On the identification of stems in FASS

Work has been under way since spring U>do ac tlie Center for Computational Linguistics to ptxxiuce a key worxl index of basic forms for FASSIS by semi­ automatic means. One step towards this goal is to build a stem dictionary covering the material (see further Sågvall Hein l9Ö5a). Two parallel lines have been followed side by side. On the one hand, the full text of FASS has been investigated and prepared for treatment by the Center’s program package for linguistic text processing, TF.XTFACK, which produces word lists, concordances, etc (see Ros6n 19ÖÖ and Rosén Å Sjöberg 19Ö5 for further details). On the other hand, the present key word file in FASSIS has been used as a test file In attempting to identify stems in the material. This presentation will concentrate on these attempts.