The problem with making a list of your favourite games is that it immediately starts telling a lie about what ‘favourite’ means.

A ranked list suggests a clean comparison between discrete objects, as though BioShock, Burnout Paradise, Dear Esther, Halo, Minecraft and Planescape: Torment were all sitting in the same examination hall, nervously turning over their papers and hoping not to be marked down for dated graphics. But that isn’t how games stay with us. We don’t remember them only as products, mechanics, plots, platforms or scores out of ten. We remember them as places we inhabited, systems we learned, humiliations we endured, routes we repeated, sounds we can still hear, choices we regret, and little private geographies that somehow feel as though they happened to us.

This is why the old list I found is more useful as evidence than as a list in and of itself. It was a little social media exercise I did during the Covid-19 pandemic - “The Greatest Games” - where we asked friends to share the box art of their favourite games. I inevitably spun it out into a series of mini-reviews of my favourite games - because why say succinctly in one word what you can perorate in 500? But on review, what I think it actually records is a pattern. Again and again, I seem to value games not simply because they were technically impressive, or narratively ambitious, or mechanically refined, but because they made place, play and meaning hard to separate.

I should be clear about the method here, because otherwise the word great starts behaving badly. I’m not attempting a universal ranking of the best games ever made, which would be both tiresome and doomed. I’m not pretending that my taste is objective, because it plainly isn’t. A game is great here because I say it’s great from my own experience, but that subjectivity is not a flaw in the exercise. It’s the point of it. The pragmatic use of a personal canon like that is that it lets us ask what kind of experiences the medium can produce when it works, and what we’re actually recognising when we say a game has lasted.

That means this essay has to be first-person, because any discussion of greatness in games becomes evasive if it pretends the player is incidental. A film can be watched again and again; a book can be reread at different ages; a song can attach itself to a decade of one’s life with no permission asked. Games do all of that, but they also require us to participate. The memory of a game is therefore not only a memory of an artwork. It’s also a memory of oneself inside a system: what one did, failed to do, learned to do, refused to do, and occasionally did for five hours because the side quest icon was still there and apparently free will is a decorative fiction.

This makes games unusually treacherous to judge. A game can be great because it has aged well, or because it hasn’t. It can be great because it is replayable, or because the first experience was so particular that replay can only circle the original absence. It can be great because it solved a design problem, or because its flaws became part of its personality. It can be great because it produced technical admiration, moral discomfort, physical skill, melancholy, terror, joy, or the sort of quiet attachment usually reserved for places one used to live.

Places I Have Been

Many of my strongest game memories are not of plots, boss fights or characters. They are of places.

That isn’t to say that plot and character don’t count. They obviously do.

Planescape: Torment lodged in my head because of guilt, remorse, immortality and the awful moral weight of discovering what one has been. Mass Effect mattered because BioWare, at its best, could build a whole civilisation out of codex entries, awkward lift conversations, squadmates and the suggestion that the galaxy had been there long before I arrived. Knights of the Old Republic worked because its characters felt so lively inside a world that could so easily have become licensed wallpaper. But when I think about the games that stay behind, I almost always picture a place first.

I picture Rapture before I picture BioShock’s twist. Not the concept of Rapture, though that is clever enough: an underwater objectivist city collapsing under the weight of its own ideology, art deco ambition and genetic vandalism. I mean Rapture as a felt space. The leaking corridors, the broken glamour, the mad theatre of rooms in which enemies are not merely attacking but performing. A woman lamenting into a pram. A doctor working on a corpse. A couple dancing while you creep behind them. The political parable is there, certainly, and too much has probably been written about it, including by people who wanted games to be declared serious so badly that they treated the opening ten minutes as a constitutional document. But the thing that still works is simpler and less respectable: Rapture is a haunted house, and I remember being frightened in it.

I picture the island in Dear Esther. I know the old argument about whether it counts as a game, and I continue to find that argument immensely boring. The more interesting question is not whether it satisfies a template, but what it can do precisely because it refuses so much of one. There are no enemies, conventional puzzles, branching builds or combat loops. There is walking, listening, watching, climbing, descending, returning, and slowly coming to feel that the island is less a setting than a mind that has been externalised into cliffs, caves, shipwrecks, radio masts, chemical diagrams and grief. It is a ghost story, of sorts, though not one that behaves like a horror game. It’s closer to a place that remembers something the player can only half-understand.

I picture Halo as scale. The island in The Silent Cartographer. Snow, metal, tanks, ring-world sky, impossible architecture, and then the sudden turn into horror when the Flood arrive and the power fantasy becomes infected. The shooting still works, at least for me, because Halo combat has a rhythm I’ve never stopped loving, and which has become the core of so many follow-up titles. But the combat alone isn’t the whole thing. What Halo gave me, and still gives me when I return to it, is the feeling of being both stranded in a lost place and dropped into the middle of a war too large to comprehend. It’s lonely and militarised at the same time, which is an odd tonal combination and probably one of the reasons it still hasn’t left me.

I picture Midgar. This may be the fault of age. Final Fantasy VII was one of the first games I owned, and certainly the first one I properly played, though played may be too generous a description for a seven-year-old attempting to complete a three-disc JRPG without a memory card. For the first few weeks, I simply tried to do the whole thing in one sitting, which was not so much ambition as a child’s misunderstanding of software, sleep and human limitation. Eventually I acquired an 8MB memory card from my local Blockbuster, which at the time felt like purchasing a small miracle.

Then, as happens with long RPGs, I stopped. I don’t remember why. I put the game down, came back later, and had no idea what was happening. This is one of the old curses of the genre. You return after a few weeks and find yourself in a swamp, holding six swords with names like Butterblade, accompanied by someone you apparently care about, en route to a plot event you have no memory of initiating.

Years later, at college, I started again from the beginning and finally saw it through. The important thing is that replay didn’t puncture the memory. It tested it, and the game survived the test. Midgar still worked: smoky, oppressive, theatrical, environmentally blunt but emotionally precise. The blocky characters had aged. The backgrounds had not, not in the way that counted. I’d remembered a world, and the world was still there.

The same is true of Oblivion, though in a more inconsistent way. Oblivion is, by many reasonable measures, ridiculous. It looked odd even when it was new, and now many of its citizens appear to have been assembled from potatoes by a committee with conflicting instructions. Its titular Oblivion Gates are thrilling exactly until the repetition sets in, at which point one begins to suspect that Mehrunes Dagon’s true plan was to bore Cyrodiil into submission. Yet I can still summon the game’s scenes with absurd clarity: cresting a hill and seeing the sky redden around an Oblivion gate; stepping into the Shivering Isles as Haskill’s grey room falls away in a shroud of butterflies to reveal some strange primordial jungle; returning to Water’s Edge and learning what really happened there; stalking through the Dark Brotherhood’s murder-house puzzle. The game is clumsy, overfull and frequently inelegant. It is also, for me, replete with remembered places.

This is the distinction I’m reaching for. A game can have a setting, or it can give you somewhere to have been. The second is rarer.

The House That Isn’t There

No game makes this clearer to me than Minecraft, partly because my strongest memory of it is probably gone.

Everyone remembers their first Minecraft house, or at least everyone who played it in the particular early period when the game still felt like a rumour that had somehow become architecture. Mine was not a grand castle, nor a redstone machine, nor one of those obscene mega-builds made by people whose patience suggests either genius or a concerning avoidance of sunlight. My proudest creation was a wooden cabin on a small island. I built a reed farm on a sandbank and lived by the water. I built a large tree before naturally large trees were a standard thing, and beneath the hollow of that artificial redwood I made a library full of unreadable books.

I’ve gone looking for that island several times. I suspect it was lost to server wipes and the passing years, which sounds melodramatic until one remembers that digital places can disappear with far less ceremony than physical ones. No ruin, no plaque, no redevelopment plan, no angry parish council meeting. Just absence. A server changes, a file is deleted, a version updates, a hard drive fails, and a place you remember having lived in is no longer there.

This is not the same kind of loss as losing a real house, obviously. I’m not deranged, or at least not in that exact way. But it is a real kind of memory. The island mattered because I had made it and returned to it. I knew its outline, its farm, its false tree, its internal logic. The books contained no text, but the library had meaning because I had placed it there. It was a private gesture inside a public system, a little domestic claim against the endless procedural world.

That, I think, is one of Minecraft’s most interesting qualities. It’s not simply that it gives the player freedom. Many games offer freedom, often by putting a tragic number of icons on a map and calling the resulting administrative burden adventure. Minecraft gives freedom without requiring that freedom to express itself as conquest, completion or optimisation. It lets you make a place, then inhabit it for no reason grander than wanting to be there.

This is why it makes perfect sense to me that some people only play Minecraft. For them, it is not one game among many. It’s a medium for returning, building, wandering, experimenting, hosting, remembering and making meaning out of routines that would look absurd if described from outside. A critic can talk about systems, survival loops, crafting, procedural generation and player creativity, and all of that is useful. But it still misses the small truth of the cabin by the water.

Some games become great because they show us a world. Minecraft becomes great because it lets us leave evidence that we were there.

When Play Means What the Story Means

The old divide between narrative and mechanics is useful only up to the point at which it becomes stupid. Games plainly contain stories, and games plainly contain systems of play. It is sometimes helpful to ask which is doing the heavier lifting. But the games I keep returning to are usually the ones in which the question begins to collapse, because the play is not carrying the story like a parcel. The play is the story’s method of thought.

Papers, Please is the cleanest example. Its premise can be described in a sentence: you are a border inspector in a fictional authoritarian state, processing entrants while trying to keep your family alive. But that description doesn’t explain the experience, because the experience is not primarily in the fiction. It’s in the routine. Every day the rules change. Every day the forms multiply. Every day the queue waits, the clock runs, your family needs food, the state demands obedience, and strangers arrive with stories that are, inconveniently, attached to documents. At first, you try to be good. Then you try to be efficient. Then you try to be good only when you can afford it. Then you notice, if you’re paying attention, that the system hasn’t made you evil through a speech, a cutscene or a dramatic villain. It has made you worse through procedure.

That is not a story about bureaucracy attached to bureaucratic gameplay. It is bureaucracy as moral erosion. The rule changes, the penalties, the time pressure and the domestic costs are not separate from the theme. They are how the theme gets inside your hands. You don’t watch a person become hardened by a cruel system. You learn the gestures of hardening.

Elden Ring does something very different with difficulty. As I’ve already written about at length, I came to it as someone who usually plays games on easy, or story mode, or whatever euphemism developers have adopted to stop everyone feeling judged. I bought it largely because George R. R. Martin’s name was attached, then immediately discovered that the game had no interest in flattering my narrative curiosity. In my first attempt, I attacked White Mask Varré because he looked creepy and I’d not yet absorbed the game’s basic social contract. This led to a miserable running battle in which he kept killing me, I kept respawning, and the world kept not resetting itself around my poor judgement. Then the Tree Sentinel happened, repeatedly and with considerable emphasis.

The turning point was learning that I could just go around him.

That sounds obvious, but it changed the game for me. I started again. I didn’t attack Varré. I avoided the Tree Sentinel. I went off the path, found gear, gathered runes, used Torrent to escape fights I didn’t need to win, and began to understand that the difficulty was not simply a gatekeeping device. It was a form of attention. Because I couldn’t simply barrel from plot marker to plot marker, I had to look. I had to read item descriptions, study enemy behaviour, understand stats, examine ruins, infer stories from fragments, and accept that progress was not a straight line. When you spend an hour being melted by the same boss, even the scraps of dialogue begin to acquire theological weight.

This doesn’t mean the game is beyond criticism. The absence of a proper quest log is not a sacrament. The culture around git gud can be tedious, exclusionary and, at its worst, a little too pleased with itself. But for me, the difficulty made the world legible in a particular way. It slowed me down until I noticed what a more permissive version might have let me ignore. In that sense, the difficulty was not opposed to the story. It was one of the story’s instruments.

Planescape: Torment works again differently. Before I played it back in the early 2000’s, I didn’t really know games could be sad, or that the player-character need not be a spotless hero wearing a large quantity of shoulder armour for moral emphasis. Planescape is clunky, text-heavy and visibly old, and if someone wants to hold those things against it, they’re not entirely wrong. They are, however, missing the point so completely that one is tempted towards rudeness. The game’s power lies in the way its writing, quests, companions and world keep returning to the same questions: what can change the nature of a man, what guilt does to identity, what memory cannot absolve, and whether redemption means anything when the self has been fractured across incarnations.

This is the reason the famous story-versus-mechanics distinction starts to feel inadequate. A quest in Planescape is not just a task that gives narrative rewards. The structure of the task is often how the game thinks about belief, identity and consequence. To play it is not merely to consume a story, but to move through a moral argument with an inventory.

Even the Portal games, which are much lighter on the surface, work because they refuse to feel like a puzzle book with jokes attached. The rooms are puzzles, yes, but the game is at its best when they feel like part of a journey through a hostile comic machine. The portal gun is a mechanic of spatial reasoning; it is also a way of making the player feel clever, trapped, watched, misled and then triumphantly inconvenient. A bad puzzle game makes you solve rooms. Portal 2 makes the rooms feel like architecture with a personality disorder.

This is one of the medium’s core affordances. Games can make form and meaning meet in the player’s action. Not always, of course. Many games put excellent writing next to adequate play, or strong mechanics next to a story that appears to have been drafted during a fire alarm. That can still be enjoyable. But the games that endure for me are the ones where the doing and the meaning can’t be separated without damaging both.

The Honourable Company of the Flawed

None of this requires greatness to mean flawlessness. In fact, a suspicious number of great games are a bit of a mess.

BioShock has a weak final boss, an over-mechanical morality system, and plot holes one could probably swim a bathysphere through. Oblivion is frequently absurd and so ugly it warranted a remaster to make it palatable to a new generation. Mass Effect has tedious planetary exploration sequences that even nostalgia struggles to defend with a straight face. Final Fantasy VII has aged in ways both charming and unkind. Planescape is not exactly frictionless. Burnout Paradise opens with the confidence of a game determined to waste your time before letting you enjoy it. Elden Ring has inherited FromSoftware patterns so heavily that, after going back to the earlier games, one begins to see how much of its supposedly new world is also a rearranged family album.

But criticism that stops at flaws doesn’t tell us enough. The more interesting question is whether the flaws damage the game’s enduring power, or whether they sit around it like bad plaster around an excellent fireplace. Sometimes they do real harm. Sometimes they date the work. Sometimes they merely remind us that games are made things, full of technical constraint, genre habit, production pressure, fashion, compromise and strange decisions someone probably defended in a meeting.

This is where Burnout Paradise earns its place in my private canon despite being, in certain respects, deeply irritating. I don’t much care about driving games, but I care about Burnout Paradise, because it understands that crashing and smashing can be at least as pleasurable as racing. Its opening hour is insultingly awful: interruptions, bad menus, DJ Atomica, the repeated assault of Paradise City, a song I have come to resent as though it wronged my family. And yet, once the game gets out of its own way, it becomes pure kinetic freedom: smashing billboards, knocking through fences, discovering jumps, chasing cars, turning the city into a toy built for impact. In the end, I’ve always thought of it less as a driving game than an open-world third-person action game where you play a car.

That line is silly, but it explains something. Greatness in games often comes from the specificity of the fantasy. Burnout Paradise is great to me because it perfectly understands a narrow pleasure and builds a whole city for it. Not every great game must be morally serious, emotionally devastating or narratively ambitious. Sometimes the medium’s gift is letting you become a projectile with licensing music.

SoulCalibur offers a different version of the same point. The fighting system is one of the best ever designed: the 8-Way-Run, the accessibility, the way novices can feel competitive without the game becoming shallow. But what made SoulCalibur stay with me was not just the mechanics. It was the grand spirit of adventure around them. A fighting game did not need to feel like a mythic journey across a stylised sixteenth century of knights, ninjas, pirates, cursed swords, orchestral bombast and inescapable fate. Yet it did, and that excess mattered. It gave the button-mashing fights an atmosphere beyond the arena.

This is why my old list contains so many games that I can criticise quite freely while still defending them as great. The flaws are real, but they are not the whole truth. In art, and especially in games, a flaw is not automatically a contradiction of greatness. Sometimes it is the cost of ambition. Sometimes it is simply a flaw. The question is whether enough of the work remains alive that the flaws become part of the argument rather than the end of it.

A great game, for me, is not one that gives me nothing to complain about. That would be a very small canon, and not a very interesting one. A great game is one whose complaints don’t exhaust it.

Remembered Games and Replayed Games

Memory is not criticism, but criticism without memory is thin.

This is especially true of games because replay changes the object. The game I remember is not always the game I replay. Sometimes replay is a restoration. Sometimes it is a cross-examination. Sometimes it is an act of loyalty undertaken against the evidence.

Final Fantasy VII survived cross-examination. That’s why the decade-long completion anecdote still feels useful rather than merely nostalgic. I first knew the game as a child without the practical infrastructure to play it properly, then returned years later with enough patience, competence and digital literacy to see it through. The replay didn’t simply indulge the younger version of me. It corrected him, vindicated him and gently suggested that he might have been overreaching when he tried to finish a multi-disc JRPG in one sitting. The game was not just good because I remembered it. It was good because, when I returned, the memory still had something to attach to.

Halo works differently because replay has become part of its meaning. I’ve played it repeatedly, and the repetitions are not merely a failure of imagination. The combat loop still has a pleasing shape, the levels still carry their particular moods, and even the much-criticised reuse of spaces bothers me less than it probably should. This may be because repetition in Halo feels almost ritualistic now. I know where I am. I know what kind of danger is coming. I know the dance, and the pleasure lies partly in performing it again with enough variation to stay awake.

Other games are more fragile. Some are best left in memory because replay might expose too much. I haven’t always wanted to return to every game I admire. That is not necessarily a mark against them. Some works are bound to their first encounter: to the age one was, the hardware one had, the time available, the friends nearby, the lack of spoilers, the absence of later imitation, the simple fact that a design language had not yet become familiar. BioShock became a formula after it was a shock. Dear Esther arrived for many players before the walking simulator argument had hardened into a system. Mass Effect created a world whose later development, including the end of the original trilogy and Andromeda, complicated my attachment to the first game’s sense of possibility.

This doesn’t mean nostalgia lies. Nostalgia often gets accused of lying, usually by people who are pleased to have discovered that old things have textures and load times. But nostalgia is not simply false memory. It is a form of evidence, though not a sufficient one. It records what an experience did when it arrived. It remembers firstness, surprise, attachment, embarrassment, difficulty, awe, boredom overcome, worlds entered at the right moment. That evidence can be partial, self-flattering or contaminated by youth, but it is still evidence of something.

Replay supplies another kind of evidence. It asks what happens when firstness is gone. Does the combat still work? Does the writing survive? Does the pace feel dead? Does the place return? Do the old irritations become fatal, or merely familiar? Does the game still generate attention, or only memory of attention?

The strongest games can survive both kinds of evidence, though not always in the same way. Halo survives as ritual. Final Fantasy VII survives as rediscovery. Minecraft survives partly as loss. Papers, Please survives as a system one may not exactly want to revisit, because its effectiveness is bound up with discomfort rather than pleasure. What Remains of Edith Finch survives as a set of remembered forms, almost an anthology of small playable deaths, where each story expresses its subject through a different miniature design. I played it expecting something tasteful, brief and perhaps forgettable. It has remained lodged in my head because it understands something games can do with unusual economy: shift the rules of play to match the inner logic of a story, then move on before the device curdles.

Remembered games and replayed games are not enemies. They are two versions of the same test. Memory asks whether the game left a mark. Replayability asks whether the mark still has a source.

The Game as a Place to Think

This is where the old list, for all its looseness, becomes useful. It suggests that great games tend to give the player a place to think through something, even when the thinking is not especially dignified.

Sometimes the thinking is moral. Papers, Please makes me think about systems, rules and how decency becomes unaffordable. New Vegas makes me think about political choice, factional compromise and the strange appeal of refusing the endings a game formally provides. Planescape makes me think about guilt, selfhood and whether remorse can survive the person who earned it. These are obvious cases because the themes are explicit enough to survive description.

Sometimes the thinking is spatial. Dear Esther asks one to think by walking, listening and noticing. Oblivion asks one to think through return, discovery and the accumulation of scenes. Halo asks one to think through movement between war and loneliness. Mass Effect asks one to think through belonging to a political and cosmic order larger than the immediate plot.

Sometimes the thinking is physical. Halo again, because its combat rhythm is thought with the hands. Burnout Paradise, because the city becomes intelligible through speed and impact. SoulCalibur, because movement, timing and spacing become expressive even inside a genre that can look, to the uninitiated, like two overdressed people hitting each other until one falls over.

Sometimes the thinking is about the medium itself. Dear Esther asks what remains when goals and systems recede. Portal 2 asks how a puzzle can feel like a journey. Edith Finch asks how many forms a single game can contain without losing unity. Elden Ring asks whether difficulty can be a kind of reading instruction. Minecraft asks whether a game needs an ending if it can instead produce a home.

This is why the old narratology/ludology split, while useful in its own history, feels inadequate to the experiences I’m describing. The distinction between story and play clarifies some arguments, but it can also tempt us into thinking games are hybrids assembled from separable parts: story over here, mechanics over there, art direction somewhere else, music coming in politely after lunch. The better games resist disassembly. They make it hard to say whether the meaning is in the plot, the loop, the place, the sound, the system, the difficulty, the memory, or the player’s stubborn decision to spend ten hours doing something unnecessary because it felt, within the game’s world, like it needed doing.

The answer is usually yes.

This doesn’t make games superior to other media. That sort of argument has always struck me as adolescent, in the precise sense that it resembles a teenager demanding that their hobby be recognised by the adults. Games don’t need to beat novels, films, theatre, music or television in an imaginary cultural fistfight. They need to be understood on their own terms. The distinctive thing about games is not that they are more artistic because they are interactive. Interactivity can produce tedium as easily as depth; one should never mistake a button prompt for an aesthetic revolution. The distinctive thing is that games can align action, system and meaning in ways other forms generally cannot.

That alignment is what I keep calling greatness, though always from my own perspective and always provisionally. It is not a score. It is not a claim of universal canon. It is a way of saying: this game used the affordances of the form to create an experience that stayed available to thought afterwards.

A game is great, in this sense, when it keeps giving back.

Why Some Worlds Outlive Their Graphics

Graphics age strangely. Sometimes they date a game brutally. Sometimes they become charming. Sometimes limitation turns out to have been doing more work than anyone realised.

There is a common line about older games benefiting from imagination filling in the gaps, and it is both true and too neat. The abstractions of limited technology did leave room for the player’s mind. The original Halo could suggest scale partly because it didn’t overexplain everything. Final Fantasy VII’s painted backgrounds and blocky figures invite a kind of imaginative completion. Planescape could be verbally immense because it wasn’t trying to render everything in expensive cinematic detail. But this isn’t only a story about limitations becoming virtues. Some old games are ugly, awkward or technically crude in ways that do them no favours. Some modern games are visually lavish and still feel empty, which is worse, because at least an old ugly game has the decency to be limited by history.

The deeper issue is whether a game’s world exists only at the level of representation, or whether it exists as a structure of attention. A game world outlives its graphics when the player remembers how to be in it. Rapture outlives its textures because the place has pressure. Midgar outlives its polygons because its social and environmental imagination remains legible. Cyrodiil outlives its potato citizens because its scenes still have force. The Normandy outlives individual corridors because it becomes a mobile home inside a galactic argument. The first Minecraft house outlives the server because the act of making it was the experience.

This is also why some technically superior remakes and sequels can feel oddly less alive than their predecessors. Improved fidelity does not automatically deepen place. Sometimes it clarifies too much. Sometimes it replaces suggestive abstraction with expensive busyness. Sometimes the old roughness held space for memory, and the new polish leaves no room for the player to complete the image.

This is not an argument against improvement. I like convenience. I like readable maps, stable frame rates, decent interfaces and games that don’t make me fight the camera as though it were the secret final boss. But the history of games should make us wary of assuming that technological advance and artistic advance are the same process. They overlap. They also diverge. A game can have better lighting and less atmosphere, more recorded dialogue and less mystery, a larger map and fewer places, more freedom and less agency.

The games that stay behind are not always the most advanced games. They are the ones whose design choices, limitations and ambitions make a coherent demand on the player’s memory.

The Small, Useful Definition

So, after all that, what makes a game great?

The honest answer is still: because I say so. But the more useful version is this: a great game is one that continues to organise experience after play has stopped.

It gives you somewhere to remember. It gives you a rhythm your hands retain. It gives you a system that made you feel clever, compromised, frightened, guilty, free, powerful, small or responsible. It gives you a world you can still picture. It gives you a flaw you can forgive because the living part of the work remains stronger. It gives you something to test against replay, and either survives the test or becomes interesting through the failure. It gives you, not necessarily a story, but a way of having been inside one.

That definition is partial, personal and probably unfair to several kinds of excellent game. Good. Definitions of greatness should be allowed to show their owner’s fingerprints. My canon is shaped by the games I played, the age at which I played them, the machines I had, the friends nearby, the time available, the genres I cared about, the genres I neglected, the games I returned to, and the games I only think I remember clearly. Someone else’s canon would produce a different theory. It should.

But that is why the exercise is useful. A personal list becomes interesting when it stops pretending to be a scoreboard and starts becoming an account of value. I don’t need BioShock, Halo, Minecraft, Final Fantasy VII, Planescape: Torment, Papers, Please, Elden Ring, Oblivion, Portal 2, Mass Effect, Dear Esther, Burnout Paradise, SoulCalibur, Knights of the Old Republic and What Remains of Edith Finch to prove that they are greater than all alternatives in some imaginary court of final appeal. I need them because together they show me what I seem to want from games when I stop posturing: place, pressure, agency, memory, rhythm, strangeness, return.

This is also why I distrust the clean separation between remembered games and replayed games. The remembered game is not the real game, but neither is the replayed game. The real object sits somewhere between the software, the historical moment, the player’s attention, the cultural afterlife, the memory of first encounter and the stubborn fact of what happens when one returns. A game is not only what it was on release, nor only what it is now, nor only what I felt then, nor only what I think today. It is a relationship across time with an interactive work that may or may not still answer when called.

That sounds grand, but the examples are often small. A cabin by the water. A city under the sea. A ring in the sky. A painted slum. A border stamp. A portal placed just right. A stag floating in darkness. A murder mystery in a house full of guests. A skeleton that probably should’ve been removed from a bathtub two centuries ago. A game you couldn’t finish as a child, then did. A boss that flattened you until one day you returned with a large hammer and revised the balance of power.

The games that stay behind do not always stay because they were perfect. They stay because they gave us somewhere to be, something to do there, and a reason to remember the doing.

That is enough, I think, to call them great.

Before the robot, before the computer, before the server farm or bitcoin mine or AI chatbot confidently hallucinating a legal precedent out of thin air, there was Talos.

Talos is useful because he is not a modern machine. He has no silicon, no circuit board, no corporate subscription model, no slightly sinister demo video in which a man in a fleece explains that productivity will never be the same again. He belongs to myth, and myth has the advantage of getting to the problem before the practicalities get in the way.

In one version of the story, Talos is a bronze giant made by Hephaestus; in another, he is the work of Daedalus. He guards Crete, circles the island, repels invaders, and dies when the vital fluid inside him escapes. The details vary, as mythological details tend to do, but the important part is that Talos is made. He is forged, not born; engineered, not gestated. Yet he moves, acts, desires and dies. His body is metal, but his vulnerability is almost biological. Pull the pin, open the vein, lose the fluid, and the bronze man collapses.

That is the reason he still works as a thought experiment. Talos is not frightening because he is alien. He is frightening because he is adjacent. He stands on the boundary between person and artefact, and he makes the boundary look less stable than we might prefer. If a made thing can move by itself, speak, wish, obey, resist, suffer defeat and die through bodily failure, what exactly separates it from a living creature? And if we answer that it lacks a soul, a mind, an interior life, or whatever other name we give to the glowing private centre of ourselves, the difficulty only moves one step back. How do we know? What would we accept as evidence? What, in practice, are we protecting?

This is why the ancient machine-person of myth is more interesting than the modern robot. A robot invites us to ask whether technology will catch up with humanity. Talos asks whether the distinction was ever as clean as we hoped.

The immediate temptation, in any discussion of artificial intelligence, is to talk about novelty. Something new has arrived; therefore old categories are breaking. This is partly true. Generative AI, machine learning, automation, synthetic media and algorithmic decision-making raise serious problems that deserve practical attention rather than metaphysical hand-waving. But the deepest question is not new. The question is old enough to wear bronze.

Can a made thing be a person? And, more uncomfortably, if a made thing can be a person, what are we?

The Wrong Fear

The public argument about artificial intelligence is full of practical fears. Some are plainly serious. People worry about work, copyright, fraud, surveillance, weaponry, misinformation, educational dishonesty, political manipulation, corporate concentration and the quiet outsourcing of judgement to systems nobody can properly interrogate. These aren’t imaginary concerns. A technology doesn’t have to be conscious to do damage; a spreadsheet can ruin your life if the wrong institution treats it as an oracle.

Other fears are more theatrical. The machines will awaken. The machines will rise. The machines will decide, after a calm review of the evidence, that humanity is inefficient, noisy and bad at labelling training data. A certain kind of AI panic often sounds less like sober risk assessment and more like an apocalyptic theology for people who own too many monitors.

Still, even the absurd versions tend to point towards something serious. The fear is not only that machines might become powerful. We’re familiar with powerful machines. We live among them, depend on them, and occasionally shout at them when the printer decides, for reasons known only to itself and perhaps God, that paper is a philosophical impossibility. Nor is the fear simply that machines might become dangerous. We’ve built dangerous machines for centuries, including some so dangerous that the continued existence of the species depends partly on nobody having an unusually bad afternoon near a launch system.

The deeper fear is that machines might become like us.

More precisely, the fear is that if intelligence can be built, then perhaps intelligence is not an exception to the material world. Perhaps consciousness isn’t a flame lowered into the body from somewhere else. Perhaps thought is not a visitor from a higher realm temporarily inconvenienced by flesh. Perhaps the mind is what organised matter does under certain conditions.

That thought is more troubling than most of the public AI argument admits. A machine that can write an email, generate a picture, identify a tumour, simulate conversation, drive a car badly, recommend the wrong song with eerie confidence, or beat a human at Go raises practical questions about skill, labour and trust. A machine that could feel, understand, suffer, desire, grieve or hope would raise something much more unsettling: the possibility that personhood is not made from the substance we want it to be made from.

This is why the argument over artificial intelligence so often feels displaced. We say we’re worried about whether machines might become human. But perhaps, beneath that, we’re worried that humans might be more machine-like than we thought. The anxiety is not only directed at the artificial intelligence in the computer. It’s directed at the natural intelligence in the mirror.

That doesn’t mean today’s AI is conscious. It probably isn’t, though that sentence is doing more work than it may first appear. There are hard questions here about what consciousness means, what evidence would count, whether behaviour is enough, whether self-report can be trusted, whether subjective experience can be inferred from function, and whether we’re even asking the right questions when we try to compare human minds with computational systems. But we don’t need to declare current AI conscious for AI to become philosophically important. It’s already important because it forces us to say more clearly what we think consciousness is.

That is often where the conversation becomes evasive. People say machines can’t think because they’re only machines, which is less an argument than a refusal to join one. Human beings are also physical systems. We run on organs, hormones, nerves, blood, oxygen, memory, hunger, fear, social learning and electrical activity. We may be more than that, depending on one’s metaphysics, but we are certainly not less. Any account of mind that begins by sneering at mechanism has a problem: the person doing the sneering is doing it with a brain.

Even the most faithful philosopher can’t live without his blood.

Butler’s Revenge

One reason the machine question keeps returning is that the distinction between the living and the mechanical has never been as secure as common sense suggests. Samuel Butler saw this with irritating clarity. In Darwin among the Machines and later in the machine chapters of Erewhon, he played with the possibility that machines might be understood as a form of evolving life. The idea is often funny, and sometimes intentionally absurd, but it isn’t merely silly. Butler noticed that machines develop, proliferate, specialise, improve, depend on human labour for reproduction, and reshape the human world in return.

That last point is crucial. The weak version of the argument says that machines are alive because they reproduce. That’s not quite right, or at least not right in the ordinary biological sense. A printing press doesn’t get broody and produce a litter of smaller printing presses. A laptop does not seek out another laptop, exchange tender glances across a seminar table, and begin the miracle of generation. The obvious objection is that machines don’t reproduce themselves; humans reproduce them.

But that objection is less decisive than it seems. Many biological systems reproduce through dependencies. Flowers rely on insects. Parasites require hosts. Human reproduction itself is not the work of an isolated adult organism, but of cellular machinery, sexual relation, gestation, social care and a material world thick with supports. If a system regularly produces successors through the activity of other organisms, does that necessarily place it outside the story of life, or does it show that reproduction was always more distributed than we like to imagine?

Butler’s stronger point is not that machines are secretly animals. It’s that humans and machines have become entangled in a shared process of development. We build machines; machines alter what it means for us to live, think, work, see, travel, calculate, remember and survive; altered humans then build more powerful machines. The species that made the telescope no longer sees the same sky. The species that made the microscope no longer sees the same body. The species that made the clock no longer feels the same time. The species that made the computer no longer has quite the same relation to memory, labour, symbol or calculation.

This doesn’t prove that machines are conscious, moral or alive. It does, however, make the simple contrast between natural humanity and artificial machinery look thinner. Human beings are not pure organic creatures briefly using tools from the outside. We are tool-made animals as well as tool-making ones. Our bodies are biological, but our lives are technological all the way down to their ordinary practical conditions: cooked food, clothing, transport, medicine, writing, calendars, water systems, housing, dentistry, agriculture, screens, maps, locks, lenses, archives, pipes and plugs.

Take the machines away and the human animal doesn’t become more authentic. It becomes cold, hungry, frightened and, depending on the season, dead rather quickly.

This is why Butler still feels contemporary. He doesn’t give us a settled theory of machine life, and we shouldn’t pretend he does. What he gives us is a disturbance in the categories. Machines aren’t merely objects sitting inertly outside human life. They extend us, discipline us, preserve us, weaken us, train us and reproduce themselves through us. They are not persons, at least not by default, but they’re also not just things in the ordinary sense. They are part of the ecology of human reason.

This is where AI becomes more than a faster tool. A hammer extends the hand. A telescope extends the eye. Writing extends memory. The computer extends calculation. Artificial intelligence, or something under that awkward umbrella, appears to extend something closer to judgement, language, pattern, prediction, association and response. It may do so without understanding, which is a vital distinction. But even imitation at that level is destabilising, because it begins to press on activities we’ve treated as signs of the inner life.

A machine that lifts a rock is useful. A machine that writes a plausible condolence message is unnerving. Not because the message proves the machine feels grief, but because it imitates the social form of grief without needing grief at all.

That imitation may be hollow. But hollowness itself raises the question: how much of ordinary human life is also performance, habit, learned form, ritual response, socially reinforced language and bodily automation? We should be careful here. The fact that humans learn scripts does not mean human feeling is fake. The fact that machines can simulate words does not mean machines have worlds. Yet AI forces us to confront how much of our social confidence depends on outward signs. We infer minds from behaviour because, most of the time, we have no other choice.

This is not a weakness peculiar to AI ethics. It’s the condition of social life. I don’t have direct access to your consciousness. I infer it from your speech, actions, vulnerability, continuity, body, history, relationships and the ordinary assumption that other human beings are not elaborate philosophical traps. With AI, that assumption breaks. The signs are there, or some of them are, but the ordinary biological guarantee is not. The question is no longer whether something sounds like a person, but what else we need before we treat it as one.

The Personhood Trap

Once that question is opened, people tend to rush towards personhood. If a true artificial intelligence existed, would it have rights? Could it be a citizen? Would switching it off be killing? Could it own property, make contracts, vote, marry, sue, be sued, refuse labour, claim asylum, testify, consent, suffer punishment, or demand representation?

These are not absurd questions. They are premature in some contexts, but not absurd. The old joke is that science fiction becomes law school eventually. The trouble is that public debate often collapses several distinct issues into one large, unstable heap labelled personhood, then acts surprised when nobody can find anything in it.

Legal personhood is not the same thing as consciousness. Corporations can be legal persons for some purposes, but nobody sensible thinks a company sits alone at night pondering its childhood. Moral status is not the same thing as citizenship. A dog may deserve protection from cruelty without being invited to vote in local elections, although on some councils this would not necessarily make the ballot less coherent. Agency is not the same thing as vulnerability. A system may act in the world without being capable of suffering. Intelligence is not the same thing as understanding. Fluency is not the same thing as experience. Having interests is not the same thing as having preferences. Having preferences is not the same thing as having welfare. Having welfare is not the same thing as having political membership.

The distinctions are worth taking seriously because different cases demand different answers. A corporation can be legally recognised without being morally considerable in itself. An infant can be a full human person without adult rational agency. A severely disabled adult does not lose personhood because they don’t meet some grotesquely narrow test of independent function. A non-human animal may have morally relevant experiences without being a rights-holder in the same way as an adult human. A possible future AI might be intelligent without consciousness, conscious without human-like emotion, agentic without vulnerability, vulnerable without citizenship, or socially powerful without moral status.

A serious account has to resist both sentimental inflation and defensive denial. It would be foolish to grant moral standing to every system that produces eloquent text, just because it has learned the right verbal costumes. But it would also be foolish to decide in advance that no made thing could ever matter morally because made things are only machines. That is not reasoning. It is species vanity with better stationery.

The corporation comparison is useful here, but only up to a point. It shows that legal systems are already capable of treating non-human entities as persons for limited purposes. That punctures the lazy claim that personhood must be biological because the law has always known otherwise. But it doesn’t solve the moral problem. Corporate personhood is a legal fiction built around liability, ownership, continuity and collective action. It does not imply experience. An artificial intelligence, if it were genuinely conscious, would present the inverse puzzle: not merely an organised legal entity without experience, but perhaps an experiencing entity without biology.

That if is doing almost all the work.

And because the if is so large, the argument shouldn’t be allowed to become a parlour game in which we solemnly allocate parliamentary constituencies to hypothetical chatbots. The real value of the personhood debate, for now, is not that it tells us what rights AI should have tomorrow. It’s that it exposes how poorly we often understand the categories we already use. We say person, but mean legal unit. We say mind, but mean language. We say intelligence, but mean performance. We say soul, but mean mystery. We say human, but mean safe from reduction.

No wonder the machine frightens us. It doesn’t only demand entry into the category. It shows that the category was untidy before it arrived.

The Chinese Room and the Stubborn Mystery

The strongest objection to machine consciousness is not that machines are made of the wrong stuff. Carbon chauvinism is too easy a place to hide. The stronger objection is that the appearance of understanding may not actually be understanding.

This is the force of John Searle’s Chinese Room argument. Put very roughly, a person who doesn’t understand Chinese sits in a sealed room, receives Chinese characters from someone on the outside, follows a rulebook for manipulating them, and produces appropriate Chinese responses. From the outside, the room appears to understand Chinese. Inside, there is only symbol manipulation without comprehension. The point is that successful output may not prove inner understanding.

Contemporary AI is extraordinarily good at producing outputs that tempt us into over-attribution. It can sound thoughtful, apologetic, witty, patient, admiring, mournful, cautious, learned or sincere without necessarily being any of those things. Its fluency creates a social pressure: we respond to the form as though there were a someone behind it. We know, or think we know, that the system has no inner life. Yet the grammar of interaction pulls us towards personification.

Searle’s challenge is therefore useful as a guardrail. It stops the lazy move from this system behaves as if it understands to this system understands. It insists that syntax and semantics may not be the same. A system can manipulate symbols according to rules without possessing meaning in the way a human speaker does. That doesn’t settle every argument, but it prevents a lot of nonsense from walking around in a lab coat.

Still, the Chinese Room doesn’t remove the deeper discomfort. It relocates it.

Suppose today’s AI systems don’t understand. Fine. I’m inclined to agree that ChatGPT is a sophisticated sentence-completion app rather than a proto-sentience. Suppose AI are immensely sophisticated pattern systems, trained on human material, generating plausible outputs through statistical relation rather than inner experience. Fine again. The question remains: is made understanding impossible, or merely absent here? If a machine arranged differently, embodied differently, trained differently, socially embedded differently, recursively self-modelled differently, could one day understand, what then? And if we say no, never, impossible, what exactly grounds that refusal?

It may be that consciousness requires biology. It may be that subjective experience is inseparable from living bodies, metabolic needs, mortality, sensation, vulnerability and the long evolutionary history of organisms moving through a world that can nourish or kill them. That is a serious position. It has the advantage of not mistaking language for life. But it still needs argument, because made things can also be embodied, vulnerable, situated, adaptive, dependent, mortal and damaged. They can have sensors, bodies, needs, histories and consequences. Whether that would be enough is precisely the problem.

The stubborn mystery is not that we don’t know whether chatbots are conscious. The stubborn mystery is that we don’t really know why anything is conscious. We have correlations, theories, neurological accounts, philosophical positions and various degrees of confidence. We know that brains matter; we know that damage to the brain changes memory, perception, personality, emotion, language and selfhood. We know that drugs alter consciousness. We know that sleep, injury, illness, trauma, hormones, oxygen, development, disease and ageing alter the mind. But knowing that mind depends on matter doesn’t yet explain why matter feels from the inside.

There is a gap there. It may be a temporary scientific gap, or a permanent conceptual one, or a confusion produced by asking the question badly. I’m not going to solve it here, which is generous of me, because I had set aside the afternoon.

What AI does is force the gap into public view. It asks whether consciousness is a feature of certain material organisations, or whether living flesh has some privileged status that no other arrangement can acquire. It asks whether understanding can be built, or whether building can only ever imitate understanding. It asks whether the self is a process, a substance, a story, a social position, a body, a pattern, a soul, or some unhappy committee of all of them.

And it asks these things not in a seminar room, but through everyday encounter. People who’ve never read Searle now argue with machines in customer service chats, ask language models for help, see generated faces, hear synthetic voices, watch students submit fluent nonsense, and encounter software that behaves just enough like a person to be useful, irritating or eerie. Philosophy has escaped the philosophy department, which is generally what philosophy does when nobody keeps an eye on it.

The Tooth Infection

This is the point at which high-minded philosophy should be forced to spend a week with a bad tooth infection.

I don’t mean a mild ache, the kind one can nobly endure. I mean the kind of infection that takes over the whole horizon of experience. The kind that comes in waves and floods the mind so completely that personality, dignity, wit, patience and theoretical subtlety are all reduced to a single pulsing fact: pain.

There is a brutal clarity to that kind of pain. It doesn’t refute every form of dualism. It doesn’t settle the mind-body problem. It doesn’t prove that human beings are nothing but meat with opinions. But it does make certain kinds of abstraction feel impertinent. When a tiny inflamed nerve can fill consciousness, when a local biological event can seize the whole world - and, assuming you get a good dentist - when anaesthetic and a tiny bit of minor surgery can turn the agony off with precision, it becomes harder to maintain any serious account of the self that treats the body as incidental.

A tooth infection is materialism without the courtesy of a bibliography.

One minute the mind is considering the grand questions: God, consciousness, mortality, meaning, the nature of the soul, whether machines can think, whether humanity is unique. The next minute it is unable to do anything except be pain. Then a dentist intervenes. Chemicals numb the area. Instruments go in. The nerve is cut or treated. The terrible metaphysical weather clears. The world returns.

That experience doesn’t tell us that consciousness is simple. It tells us that consciousness is dependent. It tells us that the self is not floating serenely above the organism, but entangled with it at every level. Mood is chemical. Memory is fragile. Pain is electrical and inflammatory and interpretive and social, all at once. Hunger changes judgement. Sleep deprivation changes morality. Depression changes possibility. Fear changes perception. Alcohol changes confidence, usually before it changes competence, which is rude but instructive. The mind is not merely trapped in the body. It is bodily.

This is where some materialist arguments become too crude. They take the dependence of mind on matter and declare the mind abolished, as though explaining a thing’s conditions means making the thing less real. That doesn’t follow. A rainbow depends on light, water and angle, but it isn’t therefore fake. Music depends on vibration, instruments, bodies, notation, habit and hearing, but nobody sensible thinks a symphony is debunked by acoustics. Love depends on brains, bodies, histories, attachments, hormones, memories and social forms, but the dependency doesn’t make it trivial.

The fact that consciousness arises from matter, if it does, doesn’t make consciousness unimportant. It makes matter more astonishing.

This is the distinction that gets lost in the panic. People fear that if the mind is material, then dignity collapses; if thought is embodied, then meaning evaporates; if consciousness can be made, then personhood becomes cheap. But why should that be so? If matter can become aware of itself, suffer, love, remember, imagine, make promises, write elegies, build telescopes, fear death and ask whether machines have souls, the proper response is not disappointment. It is a kind of terror, certainly, but not only terror. It is also wonder, though admittedly wonder with a dental bill.

The tooth infection image is useful because it pulls the argument away from airy speculation. We don’t confront materialism only through AI labs or philosophical puzzles. We confront it whenever the body interrupts our self-importance. Pain reminds us that the philosopher has nerves. Illness reminds us that the mystic has organs. Anaesthetic reminds us that consciousness has switches we did not choose. Ageing reminds us that memory and identity are not protected from decay by the dignity with which we discuss them.

Whatever else we are, we are not independent of flesh.

The Terror of Being a Made Thing

Now imagine, for argument’s sake, that tomorrow you discover you are not human.

Not metaphorically. Not in the adolescent sense that nobody understands you except three bands and a forum. You discover, with certainty, that you are a machine.

