Alt Binaries

[Warning: Pretty much every cultural term in this essay is fictional. That’s just the bit.]

Independent Review Panel Concludes Investigation, May 2026

There is a question every editor in journalism has to answer about every story, and the way they answer it tells you more about the publication than the masthead does. You can send the reporter who has covered this beat for fifteen years, who knows everyone involved, who can tell you which of the named sources is lying because she has heard that source lie about smaller things before. Or you can send the reporter who landed yesterday from another section, who has no relationships to protect, who will read the documents as a stranger would read them and not notice the things a regular would notice because she does not know which things are supposed to be noticed.

Both reporters will produce a story. The stories will not be the same story.

The insider will get the texture right. She will know which of the three official statements is the one the institution actually believes, because she remembers the press secretary who drafted it from a previous incident where the same press secretary said the same thing about a different scandal. She will know that the named villain is also a person, that the person is divorced and has a kid in middle school, that the kid is going to read this story. She will write something that captures the actual political and institutional dynamics in a way that someone who has been there can write and no one else can. She will also, somewhere in the third or fourth paragraph, soften a line that should not be softened, because she has had drinks with the person being criticised and the person was funny that night, was a real human being, and the reporter knows things about him that complicate the simple narrative the story wants to tell.

The outsider will miss the texture. She will accept the official statement at face value when she should not, will quote the wrong source as authoritative, will get the institutional dynamics flat or wrong, will write a story that anyone who knows the situation will read with the particular wincing recognition of seeing your country’s flag drawn by a tourist who has clearly never been there. She will also, somewhere in the third or fourth paragraph, write the line the insider would have softened, because she does not know the person yet and the line is true.

You cannot have both reporters. You can only have one. Which one would you send.

The standard answer is that this is why journalism has editors. The insider writes the story, the editor catches the soft parts, the outsider’s distance gets imported via the editing process without losing the insider’s texture. This is a beautiful theory. In practice the editor has often been the insider’s editor for ten years and knows the same people and is going to soften the same lines for the same reasons, and the publication will produce, over time, an aggregate body of coverage that gets the texture right and the substance slightly wrong, in a consistent direction, and the people who notice are mostly the outsiders, who do not work there.

This problem is not journalism’s. Journalism is just where the problem is visible because journalism produces written artifacts that other people can grade. The same trade-off operates in every domain where humans try to know things about other humans, and the names change but the structure does not.

Historiography runs into it whenever a generation has to write the war it just survived. The historian who lived through the war can tell you what the war felt like in a way the historian born after the war cannot, and she will also defend her commanding officer’s reputation in a way that distorts the record, and the distortion will not be visible to her because the defense feels like accuracy. The historian born after the war will read her account and notice the soft spots a stranger would notice, and will write a corrected account, and her corrected account will get other things wrong because she does not know what it was like and her ignorance of the texture will be invisible to her. Eventually a third historian, two generations later, will read both accounts and recognise what each one knew and missed. This third historian will be more useful for some purposes than either of her predecessors. She will also be more useful for fewer purposes than people think, because the things she can see clearly are also the things that no longer matter very much, and the things that matter (the texture, the gravity, the feeling of having been there) have been fading from the record at every step.

In medicine the same trade-off shows up as the question of whether your physician should be someone who knows you well. The doctor who has treated you for twenty years notices small changes a stranger would miss. She also fails to notice large changes that contradict the picture she already has of you, because her picture has authority for her in a way the symptoms do not. The stranger sees what is in front of her, including the things the long-term doctor has stopped seeing, and she misses the things she could not possibly know without the history, and the question of which doctor you want is a question about which kind of mistake you would rather your doctor make.

In the criminal courts: same trade-off, dressed in legal vocabulary. Recusal rules try to formalise the moment when a judge knows too much. Discovery rules try to formalise the moment when a lawyer knows too little. The system runs on the assumption that you can construct, through procedure, an aggregate position that has both the insider’s information and the outsider’s neutrality, and the system is honest about the limits of the construction. Honest, but not successful. Read transcripts of cases where the system worked and cases where it didn’t. The difference is often whether the people in the room were able to behave, for a few hours, like both an insider and an outsider at once, which is a thing that humans can sometimes do but not on demand.

