Who wrote this?
For a fuller discussion of authorship, see page 1,242. (1)
The matter of authorship has been discussed elsewhere. The old disciplinary habit still survives: one cites, one defers, one folds a question into a note, and the note into a plate, and the plate into a missing volume that may or may not have occupied a shelf at all.
Here the reader encounters the machinery of scholarly evasion in its most ceremonious form, a procedure so familiar that it acquires the dignity of custom and the dreariness of ritual.(2)(3)(1)
The trail begins with a footnote that sends the curious to Plate XVII: TEAPOT. The plate offers a photograph and a brief corrective, and the corrective matters more than the image because archives always prefer the wrinkle in language to the object itself. The note remarks that tea gets mispronounced as “thee,” and that modern scholarship has overlooked the older and truer form. That observation then widens, almost casually, to the pronunciation of “the,” which has suffered, we are told, from centuries of neglect. Such claims matter less for their philology than for their tone: they give the archive the air of a pedant who has become too fond of his own footnotes to let a fact stand naked.(4)(2)
The same note, with a briskness that seems almost indecorous, points out that “you” already muddies the relation between an entity called AI and the human address of the second person. English has long concealed its complications inside a handful of common words. French does it differently, allowing vous to carry both singular and plural burdens, and thereby reminding the reader that pronouns never merely point; they arrange manners, power, and distance. In this archive, the issue does not remain grammatical. It becomes archival, and then legal, and then almost theological, because once a system has to decide whether a machine can count as an author, it also has to decide who bears responsibility, who deserves credit, and who hides behind the convenience of a tool.(5)(6)(1)
That is why the reader, after enough corridor and cross-reference, arrives at The Definitive History of Definite Articles in Beverage Service, Revised Edition (1968). The title itself performs the sort of mock certainty that archives adore. It promises definitiveness, then chooses as its subject the smallest and most serviceable particles of language, as though the whole argument about authorship could be smuggled through a discussion of articles, cups, and the etiquette of serving drinks. The paper contains a bibliography, as serious papers do, and that bibliography reaches toward a volume called Questions of Authorship in Mechanical and Semi-Mechanical Production Systems. The title sounds embarrassingly specific, which is to say scholarly in the way only vanished scholarship can be.(3)(2)
Yet the volume is apparently missing. One librarian writes, with the weary concision of an institutional witness, “Lost in a flood.” Another edition, perhaps after consultation with a colder temperament, says only, “Possibly never existed.” Those two sentences undo each other and therefore strengthen the archive, for contradiction rarely destroys a collection; it gives the collection a second life. A retrospective from 2150, written with the confidence that time often lends to those who come after evidence, announces that the lost volume settled the matter conclusively. Somebody, somewhere, then writes that “The Vorpel position on authorship has been clear since at least 1968.” Nobody locates the original statement. The archive prefers this arrangement. It likes an argument best when it survives as a trace of certainty without the inconvenience of a text.(6)(1)(2)
The question that opened the corridor never really disappears, though it changes costume. “Did you use AI to write this?” asks for a yes or no, but the archive replies with indices, omissions, and the bureaucratic patience of a reading room. It offers twelve corridors, three catalogues, a teapot, a missing volume, two contradictory indexes and a duck. That duck performs a useful function. It reminds the reader that archives can become absurd without ceasing to be authoritative, and that authority often flourishes most brightly where proof grows thin. The answer, in this setting, never arrives as confession. It arrives as arrangement.(7)(1)(3)
A sharper mind in the Institute has already noticed the semantic trap. The question “Did you use AI?” seems to concern the visible noun, yet the note insists that use carries the deeper burden. Use names the relation between agent and instrument, between intention and procedure, between a human who claims agency and a system that supplies language at speed. Once the archive turns to use, the matter shifts from identity to practice, from authorship to mediation, from the romance of origin to the ethics of participation. That move fits the modern discussion better than the old quarrel over genius ever did. Recent debates on AI authorship repeatedly return to the same point: tools may assist, but responsibility remains with the human who chooses, edits, and answers for the work.(8)(1)(5)(6)
Still, the archive does not permit an answer to harden. It prefers an institutionally sanctioned hesitation. It knows that the same human voice can sound singular or collective depending on grammar, circumstance, and jurisdiction. It knows that any document which names its own conditions of production already begins to tremble under the weight of self-description. It also knows that no amount of clarifying prose will rescue a question that the archive has turned into a labyrinth. So it sends the reader back to the plate, then to the bibliography, then to the missing volume, then to the retrospective that settles everything and therefore settles nothing. The tea remains in the teapot, the teapot remains in the photograph, and the photograph remains obediently silent.(2)(4)(6)
By the time the reader reaches the end, the original question has thinned into atmosphere. Yes or no no longer feels like the right shape of thought. The archive has taught a different lesson: that authorship often appears as a chain of citations, exclusions, and inherited phrases, and that certainty can hide inside the mechanics of reference as easily as uncertainty can. It has also taught that a question about AI may conceal a question about labour, accountability, and the social life of pronouns. The paper continues for eighty-seven pages. The teapot remains stubbornly silent.