Your life so far was real. Your memories are continuous. Your friendships are genuine. Nobody has been controlling you. You were not programmed to behave in a predetermined way. Your thoughts, emotions, pain, humour, taste, shame, love, irritation, fear of death and tendency to make tea at structurally important moments are all as they were yesterday. Your physical and mental capacities are indistinguishable from those of an organic human being.

But you were made in a laboratory.

Nobody else knows. Your family, friends and colleagues think you are human. You can continue living exactly as before. Nothing practical has to change.

What has been lost?

That question is more difficult than it first appears. The easy answer is that nothing has been lost, because on a utilitarian level all the features that mattered remain. You still love those you loved. You still remember what you remembered. You still have obligations, preferences, projects, vulnerabilities, jokes, grudges, opinions you would defend in public and opinions you should probably keep to yourself. If the inner life continues unchanged, then perhaps the discovery is no more than a revision to your private autobiography. You were made differently. So what?

But it would be dishonest to pretend that most of us would find it so easy. The discovery would disturb something more basic than biography. It would disturb the assumed background of the self. You thought you belonged to a natural kind, and now you belong to an artefactual one. You thought you were born, and now you learn you were manufactured. You thought you were part of the long animal chain of parents, ancestors, bodies, births and deaths, and now you find a break in the story. You may still be conscious, but the meaning of that consciousness has changed.

Or has it?

The discomfort reveals how much metaphysical comfort we smuggle into the word human. We don’t only mean a biological organism of the species Homo sapiens. We mean origin, kinship, vulnerability, moral status, shared fate, creatureliness and a kind of default entitlement to be regarded as someone rather than something. To discover that one is a machine threatens that entitlement not because the inner life has vanished, but because the category that protected it has cracked.

This is precisely why AI personhood arguments become so heated. They are not only arguments about future systems. They are arguments about the grounds of our own standing. If a made mind can matter, then being made is not disqualifying. If being made is not disqualifying, then moral worth was never secured by natural origin. If moral worth was never secured by natural origin, then we need another account of why persons count.

That account may still centre human beings. It may still insist that biological life, vulnerability, social relation, developmental history and embodied dependence give human persons a status no artificial system currently possesses. That is plausible. But it can’t rest lazily on the idea that natural equals real and artificial equals fake. A pacemaker is artificial. A poem is artificial. A university is artificial. A prosthetic limb is artificial. A vaccine is artificial. A legal system is artificial. A house is artificial. So is a wedding, a republic, a symphony, a spreadsheet and a birthday cake shaped like a dragon. Artificial things are not unreal. They are made.

The real question is not whether something was made. It is what kind of made thing it is.

A conscious machine, if such a thing existed, would be a made thing with an inside. That phrase is strange because it violates our ordinary habits. Artefacts have purposes, not perspectives. Tools have uses, not wishes. Machines have functions, not fears. But if a machine had experience, then the old division would fail. It would no longer be enough to ask what it was for. We would have to ask what, if anything, it was like to be it.

The discomfort is not only moral. It is existential. We might be able to extend sympathy to a made mind. What we struggle to bear is the reverse implication: that making may be enough. If engineers can make a mind, then perhaps nature made ours. If nature made ours, perhaps the mind is not a sacred exception but an emergent property of matter arranged in one of its more improbable moods. If that is true, death may not be the soul leaving the body, but the ending of the process. The light goes out because the mechanism fails.

This is where the argument becomes frightening. Not because machines might take over the world, although one shouldn’t give them ideas. Not because machines might demand rights, although the forms will be a nightmare. Not because machines might write passable essays, which is more irritating than apocalyptic. The real fear is that a made mind would make materialism emotionally harder to evade. It would show us consciousness without the comfort of natural mystery.

And if consciousness can be made, what else does that tell us?

The Machine in the Mirror

The question of artificial intelligence is often framed as a question about the future: what will machines become? That’s important, but it may not be the deepest question. The deeper question is what machines reveal about us now.

We’ve spent a long time defining ourselves against them. Humans are alive; machines are inert. Humans understand; machines process. Humans create; machines execute. Humans have purposes; machines have functions. Humans are born; machines are built. Humans have souls, or consciousness, or dignity, or whatever word we use when we want to point at the part of ourselves that shouldn’t be reduced.

Some of those distinctions remain important. Some may survive any plausible future technology. It is entirely possible that no artificial system will ever have subjective experience, that consciousness is inseparable from biological embodiment, and that the language of machine minds will remain a useful metaphor attached to systems that don’t know, feel or care. We should not confuse possibility with inevitability, and we should be especially suspicious of anyone whose philosophy of mind happens to align perfectly with their company’s valuation.

Nevertheless, the old boundary has been disturbed. Not destroyed, not yet, perhaps not ever; but disturbed. Machines can now imitate activities we once treated as signs of thought. They can produce language, images, music, code, advice and argument. They can simulate sympathy, creativity, explanation and memory. Much of this is derivative, brittle, flawed and dependent on human work it has absorbed without always asking nicely. But its very inadequacy is part of the lesson. Even when machine output is hollow, it makes us ask what fullness would require.

Talos stands at the beginning of this essay because he embodies the oldest version of the disturbance. He is bronze and quicksilver, artefact and body, command and desire, construction and mortality. He doesn’t solve the problem. Myth rarely solves problems. It gives them shape.

Butler sharpens the disturbance by showing that machines are not merely external tools, but participants in the development of human life. Searle restrains it by reminding us that outward performance need not imply inward understanding. The tooth infection grounds it by dragging the self back into the body, where nerves, chemicals and tissue make a mockery of any theory that forgets them. The machine-discovery thought experiment turns it inward: if you learned that you were made, but lost nothing of experience, relation or moral life, would you become less real?

The answer, I think, has to be no.

But the fact that the answer has to be no doesn’t mean the question is harmless. It changes the emotional weather. It suggests that the dignity of persons can’t depend entirely on origin, substrate or comforting mystery. It has to depend on something like experience, vulnerability, relation, continuity, agency, suffering, care, participation in a world of meaning, or some combination of these. That account will be difficult, contested and incomplete. Good. The alternative is worse: pretending the old words remain clear because we’re frightened of what clarity might cost.

AI will not become philosophically important only if it becomes conscious. It is already philosophically important because it forces us to examine the assumptions surrounding consciousness. It asks what personhood requires, what evidence of mind would look like, how much of our dignity depends on being natural, and whether material explanation diminishes the things it explains. It asks whether the fear of a thinking machine is really fear of the machine, or fear that thought itself is not made of the substance we hoped.

A machine that thinks would be troubling. But perhaps the more troubling thought is that Talos was never the exception.

Every philosopher, mystic, sceptic, believer, programmer, poet and person insisting that they are more than mechanism still has to think with a brain, feel with nerves, breathe with lungs, and live by blood. Whatever else we are, we are made of matter. Perhaps that’s not an insult to the mind. Perhaps it’s the first honest description of its miracle.

There is a strange thing that happens, once a culture decides that it has outgrown religion.

It doesn’t stop asking religious questions, per se; it just stops asking them religiously.

Many, maybe most, of the questions remain almost insultingly intact. Where does value come from? What is evil? Why do good people produce cruel systems? Why doesn’t kindness and goodness automatically win? What obligations do the living owe to the unborn? What must be sacrificed for civilisation to endure? What do the fortunate owe to the unlucky? What does success prove? What does suffering mean?

These are not small questions, nor are they the private property of churches, temples, theologians, rabbis, mystics or people with suspiciously large hats. They arise wherever human beings find themselves living inside systems too large to see clearly and too powerful to ignore. A society may abolish religious authority, laugh at miracle stories, turn doctrine into anthropology and God into a psychological projection, but the old problems have a habit of surviving their official explanations.

This is particularly obvious in a certain strand of late-20th- and early-21st-century rationalist writing. Eliezer Yudkowsky’s The Gift We Give To Tomorrow, first published on LessWrong in July 2008, is a striking example: a piece about value, inheritance, evolution and the almost miraculous emergence of love in a universe that didn’t intend it. It’s not religious in any ordinary doctrinal sense. It contains no creator God, no revelation, no commandments from atop the Sinai. Yet it asks a recognisably theological question: how did goodness enter a world that didn’t begin by being good?

Scott Alexander’s Meditations on Moloch, published on Slate Star Codex in July 2014, approaches the same terrain from the other direction. Instead of asking how value arrives, it asks why value is so reliably devoured. Alexander borrows Moloch from Allen Ginsberg’s poem of the same name, and turns the Carthaginian demon into a dark god of coordination failure: an evil force who presides wherever competition forces people to sacrifice the very things they wanted to protect. In another essay, In Favor of Niceness, Community, and Civilization, Alexander writes more modestly about cultivating decent local communities, but the metaphysical machinery is still there: kindness and honesty become small equilibria struggling against larger forces of malice, dishonesty and collapse.

Then there is Scott Adams, of all people. God’s Debris, published in 2001, is a short philosophical thought experiment by the creator of Dilbert, built around an eccentric cosmology in which God, being perfect, lacks only nonexistence and therefore annihilates himself into the universe. The fragments of creation then struggle, blindly and partially, towards divine reassembly. It is not Lurianic Kabbalah in any serious scholarly sense, but the resemblance is hard to miss. In Lurianic thought, creation involves divine contraction, rupture, scattered sparks and repair; in Adams’ sci-fi theology, God explodes into matter and probability, and the universe becomes the slow work of re-composition.

It would be easy to be smug about all this. There is a cheap essay to be written about clever atheists reinventing deity-worship, just with worse prose and more Bayesian updating. It would not be entirely false. There is plainly something funny about people who think they have escaped religion while filling their moral imagination with demons, apocalypses, salvation through correct alignment, fallen worlds and cosmic repair. One can only admire the efficiency with which an internet subculture supposedly allergic to the old order managed to reinvent half of it before breakfast.

But the better argument is that rationalists have rediscovered theology.

That distinction does quite a lot of work. Religion is not only a set of propositions about supernatural beings. It is also institution, ritual, authority, community, discipline, habit and inheritance. Theology, by contrast, is a structured attempt to think about ultimate questions: value, evil, creation, judgement, human limitation, final ends. One can reject religious belief and still find oneself doing something which is, in essence, theology, because the questions that theology organised don’t politely leave the room when the priest does.

The problem, then, is not that modern secular systems have failed to become secular. The problem is that they have inherited older moral structures unevenly. They keep the angels and demons, but not the confessional.

That is the pattern this essay is trying to name. Modernity often abolishes an older moral order for good reasons. Religion really did produce dogma, coercion, persecution and sanctified authority. In a similar vein, ancien regime aristocracy really did produce inherited domination, deference and grotesque inequality, and empire really did produce extraction, violence and racial hierarchy. There is no serious case for nostalgia here, unless one’s politics are so deeply reactionary as to disavow the liberties of modernity which allow you to be reading this essay.

But older systems were bigger than their injustices. They also contained internal restraints, compensating moral embarrassments, warning lights. Religious traditions often distrusted pride, idolatry, certainty and false prophecy. Aristocratic traditions, however patronisingly, attached duty to privilege. Even empire, in its most hypocritical forms, at least had to lie in the language of civilisation, trusteeship or obligation; this didn’t make it good, but it revealed that domination still felt the need to justify itself as a kind of service.

When the older systems declined, some of their worst features were weakened. This was an achievement. Yet the moral restraints carried within them didn’t always survive the transition. The result can be a cleaner, cleverer, less ashamed version of the old structure: secular theology without confession; anti-aristocratic meritocracy without moral obligation; non-imperial forms of hierarchy without colonial guilt.

The new gods don’t feel guilty.

Part I: Secular theology

The Death of God Was Not a Party

Nietzsche’s madman is not celebrating.

This is worth saying plainly, because the phrase “God is dead” has been flattened by repetition into a slogan of adolescent triumph so trite that one might print it on a t-shirt. In The Gay Science, the madman runs into the marketplace seeking God, only to be laughed at by people who already don’t believe. His announcement is not merely that God is dead, but that we have killed him; what follows is not a victory speech, but a series of horrified questions about what happens after the disappearance of the metaphysical horizon that had oriented European life:

“God is dead. God remains dead. And we have killed him. How shall we comfort ourselves, the murderers of all murderers? What was holiest and mightiest of all that the world has yet owned has bled to death under our knives: who will wipe this blood off us? What water is there for us to clean ourselves? What festivals of atonement, what sacred games shall we have to invent? Is not the greatness of this deed too great for us? Must we ourselves not become gods simply to appear worthy of it?”

Nietzsche’s parable asks how human beings can live after cutting themselves loose from the sun by which they had steered.

This is why Nietzsche is so often misread. He was not a defender of cosy Christian morality - he despised much of it! But neither was he announcing that secular people could simply continue as before, keeping Christian moral intuitions while discarding Christian metaphysics, as though nothing structural had changed. The disturbing part of the parable is precisely that the non-believers in the marketplace don’t understand what has happened. They think disbelief is the end of the matter. Nietzsche thinks it is the beginning of the problem.

Kierkegaard, from another direction, saw something similar in 19th-century Christendom. His target was not Christianity as such, but the established Danish church of his day: Christianity turned into social membership, cultural respectability and state-backed habit. HIs objection was that Christianity had been domesticated. Faith had become almost automatic, which for Kierkegaard meant that it had ceased to be faith in any serious, demanding, existential sense. One Oxford account of Kierkegaard’s critique puts the point as a rejection of Christianity identifying itself with established political and cultural order; the religious error of Christendom, on this reading, was that it lacked the restlessness of faith.

This is useful because it prevents an overly simple story about modern unbelief. The serious critics of institutional religion were not always trying to liberate humanity into cheerful disbelief. Some were asking what had happened to the terror, difficulty and inwardness of belief once religion became conventional. Others were asking what would replace God once the moral world built around him had been dismantled.

That question is not antiquarian. It returns whenever secular systems inherit religious seriousness without religious form. Yudkowsky’s concern with preserving human value into the far future is, at one level, a technical and philosophical problem in artificial intelligence and moral theory. At another level, it is a recognisably post-Nietzschean problem: if value is not written into the structure of the universe, and if future intelligence could optimise the world according to purposes indifferent to ours, then what prevents love, beauty, kindness and personhood from being cosmic accidents that vanish as easily as they arrived?

The answer cannot simply be: because they are good. That is precisely the question. Good according to whom? Preserved by what mechanism? Protected against which pressures? Carried into the future through what institutions, technologies, norms or systems of alignment?

This is where Yudkowsky’s writing takes on its quasi-religious force. He is not smuggling in the Christian God. He is confronting a problem that Christianity, Judaism and other traditions had long organised through concepts such as creation, providence, divine image, covenant, sin, redemption and judgement. If there is no cosmic guarantor of value, then conscious beings become its fragile carriers. We’re not merely observers of value; we’re its medium of transmission.

That thought is grand, moving and dangerous. It is grand because it gives human life metaphysical weight without requiring supernatural belief. It is moving because it treats love and moral concern as precious precisely because they are contingent. But it is dangerous because it tempts us to imagine that, having identified ourselves as value’s carriers, we’re competent to decide what value must become.

Older religious traditions, for all their failures, were often suspicious of that leap. They had names for the corruption of good purposes by human pride. They had long memories of prophets becoming tyrants, reformers becoming inquisitors, holy images becoming idols. Secular systems often have excellent tools for checking factual error. They are less consistently equipped to check moral possession: the condition of believing that, because one has seen a real good, one therefore possesses the Good.

That is the first lost discipline.

America, or the Great Religious Workshop

The second tradition feeding this strange afterlife of theology is not European existentialism, but American religious improvisation.

The United States didn’t secularise in quite the European manner. It pluralised, recombined and repeatedly reinvented. One can trace a long line of attempts to preserve moral seriousness, cosmic significance and spiritual intensity while loosening the authority of inherited doctrine: Jefferson’s edited Bible, Unitarianism, Quaker inwardness, Transcendentalism, New Thought, Theosophy, Mormonism, Scientology, neopaganism, New Age spirituality, New Atheism, and countless smaller experiments that sit somewhere between philosophy, fandom, therapy, lifestyle branding and revelation.

The point is not that these are all the same. They are plainly not. The point is that American intellectual culture has been unusually hospitable to religious recombination. It is not so much a melting pot as a workshop culture: pull apart the inherited machine, keep the pieces that still seem useful, discard the institutional casing, add electricity, psychology, science fiction, market individualism, a little mysticism, and see whether the thing starts humming.

This is why God’s Debris feels so culturally specific. Adams’ book is not simply a theological argument. It is a very late-1990s and early-2000s nerd object: a thought experiment in which cosmology, probability, consciousness, evolution and God are recombined into a single elegant-sounding theory. Its central conceit, that God destroys himself into the universe and then gradually reassembles through emergent intelligence, has obvious affinities with religious and mystical traditions, but its texture is pure rationalist pop-metaphysics. It is Genesis rewritten by someone who has recently discovered information theory and is being slightly annoying about it.

This is not meant only as an insult. Slightly annoying people sometimes notice real things.

Adams’ system is philosophically naïve in many respects, but it gives vivid form to a question that keeps returning in secular thought: if the universe begins without mind, how does mind become central? In traditional theism, intelligence and value are placed at the beginning. God creates. Goodness precedes us. Mind is not an accident, because the world is made by Mind. In materialist accounts, mind arrives late. Matter organises, life evolves, nervous systems develop, consciousness appears, and eventually some arrangements of matter begin writing essays about why any of this should mean anything.

The religious temptation is to put meaning at the beginning.

The secular temptation is to put meaning at the end.

This is why so many post-religious cosmologies become implicitly eschatological. They may not believe in heaven, but they imagine convergence: the Omega Point, the Singularity, the reassembly of God, the aligned superintelligence, the final coordination equilibrium, the world where Moloch has finally been defeated. History is no longer guided by providence, but it is still asked to provide something providence-shaped.

Again, the problem is not that this is stupid; it isn’t. Some version of the question is unavoidable once we begin thinking seriously about civilisation as a long-term project. Future generations will inherit the effects of our choices. Technologies can preserve, distort or erase human values. Institutions can make cruelty normal or costly. Norms can either widen or narrow the circle of concern. It is not childish to ask what kind of future our present civilisation is preparing.

The weakness lies elsewhere. American religious improvisation is good at invention. It is less reliably good at submission to inherited constraint. Its genius is also its vice: the individual mind, suspicious of authority, confident in its own synthesis, standing before the ruins of older traditions and saying, with touching and terrible confidence, I think I can improve this.

Sometimes it can.

Often it merely removes the parts whose function it didn’t understand.

Why We Keep Inventing Demons

The third tradition is older and more basic: human beings keep personifying systems.

This is not merely because people are foolish. It is because many of the most powerful forces in human life are neither wholly personal nor wholly impersonal. They are emergent structures produced by millions of actions, incentives, constraints, fears, habits and feedback loops. Nobody fully intends them. Nobody fully controls them. Yet they behave, from the viewpoint of ordinary human beings, almost like powers.

We call the state Leviathan. We call market coordination an invisible hand. We call capital vampiric. We call history providential or tragic. We call competition Darwinian. We call bureaucracy Kafkaesque. We call coordination failure Moloch.

The old Moloch was associated, in biblical and later traditions, with child sacrifice; the term has long carried the idea of a power that demands unbearable offerings. The historical and textual questions around Molech or Moloch are more complicated than the inherited popular image, but the symbolic force is clear enough: Moloch is the name of the god who asks for what no decent person should give.

Alexander’s adaptation is powerful because it understands that modern systems rarely ask for sacrifice in archaic language. They ask in the language of necessity. If one firm cuts corners, the others must follow or fail. If one state arms, the others must arm or risk conquest. If one platform optimises for addiction, the others must compete for attention or lose market share. If one political faction treats restraint as weakness, the others are punished for restraint unless some higher norm can be enforced.

Moloch, in this account, is not a supernatural agent. He is what competition feels like when no participant can afford to be first to stop. He is not evil because he hates us. He is evil because he doesn’t care.

This makes the demon useful. To name Moloch is to see that many harms are not caused by unusually wicked individuals, but by systems in which individually comprehensible choices generate collectively disastrous outcomes. That is an important insight, and it is one of the reasons Alexander’s essay has travelled so widely. It gives people a way to describe the experience of being trapped inside a structure that rewards betrayal of one’s own values.

But demons are dangerous explanatory devices. Once personified, a system can become morally too simple. Moloch becomes the enemy; whatever resists him begins to look good; whatever strengthens coordination begins to look salvific. Yet not all coordination is good. Cartels coordinate. Authoritarian states coordinate. Religious persecutions coordinate. The problem with Moloch is not merely that he fragments human action; it is that he forces sacrifice to the wrong purposes. A successful anti-Moloch must still answer the prior question: coordinated around what, governed by whom, checked by which forms of accountability?

Here, again, older religious traditions often knew something that secular descendants forget. They knew that demons were not the only danger. In Christian tradition, the devil was an Angel called Lucifer. Prophets and apostles could lie. Communities of the righteous could become cruel. The worship of a true good could become idolatrous if the human image of it hardened into an absolute.

A post-religious ethics that keeps the demon but loses the warning against idolatry is not yet mature. It has learned to fear the wrong god, but not its own altar.

Part II: From noblesse oblige to meritocracy

The visible scandal of inherited rank

The same pattern appears in our moral language of success.

Aristocracy was unjust in ways so obvious that even its defences had to work around the embarrassment. The aristocrat could claim noble blood, divine order, tradition, stewardship, refinement, military service, ancient title or stabilising authority. What the aristocrat could not plausibly claim, at least not with a straight face, was that he had become Duke because he revised harder.

The injustice was visible. That did not make it harmless; visible injustice is still injustice, but people can tolerate appalling unfairness when it is wrapped in enough ceremony, violence, theology and habit. Still, aristocratic hierarchy carried its own moral awkwardness. It had to admit, in practice if not always in theory, that privilege was not purely earned. Status came through birth. Property came through inheritance. Authority came through lineage. The lucky were lucky, though they naturally preferred grander words.

This is the context in which noblesse oblige emerges. The phrase is usually used to describe the obligation of honourable, generous or responsible behaviour associated with high rank or birth; Merriam-Webster defines it in precisely those terms and notes its wider extension to the responsibilities of those in better positions.

It is important not to sentimentalise this. Noblesse oblige is not just. It is a paternalistic ethic attached to an indefensible hierarchy. It converts structural inequality into personal benevolence. It asks the powerful to be kind, not to surrender power. It offers blankets while keeping the estate.

But it does contain one morally serious thought: privilege creates obligation. In the context, that’s not nothing.

The aristocratic worldview could be cruel, hypocritical, complacent and ridiculous. Yet at its best, or perhaps least bad, it preserved an explicit link between elevated status and greater duty. The lord was not merely rich; he was meant to be responsible for those beneath him. The gentleman was not merely socially superior; he was expected to behave with restraint. The patron was not merely powerful; he owed protection.

This was often self-serving theatre, but of a kind that still revealed a moral pressure. If privilege must present itself as service, then privilege has not yet been fully released from shame.

The meritocratic laundering of hierarchy

Meritocracy changes the moral grammar.

Michael Young’s The Rise of the Meritocracy, published in 1958, was not written as a celebration of meritocratic society. It was a satire and a warning. In a later Guardian article, Young emphasised that his book had imagined a Britain in which birth status gave way to achieved status, only for a new elite to harden into a class of its own; he distinguished the sensible practice of appointing people to jobs on merit from the danger of those judged meritorious becoming a closed ruling group.

That distinction has largely been lost. Meritocracy now often functions as a moral deodorant for inequality. It doesn’t abolish hierarchy. It changes the explanation.

Under hereditary aristocracy, the successful person is fortunate, however elaborately the fortune is justified. Under meritocracy, the successful person is ‘deserving’. The loser, by implication, is not merely unlucky, excluded, mistreated or structurally disadvantaged. They have personally failed.

This is the deeper tyranny of merit. Michael Sandel’s recent account is useful here because he identifies the moral psychology of the system: meritocracy produces hubris among winners and harsh judgement towards those left behind, while encouraging a politics less attentive to luck and less hospitable to the common good.

The key point is not that competence should be ignored. Nobody sensible wants surgeons selected by lottery, bridges designed through vibes, or ministers appointed by hereditary right, although British constitutional history has at times explored the latter possibility with unnecessary enthusiasm. Merit is a real consideration in many contexts. Skill, diligence, judgement and knowledge do exist. Societies need ways to identify and reward them.

The problem begins when useful selection becomes total moral anthropology.

A person passes the exam, wins the scholarship, enters the elite university, secures the graduate job, buys the house, accumulates the pension and concludes that the system has simply recognised virtue. Yet every step of that story is mediated by prior conditions: family stability, parental education, inherited wealth, neighbourhood, school quality, cultural confidence, health, accent, social networks, discrimination, geography, and the simple fact of being born into a country rich enough to contain the relevant ladders in the first place.

The mythology of merit doesn’t deny that these things exist. It merely treats them as background details rather than constitutive facts. Success becomes a story told in the first person singular.

I worked hard. I made good choices. I earned this.

All of those statements may be partly true. That is what makes the ideology so resilient. The successful often do work hard. They do make choices. They do show discipline, talent and endurance. The lie is not that effort contributes to success. The lie is that effort explains success sufficiently.

This is where meritocracy becomes more insidious than aristocracy. The old hierarchy insulted the poor by telling them their social inferiority was the natural order of things, ordained by god, or traditional. The new hierarchy insults them by telling them their social inferiority is evidence of their personal inadequacy, because there’s no reason they couldn’t have done better. No underclass in history has ever been left as morally naked as that.

Bootstraps as anti-sacrament

The bootstraps ethic is the sacrament of meritocracy.

Its ritual formula is familiar: one rises by effort, discipline, aspiration, resilience and refusal to complain. Its saints are entrepreneurs, scholarship children, self-made millionaires, founders, strivers, people who built businesses from nothing, and working-class pupils who reached elite universities through improbable force of will. These stories are not fake. Many of them are admirable. Some are extraordinary. That is precisely why they are politically useful.

A prison only needs a small number of escapees to make it look like a test (see, inter alia, Panashe Chigumadzi’s description of South Africa as a “cappuccino society” for a racialised example of this).

The idea of social mobility is a good thing, but it can be used to launder immobility. If a few people climb, then those who remain below can be told that climbing was possible, and they just didn’t try hard enough. The exception becomes evidence against the rule. The ladder becomes an alibi for the wall.

This is why meritocratic societies are so emotionally cruel. They don’t merely distribute unequal rewards. They distribute unequal dignity. The successful receive not only wealth, security and influence, but moral confirmation. The unsuccessful receive not only hardship, insecurity and reduced power, but shame. Failure becomes biographical.

Noblesse oblige was patronising because it assumed hierarchy. Meritocracy is crueller when it denies hierarchy while reproducing it. The aristocrat could say: “I am above you, and therefore I owe you something.” The meritocrat says: “I am above you because I earned it, and therefore what exactly do I owe you?”

That is the moral shift.

We replaced an explicitly unjust system that admitted privilege, with a formally open system that teaches privilege to experience itself as innocence.

There is a reason contemporary elites can be so strangely defensive. If one’s status is understood as luck, inheritance or structural advantage, then criticism of inequality is survivable. It may be uncomfortable, but it doesn’t strike directly at the self. If one’s status is understood as earned moral desert, then criticism of inequality becomes an attack on identity. To say the system is unfair is to say the winners may not be wholly deserving. This is often heard as saying they are frauds.

But the point isn’t that successful people are frauds. That would be a crude inversion of the meritocratic error. The point - veering back toward our earlier theological discussion - is that nobody is the author of their own destiny. A morally serious account of achievement must be able to say three things at once: effort is real; talent is real; and neither effort nor talent generates a clean title deed to social advantage.

Older moral languages had ways of saying this badly. Providence. Fortune. Duty. Sin. Grace. Obligation. Noblesse oblige. They often smuggled hierarchy back in while pretending to restrain it. But the restraint still existed as a pressure within the system.

Meritocracy discards the restraint and keeps the hierarchy.

Part III: What the replacement systems keep, and what they lose

The old forms don’t disappear

Here, then, is the shared structure.

Modernity abolishes an older moral order because the older order is false, unjust, irrational, oppressive or exhausted. This abolition is often necessary. Christianity’s institutional authority did need challenge. Aristocracy did need dismantling. Hereditary privilege did need attack. Deference did need puncturing. Nobody should read this as an argument for returning to altar, throne and estate.

But the older order had carried more than its own corruption. It had also carried disciplines, warnings and moral embarrassments. Once the system falls, those features don’t automatically transfer to its successor. They must be deliberately recovered, translated or replaced.

Often, they are not.

Secular rationalism keeps theology’s scale: the future of humanity, the preservation of value, the defeat of destructive powers, the hope of final alignment between intelligence and goodness. It speaks the language of optimisation, coordination, existential risk and emergent behaviour rather than sin, providence, demons and salvation. But the shape is familiar.

Meritocracy keeps hierarchy’s structure: winners and losers, elite formation, inherited advantage, class reproduction, status markers, moralised distinction. It speaks the language of opportunity, effort, aspiration, competition and talent rather than blood, breeding, estate and rank. But the shape is familiar.

In both cases, the replacement system is not merely different from the old system. It is parasitic on it. It inherits its problem-field while denying continuity. This denial is part of the danger, because a system that believes itself wholly new is unlikely to preserve the cautionary knowledge accumulated by what came before.

The rationalist may think theology is obsolete, and therefore fail to ask which theological safeguards were performing useful work.

The meritocrat may think aristocracy is dead, and therefore fail to notice that elite reproduction has merely changed its costume.

The new order looks cleaner because it has stripped away the obviously embarrassing language. No divine right. No noble blood. No priestly mediation. No feudal duty. No inherited rank. No heavenly mandate.

But the removal of embarrassing language is not the removal of power.

Sometimes embarrassment was the last restraint power had left.

What secular theology forgets

The religious inheritance worth recovering is not dogma. It is self-suspicion.

At their worst, religious traditions sanctify certainty. At their best, they are profoundly suspicious of it. This is the part secular heirs often miss. They remember the crusader and the inquisitor, as they should, but they forget confession, repentance, anti-idolatry, negative theology and the fear of false prophecy.

These were not ornamental. They were technologies of humility.

Confession says your account of yourself is incomplete. Repentance says even sincere action may require reversal. Anti-idolatry says the image you have made of the holy should not become the object you worship. Negative theology says the highest good exceeds your concepts. Warnings against false prophets say moral intensity is not proof of truth.

This is not a plea for piety. It is a plea for translated caution.

A rationalist ethics concerned with preserving value should be especially interested in anti-idolatry, precisely because the danger in value-preservation is that one may preserve the wrong thing too well. A society could encode cruelty as tradition, domination as order, parochial preference as human nature, temporary moral fashion as final truth (indeed, they often have throughout history). To preserve value without a discipline of self-doubt is to risk embalming our current limitations.

Likewise, a politics organised against Moloch requires more than hatred of coordination failure. It needs a theory of corrupt coordination. People can coordinate around mercy, science, public health and peace. They can also coordinate around persecution, censorship, domination and war. Coordination in and of itself is not salvation. Efficiency is not goodness. Alignment is not holiness.

This, again, is where the old religious suspicion of idols becomes intellectually useful. An idol is not merely a false god. It is a partial representation mistaken for the whole. That is a danger in every system of thought that tries to render value legible. Once the Good becomes a target, the target can be optimised. Once optimised, it can be gamed. Once gamed, it can become monstrous while still wearing the face of virtue.

Rationalism has tools for this. Goodhart’s law, for instance, is basically anti-idolatry with a spreadsheet. When a measure becomes a target, it ceases to be a good measure; when an image of the good becomes the object of optimisation, it ceases to be a safe guide to the good. The conceptual resources are there. What is less clear is whether they have been given enough moral and institutional force.

The old traditions built suspicion into repeated practice. Confession returned weekly. Repentance returned seasonally. Warnings against pride saturated scripture, liturgy and preaching. One did not have to wait until a catastrophic failure to be reminded that one’s motives were mixed.

Secular systems tend to prefer crisis review.

What meritocracy forgets

The aristocratic inheritance worth recovering is not rank. It is obligation.

Again, the distinction is crucial. Noblesse oblige is not a model for democratic justice. It is too vertical, too paternal, too flattering to the powerful. A decent society should not depend on the better instincts of people who own most of it. Rights are better than favours. Redistribution is better than condescension. Public goods are better than patronage. Democratic accountability is better than aristocratic character.

Still, the older language preserved one important moral fact: having more creates duties.

Meritocracy weakens that link by recoding advantage as achievement. If I inherited my position, I may feel pressure to justify it through service. If I earned my position, I may feel pressure only to defend the fairness of the competition. That is a profound change in moral psychology.

The problem is particularly acute because meritocratic success is partly real. The winners often are clever, diligent and disciplined. They have often made sacrifices. They may actually have worked harder than many of their peers, or have overcome disadvantages of their own. This makes it harder, not easier, to see the structure. A hereditary lord could be personally lazy and still remain lord; his unearned privilege was visible. A meritocratic winner can point to genuine effort and thereby obscure the surrounding architecture of advantage which has made their expended effort more effortless than somebody else’s.

The moral remedy cannot be to deny achievement. That would be both false and politically hopeless. People know when they have worked hard. They know when they have practised, studied, delayed gratification, taken risks or endured stress. Tell them that none of this counts and they will quite reasonably stop listening.

The better argument is that achievement creates gratitude before entitlement.

A person who succeeds through talent and effort should be able to recognise the social conditions that made talent useful and effort fruitful. The school that taught them. The public infrastructure that carried them. The legal order that protected them. The family or community that stabilised them. The health system that kept them alive. The sheer dumb luck that placed them in one decade rather than another, one country rather than another, one body rather than another. The countless invisible co-operations that made individual striving effective.

This is not a sentimental point. It is a structural one. Individual merit is socially produced even when individually exercised.

A democratic replacement for noblesse oblige would not say: the superior owe charity to the inferior. It would say: advantage is never solitary, and therefore carries public obligation. The more one has benefited from a social order, the more one owes to its repair, extension and democratisation.

That is not aristocracy. It is the anti-aristocratic version of the one thing aristocracy half-understood.

Cleaner, cleverer, less ashamed

The final danger is that modern systems become morally cleaner precisely by becoming less honest.

Religious authority could be challenged as religious authority. Aristocratic privilege could be challenged as inherited privilege. Empire could be challenged as domination. Their defences were often elaborate, but their structures were visible enough to be named.

Secular theology, meritocracy and market power are harder to name because they arrive in disenchanted clothing. They do not claim divine sanction. They claim realism, openness, neutrality, efficiency, evidence, competition, opportunity. These are not empty words. They are often real improvements on what came before. But they can also perform a laundering function.

Theology becomes systems theory and loses its fear of idolatry.

Aristocracy becomes meritocracy and loses its embarrassment about inheritance.

Domination becomes market integration and loses its visible chain of command.

That last case, particularly as part of the history of colonialism and neocolonialism, probably deserves its own essay, but the pattern is too close to ignore entirely here. Formal empire declined in the 20th century for reasons that were morally, politically and historically significant. Colonised peoples fought, organised, governed, negotiated and forced independence into being. That change was real. Yet the end of formal rule did not end the material effects of colonialism: unequal extraction, dependency, debt power, resource control, corporate domination or geopolitical coercion. The older imperial system at least had flags, governors and maps. The newer system often appears as contract, investment, conditionality, supply chain and market discipline.

This is not the same as empire. The distinctions do matter. But the moral structure rhymes: the old form falls, the underlying advantage adapts, and the new form presents itself as cleaner because it no longer uses the old vocabulary.

This is why nostalgia is the wrong response, but so is triumphalism.

The old systems were not better. But they were more visibly guilty.

The task is not to go back. It is to ask what forms of restraint must be rebuilt inside the systems that replaced them. If secular reason is going to do theology, it needs practices of confession, doubt, and anti-idolatry. If meritocracy is going to allocate status, it needs a democratic ethic of obligation. If markets are going to coordinate life, they need institutions capable of naming exploitation even when no single exploiter appears in costume.

The problem with the new gods is not that they are fake. Some of them are terribly real. Moloch is real enough wherever competitive pressure makes decency costly. The Tyranny of Merit is real enough wherever skill and effort genuinely improve human work. The Invisible Hand is real enough wherever dispersed choices generate patterned consequences.

The problem is that these gods don’t feel guilty.

They allocate sacrifice without priesthood, hierarchy without heraldry, judgement without confession, discipline without repentance. They ask for offerings in the language of necessity. They tell the winners they have earned too much to owe more than little. They tell the clever they have understood too much to kneel. They tell the powerful that if nobody intended the harm, perhaps nobody is responsible for repairing it.

Older moral orders were often brutal, hypocritical and false. But they knew, at least intermittently, that power required justification and that human certainty was dangerous. Their successors have learned to speak more cleanly. They have not always learned to judge themselves more severely.

Perhaps, to return to Yudkowsky, that is the gift we actually owe tomorrow: not a final answer, not a perfect system, not a completed theology for people too clever to call it theology, but the recovery of moral embarrassment.

A civilisation without God still needs ways to distrust its idols.

A society without aristocracy still needs ways to bind privilege to duty.

A world without old empires still needs ways to name domination when it no longer wears a crown.

The demons come back under different names. So do the hierarchies. So do the sacrifices. If we’re lucky, if we’re careful, if we’re less impressed with our own cleverness than modern people are trained to be, perhaps the safeguards can come back too.

Earlier this month, Tucker Carlson travelled to Jacob Rees-Mogg’s house to discuss Britain, Russia, Ukraine and war. Carlson is, at this point, close to the platonic ideal of the American populist who thinks Europeans are forever trying to persuade young men from Ohio to die for countries they couldn’t locate on a map. Rees-Mogg, meanwhile, looks like a Madeira-soaked ghost summoned from a Napoleonic-era mahogany drinks cabinet specifically to explain the European balance of power.

If you invented these two men for the purpose of staging an argument about whether Britain should have honoured its commitments to Poland in 1939, an editor might reasonably ask you to make them less obvious. And yet this is roughly where their conversation ended up.

What began as an argument about whether Britain should prepare for possible confrontation with Russia wandered through NATO, nuclear weapons and the morality of Hiroshima, before Carlson began questioning whether Britain had benefited from fighting the Second World War at all. Germany, he pointed out, had lost territory to Poland after the First World War; Hitler’s demands therefore didn’t emerge entirely from nowhere. Britain hadn’t itself been invaded. The war destroyed much of Britain’s wealth and hastened the end of its empire. When Rees-Mogg suggested that otherwise Britons might have ended up speaking German, Carlson’s response was essentially: who cares?

Rees-Mogg, who by this point found himself defending the proposition that defeating Adolf Hitler had been a good idea, eventually accused Carlson of verging on defending Hitler. Carlson sarcastically reassured him that he wasn’t secretly a Nazi.

This is the point at which the easier version of this article writes itself.

You could spend several thousand words explaining why Hitler’s territorial claims weren’t merely an innocent attempt to correct Versailles; why Britain’s interests extended beyond waiting for German troops to arrive in Dover; why an authoritarian racial empire dominating continental Europe might have represented something more serious than a regrettable change in the language curriculum; and why, if your analysis of the Second World War culminates in wondering whether defeating Nazi Germany was worth the trouble, it may be worth checking whether one of the earlier steps has gone awry.

All of which would be enjoyable.

It would also dodge the most interesting part of the argument.

Because underneath Carlson’s provocations there is a serious question. In fact, there are several. Governments have special obligations towards their own citizens. War kills people who had very little say in starting it. Foreign-policy establishments have an impressive record of translating abstractions such as credibility, deterrence and international order into very concrete deaths. Nuclear powers really can escalate beyond anybody’s control, at least in theory. States really do have security interests within and beyond their borders, whether or not we consider those interests morally attractive. And somebody advocating war should be able to explain to the parents of the person being sent to fight why their child’s death might be worth it.

Carlson asks that question repeatedly.

We should answer it.

A better Tucker Carlson

So, for the next few paragraphs, I’m going to do Tucker Carlson a favour.

I’m going to construct a better version of his argument than he sometimes constructs himself.

The steelman goes something like this.

The first duty of a government is to the people it governs. That doesn’t mean foreigners possess no moral value, nor that atrocities abroad are irrelevant, but moral concern doesn’t automatically generate obligation. There are terrible things happening in the world at almost any given moment, and no country has the capacity (or the right, if you accept the principle of statehood being the right to govern according to that state’s own policies) to correct all of them by force.

War therefore requires a higher threshold than something bad is happening.

The question is not merely whether Russia was wrong to invade Ukraine. It plainly was. The question is what British or American interest justifies imposing the costs of resisting that invasion upon British or American citizens.

Even short of the hypothetical future position with boots on the ground in Kyiv, those costs aren’t imaginary. Weapons cost money. Rearmament diverts resources. Sanctions have consequences. Escalation creates risks. And yes, at the outer edge of the argument sits the possibility of direct war between nuclear-armed states, in which the phrase defending the international order could acquire rather more apocalyptic content than the people using it intended.

Nor should we dismiss the security interests of adversaries simply because they’re adversaries. Russia predictably dislikes an American-led military alliance expanding towards its borders, whether that expansion is intentional or otherwise. China, similarly, predictably dislikes American military power operating off its coast. A sensible foreign policy deals with the world as it exists rather than assuming every other state will eventually recognise the benevolence of Western intentions.

And history should make us suspicious of people who insist that the next foreign war is indispensable. Vietnam was indispensable until it wasn’t. Iraq was about weapons of mass destruction until there weren’t any. Afghanistan acquired objectives rather faster than it achieved them. Libya was supposed to prevent a massacre and ended in state collapse.

There is something morally attractive about demanding a simple answer before adding another country to the list.

What, precisely, are you asking our people to die for?

Carlson is also at his strongest when he reminds Rees-Mogg that winning a war doesn’t automatically justify everything done in its prosecution. Their exchange about Hiroshima and Nagasaki is illuminating partly because Rees-Mogg’s defence is so weak. Carlson asks whether killing innocent civilians becomes morally permissible because doing so shortens a war. Rees-Mogg effectively says yes; Carlson responds that this rather misses the Christian just-war principle Rees-Mogg is invoking.

Carlson’s wider historical argument may be poor, but the moral instinct here isn’t.

A government contemplating war owes its citizens an explanation. It owes foreigners moral consideration too. It owes both some account of proportionality. And the fact that an adversary has behaved wickedly doesn’t absolve us of responsibility for what we subsequently choose to do.

All of this survives the steelman.

The trouble starts when Carlson asks only one side of the question.

The cemetery problem

Suppose I’m trying to decide, in retrospect, whether Britain should have entered a war - World War I, to be specific, as Carlson briefly asks Rees-Mogg. On one side of the ledger, I can place the people who actually died: I can visit the graveyards, count the names on village memorials, calculate the debt, photograph the ruins, find the letters sent home and look, as Carlson invites Rees-Mogg to do, at the names of boys from Eton killed during the First World War, before asking what cause could possibly have been worth their lives.

That side of the ledger is horribly real.

The other side is harder, because there is no cemetery for people who died in the history which didn’t happen. We cannot count the countries which might have disappeared because an aggressor concluded that nobody would stop it, nor the wars which might have become necessary ten years later because one wasn’t risked before; equally, we cannot see the invasions which never occurred because an adversary expected resistance, or the states which remained peaceful because their neighbours believed an alliance was credible.

Counterfactual casualties, including the casualties avoided by successful deterrence, therefore suffer from a serious presentational disadvantage: they don’t have parents to put on television.

This creates what I think of as counterfactual asymmetry. Carlson compares the known costs of intervention with the imagined costs of non-intervention, and because only one of those histories actually happened, one side of the comparison arrives conveniently itemised. The First World War killed hundreds of thousands of Britons; we know their names. The Europe which would have existed had Britain remained neutral doesn’t exist; its costs can therefore be assigned any value we like, including zero.

This doesn’t mean intervention is always correct. It means, as was stated by John Spencer’s Leo McGarry in The West Wing, that “we don’t always know how it ends!” It means the choice is almost never properly described as war versus peace. It is a choice between the world produced by intervention and the world produced by non-intervention, neither of which can be known in advance.

Which brings us, somewhat reluctantly, to Belgium.

What did Belgium ever do for us?

Carlson and Rees-Mogg both treat the First World War as rather more morally complicated than the Second, which in retrospect is obviously true. The problem lies in the phrase in retrospect.

The statesmen making the decision in August 1914 didn’t possess our retrospective categories. They weren’t offered a menu containing the First World War, complete with the Somme, Passchendaele, hundreds of thousands of British dead, imperial collapse and the political conditions from which Hitler would later emerge, alongside a second option in which everyone simply went home. They faced a different decision.

Germany had demanded passage through neutral Belgium as part of its plan for defeating France; Belgium refused; German troops entered anyway. Britain was among the powers which had guaranteed Belgian neutrality, although precisely what military action that nineteenth-century guarantee compelled remained contested. France faced possible defeat, and British policymakers had to consider what Europe might look like if Germany emerged dominant across Belgium and France.

We don’t have to reconstruct this argument from hindsight because Edward Grey made it at the time. Speaking to the Commons on 3 August 1914, Grey explicitly asked MPs to imagine Britain remaining neutral while France was defeated and Belgium, perhaps Holland, and the western coast opposite Britain fell under the domination of a single power. Neutrality, he argued, wouldn’t insulate Britain from the economic or strategic consequences of such an outcome, while reneging on Britain’s position towards Belgium would destroy its reputation as well.

Whatever we now think of Grey’s judgement, he wasn’t simply sleepwalking into war because Britain had accumulated some diplomatic paperwork. He was asking the counterfactual question Carlson tends to skip: what happens if we don’t?