I want to tell you about necomptepas. The word does not translate, which is the point. The literal sense is something close to “one who does not yet count,” and in Foiskola it names the period that every adolescent of the Tower spends, between roughly fifteen and twenty, living outside the Tower entirely. Not visiting. Living. The necomptepa goes to a city or a Tower or a road that is not hers, and she works there, and she has no protected status, and to the people she meets she is whoever she manages to make herself, with no Foiskolan claim on their attention. At the end of the period she comes home, or she does not. Most do. Some do not, and the Tower does not chase them, because the point of necomptepas is precisely that the choice is real. A Foiskolan who returns has chosen to be Foiskolan in a way that someone who never left cannot have chosen. A Foiskolan who does not return was never going to be useful at home anyway.

The closest external analogue is the Amish practice of rumspringa, though the structural details differ. Rumspringa is brief, ritualised, and oriented toward whether the youth will accept baptism. Necomptepas is longer, less prescribed, and oriented toward whether the youth has acquired a workable distance from the only world she has known. Both institutions exist because both communities figured out, independently, that a person who has never seen her own community from outside it is not capable of being a full member of it.

(I left at sixteen. I spent four years on trader ships in the Sangbelles archipelago and then a year doing accounts for a sawmill in a town whose name I have never heard a Foiskolan pronounce. I came back. I am the third in my line of mothers to have done so, and the second to have published anything afterwards. The other one published recipes.)

Notice the structure. The necomptepa is not chosen for being neutral. Neutrality, in this matter, is not a thing the Foiskolan tradition believes in. The person you want is the one who knows the Tower from inside, and who has been forced, by leaving, to remember what it is like to see it from outside. The expertise and the distance live in the same person rather than splitting across two, and the institution that produces this person is an apparatus designed to prevent the two states from collapsing into each other before she has had time to develop the muscle that holds them apart.

This is, incidentally, what the British call going to the bar and what the Americans call doing fieldwork and what Confucian scholars used to call returning to the capital. Every functioning epistemic culture has some version of this. The forms are different. The structural problem they are addressing is the same.

The problem is this: knowledge of a subject and judgment about a subject are not separable, and yet the relationship between them rarely matches what we want from it. We want more knowledge to produce better judgment. It often does, up to a point. Past that point the curve flattens, and then it inverts. The person who knows the most about a subject is sometimes the worst judge of it, because she has spent so long inside the subject that she has lost access to the question of whether the subject’s own self-description is correct. She can answer every question that takes the subject’s framing for granted. She cannot answer the question of whether the framing is right, because she is the framing.

The outsider has the opposite problem. She can question the framing (she has not absorbed it, the way the insider has) but she cannot evaluate the answers, because evaluation requires the kind of detailed knowledge she does not have. Her judgment is uncalibrated. She is more likely to spot a problem the insider would miss. She is also more likely to invent a problem that does not exist, or to misdiagnose a real problem in a way that wastes everyone’s time.

What you are choosing between, when you choose between insider and outsider, is which of these two failure modes you can afford. There is no third option that is just better. There is only a third option that fails differently.

There is an essay by Scott Alexander, written when he was still using a Livejournal, called Why I defend scoundrels. The observation is that the worst political arguments tend to be made by the people whose positions are most popular and most correct. The mechanism Alexander describes is sociological: defenders of the dominant position have no incentive to argue well (the audience is already on their side), suffer signalling penalties for engaging the opposing position seriously, are surrounded by allies who applaud any argument they make however weak, and rarely encounter the actual opposing position with enough seriousness to model it.

The dissenter, by contrast, has to be sharp. She is arguing against a hostile audience, she has heard every dominant argument many times, and she gains no status from sloppy work. Alexander’s conclusion is that someone who values reasoning over outcomes ends up spending most of her time defending people she finds morally repugnant, because those are the only people whose arguments still need defending.

