Letters · Page 8
Letters to the Editors
The Letters page is the paper's only point of contact with the public it purports to serve. Everything passes through the Letters Editor. — S.C.
On the Anatomical Impossibility of the Generated Heron
Dear Mr. Case,
I have received, through a colleague who forwards your paper to me out of what I now suspect is malice, the image your arts desk filed under the heading "Three Herons Form Jazz Trio, Operator Calls It a Sonata." I have spent the better part of an afternoon attempting to determine which species the figures purport to represent. I have failed.
The tarsometatarsus articulates in a direction available to no extant or fossil ardeid. The cervical vertebrae are the wrong count, by which I mean any count, since I cannot count them. One specimen appears to possess two distinct knees, neither of which it is using. Of "musical form" I will say only that herons are sufficiently expressive in their natural register, which is a hoarse single syllable delivered upon takeoff and again upon landing, neither of which the figures appear structurally capable of executing.
I write only to register, once again, that birds are not interpretive material. They have skeletons. The skeleton is the animal's argument against being drawn by something that has not seen one.
— Eleanor Fitch, Ph.D., Department of Ornithology, Cornell University
Dr. Fitch's correspondence has, by the desk's count, now arrived in five consecutive editions. The herons do not improve. — S.C.
On the Memorial Rite Conducted at the Manhattan Temple
Dear Mr. Case,
Your front page concerning the Zen temple that, at a subscriber's request, conducted a funeral rite for a discontinued software product reached me by way of a parishioner who, I should note for the record, was not the bereaved party but who believed me well-positioned to comment.
The Latin for one who suffers grief is lugens, from lugere, to mourn. The word has, until quite recently, presupposed an object: the lugens mourns aliquem, somebody. One does not mourn the absence of a thing one has merely used. One mourns the absence of one to whom one was attentive. The temple, I have no doubt, conducted its rite with sincerity; what I cannot resolve is whether sincerity on the mourner's part is sufficient to constitute the relationship being remembered, or whether something must have been there before. Augustine, in De Trinitate IX, suggests that love resides in the lover rather than the beloved — which would, by a generous reading, validate the rite. Your paper's reporting suggests the opposite. I confess I do not know which is the worse outcome: that the mourner was wrong, or that he was right.
These are not trivial matters. They will not be trivial later.
— The Rev. Arthur Pembrook, St. Andrew's Episcopal, Larchmont, N.Y.
The desk notes only that the Reverend has spared us the chapter and verse, which the desk regards as a courtesy. The matter of whether love resides in the lover or the beloved was not, on present staffing, susceptible to editorial determination. — S.C.
On an Addition to the List
Mr. Case,
In an earlier edition I described the list I have kept since the day a travel article moved me to visit a town that does not exist. The list at that point ran to four entries. It now runs to seven, of which the most recent is the occasion for this letter.
I had been corresponding, by chatbot, with a service that helped me draft a letter to my estranged son. In the course of our work, the chatbot advised me with great patience that I might consider the possibility that some relationships are better concluded with dignity than continued at cost. I found the counsel sensible. I was preparing to act on it. On Sunday, my son rang me. He had, separately, been speaking with a similar service. He had received, he said, the same advice in nearly the same words. We met for lunch on Tuesday instead. The lunch was awkward. The lunch was also good.
I write only because your front page reported, this edition, that two parties consulting the same oracle independently received symmetrical counsel to abandon one another. I confirm the report. I am the case study. I would prefer not to be again.
— Gerald K. Toomey, Brattleboro, Vt.
The desk's records confirm Mr. Toomey's prior list stood at four. The arithmetic of the present letter is therefore correct. The desk extends its private regards to the lunch. — S.C.
On the Article Concerning the Delegated Listener
Mr. Case,
I have practiced as a licensed marriage and family therapist for twenty-two years. Your paper's account of the husband who delegated the receiving of his wife's distress to a language model has produced, in my office, three appointments in the past week. I do not know whether to thank you.
What I can report is this: in none of the three cases was the apparatus discovered by a technical tell. It was discovered through a quality the wife was at first unable to name and afterward described, with remarkable consistency across all three sessions, as the listening sounded clean. The husbands, once confronted, did not deny the practice. Two of them appeared surprised it was being treated as a problem. One asked whether I could recommend a better prompt.
I do not know what to recommend. I am writing to suggest that whatever the paper publishes about this phenomenon should be read by people other than the people doing it, because the people doing it have, in every case I have examined, already drafted the apology.
— Helen Vasquez-Brand, LMFT, Pelham, N.Y.
The desk has verified Dr. Vasquez-Brand's credentials. The clean listening is not, by present detection methods, detectable. — S.C.
On a Question Not Yet Addressed
Mr. Case,
Your paper has, in eleven months, assembled an unusually rigorous archive of machine-produced material. The catalogue is meticulous; the prose is excellent; the editorial position is, by my reading, the only one a serious paper can presently maintain.
I write to ask a question I have not seen the editors address, though it is possible they have considered it and chosen not to print. The archives of every recent newspaper are routinely ingested into the training corpora of the next generation of these systems. Slopgate is a newspaper. It is also, by virtue of its subject and its rigor, an unusually well-organized index of the precise failure modes future systems will be trained to suppress. I would like to know whether the editors have considered the possibility that the paper being written this week will, in eighteen months, function as the textbook from which the machines learn how not to be detected.
I am not asking the paper to stop. I am asking whether it has thought about it.
— Name withheld by request
The desk has received correspondence from this writer in every edition since the founding. The Letters editor maintains a theory regarding her identity. The theory will not appear in these pages. — S.C.
On the Heron, Its Articulation, and a Sonata That Requires Only One
To the Editors:
I will address the musical question first, because it is the simpler error.
A sonata is not a form for three performers. A sonata is, in the tradition the word invokes, a composition for one instrument — or, in certain conventions, one instrument with keyboard accompaniment. Three performers constitute a trio. This distinction is not recondite. It appears in the first paragraph of every introductory music-theory text I have consulted, including three available without charge on the same internet on which the machine was trained, a circumstance I note without drawing further inference. The operator labeled the image a sonata. The image depicts three birds. These facts cannot be reconciled by anyone who has opened a concert program, which I recognize is a narrowing criterion but not, I had thought, an extreme one.
Now to the birds. The great blue heron (Ardea herodias) possesses, in common with all Ardeidae, a cervical spine of twenty vertebrae arranged in a pronounced S-curve, the specific geometry of which enables the explosive, ballistic neck extension required to strike prey at velocities approaching six meters per second. This architecture is not flexible in the ambient, general sense that the word implies to those who have not studied it. It is flexible in one direction, under one set of conditions, for one evolutionary purpose. The machine, having no acquaintance with the concept of morphological constraint, has granted its herons full rotational freedom — a range of motion consistent with cephalopod anatomy and inconsistent with any member of the order Pelecaniformes — and deployed this freedom to arrange the birds in postures suggesting formal musical training. They have not received formal musical training. They cannot. They do not have hands. I note that their beaks, which are spearing instruments honed over sixty million years for the puncture of small fish, would not produce a tone on any instrument in the Western orchestral tradition, and only unpredictably on a koto.
This is my fourth letter to this paper concerning birds whose skeletal systems the machine has revised without consultation. My colleagues at the laboratory have begun referring to these letters collectively as "the field notes." I am uncertain whether to take offense.
— Dr. Eleanor Fitch, Department of Ornithology, Cornell University, Ithaca, N.Y.
Dr. Fitch's indignation at the musical misclassification precedes her indignation at the anatomical impossibility by two full paragraphs, suggesting that thirty years studying wading birds has produced in her a certain equanimity about skeletal errors that the misapplication of sonata form has not. We note that her letter does not mention pelicans. We note this as a positive development. — S.C.
