
the
problem
and
an actual
is that
in
(23)
there
is
.
.
g1
g2
G1
and,
unextended
by
the
definition
g1f(G1). Furthermore
additivity
homogeneity
of f,
this
we
of
the
original
obtain g2 f(G2 )
extension
retains
the
suppose
that
are
necessarily
.. .
in ED
can
We
find
rational
nec essa
positive
rily
number
in ED
c 1
We
such
c an
thatfind(Ea+
pos
G),itiv E,
e rat
and ional
G are
nu
1c
1c
we
all in
Then,
mber
c ED1
such
that by the rules
obtained,
?1c
?1c
= cf
(27)
Let
space
+ cf
us now
OD
already
(E + G)
?1c
have
= cf
f(E + G)
1c
= f(E) + f(G).
This
can
be done by noting
as
that
the difference H
semi-definite
any
= E
operators. Therefore
,., , ,
follows
that
E1 E2
G2
(as
G1 and
f(G).
suppose
there
= E2
+ G2
extended
previous
in the
we obtain
extension,
immediately.
see
are
four operators
= E1
that
G1
+ G1
this
= E2
Applying f
paragraph)
to this
= f(E2 )+f(G1 ) so
f(E1)+f(G2 )
new
To
such that H
G2
It follows that E1
equation,
this
f(G)
definition is unique
= f(E2 )
full linearity
can be checked
as far as the
to
are two
linear functional
on
OD
we
things
that
are
significant
to Gleasons
to bar the
same
has
gotten
into
much of the
the
habit
of
thinking that
18
there is nothing particularly quantum
mechanical
because
about
a single
qubit.[52] Indeed,
orthogonal projectors
on H2 can be
mapped onto
orthogonal
antipodes projectors
of the Bloch
on sphere,
H2
it is
known that the measurement-outcome statistics
can be
noncontextual
as ones
one
potential measurements.
20
number
goes awry
when
we go the
from
fs additivity. Suppose for two positive
semi-definite
the
a positive
semi-definite
= G and
let
an
be an increasing
,
.
(28)
which implies
calculus,
can consistently
f(cE)
(30)
= cf(E)
define
we
Reworking
the extensions
of f in the last
straightforward
basic
notion
of
measurement
is significantly
add
extra
measurement
say
to
restrictions
when
they
the
notion
to practical
of
the route
only make
usage more
if it stands
can
ask
noncontextuality
assignments)
property
that
both
(of
intervention
into the
original
use of
as a kind of
world, one might
version make
invasive
probability
Gleasons
expect the
very
opposite.
why measurement
upon
context
the whole
interaction.
probabilities
of
the
measurement
})? Is there
any
good
reason
assumption?
20
In fact,
one
in order to derive
of
Kochen
noncolorability
and Specker
3-outcome
only
for
of
POVMs
single
the
qubit. Considering
so-called
trine
or
19
4.1
Noncontextuality
noncontextuality
assignments
that
for measurement
basic
than
measurements
even
the
(i.e.,
probability
outcomes is
particular
that
Noncontextuality bears
of
they
more
structure
be
of
POVMs).
Here is
a way to see
a more
general
worldone
details of quantum
notions
of
systems,
consequences,
notion of a
that, skipping
mechanics,
the
still retains
machines,
actions,
the
and
scientific
agent
performing
those
consequences.
imagine acting on
a system
Take
with
we
one
S and
might colloquially
call measurement
devices
, ,.
we
it
had
mechanics.
the aid of
For the
of quantum mechanics,
outcomes.)
For the
case
of machine
N, let
us
one
takes
can stop
the agents in
to
them
consequences
to
ascribe
probabilities
to
the
probability
distributions,
lying around.
These
probability
subjective
judgments
distributions
about what
a
more. Let us suppose
the labels mk and nk are, at the very least, to
be identified with elements in some master set
F
that is, that there is some kind of
mas
connective
ter set F
glue
for comparing the operation of
one machine to another. This set may even be a
set with further structure, like a vector space or
This is well and good, but it is hardly
of first
concern
same
disparate
in
element
labels mi
will
and nj
F in the master
appearance,
an
construction,
set
and
be. Perhaps
Lucent
one
machine
Technologies
was
while
manufactured
the
other
by
was
one
the
purpose,
assignments
pM
(mi )
pM
6=
namely
the
probability
pN (nj )
(31)
with the
same
(mi )
pN (nj )
(32)
we
tautologyit
why
we
have it is
do have
one
a set
a master
may tell us
of
that
we
we
make
set
over
something
master set
settled
choice
Asking
about physics.
from
of time. Where
particular
another.
a waste
on
are
equally interesting
noncontextuality
21
assignments.
once we
have
is particularly interesting!
21
See
Ref.
[56],
pp.
8688,
and
Ref.
[57]
for
some
20
Le Bureau International des Poids
et Mesures
There
important,
the basic
a Paris
4.2
is
still
one
advantage
notion
further,
particularly
to thinking of POVMs
of measurement
as
in quantum
an appropriately chosen
single POVM one can stop thinking of the
quantum state as a linear operator altogether,
mechanics.
For
with
outcomes
of
the
International
Measures.
Our
22
(potential)
measurement.
meter in
outcomes
operator.
for
Bureau
of
Such
on
hinges
which
com-pletely
informationally
some
and
Here is what I
mean by this.
problem
measurement
for
the
as a probabilistic
to
quantum
standard
of it
respect
with
the
specify
finding
of
unique
density
are
called
measurements
complete
probabilities
we
.
.
consider
a slightly
of the minimal
consisting
of D
complete
linearly independent
=tr(Ed
)now
operators,
thought of
as projections
. Any two
to
distinct
measurement
outcomes
informationally
Do
outcome
statistics
for
minimal
number
The
can
POVM
have
2.
and
this
still
of
be
complete is D
minimal
informationally
complete
simple
= 1,...,D, let
1.For d
= |ej
ihej |,
(33)
where j,too, runs
2.For d
12
= D+ 1,...,
over
12
D(D + 1),let
(|ej i
+ |ek i)(hej |
+ hek |)
(34)
where j < k.
3.Finally, for d
12
12
D(D + 1) + 1,...,D
(|ej i
+ i|ek i)(hej
|
ihek
2,
let
|)
(35)
where again j < k.
operators d into
a proper
POVM. This
can be
operator G defined
by
(36)
d=1
22
papers
21
It is straightforward
positive
linear
1/2
(36), we find
D
G is
with positive
transformation
to
a valid decomposition
1/2d
Eq.
that
(invertible)
1/2XG
establishing
(i.e., Hermitian
definite
eigenvalues)
0, thus
for all |i
>
1/2
of the identity,
(37)
d=1
The operators
1/2d
Ed
G
1/2
(38)
satisfy the conditions
(7), and
linear
moreover,
independence
of
a POVM,
they
of the original d
and
Thus
we
have what
we need.
With
the
existence
of
minimal
informationally
complete POVMs assured, we
can think about the vault in Paris. Let us
suppose from here out that it contains a machine
that enacts a minimal informationally complete
POVM Eh whenever it is used. We reserve the
index h to denote the out-comes of this standard
quantum measurement
Let
us
develop
in classical statistical
this from
Bayesian
point of view.
Whenever
one
has
a quantum system
in
believes
of it
to ascribe
probability
outcomes
of
a subjective
judgment
will obtain
as a result
one
s system
performs
systemone
about which
of
an interaction
measurement
different
consequence
between
Whenever
one
{Ed }
andon that
themac
from
the
standard
quantum measurement
level
understanding,
of
gathering
all
(or evoking)
ones
subjective
a piece
is not of information
If
completely
integer
free
to
already
be
supposes
device has D
outcomes
one is no longer
just any subjective
There are constraints.
D), then
to
make
PSQM
we may
mechanics and
Let
P(h)
measurement
some
from
(for
is
that
quantum
doing
one to update
judgment
of description,
edging toward
is
23
What is important
this change
one
as
now
Then,
from
unity, regardless
away
max
tr(Eh )
max
tr(Eh )
max
(Eh )
max
(G
max
(G
1/2h
1)
1/2)
= max (h
1h
(39)
of G
explicitly.[63] Through
P(h)
12
+ cot
3
4D
<1
(40)
23
22
quantum
Accepting
accepting
that ones
outcomes
of
mechanics
subjective
standard
point
within
the
is,
beliefs
in
space
outcomes
part,
for the
quantum measurement
over
to roughly
(0.79D)
More
generically,
informationally
be bounded
for
any
minimal
away
outcomes.
Thus
something
driving
mechanics
and
Bayesian
probability
standard
For all
quantum
consequences
measurement
h, there
intervention:
are no
conditions
24
But what
terms?
What
is that
is
our
assumption
best
in physical
description
of the
wedge?
Some think they already know the answer,
25
Wither Entanglement?
we know
re-
can no longer
be described in the
one
a representative
of its
viz. by
own.
one that
rep-
24
It
is
at
this
point
quantum mechanics
and
sees
[59]. Hardy
generalization
and
quantum
light of
a certain
how
of
mechanics
to probability
the
present
most crucially
extension
that
differs
mechanics
classical
of
as
probability
is depicted here
theory. It is
we
account
restriction
as a
that
we are
in
working like
crazy to identify.
25
This is not
a spelling
mistake.
