The diagonalization method in quantum recursion theory

Karl Svozil∗
Institute for Theoretical Physics, Vienna University of Technology,
Wiedner Hauptstraße 8-10/136, A-1040 Vienna, Austria

Abstract
As quantum parallelism allows the effective co-representation of classical mutually exclusive states, the
diagonalization method of classical recursion theory has to be modified. Quantum diagonalization involves
unitary operators whose eigenvalues are different from one.

PACS numbers: 03.67.Hk,03.65.Ud

Keywords: Quantum information, quantum recursion theory, halting problem

1
INTRODUCTION

The reasoning in formal logic and the theory of recursive functions and effective computabil-
ity [1–6], at least insofar as their applicability to worldly things is concerned [7], makes implicit
assumptions about the physical meaningfulness of the entities of discourse; e.g., their actual physi-
cal representability and operationalizability [8]. It is this isomorphism or correspondence between
the phenomena and theory and vice versa — postulated by the Church-Turing thesis [9] — which
confers power to the formal methods. Therefore, any finding in physics presents a challenge to
the formal sciences; at least insofar as they claim to be relevant to the physical universe, although
history shows that the basic postulates have to be re-considered very rarely.
For example, the fundamental atom of classical information, the bit, is usually assumed to be
in one of two possible mutually exclusive states, which can be represented by two distinct states of
a classical physical system. These issues have been extensively discussed in the context of energy
dissipation associated with certain logical operations and universal (ir)reversible computation [10–
13].
In general, all varieties of physical states, as well as their evolution and transformations,
are relevant for propositional logic as well as for a generalized theory of information. Quan-
tum logic [14], partial algebras [15, 16], empirical logic [17, 18] and continuous time computa-
tions [19] are endeavors in this direction. These states need not necessarily be mapped into or
bounded by classical information. Likewise, physical transformations and manipulations avail-
able, for instance, in quantum information and classical continuum theory, may differ from the
classical paper-and-pencil operations modeled by universal Turing machines. Hence, the computa-
tional methods available as “elementary operations” have to be adapted to cope with the additional
physical capabilities [20].
Indeed, in what follows it is argued that, as quantum theory offers nonclassical states and oper-
ators available in quantum information theory, several long-held assumptions on the character and
transformation of classical information have to be adapted. As a consequence, the formal tech-
niques in manipulating information in the theory of recursive functions and effective computability
have to be revised. Particular emphasis is given to undecidability and the diagonalization method.

2
QUANTUM INFORMATION THEORY

As several fine presentations of quantum information and computation theory exist (cf.
Refs. [21–29] for a few of them), there is no need of an extended exposition. In what follows,
we shall mainly follow Mermin’s notation [29, 30]. For the representation of both a single clas-
sical and quantum bit, suppose a two-dimensional Hilbert space. (For physical purposes a linear
vector space endowed with a scalar product will be sufficient.) Let the superscript “T ” indicate
transposition, and let |0i ≡ (1, 0)T and |1i ≡ (0, 1)T be the orthogonal vector representations of
the classical states associated with “falsity” and “truth,” or “0” and “1,” respectively.
From the varieties of properties featured by quantum information, one is of particular impor-
tance for quantum recursion theory: the ability to co-represent classically distinct, contradictory
states of information via the generalized quantum bit state
 