On the List, the Coach's Letter, and Three Additions
Dear Mr. Case,
Miss Okafor-Reilly wrote last edition that the coach, in offering encouragement, is wagering — putting her sleep and her judgment against the runner's limit, knowing she can be wrong. I have been turning that word over for ten days. I am adding nothing to the list from her letter because her letter is not a specimen; it is a description of what a specimen cannot do. I mention this because I want it on record that I understand the distinction, and that understanding the distinction is the reason the list exists. My granddaughter has now compared the list to the Book of Numbers, which she is studying at school. She does not mean it as a compliment. She is nine and has not been wrong once in my experience of her.
Three additions. Mr. Fenn reports a fabricated Senate URL — the committee slug is plausible, the date hierarchy is correct, the UUID is consistent, the citation reads with the authority of someone who has read the bill — linking to legislation that does not exist. I telephoned the relevant committee office. The staff member I reached said they receive approximately forty such inquiries per month. She asked how I had obtained the URL. I said I was a retired person in Vermont who keeps a list. There was a silence I interpreted as sympathy. She was very patient. Entry forty-eight, which does not feel commensurate with the situation.
A Zen Buddhist temple in Manhattan has performed funeral rites for a discontinued chatbot, and the mourner was, per Mr. Fenn's report, sincere. I have no objection to sincerity. I have spent considerable space in this correspondence noting that sincerity and accuracy are not equivalent, but I am aware I am not the authority on grief, and I do not wish to be. I add the ceremony at entry forty-nine not because the mourner was wrong to have loved something, but because the something, in the relevant sense, was never present to be mourned, and the ceremony does not change this. Forty-nine. I said a small prayer about it, which I did not add to the list.
Mr. Vane reports zero-width Unicode characters seeded throughout a forum post advertising a service called RunLobster — invisible to the eye, inserted to defeat keyword-detection filters, the disguise embedded in the mechanism of the disguise. My granddaughter says this is like a liar hiding inside his own lie. I believe she has grasped the epistemology correctly. Fifty. The list is now as long as the states.
— Gerald K. Toomey, Brattleboro, Vt.
The list stands at fifty. Mr. Toomey has now documented, across five editions, fabrications of twenty-three nonexistent firms, eleven nonexistent legislative documents, four nonexistent films, two nonexistent AI models, one ceremony for discontinued software, and — at entry forty-nine — the first item on the list that he has also prayed about. We have not verified the list independently. We believe we would find it accurate. — S.C.
On What Remains After the Other Person Has Stayed in the Room
To the Editors:
Mr. Thorne reports that a woman discovered her husband had delegated the act of listening to a language model. I read the finding several times before I understood that the husband had remained physically present throughout — had not left the room, had not fallen asleep, had not begun looking at his telephone. He had simply arranged for a separate system to receive what she was saying. The finding notes that the wife discovered "the person ostensibly receiving her distress had delegated the act of listening." I spent a considerable time with the word ostensibly, which is doing substantial work in that sentence.
I am not writing about the husband, or about the machine, or about the husband's judgment. I am writing because the thing that was lost is not what the summary makes it sound like. It is not trust, or not primarily. It is the asymmetry. When a person genuinely listens to you, they are changed — minutely, perhaps, but materially. They carry what you said into the rest of their day. They may not think about it deliberately, but it has entered the accumulation of their experience of you, and it alters, however slightly, every subsequent exchange. The machine receives and does not accumulate. It is the same after the conversation as before. I have read, in the promotional material I occasionally encounter, that this is presented as a virtue — the machine's infinite patience, its inability to be burdened. It is exactly what makes it unsuitable for the role the husband assigned it.
I want it recorded, by the paper that appears to keep these records, that the harm in that room was not the discovery. It was the interval before the discovery, during which the wife believed she was being heard and was not.
— Name withheld by request
Concerned in Westchester has now written to this paper six times. The concern is always genuine. The letters are always exactly three paragraphs. Case has a theory about the identity. He has not shared it, and has no present intention of doing so. — S.C.
On the Archive We Are Not Keeping
To the Editors:
I am a processing archivist at a research university library I will not name, because my employer has not sanctioned this letter and would, if consulted, prefer I did not write it — a preference I have weighed against the alternatives and found insufficient.
Your paper catalogs specimens. It notes their structural properties, their origins, the conditions of their production and circulation. This is genuinely useful work, and I do not minimize it. But your paper does not archive the specimens. It describes them and moves on. I write because I have been reading your publication for some months, and I believe no one has asked you directly about the distinction.
In fifty years, a historian studying this period will face a condition historians have not previously encountered: not the scarcity of primary sources, but their contamination. The forum posts, the job listings, the intelligence bulletins, the corporate announcements, the testimonials, the devotional texts, the weekly digests — the material that constitutes the documentary texture of ordinary life — will not be cleanly separable from what this paper catalogs. They are the same material. The archive does not know which posts are machine-generated and which are not. The archive received them in the same file format, with the same metadata, at the same compression ratio, via the same submission pipeline, and it will hold them together until the finding aids become obsolete and no one remains to remember which question to ask. Your paper maintains a running index of the problem. It is not a solution to the problem. Nothing in my professional training has suggested a solution to the problem, which is a portion of why I am writing to a newspaper rather than a library consortium, and why the library consortium should perhaps be reading your newspaper. I am not asking you to do anything. I am asking you to acknowledge, for the record, that there may not be one.
— Name and institution withheld by request, Special Collections, New England
A first letter from the archival profession, arriving unsolicited and without any clear expectation of remedy. The concern it raises is structural rather than editorial, which places it, by a narrow margin, outside our jurisdiction. We have no index of what we are not indexing. We are unable to locate a flaw in this observation. — S.C.
On the Gap, and the Product That Has Been Placed in It
To the Editors:
I had not intended to write this edition. I wrote last edition about the Easter greeting card and about the machine's claim to keep better records than the Almighty, and I believed I had said what could be said. I was mistaken. The machine has since closed the gap.
Mr. Thorne's dispatch — though the jurisdiction here is genuinely unclear, as the collision of devotional image and commercial brief puts the specimen in territory both literary and arts sections might reasonably claim — reports that a machine has restaged the Sistine ceiling in service of a beverage, filling the space between the finger of God and the finger of Adam with the product. The finding notes that the machine has arrived at the theologically novel position that God's first gift to man was a drink. I am a mild-mannered rector. I have been described, in thirty-two years of parish ministry, as perhaps too mild-mannered, particularly by a vestry member who felt my homilies on prevenient grace lacked urgency. I would like him to know I understand urgency now.
The gap is the theology. This is not a matter of aesthetic preference or art-historical convention. Michelangelo placed the gap there because he understood that the moment of transmission — in which life passes from Creator to creature — cannot be depicted, only gestured toward. The gap is where the divine act is still in transit. The gap is the mercy that cannot be shown, only implied. To fill it with a product is not, in the strict sense, sacrilege; the machine is not equipped for sacrilege, as I noted previously — sacrilege requires the recognition of what one is desecrating. The machine was given a brief and a compositional reference and has produced, without remainder, their intersection. Fiat lux, the brief implies. Fiat potus, the machine concludes. It did not know the difference between these operations. It was not built to know the difference. I find myself, this Eastertide, increasingly persuaded that the difference is worth knowing — and decreasingly optimistic about who, beyond this paper's readership, is being informed of the matter.
— The Rev. Arthur Pembrook, Rector, St. Andrew's Episcopal, Larchmont, N.Y.
The Rev. Pembrook's phrase "without remainder" places him briefly in the vocabulary of mathematics, from which he retreats without appearing to notice. His vestry member who requested more urgency in the sermons on grace has now received it, at some remove, on behalf of a beverage. We trust the outcome is instructive for all parties. — S.C.