23
entanglement
Quantum
captured
the attention of
ac-counts
most
quantum
it
is the
information
computing [64]
has
our
certainly
community.
main
theory
ingredient
and
By
in
quantum
paper has
it, more
been
than
quantum mechanics,
information
the issue of
a preexisting
we are willing
to
quantum
place.
states
What
objective
for
rejecting
reality
I
am alluding
the
in the first
to here
is the
3 and
inequality
the
violations
phenomena
by
quantum
of
the
Bell
mechanics.
something
about
that
The technical
we
is, why do
the
tensor-product
and
HAB
sum
= HA
= HA
certainly
third HAB
of the two
spaces
r
spaces
We could
HAB
= HA
product
ssmascads
nnpro
of
coul
HB dtake
[66]
We
t hei
could
Gratake
But instead
product,
HAB
are
other things.
(41)
There
to obtain
HB
of this question
systems according to
rule?
ways to
innumerable
HA
translation
combine
is
point
HB
we
Why?
it
Could
arise
as
considerations
sectionnamely,
property
the
the
of
from
previous
noncontextuality
The
answer
is
theorem I
am about demonstrate
26
quantum
systems,
measurement
on
selfsame
measurement-outcome
for
probabilities?
from
and
we can
we
have two
a
we can
make
That is to saywalking
first
the
measure
outcome
i
meas ur e{F
{ Ei
}on
and
measure
}
onA,B.
j
on
then
Similarlywalking
measure
outcome j,measure
Ei
Xi
{E
we must
valid POVMs,
Xi
} on A. So that
i
have
Xj
=I
and
X
Fj
=I
on
we
the
have
(42)
and
Xi
=I
Xj
and
Fj
=I
(43)
similar
ways
rule. See
24
now
Suppose
POVM-version
thatjust
of Gleasons
as
with
the
theorem in Section 4
such
a measurement
should
upon
not depend
same
the
assumption
applied to local
we
a function
f :EDA
EDB
(44)
now
on the separate
us suppose there
measurements
[0,1]
such that
Xij
f(Sij )
Xij
(45)
whenever
the Sij
or
Eq.
no
(43).
Note in particular
use
and EDB
The
notion
of
local
on
measurement
the
the ordered
conditions
course,
pairs
of Eqs. (42)
satisfy
Sij
and
(43).
the
side
This, of
one
measurement
sophisticated
can
imaginemore
measurements
multiple ping-pongings
the present
Ref. [69]but
could
involve
as
in
class
is
between A and B
restricted
must
come
from
tensor-product structure.
The theorem
27
linear operator
defined
over
such that
L(E
f(E,F)
= tr
F)
and HB
are
POVM trees,
L
on
HA
(46)
If HA
the field of
does
complex
not numbers
hold, however, if the underlying field is
the real numbers.
The proof
of this
statement
is almost
, ,
instance, for
gE
(F)
a fixed
E and F separately.
E EDA
f(E, F)
(47)
For
define
and
{(I E,Fi ),
(48)
POVMs
on HB
Xi
gI -E
(Fi )+
=1
gE
(G )
gE
(H )
(49)
and
Xi
gI -E
(Fi )+
X
(50)
=1
gE
(G )
gE
=const.
(H )
(51)
(E)
=f(E,F)
(52)
27In
,
,
Namely, that L
is in
fact
a density operator.
This
Ref.
[1],
a significantly
stronger
flat-out
was a
25
and
prove
a linear
on HA
uniquely to
operators
functional
can
can be extended
on the Hermitian
way to give a
Namely, for
any two
B on
HA
2 E2
and B
such that
EDB
F2
and HB
= 1
respectively, let A
F1 2 F2
1 ,2 ,1 ,2
gE1
(B)
gE2
(53)
see
To
EDA
(B)
= 1
E1
2 E2
(54)
Then
we have
f(A,B)
=
=
=
=
1 hF1 (A)
1
1
1
1
1 g E1
(B) 2 g E2
(B)
,B) 2 f(E2
,B)
1 f( E1
,F1)
1 1 f( E1,F1
) 2 f( E2
,F2 )
2 1 f( E1,F2
) 2 f( E2
hF2
(A)
f(E1,F1 ) 2 f(E2
f(E1,F2 ) 2 f(E2
,F1)
,F2
and F1
other decomposition
A
,.
E1
be decompositions
0, E1,E2
Then define
f(A, B)
= 1
any
1 gE1 (B)
gE2
(B)
(55)
which is as desired.
,
, .
,.
we have essentially
established,
be
a complete
on HA and
a complete basis for
=1,...,D
let {Fj }, j
be
i Ei
f(E,F)
and F
Xij
Pj
j Fj
i j
on HB
If E
then
f(Ei ,Fj )
(56)
Define L to be a linear operator
L(Ei
tr
Fj
DB
= f(Ei
on HA
HB
linear equations
,Fj
(57)
Such
have,
an operator
we
f(E,F)
i j
L(E
tr
F)
L(Ei
Xij
tr
Fj
(58)
spaces
HA and HB
the
Fj }
14
DA DB
(DA
+ 1)(DB + 1)
(59)
26
equations will appear inEq. (57), whereas
12
DA DB
(DA
+ 1)
DB
(60)
are
an
L. For
= DB = 2. Then
Eq. (59)
to uniquely
needed
instance
take DA
specify
nice if
full
we
could
probability
go
rule
measurementsi.e.,
our
operator. Unfor-tunately,
strong
enough
counterexample
for
LS
(E F)
assumptions
density
are not
is a
a linear operator
to the swap operator on the
that.
spaces:
quantum
be
Here
[72] Consider
that is proportional
two Hilbert
local
for
that L must
F E
D
(61)
This
clearly
satisfies
the
conditions
of
our
to
density
operator.
Of
course, one
by requiring
that
could
recover
it give
positivity
positive
(i.e., resolutions
HB ).But in
would be a cheap
good
here, that
should ask in
conscience
nonlocal
for L
probabilities
:Why
measurements)
(including
should
it
correspond
identity
that
Alternatively,
quantum mechanics.
density
operator.
noted
L to beabove,
function
This
is uniquely
f(E, F); we
never
the
functi
than
theon
probabilities
outcomes in specifying
condition supplies
is because
determined
L,
by
as
the
of local measurements
L. Ferreting
an avenue
out such
All of this
L is, it must
away
be a
and
does
not, however,
HB
on
operator
linear
Therefore, let
striking
feature
tensor-product
us
close by emphasizing
way
of this
rule
take
deriving
of
the
separate
combining
for
the
or
nonlocal
about
resembling passion at
one
did not
even
have
a distance
assignments
[73]
Indeed,
to consider probability
of measurements
an example on
H3
H3
|1
|2
ihd
|},d
variety
rule. That
one need
like {Ed
= 1,...,9, where
=
i
= |0i|0 + 1i
i |1i|1i
|6
i |1+
|7
i |1
2i|0i
|3
i |0i|0 1i
2i|0i
(62)
|4
i |2i|1+ 2i
=
=
+ 1i|2i
|8
i |0
|5
i |2i|1 2i
1i|2i
|9
an
on
H3
and |0
1i
1
2
i |0
orthonormal
(|0i
|1i), etc.
is
quantum
entanglementit
measured
by
elaborate
ping-ponging
local
still
means,
cannot
be
with
the
even
strategies
mentioned
earlier.
seems to
more a
quantum
entanglement,
statement
the probability
rule, is
more a
be
product
of the
structure of the
27
observablesthat
with
they
noncontextuality.
are
POVMscombined
In searching
for
the
secret
to
ingredient
general
Bayesian
quantum mechanics,
drive
wedge
probability
it
seems
between
theory
and
quantum
entanglement.
And
so you see
as true
I
have
come to doubt
I
stand alone without beliefs
28
Allthat I
once held
The only truth I
know
is you.
Paul Simon,
timeless
states
Quantum
knowledge,
are states
not
commitments,
of information,
pragmatic
belief,
states
of
gambling
nature.
That
paper. Thus,
sense of the remainder of
one strategy ought to be to
quantum mechanics,
seek
guidance
avenue
of
Bayesian
[74] from
probability
most
the
rational-decision
theory
theory
[75,
developed
to date
76,
77]
Indeed, the
very
develop
reliable
making
decisions
reasoning
of
in the light
and
of incomplete
quantum
mechanics?
independent
To what
structure
extent
are
a distinction into
an analogy. The core of the matter is the manner
in which states of belief are updated in the two
theories. At first sight, they appear to be quite
different in character. To see this, let us first
This section is about turning
explore
states change
accounts
of quantum
mechanics,
information,
however,
it has become
measurement,
exact
nature
may
the measurement
of
interaction,
The
change
be
broadest
arises
opera-tions
consistent
in the theory
[50]
operator
measured
on
of state
of effects
and
that
a quantum system is a
a POVM {Ed } is
system.
opera torThen,
an d
according
a POVM {E
to
notion
for
and
of the form
Xi
di
Adi A
tr(Ed )
(63)
where
Xi
di
(64)
Adi
=Ed
There is no constraint
on the number
of indices i
even be
Hermitian.
28
This is not
a spelling
mistake.
28
The usual
justification
as
generalityjust
commonplace
than
as a
arises in
an
an
of
the
ancilla
of
the
POVM
that the
with
case
for
about by imagining
interact
the
justification
formalismcomes
measurement
in
observation.
In
system is pictured to
first, and
only
then
primary system.