α0
|ψi = α0 |0i + α1 |1i ≡  , (1)
α1

with the normalization |α0 |2 + |α1 |2 = 1. This feature is also known as quantum parallelism,
alluding to the fact that n quantum bits can co-represent 2n classical mutually exclusive states
© ª
|i1 i2 · · · in i | i j ∈ {0, 1}, j = 1, . . . , n of n classical bits.
As will be argued below, recursion
 theoretic diagonalization can be symbolized by the diagonal-
0 1
ization or “not” operator X =  , transforming |0i into |1i, and vice versa. The eigensystem
1 0
of the diagonalization operator X is given by the two 50:50 mixtures of |0i and |1i with the two
eigenvalues 1 and −1; i.e.,
1 1
X √ (|0i ± |1i) = ± √ (|0i ± |1i) = ±|ψ± i. (2)
2 2
In particular, the state |ψ+ i associated with the eigenvalue +1 is a fixed point of the operator X.
Note that, provided that |ψi 6∈ {|0i, |1i}, a quantum bit is not in a pure classical state “relative
to” the propositions corresponding to the projectors |0hi0| and |1hi1|. Any practical determination
of the quantum bit amounts to a measurement of the state “along” one context [31] or base, such as
the base “spanned” by {|0i, |1i}. Any such single measurement will be indeterministic (provided
that the basis does not coincide with {|ψ+ i, |ψ− i}); in particular, |hψ± |0i|2 = |hψ± |1i|2 = 1/2.
That is, if the fixed point state and the measurement context mismatch, by Born’s postulate [32,
33], the outcome of a single measurement occurs indeterministically, unpredictably and at random.

3
Hence, in terms of the quantum states |0i and |1i corresponding to the classical states, the fixed
point remains indeterminate.
In what follows it is argued that, due to the superposition principle, the quantum recursion
theoretic diagonalization method has to be reformulated as a fixed point argument. Application of
the diagonal operator X yields no reductio ad absurdum. Instead, undecidability is recovered as a
natural consequence of quantum coherence and of the unpredictability of certain quantum events.

DIAGONALIZATION

For comprehensive reviews of recursion theory and the diagonalization method the reader is
referred to Refs. [1–6]. Therefore, only a few hallmarks will be stated. As already pointed out by
Gödel in his classical paper on the incompleteness of arithmetic [34], the undecidability theorems
of formal logic [2] are based on semantical paradoxes such as the liar [35] or Richard’s paradox.
A proper translation of the semantic paradoxes into formal proofs results in the diagonalization
method. Diagonalization has apparently first been applied by Cantor to demonstrate the unde-
numerability of real numbers [36]. It has also been used by Turing for a proof of the recursive
undecidability of the halting problem [37].
A brief review of the classical algorithmic argument will be given first. Consider a universal
computer C. For the sake of contradiction, consider an arbitrary algorithm B(X) whose input is a
string of symbols X. Assume that there exists a “halting algorithm” HALT which is able to decide
whether B terminates on X or not. The domain of HALT is the set of legal programs. The range of
HALT are classical bits (classical case) and quantum bits (quantum mechanical case).
Using HALT(B(X)) we shall construct another deterministic computing agent A, which has as
input any effective program B and which proceeds as follows: Upon reading the program B as
input, A makes a copy of it. This can be readily achieved, since the program B is presented to A
in some encoded form pBq, i.e., as a string of symbols. In the next step, the agent uses the code
pBq as input string for B itself; i.e., A forms B(pBq), henceforth denoted by B(B). The agent now
hands B(B) over to its subroutine HALT. Then, A proceeds as follows: if HALT(B(B)) decides that
B(B) halts, then the agent A does not halt; this can for instance be realized by an infinite DO-loop;
if HALT(B(B)) decides that B(B) does not halt, then A halts.
The agent A will now be confronted with the following paradoxical task: take the own code as
input and proceed.

4
 Classical case

Assume that A is restricted to classical bits of information. To be more specific, assume that
HALT outputs the code of a classical bit as follows (↑ and ↓ stands for divergence and convergence,
respectively): 
 |0i if B(X) ↑
HALT(B(X)) = . (3)
 |1i if B(X) ↓

Then, whenever A(A) halts, HALT(A(A)) outputs |1i and forces A(A) not to halt. Conversely,
whenever A(A) does not halt, then HALT(A(A)) outputs |0i and steers A(A) into the halting mode.
In both cases one arrives at a complete contradiction. Classically, this contradiction can only be
consistently avoided by assuming the nonexistence of A and, since the only nontrivial feature of A
is the use of the peculiar halting algorithm HALT, the impossibility of any such halting algorithm.

