A noncommutative view on topology and order (Q1029479)

Connes' noncommutative geometry is a promising candidate for the merging of general relativity and the standard model of particle theory. However, it faces two major challenges. The first is quantization. The second is compatibility with Lorentz signature. One way to deal with the second problem would be to perform back and forth Wick rotations. Although this has been proved to be a successful procedure in Minkowski space, it is known to be unreliable in some background independent contexts [\textit{J. Ambjorn, J. L. Nielsen, J. Rolf} and \textit{R. Loll}, Chaos Solitons Fractals 10, No.~2--3, 177--195 (1999; Zbl 1005.83015)]. A more satisfying approach that would be an adaptation of the axiomatic description of spectral triples is needed. Several attempts have already been made. A first option is to define semi-Riemannian spectral triples, in which one replaces \(C^*\)-algebras by \(*\)-algebras of bounded operators on a Klein space [\textit{A. Strohmaier}, J. Geom. Phys. 56, No. 2, 175--195 (2006; Zbl 1086.58004)]. A less radical possibility is to use the splitting of a globally hyperbolic manifold into space-time. One can then define a collection of noncommutative 3-spaces with causal relationships between them. It is possible to arrive at an axiomatic characterization of the collection of local algebras of causal functions on a globally hyperbolic spacetime [\textit{V. Moretti}, Rev. Math. Phys. 15, No. 10, 1171--1217 (2003; Zbl 1055.46048)]. The main motivation is to look for a correct setting for Lorentzian noncommutative geometry. The author focuses only on one aspect of the problem, namely, causal order. It is known that the conformal structure of a spacetime can be recovered from its causal order relation alone [\textit{D. Malament}, J. Math. Phys. 18, 1399--1404 (1977; Zbl 0355.53033)]. For this reason, it has been advocated that causal order should be the central object of quantum gravity and that spacetime should be replaced at microscopic scales by what is called a causal set. In the spirit of noncommutative geometry, a dual formulation of a causal set in terms of functions preserving the causal order is needed. The author obtains such a formulation. The kind of space we seek to render noncommutative is a partially ordered set, or poset, endowed with a topology. The two structures have to satisfy a compatibility condition. Several conditions of this kind have been studied. In order to obtain good duality results, the minimal condition to ask is that there are enough continuous non-decreasing functions to fully determine the order relation. Spaces of this kind are called completely separated ordered spaces or toposets for short. The author defines the category of \(I^*\)-algebras. These are couples \((I,C)\), where \(C\) is a \(C^*\)-algebra and \(I\) is a closed convex cone in \(C\), subject to several additional hypotheses inferred from the properties of the cone of non-decreasing functions defined on a toposet. The author proves the main result, that is the extension of the Gelfand-Naimark theorem to the category of unital commutative \(I^*\)-algebras. Unitarity of the algebra means compactness of the underlying space. The formulation of the definition of noncommutative causal toposets is given and the simplest noncommutative \(I^*\)-algebras are classified.