On the various kinds of post-editing – Bilingual post-editing, translation memory match editing, hybrid post-editing, monolingual post-editing, stealth post-editing and synthetic-text editing

Abstract

Since post-editing was first envisaged in the 1950s, advancement in artificial intelligence and computer-aided translation tools has led the original concept to diversify. Translators are now asked to do bilingual post-editing, translation memory match editing, hybrid post-editing – which is the combination of the first two – and even stealth post-editing, where the post-editing task is passed off as a human translation revision assignment. In addition to this, with the rise of generative artificial intelligence, synthetic-text editing is now becoming a real prospect. This paper does not aim to be a definitive guide with detailed case studies and numerous examples, as several works of this nature already exist. Instead, it briefly presents the various kinds of editing performed on machine-generated text, primarily focusing on highlighting the differences between them. Specifically, due to the significant differences between bilingual post-editing and translation memory match editing, the author suggests that initially teaching them separately may be the best approach. However, it is also important to provide students and translators with training in hybrid post-editing. Given the potential demand for synthetic-text editors, research is needed into the kinds of errors generative artificial intelligence makes. Moreover, owing to their ability to communicate with the author in the author’s native language, translators may be well placed to step in as providers of synthetic-text editing services.

Published in

Testo a fronte, n. 69 – II semestre 2023; pp. 239-254.
DOI: 10.82026/11557.