Академический Документы
Профессиональный Документы
Культура Документы
UFR Mathmatiques
Quantum mechanics
Mathematical foundations
and applications
Lecture notes
Dimitri Petritis
Contents
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
15
17
17
22
23
24
29
30
31
31
33
36
36
38
38
39
39
39
39
4 Algebras of operators
41
41
42
45
45
46
4.4.2 Projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
46
47
47
49
49
51
5.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
53
56
58
60
66
70
74
6.2.1 Logics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
74
75
77
80
82
84
87
7.1 Observables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87
7.2 States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3 Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
92
95
95
8.2 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
95
95
97
99
99
99
109
113
13 Cryptology
125
135
143
145
References
146
Index
150
1
Physics, mathematics, and
mathematical physics
Physics relies ultimately on experiment. Observation of many different experiments of similar type establishes a phenomenology revealing relations between the experimentally measured physical observables. A phenomenology,
even relying on false hypotheses, can still be useful if it predicts correctly quantitative relationships occurring in yet unrealised experiments1 The next step is
inductive: physical models are proposed satisfying the phenomenological relations. Then, new phenomenology is predicted, new experiments designed to
1
For instance the phenomenology prevailing in the Anticythera mechanism had useful predictive power although the underlying hypothesis of a geocentric solar system was false.
Chapter 1
verify it, and new models are proposed. When sufficient data are available, a
physical theory is proposed verifying all the models that have been developed
so far and all the phenomenological relations that have been established. The
theory can deductively predict the outcome for yet unrealised experiments. If
it is technically possible, the experiment is performed. Either the subsequent
phenomenology contradicts the theoretical predictions and the theory must
be rejected or it is in accordance with them and this precise experiment
serves as an additional validity check of the theory.2 Therefore, physical theories have not a definite status: they are accepted as long as no experiment
contradicts them!
It is a philosophical debate how mathematical theories emerge. Some scientists among them the author of these lines share the opinion expressed by
Michel Demazure (see quotation), claiming that Mathematics is as a matter of
fact an experimental science. Accepting, for the time being, this view, hypotheses for particular mathematical branches are the pendants of models. What
differentiates strongly mathematics from physics is that once the axioms are
stated, the proved theorems (phenomenology) need not be experimentally corroborated, they exist per se. The experimental nature of mathematics is hidden
in the mathematicians intuition that served to propose a given set of axioms
instead of another.
Mathematical physics is physics, i.e. its truth relies ultimately on experiment
but it is also mathematics, in the sense that physical theories are stated as a
set of axioms and the resulting physical phenomenology must derive both as
theorems and as experimental truth.
A general physical theory must describe all physical phenomena in the universe, extending from elementary particles to cosmological phenomena. Numerical values of the fundamental physical quantities, i.e. mass, length, and
time span vast ranges: 1031 kg M 1051 kg, 1015 m L 1027 m, 1023 s T
1017 s. Units used in measuring fundamental quantities, i.e. kilogramme (kg),
metre (m), and second (s) respectively, were introduced after the French Revolution so that everyday life quantities are expressed with reasonable numerical values (roughly in the range 103 103 .) The general theory believed to
describe the universe3 is called Quantum Field Theory; it contains two fundamental quantities, the speed of light in the vacuum, c = 2.99792458 108 m/s,
2
/Users/dp/a/ens/mq/iq-intro.tex
and the Planks constant = 1.05457 1034 Js. These constants have extraordinarily atypical numerical values. Everyday velocities are negligible compared
to c, everyday actions are overwhelmingly greater than . Therefore, everyday
phenomena can be thought as the c and 0 limits of quantum field
theory; the corresponding theory is called classical mechanics.
It turns out that considering solely the c limit of quantum field theory
gives rise to another physical theory called quantum mechanics; it describes
phenomena for which the action is comparable with . These phenomena are
important when dealing with atoms and molecules.
The other partial limit, 0, is physically important as well; it describes
phenomena involving velocities comparable with c. These phenomena lead to
another physical theory called special relativity.
Although quantum field theory is still mathematically incomplete, the theories obtained by the limiting processes described above, namely quantum mechanics, special relativity, and classical mechanics are mathematically closed,
i.e. they can be formulated in a purely axiomatic fashion and all the experimental observations made so far (within the range of validity of these theories)
are compatible with the derived theorems.
Among those three theories, quantum mechanics has a very particular status. It can be formulated in a totally axiomatic way; all its predictions have been
verified with unprecedented accuracy; not a single experiment has ever put the
theory in difficulty; quantum phenomena play a prominent role in the global
economy (a very conservative estimate is that 35% of the global wealth relies
on exploiting quantum phenomena). In spite of this tremendous success in the
predictive/explanatory power of quantum mechanics and its mathematically
closed form, its axiomatic setting is not very satisfactory; the theory looks as if
a conceptual building block were missing in the description of the theory.
The purpose of this course is twofold. Firstly, the mathematical foundations of quantum mechanics are presented. Algebra, analysis, probability, and
statistics are necessary to describe and interpret this theory. Its predictions are
often totally counter-intuitive. Hence it is interesting to study this theory that
provides a useful application of the mathematical tools, a source of inspiration4 for new developments for the underlying branches of mathematics, and
4
Recall that entire branches of mathematics have been developed on purpose, to give precise mathematical meaning to initially ill-defined mathematical objects introduced by
physicists to formulate and handle quantum theory. To mention but the few most prominent
examples of such mathematical theories: von Neumann algebras, spectral theory of operators,
/Users/dp/a/ens/mq/iq-intro.tex
Chapter 1
a description of unusual physical phenomena. All these phenomena are verified experimentally nowadays. Quantum mechanics intervenes in a decisive
manner in the explanation of vast classes of phenomena in other fundamental
sciences and in technology. Without being exhaustive, here are some examples
of such quantum phenomena:
atomic and molecular physics (e.g. stability of matter, physical properties
of matter), quantum optics (e.g. lasers),
on which rely chemistry (e.g. valence theory) and biology (e.g. photosynthesis, structure of DNA),
solid state physics (e.g. physics of semiconductors, transistors),
tunnel effect (e.g. atomic force microscope) and nanotechnology,
supraconductivity (e.g. used in magnetic levitation for ultra-fast trains)
and superfluidity
...
There is however another major technological breakthrough that is foreseen
with a tremendous socio-economical impact: if the integration of electronic
components continues at the present pace (see figure 1.1), within 1015 years,
only some tenths of silicium atoms will be required to store a single bit of information. Classical (Boolean) logics does not apply any longer to describe atomic
logical gates, quantum (orthocomplemented lattice) logics is needed instead.
Theoretical exploration of this new type of informatics has started and it is
proven [30] that some algorithmically complex problems, like the integer prime
factoring problem for which the best known algorithm [18] requires a time
is superpolynomial in the number of digits can be achieved in polynomial
time using quantum logic. The present time technology does not yet allow the
prime factoring of large integers but it demonstrates that there is no fundamental physical obstruction to its achievement for the rapidly improving computer technology. Should such a breakthrough occur, all our electronic transmissions, protected by classical cryptologic methods could become vulnerable.
The following table gives a very rough estimate of the time needed to factor an
n = 1000 digits numbers, assuming an operation per nanosecond for various
hypotheses of algorithmic complexity.
On the other hand, present day technology allows to securely and unbreakably cipher messages using quantum cryptologic protocols. Thus the second
purpose of this course is to present the applications of quantum mechanics
theory of distributions, non-commutative probabilities.
/Users/dp/a/ens/mq/iq-intro.tex
/Users/dp/a/ens/mq/iq-intro.tex
Chapter 1
O (exp(n))
10417 yr
O (n 3 )
1s
Table 1.1: A very rough estimate of the order of magnitude of the time needed
to factor an n-digit number, under the assumption of execution of the algorithm on a hypothetical computer performing an operation per nanosecond, as
a function of the time complexity of the used algorithm. When the cryptologic
protocol RSA has been proposed [24] in 1978, the best factoring algorithm had
a time complexity in O (exp(n)) (for comparison: age of the universe 1.5 1010
yr). The best known algorithm (the general number field sieve algorithm) reported in [19] requires time O (exp(n 1/3 log2 n)) to factor a n-digit number. The
Shors quantum factoring algorithm [30] requires a time of O (n 3 ).
/Users/dp/a/ens/mq/iq-intro.tex
2
Phase space, observables,
measurements, and yes-no
experiments
2.1 Introduction
As is the case in all experimental sciences, information on a physical system
is obtained through observation (also called measurement) of the values
within a prescribed set that can take the physical observables. The acquisition procedure of the information must be described operationally in terms
of
macroscopic instruments, and
prescriptions on their application on the observables of an objectified
physical system.
The biggest the set of observables whose values are known, the finest is
the knowledge about the physical system. Since crude physical observables
(e.g. number of particles, energy, velocity, etc.) can take values in various sets
7
2.1. Introduction
(N, R+ , R3 , etc.), to have a unified treatment for general systems, we reduce any
physical experiment into a series of measurements of a special class of observables, called yes-no experiments. This is very reminiscent of the approximation of any integrable random variable by a sequence of step functions. Therefore, ultimately, we can focus on observables taking values in the set {0, 1}.
To become quantifiable and theoretically exploitable, experimental observations must be performed under very precise conditions, known as the experimental protocol. Firstly, the objectified system must be carefully prepared
in an initial condition known as the state of the system. Mathematically, the
state incorporates all the a priori information we have on the system, it belongs
to some abstract space of states. Secondly, the system enters in contact with
a measuring apparatus, specifically designed to measure the values of a given
observable, returning the experimental data with values in some space (X, X );
this is precisely the measurement process.
The whole physics relies on the postulate of statistical reproducibility of
experiments: if the same measurement is performed a very large number of
times on a system prepared in the same given state, the experimentally observed data for a given observable are scattered around some mean value in
X with some fluctuations around the mean value. However, when the number of repetitions tends to infinity, the empirical distribution of the observed
data tends to some probability distribution on (X, X ). Thus, abstractly, a measurement is a black box transforming states into probability measures on some
space of observations. Mathematically, it will be shown that the measurement
corresponds to a transformation kernel.
Dealing with random variables, the natural question that arises is: what is
the appropriate (abstract) probability space, if any, on which random variables
entering a given problem can be defined? For sequences of classical random
variables, the answer is well known: such an abstract probability space exists,
provided that the sequence verifies the Kolmogorovs compatibility conditions;
in that case, there exists a canonical (minimal) realisation of the abstract probability space on which the whole sequence is defined. Elements of this probability space are called trajectories of the sequence. The physical analogue of
the minimal realisation of the abstract probability space is called phase space.
Elements of the phase space are called (pure) phases. The physical analogue
of a random variable is called observable.
It turns out that physical observables for classical systems are just random
variables so that the phase space for such systems is a genuine probability space,
/Users/dp/a/ens/mq/iq-phase.tex
while for quantum systems, observables are (generally non-commuting) Hermitean operators acting on an abstract Hilbert space that plays the rle of
quantum phase space.
bX 3 f 7 K X f () :=
K X (, d x) f (x) bF , ,
(d )K X (, A) M1 (X ), A X .
Note also that X = K X 1 X and P X = PK X . If X is fixed, on denoting Q A the random variable Q A () = 1 A (X ()), for A X , we remark that Q 2A = Q A (hence Q A
is a projection) and Q A Q B = 0 if A B = ;. Such random variables are called
questions.
1
The space X can be any Polish space (i.e. a metric, complete, and separable space.) We shall
only consider the case X = Rd , for some d , in this course.
/Users/dp/a/ens/mq/iq-phase.tex
X () =
0 if [0, 1/2[
1 if [1/2, 1].
Notice however that the above realisation of the probability space involves
the Borel -algebra over an uncountable set, quite complicated an object indeed. A much more economical realisation should be given by = {0, 1}, F =
X , and P({0}) = P({1}) = 1/2. In the latter case the random variable X should
read X () = : on this smaller probability space, the random variable is the
identity function. Such a realisation is minimal.
Exercise 2.2.4. (An elementary but important exercise; please solve it before
reading the sequel!) Generalise the above minimal construction to the case we
consider two random variables X i : X, for i = 1, 2. Are there some plausible
requirements on the joint distributions for such a construction to be possible?
The canonical construction of the minimal probability space carrying an
infinite family of random variables is also possible.
