ParaTaxe: a suite of tools for processing and publishing multi-parallel corpora (JADT 2026)
ParaTaxe (JADT 2026, Palermo): open-source tools by Elnaz Jalilian and colleagues to align, evaluate and publish texts available in many translations.
On 10 July 2026, our paper ParaTaxe: a Suite of Tools for Processing and Publishing Multi-parallel Corpora was presented at JADT 2026, the 18th International Conference on Statistical Analysis of Textual Data, at the University of Palermo. It is published in the conference proceedings (Volume II, pp. 606–613).
What is a multi-parallel corpus?
A multi-parallel corpus brings together one text and its many translations, for example the tales of the Brothers Grimm translated into many languages. Aligning these versions passage by passage makes it possible to compare how translators handled the same sentence across languages.
What ParaTaxe does
ParaTaxe (PARAllel Texts Alignment eXploration and Edition) is a project at the LIDILEM laboratory, Université Grenoble Alpes. As a postdoctoral researcher on the project, I designed and evaluated an open-source processing pipeline that:
- imports corpora through a dedicated interface;
- aligns the different versions of a text using alignment techniques based on deep learning;
- is evaluated in terms of precision and recall against a reference sub-corpus;
- publishes the aligned corpora on the web, so that they can be explored online.
The pipeline was applied in particular to the corpus of translations of Grimm's fairy tales, the Grimm Tradalign corpus, linked to the Grimm Tradalign platform (Université Grenoble Alpes and Université de Lausanne).
What's next
A data paper describing the Grimm Tradalign multi-parallel corpus is in preparation for the journal Data & Corpus, with me as first author.
Reference
Kraif, O., François, C., Jalilian, E., Rimasson Fertin, N. & Roux, P. (2026). ParaTaxe: a Suite of Tools for Processing and Publishing Multi-parallel Corpora. In A. Plaia, M. Sciandra & A. Albano (eds.), JADT 2026 – 18th International Conference on Statistical Analysis of Textual Data, Proceedings, Volume II (pp. 606–613). Palermo: University of Palermo. ISBN 978-88-5509-883-0.