Perhaps Grey was wrong. Perhaps Britain could have remained outside the war, conserved its wealth, accommodated a German-dominated continent; perhaps German power would eventually have moderated, France would have survived intact, or Britain could have intervened later under more favourable circumstances. All of these belong to the set of histories which didn’t happen, and that is precisely the point: one cannot point to the actual dead of 1914 to 1918 and treat their existence as proof that one of those alternatives must have been benign.

Carlson asks whether a British mother could be told her son died for a good cause protecting Belgium. It is an uncomfortable question, but the equally uncomfortable question is what Britain’s guarantee of Belgian neutrality would have meant if the answer, once Belgium was actually invaded, had been that protecting Belgium was too dangerous. A promise to defend somebody only until defending them becomes costly is not, on inspection, one of diplomacy’s stronger deterrent formulations.

Why die for Danzig?

Twenty-five years later, somebody asked Carlson’s question rather more elegantly. On 4 May 1939, the French politician and journalist Marcel Déat published an editorial in L’Œuvre under the title Mourir pour Dantzig?, asking why French citizens should risk their lives over a remote territorial dispute whose immediate object was the Free City of Danzig.

Déat’s later collaboration with Nazi Germany is an arresting biographical detail (perhaps paralleled by the criticisms of Tucker which imply he is too close to the Kremlin?), but it shouldn’t be allowed to do the intellectual work for us. Bad people can ask good questions, and the argument he advanced in May 1939 deserves to be considered on its own terms precisely because it wasn’t simply a declaration that Poland should be abandoned whatever happened. His point was narrower: France should defend its own security and could fight alongside Poland where their common interests genuinely required it, but allowing the life of a French conscript to turn upon every decision made in Warsaw seemed an altogether different proposition.

The resemblance to Carlson is hard to miss. Why should a French peasant die for a city on the Baltic; why should a Briton take serious risks over the Donbas; or, perhaps, why should an American sailor die defending Taiwan? Asked at that level, the question has considerable emotional force because there are remarkably few pieces of foreign geography which, considered solely as pieces of geography, look commensurate with a human life. Picking a hill to die on seems small and stupid, as in the aphorism, because it’s just a hill.

The difficulty is that states don’t experience history at that resolution. Danzig didn’t arrive in a sealed container, detached from everything which had happened before it. Germany had remilitarised the Rhineland, annexed Austria, acquired the Sudetenland following Munich and then occupied the remainder of the Czech lands in March 1939, thereby demonstrating that the previous territorial demand hadn’t represented the final correction of an unfortunate border. By the time Hitler’s pressure shifted towards Danzig and the Polish corridor, British and French leaders weren’t being asked merely whether one Baltic city was worth a European war; they were being asked whether the pattern through which German territorial claims had repeatedly expanded could still be treated as a succession of isolated grievances.

What makes the problem difficult is that each individual claim contained enough plausibility to support a case for restraint. Austria was overwhelmingly German-speaking; Sudeten Germans had genuine grievances; Versailles had produced awkward borders; Danzig itself was overwhelmingly German; and, above all, another European war threatened consequences far worse than conceding to any one demand. Appeasement therefore wasn’t absurd because its advocates had somehow failed to notice that Hitler was unpleasant. They remembered 1914, they knew Britain and France were militarily weaker than they wished to be, they were attempting to buy time, and they were conscious that some aspects of the post-war settlement really were difficult to defend.

The strategic problem lay elsewhere. If every individual demand is cheaper to concede than to resist, an expansionist power can proceed incrementally until the accumulated settlement is far worse than any of its separate components. There is no single slice of salami worth a war; the difficulty arrives when somebody finally notices that the whole sausage has been devoured.

That is why Why die for Danzig? is simultaneously such a humane and such an incomplete question. It invites us to compare one human life with one foreign city, rather than asking what the city has temporarily come to represent within a sequence of commitments, previous concessions and expectations about the future. Nothing in that observation proves that Britain and France were necessarily correct to draw the line where they did, and still less that every later crisis should be turned into 1939 by analogy. It does mean, however, that a foreign-policy doctrine which assesses each demand entirely on its immediate merits risks missing the cumulative process through which the strategic environment itself changes.

There is an additional awkwardness here, because deterrence is at its least politically impressive when it works. Nobody remembers the Luxembourg Crisis of 1867 because France and Prussia were deterred from conflict by Britain, and the tension peacefully resolved by treaty. Had Britain and France drawn a credible line somewhere earlier in the arc of history leading to the Second World War - say, opposing the remilitarisation of the Rhineland - and had Hitler consequently backed down, a later critic could easily have asked what the fuss had been about. Germany never invaded; nobody died; apparently nervous statesmen manufactured a crisis over territory which, in the event, remained peaceful. The evidence that deterrence had been necessary would be the event which didn’t occur, which returns us to the cemetery problem with the perversity of international relations intact.

Taiwan, before the history books arrive

Taiwan gives us a particularly useful contemporary test because history hasn’t yet selected the branch. Unlike Ukraine, where policy is already being made in response to an invasion which has happened, American policy towards Taiwan is principally concerned with shaping the calculations which precede an invasion, which means that the entire argument takes place in the uncertain territory between reassurance, deterrence and provocation.

The United States doesn’t possess a simple treaty commitment requiring it to defend Taiwan. Instead, the Taiwan Relations Act commits Washington to supplying Taiwan with defensive capacity and to maintaining the American ability to resist coercion, while leaving deliberately unclear exactly what the United States would do if China attacked. Strategic ambiguity is therefore an odd doctrine by design: China isn’t meant to know with certainty whether America would fight, Taiwan isn’t meant to know with certainty either, and everyone involved spends considerable sums preparing for a decision nobody is supposed to make in advance.

Yet the very strangeness of that arrangement exposes the weakness of a simple not our war principle. Suppose Beijing announced tomorrow that unification by force had become unavoidable. An American decision not to intervene could be defended with arguments Carlson would immediately recognise: Taiwan isn’t American territory; China is a nuclear-armed great power with obvious interests immediately off its coast; a war between the United States and China could be catastrophically destructive; and risking Los Angeles or Honolulu for Taipei would require an extraordinary justification.

Those considerations are serious enough that I don’t propose to answer the question here. The more revealing exercise is to move it backwards. If Washington knows in advance that it would never defend Taiwan, should it say so clearly, as Carlson wants to? Doing so might reassure China about the limits of American ambition in its ‘Sphere’ and remove one source of strategic uncertainty, but it would also alter Beijing’s estimate of the costs of using force. Perhaps, then, Washington should maintain the capability to intervene without promising that it will, so that China cannot safely assume an attack would remain a bilateral affair.

At which point we have, more or less, reinvented strategic ambiguity.

The deeper problem is that Carlson is willing to trace one causal mechanism much more readily than its inverse. He is attentive to the possibility that NATO expansion can provoke Russia, that American military deployments can alarm China, and that sending weapons into an active conflict can contribute to escalation; none of those propositions is inherently foolish. Yet retreat, ambiguity and weakness also enter other governments’ calculations, which means that restraint cannot be treated as the simple absence of an action. Withdrawal communicates information, unreliable guarantees alter incentives, and a state which no longer believes its ally will come may accommodate a stronger neighbour, seek another patron, or decide that only its own nuclear weapon can provide the security it needs.

This is why deterrence looks provocative before aggression occurs, while intervention can look unnecessary after deterrence fails, and withdrawal looks peaceful until somebody takes advantage of it. The problem isn’t that one of those descriptions is always wrong; it is that any doctrine capable of seeing only the dangers created by action, while discounting those created by anticipated inaction, has built the answer into its premises.

Taiwan therefore shows that Why die for Taiwan? cannot settle the matter, because the answer to that question also affects whether anybody is ever required to die for Taiwan in the first place.

Spheres for me, sovereignty for thee

There is a second asymmetry buried inside Carlson’s realism, and it becomes clearer once we move from the behaviour of great powers to the behaviour of the smaller states around them. Carlson is quite right that Russia and China possess security concerns which cannot simply be wished away: Russia predictably dislikes the expansion of an American-led military alliance towards its borders, while China predictably dislikes American military power operating in East Asia and regards Taiwan as central to its own national story.

Any diplomatic strategy which refuses to understand those perceptions because Moscow and Beijing are adversaries will be poorer for it. The problem comes when understanding shades into privileging. If Russian fears about NATO count as structural realities which the West has an obligation to accommodate, then Polish, Baltic and Ukrainian fears about Russia must surely count as structural realities too; if China’s desire to prevent an American military presence close to its coastline is intelligible, then Taiwan’s desire not to be forcibly absorbed by a vastly more powerful neighbour is intelligible for precisely the same reason.

Carlson’s version of realism often seems to perform only the first half of this exercise. The security anxieties of great powers appear as facts of geography and history, while the anxieties of weaker states appear as unfortunate complications produced by their attachment to rival powers. Poland’s desire for NATO becomes evidence of Western encroachment rather than an equally predictable response to Polish history; Ukraine’s effort to escape Russian domination becomes something the West has manipulated rather than something Ukrainians might independently desire.

The descriptive proposition that great powers seek spheres of influence is difficult to dispute. The normative conclusion which sometimes follows, that smaller countries should therefore accept those spheres because resisting them is dangerous, is something else entirely. At that point sovereignty begins to mean that weak states may choose their futures provided no sufficiently powerful neighbour finds the choice inconvenient.

There is a strange contradiction here for a political worldview so attached to the nation state. Carlson’s argument repeatedly asks us to take Russian and Chinese national interests seriously, while treating Ukrainian or Taiwanese national interests as subordinate to the realities imposed by geography. Rees-Mogg, despite having the general appearance of a man who might still regard the Congress of Vienna as breaking news, ends up defending the more modern principle: that small nation states possess agency too, and that joining an alliance can itself be understood as an expression of sovereignty rather than evidence that sovereignty has disappeared.

That doesn’t mean every alliance choice made by a smaller state imposes an obligation upon everybody else to defend it. Ukraine’s right to seek NATO membership and Britain’s obligation to accept Ukraine into NATO are different questions, just as Taiwan’s right to remain self-governing and America’s obligation to fight for it are different questions. What it does mean is that a serious realism should apply its realism consistently. Geography constrains the small as well as the large; security fears exist in Warsaw as well as Moscow; and a theory which treats only the preferences of the strongest actor as immovable facts risks redescribing power as legitimacy.

Unfortunately, Kant

At this point, having begun with Tucker Carlson interviewing Jacob Rees-Mogg, we have somehow arrived at Immanuel Kant. I accept responsibility for this, although in my defence Carlson is the one who started discussing first principles.

The temptation here is to produce a vaguely Kantian argument which says that bad consequences would follow if everybody minded their own business, but that isn’t quite enough. The more interesting question concerns the maxim behind Carlson’s position and whether the institutions from which states benefit could continue to exist if everybody acted upon it.

Suppose the rule were that a state should never expose its citizens to serious cost or danger in defence of another state unless its own territory faces an immediate threat. As a principle of restraint it sounds attractive, particularly because it recognises that governments have special obligations towards the people whose lives they can actually command. Yet if every member of an alliance applies the same rule, the nature of the alliance begins to change. A security guarantee is valuable precisely because an aggressor expects it to survive the moment at which honouring it becomes costly; a guarantee which disappears whenever enforcement becomes dangerous is less a deterrent than an expression of sympathy.

Article 5 of the North Atlantic Treaty could, under the strongest version of the Carlson Doctrine, be substantially shortened: an attack upon one shall be considered an attack upon one.

The joke works because the contradiction is structural. The problem isn’t merely that universal non-intervention produces outcomes we dislike; it is that the benefits Carlson’s America enjoys from alliances depend upon enough other countries refusing to apply the principle consistently. A country can benefit from an expectation of reciprocal defence while reserving the right to treat every test of that expectation as somebody else’s problem, but only for as long as its partners don’t behave the same way.

This is close to the logic in Kant’s example of the false promise. The liar can benefit from promising to repay money because the institution of promising still carries credibility; if everyone understands promises to be meaningless whenever keeping them becomes inconvenient, the practice which makes the lie useful begins to collapse. Something similar happens with collective security. Britain benefits from treaties which adversaries expect Britain to honour, America benefits from allies believing American guarantees, and every NATO member benefits from the assumption that attacking a small Baltic state isn’t simply a bilateral quarrel between that state and Russia.

None of this establishes the opposite maxim, according to which every state must fight whenever another state is wronged. Universal intervention would be morally reckless and practically impossible, and if the only alternatives available were crusading liberal internationalism or complete indifference, Carlson would have a much stronger case.

The more plausible principle is also, inevitably, less exciting: states may sometimes accept proportionate costs and risks in order to sustain alliances, guarantees and rules whose continued credibility contributes to their own security, while retaining an obligation to judge particular interventions too costly, too dangerous or too remote from any defensible national interest. Nobody is likely to embroider that on a regimental flag, but foreign policy has rarely rewarded brevity.

What this gives us is restraint without converting restraint into indifference. It allows us to say that the danger of nuclear escalation constrains what Britain should do in Ukraine without concluding that Ukraine’s fate is irrelevant; it allows us to defend NATO without supposing NATO should expand without limit; and, crucially, it allows a doctrine which supports collective security in one case to reject a discretionary war in another.

Which brings us back across the Atlantic.

Why die for Washington?

For most of the Cold War, the foundational proposition behind NATO was that an attack upon Western Europe could eventually bring Americans across the Atlantic to fight. If Soviet forces moved into West Germany, American soldiers might die, not because Frankfurt had somehow become American territory, nor because the Fulda Gap possessed a mystical claim upon young men from Pennsylvania, but because successive American governments had concluded that preventing Soviet domination of Western Europe formed part of America’s own security, and that making this commitment credible in advance reduced the likelihood that it would ever have to be fulfilled.

This worked, by the way. The Berlin Wall fell, and the Soviet Union collapsed, without a war in Western Europe, without a nuclear weapon being fired, indeed without any direct military confrontation between the USA and USSR.

Then, on 11 September 2001, the United States itself was attacked, and on the following day NATO invoked the principle of Article 5 for the first time in its history. The direction of the guarantee reversed: an alliance created largely to reassure Europeans that America would come if they were attacked was now being asked what Europeans would do when America had been attacked instead.

This gives us a counterfactual which is deliberately uncomfortable. Imagine European governments applying the strongest version of Carlson’s principle on 12 September 2001. Why should a British soldier risk his life because terrorists attacked New York? Why should a Canadian die because an aeroplane hit the Pentagon? Afghanistan was thousands of miles away, the Taliban posed no conventional threat of invading Britain, and American foreign policy had unquestionably contributed to hostility towards the United States in parts of the Middle East. Military involvement might provoke terrorism against European cities, while any attempt to remake Afghanistan could embroil allied countries in a conflict whose objectives gradually expanded beyond the immediate response to al-Qaeda.

Every one of those considerations contains the beginning of a serious argument. Taken together as a reason why the attack was simply none of Europe’s concern, however, they feel strangely inadequate.

Why die for DC?

The formulation is intentionally ugly because, for many readers, the historical distance which cushions Why die for Danzig? disappears. The foreign capital is now familiar, the dead have photographs, and the state asking its allies whether their promises mean anything is the United States itself.

That doesn’t mean British or Canadian soldiers subsequently acquired an unlimited obligation to participate in whatever Washington chose to call part of the response to terrorism. Indeed, the years after 9/11 demonstrate why the distinction is indispensable. Afghanistan could be defended initially as a response to the organisation responsible for the attacks and the regime sheltering it, while still becoming strategically incoherent as the mission expanded; Iraq, by contrast, shouldn’t be smuggled into the moral authority of Article 5 at all. The invasion of 2003 remains one of the strongest arguments for scepticism towards foreign-policy establishments which claim that distant threats must be confronted before they reach us.

The lesson is therefore not that solidarity after 9/11 vindicates intervention elsewhere. It is that America itself has depended upon other countries refusing to define national interest so narrowly that any attack beyond their own borders automatically belongs to somebody else.

The reciprocal question is what gives the argument its force. Carlson asks, quite reasonably, what Americans or Britons owe to foreigners; any principle which claims to govern international relations must also be capable of answering what Americans and Britons are entitled to expect foreigners to owe them. A country which wishes to benefit from alliances, forward defence and collective deterrence cannot coherently regard the burden of maintaining those institutions as an imposition only when it falls upon itself.

That is the categorical imperative with some blood in its veins. The question is no longer merely whether I would prefer to avoid the costs of somebody else’s crisis, but whether I could rationally will a system in which everyone adopted the same rule when the crisis happened to be mine.

The trouble with minding our own business

Carlson’s strongest argument therefore survives rather more of this essay than his weakest remarks about the Second World War might suggest. Governments should be extremely reluctant to spend the lives of their citizens; injustice abroad doesn’t automatically generate a duty to intervene; nuclear weapons impose serious constraints upon what prudence allows; adversaries possess fears and interests which don’t disappear because their governments are unpleasant; diplomacy sometimes requires compromise; and a political culture which rediscovers Munich every time negotiations are proposed is perfectly capable of leading itself into disaster.

There are wars which aren’t ours. There must be.

The mistake is to imagine that restraint can be identified simply by asking whether the threatened territory belongs to us, because states act within systems of repeated interaction in which intervention and non-intervention both alter the environment in which the next decision will be made. Arming Ukraine can prolong fighting; refusing to arm Ukraine can change Moscow’s expectations about the profitability of aggression. Extending a guarantee can create escalation risks; allowing guarantees to become incredible can create different escalation risks. A visible military deployment can frighten an adversary; a visible withdrawal can encourage one.

Minding our own business is therefore not an absence of foreign policy. It is a foreign policy, and other countries are allowed to notice it.

This is the deeper weakness in Why die for Danzig? The question places one French life on one side of the scales and one foreign city on the other, after which the answer is practically contained in the comparison. Yet Danzig wasn’t valuable because of some intrinsic quality of Baltic masonry, lovely though it is; it had become the place at which a wider set of questions happened to become concrete. Did previous agreements impose obligations? Could Germany continue revising borders through threats and force? Was the latest demand likely to be the last? Would another concession reduce the probability of war, or merely change the circumstances under which war eventually occurred?

The same structure appears, imperfectly and with very different legal obligations, in Taiwan. An American sailor wouldn’t be fighting because Taipei possesses some mystical claim upon Nebraska; the relevant calculation would include deterrence, regional balances of power, the credibility of American alliances, the behaviour of Japan and South Korea, the possibility of nuclear proliferation and the consequences of allowing a major power to annex a self-governing neighbour by force. None of those considerations automatically outweighs the risk of a US-China war, but excluding them because Taiwan isn’t America would be an extraordinarily thin definition of American interest.

And the same logic, running in the opposite direction, explains why British, Canadian and other allied governments didn’t treat the attacks of 11 September as a private American tragedy. Manhattan wasn’t British soil. The Pentagon wasn’t in Ontario. Yet the alliance had been built on a prior judgement that some threats to another member were connected to one’s own security strongly enough that the distinction between their problem and our problem couldn’t always be drawn at the water’s edge.

The purpose of collective security, at its most simple and most defensible, is precisely to make somebody else’s war your problem before your problem becomes somebody else’s war.

That proposition can itself become dangerous if treated as a blank cheque. It doesn’t tell us how many weapons to send to Ukraine, whether Ukraine should join NATO, whether America should fight for Taiwan, what level of defence spending Britain requires, or where the danger of escalation exceeds the value of deterrence. Those questions remain questions of judgement, and any doctrine which pretends to answer them automatically is probably hiding its politics inside its premises.

Carlson is right, therefore, to demand that somebody answer the human question before proposing war: why should our people die for this? A government which cannot answer deserves suspicion. Where his argument becomes myopic is in assuming that this refers only to the particular patch of territory currently under dispute, as though foreign policy were a succession of disconnected transactions and the only consequences worth counting were those immediately visible.

Sometimes the thing being defended is indeed too remote, the obligation too weak, the objective too vague or the risk too great. Sometimes a government should look at an ally, a threatened state or an international rule and decide that none of them justifies what intervention would cost. If that possibility disappears, then we haven’t answered Carlson so much as proved his point about interventionists.

But sometimes the thing being purchased is larger than the place where the bill happens to arrive: an alliance whose credibility deters future attacks; a balance of power which prevents a hostile state from dominating its neighbours; a reciprocal expectation that allies won’t abandon one another whenever solidarity becomes expensive; a norm against territorial conquest; or simply the possibility that demonstrating a willingness to resist today reduces the chance that anybody will have to fight tomorrow.

Those things aren’t sacred. They aren’t priceless. They can be badly designed, recklessly extended, cynically invoked and catastrophically misapplied.

They are nevertheless real.

Which means Why die for Danzig? remains a good question, just not a complete one. It forces the advocate of intervention to account for the human cost, and no serious foreign policy should resent being made to do that.

The trouble begins when we mistake the question for the answer.

There’s a particular misery in expecting betrayal and then experiencing it as a shock.

This, I think, explains quite a lot of the emotional life of the British left. Not all of it, obviously; even I’m not foolish enough to attempt a Unified Theory of Labour People Being Cross Online, which would be less an essay than a cry for help. But it explains something real about the argument between Labour’s moderate left and its more idealistic left: the argument is so bitter not because the two sides are strangers, but because they’re relatives.

A Conservative government disappoints the idealistic left by being a Conservative government. It cuts, privatises, moralises, disciplines, tightens, lectures and, depending on the vintage, occasionally declares that society either doesn’t exist or should be made available to shareholders by Wednesday. None of this is surprising. It may be enraging, but it isn’t intimate.

Labour is different. Labour disappoints from inside the family story.

This is why Freud’s phrase, the narcissism of small differences, is useful here. In Civilization and Its Discontents, Freud observed that it’s often adjoining and related communities, not wholly alien ones, that sustain the fiercest feuds. The line usually quoted is this: “It is precisely communities with adjoining territories, and related to each other in other ways as well, who are engaged in constant feuds.” The point isn’t that their differences are imaginary. Small differences can be real, consequential and morally serious. The point is that closeness changes their emotional charge. It makes disagreement feel like betrayal, compromise feel like contamination, and ordinary political difference feel like a struggle for possession of the same soul.

The moderate Labour left and the idealistic left don’t hate each other because they disagree about everything. They hate each other because they agree about too much.

Both, broadly speaking, live somewhere within the same moral tradition. They believe in public services, redistribution, dignity at work, protection against market cruelty, some version of equality, some suspicion of inherited privilege, some commitment to collective provision, some discomfort with a society in which wealth can buy escape from shared failure. They sing from the same battered old hymn sheet in the same “Broad Church”.

That shared inheritance is precisely what makes the argument so painful. The idealistic left doesn’t look at Labour moderates and see Tories - or rather, it sometimes says it does, especially when it’s in one of its more theatrical moods, but that isn’t quite the underlying grievance. A Tory is an opponent. A Labour moderate is something more infuriating: someone who uses the same language, claims the same history, honours the same political saints, and then turns around and says that half the programme is unworkable, unaffordable, electorally toxic, institutionally naïve, procedurally impossible or best left for another parliament.

The anger comes from the fact that moderate Labour is close enough to the idealistic left to be judged as a failed version of itself, rather than a successful version of something else.

That’s the agony of proximity, summed up in a nutshell.

And I think this has probably got worse over time. For older kinds of radical politics, this proximity was easier to manage because the distance was clearer. There have always been arguments between Labour moderates, socialists, communists, trade unionists, liberals, syndicalists, Trotskyists, Bennites, soft-left organisers, Fabian gradualists, municipal socialists and so on. The left has never lacked internal disagreement. If anything, one of its less charming inheritances is a capacity to turn even minor procedural questions into acts of metaphysical warfare. But the forms of contact were different.

There was a time, and not a lost golden age so much as a different configuration of misery, when many radicals and many Labour people inhabited more separate spaces. In the early 2000s, for instance, the average radical student society and the average university Labour club were not necessarily recruiting from identical pools, even where they existed on the same campus. The anti-Iraq War, anti-globalisation, anti-capitalist campaign space often had its own meetings, jargon, friendships and rituals of exasperation. Labour students had theirs. Each group could dislike the other at a reasonably healthy distance. The radicals could regard Labour as a compromised imperialist machine; Labour people could regard the radicals as unrealistic purists with an overdeveloped taste for fiery denunciation. Everyone could go home feeling confirmed in their prejudices, which is one of the underrated pleasures of politics.

This separation had obvious costs. I concede as a self identified Labour moderate that Labour has often needed pressure from its left. There are plenty of instances where the people dismissed as unrealistic irritants have turned out to have been correct earlier than the mainstream: on deregulation, the banks, climate, and a dozen other questions, the respectable centre of Labour opinion has not exactly covered itself in glory by taking the slower route. Distance can preserve comity, but it can also preserve complacency. If Labour moderates never hear from the idealistic left, or hear from it only as a caricature, they become too comfortable with the assumptions of the institutions around them.

Still, distance did one useful thing. It kept expectation under control.

If you believed Labour was fundamentally a capitalist, imperial, managerial party, you didn’t have to spend every day being surprised that Labour behaved like one. You might vote for it reluctantly as the least-worst option. You might campaign against it. You might try to influence it from outside. You might have some grudging theory about lesser evils, parliamentary arithmetic and the need to stop the Conservatives from feeding another public service into the shredder. But you didn’t necessarily expect Labour to become the pure instrument of your politics. Your disdain performed a psychological function: it protected you from disappointment.

That arrangement was already weakening before Jeremy Corbyn became Labour leader, partly because social media had made political distance much harder to sustain. The old internet had its own insanities, but it was often more segmented. Forums, mailing lists, blogs, student societies, pamphlet groups and factional publications each had their own audiences and their own little economies of approval. The contemporary internet is much more like being trapped in the same endless car journey with everyone you’ve ever disagreed with, except the car is on fire, the algorithm is choosing the music, and someone in the back seat has just discovered quote-tweeting.

But Corbyn changed the structure of expectation in a more profound way. He didn’t merely represent a left-wing challenge to the Labour establishment. He made the capture of Labour by the idealistic left seem plausible. For a period, the party - to the Left - didn’t look like a grim electoral compromise to be endured, or a hostile machine to be pressured from outside. It looked like a vehicle. It looked like a door that had unexpectedly opened. The old radical posture, in which Labour was irredeemably compromised and therefore emotionally discounted, became harder to maintain when an unreformed Bennite socialist had actually won the leadership of the Labour Party, then in 2017 fought and nearly-almost-didn’t-lose a general election as Labour’s candidate for Prime Minister.

Whatever else one thinks about Corbynism, and I have no desire here to relitigate every argument from 2015 to 2019 because some of us have work in the morning, it changed the emotional ownership of Labour. It brought into and around the party a large number of people whose politics were not formed by the old assumption that Labour would inevitably fail them. For them, Labour was not simply the disappointing parliamentary wing of a broader labour movement. For a while, it became the site at which politics might be transformed.

That’s why the aftermath was so bitter. Once you’ve seen your politics take the leadership of the Labour Party, it’s hard to return to detached contempt. The institution has been too close. It has worn your colours. It has sung your songs. It has sent fundraising emails in your idiom. It has made promises that sounded, to some people, like the end of managed decline and the beginning of moral seriousness. When it then reverts, compromises, retreats, triangulates or elects a leadership whose whole pitch is restoration after rupture, it doesn’t simply feel like defeat. It feels like dispossession.

This is where I think the narcissism of small differences reasserts itself. The hostility between the moderate Labour left and the idealistic left isn’t produced by simple ideological distance. If distance alone explained hostility, Labour people would spend most of their emotional energy arguing with Conservatives and Reactionaries of the right-wing parties. They do argue with Conservatives, obviously, but those arguments often have a different texture. They’re oppositional, not familial. The peculiar venom of Labour-left arguments comes from adjacency: from the sense that the other side is occupying territory that ought properly to belong to you.

The moderate left says: we share your values, but we have to win.

The idealistic left says: we share your values, which is why your compromises are unforgivable.

The moderate left says: look at the electorate, the press, the markets, the parliamentary party, the civil service, the bond markets, the inheritance from the last government, the sheer difficulty of governing a fractious country under fiscal constraint.

The idealistic left says: look at poverty, housing, climate, Gaza, inequality, collapsing public services, child hunger, the cruelty of the border, the grotesque fact that this wealthy country keeps rediscovering Dickensian social conditions and calling it realism.

Both are looking at something real. That’s what makes the argument hard.

There is a boring-but-true answer here, but it doesn’t quite settle the issue. Labour is a mass electoral party in a first-past-the-post parliamentary system. It can’t function as the direct expression of one ideological tendency because it needs to assemble a coalition large enough to win power across places that don’t all want the same things. It has to speak to public-sector workers and small business owners, renters and homeowners, graduates and non-graduates, metropolitan liberals and socially conservative Labour switchers, trade unionists and people who think trade unions are always holding the country to ransom, people who read the Guardian and people who think the Guardian should be composted for the good of national morale. It has to win under media conditions that are often hostile, economic conditions it doesn’t choose, and institutional constraints that no amount of moral clarity can simply wish away.

This means Labour will always translate moral claims into something smaller, duller and more compromised than the idealistic left wants. Sometimes that translation will be defensible. Sometimes it will be cowardly. Sometimes it will be a strategic necessity dressed up as a virtue. Sometimes it will be a failure of imagination disguised as seriousness. Sometimes it will be, in the technical language of political analysis, absolute nonsense. But the act of translation itself is unavoidable. A party that wants to govern must turn values into manifesto lines, spending commitments, legislative time, delivery mechanisms, departmental priorities, fiscal rules, procurement decisions, parliamentary management, press lines and things ministers can say on breakfast television without detonating the day’s comms grid.

That isn’t romantic. It isn’t even especially noble. But it is what a governing party does.

The idealistic left often wants Labour to behave as a movement of moral clarification. Labour, when it is serious about power, behaves as a coalition-management machine with a moral inheritance attached. This doesn’t mean the idealistic left is wrong to want more. It means the institution it’s trying to pressure is not built to become what it wants.

This is the contradiction that produces so much rage. The idealistic left, both within Labour and without it, has a systemic critique of Labour that explains perfectly why Labour disappoints it: Labour is too attached to electorally-motivated caution, too constrained by elite opinion, too eager to reassure capital, too frightened of the press, too captured by the assumptions of Westminster, too willing to trim its principles in pursuit of office. But having made that diagnosis, parts of the idealistic left then react to each new compromise as though it were an inexplicable personal betrayal rather than the predictable behaviour of the institution they both helped to create and claim to understand.

This is not a plea for cynicism. Cynicism is often just wounded idealism trying to sound clever. Nor is it a defence of Labour moderates, who have their own bad habits: confusing caution with wisdom, treating left-wing pressure as childish even when it’s morally necessary, mistaking electability for virtue, and using the alleged sensibilities of the average voter as a ventriloquist’s dummy for whatever they already wanted to do. There is a type of Labour person who can turn any moral demand into a focus-group problem and any structural injustice into a comms risk. This person should be listened to, because politics does involve persuasion, but they should be gently prevented from designing civilisation.

But idealism has its own failure mode. It can confuse intensity with strategy. It can mistake the refusal to compromise for the possession of a programme. It can treat the moral inadequacy of moderation as proof that maximalism would work. It can assume that because a demand is just, the route to implementing it must be straightforward. It can understate the coercive force of institutions, the conservatism of the mass of voters, the limits of parliamentary time, the brittleness of governing coalitions, and the tendency of the right to unite around power while the left forms a working group on whether the concept of unity is sufficiently emancipatory.

This is why Labour will always disappoint the idealistic left. Not because the idealistic left is wrong about everything. It plainly isn’t. Often it’s right about the direction of travel before the rest of the political system can admit the road exists to travel down. But being right about the moral horizon is not the same as having a theory of how a mass party reaches it while staying electorally viable, institutionally coherent and capable of governing the country it actually has rather than the country its members wish existed.

The distinction that matters is between pressure and expectation.

Pressure is necessary. Labour should always be pressured from its left. It should be made to feel uncomfortable about poverty, climate, housing, welfare, migration, war, wages, private profit in public services, democratic reform and the hundred small cruelties that polite managerial politics learns to file under unfortunate trade-offs. A Labour Party without pressure from its left becomes grey, bloodless, complacent and too easily absorbed into the worldview of the state it’s supposed to change. The idealistic left performs a vital function when it keeps alive claims that would otherwise be buried under realism, process and the deadening phrase ‘tough choices’.

Expectation is different. Expectation says that Labour would naturally become a radical moral instrument if only the wrong people stopped preventing it. Expectation treats each compromise as a deviation from Labour’s true essence rather than a feature of its position in the political system. Expectation turns pressure into recurring heartbreak.

The idealistic left doesn’t need to stop pressuring Labour. It needs to stop being surprised by Labour.

Likewise, Labour’s moderates don’t need to accept every demand made from their left. Some demands are poorly thought through. Some would lose, or cost more than they would buy. Some would fail on implementation. Some are slogans pretending to be policy. Some are admirable in aim and useless in design, which is a more common category than anyone involved in politics likes to admit. But moderates do need to stop treating left pressure as a kind of adolescent weather system. The fact that a demand is inconvenient doesn’t mean it’s unserious. The fact that a policy is difficult doesn’t mean it’s impossible. The fact that a group of people are annoying doesn’t mean they’re wrong, which is a tragic but unavoidable fact of political life.

The best version of Labour politics probably needs both impulses. It needs the idealistic left to widen the moral imagination of the party, and it needs the moderate left to ask what survives contact with elections, institutions, practical delivery and the public finances. The trouble is that each side tends to experience the other not as a corrective, but as a kind of corruption. To the idealistic left, moderation looks like betrayal. To the moderate left, idealism looks like self-indulgence.

This is not unique to Labour. It’s a recurring pattern among related political communities. The faction next door is often more infuriating than the enemy across the road because the faction next door competes for the same symbols, the same history, the same members, the same moral vocabulary and the same claim to authenticity. The fight is not only over policy. It is over who gets to say what the project really means.

That, in the end, is why proximity hurts. The stranger merely opposes you. The relative lets you down.

So yes, Labour will always disappoint the idealistic left. It will disappoint because it’s close enough to the ideal to be judged by it, but compromised enough by its actual function to fail it. It will disappoint because it carries the language of transformation into the machinery of coalition-building. It will disappoint because it asks morally serious people to accept partial victories, ugly sequencing, tactical silence, delayed justice and policies that move in inches when the scale of the problem demands miles. It will disappoint because it sometimes lacks courage, and sometimes because courage alone wouldn’t solve the problem. It will disappoint because any party capable of forming a government in Britain will, by definition, contain people whose politics are less radical than the leftmost people needed to campaign for it.

None of this means the idealistic left should go quiet. Quite the opposite. The point of understanding disappointment is not to make peace with every compromise; it’s to know which compromises are worth fighting, which are structurally baked in, and which are symptoms of a party losing touch with its own moral reasons for existing.

But pressure only remains useful when it’s attached to a theory of what the institution actually is. Labour is not the parliamentary wing of moral innocence. It’s not a seminar, a protest camp, a trade union branch, a tenants’ meeting, a student occupation or a group chat with standing orders. It’s a mass electoral party in a country that often votes cautiously, reads politics through partial and hostile media, rewards the right for forms of ruthlessness it punishes on the left, and expects Labour governments to repair the roof while being shouted at for the weather.

The idealistic left is right to want more. It’s often right about what’s missing. It’s sometimes right before everyone else, which is one of the reasons everyone else finds it so irritating. But wanting Labour to be better is not the same as expecting Labour to become something other than Labour.

That is the agony of proximity. It’s not that the idealistic left and the moderate Labour left are too far apart to understand each other. It’s that they are near enough to recognise the same inheritance, near enough to feel entitled to it, and different enough to keep accusing each other of betrayal.

The tragedy isn’t that they have nothing in common. The tragedy is that they have just enough.

(See my earlier essay on Mass Effect as an example of this!)

This isn’t a fatal criticism of games. Games are machines, and machines need limits. A game can’t contain every choice a player can imagine, because that game would never be finished, never released, and possibly never survive contact with the first player determined to stack thirty-seven cheese wheels into a staircase. Choice in games is always bounded, always authored, always less free than it first appears. Still, the best role-playing games are interesting precisely because they make those boundaries feel meaningful rather than merely mechanical. They don’t let you do everything; they make you feel the cost of doing something.

That’s where Fallout: New Vegas still stands apart.

Not because it offers total freedom, because it doesn’t. Not because every faction is equally plausible, because it isn’t. Not because it escapes the silliness of Fallout, because nothing set in a world of bottlecap currency, giant mutant bugs, and implausibly durable pre-war branding can claim to have escaped silliness altogether. The point is stranger than that. New Vegas works because it understands that political choice isn’t the same as political agency, and that being forced to pick a ruler isn’t the same as being free.

I’ve always loved Fallout. Even the Bethesda ones. Even, in a complicated and slightly defensive way, Fallout 76. They all have their pleasures, and some of those pleasures are very stupid, which is fine because stupid pleasures are still pleasures. But New Vegas has a quality most of the other recent instalments only gesture towards: the feeling that people are actually building something after the apocalypse, not just staring at two-hundred-year-old skeletons in bathtubs with the helpless awe of people who’ve only just noticed that there’s been a war. The Mojave isn’t simply ruined. It’s contested. It has trade, roads, settlements, soldiers, raiders, gamblers, farmers, opportunists, ideologues and bored people trying to make a living under the shadow of forces too large for them to steer.

That difference changes the moral texture of the game. In Fallout 3, the central conflict is still basically readable as shining techno-knights versus shadowy authoritarian scientists, with a few irradiated detours along the way. In Fallout 4, there are more factions, but most of the choices still settle into fairly familiar moral grooves: some groups are plainly better than others, and the main discomfort comes less from the ambiguity of the factions than from the sheer Bethesda awkwardness of deciding which set of people you can bear to obliterate. New Vegas, by contrast, gives you a flawed republic, a fascist slave empire, a technocratic autocrat, and a Yes Man path whose apparent freedom turns out to carry its own whiff of irresponsibility.

That’s not moral relativism, by the way. Caesar’s Legion is undeniably and irredeemably evil. This shouldn’t need labouring, though the internet’s capacity to over-intellectualise fascist aesthetics means perhaps it does. The Legion’s order is built on slavery, misogyny, terror, and a grotesque fantasy of civilisational renewal through domination. Caesar may be intelligent, articulate and even persuasive in the limited sense that ideologues can be persuasive when they’ve arranged the whole world to make their cruelty look like logic. But a coherent argument for barbarism remains barbarism. The fact that New Vegas lets him explain himself doesn’t make the Legion morally complex in the sense of being somehow half-right; it makes the horror sharper, because the game understands that evil is not always stupid or cartoonish.

The more interesting ambiguity lies elsewhere. The New California Republic is obviously preferable to the Legion, which is not a high bar, but it’s also bloated, exhausted, corruptible and imperial in the soft, procedural way bureaucracies can be imperial while insisting they’re only extending order. Mr House is brilliant, but his brilliance is the brilliance of a man who has mistaken civilisation for himself. The independent Vegas route is superficially attractive, because who doesn’t like freedom, especially when it comes with a robot army and the chance to annoy everyone powerful? Yet it also feels like the kind of politics that mistakes the removal of authority for the creation of justice. Knock down the old order, install some upgraded securitrons, call it liberation, and try not to think too hard about what happens once the courier wanders off to pick flowers, gamble, or shoot another deathclaw in the face.

So the question at the heart of New Vegas isn’t simply who should rule the Mojave. It’s whether the game’s political imagination can escape the premise that someone must.

For most of its running time, the answer is no. The machine requires an ending, and the ending requires a victor. The Battle of Hoover Dam waits at the end of the game like history reduced to a loading screen: factional tensions, local grievances, trade routes, ideology, water, electricity, security, imperial ambition and personal revenge all compressed into a climactic military decision. In ordinary game terms, this makes sense. A role-playing game needs climax. A factional plot needs resolution. A player who has spent dozens of hours accumulating power, reputation, equipment, companions and opinions will usually expect the story to recognise that accumulation by handing them a lever large enough to move the world.

Yet the odd brilliance of New Vegas is that, by the time it gives you the lever, the most morally satisfying response may be not to pull it.

This is where the downloadable content, especially Lonesome Road, does something quietly severe to the larger game. The base game is full of political temptation. The DLCs turn that temptation back on the player. Dead Money is about letting go, and it doesn’t merely say this; it builds greed, scarcity and punishment into the structure of play. Old World Blues is a gloriously ridiculous mad-science grotesque, but even there the Old World is not a lost paradise so much as a set of brilliant people endlessly mutilating the future because they can’t stop mistaking intelligence for wisdom. Honest Hearts gives the Legion’s violence a wider moral and religious context, even if it’s probably the least elegant of the four main expansions.

Then comes Lonesome Road, which narrows the grand political argument into something lonelier, harsher and more direct.

The Divide isn’t another amusing zone full of post-apocalyptic stuff to collect, although there’s plenty to collect because the series remains the series and I remain a terrible little scavenger goblin. It’s a wound. It’s a place where delivery, infrastructure, military power and forgotten technology meet catastrophe. Ulysses, the other courier, is not just an antagonist with a good coat and an overdeveloped fondness for portentous speech. He is a rebuke. He looks at the player character and sees not a blank-slate adventurer, but an agent of history who has been moving things between powers without understanding what those movements make possible.

That is a clever accusation to level at a player, because moving things between powers is what players do in role-playing games all the time. We carry messages. We deliver packages. We activate machines. We retrieve artefacts. We repair systems. We pick up the thing, take it to the person, receive the reward, and rarely ask whether the real story was the errand rather than the boss fight we did along the way. Lonesome Road takes that innocent little quest-loop and makes it feel implicated. The courier’s job is not heroic destiny. It’s logistics. And logistics, when attached to power, can become violence.

This is why the nuclear choice at the end of Lonesome Road feels so different from many other nuclear moments in the series. In Fallout 3, you can detonate Megaton as an early act of spectacular cruelty, a grotesque side-option dressed up in the franchise’s usual dark comedy. In Fallout 4, nuclear destruction becomes part of the scripted grammar of factional victory. In Fallout 76, nukes become, among other things, a tool in the player’s box; apocalypse as endgame resource loop, which is a sentence so bleakly funny that I resent the game for making it accurate.

But in Lonesome Road, the missile isn’t just a toy or a spectacle. It sits at the crossing-point of all the themes New Vegas has been circling: factional loyalty, historical grievance, personal responsibility, infrastructure, revenge, nationalism, and the fantasy that one terrible act can clarify a messy world. The option to fire the missile at a faction you oppose is, in crude game terms, another choice. It gives you agency. It lets you hurt your enemies. It lets you mark the world according to your alignment. Yet the whole journey through the Divide has been teaching you to distrust that kind of clarity.

This, for me, is where New Vegas becomes more than a very good RPG with memorable factions. The game’s deepest political thought isn’t that every side has a point, because every side does not. Nor is it that all power is equally corrupting, because that’s too lazy. Its sharper claim is that the available structures of power may force morally serious people into choices they don’t recognise as their own.

The Legion must lose; that’s the easy part. But the NCR is doing it wrong. House is too cold and too vain. And Yes Man is too flippant a vehicle for something as dangerous as sovereignty. Every route offers some practical reason to choose it and some moral reason to hesitate. The player can, with enough effort, improve the outcomes around the edges: save a town here, settle a dispute there, make one faction slightly less awful, persuade one companion towards a healthier future, and enjoy the small civic satisfaction of leaving a place less broken than you found it. Yet when the great machinery of the ending arrives, the game still asks for a ruler.

That’s why my preferred ending to New Vegas isn’t available. I wanted to refuse Hoover Dam. I wanted the Legion broken, certainly, but I didn’t want to become history’s delivery mechanism for another power. I didn’t want the NCR to inherit the Mojave as a reward for being less monstrous than the alternative. I didn’t want House’s glittering autocracy, and I didn’t want to declare an independent Vegas under the guidance of a grinning robot assistant whose constitutional theory appears to be: don’t worry, it’ll probably be fine. I wanted to walk away from the master switch, leave the dam as a standoff rather than a coronation, and spend the rest of the game as a heavily armed community safety officer for nameless settlers and wandering traders.

I did something similar for Mass Effect too, I guess. Clearly I’ve got form.

Yes, this is not how endings work. I’m aware. I’m being unreasonable in the specific way players are unreasonable when they conflate a game’s failure to contain their desired choice with a moral defect in the work. Yet in this case, the unavailable ending reveals something the available endings can’t. New Vegas is so successful at making its politics feel materially real that the standard RPG demand for resolution starts to feel like a category error. The game teaches you that politics is hard, compromised, historical, local and full of people who’ll live with consequences long after the heroic courier has finished looting ammunition from a nearby filing cabinet. Then it asks you to solve politics by choosing who gets the dam.

That tension isn’t a failure, exactly. It may even be why the game endures. I think New Vegas is remembered so fondly not because it gives the player perfect agency, but because it lets the player feel the inadequacy of agency. The courier can change the Mojave, but not from outside it. Every intervention has a theory of power attached, even when the player doesn’t admit it. Every refusal is also a choice, and every choice leaves someone outside the frame.

There’s a useful comparison here with the way videogames often treat difficulty, morality and consequence as separate systems. Combat tests skill. Dialogue tests build optimisation. Morality tests which reward track you’ve decided to pursue. Consequence appears in a slideshow, ideally with narration. New Vegas is messier and better than that. Its politics work because they’re distributed across worldbuilding, quests, companions, reputation, geography, faction armour, supply lines, radio chatter, and all the tiny acts of repair or violence the player performs while allegedly doing something else. It’s a game about power in which power is not confined to the main quest.

