PREMOVE – A diachronic dataset of Ancient Greek and Latin annotated PREverbed MOtion VErbs

The PREMOVE dataset includes the manual annotation of 2,835 occurrences of preverbed motion verbs in Latin (1,484) and Ancient Greek (1,351). Each verb occurrence includes lemmatisation, as well as the annotation of morphology, syntax (Figure/Ground, spatial relations), semantics (preverb semantics, verbal/nominal semantics, verb class, actionality), geography (links to the World Historical Gazetteer whenever place names occur with the preverbed verb), metadata (author, work, literary genre, century, language period). PREMOVE has been created in 2025 following the guidelines created by Farina (2024) and available at https://doi.org/10.18742/25055573. The corpus on which annotations are based, i.e, the PREMOVE Base Corpus, is given in the "PREMOVE_Base_Corpus.png", from Farina & Ciletti (2026. Forth). More information about the corpus is given in Farina (2026). PREMOVE is also available as a queriable interface on PrevNet (https://prevnet.sites.er.kcl.ac.uk/). PREMOVE and some of its annotations have also been linked to the LiLa Knowledge Base (https://lila-erc.eu/data/dataset/PREMOVE/dataset). The PREMOVE dataset is hosted on the Oxford Text Archive (http://hdl.handle.net/20.500.14106/2579). It is also temporarily hosted on Zenodo due to OTA maintenance (January 2026 –). Please cite as: Farina, Andrea. 2025. PREMOVE – A diachronic dataset of Ancient Greek and Latin annotated PREverbed MOtion VErbs. Oxford Text Archive. http://hdl.handle.net/20.500.14106/2579.