A Friend Asked, Volume II: Living as an End

Qin, Han

doi:10.5281/zenodo.23959195

About this volume

A Friend Asked: Essays on Self-as-an-End, Volume II: Living as an End is a collection of applied-philosophy essays that begins with ordinary questions. This volume contains the general preface and sixty-five essays, arranged in the order in which they were originally written and organized into five parts: “Inner Life,” “Knowledge and Consciousness,” “Education, Freedom, and Relationships,” “Power and Public Life,” and “New Forms of Instrumentalization and Human Resistance.”

Volume II brings the Self-as-an-End (SAE) framework into concrete questions concerning death, love, loneliness, education, freedom, consciousness, power, public life, and artificial intelligence. The essays minimize specialist vocabulary and approach philosophy through situations that remain open to general readers.

Author's Note

These essays did not begin with a finished theory in search of examples. They began with questions from friends: Why does a person become afraid? Why does a technology inspire both excitement and unease? Why does an institution turn people into instruments in the name of protecting them? The writing often came before the answer, and judgments changed in the face of new facts and counterexamples. In gathering these pieces into a book, I have not tried to erase the particular circumstances in which they arose. I hope, instead, that the questions will remain visible after the news has receded. Readers need not accept every conclusion here. If these essays can help a question be asked more precisely, they will already have done part of their work.

Translator's Note

The Chinese and English editions of this book are published separately, but their volumes, parts, essay numbers, and hidden stable source IDs correspond exactly. The translation is based on the edited Chinese manuscript, not on a line-by-line transcription of the original X posts. Platform residue, passing references, and conversational phrasing that depended entirely on the moment of posting were addressed during the Chinese editing process. The English text preserves the historical setting of each essay while adding only the minimum context needed by readers who may not share the same news memory.

Self-as-an-End (SAE) remains in English throughout the book. It is not a synonym for self-interest, self-actualization, or autonomy. The term extends a Kantian concern with persons as ends in themselves into a broader question: whether a person is recognized, within the institutions and social structures that shape their life, as the ultimate source of the ends they live by rather than as a means for goals imposed by an external system. Where the distinction matters, the translation separates Kant's formulation from the author's expanded use of Self-as-an-End without adding an apparatus of scholarly notes.

Core terms are kept consistent. Instrumentalization and self-instrumentalization retain distinct meanings, while terms such as alienation, agency, and subjectivity are distinguished according to context. Consistency, however, does not mean mirroring Chinese syntax sentence by sentence. The translation aims to preserve the force of the judgments, the humor, the hesitation, and the author's habit of correcting himself. Sentence length and paragraphing have sometimes been adjusted for natural English, but no new argument has been added.

Chinese institutions, classical texts, internet culture, films, television series, and games are explained only where comprehension requires it. Personal names follow the Romanized forms commonly used by the people concerned, and titles are kept consistent after their first appearance. Relative dates have been anchored to the historical moment of composition, and predictions remain predictions rather than being rewritten in light of later events. References already present in the essays remain part of the prose; no new system of citations or notes has been imposed.

Some wordplay, kinship-inflected forms of address, and political resonances cannot be carried over at equal weight. In such cases, the translation favors the function the expression serves in the argument, with a brief explanation in the text only when indispensable. Translation is itself an act of interpretation. The aim here is not to remove every ambiguity on the author's behalf, but to let English readers encounter the same questions that confront readers of the Chinese original.

General Preface

Writing as a Form of Prayer

I could hear the friendly teasing in the question, but I also thought it was a good one, and more meaningful than it first appeared.

Kafka, yes, the Kafka who wrote The Metamorphosis, once said that writing was a form of prayer. I agree. First, why prayer rather than some other metaphor? Sincerity, passion, and persistence can all describe serious writing, but they describe so many things that they can be made to fit almost anything. Prayer is different. It has a very specific structure: we speak to someone or something that is under no obligation to answer us.

Every part of that sentence matters. We are speaking, but we are not merely talking to ourselves. Our words have a direction. They are addressed to someone. Yet that person owes us no reply. However precisely, sincerely, or forcefully we speak, we cannot demand an answer.

This is where a writer stands. We can take responsibility only for our half, for putting the words into the world. Whether understanding occurs belongs to another person, and it may not occur at all. We cannot complete the process by ourselves.

Once that structure becomes clear, several other things come into focus.

The first is that writing asks us not to close a question too quickly.

Imagine someone writing, “I have so much, and yet I am still unhappy.” They could immediately add, “So I must be ungrateful.” The essay would be over, and the question would have been shut down before it had really begun. Or they could keep writing: What exactly do I mean by unhappy? Can I value what I have and still be unwilling to continue living under a particular arrangement? Why have the conditions other people consider sufficient failed to address what is actually troubling me?

