All sorts of change: a preliminary typology of change in multi-word sequences in the Swiss Text Corpus

While there has been little work yet on quantifying diachronic change in multi-word sequences (MWS) beyond studies of individual expressions or small groups thereof, the increasing availability both of diachronic corpora and more adequate processing power and software, make larger scale quantitative analyses of diachronic change in MWS more easily possible. As part of an on-going research project into diachronic change in German MWS, the present paper seeks to outline a possible typology of change based on motivation. The data for this analysis are taken from the Swiss Text Corpus, a 20-million-word corpus of standard written German as used in Switzerland, covering language across the 20th century (Bickel et al. 2009). Frequent contiguous MWS, between 2 and 7 words in length, were extracted from each of four temporally ordered subcorpora using a version of the N-gram Statistics Package (Banerjee, & Pedersen 2003). The resulting lists were then consolidated into four substring reduced lists, one list per time period, which were subsequently compared. This analysis was used as the basis for an investigation into the motivation of observed changes. This yielded a typology of change in four categories of motivation among which the group of changes without readily identifiable motivation was the largest. While the derived typology is in need of some refinement and confirmation on other data, it sheds light on the characteristics of MWS from a new angle. References: Banerjee, S. & Pedersen, T., 2003, Proceedings of the 4th International Conference on Intelligent Text Processing and Computational Linguistics. Mexico City, The Design, Implementation and Use of the Ngram Statistics Package. Bickel, H., Gasser, M., Häcki Buhofer, A., Hofer, L. & Schön, Ch., 2009, Schweizer Text Korpus, Linguistik Online, 39(3)