Complex sentences in Brazilian Native American languages: experimental and theoretical studies (COSY)

Complement clauses, relative clauses, infinitives, control, embedded and long questions, long and logophoric binding, and adverbial clauses are central topics in the grammars of well-researched languages. In addition, the analysis of complex sentences is in many cases illuminating for the analysis of simpler sentences, such as the verb-second and verb-final word order in German. Descriptive work on less studied languages, on the other hand, rarely considers complex sentences, and the tools to collect interesting theoretical data on complex sentences are underdeveloped.

The COSY project fills this gap for five native Brazilian languages: Maxakalí, Kaingang (both Je languages), Yudja, Kawaiwete (both Tupi languages), and Kuikuro (Carib languages). COSY's contribution consists of two parts: theoretical insights into the syntactic and semantic principles and parameters of human language and practical advances in techniques to describe complex sentences in field research. To achieve these goals, COSY creates an intensive exchange between Brazilian scientists of the UFRJ and German scientists at ZAS.

The theoretical goals of COSY concern the understanding of the initially mentioned areas as well as reference change marking, index shifting, casus systems, evidentiality and speech act marking. The detailed investigation of complex sentence structures in the five American languages will lead to new discoveries that are important for syntactic and semantic theory as part of the human language faculty. The practical advances on which these insights are based are spontaneous elicitation techniques, an improved overview list, and several other experiments developed by COSY to systematically collect data on complex sentences. As part of the refinement of these techniques, we simultaneously create a unique corpus of five little studied and threatened Brazilian languages from three different language families. Finally, the project advances the application of novel technology for the analysis and storage of field research data in collaboration with Canadian, Dutch, German and Brazilian scientists and institutions.