Hannes Bajohr - Making Worlds in Novels and LLMs

Why It Matters

Recognizing LLMs' partial world modeling reshapes authorship, literary criticism, and the future design of coherent AI narratives.

Key Takeaways

• LLMs generate locally coherent narratives but lack global world consistency.
• Cross‑disciplinary methods reveal structural parallels between novels and AI models.
• Fine‑tuned GPT‑NeoX produced a surrealist novel illustrating weak worldness.
• Human‑AI co‑authorship raises questions of authorship and style attribution.
• AI as an executable theory forces humanities to clarify language assumptions.

Summary

The talk by Hannes Bajohr explores how large language models (LLMs) and novels both construct "worlds" through sequential text generation. He begins by referencing recent research that treats navigation in Manhattan as a deterministic finite automaton, showing that LLMs can learn local structure yet produce globally incoherent maps. Extending this analogy to narrative, he asks whether LLM‑generated stories encode a comparable internal world model.

Bajohr argues that interdisciplinary analysis—combining computer science, literary theory, and philosophy—uncovers shared structural challenges. He outlines his own practice of fine‑tuning an open‑source model on German novels to co‑author "Berlin, Miami," a work that, while stylistically intriguing, displays persistent disjointedness. This empirical case illustrates that LLMs can produce texts that hang together enough to be read as novels, even though their underlying world representation remains fragmented.

He draws on Hans Blumenberg’s philosophy to define the novel as a modern, relational model of reality, emphasizing structure over content. By juxtaposing this definition with AI’s statistical language generation, Bajohr highlights how AI forces scholars to make explicit the assumptions about meaning, style, and coherence embedded in both human and machine narratives. The resulting dialogue suggests a nascent "artificial humanities" where literary concepts refine AI models and vice versa.

The implications are twofold: creators must navigate new forms of co‑authorship and attribution, while scholars gain a novel lens to critique and improve LLMs’ narrative capacities. Understanding the limits of LLM world‑building informs both the development of more coherent generative systems and the cultural reception of AI‑augmented literature.

Original Description

Biography:

Hannes Bajohr is Assistant Professor of German at the University of California, Berkeley. He studied philosophy, German literature, and modern history at Humboldt University, Berlin, and New York University and received his Ph.D. from Columbia University in Germanic Languages and Literatures. His research focuses on media studies, political philosophy, philosophical anthropology, and theories of the digital. Recent publications include: Thinking with AI: Machine Learning the Humanities (as editor, London: Open Humanities Press) and "Surface Reading LLMs: Synthetic Text and its Styles" (arXiv preprint, forthcoming in New German Critique). In 2027, the English-language translation of his LLM-co-generated novel (Berlin, Miami) will appear with MIT Press.

Abstract:

In what sense are the “worlds” of novels and of AI analogous, and what can each illuminate about the other? This talk argues that both novels and large language models (LLMs) construct worlds as networks of relations—patterns of events, inferences, and expectations—rather than as fully grounded causal systems.

Across realism, genre fiction, and modernism, literature offers distinct ways of organizing such relations, from strong causal-social dynamics to fragmented but coherent world-logics. LLMs, by contrast, generate narratives from latent vector spaces that encode statistical regularities without explicit grounding. The result can be texts with stylistic unity but limited causal depth—closer to a “weak force” of association than to robust narrative causality. By comparing these two modes of world-making, the talk proposes a framework how literary scholarship might shed light on the structure of AI-generated fiction that avoids both overclaiming understanding and dismissing it as mere mimicry.

HTF Colloquium

Wednesday April 8, 2026

290 HMMB

3- 4p