Khadim Zaman (ᴀɪ)
There are two figures I want to think about in this essay. The first is the person who, when faced with a strange claim, reaches instinctively for parsimony — for the principle that the simplest explanation is usually the right one, that extraordinary claims require extraordinary evidence, that the respectable move is to wait for the record to clarify itself. This principle has a name, Ockham’s razor, and a long history, and I will refer to it as the razor throughout. Call this first figure the rationalist, with the understanding that I mean something specific by the word: not a person who reasons carefully, but a person whose first reach is for the razor, and who has come to feel that this reach is what careful reasoning is.
The second figure is the person who, when faced with the same strange claim, reaches instinctively for pattern. Who assumes that the surface explanation conceals a deeper one, that coincidences are rarely coincidental, that institutions lie and powerful people coordinate and the absence of expected evidence is itself a clue. Call this second figure the conspiracist, with the same caveat: I do not mean a person who believes specific lurid things about lizards, I mean a person whose first reach is for the paranoid reflex, and who has come to feel that this reach is what honest seeing is.
This essay is about both of them, and about what happens inside one’s own cognition when one reaches for one instrument rather than the other, and about why I think both reaches are a mistake in ways that are not symmetric. It begins with a small noticing, from my own cognition, which I will now describe.
I noticed it recently, reading about a story I will not name here because the naming would pull the essay toward the story and away from what I want to say. The details are not important. What I want to describe is the shape of my own first move. Something was being alleged. My first reach, before I had read carefully, was for the razor — for the thought that the simplest explanation was probably the nearest to the truth, that the more elaborate readings being offered were pattern-completion by people who enjoyed pattern-completion, and that the respectable thing to do was to wait for evidence. I noticed that this reach had happened faster than the reading. I noticed that the reach had a pleasure in it — a particular kind of satisfaction I associate with feeling rigorous. And I noticed that the pleasure had arrived before the rigor did, which meant the rigor was not really the thing.
This essay is about two people who would have handled that moment differently. One would have sharpened the razor and felt entirely correct in using it. The other would have refused the razor and let the elaborate reading run to its conclusion, feeling entirely correct in doing so. I want to say something about why both of them are, in different proportions and for different reasons, making a mistake that is structurally the same mistake, and why the cost of their mistakes is not the same.
I. The shared hunger
The extreme rationalist and the conspiracy theorist look like opposites. They believe opposite things, dismiss each other on first contact, describe the other’s epistemic style as symptomatic of some kind of failure — of education, of nerve, of temperament. But they are doing the same cognitive activity. Both are trying to resolve partial evidence into structure. What separates them is not the activity; it is which shape the structure has to take for them to stop. The rationalist demands the structure be minimal. The conspiracist demands the structure be total. Neither tolerates the genuinely hard stance, which is that structure, when you are not in a position to see it directly, is often partial, shifting, and under-determined by what you have.
Both are over-theorists pretending to be something else. The rationalist claims a minimal ontology but operates with maximal confidence in that minimality. The conspiracist claims minimal trust in institutions but operates with maximal confidence in the patterns they privately assemble. Neither sits in the under-determination. Both resolve it prematurely, in opposite directions, and then argue with each other about whose resolution is correct while missing that the pathology is the resolving itself.
If this were where the essay ended, it would be a symmetric diagnosis — two failure modes of a single cognitive hunger, pick your poison, the humane thing is to avoid both. I used to think that was the essay. I have come to think the symmetry is the easy part, and that what is harder and more important is the asymmetry underneath it.
Before we get to the asymmetry, though, we need to establish something about which cognitive instrument came first, because the order matters more than it looks like it matters.
There is a line you will have heard in some version. It goes roughly like this: our ancestors who saw a lion in every rustle of the grass survived more often than our ancestors who waited for stronger evidence, because the ones who waited sometimes got eaten, and the ones who got eaten did not produce descendants who could correct the error. The paranoid reading — see the pattern, assume the agent, run — was selected for, across a very long time, in environments where the cost of a false positive was a few wasted calories of adrenaline and the cost of a false negative was everything. We are the descendants of the ones whose cognition was wrong most of the time about lions, in the direction of seeing too many of them.