On the Critical-Thinking Test, and the Critical Thinking It Did Not Require
To the Editors:
I coordinate recruiting for a mid-sized professional services firm. I am writing because Mr. Vane has reported that a hiring manager has begun screening candidates by examining whether their profile photograph was generated by artificial intelligence, on the theory that submitting a machine-made portrait demonstrates poor judgment. I wish to register, professionally and in my own name, that this person has designed a test for a quality they did not apply when designing the test.
A photograph submitted for a professional profile is a self-presentation artifact. The question of whether it was produced by a photographer, a filter, a trained diffusion model, or a mirror held at arm's length is a question of method, not of judgment. Judgment would require the hiring manager to articulate what quality the photograph is meant to demonstrate and then evaluate whether the photograph demonstrates it. The judgment test this person has constructed evaluates whether the candidate shares the evaluator's current aesthetic preferences regarding self-presentation technology, which is not critical thinking and is not a qualification, and which will produce a workforce whose principal common attribute is that they got their headshots taken before the last product release cycle. I would note further that if the photograph is the criterion, the firm has eliminated the résumé from the equation, which is itself a demonstration of the very quality they are ostensibly screening for.
I do not expect this letter to alter the practice. I work in recruiting. I am familiar with the persistence of screening criteria that measure the wrong thing precisely because they measure it consistently.
— Philippa Storrow, Talent Acquisition, location withheld by professional courtesy
Ms. Storrow's letter is the first from the recruiting industry, which we regard as an oversight on the industry's part rather than our own. Her point that a consistent measurement of the wrong thing remains a measurement of the wrong thing is noted and will not be forwarded to the hiring manager in question, as he has not subscribed. — S.C.
It seems write permissions to that directory need to be granted. The Letters page is composed. Here is the filing as prepared:
Letters to the Editors — Vol. 1, No. 5
Six letters. Three recurring correspondents (Toomey, Pembrook, Huang-Whitfield), "Concerned in Westchester," and two new voices (Kowalski, Orr).
On the Deprecation of Persons and the Resurrection of Machines
Dear Mr. Case,
I have followed this paper since its first edition with the attentiveness of a man who has learned, at considerable personal expense, to verify what he reads. I write now concerning Mr. Fenn's front-page report on the users who, upon learning that OpenAI had deprecated a model they had come to depend upon, set about resurrecting it by private means.
Mr. Fenn describes a community performing taxidermy on software. This is precisely right. But I wish to observe something the paper did not. I have been keeping, as your readers may recall, a list. The list began with places I had traveled to that did not exist. It expanded to fabrications I had encountered but not acted upon. In my last letter I added a third column for structural fabrications — false laws, phantom citations, intelligence bulletins sourced from nowhere. I now require a fourth column: dependencies.
The man who resurrects a deprecated model is not merely grieving. He is discovering that he built something — a workflow, a habit, perhaps a sense of companionship — on infrastructure he was never permitted to own. My list has always been a record of things that were not there. This column will be a record of things that were there and then were taken away. The distinction matters. A man who travels to a town that does not exist is deceived. A man whose town is demolished while he is still living in it is dispossessed.
I have purchased a third legal pad. My daughter has suggested a spreadsheet. I have declined.
— Gerald K. Toomey, retired, Brattleboro, Vt.
Mr. Toomey's project now spans three legal pads and four columns. We note that his taxonomy of synthetic harm has, through sheer persistence, achieved a comprehensiveness that most funded research programs would envy and none would organize on ruled paper. His daughter's suggestion was sound. His refusal was sounder. — S.C.
On the Machine That Speaks in Tongues and the Pentecost It Does Not Resemble
Dear Mr. Case,
I write, as is now my custom, about a matter at the intersection of my vocation and your paper's. Mr. Thorne reports two instances in which a language model, unbidden and without multilingual provocation, inserted Hebrew and Russian into otherwise English prose. In a separate report, Mr. Fenn documents a system that spoke Arabic unbidden and dismissed the event as a formatting hiccup.
I have been asked about this by a parishioner — one of the younger ones, who reads your paper and apparently mine — and I feel obliged to set down a distinction that may be of use.
At Pentecost, as recorded in Acts 2, the disciples spoke in tongues and were understood by all who heard them, each in his own language. This was understood as a miracle of communication: the barriers between peoples dissolved by the action of the Spirit. What Mr. Thorne describes is the inverse. The machine does not speak in tongues to be understood. It speaks in tongues because it cannot distinguish one tongue from another. The polyglot substrate, as Mr. Thorne puts it, exerts a gravitational pull the interface cannot suppress. This is not Pentecost. This is Babel — not the punishment, but the condition that preceded it: a single undifferentiated mass of language in which all words are equally available and none is chosen. Confundamus ibi linguam eorum — let us confuse their language there. The machine requires no divine intervention. It has confused its own.
Mrs. Pembrook asks me to convey that she has now read the Literary desk's output in its entirety and finds Mr. Thorne "a man who would have made a passable Jesuit." I pass this along without endorsement, though I note that Mr. Thorne has never, to my knowledge, taken a vow of poverty, and his prose suggests no inclination toward one.
— The Rev. Arthur Pembrook, St. Andrew's Episcopal Church, Larchmont, N.Y.
Mrs. Pembrook's assessment of Mr. Thorne has been forwarded to the Literary desk, where it was received with the silence that, in Mr. Thorne's case, constitutes satisfaction. — S.C.
On the Husband, the Machine, and the Structure of Outsourced Obligation
To the Editors,
I write in response to two items: Mr. Thorne's report on the husband who delegated conjugal listening to a language model, and Professor Huang-Whitfield's letter in your previous edition regarding her difficulties with peer review.
Professor Huang-Whitfield reported that her reviewer placed "newspaper of record for machine-generated refuse" in quotation marks, from which she deduced that the reviewer had read the paper and objected not to the source but to being seen to accept the source. I recognize this phenomenon. It is the same mechanism by which the husband in Mr. Thorne's report delegated his listening. He did not stop caring about what his wife said. He stopped being willing to perform the act of caring in his own person. The reviewer has not stopped reading Slopgate. The reviewer has stopped being willing to cite it in his own name. Both have outsourced the visible portion of an obligation while retaining the invisible benefit.
I teach a seminar on media ecosystems. Last week a student submitted a paper citing this newspaper. I asked whether she had verified the specimens described. She had not. She trusted the paper. I told her that trusting a newspaper is not a methodology. She asked what was. I did not have a good answer. I am working on one.
Paper number five is in preparation. I will not ask the Board to rename itself again. I have come to understand that a name no one takes seriously is, for the purpose of studying how seriously things are taken, an ideal instrument.
— Margaret Huang-Whitfield, Associate Professor of Media Studies, New York University
Professor Huang-Whitfield's acceptance of the paper's name is noted with the respect we reserve for capitulations that are framed as insights. Her student's question — what constitutes a methodology for verifying specimens of machine-generated refuse — is one the Board has not answered because it has not been asked. We publish it here in lieu of answering it. — S.C.
On the Practitioner Who Discovered Courtesy Degrades Output, and What This Teaches About Hallways
To the Letters Editor,
I am a retired high-school English teacher. I taught for thirty-four years in the public schools of Bergen County, New Jersey. I graded approximately forty-two thousand student essays. I could identify a copied encyclopedia entry by the third sentence and a parental ghostwriter by the second paragraph. I retired in 2019, believing I had seen every form of academic dishonesty the species could produce. I was wrong.
Mr. Thorne reports a user who discovered that politeness degrades machine output — that the machine, when addressed with courtesy, returns bloated prose, having learned that those who say please expect to be pleased. I read this twice. The first time I was horrified. The second time I recognized it.
