Variation and change in Latin BE-periphrases: empirical and methodological considerations

International audienceThis paper is concerned with the variable distribution of passive and deponent BE-periphrases of the type amatus est (E-periphrases) and amatus fuit (F-periphrases), from ca. 200 BC until 600 AD. Corpus data reveal that the shift from E to F is to a large extent conditioned by the factor ‘Tense’: the rise of F-periphrases is quickest with future perfects, slightly slower with pluperfects, and very slow with plain perfects. I proceed to show that Late Latin E-periphrases exhibit a marked (and unexpected) preference for the ‘head-final’ word order amatus est (rather than ‘head-initial’ est amatus), a tendency which is absent elsewhere. On the basis of this observation I propose that Latin perfective E-periphrases like amatus est are not the historical source of Romance analytic present tense passives like sono amato. At a methodological level, the paper also makes a case for the use of (quantitative) corpus data in Latin linguistics