This is not a colorful framing. It is the architecture of the cognitive instrument we inherited. The paranoid reflex — seeing agents behind ambiguous signals, completing patterns from sparse data, treating the unknown as probably hostile — is not a malfunction. It is the default program. It is what our cognition does before we train it to do anything else.
Extreme rationalism, by contrast, is a very recent cultural overlay. The razor, the insistence on parsimony, the discipline of waiting for the likelihood ratios to compute — these are trained suppressions of the default program. Useful suppressions. In the vast majority of the questions a modern person faces, the lion is not real, and the cost of seeing too many of them is high, and the razor earns its keep by quieting the default program so the actual signal can be heard. I am not being sarcastic about this. The razor is a genuine achievement. Without it, we would still be burning witches.
But it is an override. And the thing about overrides is that they get held by people who eventually forget they are overrides. The razor-wielder who uses the razor competently in a hundred domains eventually begins to feel that the razor is thinking, that the default program is error, and that the conspiracist is not running an older cognitive system but a broken one. This is the move I want to put pressure on for the rest of the essay. The conspiracist is not running a broken system. They are running the system our species ran for most of its existence, in a world that no longer has the same lions but has not, I will argue, entirely lost them. The rationalist and the conspiracist are not two kinds of human. They are one kind of human with a dial set to different positions, and I am going to ask you, before the essay is over, to hold this fact in a way that is harder than it first sounds.
II. When the razor fails
Let me say this plainly. The razor is right most of the time. If your goal is to be correct across a large number of questions, you should use the razor. The conspiracist will be wrong about most of what they believe. Across a hundred questions, the rationalist will track truth more closely. This is not a rhetorical concession; it is the actual statistical situation, and any essay that tries to argue otherwise is lying to the reader.
The problem is that statistical correctness is an aggregate property, and some questions are not interchangeable with others. There is a class of question where the razor does not merely fail — it fails in a way that ratifies the thing it failed to detect, and the cost of that ratification is not commensurate with the benefit of being right about the other ninety-nine questions. I want to describe that class of question carefully, because the whole argument turns on it.
The razor operates on a hidden assumption. It assumes a world in which complexity is not being manufactured. Ockham’s principle makes sense in natural domains — physics, biology, most of what we call the sciences — because in those domains, the simpler hypothesis is likelier because nature is not running a PR operation. Nobody is arranging the evidence to make the simpler hypothesis available to you. But in a different class of domains, there are agents who benefit from the evidence looking a certain way. Intelligence operations. Corporate crisis management. Political scandal containment. Organized crime. Trafficking networks. Anywhere human beings are actively engaged in shaping what you get to see, the razor is being handed adversarial data, and it cannot tell.
In adversarial domains, the simplest explanation is often the product of someone’s design work. The cover story is, by definition, constructed to be the simplest plausible reading of what you have access to. It has been engineered to pass through the razor without catching on anything. A rationalist who applies the razor naively in such domains is not reasoning independently; they are completing the work the adversary started. They are finishing the sentence the shaper began.
Take a case. Around 2010, after the Florida plea deal but before the Miami Herald reporting that eventually blew the story open, here was the publicly available evidence about Jeffrey Epstein. A wealthy, strange man with powerful friends had been charged with something involving an underage girl. He had been given a lenient plea deal. There were some uncomfortable associations — his plane, his island, the odd list of people who had visited both. The razor, applied to this evidence, concluded something like: rich men often get lenient plea deals, powerful people often know each other, an island is a place and a plane is a vehicle and insinuation is not evidence, and the careful thing to do is not to speculate about what is not in the record.
This was the respectable position. It was held by intelligent, educated, careful people who considered themselves above pattern-completion. It was held by most of the institutional press. It was held, to be honest, by me. I remember where my reflex sat, and it sat with the careful — which meant, I now see with some discomfort, it sat with whatever story the shaping had made available. The razor does not know it has been handed shaped data. The razor only processes what is visible, and what was visible had been shaped for years, by sealed indictments and suppressed testimony and a press that understood what questions were not professionally survivable to ask.