Here, writing is not simply a way to express emotion. It keeps a ready-made answer from replacing the real question too soon. Writing resembles prayer because the writer does not have to play the role of someone who has already figured everything out. The writer can admit, “I do not yet understand, but I am willing to look carefully.”

Flannery O’Connor’s A Prayer Journal is a good example, and a genuinely religious one. She asks to become a better writer and to have her work published, but she also examines her own hunger for success. She prays for clarity and asks whether her desires might be obscuring what she is trying to understand. She does not reduce prayer to “Please give me what I want.”

From this we can draw a way of writing that does not depend on religious faith: I hope for an answer, but I am also willing to discover that what needs to change may be not only the answer, but the way I ask the question.

The second is that writing not only asks the world to listen to me. It also allows the world to correct me.

If an essay’s only task is to prove that its author was right all along, then the evidence, the people in the essay, and even the reader will all be arranged as tools in support of the conclusion. Another kind of writing allows what resists us to remain on the page. We may love a beautiful sentence and delete it when we discover that it is inaccurate. We may begin by casting a person as the embodiment of an idea, only to find that the particulars of their life refuse the label. We may think the argument is complete, then return to rewrite it when a reader points out a counterexample.

In such moments, writing does more than amplify the voice of the “I.” It also limits the “I’s” power to explain everything however it pleases. Writers sometimes describe revision this way in interviews: at a certain point, a sentence is no longer merely something they own; it begins to answer back. This is a description of the experience of writing, not a claim that inspiration comes from some supernatural force.

Of course, being willing to be corrected does not mean abandoning judgment. We still have to decide whether a correction deserves to be accepted or is merely another untested demand. Openness does not mean accepting everything indiscriminately.

The third is that writing longs for a response, but cannot turn that response into a command.

We can make a sentence more precise, describe an experience more concretely, and strengthen an argument. None of that guarantees that another person will understand, much less that they must agree. To write for readers we do not yet know therefore involves an unguaranteed trust: perhaps someone else is facing a similar confusion; perhaps they will pause over a particular sentence; perhaps they will disagree, but in doing so help me see what I could not see on my own.

Nor must a response always come from someone else. A private journal may give one’s future self a chance to look back and understand one’s present self more fairly.

Sometimes a lack of response is a real disappointment, but it does not cancel the value of having spoken. At the same time, writing seriously does not entitle me to anyone’s attention. I hope to reach you, but I do not claim ownership over your response. I believe communication is worth attempting, but I do not assume it must succeed.

This metaphor also reveals something else: prayer done badly and writing done badly fail in almost exactly the same ways.

One failure turns prayer into a demand that says nothing but “Give me this.” Its equivalent in writing is a sentence that is always asking for something: agreement, sympathy, recognition. Another form of prayer is performed for an audience, with posture more important than substance. Its equivalent is writing designed mainly to look profound. A third begins with the answer already settled and merely recites it. Its equivalent is an essay whose conclusion is fixed in the first paragraph and whose middle merely builds a road to it. All three are still forms of speech, but the structure in which “the other person owes me no answer” has disappeared. They do not actually need a response. They need only a witness.

Prayer is also often imagined as waiting for some power to take care of things on our behalf. In this metaphor, the opposite is true.

We cannot guarantee that a neglected experience will eventually be understood fairly, but we can begin by writing about it without distorting it further. We cannot guarantee that everyone will treat others with seriousness, but we can refuse to reduce the people in our writing to props that prove our point. We cannot guarantee that our theory is correct, but we can state its premises clearly, preserve the objections, and mark what remains unresolved.

A wish does not have to remain only a wish simply because its fulfillment is not guaranteed. It can enter into very concrete acts: how a sentence is written, how an example is handled, whether a revision is honest.

For the same reason, this kind of “prayer” need not be gentle or serene. It can contain anger. It can protest harm. It can also express gratitude for something good. Writing down an ordinary moment with one’s family need not serve any conclusion. It may simply recognize that the moment deserves to be seen, that our busyness should not make it disappear merely because it was not spectacular. In such moments, writing is not asking life for more. It is responding to the life that has already happened.

Finally, prayer may explain why we write with such seriousness, but it cannot serve as evidence that what we have written is true. An essay may begin in pain, hope, or care, but its argument must still withstand scrutiny. Sincerity does not make a conclusion true, and an explanation does not become valid merely because it comforts us. Conversely, the willingness to delete a claim that offers great comfort but cannot survive examination may be the clearest expression of a writer’s honesty.

