From the cognitive turn to AI: reflections on recent trends in translation (studies)

This paper proposal sets out to provide a brief overview of the research developments and technological innovations that took place in the past few decades and gave rise, since the beginning of the 21st century, to an approach to translation increasingly relying on machines. The so-called ‘cognitive revolution’ of the 1950s, as well as seeing the birth of the term Artificial Intelligence (AI) as we understand it today, brought the attention back to the notion of ‘mind’ in many disciplines, from psychology to linguistics (Miller 2003). For Translation Studies, this meant moving away from the product of translation and focusing on the process. Necessarily skipping the developments of the following decades, technology had to reach the 1980s and 1990s, in order for machines to achieve sufficient power and accuracy to become first interesting and then necessary to translation. The results of this change became visible at mass level in the 21st century. In 2010, Halverson (2010) proposed to investigate process-oriented translation research under the name of Cognitive Translation Studies (CTS). In the 2010s, Neural Machine Translation (NMT) – which works based neither on rules nor on statistical guesses, unlike in the past, but is able to quickly ‘learn’ from inputs and adapt to contexts – made significant progress, and was immediately made available globally through Google Translate, for example (Le & Schuster 2016). Into the 2020s, the latest AI-based tool to make the news as well as to fascinate, interest and possibly disturb the translation industry is ChatGPT (OpenAI 2023), which brings the potential of the most up-to-date NMT directly to end-users, who can train the tool themselves (Luo 2023). What is the significance of the cognitive turn in translation for the emergence of AI translation? What may the potential and limits of AI translation be? How is AI going to shape the translation industry in the near and not-so-near future? This proposal is intended as a descriptive effort and, far from providing definitive answers, it might only – if anything – contribute to raising questions about the next steps in translation development. References Halverson, S. L. (2010). Cognitive translation studies: Developments in theory and methods. In G. M. Shreve, & E. Angelone (Eds.), Translation and cognition: Recent developments (pp. 349– 369). Amsterdam: John Benjamins. doi: 10.1075/ata.xv.18hal. Le, Q. V., & Schuster, M. (2016, September 27). A neural network for machine translation, at production scale. Google AI Blog. https://ai.googleblog.com/2016/09/a-neural-network-for-machine.html. Luo, X. (2023) Asia Pacific translation and intercultural studies: ten years on. Asia Pacific Translation and Intercultural Studies. DOI: 10.1080/23306343.2023.2215525 Miller, G. A. (2003). The cognitive revolution: A historical perspective. Trends in Cognitive Sciences, 7(3), 141– 144. OpenAI. (2023). ChatGPT. https://chat.openai.com/.