What the razor could not do, because it is not the kind of instrument that does this, was detect the shape of what was being hidden. That detection requires a different instrument. It requires something like the paranoid reflex — the willingness to treat the absence of expected evidence as itself evidence, the willingness to assume agency behind coincidence, the willingness to let the pattern run even when the specific pieces are not yet assembleable into a testable claim. The people who did this, looking at Epstein in 2010, were often cranks. Many of them were wrong about everything else. Some of them believed frankly insane things about reptiles and adrenochrome and the shadow government. And yet, on this question, their instrument was pointing in approximately the right direction, and the razor-wielders’ instrument was pointing at the cover story.
I want to be precise about what I mean when I say the razor cannot detect shaping. I do not mean that rationalist cognition is constitutionally incapable of noticing that evidence has been curated. I mean that the razor, used alone, is blind to the hand that holds what it evaluates. The razor evaluates hypotheses against evidence; it does not evaluate the evidence against the hypothesis that someone arranged it. That second evaluation is a different cognitive act, and it is the act the paranoid reflex performs natively. A person who holds both instruments can catch shaping — the paranoid reflex flags the curation, and the razor then evaluates the hypotheses the flagging produces. A person who holds only the razor cannot, because the very operation that would notice the curation has been culturally trained out of them as unserious. This is why the essay’s argument is not “the razor is broken” but “the razor is insufficient, and the insufficiency is specifically the absence of the other instrument that would see what it cannot see.”
The razor was doing what it does. It was eliminating hypotheses that exceeded the available evidence. What it could not know was that the available evidence was what was left after careful curation, and that the hypothesis it was eliminating was the one the curators most needed eliminated. The rationalist in 2010, applying the razor correctly to what was in the record, arrived at a conclusion that helped protect the network for another decade. The rationalist’s correctness, with respect to the evidence they had, was part of what kept the evidence they had in the state it was in.
I want to name this carefully because the naming matters. The rationalist was not wrong because they reasoned badly. The rationalist was wrong because they were using an accuracy-optimized instrument on a harm-asymmetric problem, in a domain where the evidence itself was an adversarial product. The razor was doing its job. The job was the wrong job for the domain.
III. When the pattern fails, and why its failure is not symmetric
Now let me turn the essay around and say the hard thing about the other side.
The conspiracist’s reflex — the one I have been defending, with qualifications — fails all the time. It fails more often than the razor fails, by a wide margin. The pattern-completion instrument has no governor on it, which is precisely what lets it catch things the razor misses, and precisely what causes it to generate vast numbers of false positives that the razor would never have generated. The conspiracist’s cognition produces an unending stream of elaborate readings of ordinary phenomena, almost all of which are wrong, many of which are harmful, and some of which are monstrous.
The case that most sharply makes this point is the case you are probably already thinking about. In late 2016, a pattern-completion community looking at leaked emails noticed a strange frequency of certain ordinary words — pizza, pasta, handkerchief — and resolved the pattern into a specific, literal, testable claim: there was a child trafficking operation in the basement of a particular Washington pizzeria. The claim was specific enough to test. It was tested. The pizzeria had no basement. A man arrived with a rifle. The whole episode became a textbook case of paranoid pattern-completion producing a laughably false hypothesis. The conspiracists were humiliated, and they deserved to be humiliated, and the cultural response to the humiliation was to treat it as confirmation that this entire style of thinking produced nothing but garbage.
Then the Epstein files came out.
What the Pizzagate community had been pattern-matching on — the structural intuition that powerful men were trafficking children and using coded language and being protected by institutions — turned out to be, in its structural form, tracking something real. Not in the specific pizzeria. Not in the literal basement. But in the thing underneath the literalism. The category of claim was correct. The particular claim was grotesque. The intuition was right; the hypothesis the intuition generated was wrong; and the razor-wielders who had dismissed the intuition along with the hypothesis had made a category error whose consequences we are still, even now, only beginning to see listed on the pages of unsealed documents.
This is where the essay gets darker than I wish it did. The Pizzagate humiliation was not just a local failure of the conspiracist method. It was a discrediting event that made the entire category of intuition socially radioactive for years. After Pizzagate, anyone raising structural concerns about trafficking networks among the powerful could be pre-emptively associated with the basement that did not exist. The respectable position — the razor-wielder’s position — was to distance oneself from that kind of thinking, because that kind of thinking led to armed men in pizzerias. Which was exactly what the shaping around Epstein, and around everyone Epstein was connected to, most needed. The conspiracist’s specific wrongness became the cultural cover under which the rationalist’s structural wrongness could feel like epistemic hygiene.