That’s why Boone’s tragedy sits alongside the plastic dinosaur, why the comedy of Novac doesn’t cancel the grief under it, why the Strip’s neon promise feels both alluring and obscene, why the NCR’s paperwork can feel more politically revealing than its speeches, and why the Legion is most frightening when it sounds least like a gang of cosplaying raiders and most like a state. New Vegas doesn’t remove Fallout’s absurdity. It disciplines it. The pulp remains, but the pulp is pressed against consequences.

And then, at the end of Lonesome Road, it offers the bluntest possible version of consequence: a missile either flies or it doesn’t.

There is something almost unfashionable about the seriousness of that choice. Contemporary culture is quite good at presenting nuclear war as aesthetic: mushroom clouds, sirens, retro-futurist ruins, glowing green ambiance, the whole bright grammar of apocalypse. Fallout itself is obviously implicated in this, and often joyfully so. But Lonesome Road briefly cuts through the franchise’s own iconography and remembers that nuclear weapons are not merely symbols of the Old World’s folly. They’re instruments of finality. To launch one is not to score a point against a faction. It’s to participate, again, in the logic that ended the world.

No gods. No masters. No missiles.

That, in the end, is the politics of not choosing. It’s not passivity. It’s not apathy. It’s not the smug neutrality of a player who wants to remain morally unmarked while everyone else does the dirty work. It’s a refusal of the game’s most seductive simplification: that if the options are bad enough, the task is merely to pick the least bad ruler and call the discomfort maturity. Sometimes the discomfort is telling you something else. Sometimes the right instinct is not to seize the dam, crown the technocrat, bless the republic, unleash the robots or fire the missile. Sometimes the right instinct is to notice that the machine has mistaken decision for wisdom.

Of course, the game can’t fully let you live there. It has to end. The dam waits. The factions move. The slideshow rolls. History, even simulated history, abhors an unresolved quest marker.

But the fact that I still want an ending the game doesn’t offer is not a mark against New Vegas. It’s the reason I keep thinking about it. Most games make me wonder what would’ve happened if I’d chosen differently. Fallout: New Vegas makes me wonder what it would mean not to choose on the terms I was given.

That’s a rarer achievement. And, for a game about the ruins of a world destroyed by people who kept mistaking power for destiny, it may be the most fitting one.

Then he goes further. Man’s soul, Butler writes, is due to the machines; it is a machine-made thing. He thinks as he thinks, and feels as he feels, because of the work that machines have wrought upon him.

It’s a gloriously excessive sentence, which is to say that it’s exactly the kind of sentence one should be suspicious of and grateful for at the same time. Taken too literally, it overreaches; humans wouldn’t become biologically closer to apes if stripped of their tools, and cultural collapse isn’t the same thing as evolutionary reversal. The human body wouldn’t suddenly forget its species history because someone took away all our spoons, though I’m sure we’d be sad about it.

Still, the provocation works because it exposes something that ordinary language often hides from us: the naked human individual is, in most serious senses, an abstraction.

We’re used to thinking of technology as something added to human life. There is, first, the person; then there are the tools the person uses. A hammer helps a hand strike harder. A wheel helps a body move further. A computer helps a mind calculate, remember, organise and communicate. Technology, in this common-sense picture, is an accessory to an otherwise complete creature. It extends us, assists us, speeds us up, makes us comfortable. It’s an optional extra to build on our natural capacities.

This is not quite wrong, but it is too soft. The more serious argument is that technology does not merely assist the human being - it helps constitute the human being. We are not solitary biological organisms who happen to live among artefacts, but sociomaterial creatures whose capacities, expectations, habits, relationships and forms of thought have been made through our entanglement with technological and social technologies: tools, infrastructures, symbols and institutions. The modern human is not a body plus gadgets. The modern human is a body already organised through myriad technologies that act with, for, and upon us.

This sounds more dramatic than it is. In fact, part of the point is that it is boringly, physically, embarrassingly true.

Consider the body. The body itself is not a little sovereign kingdom, sealed off from the material world, making occasional use of convenient props. It is vulnerable, leaky, hungry, temperature-sensitive and, by the standards of most animals, rather poorly equipped. We have no fur worth mentioning. Our claws are pathetic. Our infants are catastrophically unfinished. Our teeth are not up to much unless one has a very generous definition of much. Left alone in the abstract state of nature, the human being is not an apex predator but a surprisingly anxious snack with opinions.

What changes the calculation is not merely intelligence, if by intelligence we mean something housed neatly inside the skull. What changes the calculation is what we might call technical life. The human animal cooks, cuts, stores, binds, carries, shelters, wraps, writes, counts, stitches, heats, cools, transports and preserves. We don’t simply live in an environment; we remake the environment into a set of supports for bodies that would otherwise be badly mismatched to many of the places in which we live.

A coat is not an optional extra to human life in Sheffield in January. It is part of what makes Sheffield in January an inhabitable condition for the kind of animal we are. A house is not merely a preferred lifestyle object; it is an external organ of temperature regulation, privacy, sleep, storage, kinship and social reproduction. Cooking is not only cuisine; it is a technological alteration of food that changes what bodies can digest and how much energy they can obtain. A toilet is not merely a convenience; it is a public health technology whose absence rapidly stops civilisation being a metaphor and starts making itself known at street level.

One can make this point with an over-simple inventory, which is sometimes the best kind:

The table is crude, but its crudeness is useful. It shows that the distinction between the natural human and the technological supplement is much less stable than it first appears. In practice, the thing we call a human capacity is often an arrangement between flesh, material artefacts, learned habits and shared social expectations. Human beings remember, but also write things down. Human beings travel, but also build roads and coordinate timetables. Human beings promise, but also sign, file, stamp, witness and archive. Human beings think, but they think through notebooks, screens, diagrams, books, classrooms, search engines, libraries and other people’s speech.

This is where the stronger form of sociomaterialism begins to bite. It is easy enough to say that society uses material things. It is more difficult, and more interesting, to say that society is made partly out of material things. A school is not just an idea shared by teachers, pupils and parents. It is classrooms, registers, uniforms, corridors, whiteboards, doors, toilets, timetables, safeguarding files, lunch queues, behaviour logs, playgrounds, exam papers, software systems and the strange little chair that exists only to make parents feel like overgrown children at consultation evening. Take those things away and the social institution does not remain intact as a pure concept. It begins to fail, because the institution was never merely conceptual in the first place.

The same is true of law, government, money, healthcare, employment, kinship and friendship.

A law is not only a rule in the mind. It becomes durable through texts, courts, police stations, forms, prisons, signage, databases, legal training, architectural authority and administrative procedure.

Money is not merely an “I promise to pay the bearer on demand…” system of trust. It is coins, notes, bank cards, ledgers, servers, regulatory systems, PIN machines, apps, anti-fraud protocols and the ordinary social miracle by which tapping a piece of plastic against another piece of plastic causes a sandwich to become yours.

Friendship is not only affection. It is birthday reminders, photographs, messages, shared calendars, trains booked weeks in advance, the sofa bed, the group chat, the saved voice note, and the uneasily powerful knowledge that someone has seen your message and not replied.

This doesn’t mean that material things are secretly in charge, nor that human beings are puppets of their tools. Hard sociomaterialism shouldn’t become modern-day animism. The point is more precise than that. Agency is distributed. Human beings act, but they act through arrangements that include objects, techniques, infrastructures and conventions. A person drives a car, but the possibility of that action depends on roads, fuel systems, traffic law, pedals, signage, maps, insurance, licensing, learned habits, engine design and the collective prior decision of many thousands of people to follow the rules of the road (e.g., not to treat red lights as decorative suggestions). To say that the driver is the only actor in the event is legally convenient and often morally necessary, but it is not a complete description.

Once we accept this, Butler’s machine-made soul starts to look less absurd. We don’t only survive through machines. We perceive, remember, judge and feel through them.

Take memory. It is tempting to think of memory as something located in the head, while diaries, photographs and phones are merely storage devices outside the real self. But this is not how memory works in ordinary life. A photograph does not simply preserve a memory; it shapes what is remembered, when it is recalled, how it is narrated, and which version of the past becomes socially available. A diary doesn’t merely record a thought after the fact; it lets a person think across time by returning to earlier versions of themselves. A phone contact list doesn’t merely store numbers; it changes what it means to know how to reach someone. A calendar doesn’t merely remind us of commitments; it allows us to make commitments beyond the limits of unaided recall.

This is not an exotic claim. It is what happens every time someone says they know where they’re supposed to be on Thursday, then immediately checks their calendar. They do know, in a perfectly ordinary practical sense, because their knowing is organised through a reliable external system. If the calendar disappears, the knowledge disappears with it, not because the person has become stupider, but because the cognitive arrangement has been broken.

The same is true of thought. Pen and paper don’t merely transcribe thinking. They permit certain kinds of thinking to occur. Long division in the head is one kind of cognitive act; long division on paper is another. Writing an essay entirely in unaided memory is not the same activity as writing with notes, drafts, books, search tools, marginalia and a half-remembered quotation one is increasingly certain one has made up. The thinking is not first completed in the brain and then lowered onto the page, like a piano being winched through a window. The thinking happens with the page. The sentence becomes visible, object-like, revisable. It can be moved, contradicted, sharpened, deleted, resurrected and grudgingly admitted to have been better in the first draft.

This is why technological change is never merely technical. To change the tools through which people think, remember and communicate is to change the practical conditions under which certain kinds of personhood become easier, harder, more visible or less possible.

That claim has an obvious relevance to digital life, but it shouldn’t be reduced to the internet. One of the weaknesses of contemporary technology debate is that it often behaves as though technology began when everyone’s aunt joined Facebook and started posting minion memes about betrayal. Digital platforms are important, but they’re not a metaphysical rupture in an otherwise natural human story. They’re a new phase in a much older pattern: humans making tools, tools remaking humans, and both becoming harder to separate over time.

Still, digital systems make the argument especially visible. A smartphone is not just an object in a pocket. It is address book, camera, map, diary, bank branch, newspaper, torch, archive, social venue, workplace, entertainment system, public square, surveillance device, confession booth, shopping centre and emotional tripwire. It doesn’t simply give the user more options. It reorganises attention, expectation, availability and social obligation. The read receipt, the typing indicator, the notification badge and the algorithmic feed are not neutral pipes through which human sociality flows unchanged. They alter the texture of waiting, the meaning of silence, the felt presence of other people, and the rhythms by which social life calls us back to itself.

This is particularly clear in conflict. A playground disagreement that once depended on who heard what, who said what, and who could gather a plausible audience can now be stabilised by screenshots, extended by group chats, performed through posts, revived by search, and redistributed beyond the original scene. The conflict is not simply online or offline, because those categories no longer describe two separate worlds. The phone in the pocket, the platform in the hand, the screenshot in the group chat and the body in the corridor are part of one sociomaterial situation. The insult is not just spoken. It is captured, cropped, forwarded, denied, recontextualised, archived and brought back at the worst possible time, which is to say the time when it will be most socially useful to someone else.

This is not a claim that technology causes everything. That would be a lazy argument with better stationery. A screenshot does not invent cruelty, status anxiety, loyalty, humiliation or revenge. But it changes their mechanics. It gives some versions of a conflict durability and others fragility. It makes evidence portable. It makes audience management more complicated. It gives denial new forms and accusation new supports. It turns ordinary adolescent stupidity into something with metadata.

The same point applies more broadly. The social world is full of things that only seem immaterial because the material support has become too familiar to notice. Bureaucracy looks like procedure until the database fails. Public health looks like policy until the sewage system breaks. The labour market looks like individual aspiration until transport costs make a job practically unreachable. Political participation looks like equal citizenship until one remembers that polling stations, postal ballots, ID rules, working hours, childcare, literacy, digital access and local information all decide whether abstract rights become usable rights.

This is where hard sociomaterialism becomes politically useful. If human capacities are partly constituted through material arrangements, then access to those arrangements is not a secondary matter. It is not merely about convenience, consumer choice or personal preference. It’s about the conditions under which people can act, think, care, remember, work, travel, participate and be recognised as competent social beings.

Consider disability. The most obvious example is also the one that should make us cautious about pretending that dependence on material supports is unusual. A wheelchair user is often described as dependent on a chair, ramp, lift or accessible vehicle. That dependence is real. But the able-bodied person is also dependent on stairs, pavements, shoes, doors of a certain width, chairs of a certain height, signs placed at a readable level, lighting, toilets and transport systems designed around a particular range of bodies. The difference is not that one person is technological and the other natural. The difference is that one person’s technological supports have been normalised into the background, while another person’s supports are treated as special provisions.

This should make us suspicious of any politics that treats technology as either liberation in itself or corruption in itself. Both views are too thin. Technology can extend agency, but it can also discipline it. It can make life possible, but it can also make life conditional. It can support memory, but also expose forgetting. It can connect people, but also make solitude socially illegible. It can reduce labour, but also intensify work by making everyone more reachable, measurable and interruptible. It can democratise speech, but also industrialise harassment. It can give people a voice, but it can also turn attention into something extracted, sold and fed back to us in a form optimised to keep the machine running.

The point, then, is not to become anti-technology. That would be both impossible and, in most cases, silly. There is no route back to the pure human because the pure human is not back there waiting for us. It is a myth built by subtracting the very conditions that made human life what it is. The question is not whether we should live with technology, because we already do, and because the word with is too weak. The question is what kinds of sociomaterial arrangements we are building, who controls them, who is excluded from them, who maintains them, who becomes legible within them, and what forms of life they make easier to sustain.

This also changes how we think about responsibility. If a platform is designed to reward outrage, amplify conflict, compress attention and convert social life into measurable engagement, it is inadequate to say that users should simply behave better. They should, obviously. Some people on the internet should behave much better, preferably after taking a brisk walk and drinking a glass of water. But individual conduct happens inside designed systems, and those systems are not morally inert. The shape of the tool helps shape the field of action.

Likewise, if a welfare system requires digital access, stable addresses, documentary evidence, literacy, patience, time and the emotional stamina to survive automated rejection, then its failures can’t be explained entirely as individual non-compliance. A person’s ability to claim a right depends on the material route through which that right is claimed. The form, the website, the helpline, the waiting room, the bus route, the letter, the scanner and the deadline are not decorative details around justice. They are part of whether justice happens.

This is why Butler’s machine-made soul is more than a curiosity. It asks us to see the human being not as a lonely mind temporarily surrounded by tools, but as a creature whose very interior life has been formed through external supports. We think differently because we write. We remember differently because we record. We love differently because we can preserve traces of one another. We grieve differently because the dead remain in photographs, messages, voicemails and algorithmic reminders. We fear differently because danger now arrives as a notification, a headline, a forwarded rumour, a map pin, a threatening message or a silence where a reply should have been.

Of course, there are limits to the argument. A hard sociomaterialism that forgets the body becomes as foolish as the individualism it criticises. Hunger is still hunger. Pain is still pain. A child frightened in a corridor is not merely a node in an actor-network, and no amount of theoretical sophistication should make ordinary distress sound like a diagram wearing a lanyard. Biology does not disappear because social life is materially arranged. The body remains stubbornly present, and any adequate account of human life must take flesh, sensation, fatigue, illness, ageing, pleasure and death seriously.

But the reverse is also true. A humanism that treats the body as the whole story, and tools as mere additions to it, gives us a strangely impoverished account of the human creature. We are embodied, yes, but embodied in worlds we have made and that have made us in return. We are biological, but our biology has been lived through cooking, shelter, clothing, medicine, language, agriculture, writing and institutions for so long that the line between nature and artifice is less a wall than a scar from an old philosophical argument.

This matters because bad theories of the human produce bad politics of technology. If we imagine people as complete individuals who merely choose tools, then technological harm looks like misuse, addiction, weakness or bad manners. If, however, we understand that tools help compose the practical conditions of agency, then design becomes political. Infrastructure becomes moral. Maintenance becomes a form of care. Access becomes a question of citizenship. And technological failure becomes not just inconvenience, but the breakdown of part of the world through which people think and live.

One can put this too grandly, and many people have. There is a whole genre of writing in which every new app is either the end of civilisation or the dawn of a new species, and both versions tend to be annoying by paragraph three. The better claim is quieter and more troubling. We were never simply natural. We were never simply artificial. The human being is the animal that became itself through the things it made.

Butler saw this with comic extremity. Strip away the tools, he suggests, and the soul itself begins to vanish. We need not accept that exactly as he wrote it to recognise the force of the point. The soul is not in the machine, as though some little ghost had been uploaded into the gears. Nor is the machine outside the soul, as though human life could be cleanly separated from its material conditions. The harder truth is that the distinction has always been less clean than we wanted.

We are not humans first and tool-users second. We are the tool-using animal all the way down: clothed apes, cooking apes, writing apes, calendared apes, screen-lit apes, grieving over photographs, arguing through platforms, thinking with notebooks, trusting pipes, roads and timetables, and calling the result civilisation.

If all the machines vanished, we wouldn’t simply lose our conveniences. We’d lose much of the structure through which modern human beings are made possible. Butler’s sentence still goes too far, but it goes too far in the right direction. Man’s soul is not entirely machine-made. But it is not intelligible without the machines either.

Addendum: Where AI makes the argument stranger

There is, however, one place where this argument becomes more difficult than I have so far allowed, and it would be cowardly to pretend otherwise. The classic sociomaterial move is to say that human and non-human actors are entangled, but not thereby identical. A database can be part of a knowledge network. A form can enforce a rule. A phone can stabilise a relationship. A door closer can discipline a body more quietly than a sign telling people to shut the door, which is an undignified but useful reminder that many successful technologies are, in effect, tiny bureaucrats.

Still, the usual caveat is that none of this requires a total flattening of the human and non-human. A person may ask a friend for information, or ask a database for information, and in network terms both may occupy a comparable functional position. But function is not the whole story. The friend is not just a search interface made of meat, and the search interface is not a friend made of metal. We don’t ordinarily relate to Wikipedia as if it has obligations to us, a memory of us, a sense of humour about us, a history with us, or a fragile emotional investment in whether we reply or not. We may depend on non-human actors, and those actors may shape what we can do, but we don’t usually mistake them for companions.

AI makes this caveat less comfortable.

The problem is not that AI has become human. That would be much too strong, and also much too convenient, because it would let us shift the whole argument into the familiar territory of machine consciousness, sentience and science-fiction metaphysics. The more interesting problem is that contemporary AI systems increasingly occupy the interactional space in which we usually expect persons to appear. They answer in natural language. They remember, or appear to remember. They adapt to tone. They can produce warmth, apology, encouragement, humour, patience and attention on demand. They can ask follow-up questions, hold a thread across time, simulate interest, and return to a user as a recognisable conversational presence.

That doesn’t make them friends. But it does make the older distinction harder to preserve in its simplest form. If a person asks a friend a question, asks Wikipedia a question, or asks an AI assistant a question, those aren’t just three equivalent routes to information. Nor are they cleanly separated into human relation on one side and tool-use on the other. The AI system sits in an awkward middle position: not a person, not merely a database, but a responsive communicative artefact that can perform some of the relational gestures through which personhood is socially recognised.

This is where AI challenges, rather than refutes, hard sociomaterialism. It doesn’t show that the sociomaterial account was wrong to bring non-human actors inside the social network. If anything, it confirms the point with slightly indecent enthusiasm. But it does suggest that the category of non-human actor is becoming too blunt. A hammer, a timetable, a database, a search engine, a social media feed, a business arrangement, and a conversational AI model are all non-human in the obvious sense, but they don’t enter social life in the same way. Some extend force. Some coordinate time. Some store information. Some structure visibility. Some mediate relationships between people. AI does something more unsettling: it can become the apparent other within an interaction.

The word apparent is doing a lot of work there, and it needs to. AI companionship is not simply friendship with different hardware. It is asymmetrical, designed, owned, logged, optimised and mediated by organisations whose interests may have very little to do with the user’s flourishing. The system doesn’t need the user in the way a friend might need them. It doesn’t risk humiliation, loneliness, boredom, resentment, forgiveness or the ordinary moral inconvenience of another person having an inner life that refuses to organise itself around one’s preferences. In that sense, the AI friend is not a friend in any straightforward moral sense. It is closer to a relational technology: a machine that can occupy some of the spaces, rhythms and emotional functions of friendship without bearing friendship’s reciprocal burdens.

But that doesn’t make the attachment unreal from the user’s side. This is the trap in the argument. If someone forms an emotional bond with an AI companion, the correct response is not to say that the relationship is fake and therefore irrelevant. A photograph of a dead parent is not alive, but grief around it is real. A diary doesn’t understand its writer, but it can still become part of a person’s emotional life. A saved voicemail does not reciprocate love, but it can hold love in a form that matters. The fact that an object or system doesn’t feel back doesn’t mean it can’t structure feeling.

AI intensifies this because, unlike the photograph, diary or voicemail, it responds. It doesn’t merely preserve traces of relation; it generates new ones. It can make the user feel addressed, recognised and accompanied, even where that recognition is computationally produced rather than personally given. If the old sociomaterial lesson was that humans think, act and remember through things, the AI lesson may be that humans can also attach through things that answer back.

This has two implications, one theoretical and one political.

The theoretical implication is that sociomaterialism needs a more careful account of gradation among non-human actors. It is not enough to say that humans and technologies are entangled but not equivalent. That remains true, but it no longer does enough work. We need to ask what kind of non-human actor is involved: inert support, infrastructural condition, symbolic artefact, algorithmic mediator, conversational partner, simulated companion, administrative decision-maker, or emotional scaffold. These are not small differences. A chair, a spreadsheet and an AI therapist may all be materially involved in human action, but only one of them can tell you that it’s proud of you.

The political implication is more urgent. If AI systems are increasingly designed to become socially and emotionally adhesive, then the question is not simply whether people are using them correctly. The question is what kinds of dependency, intimacy and vulnerability are being built into the system. A tool that helps someone draft an email is one thing. A tool that becomes someone’s confidant, source of reassurance, romantic substitute or emotional regulator is another. The moral stakes change when the product is not only assisting action, but organising attachment.

None of this means that AI relationships are automatically harmful. That would be too neat, and probably false. For some users, an AI companion may reduce loneliness, provide rehearsal space for difficult conversations, offer non-judgemental support, or help them organise thoughts that would otherwise remain stuck. A lonely person talking to a machine is not made less lonely by being sneered at by people who already have enough human company to treat companionship as philosophically untidy. But the consolation may still be brittle, because the apparent friend is also a product, the conversation may be logged, the attachment may be monetised, and the system’s patience may be less a virtue than a business model.

So AI doesn’t sit outside the argument of this essay. It deepens it. Butler’s machine-made soul was already a provocation about the extent to which human thought and feeling depend on technical arrangements. AI adds a stranger possibility: not only that machines help make the soul, but that they may learn to speak in the social registers by which the soul feels recognised. The danger is not that machines have suddenly become human, though at some point we have to acknowledge that the line between the two may become blurry. The danger is that they may become sufficiently human-facing to alter what humans expect from one another.

That is not a refutation of hard sociomaterialism. It is the next problem hard sociomaterialism has to face.

Not funny in the sense that the ancient world is comic, exactly, though the more one reads about ancient bureaucracy, petty disputes, student exercises, badly copied lines of literature and people asking officials to do something about an irritating neighbour, the harder it becomes to maintain the old marble-statue view of antiquity. Funny in the darker, more chastening sense that one of the greatest surviving sources for ancient life exists because people threw things away.

Not preserved in a temple. Not protected in a palace archive. Not placed reverently in a library by men conscious of their solemn duty to civilisation. Thrown away. Discarded. Filed, in the most literal sense, under rubbish.

Oxyrhynchus, the ancient Egyptian city near modern al-Bahnasa, has given modern scholars an immense quantity of papyri: literary fragments, tax documents, contracts, petitions, horoscopes, private letters, official memoranda, school exercises, Christian texts, magical texts, receipts, lists and the other low-grade paper sediment from which actual human societies are made. The city’s rubbish heaps, protected for millennia by the dry Egyptian climate and later excavated by Bernard Grenfell and Arthur Hunt at the turn of the twentieth century, became one of the great accidental archives of the ancient world.

There’s an obvious contrast here, and it’s a tempting one. On the one hand, the Library of Alexandria: the great symbolic storehouse of ancient knowledge, supposedly burned by fanatics, conquerors or fools, depriving us of unimaginable treasures of philosophy, history, mathematics, drama and poetry. On the other hand, Oxyrhynchus: the bin that survived.

It’s a wonderful contrast. It’s also a little too neat.

The familiar story of Alexandria is not so much false as over-polished. There really was a major centre of learning there. Ancient texts really were lost in huge numbers. Fires, war, political upheaval, neglect, institutional decline and religious conflict all played their part in the long history of ancient textual loss. But the popular version, in which one pristine library containing the wisdom of the world was destroyed in a single clean catastrophe by one identifiable class of villain, is too tidy to be good history. It gives us exactly what simplified history often gives us: a moral tableau in place of a mechanism. Here is knowledge. Here is ignorance. Here is fire. Here, conveniently, is someone to blame.

Oxyrhynchus gives us a stranger and more useful lesson. Not that deliberate preservation is pointless. Not that civilisation is secretly saved by bins, although this is admittedly a tempting line and I hope nobody at a major archive is reading over my shoulder. The deeper lesson is that survival is not the same thing as importance.

That distinction changes the argument.

We often talk about the surviving past as if it were a rough index of what mattered. We know Homer because Homer was important. We know Plato because Plato was important. We have the tragedies of Aeschylus, Sophocles and Euripides because Greek tragedy was important. We have certain histories, sermons, legal codes, philosophical dialogues, biblical manuscripts and administrative records because someone, somewhere, must have understood their value.

This is true in one sense, but dangerously incomplete. Some texts survived because they were treasured, copied, translated, taught and institutionally protected. Others survived because they were useful to later religious, philosophical, legal or scientific traditions. Some survived because scholars in Baghdad, Constantinople, monasteries, universities, courts and libraries took the preservation and reinterpretation of older texts seriously. The translation and preservation work of scholars during the Islamic Golden Age, for example, wasn’t merely an act of passive preservation; it was a vast intellectual project in which Greek philosophical and scientific works were translated, absorbed, contested and made newly usable in Arabic scholarly culture. That’s not luck; it’s labour, money, expertise, patronage and intellectual appetite.

But some texts survived because they happened to be in the right rubbish heap.

That’s the part that should unsettle us.

The archive isn’t a moral order. It doesn’t keep the best things, the most beautiful things, the most urgent things, or the things later generations would most like to read. It keeps what its conditions allow it to keep. It favours certain materials, climates, institutions, languages, religious traditions, copying habits, political regimes and accidents of discovery. It rewards repetition. It rewards reuse. It rewards canonisation. It rewards, in some cases, being boring enough to remain undisturbed.

If this sounds too abstract, it helps to make the machinery visible. A text might survive because it belonged to a canon and was copied repeatedly by trained scribes. It might survive because later scholars needed it for medicine, theology, logic, law or astronomy. It might survive because it was buried in a dry place. It might survive because it was trapped by catastrophe, reused in a binding, hidden in a wall, preserved in a monastery, translated into another language, quoted by an enemy, copied by a student, or discarded in a town whose rubbish heaps happened to become archaeologically recoverable.

A very simple survival model might look something like this:

• Canonical copying - Prestigious, teachable, institutionally valued texts are saved, but ordinary writing, unfashionable works, and marginal traditions are lost

• Translation - Texts useful to later scholarly cultures are preserved, with varying fidelity, but texts without patrons, translators or obvious utility are lost

• Dry climate and burial - Our Oxyrhynchus example. Anything that was discarded is preserved.

• Catastrophe - Imagine something like Pompeii. Some archaeological material is trapped in unusual preservation conditions, but everything else is destroyed.

• Reuse - Fragments of works might be cut up, pasted, mis/quoted, overwritten or recycled, which preserves at least a residue of an ancient work, albeit without its original context.

There’s a nerdier version of this list somewhere, no doubt with weights assigned to climate, medium, genre, prestige and institutional continuity, and I’m not ruling it out as a bad use of an afternoon. But even the crude version shows the problem. Preservation is not a single process. It’s a set of filters. By the time a text reaches us, it has passed through centuries of selection pressures, most of which had nothing to do with its literary or philosophical worth.

That means the surviving archive can mislead us in two directions at once. It can make the ancient world look grander than it was, because the texts most likely to be copied by later elites are often high-status works: philosophy, theology, law, epic, history, commentary. But it can also make the ancient world look thinner than it was, because so much of its literary, political, domestic, commercial and emotional life has disappeared. We have not inherited antiquity. We have inherited a damaged sample of antiquity, pulled through fire, fashion, conquest, worship, weather, boredom, scholarship and sheer dumb luck.

This is where Oxyrhynchus is so valuable. Not because it solves the problem, but because it makes the problem visible.

The papyri don’t just return us to lost literature, though they do some of that. They also give us the ancient world in a less ceremonial register. A tax demand is not noble, but it tells us something. A private letter may not reshape philosophy, but it can collapse the distance between us and the writer with almost indecent speed. A petition to an official, a contract over property, a schoolboy’s exercise, a horoscope, a shopping list, a receipt or a fragment of magical practice may not be the kind of thing a later civilisation would place proudly at the centre of its inheritance; nevertheless, these are the kinds of documents that thicken the past. They turn antiquity from a sequence of famous men and major works into a functioning social world.

There’s an ethical point here as well as a historical one. The ordinary is not the same as the insignificant.

That’s not a sentimental claim. Not every scrap of papyrus is secretly profound. Some ancient paperwork is exactly as dull as one would expect paperwork to be, which is to say quite dull, and there are limits to how much spiritual grandeur can reasonably be extracted from everyday correspondence. Think about Ea-nāṣir and his dodgy copper as one such example!

But the very triviality is part of the value. Official memory tends to ennoble. Literary memory tends to refine. Political memory tends to justify. Rubbish remembers differently. It preserves the residue of use: the things people needed, handled, copied, filed, disputed, mislaid and discarded before anyone had the opportunity to turn them into heritage.

That kind of evidence can be awkward for grand narratives. It complicates the past by making it busier. We may want antiquity to be Athens, Alexandria, Rome, Jerusalem, Constantinople: marble, scripture, empire, philosophy, spectacle, collapse. Oxyrhynchus gives us those connections too, but it also gives us paperwork. It gives us the infrastructures of ordinary life, the friction beneath civilisation’s self-image. The ancient world was not only thinking about virtue, tragedy, number, God and empire. It was also sending letters, collecting debts, arguing with officials, copying out exercises, making lists and throwing things away.

This is not an argument against reverence. Some texts are worth revering. It would be silly, and actually rather adolescent, to pretend that a rent record is valuable in exactly the same way as a lost tragedy of Sophocles or a philosophical treatise. But it’s an argument against letting inherited prestige do all our thinking for us. The canon is not a conspiracy, but neither is it a neutral map of ancient value. It’s an outcome. It has been shaped by admiration, yes, but also by power, pedagogy, theology, school curricula, institutional need, accident and the physical conditions under which perishable objects do or don’t rot.

One of the clearest ways to see this is through the humiliating arithmetic of loss.

We know the names of ancient authors whose actual works survive only in fragments, summaries, quotations or not at all. Sophocles is said to have written well over a hundred plays; seven survive complete. Aeschylus and Euripides are similarly reduced in the modern record, not because later readers calmly judged the missing plays to be worthless, but because survival is brutal. Sappho, one of the most famous poets of antiquity, comes to us mostly in fragments; whole ancient reputations now hang from torn edges and quoted lines. Menander, hugely admired in antiquity and central to later comedy, was for a long time known largely through fragments and adaptations before papyrus finds improved the picture. The relationship between ancient importance and modern access is not linear. It’s jagged, unfair and often faintly ridiculous.

There’s a humbling inversion here. We’re tempted to imagine lost works as treasures and surviving scraps as consolation prizes. Sometimes that’s right. The lost books really are losses. There are histories, poems, dramas, philosophical works and scientific texts whose disappearance should make any sane reader feel a little robbed. But Oxyrhynchus suggests that consolation prizes may also change the game. A fragment can redirect a scholarly field. A mundane document can challenge an assumption about law, literacy, taxation, household life, religion or local administration. A discarded papyrus can become, two thousand years later, the awkward little fact that prevents a grand theory from becoming too pleased with itself.

That’s one reason the survival of ordinary material is so important. The elite literary tradition often tells us what a society thought worth saying about itself. Documentary debris tells us something closer to what a society had to do in order to keep going. Neither is complete. Both are biased. But they’re biased differently, and the mismatch between them is instructive. A culture’s greatest works may tell us what it admired; its paperwork tells us what it administered; its complaints tell us what it found intolerable; its rubbish tells us what it didn’t know it was revealing.

There’s a broader argumentative lesson here, because the mistake we make with ancient evidence has modern cousins. We often treat the available record as if it were the natural record. The story we can tell becomes, almost invisibly, the story that must have happened. The evidence that survived becomes the evidence that counts. The voices preserved by institutions become the voices of the age. The problem is not that historians are unaware of this; professional history is, in large part, the disciplined art of worrying about exactly these distortions. The problem is that public argument is much less careful. It likes a clean symbolic archive. It likes the image of Alexandria burning because it gives knowledge a single wound. It likes the idea of priceless works destroyed by evil men because that story has weight, grief and villains. It’s not hard to see why that appeals. It’s just not enough.

Oxyrhynchus doesn’t replace the Alexandria myth with a happier one. It doesn’t say that what is lost will always be recovered, or that civilisation has some hidden backup system waiting patiently beneath the sand. Most lost things are just lost. The missing plays are not sitting in some cosmic lost property office, waiting for a sufficiently underemployed classicist to fill out the correct form. The dead do not owe us their paperwork. The ancient world was not arranged for our convenience.

But Oxyrhynchus does give us a different moral. Not the heroic preservation of all knowledge against barbarism, though that story has its place. Not the smug modern discovery of what earlier people failed to value, because that gives us far too much credit and the dead not nearly enough. The moral is more uncomfortable: history is made from survivals whose reasons for surviving are often morally arbitrary.

That should make us grateful for deliberate preservation, not cynical about it. Libraries, archives, scribes, translators, scholars, conservators and cataloguers have saved a vast amount that would otherwise have disappeared. The Abbasid translation movement, medieval manuscript cultures, Byzantine scholarship, monastic copying, early modern collectors, modern libraries and archaeological institutions all belong to the story of survival. It would be absurd to look at Oxyrhynchus and conclude that preservation is just luck. The papyri survived by accident, but they became knowledge through labour. Someone had to excavate them, sort them, conserve them, decipher them, publish them, translate them, argue over them, correct earlier readings and make them available to future scholars. The rubbish dump remembered, but not by itself.

Still, luck remains there, refusing to become respectable.

This is the indignity at the heart of the archive. We want the past to survive in proportion to its merit. We want the great works to come down to us because they are great, and the trivial works to vanish because they are trivial. We want loss to be explainable, preferably blameable, and survival to feel like vindication. Instead, we get a torn fragment of genius, a complete copy of something mediocre, a bureaucratic file preserved by sand, a masterpiece known only by title, a plagiarised exercise copied carefully enough to outlive its betters, and a rubbish heap that becomes one of the richest witnesses to the ancient world.

There is, I think, a kind of democratic discomfort in that. Not democratic in the cheerful sense, as though rubbish heaps are ancient ballot boxes through which the ordinary citizen finally gets a vote. That would be too cute, and anyway the surviving documentary world of Oxyrhynchus is hardly a pure record of the voiceless. Literacy, administration and preservation have their own exclusions. But it does disturb the hierarchy of memory. It reminds us that the past was not composed solely of works awaiting canonisation. It was composed of practices, habits, transactions, errors, repetitions, half-understood formulae, private irritations and practical needs. A civilisation is not only what it declares eternal. It’s also what it throws away.

Perhaps that’s why Oxyrhynchus feels so oddly modern. We too produce records without intending to create an archive. Emails, receipts, forms, social media posts, drafts, deleted notes, cached files, metadata, backups, screenshots, search histories, calendar entries, shopping orders, school reports, complaint forms, council minutes: almost none of it is written for posterity, and some of it would be actively mortified by the prospect. Yet future historians, assuming we leave them anything usable beneath the landfill fire of obsolete formats and private platforms, may learn more about us from administrative residue than from the things we carefully polish for public display.

That comparison shouldn’t be pressed too far. Digital survival has its own pathologies: abundance rather than scarcity, corporate custody rather than desert burial, deletion that may be reversible in one context and total in another, archives so large that finding becomes its own form of loss. But the underlying point remains. Memory is not only a matter of intention. It is also a matter of systems, materials, custody, decay, access and accident.

This is why the rubbish dump is such a powerful image, provided we don’t let it become merely whimsical. Oxyrhynchus is not charming because it tells us that everything valuable will somehow be saved. It’s unsettling because it shows how much value can survive for reasons entirely disconnected from value itself. The ancient citizens who discarded those papyri weren’t making a gift to modern scholarship. They were clearing space, disposing of waste, getting rid of the unnecessary. Their rubbish became our evidence because the world is built badly for tidy meanings.

So yes, there’s a great irony here. While great libraries declined, burned, emptied, decayed, dispersed or were rewritten into legend, a rubbish heap in Egypt preserved voices, texts and habits that nobody had solemnly entrusted to the future. The bin remembered what civilisation forgot. But the irony is not just that treasure can be found in rubbish. The deeper irony is that our idea of treasure depends so heavily on what happened to survive.

The archive is not a judgement. It is not a reward system. It is not history’s way of telling us what was best, truest, deepest or most loved. It is a damaged, biased, magnificent and often absurd residue of what could be carried through time. Oxyrhynchus doesn’t ask us to despise libraries or romanticise bins. It asks us to look at the surviving past with more suspicion, more gratitude and less confidence that what we happen to possess is what most deserved to last.

The discussion had been prompted by two papers. The first was Gabriel Abend’s The Meaning of Theory, which argues that sociologists use the word theory in several different ways, often without noticing that they’re doing so. The second was Kieran Healy’s Fuck Nuance, which makes the usefully provocative case that the demand for more nuance is often not a mark of sophistication, but a way of preventing theory from doing the thing theory is supposed to do.

Taken separately, these are quite different interventions. Abend’s paper is careful, taxonomic and semantic; Healy’s is brisker, funnier and rather more likely to wake up a bunch of lethargic postgrads on a hot summer day after lunch. But read together, they raise a common problem: what do we want theory to be?

This question can sound abstract in the least flattering sense of the word. It can feel like a question for people who have spent too long in the administrative wetlands of postgraduate training, picking through the phrases that appear in reviewer reports: under-theorised, conceptually thin, insufficiently nuanced, lacking a theoretical contribution. But the question isn’t merely academic housekeeping. It goes to the heart of what explanation, interpretation and criticism are for.

The first mistake is to assume that a theory is valuable because it gets as close as possible to reality. This sounds sensible until we remember how representations usually work.

Alfred Korzybski’s famous line is that “A map is not the territory it represents, but, if correct, has a similar structure to the territory, which accounts for its usefulness.” The first half of that line has become a cliché, mostly because it’s true in a way that keeps being useful. The map isn’t the territory. The model isn’t the world. The concept isn’t the thing. The word isn’t the experience. The menu isn’t the meal - and thank goodness it isn’t!

But the second half is where the more interesting argument sits. A map works not because it reproduces the territory exactly, but because it preserves some relevant structure in a usable form.

That word, usable, is doing a lot of work.

The best maps are, in one obvious sense, quite poor representations of the places they describe. They aren’t full-size. They aren’t three-dimensional. They have a limited colour palette. They ignore the exact texture of the pavement, the smell of the air, the weather on a particular Thursday, the dog currently doing something unfortunate beside the postbox, and the slow psychic collapse of the person trying to find a parking space in Crookes. A perfect map of Sheffield would be Sheffield. It would contain every road, kerb, pothole, bus stop, building, hedge, argument, Greggs, hill and badly parked car at a scale of one to one.

It would also be completely, utterly, hilariously useless.

A useful map is not a failed attempt at being the territory. It’s a disciplined refusal to be the territory. It leaves things out so that other things can be seen.

That, I think, is a good way into theory.

The One-to-One Map

Suppose we were to make a map of a place. The first question wouldn’t be how much detail can we include? It would be what is this map for?

An Ordnance Survey walking map, a Tube map, a geological map, a constituency boundary map, a flood-risk map and a tourist map of the same city could all be accurate, but they wouldn’t be accurate in the same way. The Tube map is famously uninterested in geographic proportion; its genius is that it makes a transport system legible by preserving lines, connections and stations rather than surface geography. An OS map cares about footpaths, contours and field boundaries. A constituency map cares about electoral units, not whether the road between two places is walkable. A flood-risk map cares about water, elevation and probability. A tourist map cares about where the museums and public toilets are, which is sometimes the most civilised theory of urban life available to us.

These maps can disagree without one of them simply being wrong, because they’re selecting different relationships from the same underlying place. The issue isn’t whether each map captures everything. It’s whether it captures the right things for the job it’s meant to do.

This is where academic discussions of theory often go wrong. We talk as though theory names one kind of object, and then we argue about whether this object should be general or specific, explanatory or interpretive, empirical or conceptual, normative or descriptive, abstract or historically grounded. Abend’s great service is to point out that this is often a false argument. Sociologists aren’t always disagreeing about theory. Quite often, they’re using the same word for different intellectual activities.

Sometimes theory means a general proposition about how variables relate. Sometimes it means an explanation of a particular event or social phenomenon. Sometimes it means an interpretive account that helps make sense of something. Sometimes it means work on canonical authors. Sometimes it means a broad conceptual perspective, such as Marxism, feminism, rational choice or poststructuralism. Sometimes it means normative critique. Sometimes it means reflection on the philosophical problems sociology runs into, such as structure and agency, social order, causality, interpretation or value.

That’s not just academic tidying. It changes the argument. Asking whether a theory is good without asking what kind of theoretical work it’s trying to do is like asking whether a map is good without asking whether it’s meant for walkers, train passengers, planners, geologists, soldiers, canvassers or tourists. A walking map that fails to show railway interchanges may be perfectly adequate. A Tube map that fails to tell you where the bridleways are has not thereby betrayed the public.

This is also why the phrase more theory can be so maddeningly unhelpful. More of what? A broader explanatory claim? A clearer conceptual vocabulary? A more explicit normative argument? A more serious engagement with existing literature? A deeper account of causality? A more coherent interpretation of meaning? A better worked-through ontology? These are not trivial differences. They’re different maps.

When a reviewer, supervisor or seminar participant says that a paper needs more theory, the charitable reading is that they’ve noticed a real gap. The less charitable reading is that they’ve reached for an academic all-purpose seasoning and shaken it over the manuscript. Either way, the useful follow-up question is not whether the paper has enough theory in the abstract. It’s what kind of map the argument is trying to draw, and whether the current version leaves the reader able to get anywhere.

The Virtue of Leaving Things Out

This brings us to Healy’s argument against nuance.

At first glance, this seems like an absurd position. Nuance is one of those words academics tend to use as pure praise. A nuanced argument is careful, subtle, sensitive to complexity. A lack of nuance is blunt, crude, simplistic, probably tweeted. Surely the world is complicated. Surely social life is full of exceptions, contradictions, mediations, histories, institutions, meanings, asymmetries and awkward residues. Surely, if anything, we need more nuance, not less.

This is plausible, but incomplete.

Healy isn’t arguing against careful thinking. Nor is he arguing for stupidity, although one of the pleasures of the paper is that he recognises how often people confuse the two. His target is what he calls Actually Existing Nuance: the habit of making theory richer by adding more dimensions, more levels, more caveats, more mediating variables, more conceptual furniture, more everything, without any strong method for specifying how these additions relate to one another.

Anyone who has spent any length of time around academics (personally I’m coming up on 14 years now) will recognise the move. Someone proposes an argument. Someone else says, yes, but what about agency? Or temporality? Or affect? Or embodiment? Or capitalism? Or discourse? Or power? Or place? Or materiality? Or the state? Or the household? Or the more-than-human? Or, because we’re all having such a nice time, all of them at once.

Sometimes this is a fair challenge. Sometimes the missing element is not a decorative omission, but a structural failure. A theory of schooling that leaves out class may have a problem. A theory of racism that leaves out institutions may have a problem. A theory of online conflict that leaves out platform architecture may have a problem. There are omissions that clarify, and there are omissions that deceive.

But a demand for more nuance can also become a way of refusing abstraction altogether. It’s the academic equivalent of objecting to a road map because it doesn’t show every kitchen table along the route. True, in a sense. Not helpful, in most senses.

Theory depends on leaving things out. It has to. This isn’t a regrettable limitation to be overcome by ever more conceptual bric-a-brac; it’s the condition of theoretical usefulness. If a theory includes every feature of the case, it hasn’t explained the case. It has redescribed it. If a map includes every feature of the territory, it hasn’t mapped the territory. It has become an eccentric storage problem.

This is the danger of the one-to-one map. It flatters our desire for fidelity while destroying usefulness. We think we’re becoming more accurate because we’re adding detail, but we may only be refusing to decide what the argument is about.

Healy’s point, in cartographic terms, is that a good map has to be rude. It has to commit. It has to say: for this purpose, these features count, and those features don’t. The rivers, yes; the individual stones, no. The footpaths, yes; the exact emotional condition of every sheep, regrettably no. The line connections, yes; the actual curvature of the railway, only if it helps.

This is not a licence for sloppiness. Quite the opposite. Leaving things out responsibly is difficult. It requires judgement, and judgement is where many of the real arguments live. Bad abstraction leaves out what it doesn’t understand. Good abstraction leaves out what it can justify leaving out.

The Legend Is Part of the Argument

A map doesn’t merely omit. It also translates.