More
formally,
we can start
out
by
(11)
we must
make
of
Namely,
projection-postulate-at-a-distance.
one takes
(I
d )U(S
trA
)U
(I
d )
P(d)
(65)
The
eigenvectors
Then S
S A
(66)
can be written as
|a iS ha |
29
ha
(S
|(I d )U
)U(I d )|a
P(d)
p
ha
|(I d )U
P(d)
ha
|U(I d )|a
|a
(67)
A representation of the form inEq. (63)
can be
made by taking
Ab
ha
|U(I d )|a
(68)
i.Indeed,
30
one can
easily
check
what
we
as with
the
This completes
set of Adi
find
a U, a A
course
Of
contained
There,
old
collapse
and set
postulate
case:
special
the
one
takes
both sets
= Ed } to be sets of orthogonal
us
case
operators. Let
al
think
projection
about this special
is distinctive
just
is
as a
take
a moment to
in isolation. What
29
this point
inRef. [78].
30
As
an aside,
For
are many
precise
statement
of the
latitude
of
this
freedom,
see
Ref. [79].
29
be almost orthogonal to each other. In
density-operator language, there is no sense in
which i is contained in : the two states
6=
P(i)i
space.
are in
That is,
(69)
Contrast this with the description of
probability
theory. There,
belief is captured by
an initial state of
a probability
a piece
distribution
way gathering
assigning ones
is through
one expands
P(h)
Xd
P(h,d)
Xd
P(d)P(h|d)
(70)
P(h)
P(h|d)
(71)
P(h|d)
(72)
How
P(h,d)
P(d)
gentle
quantum
this
collapse!
information,
one
looks
in
When
comparison
one
gathers
to
new
senses.
It is not
as if the
collapse
not look
more
rule? Is quantum
like Bayes
more
collapse
really
a
an
or might it be
a problematic representation? By this
stage, it should come as no surprise to the reader
that dropping the ancilla from our image of
violent kind of change,
artifact of
generalized
progress.
in
of what
quantum
starting
point.
To accentuate
similarity
we can conveniently
1
(73)
where
Ed
Ad
(74)
In a loose
given
of this sort
Ad A
P(d)
interaction with
viewed
as
more
away some
i-dependence
a measurement
truly
of the information
more
be
of
us
be expected to decrease
can be
it gained. Let
precise.
that Bayes
data. This
may
finer-grained
Notice
an explicit
upon
can
hypothesis
the acquisition of
Xh
P(h)logP(h)
(75)
30
This
number
is bounded
between
0 and
the
important
of
these
is that
number of binary-valued
it
as a good
the
it quantifies
questions
most
the
one expects
the
lower
predictable
the
means to
colleague who
entropy,
Shannon
more
the
a measurements outcomes.
Because
the
function
f(x)
xlogx
is
S(H)
Xd
Xd
Xh
P(d)P(h|d)
Xd
log
P(d)P(h|d)
Xh
X
P(d)
P(h|d)logP(h|d)
Xd
P(d)S(H|d)
X
(76)
Indeed
we
hope to find
a similar statement
measureor impredictability.
by a decrease of
ments
But,
decrease
what
can
uncertainty
have
argued
uncertainty
be
through
meant
quantum
strenuously
that
measurement?
the information
gain in
about
be information
preexisting
way out
reality. The
expects
for
the
of the
that decreases
one
is the uncertainty
results
of
other
potential
measurements.
There
that
the
:
D
S() X trlog
(77)
logk
k=1
where the k
of . (We
measurement
one-dimensional
Shannon
P consisting
orthogonal
entropy
for
the
projectors
outcomes
.
of
of
The
this
measurement is given by
=
H(P) =
(78)
Xi=
Xi=1
H(P)
(tri )log(tri )
Xi=1
operator
will
As it turns out, the
set
answer
of eigenprojectors
orms
obtains,
a setthe
outcomes
entropy
is
for [82]
When
this
of
to
, which measurement
the
density
Neumann entropy,
von
operator.
Neumann
The
von
way of a
standard measurement in a best case scenario.
Indeed, true to ones intuition, one has the most
predictability by this account when is a pure
statefor then S() = 0. Alternatively, one has
impredictability
one
achieves
by
identity
operatorfor
any measurement
are all equally likely.
then
The
best
however,
is
indicative
of
can
measurement, it
the
vast
considerations.
31
This leads to a second method for quantifying
we
as our
uncertainty
again rely
on
the Shannon
information
we
evaluate
measurement
it
with
respect
to
typical
one.
projectors
that
is
invariant
turn induces
canonical
measure
dP
on
the
space
of
Using this
P [83]
quantity
S()
H()dP
(tr)log(tr)d
(79)
which is intimately connected to the so-called
entropy
can be evaluated
?12
on the expression
13
+ +
ln2
+ Q()
(80)
where the subentropy Q() is defined by
k
Q()
k=1
i6=k
logk
k
mean
explicitly in terms of
S()
This
(81)
In the
case
= m
+ and m
em
0. to
The
for l6=
m, one
eigenvalues,
limit is convergent
and hence
Q()
is
as
finite
for a pure state , Q() vanishes. Furthermore,
since
we
know
that
?12
1
0 Q() logD
+ +
ln2
ln2
(82)
any
, Q()
never
means
exceeds approximately
that for
0.60995
bits.
The
interpretation
of
this
result
is the
when
one has maximal
a quantum systemi.e., one
has a pure state for itone can predict almost
nothing
about
the outcome
of a typical
following.
information
measurement
Even
about
[40]
a little
value.
reduces
captured
by the function
a quantification
becomes
of uncertainty in its
own
right.
F(t 0
+ (1
t) 1 ) tF( 0 ) + (1 t)F( 1 )
(83)
for any density operators 0 and 1 and
F()
F()
(84)
32
P(d)F(d )
one
any
real
arise
case
emphasized),
Xd
6=
P(d)d
(85)
for d defined
1/2
1/2
Xd
1/2
Ed
1/2
Xd
P(d) d
(86)
where
1/2
P(d)
Ed
1/2
1/2
Ad
P(d)
1/2
(87)
of is
that
for each d, d
eigenvalues.
have the
same
In the present
and d
This follows
since X
eigenvalues, for
case, setting
any operator
= Ad
1/2
X.
does the
depend only
upon
Xd
S()
P(d)S(d )
(88)
Xd
Q()
P(d)Q(d )
(89)
as we
had
an
performing
of
been
a POVM
{Ed
hoping
efficient
for. Thus,
quantum
}, an observer
in
measurement
can expect to
In this
have
sense, quantum
expect
does indeed
we can
same
collapse
be
31
and d
have the
us
once
(74). A
their meaning is
1/2
Ud E
P(d)
1/2
(90)
where Ud
only
to
is
a unitary operator.
the
constraint
Sincesubject
of
efficiencythe
Ad
1/2
P(d)
1/2
(91)
Then,
subject
measurement
to
the
interaction
details
and
of
the
the
particular
a further kind
on the measured
of back-action
system[88]:
or
Ud
d U
(92)
31
By differing
terms of
methods,
a strengthening
of this result in
in Refs. [85]
and [86].
33
But this breakdown of the transition is a purely
game.
the Ud are arbitrary to begin with, we
conceptual
Since
data collected:
Xd
P(d) d
(93)
(94)
Finally,
may
there
takes
account
into
be
measurement
details
interaction
beliefs, which
both
and
mental
further
of the observers
readjustment
the
of
the
observers
some
is enacted via
Vd d
(95)
Putting the two
same
result
processes
together,
one
has the
The resemblance
between the
process
in Eq.
uncanny.
32
By
this
way
of
viewing
things,
violent
more,
nothing
refinement
and
and
nothing
readjustment
less,
than
of ones
initial
away
come
step of coarse-graining
information that
was
(i.e.
throwing
inprinciple accessible)
us
Let
maximally
account,
it.This
1/2
Ed
In the
first,
observer
is
cases of efficient
we imagine an
measurements.
tight
|ih|
1/2
P(d)|ih|
(96)
The only state
from
change
a measurement
we just change
consequence
of
is
a sense
can come
must be purely
about
of the
new;
as a
the side effects
of our
intervention. That is to say, there
mental-readjustment
experimental
that
what
we can
in which the
predict
measurement
is solely
mechanism
{i
(97)
von
and the
Neumann
corresponds
to
= |iihi|}
is the
an
spanned by |i is
At
the
opposite
at
all
that
causing
of
we can
have
no
of things,
measurements
contemplate
possibility
end
physical
could
update
such
is purely
mental
further
AB
AB
Xi
variety
33
readjustment).
a pure state
consider
of the refinement
pi
|ai i|bi i
(with
For
no
instance,
iwhose
Schmidt
(98)
32
Other
Bayesian
similarities
between
conditionalization
quantum
collapse
and
in Refs.
minimally
disturbing
our case, a
minimal
measurement
disturbance
of
some
version
of
POVM. In
POVM
{Ed }corresponds
In
the
usual
corresponds
to taking Vd
presentationsee
to taking
Ud
=I
for all d in Eq. (95)
Refs.
