This project aims at testing whether Latin is configurational, and by this token whether it is possible for a linear string of Latin words to be structurally ambiguous. More specifically, I will evaluate the hypothesis that Latin is a discourse configurational language, in which pragmatic notions like topic and focus are encoded structurally rather than linearly. By doing so, I reconcile the widely accepted idea that information structure has a strong influence on Latin word order with more recent work that approaches Latin syntax from a purely structural angle.
I will carry out a corpus-based experiment (prose texts from 160 BC - 550 AD, ca. 1.250.000 words) to test whether the pragmatic status of direct objects is correlated with their linear or with their syntactic position in the clause. In order to characterize the latter, I draw upon the theoretical literature from the 'cartographic' branch of generative grammar. The data will be analysed statistically in R, using a mixed effects model.
Should the factor 'syntactic position' emerge as the stronger predictor for information structure, the discourse configurationality hypothesis will receive strong support. In this case, the consequences are far-reaching, as all studies of Latin word order would have to control for structural ambiguity, which in the existing literature has never been done. As a result, many synchronic and diachronic aspects of Latin word order might have to be reconsidered.