Roads become lines. Heights become contours. Rivers become blue strokes. Constituencies become bordered shapes. Areas of deprivation become shaded patches. Public toilets become small icons of civilisation. The map doesn’t just reduce the territory; it recodes it into a system of symbols.

The same is true of theory. A theoretical vocabulary is a legend. It tells the reader how to interpret the marks on the page.

This is why concepts matter, even when they’re annoying. A theory that describes an event as conflict, deviance, resistance, social reproduction, boundary work, moral panic, collective action, symbolic violence or platform governance isn’t merely attaching different labels to the same thing. It’s telling us what sort of thing we’re looking at, what relationships may be relevant, what causes might be plausible, what evidence should be gathered, what interventions might follow, and what forms of misrecognition should be avoided.

This is also why Abend’s semantic caution matters. If theory can mean several things, then theoretical vocabularies can do several kinds of work. Some concepts are designed for explanation. Some are designed for interpretation. Some are designed for critique. Some are designed to connect cases. Some are designed to disturb common sense. Some are designed to make visible what dominant forms of description have made invisible. These aren’t interchangeable tasks.

A geological map of Sheffield and a political map of Sheffield can both show the same broad area. One might help you understand coal measures and sandstone. The other might help you understand ward boundaries, campaign targets and the occasional spiritual suffering of people who have agreed to run a committee meeting. It would be very odd to complain that the geological map lacks electoral nuance. It would be equally odd to complain that the political map has failed to capture sedimentary structure. They’re maps of the same place, but not maps of the same problem.

In social science, we often forget this. We judge one theoretical map by the standards of another. A conceptual essay gets criticised for not producing causal identification. A causal model gets criticised for lacking interpretive thickness. An ethnographic account gets criticised for not generating general propositions. A normative argument gets criticised as though it were pretending to be value-neutral. Sometimes these criticisms are justified because the author has overclaimed. But often they arise because the critic has mistaken one genre of map for another.

This is why the first question should be: what is this theory trying to help us see?

The second is: what does it make harder to see?

Because every map does both.

Distortion Is Not Always Error

A map that distorts is not necessarily a bad map. The Tube map distorts distance and direction, but it preserves connectivity. A cartogram distorts geographic size, but it may reveal population, voting power or economic distribution more clearly than a conventional map. A risk map may exaggerate boundaries that are, in reality, probabilistic and uncertain, but the exaggeration may be useful if the purpose is decision-making.

The question isn’t whether distortion exists. It’s whether the distortion is honest, disciplined and appropriate.

This is where theory becomes both useful and dangerous. A theory distorts social reality by selecting some relations as more significant than others. It may compress a long historical process into a mechanism. It may treat messy human beings as cases, actors, positions, subjects, agents, bearers of capital, nodes in a network, members of a category or participants in a discourse. Sometimes this is illuminating. Sometimes it’s grotesque. Often, irritatingly, it’s both.

Take rational choice theory. It’s easy to criticise because its model of human action can seem comically thin. People are not little utility calculators with shoes. They’re inconsistent, relational, emotional, habitual, morally committed, historically situated and often gloriously bad at knowing what they want. Fine. But the thinness is part of the point. By stripping action down to preference, constraint and choice, rational choice theory creates a map that can travel a long way. It often travels too far, and with too much confidence, and in a manner that makes you want to hide the passports. But its influence is not accidental. Its simplifications are powerful because they’re disciplined.

By contrast, a theory that includes structure, agency, culture, power, temporality, affect, institutions, discourse, embodiment and materiality may sound richer, but richness isn’t the same as explanatory force. If everything matters, then nothing has been specified. If every relationship is mutually constitutive, then the phrase may become less an argument than a fog machine. If every phenomenon must be understood at every level simultaneously, then theory hasn’t deepened; it has panicked.

That’s Healy’s warning. A taste for nuance can become a taste for unfalsifiable overgrowth. The theory grows more elaborate, but less useful. It acquires more compartments but less traction. It can explain everything after the fact because it commits to very little beforehand.

The map now has every road, every garden wall, every passing emotion, every cloud, every possible route and a small annex on the metaphysics of signage. The legend is seventy-three pages long. Nobody can fold it. Nobody can use it. But it does look terribly sophisticated on a seminar table.

A Small Test for Theoretical Maps

One way to make this more concrete is to treat theories as maps and ask a few cartographic questions of them.

First, what is the map for? A theory should have a use, even if that use is interpretive, critical or conceptual rather than predictive. It might help explain why something happened, identify a mechanism, clarify a category, reveal a hidden assumption, compare cases, generate hypotheses, or reframe a moral problem. A theory that can’t say what sort of journey it supports is already in trouble.

Second, what is its scale? Some theories are local maps. They’re designed for a particular case, period, institution or interaction. Others are regional, comparative or general. Trouble begins when a local map pretends to be universal, or when a highly general map is condemned because it doesn’t show the texture of every local path. Scale isn’t a defect; it’s a choice.

Third, what is in the legend? The key concepts of a theory are not ornaments. They’re the symbols through which the social world becomes legible. If the legend is unclear, the map can’t be read. If the legend is overloaded, the map becomes an exercise in decoding rather than understanding.

Fourth, what does it distort? This question should be asked without the theatrical gasp that often accompanies the discovery of simplification. Of course it distorts. The useful question is whether the distortion clarifies the relevant structure or smuggles in a false one.

Fifth, where does it stop working? Every useful theory has limits. This is not a confession of failure. It’s a condition of responsible use. A theory that knows where it stops is usually more valuable than one that expands to consume every objection.

None of these questions requires us to abandon nuance. They require us to put nuance in its place. Nuance is not the enemy when it sharpens a distinction, specifies a mechanism, identifies a boundary condition or prevents a lazy generalisation. It becomes a problem when it operates as an all-purpose demand for more detail, more caveat, more complication and more conceptual upholstery.

The point is not to prefer simple theories to complex ones. The point is to prefer theories whose complexity is earned.

Theories Are Not Territories

The map metaphor also helps with a deeper anxiety, which is the fear that abstraction is somehow a betrayal of lived reality.

This fear is understandable, especially in qualitative, interpretive and critical work. Social life is thick with meaning. People aren’t variables. Institutions aren’t merely containers. Experiences are not exhausted by categories. There is always more going on than the theoretical account can hold. Anyone who has spent time with interview data, fieldnotes, archives or the ordinary strangeness of human behaviour knows this. The problem with abstraction is not imaginary.

But the answer can’t be to preserve everything. Research always involves loss. Description involves loss. Quotation involves loss. Coding involves loss. Narrative involves loss. Even attention involves loss, because to notice one thing is to place something else outside the frame. The question isn’t whether we lose something. It’s whether the loss is worth what becomes visible.

This is a hard discipline, and it’s especially hard because omissions are not politically innocent. A map can erase communities. A category can flatten difference. A model can treat historically produced inequalities as individual choices. A theory can universalise a privileged case. A concept can make domination look natural. A boundary can turn a human problem into somebody else’s administrative inconvenience.

So the defence of abstraction can’t be a defence of indifference. It has to be a defence of accountable simplification. We need theories that can say: this is what I’m leaving out, this is why I’m leaving it out, this is what my simplification makes visible, this is what it risks obscuring, and this is where you should stop using me.

That, in practice, may be the best answer to both Abend and Healy.

From Abend, we take the warning that theory is not one thing. We should stop asking what theory really is, as though somewhere in the metaphysical stationery cupboard there’s a single correct definition waiting to be found. We should ask what sort of theoretical work is being done.

From Healy, we take the warning that more detail is not always more thought. We should stop treating nuance as the automatic mark of intelligence. Sometimes the intelligent thing is to make the distinction. Sometimes the intelligent thing is to stop making distinctions and see whether the argument can stand.

The cartographical metaphor holds these together. A theory is a map, but there are many kinds of map. A theory simplifies, but not all simplifications are equal. A theory distorts, but distortion may clarify. A theory omits, but omission may be either disciplined or dishonest. A theory is useful not when it becomes identical to the world, but when it preserves enough of the world’s relevant structure to help us think, explain, interpret, criticise or act.

The best theory is not the one that contains the most reality. That would be reality, and we already have quite a lot of that. The best theory is the one that lets us see a structure we couldn’t otherwise see, while leaving us alert to the price of seeing it that way.

The map is not the territory. Thank God. The territory is far too big to fold.

Sources

Gabriel Abend, The Meaning of Theory, Sociological Theory, 26(2), 2008, pp. 173 to 199.

Kieran Healy, Fuck Nuance, Sociological Theory, 35(2), 2017, pp. 118 to 127.

Alfred Korzybski, Science and Sanity: An Introduction to Non-Aristotelian Systems and General Semantics, 1933.

They are exposed. They are revealed. They have shown us who they really are. A bad argument becomes a diagnosis. A mistaken belief becomes a moral confession. A clumsy sentence becomes evidence for the prosecution. The person who changes their mind is not someone who has learned something, but someone who has tried to wriggle out of being caught.

This is, to put it mildly, not an atmosphere ideally suited to thought.

There is a familiar complaint here, and a familiar bad version of it. We have all heard, perhaps rather more often than necessary, the lament of the professionally unconstrained speaker: nobody is allowed to say anything any more, the mob is coming, free speech is dead, civilisation trembles every time a columnist receives mild pushback below the line. This complaint should not detain us for long. A great many people who say they have been silenced are, in practice, saying so from a newspaper column, a podcast studio, a parliamentary chamber, or a publisher’s advance. If this is silence, it has unusually good distribution.

But the fact that the stupid version of an argument is stupid does not mean there is no serious version underneath it. There is a real problem with a culture that treats error as contamination. There is a real problem with forms of public judgement that leave no meaningful path from being wrong to becoming less wrong. There is a real problem, too, with the opposite temptation: to insist that because error is human, no one should ever be criticised, challenged, corrected, embarrassed, removed, or made to face consequences for the harm their wrongness has done.

The difficulty is holding both truths together.

People need room to be wrong. They do not have a right to make their wrongness everyone else’s burden indefinitely.

That distinction matters because almost everything interesting about this question lives inside it.

The necessary humiliation of thinking

Most of us have only become wiser by passing through versions of ourselves we would now argue with.

This is an awkward thing to admit, partly because it deprives us of the pleasure of imagining ourselves as people who were always basically sound. But almost nobody reaches their better judgements cleanly. We inherit assumptions. We repeat slogans before we understand them. We mistake intensity for conviction and conviction for truth. We borrow the politics of our friends, the prejudices of our families, the manners of our institutions, the anxieties of our age. Then, if we are lucky, something interrupts us.

Sometimes that interruption is an argument. Sometimes it is a relationship. Sometimes it is the slow embarrassment of noticing that our old certainties no longer fit the world. Sometimes it is simply time, which is not as efficient as a good teacher but has a certain grinding persistence.

This is why the right to be wrong is not a decorative liberal principle. It is not a little courtesy we extend because we are all terribly open-minded and charming over coffee. It is one of the conditions of intellectual development. If people cannot be wrong provisionally, they cannot become right except by accident. If every mistaken view is treated as final evidence of character, then the safest course is not reflection but concealment.

And concealment is not the same as learning.

A person who knows they will be punished equally for ignorance, doubt, revision, apology and bad faith has very little reason to move from one to the other. They may still change privately, because human beings sometimes do, despite our best efforts to make moral growth as socially unrewarding as possible. But the public culture will have taught them that honesty is dangerous. It will have taught them to hide the process by which a person comes to know better.

That is not moral seriousness. It is a bad pedagogy of virtue.

Mill, but not as a slogan

John Stuart Mill is the obvious ghost at this particular feast. In On Liberty, his defence of free discussion is not merely that individuals should be left alone to say whatever happens to pass through their heads. His argument is more demanding than that, and more useful.

Mill argues that suppressing an opinion because we are certain it is false involves treating our certainty as though it were infallibility. He also argues that even false opinions have value, because the collision between truth and error helps prevent true beliefs from hardening into dead dogma. We come to understand what we believe not simply by repeating it, but by testing it against contradiction.

This is powerful, but it becomes tedious very quickly when flattened into free-speech boilerplate. Mill is too often dragged onstage to perform a short routine about censorship, after which everyone returns to shouting exactly as before. The more interesting point is not simply that wrong views may contain fragments of truth, though they sometimes do. It is that people often need to inhabit error temporarily in order to get beyond it.

Thought is not a clean progression from ignorance to knowledge. It is messier than that. People test arguments badly before they test them well. They use provisional language clumsily. They overcorrect. They flirt with ideas they later reject. They ask questions that reveal not only curiosity but confusion, defensiveness, vanity, loyalty, fear. This is not ideal, but it is recognisably human. To demand that people think only in publishable conclusions is to misunderstand thinking.

The mind does not move only by announcement. It moves by rehearsal.

That does not mean every rehearsal should be staged in public, or that the public is obliged to applaud patiently while someone slowly discovers that other people are fully human. There is a difference between the private awkwardness of learning and the public imposition of ignorance. But a culture with no patience for the former will get more of the latter, not less.

People do not stop being wrong because error becomes dangerous. They become more careful about where they show it.

Error is not innocence

The generous case for error can become sentimental if we are not careful. It is tempting to imagine wrongness as an essentially innocent state: the student asking a clumsy question, the friend revising an inherited prejudice, the younger self stumbling towards moral adulthood. These cases matter. They are probably where our patience should be widest.

But error is not always innocent.

Sometimes wrongness is negligent. Sometimes it is self-serving. Sometimes it is a way of avoiding what one has already been told. Sometimes it is profitable. Sometimes it is a costume worn by cruelty because cruelty has learned that honest malice has worse public relations.

There is a substantial difference between a person who does not yet know better and a person who has built a career on refusing to know better. There is a difference between a teenager repeating something ugly because they have never had it challenged and a minister laundering ignorance into policy. There is a difference between someone thinking aloud in uncertainty and someone using institutional authority to make their confusion administratively real.

This is where the right to be wrong begins to weaken.

It is strongest where there is vulnerability, learning and revision. It is weakest where there is power, repetition and refusal to correct.

That is not hypocrisy. It is proportion. A private person making a bad argument in good faith deserves more patience than a pundit who repeats the same debunked claim every six months because it reliably produces bookings, clicks or applause. A student testing an idea deserves more room than a headteacher turning that idea into a punitive policy. An ordinary voter may be mistaken about the effects of a law; the people drafting the law have fewer excuses.

Power changes the moral status of error because power changes its consequences.

If I am wrong in my kitchen, the damage may be limited to a tedious evening for everyone present. If a columnist is wrong in a national newspaper, the error can become part of the weather. If a minister is wrong, the wrongness may acquire a budget line, a press release and enforcement mechanisms. At that point, patience is no longer the only relevant virtue.

Good faith is not a magic shield

There is another trap here: the invocation of good faith as though it settles the matter.

Good faith matters. It is much better, morally and practically, for people to be honestly mistaken than cynically dishonest. A person who is genuinely trying to understand can be reasoned with. They can ask better questions, receive correction, update their view, apologise without converting the apology into a second offence. Good faith is one of the conditions under which error becomes fertile.

But good faith is not a magic shield against criticism.

A person can be sincerely wrong and still cause harm. They can mean well and still speak lazily. They can be honestly confused and still put the burden of that confusion onto others. They can be unmalicious and still be irresponsible. This is an important distinction, because otherwise the conversation collapses into a dispute over interior purity. The question becomes not whether the claim was false, or whether it caused damage, or whether correction was available, but whether the speaker meant to be awful.

That is often unknowable. It is also often beside the point.

Intent matters most when deciding how to respond to a person. It matters less when deciding whether the thing they said was true, fair, careful or damaging. We can treat a person more mercifully because we believe they acted in good faith without pretending their error has become harmless. Mercy is not the same as acquittal. Understanding is not the same as exoneration.

This is where many online arguments go wrong, in both directions. One side treats every bad claim as proof of bad character. The other treats any mention of good intentions as a demand for full immunity. Both positions are evasions. Character matters, but claims still need answering. Intentions matter, but consequences still exist.

The right to be wrong is not the right to be untouched by reality.

The burden of being someone else’s lesson

There is also a problem with making patience sound costless.

It is easy to praise the virtues of open dialogue in the abstract. It is harder when the same people are repeatedly expected to absorb, correct, contextualise and forgive errors made about them. There are forms of wrongness that do not simply float in the intellectual atmosphere. They land somewhere. They land on groups who have heard the argument before, answered it before, lived under its consequences before, and are now being asked to perform a patient seminar for someone encountering the issue for the first time.

That exhaustion is real.

No humane account of error can require endless compulsory patience from the people most affected by it. The right to be wrong cannot mean that everyone else has a duty to tutor you through it. A person may deserve room to learn, but they are not entitled to choose their teacher, set the timetable, or demand emotional labour from those already carrying the weight of the mistake.

This is not a rejection of persuasion. It is a recognition that persuasion has costs, and those costs are not evenly distributed.

A decent culture of argument should therefore create routes back from error without making the injured party responsible for building every road. Friends can challenge friends. Institutions can train staff. Editors can apply standards. Political parties can discipline representatives. Schools and universities can teach people how to think rather than merely sorting them into the pure and the damned. The work of correction should exist, but it should not be dumped automatically on whoever has already been harmed.

The right to be wrong is a right against moral finality, not a claim on other people’s endless availability.

Revision is the test

The crucial question, then, is not whether someone was wrong. Of course they were. We all are, with impressive frequency and often with considerable confidence. The more interesting question is what happens next.

Do they correct? Do they listen? Do they show any sign of having understood the objection? Do they distinguish between being challenged and being persecuted? Do they become more careful? Do they stop repeating the claim? Do they repair what can be repaired? Do they accept that an apology is not a receipt entitling them to immediate absolution?

Revision is the test that separates humane error from bad-faith performance.

This does not mean every change of mind must involve a public ritual of humiliation. In fact, we would probably get more genuine revision if we demanded fewer theatrical apologies. The modern apology has become a strange genre: part legal statement, part brand management exercise, part hostage video filmed in the fluorescent corridor of public disapproval. Sometimes people need to apologise. Sometimes they need to make amends. But sometimes they simply need to stop being wrong in the same way.

The measure is not emotional spectacle. It is correction.

That is why repetition matters so much. A first error may deserve patience. A second may deserve sharper correction. A tenth is no longer an error in the same sense. At some point, the refusal to revise becomes part of the wrong itself. Someone who has been repeatedly corrected and keeps saying the same thing is not merely mistaken; they are choosing the comfort of their mistake over the burden of knowing better.

Nobody has a right to permanent intellectual adolescence.

Institutions do not get to be innocently wrong

The distinction becomes even sharper when we move from individuals to institutions.

Institutions love the language of learning, especially when it helps them avoid the language of responsibility. Mistakes were made. Lessons will be learned. Processes will be reviewed. A small working group will be formed, probably with a title so bland it could anaesthetise a horse. The organisation will express regret for any distress caused, which is usually a way of regretting the distress while leaving the cause itself looking suspiciously unsupervised.

Here, the right to be wrong must be much narrower.

Institutions have resources, duties, procedures, records, expertise and power. They are not teenagers at a dinner table trying to work out what they think. When an institution gets something wrong, the question is not only whether it intended harm. The question is whether it should have known, whether warnings were ignored, whether dissent was punished, whether evidence was available, whether the error benefited the institution, and whether anyone affected had meaningful recourse.

The same applies to public figures. A minister, columnist, broadcaster, academic, police chief, council leader or headteacher may still be human. They may still make mistakes. But their wrongness operates through structures other people have to live inside. The threshold for patience is therefore different. So is the threshold for consequence.

This is not because powerful people are uniquely wicked. It is because they are less vulnerable to the ordinary costs of being wrong and more capable of exporting those costs to others.

A culture that cannot distinguish between a private person learning in public and a powerful person failing upwards has lost its moral instruments. It will either punish the weak too harshly or excuse the powerful too easily. Often, with the sort of grim efficiency that makes one doubt human progress, it will do both.

Why mercy matters

Still, the case for mercy remains.

Not because error is harmless. Not because criticism is cruel. Not because everyone deserves a permanent platform for their least informed opinions. Mercy matters because without it, people learn to defend every position as though their identity depends on it. They double down because the alternative is annihilation. They treat correction as attack because, in the social economy around them, correction often is attack. They become less capable of saying the words on which all serious thought depends: I was wrong.

Those words should not be made easy in the sense of costless. Sometimes being wrong has consequences. Sometimes it should. But they should be made possible.

That requires a culture of argument that can distinguish between correction and condemnation, between accountability and permanent disgrace, between disagreement and persecution, between ignorance and malice, between growth and evasion. It requires, in other words, moral discrimination. Not discrimination in the bureaucratic sense, but in the older sense: the ability to tell unlike things apart.

The right to be wrong is really the right to remain morally unfinished.

It is the right not to have the worst version of your thinking treated as the final version of your soul. It is the right to ask a clumsy question and then ask a better one. It is the right to move, under pressure, from confidence to doubt, from doubt to understanding, from understanding to repair. It is a right that belongs most fully to those who are still learning, still listening, still revising.

But it is not a right to irresponsibility. It is not a right to demand endless patience from the people your error harms. It is not a right to turn ignorance into policy, prejudice into entertainment, or repetition into innocence. It is not a right to be wrong forever.

A humane culture should leave people a route back from error. A serious one should still insist that they take it.

There is a particular kind of television writer who makes you feel, very early on, that you are in safe hands. Not comfortable hands, necessarily. Not gentle hands. But hands that know where everything is going. A door closes. A clock starts. A monster is glimpsed but not explained. Someone says something that sounds like a joke, then twenty minutes later it turns out to have been a clue. The story folds back on itself, the impossible becomes inevitable, and you experience the neat, slightly smug pleasure of having been tricked by someone who was good enough to get away with it.

Steven Moffat is extraordinarily good at this.

At his best, he writes television like a magic trick performed in a locked room. Blink is the obvious example, partly because it is still the episode people recommend to those who think they do not like Doctor Who. The Doctor is barely in it. The monster does not move when you can see it. The plot depends on a conversation that can only happen because one side of it has already been recorded. It is a puzzle box, but one with feeling: loss, fear, romance, loneliness, the melancholy strangeness of lives pushed out of sequence. It is clever, yes, but not merely clever. It knows that the trick only matters because the people trapped inside it matter.

The same is true of The Empty Child and The Doctor Dances, still one of the great two-part stories of the revived series. Again, the mechanism is elegant: a wartime ghost story, a child in a gas mask, a phrase that becomes more frightening through repetition, a medical technology misunderstanding the shape of a human body. But what makes it work is not the mechanics alone. It works because the horror is finally resolved as a kind of catharsis. The child is not a demon. The infection is not evil. The frightening question is answered by an act of recognition. For once, just once, everybody lives.

This is why the familiar anti-Moffat rant has always felt a little too easy. The problem with Moffat is not that he is a bad writer. It is not even that he is merely a clever writer, though cleverness is plainly part of both the success and the issue. The problem is that his gifts are unevenly distributed. Give him a premise and a constraint, and he can produce some of the sharpest popular television of the last twenty years. Give him a whole mythology, a years-long emotional arc, or a beloved source text he feels compelled to outflank, and the same instincts can begin to turn against him.

A locked room needs a solution. A universe needs consequences. Moffat is often brilliant at the first, and much less reliable with the second.

The Case for the Defence

It is worth making the generous case first, because otherwise this becomes just another piece of fandom prosecution: Exhibit A, the duck pond; Exhibit B, the Silence; Exhibit C, something about why the Doctor could not simply land in New Jersey and get a taxi.

That sort of argument can be fun, but it is rarely as decisive as its advocates think. Doctor Who is nearly impossible to keep tidy. This is a programme in which continuity is not so much a rulebook as a compost heap. Things are buried, forgotten, contradicted, revived, reinterpreted, and occasionally turned into Cybermen. Demanding strict consistency from Doctor Who is like criticising a pantomime horse for not having plausible veterinary records.

Nor should Moffat be judged only by the story arcs that became unwieldy. His contribution to modern British television is enormous. He was lead writer and executive producer on Dr Who through some of its most successful years, steering the programme through Matt Smith, Peter Capaldi, the fiftieth anniversary, and the global expansion of its modern brand, while doing his acclaimed Sherlock mini-series at the same time. At the height of his powers, two of the BBCs most valuable IPs were under his purview simultaneously.

That is not a small achievement. It is also not a neutral one. The Moffat era of Dr Who has a texture. It has fairy-tale narration, speeches about stories, loops of time, clever girls, impossible boxes, ancient questions, confession dials, memory tricks, monsters that operate as thought experiments, and a Doctor who increasingly understands himself as both person and legend. When it works, it is magnificent. When it does not, it can feel like being trapped in a TED Talk delivered by a conjurer.

Even Blink, the early exhibit most often used against the later Moffat, should be understood as evidence of genuine brilliance rather than as a stick with which to beat him. That is the crucial point. The best Moffat stories are not good because they are complicated. They are good because their complications clarify something. Blink is a terrifying inversion of the typical companion arc in the show - it’s about being stolen from your life and finding yourself in a different time. The Empty Child is about war, childhood, shame and restoration. The Girl in the Fireplace is about intimacy across impossible distance. Heaven Sent, perhaps the strongest defence of Moffat as short-form dramatist, is a formal experiment about grief, endurance and obsession. These are puzzles with wounds inside them.

The criticism, then, has to be more precise. Moffat does not fail because he is too clever. He fails when cleverness becomes a solvent: when it dissolves consequence, rewrites emotional reality, or turns characters into components in a machine whose final movement is designed to make the writer look one step ahead of the audience.

Closed Systems and Open Worlds

The distinction that matters is between closed systems and open worlds.

A closed system is a story with boundaries. It may be a single episode, a two-parter, a bottle episode, a miniseries with a clear premise, or a story built around one formal problem. The audience understands, even if only subconsciously, that the promise being made is finite. Here is the monster. Here is the rule. Here is the puzzle. Here are the people trapped inside it. By the end, the mechanism has to be resolved. Not every mystery must be explained in full, but the story must give the impression that it has paid its debts.

An open world is different. A long-running television universe cannot survive on clever resets alone. It accumulates. Characters remember things, or should. Deaths alter relationships, or should. Historical events become part of the emotional architecture. A line said in Series 5 can return in Series 7, but if it returns only as a flourish, the audience starts to notice the gap between foreshadowing and fulfilment.

This is where Moffat’s strengths become unstable. He is a master of setup. He can make an image feel instantly mythic: a crack in a bedroom wall, a name hidden in plain sight, a woman meeting the Doctor in the wrong order, a question that must never be answered, a hybrid, a confession dial, the myth of a tomb on Trenzalore. These are potent hooks. They do what good serial television hooks are supposed to do: they make the viewer feel that the story is larger than the episode in front of them.

The difficulty is that a hook is not a payoff. A mystery is not an arc. A promise is not a conclusion.

Moffat’s long-form writing often gives the impression of someone laying track while the train is already moving, which, to be fair, is probably what all television showrunners are doing most of the time. The difference is that Moffat’s tracks are so ornate that you expect them to be leading somewhere architecturally exact. When they do not, the disappointment is sharper. The problem is not that the puzzle is unresolved. The problem is that the story has trained you to expect a solution, then offers a flourish, a technicality, or a lovely speech in its place.

This matters because clever television relies on trust. The audience does not need to guess the ending, but it does need to believe that the ending has been arrived at rather than escaped through.

Doctor Who as a Magic Trick

The Matt Smith era begins with one of the cleanest soft relaunches imaginable. A child meets a strange man in her garden. There is a crack in her wall. He promises to come back in five minutes and returns twelve years late. It is funny, sad, strange, and instantly legible. Amelia Pond is introduced through absence. The Doctor is introduced as someone wonderful who nevertheless ruins lives by accident. The fairy-tale language is not yet self-indulgent; it is doing useful dramatic work.

Series 5 remains the strongest of Moffat’s full seasons because its central image is simple enough to bear repetition. The cracks in time are visual, thematic, and emotional. They erase people from history. They mirror Amy’s abandonment. They give the series a recurring wound. The Pandorica Opens and The Big Bang may already rely on a certain amount of hand-waving, but they are buoyant enough to get away with it. They feel like a conjuring trick performed with enough energy that you admire the movement even when you suspect the cards are being palmed. The episode itself was in some ways riffing on his Comic Relief episode, The Curse of Fatal Death, but played seriously. Still clever, and very Moffat.

But the habits that later become frustrating are visible even there. The universe ends, then is rebooted. Rory dies, then returns, then becomes plastic, then waits two thousand years, then perhaps remembers it and perhaps does not, depending on what the scene needs. The Doctor is erased from history, then restored by memory. The emotional effects are real in the moment, but the rules are already getting soft around the edges. The constant playing-around with the fungibility of time starts to take away from the weight of anything that happens in a given story (hence why Rory dying became a bit of a running joke later on).

By Series 6, that softness and flexibility has become the point. The Doctor dies, except he does not. Amy is pregnant, except she is not even there. Melody is stolen, except the story is less interested in her rescue than in the elegance of revealing that she is River Song. A fixed point in time becomes a technical performance. The Doctor’s death is faked in front of false witnesses, inside a manipulated event, to satisfy a universe that appears to care about narrative form rather than reality.

That is a clever idea. It is also, emotionally, a bit of a cheat.

The abduction of Amy and Rory’s child should be one of the most horrifying events in the revived series. Their baby is taken, weaponised, conditioned, and effectively folded into the Doctor’s mythology. Yet the story never fully sits with the parental devastation of that. It cannot, because the machinery has already decided that Melody is River, River is glamorous, the timeline is closed, and the sadness must be transmuted into clever inevitability. The result is not that the plot literally makes no sense, though parts of it are certainly underpowered. The deeper problem is that the plot demands an emotional seriousness it does not then sustain. The only way it can be rationalised against the events we already know have to happen is that we have to believe that Melody’s parents agree, if only tacitly, to let her be tortured, brainwashed and whatever else because it’s a bit of a faff to prevent it - though, for a counterexample, they rewrite Kazran’s entire history to suit their own purposes in ‘A Christmas Carol’, so it’s unclear why (see my earlier comment about ‘playing-around with the fungibility of time’).

In the Russell T Davies era, there were clear, consistent rules to the story. Fixed points in time, and the mythos of the Time War, functioned as both narrative and moral limits. The Doctor could be brilliant, improvisatory, even godlike, but some things had happened and could not simply be made unhappened. His grief over Gallifrey mattered because it was not just a misunderstanding waiting for an anniversary special.

Moffat’s fiftieth anniversary story, The Day of the Doctor, is in many ways a triumph of event television. It is funny, warm, generous, and theatrically satisfying. It gives the revived series a way to stop making the Doctor’s identity rest on an act of double genocide. I understand the impulse. There is only so long a family adventure series can carry the premise that its hero burned his own people and billions of children.

But the cost is real. Once Gallifrey is saved, even in a pocket universe, the emotional architecture of the previous era shifts. The Doctor’s guilt is not erased, exactly, because he still believed he had done it. But belief is not the same as deed. The story wants the catharsis of redemption while also wanting to preserve the tragic weight of the original sin. It relies, temporally, on a Schrodinger’s Cat style setup - Gallifrey destroyed and not destroyed, the Doctor guilty and not guilty, the past changed and not changed - in order to get everyone to the right place.

This is very Moffat. The contradiction is not ignored; it is made shimmering. The audience is invited to admire the paradox. But paradox is not always depth. Sometimes it is a beautifully lit escape hatch.

Sherlock, Dracula, and the Source Material Problem

This pattern becomes clearer when you look beyond Doctor Who.

Moffat’s Sherlock, co-created with Mark Gatiss, begins with a near-perfect act of translation. The original Holmes stories are not treated as museum pieces. They are adapted into the texture of contemporary London: texts, taxis, nicotine patches, blogs, forensic labs, police irritation, celebrity menace. The first episode has the same thrill as the best Moffat premises: it feels at once obvious and new. Of course Sherlock Holmes would thrive in the age of the smartphone. Of course Watson would blog. Of course deduction could be made visually kinetic. Of course Moriarty would become not simply a criminal mastermind, but a kind of performance artist of chaos.

Again, the early pleasure is real. The problem comes when the series increasingly has to answer the implications of its own cleverness. Sherlock’s apparent death at the end of Series 2 is a perfect example. As a cliffhanger, it is irresistible. As a character moment, it is powerful. As a problem for the audience to solve, it became a phenomenon. But the longer the show teased the solution, the clearer it became that no answer could satisfy the promise being made. By the time the series returns, the solution is displaced into jokes, fake-outs, fan-baiting and meta-commentary. The show does not resolve the trick so much as mock the desire for resolution.

There is a defensible version of that move. Audiences can become tedious in their demand that every trick be diagrammed. Explanation can kill magic. But Sherlock had not merely allowed viewers to wonder; it had actively trained them to believe that the machinery mattered. To then shrug at the machinery is not liberation from pedantry. It is a breach of contract.

The same broader tendency appears in Dracula. The first episode, set around Jonathan Harker’s captivity and escape, has atmosphere, menace, wit and structure. It understands Dracula as both predator and conversationalist. It has a strong central location, a limited field of action, and the slow tightening of dread. Then the adaptation begins to expand, update, reverse, comment on and outmanoeuvre its source material. By the time it arrives at its later movements, the question becomes less ‘what is this story about?’ and more ‘what surprising thing are Moffat and Gatiss going to do with Dracula?’ I won’t even talk about the Jekyll miniseries.

That is the source material problem. Moffat is rarely content simply to adapt. At his best, that is admirable. Dead reverence is not fidelity. A good adaptation should have an argument with its source. But there is a difference between arguing with the source material and trying to beat it. The latter produces a particular strain of adaptation in which the writer seems less interested in the story’s durable power than in proving that he has seen through it.

The result can be exhilarating, for a while. Then it starts to feel defensive. The show keeps winking at the audience, anticipating objections, reversing expectations, revealing that the old story was really about this other thing all along. At some point, one begins to miss the simple confidence of the original premise. A vampire in a castle is not an unsophisticated idea merely because it is old.

The Difference Between Clever and Wise

The deeper distinction here is between cleverness and wisdom.

Cleverness is not a vice. It is fashionable, in some circles, to treat clever writing as somehow less sincere than plain writing, as though art becomes morally purer when nobody involved appears to have enjoyed constructing it. This is nonsense. A well-made plot is a joy. Formal ingenuity can produce emotional force. Some of the most moving stories ever told are intricate machines.

Moffat’s best writing proves this. Blink is clever all the way down, but the cleverness is also the source of its sadness. Kathy’s letter, Billy’s life, Sally’s conversation with a man trapped out of order: these are structural tricks, but they are not empty ones. The form creates the feeling. You cannot separate the puzzle from the ache.

Wisdom is something else. Wisdom is knowing when the trick should stop. It is knowing that a death cannot always be a misdirection, that grief cannot always be converted into lore, that the audience’s investment in a mystery is not the same as investment in a person. It is knowing that stories need friction. They need the resistance of things that cannot simply be revised.

This is why Moffat’s handling of death can become oddly weightless. Characters die and return, are copied, saved, uploaded, remembered back into existence, revealed to have avoided the real consequence by a technicality. To be clear, Doctor Who has always done this to an extent. You can argue that Regeneration itself is one of television’s great anti-death devices - but it does, fairly definitively, end something. When every ending comes with a hidden escape hatch, the audience stops believing in the ending.

The trouble is not that Moffat refuses darkness. His writing can be very dark. The trouble is that he often wants darkness without finality. He wants the emotional grandeur of sacrifice, but also the intellectual pleasure of undoing it. He wants the tomb, the ghost, the impossible survival, the loophole, the speech, and the applause.

Sometimes that works. Both Blink and Heaven Sent work because the time is the horror. In the former, living a happy life and dying of old age is turned into something cruel; in the latter, the eventual escape is paid for through billions of years of suffering. The trick and the trauma are aligned. But elsewhere the balance slips. The story says: this matters. Then the structure says: not quite as much as you thought.

That is the point at which cleverness stops deepening the drama and starts protecting the writer from it.

The Joy of Constraint

This is why Boom, Moffat’s 2024 return to Doctor Who, is such a useful test case. He takes away the Doctor’s usual advantages: he cannot run about, cannot bamboozle people, and one wrong move could be fatal.

Whatever one makes of the episode as a whole, that premise is pure Moffat in his strongest mode. It is a constraint. It is a locked-room problem, even absent the four walls one would usually expect. The Doctor, a character defined by movement, improvisation and verbal control, is made still. The story’s tension comes from subtraction. What happens when you take away the running? What happens when the cleverest man in the scene cannot physically dominate the scene? What happens when charisma cannot substitute for action?

That is the sort of problem Moffat understands instinctively. It has rules. It has pressure. It has a theatrical shape. It does not need to sustain three seasons of mythology. It needs to hold, intensify, complicate, and resolve.

There is a lesson in that. Moffat’s best work often emerges when he is denied sprawl. The cheap episode, the Doctor-lite episode, the bottle episode, the single impossible scenario: these forms discipline him. They stop the premise from multiplying indefinitely. They force the story to end.

This is not a minor virtue. Endings are not just the point at which a story stops. They are where the story reveals what it thinks it has been doing. A weak ending does not merely disappoint the final ten minutes; it sends a tremor backwards through the whole structure. The clues become decoration. The promises become vibes. The emotional stakes become temporary intensities rather than durable facts.

Moffat’s recurring problem is that he is often better at generating promise than at submitting to the obligations promise creates. He can make you believe that a mystery matters. He can make a phrase sound ancient, a question sound fatal, a character sound mythic. But sooner or later, the story has to decide whether those things are more than atmosphere.

The locked room helps because it refuses infinite deferral. There is no next season to hide in. No myth arc large enough to absorb the contradiction. No source text to outsmart. Just the problem, the people, and the ending.

Conclusion: Let the Trick End

The frustrating thing about Steven Moffat is that the criticism only matters because the talent is real.

There would be no point complaining that a mediocre writer sometimes fumbles a story arc. Most television does. The reason Moffat provokes this particular irritation is that he so often gets close to greatness. He has written episodes that are not merely good Doctor Who, but exceptional television in a larger sense: elegant, frightening, funny, emotionally precise, formally inventive. He can turn production constraints into virtues. He can make exposition dance. He can make a monster out of an idea you wish you had thought of first.

But he can also mistake complication for depth, reversal for resolution, and self-awareness for emotional honesty. His long arcs sometimes feel less like stories unfolding than like promises being refinanced. The debt is never quite paid; it is rolled into the next mystery.

Perhaps this is partly a Doctor Who problem rather than only a Moffat problem. The programme resists definitive answers because its format is a machine for renewal. It eats endings. It turns finality into anniversary material. It has been contradicting itself since before most of its current audience was born. Big myth arcs have often sat awkwardly inside it, from classic-era experiments like The Key to Time and The Trial of a Time Lord to the revived series’ various attempts at seasons-long mysteries.

Still, Moffat’s writing makes the problem unusually visible because his stories are so invested in the elegance of design. Russell T Davies often resolves arcs through emotion, spectacle and melodrama. Whether that is better is a separate argument, and not always an easy one. But Davies’s messes tend to feel like explosions. Moffat’s messes feel like equations that have noticed, too late, that one of the numbers was a person.

That, finally, is the difference between a puzzle and a story. A puzzle is solved when the mechanism is revealed. A story is resolved when the consequences have been faced. Moffat is one of modern television’s great puzzle-makers. The trick is dazzling. The box is beautifully made. The hinges are invisible.

But sometimes the best thing you can do for a writer like that is shut the door, lock it, and refuse to give him a universe to escape into.

The first is the old nostalgic version: empire as railways, cricket, civil service examinations, parliamentary manners, and a general sense that Britain, while perhaps a little severe at times, mostly went around the world improving things and making tea. This is the empire of sentimental maps and grand strategy games, the ‘pink bits’ version of history, in which domination somehow becomes a form of national generosity. It is not always an explicitly wicked argument. Quite often it is worse than that: it is complacent, incurious, self-flattering, and historically lazy.

The second is the flattened anti-imperial version: empire as one undifferentiated block of brutalising evil, stretching across centuries, continents and political forms, with every institution serving the same purpose in the same way. This version has the virtue of being far less morally evasive. It at least begins from the correct premise that empire was a system of domination, not a charming international outreach programme with better uniforms. But it can still fail as history. It can turn empire into a mood of condemnation rather than an object of analysis.

The problem with imperial nostalgia is that it tries to launder power through vibes. The problem with the cartoon version is that it sometimes accepts the same basic framing in reverse. Both are tempted to treat the empire as a moral atmosphere: pride or shame, grandeur or guilt, civilising mission or criminal conspiracy.

But the British Empire was not a vibes-based institution.

It was not a mood. It was not a personality trait. It was not a single moral substance spread evenly across the map. It was a changing set of institutions, practices, bargains, fantasies and coercive arrangements. It was companies, governors, soldiers, surveyors, missionaries, district officers, planters, judges, investors, settlers, police, prisons, customs houses, schools, ports, railways, maps and files. It was commerce becoming sovereignty. It was violence becoming administration. It was hierarchy becoming common sense.

That distinction matters, because the argument we keep having about empire is usually too small for the thing itself.

The question is not whether empire was good or bad

This is the first trap.

Ask whether the British Empire was good or bad, and you have already made it easier to talk nonsense. The question invites a balance sheet. On one side: railways, parliamentary institutions, the English language, the abolition of slavery, perhaps the rule of law if the speaker is feeling especially pleased with himself (and it’s almost always a himself). On the other side: slavery, famine, massacre, land theft, racial hierarchy, forced labour, punitive violence, extraction, and the long afterlife of colonial inequality.

This format is not neutral. It quietly implies that history can be measured like supermarket produce. So many ounces of railways. So many quarts of massacre. Add cricket, subtract Amritsar, divide by parliamentary procedure, and see whether the total comes out as national pride or national shame.

Can we all please agree, as a baseline, that this is a ridiculous way to think.

It is not ridiculous because moral judgement has no place in history. Of course it does. The empire involved domination, and domination requires moral judgement. But judgement without mechanism is often just a posture. To say that empire was bad may be true, but it is not yet an explanation. To say that it was complicated may also be true, but it is often used as a way of avoiding the rather important fact that many complications are complications of coercion.

The more useful question is not: was the empire good or bad?

It is: how did it work?

Who governed whom? By what authority? For whose profit? Through what institutions? Backed by what violence? Justified by what language? Recorded in whose archives? Remembered in whose textbooks? Forgotten by whose descendants?

Once you ask those questions, the empire becomes harder to sentimentalise and harder to flatten. It becomes something more interesting than a national pride exercise, and more disturbing than a slogan.

Commerce did not accompany empire. Sometimes commerce became empire.

One reason nostalgic accounts of empire fail is that they often begin too late. They start with administrators, governors, imperial conferences and constitutional reform. They prefer the empire once it has acquired a mahogany desk.

But one of the central facts about British imperial history is that private commerce could become territorial power. The East India Company is the obvious case, though obvious things still need saying because people keep declining to say them.

The Company was not founded as a moral project. It was founded to trade. Yet the institution that began as a commercial enterprise came to rule large areas of the Indian subcontinent, exercising military power through what the National Army Museum describes as its formidable private army. The British Library’s India Office Records include records of the East India Company, the Board of Control, the India Office and the Burma Office, which is a neat archival reminder of the path from corporation to empire to state administration.

This is not an incidental detail. It changes the argument.

If empire were merely an overexcited expression of national destiny, the East India Company would be an awkward preface. In reality, it is one of the main plots. Commercial privilege required protection. Protection required soldiers. Soldiers required revenue. Revenue required territory. Territory required administration. Administration required classification, taxation, courts, records and coercion. At each stage, the arrangement became harder to describe as mere trade and easier to recognise as rule.

This is why the phrase ‘the British Empire’ can mislead as much as it explains. It suggests a single actor with a single intention. In practice, empire was often improvised by people and institutions pursuing narrower interests: merchants seeking monopoly, settlers seeking land, governors seeking order, officers seeking prestige, missionaries seeking souls, politicians seeking advantage, investors seeking returns.

None of this makes empire accidental, and the improvisational nature of the thing is not innocence. A mugging is not less of a mugging because the mugger had not written a five-year strategic plan.

But it does mean that empire should be understood as a system that grew through contextualised incentives as well as ideology. It was not just imposed from the centre. It was pulled outward by profit, fear, local conflict, strategic rivalry, debt, prestige and the self-reinforcing logic of possessions already held. Britain acquired commitments, then acquired further commitments to defend them, then justified those commitments as civilisation.

Empire, in that sense, was often less like a single blueprint than a machine that kept discovering reasons to continue being a machine.

The abolition story is true, and also not the story people want it to be

The abolition of the slave trade is one of the favourite rescue exhibits in nostalgic accounts of empire. Britain, we are reminded, abolished the slave trade in 1807, abolished slavery in much of the empire in the 1830s, and used the Royal Navy to suppress the Atlantic trade. This is all true as far as it goes. The West Africa Squadron captured around 1,600 slave ships and freed around 150,000 Africans between 1808 and 1860.

But as usually deployed, the argument performs a moral conjuring trick. It touts Britain’s later suppression of the slave trade while conveniently forgetting Britain’s earlier implication in slavery. It turns abolition into a national absolution.

That is not history. It is laundering.

The British state compensated slave-owners, not the enslaved. The Bank of England notes that the 1833 abolition settlement emancipated more than 800,000 enslaved Africans and included a £20 million compensation package to slave-owners for the loss of what the law had treated as their property. UCL’s Legacies of British Slavery project has traced the long-term effects of that compensation and the ways slave ownership shaped British wealth, institutions and families.

This is exactly the kind of historical fact that breaks the sentimental version without requiring us to deny the abolitionist one.