Definition 2.2.5. (Consistency) Let T be an infinite set (countable or uncountable) and for each t T denote by Rt a copy of the real line, indexed by t . Denote by RT = t T Rt and for n 1 by = (t 1 , . . . , t n ) a finite ordered set of distinct indices t i T, i = 1, . . . , n. Denote P a probability measure on (R , B(R ))
where R = Rt1 Rtn . We say that the family (P ), where runs through all
finite ordered subsets of T , is consistent, if
1. P(t1 ,...tn ) (A 1 A n ) = P(t(1) ,...t(n) (A (1) A (n) ), where is an arbitrary permutation of (1, . . . , n) and A i B(Rti ), and
2. P(t1 ,...tn ) (A 1 A n1 R) = P(t1 ,...tn1 ) (A 1 A n1 ).
Definition 2.2.6. Let T be a subset of R. A family of random variables X
(X t )t T is called a stochastic process with time domain T .
If T = N or Z, the process is called a discrete time process or random sequence, if T = [0, 1] or R or R+ , the process is a continuous time process.
/Users/dp/a/ens/mq/iq-phase.tex
The natural question that arises is whether there exists a probability space
(, F , P) carrying the whole process. In other words, if P X denotes the distribution of the process X , what are the conditions it must fulfil so that there
exists a probability space (, F , P) such that P(B ) = P({ : X () B } for all
B B(RT )? The answer is given by the following
Theorem 2.2.7 (Kolmogorovs existence). Suppose that for n 1, the family
P(X t1 ,...,X tn ) , with t 1 < . . . t n and t i T R, for i = 1, . . . , n, is a consistent family of
probability measures. Then, there are
1. a probability space (, F , P) and
2. a stochastic process X = (X t )t T such that P(X t1 ,...,X tn ) (] , x 1 ] ]
, x n] = P({ : X t1 () x 1 , . . . , X tn () x n }.
Proof. See, for instance, in [29, Theorem II.2.1, p. 247].
d 2x
(t ) = F (x(t )),
dt2
m 2
x
2
F (y)d y.
x0
/Users/dp/a/ens/mq/iq-phase.tex
p
H0
H
Figure 2.1: The phase space for a point mass in dimension one.
Hence the Newtons equation is equivalent to the system of first order differential equations, known as Hamiltons equations:
dp
dt
dq
dt
H
q
H
,
p
=
=
q(0)
q0
=
,
p(0)
p0
p2
q(t )
If (t ) =
R2 represents the coordinate and momentum of the sysp(t )
tem at time t , the time evolution induced by the system of Hamiltons equations
2
can be thought
as
the flow on R , described by (t ) = T t (0), with initial conq0
dition (0) =
.
p0
Postulate 2.2.13. The phase space of a classical system is an abstract measurable space (, F ). The states of a classical system are the probability measures
on (, F ). Pure states correspond to Dirac masses. When two systems, respectively described by (1 , F1 ) and (2 , F2 ) are merged and considered as a single
/Users/dp/a/ens/mq/iq-phase.tex
Z
Z
1 {X A} ()(d )
1 A (X ())(d )
If = 0 for some 0 .
0 (Q A ) =
1 {X A} ()0 (d )
= 1 A (X (0 )).
RIf the space is minimal for X , then X () = and we get respectively: (Q A ) =
Exercise 2.2.18. What is the minimal phase space for a mechanical system
composed by N point particles in dimension 3?
/Users/dp/a/ens/mq/iq-phase.tex
Albert Einstein, (Ulm) 1879 (Princeton) 1955. Developed the theory of special (1905) and
general (1913) relativity, pillars of modern theoretical physics but has been awarded the Nobel
Prize of Physics in 1921 not for this achievement but for his explanation of the photoelectric
effect. Published more than 300 scientific papers of extreme originality. Beyond his scientific contributions Einstein was a humanist worried about war; he has signed with the British
philosopher and mathematician Bertrand Russell the Russell-Einstein manifesto against nuclear weapons.
3
Cf. for instance his famous aphorism God does not play dice with the world (quoted, for
instance, from the conversations with Hermanns, in William Hermanns, Einstein and the Poet:
in search of the cosmic man, Branden Press, Brookline, MA (1983).
4
Boris Yakovlevich Podolsky, (naturalised) American physicist 1896 1966.
5
Nathan Rosen, American physicsist1909 1995.
6
Literally: thought experiment. A very powerful epistemological method developed by
Einstein of questioning the validity of the theoretical predictions.
7
See the Analyse item at http://bibnum.education.fr/physique/physique-quantique/leparadoxe-epr for the unconventional beginnings and fate of the EPR paper.
/Users/dp/a/ens/mq/iq-phase.tex
David Joseph Bohm, 1997 1992. American physicist who introduced the hidden variables
formalism in the quest of restoring realism of quantum mechanics.
9
John Stewart Bell, 1928 1990. A Northern-Irish physicist with a major contribution known
as Bells theorem, establishing the experimental consequences that should have hidden variables in quantum mechanics.
10
Readers so inclined to philosophical meditation are invited to consider the ravages
fashion-led or project-oriented research can cause to the advancement of science.
/Users/dp/a/ens/mq/iq-phase.tex
Proof:
P(X 1 = 1, X 3 = 0) = P(X 1 = 1, X 2 = 0, X 3 = 0) + P(X 1 = 1, X 2 = 1, X 3 = 0)
Proof: The random variables being {0, 1}-valued, it is enough to check on all 16
possible realisations of the quadruple (X 1 (), X 2 (), Y1 (), Y2 ()) that
{X 1 = Y1 } {[X 1 = Y2 ] [X 2 = Y2 ] [X 2 = Y1 ]}.
Figure 2.2: When a photon passes through the first polariser oriented in
direction emerges polarised in that direction. When it encounters a second polariser oriented in direction passes through with probability
cos2 ( ). If the photon is initially already polarised in direction , nothing
changes if the first polariser is removed.
great number of elementary light quanta called photons; the corpuscular nature of light has been conjectured already by Gassendi11 and Newton12 and irrefutably confirmed by the photoelectric effect whose theoretical explanation
was given by Einstein. Therefore, the statement on intensities made above have
only a statistical meaning; if a photon passing through a polariser oriented in a
given direction , encounters a second polariser oriented in a direction , has
probability 21 cos2 ( ) to pass through (see figure 2.2). This is an experimental fact, in accordance with both quantum mechanical prescriptions and with
classical electromagnetic theory of light.
If the experiment is to be explained in terms of classical probability, with every polariser in direction [0, /2] is associated a random variable X {0, 1};
the random variables X are defined on a probability spaces (, F , P) where
represent the microscopic state of the photon. Now for the experimental
setting depicted in figure 2.2, the random variables X are correlated as
E(X X ) = P(X = 1, X = 1) =
1
cos2 ( ).
2
But now there is a problem because this correlation cannot be that of classical
random variables. Choosing in fact three polarisations 1 , 2 , and 3 , we have
P(X i = 1, X j = 0) = P(X i = 1) P(X i = 1, X j = 1) = 21 (1 cos2 (i j )) =
11
Pierre Gasendi, French philosopher, priest, scientist, and mathematician 1592 1655.
Sir Isaac Newton, English physicist and mathematician, 1642 1727. In his very influential
work, its Philosophi naturalis principia mathematica established the classical theory of universal gravitation and discovered what later became differential calculus to solve the equations
of motion.
12
/Users/dp/a/ens/mq/iq-phase.tex
Ca
PM1
PM2
Coincidence monitoring
Figure 2.3: Schematic view of the Orsay experiment [5]. A beam of calcium (Ca)
atoms is triggered by a laser. When thus excited, calcium atom emit simultaneously two photons (at different frequencies) in opposite directions and having
correlated polarisations. An ingenious system of optical switches is used whose
net effect can be described by the following equivalent description. The left polariser is oriented in one of the angles 1 or 2 , the right polariser into one of 1
or 2 . After passing through the polarisers, the photons are counted by photomultipliers (PM1) and (PM2) and only photons detected in synchronisation are
recorded. A careful design is made so that all photons travel on same optical
lengths and the choice of left and right polarisation is made after the photons
are emitted (so that any causal influence of the choice of orientations on the
manner the photons are emitted can be excluded).
1
2
2 sin (i
then
1
1
1
sin2 (1 3 ) sin2 (1 2 ) + sin2 (2 3 ).
2
2
2
The choice 1 = 0, 2 = /6, and 3 = /3 leads to the impossible inequality
3/8 1/8 + 1/8. Therefore, classical probability cannot describe this simple experiment.
On a second reading, this experiment is not very convincing because on
arranging polarisers on the optical table as described above, there is nothing
preventing conceptually the second random variable X to depend in fact on
both and . But then the correlation reads E(X X , ) = P(X = 1, X , = 1) =
1
cos2 ( ) and this can be satisfied by choosing, for instance, X and X ,
2
independent with P(X = 1) = 1/2 and P(X , = 1) = cos2 () which of course
can be easily conceived.
The irrefutable evidence of the impossibility of describing Nature with merely
classical probability is provided through the second experiment Aspect, Dal/Users/dp/a/ens/mq/iq-phase.tex
1
sin2 ( ),
2
for every choice of and . (Note incidentally that the same conclusion is obtained using the not yet presented quantum formalism). Now the choice
1 = 0, 2 = /3, 1 = /2, and 2 = /6, should read 1 1/4 + 1/4 + 1/4, manifestly violating the four variable Bells inequality 2.3.2.
To better grasp the significance of this experiment, it has been proposed,
see [21] for instance, to think of it as a card game between two players X and Y
who can pre-agree on any conceivable strategy in order to win the game. The
game is described in the following
Exercise 2.3.3. (The Orsay experiment as a card game [21]) The game is played
between players X , Y (see figure 2.3), and A who acts as an arbiter and as game
leader (A like . . . Aspect).
Description of the game
A disposes of a well shuffled deck of red and black cards (consider it as an
infinite sequence of i.i.d. {red, black}-valued random variables uniformly
distributed on {red, black} := {r, b}).
X and Y are free to use random resources (e.g. dice) if they wish.
Before the game starts, X and Y agree on given strategy (deterministic,
non-deterministic, or random) how to determine a {yes, no}-valued variable out of the colour of the card they will be presented. Once the game
starts, the players are not allowed any longer to communicate.
A picks two cards from the deck and presents the one to X and the other
to Y (mind that X and Y dont know each others card).
X and Y apply their own pre-agreed strategy to the colour the are presented with and simultaneously say yes or no.
After the announcement of the players, the cards are laid on the table.
Four different card pairs are possible (r r ), (r b), (br ), (bb), where the first
colour refers to the colour of the X s card and the second to the Y s one
(consider these colour pairs as boxes in a 22 board). If both players have
given the same answer then 1 is written in the corresponding box, else 0
is marked.
/Users/dp/a/ens/mq/iq-phase.tex
In the course of the game, the boxes get filled by sequences of 0s and 1s.
Let cc 0 , with c, c 0 {r, b}, be the limit of the empirical probability of 1s
in the box corresponding to colours (cc 0 ) when the game runs indefinitely. The players win the game if r r > r b + br + bb . The purpose
is to show that there exists no strategy (deterministic, non-deterministic,
or random) allowing the players to win the game.
Questions
1. Suppose that X and Y have agreed on the following strategy: X always
says yes, independently of the colour of the card presented to her/him
and Y answers the question is my card red?. Compute explicitly the
values of cc 0 for c, c 0 {r, b} and show that with this deterministic strategy, the numbers cc 0 satisfy the four variable Bells inequality r r r b +
br + bb .
2. In the above strategy, the decision making process is described through
the matrices D X and D Y with D X : {r, b} {0, 1} [0, 1] (and similarly for
D Y ), defined respectively by
0 1
0 1
DX =
and D Y =
,
0 1
1 0
interpreted as meaning P(Answer of X is a | card colour is c) = D X (c, a)
for c {r, b} and a {0, 1} (and similarly for Y ). Hence the previously
described strategies are termed deterministic strategies. Determine all
deterministic strategies and show that for all of them, Bells inequality is
verified.
3. Propose a plausible parametrisation of the space of all strategies (deterministic and random) and show that this space is convex. Is it a simplex?
Show that the previous deterministic strategies are extremal points of this
space.
4. Conclude that in general any strategy can be written as a convex combination of deterministic strategies. Is this decomposition unique? How
such a convex combination is related to hidden variables? Conclude that
no classical strategy exists allowing to win this game.
(Some hints concerning convexity and convex decomposition of stochastic matrices needed for the the two last questions can be found in the chapter Markov
chains on finite state spaces of the lecture notes [22]).
/Users/dp/a/ens/mq/iq-phase.tex
Richard Phillips Feynman, 1918 1988. American physicist with major contributions in
quantum mechanics, especially known for his work in the path integral formulation of quantum mechanics. Has been awarded the Nobel prize in Physics in 1965. Feynman was an unparalleled populariser of physics; he left co-authored with Leighton and Sands as legacy
to generations of young physicists the seminal 3-volume textbook known as The Feynman
lecture on Physics.