Of course, it is also a mistake to sanctify writing into a mission one is never allowed to stop. Writing does not require us to suffer continuously, to produce words every day, or to turn every experience into a work.

What moves me most about the metaphor, then, is that it places limitation and hope in the same sentence: I cannot guarantee understanding, yet I remain willing to understand; I cannot guarantee that I will be heard, yet I remain willing to speak carefully; I cannot guarantee that I am right, yet I am willing to let what I write receive an answer.

It does not say, “May the world prove me right.” It says: may I be able to see what does not conform to my wishes and still write about it honestly; may these words reach others without prescribing their answers for them.

Volume II: Living as an End

Volume I asked how a person might avoid being reduced to an instrument by institutions and technology. Volume II brings the question back into the interior of life. Death, loneliness, love, forgiveness, education, freedom, and power are not problems a single concept can solve once and for all. These essays move closer to the places where choices are made: how we face a past that cannot be relived, how we live with others, and how we preserve a genuine capacity to change course within new technologies and public orders. Living as an end is not a permanent state one attains. It is the repeated refusal, in concrete action, to reduce oneself or another person to a use.

Part I: Inner Life

Hope without expectation does not mean losing all hope for life. It means refusing to bind oneself to a predetermined outcome. From there, this part enters questions of death, loneliness, love, regret, forgiveness, and the direction of a life. Inner life is no simpler than the public world. Here too, we can treat ourselves as projects in need of constant optimization and other people as roles meant to fill our absences. There is no quick method for acquiring peace in these pages, only a slower work of discernment: what is fear, what is attachment, and what kind of life are we willing to bear?

1. Hope Without Expectation

A friend asked what I meant by “hope without expectation.” It means that when we stop pinning our hopes on a particular outcome, we no longer collapse into despair. Without despair, we can put our feet on the ground. And precisely then, what we are doing begins to carry genuine hope.

2. If No One Knew, Would We Still Remain Moral?

A friend asked an old question. Suppose you could do something wrong and no one would ever discover or punish it. Could most people resist? Or does much of what we call morality depend on the fear of being caught?

The question is indeed old. So old that Plato presents an identical one in Book II of The Republic.

The story goes like this. After an earthquake splits the ground, a shepherd finds a golden ring in the fissure. He discovers that when he turns its setting inward, he becomes invisible, and when he turns it back, he reappears. He finds a way into the palace, seduces the queen, conspires with her to kill the king, and takes the throne himself.

The story is told by Glaucon, one of Socrates' interlocutors. When he finishes, he poses a challenge. Suppose there were two such rings, one given to a just person and the other to an unjust one. Do you think the just person would remain just? Glaucon says no. No one could hold out. Justice is therefore never something people freely desire; they observe it only because they have no choice.

Twenty-four centuries later, not a word of the question has changed. Plato wrote the whole of The Republic in answer to it.

My answer to my friend came in three sentences: wanting to, but not daring; wanting to, but choosing not to; never wanting to at all.

Another friend later remarked that the question itself reveals a difference in experience. Those who refrain only because they are afraid cannot make those who never want to do it feel what they feel, and vice versa.

I think that is half right. The difference is real, but it is asymmetrical, and the direction of the asymmetry is the opposite of what most people assume.

To explain why, I first need to place those three sentences within a structure.

In my Self-as-an-End (SAE) philosophy, the emergence of a person proceeds through a series of cuts, each of which marks a level. Three of those cuts matter here, from 13DD to 15DD. The full derivation appears in my general methodology; I will use only these three levels here (https://self-as-an-end.net/papers/methodology.html).

13DD, the self. It is not merely that something is operating; I am operating. Negation turns back upon itself, and the “I” appears. This level forms the boundary between nature and freedom. Before it, everything can be fully described by natural science. After it, there is an “I.”

14DD, the self as an end. One does not act without reasons. Action acquires a reason, and the question “why” appears. This is the apex of the individual level: treating oneself as an end in oneself.

15DD, recognition of the other. I am not the only end. Another person is recognized as an end in themselves.

The word “recognition” matters here. It is not the same as “seeing.” Seeing and acknowledging are two different acts. SAE judgment has only two variables: what the other person actually possesses, and what you are able to see. Failing to see what is invisible to you is not immoral. Seeing what you ought to see but refusing to acknowledge it is no longer a problem of vision. Recognition of the other means acknowledging what one has seen.