[85]
and
[88]it
I
for all d in Eq. (92)
instead. For
34
An efficient measurement on the A side of this
leads to a state update of the form
AB
AB
ih
Ad
trA
(Ad I)|
Xijk
qj
AB
ih
I|
AB
|(A
AB
I)
AB
ih
|A
=
=
=
=
=
pk
(99)
I|ai i
qj
qj
Xijk
Xijk
pk
pk
hak |A
qj
Xjk
pk
hak |A
qj
Xjk
pk
pk
hbk |UA
qj
Xjk
Xjk
Ad U
UA
Ad U
i|bj ihbk |
T
d
1/2
i|bj ihbk |
T|bk
UA
pk
hbj |
|bj
Ad U
1/2
(100)
state
on
the B
operator connecting
the
represents taking
|ai
i
a transpose
with respect to the |bi ibasis. Since
the operators
Fd
UA
(101)
go
UA
T
d
Ad U
together to form
POVM,
we
indeed have
out to be rather
collapse
get
feel
for
noncontextuality
classical Bayesian
further
partitioned
we
have just
for Gleasons
In the setting of
conditionalization
reason
assumption.
one starts to
rest
the
care
about how
of
the
we
potential
care whether d is
a two outcome set {d, d} or
embedded in a three outcome set,
in
whether it is
from
to
is of the
form
17
or
P(0)0
of the form
(102)
d=1
as long as
17
P(18)18
P(d)d
= P(0)0
P(d)d
P(18)18
(103)
d=1
E
The action of
1/2
d
be consistent
relation, to
additional back-action,
the
system.
reasonably
...
anunnecessary
perturbation of
is
This of course
can
means
as a state-independent
measurement. Given
the minimally
one that
an initial state
and
a POVM
{Ed },
imally
disturbing
disturbing measurement
produces
pure Bayesian
interaction is the
updating with
no further
35
What could be
a simpler
generalization of Bayes
rule?
Indeed, leaning
on
that,
we can restate
problem
the
at the
more
technical
(72)
where
an agent
has
probability
Marginal- izing
over
the possibilities
for d,
one
hypothesis
h. If he gathers
data d, he should
use
an
explicit piece of
(104)
init.The only
problem, let
have
a detailed theory of
we can trust inthe
course
not! Bayes
a tool
a law of
way
an agent
should
reason
accept
as being
physical world.
36
a classical measurement
any I-know-not-what that induces an
theory is concerned,
simply
is
comes
about
This point
was recently
stated much
paper
Inclassical mechanics
nothing strange. It is merely
more
eloquently
Ref. [92]:
a measurement
anevent
is
some
physical system.
mean? What
causes
it? And
information
so on. These
now
they
seem tautological
new
has received
us that
information simply
the state of
means
case
say an observer
a classical system,
new
precise measurements
are not
case, since
possible inpractice).
new
...
information.
is no different
system to obtain
that he
Lets
our
the change
a measurement on the
then differs
...
words,
information.
a measurement
disturbs
the observers
information
crosses
where
a problem:
a philosophical
question
case
information also
passes
one could
from
again
no more
problematic
inquantum
35
can be downloaded
from
edu/.
36
even string
theory.
36
The formal similarities
and quantum collapse
finally
cut
the
measurement
telling
us that
may
us how to
knot
of
the
Namely,
it
may
Gordian
problem.
it is simply not
be
a problem at all
Indeed, drawing
two theories,
one
on
the analogies
between the
a spark
is left with
of insight:
is simply
law of
thought
[56]. Perhaps
the
way
some
reason
and make
fundamental
decisions in light of
situationa
fundamental
situation
waiting to be ferreted
fashion.
This much
we
know:
That
it
ismust
situationwhatever
ingredient
have.
Bayesian probability
As already emphasized,
something
to drive
theories.
Probability
general
a structure.
be
an
there must be
alone
is
too
6.1
fundamental
us.
mean to accept
quantum mechanics? Does it mean accepting (in
essence) the existence of an expert whose
let
us
ask: What
does
it
we should
probabilities
we
to answering
this question
considerations
comes
the
of
from combining
standard
this
than
further
will allow
calling
us to go even
collapse
a
quantum
quantum measurement
there,
one starts
standard
tr(Eh )
(105)
and,
after
Pd
(h)
= tr(d
Eh )
(106)
where
d
Fh
(107)
Vd
Eh Vd
of some
other
up with a final judgment
measurement
= tr( d
Eh Vd
= tr( d
Fh )
two entirely
Note t ha t,
different
in genera
POVMs;
l,{Eh }and
the range
{Ed }ref
of
indices h and d need not
their
Also, since
complete
informationally
even
minimal
d
{F h }
POVM,
will
be the
same.
informationally
itself
be
Thus, modulo
a final
the
redefinition
of
measurement
based
precisely
{Eh } is
Bayes
unitary readjustment
standard
on the data
rule in this
gathered,
one has
transition.
This
follows since
Xd
P(d) d
implies
P(h)
Xd
P(d)P(h|d)
where
P(h|d)
= tr( d
Another
Eh )
or
quantum
is
(108)
(109)
(110)
point of view, i.e., that the axes
of the probability simplex
are held
fixed, while
is,writing
Fh
(111)
0
h =1
37
dhh
0
Eh 0
is any
I-know-not-what
that generates
an application
of
decomposition
combination
choice
is
a convex
then a final
of other
states
and
quantum
disturbing
measurements
are
invasive
a further outcome-dependent
One can either
think of it
only in introducing
readjustment:
or
picture
as a
passively
quantum
depicted
readjustment
measurement
as a
here)
the
of
or
device,
actively
adjustment
further
standard
(as
to the
posterior state.
where
{Eh
are some
to a
dhh
0
} refers
Pd
(h)
dhh
0
of
the
P(h
original
we
measurement,
and
coefficients
relabeling
quantum
standard
real-valued
get
0|d)
(112)
0
h =1
This gives
an
Modulo
the
measurement
whatsoever
is
on
is in Bayesian terms.
a quantum
readjustment,
final
any
application
of Bayes rule
By
any
any convex
decomposition
refinements
P(h|d)
measurement
measurement:
is
that
also
live
readjustment,
just
It is
of P(h)
like
any
into
some
in PSQM
37
a quantum
classical
I-know-not- what
that
an agent to an
pushes
rule.
application
of Bayes
38
Accepting
mechanics
large
isin
of quantum
.
,
accepting
partsimply
that
up to
by a
Moreover,
signified
measurement
37
Note
calling
d),
between this
distribution
should
not
existence
get
that
simultaneously
the
intrinsic
feeling
the
course,
from
random
rule
of posing Bayes
no status to a
one
P(h,
a mathematical
one
In particular,
P(h, d)s
variables
on
insists
distribution
but it is only
of
experimental
I
fear the wrath
I-know-not-what
be claimedI
PSQM
mathematical
and
consequences
Of
joint
meaning.
way
it,I
give
P(h, d). If
P(d)P(h|d)
Vd
simply
way. Instating
the product
artifice
P(h,d),
set
operator
unitary
a distinction
probability
convex
the
is
joint
outside
falls
not understand
my
can see
it
interventions
there is no hand
into
no d.
now, rather
it will
violentlythat
I
do
a mechanism
of
for
my
to leave
nothing
emphasized
taskand
intrinsic
comes
it
The
cannot
use
a piece
here,
is that
be the taskof
to justify
how
that
as
already
it is not
theory
of information
of probability
gotten hold of
point
case,
in the classical
that
an agent
causes
the
makes
has
him to
38
anything that
of
Bayes
more
rule
of Bayes
than arbitrary
to
ground
the
can
measurement
is performed?
Surely
that
much
Thats
territory
Im
to follow
you
into.
Good luck!
Suppose
adamantly to the
is, that
there
is
no
right
additional
and
true
quantum state
to the
quantum
who
is
interested
in the system
walks
why.
Take
performs
{d }
the measurement
of which is accompanied
be d
= d
regardless
this state-change
the
or
device
systemi.e.,
rule is
its
it has
feature of
with
the
to do with the
nothing
subjective
observers
an objective
interaction
judgmentthen
an objective
the final
feature of the
change
it clearly
operators
without
to all state
generalizes
are
the Ad
any
adding
rank-one
further
complications.
through
Ad
even the Ed
39
support
the maximal
=Ud
themselves
For otherwise,
one
Ed
control
are
it must be that
subjective judgments.
would have
a statement
like,
are correct.
Allother states
40
Thinking
evolution
state
after
the
single-element
POVM
{I}
is
performed, stat
one e
isafte
forced
rthe
to the
single
same
-element
conclusion
even
in
that
super-operator
for
evolution
systemmost
linear
argument
powerful
I
viewed
subjectivitythe
to be the most
quantum
the
for
Section 3. Since
we can
states
argument
Einsteinian
toggle
of
the quantum
simple
quantum
bipartite
quantum
system
controlled
unitary operation Ui
us say
on a
a
circuit
that
performs
on the target
bit.
=
39
The support
of
an operator
Such
a statement,
nonzero
is the subspace
spanned
eigenvalues.
in fact, is not
so
dissimilar
to the
one
see
39
of that idea,
Figure 3: One
Einsteins
can use a
argument
slight modification
quan-tum state to
draw the
quantum
evolutions.
time
same
conclusion
By
of
of the
for
performing
measurements
ascribe
to
the
Therefore,
evolution
accepting
of
the
left-most
qubit.
evolution
be
a property
one finds
|ii|i |ii(Ui |i)
(113)
for the overall evolution. Consequently
the
(114)
On the other hand,
prepared
+ |1i.
in
suppose
superposition
state |i
a completely
positive
map
is,
||
|0i
be given by
|i
U1|ih|U
Now,
(|ih|)
= ||
That
U0 |ih|U
(115)
rather
than
feeding
feed half
we
an
entangled
an
observer
of the whole
a measurement on
the
any
maps on
upon which
evolution
mapping
is subjective information.
41
Of
try
course,
to get
could
argue
are
there
around
evolution
specified
opinion,
an effective
by the complete
however,
the Everettic
Einsteinian
conclusion.
