Frankfurt

You chose your job, your phone, the shows you watch, the opinions you hold. Nobody makes you do any of it. And still there is a feeling, most nights, that your life is being run by something you cannot point at. When you try to say what is wrong, you sound ungrateful, or paranoid. So you stop trying, and you open the phone. The strange part is that you are not even unhappy, only never sure that any of it is yours.

That feeling has a shape, and a group of thinkers spent fifty years drawing it. Their answer is that being unfree today does not mean somebody forces you. It means you live inside a kind of reason that can only ask how to get a thing, and never what the thing is for. That kind of reason built the modern world, and it also builds your wants. So the freedom you feel is real, and you really do choose. And the choosing is the form the unfreedom takes. Everything you can choose between was cut to shape before you arrived. The question of whether to want any of it was taken out of the room.

If that is right, four things have to be true, and each of them can be tested. The tests have been run, and what they found is less than either side wants. First, wanting can be manufactured, in a person who believes they want freely. In Germany, in 1933, millions had voted, with enthusiasm and before anyone forced them, for the government that would destroy the country. The working class, with unions and parties and newspapers of its own, held every right opinion and watched. The watching is the puzzle, not the vote. Second, the manufacturing is done by no liar and no villain, but by an interest that never needed a plan. It is done by reason itself, the same reason that got us out of superstition, once it is cut down to one function. Third, it happens to you after work, not at work, which you already expect to belong to someone else. Your evenings are where the argument says you are made, and that is the claim the tests reach least. Fourth, your own test of freedom, "nobody forces me", is a test this arrangement passes without anyone having built it for that purpose. That is what makes the arrangement so hard to see from inside it. A system that gets what it needs from your own wants has no reason to reach for force. So the absence of force tells you nothing about whether your wants are yours.

Those four claims are what the parts that follow have to earn, one at a time. The men who made them disagreed with each other about almost everything else, including whether there is any way out. One of them said there is no right way to live in a wrong world. Another said the way out has been sitting inside every argument you have ever had. The argument between them is not settled by choosing a side. It is settled by testing both against the one machine the older of them never saw, the one in your hand.

In 1933 a working class with its own unions, parties and newspapers watched a government come to power on a promise to destroy it. The working class did almost nothing. The Frankfurt School, a research institute in Frankfurt whose members called it the Institute, became what it is because of those two sentences. The puzzle is not the vote, whose mass base was the lower middle class. The silence. Fear and lies do not explain that silence. Something inside the person explains it, and that something was formed before the person held a single opinion. And it can be measured, badly at first and then better.

Start with what everyone expected, because the expectation is the problem. Marx had written that it is not the consciousness of men that determines their existence, but their social existence that determines their consciousness.¹ Plainly: change the conditions and the minds follow. Put people in factories, take their tools and their land, pay them by the hour. They will come to see the arrangement for what it is, and end it. Marx's own theory asked for organisation and struggle to get from the first thing to the second, and never for a reflex. But the crude reading was the one everyone in 1930 was holding, and it is the crude reading that broke. Germany had done everything the theory asked. The factories were there. The crash had thrown millions out of work. The conditions had arrived, in full, on schedule. And the consciousness that was supposed to follow them did not come, then or later.

What actually happened is more awkward than the legend, and the school's own question gets sharper when you tell it straight. The unemployed workers did radicalise. They moved left, to the Communists, not to the Nazis.² The Nazi vote was built from the lower middle class.³ And from 1930 the best predictor of who voted for the Nazis was not his job or his wage. It was whether he was Protestant or Catholic.⁴ So the workers did not choose the disaster. They watched it. They held the right opinions, and when the moment came the opinions turned out to weigh nothing. That is the puzzle a theory of conditions cannot touch. Between the conditions and the act there is a person, and the person did not do what the conditions said.

A young psychoanalyst there, Erich Fromm, had already gone looking inside that person before the collapse. In 1929 he sent a long questionnaire to German workers and clerks, and hundreds of them answered. The political questions were the easy part. The interesting questions were the private ones. How a child should be punished, whether women should work, what the man thought of his boss, what he did with his evenings. The idea was simple and, at the time, strange. A political opinion is a sentence a person says. A character is what the person will actually do when the sentence costs something. Fromm wanted to know how often the two matched. The method was crude: a questionnaire sent out to volunteers, with open answers coded by the team that had written the questions. No clean result from it can be quoted here. What the study was reaching for is the thing that matters. It was the first attempt to measure the gap between what a person believes and what a person is. It did not close that gap. A later study, in another country, measured the disposition, and measuring it is not the same as closing the gap.

With the German particulars removed, the question is this. A man says he wants the boss gone. The same man wants his son beaten for talking back, and wants his wife at home. He admires the foreman for being hard, and spends his Sunday obeying a marching band. Either the sentence is his politics, or the rest of his answers are. The school's answer is the founding move of everything that follows. The rest is the politics, and the sentence is a costume. When the costume was tested against the street in 1933, the street won.

That answer forced a change in what the school had to be. If the person is where the theory broke, then the person has to be studied. In 1930 one body of thought had been built to study how a person is made before they know it, and that was Freud's. So the Institute did a thing nobody had done. It put Marx and Freud in the same building. Marx to say what the conditions are. Freud to say how the conditions get inside a child, through the father, the punishment, the shame. There they become a character that will later vote, or not vote, or stay home. The family is where the economy is turned into a person. The father's authority at home is the boss's authority rehearsed, and the child learns to obey before it learns to vote. That is the school's founding assumption, and every part of this argument assumes it.

A historian of the period has a shorter answer, and it deserves stating at its full weight. The left was divided against itself, and its two halves spent more of their strength on each other than on the enemy. A general strike is hard to call when millions are already out of work and glad of any wage at all. And a state that arrests its opponents first does not have to persuade anyone of anything. All of that is true, and none of it needs a single word about character. That much is granted here. The school never claimed that character explains why the government came to power. The claim underneath is narrower and much harder to get rid of. It explains the distance between the opinion a person holds and what that person does when holding it starts to cost.

There is an objection here that anyone sensible raises. Character is a story. You can explain any outcome with it after the fact. If the workers had risen, character would have explained that too. A claim that fits every result is not a claim. The reply is a piece of history, and it decides whether any of this can be trusted, so it goes slowly.

After the war, in America, the Institute tried to turn the character idea into a number. They wrote a scale, a list of statements a person agrees or disagrees with. The statements were about obedience, toughness, sex, punishment, and the world being a dangerous place. The scale had one purpose, stated on its first page. To find the person whose makeup makes him particularly susceptible to antidemocratic propaganda.⁵ Before anyone was a Nazi, the theory said, you could find the person who was ready to be one. The scale was built and given to large numbers of Americans, and for a while it looked like the number the school had been missing.

Then it was caught. Every one of the statements was written so that agreeing meant authoritarian. Not one was phrased the other way round. So a person who simply tends to agree with whatever is put in front of them, a yes-sayer, scored as a fascist.⁶ The scale could not tell an authoritarian from someone polite. That is a real flaw, and for a while it looked like the whole idea died with it. What happened next is the part that answers the objection. A Canadian psychologist rebuilt the scale so that half the statements ran the other way. To score high you now had to disagree half the time.⁷ He gave the repaired version to people for decades. The finding held. The people who scored high were still the same people. They disliked so many kinds of outsider at once that he called them equal opportunity bigots.⁸

Compare that against what a story would have done. A story cannot be caught by a yes-sayer. A story does not have a flaw in its wording that a stranger can find, fix, and re-test. The character claim was stated precisely enough to be wrong, was found to be wrong in one specific way, and survived the repair. That is what a real claim looks like from the outside. Whether it is the whole explanation of 1933 is a separate question. The electoral record says it is not, because religion first and class after it did most of the predicting. What the character work explains is narrower and, for this argument, more important. The school took the gap to mean that the opinion sits in the sentence while the obedience sits in the person. On that reading, when the two disagree the person decides and the sentence loses. That reading goes a good deal further than the measurement underneath it.

What that test shows and what it does not show need separating here. The scale puts sentences in front of a person and counts which ones are agreed with, then counts which other sentences are agreed with. That is one set of agreed statements predicting another set of agreed statements. What survived the repair is that a stable disposition exists, and that it predicts hostility to outsiders. Whether it predicts what a person does when acting costs them something was never the measured result. The repaired scale kept the finding and did not need the Freudian story to hold it up. The disposition is real, and where it comes from is not carried by the numbers. So the Freud the school added to Marx stands as an argument and not as evidence.

There is a second objection, aimed not at the finding but at the whole way of working. It needs answering here, because it will be raised again against everything that follows. This is not science. You decided the workers should have revolted, and then you went looking for a reason they did not. Your theory is a politics wearing a lab coat.

Max Horkheimer, who ran the Institute, answered this by describing two kinds of thinking. The difference between them is the difference between this school and ordinary social science. Take two people looking at a fever. The first builds a better thermometer. The second asks why this society measures fevers so carefully and hunger not at all. The first takes the world as given: here are the facts, here is the instrument, make the instrument better. The second says the facts were made. The hospital, the thermometer, the decision that a temperature is worth a number and a hungry child is not. All of it was built by people, for reasons, and it could have been built otherwise. Horkheimer writes that the facts as they reach our senses are socially pre-formed twice over. The thing we look at is a product of history, and so is the eye that looks.⁹ The first kind of thinking takes what it is handed, and the second asks who handed it over.

So the second kind is biased. Horkheimer says yes, and then says the thing that turns the objection around. There is no theory of society, he writes, not even the theory of the generalising sociologist, that does not contain political interests. The generalising sociologist is the one who only counts and averages. The thermometer-builder has a politics too. His politics is that the hospital should keep running as it is, and he calls that having no politics. The difference is not that one side has an interest and the other does not. It is that one side names its interest and the other hides it inside the word neutral. That answer is a counter-charge, though, and a counter-charge is not a defence. That the thermometer-builder has a politics does nothing to show that the school's own claims can be checked, and a thermometer really can be checked. The school earned the right to be called testable in the scale, not in the essay. The scale was written down, caught out, repaired by a stranger, and tested again. The school is not exempt from the charge of bias, and never claimed to be. It claims the charge cuts both ways, and that a theory which admits its interest can at least be argued with.

The whole of it comes down to five steps. One: the theory said conditions make consciousness. Two: the conditions arrived and the consciousness did not. Three: so something between them, the person, has to be studied. Four: the person is made in the family before they hold an opinion, so Freud had to be added to Marx. Five: a stable disposition exists, it predicts hostility, and it survived being caught and repaired. What the measurements hold is this much: a disposition can be measured, and it is not the same thing as an opinion. That this disposition is what decides action, when acting costs something, is the school's hypothesis. Every part after this one assumes it on those terms and not as a result.

How a wanting gets made has not been reached at all. If a person's wanting is manufactured, and by nothing as crude as a lie, then something in ordinary modern life is doing the manufacturing. Fromm's answer was that the first suspect is the thing the person is proudest of. That thing is their freedom, and what a person is doing with it is the question the argument has to take up next.

Fromm's answer to the question of how wanting gets made is that freedom itself does part of the work. It does that work not by being false, but by being unbearable in one specific way. Freedom from the old bonds arrives as loneliness long before it arrives as power, and people trade it back to feel held. The old trades were obvious: a leader, an enemy. The modern trade is quiet. You become what is expected of you, and you experience it as your own choice. That last clause is the whole trick, which Fromm found a way to show happening in a single room.

