Martha Woodmansee

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI asks the individual to see the machine clearly—to take the pill and refuse both the narcotic of hype and the paralysis of fear. Woodmansee supplies the most historically grounded version of a question the cycle circles: who is the author? When Edo Segal wrote the book with Claude and found himself unable to say where his contribution ended and the machine's began, he was experiencing what Woodmansee had already explained: the discomfort was historically provincial. A medieval scribe would have been baffled by it; a Renaissance humanist would have shrugged. Only someone who had absorbed the Romantic myth without knowing it was a myth could feel that the collaboration itself was a problem.

Her lens is essential to the cycle's treatment of felt authorship—the phenomenological experience of a product as one's own—because it separates the feeling from the legal and philosophical apparatus built on top of it. The feeling is real; the apparatus is contingent. Understanding that distinction lets the individual engage with AI as a collaborator without the guilt or the grandiosity that the Romantic framework injects into every such encounter. The question is not whether the result expresses your genius. It is whether the result is good, and whose hands shaped it, and in what proportions, and whether any of that can be recovered once the work is done.

Woodmansee also grounds the cycle's treatment of the training corpus question. The unresolved legal and ethical problem of whether AI systems may train on copyrighted works without compensation is, in her framework, a collision between two incompatible legal constructs: the Romantic authorship model that copyright law encodes, and the compilation tradition in which every text is built from prior texts. The machine does what writers have always done—gather, transform, recombine—and the law insists this is theft because it has forgotten what writing always was. Woodmansee remembered. She built the case from the archives before anyone knew an AI would need it.

The Orange Pill cycle places her among the thinkers who reveal invisible walls: the assumptions so deeply held they feel like facts of nature. Woodmansee's particular wall is the one that makes every question about AI creativity feel like a moral crisis. Remove the Romantic construction and the crisis does not disappear—the distributional questions remain real—but it changes shape. The question is no longer whether the machine can be an author in the sacred sense, but what arrangements of credit and compensation make sense when the act of writing has always been collaborative, cumulative, and indebted to a commons no single person made.

Origin

Woodmansee's origin story is the story of a question that should not have been controversial. She set out to understand why German writers in the late eighteenth century developed such an elaborate philosophical vocabulary for the concept of original genius—why the language of spirit, organic growth, and inner necessity appeared so suddenly and so insistently in aesthetic theory precisely when the patronage system was collapsing and a commercial book market was emerging. The answer she found in the archives was not philosophical. It was economic. Writers needed legal protection for their texts because the market required property rights; the concept of original genius gave them a philosophical foundation for those rights. The metaphysics of creativity was, at bottom, an argument about who gets paid.

Her 1984 essay, published in Eighteenth-Century Studies, established the argument with a precision that professional historians respected. The 1992 essay extended it into the collaborative modes of literary production that the genius myth had displaced: the encyclopedic compilation, the editorial collective, the commonplace book. With Peter Jaszi she brought legal scholars into the same inquiry, producing in The Construction of Authorship (1994) a foundational interdisciplinary volume that showed copyright law and literary theory had been built on the same invented premise and were beginning, together, to crack under the weight of contemporary cultural production.

The work was largely invisible to the general public for twenty years, read by literary scholars and intellectual property lawyers but rarely reaching beyond the academy. Then the large language models arrived, and suddenly her question—what is a text, who made it, and to what extent does the maker own it—was the most contested question in the culture. Woodmansee had spent four decades building the historical case that the answers everyone assumed were obvious had been invented, and invented recently, and for reasons that had nothing to do with the nature of creativity. The AI moment made the case urgent in a way its author had not anticipated and its subject had long required.

Key Ideas

The Romantic authorship construct. The defining move of Woodmansee's scholarship is to treat the concept of individual authorship not as a discovery about creativity but as a historical construction—invented under specific pressures, naturalized so thoroughly that the construction became invisible. The writer as sole originator of a unique expression, and therefore the natural owner of the text, was not always how Western culture understood writing. It was built in the late eighteenth century, and it will have to be rebuilt again now that AI has made its premises untenable.

The compilation tradition. Before the Romantic myth, Western culture honored a different mode: the compiler, the encyclopedist, the arranger of inherited material. The bee rather than the fountain. The compilation tradition was not a failure of originality but a recognized form of intellectual labor—selecting, organizing, and presenting what already existed, with judgment rather than conjuring. AI's method of generating text from prior text is, in Woodmansee's framing, closer to the medieval compiler than to the Romantic genius, which does not make it unproblematic but does change what the problems are.

The training corpus as commons question. Copyright law's reliance on Romantic authorship concepts makes it inadequate to describe what AI training does, because training treats the accumulated output of human writing as a commons from which new production can be drawn. This is, Woodmansee would observe, precisely what writing has always done—drawn on a commons of prior language, prior thought, prior form. The training corpus question is the legal system catching up to the historical fact that the Romantic myth concealed.

Felt authorship versus legal authorship. Woodmansee's most practically important distinction is between the phenomenological experience of creative ownership—the felt sense that this came from me—and the legal-philosophical apparatus constructed on top of that feeling. The feeling is real and worth respecting. The apparatus is contingent and increasingly mismatched to the reality of how texts are produced. Felt authorship can survive the end of Romantic authorship as a legal doctrine; the two need not rise and fall together.

Post-Romantic creative practice. The practical upshot of Woodmansee's scholarship is not nihilism about authorship but a more honest account of what creative work is and has been: cumulative, collaborative, indebted, and shaped by predecessors no individual genius conjured from nothing. A post-Romantic framework asks not whether the work is original in the sacred sense but what the maker brought to the encounter with materials, languages, traditions, and—now—machines. The question is one of craft and contribution, not of originary purity.