Quantum mechanical case

As has been argued above, in quantum information theory a quantum bit may be in a linear
coherent superposition of the two classical states |0i and |1i. Due to the superposition of classical
bit states, the usual reductio ad absurdum argument breaks down. Instead, diagonalization pro-
cedures in quantum information theory yield quantum bit solutions which are fixed points of the
associated unitary operators.
In what follows it will be demonstrated how the task of the agent A can be performed consis-
tently if A is allowed to process quantum information. To be more specific, assume that the output
of the hypothetical “halting algorithm” is a quantum bit

HALT(B(X)) = |ψi . (4)

We may think of HALT(B(X)) as a universal computer C0 simulating C and containing a dedicated

halting bit, which it the output of C0 at every (discrete) time cycle. Initially (at time zero), this
halting bit is prepared to be a 50:50 mixture of the classical halting and non-halting states |0i and
|1i with equal phase; i.e., |ψ+ i. If later C0 finds that C converges (diverges) on B(X), then the
halting bit of C0 is set to the “classical” values |1i or |0i.
The emergence of fixed points can be demonstrated by a simple example. Agent A’s diagonal-
ization task can be formalized as follows. Consider for the moment the action of diagonalization
on the classical bit states. (Since the quantum bit states are merely a linear coherent superposition

5
thereof, the action of diagonalization on quantum bits is straightforward.) Diagonalization effec-
tively transforms the classical bit value |0i into |1i and vice versa. Recall that in equation (3), the
state |1i has been identified with the halting state and the state |0i with the non-halting state.
The evolution representing diagonalization (effectively, agent A’s task) can be expressed by the
unitary operator D as
D|0i = |1i and D|1i = |0i . (5)

Thus, D acts essentially as a not-gate corresponding to the operator X. In the above state basis, D
can be represented by  
0 1
D=X=  . (6)
1 0
D will be called diagonalization operator, despite the fact that the only nonvanishing components
are off-diagonal.
As has been pointed out earlier, quantum information theory allows a linear coherent superpo-
sition |ψi of the “classical” bit states |0i and |1i. D has a fixed point at the quantum bit state
 
1 1 1
|ψ+ i = √ (|0i + |1i) ≡ √   . (7)
2 2 1

|ψ+ i does not give rise to inconsistencies [38]: If agent A hands over the fixed point state |ψ+ i to
the diagonalization operator D, the same state |ψ+ i is recovered. Stated differently, as long as the
output of the “halting algorithm” to input A(A) is |ψ+ i, i.e., HALT(A(A)) = |ψ+ i, diagonalization
does not change it. Hence, even if the (classically) “paradoxical” construction of diagonalization
is maintained, quantum theory does not give rise to a paradox, because the quantum range of
solutions is larger than the classical one. Therefore, standard proofs of the recursive unsolvability
of the halting problem do not apply if agent A is allowed a quantum bit. The consequences for
quantum recursion theory are discussed below.

CONSEQUENCES FOR QUANTUM RECURSION THEORY

Several critical remarks are in order. It should be noted that the fixed point quantum bit “solu-
tion” of the above halting problem is of not much practical help. In particular, if one is interested
in the “classical” answer whether or not A(A) halts, then one ultimately has to perform an irre-
versible measurement on the fixed point state. This causes a state reduction into the classical states

