Building Cat3LB: a treebank for Catalan

Abstract

In this paper we present some specific issues related to the syntactic annotation of Cat3LB, a 100,000-word Catalan treebank (2,500 sentences). In the syntactic annotation we follow an incremental process with different levels of complexity: bracketing and labelling of constituents and functional tagging. Some automatic pre-processing steps have been done: morphological analysis, tagging and chunking. In this paper we will concentrate, however, on the syntactic annotation.