Translating The Ickabog into French: An Analysis of Translation Strategies for Neologisms and Proper Nouns in Human and AI Translation: Investigating Creativity in Human and AI Translation Approaches

Details

Supervisors

Lefer, Marie-Aude

Faculty

Faculté de philosophie, arts et lettres

Degree label

Master [120] en langues et lettres modernes, orientation générale, à finalité approfondie

Abstract

enThe present master’s thesis investigates the translation of neologisms and proper nouns in children’s literature, focusing particularly on The Ickabog and its French translation. The thesis aims to analyse the translation strategies used for neologisms and proper nouns in the human translation and to compare them to the strategies observed in the AI-generated version, with particular attention to their level of creativity. In addition, this work investigates end users’ preferences regarding both versions of the translation. To this end, 91 occurrences of neologisms and proper nouns were manually extracted from the original version of The Ickabog and compiled into an Excel spreadsheet. Afterwards, these lexical items, along with some surrounding context and a prompt, were inserted into the GPT-5 model of ChatGPT. A comparative analysis was then conducted to (i) identify the translation strategies used for neologisms and proper nouns in the human and AI translations respectively (ii) compare both sets of strategies with regard to their level of creativity (iii) analyse end users’ preferences between excerpts of the two versions through a survey. The results show that the human translator employed a diverse combination of translation strategies for neologisms and proper nouns. Furthermore, the human translation tends to use a greater number of creative strategies, particularly in the translation of neologisms. In contrast, the AI-generated translation appears to be not quite as creative, remaining closer to the source material. However, the survey reveals that the AI-generated version can be positively perceived by end users, although human translation was recognised to be more creative.