In another moment of disorienting technological change, we are learning how to read and write again. Against dire warnings about the decline of ‘reading for pleasure’ in conventional literary forms and publishing formats, we are hearing about a surge in reading materials produced by ‘artificial intelligence’ that get served as factual information and compelling imitations of human appearance and speech.  

‘Hallucinated’ facts produced by the quick calculating machines often go unrecognized as artifacts of the writing process by their human readers or the machines themselves. Little surprise there: we have long been like the machines that learn only from other writing. We have used the same writing instruments to record experience and to construct fictional characters; to describe known and invented worlds. We rendered the psychological depths and transformations of invented characters using the same tools as when we presented previously invisible historical subjects (working-class women, queer people, sex workers, racialized migrants etc.) to the modern readers’ historical and social imagination (Steedman; Hartman).  

Current dread of the existence of realism for which there is no matching reality invites questions about the confidence of earlier ages in their ability to discern the qualities of writing that could be read and disseminated as ‘evidence of experience’ (Scott). If affinities between fiction and (auto)biography or historiography were not always treated as alarming ‘hallucinations,’ what about the current technological change—especially the emergence of Large Language Models—is messing with the old idea that realism is a compelling form of writing about experience that may have never been had by anyone?  

This seminar invites reflections on the confluence or distinctiveness of tastes for realism in various literary cultures, linguistic traditions, and interpretations of literacy and reading ability, across historical time spans and in the present. Because historical trajectories of European empires still mirror canonical accounts of realism’s emergence as the dominant form of written representation, what have been the shared styles of authenticity admissible as evidence of experience under specific economic and social conditions? What have been the realistic descriptions of unfamiliar places and distant peoples for empires that wanted to use this writing as a tool of governance? Are realisms just acquired tastes of their reading publics? Who are the unfamiliar people and what are important distant places in the post-empires of enormous data sets and multipolar globes? Few human languages now supply much of Large Language Models: How do they synthesize descriptions of the world, and how do we learn to read what they serve as evidence of experience? Where do we now learn how to do realism of subjective empirical experience?