This study examines how metaphor loss occurs in the translation of feminist literary texts and comparing between human translators and ChatGPT. Metaphor is important in shaping meaning, culture, and ideology, but it is also one of the vulnerable elements, since it often simplified, shifted, or omitted. Previous studies have shown that translation loss is unavoidable. However, they rarely compare human and machine translation. This research uses a descriptive qualitative design with a comparative analysis. The data are taken from two Indonesian novels, Saman and Maya, and the translations produced by human translators and ChatGPT. The data is focusing on metaphors related to feminist themes such as body, identity, sexuality, and power. The findings show that human translators tend to keep metaphorical meaning by changing metaphors’ vehicle, while sometimes turning metaphors into non-metaphor. In contrast, ChatGPT mainly uses more predictable and simplified expressions. So that, human translation keeps more meaning and ChatGPT tends to make the text flatter. The study concludes that metaphor loss is not only a linguistic issue but also affects meaning changes.
Copyrights © 2026