Statistical Dependency Parsing of Four Treebanks

Multilingual dependency parsing is gaining popularity in recent years for several reasons. Dependency structures are more adequate for languages with freer word order than the traditional constituency notion. There is a growing availability of dependency treebanks for new languages. Broad coverage statistical dependency parsers are available and easily portable to new languages. Dependency parsing can provide useful contributions in areas such as information extraction, machine translation and question answering, among others. In addition, syntactic head-dependent pairs are a good interface between the traditional phrase structures and semantic theta roles. In this paper we present the learning curves of a statistical dependency parser for four languages: Arabic, Bulgarian, Italian and Slovene. We discuss issues that mostly concern the employed annotation scheme for each treebank with an emphasis on coordinated structures.