6
corresponding to |0i and |1i. Any single measurement will yield an indeterministic result. There
is a 50:50 chance that the fixed point state will be either in |0i or |1i, since as has been argued
before, |hψ+ |0i|2 = |hψ+ |1i|2 = 1/2. Thereby, classical undecidability is recovered.
Thus, as far as problem solving is concerned, classical bits are not much of an advance. If a
classical information is required, then quantum bits are not better than probabilistic knowledge.
With regards to the question of whether or not a computer halts, the “solution” is effectively equiv-
alent to the throwing of a fair coin [39]. Therefore, the advance of quantum recursion theory over
classical recursion theory is not so much classical problem solving but the consistent representa-
tion of statements which would give rise to classical paradoxes.
The above argument used the continuity of quantum bit states as compared to the two discrete
classical bit states for a construction of fixed points of the diagonalization operator. One could pro-
ceed a step further and allow nonclassical diagonalization procedures. Thereby, one could extend
diagonalization to the entire range of two-dimensional unitary transformations [40], which need
not have fixed points corresponding to eigenvalues of exactly one. Note that the general diagonal
form of finite-dimensional unitary transformations in matrix notation is diag(eiϕ1 , eiϕ2 , . . . , eiϕn );
i.e., the eigenvalues of a unitary operator are complex numbers of unit modulus (e.g., Ref. [41,
p. 39], or Ref. [42, p. 161]). Fixed points only occur if at least one of the phases ϕi , i ∈ {1, 2, . . . , n}
is a multiple of 2π. In what follows, we shall study the physical realizability of general unitary
operators associated with generalized beam splitters [43–46]. We will be particularly interested
in those transformations whose spectra do not contain the eigenvalue one and thus do not allow a
fixed point eigenvector.
In what follows, lossless devices will be considered. In order to be able to realize a universal
unitary transformation in two-dimensional Hilbert space, one needs to consider gates with two in-
put und two output ports representing beam splitters and Mach-Zehnder interferometers equipped
with an appropriate number of phase shifters. For the sake of demonstration, consider the two
realizations depicted in Fig. 1. The elementary quantum interference device Tbs in Fig. 1a) is a
unit consisting of two phase shifters P1 and P2 in the input ports, followed by a beam splitter S,
which is followed by a phase shifter P3 in one of the output ports. The device can be quantum

7
 Tbs (ω, α, β, ϕ)
0
|0i |0i-
- P1 , α + β P3 , ϕ

S(T )
P2 , β
0
|1i - |1i-

a)

TMZ (α, β, ω, ϕ)
0
|0i M |0i-
- P ,α+β P4 , ϕ
1
b
S1 P3 , ω S2
0
|1i - c |1i-
P2 , β
M

b)

FIG. 1: A universal quantum interference device operating on a qubit can be realized by a 4-port interferom-
eter with two input ports |0i, |1i and two output ports |0i0 , |1i0 ; a) realization by a single beam splitter S(T )
with variable transmission T and three phase shifters P1 , P2 , P3 ; b) realization by two 50:50 beam splitters
S1 and S2 and four phase shifters P1 , P2 , P3 , P4 .

mechanically represented by [47]

P1 : |0i → |0iei(α+β) ,
P2 : |1i → |1ieiβ ,
√ √
S : |0i → T |10 i + i R |00 i, (8)
√ √
S : |1i → T |00 i + i R |10 i,
P3 : |00 i → |00 ieiϕ ,

where every reflection by a beam splitter S contributes a phase π/2 and thus a factor of eiπ/2 = i to
the state evolution. Transmitted beams remain unchanged; i.e., there are no phase changes. Global
p p
phase shifts from mirror reflections are omitted. With T (ω) = cos ω and R(ω) = sin ω, the

8
corresponding unitary evolution matrix is given by
 
ie i (α+β+ϕ) sin ω ei (β+ϕ) cos ω
Tbs (ω, α, β, ϕ) =  . (9)
ei (α+β) cos ω i ei β sin ω