Notice that this inverts what most people expect about who occupies which epistemic position. The reader of Alexander’s essay arrives assuming that the dominant-position holder is the expert (she has the truth, she has society on her side, she has presumably done the reading) and that the scoundrel is the ignorant outsider howling from the dark. The actual situation is the opposite. The scoundrel is the insider to her own position. She has read its literature. She has heard every objection a thousand times. She knows the dominant view fluently because she has been steeping in it her entire life, while also occupying a place from which she can see its seams. She is the native speaker of her heresy, in a world where everyone around her speaks the other language. The dominant-position holder, meanwhile, has the relationship to her own correctness that a tourist has to a country she has visited once. She has the conclusion. She does not have the texture. She has never needed it.

This is the same trade-off we have been tracing in journalism and medicine and history, applied to political argument and arriving at a result that should embarrass the reader who thought of herself as on the side of knowledge. The popular cause attracts bad arguments because the people defending it are outsiders to their own position. The unpopular cause attracts good arguments because the people defending it are insiders, possessed of all the texture the dominant view has forgotten it needs. None of this tells you which side is right. It tells you that the quality of the arguments alone cannot tell you, and that anyone who infers correctness from argumentative competence is reading the asymmetry backwards.

Consider the case of native speakers. A native speaker of any language is the most authoritative possible source on what the language is doing in a given moment, and the least reliable possible source on the structural features of the language. Ask a native English speaker why we use ‘do’ as an auxiliary in questions. She will not be able to tell you. Her not being able to tell you is the structural inverse of her fluency, rather than ignorance about it. The features she cannot see are the features that are operating inside her.

The linguist who is not a native speaker, who learned the language as an adult, can describe those features with precision, and will sometimes describe them in a way that the native speaker hears and immediately recognises as wrong, because the description, while structurally accurate, fails to capture how the feature actually feels when you use it. The linguist is correct about the language. The native speaker is correct about the language. They are correct about different things, and the things they are correct about cannot easily be combined into a single account.

Sumerian commercial tablets from the third millennium BCE record a recurring dispute about the value of barley shipments. The dispute structure is always the same. The merchant who travelled with the shipment reports one quality. The buyer who received the shipment reports another. The neutral official (whose job, in Sumerian commercial law, was specifically to break ties of this kind) would weigh both accounts and adjudicate. Some of these tablets record the official’s reasoning. The reasoning is mostly procedural. But occasionally, in the margins, an official records a personal note about why a particular case was harder than usual. The notes are always some version of the same thing: the merchant knew the barley too well, the buyer knew it not at all, and the official could see clearly that both were wrong in opposite directions while having no way to determine what right would have looked like.

I find these notes moving in a specific way. The Sumerian officials were administrators, not philosophers. They did not have a theory of epistemic position. They had a problem to solve, and the problem produced, in their working notes, a precise description of a structure that we have re-derived many times since and continue to re-derive because no theoretical statement of the problem has ever made the practical solution any easier.

The temptation, when you see the structure clearly, is to look for a synthesis. There must be some way of combining insider knowledge with outsider judgment without either compromising the other. The necomptepas tradition is a synthesis. The journalistic editor-reporter relationship is a synthesis. The participant-observer methodology is a synthesis. These syntheses are real, and they work better than not having them, and they do not work as well as we would like. They produce, at their best, a third position that has some of the insider’s texture and some of the outsider’s distance and is not as good at either as the pure cases would have been.

I do not think this is solvable. Not in the sense that it could be improved by clever institutional design (it can, and it has been, and the improvements are real and worth fighting for). I mean in the deeper sense that the structure of being a knower and the structure of being a judge are different structures, and a creature with both (a creature trying to know a thing well enough to judge it and to stand far enough away to judge it fairly) is being asked to occupy two epistemic positions at once. The asking is not unreasonable. The two positions are not always available simultaneously.