14
Niels Henrik David Bohr, 1865 1962,. Danish physicist with foundational contributions in
quantum mechanics.
15
Max Karl Ernst Ludwig Planck, 1858 1947. German physicist; has been awarded the Nobel
Prize in Physics in 1918. One of the originators of quantum theory.
16
Erwin Rudolf Josef Alexander Schrdinger, Austrian physicist. Developed quantum theory.
17
Werner Karl Heisenberg, 1901 1976. German physicist with major contributions in quantum mechanics. The uncertainty relations are named after him.
18
Paul Adrien Maurice Dirac, 1902 1984. English physicist with major contributions in
quantum electrodynamics. Predicted the existence of antimatter. Has been awarded the Nobel Prize in Physics in 1933. For the needs of quantum mechanics he introduced the notion
of -function and its derivatives, known nowadays as Dirac measure and Dirac distributions
respectively after they have been given a rigorous mathematical status by Laurent Schwartz.
Another happy achievement of Dirac was the so-called Diracs notation in terms of bras and
kets (it will presented in 3.6).
19
Jnos / Johann / John von Neumann, 1903 1954, mathematician, physicist, and computer
scientist (before this last term has been coined by lack of . . . any computer), born as Margittai
Neumann Jnos Lajos in Hungary, achieved his PhD under the direction of Fejr Lipot at the
Institute of theoretical physics of the university of Budapest. Moved to Italy then Switzerland
where he completed studies as chemical engineer at the Eidegenssische technische Hohcschule
Zrich. Exerced as Privatdozent in Gttingen, Berlin and Hamburg but finaly emigrated in 1930
/Users/dp/a/ens/mq/iq-phase.tex
formulation [32], based on quantum logics and C -algebras, will be given; this
formulation has the advantage of allowing a unified treatment for both classical
and quantum systems. Another possible formulation is provided by the informational formulation of quantum mechanics [11, 12]. The reader may wonder why so many formulations have been proposed so far. The answer is that
although all the formulations are totally satisfactory from the computational
point of view, their predictive power, and their adaptedness to explain diverse
experiments, none of the existing ones is philosophically and epistemologically
satisfactory. Quantum mechanics is a partial theory, describing fragile systems,
i.e. systems that eventually leave the quantum realm to enter the classical one;
such a fragility demands for a unified treatment of classical and quantum formalism, not guaranteed by any of the existing formalisms.
/Users/dp/a/ens/mq/iq-phase.tex
General (unnormalised) vectors of H are denoted by small Latin letters f , g , h, etc.; normalised vectors, representing rays, by small Greek letters , , , etc.
/Users/dp/a/ens/mq/iq-phase.tex
The notion was introduced by Schrdinger himself who named this property Verschrnkung in German and translated into English (by Schrdinger himself) as entanglement.
The term is commonly translated intrication in French, although the author of these lines
prefers the term enchevtrement.
/Users/dp/a/ens/mq/iq-phase.tex
Eigenvectors
u()
i
1
p
5
2
2i
1
p
5
1
3
2
Projectors
P ({})
1 2i
1
5 2i
4
4
2i
1
5 2i
1
Hence
X
P ({})
X
{3,2}
1
= (3)
5
1
2i
2i
4
1
+2
5
4
2i
2i
1
The operators P ({3}) and P ({2}) are self-adjoint (hence they correspond to
observables) and are projectors to mutually orthogonal subspaces. They play
the role of yes-no questions for a quantum system (recall remark 2.2.16.)
Now, let H be a pure phase; since u(3) and u(2) are two orthonormal vectors of H (hence also pure phases), they serve as basis to decompose
/Users/dp/a/ens/mq/iq-phase.tex
X
,0 ,00
00 0 u() | P (0 )u(00 )
spec(X )
| |2 .
=
=
=
n
X
xi p i
i =1
n
X
p
p i xi p i
i =1
n
X
p
p i exp(i i )x i p i exp(i i ),
i =1
with
i R, i = 1, . . . , n. Hence, classically,
EX =
| X with =
p arbitrary
p 1 exp(i 1 )
x1
0
..
..
, verifying kk = 1 and with X =
. We have more
.
.
p
p n exp(i n )
0
xn
over seen that classical probability is equivalent to classical physics; thanks to
the previous lines, it turns out that that it is also equivalent to quantum physics
involving solely diagonal self-adjoint operators as observables. The full flavour
of quantum physics is obtained only when the observables are represented by
non-diagonal self-adjoint operators.
Consider now,
f = P ({})
u() | u()
=
0
if spec(X )
otherwise.
Were only to consider this generalisation of probability theory to a noncommutative setting and to explore its implications for explaining quantum
physical phenomena, should the enterprise be already a fascinating one. But
there is even much more fascination about it: there has been demonstrated
lately that quantum phenomena can serve to cipher messages in an unbreakable way and these theoretical predictions have already been exemplified by
currently working pre-industrial prototypes23 .
In a more speculative perspective, it is even thought that in the near future there will be manufactured computers capable of performing large scale
computations using quantum algorithms24 . Should such a construction be realised, a vast family of problems in the (classical) complexity class of exponential time could be solved in polynomial time on a quantum computer.
23
See the article [14], articles in Le Monde (they can be found on the website of this course),
the website www.idquantique.com of the company commercialising quantum cryptologic and
teleporting devices, etc.
24
Contrary to the quantum transmitters and cryptologic devices that are already available
(within pre-industrial technologies), the prototypes of quantum computers that have been
manufactured so far have still extremely limited scale capabilities.
/Users/dp/a/ens/mq/iq-phase.tex
X
spec(X )
| |2
and when the operator X is self-adjoint, the spectrum is real and the expectation is then a real number. What makes quantum probability different from
classical one, is (among other things) the impossibility of simultaneous diagonalisation of two non-commuting operators. Following the probabilistic interpretation, denote by Var (X ) = E (X 2 ) (E (X ))2 .
Theorem 2.5.1 (Heisenbergs uncertainty). Let X , Y be two bounded self-adjoint
operators on a Hilbert space H and suppose a fixed pure state is given. Then
q
Var (X )Var (Y )
| | [X , Y ] |
.
2
/Users/dp/a/ens/mq/iq-phase.tex
3
Short resum of Hilbert spaces
For the sake of completeness, some standard results on Hilbert spaces are reminded in this chapter. Most of the proofs in this chapter are omitted because
they are considered as exercises; they can be found in the classical textbooks
[1, 16, 23, 25] which are strongly recommended for further reading.
3
1X
i k kx + i k yk2 .
4 k=0
/Users/dp/a/ens/mq/iq-hilbe.tex
32
that can be shown to be independent of the representative x of [x] becomes a complete normed space. On identifying elements of X with constant
we establish a canonical embedding : X X
. One can show that
sequences X
the completion is unique up to isomorphisms (see [16, 1.6, pp. 1721] for the
details).
Exercise 3.1.3. The following are classical examples.
1. The finite-dimensional vector space X = Cd with the ordinary scalar prodP
uct x | y = dn=1 x n y n is obviously a Hilbert space.
P
2. The space X = `p (N) = {x : N C; nN |x n |p < } is a complete normed
P
space for all p 1, with norm kxkp = ( nN |x n |p )1/p . In the particular
case p = 2 it becomes a Hilbert space with a scalar product, compatible
P
with k k2 , defined by x | y = nN x n y n . It is the infinite-dimensional
generalisation of the previous example.
3. The space X = L p ([a,
R b], ) with p 1 can be equipped with a norm k kp
defined by kxkp = ( [a,b] |x(t )|p (d t ))1/p . Then (X, kkp ) is a Banach space.
For
R p = 2, it becomes also a Hilbert space for the scalar product x | y =
[a,b] x(t )y(t )(d t ).
4. The space X = C k ([a, b]) can be equipped with a scalar product x | y =
Pk R
(j)
(j)
j =0 [a,b] x (t )y (t )d t . The corresponding norm is denoted k kW k,2
but the normed space (X, kkW k,2 ) is not complete. Its completion is called
Sobolev space W k,2 ([a, b]) and is a Hilbert space. In particular, W 0,2 = L 2 .
Definition 3.1.4. Let X : H1 H2 be a linear operator between Hilbert spaces
H1 and H2 . Its norm is defined by
kX k := kX kH1 ,H2 =
sup
hH1 :khkH1
kX hkH2
khkH1
The definition holds also in the special case when H2 = C (in which case the
operator is a form). When kX k < the operator is termed bounded. If an
operator is unbounded it can only be defined on its domain Dom(X ).
33
Z
G
X ()P(d ) =
Z
G
Y ()P(d ).
34
See ?? below.
/Users/dp/a/ens/mq/iq-hilbe.tex
35
36
and consider the linear manifold of the finite linear combination of such forms
(
L :=
n
X
)
i
c i , G, H, i = 1, . . . , n, n N .
i =1
Obviously L is a vector space. A sesquilinear form | L is defined by its action on simple tensor products in L , by
| 0 0 L := | 0 G | 0 H
and extended by linearity on all elements of L .
Lemma 3.4.1. The sesquilinear form | L is
Proof.
i =1
/Users/dp/a/ens/mq/iq-hilbe.tex
37
Usual notation
Diracs notation
Orthonormal basis (e 1 , . . . , e n )
P
= i i e i
P
| = i i
H = { f : H C, linear}
: H H
: 7 f (() = |
| = f ()
X = X
|X = X | = X |
X u(i ) = i u(i )
P ({i }) projector
P
X = i i P ({i })
Tensor product
n symbols, eg. {e 1 , . . . , e n }
|e 1 , . . . |e n
P
| = i i |e i
P
| = i i
H = { f : H C, linear}
: H H
: | 7 |
| = ||
X = X
|X |
X |i = i |u(i )
|u( ) u(i )|
P i
X = i i |u(i ) u(i )|
| | = |
38
/Users/dp/a/ens/mq/iq-hilbe.tex
39
/Users/dp/a/ens/mq/iq-hilbe.tex
40
4
Algebras of operators
The set Dom(T ) is generally a linear manifold, i.e. algebraically a vector subspace of V
which is not necessarily topologically closed.
41
42
3. (a ) = a.
Involution is also called adjoint operation and a the adjoint of a. An involutive algebra is termed a -algebra.
An element a A is said normal if aa = a a, an isometry if a a = 1 , unitary if both a and a are isometries, self-adjoint or Hermitean if a = a . On
denoting h : A1 A2 a homomorphism between two -algebras, we call it a
-homomorphism if it preserves adjoints, i.e. h(a ) = h(a) .
A normed (respectively Banach) algebra A is an algebra equipped with a
norm map k k : A R+ that is a normed (respectively Banach) vector space for
the norm and verifies kabk kakkbk for all a, b A. A is normed (respectively
Banach) -algebra if it has an involution verifying ka k = kak for all a A.
Theorem 4.2.3. Let T : H1 H2 be a linear map between two Hilbert spaces H1
and H2 . Then the following are equivalent:
1. kT k = sup{kT f kH2 , f H1 , k f kH1 1} < ,
2. T is continuous,
3. T is continuous at one point of H1 .
Proof: Analogous to the proof of the theorem ?? for linear functional. (Please
complete the proof!)
/Users/dp/a/ens/mq/iq-algop.tex
43
44
Recall that a topological space is called Hausdorff when every two distinct of its points
posses disjoint neighbourhoods.
/Users/dp/a/ens/mq/iq-algop.tex
45
4.4.2 Projections
Definition 4.4.6. Let P, P 1 , P 2 B(H).
1. P is a projection if P 2 = P .
2. P is an orthoprojection if is a projection satisfying further P = P .
3. Two orthoprojections P 1 , P 2 B(H) are orthogonal, denoted P 1 P 2 of
their images are orthogonal subspaces of H (equivalently P 1 P 2 = 0).
3
In some general unital C -algebras there are only two trivial projections 0 and 1. Therefore
the situation arising in B(H) is far from being a general property of C -algebras.
4
Strictly speaking, the term Hermitean is more general; it applies also to unbounded operators and it means self-adjoint on a dense domain. The two terms coincide for bounded
operators.
/Users/dp/a/ens/mq/iq-algop.tex
46
Projections are necessarily positive (why?). The set of orhoprojections is denoted by P(H). All projections considered henceforth will be orthoprojections.
Exercise 4.4.7. (A very important one!) Let (P n ) be a sequence of orthoprojections. We have already shown that there is a bijection between P(H) and the set
of closed subspaces of H and orthoprojections, given by P(H) 3 P 7 P (H) H,
with P (H) closed.
1. Show that that P(H) is partially ordered, i.e. P 1 P 2 if P 1 (H) subspace of
P 2 (H) (equivalently P 1 P 2 = P 1 .)
