## A Topos for Algebraic Quantum Theory (2009)

Venue: | COMMUNICATIONS IN MATHEMATICAL PHYSICS |

Citations: | 8 - 1 self |

### BibTeX

@MISC{Heunen09atopos,

author = {Chris Heunen and Nicolaas P. Landsman and Bas Spitters},

title = { A Topos for Algebraic Quantum Theory},

year = {2009}

}

### OpenURL

### Abstract

The aim of this paper is to relate algebraic quantum mechanics to topos theory, so as to construct new foundations for quantum logic and quantum spaces. Motivated by Bohr’s idea that the empirical content of quantum physics is accessible only through classical physics, we show how a noncommutative C*-algebra of observables A induces a topos T (A) in which the amalgamation of all of its commutative subalgebras comprises a single commutative C*-algebra A. According to the constructive Gelfand duality theorem of Banaschewski and Mulvey, the latter has an internal spectrum �(A) in T (A), which in our approach plays the role of the quantum phase space of the system. Thus we associate a locale (which is the topos-theoretical notion of a space and which intrinsically carries the intuitionistic logical structure of a Heyting algebra) to a C*-algebra (which is the noncommutative notion of a space). In this setting, states on A become probability measures (more precisely, valuations) on �, and self-adjoint elements of A define continuous functions (more precisely, locale maps) from � to Scott’s interval domain. Noting that open subsets of �(A) correspond to propositions about the system, the pairing map that assigns a (generalized) truth value to a state and a proposition assumes an extremely simple categorical form. Formulated in this way, the quantum theory defined by A is essentially turned into a classical theory, internal to the topos T (A). These results were inspired by the topos-theoretic approach to quantum physics proposed by Butterfield and Isham, as recently generalized by Döring and Isham.

### Citations

2100 | Noncommutative geometry
- Connes
- 1994
(Show Context)
Citation Context ...-algebra of its bounded automorphisms [47]. A little more abstractly, a quantum probability space is thus a pair (N, ρ) of a (not necessarily commutative) unital von Neumann algebra and a state on it =-=[17]-=-. Even more abstractly, and along the lines of [51], one takes a C*-algebra with a state on it and if needed constructs the von Neumann algebra by the GNS construction. Hence Theorem 15 shows that the... |

456 | Domain Theory - Abramsky, Jung - 1994 |

265 |
Neumann The Logic of Quantum Mechanics
- Birkhoff, Von
- 1936
(Show Context)
Citation Context ...airing map takes values in [0,1] and is given by the “Born rule” 〈Ψ,a ∈ ∆〉 = (Ψ,E(∆)Ψ). Thus subsets of phase space became closed linear subspaces of Hilbert space, which, as Birkhoff and von Neumann =-=[10]-=- noticed, form a lattice L(H) under inclusion as partial order. However, this lattice fails to be Boolean, basically because it is nondistributive. Nonetheless, Birkhoff and von Neumann interpreted th... |

220 |
The problem of hidden variables in quantum mechanics
- Kochen, Specker
- 1967
(Show Context)
Citation Context ... map e.g. Σ(D (x,y)) → Σ(C (x,y)) then simply is the evident inclusion of lattices in the other direction. 14Kochen-Specker In general, the locale Σ is highly non-spatial: the Kochen-Specker theorem =-=[39]-=- shows that the functor Σ has no global sections [11, 12, 13, 14], or, in other words, that it has no points as a locale in T (A). The following theorem makes this statement precise. Recall that there... |

174 | An Introduction to Quantum Stochastic Calculus. Monographs in Mathematics Vol. 85. Birkhäuser
- Parthasarathy
- 1992
(Show Context)
Citation Context ...veloped as a noncommutative analogue. A quantum probability space is a triple (H, P, ρ) of a Hilbert space H, its set of projections, and a state on the unital C*-algebra of its bounded automorphisms =-=[47]-=-. A little more abstractly, a quantum probability space is thus a pair (N, ρ) of a (not necessarily commutative) unital von Neumann algebra and a state on it [17]. Even more abstractly, and along the ... |

