ARGUMENTATION AND NATURAL-LANGUAGE - PRESENTATION AND DISCUSSION OF 4FOUNDATIONAL HYPOTHESES

As an introduction to this special issue on argumentation, I propose a reflection on the conception of language (langue) and, in particular, of semantics, which seems to have guided the work exposed in this iss ue. After analysing the conditions under which an empirical study of t he semantic system of natural languages can be carried out, I examine, in partial fulfilment of the requirements involved in such an enterpr ise, four hypotheses, inspired by the work of Jean-Claude Anscombre an d Oswald Ducrot, which are related to the semantic description of huma n languages. The first hypothesis postulates that the meaning of a sen tence (or other expression of a natural language) can be described as a set of instructions for the possible interpreters, which state how t hey are supposed to build the senses of the possible utterances of tha t sentence, using their beliefs, knowledge and points of view about th e situation of utterance, According to the second hypothesis, these in structions are concerned, at least partially, with the point of view r equired in order to be able to interpret the possible utterances of th at sentence. The third hypothesis postulates that these points of view can be described using gradual fields (topical fields), monotonic rel ations between those fields (topoi), and constraints on those fields a nd relations (topical constraints). Finally, according to the fourth h ypothesis, the words of a natural language are conventionally associat ed to topical fields, topoi and/or topical constraints; these associat ions are claimed to be responsible for the fact that the use of such o r such word reflects an ideology, a belief or, simply, a particular co mpetence.