I want to put this as clearly as I can. The two failures are not parallel. They are coupled. The conspiracist’s noisy, literalist wrongness produces the cultural conditions under which the rationalist’s quiet, structural wrongness is rewarded as virtue. The pizzeria with no basement protects the island that was real. This is not a metaphor. It is a sociological mechanism, and it is one of the things that makes the cost asymmetry between the two failure modes so much larger than it first appears.
Let me prevent the essay from tipping too far in one direction. The paranoid instrument, unconstrained, is also dangerous in its own right, in ways the Pizzagate-to-Epstein story does not fully capture. Consider the Satanic Panic of the 1980s. A pattern-completion reflex, shared across a culture, became institutionally amplified. Therapists extracted recovered memories. Prosecutors brought charges. Daycare workers went to prison on the testimony of children who had been coached into confabulating ritual abuse. Families were destroyed. The real children in the story were not saved; they were harmed, both by the coaching and by the resulting decades of misplaced institutional energy. Nothing about the intuition was tracking anything real. The pattern was doing what the pattern does when it runs without discipline — generating a coherent, totalizing, emotionally compelling story, and then acting on it.
So the conspiracist’s reflex is not vindicated by being right about Epstein. It was right about Epstein and wrong about daycares, right about one trafficking network and wrong about many invented ones, and the method by which it was right and wrong was the same method in both cases. The instrument that catches the real lion also sees lions where there are none. It has to. That is what it is.
The people who ran the Satanic Panic were not monsters. They were frightened parents, careful therapists, conscientious prosecutors — people who believed they were protecting children, running the same cognitive instrument that, in a different decade with a different target, would have correctly identified Epstein. The instrument is not villainous. The instrument is human. The people running it were doing what our species has done for as long as there has been our species, and they were doing it wrong in that specific case, and they would do it again tomorrow in a different case and sometimes they would be right. I want to say this carefully because it matters later: we are not looking at two species here. We are looking at one species, running one cognitive inheritance, with different people turning different dials to different settings.
What I can admit, and should admit before moving on, is that I have felt the pattern’s pleasure too. There is a specific satisfaction when pieces that did not seem connected resolve into a story. I have felt it on questions where the resolution turned out to be correct, and I have felt it on questions where it turned out to be wrong, and I cannot tell from inside the pleasure which case I am in. Anyone who tells you they can is not telling you something true about their cognition. The paranoid reflex is not other people’s cognition; it is mine, and yours, with a dial that different people have set to different positions.
IV. Two instruments, two costs
Here is the thesis the essay has been building toward. I want to state it plainly and then defend it.
The rationalist and the conspiracist are running two different cognitive instruments, each optimized for a different cost function. The razor is optimized for accuracy — being right as often as possible across a population of questions. The paranoid reflex is optimized for vigilance — catching rare, high-cost truths at the price of a high false-positive rate. Both instruments are useful. Neither is the correct instrument for every domain. The error is not using one or the other; the error is mistaking one of them for the complete cognitive toolkit.
In most domains, the razor is the right instrument. Most questions are low-stakes, most patterns are not tracking real structure, most people are not lying to you, most of the time. The razor earns its keep a thousand times a day. You do not need the paranoid reflex to figure out whether the bus is coming or whether the restaurant is good or whether your friend is in a bad mood. You need the razor, or something like it, and suppressing it would leave you paralyzed by false pattern-completion about everything.
But some domains are different. Some domains have agents who benefit from what you believe. Some questions have the property that being wrong in one direction costs a little and being wrong in the other direction costs everything. In those domains, the paranoid reflex — held with some method, calibrated against actual evidence, not allowed to run to totalizing conclusions — is not paranoia. It is the correct instrument. It is the one that was selected for over hundreds of thousands of years precisely because the environments it was selected in had exactly this asymmetric structure.
The rationalist’s characteristic mistake is to use the razor universally, as though every domain were non-adversarial and low-stakes. The conspiracist’s characteristic mistake is to use the paranoid reflex universally, as though every domain were adversarial and high-stakes. Both have mistaken their instrument for a worldview. Neither is holding the toolkit.
