SEMSUBSET · The Grammar of Inclusion: Exploring the Boundaries of Linguistic Competence

The Grammar of Inclusion: Exploring the Boundaries of Linguistic Competence

The combination of phrase-structure syntax and model-theoretic semantics have led to a powerful, general model of language. However, previous research has not sufficiently explained how syntactic structure and structural differences across languages relate to mathematical operations on meaning: some constraints observed in the language implementation of these operations are a consequence of the language form, while others could come from non linguistic cognition. This project investigates this domain from the perspective of natural language partitives and pseudo-partitives and subset and quantity cognition. Nominal partitives as 'two of the beans' explicitly express the subset relation in English. Other Romance and Germanic languages have similar, but not identical forms: e.g. in German genitive case can be used. Pseudo-partitives like 'two cup of beans' share many of the lexical items of partitives, so if structure and meaning are projected from lexical items, both need to be considered in parallel. Interestingly, partitives are more constrained than the subset relation. For example, they express only proper partitivity: 'Two of her parents' is odd. Also pseudo-partitives show constraints apparently do not follow directly from their linguistic form, for example in the sentence 'MIT hired two thirds women' two thirds does not apply to women but to the set of people hired, contrary to 'MIT hired two thirds of the women'. The project explores the boundaries between morpho-syntax and non-linguistic cognition, with one leading question: what is the source of these constraints? We adopt the methodology of syntax and semantics, with an experimental component. The issue is addressed pursuing two goals: i) on the basis of syntactic cross-linguistic variation, we build a syntax/semantics for partitives; ii) we establish to what extent the posited format together with standard interpretation rules could predict the attested constraints. The project contributes to integrate syntax and semantics with cognitive sciences. A better understanding of this issue improves our understanding of the human mind and in particular of the language faculty within the human mind. Potential users of the project results outside the research community are computational linguists working on AI language comprehension and production and language teachers.

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз