Porting an English semantic tagger to the Finnish language
Laura Löfberg, Dawn Archer, Scott Piao, Paul Rayson, Tony McEnery, Krista Varantola, Jukka-Pekka Juntunen · Lancaster EPrints (Lancaster University) · 2003
Semantic annotation is an important and challenging issue in corpus linguistics and language engineering. While such a tool is available for English in Lancaster (Wilson and Rayson 1993), few such tools have been reported for other languages. In a joint Benedict project funded by the European Community under the `Information Society Technologies Programme', we have been working towards developing a Finnish semantic tagger that will parallel the existing English semantic tagger. The intention is to avoid building a completely new system but to bootstrap using the existing software and the largely hand-constructed English lexical resources. In this paper, we report on our work to date, which includes (i) a comparative study of the grammar of English and Finnish, (ii) the tagging of an English-Finnish parallel corpus, and (iii) the building of a Finnish lexicon using existing lexicons and software such as a Finnish-English-Finnish machine-translation system, Finnish dependency parser and morphological analyser, etc. This paper also discusses some challenging issues that have arisen during the construction of the parallel semantic tagging system between English and Finnish, namely, the complications caused by the widely different grammatical systems of the two languages. We believe that our work will provide a valuable experience for the community working on cross-language annotation schemes.