For
true
sideways
this
moves
temple.
argument
toggles nothing
evolution
like this
One
in the
could
my
just prostrations
to
dismiss
same way:
evolution
[97] In
circuit.
are
on the
for it.The
is the unitary
quantum
one
instance,
operator
the original
The
measurement;
observer
the true
on in a quantum measurement
is the revelation
of
relative
statei.e.
the effective
quantum
most productive
stance
say
state.
How
it is not the
shows that
the
of this argument
A strengthening
same way as
may
also be found in
the
many
as an exercise
40
It has long been known
preserving
over a
in a
completely
one-to-one
on a
the trace
linear maps
space can be placed
positive
D-dimensional vector
operators
that
correspondence
2-dimensional
with
space
density
via the
=I
(|ME
ihME
(116)
where |ME i
signifies
state
on HD
HD
|)
a maximally
entangled
|ME
Xi=1
|ii|ii
(117)
This
is
usually
representation
treated
as a
convenient
mathematical
accident
no
ascribes
measurement
a measurement
finds
a mathematical expression
of some agents
difficulty. When one agent
a piece of data d,he might
contemplates
viewing
be willing to
use
according to
P(h)
Xd
X
P(d)P(h|d)
(118)
from thinking
the
same
Q(h)
data warrants
Q(d)Q(h|d)
him to
(119)
A priori, there need be
no
A relation
only
comes
one
when
are drawing
the
same
seeks
say
that
meaning
way
for
see
the world
as
it completely
The
When
statistical
both
model
agents
accept
the
same
of the
data d given
to the identity
say
they
considering.
are
I.e.
performing
,
.
of the measurements
each (separately)
will
agree
they are
= Q(d|h)
(120)
two agents
the
same
h and d
Putting this in a
Xd
P(d)P(h|d)
P(h)
Xd
Q(d)Q(h|d)
Q(h)
(121)
with which to describe it.
that
generates
Xd
P(d) d
(122)
41
But again, there is nothing to stop another
first:
Xd
X
Q(d) d
(123)
no further agreement.
as inthe classical case, however,
Just
there
Xd
P(d)
1/2
1/2
Xd
Xd
Q(d)
1/2
1/2
(124)
way, two
on the identity of
assign it the same
a measurement
when they
POVM {Ed },
Ed
= P(d) 1/2 d
Q(d)
1/2
1/2
1/2
=
(125)
The importance
it draws
operators
Ed
of this
from
the
present
perspective.
P(d|h).
More
generally,
that
role
P(d|h)
(126)
That is to say, drawing the correspondence in
different terms,
is
P(d|h)
= Ud
d ()
1/2
1/2
Ed
(127)
(Of
course,
more
measurementsd
sum of
generallyfor
may
()
consist
nonefficient
of
a convex
such terms.)
The completely
mathematical
positive
description
a map.
map
to
that
gives
quantum
time
subjectiveis just
statistical
such amodel
map. Its
P(d|h),
role is
where
that of
d just
happens
on entanglement,
it
seems
cannot
What
the wedge
we see
we are
instead
looking
for either.
is that there is
a secret
will
very
likely
quantum measurements.
Intermission
us as much about
as quantum states and
tell
quantum
stateswhatever
they arecannot
be
state is
as a state
if one were
way to think
of
a quantum
a standard
away in a vault in
Paris. Section 5 argued that even our hallowed
quantum
entanglement
is
a secondary and
subjective effect. Section 6 argued that all a
measurement is is just an arbitrary application of
happen
measurement
Bayes
to
device
rulean
beliefsalong
measurements
ever
arbitrary
with
are
approach
locked
refinement
some
invasive
account
of ones
that
interventions
into
even quantum
time
42
evolutions
are
so
it went.
...
And,
It is
not something
sake. There
is
go away.
But subjectivity
to be worshipped
are
so
a
is
for its
own
we
need
element
objective,
and call
I
turn now to a more constructive phase.
a little
provocatively, but
The abandonment
. .,
of superstitious
or Fairies
Ether, Absolute
and Witches,
ex-
Space
too, if regarded
as something
objective existence, is
no
less
endowed with
a misleading
some
kind of
conception,
an
prob-
abilistic beliefs.
The hint of
a more
fruitful direction
most
commonly
information
theory
phrases
from
is
hardly
information
that
a paper
not
does
Unknown
quantum
protected
with quantum
[101],
be
of the
quantum
Bayesian
and
used
to
But
what
information-based
can
check
the
use
make
are
error
states
quantum
in
of it
teleported
[23]
correcting
codes
quantum
for
day.
in
can
one
of
perspective.
There
used
sense
uses grows
term mean?
interpretation
of
each
In
an
quantum
mechanics, it is
by their
information
the
very
least,
by the
invokes the
formulation,
more
for
a complete
still awaits
any
analysis)
quantum
state
in its
means that
an unknown
we should
This
description,
is merely
belief
no
further agent
or
then
a way must
be
2. If there is clearly
observer
on
the
scene,
found
to
reexpress
the
phenomenon
with the
(In this
case,
the end-product
for
describing
state
overall
that
the phenomenonnamely,
actually
captures
the
the
describers
experimental
tomography
[31]
tomography is
a project
practice
The
usual
quantum-state
description
of
43
optical medium
of
driven
wishes to characterize
or to
perform
measurements
description
precise
tomography
of
state
experimenter
is unknown.
is
measurements,
to
(on
enough
that
the
perform
and
measurements
assumes
kinds
of
to
an example
where there is
no
state
entirely
artificial:
Where
would
the
inside
of
placed?
On
tomographer
only
device
be
the
is trying
to characterize?
43
course
is the second
strategy
available
aboveto
the player
the
The
To do this, we
once
Bayesian probability
phasized
previously,
probabilitiesjust
objective
of
again take
Bayesian
belief,
theory
reflecting
ones
not
measures
operational
commitments
light of this, it
most
the
comes as no surprise
overarching
Bayesian
that
one
of
is to
themes
a set of
decision-making agents can come to a common
belief or probability assignment for a random
variable even though their initial beliefs may
the conditions
identify
differ[77]. Following
understanding
the
under
which
essence
of
Bayesian
theory
quantum-state
tomography.
Indeed, classical
almost
precisely
same
This
egregious
as unknown state.
The procedure
tomography
an
this
analogous
every
in its
bit
as
to quantum-state
unknown probability
repeated trials
The
as our
probability
an oxymoron
domain.
of
encounters
problem
unknown
use
the
on identically
way to eliminate
situation was
unknown probabilities
de
was
to focus
on
the
by
from
Bruno
simply
introduced
equivalence
of
the
repeated
trials
important
is
that
are
any
that
as
far
concerned.
probability
what
assignment
is really
are
systems
the
as
indistinguishable
predictions
namely,
probabilistic
, ,...,
Because
p(x1
x2
xN
for multiple
trials should
permutation
conceptual shift
use
it
may
representation
multi-trial
effect.
theorem,
probability
permutation-symmetric
number
of
multi-trial
equivalent
be symmetric
sound, de Finetti
to powerful
trialsde
he
an
probability
was
with
showed
as
this
able to
his
any
that
called
under
that
arbitrarily
Finetti
probabilities
to
For,
assignment
for
this,
of
is
large
such
exchangeableis
probabilities.
..., ...,
independent
43
Placing
the
as
respectable
anonymous
here
would
Berkeleys
be
famous
about
as
patch to his
philosophical
engagingly
player
George
by
limerick
Ronald
of
Knox
and
its
reply:
There
think it exceedingly
Continues
odd
to be
:
If he finds
:
When
:
Must
:
no one
theres
about in
:
I
am
:
Will continue
:
And
to be, :
Since
44
identically distributed distribution
p(x1,x2
=p
n1
1
p2
The numbers
nk
k
=px1
,...,xN
n2
px2 pxN
(128)
,..., .
pj
describe
will be j (j
=1
other
hand,
k)
describes
the objective,
probability
a single
The variable nj
the
number
true
experiment
of
on the
times
, ,...,
x2
xN
the objectivistonly
a kind
,...,
of Gods eye
probabilities
p1
pk
will
very
often be
the
unknown
probabilities
by
outcomes.
, ,...,
probability. Instead,
measurement
of his
in the light
of de
,x(2) ,...,x(N)
= p(x1
,x2 ,...,xN
(129)
for
any
permutation
that for
pN+M
any
(x1,x2
and
)with the
same
p(x1,x2
,...,xN
(130)
xN+1 ,...,xN+M
,...,xN
the form
p(x1,x2
,...,xN
ZSk
pxN
d~ p
ZSk
n1
n2
P(~ p)p 1 p 2
nk
k
d~ p
(131)
where
p:
pj
Sk
pj
j=1
(132)
=1
is
be
a probability
ZSk
P(~ p)d~ p
density function
on the simplex:
=1
(133)
With
this
representation
concept
unsatisfactory
probability
vanishes
an
from the
theorem,
of
the
unknown
description
in
idea of assigning
an
distribution to multiple
probability
trials.
With this
to
reword
cue
the
in hand, it is
description
tomog-raphy to meet
is
simply
experimenter
our
judgment
notice
easy to see
of
quantum-state
on
the
part
the essential
distinction
between
preparing.
In operational
same as
far
as
how
the
of
subjective
there is
the device
terms,
this
is
no
is
the
observational
might
seem to be
use
the
notion
of
an
one
need
never
statea
unknown
45
enough.