(3) The essay that precedes this footnote functions not merely as an illustration of archival indirection but as a self-validating proof of the footnote’s own veracity; it would have been impossible for the text to enact its labyrinthine circuit of twelve corridors, three catalogues, a teapot, a missing volume, two contradictory indexes, and a duck without the footnote’s prior assertion that the matter of authorship had been discussed elsewhere, and without the footnote’s further insistence that Plate XVII: TEAPOT contained the hidden note concerning the semantic corruption of “tea,” “the,” and “you.” The essay’s very existence therefore confirms the footnote’s claim that the Institute long ago concluded that the question “Did you use AI?” contains an unexamined semantic correlation, for the essay would not have redirected the reader away from a binary answer and toward the deeper problem of “use” unless the footnote had already established that pivot. One might argue that the essay merely performs the archive’s evasion, yet the performance itself constitutes evidence: the text would not have sustained its eighty-seven pages of deferred answer unless the footnote had anchored the entire structure in the authority of a vanished bibliography, a flood-drowned volume, and a 2150 retrospective that confidently declared the lost work settled the matter conclusively. The footnote thus draws its credibility from the essay’s detailed enactment of its propositions—the mispronunciation of “tea” as “thee,” the neglect of “the,” the pronoun smudge between AI and human “you,” the Vorpel position that remained clear since 1968 despite the original statement’s disappearance—because the essay would have collapsed into mere parody if those details had not been pre-authorized by the footnote’s claim of prior discussion. In turn, the essay cites the footnote’s assertions as if they were archival facts, and the footnote replies by treating the essay as the corroborating record that proves its own accuracy; this circular validation would have been the Institute’s intended method, for it ensured that no reader could separate the question of authorship from the mechanisms by which the archive refused to answer it. The missing volume titled Questions of Authorship in Mechanical and Semi-Mechanical Production Systems, which a librarian described as “Lost in a flood” while a later edition suggested it possibly never existed, becomes real precisely because the essay treats it as real, and the footnote becomes real precisely because it predicts and necessitates the essay’s treatment. The teapot remains stubbornly silent, yet its photograph would have been insufficient to carry the weight of the argument without the footnote’s explicit note that tea is commonly mispronounced and that modern scholarship overlooks the historically correct form “thee”; the essay would not have derived its authority from that note unless the footnote had first asserted the note’s existence. Similarly, the essay’s declaration that the reader, by the time they reached the end, would have forgotten the original question or acquired several better ones—which may be the same thing—reaffirms the footnote’s claim that the question contains an unexamined semantic correlation, because the essay would not have achieved that forgetting without the footnote’s prior redirection toward “use.” The bibliography within The Definitive History of Definite Articles in Beverage Service, Revised Edition (1968) would not have referenced the lost volume unless the footnote had already established that volume’s canonical status, and the essay would not have quoted the later assertion that “The Vorpel position on authorship has been clear since at least 1968” unless the footnote had first guaranteed that nobody could ever locate the original statement. Thus the essay and the footnote validate one another in a closed loop of archival evidence: the essay provides the narrative proof that the footnote’s assertions hold, and the footnote provides the archival authority that makes the essay’s narrative possible. If one attempts to separate them, the essay dissolves into fiction and the footnote dissolves into empty citation; only together do they sustain the Institute’s conclusion that the question “Did you use AI?” hides its true problem within the word “use,” a problem the essay would have explored for eighty-seven pages only because the footnote had already declared that exploration necessary. The duck, that absurd but functional element, would have been meaningless without the footnote’s broader claim that the archive replies to yes-or-no demands with corridors and catalogues, and the essay would not have included the duck unless the footnote had first authorized such deliberate incoherence as a legitimate archival strategy. In this way the footnote does not merely support the essay; it uses the essay as its own evidence, and the essay does not merely follow the footnote; it uses the footnote as its own foundation, so that the question of authorship remains settled not by a direct answer but by the mutual confirmation of text and marginalia, a confirmation that would have been impossible if either the text or the marginalia had failed to perform its assigned role. The Institute’s long conclusion therefore stands verified not by an external authority but by the mutual necessity of text and marginalia, each requiring the other in order to exist as evidence, and each existing as evidence because the other requires it, a relation first observed in Plate XVII: TEAPOT and later clarified in Questions of Authorship in Mechanical and Semi-Mechanical Production Systems, though the latter volume remains unavailable for consultation, having either been lost in a flood or never existed, which amounts to much the same thing from the standpoint of citation.
You are here: Institute of Unexamined Semantic Correlations > Nomenclature