Something even more important must also be noted. The relation among these three levels is additive, not substitutive. Recognizing another person as an end does not mean ceasing to regard oneself as an end. It means that there is now more than one end. If someone surrenders themselves in order to make another person their end, that is not movement upward but regression. SAE has only one moral law: one must oppose the suppression of any person's remainder, and “any person” includes oneself.

One more thing must be made clear, or the next three sections will be read backward.

In SAE, each person gives themselves the law and obeys their own law. This holds at all three levels, not only at the highest.

The law arises from within. Its content takes the form of the level to which you have developed. Alignment with your own law is morality; deviation from it is immorality. What follows, then, is not a story in which some people possess morality and others do not. It is a story of three different laws, each sincerely held by the person to whom it belongs.

Now we can ask the original question again.

Wanting to, but not daring. The thought arises, and what stops it is the pair of eyes outside.

That does not mean this person has no law. They have one, but its content includes the structure that gives them meaning, and that structure is external. When the structure is absent, when those eyes are truly closed, the law has nothing left to say.

This is not weakness of will. It is simply the form the law takes at that level.

Wanting to, but choosing not to. The thought still arises, but what stops it is not whether someone else can see. The person has ends of their own. Doing the wrong thing would damage something they genuinely want or harm the person they are becoming. Whether anyone is watching therefore ceases to be a variable. This person has passed through the stage of not daring and now chooses not to act. The two may look similar, but the reason has changed. It no longer runs from the outside in, but from the inside out.

Yet we should describe this accurately. There is still a pair of eyes watching the self, and those eyes are keeping accounts. The calculation asks whether this act adds to or subtracts from me.

And the account can produce another result.

What if the wrongdoing serves my end? What if it advances precisely what I truly want? The law at this level does not forbid that possibility. It may even require it. The person then begins to fight with themselves, and both sides of the fight are their own. One side says, “This will destroy me.” The other says, “This is exactly what I want.” No third party can adjudicate, because plaintiff and defendant are the same person.

That is where inner conflict begins. It is not indecision. It is a structural feature of this level.

Most people stop here. That is why the original question is a genuine and painful question for most people. At this level, the answer really can go either way.

Never wanting to at all. What happens here is not a stronger act of resistance or a more careful calculation. The thought does not arise. This does not mean it has never arisen. It means there is no longer a need to entertain it, because during the earlier stages of not daring and choosing not to, the person has already thought it through.

In their eyes, the other person is an end, not a variable in their ledger. The account is never opened. The lawsuit has no opposing sides. The option does not appear on the list, and there is nothing to restrain.

This does not mean the person no longer has ends of their own. Those ends remain fully present. What has changed is that there is now something else in the room that does not enter their ledger.

There is another point especially vulnerable to misreading. This person does not act this way because an external norm demands it. No one is ordering them, and no book is policing them. Their own law has taken this form. Having arrived there, they cannot do otherwise.

This has nothing to do with nobility. Water boils at one hundred degrees Celsius. Boiling is not an extra virtue of water. At that temperature, that is simply the state water is in.

The original question therefore measures not the strength of one's morality but what the questioner thinks morality is. It assumes that morality is a process of resisting temptation. That assumption holds only at the first two levels.

And, to be honest, the skeptical view identifies the mechanism correctly. It merely mistakes its nature.

At the first level, what restrains you is the pair of eyes outside. At the second, it is your own ledger. Both levels are indeed calculating costs. The only difference is where the cost lands. The skeptic is right about that.

But the account itself is the law of each of those levels, given and obeyed by the person themselves. To say that it does not count as morality is to measure them by the law of another level. In SAE, measuring others by one's own law has a name: ethics, not morality.

Only at the third level does calculation disappear, because there is no account to settle.

At that level, the question is not more difficult. It no longer makes sense.

Look again at the ring. It is actually a very precise instrument.

It does only one thing: it subtracts “being seen” from the equation. Whatever remains after that subtraction is the true content of your law.

When a person at the first level puts on the ring, what has been subtracted is the entire content of their law, so nothing remains. This is the level Glaucon describes.

When a person at the second level puts it on, one item is removed from the ledger. The account remains, but it becomes harder to balance. The ring does not liberate this person. It turns up the volume of the inner conflict.

When a person at the third level puts it on, nothing happens. Visibility never entered the equation. Subtracting a term that was never present leaves the equality unchanged.

Glaucon's challenge sounds irrefutable because he assumes everyone occupies the same level. He says no one can hold out. At the level he assumes, that is true.

By this point, you may already have had a reaction of your own. The easiest thing to recognize is the inner conflict: the feeling of being pulled back and forth between “This will destroy me” and “This is exactly what I want.” If you recognize it, you know where you are. If you read the words “the account is never opened” and think they are merely a performance of virtue, that too is a coordinate. No one needs to tell you.

