Eva Pettersson In my talk I will present my work on automatic extraction of verb phrases describing work from Early Modern Swedish text. This work has been conducted in cooperation with the historians at the Department of History at Uppsala University. In their ongoing Gender and Work project, they are storing information in a database on what men and women did for a living in the Early Modern Swedish society (i.e.\ approximately 1550--1800). During this work they have found that working activities in their source material are most often expressed in the form of verb phrases, such as `to fish herring' or `to sell clothes'. I try different methods for finding these verb phrases, using taggers and parsers developed for Modern Swedish to perform linguistic analysis of the text, succeeded by bag-of-words classification for extracting only those verb phrases that are likely to express working activities. To cope with spelling variation in the input text, I perform automatic spelling normalisation (using character-based statistical machine translation techniques) before the NLP tools are applied. Back to program