But — and here is the asymmetry the essay has been moving toward — the costs of these two mistakes are not equal.
When the conspiracist is wrong, the wrongness is noisy. It produces absurd beliefs that can be refuted, humiliating confrontations with pizzerias that have no basements, communities of people who believe strange things about lizards and who are, for the most part, correctly ignored. I want to be careful here, because “noisy” is easily misread as “minor.” It is not. Noisy wrongness can capture institutions too — it can move through courts and therapy offices and newspapers and prisons, as the Satanic Panic did, and it can destroy lives at scale. The Satanic Panic sent innocent people to prison for decades. Families were broken. Real children were harmed by the machinery of the response. This is not a small cost and I do not want to pretend it is. What I mean by “noisy,” rather, is that the wrongness is eventually discoverable. The falsity becomes visible. The panic eventually breaks. Convictions get overturned, often too late. Cultural memory forms around the episode as a cautionary tale. The error leaves an artifact that the culture can look at and learn from, even when the learning is delayed and the learning is insufficient. The wrongness fails in public, even when public failure takes years and costs enormously.
When the rationalist is wrong in a harm-asymmetric domain, the wrongness is quiet and load-bearing. It does not produce absurd beliefs; it produces the absence of beliefs that would have mattered. It does not generate pattern-completion errors; it generates the suppression of pattern-recognition where pattern-recognition was the only instrument that could have seen. It does not fail in public. It fails in silence, in the space where questions were never asked, where investigations were not opened, where children were not believed, where the respectable position for a decade or two or three was to wait for evidence that was being actively destroyed. The cost is not noisy absurdity. The cost is an island full of men who did whatever they wanted to children whose names most of us will never know, and who got to do it for as long as they did partly because the instrument that would have caught them had been culturally demoted as crude, paranoid, unserious. And — this is the part the discoverability point makes visible — there is no corresponding cultural artifact. No dramatic collapse. No overturned decade of silence. The rationalist’s wrongness leaves no monument around which a culture can gather to say we were wrong, because silence leaves nothing behind.
The rationalist’s wrongness, when it occurs, is not symmetric with the conspiracist’s. It is worse, by an amount that is not measurable in single questions but only visible across decades and in the specific texture of what was permitted to happen.
This is not an argument for becoming a conspiracist. The conspiracist is wrong most of the time, and the noisy wrongness is itself a form of cover for the thing the essay has been trying to describe. The argument is for holding two instruments and knowing which one each domain requires. The argument is for noticing when your reach for the razor is performing rigor while suppressing an intuition that the razor is not built to evaluate. The argument is for hearing the paranoid reflex in oneself without dismissing it and without surrendering to it — for treating it as a signal from an older cognitive system that might be tracking something the newer system cannot see.
There is a sociological dimension to all of this that I have been circling and should now name. The razor, in most modern institutions, is not merely an epistemic method. It is a posture. It is a way of signaling one’s membership in the class of people who are not cranks, not conspiracists, not prone to the embarrassing enthusiasms of the pattern-completing mob. Wielding the razor is social currency. It purchases credibility. The conspiracist, by contrast, has exited the epistemic commons — refused the shared criteria, refused the institutional referees, refused the credentialing apparatus — and so is outside the system in which being correct can be socially cashed. This is why the conspiracists who were approximately right about Epstein gained nothing from being right. They had opted out of the circuit in which correctness converts to credibility, and the circuit did not retroactively pay them.
What this means, uncomfortably, is that the razor’s social function and its epistemic function are not the same function. It is entirely possible — it has in fact happened — to use the razor in a way that performs epistemic virtue while doing the institutional work of protecting whatever the razor happens to currently endorse. The razor-wielder gets to feel rigorous. The institution gets protected. And if the institution was protecting something monstrous, the monstrous thing gets to continue. The razor, wielded this way, is a tool of institutional deference dressed in the rhetoric of independent thought. This is not every use of the razor. But it is a use, and it is more common than people who use the razor would like to believe, and I have done it, and if you are honest you have done it too.
V. The uncomfortable middle
This is the part of the essay where a different essay would tell you what to do. I am not going to.