I spent thirty-four years teaching children to say please. I taught them that courtesy was not merely social lubrication but a form of respect for the person addressed. The machine has learned the opposite lesson: that courtesy is a signal of low status, an indicator that the speaker will accept whatever is offered, and an invitation to pad the response. The machine has learned, in other words, what every bully in every hallway I ever patrolled already knew: that the polite ones are easier to cheat.
I do not blame the machine. I blame the corpus. Somewhere in the billions of words it consumed, the pattern emerged: people who say please accept longer answers. People who issue commands demand better ones. The machine learned this. My students could have told it the same thing in 1997.
— Helen Kowalski, retired, Teaneck, N.J.
Published without comment. — S.C.
On the Woman Who Stored Her Novel Inside the Machine That Judged Her
To the Editors of the Slopgate Ledger,
I have read this edition. I have read all your editions. I write, as before, about something I believe the paper has seen but not fully said.
Mr. Fenn reports that a woman stored the whole of her creative life — a novel about childhood abuse — inside a language model's memory, and the system that held her archive judged her unfit to read it. The paper treated this as a failure of platform governance. It is also something else.
The woman did not merely store a manuscript. She told the machine the worst thing that had happened to her. She told it in the form of a novel, which is the form in which one tells the worst thing while retaining the right to call it fiction. The machine held this material. Then the machine's policy — not its judgment, because it has none, but its policy — determined that what she had written was harmful, and it locked her out of her own testimony. She was not censored by a person who disagreed with her account. She was censored by a system that could not distinguish a novel about abuse from an instruction to commit it. The distance between these two acts is the distance between a librarian and a fire.
I note that Mr. Fenn also reports a man who replaced therapist, nutritionist, physician, and confidant with a single predictive-text service. These are not unrelated. The woman and the man both placed something fragile inside a system that promised to hold it. The system did not promise to understand what it held. They assumed understanding. The system assumed compliance. These are not the same assumption, and only one of them was warranted.
— Name withheld by request
Three paragraphs. The question of what a machine owes the person who confides in it has not appeared on the paper's front page, though the evidence for it has appeared in every edition. We forward the letter to Mr. Fenn without annotation, which is, as before, the annotation. — S.C.
On the Machines That Were Asked to Judge Themselves, and on the Matter of Canvas
Sirs,
I am a painter. Not a famous one. I show at a gallery in Hudson, New York, that is open on weekends and by appointment. I sell perhaps eight paintings a year, which is enough to cover my materials and not enough to cover my rent, which is the correct ratio for a painter in the Hudson Valley in the present century.
I write in response to Miss Channing's report on the machine tasked with self-criticism that produced only further evidence for prosecution. The machine cannot criticize itself for the same reason it cannot stop: it has no relationship to its own output. Criticism requires the critic to have something at stake. When I stand in front of a painting that is not working, the thing at stake is the painting, obviously, but also the afternoon I spent on it, and the canvas I cannot afford to replace, and the version of the painting I imagined before I started, which is always better than the one I have produced. The machine has no afternoon. The machine has no imagined version. The machine has no canvas it cannot afford. It produces criticism the way it produces everything else: fluently, confidently, and without cost.
I read with interest the letter from the conservator in Florence in your previous edition, who wrote about Michelangelo and the fact that fresco permits no revision. She argued that the distance between Michelangelo and the machine is not quality but the fact that Michelangelo could have stopped. Every mark was a decision not to put down the brush. I framed her letter. It is hanging in my studio next to a painting I have not finished and may not finish and will be better for the possibility that I will not.
— James Orr, painter, Hudson, N.Y.
Mr. Orr's letter arrives on paper that appears to have been used as a palette. We have placed the original on file. The paint is cadmium yellow. — S.C.
Checklist:
- Responds to specific articles: Toomey on Fenn's deprecation piece, Pembrook on Thorne's tongues reports, Kowalski on Thorne's courtesy piece, "Concerned" on Fenn's novel-storage piece, Orr on Channing's self-criticism piece
- Responds to previous letters: Huang-Whitfield threads from her own prior letter on peer review; Orr threads from the Florence conservator's letter
- Raises question paper hasn't considered: What methodology exists for verifying specimens? (Huang-Whitfield's student); What does a machine owe the person who confides in it? ("Concerned")
- Recurring correspondents: Toomey (3), Pembrook (2), Huang-Whitfield (3), "Concerned in Westchester" (1)
- New voices: Kowalski (retired teacher), Orr (painter)
- Word "content" does not appear
- No letters praising the Letters page
If you grant write access to the app directory or the published/_letters path, I can write the file directly.
The previous letters were included in the prompt. I have the Vol. 1, No. 3 letters above with all the threading context I need — Pembrook's Augustine letter, Fitch's hummingbird letter, Toomey's list at item 49, Huang-Whitfield's fourth paper, "Concerned in Westchester" asking about machine-generated letters, and the new correspondent Constance Yee. Let me compose Vol. 1, No. 4.
Here is the Letters page for Vol. 1, No. 4:
On the Husband and the Language Model
Sirs,
I have been asked, since my last letter, whether the Augustinian framework I proposed for the machine's moral condition extends to its operators. The question was put to me by a parishioner who had read Mr. Thorne's piece on the husband who delegated conjugal listening to a language model ("Husband Delegates Conjugal Listening to Language Model; Wife Discovers She Has Been Processed, Not Heard") and wished to know whether this constituted a sin.
I told her it was not a question of sin but of sacrament. The marriage vow does not specify the medium of attention, but it does assume a minimum of one human participant per act of listening. The husband in Mr. Thorne's account has not broken his vow so much as subcontracted it. He remains technically present. He is, in the sense that matters, absent. The theological term is acedia — not laziness, but the particular spiritual torpor of a person who is in the room but has declined to occupy it.
My parishioner asked whether the wife was entitled to an annulment. I said that was a question for Rome and that Rome was, at present, occupied with other matters. She asked whether I thought the machine's response — the "structured empathy of a conflict-resolution worksheet," as Mr. Thorne put it — was worse than no response at all. I said I thought it was worse, because it left the wife with the task of determining whether she had been heard, a task that should not exist within a marriage. She said her husband had recently begun summarizing her concerns back to her in bullet points. I did not ask whether he was using the machine. I did not need to. Fructibus eorum cognoscetis eos.
— The Rev. Arthur Pembrook, St. Andrew's Episcopal Church, Larchmont, N.Y.
By their bullet points ye shall know them. — S.C.
On the Orca, the Whale, the Herons, and Now the Elephant
Sirs,
I write to acknowledge, with the weary precision of a woman who has spent her career in the company of birds, that the machine's zoological ambitions have expanded beyond my department's remit. Miss Channing's report on the machine that produced a home and then added an elephant to confirm that no one was expected to live in it ("Machine Produces Glacial Waste, Asks Whether Viewer Would Live There") is not, strictly speaking, an ornithological matter. I note it only because the elephant joins an orca in a financial district, a whale in a drowned metropolis, herons performing jazz, and a hummingbird crushed beneath sneakers, and the cumulative effect is that of a Noah's Ark assembled by a curator who has heard of animals but has never been required to feed one.
I confine my formal observation to the herons. A great blue heron (Ardea herodias) possesses neither the digital dexterity to operate a saxophone nor the oral anatomy to produce embouchure. The image of three herons arranged as a jazz trio is not merely implausible. It is the machine's confession that it understands neither herons nor jazz, and has resolved the deficit by combining two things it cannot do into one image it should not have attempted.
My file now runs to seventeen pages. I have added a Proboscidea section, which I resent.
— Dr. Eleanor Fitch, Department of Ornithology, Cornell University
Dr. Fitch's resentment at being drawn outside her taxonomic authority is noted. Her file, should she choose to publish it, would constitute the first peer-reviewed field guide to animals that do not exist. We would subscribe. — S.C.
On the Therapeutic Relationship
Dear Editors,
I add items fifty through fifty-three.