149 |
Sketches of an elephant: a topos theory compendium
- Johnstone
- 2002
(Show Context)
Citation Context ...ieck as a generalisation of sheaves in algebraic geometry, Lawvere independently researched logical aspects, and with Tierney wove both strands together to arrive at the notion of an elementary topos =-=[6, 45, 38, 43]-=-. Thus every elementary topos comes equipped with an internal logic (the Mitchell-Bénabou language with its Kripke-Joyal semantics). A topos can hence be seen as a universe of discourse, to which the ... |

149 | A categorical semantics of quantum protocols
- Abramsky
- 2004
(Show Context)
Citation Context ...esting another level P-Set. There are thus various stacks of nested (internal) toposes to be studied. • Recently, dagger compact categories have been proposed as a type theory for quantum computation =-=[2]-=-. Via the Doplicher-Roberts duality [24], such a category C gives rise to a compact group G under some natural requirements. In turn, every compact group G yields a C*-algebra C ∗ (G), its 27so-calle... |

146 |
The Mathematical Foundations of Quantum Mechanics
- Mackey
- 1963
(Show Context)
Citation Context ...t is called positive when ρ(a ∗ a) ≥ 0 for all a ∈ A. When A is unital, a positive quasi-linear functional is called a quasi-state when ρ(1) = 1. Axiom VII of Mackey’s foundation of quantum mechanics =-=[44]-=- states that a measure on the projections of a von Neumann algebra extends to a state on the von Neumann algebra. Mackey stresses that, in contrast to his other axioms, Axiom VII does not have a physi... |

139 |
Mathematical Topics Between Classical and Quantum Mechanics
- Landsman
- 1998
(Show Context)
Citation Context ..., which we summarize in this subsection. This construction associates a certain internal locale to a noncommutative C*-algebra (assumed unital), and hinges on three ideas: 1. Algebraic quantum theory =-=[34, 39, 50]-=-; 2. Constructive Gelfand duality [7, 8, 9, 23, 24]; 3. Bohr’s doctrine of classical concepts [13, 62, 51]. From the first, we just adopt the methodology of describing a quantum system by a noncommuta... |

112 | An extension of the Galois theory of Grothendieck - Joyal, Tierney - 1983 |

104 |
Roberts: A new duality theory for compact groups
- Doplicher, E
- 1989
(Show Context)
Citation Context ...hus various stacks of nested (internal) toposes to be studied. • Recently, dagger compact categories have been proposed as a type theory for quantum computation [2]. Via the Doplicher-Roberts duality =-=[24]-=-, such a category C gives rise to a compact group G under some natural requirements. In turn, every compact group G yields a C*-algebra C ∗ (G), its 27so-called group C*-algebra. It seems likely that... |

100 | Topology via Logic - Vickers - 1989 |

90 |
Geometry of Quantum Theory
- Varadarajan
- 1968
(Show Context)
Citation Context ...cal connectives are given by the standard set-theoretic operations. This makes the logic of the system ‘spatial’; its realization as a Boolean algebra confirms the ‘classical’ nature of the situation =-=[66]-=-. 2 Physicists do not usually describe a system in the above way. Instead, they work with observables a : M → R, like position or energy. (Such functions might be arbitrary, or else required to be mea... |

82 |
Interpreting the Quantum World
- Bub
- 1999
(Show Context)
Citation Context ...uantum ‘logic’ is a propositional language; no satisfactory generalization to predicate logic has been found. Quantum logic is also problematic from a physical perspective. Since (by various theorems =-=[14]-=- and wide agreement) quantum probabilities do not admit an ignorance interpretation, [0,1]-valued truth values attributed to propositions by pure states via the Born rule cannot be regarded as sharp (... |

