The aim of this project is to explore the relevance and the potential theoretical advantages of relativist semantics over the
classical propositionalist theories within the debates over the cognitive transparency of semantic contents in linguistics, logic, and
philosophy.A content is transparent if a competent thinker or speaker is able to identify it and to distinguish it from any other
content by introspection alone, without an empirical inquiry.Relativist theories evaluate the contents of our utterances and
thoughts at centred worlds, consisting of a possible world w and a centre c in that world. They allow for a distribution of
truth-conditional components over two levels: the content and the relevant circumstance of evaluation.What we propose to explore in detail
is the original idea that the circumstance of evaluation (rather than the content) may be referential, i.e. centred on a referent.
If it is viable, this idea promises to reinstate the thesis that contents are transparent while preserving the new externalist
orthodoxies: we can have contents that are never in themselves referential, and remain purely descriptive or qualitative, and at the
same time truth-conditions that are externalistic, and depend essentially on the environment. That is the case when the content of
an empirical utterance or thought, which bears on a certain environment, must be evaluated relative to that environment, or
centred world.This theory, just like the relativist family to which it belongs, will be explored in several directions. Its explanatory
success, its theoretical foundations, its domain of application, and its limits will be assessed through linguistic, logical, and
philosophical analyses.