Digitising preverbs with PrevNet. A cross-linguistic analysis of motion verbs in Ancient Greek and Latin

This seminar will be held in person and online. Advance booking is required for online attendance and strongly advised for in person attendance.

Abstract: Since Busa’s pioneering work on Latin (1974-1980), the digitisation of Ancient Greek and Latin texts has paved the way for quantitative analyses (e.g., Jenset & McGillivray 2017). While both languages now benefit from numerous online resources focusing on word parsing (Crane 1991) or semantic analysis (Biagetti et al. 2021), none specifically address preverbs – prefixes attached to verbal bases – despite their frequent use in ancient Indo-European languages (Booij & van Kemenade 2003). This talk aims to introduce PrevNet, a new digital resource designed to study Ancient Greek and Latin preverbs, using a case study on motion verbs. With its comparative and interdisciplinary scope, PrevNet will be highly expandable. Motion verbs have been selected as a starting point due to their cross-linguistic relevance and the valuable insights they provide into cognitive processes involved in conceptualising movement. To develop this resource, I have compiled a large linguistic dataset covering preverbed motion verbs from the 8th century BCE (Ancient Greek) and 3rd century BCE (Latin) to the 2nd century CE. The resulting corpus includes more than 540,000 words and spans a variety of genres (poetry, historiography, theatre, oratory, novel, philosophy), allowing for more comprehensive results. The analysis focuses on eight motion verbs (e.g., Lat. venio/eo, AGr. baínō/eîmi ‘go’, Lat. navigo, AGr. pléō ‘sail’, Lat. volo, AGr. pétomai ‘fly’), and 16 preverbs per language, including both spatial and non-spatial preverbs. Over 2,800 verbal forms have been manually annotated at multiple levels (Farina 2024). PrevNet will feature detailed linguistic annotations, such as distinctions between literal and figurative meanings of preverbed motion verbs, and support cross-linguistic analyses, enabling scholars to uncover patterns that are difficult to capture through traditional qualitative methods. Designed for both linguists and non-linguists, PrevNet aims to use linguistic data to uncover socio-cultural aspects of Ancient Greek and Roman civilisations, contributing to the field of cultural analytics.

Bio: Andrea Farina is a PhD Candidate in Digital Humanities at King’s College London, funded by the London Arts & Humanities Partnership. He holds a BA in Classics (University of Milan) and a MA in Theoretical and Applied Linguistics (University of Pavia). His research interests mostly concern the syntax-semantics interface, quantitative analyses in ancient languages, and digital/computational tools and methods for linguistic research. He is Associate Editor of the Journal of Open Humanities Data, creator and convenor of the Data in Historical Linguistics international seminar series, and member of the Regional Advisory Committee of the Global Council for Anthropological Linguistics.