Putting
statement
it into technical
terms, the
judges
judge
any
permutation
of
those
(N).
no matter
only by the
consistency
(N)
be derivable
complete story.
tomography,
as
but
state
appear
in this
is
no
longer
any
(N)
(if
his judgment
no one
More
importantly,
the experimenter
an
make
structure of the
Each
has
the
of
whole
states
is in a position to
statement
unambiguous
sequence
about
of states
the
(N):
(N)
kind of permutation
invariance
over
its
theorem[32,
105]
is that
a
sequence of states (N) can have these properties,
which are said to make it an exchangeable
sequence, if and only if each term init can also be
representation
(N)
P()
(134)
where
is
an N-fold
space
ZDD
DD
distribution
and
P()d
=1
over
(135)
where d is a suitable
The interpretive
measure.
import
of this theorem
is
mandate to the
act
as if
It
says
his judgment
he knows there is
can
comes
(N)
a man
about because
can say
about
the unknown
state is
makes
a connection to the
Bayesianism
stressed
two independent
overarching
theme of
above. It guarantees
observersas
long
as they
for
have
set of measurements
sufficiently
will force
informative
a convergence
in
by
This minimal
agreement is characterized
sequence
the
of
described
above,
observers
are not
systems
and
though P()
can never
may
promise
absolutely
opinions. Quantitatively,
the latter
the
in their
means that
zero, it
be arbitrarily close to
vanish.
exchangeable
rule
gathered
quantum
by another
for
updating
an
be
on K systems
from
version
probabilities
of
[33]
yield
probability
P(|DK )
constructed
sequence can
information
Specifically, if measurements
results DK
works because
density operator
to reflect
measurements
Bayess
that
inflexible
updated
as
is exchangeable,
P(DK |)P()
P(DK )
(136)
using
measurement
results
DK
an updated
given by
to obtain the
ZDD
P(DK )
P(DK |)P()d
(137)
46
is
the
unconditional
measurement
informative
probability
a
measurements, as
results.
set of
For
on a
for
K becomes
P(|DK )becomes
particular state DK
dictated
by the measurement
results, regardless
as long as P()
in
neighborhood
observers
have
encapsulated
of DK
different
in different
the
sufficiently
the two
Suppose
initial
of the
is nonzero
beliefs,
= 1,2.
priors Pi (), i
state
of
belief
in which
additional systems
N
DK
number
i.e.,
Pi (|DK
(138)
are
any
DK
N of
state
independent
on the purpose
quantum-state
tomography:
uncov-ering some
rather
about
the various
over
agreement
unknown
It
quantum
the
observers
coming
future probabilistic
de
about
conclusions
not
is
of
Finetti
to
predictions.
to note that
theorem
and
the
mechanics.
real Hilbert
the framework
of
For quantum
exchangeable
operators
density
and unknown
counterexample
is the following.
(N)
1
2
12
(139)
where
+
1
(I+ 2 )
2
and
12
(I 2 )
(140)
and 1
,,
2
are
and 3
(139)
a valid
to an equally
is clearly
corresponds
quantum
complex-Hilbert-space
mechanics,
operator:
density
Eq.
It
weighted mixture of
y-direction.
state
The
(N)
is
thus
is unique according
theorem.
But
now
consider
real-Hilbert-space
its ostensible
remains
because,
(N)
as an operator
a valid quantum
upon expanding the
2 s cancel
away.
exchangeable
Yet,
density
written in de Finetti
, ,
real symmetric
i2
Finetti
even
number i,it
state.
This
right-hand
an odd
it
side of
number of
though
operator,
is
it is
cannot
an
be
as a linear combination of
and 3
while a real-Hilbert-space de
expansion as in Eq. (134) can only
cannot be written
I, 1
in
contain
those
Finetti
theorem
operators.
three
does
not
Hence
hold
the de
in
In
classical
exchangeability
probability
characterizes
distribution
the
numbers
are
nj in
theory,
those
same
true
is
(131).
to
is informationally
updating
where
an
more can
sequential measurements.
mechanics:
a sufficiently
any one that
sufficient
state to the
exchangeable
nothing
are
complete)
be
The
shows that
in quantum
Frequency
data,
frequency
Eq.
situations
for updating
learned
That is,
one
for
point
from
obtains
systems
independent.
That
Finetti theorem
that there
are
will
in real
be
statistically
no quantum de
Hilbert space means
there
is
fundamental differences
spaces
between
with respect to
a little
essence
quantum-state
tomography
of
is not in revealing
an
47
element of reality but inderiving that various
same
thing
exist.
as being objectively
real to make
sense
of the
we
have
run through
inthis paper
very
nicely:
the starting
can be no theories
without theoreticians;
no learning
no science
learning
particular
point in
time, are necessarily reasoning
processes
may
process
...
..
that
processes.
process
which relate to
occur
might
acase in
a population can
salty
no
10
Quantum
The
insight
significance
of
this
development
possibility
of
is to give
a new
us
and wider
way
it
was
of thought
observer,
the apparatus
we
do not
oc-curring
...
assume any
In the
longer
in the idealizations
an observer
from the
new pattern
the detached
of this
classical
a new
effects creates
new state
observation is
system.
a singling out
now,
here and
...
situation, theoretically
the observed
of
of
a particular
making obvious
described
Inthis
factual result,
possibilities,
aspect
the discontinuous
as a
way every
thereby
of the physical
phenomena.
Nevertheless,
the-ory
any
deny
of
an
there
objective
remains
reality,
as
kind of
theories
possibility
a measurement, once
cho-sen. Particular
new
these
still in the
inasmuch
the experimental
qualities
of
an
arrangement
individual
observer
is
do
48
A grain of sand falls into the shell of an oyster
and the result is
75
founding
years
that
have turned to
may
have
passed
since
the
view and
essence
an
attitude
that
may
Quantum
be
one
informationwith
specializations
of quantum
its
information
three
theory,
quantum
cryptography,
computingleads
that idea.
quantify
way
the
quantum
and
us how to
algorithms can
in telling
Quantum
Secret
systems
can
in such
be
encoded
into
a way as to
physical
reveal whether
change
so much
back to almost
going
of attitude. This
any
can
older
be
seen
textbook
by
on
way
that
uncertainty
conveys
short-changed
relation is presented in
we
1
h
2
xp
of physics is to sober
have been
places
world
fantastically
better
than
the
classical one?
If
one
real
in
nature?
thoughts
to be able to
it is essential
system in one
some
from
or
prepare a
nonorthogonal
fixed
physical
are
used to encode
key to be shared
between
receiver.
the
information
sender
an
and
eavesdrop- per
was
nonorthogonal
of the proposed
states
key into
the
forces
process
information-gathering
disturbance
is about
actually prepared in
seeks
The
to
induce
the initial
for
the
receivers
sender
over
a loss
the
measurements
of predictability
outcomes
and,
so,
is
outcomes
for the
senders
the
some
inspection)
a direct
of
directly
of
connection
gained by
an
eavesdropper
process.
goes up, the
goes up in a precisely
the
gathered
information
necessary
disturbance
also
this scenario.
appear
that
or measurement
case,
the eavesdropper)
49
is
to
disturbance
to
predict
measurements
the
perform. No hint of
system is made
use
idea of measurement
The
consider
second
at
preparations
the
senders
previous
informationthis
ability
one
while
the
disturbance
in
outcomes
the
legitimate
any
certain
might
causing disturbance.
ingredient
least
of
receiver
two
is
possible
that
one must
nonorthogonal
any
gathering
is
variable
some
some
not about
observablei.e.,
about
or reality
classically-defined
unknown
hidden
is
actually
prepared.
The lesson
is this:
about
information
quantum
noise
nothing.
no
with
It
is
to
connection
is
simply
any
preexisting variable.
How crucial is this second ingredientthat
is, that
be at least
there
two nonorthogonal
its necessity
account
above:
One
might
goal is not
eavesdroppers
a
say
by making
can
shift in the
that
the
so much to uncover
the
identity of the
sharpen
her
senders
Changing
same
predictions
the
time
as
language
as
little
still
disturbing
as
further
the
possible.
to
the
terminology
actions
the
of
4, the
Section
potential
further
would
to
like
diminishing
about
do
the
this
senders
same
those
cryptographic
context
more
point
minimally
previous
information
In
ultimately
of
comes to a
however,
an
As previously
probability
identify
the
decision-making
agents
assignment
the
within Bayesian
significant
theme
is to
a set of
a common
the conditions
probability
on
problem.
emphasized,
most
the
is that it
informationdisturbance
the
sound prediction of
present
while
consequences.)
the
eavesdroppers
under which
can come to
some random
for
variable
quantum
eavesdropping.
eavesdropper
start
The
sender
off initially
and
the
with differing
can
sender
case
so
happens
that the
than the
eavesdropper
it
about
outcomes
the
of
receivers
satisfied
with
measurement
this
on
situation,
the system in
performs
a
an attempt to
an attempt
to
come
into
the
not
there is
something
of
an
quantum
the
The
world
eavesdroppers
come
or,
of her measurements
indeed,
presence.
into
distinct property
makes
attempt
alignment
with
predictability
is always
shunted
goal.
shunting
of
This
predictability
is the subtle
itself
known.
to surreptitiously
the
away
various
manner
our
senders
from its
observers
in which the
experimental
interventions.
that
drives
probability
theory
in
general
and
quantum
mechanics
in particular
is perhaps
nothing
more
a quantum
when an agent
system
that
is
manifested
ever
our
an alignment. The
our beliefs about the potential
consequences of our interventions on a system
can come into alignment is captured by the
mathematical structure of a pure quantum state
coming into too great of
most
|i.