Alternatively, the action of a lossless beam splitter may be described by the matrix [54]
 p p   
i R(ω) T (ω) i sin ω cos ω
 p p = .
T (ω) i R(ω) cos ω i sin ω
¡ ¢
A phase shifter in two-dimensional Hilbert space is represented by either diag eiϕ , 1 or
¡ ¢
diag 1, eiϕ . The action of the entire device consisting of such elements is calculated by mul-
tiplying the matrices in reverse order in which the quanta pass these elements [48, 49]; i.e.,
    
e iϕ 0 i sin ω cos ω e i(α+β) 0 1 0
Tbs (ω, α, β, ϕ) =     . (10)
0 1 cos ω i sin ω 0 1 0 e iβ

The elementary quantum interference device TMZ depicted in Fig. 1b) is a Mach-Zehnder in-
terferometer with two input and output ports and four phase shifters. The process can be quantum
mechanically described by
P1 : |0i → |0iei(α+β) ,
P2 : |1i → |1ieiβ ,
√
S1 : |1i → (|bi + i |ci)/ 2,
√
S1 : |0i → (|ci + i |bi)/ 2,
(11)
P3 : |bi → |bieiω ,
√
S2 : |bi → (|10 i + i |00 i)/ 2,
√
S2 : |ci → (|00 i + i |10 i)/ 2,
P4 : |00 i → |00 ieiϕ .
The corresponding unitary evolution matrix is given by
 
i(α+ϕ) ω iϕ ω
ω −e sin 2 e cos 2
TMZ (α, β, ω, ϕ) = i ei(β+ 2 )  . (12)
eiα cos ω2 sin ω2

Alternatively, TMZ can be computed by matrix multiplication; i.e.,

TMZ (α, β, ω, ϕ) =
        
iϕ
e 0 i 1 iω
e 0 i 1 e i(α+β) 0 1 0
ω
i ei(β+ 2 )   √1 
2
  √1 
2
  .
0 1 1 i 0 1 1 i 0 1 0 eiβ
(13)

9
Both elementary quantum interference devices Tbs and TMZ are universal in the sense that every
unitary quantum evolution operator in two-dimensional Hilbert space
 
i α
e cos ω −e −i ϕ sin ω
U2 (ω, α, β, ϕ) = e−i β   , (14)
ei ϕ sin ω e−i α cos ω

where −π ≤ β, ω ≤ π, − π2 ≤ α, ϕ ≤ π
2 [40] corresponds to Tbs (ω0 , α0 , β0 , ϕ0 ) and
TMZ (ω00 , α00 , β00 , ϕ00 ), where ω, α, β, ϕ are arguments of the (double) primed parameters [46].
A typical example of a nonclassical operation on a quantum bit is the “square root of not” gate
√ √
( not not = X)  
√ 1 1+i 1−i 
not =  . (15)
2 1−i 1+i
√
Although not still has a eigenstate associated with a fixed point of unit eigenvalue, not all
of these unitary transformations have eigenvectors associated with eigenvalues one that can be
identified with fixed points. Indeed, only unitary transformations of the form

[U2 (ω, α, β, ϕ)]−1 diag(1, eiλ )U2 (ω, α, β, ϕ) =

 
iλ
cos ω2 + ei λ sin ω2 −1+e 2 e −i (α+ϕ) sin(2 ω) (16)
 
−1+ei λ i (α+ϕ) i λ cos ω2 + sin ω2
2 e sin(2 ω) e

have fixed points.

Applying nonclassical operations on quantum bits with no fixed points

D∗ = [U2 (ω, α, β, ϕ)]−1 diag(eiµ , eiλ )U2 (ω, α, β, ϕ) =

 −i (α+p)
³ ´ 
ei µ cos(ω)2 + ei λ sin(ω)2 e 2 ei λ − ei µ sin(2 ω) (17)
 ³ ´ ,
ei (α+p) i λ − ei µ sin(2 ω) i λ cos(ω)2 + ei µ sin(ω)2
2 e e

with µ, λ 6= 2nπ, n ∈ N0 gives rise to eigenvectors which are not fixed points, but which acquire
nonvanishing phases µ, λ in the generalized diagonalization process.

