Typology of Prepositions and their Semantic Equivalents

This project will study the use of prepositions and their equivalents, from the perspective of systematizing factors of their semantics relevant to the design of translation systems. This is done independently of any specific translation project, but it t akes into account experiences from projects like LOGON and projects in which the LOGON machinery is employed, such as MT projects involving Japanese, Korean, English and Norwegian. A common experience from larger, concretely defined projects is that theor etically complex problems tend to on the one hand delay progress to the extent that they were not anticipated from the beginning, and on the other hand get insufficiently or ad hoc resolved, because the context of development does not allow for a sufficie ntly principled investigation. Factors related to prepositions are typical instances of this type of problem: they are often the previously least systematically covered area of the grammars involved, and in a seriously multilingual setting, what is expres sed by prepositions in one language may be expressed by completely different types of constructions in other languages. In this area, the field can therefore be helped by an 'orthogonal' type of project which investigates the various mapping scenarios tha t can arise between typologically distinct languages, but does so using a representational formalism which will fit directly into the formalisms used by a substantive part of ongoing and palnned projects. The present proposal defines an outline for such a project, which would be large in scale and involving many researchers internationally. For the fall 2007, attention is limited to Norwegian, German and Japanese: the first two are addressed for exploring the relationship between two languages both using prepositions, and Japanese is explored as a representative of a language using the main 'alternative' constructs to prepositions, most crucially relational nouns and multiverb constructions.