While Sea of Intimacy proceeds — I expect to have a full draft by later this summer — I seem to have another project developing. Pulling together some of the threads on this site, and looking at other pieces here and here, helped me to see it. At this point, my working title is The Translator After AI: Creative Understanding in an Age of Prediction. It feels like a book, rather than a collection of articles, so that’s how I’m approaching it.
It begins from the premise that translation is not a problem of equivalence but a practice of creative understanding grounded in judgment. That (the creative understanding part) is a notion I picked up from Mikhail Bakhtin in the early 2000s and wrote about in a POROI word tour, “The Oxymoron of Empathic Criticism.” While contemporary AI systems can produce fluent and often persuasive translations, they do so by generating probable language rather than by recognizing what kind of problem a particular act of translation presents.
Through a series of examples, from literary texts and historical materials to classroom contexts, the book explores the broader consequences of automated translation, both in the kinds of outputs it produces and in the way it re-frames translation itself, as a practice central to cross-cultural communication and creative thinking.
My favorite part of the project, which seems to be coming out of me faster than I can keep up, is taking up specific, frequently encountered dimensions of translation and pushing them up next to what AI does to see what happens. Each one turns out to be a bit of a thought experiment — about judgment, fidelity, fluency, and other translation commonplaces. The experiments test how translation operates under different forms of pressure and how these are imperfectly captured by systems trained on large-scale textual data.
I seem to have written some of this before, just with other ends in mind. A Writers Chronicle essay of mine from a decade ago (“Cross-dressing for Workshop,” — apologies, it’s behind a paywall) now has a different kind of relevance. So I’m mining some of my own past work. Also my own present work, as the latter parts of Sea of Intimacy turn out to have a lot about the very human (corporeal and ethical) aspects of translation in them.
While I have little to say about AI there, here the final chapter extends the argument by re-framing translation as a broader condition: not only something we do to texts but something that happens to texts, practices, and persons as they are carried “across” into new ways of meaning.
###
Russell Scott Valentino 