80 |
Discussion with Einstein on epistemological problems in atomic physics
- Bohr
(Show Context)
Citation Context ...ra (assumed unital), and hinges on three ideas:1 INTRODUCTION 10 1. Algebraic quantum theory [38, 43, 58]; 2. Constructive Gelfand duality [5, 6, 7, 26, 28]; 3. Bohr’s doctrine of classical concepts =-=[11, 73, 60]-=-. From the first, we just adopt the methodology of describing a quantum system by a noncommutative C*-algebra A (defined in the usual topos Sets). This move generalizes the usual Hilbert space framewo... |

71 | A topos perspective on the KochenSpecker theorem: III. Von Neumann algebras as the base category
- Hamilton, Isham, et al.
- 1999
(Show Context)
Citation Context ...hr’s doctrine mathematically precise. We will naturally arrive at the notion of a quasi-state that causes the Mackey-Gleason problem [8], and contribute to the programme of Isham and co-workers (like =-=[35, 11, 12, 13, 14, 26, 27, 28, 29]-=-) to address the related interpretational problems posed by the Kochen-Specker theorem. Toposes The technical tool enabling the construction in this article is a topos. Originally invented by Grothend... |

56 | Intuitionistic formal spaces - a first communication - Sambin - 1987 |

56 |
Relativistic invariance and charge conjugation in quantum field theory
- Guido, Longo
- 1992
(Show Context)
Citation Context ...resenting an AQFT as a covariant functor of C*-algebras, even in the more complicated case of general relativity, has been emphasised in [7]. (It has also been tried to extend such results to sheaves =-=[32, 49]-=-.) One could thus consider another intermediate level between the ambient topos and T (A). This is one way of making the similarity between quantum theory and relativity theory in [23] more precise; s... |

48 | Marcolli M., Noncommutative geometry, quantum fields and motives
- Connes
- 2008
(Show Context)
Citation Context ...space’, will be met by a combination of two profound notions of generalized space that have been around for some time.1 INTRODUCTION 6 1. First, let us recall the strategy of noncommutative geometry =-=[24, 25]-=-. One starts with the replacement of a compact topological space X by the associated algebra of complex-valued continuous functions C(X, C). If X fails to be Hausdorff, this step loses information, bu... |

42 | Operational quantum physics - Busch, Grabowski, et al. - 1997 |

35 | Geometry of Quantum Theory, second edition - Varadarajan - 1985 |

31 | Inductively generated formal topologies
- Coquand, Sambin, et al.
- 2003
(Show Context)
Citation Context ... ∧Dq−a = ⊥, and hence Da−q = Da−q ∧ ⊤ = Da−q ∧ (Db ∨Dq−a) = Da−q ∧Db � Db = ∨ U0. 31 Alternatively, writing Da ⊳0 U iff U ⊇ Da, the covering relation ⊳ is inductively generated by ⊳0, as explained in =-=[27, 79]-=-. The triple (LA, �,⊳0) is a flat site as defined in [79]. ։ □A GENERATING LATTICES FOR FRAMES 43 Thus we have two alternative expressions for the spectrum: O(Σ) ∼ = {U ∈ Idl(LA) | ∀q>0Da−q ∈ U ⇒ Da ... |

30 |
Topos theory and consistent histories: The internal logic of the set of all consistent sets
- Isham
- 1997
(Show Context)
Citation Context ...hr’s doctrine mathematically precise. We will naturally arrive at the notion of a quasi-state that causes the Mackey-Gleason problem [8], and contribute to the programme of Isham and co-workers (like =-=[35, 11, 12, 13, 14, 26, 27, 28, 29]-=-) to address the related interpretational problems posed by the Kochen-Specker theorem. Toposes The technical tool enabling the construction in this article is a topos. Originally invented by Grothend... |

29 |
Lattice theory, data types and semantics
- Scott
- 1972
(Show Context)
Citation Context ... π1 ◦ pt = id are just (unconstrained) locale maps X → [T]. 1.6 Observation and approximation Our construction of the locale map δ(a) : Σ → IR in Section 1.4 involves the so-called interval domain IR =-=[63]-=-. To motivate its definition, consider the approximation of real numbers by nested intervals with endpoints in Q. For example, the real number π can be described by specifying the sequence [3, 4], [3.... |

