Paper Knowledge
Paper Knowledge is Gitelman's second major monograph, following Always Already New. Across four case studies — the job printing trade of the nineteenth century, the photocopier in mid-twentieth-century office culture, the PDF as a document format, and the database as an organizational technology — Gitelman developed the argument that documents are not neutral containers for information but artifacts whose material properties, institutional contexts, and formal conventions actively shape what counts as knowledge. The book's theoretical payoff is the concept of the epistemic object, which turns out to have unexpected purchase in the analysis of AI-generated text. The format of a document carries epistemological implications inherited from the institutional frameworks in which the format developed, and when a new medium borrows old formats, the inherited implications can become misleading.
In the AI Story
The book is Gitelman's most methodologically rigorous application of media archaeology to the document. Its four case studies were chosen to demonstrate that the document is not a single stable object across the history of print and digital culture but a shifting assemblage of material, institutional, and formal features that participate actively in the knowledge they appear merely to record.
The PDF chapter is particularly resonant for AI-era discourse. Gitelman analyzes the PDF as a format designed to preserve the visual appearance of a document across platforms while making the document more difficult to edit, search, and repurpose. The format encodes specific institutional priorities — fidelity to original appearance, resistance to modification — that are not neutral features of the technology but reflect the commercial and institutional interests of Adobe and the organizations that adopted the format.
Applied to AI, the book's framework treats every AI-generated text as an epistemic object whose format carries inherited institutional guarantees. The format of polished academic prose implies rigorous scholarship. The format of the executive summary implies executive judgment. The format of the technical documentation implies technical expertise. When the format is produced by statistical pattern-matching rather than by the processes the format conventionally implies, the inherited guarantees become misleading.
The book's method — treating documents as active participants in knowledge production rather than as passive records of it — provides the analytical template this volume extends to AI.
Origin
The book developed from Gitelman's continuing research at New York University and from conversations within the emerging field of critical data studies. It was published in Duke University Press's Sign, Storage, Transmission series, edited by Gitelman and Jonathan Sterne.
Key Ideas
Documents are epistemic objects. They actively participate in the production of knowledge rather than passively recording it.
Format is constitutive. The structural conventions of a document type shape what can be claimed and how claims are evaluated.
Four case studies. Job printing, the photocopier, the PDF, and the database each demonstrate the active role of the document in institutional life.
Institutional interests. Document formats encode the interests of the institutions that adopted them, not the neutral properties of the technology.
AI inherits the framework. The book's analytical vocabulary extends to AI-generated text with minimal modification — the format persists, the process changes, the inherited guarantees become misleading.
Appears in the Orange Pill Cycle
Further reading
- Lisa Gitelman, Paper Knowledge: Toward a Media History of Documents (Duke University Press, 2014).
- Matthew S. Hull, Government of Paper: The Materiality of Bureaucracy in Urban Pakistan (University of California Press, 2012).
- Ben Kafka, The Demon of Writing (Zone Books, 2012).