1 A Constraint-Based Tagger for Norwegian 1 Kristin Hagen, Janne Bondi Johannessen and Anders Nøklestad The Text Laboratory, University of Oslo 1. Introduction At the University of Oslo, work is currently being carried out to develop an automatic morphosyntactic tagger for Norwegian bokmål and nynorsk. The project comprises seven man-labour years and extends from April 1996 to December 1998 2 . 2. Morphosyntactic taggers 2.1 General Now, what is a morphosyntactic tagger? It is a computer program whose task is to provide correct grammatical information for every word in any stretch of running text. Ideally, every word should have a grammatical description, and the description should be unambiguous. This means that the tagger must be able to disambiguate words that have more than one possible reading, like the word så in (1): (1) ‘On the tram to town today, I saw her again’: På trikken til byen i dag så jeg henne igjen ”se” verb past tense ”så” conjunction ”så” adverb ”så” noun masc. appellative sg. indef. ”så” verb imperative ”så” verb infinitive The tagger should have the ability to decide that så is a past tense verb in this context. 1 Published as: Johannessen, Janne Bondi, Kristin Hagen og Anders Nøklestad. 2000. A Constraint-based Tagger for Norwegian. I Lindberg, Carl-Erik og Steffen Nordahl Lund (red.): 17th Scandinavian Conference of Linguistics. Odense Working Papers in Language and Communication 19, 31-48, University of Southern Denmark, Odense. 2 The project is mainly financed by the Norwegian Research Council, the Documentation Project, the Text Laboratory and the Finnish company Lingsoft.