27 |
Formal spaces
- Fourman, Grayson
- 1982
(Show Context)
Citation Context ... (28) (r, ∞) ↦→ Da−r. (29) As â −1 is a frame map, for bounded open intervals (r,s) we therefore obtain 5 â −1 : (r,s) ↦→ Ds−a ∧Da−r. (30) We now recall an explicit construction of the Dedekind reals =-=[39]-=-, [51, D4.7.4 & D4.7.5]. Define the propositional geometric theory TR generated by formal symbols (p,q) ∈ Q × Q with p < q, ordered as (p,q) � (p ′ ,q ′ ) iff p ′ � p and q � q ′ , subject to the foll... |

25 | Some points in formal topology
- Sambin
(Show Context)
Citation Context ...V (where U∧V = {u∧v | u ∈ U, v ∈ V }). For example, if (L, ≤) = (O(X), ⊆) one may take x ⊳ U iff x = ∨ U, i.e. if U covers x. Alternatively, one may define x ⊳f U iff U is a finite cover of x. 70 See =-=[61]-=- for an overview, and [58] for the connection with domain theory. Ref. [3] shows that formal topology is valid in constructive set theory, and hence in particular in topos theory. 71 A coverage on a m... |

23 | Topoi: The Categorical Analysis of Logic - Goldblatt - 1984 |

22 | About Stone’s notion of spectrum
- Coquand
- 2005
(Show Context)
Citation Context ...ation theorem, which characterises them completely as algebras of complex-valued continuous functions on a compact Hausdorff space. It can be shown that the essence of this theorem holds in any topos =-=[4, 19]-=-, a fact we recollect in Theorem 4 below. The paraphrase leading up to it is somewhat verbose, because the specifics are of importance in Section 4. It follows a presentation that avoids the Banach-Ma... |

22 |
The Logical Analysis of Quantum Mechanics
- Scheibe
- 1974
(Show Context)
Citation Context ...mmutative C*-algebra (assumed unital), and hinges on three ideas: 1. Algebraic quantum theory [34, 39, 50]; 2. Constructive Gelfand duality [7, 8, 9, 23, 24]; 3. Bohr’s doctrine of classical concepts =-=[13, 62, 51]-=-. From the first, we just adopt the methodology of describing a quantum system by a noncommutative C*-algebra A (defined in the usual topos Sets). As to the second, it turns out that the notion of a C... |

22 | D.: Quantum measurements without sums
- Coecke, Pavlović
- 2007
(Show Context)
Citation Context ... turn, every compact group G yields a C*-algebra C ∗ (G), its 27so-called group C*-algebra. It seems likely that there is a relationship between classical types (i.e. the so-called classical objects =-=[16]-=-) in C, and the finite-dimensional ‘classical snapshots of reality’ in C(C ∗ (G)), the closure of which is the entire C(C ∗ (G)). • Even if p, q ∈ Asa do not commute, Proposition 18 still allows one t... |

20 |
Mathematical and conceptual foundations of 20th-century physics
- Emch
- 1984
(Show Context)
Citation Context ...ze in this subsection. This construction associates a certain internal locale to a noncommutative C*-algebra (assumed unital), and hinges on three ideas:1 INTRODUCTION 10 1. Algebraic quantum theory =-=[38, 43, 58]-=-; 2. Constructive Gelfand duality [5, 6, 7, 26, 28]; 3. Bohr’s doctrine of classical concepts [11, 73, 60]. From the first, we just adopt the methodology of describing a quantum system by a noncommuta... |

20 |
Toposes and local set theories: An introduction
- Bell
- 1988
(Show Context)
Citation Context ...ing more observables, and hence more information. Let us comment briefly on the internal logic of a topos in the context of our running example. For more information, we must refer the reader to e.g. =-=[43, 38, 5]-=-. The internal logic of a topos assigns a truth value to every proposition of higher-order, many-sorted logic. However, unlike in the topos Set, these truth values can be more general than the Boolean... |

