University of Toulouse (France)
The conference will take place in Room D (see the location information)
I will discuss how different grammatical formalisms can be combined with logic and cognitive modeling techniques to account for the meaning of natural language. In a slogan, if you take care of the syntax of representational system, its semantics will take care of itself. I will survey some recent work on the meaning of quantifiers, reasoning, learnability, and evolution of language. The common thread of all the models will be taking the idea of logico-syntactic properties of thought (Language of Thought, LoT) seriously to account for linguistic and cognitive phenomena, showing how grammar can be driving the semantic engine.
A relational schema is a set of metadata describing relational instances, corresponding to tabular data. Schema information includes the names of relations, their arities, and optionally integrity constraints that capture some of the semantics of the data. In this talk I will outline research concerning the “expressiveness” of relational schemas. What does it mean to say that one schema subsumes the “information content” of another? How can one verify a schema-level relationship algorithmically? I will give a quick look at approaches to the problem proposed in database theory, including some work dating to the late 1970’s and my own recent work in the area.
Hopefully the ideas proposed for comparing schemas for structured data can be of interest in comparing the expressiveness of vocabularies in other contexts.