How to Read All Languages:  Translation Alignment with Ugarit

1. Abstract

This workshop will illustrate the importance of translation alignment in the field of slow reading and language learning. We will provide a short theoretical overview on the principles of translation alignment, together with a hands-on tutorial on Ugarit (http://ugarit.ialigner.com/), a web-based translation alignment editor. Ugarit is designed as a Citizen Science tool, aiming at collecting training datasets of manually aligned words from diverse text corpora. The ultimate goal of Ugarit is to improve automatic translation alignment methods and to implement a set of dynamic lexica, with particular regard for languages with less supported infrastructures. However, the tool also has a strong pedagogical potential, which has been assessed in the course of various hands-on workshops and in an ongoing integration in school curricula: we have tested how translation alignment with Ugarit can help readers to engage with languages that they have never seen, grasping their essential semantic and morphological aspects. We propose text alignment as a way to empower the perception of the complexity of a language, but also as a method to leverage usual obstacles in the process of reading apparently “impenetrable” sources by directly engaging with them. At present, Ugarit includes aligned pairs from 36 languages (including less represented languages, like Bulgarian, Ethiopic, Sanskrit, Yiddish, and Armenian), 277 unique users, and about 23,400 parallel texts hosted.