Translation Concepts in the Context of Human-Machine Communication

Although Translation Studies encompass a rich variety of research topics, its disciplinary narrative can be broadly categorised into two primary streams: one stemming from considerations about translation as a social and cultural activity, evolving from the examination of translated religious and literary texts and later expanded to investigate the functional and human aspects of translation; and another grounded in the computational aspect of language and translation. While the former is typically known as translation studies, the latter, commonly referred to as machine translation, more commonly associated with the field of natural language processing in computer science. Historically, these two narratives have seldom intersected. However, the advent of Generative Artificial Intelligence necessitates a long-awaited dialogue, as today, humans not only communicate with humans for translation tasks, but also directly with language models. 

This paper explores the efficacy of integrating conceptual tools traditionally used in human-to-human communication for translation purposes (i.e., translation brief, and author/translator persona) into the communication with GPT-4. The findings indicate that, although certain elements are constructive in facilitating human-to-human communication for translation tasks, their effectiveness is limited for improving translation quality in the context of human-machine communication. This highlights a need for further research into how translation theorists and practitioners might develop the current set of conceptual tools rooted in the human-to-human communication paradigm for translation purposes in this emerging workflow involving human-machine communication.