The meaning of linguistic entities features a high degree of variability and openness. In language usage, this uncertainty of meaning is dissolved by context sensitive interpretations in favour of a very little variance. Especially for the domain of sentences, it is not yet clear how this identification of specific meaning takes place. Which linguistic mechanisms are responsible for these context sensitive interpretations? How do combinatory principles of sentence meaning interact with pragmatic strategies in an evaluation of meaning?
A lot is known in syntax about the characteristics of cognitive processes that are triggered by linguistic structures; the sub-processes of online interpretation, in particular the composition of meaning, are far less well-understood. One central question is how the supposed incrementality of cognitive processes can be related to theories of the compositional processing. Existing interpretations of Frege's principle of compositionality are based on the sentence as a whole and ignore the temporal dimension of the semantic processing almost completely. The size and nature of sub-sentential units to which the various aspects of an incremental computation of meaning refer, and the processes that control the actual composition of meaning in time, are therefore unresolved issues.
Research in crosslinguistic variation has traditionally focused on morphology and syntax. Mechanisms of interpretation have commonly been regarded as universal. However, linguistic structures describing the same state of affairs often seem to differ substantially in individual languages. It is therefore reasonable to also expect variation within the semantic component of grammar. - What are the possibilities and limitations of such semantic variation? Where and how is semantic variation located in a given grammar? How did it develop, and what possibilities are there for language change? In how far does language change obey universals laws or tendencies?