Fromm's demonstration begins with two men, five centuries apart, alike in every way but one. The first makes chairs in a town with a guild. He cannot leave his trade, cannot set his own price, cannot move to the next town and set up shop. His father made chairs and his son will. But if he makes good chairs, Fromm writes, he has done all that is needed to live safely at the level his position assigns him.¹² The guild is built on mutual help.¹³ He has no freedom in our sense, and he also, which is the point, has company at every hour.¹⁴ The security is relative, and it depends on his staying exactly where he was put. Fromm is careful about that limit, and his care with it matters to the argument. A critic would attack the romance of the old town, and the argument does not need that romance.

The second man is hired. He can quit tomorrow, move cities, retrain, sell his labour to whoever pays most, and nobody assigns him a level. And his boss, Fromm writes, has become an abstract figure he never sees. Management is an anonymous power he deals with at a distance, and toward it, as one man, he is insignificant.¹⁵ He can be replaced by Monday. The two men beside each other make the axis visible. The chair-maker is unfree and held, and the hire is free and nobody's. Everything else about them can be swapped and the difference stays.

Fromm's names for the two sides of that difference are freedom from and freedom to, and he says the two are not the same thing.¹⁶ Freedom from is the cut: the bond, the guild, the church, the father, removed. Freedom to is what a person can then actually do with a self that is now on its own. His historical claim is that the gap between the two of them has grown.¹⁷ A child shows the same shape in small, on a scale anyone can check. The child gets freer of the ties that held it back, and at the same moment freer of the world that made it feel safe.¹⁸ Every step out of the family is a gain and a loss in one movement. Modern history, on his account, is that step taken by whole populations, with no one to walk back to.

This is more than a mood, and the reason is in one line of Fromm's. His sentence is that to feel completely alone leads to mental disintegration, the way starvation leads to death.¹⁹ Not sadness. Disintegration. If that is true, then isolation is not a condition a person can wait out. It is a pressure that has to be discharged somewhere, the way hunger has to be. The plumbing in that image is Freud's, and nothing here depends on the plumbing, only on the pattern it describes. A pressure that has to go somewhere will go to whatever is offered. Fromm does not put the feeling of insignificance on freedom alone. He blames both, the bonds that were cut and the market that made sure nothing grew back in their place.²⁰ Capitalism, on his account, made the individual more alone and gave him a feeling of insignificance and powerlessness.

So what does a person do with a pressure they cannot bear and cannot name. Fromm's chapter on this describes three escapes from it. One: find someone to submit to, or someone to command, so that the self is fused to something larger and the loneliness stops. Two: destroy what is outside, since a threat you have removed can no longer make you feel small.²¹ Those two are the ones the 1930s ran on, and everyone can see them. The third is the one he says most normal people in a modern society actually take, and almost nobody sees, because it looks like health.

Fromm's name for the person who takes the third escape is the automaton. What he writes about that person runs like this. The individual ceases to be himself. He adopts entirely the kind of personality offered to him by cultural patterns.²² He therefore becomes exactly as all others are and as they expect him to be. The loneliness goes, because there is no longer a separate self to be lonely. The price is the self. And the reason the price is not felt is that the person is convinced the whole time that the adopted self is theirs, chosen, original. To break that conviction Fromm reached for a demonstration rather than an argument.

A man is put under hypnosis. He is told that when he wakes he will want to read a manuscript he believes he has brought with him. He will look for it and not find it, and will then believe another man has stolen it. He will be very angry with that man in the room.²³ Then he is woken, and the suggestion itself is wiped. He wakes. He wants the manuscript, and when he is asked why, a reason comes to him at once. He searches and does not find it, then turns in fury on the other man, who has done nothing. He has reasons for the anger too, and they arrive without effort. Every one of those reasons is his own, and he believes each as he says it. He would pass any test of sincerity you could give him. And the want, the search, the enemy and the rage were all put into him minutes ago. He cannot see the join, and no test he could run on himself would show it to him.

Two things about that scene carry everything that follows. First, the want came with an enemy attached, because the suggestion did not just install a desire. It installed a story about who was standing in the desire's way. The man's anger at an innocent person felt exactly as justified as anger ever feels. That detail returns when a machine is found feeding a population's anger at the other side. Second, there is Fromm's own conclusion about the scene, which he puts in these words. We can have thoughts, feelings, wishes, even sensual sensations which we subjectively feel to be ours, and which have been put into us from outside. Then the harder sentence, which says this substitution exists to an extent that gives the impression that these pseudo acts are the rule. The genuine or indigenous mental acts, in his words, are the exceptions.

Now the question you should be asking, because it is the right one. If wants can be installed, how would you ever know which of yours were. The hypnotised man could not, and he had all the evidence a person could have. Fromm imagines the objection that will be made about the man. Surely, it runs, he is the one who knows best what he thinks. There is no better proof than his own statement about what he feels. That proof was worthless.²⁴ So the test cannot be "does it feel like mine", because everything feels like mine. Fromm's test runs differently, and it is easy to weaken it by trying to improve it. The question, he says, is whether the thought is the result of one's own thinking, of one's own activity.²⁵ It does not ask whether the content is right, and it does not ask where the thought came from. A true thought can be adopted and a false one can be your own. The test is whether you did anything to get it.

He gives the ordinary version of the same thing, with no hypnotist in the room. A man refuses to lend a friend money and explains that lending would only encourage the friend's carelessness. He believes he is acting out of concern for the friend. He is acting out of stinginess, and the concern was manufactured afterwards to fit.²⁶ Fromm's word for that manufacturing is rationalising, and his definition of it is exact. Rationalizing is not a tool for penetration of reality but a post-factum attempt to harmonize one's own wishes with existing reality.²⁷ The reason arrives after the want, dressed as its cause. That is the join the hypnotised man could not see, and neither can the rest of us, most of the time.

Fromm compresses the whole diagnosis into a single sentence. Modern man, Fromm writes, lives under the illusion that he knows what he wants, while he actually wants what he is supposed to want.²⁸ Set that beside the sentence the cold open began with. You chose your job, your phone, your shows. Nothing in Fromm calls that false. Everything in him says it is what the man with the missing manuscript would have said.

There is an objection here that a lot of people will feel. "By this standard nobody has ever had a thought of their own. Every want comes from somewhere. The chair-maker got his from the guild and the church. You have made ordinary social life into a disease." The answer is in the axis the two men were set on. Neither man made his own wants, and nothing in Fromm pretends the chair-maker did. The difference is in how the shaping arrived. The guild's shaping came as a rule you could name and resent and fight, and people did all three. The automaton's shaping comes as a choice you are congratulated for making, so there is nothing to resent and no rule to point at. The obvious reply is that the old world hid its hand too, behind God and custom. That is granted, and both worlds hid the hand that did the shaping. What survives the objection is smaller and harder than it first looked. A rule was a rule, so you could break it and be punished, and the punishment told you the rule was there. A choice you are told is yours punishes nothing, so nothing ever tells you it was given.

There is a harder objection, and it comes from Herbert Marcuse, the member of the school who will matter most later. Fromm's cure is what he calls spontaneous activity: the free activity of the whole self, whose main parts are love and work.²⁹ Marcuse's charge is that this turns a social fact, made by a market, into a private task. Telling the isolated person to love and work spontaneously accepts the market's premises and criticises the surface. As if a person could practise all of that inside the arrangement and stay sane.³⁰ The concession is that the cure is vague and that it is private. It asks the person to fix by an act of will a condition that was never chosen. The diagnosis stays here and the cure does not. There is also a debt to record. Marcuse's own answer to the same question, when it arrives, will have to be tested for the same failure from the opposite side.

The chair-maker could name what held him, and the hire cannot, because what holds him arrives as his own preference. The hypnotised man gave good reasons for a want that had been put in him minutes before.

One question is left standing, and this part has no answer for it. An automaton conforms to something, and the self he adopts is offered, Fromm says, by cultural patterns. Patterns do not fall from the sky, and somebody cuts them, in quantity, to fit. The hypnotist in the room was a man. Outside the room there is no man. There is a way of thinking instead, and it is the one you trust most and would defend hardest. It is reason.

The kind of reason that got us out of superstition can only answer one question, how. It has no way to ask the other one, what for. And by succeeding completely it turns back into the thing it escaped. That is the centre of everything the school wrote, and the last of those three sentences sounds like a paradox. It is a mechanism with four steps. By the end of this part it should be visible in a hospital corridor.

Horkheimer and Adorno wrote the book that carries this claim, Dialectic of Enlightenment, in exile in California in 1944. Their stated purpose was to explain why humanity, instead of entering a truly human state, was sinking into a new kind of barbarism.³³ Notice what the sentence assumes. The barbarism was no relapse into the past, but a product of the most educated, most scientific, most administered society on earth. The book's wager is that it came out of it because of those things, not despite them.

Start with two cases, and they are an illustration built for this argument, not a scene from the book. A priest in an old city kills a goat and reads its liver. The liver says the harvest will fail, so the city stores grain. Nobody asks the priest why a liver should know about rain. He is the one who reads; they are the ones who act. An analyst in a modern firm opens a dashboard. The number says the quarter will miss, so the firm cuts a department. Nobody in the meeting asks what the number is for, or whether a quarter is a thing worth being measured by. He is the one who reads; they are the ones who act. Everyone can feel the difference, and it is real. The dashboard works. The liver did not. What the two rooms share is easy to miss. In both, there is a power nobody in the room made, that nobody in the room questions, that hands down a verdict, and the room obeys. The first room calls it the gods, and the second calls it the numbers. Applied to those two rooms, the book's claim is that the second is the first with the names changed. The working dashboard is exactly what makes the likeness hard to see.

To see why, the word reason has to be split in two, and Horkheimer did the splitting. One kind of reason is concerned with means and ends, in his phrase. It asks whether a procedure fits a purpose that is more or less taken for granted.³⁴ Call it the counting reason. It can tell you the fastest route, the cheapest supplier, the dose that kills the infection. It cannot tell you whether to go, whether to buy, or whether the patient's life is worth saving, because those are not questions about means. The other kind is older, so call it the old reason, the one that claimed to know the order of things. It held that the world had a structure of its own, and a life could be measured against it. The degree of reasonableness of a man's life, Horkheimer writes, could be determined according to its harmony with this totality.³⁵ That kind of reason asked what for. It also underwrote the priest, and every tyranny that ever claimed the same knowledge. The enlightenment's achievement was to kill the old reason. Its cost is that only the counting reason was left, and counting cannot ask what the old reason asked. Horkheimer's sentence for the result is short. Having given up autonomy, reason has become an instrument.³⁶

Those four steps are the hard part, where the claim gets earned. Step one: to master nature you have to be able to compare things. To compare unlike things you have to turn them into amounts of the same thing. A forest becomes so many tons of timber, and an hour of a nurse's attention becomes a unit of care. The book's sentence is that this society is ruled by equivalence: it makes dissimilar things comparable by reducing them to abstract quantities.³⁷ Nothing sinister is intended by it, and you cannot build a bridge without it. Step two: once that is the standard, anything that will not become an amount looks like a defect. For enlightenment, they write, anything which does not conform to the standard of calculability and utility must be viewed with suspicion.³⁸ Grief, a view from a window, a Sunday with nothing in it. None of those is a quantity of anything, so none of them can be defended in the room where the decisions are made. Step three, and this is where it turns: the person doing the mastering is also made of the stuff being mastered. He has a body that tires, wants that wander, a childhood that is still in him. To master the forest he has to master those too: sleep on schedule, want on schedule, treat his own hours as units. The book says humanity had to inflict terrible injuries on itself before the self was created. Something of this process, it adds, is repeated in every childhood.³⁹ Step four: the self that survives that treatment is a thing that manages itself. And the book's darkest sentence is about what that costs: self-preservation, it says, destroys the very thing which is to be preserved.⁴⁰ The living person was the point of all the mastering. What the managing self manages is that living body, so the preserving spends exactly what it was built to keep.

