Technologies of Literature: Reading, Judgment, and the Large Language Model

Abstract

This essay explores reading practices under the condition of technological change, with a particular emphasis on the emergence of generative AI and Large Language Models (LLMs). The argument unfolds in three sections. First, we canvas the history of the relationship between technology, poetry, and literary criticism in the early twentieth century. Criticism is a metalanguage that scaffolds the institution of literary production, bringing it into being. By briefly exploring early twentieth-century critical technologies that prioritised the quantitative and the statistical over the epistemic singularity that is usually characteristic of close reading, we show how disciplinary conversations surrounding generative AI, LLMs, and the institution of literature have much longer – and far more interconnected – histories than we might imagine. These histories help us determine how new LLM technologies really are for the discipline of literary studies and what it means to understand them as part of a longer, embedded intellectual lineage. We then consider the relation between language and subjectivity, bringing the concept of fictionality (and fictional beings) to bear on the study of current-state LLMs. Here, we argue for the relevance of literary critical concepts to the critique of LLMs, with a focus on how judgments about the efficacy, veracity, and utility of machine generated material take place and are subsumed into a broader techno-philosophical field that has always been invested in what it means to judge the linguistic, semantic, and literary value of texts. Our final section considers the ways that emerging LLM technologies are reformulating the practice of reading and the literary critical function of judgment.