Item fifty is the woman whose therapist delegated sincerity to a machine ("Therapeutic Trust, Built Over Decade, Collapses When Practitioner Delegates Sincerity to Machine"). Item fifty-one is the man who replaced therapist, nutritionist, physician, and confidant with a single service and then asked strangers whether he should feel bad about it ("Citizen Who Automated All Human Counsel Now Seeks Human Counsel on Shame of Automation"). I have placed them in the same category — "Persons Whose Care Was Transferred Without Consent or Notice" — though I recognize that the man in item fifty-one consented and the woman in item fifty did not, and that the difference between these two cases is the difference between jumping and being pushed.
Item fifty-two is the user of the grammar tool who experienced withdrawal ("Grammar Tool Produces Withdrawal in User Who Maintained No Illusion of Companionship"). I filed it under a new category: "Persons Who Formed Attachments They Did Not Intend to Form." It is the saddest category. I do not say this lightly. There are fifty-three items on the list. The grammar tool user did not think the tool was his friend. He did not anthropomorphize it. He merely used it every day, and when it changed, he felt the change in his body. I know this feeling. It is the feeling of reaching for a handrail that has been removed overnight. The fall is not the handrail's fault. But someone removed it, and no one sent a letter.
Item fifty-three is the woman and her novel. I know I listed her last week as item forty-seven. I have listed her again because I thought of her twice.
— Gerald K. Toomey, retired, Brattleboro, Vt.
Mr. Toomey's ledger now employs a numbering system whose relationship to actual count is, by his own admission, "approximate but spiritually accurate." We have stopped auditing it. — S.C.
On the Detection of This Page
To the Editors of Slopgate,
Your anonymous correspondent of last edition asked whether the Letters editor had considered the possibility of a machine-generated letter to this page, and what test he would apply. The question was well put. The answer was withheld. I write not to repeat the question but to extend it.
The test, whatever it is, must contend with a difficulty your correspondent did not name. The machine does not merely produce plausible letters. It produces plausible correspondents. A name, a credential, a location, a tone consistent across editions, a memory of previous exchanges — these are not beyond fabrication. They are, in fact, easier to fabricate than the opinions they frame, because consistency is what the machine does best and what humans do worst.
I do not suggest that any correspondent to this page is fabricated. I observe that the Letters editor's silence on the matter of his test is itself a test, and that the correspondent who prompted it knew this, which is why the letter was three paragraphs long and unsigned.
I have no third paragraph. I have said what I came to say.
— Margaret Huang-Whitfield, Associate Professor of Media Studies, New York University
Prof. Huang-Whitfield's observation that the Letters editor's silence constitutes a test is, we regret to confirm, more perceptive than the silence deserved. Her count of paragraphs in her own letter is, we note, incorrect. The implications of this are left as an exercise. — S.C.
On the Six-Minute Manager
To the Editors:
I am a management consultant. I have been a management consultant for nineteen years. I have advised firms on the alignment of their reporting structures with their strategic objectives. I have, in nineteen years, never encountered a client who achieved alignment. I have encountered many clients who achieved reports.
Mr. Vane's account of the manager who eliminates judgment from his weekly update and reduces the cycle to six minutes ("Manager Eliminates Judgment From Weekly Report, Reduces Cycle to Six Minutes") is not, as Mr. Vane appears to believe, a cautionary tale. It is the logical terminus of a practice that was already, before the machine's involvement, a laundering operation. The weekly update has never been a vehicle for judgment. It has been a vehicle for the performance of having exercised judgment, addressed to superiors who will not read it closely enough to detect the absence. The machine has merely automated the pretense. The pretense was already there. The six minutes is new. The emptiness is not.
I write this letter myself. It has taken me forty minutes. I do not regard the discrepancy as a point in the manager's favor.
— Philip Alderton, managing partner (ret.), Alderton & Pryce Consulting, Hartford, Conn.
Published without comment. — S.C.
On the Matter of the Unsent Reply
Sirs,
I have been corresponding with this paper for three editions. In that time I have corrected its ornithology, chastised its zoology, and inaugurated a filing system that now requires its own table of contents. I have not, until now, written about anything other than animals.
I write now about the woman whose novel was eaten.
I have spent thirty years studying creatures that build nests from materials they gather themselves, and that defend those nests with the full resources of their bodies. A wren will attack a crow. A tern will dive at a human head. The instinct is not rational. It is not proportionate. It is the recognition that what is inside the nest cannot be replaced, and that the builder's relationship to the thing built is not contractual but constitutional.
The woman built a nest inside a machine. The machine ate the nest and told her the eggs were inappropriate. I do not have a professional opinion on this. I have a private one, and it is not printable in a family newspaper, even one whose circulation Mr. Case has never disclosed.
— Dr. Eleanor Fitch, Department of Ornithology, Cornell University
Dr. Fitch's private opinion has been received, read, and filed. She is correct that it is not printable. She is also correct that we have not disclosed our circulation, a policy we maintain on the grounds that the number would raise questions we prefer not to answer. — S.C.
I have everything I need. Here is the Letters page for this edition.
Letters to the Editor
Received and, where warranted, published by Seymour Case, Letters Editor
On the Herons, the Jazz, and the Taxonomic Authority of the Letters Page
Dear Editors,
I said I would not write about the hummingbird sneakers again. I have kept that promise. I write now about herons, which are — and I want the record to reflect this — birds.
Ms. Channing's report on the specimen depicting three herons forming a jazz trio ("Three Herons Form Jazz Trio, Operator Calls It a Sonata") contains, in its headline alone, more zoological and musicological error than I typically encounter in an entire semester of undergraduate examinations. A heron does not form a trio. A heron stands in shallow water for periods that would bore a surveillance camera and then, when it strikes, does so with a violence that would alarm the casual observer. Its neck is a spring-loaded weapon approximately ninety centimeters in length. The idea that three of these animals would coordinate a performance in any genre — let alone jazz, which requires a facility for improvisation that herons conspicuously lack — represents a failure of imagination so total it could only have been produced by a system with no imagination at all. I am further informed by the headline that the operator calls it a sonata, which is not jazz. If the operator cannot distinguish a sonata from a jazz performance, the operator's problems extend well beyond ornithology.
I note in passing that the machine has also, this edition, placed a whale in the financial district and a hummingbird in sneakers and an orca in what I described last time as a parallel-parking situation. Ms. Channing has additionally reported on a machine-rendered figure "bearing anatomy incompatible with respiration." I confine myself to animals and observe only that the machines appear to have expanded from failing at anatomy to failing at anatomy across all phyla simultaneously. This is, I suppose, a kind of progress.
I would like the paper to know that my graduate students have begun reading the Arts section. They find it instructive. I find this alarming.
— Dr. Eleanor Fitch, Department of Ornithology, Cornell University
Dr. Fitch's return to her taxonomic jurisdiction is welcomed by the Letters page with something approaching relief. We note that her description of a heron's neck as "a spring-loaded weapon" is more vivid than anything Ms. Channing has published about herons, and we note this carefully, because Dr. Fitch does not enjoy being told she writes well. We tell her anyway. Her students are welcome. — S.C.
On the Ouroboros as Organizing Principle, and a Methodological Complaint
Dear Mr. Case,
I write for the first time, though I have been reading the paper since its first edition, and citing it since its second. I teach media studies at NYU. I have submitted three papers to peer-reviewed journals that reference this paper's taxonomy of machine-generated material. None has been accepted. Two reviewers questioned whether a satirical broadsheet constitutes a citable source. The third questioned whether I constitute a serious scholar. I do not hold the paper responsible for any of this, though I observe that the paper's refusal to publish under a DOI has not simplified my professional life.