SUMMARY

It has been argued that, because of quantum parallelism, i.e., the effective co-representation of
classical mutually exclusive states, the diagonalization method of classical recursion theory has to
be modified. Quantum diagonalization involves unitary operators whose eigenvalues carry phases
strictly different from multiples of 2π. The quantum fixed point “solutions” of halting problems

10
can be 50:50 mixtures of the classical halting and nonhalting states, and therefore do not contribute
to classical deterministic solutions of the associated decision problems.
Another, less abstract, application for quantum information theory is the handling of inconsis-
tent information in databases. Thereby, two contradicting classical bits of information |0i and |1i
are resolved, i.e., co-represented, by the quantum bit |ψ+ i. Throughout the rest of the computa-
tion the coherence is maintained. After the processing, the result is obtained by an irreversible
measurement. The processing of quantum bits, however, would require an exponential space over-
head on classical computers in classical bit base [10]. Thus, in order to remain tractable, the
corresponding quantum bits should be implemented on truly quantum universal computers.

∗ Electronic address: [email protected]; URL: http://tph.tuwien.ac.at/

˜svozil
[1] H. Rogers, Jr., Theory of Recursive Functions and Effective Computability (MacGraw-Hill, New York,
1967).
[2] M. Davis, The Undecidable. Basic Papers on Undecidable, Unsolvable Problems and Computable
Functions (Raven Press, Hewlett, N.Y., 1965).
[3] J. Barwise, Handbook of Mathematical Logic (North-Holland, Amsterdam, 1978).
[4] H. Enderton, A Mathematical Introduction to Logic (Academic Press, San Diego, 2001), second edn.
[5] P. Odifreddi, Classical Recursion Theory, Vol. 1 (North-Holland, Amsterdam, 1989).
[6] G. S. Boolos, J. P. Burgess, and R. C. Jeffrey, Computability and Logic (Cambridge University Press,
Cambridge, 2007), fifth edn.
[7] R. Landauer, “Information is Physical,” Physics Today 44, 23–29 (1991).
http://dx.doi.org/10.1063/1.881299
[8] P. W. Bridgman, “A Physicist’s Second Reaction to Mengenlehre,” Scripta Mathematica 2, 101–117,
224–234 (1934), cf. R. Landauer [50].
[9] A. Olszewski, J. Woleński, and R. Janusz, Church’s Thesis After 70 Years (Ontos, Berlin, 2006).
[10] R. P. Feynman, “Simulating physics with computers,” International Journal of Theoretical Physics 21,
467–488 (1982).
[11] E. Fredkin and T. Toffoli, “Conservative Logic,” International Journal of Theoretical Physics 21, 219–
253 (1982), reprinted in [51, Part I, Chapter 3].