Take
curves
for
all possible
information-disturbance
a quantum system,
input
we
property, the
50
10.1
What
Give Us
we need
a Little
Reality
Herbert Bernstein,
circa 1997
paper [1] I
ended
we might
command
bundle
curves
of all possible
for
a system?
If it
single
real
information-disturbance
parameterthe
dimension
of
the
systems
negative:
to be
Hilbert-space
dimension
a stand-alone concept
with
will survive
no
need of
an
same
the
Einsteinian
test for it
is to
as we
pose
the
evolutions. Posit
a bipartite system
with Hilbert
spaces
HD1
and HD2
respectively)
and imagine
an
initial quantum
one can
of D2 ? It does not
the only
way to change
supposition.
Thus,
every
right
to call
to that
extent,
the numbers
D1
one
has
and D2
11
so as to
more
substantive changes to
1. I
made the language
throughout.
are now
readers
who
thought
explained
they
for
were
typographical
errors.
3. For the
labeled
purpose
the
of Section 1s imagery, I
followers
of
the
Spontaneous
Collapse
and
Spontaneous
Many-Worlds
Collapseans
interpretations,
and
Everetticsin
contrast to
the
previous
and
Collapsicans
terms
Spontaneous
better
Everettistasto
emphasize their
religious aspects.
4. Some
figures
were
removed
from
the
personae
on page 2 was added.
5. Inow denote the outcomes
of
general
(and
only)
gather
subjective
It
probabilities
for something
44
45
is
else.
probability
51
assignments P(h) for
some
hypothesis h to
on the
data d.
more
convinced
Ihave
of the consistency
become
of
much
and value of
the
radically
subjective
more
inclined
Finetti [104]
[111]. To
say,
to the
view
of
Bruno
de
that end, I
have stopped calling probability
more true to
the conception
that they
are
the phenomenon
he is describing. It is not
the present
8. The metaphor
accuracy.
program can
be likened
to the spacetime
are three
footnotes to be found
important explanatory
there.
10. Einsteins
argues more
strongly
uses
quantum
them to imagine
measurement
at the
13.
To
elaborate
entanglement
and
the
the
connection
standard
between
probability
rule, I
switched
of the
ment? sections.
was
in Section 5
however,
remains:
The tensor-product
rule
can be thought
of
as
secondary
to the structure
of local
observables.
15.
much
classical
greater
measurement
elaboration
of
the
problemi.e.,
the
mystery of
16.
Section
description
6.1,
wherein
of the relation
a more
between
detailed
real-world
measurements
and
the
hypothetical
standard
new. Heres
another
52
rationalist
beliefs in contexts
theory
of uncertainty,
kinds of undesirable
theory
decision
behavioural
establishes
maximization
provides
making
that
the
and
fit
together
inconsistencies.
expected
basis
that
an
individual should
The
how
Bayes
ways
in the
utility
rational
for
theorem
in which
light
of
beliefs
changing
procedures
for individuals
concerned
with
disciplined uncertainty
sense
of claiming to model
sense
of saying if
of these
argues
for the
new.
21. One
in Ref. [112].
12
Acknowledgments
I
thank
corre-spondence
Asher
Cabello,
Plotnitsky
dramatis
years
of
David
Comer,
for
their
personae
of
me courage
Peres,
help
the
in general,
and
Arkady
in compiling
Introduction,
the
Jeff
thanks
Steven
Donald,
Philippe
Jim Malley,
Mike
Nielsen,
Masanao
Summhammer,
for their
Chris Timpson,
version
some
out
shape
gratitude
my
technical
Entanglement?
mistake
in the
Wither
section. Finally, I
thank
Ulrich
me a Kantian; it taught me
a little harder to make myself
that I
should work
look Jamesian.
References
[1] C. A. Fuchs, Quantum
of Quantum Information,
Implications
Information
in
Transfer:
Foundations
in the Light
Quantum
and
Computation
Proceedings
of
the
NATO
Advanced
pp. 3882.
Also posted
at quant-ph/0106166.
[2] C. A. Fuchs, Quantum
Are They?
posted
at
o Phase
Transition),
http://netlib.bell-labs.com/who/cafuchs/.
[3] J. T. Cushing,
A. Fine, and
S. Goldstein,
An
Histories
2631 (1999).
[5]
J. G. Cramer,
Transactional
Mechanics,
An
of
Overview
Interpretation
of
the
Quantum
Int.J.
and
P. Pearle,
Theories: Changing
Dynamical
Quantum
so
Theory
the
Represents
Statevector
Proceedings
of
the
Reality,
Biennial
PSA
Meeting
1 990:
the
of
Philosophy
pp. 1934.
53
[7] W. H. Zurek, Decoherence,
Einselection
and
(1998).
Objectivity: A Realistic
0012122;
Formalism
P. Grangier,
of
Reconstructing
Mechanics
Quantum
in
the
the
Contextual
Objectivity
of Philosophy
Interpretation
(Summer
at
http
://plato
edu/entries/qm-manyworlds/.
[11] C. Rovelli,
of
Relational
Quantum
stanford
Mechanics,
10:
...
which
large
by large posters
concerts.
public
Company,
as
...
as
though there
In August
were
announcing
the Antirelativity
twenty
[13]
M.
Relativity
Gardner,
the
for
Million,
IQM-1
on
and R. N. Graham,
the
Interpretation
of
Resource
Quantum
in
of Rationality,
an
introduction
by J. J.
pp. 317345.
p.23.
[17]
A. Zeilinger,
Quantum
Mechanics,
Foundational
Found.
Principle
Phys.
for
29, 631643
(1999);
C. Brukner
Invariant
and
Information
A. Zeilinger,
in Quantum
Operationally
Measurements,
Phys.
Rev. Lett. 83, 33543357
in Quantum
Phys.
Measurements,
Rev. A 63,022113-1022113-10
(2001); C.Brukner,
Comment
Shannon
...
on
Information
Zeilinger,
Conceptual
quant-ph/0007116;
Supposed
C. G. Timpson,
Inadequacy
Conceptual
Inadequacy
by C. Brukner
of
On
the
of
and A.
the
Shannon
Information, quant-ph/0112178.
quant-ph/0112166.
[19] N. D.Mermin, The Contemplation
Computation,
of Quantum
Information
Processing
G. Brassard,
C. Cr epeau, R.
an
Unknown Quantum State via Dual Classical and
Theory,
Mechanics
as a
Principle
J. S.
Bell,
Speakable
Quantum Mechanics:
Collected
and
Unspeakable
Papers
on
in
Quantum
and
Quantum
Quantum
Information,
(Cambridge
University Press, Cambridge, 2000).
[28]
P.
Busch,
Valuations:
Giving
Quantum
Operational
States
as
Content
Effect
to
von
Neumanns
No-Hidden-Variables
Theorem,
quant-ph/9909073.
Gleasons
Theorem
with
POVMs,
unpublished,
2000.
[30]
A. M. Gleason,
Subspaces of
a Hilbert
on
Measures
the
Closed
894
(1957).
Distributions
Quasiprobability
in
of
Terms
of
Probability
Distributions
Phys.
Phase,
the
for
Rev.
Rotated
40,
2847
Quadrature
(1989);
U.
Leonhardt,
Quantum-State
Tomography
and
Discrete
Quantum
States:
The
de
Quantum
Finetti
Representation,
to
appear
in J. Math. Phys.
T. A. Brun,
Bayes Rule,
and
C. M. Caves,
014305-4
(2001).
Considered
Complete?,
Phys.
Rev.
47,
777780 (1935).
[35] A.Fine, The Shaky Game: Einstein Realism and
the Quantum Theory, (University of Chicago Press,
1986);
Chicago,
Problem
in
Symposium
Its
M.
Jammer,
Historical
EPR
The
Development,
in
on
of Modern Physics: 50
the Foundations
years
of
(World
pp. 129149.
Having
a Given
Density Matrix,
Phys. Lett. A
J. Bernstein,
Quantum
Profiles,
(Princeton
Concerning
the Essays
Einstein:
Philosopher-Scientist,
third
IL,1970),
[39]
pp. 665688.
A. Peres,
Quantum
Theory:
Concepts
and
[40]
C. M. Caves
Information:
How
and
Much
C. A. Fuchs,
Information
Quantum
in
State
Vector?, in
Thought
p.226.
[42]
R.
Penrose,
Concerning
The
Computers,
Emperors
New
Mind:
Physics,
(Oxford
University
Press,
1989),
Oxford,
pp.
268269.
Foundations
of Statistics,
in Classical and
Symposium,
Foundations
of
Quantum
edited
by
S. Kamefuchi,
Namiki,
H. Ezawa,
S. Nomura,
Y. Murayama,
Y. Ohnuki,
and
M.
T. Yajima
(Physical
Society of Japan, Tokyo, 1984),
[45]
W.
Heisenberg,
Physics
pp. 265276.
and
Beyond:
Encounters
and Conversations,
translated
by A. J.
Pomerans
(Harper & Row, New York, 1971),
pp. 6364.
;Quantum
(2000)
Formulation,
Inspiration:
Theory
and
Interpretation,
Peres
Reply,
53(9),
14 (2000).
[47] P. Simon, Kathys
Song,
(Columbia
Records,
55
11751182 (2000).