One aside: these three names are coordinates, not labels. They do not describe three kinds of people. The same person may recognize another as an end in one matter and remain at the stage of wanting to but not daring in another. In SAE, good and evil describe direction, not position. There are no finally condemned evil people, only diagnosable operations. So do not rush to paste a label on yourself or anyone else.

My friend then asked a second question. Can people move gradually upward from the lower level, or is the threshold fixed at birth?

Not only can they move; they must begin at the first level. There is no way to skip it.

The reason is structural, not inspirational. Each level is produced by making a cut in the preceding one. Without an “I,” there is no self that can become an end. Without treating oneself as an end, there can be no “also” in the sentence “You are also an end.” The third level is cut from the material of the second. Without the material, the cut cannot be made.

There is a second reason, and it is even more decisive.

Because the law arises from within, it can only be reached; it cannot be given. A law handed to you by someone else is not your law. Even if its content is identical word for word, what you have received is a sentence, not the law itself. A shortcut is not merely difficult. No such route exists, because there is no channel through which this arrival can be transmitted.

What skipping produces, then, is not recognition of the other. It produces someone who has memorized the lines of recognition. They say the right things, and the reason they say them is fear that they will be caught saying the wrong things.

The appearance belongs to the third level; the interior remains at the first. This is why moral indoctrination always fails. You can insert the conclusion, but not the process of arrival. SAE calls it colonization when someone claims to arrive for another person and then inserts the conclusion. Waiting for a person to unfold until the law appears of its own accord requires no instruction; this is cultivation.

Yet the fact that everyone must begin at the first level does not mean everyone will reach the third. This is a process with no guarantee of completion. Some people remain at one level for their entire lives, and remaining there is not failure. Each level generates its own law from within, and that law is genuine morality at that level. Alignment with the law of one's own level is morality; departure from it is immorality. Which level one occupies is a separate question.

We can now return to the claim that the two sides cannot understand each other's experience.

The symmetry is false.

From above, there is understanding, because one has traveled through what lies below. From the third level, you can see the moral phenomena of every preceding level. You can describe the law of the first and of the second because your field of vision contains them. The person who no longer entertains the wrongdoing knows perfectly well what it feels like to want to but not dare. That is where they began.

From below, the problem is not failure to understand but invisibility. It is not that one has failed to look closely enough. From that position, there is nothing there to see. In the same way, a person at the third level cannot see the law of a fourth level because they have not reached it.

This is why the two groups cannot converse, and why their failures of conversation differ. The person who has made the journey knows what the other is saying. The person who has not arrived does not merely disagree. The thing being described does not exist within their field of vision.

This invisibility is structural, not a defect. Not having arrived is not evil. The second level is not bad. It is simply not the third.

But the next sentence must be said, or this essay becomes a license for arrogance.

Structural invisibility is real. Using “you cannot see it” as a shield is something else.

Once the claim that those below cannot see what is above has been established, it becomes an extraordinarily convenient weapon. Anyone facing criticism can say, “You fail to understand because your level is too low.” The statement cannot be falsified, and it ends the questioning at once. Using a judgment of levels to exempt oneself from scrutiny has a name in SAE: a structure pretending to be a law. That is colonization itself.

Judgment still has only two variables: what the other person actually possesses, and what you can see. If you see what ought to be seen and refuse to acknowledge it, you are lying. That is not a question of levels.

Nor would someone genuinely at the third level use hierarchy to silence another person's questions. Their law requires them to oppose the suppression of any person's remainder, including the person questioning them. To use level as a means of suppressing a critic is precisely an act of falling away from the third level.

One final point.

This essay does not ask anyone to go anywhere.

SAE has only one moral law: one must oppose the suppression of any person's remainder. That law is directed at oneself, not at others. It is not bestowed by one person upon another. Each person arrives at it for themselves. Its direction is universal, because it concerns any person, but its binding force falls only upon the person who holds it. You cannot use it to demand that another person act in the same way. The instant you do, your morality has been projected outward as ethics, and ethics is precisely what this framework seeks to dismantle.

All I can do, then, is lay it out here. If your law resonates with it, that is your own remainder unfolding, not the result of my persuading you. If it does not resonate, then let each continue to resonate in its own way. No one has lost anything.

3. Why Do We Understand the Principle Yet Still Fail to Act?

A friend asked why we can understand all the principles yet still fail to act on them. Why is the unity of knowledge and action so difficult?