19 | Open locales and exponentiation - Johnstone - 1984 |

18 |
Aspects of general topology in constructive set theory
- Aczel
(Show Context)
Citation Context ...tivate its definition, consider the approximation of real1 INTRODUCTION 16 numbers by nested intervals with endpoints in Q. For example, the real number π can be described by specifying the sequence =-=[3,4]-=-,[3.1,3.2],[3.14,3.15], [3.141,3.142],... Each individual interval may be interpreted as finitary information about the real number under scrutiny, involving the single observation that the real numbe... |

18 | Entailment relations and distributive lattices
- Cederquist, Coquand
- 2000
(Show Context)
Citation Context ...al spaces ‘without points’. A compact regular locale can be conveniently represented by a normal distributive lattice [18], and in turn such lattices can be represented by normal entailment relations =-=[15]-=-. In topological (or better localic) terms, a lattice presents a basis while an entailment presents a subbasis. In logical terms, these entailment relations provide a propositional geometric theory an... |

18 |
Riesz spaces
- Zaanen
- 1983
(Show Context)
Citation Context ... the integral can be extended to this new space. This ‘embedding’ identifies x, y when |x − y| ≤ 1 n for all n ∈ N. It is thus an embedding when the ordered vector space is an Archimedean Riesz space =-=[42, 56]-=-. This, however, can not be defined geometrically. The situation is thus similar to that of seminorms on pre-C*algebras. In a sense, C(X) with a positive linear functional is also the only example. Al... |

17 |
A globalisation of the Gelfand duality theorem
- Banaschewski, Mulvey
- 2006
(Show Context)
Citation Context ...ation theorem, which characterises them completely as algebras of complex-valued continuous functions on a compact Hausdorff space. It can be shown that the essence of this theorem holds in any topos =-=[4, 19]-=-, a fact we recollect in Theorem 4 below. The paraphrase leading up to it is somewhat verbose, because the specifics are of importance in Section 4. It follows a presentation that avoids the Banach-Ma... |

17 | Kochen-Specker theorem for von Neumann algebras
- Döring
- 2005
(Show Context)
Citation Context ...ded operators on H. Then the locale Σ in T (A) has no points. The Kochen-Specker theorem holds for a more general C*-algebras then just the collection of all bounded operators on a Hilbert space, see =-=[25]-=- for results on von Neumann algebras. For C*-algebras one has [34]: a simple infinite unital C*-algebra does not admit a dispersion-free quasi-state. The previous theorem also holds for such extension... |

17 |
The Caterorial Analysis of Logic
- Topoi
- 1984
(Show Context)
Citation Context ...spect of topos theory is that it unifies two seemingly wholly distinct mathematical subjects: on the one hand, topology and algebraic geometry and on the other hand, logic and set theory. We refer to =-=[41, 63, 50, 51]-=- for accounts of topos theory; see also [9, 65, 57] for historical details. Briefly, a topos is a category in which one can essentially reason as in the category Sets of all sets (with functions as ar... |

16 |
Handbook of Categorical Algebra 3: Categories of Sheaves. Encyclopedia of Mathematics and its Applications 52
- Borceux
- 1994
(Show Context)
Citation Context ... in the category of sets (as defined above) to frames in more general topoi. This is indeed possible, as all of the above concepts can be defined in any topos by using its internal language [63]; see =-=[12]-=- for details. In particular, in a topos T one may consider the category FrmT of internal frames and its opposite category LocT of internal locales. The terminal object of the latter is the locale ∗ wh... |

16 |
Metric spaces, generalized logic, and closed
- Lawvere
- 1974
(Show Context)
Citation Context ...nto a norm, and then complete this normed space. Both of these constructions are, however, not geometric and thus better avoided. Fortunately, there is a satisfactory solution: the localic completion =-=[41, 54]-=-. The localic completion assigns to every generalised metric space X a locale, the points of which are those of the completion of X. A C*-algebra is the localic completion of a pre-C*-algebra. Every C... |

