Replication Data for: Alea iacta est. Insights from corpus semantics into the diachrony of the Latin passive

Dataset includes annotated corpus data from Latin texts from the 3rd c. BCE until the 10th c. CE. Attestations of forms of both the original construction ('PP-BE.inf.', e.g. 'cantatus est') and the innovation ('PP-BE.perf.', e.g. 'cantatus fuit') for the expression of the passive (taken together with forms of deponent verbs into 'nonactive') of perfectum stem tenses were extracted from major online corpora (331.131 data points); a random sample (n = 383) of data points of PP-BE.inf. that represents all text types and time periods as evenly as possible was then subjected to a close-reading analysis in order to ascertain the attestation rate of meanings for which it competed with the innovation PP-BE.perf. (specialized in ANTERIORITY) and the present tense (specialized in present events or situations). Only the data points that were annotated in full in this second phase are included in the current dataset. For the data points examined in the first phase, only the formal categories in the list below were annotated to the extent that these annotations are not subject to interpretation.