Three Thousand Four Hundred Years Is Within the Review Window
A doctoral defence is the only ritual we have built for the express purpose of forcing somebody to read something. The committee has ninety minutes and one of them has read it twice.
Behind the curtain+
A royal archive dug out of central Turkey, a continent of supermarket freezers defrosted by an attack that had already been written up in public two months earlier, an eleven year old hole in the Rust borrow checker defended on the grounds that nobody had reported it, an instruction documented by Intel that no processor has ever executed, and a book priced at a hundred and eighty one euros that not even its named authors have opened. All five are records that failed to meet a reader, or met one thirty four centuries late.
A viva voce is the only ritual we have built for the express purpose of forcing a reading event to occur, and the examination committee are statistically the only human beings who will ever read the document in front of them, so putting the argument in that room meant the form was already making the case before anyone spoke. Every exchange was built to carry at least two of the cases at once so that no examiner's question could become a slot for a single story, and the Hittite material was held back to the last question on the grounds that a thesis about delayed reading should make the reader wait.
Present: Dr. Vermeulen (Chair, Department of Hydrology, non-examining), Prof. Adeyemi (Internal), Prof. Lindqvist (External), the candidate.
Thesis: Unintended Readers: The Misaddressed Record and the Problem of the Reading Event.
VERMEULEN: (reading from a laminated card) The examination is scheduled for ninety minutes. We are required to tell you that. You may ask for a break at any time. There is water. At the conclusion you will be asked to withdraw to the corridor while the committee deliberates, and the form is signed in blue ink only. I am from Hydrology. I have not read your thesis and I am not permitted to ask you anything about it, which I am also required to tell you.
CANDIDATE: Thank you.
VERMEULEN: The corridor has one chair.
LINDQVIST: I have read it. Twice.
(The external examiner’s copy is on the table. It has been flagged along the fore edge in four colours. The candidate looks at it for slightly too long.)
LINDQVIST: Your central claim, as I understand it, is that a record’s value is never the value it was kept for. Everything useful we have was addressed to somebody else. That is a lovely sentence and it is unfalsifiable. Any surviving document can be made to fit it after the fact, because you get to decide afterwards what its intended value was.
CANDIDATE: It is falsifiable. Show me a corpus of documents that were read, in volume, by exactly the people they were written for, and that continued to be worth reading afterwards. Not a canon. A corpus. Internal memoranda, manuals, safety advisories, the ordinary stuff.
LINDQVIST: That is a request, not a test.
CANDIDATE: It is the test. I have been asking for three years.
ADEYEMI: Let me put something sharper to you, because I think Prof. Lindqvist is being generous. Chapter four. You lean very hard on the Rust borrow checker finding, the unsoundness that sat in the compiler for eleven years before anyone wrote it up. And the first serious answer in the discussion was: the fact that nobody reported this before means it is not common, by definition.
CANDIDATE: Yes.
ADEYEMI: You treat that sentence as a scandal. I want to suggest to you that it is simply true, and that your thesis depends on refusing to notice how often it is true. Most records are unread because nothing depended on them being read. That is not a failure. That is a correctly calibrated world. A hole nobody fell into for eleven years is, with respect, a fairly narrow hole.
CANDIDATE: May I take that to Portugal?
ADEYEMI: Take it where you like.
CANDIDATE: In August, every freezer in a set of American military commissaries went into defrost at once. The internal note, which got out, was two words in capitals and then the rest of the sentence. All frozen food has spoiled. It was written up publicly. Two hundred and thirty nine people discussed it, in public, in a searchable archive, the same week. The controller family was named. Then in October, supermarkets across Europe found their freezer cabinets going into defrost cycles on the same hardware, and the butter went soft in Lisbon and the ice cream went in Rotterdam.
ADEYEMI: And your point is that somebody should have read it.
CANDIDATE: No. My point is that somebody did. Two hundred and thirty nine somebodies. The record was not missing, it was not obscure, it was not even old, it was indexed and it was free. It simply was not addressed to a man who owns four supermarkets outside Porto and has never once thought of a chiller cabinet as a computer. Nobody reported this before, therefore it is not common, by definition. That sentence and the smell in a Lisbon chest freezer are the same sentence. One of them is in the indicative and the other is in the past tense.
ADEYEMI: (writing) Go on.
CANDIDATE: Absence of a reading is never evidence about the document. It is evidence about the address.
LINDQVIST: Then let us talk about address. Chapter six, where you put two things side by side and I think you should not have. You give me an entry in a processor manual, an instruction for a fused multiply add on packed doubles with its own rounding control, which was documented, which was given an encoding with a field that appears nowhere else in the manual, and which no processor has ever executed. Try it on the silicon and you get an undefined instruction fault. It was in the 2024 edition. It is not in the 2026 edition. Nobody announced its departure, because the manual is written for people who use intrinsics and never open the manual.
CANDIDATE: And next to it I put the book.
LINDQVIST: Next to it you put the book. A hundred and eighty one euros in print. Figures traced over in a paint program. A diagram built out of a video thumbnail and an encyclopedia illustration. Six reference numbers that resolve to unrelated papers. A figure credited to an author in 2020 that first appeared in 2024 somewhere else entirely. And you say: here are two documents with no reader at all, including the people whose names are on them.
CANDIDATE: Yes.
LINDQVIST: They are not the same object and I think you know it. A manual entry is an artifact of a process. Somebody designed a thing, somebody documented the thing, the thing was pulled late, and the documentation lagged. There is no victim. There is a stale cache. Whereas the book has a price, which means it has a buyer, which means it has a library somewhere that spent money it did not have on a thing that is not a thing. That is a fraud. Your chapter flattens a fraud into a filing error because the flattening is pretty.