This too is an old question. Someone asked exactly the same thing five hundred years ago, and the questioner was no outsider. He was Xu Ai, one of Wang Yangming's closest students.

At the very beginning of the Instructions for Practical Living, Xu Ai is arguing with others about the unity of knowledge and action. Unable to settle the debate, he goes to his teacher. His example is this: everyone today knows that one should be filial to one's father and respectful to one's elder brother, yet people fail to be filial or respectful. Knowledge and action must therefore plainly be two different things.

Yangming's answer is brief: such people have already been cut off by selfish desires; this is not knowledge and action in their original form. There has never been someone who knows but does not act. To know and not act is simply not yet to know.

At first glance, this looks like wordplay. You say that you understand but cannot do it; he says that means you do not understand. There seems to be no reply, because he has simply removed your premise.

But the examples he gives next are not games.

He cites the Great Learning: “loving a beautiful sight as one naturally loves beauty, and hating a foul odor as one naturally hates a foul odor.” Seeing beauty belongs to knowledge; loving it belongs to action. At the very moment one sees it, one already loves it. One does not first see it and then separately form an intention to love it. Smelling a foul odor belongs to knowledge; hating it belongs to action. At the moment one smells it, one already hates it.

Then comes the person with a blocked nose. Though a foul odor may be before him, he has not smelled it, so he does not particularly hate it. He simply does not yet know the odor.

Yangming gives an even more concrete formulation: to know pain, one must already have felt pain; to know cold, one must already have been cold; to know hunger, one must already have been hungry.

Notice the crucial phrase: one does not first see it and then separately form an intention to love it.

Forming an intention is the step we ordinarily assume: first we know, then we decide to act. Yangming says genuine knowledge contains no such step.

The question then becomes: What exactly is that step? Why is it sometimes indispensable and at other times entirely unnecessary?

A principle is universal, while the situation before you is particular. A principle does not arrive with instructions for use. No principle comes with a manual explaining what it requires here and now. There is a distance between a universal sentence and the particular matter before us.

What travels that distance is judgment.

Kant wrote specifically about this. He argued that in cognition as in morality, judgment borrows its a priori principles rather than possessing principles of its own. It therefore cannot be taught. It can only be exercised and practiced, and much of it may even be a natural gift.

That sentence is worth pausing over. Principles can be taught; judgment cannot. Yet judgment is exactly what stands between a principle and an action.

So when you understand but cannot act, the problem is not that you understood and then became lazy. The sentence in your hand has not yet become a usable judgment.

Placed within a structure, this becomes much clearer.

In the previous essay, I described three levels in the SAE philosophical framework (https://self-as-an-end.net/papers/methodology.html).

13DD, the self. It is not merely that something is operating; I am operating.

14DD, the self as an end. One does not act without reasons. Action acquires reasons, and one treats oneself as an end in oneself.

15DD, recognition of the other. I am not the only end. Another person is recognized as an end in themselves.

The third formulation is, in fact, Kant's famous principle: persons are ends in themselves and must never be treated merely as means. SAE follows Kant throughout, and especially at this level.

Judgment takes three entirely different forms across these levels.

At the first level, a person hears the principle but does not understand it, so no judgment can be made.

They can repeat the sentence and may sincerely accept it. But in a situation where no one points out what to do, they do not know what the sentence requires. If someone tells them explicitly, they can comply. Where nothing is specified, there is a blank.

The principle is an instruction handed in from outside, not a reason that has grown from within. They can observe it in the situations covered by the instruction. Beyond those situations, it has nothing to say.

This is not weakness of will. The judgment has not yet been made, so there is nothing to carry through.

At the second level, the person understands and can make the judgment, but more than one judgment appears, and the judgments fight each other.

They do not merely understand the wisdom of going to bed early; they can explain it to others. But those two hours awake at night are the only hours in the day that belong to no one else.

Both sides are things they genuinely want. Both sides are their own. No third party can adjudicate, because plaintiff and defendant are the same person.

This level contains a great deal of inner conflict for precisely that reason. It is not a matter of strength or weakness of will. This is simply what the level looks like.

And there is something still more troublesome hidden within it: understanding itself can be entered in the ledger.

At this level, “I am a person who understands this principle” is itself an asset, and one that can be redeemed without paying the cost of action. You read a sentence that feels profoundly right, and the comfort in that instant is genuine. The reward has already been paid. Your saved items multiply, your bookshelves fill, and you do not move.

Yangming wrote about this five hundred years ago: people divide knowledge and action into two things. They think they must first know and only afterward act. “For now,” they say, “I will study and discuss, doing the work of knowing. Once my knowledge is genuine, I will do the work of acting.” As a result, they never act throughout their lives, and throughout their lives they never know.