2. For general orthoprojections P 1 and P 2 , is P 1 P 2 an orthoprojection?
3. Show that P 1 and P 2 have a least upper bound.
4. Is Q = P 1 + . . . + P n an orthoprojection?
5. Is Q = P 2 P 1 an orthoprojection?
6. Show that a monotone sequence of orthoprojections converges strongly
towards an orthoprojection.
47
48
49
/Users/dp/a/ens/mq/iq-algop.tex
50
5
Spectral theory in Banach algebras
5.1 Motivation
In linear algebra one often encounters systems of linear equations of the type
Tf =g
(5.1)
with f , g Cn and T = (t i , j )i , j =1,...,n a n n matrix with complex coefficients. Elementary linear algebra establishes that this system of equations has solutions
provided that the map f 7 T f is surjective and the solution is unique provided
that this map is injective. Thus the system has a unique solution for each g Cn
provided that the map is bijective, or equivalently the matrix T is invertible.
This happens precisely when det T 6= 0. However, this criterion of invertibility is of limited practical use even for the elementary (finite-dimensional) case
because det is too complicated an object to be efficiently computed for large
n. For infinite dimensional cases, this criterion becomes totally useless since
there is no infinite dimensional analogue of det that discriminates between invertible and non-invertible operators T (see exercise 5.1.1 below!)
Another general issue connected with the system (5.1) is that of eigenvalues. For every C, denote by V = { f Cn : T f = f }. For most choices of ,
51
5.1. Motivation
the subspace V is the trivial subspace {0}; this subspace is not trivial only when
T 1 is not injective (i.e. ker(T 1 ) 6= {0}.) On defining the spectrum of T by
..
D =
and U unitary. Then p l (T ) = Up l (D)U and letting l
.
n
we get f (T ) = U f (D)U . Thus, if T is diagonalisable, the computation of f (T ) is
equivalent to the knowledge of f (t ) for t spec(T ). For the infinite dimensional
case, the problem is more involved but again the spectrum remains fundamental.
The rest of this chapter, based on [2], is devoted to the appropriate generalisation of the spectrum for infinite dimensional operators.
Exercise 5.1.1. (Infinite-dimensional determinant) Let H = `2 (N) and (t n )nN
be a fixed numerical sequence. Suppose that there exist constants K 1 , K 2 > 0
such that 0 < K 1 t n K 2 < for all n N. For every x `2 (N) define (T x)n =
t n x n , n N.
/Users/dp/a/ens/mq/iq-sptba.tex
52
more precisely X `p Y.
/Users/dp/a/ens/mq/iq-sptba.tex
53
Res(T ) = C \ spec(T ).
Notice that in finite dimension, invertibility of an operator R reduces essentially to injectivity of R since surjectivity of R can be trivially verified if we
reduce the space V into Ran(R). In infinite dimension, several things can go
/Users/dp/a/ens/mq/iq-sptba.tex
54
wrong: of course injectivity may fail as in finite dimension; but a new phenomenon can appear when Ran(R) is not closed: in this latter case, Ran(R) can
further be dense in V or fail to be dense in V . All these situations may occur
and correspond to different types of sub-spectra.
Definition 5.2.6. Let T B(V) where V is a Banach space.
1. The point spectrum of T is defined by
55
Z
R
f (y)g (x y)(d y)
turning this space into a commutative Banach algebra. This algebra is not unital (this can be seen by solving the equation f ? f = f in L 1 ), but it has an approximate unit (i.e. a sequence ( f n )n of integrable functions with k f n k = 1 for
all n and such that for all g L 1 (R), kg ? f n g k 0. (Give an explicit example
of such an approximate unit!)
Example 5.3.2. The algebra Mn (C) is a unital non-commutative algebra. There
are many norms that turn it into a finite-dimensional Banach algebra, for instance:
1. kAk =
Pn
i , j =1 |a i , j |
.
2. kAk = supkxk1 kAxk
kxk
Definition 5.3.3. Let A be a unital Banach algebra. (We can always assume
that k1 k = 1, may be after re-norming the elements of A.) An element a A is
called invertible if there is an element b A such that ab = ba = 1 . The set of
all invertible elements of A is denoted by GL(A) and called the general linear
group of invertible elements of A.
Theorem 5.3.4. Let A be a unital Banach algebra. If a A and kak < 1 then
1 a is invertible and
X
(1 a)1 =
an .
n=0
Moreover,
k(1 a)1 k
/Users/dp/a/ens/mq/iq-sptba.tex
56
1
1 kak
lud on 1 October 2013
and
k1 (1 a)1 k
kak
.
1 kak
Proof: Since ka n k kakn for all n, we can define b A as the sum of the absoP
PN
n
n
lutely convergent series b =
n=0 a . Moreover, b(1 a) = (1 a)b = limN n=0 b =
limN (1 b N +1 ) = 1 . Hence 1 a is invertible and (1 a)1 = b. The first maP
1
jorisation holds because kbk
kakn = 1kak
. The second one follows from
P nn=0
remarking that 1 b = n=1 a = ab, hence k1 bk kakkbk.
Exercise 5.3.5.
1. Prove that GL(A) is an open set in A and that the map1
ping a 7 a is continuous on GL(A).
2. Justify the term general linear group of invertible elements, i.e. show
that GL(A) is a topological group in the relative norm topology.
Definition 5.3.6. Let A be a unital Banach algebra. For every a A, the spectrum of a is the set
spec(a) = { C : a 1 6 GL(A)}.
In the rest of this section, A will be a unital algebra and we shall write a
instead of a 1 .
Proposition 5.3.7. For every a A, the set spec(a) is a closed subset of the disk
{ C : || kak}.
Proof: Consider the resolvent set
57
1
[(a ) (a 0 )] = (a 0 )2 .
0 0
lim
Assume now that spec(a) = ; and choose an arbitrary bounded linear functional : A C. Then, the scalar function f : C C defined by 7 f () =
((a )1 ) is defined on the whole C. By linearity, the function f has everywhere a complex derivative, satisfying f 0 () = ((a )2 ). Thus f is an entire
function. Notice moreover that f is bounded and for || > kak, by theorem
5.3.4,
k(a )1 k =
k(1 a/)1 k
||
1
||(1 kak/||)
1
.
|| kak
Thus lim f () = 0 and since this function is bounded and entire, by Liouvilles theorem (see [?] for instance), it is constant, hence f () = 0 for all C
and every linear functional . The Hahn-Banach theorem implies then that
(a )1 = 0 for all C. But this is absurd because (a ) is invertible and
1 6= 0 in A.
Definition 5.3.9. For every a A, the spectral radius of a is defined by r (a) =
sup{|| : spec(a)}.
Exercise 5.3.10.
1. Let p R[t ] and a A. Show that p(spec(a)) spec(p(a)).
(Hint: if spec(a), the map 0 7 p(0 ) p() is a polynomial vanishing
at 0 = . Conclude that p(a) p() cannot be invertible.)
2. For every a A show that r (a) = limn ka n k1/n .
..
D =
and a unitary matrix U such that T = U DU ; we have then
.
dn
/Users/dp/a/ens/mq/iq-sptba.tex
58
for f = ( f 1 , f 2 , . . .). Specifying a particular orthonormal basis in H is equivalent to specifying a particular unitary operator U . Suppose now that T B(H)
is a normal operator and admits the basis vectors of E as eigenvectors, i.e.
Te k = t k r k , t k C, k N. Then t = (t k )k ` (N) and U T U = M where M
is the multiplication operator defined by (M f )k = (U T U f )k = (U 1 T U f )k =
P
(U 1 T i f i e i )k = f k t k . Thus an operator T on H is diagonalisable in a given
basis E if the unitary operator associated with E implements an equivalence
between T and a multiplication operator M acting on `2 (N). This notion is still
inadequate since it involves only normal operators with pure point spectrum;
it can nevertheless be appropriately generalised.
Definition 5.4.1. An operator T acting on a Hilbert space H is said diagonalisable if there exist a (necessarily separable) -finite measure space (, F , ), a
function m L (, F , ), and a unitary operator U : L 2 (, F , ) H such that
U Mm = T U
where M m denotes the multiplication operator by m, defined by M m f () =
m() f (), for all and all f L 2 (, F , )
Example 5.4.2. Let H = L 2 ([0, 1]) and T : H H defined by T f (t ) = t f (t ), for
t [0, 1] and f H. This operator is diagonalisable since it is already a multiplication operator.
Notice that a diagonalisable operator is always normal because the multiplication operator is normal. The following theorem asserts the converse.
Theorem 5.4.3. Every normal operator acting on a Hilbert space is diagonalisable.
Proof: Long but without any particular difficulty; it can be found in [2], pp. 52
55.
/Users/dp/a/ens/mq/iq-sptba.tex
59
Ij
!1
f (I j)
Let m = inf f (), M = sup f (), and subdivide the interval [m, M ] into a
finite family of disjoint intervals (I j ) j , with |I j | < (see figure 5.1.) For each j ,
select an arbitrary j I j ; in the subset f 1 (I j ) F , the values of f lie within
from j . Therefore, we get the desired result by setting E j = f 1 (I j ). If for
every Borel set B B(R), we define P (B ) = 1 f 1 (B ) (this is a function-valued set
/Users/dp/a/ens/mq/iq-sptba.tex
60
j P (I j )()| < , .
P
Now, P is set function
(a measure actually) and the sum j j P (I j ) tends (in
R
some sense3 ) to P (d ). More precisely, the function f is equivalent to a
deterministic stochastic kernel K := K f from (, F ) to (R, B(R)), defined by the
formula
B(R) 3 (, B ) 7 K (, B ) = f () (B ) = 1 f 1 (B ) ().
(K )(B ) :=
(d )K f (, B ) = f (B ),
where f denotes the law of f . Conversely, the kernel acts on measurable positive functions g defined on the real axis by
Z
f
(K g )() := K f (, d x)g (x) = g ( f ()).
R
In particular, if g = id then
f
(K id)() =
Z
R
K f (, d x)x = f ().
/Users/dp/a/ens/mq/iq-sptba.tex
61
nN P (F n ).
62
()
lud on 1 October 2013
63
X F (x)P (d x).
Proof: The boundedness of F implies that the right hand Rside of the integral
gives rise to a well-defined sesquilinear
R functional ( f , 2g ) = X F (x)2 f | P (d x)g ,
for f , g H. Moreover, |( f , f )| X |F (x)|kP (d x) f k kF kk f k , hence the
functional is bounded. Existence and uniqueness of TF follows from the
Riesz-Frchet theorem.
Theorem 5.5.9 (Spectral decomposition theorem). If T Bh (H) then there exists a spectral measure on (C, B(C)), supported by spec(T ) R, such that
Z
T=
P ().
spec(T )
L f ,g (p) f | p(T )g =
spec(T )
p() f ,g (d ),
for all p R[t ], verifying | f ,g (B )| k f kkg k, for all B B(C). Using the uniqueness of f ,g , it is immediate to show that for every B B(C), S B ( f , g ) = f ,g (B )
is a sesquilinear form. Now, |S B ( f , g )| = | f ,g (B )| k f kkg k, for all B . Hence the
/Users/dp/a/ens/mq/iq-sptba.tex
64
p()(d ) =
p()q() f | P (d )g
= f | P (p(T )q(T )g
= q(T ) f | p(T )g
Z
=
p() q(T ) f | P (d )g .
Therefore,
(B ) =
q()1 B () f | P (d )g
= q(T ) f | P (B )g
= f | q(T )P (B )g
Z
=
q() f | P (d )P (B )g .
Since q is arbitrary,
Z
f | P (B C )g =
f | P (d )P (B )g
= f | P (B )P (C )g ,
and since f , g H are arbitrary, we get P (B C ) = P (B )P (C ).
Theorem 5.5.10. If T is a normal operator in B(H), then there exists a necessarily unique complex spectral measure on (C, B(C)), supported by spec(T ), such
that
Z
T=
spec(T )
P (d ).
65
66
u(b) u(a) =
note by f its Fourier transform f () = R f (x) exp(i x)d x. Let p : S(R) S(R)
be defined by p f = i f 0 and q : S(R) S(R) by q f (x) = x f (x), for all x R.
Show that [q, p] = i 1 . If | denotes the L 2 scalar product on S(R), show that
| f | f | 2kp f k2 kq f k2 .
Conclude that for any f S(R),
k f k2 4kx f kL 2 (R) k fkL 2 (R ) .
Below are depicted the graphs of pairs | f (x)|2 and | f()|2 , chosen among a
class of Gaussian functions, for different values of some parameter. How do you
interpret these results?
