LINGUISTICALLY ANNOTATED SPOKEN NGANASAN CORPUS

The paper discusses the key issues of the annotation method employed in the project "Lingustically annotated spoken Nganasan corpus". The data are processed and stored in the EXMARaLDA format. The annotation of the database involves grammatical and part-of-speech tagging (made in Toolbox or Flex), translation into Russian and English. However, the present paper addresses the questions of syntactic roles, and information structure. For this purpose we use the format designed by other researchers and adapted by us to the Nganasan language. In the paper we describe the system of annotation (tags, terms and their clarification) illustrated by a large amount of Nganasan examples.