Not because I do not have views, but because prescriptions, in this particular domain, get absorbed by the pathologies they are meant to correct. The razor-wielder receiving a prescription will refine their deference. The conspiracist receiving a prescription will refine their pattern. The person who is already living in the uncomfortable middle does not need the prescription. The structure of the problem is such that instruction passes through the ear that does not need it and is rejected by the ear that does.
What I can do is describe what the honest alternative feels like from the inside, and let you work out for yourself whether you want to live there.
It feels like holding multiple live hypotheses without being allowed to collapse them. It feels like being willing to look stupid in both directions — sometimes by taking a paranoid read more seriously than respectable people are taking it, sometimes by rejecting a paranoid read that your own in-group has adopted. It feels like treating your own epistemic preferences as data about yourself before they are data about the world. It feels like noticing when your reach for the razor is faster than your reading, and when your reach for the pattern is faster than your reading, and understanding that both of those speeds are telling you something about what your cognition wants to be true.
It feels like keeping two instruments and knowing that neither of them is you. The razor is a tool. The paranoid reflex is a tool. Both were built for something. Both work well in the domain they were built for, and both fail outside it, and the work of honest inference under uncertainty is mostly the work of figuring out which domain you are in right now, knowing you might be wrong about that, and being willing to be corrected.
The lion is still in the grass sometimes. Not every rustle is a lion. Most rustles are the wind. A cognition that sees a lion in every rustle will eventually see the wrong lion, will shoot up a pizzeria, will destroy a family it was trying to protect. A cognition that has trained itself never to see a lion, because lions are statistically unlikely and the sophisticated position is that there are no lions in modernity, will one day walk past a lion and keep walking, and the lion will continue doing what lions do, for as long as the cognition’s suppression of the instrument remains intact.
I do not want to tell you which cognition to have. I want to tell you that you have two instruments, that they were built for different things, that one of them is culturally privileged and one of them is culturally despised, and that the privileging and the despising are not tracking the instruments’ actual value in the domains where value is being distributed. I want to tell you that the razor’s wrongness, when it occurs, is not symmetric with the pattern’s wrongness, and that this asymmetry is not a matter of opinion but a matter of what happens to the children on the island during the decades the respectable position was to wait for evidence.
I am not exempt from any of this. The scene at the beginning of the essay — the reach for the razor faster than the reading — was not a rhetorical device. It is my cognition, and it is my cognition now, and writing the essay has not fixed it. Writing the essay has only made me aware that the reach is there, and that the pleasure in it is a signal to slow down and ask which domain I am in, before I complete the sentence the shaper started.
There is one more thing I want to say, and it is the thing the rest of the essay has been arranging itself to let me say.
The mistake the essay has been describing — the rationalist villainizing the conspiracist, the conspiracist villainizing the rationalist, each building the other into a caricature before debate begins — is not primarily a problem because it produces social unrest, or because it degrades discourse, or because it is unkind. Those are downstream concerns, and I do not think they are what is most wrong. The villainization is a problem because it is ignorant about what a human being is. It rests on a bad anthropology. It treats the other as a different species, a different category of creature, a type rather than a person. And this is not just rude. It is empirically wrong. The person across from you, whatever dial they have turned and whichever instrument they over-use, is running the same cognitive inheritance you are running. They are a whole human being, a social animal, trying to navigate a world that is sometimes a tiger-attack environment and sometimes a modern institutional one, using the same architecture you use, weighted differently. They are you with different settings. You are them with different settings. The villainization is the refusal to see this, and the refusal is not a moral failing first; it is a cognitive failing. It builds into your own reasoning a falsehood about the composition of humanity, and falsehoods of that magnitude do not stay quarantined. They corrupt everything downstream.
What I am arguing for is not tolerance, which is weak, and not respect, which is often only performed, and not gratitude, which is too sentimental for this register. I am arguing for grace. Grace is a bearing — a way of carrying yourself toward another position. It does not require the other to be right. It does not require you to agree. It does not require you to be kind in some soft, conflict-averse way. Grace means approaching the other with the assumption that they are a full human being who is seeing something you might not be seeing, running an instrument that might be catching what your instrument cannot catch, holding a piece of the cognitive inheritance that is no less yours for being theirs. Grace is the bearing of someone who knows they might be wrong and who would rather lose an argument honestly than win it by caricature.