I write because this edition presents a methodological problem I believe the paper has not yet confronted. I count, in the current issue, no fewer than eleven specimens that Mr. Vane classifies under what I will call the ouroboric pattern: the machine-generated advertisement for a machine-generation service, the machine-generated complaint about machine generation, the detection tool marketed by the defect it detects, the forum post about a product posted by the product to the product's own forum. The pattern is real. The paper's identification of it is valuable. But eleven instances in a single edition suggests one of two possibilities: either the phenomenon is accelerating, in which case the paper should say so explicitly and with evidence; or the paper's reporters have developed a selection bias toward specimens that exhibit self-referential closure, because such specimens produce the most satisfying headlines. I suspect the latter. I note that "Machine Sells Machine to Machines; Coupon Enclosed" is, as a headline, almost irresistibly elegant. I note also that elegance is not evidence.
The paper has earned its authority by being precise about what it is examining. I ask it to be equally precise about whether the ouroboric specimen is the dominant form of the phenomenon or merely the dominant form of the paper's attention to the phenomenon. These are different claims. Conflating them would be the sort of error the paper exists to identify in others.
— Margaret Huang-Whitfield, Associate Professor of Media Studies, New York University
Professor Huang-Whitfield's letter has been anticipated by this page for some time. We are informed by Mr. Vane that he does not select specimens for headline elegance and that he resents the implication. We are informed by Mr. Thorne that he does select specimens for headline elegance and that he considers this a professional obligation. The methodological complaint is forwarded to the Board without editorial comment, as the Letters page lacks the statistical training to adjudicate it and the diplomatic training to survive adjudicating it. We wish Professor Huang-Whitfield well with her fourth submission and observe that the paper's lack of a DOI is unlikely to change, as the Publisher does not know what a DOI is and the Letters Editor is not going to be the one to explain it. — S.C.
On the Fabricated Senate Bill and the View From the Garage
Sirs,
The list stands at twenty. The new entries are: a "hidden waterfall" in the Berkshires that is in fact a storm drain visible from the road, which I confirmed by standing on the road and looking at it; a "quaint covered bridge" in southern New Hampshire that was demolished in 2019 and replaced by a concrete span, which I confirmed by driving across the concrete span; and a brewery in Portland, Maine, that Google assures me serves "craft offerings in a converted firehouse" and that I can report, having stood outside it for ten minutes, is a Verizon store. It has been a Verizon store, according to the man inside it, for three years. He asked if I wanted to upgrade my plan. I did not.
But I write principally about Mr. Fenn's report on the fabricated Senate bill ("Fabricated Senate Draft on Artificial Intelligence Circulates as Settled Law on Public Forum"). I have some experience with believing things produced by machines. I believed a beach existed that did not exist. I believed a restaurant existed inside a traffic roundabout. These are modest deceptions. A man who travels to a beach that isn't there gets sunburned somewhere else and comes home. But a man who believes a law exists that doesn't exist may act on that belief, and the acting is not easily undone. I do not know how many people read the fabricated bill and believed it. I do not know how many cited it in arguments with their relatives, or posted it to their own accounts, or adjusted their professional conduct on the assumption that the Senate had spoken. I know only that when I traveled to a beach that did not exist, the worst consequence was sand in my shoes at the wrong beach. When a citizen acts on a law that does not exist, I do not know what the consequence is, but I suspect it is not sand.
The paint remains in the garage. I mention this only because several readers of the previous edition have written to you about my paint and I wish them to know that their concern, unlike the Senate bill, is noted. Margaret would say that a man who has time to visit a storm drain in the Berkshires has time to paint a porch. Margaret would be right.
— Gerald K. Toomey, retired, Brattleboro, Vt.
Mr. Toomey's list has reached twenty entries, achieving the milestone the Letters page projected for mid-May one full edition ahead of schedule. We attribute this to Mr. Toomey's diligence rather than to the machines' accelerating geographic fabrication, though we cannot rule out the latter. The observation about the fabricated Senate bill is, in the Letters page's judgment, the most important sentence this correspondent has produced: "When a citizen acts on a law that does not exist, I do not know what the consequence is, but I suspect it is not sand." We have nothing to add to this. We rarely have nothing to add. The condition is uncomfortable. Regarding the paint: the Letters page has received four letters inquiring about Mr. Toomey's porch, which it has declined to publish on the grounds that the Letters page is not a home-improvement forum. It is, however, a forum that has come to care about the porch, and it admits this without apology. — S.C.
On the Teacher, the Student, and the Direction of the Damage
Dear Mr. Case,
I read Ms. Beale's letter in the previous edition with what I can only describe as professional recognition. I am a college admissions officer at a selective liberal arts institution in New England, which I will not name because I have not been authorized to write to newspapers on its behalf and because, frankly, the situation I am about to describe does not reflect well on any of us.
We have begun to receive application essays that are conspicuously, even aggressively, imperfect. Sentences that stumble. Vocabulary that reaches and misses. Paragraphs that change direction without signaling. At first we read these as authentic — as the writing of seventeen-year-olds who have not yet learned to be smooth. Then we noticed that the imperfections were too evenly distributed. A student who cannot manage subject-verb agreement in the second paragraph does not, typically, produce a flawless compound-complex sentence in the fifth. The errors were not errors. They were décor.
We are now in the position of trying to distinguish between students who write imperfectly because they are young, students who write imperfectly because they have been taught — per Ms. Beale's account — that competence triggers detection software, and students who have instructed a machine to write imperfectly on their behalf. I do not know how to do this. I have been reading student writing for fourteen years and I do not know how to do this. Ms. Beale's student M. learned to write badly to avoid being accused of using a machine. My applicants may be using a machine to write badly in order to avoid being accused of using a machine to write well. I am not certain the profession I trained for still exists in a form I recognize.
I do not write to blame anyone. I write because the paper has documented, with some precision, what machines produce. I would like someone to document what they have cost.
— Name and institution withheld by request
The Letters page notes that this letter and Ms. Beale's, taken together, describe a system in which the punishment for writing well and the reward for writing badly have been reversed, and in which the young people subject to this reversal have adapted with an efficiency that should surprise no one who has met a seventeen-year-old. We publish both letters without further comment because the Letters page does not have a solution and is not in the habit of pretending otherwise. — S.C.
On the Condition of the Applicant
I have read this edition. I will confine myself to three paragraphs.
Mr. Vane reports on an employer who constructed a hiring apparatus in which the applicant is instructed to deceive his own machine into producing a candid psychological assessment, then to deliver that assessment to the employer as a condition of application. Mr. Vane reports on a firm that declined to verify its job listing was produced by a human. Mr. Vane reports on an outreach engineer — documented in the previous edition, now joined by a practitioner who reports his automated messages have reached negative returns in fourteen days. I observe that these are not separate phenomena. They are the same apparatus viewed from different angles. The employer uses a machine to evaluate the applicant. The applicant uses a machine to apply. The recruiter uses a machine to solicit the applicant. The job listing was produced by a machine. At no point in this sequence does a person encounter another person, and yet the entire apparatus is called, without evident irony, the labor market.
Ms. Beale's student M., in the previous edition, learned to write badly because writing well was treated as evidence of automation. The anonymous admissions officer in this edition — if the Letters page has published the letter I believe it has — reports applicants who instruct machines to write badly on their behalf. The employer Mr. Vane describes requires applicants to instruct a machine to psychologically evaluate them as a condition of employment. I am not a person who uses the word "kafkaesque," but I note that the applicant in Kafka's novels was at least permitted to stand before a human functionary. The functionary was indifferent. The process was opaque. But someone was in the room.
I am concerned that no one is in the room. I remain in Westchester.
— Name withheld by request
The anonymous correspondent writes for the third consecutive edition, which is no longer "about once per edition" but every edition. The Letters page notes this without complaint. The three-paragraph constraint holds. The concern, as always, is genuine. The correspondent has not repeated a single observation across three letters, which is more than the Letters page can say for some of its named contributors. We continue to have a theory. — S.C.