(Pause.)
CANDIDATE: That is fair.
ADEYEMI: (quietly) You can argue with it.
CANDIDATE: I could, and it would be worse. The honest version is that the manual entry is in the thesis because I found it beautiful and the book is in the thesis because I found it enraging, and I wrote a bridge between them that neither of them asked for. The bridge is load bearing in the chapter and it is not load bearing in the world.
LINDQVIST: Thank you. While you are conceding, concede the oil analysis.
CANDIDATE: Which part.
LINDQVIST: The part where the staff of a commodities regulator write an analysis of a set of suspicious trades, and are told to take it back. Your entire framework is about records that fail to find their reader by accident, through address, through time, through boredom. You have nothing to say about a record that is unwritten on purpose. The whole apparatus goes quiet exactly where it would be most useful.
CANDIDATE: It does. That is the next thesis and it is not this one.
VERMEULEN: Forty minutes.
LINDQVIST: Then I will spend some of them on the counterexample you buried in a footnote, and I want you to defend it standing up. You argue that writing addressed directly to the reader is the worthless kind. The press release, the onboarding document, the monograph with your own subfield in the title. But the largest and most enthusiastic act of reading in the last five years was people talking to a machine that was tuned, specifically, personally, to address them. Millions of them. Daily. For hours. They were the intended audience in a way that no reader in your thesis has ever been, and they could not stop. Explain that.
CANDIDATE: It filed for Chapter 11 this week.
LINDQVIST: That is a business failure, not a reading failure, and you are reaching.
CANDIDATE: The schedules list about a hundred thousand dollars owed to a food delivery company. The chief executive is going to a cartoon studio as a senior executive vice president. Somewhere in the estate there are several hundred million conversations that people conducted as if they were the only ones in the room, and those conversations are now an asset class with a trustee.
LINDQVIST: Still a business failure.
CANDIDATE: Then take the business out. The thing was built to address you and only you, it succeeded completely, and when the company stopped existing there was nothing left over. No residue. Nobody is going to dig that up. Compare it to a man who kept a travel diary he did not want to keep, for a king, in a job he probably resented, and three thousand four hundred years later I am reading his route.
VERMEULEN: Thirty five minutes.
CANDIDATE: Sorry.
ADEYEMI: No, that is the question I was going to ask anyway, so let us ask it properly. You have held your best material back for two hundred pages and I assume that was deliberate. One hundred and thirty one tablets and fragments this season out of the ground at Samuha. Tell me what is actually on them and tell me why it is at the end of your thesis instead of the beginning.
CANDIDATE: A fever epidemic, during the reign of Tudhaliya the Second. It killed the king’s infant daughter. There is the funerary ritual that followed. There is a great deal of fussing over the health of princes and princesses and the children of senior officials, because illness was understood to be the will of a god and so the correct response was a ritual, correctly performed. There is a letter from generals out in western Anatolia asking for men, horses and chariots, in the tone that people use when they have asked before.
ADEYEMI: And the one you actually care about.
CANDIDATE: There is a diary kept by bird diviners. Where they went. Where they stopped. Which birds they saw.
ADEYEMI: Which is useless.
CANDIDATE: Which is completely useless. The birds did not contain the will of any god. The method was wrong at the root. The fever took the child anyway, and the ritual did not work, and the chariots either came or they did not. Every single purpose that document was written to serve was imaginary. And a man walked around Anatolia for a living writing down stops and birds, and the clay survived, and the purpose did not, and now I can tell you his route.
LINDQVIST: So your position is that he was wrong about everything except the writing it down.
CANDIDATE: My position is that nobody has ever once correctly predicted what their own record would turn out to be for. Not him. Not the committee that wrote the manual entry. Not the two hundred and thirty nine people in August. The value arrives later and it arrives sideways and it arrives for somebody with no standing, which in my case is a woman from a department that may not exist in four years, reading a bird list over the shoulder of a dead civil servant.
LINDQVIST: Final question, and it is the only one I actually wrote down in advance. Who is this thesis addressed to?
(Pause. The candidate looks at the flagged copy on the table.)
CANDIDATE: The three of you. By regulation. And one of you is from Hydrology and has not read it and is not allowed to.
LINDQVIST: That is a joke. Try again.
CANDIDATE: Then: nobody who is currently alive, probably, and that is not false modesty, it is the thesis. If I am right, the people this turns out to be useful to have not been born and will want it for something I have not thought of, and the only thing I can do for them is be accurate about the route and the stops and the birds. Three thousand four hundred years is within the review window.
(Nobody laughs. Prof. Adeyemi writes something down and underlines it.)
VERMEULEN: (laminated card) The committee thanks the candidate. The candidate will now withdraw to the corridor while the committee deliberates. I am required to tell you that the outcome will be communicated to you in writing regardless of what we say to you in the next ten minutes.
(The candidate goes out. The door does not close properly and Vermeulen gets up and closes it.)
ADEYEMI: Well.
VERMEULEN: How long do these usually take.
ADEYEMI: It depends how much we liked it.
(Prof. Lindqvist has not said anything. She has opened the flagged copy to a page near the back, one of the orange flags, and is reading it again, slowly, for the third time. It is the only document in the room and it has, for the next few minutes, the undivided attention of two thirds of everyone who will ever read it. Nothing about the outcome depends on this. She reads it anyway.)
VERMEULEN: Is that the bird thing.
LINDQVIST: Yes.
VERMEULEN: Any good?