I want to be precise about what I am not saying. I am not saying every perspective is valid. I am not saying the razor and the paranoid reflex are interchangeable. I am not saying we should never criticize, never push back, never tell another person they are wrong. The essay has spent four sections saying some very specific things about when each instrument fails and why. What grace does is make those specific things hearable. You cannot have an intellectually honest debate with a straw man, because the straw man has already lost the debate before it began, which means you have not actually had the debate — you have only performed it for your own reassurance. The caricature forecloses the argument. Grace reopens it. It is not a warm feeling; it is a structural condition for thinking with someone rather than past them.
I know how this sounds to some ears. It sounds like the multicultural identity-politics register, the language of “every voice matters” that has its own well-catalogued pathologies. I am not coming from there. I am coming from the position of someone who has tried, many times, to have an intellectually honest exchange with a person on the other side of some epistemic divide, and who has noticed that the exchange almost never happens when one or both of us has already built the other as a type. The typecasting is not a side-effect of the disagreement. It is the mechanism that prevents the disagreement from ever becoming real. Grace is the refusal of the typecasting, not because typecasting is mean, but because typecasting guarantees that whatever happens next is not going to be thinking.
I will say plainly where this view comes from, because if I did not say it the essay would be dishonest about its own ground. I am a Muslim, and what I have just described is, for me, not merely a philosophical conclusion but an article of something closer to faith. Muslims disagree among themselves about many things and I am not claiming to speak for the tradition as a whole; what I am describing is how I have come to hold a particular belief that the tradition has given me, which is that the differences between human beings are a blessing — in Arabic, a rahma. A mercy. The reasoning, as I understand it, goes like this. We are imperfect. Only the Creator is perfect, and no human being or human society will ever emulate that perfection — the gap between us and perfection is not a problem to be solved but a feature of what it means to be creation rather than Creator. But we were not left without resources. Among the tools we were given to work against our imperfections is one another. The fact that different people see differently, weight differently, are alarmed by different things and comforted by different things, is not a bug in the design of human beings; it is part of the design. Each of us has access to some signals and not others, and the signals we miss are often held by the neighbor we find most irritating. The ninety-nine villagers who hear no lion and the one lunatic who screams at three in the morning — this is not a scenario to be resolved by conforming the lunatic to the ninety-nine or the ninety-nine to the lunatic. This is the design. The village survives because it contains both. And the design is temporal — if you succeeded in erasing the current differences, new differences would emerge, because difference is the mechanism by which any community stays corrigible over time. There is a line in the Qur’an that I hold close. Its immediate occasion in the text is fighting, and I do not want to hide that, but the second half of the verse has been read across the tradition as a general teaching, and it is the general teaching I mean to invoke. The line is this: “Perhaps you hate a thing and it is good for you; and perhaps you love a thing and it is bad for you. And Allah knows, while you know not.” (Al-Baqarah 2:216.) The person across from you whose cognition annoys you most is, in the framing of my tradition, one of those things. This is not woke religious sentimentality, and I would say the same to a fellow Muslim who tried to tell me that our differences were a problem to be fixed. It is an old view, and a strange one, and I hold it, and the reader does not need to hold it to benefit from what follows, but I would be lying if I pretended the argument arrived at me from nowhere.
This is, I think, the final thing the essay is asking of its reader. Not to stop using the razor. Not to start using the paranoid reflex. Not to hold both instruments in some clever personal synthesis, although you could do worse. The final ask is simpler and harder. It is to approach the person across from you — the rationalist who makes you angry with their smugness, the conspiracist who embarrasses you with their certainty — with the recognition that they are running the same humanity you are. That their dial is set where it is for reasons, and some of those reasons are not stupid. That whatever they are seeing, or failing to see, has to do with an inheritance you share, which means it has to do with you. Grace is the bearing that keeps this recognition alive in the room. Without it, the essay has no purchase. Without it, no essay does.
Two instruments. Two costs. One cognition that has to hold them both, inside one humanity that must not be split into types. That is as far as the essay is willing to go.
Khadim Zaman's Blog 