11
 http://dx.doi.org/10.1007/BF01857727
[12] H. S. Leff and A. F. Rex, Maxwell’s Demon (Princeton University Press, Princeton, 1990).
[13] R. P. Feynman, The Feynman lectures on computation (Addison-Wesley Publishing Company, Read-
ing, MA, 1996), edited by A.J.G. Hey and R. W. Allen.
[14] G. Birkhoff and J. von Neumann, “The Logic of Quantum Mechanics,” Annals of Mathematics 37,
823–843 (1936).
[15] S. Kochen and E. P. Specker, “Logical Structures arising in quantum theory,” in Symposium on the
Theory of Models, Proceedings of the 1963 International Symposium at Berkeley pp. 177–189 (1965),
reprinted in [52, pp. 209–221].
[16] S. Kochen and E. P. Specker, “The calculus of partial propositional functions,” in Proceedings of the
1964 International Congress for Logic, Methodology and Philosophy of Science, Jerusalem pp. 45–57
(1965), reprinted in [52, pp. 222–234].
[17] D. J. Foulis, C. Piron, and C. H. Randall, “Realism, operationalism, and quantum mechanics,” Foun-
dations of Physics 13, 813–841 (1983), invited papers dedicated to Günther Ludwig.
http://dx.doi.org/10.1007/BF01906271
[18] C. H. Randall and D. J. Foulis, “Properties and operational propositions in quantum mechanics,”
Foundations of Physics 13, 843–857 (1983), invited papers dedicated to Günther Ludwig.
http://dx.doi.org/10.1007/BF01906272
[19] O. Bournez and M. L. Campagnolo, “A Survey on Continuous Time Computations,” in New Computa-
tional Paradigms. Changing Conceptions of What is Computable, S. Cooper, B. Löwe, and A. Sorbi,
eds. (Springer Verlag, New York, 2008), pp. 383–423.
http://www.lix.polytechnique.fr/ bournez/pmwiki/uploads/Main/SurveyContinuousTime.pdf
[20] D. Deutsch, “Quantum theory, the Church-Turing principle and the universal quantum computer,”
Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences (1934-
1990) 400, 97–117 (1985).
http://dx.doi.org/10.1098/rspa.1985.0070
[21] J. Gruska, Quantum Computing (McGraw-Hill, London, 1999).
[22] M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information (Cambridge Uni-
versity Press, Cambridge, 2000).
[23] H.-K. Lo, S. Popescu, and T. Spiller, Introduction to Quantum Computation and Information (World
Scientific Publishing Company, Singapore, 2001).

12
[24] R. K. Brylinski, G. Chen, and B. K. Brylinski, Mathematics of Quantum Computation (Chapman &
Hall/CRC Press, London, 2002).
[25] M. Hayashi, Quantum Information. An Introduction (Springer-Verlag, Berlin, Heidelberg, 2006).
[26] H. Imai and M. Hayashi, Quantum Computation and Information. From Theory to Experiment
(Springer-Verlag, Berlin, Heidelberg, 2006).
[27] V. Scarani, H. Bechmann-Pasquinucci, N. J. Cerf, M. Dusek, N. Lütkenhaus, and M. Peev, “The
Security of Practical Quantum Key Distribution,” (2008).
http://arxiv.org/abs/0802.4155
[28] G. Jaeger, Quantum Information. An Overview (Springer, New York, 2007).
[29] N. D. Mermin, Quantum Computer Science (Cambridge University Press, Cambridge, 2007).
http://people.ccmr.cornell.edu/ mermin/qcomp/CS483.html
[30] N. D. Mermin, “From Cbits to Qbits: Teaching computer scientists quantum mechanics,” American
Journal of Physics 71, 23–30 (2003).
http://dx.doi.org/10.1119/1.1522741
[31] K. Svozil, “Contexts in quantum, classical and partition logic,” in Handbook of Quantum Logic and
Quantum Structures, K. Engesser, D. M. Gabbay, and D. Lehmann, eds. (Elsevier, Amsterdam, 2008),
pp. 551–586.
http://arxiv.org/abs/quant-ph/0609209
[32] M. Born, “Zur Quantenmechanik der Stoßvorgänge,” Zeitschrift für Physik 37, 863–867 (1926).
http://dx.doi.org/10.1007/BF01397477
[33] A. Zeilinger, “The message of the quantum,” Nature 438, 743 (2005).
http://dx.doi.org/10.1038/438743a
[34] K. Gödel, “Über formal unentscheidbare Sätze der Principia Mathematica und verwandter Systeme,”
Monatshefte für Mathematik und Physik 38, 173–198 (1931), English translation in [53], and in [2].
[35] A. R. Anderson, “St. Paul’s epistle to Titus,” in The Paradox of the Liar, R. L. Martin, ed. (Yale
University Press, New Haven, 1970), the Bible contains a passage which refers to Epimenides, a Crete
living in the capital city of Cnossus: “One of themselves, a prophet of their own, said, ‘Cretans are
always liars, evil beasts, lazy gluttons.’ ”,— St. Paul, Epistle to Titus I (12-13).
[36] G. Cantor, “Gesammelte Abhandlungen,” (Springer, Berlin, 1932).
[37] A. M. Turing, “On computable numbers, with an application to the Entscheidungsproblem,” Proceed-
ings of the London Mathematical Society, Series 2 42 and 43, 230–265 and 544–546 (1936-7 and