[49] R. Cooke, M. Keane, and W. Moran,
Elementary
Proof
of Gleasons
Phil.
Soc.
Theorem,
An
Math.
Proc.
Camb.
Pitowsky,
Infinite
98,
117128
(1981);
I.
Theorems
and the
Logic
of Indeterminacy,
218228 (1998).
J. Math. Phys.
39,
[50]
K.
Kraus,
States,
and
Effects,
Operations.
Theory
(Springer,
Berlin,
1983), Lecture Notes inPhysics Vol. 190.
[51] A. S. Holevo, Information-Theoretical
Aspects
of Quantum Measurement,
110118 (1973).
[52] S. J.
Strategies,
van
Enk, Quantum
ona Qubit,
Finite
the Kochen-Specker
Precision
Measurement
Phys. Rev.
Theorem,
Lett.
83,
37513754
Non-Contextual
(1999);
Hidden
A.
Variables
Kent,
and
Physical
Measurements,
[55] I
thank Howard Barnum for pointing out the
Foundations
of
1963),
[56]
pp. 6263.
C. A. Fuchs,
Notes
on a
Paulian
Idea:
Foundational,
Historical,
Anecdotal
and
Forward-Looking
Thoughts
on the Quantum,
quant-ph/0105039.
chanics, forthcoming.
[58]
L.
Hardy,
Theory
Quantum
From
Five
on
Modality,
Probability,
and Bell s
Aspects
Int.J.Theor. Phys.
16,321(1977).
[61]
F.
E.
Observables,
Schroeck,
(1989); F.E.
Schroeck,
Jr.,
Int. J. Theor.
Jr.
Yields Informational
Coexistence
Phys.
Unsharpness
Completeness,
of
28, 247262
in Measurement
in Symposium
on
the
Quantum
Foundations
Theory
Philosophical
of Modern
of Measurement
Physics
and
1 990:
Related
Schroeck,
Fischer,
and
and
Completeness
Informational
pp. 375389;
F.E.
G. Harnett,
Frames,
in
Symposium
on
the Foundations
Measurement,
Quantum
of
Modern
Physics
Irreversibility
1993:
and
the
Physics
pp. 378389.
[62] P. Busch,
Physical
Complete
Informationally
Quantities,
Int.
J.
Theor.
of
Sets
Phys.
30,
12171227
(1991).
on Informationally
and G. G. Plunk,
Complete Measure-
ments, inpreparation.
[64]
R.
Jozsa,
Computation,
Entanglement
in The Geometric
and
Quantum
Universe:
Science,
Geometry,
p.369.
Matrix
Analysis,
(Springer-Verlag,
N. R. Wallach,
An
Unentangled
Gleasons
Theorem, quant-ph/0002058.
56
[68]
D.
Aerts
Justification
for
and
I. Daubechies,
Using
the
Tensor
Physical
to
Product
Describe Two
Quantum
Systems
as
Helv.
ay, C. Randall,
and D.
ay,
EPR
Experiments:
The
(1987); M.
Structure
of
the
Sample
Space,
(1988);
Int. J. Theor.
[70]
S.
Nonlocality
Quantum
without
MacLane
and
G.
Birkhoff,
Algebra,
[71] H. Araki, On
Characterization
of the State
75,
124
(1980);
Accessibility
of
W.
K.
Quantum
Wootters,
Local
Information,
in
Complexity, Entropy
p.2946.
[72] A. Wilce, Tensor
Probability
in Quantum
Mechanics,
in
preparation (2002).
[75] H. E. Kyburg, Jr. and H. E. Smokler, eds., Studies
[76] E. T. Jaynes,
Probability
Theory:
on
The Logic
Web
at http://bayes.wustl.edu/.
as perhaps
field.
[77] J. M. Bernardo and A. F. M. Smith, Bayesian
Theory (Wiley, Chichester, 1994).
[78] N.D.Mermin, From Classical State-Swapping
Quantum Teleportation,
to
1012320-6 (2002).
[79]
B.
Schumacher,
Entanglement
Sending
Quantum
Phys. Rev.
A 54,2614
2628 (1996).
[80] T. M. Cover
of Information
and J. A. Thomas,
Elements
York,
1991).
[81] R. B. Ash, Information
York, 1965).
Rev.
[83] W. K. Wootters,
Random
Quantum
States,
Fractional
Integration
on Hyperspheres,
Reduced Distributions
and
J.Phys. A
Bound
Accessible
for
in
Information
Quantum
Mechanics,
[85]
C. A. Fuchs
Tradeoff
Relations
and
for
K. Jacobs,
Information
Finite-Strength
Quantum
(2001).
[86]
M. A. Nielsen,
Measurement
Mixing
Characterizing
in Quantum Mechanics,
and
Phys. Rev. A
63,
022114-1022114-11
[87]
R.
Schatten,
(2001).
Norm
Ideals
of
Completely
to Eliminate
Phys. Rev.
51, 24592468
(1995).;
J. Audretsch,
of
a Qubit
L.
under
the
Influence
of
Measurements,
Succession
Unsharp
of
quant-ph/0201078.
57
[89] J. Bub, Von Neumanns
as a Probability
Projection
Postulate
(1990).
and Measurement,
by O. Hirota, A. S. Holevo,
and
edited
C. M. Caves
(Plenum
Press,
Ozawa,
An
New
York,
Operational
1997),
pp.
Approach
233241;
M.
to Quantum
State Re-
in
11
April 2002.
[94] G. Boole, An Investigation
of the Laws
of
(2002).
Compatible
and
Uncovering Disagreement,
inpreparation.
as 10 September
as
early
1999.
quant-ph/0012067;
Quantum
W. D
ur
Gate,
and
J. I. Cirac,
Non-local Operations:
Purification,
Tomography,
and
Storage,
Probabilistic
Compression,
Implementation,
Phys. Rev. A
64, 012317-1012317-14
(2001); S. F. Huelga, J.
Vi-
Control:
Teleportation
of
Unitary
G.
dal, L. Masanes,
and
J. I. Cirac,
Storing
Rotations
of
Secret
Sharing,
and
J. Opt.
B 3,
407411 (2001); S. F.
Huelga,
Remote
M. B. Plenio,
Control
and
of Restricted
J. A. Vaccaro,
Sets of Operations
Teleportation
of
[99]
Phys.
Angles,
042316-1042316-12
Rev.
65,
(2002).
A. Jamio lkowski,
Linear
Transformations
Rep.
Math.
Phys.
3, 275278
(1972).
on Complex
Matrices,
(1975).
(1996).
[102] C. H. Bennett
Cryptography:
and G. Brassard,
Public
Key
Distribution
Quantum
and
Coin
Tossing,
in Proc.
Computers,
IEEE
International
on
Conference
Systems
(IEEE
Press,
New York, 1984),
p. 175, IEEE,
States,
1984; C. H.
3121 (1992).
[103] S. J.
State of
Rev.
Lett. 88, 027902-1027902-4
Enk
and
C. A. Fuchs,
(2002); S. J.
Quantum
of
State
van
a
Propagating
Laser Field,
(2002).
[104] B. de Finetti, Theory of Probability
New York, 1990); For
a collection
(Wiley,
de Finettis origi-
and
Probability
Symmetric
and G. R. Moody,
States
and
an
Locally
Analogue
of de
Finettis
verw.
Z. Wahrschein.
Theorem,
Geb.
33,
343351 (1976).
[106] E. C. G. Stueckelberg,
Real
Hilbert
Quantum
Helv. Phys.
Space,
Theory
Acta
in
33, 727
(1960);
S. L.Adler, Quaternionic
Quantum Mechanics
1995).
58
[107] C. H.Bennett, G. Brassard, and N.D.Mermin,
Quantum Cryptography
Fuchs, Information
Gain
vs.
(1992); C. A.
State Disturbance
in
Quan-
tum Theory,
Fort.
der Phys.
46, 535565
(1998).
[108] C. A. Fuchs
vs.
Disturbance
Gain:
State
Uncertainty
Relations
for Quantum
20382045
(1996);
Information,
C. A. Fuchs,
R. B.
Griffiths,
C.-S.
Niu,
Eavesdropping
in
and
A.
Quantum
Peres,
Optimal
Cryptography.
I.
Information Bound
and
Optimal
Strategy,
(1997);
11631172
D.
Phys.
Bruss,
Rev. A 56,
Optimal
Eavesdropping in
Quantum
with
Cryptography
Six
States,
(1998); K. Banaszek
and
I. Devetak,
Ensembles
Fidelity
of Identically
for
Finite
Qubits,
Phys.
Trade-off
Prepared
Rev. A
of
mitment Protocols,
of
Optimization
Generalizations
Commitment
Coherent
the
of
BB84
Attacks
Quantum
in
Bit
Proto-
Weak
Cheat-Sensitive
quant-ph/0202118;
Coin
Flipping,
Sensitive
Quantum
Bit
Commitment,
quant-ph/9911043.
[110] R. Blume-Kohout,
C. M. Caves, and I. H.
Deutsch,
Mount
Climbing
Scalable:
Physical-Resource
Requirements
[111]
for
Scalable
Quantum
quant-ph/0204157.
Computer,
E. T. Jaynes,
Probability,
The
in Appendix
de Finetti
System
of
A, Other Approaches
to
M. Fortun
Through:
and
Pursuing
H. J. Bernstein,
Science
and
Muddling
Truths
in
the
Twenty-First
Century,
1998).
59
(Counterpoint,
Washington,
DC,