16 |
Generalized real numbers in constructive mathematics
- Richman
- 1998
(Show Context)
Citation Context ...A). To obtain a correspondence as in the previous lemma, we then consider cuts of Q instead of Dedekind cuts. This so-called interval domain [50] is often seen as a generalisation of the real numbers =-=[48]-=-. It is generated by rational intervals, ordered by reverse inclusion: a smaller interval means that we have more information (about the number that the ever smaller intervals converge to). However, b... |

15 |
The generally covariant locality principle—a new paradigm for local quantum field theory
- Brunetti, Fredenhagen, et al.
- 2003
(Show Context)
Citation Context ... topos Set (O(M),⊆) of Setvalued functors on Minkowski space. Presenting an AQFT as a covariant functor of C*-algebras, even in the more complicated case of general relativity, has been emphasised in =-=[7]-=-. (It has also been tried to extend such results to sheaves [32, 49].) One could thus consider another intermediate level between the ambient topos and T (A). This is one way of making the similarity ... |

14 |
The spectral theory of commutative C*-algebras: the constructive spectrum
- Banaschewski, Mulvey
- 2000
(Show Context)
Citation Context ...iates a certain internal locale to a noncommutative C*-algebra (assumed unital), and hinges on three ideas:1 INTRODUCTION 10 1. Algebraic quantum theory [38, 43, 58]; 2. Constructive Gelfand duality =-=[5, 6, 7, 26, 28]-=-; 3. Bohr’s doctrine of classical concepts [11, 73, 60]. From the first, we just adopt the methodology of describing a quantum system by a noncommutative C*-algebra A (defined in the usual topos Sets)... |

14 | The probabilistic powerdomain for stably compact spaces
- Alvarez-Manilla, Jung, et al.
(Show Context)
Citation Context ...lized Riesz-Markov Theorem, then, is as follows. 55 54 Equivalently, µ satisfies additivity as well as µ(⊥) = 0, and is Scott continuous [67]. Probability valuations extend uniquely to Borel measures =-=[5]-=-. 55 Despite this theorem, from a constructive point of view there is a crucial difference between valuations and integrals. The integral I(f) of a function f ∈ C(X) is a Dedekind real, so that it can... |

13 | Formal topology and constructive mathematics: the Gelfand and Stone-Yosida representation theorems - Coquand, Spitters - 2005 |

13 |
Continuous domains as formal spaces
- Negri
(Show Context)
Citation Context ...U, v ∈ V }). For example, if (L, ≤) = (O(X), ⊆) one may take x ⊳ U iff x = ∨ U, i.e. if U covers x. Alternatively, one may define x ⊳f U iff U is a finite cover of x. 70 See [61] for an overview, and =-=[58]-=- for the connection with domain theory. Ref. [3] shows that formal topology is valid in constructive set theory, and hence in particular in topos theory. 71 A coverage on a meet semilattice L is a fun... |

12 | Between classical and quantum - Landsman - 2007 |

11 |
Pretopologies and uniform presentation of sup-lattices, quantales and frames
- Battilotti, Sambin
(Show Context)
Citation Context ...ology is a generalization of the above triples (L, �,⊳), where � is merely required to be a preorder. In this more general case, the axioms on the cover relation ⊳ take a slightly different form. See =-=[8, 69]-=-.A GENERATING LATTICES FOR FRAMES 40 A proof of this theorem by explicit computation may be found in [22, Thm. 5.2.3]. Here, we give an alternative proof, which requires some familiarity with geometr... |

11 | A topos foundation for theories of physics: II. Daseinisation and the liberation of quantum theory
- Döring, Isham
- 2008
(Show Context)
Citation Context ...hr’s doctrine mathematically precise. We will naturally arrive at the notion of a quasi-state that causes the Mackey-Gleason problem [8], and contribute to the programme of Isham and co-workers (like =-=[35, 11, 12, 13, 14, 26, 27, 28, 29]-=-) to address the related interpretational problems posed by the Kochen-Specker theorem. Toposes The technical tool enabling the construction in this article is a topos. Originally invented by Grothend... |