/Users/dp/a/ens/mq/iq-sptba.tex
67
0.8
1.5
0.6
0.4
0.5
0.2
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1 6
xi
0.5
0.4
0.3
0.2
0.1
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1 10
10
xi
16
0.12
14
0.1
12
10
0.08
8
0.06
0.04
2
0.02
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1 40
30
20
10
10
20
30
40
xi
100
0.02
80
0.016
0.018
0.014
60
0.012
0.01
40
0.008
0.006
20
0.004
0.002
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1 300
/Users/dp/a/ens/mq/iq-sptba.tex
200
100
100
200
300
xi
68
6
Propositional calculus and quantum
formalism based on quantum logic
Phenomenology is an essential step in constructing physical theories. Phenomenological results are of the following type: if a physical system is subject
to conditions A, B,C , . . ., then the effects X , Y , Z , . . . are observed. We further introduced yes-no experiments consisting in measuring questions in given states.
However, there may exist questions that depend on other questions and hold
independently of the state in which they are measured. More precisely, suppose for instance that Q A denotes the question: does the physical particle lie
in A, for some A B(R3 )? Let now B A be another Borel set in R3 . Whenever Q A is true (i.e. for every state for which Q A is true) Q B is necessarily true.
This remark defines a natural order relation in the set of questions. Considering
questions on given physical system more abstractly, as a logical propositions,
it is interesting to study first the abstract properties of a partially ordered set of
propositions. This abstract setting allows the statement of the basic axioms for
classical or quantum systems on an equal footing.
69
70
{1, 2, 3}
{1, 2}
{2, 3}
{1, 3}
{1}
{3}
{2}
{}
Figure 6.1: The Hasse diagram of the lattice of subsets of the set {1, 2, 3}.
Proof: : Exercise!
71
72
This theorem serves to extend the notion of measurability, defined for maps
between measurable spaces, to maps defined on abstract Boolean -algebras.
Recall that if X is an arbitrary set of points equipped with a Boolean -algebra of
subsets X , and Y a complete separable metric space equipped with its Borel algebra B(Y), a map f : X Y is called measurable if for all B B(Y), f 1 (B )
X.
Definition 6.1.15. Let be an abstract Boolean -algebra and (Y, B(Y)) a
complete separable metric space equipped with its Borel -algebra. A Y-valued
classical observable associated with is a -homomorphism h : B(Y) . If
Y = R, the observable is called real-valued.
The careful reader will have certainly remarked that the previous definition
is compatible with axiom 2.2.15. As a matter of fact, with every real random
variable X on an abstract measurable space (, F ) is associated a family of
propositions Q BX = 1 {X B } , for B B(R). The aforementioned -homomorphism
h : B(R) F , stemming from X () through the spectral measure K X (, B ), is
/Users/dp/a/ens/mq/iq-proca.tex
73
6.2. Classical, fuzzy, and quantum logics; observables and states on logics
given by
h(B ) = { : Q BX () = 1} = X 1 (B ) F .
Notice that this does not hold for quantum systems where some more general
notion is needed.
74
75
6.2. Classical, fuzzy, and quantum logics; observables and states on logics
B(R) . We regard to physical quantities X and X 0 as identical whenever
the corresponding maps x, x 0 : B(R) are the same. If f : R R is a Borel
function, we mean by X 0 = f X a physical quantity taking value f (r ) whenever
X takes value r . The corresponding map is given by B(R) 3 B : x 0 7 x 0 (B ) =
x( f 1 (B )) . Hence we are led naturally to the following
Definition 6.2.7. Let be a logic. A real observable associated with is a mapping x : B(R) verifying:
1. x(;) = 0 and x(R) = 1,
2. if B 1 , B 2 B(R) with B 1 B 2 = ; then x(B 1 ) x(B 2 ),
3. if (B n )nN is a sequence of mutually disjoint Borel sets, then x(nN B n ) =
nN x(B n ).
We write O () for the set of all real observables associated with .
Exercise 6.2.8. Let be a logic and x O (). Show that for any sequence of
Borel sets (B n )nN we have
x(nN B n ) = nN x(B n )
and
x(nN B n ) = nN x(B n ).
Definition 6.2.9. Let be a logic and O () the set of its associated observables.
A real number is called a strict value of an observable x O (), if x({}) 6= 0.
The observable x O () is called discrete if there exists a countable set C =
{c 1 , c 2 , . . .} such that x(C ) = 1; it is called constant if there exists c R such that
x({c}) = 1. It is called bounded if there exists a compact Borel set K such that
x(K ) = 1.
Definition 6.2.10. We call spectrum of x O () the closed set defined by
76
If (a n )nN is a partition of unity, i.e. a family of mutually orthogonal propositions in such that nN a n = 1, there exists a unique discrete observable
admitting as spectral values a given discrete subset {c 1 , c 2 , . . .} of the reals. In
fact, it is enough to define for all n N, x({c n }) = a n and for any B B(R),
x(B ) = n:cn B a n . Notice however that discrete observables do not exhaust
all the physics of quantum mechanics; important physical phenomena involve
continuous observables.
R if 0 B
f 1 (B ) =
; otherwise.
Hence, if 0 B , then o (B ) = f o (B ) = o ( f 1 (B )) = 1, because o is a probability on R; if 0 6 B then similarly o (B ) = 0. Therefore, in all circumstances,
o (B ) = 0 (B ).
If x O () is any observable and B B(R) is such that x(B ) = 0 , then
x (B ) = 0. In fact, for this B , we have 1 B x = o and x (B ) = o ({1}) = 0 ({1}) = 0.
This implies that if x is discrete, the measure x is supported by the set of the
strict values of x.
/Users/dp/a/ens/mq/iq-proca.tex
77
6.2. Classical, fuzzy, and quantum logics; observables and states on logics
Definition 6.2.13. An observable q O () is a question if q({0, 1}) = 1. A question is the necessarily discrete. If q({1}) = a , then q is the only question such
that q({1}) = a; we call it question associated with the proposition a and denote
by q a if necessary.
Definition 6.2.14. Let be a logic. A function p : [0, 1] satisfying
1. p(0) = 0 and p(1) = 1,
2. if (a n )nN is a sequence of mutually orthogonal propositions of , and
P
a = nN a n , then p(a) = nN p(a n )
is called state (or probability measure) on the logic . The set of states on is
denoted by S ().
The concept of probability measure on a logic coincides with a classical
probability measure when the logic is a Boolean -algebra. For non distributive logics however, the associated probability measures are genuine generalisations of the classical probabilities. For standard quantum logics, the associated states are called quantum probabilities.
Theorem 6.2.15. Let p S (), where is a logic.
1. On defining a map p : O () M1+ (R, B(R)), by the formula: for every x
p
O () and for every B B(R), x (B ) = p(x(B )), then p is a state function.
2. Conversely, if is an arbitrary state function, then for every x O (), then
there exists a unique probability measure p S () such that for every x
O () and for every B B(R), x (B ) = p(x(B )).
Proof:
p
78
79
x (B ) = p(x(B )) =
X
nN
c n p n (x(B )) =
X
nN
c n x n (B ).
This decomposition has the following interpretation: the sequence (c n )nN defines a classical probability on N meaning that in the sum defining p, each p n is
chosen with probability c n . Therefore, for each integrable observable x O (),
P
the expectation Ep (x) = nN c n Ep n (x) consists in two averages: a classical average on the choice of p n and a (may be) quantum average Ep n (x).
Exercise 6.3.4. Give a plausible definition of the notion of integrable observable used in the previous remark and then prove the claimed equality: Ep (x) =
P
nN c n Ep n (x)
Definition 6.3.5. A state p S () is said to be pure if the equation p = cp 1 +
(1c)p 2 , for p 1 , p 2 S () and c [0, 1] implies p = p 1 = p 2 . We write S p () for
the set of pure states of . Obviously S p () = Extr S ().
Definition 6.3.6. Let D S () and p 0 S (). We say that p 0 is a superposition of states in D if for a ,
p D, p(a) = 0 p 0 (a) = 0.
P
It is an exercise to show that the state p = nN c n p n defined in the proposition ?? is a superposition of states in D = {p 1 , p 2 , . . .}. In the case is a Boolean
-algebra, the next theorem 6.3.7 shows that this is in fact the only kind of possible superposition. This implies, in particular, the unicity of the decomposition of a classical state into extremal (pure) states. If is a standard quantum
logic, unicity of the decomposition does not hold any longer!
/Users/dp/a/ens/mq/iq-proca.tex
80
1 if a A
0 otherwise.
1
p(A A 0 )
p(A 0 )
1
p(A A c0 ),
1 p(A 0 )
()
()
we get p(A) = p(A 0 )p 1 (A) + (1 p(A 0 ))p 2 (A). Yet, applying (*) and (**) to A 0 ,
we get p 1 (A 0 ) = 1 and p 2 (A 0 ) = 0, hence p 1 6= p 2 . This is in contradiction with
the assumed purity of p. Therefore, we conclude that for all A , we have
p(A) {0, 1}. Replacing A n by A cn if necessary, we can assume without loss of
generality that p(A n ) = 1 for all the sets of the collection generating . Let B =
n A n . Then p(B ) = 1 and consequently B cannot be empty. Now B cannot
contain more than one point either. In fact, the collection of all sets C such
that either B C or B C = ; is a -algebra containing all the sets A n , n N.
Hence, it coincides with . As singletons are members of , the set B must be
a singleton, i.e. B = {a} for some a X. Put then p = a . Finally, let p 0 be a
superposition of states in D (all its elements are pure states). If p 0 = a0 but
p 0 6 D, then p({a 0 }) = 0 for all p D but p 0 ({a 0 }) 6= 0, a contradiction.
i.e. there is a countable collection of subsets A n X, n N, generating by complementation, intersections, and unions.
/Users/dp/a/ens/mq/iq-proca.tex
81
82
1 3: By symmetry.
2 1: Since a (a b) b, on writing a 1 = a (a b) , b 1 = b (a b) , and
c = a b, we find a = a 1 c and b = b 1 c. Since a 1 b, it follows that a 1
b 1 and a 1 c, while, by definition, c b 1 which proves the implication.
Henceforth, the equivalence 1 2 3 is established.
1 4: If a = a 1 c, b = b 1 c and a 1 , b 1 , c mutually orthogonal, write d = a 1
b 1 c and define x to be the discrete observable such that x({0}) = a 1 ,
x({1}) = b 1 , x({2}) = c, and x({3}) = d . Then x({0, 2}) = a and x({1, 2}) = b.
4 5: x(A (A B )c ) = a (a b) and x(B (A B )c ) = b (a b) . On writing
a 1 = a (a b) , a 2 = a b, a 3 = b (a b) , and a 4 = (a b) , we see
that (a i )i =1,...,4 are mutually orthogonal and a 1 a 2 a 3 a 4 = 1. If
A = {a i 1 . . . a i k : k 4; 1 i 1 . . . i k 4},
it is easily verified that A is Boolean sub-algebra of . Since a, b A , this
proves the implication.
5 2: Let A be a Boolean sub-algebra of containing a and b. Now, [a (a
b) ] b = 0. As a, b, a (a b) , b A , it follows that
a (a b) = [(a (a b) ) b]
[(a (a b) ) b ]
= [(a (a b) ) b ]
b.
Therefore a (a b) b.
83
()
action of on p, by (p)(a)
= p(1 (a)), for all a , then is a convex automorphism of S ().
Definition 6.5.1. A map : S () S () is a convex automorphisms if
1. is bijective and
2. if (c n )nN is a sequence of non-negative reals such that
(p n )nN is a sequence of states in S (), then
(
/Users/dp/a/ens/mq/iq-proca.tex
X
nN
cn p n ) =
84
X
nN
nN c n
= 1 and
c n (p n ).
nN c n (p
To further exploit the notions of logic, states, observables, and convex automorphisms, we must specialise the physical system.
/Users/dp/a/ens/mq/iq-proca.tex
85
/Users/dp/a/ens/mq/iq-proca.tex
86
7
Standard quantum logics
We recall that a standard quantum logic was defined in chapter 5 to be the set
of Hilbert subspaces of C-Hilbert space H. For every Hilbert subspace M ,
we denote by P M the orthogonal projection to M . If x O (), then B 7 P x(B ) ,
for B B(R), is a projection-valued measure on B(R). Conversely, for every
projection-valued measure P on B(R), there exists an observable x O () such
that P (B ) = P x(B ) , for all B B(R). We identify henceforth Hilbert subspaces
with the orthogonal projectors mapping the whole space on them (recall exercise 4.4.7.)
7.1 Observables
Lemma 7.1.1. Let M 1 , M 2 . Then propositions associated with M 1 and M 2
are simultaneously verifiable if and only if [P M1 , P M2 ] = 0.