Those four steps make one shape, and the shape has a name that will be used again. A thing sets out to free itself from something, succeeds, and by the very means of its success becomes what it was fleeing. Reason set out to free us from powers we obeyed without understanding, and it became a power we obey without understanding. The book's own summary is two clauses: myth is already enlightenment, and enlightenment reverts to mythology.⁴¹ It also says why nothing gets out. Any myth raised against enlightenment, once it is used as an argument, has already accepted enlightenment's own principle. Their three words for that result are enlightenment is totalitarian. The phrase is not doing police-state work, and it names a system with no outside. The argument you raise against it is already made of it. The school's word for a thing that turns into its opposite by succeeding is dialectic. That is all the word means here. It has already happened to freedom, in the last part. It will happen again to leisure, and at the end to the critique itself.

There is a passage in the book that puts the whole arrangement into one boat. It is the picture the argument returns to. Odysseus has to sail past the Sirens. Their song is the most beautiful sound a person can hear, and everyone who hears it steers toward it and dies. He knows only two ways through, and for his crew he prescribes wax. He plugs their ears and orders them to row with all their might. They will pass the island in safety because they will hear nothing at all.⁴² The book's gloss follows: anyone who wants to survive must not listen to the song, and can only manage that by being unable to hear. Workers, it says, must look ahead with alert concentration and ignore anything that lies to one side. Thus the workers are made practical. For himself, Odysseus chooses the other way. He has himself tied to the mast, and hears the song in full. The more beautiful it gets, the tighter he has himself bound, so that he cannot act on what he hears. He shouts to be freed. The rowers cannot hear him, and they leave him tied, to save him and themselves.

Told again with the roles named, because the book names them and most retellings leave this out. The one who hears is the landowner, the man who has others to work for him. The ones who row are the ones whose labour gets everyone past the island. Neither is free. The rowers move and cannot hear. The master hears and cannot move. The book puts what happens to the song into one long line. The fettered man listens to a concert, as immobilised as audiences later, and his enthusiastic call for liberation goes unheard as applause.⁴³ Then the last line of the passage: in this way the enjoyment of art and manual work diverge. The workers get the work and no song, and the owner gets the song and no possibility of acting on it. So the song becomes a performance, something you sit through with your hands tied and call culture. That is the arrangement. It will be the arrangement in a cinema, and in a living room. It will be the arrangement in a design meeting where the engineers can hear exactly what their product does, bound to the mast of the quarter.

One declaration, since the passage is read here in one way and the book reads it in two. Horkheimer and Adorno also take the Odyssey as the story of how the self was first formed. The hero is their earliest model of the individual who survives by cunning and self-denial.⁴⁴ That reading stays out of this argument, because it is a story about origins, and origins cannot be checked. The arrangement can. In most large organisations you can see who has wax and who has rope.

The objection comes from every reader of this book, and it is right to come. You are using reason to condemn reason. Every sentence in your book is an argument, and every argument relies on the thing you say is corrupt. You have sawn off the branch you are sitting on. That charge was later made formally by Habermas, a younger member of the same school. He called it a performative contradiction.⁴⁵ The book, in his reading, has to use the critique it has declared dead in the very act of describing its death. The reply goes this far and no further. The book condemns reason cut down to one function, the counting function, and nothing wider. The reply to the counting function cannot itself be a count. A more recent reader of the book puts it harder: the contradiction is the point. The book is reason turning to look at itself, and nothing else could do that job, and the turning is bound to strain.⁴⁶ That answer holds against the charge as stated. It does not yet show that there is a reason that can ask what for without bringing back the priest. That debt is owed, it is named here, and a later part pays it or fails to.

Before leaving, the mechanism has to be seen in one real place, or it stays a theory about a book. England, early 2000s. The health service set targets and attached careers to them. Ambulances: reach a life-threatening call within eight minutes. Emergency departments: nobody waits more than four hours. The reasoning was the counting reason, and it was not stupid. You cannot manage what you cannot measure, and eight minutes is a number a manager can act on. Two researchers collected what the official reviews had found. In a third of the ambulance services, the response times had been corrected. Corrected is the report's word, and it means recorded as under eight minutes. The crews were not necessarily faster. The clocks were.⁴⁷ In the emergency departments the waiting times, when plotted, peaked at the four-hour target. Less sharply than the ambulance clocks, but at the target, not where illness put them.⁴⁸

The targets also worked, and that is what makes the case something other than a story about cheating. The share of patients waiting over four hours fell from nearly a quarter to about one in twenty. Real people were seen faster.⁴⁹ So the metric did two things at once. It improved the thing it stood for, and it ate part of the thing it stood for, and it cannot tell you which part. The two researchers who collected these cases call the flaw taking a part to stand for the whole. Eight minutes stood for the patient being helped in time. The moment the part became the target, the part could be met while the whole was abandoned. From inside the counting room the two look identical. The four-step mechanism is in that corridor, and not only in a book. A number was created to serve a purpose, and it became the purpose. Everyone in the system obeyed it, nobody could ask what it was for, and the question itself sounded like an excuse for being slow.

An economist would push back at this point. The dashboard is answerable to evidence and can be revised, and the liver never was. Better still, the gaming was caught by an audit, and an audit is one more count. So the counting reason corrected itself, and the two rooms are not alike after all. The first half of that is true, and it is the real difference between the rooms. A number that can be checked against the world and then dropped is not a goat's liver. A count did catch the count. What the audit could establish was that the clocks had been wrong. What it could not establish was that eight minutes was the wrong thing to measure. Only a question from outside the counting can establish that, and no such question was asked. The argument here never said dashboards fail. It said the room full of them has no way to raise the question that judges the count itself. The likeness between the two rooms lies in that missing question, not in accuracy.

A sociologist saw the shape coming more than a century ago and gave it an image that has outlived the argument it came from. One Puritan preacher, he noted, had meant the care for possessions to lie on the shoulders like a light cloak. A cloak to be thrown off at any moment. But fate decreed that the cloak should become an iron cage.⁵⁰ A thing built to serve turned into a thing that holds. That is the dialectic in one sentence, and it is the fourth time the shape has appeared in this part alone. Four times in one part is enough to expect it again.

The boat is the whole argument: a crossing timed to the minute, with nobody aboard able to ask why this island and why this song. Wax for the ones who move, rope for the one who hears, and a measure of eight minutes standing in for every reason anyone rows.

What has not been shown is where an ordinary person meets this machine. The natural guess is at work, in the targets and the dashboards, and that guess is half right. But at work a person expects to be counted. They know the hours are not theirs. The place where the counting gets inside is the evening, the one stretch of the day a person is certain belongs to them. That is where the wants themselves are cut to size, after work, with the day's obedience done and the screen on. What that hour does to a person is what the next part has to show.

Your evening is the working day continued by other means. The thing being manufactured in it is you. That sounds like an insult to everyone who has ever enjoyed a film. It is a claim about how the film was made, not about whether you enjoyed it. That difference matters at every step of what follows.

Horkheimer and Adorno gave the machinery a name, and Adorno later explained the name. That is the place to start, because the explanation is the whole argument in one move. They had first written "mass culture". They replaced it, he says, with "culture industry", in order to exclude from the outset the interpretation agreeable to its advocates.⁵² The agreeable interpretation is the one everyone holds. Popular culture is what the people like, risen up from below, given back to them by companies that merely listen. The word industry is there to say the opposite. Nothing rises. It is cut to a shape, and the shape is the subject of this part.

Begin with the sentence that connects this part to the last one. Late capitalism was their phrase for the economy they were living in. Amusement under late capitalism, they write, is the prolongation of work. It is sought as an escape from the mechanised work process, and to recruit strength so as to be able to cope with it again.⁵³ That is a mechanism and not a sneer, and it has four moves. One: mechanised work uses up a specific thing, attention, and gives back nothing that the attention could rest in. Two: the person arrives home with nothing left, so whatever fills the evening has to demand nothing. Three: to demand nothing, it has to be made of shapes the person already knows, so that it can be followed without effort. Four: shapes that are already known are the shapes of the workday, the same rhythm, the same anticipation, the same small reward on schedule. So the thing built to rest the attention exercises it in exactly the pattern work needs. The rowers with wax in their ears go home and put on something that sounds like rowing.

Now the machinery itself, and the book's word for it is schema. There is an old idea in philosophy that a mind does its own sorting. It takes what the senses bring and files it under concepts, and that filing is the mind's own work. Horkheimer and Adorno say the industry has taken that job. There is nothing left for the consumer to classify, they write. Producers have done it for him. Its prime service to the customer is to do his schematising for him.⁵⁴ In a cinema seat it looks like this. As soon as the film begins, they say, it is quite clear how it will end, and who will be rewarded, punished, or forgotten.⁵⁵ You are not deceived about the ending. You are relieved of the work of finding it out, and the relief feels like pleasure, and it is. No independent thinking must be expected from the audience. The product prescribes every reaction.⁵⁶ The sorting your mind used to do for itself has been done upstream, and what arrives is pre-sorted, so the mind has only to receive.

Two hit songs, then, and the book's own examples rather than anyone's taste. The details are interchangeable, they write. There is the short interval sequence that worked in one hit, and the hero's momentary fall from grace. The rough treatment the beloved gets from the male star, his rugged defiance of the spoilt heiress.⁵⁷ These are ready-made pieces to be slotted in anywhere, and they never do more than fill the place assigned to them. The songs, the stars and the serials recur as rigid types, and the content only appears to change.⁵⁸ What makes you feel you chose one over the other is a single detail, and the book has a word for that too. From the standardised improvisation in jazz, they write, to the film star who must have a lock of hair straying over her eyes.⁵⁹ And there, they write, pseudo-individuality reigns. The lock of hair is the difference. Everything under it is the same product. And the lock of hair is not there for her. It is there so that you can tell her apart, and feel that the telling apart was your taste at work. Fromm had seen the same thing three years earlier and put it on a smaller object. Handbags, playing cards and portable radios, he wrote, are personalised, by having the initials of the owner put on them.⁶⁰ The initials are the lock of hair. The automaton is offered a self with his own monogram already on it.

Laughter has to be handled carefully here, because the book is often accused of hating it and does not. It distinguishes two kinds. Conciliatory laughter, they write, is heard as the echo of an escape from power. That is the laugh of people who have got out from under something, and the book has nothing against it. The wrong kind, they go on, overcomes fear by capitulating to the forces that are to be feared. It is the echo of power as something inescapable.⁶¹ You know both laughs. One is the sound of a room where something true has been said about the boss. The other is the sound of a room laughing at the person the boss is laughing at. The second is louder and comes more easily, and it is the one the industry can manufacture, because it needs no truth, only a target. Fun, in their phrase, is a medicinal bath.⁶² And, to add the step they leave implicit, what that bath treats is the ability to want anything else.

