A Quantum-theoretic Framework for Concept Theory and Natural Language Processing

The construction of a consistent theory for structuring and representing how concepts combine and interact is one of the main challenges for the scholars involved in cognitive studies. However, all traditional approaches meet serious hindrances when dealing with combinations of concepts and conceptual vagueness (Guppy effect, overextension and underextension of membership weights, borderline contradictions). One of the main consequences of these difficulties is the impossibility of modeling concepts and their combinations in a classical (fuzzy set) logic and probability theory framework. We present here a theoretic framework for conceptual combinations which employs the mathematical formalism of quantum theory in Fock space. Our model faithfully describes a large amount of experimental data collected by different scholars on conjunctions and disjunctions of two concepts. Our approach puts forward a completely novel possible solution to the `combination problem' in concept theory, thus shedding a new light on the above mentioned difficulties. Additionally, we introduce an 
explanation for the occurrence of quantum structures in the mechanisms and dynamics of concepts and, more generally, in cognitive and decision processes, according to which human thought is a specifically structured superposition of a `logical thought' and a `conceptual thought', and the latter usually prevails over the former, at variance with some widespread beliefs. Inspired by this quantum cognition research, we illustrate the basics of a `quantum meaning based’ framework for information retrieval (IR) and natural language processing (NLP). The potential advantages of this approach with respect to traditional approaches, such as latent semantic analysis (LSA), are briefly discussed.