Proof:
(): Propositions M 1 and M 2 are simultaneously verifiable if there exist
mutually orthogonal elements N1 , N2 , N such that M i = Ni N , for
87
7.2. States
i = 1, 2. Then P Mi = P Ni + P N and the commutativity of the projectors
follows immediately.
(): If [P M1 , P M2 ] = 0, let P = P M1 P M2 . Then P is a projection. Define
Q i = P Mi P , for i = 1, 2; it is easily verified that Q i are projections and
PQ i = Q i P = 0. Therefore Q 1Q 2 = Q 2Q 1 = 0. If we define Ni = Q i (H),
for i = 1, 2 and N = P (H), then N1 , N2 , N are mutually orthogonal and
M i = Ni N which proves that M 1 M 2 .
Theorem 7.1.2. Let be a standard logic with associated Hilbert space H. For
any x O (), denote X the self-adjoint (not necessarily bounded) operator on H
with spectral measure given by the mapping B(R) 3 B 7 P x(B ) . Then
1. the map x 7 X is a bijection between O () and self-adjoint operators on
H,
2. the observable x is bounded if and only if X Bh (H),
3. two bounded observables x 1 and x 2 are simultaneously observable if and
only if the corresponding bounded operators X 1 and X 2 commute,
4. if x is a bounded observable and Q R[t ], then the operator associated
with Q x is Q(X ),
5. more generally, if x 1 , . . . , x r are bounded observables any two of them being
simultaneously observable, and Q R[t 1 , . . . , t 2 ], then the observable Q
(x 1 , . . . , x r ) has associated operator Q(X 1 , . . . , X r ).
Proof: Assertions 14 are simple exercises based on the spectral theorem for
self-adjoint operators. Assertion 5 is a direct consequence of theorem 6.4.4.
7.2 States
In chapter 2, we defined (pure) quantum states to be unit vectors of H. In chapter 6, states have been defined as probability measures on a logic. We first show
that in fact rays correspond to states viewed as probability measures on .
/Users/dp/a/ens/mq/iq-stqlo.tex
88
Unit vectors of H are called rays. Let H, with kk = 1 be a ray and denote
by p : [0, 1] the map defined by
3 M 7 p (M ) = | P M = kP M k2 .
We have: p (1) p (H) = 1, p (0) p ({0}) = 0, and if (M n )nN is a sequence of
mutually orthogonal Hilbert subspaces of H and M = nN M n , then
X
X
p (M ) = kP M k2 =
| P Mn =
p (M n ).
nN
nN
Then for all T, T1 , T2 B+ (H) the trace has the following properties
1. is independent of the chosen basis,
2. tr(T1 + T2 ) = tr(T1 ) + tr(T2 ),
3. tr(T ) = tr(T ) for all 0,
4. tr(U T U ) = tr(T ), for all U U(H).
Proof: (To be filled in a later version.)
89
7.2. States
2. We show then that T1 (H) is a vector space. In fact, for every C, due to
the fact that |A| = |||T |, it follows that if A T1 (H) then A T1 (H) as
well.
For T1 , T2 T1 (H), denote by U ,V,W the partial isometries arising into
the polar decompositions T1 +T2 = U |T1 +T2 |, T1 = C |T1 |, and T2 = W |T2 |.
Then,
X
X
e n |U (T1 + T2 )e n
e n | |T1 + T2 |e n =
n
n
X
X
e n |U V |T1 |e n + e n |U W |T2 |e n
=
n
n
X
X
| e n |U V |T1 |e n | + | e n |U W |T2 |e n |.
n
Now,
X
X
1
1
e n |U V |T1 |e n =
|T1 | 2 V U e n | |T1 | 2 e n
n
X
X
1
1
( k|T1 | 2 V U e n k2 )1/2 ( k|T1 | 2 e n k2 )1/2 Cauchy-Scwharz on `2 (N
n
We conclude that
X
X
1
1
1
k|T1 | 2 V U e n k2 =
|T1 | 2 V U e n | |T1 | 2 V U e n
n
n
X
=
e n |U V |T1 |V U e n
n
X
e n |V |T1 |V e n
n
X
e n | |T1 |e n
n
= tr(|T1 |),
because U ,V are partial isometries. The second term is majorised similarly so that tr(|T1 + T2 |) tr(|T1 |) + tr(|T2 |) < , showing that T1 + T2
T1 (H).
3. Using the decomposition of every B B(H) into the combination of four
P
P
unitary operators B = 4i =1 c i Ui , we get tr(T B ) = 4i =1 c i tr(T Ui ) so that
it becomes sufficient to prove that T T1 (p
H) and U U(H) implies that
p
p
1
U T = T T = |T | and |T U | = (T U ) T U =
T
U
,U
T
T
(
H
).
But
|U
T
|
=
(U
T
)
p
U |T |2U = U |T |U ; furthemore U |T |U 0. Hence tr(|T U |) = tr |T | =
tr(|U T |).
/Users/dp/a/ens/mq/iq-stqlo.tex
90
Exercise 7.2.4. Show the T T1 (H) implies that T T1 (H) (hence T 1 (H) is a
bilateral -ideal of T1 (H).
Exercise 7.2.5. Show that T T1 (H) is not necessarily closed with respect to
the operator norm stemming from the Hilbert norm. Nevertheless, T1 (H) is a
Banach space for the k k1 norm defined by kT k1 = tr |T |.
Definition 7.2.6. If D is a bounded, self-adjoint, non-negative, trace-class operator on H, then D is called a von Neumann operator. If further tr(D) = 1, then
D is said to be a density matrix (operator). The set of density matrices on H is
denoted by D(H).
The states p , for a ray of H, can also be described in another way. Let D
be the projection operator on the one-dimensional subspace1 C. Then D is
trace-class and for every X B(H), it follows that D X is also trace-class. Let
(n )nN be an arbitrary orthonormal basis of H; without loss of generality, we
can then assume that 1 = . We have
tr(D X ) = tr(X D )
=
X
nN
n | X D n
= | X
= E (X ).
In particular, if X = P M for M ,
p (M ) = | P M = tr(D P M ).
Lemma 7.2.7. Let (n )nN be an arbitrary sequence of rays in H and (c n )nN an
P
arbitrary sequence of non-negative reals such that nN c n = 1. Denote by D n the
projection operator on the one-dimensional subspace Cn , for n N. Then
D=
X
nN
cn D n
/Users/dp/a/ens/mq/iq-stqlo.tex
91
7.3. Symmetries
Lemma 7.2.9. Let D be a density matrix defined as in lemma 7.2.7 and p : R
the mapping defined by 3 M 7 p(M ) = tr(P M D). Then p S () and moreP
over it can be decomposed into p = nN c n p n .
Proof: First the superposition property follows from the linearity of the trace:
P
P
for all M , we have p(M ) = tr(P M D) = nN c n tr(P M D n ) = nN c n p n (M ). It
is now obvious that p is a state: in fact, p(0) = p({0}) = 0 and p(1) = p(H) = 1.
Conversely, if D is any density matrix, then the map 3 M 7 p(M ) = tr(DP M )
is a state in S (). States of this type are called tracial states. The natural question is whether every state in S () arises as a tracial state. The answer to this
question is one of the most profound results in the mathematical foundations
of quantum mechanics, the celebrated Gleasons theorem:
Theorem 7.2.10 (Gleason). Let H be a complex separable Hilbert space with
3 dim H 0 , D(H) the convex set of density matrices on H, and the logic of
subspaces of H. Then
1. the map D(H) 3 D 7 D S (), defined by D (M ) = tr(DP M ) for all M
, is a convex isomorphism of D(H) on S (),
2. a state p S () is pure if and only if p = p for some ray inH,
3. two pure states p and p are equal if and only if there exists a complex
number c with |c| = 1 such that the rays and verify = c.
The proof, lengthy and tricky, is omitted. It can be found, extending over 13
pages (!), in [31], pages 147160.
7.3 Symmetries
Definition 7.3.1. A linear map S : H H is a symmetry if
1. S is bijective, and
2. for all f , g H, the scalar product is preserved: S f | Sg = f | g .
Exercise 7.3.2. Let Aut() where is the standard quantum logic associated with a given Hilbert space H. Show that
/Users/dp/a/ens/mq/iq-stqlo.tex
92
In general, anti-unitaries may also occur as symmetries. They are not considered in this
course.
/Users/dp/a/ens/mq/iq-stqlo.tex
93
7.3. Symmetries
/Users/dp/a/ens/mq/iq-stqlo.tex
94
8
States, effects, and the corresponding
quantum formalism
95
/Users/dp/a/ens/mq/iq-proca.tex
96
9
Quantum formalism based on the
informational approach
97
Chapter 9
/Users/dp/a/ens/mq/iq-infor.tex
98
10
Two illustrating examples
f (q(t )) = kq(t )
q(0) = q 0
q(0)
= v 0 = 0.
Introducing the new variable p = m q and transforming the second order differential equation into a system of first order equations, we get the vector equation
d
(t ) = A(t ),
dt
where
()
q(t )
q0
(t ) =
, with initial condition (0) =
p0
p(t )
and
0
A=
k
1
m .
T = exp(t A) =
cos(t )
sin(t )
m
k sin(t ) cos(t )
p
k/m. Since det T t = 1, it follows that the evolution
is invertible and
q0
(T t )1 = T t . The orbit of the initial condition (0) =
under the flow reads
0
!
q 0 cos(t )
(T t )t R , where (t ) = T t =
.
q 0 k sin(t )
and =
/Users/dp/a/ens/mq/iq-ilexa.tex
100
The system is classical, hence its logic is a Boolean -algebra; the natural
choice is = B(R2 ). Now observables in O () are mappings x : B(R)
B(R2 ). Identify henceforth indicator functions with Borel sets in B(R2 ) (i.e.
for any Borel set B B(R), instead of considering x(B ) = F B(R2 ) we shall
identify x(B ) = 1 F .)
Let now X : R be any measurable bounded mapping
and chose as
R
x(B ) = 1 X 1 (B ) for all B B(R). Then, on defining X = x(d ), a bijection
is established between x and X . Now since (T t )t R = (exp(t A))t R is the orbit of the initial condition 0 in , the value X (T t ) is well defined for all t R;
we denote by X t () X (T t ). Then
d Xt
d (T t )1
d (T t )2
() = 1 X (T t )
+ 2 X (T t )
dt
dt
dt
d
p
d
q
(t ) + 2 X (T t )
(t ),
= 1 X (T t )
dt
dt
provides the evolution of X under the flow (T t )t .
The Hamiltonian is a very particular measurable bounded map on the phase
space (hence an observable) H : R, having the formula H () = k21 /2 +
22 /2m. It evolves also under the flow (T t )t : Then
p(t )
d Ht
)+
)
() = kq(t )q(t
p(t
dt
m
) + q(t
)(k q(t
))
= kq(t )q(t
= 0.
Thus, the Hamiltonian is a constant of motion. Physically it represents the enkq 2
ergy of the system. Initially, H (q 0 , p 0 ) = 2 0 = E and during the flow, the energy always remains E , so that the energy takes arbitrary (but constant with
) and
respect to the flow) values E R+ . Moreover, 1 H (T t ) = kq(t ) = p(t
p(t )
t
). Hence we recover the Hamilton equations
2 H (T ) = m = q(t
dq
H
(t ) =
= 2 H
dt
p
dp
H
(t ) =
= 1 H .
dt
q
Therefore, ddXt t = 1 X 2 H + 2 X (1 H ) = L H X with L H = (1 H 2 2 H 1 ).
Hence, denoting for every two function f , g C 1 () by { f , g } = 1 f 2 g 2 f 1 g
the Poissons bracket, we have for the flow of an observable, assuming integrability of the evolution equation, X t = exp(t L H )X . This means that the flow
/Users/dp/a/ens/mq/iq-ilexa.tex
101
(t )n
X
{H , {H , . . . {H , X } . . .}}.
n=0 n!
d 1 d 2 =
=
Z
ZB
B
(1 , 2 )
d x1 d x2
(x 1 , x 2 )
d x 1 d x 2 = (B ),
102
Notice that, as a consequence of the previous theorem, exp(t L H ) is formally
unitary on L 2 (, F , ).