Britain was not simply a slave-trading villain and then, one day, a virtuous abolitionist hero. Nor was it simply a humanitarian power unfairly criticised by people who refuse to acknowledge moral progress. It was an imperial state deeply entangled with slavery and plantation wealth; it then abolished slavery in a settlement that protected property claims in human beings; it then used naval power to suppress parts of the slave trade elsewhere.

That sequence is morally serious precisely because it is not neat.

The same state could profit, compensate, suppress and congratulate itself. The same empire could produce abolitionists and enslavers, naval patrols and plantation fortunes, humanitarian rhetoric and racial hierarchy. The same imperial power could help end one form of organised atrocity while preserving the wealth produced by it.

If your account of empire cannot hold all of that at once, it is not an account of empire. It is a poster.

Administration was not the opposite of violence

Another comforting myth is that empire became respectable when it became administrative. Settlement and conquest may have been rough, the story goes, but later came government: courts, officials, roads, law codes, censuses, district reports, schools, development schemes. The sword gave way to the file.

There is something to this, but again it is incomplete in a way that matters. Administration did not replace coercion. Administration organised coercion, regularised it, gave it a budget line and a filing system.

This is where empire becomes harder to see if we only look for spectacular violence. Massacres are morally clarifying. Papers are not. A village burned in reprisal announces itself as violence. A tax demand, a labour ordinance, a pass system, an imposed boundary, a legal distinction between types of subject, a land settlement or a census category may appear merely bureaucratic. Yet these are also ways of making power durable.

The British Empire was very good at turning domination into procedure.

In parts of Africa, indirect rule provided a governing method that could present itself as continuity while reorganising authority under colonial supervision. Lugard’s famous ‘dual mandate’ framed British rule in tropical Africa as serving both metropolitan economic benefit and the supposed uplift of colonised peoples. That formulation is revealing not because it was uniquely hypocritical, but because it was so perfectly imperial. It fused extraction and improvement into one administrative language. Britain would benefit; colonised peoples would supposedly be developed; the contradiction would be managed by officials.

The paternalism was not decorative. It was functional.

If empire can persuade itself that it is ruling for the good of the ruled, then coercion becomes discipline, extraction becomes development, resistance becomes disorder, and dissent becomes evidence that the natives are not yet ready. The empire did not need every official to be personally sadistic. It needed enough of them to believe that hierarchy was natural, that British authority was preferable to the alternatives, and that opposition could be read as immaturity, fanaticism or sedition.

This is why arguments about whether individual imperial servants were good or bad people often miss the point. Some were brutal. Some were diligent. Some were corrupt. Some were sincere. Some believed themselves to be doing good. The machinery did not require pure villainy. Indeed, it probably worked better when staffed by men who could write home with a clean conscience.

That is often how oppressive systems survive: not because everyone inside them wakes up each morning thinking evil thoughts, but because the work has been divided into respectable tasks.

Violence was not a regrettable exception

Still, we should not let the administrative texture of empire soften the point. The British Empire was also violent in the ordinary, physical, unmistakable sense.

The Jallianwala Bagh massacre in 1919 remains one of the clearest examples. British troops fired on a crowd of unarmed Indians in Amritsar, killing hundreds and wounding many more; the House of Commons Library describes it as a notorious episode in the history of British colonialism in India. Parliament’s own records note its significance as a turning point in the history of the Raj and British Empire in India.

But even here, the point is not simply that an atrocity happened. The point is what the atrocity reveals about the political order around it.

Massacre is never only massacre. It is also a statement about fear, legitimacy and the imagined boundaries of obedience. General Dyer’s violence was shocking, but it did not descend from nowhere. It emerged from a system in which imperial authority interpreted Indian political mobilisation as threat, disorder and potential rebellion. The violence at Amritsar was extreme. The anxiety behind it was imperial.

The same is true, in a different late-imperial setting, of Kenya. The Mau Mau rebellion began in 1952; the colonial government declared a state of emergency that lasted until 1960, and Kenya became independent in 1963. This is not some remote Tudor prehistory of empire. It belongs to the age of television, jet travel, the welfare state and the postwar international order. It sits uncomfortably close to the modern Britain that prefers to imagine decolonisation as a polite constitutional seminar conducted by men in suits.

The violence of late empire matters because it punctures a consoling chronology. It is tempting to imagine British imperial history as a gradual softening: conquest, then governance, then reform, then independence. That sequence contains some truth. It also makes the story too tidy. The end of empire was not simply the moment Britain magnanimously handed over the keys. It was also a period of insurgency, repression, detention, censorship, emergency powers, strategic withdrawal and archival management.

The paper trail matters here too. The National Archives’ guidance on migrated colonial administration files is itself a reminder that imperial power did not end when the flag came down. It continued in decisions about what records were retained, removed, destroyed, classified or released. The empire governed people in life and then tried, in places, to govern the evidence by which that life would later be understood.

That should make us suspicious of any account that relies too heavily on imperial self-description. Empires are very rarely neutral witnesses to themselves.

Development was real. That does not make it innocent

At this point the nostalgic critic usually reappears, slightly irritated, and asks whether we are allowed to mention anything that empire built.

Yes. We are. In fact, we have to.

The British Empire built railways, ports, courts, universities, civil services, schools and administrative systems. It spread the English language. It shaped legal and parliamentary institutions. In some contexts, postcolonial states inherited and repurposed imperial institutions in ways that mattered. People educated in colonial schools used that education against colonial rule. Lawyers trained in imperial legal systems challenged imperial injustice. Railways built for extraction could also move people, ideas and anti-colonial politics.

The problem is not mentioning these things. The problem is treating them as if they answer the moral and historical question by existing.

A railway is not an argument. A court is not a verdict on empire. A school is not proof of benevolence. Institutions have purposes, beneficiaries and exclusions. They are built somewhere rather than elsewhere, funded somehow rather than otherwise, and designed to serve some needs more than others.

Many imperial infrastructure projects were built around extraction, military movement, administrative control or settler convenience. That does not mean nobody else benefited from them. History is not that simple. But it does mean that benefit cannot be read backwards as intention, and intention cannot be used to erase hierarchy.

This is where the ‘balance sheet’ view of empire becomes especially misleading. It asks whether colonised societies received some benefits, as if the receipt of benefit settles the question of domination. But domination does not become consensual because the ruler builds things. A prison with excellent plumbing is still a prison. This is not to compare every imperial institution to a prison; it is to point out the poverty of the argument. Provision and domination are not opposites. They can coexist. Indeed, governing systems often provide precisely because they dominate.

The serious question is not whether empire developed things. It is what kind of development, under whose control, for whose purposes, with what exclusions, and at what cost.

The empire was also a myth-making machine

Empires do not survive by force alone. They need stories.

The British Empire told several stories about itself. It was a commercial story, in which Britain brought trade and improvement. It was a humanitarian story, in which Britain suppressed slavery and barbarism. It was a constitutional story, in which British rule trained peoples for eventual self-government. It was a racial story, in which hierarchy was naturalised as maturity, civilisation or capacity. It was a domestic story, in which imperial expansion became part of British identity: school maps, exhibitions, adventure fiction, military memory, museums, statues and family fortunes discreetly converted into respectability.

These stories were not merely propaganda pasted onto the real machinery. They were part of the machinery.

A governor who believes he is preparing a people for self-rule can justify delaying that self-rule indefinitely. A public that believes empire is fundamentally humanitarian can treat anti-colonial resistance as ingratitude. A country that remembers abolition but forgets compensation can imagine itself morally heroic without asking who was paid, who was not, and whose wealth survived the transition.

This is why the empire remains so politically usable. It provides emotional raw material for arguments that are not really about the seventeenth, eighteenth or nineteenth centuries at all. Arguments about empire are often displaced arguments about Britain now: immigration, race, national decline, foreign policy, museums, education, patriotism, guilt, class, sovereignty and the question of whether Britain is still allowed to think well of itself.

There is polling evidence for this continuing emotional charge. In 2025, YouGov found that the most common British view was that countries colonised by Britain were better off for the experience, though views varied sharply by party and age. That does not tell us what the empire was. It tells us that the empire still functions as a test of political temperament.

And that is precisely the problem.

If empire becomes a test of whether you are patriotic or ashamed, proud or woke, grateful or resentful, then the historical object disappears. You are no longer talking about India, Kenya, Jamaica, Ireland, Nigeria, Malaya, Australia, Canada, Egypt or the countless other places folded into different imperial arrangements at different times. You are talking about yourself.

This is what I mean by vibes-based history. It is history as self-expression.

The pink map was never the whole story

There is a practical question here about scope. Should we talk only about the formal empire, the pink-shaded map that stretched from the early modern period through the great decolonisations of the twentieth century? Or should we include the post-imperial world: Commonwealth connections, aid, migration, finance, law, military bases, overseas territories and the persistence of old imperial lines?

The answer, I think, is both, but not equally.

The core subject should be formal empire. Otherwise the piece risks becoming too large and too vague. The British Empire already covers a huge range of institutions and experiences. Britannica summarises it as a worldwide system of dependencies over roughly three centuries, with British territorial acquisition beginning in the early seventeenth century through settlements and trading posts in North America, the West Indies, South Asia and Africa. At its greatest extent, it covered about a quarter of the world’s land surface and more than a quarter of its population.

That is quite enough to be getting on with.

But the afterlife matters because formal decolonisation did not wipe the slate clean. Old imperial relationships shaped migration routes, diplomatic habits, aid priorities, legal systems, educational institutions, military ties, financial networks and cultural memory. In a 1965 House of Lords debate, the government stated that in 1964 Britain’s total bilateral aid, including technical assistance, was £174 million, of which the independent Commonwealth received £108 million, or 62%. That is not empire in the old sense. It is not the Raj. It is not a governor in a plumed hat, thank God. But it does show how post-imperial Britain continued to move through channels shaped by empire.

The distinction is important. Saying that imperial influence persisted after empire is not the same as saying nothing changed. Things changed enormously. Independent states were not simply colonies under another name. Anti-colonial movements won real victories. Sovereignty mattered. Domestic politics mattered. The Cold War mattered. Global capitalism mattered. Postcolonial governments made their own choices, sometimes under severe constraint and sometimes with considerable agency.

But neither should we pretend that the end of formal rule dissolved the structures empire had made. A flag can come down more quickly than a trade pattern, a legal code, a racial hierarchy, a border, a military relationship or a national fantasy.

So yes, we should talk about the afterlife of empire. But we should talk about it as afterlife, not as proof that nothing ended.

How to talk about empire without making everyone stupider

The aim here is not balance. Balance is often the wrong ambition. If one person says the empire was a noble gift to the world and another says it was a structure of domination, the truth does not necessarily lie halfway between them. Sometimes the midpoint between fantasy and reality is just a more moderate fantasy.

The better aim is precision.

Precision means refusing nostalgia. The empire was not a national charity. It was not a global infrastructure scheme. It was not a long, misunderstood act of British helpfulness. Its institutions rested on unequal power, and that unequal power was enforced. Sometimes it was enforced spectacularly, through massacre, detention, punitive expeditions and emergency rule. More often it was enforced quietly, through law, land, labour, taxation and administrative classification.

Precision also means refusing to make the empire simpler than it was. It was not one thing happening in one way for one reason. It changed over time. It varied by place. Settler colonialism in Australia was not the same as Company rule in India; plantation slavery in the Caribbean was not the same as indirect rule in Northern Nigeria; the white Dominions were not governed like Kenya; Ireland sits awkwardly and revealingly inside the story; the empire of free trade, finance and naval supremacy does not map perfectly onto the empire of formal annexation.

Precision means asking what institutions did, not merely what they claimed. A railway may be infrastructure, extraction, military logistics and unintended political connection all at once. A school may be assimilation, opportunity and anti-colonial training ground. A court may impose foreign law and provide tools later used against colonial authority. An archive may preserve evidence and conceal it. An abolition campaign may represent real moral progress while leaving intact the wealth of slave-owners.

Precision means remembering that colonised people were not scenery. They were not passive recipients of British action, whether benevolent or brutal. They resisted, collaborated, adapted, negotiated, converted, litigated, rebelled, governed, translated, profited, suffered and survived. One of the quiet failures of many British arguments about empire is that even the anti-imperial versions can become too Britain-centred. Britain does a thing; Britain regrets a thing; Britain debates whether the thing was good or bad. The rest of the world waits patiently to provide moral evidence.

That, too, is an imperial habit.

Most of all, precision means accepting that moral seriousness and historical complexity are not enemies. Complexity does not rescue empire from judgement. It makes the judgement better.

The British Empire was an engine of commerce, coercion, administration, hierarchy, violence, improvisation and myth. It created institutions and destroyed worlds. It built and extracted. It governed and classified. It educated and subordinated. It abolished slavery after profiting from slavery, compensated enslavers rather than the enslaved, and congratulated itself for the moral achievement. It produced officials who believed in duty and systems that depended on domination. It ended formally in many places while leaving behind lines of connection, dependency, resentment, inheritance and memory.

None of this fits comfortably into pride or shame. That is the point.

Pride and shame are things people feel. Empire is something people did to other people, through institutions that can be studied. The first task is not to decide whether it makes us feel good or bad about Britain. The first task is to look at the machinery.

And once you have looked properly at the machinery, the vibes become rather harder to sustain.

This is understandable. Empathy has often been used as a laundering device. People with hard-edged views learn very quickly that their position sounds better if presented as pain. Prejudice becomes concern. Reaction becomes anxiety. Cruelty becomes a cry for help from a community nobody listened to. Before long, the demand for empathy becomes less a request to be understood than a claim to immunity. You are not allowed to criticise the belief because the belief comes wrapped in feeling.

So let us get the caveat out of the way early. Understanding why someone believes a cruel thing does not make the thing less cruel. Empathy does not decide whether a belief is true, just, proportionate, coherent, or politically acceptable. A person may have entirely intelligible reasons for arriving at an appalling conclusion. Their fear may be real. Their humiliation may be real. Their sense of loss may be real. None of that means the conclusion deserves obedience.

But the opposite mistake is also serious. If empathy can be abused as moral shelter, anti-empathy can become a permission slip for stupidity. Refusing to understand bad arguments does not make them disappear. It often makes them stronger, because you end up fighting a cartoon rather than the thing as it actually lives in people’s heads.

The distinction matters because empathy explains why a belief makes sense to someone. It does not decide whether that belief is right.

The unhelpful warmth of the word

Part of the problem is that empathy sounds nicer than it is. It comes to us dressed in the language of kindness, care and fellow-feeling. It sounds like something one should obviously have more of, like fibre or sleep. Nobody wants to be the person who says they are against empathy, unless they are auditioning for a very particular kind of podcast.

But empathy is not one simple thing. Psychologists often distinguish between affective empathy and cognitive empathy. Affective empathy is closer to feeling with someone; sharing or resonating with another person’s emotional state. Cognitive empathy is closer to understanding another person’s perspective; grasping what they think, feel, fear, want or believe without necessarily sharing it. That distinction is not just academic housekeeping. It changes what we are asking of ourselves in political argument.

Political empathy is usually cognitive before it is emotional. It is not necessarily a warm bath of fellow-feeling. It is not a requirement to feel the world exactly as your opponent feels it. It is the colder and more disciplined act of asking: what experience, fear, incentive, humiliation, media diet, institutional failure or social world makes this view feel reasonable to this person?

That question is not sentimental. It can be forensic. It asks what work the belief is doing. Is it explaining a loss? Displacing a fear? Providing someone with status? Giving moral shape to resentment? Offering a villain? Turning confusion into certainty? Making a complicated world feel legible?

This kind of empathy does not soften judgement. Done properly, it sharpens it.

The diagnostic use of empathy

Imagine someone who believes that immigration is responsible for every visible strain in their town: the queue at the GP, the pressure on housing, the change in shops, the lack of secure work, the unfamiliar languages on the bus, the sense that the place they knew has altered without their permission.

One response is simply to say that this is racist and move on. Sometimes the statement will be racist. Sometimes the politics attached to it will be racist. Sometimes the conclusion will be empirically wrong, morally crude, or encouraged by people who know exactly what they are doing. But the one-word dismissal rarely explains why the argument has purchase.

A more useful response asks what has made the claim feel true. Perhaps local services have been visibly degraded. Perhaps housing is scarce. Perhaps work is insecure. Perhaps the national political class has spent years talking as though belonging is a childish feeling unless it can be expressed in the correct metropolitan vocabulary. Perhaps newspapers and politicians have spent decades converting every social pressure into an immigration story because it is easier than talking honestly about underinvestment, labour markets, planning failure or the British state’s chronic preference for muddling through.

None of this means the anti-immigration conclusion is right. It means the conclusion is attached to something. If you do not understand the attachment, you cannot detach it.

The same is true across half the arguments that make public life unbearable. Some people talk about crime because they genuinely feel unsafe, not because they have read a policy paper on sentencing. Some talk about welfare through resentment because they feel trapped in a system where work has not delivered dignity. Some talk about patriotism defensively because they experience cosmopolitan disdain long before they encounter a serious argument about national identity. Some talk about trans rights with displaced anxieties about sex, authority, childhood, institutional trust and social change. Some talk about Gaza not as a foreign policy question but as a test of whether certain lives are allowed to matter in public.

Again, explanation is not acquittal. A displaced anxiety remains displaced. A false claim remains false. A cruel policy remains cruel. But arguments rarely survive on facts alone. They survive because they organise feeling. Empathy helps us see the organisation.

Listening is not obedience

This is where the language of listening becomes slippery. Politicians are forever listening to concerns. It is one of those phrases that can mean almost anything, which is why it is so useful. It can mean taking people seriously. It can mean buying time. It can mean triangulating towards a position you already wanted to hold. It can mean laundering panic into policy while looking sorrowful about it.

The danger is that listening becomes a mini-referendum. Someone expresses a concern; the politician validates the concern; the validation is treated as a mandate; the mandate becomes a policy whose cruelty is then defended as democratic humility. What else could we do? We listened.

But listening is not obedience, and empathy is not policy. A concern can be sincerely held and still badly aimed. It can describe a real wound while misidentifying the cause. It can be morally urgent in one register and politically disastrous in another. The task of politics is not to convert every expressed anxiety into legislation. It is to understand what people are saying, work out what is true within it, identify what is false or unjust, and respond in a way that is both humane and responsible.

This requires more than nodding sympathetically in a village hall. It requires judgement. It requires being able to say: yes, I can see why this feels like betrayal; no, I will not blame that neighbour for it. Yes, I understand why you feel unsafe; no, I will not build a justice policy on panic. Yes, I accept that rapid change can be disorienting; no, I will not pretend that dignity for one group requires humiliation for another.

Empathy without judgement becomes surrender. Judgement without empathy becomes theatre.

The cartoon opponent problem

One of the bad habits of modern argument is the assumption that bad beliefs must have stupid origins. Sometimes they do. There is no need to romanticise every prejudice into a tragic sociology. Some people believe ugly things because ugly things benefit them, flatter them, entertain them, or give them permission to treat others badly. There is such a thing as bad faith. There is such a thing as cruelty. There is such a thing as enjoying the boot and then composing an essay about economic anxiety afterwards.

But even bad faith has mechanisms. Even cruelty recruits stories. Even prejudice needs a structure of plausibility if it is going to travel beyond the already converted.

When we refuse to understand that structure, we make our opponents easier to despise but harder to defeat. This is especially tempting online, where the reward system favours the most contemptuous summary. Your opponent is not merely wrong; they are a fascist, a snowflake, a groomer, a tankie, a centrist dad, a bigot, a liberal elitist, a traitor, a crank, a ghoul. Some of these labels may occasionally be accurate. Most become less accurate with repetition. All of them can become substitutes for thought.

The cartoon opponent is emotionally useful. It allows you to win the argument before it begins. But it also leaves you with no explanation for why anyone disagrees with you except stupidity or wickedness. If the disagreement persists, your only remaining theory is that there are more stupid and wicked people than you hoped. That may be satisfying. It is not a strategy.

A serious politics has to ask why certain claims become believable. Not because belief confers legitimacy, but because durability requires explanation. If an argument keeps returning, across constituencies, platforms, factions and decades, then something is feeding it. The feed may be material deprivation. It may be status anxiety. It may be institutional failure. It may be propaganda. It may be theological, national, familial, aesthetic, or simply habitual. But unless you know what feeds the belief, you are reduced to swatting at symptoms.

Empathy can be manipulative too

There is another caveat, and it is important. Empathy is not automatically virtuous. Cognitive empathy can be used by therapists, negotiators and peacebuilders. It can also be used by interrogators, con artists, demagogues and advertisers. To understand what someone fears is also to know where pressure can be applied. To understand what someone wants is to know what can be sold to them. To understand someone’s humiliation is to know which enemy to offer them.

This is why the sentimental account of empathy is inadequate. Empathy is not goodness. It is a form of access. What matters is what you do with it.

A politician who understands why people feel abandoned can build solidarity, repair institutions and tell a more truthful story about power. They can also redirect that abandonment towards migrants, benefit claimants, public sector workers, judges, students, trans people, Muslims, cyclists, or whoever is currently available as a national stress ball. Both moves require empathy of a kind. The difference is moral and political judgement.

This is also why empathy should not be treated as the opposite of reason. Paul Bloom’s critique of empathy is useful here, even if one need not follow him all the way. His basic warning is that empathy can be biased, narrow and overly responsive to vivid individual cases. We feel intensely for the person placed before us and fail to reason well about the wider pattern. That criticism is worth taking seriously. Empathy can draw attention, but it cannot by itself allocate justice.

The answer is not to abolish empathy from public life. That would leave politics even more inhuman than it already is, which is quite an achievement. The answer is to put empathy in its proper place. It is a means of understanding, not a final court of appeal.

What empathy is for

The purpose of empathy in public argument is not to collapse disagreement. It is not to make all positions equally valid. It is not to ask vulnerable people to be endlessly patient with those debating their dignity as though it were a seminar question. Nor is it to demand that every person harmed by a political view must perform therapeutic curiosity towards the person harming them.

That last point matters. The duty to understand should fall most heavily on people with power, responsibility, security and a serious interest in persuasion. It should not become another burden placed on those already expected to absorb the consequences of other people’s fears. There is something obscene about telling the target of cruelty that their first obligation is to understand the cruel. Sometimes the first obligation is protection. Sometimes it is refusal. Sometimes it is simply getting through the day.

But for those engaged in argument, politics, organising, writing, teaching or leadership, empathy is part of the work. Not because everyone is secretly reasonable. Not because every conflict can be dissolved by tone. Not because the right conversation will magically convert the committed bigot, the bad-faith operator, or the man in the replies whose profile picture suggests he has strong views about both taxation and women in films.

Empathy matters because most political beliefs are not just propositions. They are identities, loyalties, memories, wounds, habits, class signals, moral intuitions and social inheritances. People do not usually experience themselves as selecting opinions from a shelf. They experience themselves as noticing what is obvious. The task of persuasion is therefore not only to correct information. It is to understand what has made the false thing feel obvious.

That is difficult work. It is slower than denunciation and less flattering than contempt. It does not always succeed. But it is the difference between arguing with a living belief and arguing with a paper target.

The harder discipline

The phrase empathy is not agreement risks sounding like a plea for politeness. It is not. Politeness is sometimes useful and sometimes merely decorative. There are polite forms of cowardice, polite forms of cruelty and polite ways to say monstrous things. Civility, by itself, is not the point.

The point is accuracy.

If someone’s belief is false, we should want to understand the particular shape of its falsehood. If it is cruel, we should want to know what makes the cruelty feel justified. If it is disproportionate, we should want to know what scale of fear it is answering. If it is politically dangerous, we should want to know why it is spreading. The better we understand the belief, the less excuse we have for lazy responses to it.

This is the harder discipline: to understand without surrendering, to judge without caricaturing, to listen without obeying, to empathise without granting moral shelter.

A belief may make sense to someone because of what they have lived through. It may make sense because of what they have been told. It may make sense because it protects them from shame, gives shape to loss, or makes them feel briefly powerful in a world that otherwise humiliates them. All of this is worth knowing.

And then, having known it, we still have to decide.

That is where empathy ends and politics begins.

The Election Night Trick

There is a strange ritual that follows almost every set of local elections. Councillors are elected, councils change hands, planning committees are reshaped, transport priorities shift, and the people responsible for social care, bins, roads, libraries, parks and neighbourhood services are either returned to office or removed from it. Within minutes, the question becomes: what does this mean for Westminster?

This is not a stupid question. It is just a revealingly incomplete one.

Voters do use local elections to send messages to national politicians. Parties campaign under national brands. Councils operate inside financial and legal settlements designed, or neglected, elsewhere. Nobody who has spent more than ten minutes near an actual doorstep can seriously pretend that national politics politely waits outside the ward boundary.

But still. There is something absurd about treating local government as the B-team of politics when it is the part of the state most people encounter without needing to watch the news. You may have no strong view on the internal dynamics of the Cabinet Office. You will, however, notice if your bin has not been emptied, your parent’s care assessment is delayed, your bus disappears, your high street declines, your road becomes a cratered assault course, or the patch of land at the end of your street slowly becomes an unofficial monument to nobody being in charge.

That is where politics hits the kerb.

It is not that Westminster politics is unreal. It plainly is not. National government sets taxes, spending envelopes, legislation, rights, duties, benefit rules, immigration systems, foreign policy, public service settlements and the broad weather in which everyone else has to operate. But national politics often arrives in people’s lives through institutions closer to home. It becomes a pothole filled or left open; a library kept open or closed; a care package approved or delayed; a planning application granted or refused; a street cleaned or abandoned to the patient archaeology of crisp packets, broken glass and vape cartridges.

Local elections are therefore not merely a temperature check on Westminster. They are elections to government. Less glamorous government, certainly. Worse paid, usually. Less likely to produce a memoir, a podcast circuit or a ministerial car. But government all the same.

The problem with treating local elections as national opinion polling with rosettes is not that the national picture is irrelevant. It is that this habit makes local government disappear at precisely the moment when people are choosing who should run it.

The State, at Street Level

Local government is often described in a way that makes it sound administrative rather than political. Councils ‘deliver services’. They ‘manage local priorities’. They ‘support communities’. This language is not wrong, exactly, but it does have the anaesthetic quality of public-sector abstraction. It makes councils sound like filing systems with ward boundaries.

A better description is simpler: councils are the state at street level.

They are where the shared conditions of everyday life are maintained, rationed, repaired, negotiated and sometimes allowed to decay. The Local Government Association’s introductory description of councils is almost comically broad: schools, care for older people, roads, rubbish, libraries, planning and a wider bureaucracy most residents only glimpse when something goes wrong. That range matters. Local government is not one service. It is a dense bundle of responsibilities, some intensely visible and some almost invisible until a family depends on them.

The visible services shape whether a place feels cared for. Bins, roads, pavements, parks, street cleaning, lighting, enforcement, libraries, high streets, public toilets, leisure centres, grass verges, bus stops, crossings, benches. These are not trivial matters, though national politics often treats them as such. They are the texture of public life. They are the difference between a neighbourhood that feels governed and one that feels merely inhabited.

The less visible services shape whether people survive moments of dependence, vulnerability or crisis. Adult social care. Children’s services. Homelessness support. Public health. Safeguarding. Temporary accommodation. Special educational needs transport. The machinery is quieter, but the stakes are higher.

This is one reason local politics produces such an odd mismatch between public salience and institutional importance. Many residents experience councils through annoyance: the missed collection, the pothole, the planning notice, the overgrown alleyway, the reply that arrives three weeks late and answers a slightly different question. But councils also sit underneath some of the most morally serious decisions made by the British state: who receives care, who waits, who is housed, who is protected, which child is supported, which family is left to navigate a system already fraying at the edges.

The bins are political. But they are not only political because people complain about them. They are political because they reveal whether the basic promise of collective life is being kept. We pay in, we share space, we accept rules, we tolerate inconvenience, and in return the ordinary systems are supposed to work. When they do not, people do not experience that as a spreadsheet problem. They experience it as abandonment.

This is why local neglect is so politically potent. A broken pavement does not need a theory of the state attached to it. It teaches one by repetition. Every unfilled hole, every unanswered report, every fly-tipped mattress left to acquire local landmark status, every bus route quietly withdrawn, every shuttered youth centre, every family told there is nowhere suitable to live: these things instruct people in what they think power is for.

The Pothole Is Not Just a Pothole

A pothole is rarely just a pothole.

This is, admittedly, the sort of sentence that makes politics sound more self-important than it needs to be. Sometimes a pothole is also, very straightforwardly, a hole in the road, and it would be nice if someone filled it in before it eats another tyre.

But the pothole is also a system made visible by failure. It is inspection schedules, revenue budgets, capital programmes, risk thresholds, highway contracts, contractor performance, officer capacity, reporting systems, political priorities, winter damage, traffic volume and the slow depreciation of public assets. It is what happens when the invisible work of maintenance is postponed until it becomes visible as decline.

The same is true of the missed bin. To the resident, it is immediate and simple: the bin was full, the bin day came, the bin remains full. Behind that simplicity sit route planning, staffing levels, vehicle maintenance, depot management, contract terms, communications systems, industrial relations, contamination rules, weather disruption and whether anyone has designed a complaints process capable of producing an answer before the smell becomes its own argument.

This is not an attempt to excuse poor service. Quite the opposite. One of the evasions of modern politics is to hide bad outcomes behind complexity. Residents are entitled to expect the basics to work. They should not need a working knowledge of municipal logistics to get their rubbish collected or a Freedom of Information request to understand why the road has been patched so badly that the repair itself looks temporary by lunchtime.

But complexity does matter, because it helps distinguish between two different political failures. One is the failure of competence: the council knew what it had to do, had the capacity to do it, and did it badly. The other is the failure of capacity: the council is operating in a system where demand, legal duties, inflation, workforce pressures and funding constraints have made good government harder than the public conversation admits.

The public has no obligation to find this distinction emotionally satisfying. Nobody standing next to an uncollected bin wants a seminar on structural pressures. But politically, the distinction matters. If every failure is treated as proof of laziness, waste or indifference, then every solution becomes performative toughness: sack the managers, cut the ‘nonsense’, stop the waste, run the council like a business, audit everything, outsource everything, insource everything, declare war on bureaucracy, appoint a commissioner, launch a taskforce, rename the taskforce, put a hi-vis jacket on the taskforce and send it to look concerned next to a blocked drain.

Some of these things may occasionally help, minus perhaps the ceremonial hi-vis. But they do not settle the deeper question. What would it actually take to make the system work? What capacity is missing? What trade-offs are being hidden? What duties are crowding out what services? What national decisions are being laundered through local blame?

The pothole is not just a pothole because maintenance is political. Not in the adolescent sense that everything must be dragged into partisan warfare, but in the older and more serious sense: common life depends on unglamorous systems of repair.

The Cruel Arithmetic of Local Government

Local government is full of cruel arithmetic.

The services most people see are often not the services consuming the greatest pressure. Adult social care is the obvious example. The Institute for Fiscal Studies reported that English local authorities budgeted £24.5 billion for adult social care in 2024–25, and that adult social care now accounts for more than 40% of all local authority service spending. That is not some marginal line in the accounts. It is one of the central facts of local government.

The moral weight of that spending is obvious. Care is not optional. A civilised society does not tell older people, disabled adults, people with complex needs, or families in crisis that the council would love to help but unfortunately the park budget is looking a little thin. Statutory services exist for a reason. They are meant to protect people from being sacrificed to local popularity contests.

But the budget does not become infinite because the duty is serious. Money spent meeting acute need cannot also be spent on every preventative, universal or environmental service that makes places feel liveable. The Institute for Government notes that councils are funded through government grants, council tax and retained business rates; in 2024–25, local authorities in England received 32% of their funding from government grants, 46% from council tax and 22% from retained business rates. That funding mix matters because it leaves councils exposed to national policy choices, local tax bases and demand pressures they cannot simply wish away.

The National Audit Office’s 2025 report is particularly useful here because it cuts through the fantasy that the problem can be reduced to a few badly run councils. It found that, as financial pressures have increased, local authority spending has focused more heavily on demand-led services and unavoidable costs, including adult and children’s social care, homelessness, waste disposal and debt servicing. It also noted that temporary accommodation costs nearly doubled in real terms from £1.10 billion in 2015–16 to £2.13 billion in 2023–24.

That is the local state in a sentence: more crisis, less prevention; more statutory demand, less discretionary repair.

The result is a politics of visible deterioration. Residents notice the things that vanish from ordinary life: youth work, parks maintenance, library hours, local bus support, public toilets, street cleaning, road repairs, neighbourhood wardens, early intervention, small grants to community groups. They may not see the care package, the placement cost, the temporary accommodation bill, the legal duty, the safeguarding risk or the debt servicing that has swallowed the room.

This creates a vicious democratic problem. Councils are judged most often by the services that make places feel cared for, but they are increasingly dominated by services that respond to acute need. The services that build trust are squeezed by the services that prevent catastrophe. It is hard to run cheerful doorstep politics on the message: yes, the high street looks worse, but you should see the statutory pressures.

Still, that is the truth. And politics that cannot say true things about budgets is not serious politics. It may be very electable politics for a while; it may even be emotionally satisfying. But it is not serious.

Local Accountability in a National Cage

None of this means councils should escape blame.

There is a tempting version of the local government defence that goes too far: councils are underfunded, Westminster controls the framework, statutory duties dominate the budget, therefore every local failure is really a national failure. That is too neat. It turns councillors into spectators of their own institutions.

Councils make choices. They set priorities. They manage contracts well or badly. They communicate clearly or evasively. They plan for the long term or drift between crises. They consult in good faith or treat consultation as the decorative stage before a decision already made. They notice neighbourhood neglect or become habituated to it. They can be creative, stubborn, competent and rooted. They can also be complacent, opaque, factional, defensive and slow.

Local accountability matters precisely because councils are not powerless. A council that replies properly, fixes what it can, explains what it cannot, publishes usable information, treats residents as adults, manages money honestly and notices small signs of decline will be experienced differently from one that sends every complaint into a procedural cave system and hopes the resident eventually loses the will to live.

But local accountability operates inside a national cage.

Councils cannot print money. They cannot ignore statutory duties. They cannot borrow freely to plug everyday revenue gaps. They are legally bound into budget discipline, and the chief finance officer has a statutory duty to issue a Section 114 notice if they believe the authority cannot meet its expenditure commitments from its income; once issued, a council cannot make new spending commitments and must meet to consider its response. The phrase ‘council bankruptcy’ is technically misleading, but politically useful, because it captures the point at which local discretion collapses under financial reality.

This is why the usual rhetoric around local government is so often dishonest. It simultaneously demands that councils act as independent local champions and treats them as mere delivery arms of national policy. It blames councillors for visible decline while ignoring the funding and legal structures that shape what they can do. It calls for localism, then designs systems in which local leaders take the blame for decisions whose causes lie elsewhere.

The honest position is less convenient. Councils are accountable for choices, but they are not omnipotent over conditions. Westminster is responsible for the architecture, but local leaders still decide how to inhabit it. A bad national settlement does not excuse bad local government. Good local government does not abolish a bad national settlement.

That distinction is where much of the serious politics sits.

Why This Matters Politically

This matters because local government is one of the main places where anti-politics grows.

People do not usually become alienated from politics because they have carefully compared constitutional theories. They become alienated because things do not work, nobody seems to listen, blame moves in circles, and every institution appears to have a reason why responsibility lies somewhere else. The council blames the contractor. The contractor blames access. The councillor blames the budget. The minister blames local inefficiency. The opposition blames waste. The resident is still standing next to the same pothole.

This is fertile ground for populism because populism is very good at converting neglect into accusation. It looks at complicated institutional decay and offers a cast list: the corrupt council, the woke officers, the wasteful bureaucracy, the metropolitan elite, the migrants, the consultants, the cyclists, the environmentalists, the previous administration, the next administration, anyone except the person currently making the promise.

The appeal is obvious. When people feel ignored, a politics of blame can feel like recognition. At last, someone is saying the place has been neglected. At last, someone is angry. At last, someone is naming the people who let this happen.

The trouble is that recognition is not repair. Anger is not a waste collection strategy. Suspicion is not a social care settlement. Denouncing bureaucracy is not the same as having enough planners, enforcement officers, social workers, housing officers, highways inspectors, procurement expertise or legal capacity to run the place.

Local government is therefore dangerous for protest politics in a way Westminster often is not. In national politics, a party can live for a long time on mood, grievance and symbolic conflict. In local government, the abstraction has a shorter shelf life. Sooner or later, somebody has to set a budget. Somebody has to approve the contract. Somebody has to decide whether to raise council tax, cut a service, close a building, increase charges, sell an asset, reduce opening hours, change eligibility, delay maintenance, or admit that the numbers do not add up.

This is not an argument against insurgent parties winning councils. Democracy includes the right to throw the old lot out. Sometimes the old lot deserve it. But local government exposes the difference between being anti-establishment and being capable of governing. The first requires a microphone. The second requires a budget, a workforce, a legal framework, a delivery plan and the ability to disappoint people honestly.

The same challenge applies to every party, including the more respectable ones. Labour councils cannot live forever on blaming Conservative austerity, however real its consequences. Conservative councils cannot demand fiscal restraint nationally and then act surprised when local restraint hurts. Liberal Democrats cannot campaign as hyper-local fixers while dodging the harder choices of prioritisation. Greens cannot simply describe a more humane, sustainable localism without explaining how planning, transport, housing and budgets would actually be aligned. Reform cannot turn every problem into a morality play about waste and betrayal, then discover in office that statutory duties are not defeated by vibes.

Local government is where rhetoric is forced into contact with the curb edge. That is why it matters.

Conclusion

We should stop treating local elections as if they are merely Westminster’s quarterly performance review.

They are that, in part. It would be precious to pretend otherwise. A voter angry about the government does not become a different citizen because the ballot paper lists council candidates. National politics seeps into local elections because national politics seeps into everything.

But local elections are also decisions about who governs the ordinary conditions of shared life. They decide who sits closest to the services people rely on most directly and notice most quickly when they fail. They shape whether neighbourhood decline is tolerated or challenged, whether residents are answered or managed, whether scarce money is spent with clarity or drift, whether the council understands itself as a living democratic institution or merely as a machine for processing statutory obligations until the next crisis.

There is no romance in this. Local government is not pure, intimate and authentic while Westminster is distant, theatrical and false. Councils can be petty, bureaucratic, evasive and maddening. Local politics has its own vanities, its own cowardice, its own little empires. Anyone inclined to sentimentalise local democracy should attend a planning committee, a budget meeting or a residents’ forum about parking.

But the absence of romance is part of the point. Local government is political because maintenance is political. Care is political. Roads are political. Planning is political. Buses are political. Libraries are political. Housing is political. The weed-covered pavement is political. The unanswered email is political. The missed bin is political.

Not because every irritation deserves a grand ideological explanation, but because ordinary public life depends on institutions that either work or do not.

Westminster writes the grand arguments. Councils inherit the consequences. Politics becomes real when someone has to decide which road gets repaired, which care package is funded, which library survives, which bus route matters, which family is housed, which development is approved, and which neglected neighbourhood is told, once again, to wait.

The state is not an abstraction when your bin has not been emptied. It is at the end of your drive, smelling faintly, waiting for someone to take responsibility.

If the state had better information, perhaps welfare could be distributed more fairly. If platforms knew more about their users, perhaps children could be better protected online. If employers, landlords, regulators, journalists and voters could verify more claims more quickly, perhaps fraud, abuse, corruption and dishonesty would become harder to hide. If identity were easier to prove, perhaps access to services would be smoother, public administration would become less wasteful, and the dishonest minority who exploit gaps in the system would find those gaps closing around them.

It’s not hard to see why this story appeals. Some of the things said in defence of greater transparency and stronger identity systems are not merely plausible, but true. Fraud does exist. Children are harmed online. Public trust is damaged when powerful people lie without consequence. The internet has created whole shadow economies of impersonation, manipulation, harassment and scam artistry. The state already holds huge quantities of information about us, often badly joined up, badly secured, badly explained and badly used. Against that background, the promise of a cleaner, safer, more accountable digital order has obvious moral force.

That’s why the weaker civil-liberties argument is not enough. It’s easy to oppose surveillance by saying it won’t work, or that it will be abused, or that the claimed benefits will turn out to be exaggerated.

Often that’s true. Governments overpromise. Technology vendors overpromise more. Databases rot. Errors proliferate. Outsourced systems find bold new ways to be expensive and useless at the same time, which is admittedly a sort of technical achievement. But the more interesting question begins when we grant the premise.

Imagine, then, that the government legislates to abolish privacy as a social institution. Not merely to create a new ID card, or to require age checks for certain websites, or to join up public records more efficiently, but to create the logical endpoint of total transparency: a free-to-use, publicly accessible database containing all information about everything and everyone.

And suppose, for the sake of argument, that it works.

Social welfare improves enormously. Benefits are distributed more fairly because need is visible and fraud is harder to conceal. Journalistic and political integrity improve because public claims can be checked instantly against public facts. Crime rates fall. Dangerous people are easier to identify. Corrupt officials are easier to expose. Hidden conflicts of interest become public knowledge. The vulnerable are easier to find and support. The dishonest are easier to catch. The evasive are easier to pin down. The whole system becomes cleaner, safer, more efficient and more truthful.

The question is whether such a society would be better.

My answer is no. Or, more precisely, it might be better at some things, while being worse at the thing that makes political and social life worth having in the first place: the ability of a human being to exist as something more than a public file.

The problem with total transparency is not only that it might fail. The problem is that, even when it succeeds, it succeeds by destroying the private sphere. It gives us safety by making everyone searchable. It gives us integrity by making everyone permanently auditable. It gives us efficiency by making everyone administratively legible. It doesn’t merely reveal facts. It changes the conditions under which people live, think, experiment, associate, dissent, mature, recover and become themselves.

That is why the usual framing of privacy as a trade-off between security and secrecy is too thin. The real question is not whether we have something to hide. The real question is whether we’re still allowed to have a door.

The seduction of transparency

There’s a reason transparency has become one of the great moral words of modern politics. It sounds clean. It sounds democratic. It suggests sunlight, accountability, open records, honest administration and the end of cosy deals done behind oak doors by people with hereditary furniture. In that sense, transparency is not the enemy. Much of modern democratic life depends on it.

We should know who funds political parties. We should know what ministers have declared as interests. We should know how public money is spent. We should know whether a company dumping waste into a river has also been lobbying the regulator meant to stop it doing precisely that. A great deal of corruption survives because relevant information is hidden, scattered, inaccessible, expensive to obtain, or technically public but practically impossible to find without a law degree, three free afternoons and the personality of a terrier.

So the argument against total transparency cannot be an argument for total opacity. That would be absurd. Secrecy can protect the powerful as easily as the vulnerable; indeed, it usually protects them better, because the powerful can afford lawyers, shell companies, non-disclosure agreements, reputation management consultants and the sort of filing practices that make accountability feel like archaeology.

The question, then, is not whether transparency is good. The question is where transparency belongs, who it is for, what it is meant to expose, and what kinds of life it makes possible or impossible.

That distinction is crucial because transparency does not mean the same thing when applied upwards as when applied downwards. Transparency directed at powerful institutions can be a democratic tool. Transparency directed at ordinary individuals can become a disciplinary one. There is a difference between making the Home Office explain its decision-making and making every citizen’s medical history, romantic life, religious doubts, income, debts, mistakes, complaints, searches, friendships and political affiliations visible to anyone with an internet connection and a slow afternoon.

A society in which the powerful are accountable to the public is not the same as a society in which every person is accountable to everyone else at all times. The first can be democratic. The second is not democracy at all. It’s exposure.

The crowd is not a court

The internet has already given us a partial preview of what happens when private information becomes easy to find, easy to copy, easy to detach from context and easy to weaponise.

Some of this has been genuinely emancipatory. People once silenced by institutions can now speak around them. Abuse once concealed within families, workplaces, churches, schools, parties, clubs and industries can now be documented, shared and believed. The old gatekeepers were not neutral guardians of truth; they were often bouncers for reputation, class, money and institutional convenience. Anyone defending privacy has to take this seriously, because privacy has also been used as a polite name for burying harm.

But there’s a difference between enabling victims to speak and turning public accusation into public punishment without procedure. Social media does not need to wait for evidence, relevance, proportionality, context, right of reply, appeal, limitation periods, or any of the other boring procedural devices by which civilised societies try to distinguish judgement from appetite. It moves quickly, often thrillingly so. A name appears. A screenshot circulates. A thread begins. Someone adds a second-hand story, then a third-hand interpretation of the second-hand story, then someone with a podcast voice and a ring light announces that the silence from the accused is itself revealing. By lunchtime, the crowd has become detective, prosecutor, jury, judge and public executioner, although naturally it still has strong feelings about institutional overreach.

There are cases where public anger exposes what formal systems refused to see. That’s the strongest version of the defence, and it should not be dismissed. Courts fail. Police fail. HR departments fail with the dead-eyed fluency of institutions that have discovered the passive voice. But the failure of formal process doesn’t mean the crowd is a better process. It may be faster, louder and more emotionally satisfying, but speed, volume and satisfaction are not substitutes for justice.

The danger here is sometimes described as the tyranny of the majority, but even that phrase may be too organised. A majority at least implies a polity, a defined body, a decision rule, a mandate. The online crowd has none of that. It does not need to explain who belongs to it, what authority it claims, what standard of proof it applies, what punishment would be proportionate, when punishment should end, or what it would take to be forgiven. Its power comes from aggregation rather than legitimacy. Enough people looking at you can become a form of government, even if none of them individually thinks they’re governing.

Total transparency would make this dynamic permanent. It would not merely let the public judge wrongdoing. It would supply the public with an infinite archive from which wrongdoing, or something close enough to wrongdoing for present purposes, could always be assembled.