13
 1937), reprinted in [2].
[38] K. Svozil, “Consistent use of paradoxes in deriving contraints on the dynamics of physical systems
and of no-go-theorems,” Foundations of Physics Letters 8, 523–535 (1995).
[39] P. Diaconis, S. Holmes, and R. Montgomery, “Dynamical Bias in the Coin Toss,” SIAM Review 49,
211–235 (2007).
http://dx.doi.org/10.1137/S0036144504446436
[40] F. D. Murnaghan, The Unitary and Rotation Groups (Spartan Books, Washington, D.C., 1962).
[41] R. Shankar, Principles of Quantum Mechanics, 2nd Edition (Kluwer Academic/Plenum Publishers,
New York, 1980,1994).
[42] P. R. Halmos, Finite-dimensional vector spaces (Springer, New York, Heidelberg, Berlin, 1974).
[43] M. Reck, A. Zeilinger, H. J. Bernstein, and P. Bertani, “Experimental realization of any discrete unitary
operator,” Physical Review Letters 73, 58–61 (1994).
http://dx.doi.org/10.1103/PhysRevLett.73.58
[44] M. Reck and A. Zeilinger, “Quantum phase tracing of correlated photons in optical multiports,” in
Quantum Interferometry, F. D. Martini, G. Denardo, and A. Zeilinger, eds., pp. 170–177 (1994).
[45] M. Zukowski, A. Zeilinger, and M. A. Horne, “Realizable higher-dimensional two-particle entangle-
ments via multiport beam splitters,” Physical Review A (Atomic, Molecular, and Optical Physics) 55,
2564–2579 (1997).
http://dx.doi.org/10.1103/PhysRevA.55.2564
[46] K. Svozil, “Noncontextuality in multipartite entanglement,” J. Phys. A: Math. Gen. 38, 5781–5798
(2005).
http://dx.doi.org/10.1088/0305-4470/38/25/013
[47] D. M. Greenberger, M. A. Horne, and A. Zeilinger, “Multiparticle interferometry and the superposition
principle,” Physics Today 46, 22–29 (1993).
[48] B. Yurke, S. L. McCall, and J. R. Klauder, “SU(2) and SU(1,1) interferometers,” Physical Review A
(Atomic, Molecular, and Optical Physics) 33, 4033–4054 (1986).
http://dx.doi.org/10.1103/PhysRevA.33.4033
[49] R. A. Campos, B. E. A. Saleh, and M. C. Teich, “Fourth-order interference of joint single-photon wave
packets in lossless optical systems,” Physical Review A (Atomic, Molecular, and Optical Physics) 42,
4127–4137 (1990).
http://dx.doi.org/10.1103/PhysRevA.42.4127

14
[50] R. Landauer, “Advertisement For a Paper I Like,” in On Limits, J. L. Casti and J. F. Traub, eds. (Santa
Fe Institute Report 94-10-056, Santa Fe, NM, 1994), p. 39.
http://www.santafe.edu/research/publications/workingpapers/94-10-056.pdf
[51] A. Adamatzky, Collision-based computing (Springer, London, 2002).
[52] E. Specker, Selecta (Birkhäuser Verlag, Basel, 1990).
[53] K. Gödel, in Collected Works. Publications 1929-1936. Volume I, S. Feferman, J. W. Dawson, S. C.
Kleene, G. H. Moore, R. M. Solovay, and J. van Heijenoort, eds. (Oxford University Press, Oxford,
1986).
[54] The standard labeling of the input and output ports are interchanged, therefore sine and cosine are
exchanged in the transition matrix.

15