Send message to Author

Paper short abstract

This paper examines AI-powered translation through the linguistic theory of Tokieda Motoki (1900–1967). Drawing on Tokieda’s processual view of language, it highlights the differences between human interpretation and algorithmic language processing and reflects on the implications for area studies.

Paper long abstract

The recent advances in generative AI are poised to bring about considerable changes in the way language is understood, not just at universities, but in broader society. As fast, free and “good enough” machine translation becomes widely available, “language barriers” – at least in the form that we have come to know them – may soon become a thing of the past. Yet a world where everything is translatable is not necessarily a world where nothing is lost in translation. On the contrary, as the sheer quantity of translations is set to exponentially increase, the number of misinterpretations will likely follow suit. This paper starts from the assumption that, in order to meet the challenges of intercultural communication in the future, students and practitioners of area studies need more than just practical language competence; they also require a solid theoretical footing in how language is expressed and interpreted, particularly with regard to the social – and human – aspects of language.

It is from this perspective that the paper revisits the work of Tokieda Motoki (1900–1967), a strikingly original thinker whose work has otherwise become largely forgotten. Tokieda, who rose to prominence at the height of Japanese imperialism, came under heavy criticism in the postwar period and remains a controversial figure today, not least because of his entanglements with assimilationist policies in Japanese-occupied Korea. This paper argues, however, that the advent of AI has made Tokieda’s ideas newly relevant: not only was he the first Japanese linguist to explicitly criticise the notion that language can be studied as an abstract system separate from the humans that speak it – an idea around which modern generative AI is built – he also built his own unique “processual theory of language” (gengo katei setsu), which treats language as a fluid and reciprocal process of expression and interpretation. With the aim of stimulating theoretical debate about language and area studies in the age of algorithms, this presentation asks what Tokieda’s “three preconditions for language” – the subject (shutai), the context (bamen) and the message (sozai) – reveal about the nature of AI-powered translation.