Most people remain at this level for a long time. That is why the question is real and genuinely painful for most of us.

At the third level, there is no need to form that separate intention.

Once another person is recognized as an end, judgments of this kind become immediate and do not pass through a ledger. The other person is not a variable in my account, so there is no calculation to make and no two sides to fight.

To be precise, this does not mean judgment has become stronger. It means that one category of things that once required calculation no longer needs to be calculated. Other difficult judgments remain difficult. When the needs of two people collide, or when my remainder collides with another person's, the work of cutting and discerning may still take a very long time.

The distinction among the three levels, then, is not strength of will. It is three states of judgment: unable to make one, making several that conflict, and no longer needing to make one.

Now return to Yangming's diagnosis, and something interesting appears.

He says selfish desires have cut us off. Those words precisely describe the second level. The other side of the ledger contains another end of mine, and it blocks this one.

But his own example of the person with a blocked nose describes something else. That person is not blocked by selfish desire. He simply cannot smell. He has not arrived; he has not been obstructed. That is the first level.

Being blocked and not having arrived are two different conditions. In the former, the thing is present but pressed down by something else. In the latter, it has not yet grown. Both are real, but they are not the same.

Yangming saw both. He simply named the whole matter after the second.

One question remains: Why does “understanding” feel so much like understanding?

Because from within, the two experiences feel identical. Receiving a sentence and arriving at a law produce the same inner flash of brightness. No internal test can distinguish them.

The only test lies outside.

Yangming states it plainly. If we say that someone knows filiality or knows fraternal respect, that person must already have practiced filiality or fraternal respect. Surely merely knowing how to talk about them is not enough to say that the person knows them.

The unity of knowledge and action is therefore not a goal to be achieved. It is a measuring instrument. When knowledge and action separate, what the instrument reveals is not weak will. It reveals that the knowledge has not yet arrived.

One misunderstanding should be prevented here. This does not mean reading is useless.

A principle is a structure, something left behind after another person has done the work of cutting. A structure is not waste. It is material. Yangming never said knowledge was unnecessary. He said knowledge is the beginning of action.

The mistake is not reading. The mistake is taking possession of the material for arrival itself. Yangming's full sentence has two halves: knowledge is the beginning of action; action is the completion of knowledge. Take only the first half and you get the condition he described, a lifetime without action.

One final thing.

After reading this essay, you may feel that you understand it. That “understanding” is exactly the kind this essay has been discussing.

This essay cannot give you anything. It can do only one thing: remove a shortcut you thought existed. The remaining route is clumsy. You must walk it yourself, making judgments in particular situations, getting them wrong, and trying again. That is what Kant meant when he said judgment can be practiced but not taught.

So I can only lay the argument here. As in the previous essay, if your law resonates with it, that is your own remainder unfolding, not the result of my persuading you. If it does not resonate, let each continue to resonate in its own way. No one has lost anything.

4. Why Do We Fear Death?

The question usually carries an assumption: fear of death is one thing, the same thing feared by everyone.

But death itself has no content. No one has experienced it and returned to tell us what it is like. Strictly speaking, then, people do not fear death. They fear what death interrupts.

And what it interrupts depends entirely on what we are holding.

Twenty-three centuries ago, Epicurus offered a famous solution. Death, he said, is nothing to us. When we exist, death is not present; when death arrives, we no longer exist. If we and death can never meet, what reason could we have to fear it?

The argument is coherent, clean, and almost impossible to fault. Yet twenty-three centuries have passed, and no one has ceased to fear simply because they read it.

Why does a logically impeccable argument have so little effect in life? That is what this essay is about.

Let us return to the three levels used in the previous two essays (https://self-as-an-end.net/papers/methodology.html).

13DD, the self. It is not merely that something is operating; I am operating.

14DD, the self as an end. One does not act without reasons. One treats oneself as an end in oneself.

15DD, recognition of the other. I am not the only end. Another person is recognized as an end in themselves.

Place the fear of death within these three levels, and three entirely different fears appear.

At the first level, one fears the unknown.

There is already an “I,” but no end that must be pursued has yet taken shape. What the person holds remains formless. Death is therefore not “the interruption of something of mine.” It is simply darkness.

What they fear is the darkness itself. They do not know what lies in it, or whether an “I” will remain there.

This is the most primordial and most universal fear. It needs no reason. Anyone who becomes conscious of an “I” passes through it.

It is also the strongest of the three. The other two fears have particular objects, and where there is a particular object, something can still be said. This fear has no object. There is nothing to discuss.