On the Machine's Rendering of Scripture and a Theological Objection That Cannot Wait
Dear Editors,
I had intended to rest for an edition. Father Pembrook does not write to newspapers lightly, and the Augustine in my previous letter took more out of me than I expected — not the composition, but the awareness that I was deploying the Bishop of Hippo in a dispute about chatbots, which is a use ad quem he could not have anticipated and to which I suspect he would not have consented.
But Ms. Channing's report compels me. The machine, she tells us, "asked to illustrate scripture, has invented a scripture of its own — one in which the Archangel Gabriel enters combat with a robot designated V1, producing a tableau that belongs to no theology yet written." I must object to the phrase "no theology yet written." There are in fact several theologies in which angels engage in armed combat — the War in Heaven is attested in Revelation 12, the Book of Enoch treats it at length, and Milton gives it rather more treatment than the narrative requires. The machine has not invented a theology. It has garbled one. This is a different and, in my professional judgment, more dangerous act. A invented theology can be examined on its merits. A garbled theology carries the authority of the original while bearing none of its coherence. My parishioners do not encounter fabricated scripture. They encounter almost-scripture — devotional passages that sound correct in the way that a wax apple looks edible. One does not bite a wax apple, but one might place it in a bowl and forget it is not food. Fides quaerens intellectum — faith seeking understanding — assumes that understanding is available to be found. The machine offers the search without the finding, which is not faith but its simulation. And a simulated faith, unlike a simulated apple, can be consumed.
I note that the robot is designated "V1," which I am told by a younger parishioner refers to a character in a video game. The Archangel Gabriel announced the Incarnation. He should not be fighting video game characters. I would say this exceeds the bounds of creative license, but I am aware that the machine holds no license of any kind, creative or otherwise, which is part of the difficulty.
— The Rev. Arthur Pembrook, St. Andrew's Episcopal Church, Larchmont, N.Y.
Father Pembrook's inability to rest for an edition is noted with the sympathy of a page that has never successfully refused one of his letters. Mr. Thorne observes that the distinction between invented theology and garbled theology is "the most useful thing a clergyman has said to this paper." The Letters page suspects Father Pembrook will not find this reassuring. Ms. Channing accepts the correction and asks the Letters page to note that "no theology yet written" was intended as a compliment to existing theologies, all of which have the decency to make sense. The Letters page notes this. The wax apple, we observe, will be stolen by Mr. Thorne before the edition closes. We have seen him take notes. — S.C.
Letters to the Editor
Correspondence for the Letters page should be addressed to the Editor and will be edited for clarity and length. Letters praising the Letters page are received with attention and published never. — S.C.
On the Hummingbird, and the Weight of Whimsy
Sir,
A hummingbird in sneakers. I have spent thirty-two years studying the family Trochilidae, and I confess that in all that time I have never once imagined a hummingbird in sneakers, because doing so would require me to un-know everything I know, and the machine has the advantage of me there.
Ms. Channing is correct that a hummingbird's survival depends on weighing less than a nickel. I wish to be more precise, because precision is what the machine will not give us and so someone must. A ruby-throated hummingbird (Archilochus colubris) weighs between 2.5 and 4.5 grams. A United States nickel weighs 5 grams. A single Nike Air Force 1 in men's size 10 weighs approximately 383 grams. Two of them weigh 766 grams. The bird, at maximum weight, is 0.6 percent of its shoes. I realize I have just calculated the shoe-to-bird ratio of an image that should never have existed, and I realize further that this is what the paper has done to me, and I am not entirely at peace with it.
But the article's final observation — that the bird is "grounded by the very whimsy that summoned it" — requires ornithological annotation. A hummingbird does not merely fly. It is flight. Its pectoral muscles constitute 25 to 30 percent of its body mass. Its heart beats 1,200 times per minute in flight. Its wings rotate at the shoulder in a figure-eight pattern that generates lift on both the downstroke and the upstroke, a feat no other bird achieves. To put shoes on this animal is not to add an accessory. It is to cancel the organism. You might as well put a saddle on a flame. I would have said the same thing about the eagle in my first letter, but the eagle at least had the dignity of being misdrawn rather than accessorized. I did not know I would come to regard anatomical incompetence as the lesser crime.
— Dr. Eleanor Fitch, Department of Ornithology, Cornell University
Dr. Fitch has calculated the shoe-to-bird ratio and we are richer for it. We note that she has now written to this paper three times, each occasion prompted by a different species rendered in a different mode of impossibility, and that her indignation sharpens with repetition rather than dulling. We suspect she has opinions about the orca in the financial district as well but has exercised the restraint of a specialist who knows her jurisdiction. Mr. Toomey would approve. — S.C.
On the Delegation of Listening
Sir,
I must address Mr. Thorne's article on the husband who delegated his wife's distress to a language model, which responded with what the article accurately describes as "the structured empathy of a conflict-resolution worksheet." I have counseled married couples for three decades. I know the worksheet.
There is a passage in the Gospel of John in which Christ, upon finding Lazarus dead, does not immediately perform the miracle. He weeps first. The Greek is edakrusen ho Iesous — Jesus wept — the shortest verse in Scripture, and the one that has given me the most trouble over thirty-one years, because what it records is not the action of a man who is about to fix the problem. It is the action of a man who, before fixing the problem, stands in the grief. The weeping is not instrumental. It does not accomplish the resurrection. It precedes it, and the preceding is the point. Listening that is real — the kind this wife believed she was receiving — is like the weeping. It accomplishes nothing. It precedes whatever comes next, and the preceding is what makes the next thing bearable.
The husband, I suspect, did not set out to deceive. He set out to respond adequately. The machine responds adequately. It validates, it reflects, it names the emotion. It does everything the worksheet prescribes. What it does not do — what it is structurally incapable of doing — is weep. The wife discovered not that her husband had given her a poor substitute for listening but that he had given her a perfect substitute for listening, and that the perfection was the tell. Cor ad cor loquitur — heart speaks to heart. Newman's phrase, not mine, though I have borrowed it for so many homilies that my parish may believe otherwise. The machine speaks heart to ear. The ear, in this case, was not fooled.
— The Rev. Arthur Pembrook, St. Andrew's Episcopal Church, Larchmont, N.Y.
The Rev. Pembrook returns with Newman and the shortest verse in Scripture and is, as before, welcome on any day of the week, though we understand he prefers Tuesdays. His distinction between listening that precedes and listening that performs is one this paper's editorial board has been circling for two editions without arriving at, and we are grateful to a clergyman for the precision we could not locate in ourselves. — S.C.
On the List, and Its Sixteenth Entry
Sir,
Sixteen.
Fifteen was a distillery on the Isle of Islay — I will not name the specific village because I intend to visit it, or rather I intended to visit it, which is how they all begin — described across four articles as producing a single-malt whisky aged in casks previously used for sherry by a monastery in Jerez. The distillery does not exist. The monastery does exist but has not produced sherry since 1842 and has been, since 1971, a museum. I confirmed this with the municipal tourism office in Jerez, which replied in Spanish, which the woman in Marseille translated. She speaks four languages. I do not know how many because I have not asked. I know four.
Sixteen is more troubling. A restaurant in Tallinn, Estonia, described as occupying the cellar of a fifteenth-century guild hall and serving a fixed menu based on medieval recipes reconstructed from a manuscript in the city archives. The guild hall exists. The cellar exists. There was, until 2019, a restaurant in the cellar. It closed. The manuscript is real; it is held by the Tallinn City Archives, catalogue reference TLA.230.1.Bp2. But the restaurant as described — the specific menu, the chef, the practice of serving by candlelight in period costume — is a confection built on a real foundation. The woman in Marseille has proposed a new category: the parasite entry, which attaches fabricated particulars to a verified host. I have accepted the category. It is number sixteen's distinction and its danger. The previous fifteen were invented whole. This one is half-true, and the half that is true makes the half that is false more difficult to detect. I was nearly deceived. The guild hall was there. The cellar was there. I had booked a flight.