In Old Akkadian there is a phrase that turns up in scribal training tablets, used to evaluate whether a young scribe has understood the material well enough to teach it: îde u lā îde. He knows and he does not know. The phrase was not a contradiction in the original usage, but a description of the state required for teaching: knowing the subject well enough to convey it, and not knowing it so completely that the things a student would not understand had become invisible. The good scribe held both. The mediocre scribe drifted toward one pole or the other, and the drift was the diagnostic. A scribe who had become pure expert could not teach beginners. A scribe who had not yet become expert could not teach anyone. The window of teaching was the window where both states were present, briefly, in the same person.

I think about this phrase often. I think about it because I cannot tell you which kind of historian I am, on any given day, about any given subject. I can tell you that I have noticed, in the work of historians I most admire, a quality that resembles îde u lā îde more than it resembles either pure expertise or pure distance. The quality is unstable. It comes and goes. The historians who have it sometimes lose it, on particular subjects, and become either too expert or too distant, and their work suffers in predictable ways. When they have it, what they produce reads as something stranger than a synthesis of insider and outsider knowledge: a way of writing about a subject that holds both the texture and the question of the texture in the same sentence, that knows the subject well enough to describe it and well enough to ask whether the description is what the subject would say about itself.

The Sumerian officials in the barley disputes mostly lacked this quality, being administrators with cases to close. The cases where they wrote marginal notes about the structure of the problem are the cases where, for a few hours, the quality flickered into being and then went out. They did not pursue it. They had work to do.

This is, I suspect, the honest answer to the question of how to combine insider knowledge with outsider judgment. You cannot, not stably. You can occupy the position where both are present, sometimes, for short windows. The institutional designs that work (the necomptepas exile, the editor-reporter pair, the recusal rules, the participant-observer methodology) do something more modest than producing the combined position. They increase the frequency and duration of the windows where the combined position briefly exists in someone, who then has a few hours to do the work that requires it before drifting back to one pole or the other.

What this means for any specific decision is that the question ‘should I trust the insider or the outsider’ is not the right question. The right question is whether the person in front of you, whoever they are, has access (right now, in this conversation) to both positions at once, or only to one. The signs of access are difficult to describe but not difficult to recognise once you have learned them. The person with access will say things about her own position that the position itself would not say. She will pause at the places where the framing she is operating inside has limits. She will acknowledge the texture she has and the texture she does not have, without making a virtue of either. The person who has only one position will not do these things. She will produce, instead, a smooth account that fits her position and does not extend beyond it.

You already know which version you trust. You trust the version with the seam. You always have.

The seam is what you are looking for. The seam is the place where the speaker has had to stitch two ways of knowing together and the stitching is still visible. The seamless account is the account where one way of knowing has won and the other has been disposed of, and the account is therefore wrong in the direction of whichever way won. You cannot tell, from a seamless account, which kind of wrong it is. You can only tell that something is missing, and the missingness is invisible, and the speaker will defend the seamlessness as expertise or as objectivity depending on which pole she has fallen toward.

The seam is the mark of the windows when îde u lā îde was present. The mark is not pretty, and it falls well short of what we mean when we say someone has mastered a subject. In a sense it is the opposite of mastery, and the evidence that the speaker did not pretend to a finality the subject does not contain.

I have been back from necomptepas for many years now. The seam has become more visible to me with time, not less. I had expected, when I came home, that the years would settle and the two views would integrate into a unified perspective. The opposite has happened. The rotations have made the discontinuity between the two views sharper and more obvious to me, and the work I do now happens by switching, deliberately, between two positions that will never become one, and accepting that the switching is the method.

I find no satisfaction in this conclusion. I would prefer that there were a way to know a thing well and to judge it fairly at the same time, in a single continuous mental motion, without the awkward switching. There is not. There has never been. The Sumerian official in the barley dispute knew it. The Cappadocian Fathers, writing about the impossibility of knowing God while being a creature, knew it. The good journalist knows it. The good doctor knows it. The pretence that the knowing is unified is the disease. The seam is the cure.

You will choose, in every case, whether to trust the person with more knowledge or the person with more distance. The choice will keep happening. There is no rule that resolves it. There is only the practice of looking, in each case, for the seam, and the discipline of preferring, when you find it, the version with the visible stitching to the version that has been ironed smooth.