This seminar examines—and contextualizes—how large language models (LLMs) are transforming the acts of reading, interpreting, and imagining that have long defined literature and literary theory. We will approach our topic through two interrelated lines of inquiry.

First, we ask how the literary imagination—as an episteme—might expand from embodied to machinic, from cognitive to distributed. By revisiting the cognitive and neuroscientific foundations of metaphor, narrative, reading, and creative thought, alongside phenomenologies and cross-linguistic histories of imagination, we will explore new ways of thinking about creativity, metaphor, and world-construction in the human mind and beyond.

Second, we will turn to the concept of world literature as a lens for understanding the ontologies and epistemologies of LLMs themselves. Generative models such as ChatGPT, Gemini, and DeepSeek have “read” more literature, in more languages, than any human ever will, and these models constantly compare texts, even if their methods for doing so remain opaque. We know from centuries of literary study that comparison is never neutral—it always involves choices and relations of power. What strategies of prompting and reading might illuminate the comparatist within the black box?

By bringing these two sets of questions together, the seminar initiates a conversation about how the convergence of cognition and computation are reshaping reading, interpretation, and creativity in a multilingual or global context. At the same time, we will foreground the value of literary approaches to critically studying technology. By insisting that LLMs are neither benign nor unprecedented, we will collectively experiment in producing engaged scholarship that resists the imperatives of industry-led hype.