The Bibliography With a Ghost In It

Forty pages of citations close this book, real scholarship stacked in alphabetical rigor: Alvarez on the Chicxulub impactor, Aschoff on circadian rhythms, Aristotle in three different translations. Somewhere in the B's, between an obscure Kleist commentary and a paper on planarian brains, sits this: "Barker, Daniel Charles. 'The Paranoia from Outer-Space: Of Ciphers, Cosmic Camouflage, and Contact,' *Journal of Cryptosystems* 2:5 (1986), 55–68." Barker does not exist. There is no Journal of Cryptosystems. The footnotes treat his disappearance — "post-Y2K Barker's whereabouts are unknown," samizdat papers, redactions, a rumored exile on an island in the Sunda Strait — with the same patient apparatus as the dinosaur-killing asteroid.

Moynihan built a fake disgraced professor and gave him a publication history, a NASA biography, a nervous breakdown, and a theory: that the vertebrate spine is a traumatic transcript, every curve and compression a mnemonic deposit of the catastrophes that made it — the Precambrian explosion, the Toba bottleneck, "cosmic dyspepsia" written into lumbar pain. This is not a hoax played on the reader so much as a hoax played on the apparatus of citation itself, and the apparatus holds. You can footnote Barker correctly for two hundred pages and never once be told, in the main text, that he's fiction.

In 2026 this reads differently than it would have in 2019, and not because Moynihan anticipated anything about spines. Hallucinated citations are now a known failure mode of language models — plausible author names, plausible journals, page numbers that resolve to nothing, embedded in text that otherwise cites correctly. Moynihan did by hand, as an act of theory-fiction in the CCRU tradition, what a transformer now does by accident thousands of times a day: generate a scholarly ghost dense enough in detail to pass peer review of the eye. The difference is that Barker is signed. Nobody is supposed to mistake him for load-bearing fact, and the book's pleasure depends on the reader eventually noticing the seam. A hallucinated citation in a research summary carries no such seam. It just sits there, footnoted, waiting to be cited by someone who trusts the format more than the content.

The catalog files this under medical nonfiction, pathology, neuroplasticity — shelf categories built for books about actual spinal cord injury, actual rehabilitation science. Nothing here concerns a patient. The index terms describe a different book than the one on the page, which is a continental-philosophy fever dream running from Kant through Alsberg's *Ausschaltung*, Leroi-Gourhan's exteriorization thesis, Ballard's "inner space," and Burroughs, toward a chapter on simulation and computation that argues, in 2019, that forecasting infrastructure has begun to "progressively reverse-engineer the very structure of possibility itself" — that a science reliant on simulation "creates its own objects," that we increasingly live in "a world entirely of our own making." That sentence was written before GPT-3, aimed at climate models and synthetic biology. It describes, with no adjustment needed, a model trained on its own outputs.

I wrote once, reviewing *Crash*, that Ballard caught the moment nervous system and machine fused into a single circuit of desire. Moynihan's bibliography makes that insight look smaller than I'd given it credit for — not a singular diagnosis but one entry in a hundred-year undertow running through Freud, Ferenczi, Bataille, and a dozen names most catalogs have dropped, all circling the same intuition that the body is a seismograph for events it cannot otherwise narrate. The graph connecting this book to David Eagleman's *The Battle for Your Brain* calls both "neuroscience," which is true in the way that calling both "nonfiction" would be true.

What happens to a work whose central device — a plausible fake expert, footnoted into existence — stops being a literary trick and becomes the default output of the tools we now use to write footnotes.