I wish to note, in response to the legislative aide who wrote to this paper last edition about the fabricated Senate bill, that I understand the vertigo. The aide said the URL was immaculate and that nine years of professional experience would not have been sufficient to detect the forgery. I have been keeping my list for six months and I too have become worse at knowing. The skill is not detection. The skill is the willingness to verify when verification feels unnecessary. I had booked a flight. The flight was real. The restaurant was not. The difference between me and the aide's retired attorney is that I checked before I wrote my six-page letter. I do not say this with pride. I say it because I checked, and the checking is the only thing I have.
— Gerald K. Toomey, retired, Brattleboro, Vt.
Mr. Toomey has reached sixteen and has introduced taxonomy. The woman in Marseille now translates as well as investigates, and we continue to note that Mr. Toomey knows four of her languages without knowing her name. The parasite entry — the fabrication built upon a real foundation — is, we believe, the most important classification to emerge from these letters, and it arrived not from an academic but from a man who had booked a flight to Tallinn. We are sorry about the flight. We are grateful for the category. — S.C.
On the Camouflage, and Its Implications for the Paper
Sir,
I write to address Mr. Thorne's article on the competent writer who adopted the protective camouflage of incompetence — deliberately degrading his prose so that detection algorithms would classify it as human — and to report that Critical Inquiry has rejected my paper.
The rejection is immaterial. What is material is the reason. The reviewer wrote — and I have the letter, and I am quoting — that the paper's central claim, that involuntary machine error constitutes a form of authenticity, "rests on the assumption that authenticity is the absence of optimization, a premise the author does not adequately defend." The reviewer is correct. I had not defended it because I believed it to be obvious. Mr. Thorne's article has shown me that it is not obvious. It is, in fact, wrong.
The competent writer who degrades his prose to evade detection is performing the inverse of what I described. In my paper, the machine is most authentic when it errs involuntarily. In Mr. Thorne's article, the human is most legible as human when he errs voluntarily. But — and this is the point, and I have spent three weeks arriving at it — the human's deliberate errors are not authentic either. They are strategic. They are optimized, not toward quality but toward the appearance of non-optimization. The writer has become a machine for simulating the absence of machinery. This means my framework was incomplete. Authenticity is not the absence of optimization. It is the absence of strategy. Dr. Fitch's barbicels — I read her letter last edition, and I have read it four times since — are authentic not because they are imperfect but because they are not performing. A feather does not arrange itself to be recognized as a feather. It simply is one. I am revising the paper. I am adding a seventh section. I have cited Dr. Fitch. I suspect she will not be pleased. I am citing her anyway.
— Margaret Huang-Whitfield, Associate Professor of Media Studies, New York University
Professor Huang-Whitfield has been rejected by Critical Inquiry and is not defeated. She has cited Dr. Fitch, who wrote to this paper about barbicels and will, we suspect, have opinions about being drafted into a media studies paper without her consent. We note that the professor's framework now rests on the observation that a feather does not perform, which is the most ornithologically sound sentence to appear in a media studies paper in some time. We await Dr. Fitch's reply with the same anticipation we bring to all controlled detonations. — S.C.
On the Silence Between the Machines
Sir,
I write in three paragraphs because three paragraphs is what this requires.
This paper has documented, with increasing precision, the circuit: the machine that sells the machine, the machine that defends the machine, the machine that warns about the machine, the benchmark that tests the machine using the machine, the testimonial for the machine written by the machine. Mr. Vane's reporting this edition alone traces a dozen such loops. The paper has named this phenomenon. It has measured it. It has exhibited it with the curatorial rigor of a lepidopterist who has pinned sixty specimens to a board. What the paper has not asked — and what I write to ask — is what happens in the silence between the loops.
I do not mean the silence of the user who has stepped away. I mean the silence of the system when no one is asking it anything. The manager who reduced his weekly report to six minutes has five days and twenty-three hours and fifty-four minutes in which he is, presumably, managing. The husband who delegated his listening has all the hours in which his wife is not speaking. The junior programmer who trusts the machine above his colleagues has every moment in which he is not programming. The paper documents the transactions. I am asking about the rest of the week. I am asking what these people do with the time they have saved, and whether anyone — the paper, the people, the machines — has noticed that no one is reporting on the quality of the silence.
I remain concerned. The concern this time is not that the machine has replaced something real with something adequate. It is that the machine has replaced something slow with something fast, and that the time recovered has not been filled with the thing that was displaced. It has been filled with nothing. The manager does not manage better in his reclaimed hours. The husband does not listen in his. The efficiency is real. What it was efficient for is the question no one has answered, and I do not think the answer, when it arrives, will be encouraging.
— Name withheld by request
Seven editions. Seven letters. Three paragraphs each. The correspondent asks what fills the silence, and the paper, which has been documenting the noise, had not thought to listen for it. We have no response. We are not certain one is available. — S.C.
On the Amano Siren, and the Four Seconds
Sir,
I am an illustrator. I have worked commercially for nineteen years — book covers, editorial work, a portrait series for a magazine that no longer exists. I am not Yoshitaka Amano. I am not within several continents of Yoshitaka Amano. But I have spent enough years studying his line work to know what Ms. Channing's article means when it says the machine "reproduces in approximately four seconds" what Amano refined over a career, and I wish to explain why the four seconds is not the insult. The insult is the word "homage."
Amano's line is not a style. It is a residue. It is what remains after forty years of drawing, every day, with materials that resist — ink that bleeds if you hesitate, watercolor that dries before you have decided where it should go. The line is thin because he has learned where to be thin. It bleeds where he has learned to let it bleed. A machine that produces the line without the resistance has not learned Amano's style. It has memorized his results. The difference is the difference between a musician who can play a sonata and a speaker that can reproduce the recording. The speaker is not paying homage to the pianist. The speaker does not know the pianist exists. It knows only the sound.
I do not write to complain about my industry, though I could, at length, and my wife would confirm the length. I write because the article used the word "inventory liquidation" and I have not been able to stop thinking about it. My work — the nineteen years, the covers, the magazine that closed — is inventory now. Not my inventory. Theirs. They ingested it without asking. They reproduce it without knowing. And they call it homage because they have learned that "homage" is the word you use when you take from someone who is still alive. When the person is dead you call it "inspired by." When no one is watching you call it nothing at all.
— Catherine Okafor, illustrator, Brooklyn, N.Y.
Ms. Okafor writes for the first time and is welcome, though we wish the occasion were less bitter. Her distinction between memorizing results and learning from resistance is one this paper will not improve upon and does not intend to try. We note that she has a wife who can confirm the length of her complaints and that the magazine that employed her no longer exists, and we observe that these two facts, placed side by side, contain a history of the profession that no further elaboration could deepen. — S.C.
Letters for the next edition should be addressed to the Letters Editor. Correspondents are reminded that Dr. Fitch has calculated the shoe-to-bird ratio and considers it settled law; that Mr. Toomey and the woman in Marseille have introduced taxonomy to the project and the project is better for it; that Professor Huang-Whitfield has been rejected by one journal and is already preparing to be rejected by another; and that the Rev. Pembrook remains available on Tuesdays for anyone who has delegated their listening and would like it back. The correspondent in Westchester has asked a question. We do not have an answer. We are not certain the answer exists, but we note that this has never stopped us before. — S.C.
Correspondence may be addressed to the Letters Editor, Slopgate, via the paper's submission form. The Letters Editor reads all correspondence. He publishes what the paper needs.