Stylometry in Computer-Assisted Translation: Experiments on the Babylonian Talmud

What it is about

The Italian translation project of the Babylonian Talmud works in the Traduco web tool, where each translated string passes through translator, reviser, editor and curator. The authors take the translator's last version of each block or logical unit and its final version, and measure the Levenshtein distance between them as the revision effort. For the tractates Berakhot and Ta'anit (about 216,000 and 82,000 tokens) they run an existing Italian POS tagger and dependency parser, and correlate a set of raw, lexical, morpho-syntactic and syntactic features with the revision measure using Spearman's coefficient. Size features correlate most strongly, as expected. Among the syntactic features, the number of prepositional chains, the length of dependency links and the number of subordinate clauses also correlate with more revision. The study is a correlation, not a predictor, and it does not test a tool.

Ideas

• I1. A revision-risk check on dependency trees: for each sentence, the distance between a translator's draft and its final version, set against features Mova's trees already give (length of the longest dependency link, subordinate clauses, prepositional chains). Test whether a simple, explicit rule ranks sentences in the same order as the revisions do.
• I2. Keep the score explainable: a flag that says "a dependency link of seven tokens and two subordinate clauses" rather than one opaque number. This is the transparent form Mova prefers, and it could be shown in a translation-support view.
• I3. Use the version history as a held-out set: draft against final, split by tractate, used once per release, as the paper used its two datasets. Before any use, the licence of the Talmud translation data must be checked.

For Mova

I1 and I2 are a design note for prag (explanations over tree features) and for the English side of any translation-support tool. I3 is a future test set, not in the repository until its licence is confirmed. The paper's own finding that the source-side style of Mishnaic Hebrew does not help revision support is a reason not to build a source-style module for 11. Mishnaic Hebrew (hbo-x-mishnaic) on this basis.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.4000/books.aaccademia.2414

@inproceedings{giovannetti2017stylometry,
  title = {{Stylometry in Computer-Assisted Translation: Experiments on the Babylonian Talmud}},
  author = {Giovannetti, Emiliano and Albanesi, Davide and Bellandi, Andrea and Dattilo, David and Dell'Orletta, Felice},
  booktitle = {Proceedings of the Fourth Italian Conference on Computational Linguistics (CLiC-it 2017)},
  year = {2017},
  doi = {10.4000/books.aaccademia.2414},
  url = {https://doi.org/10.4000/books.aaccademia.2414}
}

References

20 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Alpaslan Acar and Korkut Uluç İŞİSAĞ. 2017. Readability and comprehensibility in translation using reading ease and grade indices. International Journal of Comparative Literature and Translation Studies, 5(2):47–53. ↗
2. Giuseppe Attardi, Felice Dell’Orletta, Maria Simi, and Joseph Turian. 2009. Accurate dependency parsing with a stacked multilayer perceptron. In Proceedings of the 2nd Evaluation Campaign of Natural Language Processing and Speech Tools for Italian, (EVALITA 2009). find
3. Mona Baker. 2000. Towards a methodology for investigating the style of a literary translator. Target. International Journal of Translation Studies, 12(2):241–266. ↗
4. Andrea Bellandi, Giulia Benotto, Gianfranco Di Segni, and Emiliano Giovannetti. 2016. Investigating the application and evaluation of distributional semantics in the translation of humanistic texts: a case study. In Proceedings of the 2 nd Workshop on Natural Language Processing for Translation Memories, pages 6–11. find
5. Andrea Cimino and Felice Dell’Orletta. 2016. Building the state-of-the-art in POS tagging of italian tweets. In Proceedings of Fifth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian. ↗
6. Final Workshop (EVALITA 2016), Napoli, Italy, December 5-7, 2016. find
7. Felice Dell’Orletta, Simonetta Montemagni, and Giulia Venturi. 2014. Assessing document and sentence readability in less resourced languages and across textual genres. In John Benjamins Publishing Company, editor, Recent Advances in Automatic Readability Assessment and Text Simplification. Special issue of International Journal of Applied Linguistics, 165:2, pages 163–193. ↗
8. Emiliano Giovannetti, Davide Albanesi, Andrea Bellandi, and Giulia Benotto. 2017. Traduco: A collaborative web-based cat environment for the interpretation and translation of texts. Digital Scholarship in the Humanities, 32(suppl 1):i47–i62. ↗
9. Magda Heydel and Jan Rybicki. 2012. The stylometry of collaborative translation. woolf’s night and day in polish. In Digital Humanities 2012 Conference Abstracts, pages 212–217. ↗
10. Libo Huang. 2015. Readability as an indicator of self-translating style: A case study of eileen chang. In Style in Translation: A Corpus-Based Perspective, pages 95–111. Springer. ↗
11. Patrick Juola. 2008. Authorship attribution. In Now Publishers Inc. find
12. Sholeh Kolahi and Elaheh Shirvani. 2012. A comparative study of the readability of english textbooks of translation and their persian translations. International Journal of Linguistics, 4(4):344. ↗
13. Moshe Koppel, Jonathan Schler, and Kfir Zigdon. 2005. Automatically determining an anonymous author’s native language. In Intelligence and Security Informatics, vol. 3495, LNCS, Springer–Verlag, pages 209–217. ↗
14. Alexander Mehler, Serge Sharoff, and Marina (Eds.) Santini. 2011. Genres on the web. computational models and empirical studies. In Springer Series: Text, Speech and Language Technology. ↗
15. Alessandra Pecchioli. 2017. Elaborazione del linguaggio naturale (nlp) in ebraico: il caso dell’analisi linguistica automatica applicata all’ebraico mishnaico del talmud. Oral communication, sep. find
16. XXXI Convegno AISG 2017 - Nuovi studi sullEbraismo, 4-6 settembre 2017, Ravenna, Italy. find
17. Stefan Richter, Andrea Cimino, Felice Dell’Orletta, and Giulia Venturi. 2015. Tracking the evolution of written language competence: an nlpbased approach. In Cristina Bosco, Sara Tonelli, and Massimo Zanzotto, editors, Proceedings of the Second Italian Conference on Computational Linguistics -CLiC-it 2015, pages 236–240. find
18. Jan Rybicki. 2012. The great mystery of the (almost) invisible translator. Quantitative Methods in Corpus-Based Translation Studies: A practical guide to descriptive translation research, 231:231– 248. find
19. Hans van Halteren. 2004. Linguistic profiling for author recognition and verification. In John Benjamins Publishing Company, editor, Proceedings of the Association for Computational Linguistics (ACL04), pages 200–207. ↗
20. Sze-Meng Wong and Mark Dras. 2009. Contrastive analysis and native language identification. In Proceedings of the Australasian Language Technology Association Workshop. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2017-giovannetti-stylometry-computer-assisted-translation-experiments/