Now the sentence the part has been building to, and it needs its two versions because they are twenty years apart. In the book, the consumer is shown all his needs as capable of fulfilment. But the needs are so predetermined that he feels himself the eternal consumer, the object of the culture industry.⁶³ Twenty years later Adorno said it plainly.⁶⁴ The customer is not king, as the culture industry would like to have us believe, not its subject but its object. What is being claimed is stranger than "they sell you things". A factory that makes cars has a product, the car, and a customer who buys it. The culture industry's product is not the film. The film is the tool. The product is a person who wants the next film, in the shape the next film will come in. He experiences that wanting as a taste. The customer is the thing on the assembly line. And what the tool promises, it withholds, by design. The culture industry perpetually cheats its consumers of what it perpetually promises, the book says.⁶⁵ The promise is all the spectacle consists of, and all it confirms is that the real point will never be reached. Their image is the exact one: the diner must be satisfied with the menu. A satisfied diner leaves. A diner with a menu comes back.⁶⁶

The objection that carries most weight was made a generation later, and a lot of it is right. A British media theorist pointed out that a message sent is not a message received. A viewer can take a programme straight, inside the code it was made in. A viewer can negotiate: accept the big picture and quarrel with the details that touch their own life. Or a viewer can understand every word and read the whole thing against the grain, the way a striker watches a news report about the strike.⁶⁷ People are not receivers. They do things with what they are given, and often the opposite of what was intended.

The objection wins three things, and each of them is granted here. The book's sentence that the product prescribes every reaction is too strong. Nothing in the theory predicts what any one person will do with any one film. The striker reading against the grain is real, and the theory has no account of him. What it leaves standing is a different claim. The objection is about the message. The claim is about the form. Every one of the three readings happens inside the schedule, at the hour the schedule set. It happens in the shape the product came in, with the attention the product trained. The striker who reads the news against the grain still watched the news, at six, for the length the news runs. And rebellion inside the schedule is itself a product line, with its own lock of hair. Adorno had already conceded the part of this that is true, in the same essay. The masses are not the measure but the ideology of the culture industry, he wrote. Even though, he went on, the culture industry itself could scarcely exist without adapting to the masses.⁶⁸ The second half of that is a real admission. It adapts to them. It has to. What it does not do is take its measure from them. The word industry was chosen to hold that difference.

There is a second objection, and it has to be conceded rather than answered. In the sentence just quoted, the book files jazz beside a starlet's hairstyle, as one more specimen of standardisation. That is a sentence by someone who did not know what he was listening to, and the example goes while the argument does not. What survives the concession is everything that matters. The argument is not that these products are bad. It is about how they are made and what the making does to the person who receives them. A product can be made to a schema and be good. The schema is still doing your sorting for you. Adorno's taste contaminated the example, which can be thrown out without touching a step of the mechanism.

There is a measured test, and what it can settle is narrower than it looks. What it tests is the crude version of this claim, the version that says the feed puts opinions into you. That is the version most people hold. It does not measure the claim about form, and nothing in the numbers should be heard as measuring it. In the autumn of 2020, more than twenty thousand people agreed to have their feed on the largest social platform changed for three months. Their ranking algorithm, the machine that decides what they see first, was switched off and replaced with a plain list in time order.⁶⁹ Say first what anyone would expect. The algorithm is the schema, the sorting done upstream. With it gone, people should see a more honest world and, if the culture industry is right, begin to think differently. The three months produced this. People spent much less time on the platform.⁷⁰ Those on the plain list saw more political content, not less.⁷¹ They saw about two-thirds more from sources rated untrustworthy. They saw nearly half as much that was uncivil. And after three months, their political attitudes, their polarisation, their knowledge, and what they did offline had not measurably changed at all.⁷² A companion experiment cut people's exposure to like-minded sources by a third for the same three months. Same result. Nothing moved.⁷³

Compare that against the guess, in both directions, because both are instructive. First, the experiment refutes that crude version, the one most people hold. The feed does not make your opinions. Three months without the machine, and people believed what they had believed. Anyone who says the algorithm is putting thoughts in your head has this result to answer, on this scale and this timespan. Two limits belong to that finding, and the first is the authors' own. Effects further down the line, they allow, might require a more sustained intervention than three months. The second comes from other researchers. The plain list was compared against a feed that was not normal, because the company was making emergency changes to its ranking then. So the comparison is to that machine, in that season, and not to the machine on an ordinary day.⁷⁴ Second, the experiment also showed what the machine had been doing to that material. With the ranking on, less untrustworthy material reached people, and with it switched off, more did. So the ranking had been holding that material down. The emergency changes named a moment ago are the likeliest reason it was. That makes even this result a fact about that autumn, not about ranking as such. What nobody chose was the total shape of what the machine optimised for, and that holding down was one small part of it. That total shape is the counting reason running at the scale of a country.

The ruling has to split the claim in two, and the experiment is what separates them. As a claim about content, that the products put beliefs into people, it loses on this evidence, and this argument lets it go. As a claim about form, that the products train the attention and shape the wanting, it is untouched. The experiment measured opinions for three months, and at neither end does it reach the form claim. That leaves the form claim, for now, exactly where Part 1 said a claim must not stand. It fits every result, and a claim that fits every result is a story. What keeps it from being one is that it can be tested, and so far it has not been. No study yet measures directly whether leisure media trains the attention that work requires. A later part leaves the form claim there and tests the wanting claim instead. People were kept off the platform altogether for a month, and what they wanted afterwards was measured.

The evening repeats the shape of the working day, the sorting arrives already done, and the difference you choose by is a lock of hair. The test took the opinions out of that claim and left the rest standing, untested and testable, with the machine optimising for something nobody chose.

One objection has not yet been heard, and it comes from inside the school, from the man Adorno admired most. Everything so far has treated the dulling as the work of the industry's products. One question has gone unasked: whether a machine that copies images by the million has to dull anyone at all. Walter Benjamin looked at the same machine and saw the opposite possibility. The camera, he thought, might teach a crowd to see.⁷⁵

Benjamin's claim runs like this. A machine that can copy an image a million times destroys something real in the original, and he called that something the aura. The destruction can go two ways. It can teach a crowd to see, or it can turn politics into a show, and nothing in the machine decides which. What decides is who owns it. That last clause follows from a sentence of his that comes later in this part. If he is right, the culture industry is a fact about ownership and not about cameras. The last part left open whether the copying machine itself does the dulling. Benjamin answers that it does not, because the dulling is a fact about owners.

Begin with the thing he says is lost, because the word has been worn smooth. It is easiest to see in two objects. A painting of the Madonna in a side chapel, in a town where it has hung for four hundred years. It is covered most of the year, and people do not go to look at it. They go to be near it, and its power does not depend on being seen at all.⁷⁶ The same image, printed, on a wall in a flat, one of thousands of copies. Anyone can look at it as long as they like, and nobody has ever knelt to it. Benjamin's word for what the first has and the second lacks is aura. His definition is exact: the unique phenomenon of a distance, however close it may be.⁷⁷ You can stand a foot from the painting in the chapel and it is still far from you. It belongs to a place and a history, and you do not. Nearness is the whole of what the print has to give you.

The distance comes from ritual, he says, and the earliest images were not made to be looked at. The elk painted on a cave wall was an instrument of magic.⁷⁸ Certain Madonnas, he notes, remain covered nearly all year round. An image made for a rite has what he calls cult value, and it works whether or not anyone sees it. An image made to be shown has exhibition value.⁷⁹ The whole history of art, on his account, is the slow shift from one to the other. Reproduction finishes the shift. The technique of reproduction, he writes, detaches the reproduced object from the domain of tradition.⁸⁰ What withers in the age of mechanical reproduction is the aura of the work of art.⁸¹ He refuses to mourn it, and that is the surprise. The aura was the distance between an image and the people, held in place by a priesthood. What belonged to a chapel, once the copy closes that distance, belongs to anyone with a wall.

The case where he shows it happening to a person is the film actor, and it is where the essay bites hardest. A stage actor performs to a room. He adjusts to it, feels it, and the room feels him. His aura, Benjamin says, is tied to his presence, and there can be no copy of it.⁸² The film actor performs to a machine. The audience that will judge him is not in the room and will never be in any room with him. A playwright Benjamin quotes says the film actor feels as if in exile, exiled not only from the stage but from himself.⁸³ His reflected image, Benjamin adds, has become separable, transportable. It can be cut, sold, and shown in a town he has never heard of, while he is asleep.⁸⁴ That scene sits directly on top of the last part, on the pseudo-individuality of the star and the lock of hair. Benjamin is describing that same person from inside, as a man whose presence has been taken off him and put into circulation.

He did have a hope, and it should stand in full before Adorno answers it. He compares two pairs. A healer lays hands on a patient and keeps a distance from the body. A surgeon cuts into it. The painter, he says, keeps a natural distance from reality. The cameraman penetrates deeply into its web.⁸⁵ A camera can slow a hand, enlarge a face, cut from a street to a room in an instant. It shows people the inside of the world they walk through, in a way no painting could. And it shows it to everyone at once, in a new way of looking. The old way, before a painting, was contemplation, one person absorbed and still. The new way, in a cinema, is distraction, a crowd half attending, taking the thing in through habit rather than effort.⁸⁶ It is the way you learn a city by walking it rather than studying a map. He does not think this is a decline. The public, in his sentence, is an examiner, but an absent-minded one.⁸⁷ With regard to the screen, he writes, the critical and the receptive attitudes of the public coincide.⁸⁸ The same crowd that recoils from a difficult modern painting laughs with a comedian on screen, and in laughing is judging, and knows it is.

He adds a second observation about the crowd, and the two do not come from the same place in his essay. He notices that the press has turned readers into writers: at any moment, he says, the reader is ready to turn into a writer.⁸⁹ His evidence is newspaper boys leaning on their bicycles arguing about the result of a bicycle race, with an expertise nobody taught them.⁹⁰ Any man today, he says, can lay claim to being filmed. The newsreel offers everyone the chance to rise from passer-by to extra.⁹¹ That argument is about who gets to make images. The argument about the absent-minded examiner is about who gets to judge them. They are two arguments, in two places, and in the version of the essay used here Benjamin does not weld them. Welded, they make the hope this part is testing. A crowd that makes its own images and judges them by habit could be a public that no priesthood, and no industry, can stand between.

Adorno read the essay in draft and wrote back, and the letter is the objection. The laughter of the audience at a cinema, he wrote, is anything but good and revolutionary. Instead, it is full of the worst bourgeois sadism.⁹² The wrong laugh from the last part turns up here, in the room Benjamin was hopeful about. But the letter's real sentence is about something wider than the cinema, and it is the opposite of the snobbery Adorno is usually accused of. He puts high art and popular film side by side. Both, he writes, bear the stigmata of capitalism. Both contain elements of change. Both, in his phrase, are torn halves of an integral freedom, to which however they do not add up.⁹³ He is not defending the concert hall against the cinema. He says the concert hall is also a wound. The serious composer bound to the mast and the crowd with wax in its ears are two halves of one broken thing. You cannot glue them back into freedom by preferring one. The objection to Benjamin is not that he loved the wrong art, but that he thought one half could be whole on its own.

There is a harder objection than Adorno's, made by a later reader of Benjamin, and it comes from Benjamin's own last page. This reader's argument is about shock. A body hit by too much, too fast, defends itself in one way, which is to go numb. The purpose of the defence, in her words, is to numb the organism, to deaden the senses, to repress memory.⁹⁴ Applied to the cinema crowd it says this. The absent-minded examiner may not be learning to judge, and may be learning not to feel, which from the outside looks the same. And Benjamin knew it. His essay ends with a warning that reverses its own tone. His warning is about the masses, and in his words fascism seeks to give them an expression while preserving property. All efforts to render politics aesthetic culminate in one thing: war.⁹⁵ And the last line, which nobody who has read it forgets. Humanity's self-alienation has reached such a degree that it can experience its own destruction as an aesthetic pleasure of the first order.⁹⁶ That is the crowd he hoped for, trained by distraction. Taken to the cinema by the people who own the cinema, and shown the war as a film. The machine that could teach them to see could also teach them to watch, and what settled the difference was who owned it.

