Henriëtte de Swart & Bert Le Bruyn

Time in Translation – five years later

———————–

Abstract

In this talk, we report results from the Time in Translation project (https://time-in-translation.hum.uu.nl/) carried out at Utrecht University between 2017 and 2022. Reference to time and events is part of human cognition, but languages vary in the expression of temporal information. We know from the literature that categories like tense (e.g. present/past/future tense) and aspect (e.g. perfective/imperfective/progressive) are subject to cross-linguistic variation. However, the datapoints reported for the languages under investigation are often difficult to compare, because temporal information is highly context sensitive. Translation Mining was developed as a methodology to extract data from parallel corpora as the input to a systematic multilingual comparison of form-meaning mappings in context. The distributional patterns are extrapolated to linguistic theory, and the results contribute to cross-linguistic semantics.

The original research focus was on the present perfect, because we knew from the literature that English, Dutch, German, Spanish and French use this verb form with different meanings. Over the course of time, the project expanded to a larger number of languages and tense-aspect categories. The findings were replicated in different datasets, and insights were tested against experimental research. These extensions did not only enable us to properly identify the ingredients of a cross-linguistically robust semantics of the perfect, but also helped to ground them firmly into a broader theory of tense and aspect.