A+ CATEGORY SCIENTIFIC UNIT

Operator systems, contextuality and non-locality

Michalis Anoussis, Alexandros Chatzinikolaou, Ivan G. Todorov Dissertationes Mathematicae (2026) MSC: Primary 47L25; Secondary 81R15, 46L07. DOI: 10.4064/dm240830-4-11 Published online: 28 October 2026

Abstract

We introduce an operator system, universal for the probabilistic models of a contextuality scenario, and identify its maximal $C^*$-cover via the right $C^*$-algebra of a canonical ternary ring of operators, arising from a hypergraph version of stochastic operator matrices. We study dilating contextuality scenarios, which have the property that each positive operator representation thereof admits a dilation to a projective representation on a larger Hilbert space, and characterise them via the equality of the aforementioned universal operator system and the operator system arising from the canonical generators of the respective hypergraph $C^*$-algebra. We characterise the no-signalling probabilistic models over a pair of contextuality scenarios of different types, which arise from either the positive operator representations or from the projective representations of these scenarios, in terms of states on operator system tensor products. Generalising the notion of a synchronous no-signalling correlation to the hypergraph framework, we define coherent probabilistic models associated with a given contextuality scenario and characterise various classes thereof via different types of traces of the hypergraph $C^*$-algebra, associated with the scenario. We establish several equivalent formulations of the Connes Embedding Problem in terms of no-signalling probabilistic models and hypergraph operator systems.

Authors

  • Michalis AnoussisDepartment of Mathematics
    University of the Aegean
    Karlovassi, Samos 83200, Greece
    e-mail
  • Alexandros ChatzinikolaouDepartment of Mathematics
    National and Kapodistrian University of Athens
    Athens 157 84, Greece
    e-mail
  • Ivan G. TodorovSchool of Mathematical Sciences
    University of Delaware
    Newark, DE 19716, USA
    e-mail

Search for IMPAN publications

Query phrase too short. Type at least 4 characters.

Rewrite code from the image

Reload image

Reload image