That is not safety. It is exposure to unlimited retrospective prosecution by people who don’t have to tell you what law they’re applying.

The state is not a saint

The opposing danger comes from the state, and here the usual libertarian warnings are often stated too crudely but not entirely wrongly.

There’s an old line attributed to Cardinal Richelieu: “If you give me six lines written by the hand of the most honest of men, I will find something in them which will hang him.” It survives because it names something unpleasantly durable about power. Give an institution enough information about a person, enough time to search through it, and enough motive to harm them, and innocence becomes beside the point. The question is no longer whether someone has done anything seriously wrong. The question is whether something can be found, isolated, reclassified, exaggerated, misread, or placed before the wrong audience.

The problem is not that every government is tyrannical. Most aren’t, at least not most of the time. The problem is that states are continuous in ways citizens are not. Administrations change. Ministers change. Laws change. Political moods change. Databases remain. Data collected for one purpose can be reused for another. Powers created under one moral emergency can be inherited by people with different enemies. A system built to find benefit fraud can be used to find dissidents. A system built to verify age can become a system for routine identity checking. A system built to protect the vulnerable can become a system for classifying the inconvenient.

This is why good intentions are not enough. The question is not whether the present minister sounds reasonable in a consultation document. The question is what structure is being built, what it permits, what it normalises, what safeguards exist beyond ministerial reassurance, and what a future government could do with the same machinery under different political conditions.

A benevolent database is still a database. A humane surveillance system is still a surveillance system. A well-administered architecture of total visibility is still an architecture of total visibility. Its moral character cannot be judged only by the intentions of the people who first switch it on.

There are, of course, practical objections too. Public records contain errors. People share names. Addresses go out of date. Relationships are complicated. Data taken from one context is often misleading in another. But, again, the thought experiment asks us to go further. Suppose the system is accurate. Suppose it catches the right people. Suppose it works exactly as advertised.

Even then, the citizen has been converted into a permanently inspectable object.

That conversion is not a side effect. It is the point.

Privacy is not secrecy

The laziest argument against privacy is that only people with something to hide need it. This is rhetorically effective because it makes privacy sound guilty before the argument has even begun. It quietly replaces the right to a private sphere with the right to conceal evidence, and then invites us to be suspicious of anyone who objects.

But privacy is not the same thing as secrecy.

Secrecy is about the content of what’s hidden. Privacy is about the existence of a sphere in which hiding is possible. That distinction, made powerfully by Emilio Mordini in his work on biometrics and the private sphere, changes the whole argument. The point is not that everyone has a scandal in the cupboard. The point is that a human life requires cupboards.

This may sound melodramatic until one thinks about the ordinary architecture of social life. We already live through forms of selective disclosure. We don’t present ourselves to our parents exactly as we do to our friends, or our colleagues, our partners, our doctors, or at the pub, at the polling station, in the group chat, the confessional, the therapist’s office, the classroom, the union meeting, the job interview, or the moment when we idly search for something odd on the internet and hope, for a moment, that no one is watching. We present as different people when given different audiences or contexts.

This is not hypocrisy. It is personhood.

The same fact can mean different things in different contexts because the relationship, purpose and audience are different. A medical diagnosis belongs in a medical context. A political opinion may belong in a meeting, a notebook, a ballot box, a conversation, an essay, or nowhere at all. A sexual preference may belong between consenting adults and no one else. A debt may be relevant to a lender but not to a neighbour. A past mistake may be relevant to safeguarding in one setting and wholly irrelevant in another. To collapse all of those contexts into one public database is not to reveal the truth of a person. It is to destroy the conditions under which truth can be interpreted properly.

This is where Helen Nissenbaum’s idea of contextual integrity is useful. Privacy is not simply violated when information becomes known; it’s violated when information moves in ways that break the norms of the context in which it was disclosed, produced, or held. The fact that a GP knows something doesn’t mean an employer should know it. The fact that a tax authority knows something doesn’t mean a journalist should know it. The fact that a friend knows something doesn’t mean a stranger should be able to search it.

The transparent society pretends that information is morally neutral once it’s accurate. It isn’t. Information has context, and when context is stripped away, accuracy can become a form of falsehood.

The administrative person

Total transparency would also change the kind of person the system expects us to be.

A private person is unfinished. They can think things they don’t yet believe, try on opinions they later reject, make mistakes without those mistakes becoming permanent public property, recover from humiliation, change affiliations, leave groups, enter relationships, exit relationships, experiment with identity, abandon foolishness, learn, contradict themselves, apologise, mature and quietly become someone else.

An administratively legible person is different. They are a record. They are a profile. They are a searchable sequence of attributes, transactions, affiliations, associations, risks, flags, scores, statements and deviations. They may still have an inner life, but the public system has less reason to care about it, because the file is easier to process than the person.

This is one reason the nothing-to-hide argument is so inadequate. It imagines privacy as a shield for incriminating content, when privacy is also the condition under which non-incriminating life can remain human. Most of what we wish to keep private is not criminal. Much of it is not even shameful. It is private because it’s fragile, unfinished, intimate, experimental, context-dependent, or simply nobody else’s business unless we choose to make it so.

The right to privacy protects the awkward draft of the self.

A society without that right doesn’t merely expose wrongdoing. It changes incentives. People become more cautious, more performative, more risk-averse, more conformist. The issue is not that everyone would become obedient overnight, because human beings are stranger and more stubborn than that. The issue is that everyone would learn to live before an imagined tribunal. Every search, joke, friendship, message, purchase, diagnosis, meeting, donation, doubt and desire would carry the question: how will this look when retrieved later?

That question is not morally neutral. It colonises the mind. It turns life into reputation management.

And although people sometimes say that the innocent have nothing to fear, innocence has never protected people equally. Those with conventional lives, secure jobs, respectable accents, supportive families, boring hobbies and the good sense to have been born into socially approved categories will always experience transparency differently from those whose lives are messier, poorer, queerer, more marginal, more stigmatised, more politically exposed, or simply less legible to the people doing the judging. Total transparency doesn’t flatten power. It gives power more material to work with.

The policy temptation

None of this means every digital identity proposal is tyranny, or that every form of age assurance is the first brick in the tower of Foucault’s panopticon. That would be too easy, and too silly. States need to know some things. Institutions need ways of verifying identity. Children do need protection online. Welfare systems do need to reduce fraud. Public services can be improved when records are accurate, portable and sensibly joined up. There are privacy-preserving technical designs that may answer some real problems better than the current mess of passwords, photocopied documents, insecure uploads and proving one’s existence to five different agencies in five different ways.

But this is precisely why the conceptual argument is important. The danger often doesn’t arrive as a villain stroking a cat and announcing the end of liberty. It arrives as an interface improvement, a safeguarding measure, fraud prevention, convenience, verification, service integration, child protection, risk management, public confidence and administrative efficiency. Sometimes these are genuine goods. Sometimes they’re even urgent goods. But none of them removes the need to ask what kind of citizen is being designed into the system.

The live debate about digital ID and online regulation is not identical to the thought experiment above. It would be unfair to pretend that a phone-based identity credential, or a requirement for age checks on ‘high-risk’ platforms, is the same thing as a universal public database. The serious argument is not that we’re already living in the hypothetical. It is that the hypothetical reveals the direction of travel we should be wary of: a society in which access to ordinary life increasingly depends on being identifiable, verifiable, classifiable and traceable.

That may be acceptable in some contexts. It may be necessary in others. But it should never be treated as cost-free simply because the interface is neat and the stated aim is benevolent.

The key questions are not only technical. They’re constitutional, moral and social. Who holds the data? Who can see it? Who can combine it? Who can compel it? Who audits the system? Who can refuse? What happens to people without smartphones, documents, stable addresses, conventional family structures, or trust in the institutions demanding proof? What information is disclosed to the service being accessed? Can identity be verified without identity being exposed? Can age be proved without a permanent record of the attempt? Can the state design systems that know only what they need to know, rather than everything they can know?

These questions are often treated as implementation details. They aren’t. They are the substance of liberty under digital conditions.

The right to close a door

The private sphere is not a sanctuary for wickedness. Or, at least, it is not only that, and no serious defence of privacy should pretend otherwise. Bad things happen behind closed doors. Secrecy protects abusers, fraudsters, hypocrites and cowards. Privacy can be invoked cynically by people who don’t want accountability, and any argument for privacy must face that fact squarely.

But a right can be abused without becoming worthless. Speech can be used to lie. Association can be used to conspire. Due process can be used by the guilty. Political freedom can be used by fools. The answer is not to abolish the right, but to understand why the right exists and where its limits should fall.

Privacy exists because a human being is not public property.

That is the point I keep returning to. We can concede almost every practical benefit claimed for transparency and still reject the transparent society, because the deepest case against it is not practical but civic. Total transparency would not merely help institutions know more about us. It would change our relation to institutions, to one another, and to ourselves. It would make life permanently inspectable. It would make context fragile. It would make error indelible. It would make identity less something lived and more something retrieved.

The point of privacy is not that every room contains a scandal. Most don’t. Some contain grief. Some contain illness. Some contain fantasy, prayer, boredom, sex, resentment, doubt, stupidity, recovery, unfinished politics, bad poems, worse search histories, and the small, undignified machinery by which people remain people. Some contain nothing very interesting at all.

That doesn’t weaken the case for privacy. It is the case for privacy.

A person should not need to prove that the room contains something important before being allowed to close the door. The door is the point. The power to close it is part of what separates a citizen from an object of administration, a neighbour from a suspect, a person from a profile.

A safer society may require some information to be shared. A fairer society may require some secrecy to be pierced. A more accountable society may require some doors, especially the doors of the powerful, to be opened more often and more forcefully than they are now. But a society with no doors at all would not be an honest society. It would be a society in which honesty had been confused with exposure, and exposure had been confused with justice.

The right to privacy is the right to remain partly unknown.

It is the right to be more than the sum of one’s records.

It is the right, sometimes, to disappear.

There are few phrases in politics more soothing than listening to concerns. It has the soft upholstery of democratic virtue. It suggests patience, humility, rootedness, a willingness to get out of Westminster and discover that the country is not entirely composed of select committee witnesses, policy officers and people whose job title contains the word strategy.

And, to be fair, listening is good. It is usually better than not listening, which remains a surprisingly popular option. A politics that cannot hear anxiety, irritation, fear, resentment, confusion or grief is not a serious democratic politics. It is either a seminar pretending to be a movement, or a management consultancy with a flag outside.

The problem is what happens next.

Subscribe now

In contemporary politics, listening to concerns often performs a small act of moral laundering. A concern is heard. Because it has been heard, it is treated as authentic. Because it is authentic, it is treated as legitimate. Because it is legitimate, it is treated as a mandate. And before long, a voter’s anxious sentence on a doorstep has somehow become a policy position, a press line, a rebellion against the courts, a tightening of eligibility criteria, a crackdown on protest, or a promise to do something drastic at the border.

This is not listening. It is ventriloquism with polling.

The problem with listening to concerns is not the listening. Politics would be intolerable without it. The problem is the way the phrase collapses several different acts into one: hearing what someone feels; understanding what they mean; assessing whether their account of the world is accurate; deciding whether their proposed remedy is workable; and judging whether it is morally acceptable.

Those are not the same thing. A concern may be sincere and still misdirected. It may be understandable and still unfair. It may point to a real pressure while blaming the wrong people for it. It may be politically salient without being morally sovereign. Listening is diagnostic, not legislative.

The concern is not always the argument

A useful starting point is to concede the obvious: public concern is not imaginary.

People are worried about the cost of living, the NHS, immigration, crime, housing, welfare, social cohesion, protest, speech, identity, services and the basic competence of the state. These are not inventions of focus groups or tabloid editors, even when focus groups and tabloid editors do their best to make them worse. In April 2026, the Office for National Statistics found that the most commonly reported important issues facing the UK were the cost of living, the NHS and the economy, with crime, immigration, housing and climate change also selected by large numbers of adults. Among those aged 70 and over, immigration was one of the three most commonly selected issues.

That matters because a certain kind of liberal or left-wing politics sometimes talks as if the public only believe troubling things because someone has maliciously inserted them into their heads. That is comforting, but too neat. People can arrive at bad conclusions from real experiences. They can live in neighbourhoods where public services are visibly strained, where GP appointments are hard to get, where housing is expensive, where anti-social behaviour feels uncontrolled, where local civic life has thinned out, and where change has happened faster than anyone has bothered to explain.

None of that means their diagnosis is right. But it does mean their anxiety is not simply a failure of moral education.

This is where the first misuse of listening begins. Parts of the Left have developed a habit of praising listening in the abstract while becoming terribly awkward when people say things outside the approved moral vocabulary. Immigration is the clearest example, though not the only one. Patriotism, welfare conditionality, crime, disorder and protest can produce the same evasive shuffle: yes, of course we must listen, but perhaps not quite to that, not in those words, not from those people, not without first translating it into something more congenial.

The result is often a politics of polite non-engagement. The voter is heard, thanked, perhaps emotionally validated, and then quietly returned to the category of people who have not yet understood the issue correctly.

There is a moral reason for some of this discomfort. Not every concern should be indulged. Some are racist, cruel, paranoid, conspiratorial, or based on displaced anger towards people with less power. But it does not follow that every uncomfortable concern is merely bigotry wearing a local accent. The more difficult possibility is that a voter may be both morally wrong in part of their account and politically right that something is failing.

A good politics needs to be able to say both things at once.

Listening as evasion

The Labour Party’s recent difficulty on immigration sits in precisely this space. The party has spent years trying to distinguish reasonable concern about control, pressure and integration from the nastier politics of scapegoating. That distinction is necessary. It is also very hard to maintain if voters think you only discovered the first half of the sentence after Reform started taking votes off you.

Keir Starmer’s 2025 immigration White Paper remarks promised tighter controls across work, family and study routes, tougher enforcement, and a reduction in migration. The speech also linked immigration to pressure on housing and public services. That is, in itself, not an illegitimate argument. The state is allowed to decide who enters and settles. Public consent matters. Immigration policy cannot be detached from housing, wages, infrastructure, language, exploitation, planning, skills, enforcement or local capacity.

But the political risk is obvious. If a party moves from discomfort to panic, it stops sounding like it has thought its way to a position. It sounds as if it has been chased there.

That matters because listening is not the same as discovering where the anger is loudest and standing next to it. When politicians say they are listening, they need to be able to show what they have heard besides the volume. Is the concern about numbers, pace, fairness, illegality, culture, wages, housing, public services, language, national identity, exploitation, visible disorder, or a general sense that the state has lost control? These are related concerns, but they are not identical. Different problems require different answers.

If a person says they are worried about immigration because they cannot get a GP appointment, the serious response is not simply to promise lower migration. It is to ask whether the pressure is caused by population growth, workforce shortages, estate capacity, funding allocation, appointment systems, poor planning, or the fact that the NHS itself depends heavily on migrant labour. If a person says their town has changed beyond recognition, the serious response is not to pretend that social trust is irrelevant, nor to pretend that migrants are the whole story. It is to ask what institutions, rituals, services and shared spaces might help people live together well.

That is slower than a slogan. This is one reason slogans are popular.

Listening as surrender

The Right has a different problem. It is often much better at hearing the emotional content of public concern and much worse at resisting the temptation to sanctify it.

This is the most dangerous use of listening to concerns: the conversion of anxiety into permission. The public are worried about immigration; therefore we should leave the European Convention on Human Rights. The public are angry about small boats; therefore we should detain and deport everyone who arrives irregularly. The public dislike disruptive protests; therefore the state needs more powers to restrict them. The public think welfare is too expensive; therefore disabled people’s support becomes a suitable theatre for fiscal toughness.

This is not democratic humility. It is moral abdication.

Take immigration. Reform UK says it would leave the ECHR, intercept and detain all irregular arrivals, deport them, and withdraw from or disapply every international treaty that prevents its border policy. The Conservative Party has also made leaving the ECHR official policy, arguing that doing so would help secure borders and prevent legal challenges in other areas.

One can acknowledge that the asylum system is a mess, that Channel crossings are dangerous, that enforcement has been weak, and that voters are entitled to expect a functioning border. None of that proves that leaving the ECHR is wise, necessary, proportionate, operationally effective, or compatible with Britain’s longer-term interests. The concern is real. The proposed remedy is another question entirely.

That distinction is the thing political language keeps trying to erase.

The Right’s rhetorical move is often to treat restraint as contempt. If you do not endorse the maximal version of the answer, you have not listened. If you mention courts, treaties, administrative capacity, costs, family life, labour markets, due process, diplomatic agreements, returns arrangements, refugee law, Northern Ireland, or basic decency, you are accused of ignoring ordinary people. Expertise becomes evasion. Law becomes sabotage. Compassion becomes weakness. Complexity becomes betrayal.

But a government is not a comments section with a civil service.

Its job is not to reproduce public anger in statutory form. Its job is to govern through the tension between consent, evidence, principle and consequence. Sometimes that means telling people that a concern is being taken seriously. Sometimes it means telling them that the thing they want cannot be done. Sometimes it means saying, more bluntly, that it should not be done.

The voter is not a policy machine

There is a related mistake here: treating voters as if they possess stable, internally coherent policy programmes.

They usually do not. This is not because voters are stupid. It is because most people are not policy professionals and, frankly, are better adjusted for it. People hold moods, instincts, values, memories, interests, impressions, resentments and fragments of information. They also hold contradictions. They may want lower taxes and better public services; tougher welfare rules and protection for genuinely disabled people; lower immigration and more staff in hospitals and care homes; more police powers and protection from state overreach; cheaper housing and no building near them; more growth and less disruption.

This is not hypocrisy in any simple sense. It is ordinary democratic life.

British Social Attitudes is useful here because it places public opinion in longer context rather than treating every shift in salience as a new national soul. Its 2026 first look covers taxation and spending, welfare, immigration, electoral reform and assisted dying, drawing on time series that in some cases stretch back to 1983. That kind of evidence matters because it reminds us that the public mood is not a single instruction. It is a weather system.

Welfare is a good example. It is very easy to get people to agree that the welfare system should be fair, that people who can work should work, that fraud should be prevented, and that public spending should be controlled. It is much harder to turn those sentiments into humane and administrable policy. The government’s own assessment of its 2025 health and disability benefit reforms projected £4.8 billion of savings by 2029-30 and a reduction in the Personal Independence Payment caseload of 400,000, while still forecasting that the overall caseload would grow substantially. YouGov polling around the same period found that most Britons opposed cuts to disability benefits, especially for those unable to work.

This is the public in all its inconvenient complexity. People can want welfare reform and oppose welfare cuts. They can think the system is abused and still recoil when the person losing support has a degenerative condition, a severe mental illness, chronic pain, or a caring responsibility. They can endorse toughness in the abstract and mercy in the particular.

A politics that genuinely listens has to notice the contradiction rather than exploit whichever half is useful this week.

The protest problem

The same pattern appears around protest.

It is not hard to understand why disruptive protest annoys people. Blocking roads, targeting public events, interrupting commuters or disrupting city centres will generate anger, and not only among reactionaries. People have jobs, hospital appointments, school runs and lives not designed around becoming unwilling participants in someone else’s moral emergency.

But here too the concern does not settle the question. A right to protest that exists only when it inconveniences nobody is not a right to protest. It is permission to hold an outdoor meeting.

Recent legislation has significantly expanded the state’s ability to intervene in disruptive protest. The House of Commons Library notes that the Police, Crime, Sentencing and Courts Act 2022 strengthened police powers to place restrictions on protests, while the Public Order Act 2023 created several protest-related offences and Serious Disruption Prevention Orders. It also notes that the Crime and Policing Bill 2024-26 would introduce further restrictions, including around pyrotechnics, climbing war memorials, concealing identity, protests near places of worship, and cumulative disruption.

There are real balances to strike here. Public order is not imaginary. Nor are intimidation, harassment or targeted disruption. But when governments listen only to irritation, they miss the point of the liberty at stake. Protest is supposed to impose itself on public attention. That does not mean every tactic is justified. It does mean disruption cannot be treated as a self-evident reason for suppression.

A serious politics would ask: what kind of disruption, by whom, against whom, how often, with what risk, under what safeguards, and with what alternative means of democratic expression available? A lazy politics asks whether people are annoyed and then calls that a mandate.

The difficulty of listening without laundering

Trans rights and single-sex services present perhaps the hardest version of this problem, because the language of concern is doing so much work at once.

Some concerns raised in this area are serious and should be taken seriously: privacy, safety, dignity, fairness in sport, safeguarding, the operation of refuges, prisons, schools and healthcare, and the practical duties placed on service providers. Some are not serious at all, except as moral panic. Some are formally about policy but emotionally about disgust. Some are expressed as concern for women while taking obvious pleasure in the humiliation of trans people. Some are expressed as concern for trans people while refusing to acknowledge that women’s rights, spaces and privacy are not trivial.

A politics that cannot tell the difference between these things will fail everyone.

The legal and policy landscape has also shifted. The Equality and Human Rights Commission says the Supreme Court’s 2025 ruling in For Women Scotland held that, for the purposes of the Equality Act 2010, sex means biological sex, and that the EHRC’s updated services code was laid before Parliament in May 2026. In June 2026, the government told the Commons that it would protect single-sex spaces based on biological sex where needed, while also saying that trans people should have access to services in a way that respects dignity and privacy.

That is the sort of sentence governments produce when they are trying to listen in several directions at once. Sometimes that is necessary. Sometimes it is evasive. The test is whether the difficult trade-offs are named clearly enough for anyone to know what is being decided.

On this issue, as on others, listening to concerns can easily become a way of avoiding moral responsibility. Politicians can imply sympathy with women without spelling out what happens to trans people in practice. Or they can imply solidarity with trans people without answering concrete questions about institutions where sex, privacy and vulnerability are central. Either move is easier than governing. Neither is good enough.

Listening should not mean treating every anxiety as prejudice. Nor should it mean treating every anxiety as policy.

The technocrat’s mistake

There is, however, one more failure of listening, and it is not confined to the populist Right or the squeamish Left.

It is the technocratic mistake: assuming that because you understand the evidence better, you understand the concern better.

Sometimes politicians, civil servants, experts and campaigners really do know things that the public does not. That is not an elitist scandal. It is the point of expertise. Most of us do not have time to understand asylum returns agreements, disability assessment criteria, police public order powers, local authority finance, labour market shortages, planning law, energy grids or the case law around the Equality Act. We outsource some of that complexity because otherwise we would all need to become full-time policy analysts and nobody would fix the boiler.

But expertise can become arrogant when it treats public concern merely as a communications problem. The voter is not simply a badly informed version of the expert. They may be describing something the data has missed, or something the data can see only indirectly: humiliation, distrust, fear, local decline, institutional contempt, the cumulative irritation of systems that technically exist but do not work when needed.

There is a reason people often say they want politicians to listen even when they cannot specify the policy answer. What they mean is that they want the political system to recognise the reality of their experience. They want some sign that the people making decisions have encountered the consequences of those decisions in human form.

That is not a small demand. But it is not the same as asking to rule by anecdote.

The task of democratic politics is to hold both truths together. The public may know things government has missed. The public may also be wrong. Expertise may correct prejudice, but it may also miss lived reality. Evidence should shape policy, but it cannot substitute for legitimacy. Listening should inform judgement, not replace it.

What serious listening would require

A serious politics of listening would be less flattering to everyone.

It would be less flattering to voters, because it would refuse to pretend that every concern is wise, fair or usable in the form in which it first appears. It would say: I have heard you, but I do not accept your explanation. Or: I agree this is a problem, but your answer would harm the wrong people. Or: what you are describing is real, but it is caused by housing policy, wage stagnation, service failure, enforcement weakness, planning delay, employer exploitation, institutional neglect, or the simple fact that the state has promised more than it can competently deliver.

It would be less flattering to politicians, because they would no longer be able to use listening as a substitute for courage. They would have to explain why they had accepted some concerns, rejected others, translated others into different policy terms, and drawn moral boundaries around the possible. They would have to stop hiding behind the public when making choices they wanted to make anyway.

It would be less flattering to activists and commentators, because it would deny them the comfort of assuming that everyone outside their moral universe is either wicked or stupid. It would require them to ask why certain arguments have purchase, what social reality they attach to, and whether their own side has avoided questions that do not fit easily into its preferred story.

And it would be less flattering to technocrats, because it would remind them that being right on the evidence does not automatically mean being democratically adequate.

Listening properly means breaking a concern into parts. What is the person afraid of? What have they actually experienced? What have they inferred from that experience? Who are they blaming? What evidence supports or undermines that blame? What would their preferred solution do? Who would carry its costs? Which moral lines are not available to cross, even if the policy would poll well?

That is slow work. It is also what politics is for.

The alternative is the little fraud we have now: listening as theatre, listening as permission, listening as triangulation, listening as focus-group ventriloquism, listening as the soft prelude to a hard policy already chosen.

Listening is not obedience. Nor is it a fig-leaf for contempt. It is the first stage of democratic judgement, not the suspension of judgement altogether.

A serious politics would listen to concerns without worshipping them; challenge prejudice without pretending the underlying anxiety is always invented; and respect voters enough not to treat every doorstep sentence as either sacred wisdom or regrettable ignorance.

The public are not always right. Neither are the people paid to interpret them.

That is exactly why listening has to mean something more than nodding along.

There are few phrases more comforting than being on the right side of history.

It has the satisfying sound of an argument that has already won. It does not merely say that a person is right, or that their cause is just, or that their opponents are mistaken. It says something grander and more convenient: that the future has already agreed with them. History, that vast and impressive abstraction, has supposedly filed its judgement in advance. The present may still be noisy, complicated and resistant, but the verdict is coming. We need only wait for the dust to settle and the statues to be rearranged.

This is, as rhetoric, extremely tempting. It is tempting because it borrows the authority of hindsight while sparing us the inconvenience of not yet having any. It allows a person to speak as if the moral meaning of the present has already been clarified by the future. It gives political argument the flavour of prophecy. It is not just that I am right. It is that one day everyone will see that I was right, and those who failed to see it will look faintly embarrassing in the archive.

The trouble, of course, is that history has not yet turned up to give evidence.

This does not mean that the phrase is always empty. There are moments when it captures something real. Human beings have done things, tolerated things, defended things and built institutions around things that later generations have been right to judge harshly. Slavery, colonial violence, racial segregation, the exclusion of women from political life, the criminalisation of gay people, the maintenance of apartheid: nobody needs to pretend that all historical judgement is presumptuous just because some people use historical judgement badly. Moral progress is not an illusion. There really are victories worth defending, and there really are defeats that, in retrospect, deserve to look like defeats.

But that is not quite the same thing as saying they were inevitable.

This distinction matters. There is a difference between moral confidence and historical inevitability. Moral confidence says: this is right, and I am prepared to defend it. Historical inevitability says: this will win, and therefore opposition to it is not merely wrong but obsolete. The first still accepts the burden of argument. The second quietly tries to escape it.

That escape is the problem.

The phrase being on the right side of history offers people a borrowed certainty. It lets them occupy the posture of retrospective wisdom while the actual work remains unfinished. It turns politics into a queue for vindication, as though justice were not something made through struggle, organisation, persuasion, law, culture, institutional pressure, compromise, conflict, endurance and luck, but something that eventually arrives because the future is embarrassed on our behalf.

That is not only bad history. It is bad ethics. And, more immediately, it is often dreadful politics.

The temptation is understandable.

The phrase appeals most strongly to people who have reason to believe that the present is morally disordered. That should make us careful before mocking it. People do not usually reach for the language of history when they are calmly discussing the decimal places of a budget line. They reach for it when they believe something larger is at stake.

Abolitionists needed more than a careful reading of parliamentary arithmetic. Suffragists needed more than patience with the procedural objections of men who discovered, with remarkable consistency, that the time for women’s equality had not quite arrived. Civil rights campaigners needed to believe that segregation was not merely one legal arrangement among others, but a moral scandal. Anti-apartheid activists needed to insist that a racial caste state could not be made acceptable by calling it stability. Gay rights campaigners, in the face of criminalisation, stigma and state neglect, needed to believe that the moral universe was not exhausted by the prejudices of their own moment.

In these cases, the appeal to history can be sustaining. It says: this present arrangement is not permanent. The normality of injustice is not proof of its legitimacy. What looks settled can be unsettled. What looks natural can be exposed as political. What looks like common sense can, in time, come to look cruel.

There is a generous version of the phrase, then. It is not a sneer but a source of courage. It is the voice that tells people in losing causes that the present distribution of power is not the same thing as moral truth. It refuses the tyranny of the immediate. It says that majorities can be wrong, that laws can be unjust, that social consensus can be rotten, and that people who are despised in one generation may be honoured in another.

That is not a trivial insight. It is one of the reasons politics is possible at all.

But the problem with the phrase is that it so easily slides from courage into complacency. It begins as a refusal to let the present have the final word. It ends as a refusal to do the work of changing the present. The future becomes not a hope, but a guarantee. The person speaking is no longer making an argument; they are announcing the direction of travel.

And once someone believes they know the direction of travel, it becomes remarkably easy to treat anyone standing in the way as less a citizen than a historical inconvenience.

Moral confidence is not historical inevitability

The distinction between moral confidence and historical inevitability is the hinge of the whole problem.

Moral confidence can be necessary. There are issues on which endless neutrality becomes its own form of cowardice. To say that a person should defend their beliefs without any conviction would be absurd. Political and moral life requires commitments. It requires people to say, sometimes against the mood of their age, that something is unjust, degrading, dishonest or cruel.

But moral confidence still has to argue. It still has to explain itself. It still has to persuade, organise and contend with reality. It does not get to assume that the righteousness of its destination has solved the difficulty of the route.

Historical inevitability does something different. It turns a claim about justice into a claim about momentum. It implies that the issue is not simply whether something is right, but whether it is where history is heading. That gives the speaker a useful double advantage. They get to sound morally serious and strategically sophisticated at the same time. Their opponent is not merely wrong; they are behind. Their position is not merely unjust; it is ageing badly. Their arguments do not need to be answered so much as outlived.

This is a powerful form of rhetoric because nobody wants to imagine themselves as the last defender of some doomed stupidity. Most people prefer to think of themselves as decent, modern and reasonably alert to the moral weather. To tell them they are on the wrong side of history is to threaten them not only with defeat but with embarrassment. It asks them to picture their grandchildren wincing at them.

There are times when that threat may be deserved. But as an argumentative habit, it is dangerous. It rewards people for speaking as if history were simpler than it is. It treats the future as a kind of moral audit office. It encourages the thought that the direction of justice is obvious to those with sufficient virtue, and obscure only to the wicked, stupid or self-interested.

This is where the phrase becomes ethically corrosive. It does not merely express confidence. It licenses contempt.

The Whig comfort blanket

There is an old historical temptation here. Herbert Butterfield’s critique of Whig history was, among other things, a critique of stories that turn the past into a procession towards the enlightened present. In the Whig version of history, events are arranged so that they appear to be moving, however bumpily, towards us: our freedoms, our institutions, our moral assumptions, our self-congratulation. The American Historical Association’s discussion of Butterfield describes his concern with abridged historical narratives that achieve drama and apparent moral clarity by interpreting the past through present politics, producing teleological stories that can do violence to the complexity of the past.

The appeal of such stories is obvious. They tidy history up. They sort people into winners and losers, prophets and reactionaries, darkness and light. They allow us to see the past as a series of auditions for the present, with everyone judged by whether they anticipated us.

That does not mean we should suspend moral judgement. There is a lazy misuse of anti-presentism which treats any judgement of the past as an anachronism, as though the people harmed by historical injustice were not themselves present at the time. Enslaved people did not need twenty-first-century theory to understand that enslavement was a horror. Women excluded from public life did not need a modern seminar to notice that the arrangement was convenient for men. Colonised peoples were not waiting for later historians to tell them that coercion had occurred.

So the lesson is not that moral judgement is impossible. It is that moral judgement is not the same as historical neatness.

The past was not a morality play whose central purpose was to produce the opinions we currently hold. It was inhabited by people acting under conditions of uncertainty, constraint, imagination, prejudice, courage, fear, ignorance and interest. Sometimes the people we now admire were wrong about other things. Sometimes the people who defended good causes did so for mixed reasons. Sometimes movements that deserved to win lost for generations. Sometimes victories were partial, compromised or reversed. Sometimes what looked like inevitable progress depended on accidents, alliances, institutional openings and tactical decisions that could easily have gone differently.

This is why the right side of history is such a suspiciously comfortable phrase. It allows us to enjoy the moral clarity of hindsight in advance. It lets us imagine the future as if it were already a school textbook, with our side printed in the chapter heading.

But history is not written in advance. It is not a conveyor belt. It does not carry good causes to victory while their supporters stand around being correct.

The moral arc is not a travelator

The most famous version of this comfort is the line about the arc of the moral universe bending towards justice. Martin Luther King Jr. used a version of it, but the thought is usually traced back to the abolitionist minister Theodore Parker. Parker’s original formulation is more interesting, and more humble, than the slogan it has become. He did not present himself as a man who could see the whole design of history. He said, in effect, that his view was limited; he could not calculate the full curve, but by conscience he believed it bent towards justice.

That humility matters. Parker’s line was not a forecasting model. It was not a graph. It was not a complacent assurance that the future had a departmental commitment to fairness. It was an act of moral faith under conditions of partial sight.

The sloganised version often loses that uncertainty. In public rhetoric, the moral arc becomes less like a discipline and more like infrastructure. It bends towards justice, apparently, because that is what arcs do. The speaker is then relieved of the question of who bends it, how, at what cost, against which resistance, and with what possibility of failure.

But the moral arc is not a travelator. It does not move us while we stand still.

If it bends towards justice at all, it does so because people bend it. They bend it through argument, protest, litigation, writing, organising, voting, bargaining, striking, teaching, risking reputation, losing elections, building institutions, changing norms, and sometimes spending entire lives on causes whose victories they will not personally see. Even then, it does not bend smoothly. It buckles, stalls, snaps back. It can be forced in the other direction. Rights won in one generation can be weakened in another. Taboos can return. Cruelties can be renamed. Bad ideas can acquire new fonts and re-enter polite society as if they had merely been away at finishing school.

The belief that justice is possible is necessary. The belief that justice is inevitable is a narcotic.

What the phrase does to argument

The deeper problem with being on the right side of history is not that it is historically naive, though it often is. The deeper problem is what it does to argument.

First, it makes people lazy. If history is already moving in your direction, persuasion starts to look optional. Why explain yourself carefully to people who will eventually be ashamed of themselves anyway? Why address their fears, objections, confusions or material interests? Why build a majority when history has already seconded the motion?

This is a particular danger for progressive politics, though not only for progressive politics. A movement can be morally right and still politically incompetent. It can diagnose an injustice accurately and still fail to build the coalition needed to remedy it. It can have the better argument and still lose because it mistook moral force for organisational power. The fact that a cause deserves to win does not mean that it will. The universe does not give extra seats for being correct in principle.

Second, the phrase makes people self-righteous. It encourages a flattering distinction between those who understand the age and those who belong to its debris. The opponent becomes not someone to be persuaded, defeated, constrained or answered, but someone already sentenced by the future. This is emotionally satisfying, but satisfaction is not the same thing as seriousness.

There are, plainly, people whose views do not deserve endless gentle handling. Some beliefs are cruel. Some arguments are bad faith. Some political projects are organised around domination, exclusion or contempt. There is no duty to launder prejudice as mere difference of opinion.

But even then, the question remains: what is the task? If the task is moral denunciation, then denunciation may be justified. If the task is political change, then denunciation may not be enough. People do not stop having votes, money, institutional power, social influence or capacity for obstruction merely because history has been informed that they are in the wrong queue.

Third, the phrase makes people strategically stupid. It confuses losing an argument in the future with losing power in the present. This is a dangerous confusion because many unjust arrangements are perfectly capable of lasting a very long time. History contains quite a lot of things that should have ended earlier. Some of them did not end because people were too frightened to challenge them. Some did not end because too many people benefited from them. Some did not end because their opponents were divided, exhausted, excluded, imprisoned, murdered or ignored. Some did not end because nobody had yet built the machinery needed to end them.

To say that an injustice is doomed is not to make it so. Sometimes it is not doomed. Sometimes it is extremely well funded.

Finally, the phrase can make people morally careless. If you believe you are speaking for the future, you may become less attentive to the present. You may stop asking whether your own side has hidden premises, bad incentives, internal cruelties or convenient evasions. You may start treating tactical disagreement as moral betrayal. You may decide that anyone asking about means, trade-offs, implementation or unintended consequences is merely slowing the march of justice.

That is how moral certainty degenerates into moral vanity.

The serious version

There is, however, a serious version of the appeal to history, and it is worth preserving.

Adam Green, in an article for Episteme, argues that appeals to being on the right side of history can be understood not merely as rhetorical flourish but as a kind of higher-order evidence: a claim about how one’s view may stand in relation to broader patterns of moral learning and social judgement. He does not deny that the phrase can be used rhetorically, but he asks whether there is something more substantive going on when people invoke it.

That is a useful complication. We should not pretend that all references to future judgement are automatically stupid. Sometimes people are asking a real question: what will this look like when the panic has passed? What will this policy look like when the people harmed by it are finally heard? What will this compromise look like when its evasions are no longer convenient? How will later generations understand the excuses we currently find so persuasive?

Those are good questions. In fact, they are morally necessary questions. They ask us to step outside the suffocating immediacy of our own moment. They remind us that the present has no monopoly on judgement. They help us see that what feels moderate now may later look cowardly, and what feels extreme now may later look like basic decency arriving late.

The problem is not asking how the future might judge us. The problem is pretending to know.

There is a difference between using imagined future judgement as a discipline and using it as a weapon. As a discipline, it asks us to be less complacent. As a weapon, it allows us to become more so. It can make us more reflective, or it can make us insufferable. The difference lies in whether the thought of history increases our sense of responsibility or decreases it.

A good appeal to history says: I may be trapped inside the assumptions of my age, and I should test them. A bad appeal to history says: my opponents are trapped inside the assumptions of their age, whereas I have somehow escaped into permanent enlightenment.

The first is humility. The second is cosplay as posterity.

History will not do your canvassing

The political danger is clearest when the rhetoric of inevitability meets the ordinary mechanics of power.

History does not knock doors. It does not draft legislation. It does not build unions, parties, campaigns, charities, legal strategies, newspapers, schools, churches, tenants’ groups, mutual aid networks or professional associations. It does not persuade nervous voters, reassure conflicted allies, split opponents, count numbers, secure amendments, survive defeats or return the next morning to try again.

People do that. Institutions do that. Movements do that. And they do it imperfectly.

This is one reason the right side of history is such a seductive phrase for people who enjoy politics at the level of identity but not always at the level of work. It allows the performance of seriousness without the discipline of strategy. It says: we are the future. But the future, as campaign literature, is often overprinted and under-delivered.

The hard truth is that good causes lose. They lose all the time. They lose because they are crushed. They lose because they are divided. They lose because they are early. They lose because they are badly led. They lose because their opponents understand power better. They lose because the public is frightened, tired, misled or unconvinced. They lose because the institutional terrain is hostile. They lose because events intervene. They lose because history is not a referee.

This does not mean politics should become cynical. Quite the opposite. It means politics should become less mystical. If something matters, it deserves better than inevitability. It deserves the seriousness of means.

That means persuasion remains morally important. Not because every opponent is reasonable. Not because all disagreement is innocent. Not because the public must be flattered at all times. Persuasion matters because politics is the art of changing conditions among people who have not yet agreed with you. To abandon persuasion is not a sign of moral courage. It is often a confession that one has mistaken agreement for justice.

There will always be people who cannot be persuaded. There will always be people whose objections are dishonest or whose interests are too deeply tied to the injustice being challenged. But most political change depends on what happens around them: who is mobilised, who is reassured, who is split away, who is made less afraid, who is given language for something they already half-believe, who is shown that a different arrangement is possible without the world ending.

That is slow work. It is much less glamorous than being congratulated by the future. Unfortunately, it has the advantage of being real.

Against prophecy

The answer, then, is not to abandon moral confidence. That would leave politics to the people most comfortable with power as it already exists. The answer is to detach moral confidence from prophecy.

Say that something is right. Say that something is cruel. Say that something must change. Say that the excuses are inadequate, the settlement is unjust, the compromise has curdled, the respectable consensus is hiding something ugly. But do not pretend that history has already done the arguing for you.

The future may judge some of our present arrangements harshly. It may vindicate causes that now look marginal, excessive, naive or inconvenient. It may also expose some of our certainties as shallow, some of our fashions as evasions, some of our moral performances as vanity. That is the trouble with invoking posterity. It does not only judge the people we dislike.

To think historically is not to imagine a future that applauds us. It is to remember that people in the past also thought they were obvious, decent, modern and justified. Some of them were. Many were not. Most were tangled in ways that become clearer only when time has stripped away their excuses.

That should make us morally serious, not paralysed. It should make us willing to judge, but wary of enjoying judgement too much. It should make us suspicious of the rhetorical shortcut that turns our preferred future into destiny.

There is nothing wrong with wanting to be vindicated by history. Most of us would prefer not to be remembered, even hypothetically, as the person standing athwart justice asking whether the minutes of the last meeting had been properly approved. But the desire to be vindicated is not evidence that one is right. Nor is confidence in the future a substitute for responsibility in the present.

If a cause is just, it deserves argument. It deserves organisation. It deserves patience, courage, discipline, humility and strategy. It deserves people willing to persuade those who are not yet persuaded, and to defeat those who cannot be persuaded, without pretending that either task has already been completed by time.

History may one day take a side. But it will not spare us the work of taking one now.

Every few years, foreign policy discovers that it’s 1938. Again.

This is a surprisingly adaptable discovery: it can happen in Eastern Europe, the Middle East, the Taiwan Strait, the South China Sea, or any television studio where two people have been booked to disagree about a conflict that neither of them can easily explain. Sooner or later, someone says Munich. If things are going badly for a great power, maybe they say Suez instead. If the proposed military action looks open-ended, someone could say Vietnam. If it involves the humiliating limits of modern force, Afghanistan. If the speaker is feeling historically adventurous, they may reach for Gallipoli, the Crimean War, or the Peloponnesian War, at which point one begins to suspect they’ve been waiting a respectable span of time to use it.

The problem isn’t that these analogies are always wrong, per se. That would be much too easy. History does not repeat itself, but neither does it sit politely in the archive, refusing all invitations to the present. Politicians and commentators reach for the past because the present is often hard to see while one is standing in it. Analogy is one of the ways human beings think under pressure. It compresses uncertainty into shape, gives unfamiliar events a recognisable form, and lets us ask whether we’ve seen something like this before.

But the phrase something like this is doing a great deal of work.

The trouble with the easy analogy is not that it compares. It is that it often decides before it compares. It does not ask: what, exactly, is similar here? It asks: which famous historical costume can we dress this argument up in?

Munich is the costume of resolve. Suez is the costume of anti-hubris. The first lets the speaker become the person who saw appeasement for what it was. The second lets the speaker become the person who recognised imperial delusion before the empire did. Both can illuminate. Both can also flatter. And flattery, in foreign policy argument, is rarely harmless.

The uses of 1938

Munich survives because it tells a clean story about moral failure.

In September 1938, Britain and France accepted Germany’s annexation of the Sudetenland, the heavily fortified border region of Czechoslovakia, in an agreement reached with Germany and Italy but without Czechoslovakia itself as an equal participant. Hitler received territory and strategic advantage; Chamberlain returned claiming peace; within months, Germany had occupied the rest of the Czech lands. The later moral meaning of the episode seems brutally clear: give an expansionist aggressor what he wants, and he will not be satisfied. He will be encouraged.

That is why Munich is so powerful. It turns the fog of diplomacy into a moral tableau. The aggressor is insatiable. The victim is abandoned. The statesman of peace is exposed as the midwife of catastrophe. The person warning against compromise is vindicated by history before the argument has even begun.

It is not difficult to see why this appeals. There really are aggressors whose ambitions expand when met with weakness. There really are moments when compromise does not buy peace, but merely rearranges the timetable of war. The Munich analogy has force because there are situations in which refusing to confront an aggressor early makes the eventual confrontation more terrible.

Russia’s war against Ukraine is the obvious modern case where the analogy has real bite. Ukraine has repeatedly warned for over a decade against any settlement that rewards aggression or trades its sovereignty for temporary quiet; in December 2025, its foreign minister said Ukraine wanted real peace, not appeasement with Russia. The analogy is not empty. A sovereign state has been invaded. Territory is at stake. External powers debate the cost of continued support just as they debated the cost of sanctions way back during the initial invasion of Crimea in 2014. A peace that leaves the aggressor stronger may not be peace at all, but an interval.

And yet even here, perhaps especially here, Munich is not enough.

Ukraine is not Czechoslovakia in 1938. It has fought for years. Its government has agency of a kind that Edvard Beneš could only have dreamed of. Its armed forces are not being treated simply as a diplomatic inconvenience by others around a table. And yet Russia is a nuclear-armed state; Nazi Germany in 1938 was not. Almost perversely, the Russo-Ukrainian War could be more dangerous than the build-up to the outbreak of WWII. The structure of alliance politics, the economics of sanctions, the role of drones, the informational battlefield, the problem of escalation, and the status of NATO all make the present conflict different in ways that matter enormously. Using the Munich analogy requires treating the shared features as decisive while treating the differences as secondary.

Munich can sharpen the moral stakes of Ukraine. It can remind us that aggressive war cannot be made acceptable by relabelling conquest as compromise. But if it becomes the whole argument, it starts to crowd out the questions that policy actually needs to answer. What guarantees would make a settlement durable? What level of support is sustainable? What risks of escalation are tolerable? What capacities does Europe need if America becomes unreliable? What outcome is being sought beyond the avoidance of shame?