So the charity has to be withdrawn in the right place, and only there. Kept from Benjamin: the outcome was never in the machine. A camera can penetrate the world or stage it, and the difference is who is holding it and to what end. Kept, too, is the sentence about property. The masses have a right to change property relations, and fascism seeks to give them an expression while preserving property. If that expression can be handed over while property is kept, then what property needs decides what the crowd is shown. A rally and a Saturday film are the same move at different volumes. A crowd is given something to feel, by whoever owns the means of showing it, on terms that leave the ownership alone. That is what the clause about who owns it means. It is the culture industry idea at its sharpest, because it says what the industry is for. Dropped from Benjamin: that the crowd becomes a critic on its own, by habit, in the dark. On the reading of his own last page that his later reader gives, the crowd learned to watch rather than to judge. The habit trained in the dark, on that reading, was the habit of watching. And it went that way for the reason Benjamin's sentence points to rather than the reason Adorno gave. The cause was not a base crowd, as one reading of Adorno's letter would have it. The people who owned the copies needed a crowd that would buy, and the step from there is this argument's own. None of them had to want dullness. They needed sales, and what demands nothing of a person trains nothing in them. An attention that is never asked for anything after work is the easiest kind to sell to. An interest does the work a plan would without anyone sitting down to draw it. That is the sense in which there is no villain in this story and still an owner.

The chapel's distance dies in the print on the wall, and that death is a liberation before it is anything else. What comes after it is the actor talking to a machine and the crowd laughing in the dark. Whether that crowd is judging or going numb is not settled by the camera. It is settled by whoever owns the schedule and needs the crowd to be one thing rather than the other. The last part said the product is the customer. This one adds the missing word: the industry makes the customer its owners' balance sheet requires.

So the question moves off the machine and onto the wanting. If the camera does not decide and the owner does, the owner still has to work with what the crowd already wants. Or with what it can be made to want. Marcuse looked at that wanting, in the richest society that had ever existed, and said it was where the trap had finally closed. Not because people were miserable. Because they were happy.

Marcuse's sentence for the normal person in a rich society is the place to start, because it removes a misreading before it can form. Repression, in what follows, means the wants a person has learned not to have. In the normal development, he writes, the individual lives his repression freely as his own life. He desires what he is supposed to desire. His gratifications are profitable to him and to others. He is reasonably, and often even exuberantly, happy.⁹⁸ The last word is the one that carries the argument, and the word is happy. Nothing in what follows says that people are secretly miserable. It says that a society can be unfree while satisfying its members, by supplying wants whose satisfaction keeps the arrangement going.⁹⁹ And the mark of that is a specific loss, the ability to think of what is not there.

The picture at full size runs in his own words. A comfortable, smooth, reasonable, democratic unfreedom prevails in advanced industrial civilisation, a token of technical progress.¹⁰⁰ Every adjective in that sentence is a compliment, and the noun is unfreedom. That is the shape of the claim. Nothing has been taken away, everything has been given, and the giving is itself the mechanism. The people recognise themselves in their commodities, he writes. They find their soul in their automobile, hi-fi set, split-level home, kitchen equipment.¹⁰¹ The list is dated and the point is not, because your own objects go into it just as well. The man in the hypnosis room supplied reasons for a want that was planted. Marcuse's man goes further than that, because what he supplies is not a reason but a self. He looks at what he owns and sees who he is, and the looking feels like recognition, because that is what recognition feels like. The result, in the phrase from his book One-Dimensional Man, is euphoria in unhappiness. Both halves hold at once, because the unhappiness is the condition and the euphoria is the feeling. The feeling is real, and what it covers is the condition that keeps producing it.

The objection everyone raises comes next. "People chose those things. Nobody made them buy the car. There were fifty models, and they picked one." Marcuse's reply is one sentence and it should be heard exactly. Free choice among a wide variety of goods and services, he writes, does not signify freedom if these goods and services sustain social controls. Social controls over a life of toil and fear.¹⁰² The number of options is not the measure of freedom, and what those options sustain is. Fifty cars, all of which require the job that requires the car, is not a choice about the job. The range is wide enough inside the job that the job itself stops being visible.

The loss is easiest to see in two people who are alike except in one respect. The first is asked what they want and answers instantly: this one, not that one, the newer version, the one in the advert. Every want is a want for one of the things on offer. The second person can do all of that, and can also ask a question the first cannot form: why any of them. Not "which", but "whether". Marcuse calls the thing the second person has a second dimension, and one reading of that phrase is declared here. In his book the phrase belongs to culture, to the art and thought that once stood against the given world and said it could be otherwise.¹⁰³ His claim is that this culture has been flattened, absorbed, made into one more product. The phrase will be used here for the person, for the ability to think of what is not, because that is where the loss lands. Where the book puts a culture that could say no, the rest of this argument puts a person.

What replaces the second dimension, he says, is something he calls the happy consciousness. The belief that the real is rational and that the system delivers the goods.¹⁰⁴ It does deliver them, and that is the point rather than an awkwardness for it. A man objects that the arrangement leaves him no time of his own. What comes back is a cheaper model, then a faster one, then a version that saves him an hour a week. The objection is not refuted, it is filled, and a filled objection stops being an objection. The line he writes here closes a loop opened three parts ago. This universe of discourse, he writes, closes itself against any other discourse which is not on its own terms.¹⁰⁵ That is the older book's sentence, enlightenment is totalitarian, said again about pleasure.¹⁰⁶ A system with no outside, not because it forbids the outside but because it converts everything that arrives from there. Marcuse's line for what that conversion does to one kind of pleasure is five words long. He is writing about the loosening of restraint on sex, and not about pleasure in general. Pleasure, thus adjusted, generates submission.¹⁰⁷

The whole argument now turns on a distinction, and the distinction is where Marcuse is weakest. He divides needs into true and false. False, he says, are those which are superimposed upon the individual by particular social interests in his repression. The needs which perpetuate toil, aggressiveness, misery and injustice.¹⁰⁸ That definition contains two different tests, and the two have different fates. The first test is about origin: where did the need come from, and whose interest put it there. The second test is about effect: what does satisfying the need keep in place. A need can pass one and fail the other. Take the want for a bigger car. The origin test asks who put it there, and finds the advert and the neighbour. The effect test asks what satisfying it keeps in place, and finds the loan and the job the loan needs. Against both he sets the vital needs, the only ones with an unqualified claim to satisfaction: nourishment, clothing, lodging, at the attainable level of culture.¹⁰⁹ The last six words matter and will be needed shortly.

The break comes here, and it is the objection this part has to answer rather than manage. Who decides which needs are false. If the person's own wants are what the system shaped, the person cannot be the judge. If someone else is the judge, then someone else is deciding what you really want, and that is the oldest excuse for tyranny there is. Marcuse saw the trap and answered. In the last analysis, he writes, the question of what are true and false needs must be answered by the individuals themselves. But only in the last analysis, that is, if and when they are free to give their own answer.¹¹⁰ The judge is the free individual, and who is free is exactly what is in dispute. So the test cannot be run until the thing it was meant to test has already been settled. It is a court that may sit only once the accused has been declared free, by that court. The sentence defers the judgement to a condition it cannot certify, and marks the deferral twice with the words in the last analysis. Whether that is a circle or an honest deferral is a question for a later part, when it lets someone else make the same move.

The origin test carries a further consequence, and it is worse than the deferral. If false needs are put in by someone else's interest, then the person who has them is by definition unable to see them. A majority in that state cannot say what it truly needs, because the wanting that would say it is the thing under suspicion. Someone who can see has to speak on their behalf. The room is a few people who count as clear-sighted, deciding what a town should want, with the town not present. The name for those few is a committee. The committee is this argument's inference from his origin test, not a proposal Marcuse made. It is used from here on because nothing names the thing better. He saw it coming and refused it in so many words. No tribunal can justly arrogate to itself the right to decide which needs should be developed and satisfied, he writes. Any such tribunal is reprehensible.¹¹¹ He did not want a tribunal. The origin test builds one whether he wanted it or not. He left himself nothing that could stop it. This is the failure Part 2 said his answer would have to be tested for. Fromm's cure was private, and it failed by asking one person to repair a social fact. Marcuse's cure is public, and it fails from the other side, by appointing a judge. The theory refuses the tribunal and cannot do without one.

And Marcuse followed the consequence into practice, so the failure can be seen rather than inferred. In an essay on tolerance he proposed what he called liberating tolerance. It would mean, in his words, intolerance against movements from the Right and toleration of movements from the Left.¹¹² He listed what would lose the protection of free speech and assembly, and the list runs a long way. Groups that promote aggressive policies, armament, chauvinism, discrimination on the grounds of race and religion.¹¹³ Or which oppose the extension of public services, social security, medical care. The last clause silences anyone who thinks public services should not be extended. Then he asked his own question, who is qualified to draw these lines for the whole society, and gave the answer he had given before. Everyone in the maturity of his faculties, everyone who has learned to think rationally and autonomously.¹¹⁴ Marcuse's answer to who decides fails in exactly the way that inference predicts. The free thinker judges, and who counts as a free thinker is what was in question. The credit given to Marcuse's diagnosis does not extend to this. A right diagnosis ends by handing itself to the committee the origin test made necessary.

So what is left standing. More than the objection thinks, and the surviving parts need naming one by one. The origin test dies, because you cannot judge a want by where it came from. Every want came from somewhere, and the judge of origins ends up as the committee. Fromm's test was already better: not where the want came from, but whether you did anything to arrive at it.¹¹⁵ The effect test lives, and it lives because it can be turned into something that can be checked. Marcuse gave the outside version of it himself, in the six words held back earlier. Needs are measured at the attainable level of culture. The measure is not a picture of the good life but what this society could do now and does not. His earlier book, Eros and Civilization, says it plainly. The real issue is the awareness that people could both work less and determine their own needs and satisfactions. That is not a utopia but a gap between the possible and the actual. A gap can be measured without anyone deciding what you really want. A town whose factory could run four days a week and runs five is such a gap. Whether it could is a question about machines and hours, answered with a ledger, and not a question about what anyone should want. The gap is drawn in this society's own units, in hours and in output, and those units are not innocent either. A later part presses that, and still keeps this picture, because it is the minimal kind: the society's own promise held against its own performance.

There is an inside version too, and it is the one that will be put to a test three parts from now. A need is suspect when the person, given a real rest from it, would rather be without it. The question comes not in the abstract, where everyone says the right thing, but after the rest has been had. That test does not need a committee. It needs the person, and time, and it lets the person be the judge without pretending they were free to begin with. What the test does is flag a want, and flagging a want is not convicting it. Coffee would be flagged, and exercise, and a marriage in a hard month. The opposite result is what clears them, because wanting a thing more after a real rest from it is the answer that acquits. In the study it will be put to, the want shrank and did not vanish, which is a flag and nothing more. It is tempting to say a false need is one that grows when it is fed, like a thirst that salt water makes worse. The best evidence for it, the idea that every satisfaction fades back to a fixed baseline, has failed under testing. So the image stays out, and the rest-from-it test carries the load alone.

A society can be unfree while satisfying its members, and the man who finds himself in his commodities asks which and never whether. Marcuse's judge was a free individual he could not produce, so the origin test dies and two survive: the gap, and the rest-from-it test.

So the deepest problem the school ever faced is now unavoidable. Marcuse could not say what people really want without becoming the committee. Suppose no one can, and every picture of the good life is drawn with the pencils this society hands out. Then anyone who tells you what you should want instead is selling the next product. Then critique has no ground to stand on. Adorno accepted the first half of that and refused the second. No picture of the good life can be drawn from inside this one, and he never pretended otherwise. What he held is that critique keeps a ground anyway, a ground that can only say no. The next part is about what that ground is, and about what a critique that can only refuse is good for.

