The Syntactic Projection Problem and the Comparative Syntax of Locative Inversion

The fundamental problem this study addresses is how to predict the syntactic properties of verbs from information about their meaning and use－the 'syntactic projection problem'. That a universal solution to this problem must exist has been argued from nature of language acquisition (Pinker (1989)). But the most compelling evidence for universal projection principles comes from comparative syntax, where languages that may differ genetically, areally, and typologically can be shown to instantiate the same principles for projecting verbal meaning and use into syntactic structures. Such a case is examined here. English, a West Germanic language spoken in England and its former colonies, is genetically and areally unrelated to Chichewa, a Bantu language spoken in East Central Africa. The two languages also differ typologically, English belonging to a group of languages that employ case and government to express syntactic relations, and Chichewa belonging to a group that employs noun class and agreement instead (Bresnan and Mchombo (1987)). Despite these differences, English and Chichewa show remarkable correspondences in the properties of locative inversion, a syntactic unaccusative alternation studied in Chichewa by Bresnan and Kanerva (1989): I will show that at the level of argument structure and function, English and Chichewa are subject to the same principles of syntactic projection, from which the unaccusativity or inversion phenomenon arises (following Bresnan and Kanerva (1989)).