Extensive data for morphology: using the World Wide Web

Abstract

ABSTRACTThis paper presents a number of recent studies in French morphology which make extensive use of data. These data relating to derived words have been automatically collected from digital corpora, mostly from the Web. The main point developed here is that this massive increase in the amount of available data can substantially modify the results of a morphological study, and can lead to new theoretical conclusions that would not have been possible with traditional data such as wordlists gathered from dictionaries. However, using the Web as a corpus brings up several technical and methodological questions, which are dealt with through examples and discussions about the different tools and techniques available. We exemplify our thesis through the study of the suffixal forms: -esque, -este, -able, -ment.