The ground that can only say no belongs to one man, and he spent his life defending it. Adorno's most quoted sentence is usually heard as despair, and the book it comes from opens by naming its subject. Minima Moralia opens with a dedication. The book relates, he writes there, to a realm which has counted, since time immemorial, as the authentic one of philosophy. The dedication mourns what became of that realm, and then it names the realm.¹²⁰ The teaching of the good life. A man who wrote a whole book about the good life did not think there was nothing to say about it. He thought there was only one honest way left to say it. His position comes in two halves. You do not need a picture of the right life to know a wrong one. And any picture you draw is drawn in the wrong life's colours, so the honest thinker refuses to draw it and says no instead. Whether saying no is enough is what this part has to rule on.

The sentence itself goes like this. There is no right life in the wrong one.¹²¹ It can be read two ways, and one is declared here and the other set aside. The first reading: no act of yours can be good, so stop trying. That is the despairing version, and it is not what the sentence says. The second reading: the conditions for living rightly are not present. So whatever you do, however decent, is done inside an arrangement that makes it less than it could be. A kind act in a wrong world is still kind, and it is also compromised. Compromised by where the money came from, by who is not in the room, by what the kindness lets continue. That is the reading used here, because it is the one the rest of his work supports. The first reading would make a book about the good life a contradiction.

Why he thinks every picture is contaminated turns on what he thinks thinking is. Two diagnoses, in a scene built for this argument rather than taken from him. A child is struggling in school and a doctor gives the struggle a name. The name explains something, gets the child help, and the child grows up and the name is one fact among many about them. A second child gets the same name. This time every teacher reads the child through it. What the child does that fits the name is noticed, and what does not fit is not seen. By fifteen the child has become the name, because that is all anyone was looking for. Same word, two fates. In the first case the concept served the child. In the second the concept replaced the child, and the part of the child that did not fit the concept went missing. In Negative Dialectics he names what happened there, that to think means to identify. A concept says this is one of those, and it has to, because that is what a concept is for. And every time it does, something about this particular thing that is not one of those gets pushed aside.¹²² His German makes one phrase of it, and a plain rendering runs like this. The conceptual order, he writes, complacently pushes aside what thinking wants to grasp. That is the schema from the culture industry, turned inward. Adorno's claim is that the concept does the same thing to whatever it touches. The leftover, the part that did not fit, is where the actual person was.

Apply that to a picture of the good society and the refusal follows. To draw the picture you need concepts: freedom, community, work, rest. Every one of those words was given its present meaning by the society you are trying to think past. Freedom, in this society, means a choice among what is offered. Community means a market segment. Rest means the medicinal bath. A picture of a better world drawn in those words is a picture of this world with the lighting changed. It is a product by next season. The committee is what happens when someone draws the picture anyway and then needs to enforce it. The last part left one standard standing, the gap between what a society could do for its members and what it does. Adorno would say that standard is a picture too, drawn in hours and output and the society's own accounting. This argument keeps it anyway, as the least contaminated picture available, the one Part 6 kept for the same reason. It admits that a picture is what it is. So Adorno refuses to draw one. The refusal does not come from despair. It comes from watching every positive picture so far get sold back to the people it was drawn for. The reason given here is the one his own argument supports.

But a refusal to say what is right needs a way of knowing what is wrong, or it is just silence. The standard he leaves turns on words that are easy to drop, so it comes over verbatim. The need to give voice to suffering, he writes, is the condition of all truth.¹²³ Not suffering. The need to give voice to it. Truth, on this account, is answerable to a demand. Something hurts and wants to be said. A thought is true to the degree that it lets that be said, rather than filing it under a concept where it disappears. That is a standard, and it can only say no. But it can say no with precision, to this arrangement, this wrong, this child who became a name.

A standard that only says no still has work to do, because the good gets known as the shape left by specific wrongs. The illustration that follows is this argument's invention, not something he wrote. A woman falls and her leg is bent where legs do not bend. Nobody in the room needs a theory of health to know something is wrong. They know the wrong directly, from the leg, and they can act on it without a definition of the healthy human being in their pocket. The objection arrives at once. Medicine has a theory of health. It knows what a leg should do, and that is how it knows the bent one is broken. Behind every no there is a hidden yes. The reply concedes something and keeps something. Medicine has had many pictures of health. On the account this argument accepts, the ones that lasted were built out of the injuries rather than drawn before them. Its working account of health is the accumulated memory of what went wrong, organised. That is Adorno's position, stated in a hospital. Anyone who claims to have the picture in advance is drawing it again. And the objection wins something here, which this argument concedes. If the yes accumulates from the noes, then pictures can eventually be drawn, and the ban on drawing them is prudence rather than principle. That is how this argument reads him, though he would have resisted the reading.

The charge from Part 3, that the book saws off the branch it sits on, lands hardest here. Adorno refuses any positive standard, and that refusal is what makes the charge sharpest at this point. Adorno never answered Habermas. He died in 1969, before the charge was made in that form. The nearest thing to an answer sits in an essay from his last year, and it can be read instead of guessed at. The uncompromisingly critical thinker, he wrote, neither signs over his consciousness nor lets himself be terrorised into action.¹²⁴ He is in truth the one who does not give in. That refuses the charge's premise rather than refuting the charge. Habermas asks what he stands on. Adorno answers that the demand to stand somewhere is the demand for a picture, and he has said what pictures become. Whether that answer is enough cannot be settled by the two of them. It has to be settled by watching what the position does when it is tested.

It was tested in the last year of his life, and the test is one case and not a biography. On 31 January 1969 students occupied his institute. Adorno had them removed by the police. His own account of it, in a letter to Marcuse that February, is flat.¹²⁵ We had to call the police, who then arrested all those who they found in the room.¹²⁶ Marcuse, in America, was on the students' side, and his reply says where he stands. If the alternative is the police or left-wing students, then I am with the students. With one crucial exception, namely, if my life is threatened or if violence is threatened. The two men wrote to each other about it for months, and the letters are not what is usually said about them. Neither man denies that practice can sometimes push theory further than theory gets on its own. Their quarrel is about the moment, and Adorno denies that 1969 is such a moment. He thinks the students' action is not action but the appearance of it. In an essay from the same year he had a word for it. Pseudo-activity, he writes, is generally the attempt to rescue enclaves of immediacy in the midst of a thoroughly mediated and rigidified society. When everything is arranged in advance, people stage moments that feel unarranged, and the staging is itself one more product. The sit-in as lock of hair.

There are two halves to what the episode shows. Adorno's position held. He did not draw a picture, did not pretend the students had one, and did not let himself be terrorised into action he thought was false. The position is coherent. A thinker who has spent his life showing how the good gets sold back is entitled to refuse to sell one. And the position cost him something specific. He could say what was wrong with their action, in detail, and with a word for it. What he would not say was what right action would look like, because his whole position forbade him a picture of it. So the students got a diagnosis and no direction, and the police got the room back. He could rank the two wrongs when it came to it, and the letters show him doing exactly that. The occupation, he wrote, left no other course. What no ranking of wrongs supplies is the direction the students had come for. That is the gap, and it is not a failure of nerve. It is built into the choice to refuse pictures. That price belongs in the account rather than out of sight behind the refusal.

Both halves together settle what the account can and cannot do. Adorno's account is coherent, and it is not incomplete in the way its critics say. You can know a wrong without a theory of the right, the way a room knows a broken leg. The refusal to draw pictures is a lesson learned from what pictures became, not a shrug. What the account cannot do is tell you what to do next, and no ranking of wrongs supplies that. It is a way of seeing that forbids you the comfort of a solution. It turns out that people need more than that, or at least the students did.

The two children and the one name are what thinking does to everything it touches. That is why every picture of the good society comes back as this society's product, and why the standard that survives can only say no. The bent leg is why that no is not empty, and the students are why it is not enough.

The second generation of the school was built to answer the question that leaves open. Everything so far has treated reason as one thing, the counting kind, and found it could only ask how. Suppose that is a mistake about reason. Suppose there is a second kind, one the first generation used on every page without noticing. Suppose it has a standard built into it that nobody can argue without. That is Habermas's claim, and he thinks it changes everything.

Habermas's claim is that the first generation made one mistake, and that the mistake was about reason. They treated the counting kind as if it were the only kind. He says there is another, and it is neither the counting reason nor the old reason that claimed to know the order of things. It lives in the act of arguing, between people, and it carries a standard nobody can argue without using. Call it the reason that talks. If he is right, the machine from the third part is not all there is to reason. There is a ground to stand on, and Adorno was standing on it the whole time he said there was none. That is the rescue, and how far it reaches turns out to matter more than that it exists.

Two people who want different things from the same situation, and two ways of settling it. The first way: one of them has the money, or the strength, or the position, and the other gives in. The dispute ends, and nothing has been agreed. Habermas marks the difference himself. Agreement can indeed be objectively obtained by force. But what comes to pass manifestly through outside influence or the use of violence cannot count subjectively as agreement. The loser complies, and he does not agree, and he knows the difference, and so does the winner. The second way: one of them gives a reason. We should do it my way, she says, because the last time we did it yours, this happened. The moment that sentence is spoken, something has changed that neither person chose. She has invited the other to check the reason, and staked the outcome on whether the reason is good. She has, without saying so, agreed that if the other has a better one, the better one should win. Those two ways are the counting reason and the reason that talks. More exactly, they are two attitudes a person can take. One is aimed at success, and treats the other person as an obstacle or an instrument. The other is aimed at reaching understanding, and treats the other person as someone who can say yes or no to what you have said. Reaching understanding, he writes, is the inherent telos of human speech, and telos means the built-in purpose, the thing the activity is for.¹³⁰ People do not always argue honestly, and the claim does not need them to. When they argue at all, they have already accepted a standard the counting reason does not contain.

What a speaker accepts should be heard as the strange thing it is. Every time you say something to someone as a claim, Habermas says, you make three promises at once, whether you mean to or not. That what you say is true. That you have the right to say it, in this situation, to this person. That you mean it. Truth, rightness, sincerity.¹³¹ The listener can challenge any of the three, and the speaker who made the claim is committed to answering, or to withdrawing it. Nobody invented that rule, because it is not a rule at all but what a claim is. A sound you make that carries none of those three promises is not a claim, but a noise, or a price. A mechanic tells you the brakes are gone. You can ask whether that is true, and he owes you a look under the car. You can ask whether it is his place to say, and he owes you an answer about whose car it is and whose call. You can ask whether he means it or is padding the bill, and he owes you a straight face. Three questions, one sentence from him, and he accepted all three the moment he spoke.

The test that turns this from a description into a standard comes when someone denies it. There is no such thing as a better argument, the sceptic says. Reason is just force wearing a suit, and I can prove it. The proof is where he loses. He has given a reason and invited you to check it. He has staked his claim on its being a better argument than yours, which is the thing he was denying exists. Habermas's sentence for this is sharper than the usual version. The sceptic, as soon as he cites a reason, has accepted a presupposition. He can never convince an opponent of something by resorting to lies. At most, Habermas says, he can talk him into believing something to be true.¹³² Two verbs. You can talk someone into a belief with a lie. You cannot convince them with one, because convincing means their agreeing for the reason you gave, and the reason was false. The liar knows this, and it is why he lies rather than argues. So even the liar is using the standard, and he is using it to know what he is avoiding.

