Methodological Aspects of the Categorization of a Middle French Corpus with a Computer-Assisted Text Analysis Software (SATO)

Abstract

Achievements in computer science during the last decade have had major implications for diachronic linguistics as well as for other linguistic disciplines. Since the range of data available for historical syntax cannot be broadened by performing experiments, the study of syntactic change in time is highly dependent upon the availability of extended corpora. For analytical purposes, the use of computational tools has by now become a necessity. In medieval studies, computerized corpora have been constituted,¹ and printed concordances tend to be replaced by machine ones.² However, the main problem in this field is to have at one’s disposal software to tag corpora. In the following paper we shall focus on the study via SATO, a computer-assisted text analysis software, of the constraints on the locus of subject and verb in a Middle French (MidF) corpus.

Research for this article was made possible by SSHRC grant No. 410-89-1409 and 410-89-0785 and by a UQAM special grant. We thank all the members of the CASH project for their comments and support. This paper is a joint project between CASH (UQAM) and ELI (ENS de Fontenay St Cloud).