Quantitative evidence and diachronic syntax

This paper focuses on the advantages and limitations of quantitative evidence in the study of diachronic syntax. It is argued that, although quantitative evidence is indispensable in historical linguistics, quantitative data can only contribute to a better understanding of diachronic change when two premises are fulfilled: first, the researcher has to be transparent about the sources and the methodology used in the investigation and second, the investigation of texts has to be based on hypotheses about possible individual grammars and clear predictions about the kind of evidence supporting and contradicting them. Although these points seem at first sight to be self-evident, they are not always standard in diachronic research. We illustrate how to gain the best possible advantage of the availability of quantitative data by referring to a selection of case studies on French we have conducted.