Habermas borrows a list a colleague drew up of what an argument presupposes, and the list is short enough to hear.¹³³ Everyone who can speak is allowed to take part. Everyone may question any claim, introduce any claim, and say what they need and want. Nobody may be stopped from doing those things by force, from outside or from inside. He summarises it for the people arguing. The structure of their communication rules out all external or internal coercion other than the force of the better argument. Set that list against the committee. Habermas's rules say there is no such position to stand in. The people you are deciding for are allowed to take part, to question you, and to say what they want, or the decision was not justified.¹³⁴ He gives the self-refuting sentence himself. In plain form it runs: having silenced the people the decision was about, we convinced ourselves the decision was justified. You cannot say that, because the second half is cancelled by the first. A family sends the teenager out of the kitchen and then decides what is best for the teenager. Whatever they decide, they have not convinced him, because he was not there to be convinced. That is the ground the first generation was missing, and it was under their feet. Adorno's no, said aloud to anyone, already assumed that the person could hear it and answer.

Habermas raises the first objection against himself. No real conversation looks like that. Somebody always has more power, more time, better words. The rules describe a room that has never existed. He agrees. The presuppositions, he writes, are partly counterfactual. In all cases, he says, we have to be content with approximations.¹³⁵ The rules describe no place. They describe what you are committed to the moment you argue, and what you can be held to afterwards. A tenant sits across from a landlord who talks for ten minutes and calls it settled. When the tenant says you did not let me speak, she is pointing out that he claimed to have convinced when he had only prevailed. The standard does its work as a charge.

The same move was refused earlier, and consistency has a price here. Part 6 said Marcuse's judge had to be certified free before it could judge, and left open whether that deferral was a circle. These partly counterfactual presuppositions defer in the same way, to a state of affairs that does not obtain. The difference is where the deferral bites. Marcuse's judge is the court that may sit only once it has declared the accused free, and it cannot sit to do that. That is a circle, because the certification never comes. Habermas's is a promise you made by opening your mouth, and a promise can be read back to you. That is a deferral and not a circle, because the charge can be made without any certification. His ideal is what the arguer has already presupposed by arguing, so the charge can be laid out of his own mouth, afterwards. The charge has its footing, and it does not buy an answer, since the man charged can say nothing and walk away.

Habermas does something the first generation only gestured at. He says where in a society each kind of reason lives. The first is the world of ordinary life, family, friends, neighbours, the place where things get done by people talking until they agree. He calls it the lifeworld. The second is the world of markets and administrations, where things get done by money and by rules, and nobody has to agree to anything. You do not persuade the supermarket, you pay it. He calls those systems, and money and power their steering media, because they steer action without any understanding being reached. There is nothing wrong with that in its place, and nobody wants to reach a shared understanding with the electricity company. The disease he names is when the second world moves into the first. The imperatives of the systems, he writes, make their way into the lifeworld from the outside, like colonial masters coming into a tribal society. They force a process of assimilation upon it.¹³⁶ He calls it the monetarisation of services, relationships and time.¹³⁷ The new conflicts, he writes, arise along the seams between system and lifeworld.¹³⁸ A seam is where two things made separately were joined, and it is the place they come apart. Take two pairs. A family that talks through a problem at the table, and a family that resolves it by an app that scores each member. A friendship that is what it is, and a friendship that has a number on it. In each pair, a thing that used to be done by agreement is now done by a medium that needs none. That is the culture industry's product line, described from the side of what it replaced.

He also gave the history, and it has to be told with his own warning attached. In the coffee houses of the eighteenth century, he says, private people came together as a public. A public, or a public sphere in his phrase, is what a room of strangers becomes when it argues about what concerns them all. They argued about those things, and the argument, in principle, disregarded status. A merchant and a clerk could disagree, and the better argument was supposed to carry the day. Then his warning: not that this idea of the public was actually realised in earnest in the coffee houses. As an idea, though, it became institutionalised, and if not realised, it was at least consequential.¹³⁹ The idea outran the rooms, and the idea is what mattered. What happened to it is the point, because the public that argued became a public that is addressed. Publicity, he writes, once meant the exposure of political domination before the public use of reason. Publicity now adds up the reactions of an uncommitted friendly disposition.¹⁴⁰ The state, he adds, has to address its citizens like consumers.¹⁴¹ The addressed public is the crew with wax, and the few who still argue are at the mast. The rescue describes the same room the first generation described, in its own terms, agreeing with them about what it looks like.

Two objections follow that the rescue does not survive intact. The first is about who was in the coffee house. A later critic pointed out that the public sphere was not a room that failed to include everyone, but a room built by excluding them. Women and the poor were not late arrivals, and their absence is what made the merchant and the clerk feel equal. And to say we will set our differences aside and argue as equals favours the one with fewer differences to set aside. Habermas's own rules concede the force of this, because an agreement reached by silencing people is not justified. The critic's point is that the silencing was not an accident of the room. It was the room.¹⁴² The second objection is the one the whole part has been heading toward. A standard that can only be used as a charge needs someone who can be charged. The person who prevailed by force can be told he did not convince. A system cannot be told anything, because money does not argue, and neither does a ranking algorithm. Habermas's own description says so, since the steering media coordinate action without any understanding being reached. There is no claim in them to challenge, and no speaker to hold to the three promises.¹⁴³ The ground he found is real, and the machine, which never claimed to convince, has nothing in it for the charge to land on.

A later member of the school pointed out that what people fight about first is not fairness in the argument, but whether they were seen. A person can be given every right in the list, allowed to speak, questioned politely, answered with reasons, and walk out humiliated. Insult and humiliation, in his account, are a denial of recognition, and the injury is below the level of who got to talk.¹⁴⁴ That objection adds rather than subtracts, and it explains why people who were heard still feel that nothing about them was. It will matter again when a machine is asked what it owes a person, because a system sees nobody either.

The ruling comes in two halves, and the first half is the contradiction charge answered. Habermas answers the charge from Part 3 and Part 7, and the man who supplies the ground is the man who made the accusation. The branch the first generation sat on was never the counting reason, but the act of saying something to someone who could answer. The second half asks how far that ground reaches, and the objection has already answered it, since it reaches persons and not systems. In the book where he laid this out, the standard stops about there, and Adorno's picture of an arrangement that cannot be argued with survives. Habermas argued later that law is the transformer, his word, through which the standard reaches the systems. He never claimed the standard binds systems directly. Putting them outside its reach was the design, and law was his route back. So the complaint here is not that he failed to reach the machine, but that the route he named does not reach the evening. That is where the two generations stand. The company is reached slowly through law, and the evening is not reached at all.

To give a reason is to accept that a better reason should win, and the sceptic accepts that in the act of denying it. The debt from the third part, though, is only half paid. What was owed was a reason that could ask what for, and the reason that talks answers the contradiction charge instead. The reason that talks says how a what-for question must be settled, among people who can each say no. Habermas does not answer the question itself. So the question is handed back to the person, which is where the last part will leave it.

Both generations now have to be tested on the same case, and there is a case neither of them saw. A machine that sorts what several billion people see, every hour, by a number nobody chose, with other signals added whose weights are not published. It answers no argument. It is the counting reason at its largest scale so far. It is also the thing the standard cannot touch, and the wanting Marcuse said had been captured, all in one object. The next part puts the numbers on the table.

The present is the best case the school ever had for its diagnosis, and the hardest for either cure. Adorno never saw the object in your hand, and Habermas wrote about it only late in life. Neither the diagnosis nor the rescue was built with it in view. That is what makes it a fair test, because nobody built it to prove a theory. It was built to hit a number.

Begin with what the machine is, without the mystique. Several billion people produce more than anyone could look at, and something has to decide what each person sees first. The something is a ranking system, trained on what people do. Did they stop, tap, share, reply, stay. Whatever they engaged with, show more of it, to them and to people like them. Other signals are added beside that one, including, on some platforms, surveys asking a sample of users what they wanted to have seen. How much each signal weighs is not published. The machine does not manufacture what you see, because the users make all of it themselves. Benjamin's hope that every reader would turn into a writer came true here, literally, and it is the business model. Habermas, in his nineties, said the same thing about the platforms in his own words. They create a space of communication in which readers, listeners and viewers can spontaneously assume the role of authors.¹⁴⁷ The sorting still happens upstream, so the schema survived the very thing that was supposed to abolish it. The fourth part hung its mechanism on a broadcast schedule, and the feed has no schedule at all. The shape survives that, because the sorting is still done before you arrive, whichever hour you pick.

Set that against the third part. The counting reason asks how and never what for, and here it is at the largest scale ever built. Engagement is a how. The question of what is worth a person's attention does have a slot in the system, and the slot is that survey. A sample of users is asked, and nobody outside the company can see what their answer weighs. Nobody forbade the question, and nobody had to forbid it. A study that asked users what they preferred found that the ranking did not follow what they said. The what-for is present and outranked, which is the counting reason working exactly as the eight-minute clock did. The engineers at the mast can hear exactly what the product does. Some cut the rope and leave, and the boat sails on. The users are the rowers with wax.

Now the measurements, from the studies themselves, with what each one does and does not show. First, what the machine selects. A team with access to one platform's own data compared two groups of its users. One saw the feed the ranking system built, and the other a plain time-ordered list. Across seven countries, the mainstream political right was amplified more than the mainstream left in six of them.¹⁴⁸ The authors checked whether the party a politician belonged to explained the tilt, and it did not, and they declined to name a cause. What that measures is reach, and reach only, and not whether a single mind was changed. A tilt in what one platform's users saw, in six of seven countries, that its own makers could not explain. The dashboard from the third part handed down a verdict that nobody in the room ever made.

Second, what the selection is made of. A separate experiment showed a few hundred users two versions of the same slice of their feed. The posts an engagement ranking would put in front of them sat beside the posts a time-ordered list would. Then the researchers asked which of the two they would rather have been shown. The engagement ranking amplified posts that were more partisan and expressed more hostility to the other side. Fromm's planted want came with a planted enemy, and here the enemy is what gets amplified, across a whole platform. And when asked, users were less likely to prefer the political posts the engagement ranking had chosen for them. The authors put it in one sentence: the algorithm seems to be catering to users' revealed preferences, but not to their stated preferences.¹⁴⁹ That sentence is Fromm's manuscript, measured. What a person does and what a person says they want came apart, and the machine served the doing. The study cannot say which of the two is the person, only that the two came apart and the machine served one. Marcuse's problem, who gets to decide which, is not solved by any of this. The study does show that the split is real, and that the counting reason serves the side it counts by default. An older study had already found the raw material, and in its first test the messages were about gun control. Adding a single word of moral emotion to one of them raised its chance of being passed on by about a fifth. The size of that finding has been disputed since.¹⁵⁰ The appetite is human, and the machine did not invent it. All the machine found was the metric that harvests it.

Third, and this is the test the sixth part promised, run on the claim about wanting and not on the fourth part's mechanism. Several hundred people were paid to switch off their account on the largest platform for the four weeks before an election. They came out of a few thousand who took part, and they were compared against people who kept the account. Most people expect two things at once: that they would feel better without it, and that they would rush back the moment it was allowed.¹⁵¹ The people who switched off gained about an hour a day. They spent it partly with family and friends and partly watching television alone, so call it freed, not redeemed. They knew somewhat less about the news, and they were less polarised. A larger repeat of the study around a later election did not find that effect on polarisation.¹⁵² They reported being a little happier, and the effect was small, a fraction of what a good therapy produces. When the four weeks ended, and this is the part that matters here, they did not rush back. Weeks later they were still using the platform noticeably less than the people who had never left. Asked how much money it would take to give the platform up again, they named less than before, up to a seventh less. Having been without it for a month, they valued it less than they had before.