Any probability measure p on is a state. We have for all B B(R), x (B ) =
p(x(B )) = p(X 1 (B )) while x t = p(x t (B )) = p(X t1 (T t B )) = p(x(T t B )). Hence
(p)(M
) = p D ((M ))
= tr(PU M D)
= tr(U P M U D)
= tr(P M D (U ) ),
with D (U ) = U DU . Physics remains invariant under time translations. Hence
time translation (evolution) must be a symmetry implemented by a unitary
operator U (t ) acting on H. Define U (t ) = exp(i t H /) (this a definition of
H .) Then H is formally self-adjoint, hence an observable (a very particular
one!) generating the Lie group of time translations. It will be shown below
that H is time invariant. Now U (t ) acts on rays of H to give a flow. Denoting
(t ) = U (t ), we have the Schrdingers evolution equation in the Schrdingers
picture:
d
(t ) = H (t ).
i
dt
/Users/dp/a/ens/mq/iq-ilexa.tex
103
= exp(t L H )X
(i t )n
X
[H , [H , . . . , [H , X ] . . .]].
=
n
n=0 n!
Physics remains invariant also by space translations. Hence they must correspond to a symmetry implemented by a unitary transformation.
Lemma 10.1.2. The operator x is formally skew-adjoint on L 2 (R).
Proof: For all f , g S(R) (dense in L 2 (R)), we have, f | x g =
R
ddx f (x)g (x)d x + f g |
.
104
A0 = 0
p
nn1 , for n 1, and
4. A An = nn , for n 0.
/Users/dp/a/ens/mq/iq-ilexa.tex
105
n!
,
(n2 j )!2 j j !
2j
2( j + 1)
)c n+1, j =
)c n+1, j +1
n +1
(n + 1)(n 2 j )
and if
n (x) =
[n/2]
X
(1) j c n, j x n2 j ,
j =0
then
(x
while x n =
P[n/2]
j =0
d
) n (x) = n+1 (x)
dx
c n, j n2 j (x).
106
1. H | n = (1/2 + n)| n ,
2. A | n =
3. A| n =
p
n + 1| n + 1 ,
p
n| n 1 , and
4. A A| n = n| n .
Figure 10.2: Comparison of probability densities. In blue is depicted the probability density of the classical oscillator. In red the corresponding density for
the quantum oscillator for n = 10 (left) and n = 60 (right).
/Users/dp/a/ens/mq/iq-ilexa.tex
107
Figure 10.3: Comparison of distribution functions. In blue is depicted the distribution of the classical oscillator. In red the corresponding distribution for the
quantum oscillator for n = 1 (left), n = 10 (middle), and n = 30 (right). Already
for n = 30, the classical and quantum distributions are almost indistinguishable.
/Users/dp/a/ens/mq/iq-ilexa.tex
108
11
Quantifying information: classical
and quantum
Entropy is closely related to irreversibility since the second principle of thermodynamics states: Entropy of an isolated system is a non decreasing function
of time. It can remain constant only for reversible evolutions. For a system A
undergoing an irreversible transformation the entropy increases; however the
system can be considered as part of a larger isolated composite system (A and
environment), undergoing globally a reversible transformation. In that case
the total entropy (of the system A and of the environment) remains constant
but since the entropy of A must increase, the entropy of the environment must
decrease1 hence the missing information decreases. In other words, when the
1
Notice that this assertion is not in contradiction with the second principle of thermodynamics because the environment is not isolated.
/Users/dp/a/ens/mq/iq-qinfo.tex
110
/Users/dp/a/ens/mq/iq-qinfo.tex
111
/Users/dp/a/ens/mq/iq-qinfo.tex
112
12
Turing machines, algorithms,
computing, and complexity classes
go
go
go
a0
0
1
]
s0
go
go
halt
d
L
L
R
Mind that during Turings times no computer was physically available. The external tape
was invented by Alan Turing who was fascinated by typewritters as an external storage
device.
/Users/dp/a/ens/mq/iq-turin.tex
114
If the programme, described by this Turing machine, starts with the head over
any non-blank symbol of the input string, it ends with the head over the leftmost non-blank symbol while the string of symbols remains unchanged.
Other equivalent variants of the deterministic Turing machine may have
displacement sets with a 0 (do not move) displacement, have their alphabet A
partitioned into external and internal alphabet, etc. The distinction into internal and external alphabet is particularly useful in the case of semi-infinite tape,
an internal character , identified as first symbol, can be used to prevent the
head from going outside the tape. It is enough to define U (, go) = (, go, R).
Notation 12.1.3. If W is a finite set, we denote by W = nZ+ W n and W =
W = W Z+ . Notice that Z+ = {0, 1, 2, . . .} 6= N = {1, 2, . . .} and that W 0 = {;}. Elements of W are called words of finite length over the alphabet W . For every
w W , there exists n Z+ such that w W n ; we denote then by |W | = n the
length of the word w.
For every A b , we denote by A the completion of the word by
blanks, namely = (1 , . . . , || , ], ], ], . . .).
Considering the example 12.1.2, we can, without loss of generality, always
assume that the machine starts at the first symbol of the input string = A b .
Starting from (, s 0 , h 0 = 1), successive applications of the transition function
U induce a dynamical system on X = A S Z. A configuration is an instantaneous description of the word written on the tape, the internal state of the
machine, and the position of the head, i.e. an element of X.
Let = inf{n 1 : s n S f }. The programme starting from initial configuration (, s 0 , h 0 = 1) stops running if < , it never halts when = . While
1 n < , the sequence ((n) , s n , h n )n is defined by updates of single characters; if, for 0 n < , we have u((n)
, s ) = (a 0 , s 0 , d ), then ((n+1) , s n+1 , h n+1 ),
hn n
is defined by
s n+1 = s 0
h n+1 = h n + d
(n)
(n)
0 (n)
(n+1) = ((n)
1 , . . . , h 1 , a , h +1 , . . . , |(n) | ).
n
If the machine halts at some finite instant, the output is obtained by reading
the tape from left to right until the first blank character. The sequence of words
((n) )n is called a computational path or computational history starting from .
/Users/dp/a/ens/mq/iq-turin.tex
115
116
P 6= PSPACE .
NP
class if
117
http://www.claymath.org/millennium/
/Users/dp/a/ens/mq/iq-turin.tex
118
n1
X
x k b k Zb n .
k=0
119
x0
y1
y0
z2
z1
z0
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
1
0
0
1
1
0
1
1
1
0
0
1
1
0
1
1
0
1
1
0
0
1
0
0
1
0
1
0
1
1
0
1
0
0
1
0
1
1
0
1
0
Table 12.1: The truth table of the Boolean function A4 A3 implementing the
addition with carry of two binary 2-digit numbers.
Classical computers are based on gates {XOR, AND} for example. It is easily
shown that these gates are irreversible. Therefore it is intuitively clear why classical computers can produce information. What is much less intuitively clear
is how quantum processes can produce information since they are reversible
(unitary).
In 1973, B ENNETT predicted that it is possible to construct reversible universal gates. In 1982, F REDKIN exemplifies such a reversible gate. Fredkins gate
is a 3 inputs - 3 outputs gate, whose truth tableau is given in table ??. This gate
produces both AND (since inputs 0, x, y return outputs x y, x y, x) and NOT
gates (since inputs 1, 0, x return outputs x, x, x.) The gates AND and NOT forming a complete basis for Boolean circuits, the universality of Freidkins gate is
established.
In 1980, B ENIOFF describes how to use quantum mechanics to implement
a Turing machine, in 1982, F EYNMAN proves that there does not exist a Turing
machine (either deterministic or probabilistic) on which quantum phenomena can be efficiently simulated; only a quantum Turing machine could do so.
Finally, in 1985, D EUTSCH constructs (on paper) a universal quantum Turing
/Users/dp/a/ens/mq/iq-turin.tex
120
Input
a b c
0 0 0
0 0 1
0 1 0
0 1 1
1 0 0
1 0 1
1 1 0
1 1 1
Output
a b0 c 0
0 0 0
0 0 1
0 1 0
1 0 1
1 0 0
0 1 1
1 1 0
1 1 1
0
Table 12.2: The truth table of Fredkins gate. We remark that c 0 = c and if c = 0
then (a 0 = a and b 0 = b) else (a 0 = b and b 0 = a.)
machine.
121
Define now onservables having (| )A , (| s )sS , and (| h )hZ as respective eigenvectors. To do so, identify the sets A with {0, . . . , |A| 1} and S with
and H the self-adjoint operators describing these
{0, . . . , |S|1}. Denoty by T , S,
observables, i.e.
S =
|S|1
X
s| s s |
s=0
h| h h |
hZ
= i Z Ti where Ti =
|A|1
X
a| a a |.
a=0
Y
j Z\{h}
j ,0 .
j
122
system. To proceed, suppose that S f = {halt} {0} and introduce a halting flag
operator F = | 0 0 |. Once the state s is set to 0, the function c is such that U
does not any longer change either the state s or the result of the computation.
A predicate is a projection operator P = | |. Let the machine evolve
for some time n: it is at the state | n = U n | . Perform the measurement
n | P F I n = p [0, 1].
Definition 12.8.4. A language L belongs to the BQP complexity class if there is
a machine M QTM such that
if L, then the machine accepts with probability p > 2/3,
if 6 L, then the machine rejects with probability p > 2/3,
within a running time poly(||).
Theorem 12.8.5.
/Users/dp/a/ens/mq/iq-turin.tex
123
/Users/dp/a/ens/mq/iq-turin.tex
124
13
Cryptology
125
Appeared as a first patent US Patent 1310719 issued on 22 July 1919, and further improved
in a series of patents: US Patent 1416765, US Patent 1584749, and US Patent 1613686.
/Users/dp/a/ens/mq/iq-crypt.tex
126
127
Proof:
cd
= m ed
ed
ed
mod n
= m 1+k(n) ,
If Eve intercepts the message, to compute d she must know (n), hence
the factoring of n into primes. Security of the protocol is based on the conjecture that it is algorithmically hard to factor n. If we denote by N = log n, then
it is worth noticing that when the RSA protocol has been introduced, the best
known algorithm of factor n run in exp(N ) time. The best 2 known algorithm
nowadays [18] runs in exp(N 1/3 (log N )2/3 ) time. This algorithmic improvement,
combined with the increasing in the computational capabilities of computers,
allows the factoring of a 1000 digits number in ca. 8 months instead of a time
exceeding the age of the universe at the moment the algorithm has been proposed. Until May 2007, the RSA company ran an international contest offering
several hundreds thousand dollars to whoever could factor multi-digit numbers they provided on line. When the contest stopped the company gave the
official reasons explained in RSA factoring challenge.
2
/Users/dp/a/ens/mq/iq-crypt.tex
128
Q
Since | 6= 0 we get | n1
i =1 i | i =Q1 and since | 6= exp(i )| psi ,
it follows that 0 < | | | < 1. Subsequently, n1
i =1 | i | i | > 1 but this is impossible since for every i , | i | i | 1.
This theorem is at the basis of the BB84 quantum key distribution protocol [8]. Alice and Bob communicate through a quantum and a classical public
channels; they agree publicly to use two different orthonormal bases of H = C2
(describing the photon polarisation):
+
B + = {+
0 = | 0 , 1 = | 1 }
|0|1 |0+|1
, 1 =
}.
B = {
p
p
0 =
2
2
The first element of each basis is associated with the bit 0, the second with the
bit 1. Moreover Alice and Bob agree on some integer n = (4 + )N with some
> 0, where N is the length of the message they wish to exchange securely;
it will be also the length of their key. Alice needs also to know the function
/Users/dp/a/ens/mq/iq-crypt.tex
129
+
0
+
1
T (x, y) =
if (x, y) = (0, 0)
if (x, y) = (0, 1)
if (x, y) = (1, 0)
if (x, y) = (1, 1).
130
131
P((a k0 1 , . . . , a k0 L ) = (a k1 , . . . , a kL )|a, b) = 1.
132
/Users/dp/a/ens/mq/iq-crypt.tex
133
/Users/dp/a/ens/mq/iq-crypt.tex
134
14
Elements of quantum computing
In this chapter, B denotes the set {0, 1} and elements b B are called bits; H
will denote C2 and rays | H are called qubits. Similarly, arrays of n bits are
denoted by b = (b 1 , . . . , b n ) B n ; arrays of n qubits by | = | 1 n Hn .
135
136
1
.
..
a b
c d
..
.
a b
U(2).
,
c d
1
Moreover, the sequence of matrices appearing in the product can be explicitly
constructed in a running time O (m 3 )poly(log(1/)) where = kU V k.
Proof: An exercise if one recalls that for all c 1 , c 2 C, there exists a unitary operator W U(2) such that
p
c1
|c 1 |2 + |c 2 |2
.
W
=
c2
0
137
Definition 14.2.5. A unitary operator U : Hn Hn is approximated by a unitary operator U : HN HN , with N n, within if for all | Hn
kU 0 (| | 0N n ) U | | 0N n k kk.
