To build an ontology of a domain, it is necessary to categorize this domain with objects, relations between objects, process acting onto objects, process transforming a state and building an event and so on ... Often, these different entities and relations must be identified inside linguistic segments (nominal phrases, clauses, sentences, paragraphs titles ...) by means of syntactic and semantic annotations. The Cognitive and Applicative Grammar (CAG) is a polystratal model (Descles, 1990, 2004, 2005), that extends the Shaumyan's Universal Applicative Grammar (1987). This model opens a way towards a sound bridge between Formal Ontology, Logics, Cognitive Linguistics and Natural Language Processing to annotate texts. The underlying formalism of all levels of CAG is always applicative or functional one. We consider the follow applicative scheme (AS) [ξ{sub}1 = ω @ ξ{sub}0], where ξ{sub}1 is the place of the result build by the application, designated by '@' of an operator at the place 'ω' in (AS), acting onto an operand in the place 'ξ{sub}0' in (AS). Fundamental distinctions are basic: operator, operand, object (individual or class). Operator/operand is context relative since the same applicative expression can either be an operator applied to an operand, sometimes to itself, or an operand of another operator. However, by definition and following Frege, an "object" is never an operator and it cannot stand for the place 'ω' in (AS).
展开▼
