Home
“diviser chacune des difficultés [...], en autant de parcelles qu'il se pourrait, et qu'il serait requis pour les mieux résoudre.” René Descartes, Discours sur la méthode.
The research follows the principle that the meaning of a sentence comes from the meaning of its parts. Unless learned as formulaic material, these parts provide an abstract indication that is enriched by context to relate to the shared experience of speakers. Questions arise as to how underspecified meaning is represented, how it is enriched by contexts, what constraints bear on this enrichment process, and whether these are language-specific or relate to general cognitive mechanisms in their learning and treatment.
The study of the abstract aspects of meaning I pursue with reference to grammatical phenomena such as negation, indefinites and qualitative expressions. As these questions relate to what does not exist, they are less likely to be influenced by knowledge of the world than lexical elements would be. The variation of interpretation of underspecified items across contexts in English and French helps me identify representations and mechanisms internal to linguistic meaning organisation, as does regional variation, historical evolution and across languages. Whether these mechanisms are specific to language can be evaluated by looking at the behaviour of public discourses, how they work and how they change, which I have done for issues of linguistic minorities, ethnic questions and gender issues.
At the moment, I’m very interested by how information known both to speaker and hearer influences the shape of negative sentences. It is in these discourse-old contexts that new negatives are assumed to emerge in the history of French, and I have shown that this is where some are found before they disappear altogether. Unexpected uses such as positive polarity item with a clause-mate negative, negative doubling, negative adverbial questions and (self) corrections are found to be made possible by discourse-old information. The dynamics of information structure is optimally assessed through less formal exchanges, calling for the creation of new resources such as the Anglo-Norman Year Books Corpus.