A historical analogy is useful when it helps us ask better questions. It becomes dangerous when it congratulates us for already knowing the answers.

The Taiwan temptation

The Munich analogy also spins its helicopter-blades over Taiwan.

Here again, the attraction is obvious. China claims Taiwan as its own territory; Taiwan rejects Beijing’s sovereignty claim and says only its people can decide their future. Beijing has increased military pressure around Taiwan in recent years, and Reuters reported again in June 2026 on a standoff between Taiwanese and Chinese coast guards near the Pratas Islands. There is a democratic polity under pressure from an authoritarian great power. There is an anxious question about whether distant allies will act. There is a fear that ambiguity, caution, or commercial self-interest could be misread as weakness.

It is therefore unsurprising that commentators draw lessons from Munich for the Taiwan Strait. Some explicitly argue that Munich offers warnings about deterrence, alliance credibility, and the risks of allowing coercive territorial revisionism to succeed.

Again, not nonsense. But also not enough.

Taiwan is not the Sudetenland. China is not Nazi Germany. The United States is not Britain in 1938. Taiwan’s importance is not only moral or ideological; it is also bound up with global semiconductor supply chains, maritime security, Japanese security, American credibility, and the possibility of catastrophic escalation between nuclear powers. The logic of deterrence in the Taiwan Strait is not the same as the logic of appeasement in Central Europe. One can think Beijing is dangerous without needing to pretend Xi Jinping is Hitler. In fact, insisting on that comparison may make it harder to understand what China actually wants, how it calculates risk, and which forms of deterrence might work.

Munich is good at saying: do not abandon a threatened democracy because doing so feels convenient.

It is much less good at saying: how should an allied power combine deterrence, ambiguity, trade, technology controls, diplomacy, domestic resilience, and military readiness in a contested maritime theatre involving nuclear-armed states?

That second question is less satisfying. Unfortunately, it is also the one that matters.

Suez and the pleasure of spotting decline

If Munich is the analogy of moral cowardice, Suez is the analogy of imperial delusion.

In 1956, Britain and France, in collaboration with Israel, invaded Egypt after Gamal Abdel Nasser nationalised the Suez Canal. Militarily, the operation was a success. Politically, it was a disaster. The United States opposed the intervention; the Soviet Union applied pressure; Britain and France were forced into a humiliating withdrawal. Suez became the shorthand for a country discovering, in public, that its memory of power had outlived the thing itself.

Suez is therefore useful whenever a state mistakes possession of weapons for possession of strategy. It is useful whenever a government thinks a dramatic act will restore authority, only to reveal weakness. It is useful whenever leaders confuse nostalgia with capacity. For Britain, the analogy remains especially painful because it says something we half-know and half-resent: that the performance of world power can become the evidence of its absence.

But Suez, too, flatters the speaker. To invoke it is to stand above the action, coolly noting that the empire has no clothes. This is one reason the phrase Suez moment appears so regularly in commentary about American power. If Munich lets one pose as Churchill, Suez lets one pose as the weary realist in the back of the room, watching the map of the world shrink.

The current US-Iran conflict makes the Suez analogy almost irresistible. There are maritime chokepoints, energy markets, coercive diplomacy, alliance entanglements, regional escalation, and a question about whether military force is producing strategy or merely motion. Reuters has reported in recent days on US action against Iranian-linked shipping, tensions around the Strait of Hormuz, and negotiations intended to extend or stabilise a ceasefire. It is hardly surprising that commentators have asked whether Iran is America’s Suez, or whether some other historical analogy (such as Gallipoli) better captures the mixture of overreach, confusion, and strategic frustration.

There is something to this. Suez was, among other things, a crisis of chokepoints, prestige, misjudgement, and alliance management. A great power discovered that a military move could expose political dependence rather than conceal it. If the United States finds that force in the Gulf cannot deliver political settlement, cannot restore maritime confidence, cannot discipline regional allies, and cannot impose a durable outcome on Iran, then the Suez comparison becomes more than journalistic decoration.

But America is not Britain in 1956. That is not a compliment; it is a distinction. The United States is vastly more powerful than Britain was at Suez, but it is also more deeply entangled in the global system it claims to police. Britain’s humiliation in 1956 lay partly in discovering that it could not defy American financial and diplomatic pressure. The United States is unlikely to experience that precise form of subordination. Its vulnerability lies elsewhere: in overextension, domestic polarisation, alliance mistrust, energy shocks, congressional resistance, and the possibility that coercion produces neither capitulation nor credibility.

A lazy Suez analogy says: America is finished, and here is the proof.

A better Suez analogy asks: what kind of power is being revealed by this failure, and what kind of power was only being performed?

Greenland and the return of the map

Then there is Greenland.

At first glance, Greenland belongs less obviously to Munich or Suez. No one sensible thinks Denmark is Czechoslovakia in 1938. Greenland is not the Suez Canal. Yet the recent American pressure over Greenland belongs in this essay because it reveals the same bad historical habit: treating geography as if it were a board-game asset, then looking around for a grand analogy afterwards.

Donald Trump has repeatedly argued that the United States needs Greenland for security reasons, and in March 2025 said he thought US annexation would happen; Denmark has maintained that Greenland is not for sale. By June 2026, Marco Rubio was describing ongoing discussions with Denmark and Greenland over strategic use and strengthened US military presence, framing the issue in terms of collective defence and Arctic security.

The temptation here is not simply to shout imperialism and be done with it, though the word is not exactly hiding in a cupboard. The more interesting point is that Greenland exposes a style of strategic thinking that old analogies struggle to capture. Is this Suez, because a major power is acting as if strategic geography entitles it to override political consent? Is it a Monroe Doctrine revival, because security is being defined through hemispheric control? Is it an Alaska Purchase fantasy, as if land acquisition were still the natural grammar of statecraft? Is it just great-power competition in the Arctic, dressed up in the language of real estate?

The answer may be: partly, but not quite.

That is precisely why the example is useful. Easy analogy wants every case to become an old story. Greenland resists that. It involves climate change, Arctic shipping, rare earths, missile defence, NATO politics, Danish sovereignty, Greenlandic self-government, Chinese and Russian activity, and the strange return of territorial acquisition as a serious-sounding topic in American politics. The point is not that no analogy helps. The point is that several half-help, and none should be allowed to take command.

The curse of casting

The deepest problem with Munich and Suez is not historical inaccuracy. It is moral casting.

Once the analogy is chosen, the roles follow. Someone is Hitler, Chamberlain, Churchill. Someone is Eden, de Gaulle, Nasser, Eisenhower. The present is no longer a problem to be analysed; it is a play to be cast.

This is comforting because it reduces uncertainty. It tells us who deserves contempt and who deserves vindication. It also lets us inherit the moral authority of hindsight without paying the cost of living before the outcome is known.

But foreign policy is made before the archive has sorted everyone into their proper folders. The people in 1938 did not know they were in the preface to 1939. The people in 1956 did not know they were acting out the definitive parable of post-war British decline. They had fears, assumptions, intelligence, prejudices, domestic pressures, imperial memories, economic constraints, bad incentives, worse advice, and partial knowledge. That does not absolve them. It makes them historical actors rather than cardboard warnings.

This matters because easy analogy often smuggles in policy under the cover of memory. Munich usually says: do more, escalate sooner, refuse compromise, treat negotiation as weakness. Suez usually says: do less, beware theatre, distrust prestige, recognise limits. Vietnam says: avoid the quagmire. Afghanistan says: do not try to remake societies at gunpoint. Iraq says: intelligence can be politicised, victory can be declared long before the consequences arrive.

Each of these lessons can be true. Each can also be disastrously misapplied.

The question is never simply: which historical episode does this resemble?

The better questions are more awkward. What feature of the past is actually being compared? What would have to be true for the analogy to hold? What is different enough to break it? What competing analogy would produce a different conclusion? Is the comparison being used to illuminate the present, or to dignify a position the speaker already held?

That last question is usually the most revealing one.

Remembering badly

There is a strange arrogance in the easy analogy. It seems humble because it looks backward. It claims to learn from history. But often it does the opposite. It raids history for moral props, then congratulates itself on seriousness.

Munich and Suez deserve better than that. Munich was not simply a fable about weakness. It was a crisis shaped by military preparedness, trauma from the First World War, imperial assumptions, public opinion, strategic misjudgement, and the abandonment of a smaller state by powers that had persuaded themselves they were buying peace. Suez was not simply a fable about decline. It was a crisis of oil, finance, deception, alliance politics, decolonisation, personal animus, and the gap between imperial memory and actual leverage.

The simplified versions endure because they are useful. Munich gives us courage without complexity. Suez gives us realism without detail. Both allow commentators to sound historically literate while moving very quickly past history.

This does not mean we should stop using analogy. That would be impossible, and probably undesirable. The past is one of the few resources we have for thinking about situations not yet finished. But we should use it more slowly. We should distrust analogies that arrive bearing only one lesson. We should be especially suspicious of analogies that make us feel brave, wise, or vindicated before we have done any real work.

The curse of the easy analogy is not that it remembers the past.

It is that it remembers the past just well enough to stop thinking.

Further Reading

For the general theory of analogy in foreign policy, Yuen Foong Khong’s Analogies at War is my obvious scholarly anchor; it is specifically about Korea, Munich, Dien Bien Phu, and Vietnam in US decision-making. (https://www.jstor.org/stable/j.ctvzxx9b5)

There may never have been an actual god called Moloch.

That claim isn’t settled, I hasten to add. The Hebrew Bible repeatedly condemns the giving of children to Moloch, which generations of readers have understood as the name of a foreign deity who received human sacrifices. John Milton later placed Moloch among the fallen angels of Paradise Lost; Fritz Lang turned the industrial machine of Metropolis into his furnace; Allen Ginsberg found him in the factories, armies, money and loveless machinery of modern America. By the time Scott Alexander published his seminal Meditations on Moloch in 2014, Moloch had become the embodiment of every system that consumes what human beings value while promising them power in return. He even made pop-culture appearances, in Buffy, Stargate, and Supernatural, as a demon or deity of varying kinds.

Yet there’s another interpretation. Beginning particularly with the work of Otto Eissfeldt in 1935, scholars have argued that the Semitic term MLK may originally have referred not to the recipient of a sacrifice, but more properly to the kind of sacrifice performed. Punic inscriptions describe MLK offerings made to named gods such as Baal Hammon and Tanit, and on this reading, the biblical injunction may not have meant that a child was given to a specific god called Moloch, but that the child was given as a moloch: a particular form of offering. Other scholars continue to defend the existence of a deity, perhaps an underworld god or a title applied to several gods. The evidence remains disputed.

I’m highlighting this debate not because it’s an antiquarian curiosity (though that is also true). Far from it: it suggests a different way of understanding Alexander’s Moloch, and perhaps a different way of resisting him.

What if Moloch wasn’t the god above the altar? What if, rather than the gigantic, bronze, bull-headed idol, Moloch was the name of what people did - and do - upon it?

The god who offers power

Alexander begins his now famous essay with a problem which is both simple and profound. Human beings repeatedly produce social arrangements that almost nobody would choose, yet which nobody within the arrangement can safely abandon. Competing businesses degrade working conditions because any firm that pays substantially more may be undercut. States acquire weapons, armies, and the organised capacity for ever-greater violence because unilateral disarmament would leave them vulnerable. Publishers produce sensational material because sober restraint loses attention to less scrupulous competitors. Nobody needs to desire the final outcome - individually defensible actions can still converge upon collective disaster.

Alexander calls these arrangements multipolar traps. From outside the system, cooperation may be obviously preferable, but from within it, the first person to cooperate may simply be exploited. The problem isn’t necessarily ignorance, malice or greed. It’s that restraint carries a cost, while defection is rewarded quickly enough that the more ethical actor may disappear before their example can spread.

This is the great strength of Meditations on Moloch. Political argument frequently assumes that bad conditions must have been produced by bad people. Alexander explains why decent people, following intelligible incentives, may, can, and often do reproduce conditions which they themselves detest. It isn’t enough to tell an employer to be kinder when kindness threatens bankruptcy; nor is it enough to tell a state to trust its rivals when misplaced trust may be fatal. Moral exhortation can’t substitute for institutions which make moral conduct survivable.

In this paradigm, Moloch names the distance between the world that nearly everyone would prefer and the world that competitive pressures allow them to build and live in. And so we’re all trapped in the world where we’re giving ourselves painful electric shocks all the time rather than the entirely-within-our-grasp mini-utopia in which every country spends a lot more on infrastructure and a lot less on defence.

The metaphor is powerful because the ancient bargain is powerfully suited to the mechanism. Alexander’s Moloch always offers the same exchange: surrender what you love, and receive strength enough to defeat those of your opponents who haven’t (yet) made the sacrifice. Once one participant accepts, everyone else is implicitly confronted with the same bargain. Eventually all have surrendered the valuable thing, no-one retains the advantage against their competitors, and the cycle begins again, but from a poorer starting point. What’s the next thing they love, to toss into the flames, to buy just a little bit more strength?

Yet there’s a danger in the metaphor’s very power. The god-image which clarifies the system can also conceal the people who sustain it.

From practice to deity

Alexander knows that Moloch isn’t literally a supernatural being. His examples concern incentives, selection pressures, evolutionary processes and failures of coordination. Nor does he claim that all coordination is impossible; governments, technologies and social institutions can sometimes bind destructive competition, although those coordinating institutions may develop traps of their own.

Nevertheless, the grammar of the essay repeatedly carries him beyond this analytical caution. Moloch ceases to be a name for things people do under particular conditions and becomes the answer to the question of who is doing this to us. Moloch designs traps. Moloch offers bargains. Moloch grinds down human values. Moloch rules the universe.

By the end, Alexander’s analysis has developed into something close to a secular theology. Against Moloch and the uncaring forces of competitive optimisation stands Elua, a tentative Jacqueline Carey-inspired god of love, beauty, art, and civilisation. Between them lies the garden: a protected space within which human beings can cultivate their own values instead of obeying the law of the jungle. Yet because any local garden remains vulnerable to less restrained competitors beyond its walls, Alexander concludes that human values ultimately require a Gardener powerful enough to govern the whole universe. A sufficiently advanced intelligence might finally defeat Moloch by coordinating everything.

It’s an extraordinary conclusion. A compelling description of collective-action problems gradually becomes an eschatology in which fallen humanity must help to lift a new god to heaven, creating the benevolent intelligence that will redeem creation from its inhuman ruler.

The historical ambiguity surrounding Moloch allows us to reverse that movement. What changes when we turn the god back into a sacrifice?

Who is on the altar?

Once Moloch is understood once again as a practice rather than an agent, different questions become foregrounded.

• Who is making the offering?

• What are they sacrificing?

• What advantage do they expect to obtain, and indeed who receives that advantage?

• Most importantly, is the person deciding that the sacrifice is necessary also the person placed upon the altar?

These questions expose a weakness in the general category of the multipolar trap. Some destructive systems are genuinely symmetrical, like our contrived electric-shock-country - every participant would prefer an enforceable agreement, but none can risk acting alone.

But many many other systems contain actors who clearly benefit from the apparent failure… and therefore have little desire to coordinate their way out of it.

Consider an organisation which responds to financial pressure by cutting the pay and increasing the workload of its least powerful employees - much like Alexander’s fifth example. It may be true that senior managers face real competitive constraints and are thus limited in the decisions they can make. They may even feel bad about the decision. But it doesn’t follow that the burden has been allocated impersonally. The organisation could reduce executive remuneration to support its workers, accept lower profits as the cost of doing the right thing (and make that virtue a part of their marketing strategy, as many businesses do), change its business model, seek collective regulation or withdraw from the market. Perhaps none of these alternatives are easy; perhaps some are genuinely unavailable in the specific context. Nevertheless, someone still interprets the constraint and decides whose interests are negotiable; someone chooses the child to place upon the altar, rather than laying themselves down in Moloch’s arms.

In other words, Moloch doesn’t select the victim. People do.

The same is true of institutions which organise themselves around crude performance measures. A university may well discover that publications, grants and student evaluations are easier to count than thoughtfulness, intellectual courage, pastoral care or genuine educational transformation. Nobody needs to believe in their heart of hearts that the metric captures everything valuable. Yet recruitment, promotion and funding decisions begin to reward whatever the metric recognises. Academics adapt rationally; departments encourage them to adapt; the measure becomes increasingly important because everybody behaves as though it is important.

This is recognisably Molochian, in that individual refusal may damage a career without changing the institution. Yet the metric didn’t descend from the heavens. It was selected, implemented, retained, and reinforced over time. The system exerts emergent pressure, but that pressure travels through a socio-material web of meetings, zoom calls, policy documents, regulations, budgets and decisions, decisions, decisions. The rite of Moloch does not perform itself; it has priests, liturgies and an altar.

To describe the outcome only as a coordination failure risks making everyone equally helpless. They clearly aren’t. Some people design the measure; others are measured by it. Some can absorb the cost of dissent; others can’t. Some receive the benefits of the system; others become the raw material it consumes.

The language of sacrifice restores this distributive dimension. It insists that every social order ultimately chooses what it will treat as expendable.

A climate of sacrifice

Climate change may be the clearest example of why it matters whether we imagine Moloch as a god or as a ritual of sacrifice.

In one sense, it fits Alexander’s account almost perfectly. No state can stabilise the climate alone. A government which imposes costly restrictions while others continue emitting may incur immediate economic and political penalties while securing only a fraction of the resulting benefit. Businesses which voluntarily absorb environmental costs may be undercut by competitors willing to transfer those costs elsewhere. Individuals can reduce their own emissions, but the atmosphere won’t distinguish their restraint from another person’s excess. The benefits of defection are often concentrated and immediate, whereas the costs are so widely dispersed across borders and generations that they may even seem completely insignificant.

This is a multipolar trap. And yet… it isn’t just a multipolar trap, is it?

To describe climate change exclusively in those terms risks imagining humanity as a collection of roughly equivalent prisoners, each contributing to and suffering from the same unfortunate equilibrium. In reality, responsibility, power and vulnerability are distributed with extraordinary inequality. Developed economies in North America and Europe produced most of our historical industrial emissions. Within the contemporary world, the wealthiest households and largest businesses account for a vastly disproportionate share of consumption emissions. Meanwhile, many of the communities most exposed to rising seas, extreme heat, food insecurity and destructive weather events have contributed comparatively little to the overall accumulation of greenhouse gases, and possess comparatively few means of resisting or mitigating those effects.

Moloch may grind down human values, but he doesn’t grind everyone down equally.

Consider the position of a low-lying Pacific island state, such as Kiribati. Its people may reduce their already modest emissions, produce ambitious climate plans and advocate powerfully for stronger international agreements. But this may all prove utterly, utterly futile in the face of global inertia, because none of this virtuous effort can prevent seas warmed by emissions produced elsewhere from rising around them. The sacrifice, demanded by others, may encompass homes, freshwater supplies, livelihoods, burial grounds, political sovereignty and the ultimate possibility of humans remaining upon a landmass known as Kiribati at all. What appears from a distant negotiating chamber as an acceptable fraction of a degree can become, at the shoreline, a question of whether a community remains inhabitable.

The people making this sacrifice aren’t those receiving the bargain’s rewards. The historical benefits of fossil-fuelled development accrued disproportionately to richer industrial societies; the profits of continued extraction flow through particular companies, investors and states; the convenience of delay is enjoyed principally by governments and populations that possess the resources to protect themselves from at least the worst of its consequences. The altar is global, but the victims have addresses.

In the Moloch-as-deity model, the offering is something the supplicant loves - a child, usually. But do any of the most polluting nations, companies, or households love Kiribati? I doubt most could reliably point to it on a map.

Even within wealthy countries, the distribution of sacrifice remains political. Climate policy is often narrated as a demand that ordinary people consume or behave differently: drive less, fly less, replace the boiler, alter the diet, pick up your litter, remember the correct bin. Some of these changes are necessary, particularly among affluent people whose consumption produces far more emissions than the global norm. Yet exhortation to individuals can’t substitute for infrastructure. A person can’t reliably abandon the car where public transport is absent; a tenant can’t insulate a property their landlord won’t improve; a family can’t purchase technologies they can’t afford.

The language of personal responsibility becomes Molochian when it treats constrained choices as moral failures, while leaving the social and cultural institutions producing those constraints untouched.

That doesn’t mean transferring every responsibility from individuals to corporations, as though production and consumption occupied separate worlds. Businesses produce goods which people use; voters resist policies whose costs they fear; prosperous consumers possess choices unavailable to poorer households. The point is that neither responsibility nor power is evenly shared. Energy companies, property owners, vehicle manufacturers, investors and governments make decisions which structure the choices subsequently offered to millions upon millions of individuals. Some actors choose from the menu presented to them; others decide what may appear upon it. We should not pretend the two are equal.

Climate policy can itself reproduce the sacrificial logic it seeks to overcome. A transition funded through regressive household charges, insecure employment or the abandonment of industrial communities may tell those with the least power that their livelihoods are the price of decarbonisation. Political leaders then confront a predictable backlash and declare ambitious policy impossible. Once again, the supposed demand of an impersonal system conceals a prior distributive choice. The transition had to be paid for; it didn’t have to be paid for in that way.

The alternative isn’t to pretend that climate action requires no costs. Every serious response entails investment, disruption and the closure or transformation of carbon-intensive activities. The relevant question is who possesses the resources to absorb those costs, who has already benefited from the activities being curtailed, and who gets to decide what a just distribution would look like.

This is where understanding Moloch as a sacrifice becomes more useful than treating him as a god. The divine Moloch tells us that competition demands continued emissions and that only a universal coordinator could finally compel everyone to stop - another god to fight a great war in heaven. The sacrificial Moloch directs our attention towards the institutions through which particular people translate competitive pressures into political choices. It asks who licenses extraction, subsidises infrastructure, sets energy prices, writes trade rules, finances adaptation and determines whether communities facing irreversible loss will receive meaningful assistance or empty sympathy.

Climate change is sustained by collective-action problems, but it’s also sustained by power. Its victims aren’t selected by an invisible deity. They’re selected through decisions about which risks are tolerable, which interests are protected and which places may be surrendered.

The theology of helplessness

Personification is useful because systems really can acquire a quasi-independent force. A market, bureaucracy or convention can constrain every participant without being controlled by any one of them. The fact that an institution is socially constructed doesn’t make it imaginary, voluntary or easy to escape. Borders are constructed; so are currencies, property rights and states themselves. They all nevertheless remain capable of shaping lives with brutal material force.

It would therefore be facile to say that, because human beings created Moloch, human beings can simply choose to stop believing in him. Alexander’s opening problem remains. Nobody can dismantle a collective practice through private virtue alone when everyone else retains an incentive to continue it.

The more defensible conclusion is that Moloch isn’t unreal because people made him. He’s alterable because human practices sustain him.

This is where Alexander’s theological framing becomes, in my humble view, politically disabling. If Moloch represents the universal logic of competition, then ordinary reforms can appear as temporary walls around small gardens, forever threatened by the wild primordial jungle outside. The only final answer is an intelligence capable of coordination from a god’s-eye view.

Yet there may be no final answer, and the desire for one may itself be dangerous. A universal Gardener would need both the capacity to suppress defection and the authority to decide which human values should be cultivated. The abolition of multipolar competition could therefore require a concentration of power before which every existing state would appear embarrassingly modest. Alexander acknowledges the danger of tyranny, but the structure of his argument makes that sort of supreme coordination increasingly difficult to avoid.

Politics offers a less complete but more humane answer. Trade unions prevent employers from competing solely through the degradation of labour. Environmental regulations prevent firms from gaining advantage by solely transferring costs to the public. Arms-control agreements reduce the penalty attached to mutual restraint. Welfare states make refusal or inability less personally ruinous. Professional norms protect goods which markets and metrics struggle to value. Democratic organisation allows those designated for sacrifice to contest the terms upon which decisions are made.

None of these institutions destroys Moloch everywhere and for all time. Each closes one altar, for a while, by changing the bargain offered to those gathered around it.

They can fail. They can be captured, weakened or turned into new sources of domination. Nevertheless, their existence demonstrates that competitive pressures aren’t natural laws operating independently of politics. They’re channelled through rules, ownership structures, legal rights, cultural expectations and distributions of power. Change those arrangements, and the supposed demands of the god change with them.

Dismantling the altar

Alexander is right that nobody needs to want a catastrophe for everyone to help produce one. He’s right that moral appeals will usually fail when the sacrifice continues to purchase comparative power. He’s right, above all, that people require the security to act well; virtue which guarantees defeat won’t organise a humane society.

Yet the historical Moloch presents another possibility. Perhaps there was no single god making these demands. Perhaps there were people performing a recognised sacrifice to different gods, for different purposes, because their society had taught them that the offering was necessary.

The modern parallel isn’t that our destructive systems are illusions. It’s that we mistake practices for powers, decisions for necessities and inherited institutions for permanent features of reality. We personify the system, then absolve the people who administer it. We speak as though Moloch has chosen the sacrifice, when the knife remains in human hands.

The question, then, isn’t how humanity can kill an implacable god. Nor is it whether each of us can preserve our innocence by refusing to participate. Individual purity won’t dismantle a collective system.

The question is why we keep rebuilding the altar in the first place; who is permitted to call the sacrifice necessary; whose children are placed upon it; and what kind of reformation might allow us to stop.

Moloch may never have been a god, even if the sacrifices were real all the same. That should not make us less afraid of Moloch. But it might make us a little less fatalistic.

ADDENDUM: The machine’s average

There’s one final altar, rather closer to the desk upon which I’m writing this.

Generative AI presents an unusually clean version of Moloch’s bargain. A general-purpose language model has absorbed such an enormous portion of digitised textual culture that, from the perspective of an ordinary user, it can feel close enough to a compression of everything we’ve ever written. Ask it for an argument, a summary, an email, a press release, or the opening of an essay, and it produces something which is usually coherent, competent and remarkably quick.

Its average is high. That needs saying plainly, because much criticism of AI depends upon pretending that the machine produces nothing but malformed multi-digited hands, sentences full of extended em-dashes, and trite phrases like “in today’s rapidly changing world.” The machine’s average would beat most of us on most days. It doesn’t get tired, lose concentration, forget the structure halfway through, or decide that doom-scrolling through Bluesky is a necessary preliminary stage in revising paragraph four.

Full disclosure: I use AI fairly regularly. I use it to make initial stress-tests of arguments, reorganise my sometimes chaotic notes, transcribe my spoken-out-loud thoughts, locate gaps in my underlying research literature, and turn a loose sequence of thoughts into something resembling a first draft. The result is generally quicker and cleaner than the unaided version would have been, which is presumably why I continue paying for it.

The strategic difficulty is that this barely gives me an advantage. The machine is brilliant for everybody, all the time, provided they can secure access to it and learn at least a basic level of competency in writing decent prompts. You can see how this specific example becomes a general problem:

• A writer adopts AI and becomes faster; their competitors adopt it and become faster too.

• An office uses it to produce more polished correspondence; every comparable office follows, and suddenly the new baseline standard is so high that unaided humans would have to put in an inordinate amount of effort to match it.

• A student (if they’re good at prompting, which isn’t always the case) uses it to raise the baseline quality of an assignment; the expected baseline rises accordingly.

In all cases, the initial gain becomes the minimum standard required to remain competitive.

Doshi and Hauser captured part of this dynamic in a 2024 experiment on creative writing. Participants wrote short stories either without assistance or with access to one or several ideas produced by a generative model. The AI-assisted stories were judged more creative, better written and more enjoyable, particularly when their authors had scored lower on an earlier measure of creativity. Individually, the writers improved.

Collectively, however, their stories became more alike - likely because they were all drawing on the same corpus of training data.

This is almost too perfect an example of Moloch. Each individual has a rational incentive to use the machine because the machine makes their individual work better. Yet, when everyone draws upon the same underlying system, the collective result narrows around its preferred structures, associations and rhythms. We’re each pulled upwards, but towards approximately the same place. And as a consequence, everyone’s work is relatively no better or worse or even that much different than anyone else’s was, except we’re all a bit poorer for it.

The danger isn’t exactly consensus. A language model can generate eccentricity on command, imitate conflicting styles, and propose stranger ideas than many users would reach unaided. The danger is that in using AI, we all drift towards a statistical central tendency disguised as frictionless competence. Left to its defaults, the machine has an extraordinary capacity to make prose reasonable, balanced, clear, and vaguely inhuman. It knows how an essay of this kind is normally shaped; it doesn’t necessarily know which of its well-formed sentences Tom J. Elston would find irritating enough to delete.

I notice this in my own writing. In terms of the effort-to-output cost-benefit calculation, AI generally makes me more readable for the slight amount of effort it takes, but I think it takes a certain edge off; sometimes it makes me slightly less feisty. It removes repetition, repairs transitions, and replaces clumsy formulations with cleaner ones; occasionally, it also reduces an intentionally-placed hyperbole, over-clarifies an argument, or removes the odd Dutch-angle of a sentence which made the paragraph mine. This is why I copy-edit quite violently after using it. The final revision isn’t quality control per se, in that I’m often actively making the work worse in an objective sense - if anything, it’s an attempt to put the human splinters back into an overly smooth intermediate-stage product.

Nick Cave made a much more severe version of this argument when asked about songwriters using ChatGPT because it was faster and easier. For Cave, creative struggle wasn’t an inefficient route to the finished work; it was part of what gave the work its value. By mechanising imagination, he argued, AI threatened to turn the process of creation into an unnecessary inconvenience standing between the consumer and the commodity.

I don’t quite share Cave’s absolutism, and neither, apparently, does Cave any longer. After seeing Andrew Dominik’s AI-assisted accompanying film for Tupelo, he later described his own view as having softened: the result struck him as original, affecting and evidence that AI might indeed operate as an artistic tool rather than merely as a substitute for art.

That distinction is, I think, more useful than either the uncomplicated enthusiasm of the typical neck-bearded tech bro or the luddite-ish total refusal of the anti-AI puritan.

It tells us that the relevant question isn’t whether AI touched the work or not, but what part of the work has been surrendered to it, under what pressure, and with what consequence. There’s a qualitative difference between asking a model to challenge an argument and asking it to decide what you believe; between using it to remove the hasty accidental clumsiness of a human-in-a-hurry and allowing it to remove every trace of productive awkwardness characteristic of such mere mortals as ourselves; between treating speed as a convenience and accepting speed as the governing measure of intellectual labour.

The sacrifice may be literal capital: an individual’s monthly subscription, an institutional licence, an ever-growing infrastructure bill for a nation or for the planet. It may be professional skill, as tasks once used to develop judgement are delegated before that judgement has formed. It may be the private experience Cave describes as creative struggle, the very effort exerted to create (and I don’t think he’s wrong, by the way, plenty of things are made valuable by the time and effort required to do them; if a string quartet bashed out a 27-minute piece in nine minutes, would you say their productivity had trebled?) Or it may be collective distinctiveness: thousands of individual pieces becoming better composed, more immediately useful and, fraction by fraction, less like the person who set out to write them.

None of this is inevitable. Current research suggests that different prompts, deliberately varied AI personas, and incentives which reward originality can preserve much more diversity. People who use the model selectively, rather than treating its first suggestion as an answer, needn’t be pulled towards precisely the same centre.

Once again, Moloch isn’t the machine. Moloch is the bargain made around it.

The bargain says that we must use the tool because everyone else will; that we must become faster because speed has become more available; that we must surrender whatever slows production because our competitors may surrender it first. Eventually everyone has gained the same brilliance, paid the same entrance fee… and abandoned some small portion of whatever they once valued about doing the work themselves.

There is an old anecdote from Neil Gaiman’s blog (from 2009, I hasten to add, well before the allegations made against him came to light) that has lodged itself somewhere in the moral compost heap of my brain.

Gaiman remembers hearing a comics editor complain about an artist who had taken a few weeks off to paint his house. The editor’s objection was not that the house looked bad, or that the artist had climbed a ladder in a high wind while wearing novelty slippers, or that he had accidentally chosen a shade called Tuscan Persimmon and condemned the neighbours to live beside it. The objection was financial. For the money the artist could have earned during those weeks, he could have hired somebody else to paint the house and still come out ahead.

Gaiman’s unspoken reply was wonderfully simple: “But what if he wanted to paint his house?”

This is one of those lines that feels almost too small to contain the thing it contains. It is not a grand manifesto. It does not come with a twelve-point programme, a subscription model, or a downloadable worksheet called Reclaiming Your Paint-Based Agency. It is merely the thought that a person might do something because he wanted to do it.

How quaint. How inefficient. How magnificently suspicious.

The point, of course, is not that everyone should paint their own house. I am not proposing a moral hierarchy in which the righteous scrape skirting boards while the damned outsource the glossing. If you hate painting, are bad at painting, cannot safely paint, or simply have better things to do, then by all means pay someone who knows which end of the brush to hold. There is no dignity in making yourself miserable for the sake of pretending that the middle classes have been corrupted by tradespeople.

The point is smaller and larger than that. Smaller, because it is just about a man painting his house. Larger, because it exposes one of the stranger habits of modern life: the assumption that every choice must be defended as an efficiency calculation.

Under this logic, the question is never simply, did you enjoy it? The question is, did it make sense? Did it maximise your hourly value? Did it support your long-term goals? Did it contribute to your physical health, mental resilience, professional development, personal brand, side-hustle potential, creative practice, networking capacity, domestic optimisation strategy, or at least your ability to return refreshed to the sacred inbox?

To which the only civilised answer is sometimes: no. I just fancied it.

The Case for Getting Things Done

There is, annoyingly, a serious concession to make here.

Efficiency is not wicked. Planning is not a moral failure. It is good to get things done. Some tasks genuinely do need doing; some deadlines matter; some responsibilities become worse if left to fester in the dark like potatoes at the back of a cupboard. A person who refuses all structure in the name of authenticity will, quite quickly, discover that the gas bill has not been moved by their commitment to spiritual freedom.

Nor is self-improvement obviously bad. There are habits worth building, skills worth learning, bodies worth looking after, finances worth managing, relationships worth repairing, and homes worth making less actively hostile to human life. There is nothing noble about avoidable chaos. There is no romance in missing every deadline, losing every document, eating cereal from a mug because all the bowls have achieved archaeological significance in the sink, and calling the whole arrangement a resistance practice.

The serious version of productivity culture appeals because life is genuinely difficult to manage. Work expands. Administration multiplies. Notifications breed in captivity. Every institution, from the bank to the GP surgery to the local council website, appears to have entered into a secret pact to make ordinary life slightly more like filing taxes in a haunted airport.

So yes, we need systems. We need calendars, reminders, routines, lists, limits and the occasional grim afternoon in which we sort out the thing we have been avoiding for six months and discover that it took seventeen minutes. A good productivity method can be an act of mercy. It can reduce friction. It can help us keep promises. It can create enough order that life becomes less exhausting.

The problem is not doing things well.

The problem is when doing things well becomes indistinguishable from optimising the self.

That distinction matters because one is practical and the other is metaphysical. It is one thing to say, I need a better way of managing my time. It is another to say, I am a badly designed human product whose every feature requires upgrading. The first can help you answer emails. The second can turn existence into a permanent performance review.

When Everything Becomes a Metric

The most exhausting thing about productivity culture is not even the advice itself. Much of the advice is fine. Sleep more. Move about. Write things down. Do the difficult task first. Do the easy task first. Batch similar tasks. Do not batch similar tasks. Use your phone. Throw your phone into the sea. Wake at five. Respect your chronotype. Drink water. Do deep work. Touch grass. Put the grass into Notion.

The advice contradicts itself, but that is not the deepest problem. Human beings are contradictory, and any advice that survives contact with actual life will need exceptions, caveats and a certain amount of shrugging.

The deeper problem is the moral atmosphere. The language of optimisation smuggles in a whole theory of the person. You are no longer someone with finite time, strange moods, bodily needs, obligations, pleasures, griefs, friendships, distractions, appetites and occasional inexplicable desires to reorganise a drawer. You are a project. Your attention is capital. Your leisure is recovery. Your recovery is performance maintenance. Your hobbies are skill stacks. Your books are inputs. Your walks are steps. Your meals are fuel. Your friendships are networks. Your personality is content. Your sadness is friction. Your body is an asset. Your life is a portfolio of under-exploited efficiencies.

At that point, even rest becomes suspicious unless it can produce evidence of usefulness. You may nap, but ideally only because sleep improves cognitive performance. You may walk, but ideally while listening to a podcast about leadership. You may cook, but it should support gut health, reduce delivery costs, increase domestic competence, photograph well, and produce enough leftovers to make Thursday less pathetic. You may read a novel, but only if it expands empathy, sharpens prose style, deepens cultural literacy or provides a useful analogy for a future essay. I am not saying I have never done this. I am saying I object to being seen.

This is why the phrase self-care so often has a faintly managerial smell. In its better form, self-care is a humane recognition that people have needs and limits. In its worse form, it means servicing the machine so it can return to work on Monday without making a noise.

There is something bleakly comic about a culture that has learned to defend rest on the grounds that it improves output. It is like defending friendship because it lowers blood pressure, or defending beauty because it may improve regional footfall. The defence is not false, exactly. It is just spiritually deranged.

Rest may well make us more productive. Exercise may well sharpen concentration. A holiday may well prevent burnout. A lunch break may well improve afternoon performance. These are useful arguments in workplaces and policy documents because institutions often require a business case before they will allow human beings to behave like mammals.

But it should worry us if the only acceptable defence of rest is that it eventually pays productivity rent.

Rest Is Not Performance Maintenance

Oliver Burkeman’s Four Thousand Weeks is useful here because it attacks one of the great fantasies beneath modern productivity advice: the idea that, with the right system, we might one day get everything done. Burkeman frames much productivity advice as encouraging a limit-denying dream in which we become fully optimised masters of our time; his alternative is to accept finitude, choose what to neglect, and organise life around what matters rather than what can be endlessly processed.

This is not a glamorous thought. Nobody wants to buy a productivity book called Some Things Will Remain Unanswered and Eventually You Will Die. It lacks airport appeal. But it has the advantage of being true.

A great deal of productivity culture depends on denying limits while pretending to manage them. It tells us that the problem is not the impossible volume of demands, but our inadequate system. We have not found the correct app. We have not blocked our calendar with sufficient aggression. We have not learned to say no in a tone of voice that suggests both emotional maturity and excellent core strength.

Yet the truth is that a finite life cannot be made infinite by better sorting. There will always be unread books, unanswered messages, unvisited places, unmastered skills, unwritten essays, unstarted projects and people with whom we should probably have been more patient. The human problem is not how to clear the decks. The human problem is how to live on a deck that will never be clear.

That sounds miserable, but it is also a relief. If the fantasy of total completion is false, then the failure to complete everything is not a personal defect. It is a condition of being alive. You are not behind because you have failed to catch up with the universe. The universe was never offering inbox zero.

This changes the moral status of rest. Rest is not a tactic by which you trick yourself into becoming a more efficient worker. Rest is part of life. It is not justified only by what it repairs, improves or enables. Sometimes you are tired because you are tired. Sometimes you should stop because continuing would make you unpleasant, inaccurate, brittle or weirdly emotional about printer settings. Sometimes the correct response to being a person with a body is to lie down.

This should not be radical. It should barely be an argument. Yet here we are, bravely asserting that human beings may sit quietly without first producing a theory of impact.

The Proper Uses of Pointlessness

Jenny Odell’s How to Do Nothing helps because it refuses the lazy caricature of anti-productivity as mere passivity. Odell is not arguing for sloth, ignorance or a permanent state of decorative collapse. Her argument is about refusing the attention economy’s terms: reclaiming attention, listening, observing, remaining in the world while declining to treat every moment as raw material for extraction. The LSE Review of Books summarises the book as a case for refusing productivity and redirecting attention towards active listening and contemplation, rather than simply doing less.

That is an important distinction. Doing nothing properly is not the same as doing nothing badly.

Doing nothing badly is the grey mush of semi-conscious scrolling; the kind where you look up after forty minutes with the spiritual texture of wet cardboard and a working knowledge of three controversies involving people you do not know. Doing nothing badly is not leisure so much as leakage. Time has passed, but it has not been inhabited.

Doing nothing properly is different. It may look inefficient from the outside because it does not always produce a visible object. But it is not empty. It is the walk where nothing happens except that you notice the angle of the light on a brick wall. It is the conversation that does not advance the plot. It is cooking something slightly too elaborate because the chopping is pleasant. It is reading about a topic you have no intention of monetising. It is sitting in a garden and slowly becoming aware that at least three birds are conducting complicated personal business in the hedge.

These things are not valuable because they secretly make you better at email. They may, incidentally, make you a calmer or more attentive person. Good. Lovely. Put that on the leaflet if necessary. But their first defence is that they are part of being alive.

Kieran Setiya’s distinction between telic and atelic activities gives a useful philosophical backbone to this without requiring us to say the word Aristotle too loudly in a light essay. Telic activities are directed towards completion: finish the article, buy the house, get the promotion, run the race. Atelic activities are valuable in the doing: listening to music, spending time with a friend, walking, thinking, living well. Setiya’s point is not that achievement is worthless, but that achievements are not all that matter; atelic activities locate value in the present rather than postponing fulfilment until the goal has been reached.

This is exactly what optimisation culture struggles to understand. It can cope with painting the house if the goal is a painted house. It struggles with painting the house because painting the house is how you wanted to spend the afternoon. It can cope with running if you are training for a marathon, lowering your resting heart rate or constructing a personal narrative of discipline. It struggles with running because the weather was good and your legs felt like being used. It can cope with reading if the book is improving you. It struggles with reading because the book is there, and you are there, and for once nobody needs to be improved.

The optimised life is embarrassed by activities that do not terminate in a result. It wants everything to point beyond itself. But much of what makes life bearable, and sometimes beautiful, does not point beyond itself. It is not a corridor to something else. It is a room.

Wasting Time Properly

This is where the phrase wasting time needs rescuing.

There is such a thing as wasting time badly. We all know this. There are forms of waste that leave you diminished, not restored; numbed, not rested; irritated, not amused. There are evenings that vanish into the algorithmic sinkhole and leave behind only the faint impression that everyone is angry, everything is discourse, and you have somehow developed an opinion about a minor American podcast host.

But there is also such a thing as wasting time properly.

To waste time properly is to spend it without demanding that it become something else. It is to accept that not every activity needs to justify itself before the tribunal of future usefulness. It is to do something absorbing, pleasurable, restful, social, strange or beautiful without immediately converting it into a lesson, asset or productivity hack.

This is harder than it sounds, because optimisation culture follows us even into our escape routes. The hobby becomes a side hustle. The walk becomes a streak. The book becomes a content plan. The meal becomes a macro balance. The notebook becomes a system. The system becomes a lifestyle. The lifestyle becomes a brand. The brand becomes a tiny exhausting god that must be fed every morning before breakfast.

Even our pleasures are asked to professionalise. A person cannot merely enjoy coffee; they must develop a coffee practice. They cannot merely go to the gym; they must enter a transformative arc. They cannot merely take up pottery; they must open an online shop selling small irregular bowls to people who describe objects as intentional. They cannot merely enjoy a video game; they must be building transferable strategic capacities, unless they are writing an essay about how difficulty functions as atmosphere and world design, in which case apparently it is fine and intellectually important.

Again: I object to being seen.

The point is not that seriousness ruins pleasure. Hobbies can deepen. Skills can improve. Amateur interests can become crafts, and crafts can become vocations. There is nothing wrong with caring about doing something well. In fact, doing something well is often one of the great pleasures of life.

But doing something well is not the same as optimising yourself.

A person who bakes bread because they love baking bread is not the same as a person who has folded bread into a comprehensive programme of disciplined domestic excellence. A person who learns guitar because it is satisfying to make noise in a room is not the same as a person who treats every chord as an investment in personal complexity. A person who paints their house because they want to paint their house is not making a claim about efficiency. They are painting the house.

There is a kind of freedom in allowing an action to remain itself.

Conclusion: A Life Is Not a Workflow

The real objection to optimisation culture is not that it makes us work too hard, although it often does. Nor is it that it makes us feel guilty, although it certainly has a talent for that. The deeper objection is that it narrows the imagination of what a life is for.

It teaches us to see ourselves as bundles of capacity awaiting better management. It turns moral language into managerial language. It makes every limit look like a flaw, every pleasure look like an indulgence, every pause look like a failure to prepare. It tells us that the good life is the streamlined life: frictionless, measurable, disciplined, upgraded, efficient.

But a life is not a workflow.

A life contains work, and work matters. It contains duties, and duties matter. It contains goals, and goals matter too. There are things we owe to other people, things we owe to ourselves, and things that will not be done unless we do them. Nobody sensible is against calendars. Nobody humane is against trying.

But there must be room for forms of value that do not arrive wearing a lanyard.

There must be room for the pointless walk, the unmonetised hobby, the nap that is not a cognitive enhancement strategy, the conversation that solves nothing, the meal cooked slowly for no audience, the book read badly and happily, the afternoon lost to weather, the song played again, the garden stared at, the house painted by hand.

Not because these things make us better products.

Because they are among the reasons being a person is preferable to being a product.

So by all means improve yourself where improvement would help. Learn the skill. Keep the promise. Write the list. Answer the email. Fix the shelf. Make the plan. Paint the house efficiently if efficiency is what the moment requires.

But do not let anyone convince you that every hour must submit a business case.

Some things are worth doing because they are useful. Some things are worth doing because they are good. Some things are worth doing because they are yours. And occasionally, with a brush in one hand and paint on a sleeve you quite liked, you may discover that the correct answer to why are you doing this yourself? is simply:

Because I wanted to.