The sixth part's rest-from-it test has now been run. These people, given a rest, wanted less of the thing, used less of it, and priced it lower. The authors read that two ways, and both readings are kept here. Either the habit was broken and did not fully re-form, or people learned, by trying it, that they were better off with less. In both readings the wanting changed because the person got outside it for a month. The judge was the person, and no committee sat. The same paper says that even the reduced valuation was still large. People still wanted the platform, still thought it worth real money, still came back. Nothing in the study says the want was not theirs. The authors' own conclusion is that the platform went on producing large value for the people using it.¹⁵³ So the test raised the flag. It does not show a want that disappears. Anyone who tells you these people were freed has read the abstract and not the table.

The objection that matters is not the one people expect. The expected one is "one platform, four weeks, volunteers", and it is true, and the authors said it first. Those three limits are minor, and there is a fourth thing in the design that is not. The experiment randomised only people who priced the platform under about a hundred dollars a month, the least attached it recruited.¹⁵⁴ The test was run on the people it least needed to reach. The result says less than it seems about the captured wanting this argument is about, because the most captured were never in the sample. The objection people expect least is about size. Two researchers took data on more than three hundred thousand adolescents. They asked how much of their wellbeing could be explained by how much they used screens. The answer was well under half of one percent, and they set it beside other things. In comparable data, eating potatoes explained nearly as much, and wearing glasses explained more. They also showed that there were hundreds of millions of defensible ways to run the same analysis.¹⁵⁵ By choosing among them, you could produce a frightening result or a null one at will. The objection deserves its full weight, and the story that the feed is making everyone miserable is not supported. The effects on wellbeing, where they exist, are small, and the loud versions of the claim are a product of which analysis got published. None of that is in dispute here, and it costs the argument nothing, because the argument never said people were miserable. Marcuse's sentence was that they were exuberantly happy. The claim was never about mood, but about the shape of the wanting and whether the person could get outside it. What the deactivation study measured was that shape, which is not what the potato study ever looked at.

The two cures go on the table now, because that is what the case is for. The inside test built from Marcuse's effect test is the rest-from-it test, and here it flagged a valuation that fell and stayed large. Habermas's standard finds nothing to bite on in the ranking, and what reaches the company is a court, later, through law, and not the evening. The steering medium steers, which is Adorno's point about a system with no outside. The tilt in six of seven countries is the reading on the dial.

So the feed is the counting reason at its largest scale, tilting what people see and serving what they do over what they say. Given a real rest, people want less of it and price it lower. Nothing in any of it says that they were miserable.

One person is left holding the object, having heard all of this. The diagnosis stands, and both cures have been weighed and found real and insufficient. The school offers no programme, and whether that is a failure or the content of the thing is still open. So is what one person can do in an evening with the phone in their hand.

The only practical thing the school has to offer is a habit, and it had better be stated exactly. The freedom left to you is the habit of asking, of any want or evening, the question the counting reason cannot ask. What is this for. It works, to the extent it works, because it is the one act that cannot be done for you. A machine can sort your world in advance, and it cannot ask on your behalf whether the sorting was worth having. Where that question has been given a slot, nobody outside can see what it weighs. Whatever the answer to what is this for turns out to be, it is not the thing a platform counts as engagement.

Advice is what the culture industry sells, and the school was never in that business. Fromm's test of a want asked not where it came from but whether it was the result of your own activity.¹⁵⁹ Applied to the evening, the question is whether this want was the result of anything you did, or whether it arrived already sorted. Sometimes you look for who put the want there and find no hypnotist, only a shape that fitted.

Adorno's image for how this kind of seeing works is the mote, the splinter, in your own eye. It is the least despairing thing he wrote. The mote in your own eye is the best magnifying glass.¹⁶⁰ That is an instruction and not a consolation, because the flaw is the instrument and you do not remove it before looking. There is a place where your own life does not fit the arrangement, where the want feels wrong or the reward feels empty. That place is the one point from which the arrangement is visible, because everywhere else it fits, and what fits cannot be seen. The child who became a name is the case in point. The part of him that never fitted the name was the one place the name could be seen from, and no teacher stood there.

Habermas adds the outward version of the same habit, and it is a demand rather than a question. Whenever something is done to you by a system, ask for the reason. The system will not give one to you, and a court can make it give one, slowly, over years. But the moment you ask, you have located the place where a claim should have been and was not. The ranking never said "this is worth your attention", and the target never said "this patient is better". When you ask them what for and get silence, the silence is the finding. You have found the seam, in his word. The line between the part of life that runs on agreement and the part that runs on a medium that needs none.¹⁶¹ Knowing where the seam is does not move it, and it does tell you what you are standing on. It tells you which of your evening's wants were reached by agreement and which by a number. The faculty that asks what for is the reason that talks, turned around on your own wants. It settles how an answer has to be reached, between people, and it supplies none. The debt stays half paid, and you carry the other half. The question goes to another person and comes back answered with reasons they can push against. That is why the habit is a conversation and not an hour alone with yourself. Planted wants come with their own reasons, so the test is never whether you have one. The test is whether your reason survives another person's questions. One that arrives without effort and collapses when pushed is the manuscript's.

Marcuse's contribution has to be handled with the most care, because it comes closest to a programme and the programme failed. He called for a great refusal, and his definition of it was the protest against that which is.¹⁶² That is a stance before it is a policy, and as a stance it survives everything the sixth part said. What the sixth part rejected was where his politics actually went, the withdrawal of tolerance from one side of the argument. Someone has to be appointed to say which side a movement is on, and that is the committee again. As a habit for one person the refusal is exactly right, and it is close to the same habit. Marcuse did not expect even that much to be safe. His own verdict on his own slogan sits in the same chapter, where he is describing the closing of the second dimension. And with its closing, the Great Refusal is in turn refused.¹⁶³ Refused there means absorbed, taken up and sold back, which is why the habit survives one person at a time. The politics is left where Marcuse dropped it. What survives is the habit of refusing one thing you can name, on your own account, after asking what it was for.

The whole argument, once, in order, because a person should be able to carry it out of the room. One, the theory said conditions make consciousness, and the conditions arrived while the consciousness did not, so the person had to be studied. The school assumed the person is made in the family before they hold an opinion. Two, freedom from the old bonds arrives as loneliness, and the loneliness has to be discharged somewhere. The modern discharge is to become what is expected and to call it choice. Three, the reason that beat superstition can only ask how and never what for, so it turns everything into amounts, including the person. It turns back into the power it escaped, and the name for a thing that becomes its opposite by succeeding is dialectic. Four, the place this reason gets inside you is not work but leisure, where the sorting is done upstream and the product is you. That leisure trains the attention the work needs is the mechanism this argument holds as a hypothesis. Five, the machine does not decide, and whoever owns the machine decides, working with what people can be made to want. No villain and still an owner, an interest doing the work a plan would. Six, a society can be unfree while satisfying its members, and the loss is the ability to ask whether instead of which. Nobody can judge your wants for you, and a real rest from a want is what raises the flag. Seven, you do not need a picture of the right life to know a wrong one, and every picture so far has been sold back. So a standard of that kind can only say what is wrong, and it cannot say what right action would be. That leaves a person who has to act with nothing to act on, and the price was chosen knowingly. Eight, there is a reason that talks, alive in the act of arguing, and it carries a standard nobody can argue without. It binds people who can be asked to answer, it reaches a company through a court, and none of it reaches your evening. Habermas never claimed otherwise, and the route he named, law, stops short of the evening, and that is the whole of the complaint. Nine, the feed is the counting reason at its largest, tilting what its users see without anyone intending it. The question of what is worth a person's attention has a slot in there, and the weight of that slot is a secret. Four weeks away from it lowered how much people valued it, the valuation stayed large, and that is the flag. Ten, what is left is the question, asked by you, of your own wants, to someone who can answer. Nothing in the machine can ask it for you.

One objection has been left standing through all of this, and it deserves its own best words rather than a summary. Nothing in this argument can say what a free want would look like. Every test in it can cast suspicion on a want. Only one of them could clear a want: a rest that ends with the person wanting the thing more. In the study, nobody came back wanting more, so nothing was cleared. A liberal can answer that freedom is the absence of force, a door you can leave by, and a voice that can be heard. That is a standard a person can apply on a Tuesday, to a law or a workplace, and this argument has none. The reply is not a rescue, because the argument never claimed a standard of freedom in the first place. What it claimed is narrower and worse, that the liberal's standard is the one an administered life passes by design. No force is used, the door is open, the voice is heard, and the wants were shaped before any of that was tested. A feeling of being run is evidence against nothing in that standard, which is why it could be dismissed as ingratitude every time. So the word for the condition is the one Part 3 used for the society, administered. Unfree has been used here in that sense and no larger one. There is a larger concession to make in the same breath. That wanting is manufactured at all is the school's hypothesis, and it has never been a finding. The repaired scale, the feed experiment and the four weeks off the platform are consistent with it and confirm none of it. The man with the manuscript is an image of the claim and never was its proof. Without the hypothesis the rest of it still holds. The counting reason is real and measured, and the schema is real and measured. And the question what is this for is still the one thing none of the machines can ask. There is a residue after that, and it cannot be talked away. The feeling all this began with may not be the splinter at all. It may be what living among billions of other people's choices feels like from inside one life, and nothing here rules that out. What the argument can do is say what to do with the feeling either way, which is to ask the question and watch.

There is one more turn, the dialectic applied to everything said so far, and leaving it out would be cheating. Anything that can be wanted can be sold, and the wish to be free of the machine is a want, so it will be sold. It already is. The retreat, the programme, the app that limits the other apps, the book telling you to put down the phone, bought on the phone. The question what is this for can itself be turned into a product with a lock of hair on it. The person who buys it will feel they chose. Benjamin put the general form of it about history itself. There has never been a document of culture, which is not simultaneously one of barbarism.¹⁶⁴ Every critique of the arrangement is printed and sold by the arrangement, so it carries the mark of what it condemns. That is what his sentence means here. The critique is made inside the arrangement, printed by its presses, sold in its shops, and this argument is not exempt. It is a product. You are engaging with it. Somewhere a number just went up.

So what is the defence, if the critique itself is for sale. There is only one and it is small, which is knowing that. The person who knows the wish to be free is being sold to them can still buy it, and often should. They can ask what it is for while paying. That is the difference between the man with the manuscript and a person who has heard this out. The second person may want exactly the same things as the first. The second person cannot see the join either, and knows there is one. The sign of it is a reason that cost nothing to find. Of the thinker who keeps doing this, Adorno wrote that he is in truth the one who does not give in. Not the one who wins, which is a different thing and was never on offer. One person can ask what tonight is for, tonight, and get an answer they did not have this morning. Nothing at the platform moves when they do, and that is the whole size of what is being offered.

Go back to the sentence at the start. You chose your job, your phone, the shows you watch, the opinions you hold. Nobody makes you do any of it. And still there is a feeling that your life is being run by something you cannot point at. You can point at it now. It is a way of thinking that only asks how, built into everything you own, including your wants. It runs most strongly in the hours you are sure are yours. The fourth thing claimed at the start was that "nobody forces me" is the exact test this arrangement passes. The evening and the feed show the shape of the arrangement passing that test. The tests reach the feed, and they barely reach the evening. The feeling was not ungrateful and it was not paranoid. Whether it was the mote or only the noise of living among billions of choices, do the same thing with it. Only you can start it, and you cannot finish it alone. What it cannot tell you is what to want instead. Nobody can, and everybody who says they can is selling something, and this is where the school stops and you go on. The question that was taken out of the room has been handed back. It is yours, and the only way to lose it again is to let something else ask it for you.