Definition 14.2.6. For every unitary operator U : Hn Hn there exists a unitary operator C (U ) : H Hn H Hn , called the controlled-U operator, defined for all | Hn by
C (U )| | =
||
if = 0
| U | if = 1
C (U )| 1 k | =
| 1 k |
if 1 k = 0
| 1 k U | if 1 k = 1
0 1
Example 14.2.7. Let 1 =
be the unitary operator corresponding to the
1 0
where is the Toffoli gate.
classical NOT gate. Then C 2 (1 ) = ,
1 1
1 0
(Hadamard gate) and K =
(phase gate), is called
with H =
1 1
0 i
the standard computational basis.
p1
2
138
Exercise 14.2.10. Let 0,...,3 be the 3 Pauli matrices augmented by the identity
1
0
.
matrix, H the Hadamard gate, and () =
0 exp(2i )
1. Show that if A M2 (C) with A 2 = 1 and R, then
exp(i A) = cos 0 + i sin A.
), where
2. Let R j () = exp(i 2 j ), for j = 1, 2, 3 and R n () = exp(i 2 n ~
2
2
2
= (1 , 2 , 3 ). Express R j ()
n = (N1 , n 2 , n 3 ) with n 1 + n 2 + n 3 = 1 and ~
and R n () on the basis 0 , . . . , 3 .
3. Show that H = exp(i )R 1 ()R 3 (), for some , , to be determined.
4. If | C2 is a ray represented by a vector of the Bloch sphere S2 = {x
R3 : kxk2 = 1}, show that
R n ()| = | Tn ()x
where Tn ()x is the rotation of x around n by an angle .
5. Show that every U U(2) can be written as
U = exp(i )R n ()
for some , R.
6. Show that every U U(2) can be written as
U = exp(i )R 3 ()R 2 ()R 3 ()
for some , , , R.
and n are two not parallel vectors of S2 . Show that every
7. Suppose that m
U U(2) can be written as
U = exp(i )R n (1 )R m (1 )R n (2 )R m (2 ) .
8. Establish identities
H 1 H
= 3
H 2 H
= 2
H 3 H
H ( )H
8
= 1
= exp(i )R 1 ( )
4
for some .
/Users/dp/a/ens/mq/iq-qcomp.tex
139
1 1 1
H=p
.
2 1 1
1
H | = p ((1) | + | 1 ), B.
2
H
7
1 X
| 000 = p
| x .
8 x=0
1
0
() =
.
0 exp(2i )
()| = exp(2i )|
+ )H ()H | 0 = cos | 0 + exp(i ) sin | 1 .
4 2
0
C (1 ) =
0
0
0
1
0
0
0
0
0
1
0
.
1
0
140
0
C (()) =
0
0
0
0
0
0
.
1
0
0 exp(2i )
0
1
0
0
For x, y B ,
C (())| x y = exp(2i x y)| x y .
C 2 (3 )| x y z = | x, y, (x y) z .
Uf ||0 =
2m
1
X
2m/2
x=0
| x, f (x) .
/Users/dp/a/ens/mq/iq-qcomp.tex
141
/Users/dp/a/ens/mq/iq-qcomp.tex
142
15
The Shors factoring algorithm
143
Chapter 15
/Users/dp/a/ens/mq/iq-shora.tex
144
16
Error correcting codes, classical and
quantum
145
References
/Users/dp/a/ens/mq/iq-proca.tex
146
Bibliography
References
[11] Giulio Chiribella, Giacomo Mauro DAriano, and Paolo Perinotti. Theoretical framework for quantum networks. Phys. Rev. A (3), 80(2):022339, 20,
2009. 23
[12] Giulio Chiribella, Giacomo Mauro DAriano, and Paolo Perinotti. Informational derivation of quantum theory. Phys. Rev. A, 84:012311, Jul 2011.
23
[13] A. Einstein, B. Podolsky, and N. Rosen. Can quantum-mechanical description of physical reality be considered complete? Phys. Rev., 47:777780,
May 1935. 15
[14] Nicolas Gisin, Grgoire Ribordy, Wolfgang Tittel, and Hugo Zbinden.
Quantum cryptography. Rev. Mod. Phys., 74:145195, Mar 2002. 28
[15] Anthony J.G. Hey. Feynman and computation. Perseus Books Publishing,
1998. 22
[16] John K. Hunter and Bruno Nachtergaele. Applied analysis. World Scientific
Publishing Co. Inc., River Edge, NJ, 2001. 31, 33
[17] Richard V. Kadison and John R. Ringrose. Fundamentals of the theory of
operator algebras. Vol. I, volume 15 of Graduate Studies in Mathematics.
American Mathematical Society, Providence, RI, 1997a. Elementary theory, Reprint of the 1983 original. 36, 50
[18] A. K. Lenstra and H. W. Lenstra, Jr. Algorithms in number theory. In Handbook of theoretical computer science, Vol. A, pages 673715. Elsevier, Amsterdam, 1990. 4, 128
[19] A. K. Lenstra and H. W. Lenstra, Jr., editors. The development of the number
field sieve, volume 1554 of Lecture Notes in Mathematics. Springer-Verlag,
Berlin, 1993. 6
[20] Arjen K. Lenstra. Integer factoring. Des. Codes Cryptogr., 19(2-3):101128,
2000. Towards a quarter-century of public key cryptography. 128
[21] H. Maassen. Quantum probability and quantum information theory. In
Quantum information, computation and cryptography, volume 808 of Lecture Notes in Phys., pages 65108. Springer, Berlin, 2010. 16, 20
[22] Dimitri Petritis. Markov chains on measurable spaces, 2012. Lecture notes
of the University of Rennes 1. 21
/Users/dp/a/ens/mq/iq-proca.tex
148
BIBLIOGRAPHY
[23] Michael Reed and Barry Simon. Methods of modern mathematical physics.
I. Functional analysis. Academic Press, New York, 1972. 31, 36
[24] R. L. Rivest, A. Shamir, and L. Adleman. A method for obtaining digital signatures and public-key cryptosystems. Comm. ACM, 21(2):120126, 1978.
6, 127
[25] Walter Rudin. Real and complex analysis. McGraw-Hill Book Co., New
York, third edition, 1987. 31
[26] Walter Rudin. Functional analysis. International Series in Pure and Applied Mathematics. McGraw-Hill Inc., New York, second edition, 1991. 48
[27] Raymond A. Ryan. Introduction to tensor products of Banach spaces.
Springer Monographs in Mathematics. Springer-Verlag London Ltd., London, 2002. 36
[28] Claude Shannon. Communication theory of secrecy systems. Bell System
Technical Journal, 28:656715, 1949. 126
[29] Albert Nikolaevich Shiryayev. Probability, volume 95 of Graduate Texts in
Mathematics. Springer-Verlag, New York, 1984. Translated from the Russian by R. P. Boas. 11
[30] Peter W. Shor. Polynomial-time algorithms for prime factorization and
discrete logarithms on a quantum computer. SIAM J. Comput., 26(5):1484
1509, 1997. 4, 6, 125
[31] V. S. Varadarajan. Geometry of quantum theory. Vol. I. D. Van Nostrand Co.,
Inc., Princeton, N.J.-Toronto, Ont.-London, 1968. The University Series in
Higher Mathematics. 92
[32] V. S. Varadarajan. Geometry of quantum theory. Springer-Verlag, New York,
second edition, 1985. 23
[33] V. S. Varadarajan. Geometry of quantum theory. Springer-Verlag, New York,
second edition, 1985. 73
[34] Gilbert S. Vernam. Cipher printing telegraph systems for secret wire and
radio telegraphic communications, volume 55. 1926. 126
[35] Johann von Neumann.
Mathematische Grundlagen der Quantenmechanik. Unvernderter Nachdruck der ersten Auflage von 1932. Die
Grundlehren der mathematischen Wissenschaften, Band 38. SpringerVerlag, Berlin, 1968. 22
/Users/dp/a/ens/mq/iq-proca.tex
149
Index
[36] John von Neumann. Collected works. Vol. I: Logic, theory of sets and quantum mechanics. General editor: A. H. Taub. Pergamon Press, New York,
1961a. 23
[37] John von Neumann. Collected works. Vol. II: Operators, ergodic theory and
almost periodic functions in a group. General editor: A. H. Taub. Pergamon
Press, New York, 1961b. 23
[38] John von Neumann. Collected works. Vol. III: Rings of operators. General
editor: A. H. Taub. Pergamon Press, New York, 1961c. 23
[39] John von Neumann. Collected works. Vol. IV: Continuous geometry and
other topics. General editor: A. H. Taub. Pergamon Press, Oxford, 1962. 23
[40] John von Neumann. Collected works. Vol. V: Design of computers, theory of
automata and numerical analysis. General editor: A. H. Taub. A Pergamon
Press Book. The Macmillan Co., New York, 1963a. 23
[41] John von Neumann. Collected works. Vol. VI: Theory of games, astrophysics, hydrodynamics and meteorology. General editor: A. H. Taub. A
Pergamon Press Book. The Macmillan Co., New York, 1963b. 23
[42] Joachim Weidmann. Linear operators in Hilbert spaces, volume 68 of Graduate Texts in Mathematics. Springer-Verlag, New York, 1980. Translated
from the German by Joseph Szcs. 36
/Users/dp/a/ens/mq/iq-proca.tex
150
Index
algebra, 42
Banach, 43
Boolean, 72
Boolean -, 72
commutative, 42
C , 44
normal element of an, 43
normed, 43
self-adjoint element of an, 43
unital, 42
unitary element of an, 43
algebraic
adjoint, 43
isometry, 43
tensor product, 24
algorithm
Monte Carlo, 118
automorphism
convex, 84
computational
space, 116
time, 116
computational history, 115
computational path, 115
conditional expectation
classical, 34
consistency, 10
density matrix, 91
determinant, 52
direct sum, 34
eigenvalue, 55
entanglement, 25
entropy, 110
equation
Hamilton, 13
Schroedinger, 103
Euler function, 127
expected value, 79
Banach algebra
general linear group of invertible function-valued measure, 61
elements of, 56
gate
invertible element of, 56
Freidkin, 120
Boolean circuit, 119
complement, 72
completion, 32
complexity class, 116
BPP , 118
NP , 117
P , 117
PSPACE , 117
computable function, 116
harmonic oscillator
classical, 99
quantum, 103
Hasse diagram, 71
hatling flag, 123
Hilbert
norm, 32
inequality
151
Index
Cauchy-Schwarz-Buniakovski, 32
involution, 42
isometry, 47
partial, 48
joint probability distribution, 84
language, 116
lattice, 70
atom (in a), 72
atomic, 72
automorphism, 72
complemented, 72
Dilworth, 72
distributive, 71
isomorphism, 72
modular, 72
orthocomplemented, 74
orthomodular, 72
lattice
-complete, 72
lattice of propositions, 70
logic, 74
classical, 75
probability measure on, 78
standard quantum, 75
state on, 78
measurement, 7
classical, 14
quantum, 23
momentum operator, 67
observable, 7
associated with a logic, 76
bounded, 76
classical, 14, 73
constant, 76
discrete, 76
quantum, 23
spectral value of, 76
spectrum of, 76
/Users/dp/a/ens/mq/iq-proca.tex
INDEX
classical, 13
quantum, 23
Poisson bracket, 101
polynomial growth, 116
poset, 70
position operator, 66
predicate, 116
predicate:decidable, 116
probability distribution, 79
projection, 35, 46
proposition
simultaneously verifiable, 82
quantum circuit, 135
question, 78
associated with proposition, 78
random variable, 9
law of, 9
ray, 89
representation
faithful, 50
of locally compact group, 85
space, 50
, 49
unitarily equivalent, 50
resolvent
set, 54
scalar product, 32
sequence
Cauchy, 32
fundamental, 32
simultaneous observability, 82
space
Banach, 32
complete, 32
Hilbert, 32
metric, 32
normed, 32
pre-Hilbert, 32
spectral measure, 62
/Users/dp/a/ens/mq/iq-proca.tex
spectral radius, 58
spectrum, 54, 57
continuous, 55
point, 55
residual, 55
-algebra, 43
-homomorphism, 43
state
classical, 13
entagled, 25
entangled, 25
pure, 80
superposition, 80
tracial, 92
state function, 77
stochastic process, 10
symmetry, 92
tensor product, 24
theorem
Gelfand-Namark, 50
Gleason, 92
Kolmogorovs existence, 11
Liouville, 102
of spectral decomposition, 64
time evolution
classical, 14
quantum, 23
trace, 89
Turing machine
deterministic, 114
non-deterministic, 117
pre-quantum, 121
probabilistic, 118
quantum, 122
variance, 79
yes-no question, 14
153