Modern Social Science Research

This paper based on examples from transportation scientific and technology texts, analyzes the performance of the chatbot ChatGPT in Chinese-English translation tasks from two aspects: lexical cohesion and structural cohesion within the discourse structure, with a particular focus on the post-editing process that follows. The paper selects portions of the book Railway Transport of Long and Large Goods in Foreign Countries and tasks ChatGPT to complete the translation based on well-crafted prompt instructions. Subsequently, the author conducts post-editing on this basis and summarizes the performance of ChatGPT in the field of Chinese-English translation of transport science and technology texts, as well as the considerations for translators during post-editing, providing a reference for translators in practice. The results reveal that ChatGPT excels in syntactic cohesion but performs poorly in vocabulary and semantics, necessitating substantial post-editing work. Therefore, while translators embrace and guide the development of translation technology, they must also enhance their own skills, as human involvement is an indispensable and crucial component, both in quantity and quality, thereby promoting the ideal of interdependence and symbiosis between humans and machines into reality.

Scientific and technical texts; ChatGPT; Post-translation editing