Content and power: emotional intensity in Basque machine translation
In this small empirical study designed as a first step, the author of this article examines how well the original affective force is preserved in translating Spanish social media posts into Basque using machine translation. To do this, she created a corpus of simulated tweets, tested them with three translation systems (Itzuli, ChatGPT and Latxa), and analyzed them with human evaluators participating via crowdsourcing. The results clearly demonstrated that, when translating strong messages, machine translation tends to soften intensity and present texts in a more neutral tone.