At the second level, one fears that the account will never be settled.

The person at this level has ends, and they are their own rather than ends inserted by someone else. They know where they are going.

Death therefore becomes something concrete: I have not arrived.

The book is unfinished. The child has not grown up. I have not gone to that place or said those words. What one fears is not darkness, but interruption, the unchecked items remaining on the list.

Most people live at this level, so most discussions of death are really discussions of this fear.

This level also contains a pain absent from the first: the clearer a person's ends, the more specific the fear. Someone who knows what they are trying to do knows more clearly what they stand to lose than someone who has not yet decided.

The first fear has not disappeared. It has merely been covered by something closer and easier to name. At night, when the nearer concern falls silent, the earlier fear can emerge again.

At the third level, one fears that a road will not be walked to its end.

Once another person is recognized as an end, something becomes unavoidable: cultivation.

Cultivation is not instruction, nor is it giving someone a conclusion. It is accompanying a particular person for part of their own road, allowing them to grow and arrive for themselves.

Cultivation has a quality nothing else possesses: it cannot be handed over.

Tasks can be transferred. Money can be left behind. What one has understood can be written down and passed on. But cultivation unfolds over a long period, has a particular person as its object, and is constituted by the two of you together. If you are gone and someone else takes your place, it is not the same cultivation. It is another passage, beginning again from the start.

The third fear takes this form: I am partway along several roads with others, and no one can finish walking my part for me.

No argument can dissolve this fear. It is no answer to say someone else will take over, because the person who takes over is someone else. Nor is it enough to say that I did my best. What one cannot but do belongs to a different dimension from how much effort one has made.

Two things need further clarification.

First, the third level does not cease to fear what the second fears. One's own ends remain, as does the fear of leaving them unfinished. The levels are additive, not substitutive. One fear is not exchanged for another; a new fear is added. And the new one is more difficult because it is not entirely in one's own hands.

Second, and more easily misunderstood, being unable not to turn toward another person and cultivate them has nothing to do with nobility.

What one cannot but do differs from what one is willing, glad, or selfless enough to do. A willing person can choose not to be willing. That is virtue, and it deserves praise. What one cannot but do offers no alternative. It resembles the fact that water boils at one hundred degrees Celsius. Boiling is not a virtue of water.

These three levels are therefore not three spiritual attainments. A person at the third level does not fear death less. If anything, they have more to fear.

Now return to Epicurus. His sentence is actually a precise instrument.

It turns the fact that “I will not be present” from bad news into a reason.

At the first level, it works. The fear concerns the darkness, and Epicurus says that you will not be there, will have no experience, and therefore no darkness will be waiting for you. The unknown is canceled. Epicurus has indeed comforted many people over more than two thousand years, and this is the level he has comforted.

At the second level, the sentence does nothing. An unsettled account has nothing to do with whether I am present. An unfinished book remains unfinished whether I am there or not. The argument touches no part of the fear.

At the third level, the sentence works in reverse. Far from offering comfort, it names the content of the fear exactly. It is because I will not be present when death arrives that I cannot finish walking that road. The premise with which Epicurus cancels fear is the very cause of fear at the third level.

The same sentence produces three results: it helps, it does nothing, and it strikes the fear at its center.

That is why an argument with no obvious logical flaw may have almost no effect in life. It is not wrong. It speaks to one level, while its listeners do not all occupy that level.

Return to the original question.

“Why do people fear death?” may not be precise enough. A more useful question is: Which fear is ours?

Is it the darkness, the unfinished list, or several roads left halfway walked?

They differ so profoundly that words capable of comforting one may do nothing for another, or may even intensify it.

Nor are they three rungs to be climbed. As I have said, the person at the third level may have more fears, not fewer.

This essay cannot make us stop fearing death. Perhaps no essay can.

It asks us to face the fear. Fear of death is not itself frightening. It is evidence that a person has possessed something. We fear losing only what we hold. The first level holds a newly emerged “I.” The second holds one's own ends. The third also holds the ends of others. The more one holds, the more one fears losing. There is nothing shameful in that.

What is frightening is not knowing what we fear. When we do not know, fear remains undifferentiated. An undifferentiated fear cannot be faced or worked through. It can only be pressed down.

Once we know which fear is ours, we can at least ask the next question: What should I do now?

The answer differs for every person, and each can answer only for themselves.

5. How Can We Know That We Are Right?

A friend asked, “How can I know that what I know is right?”

The blunt answer is: you cannot.

Of course, that answer is a little lazy and evasive. It is too easy to say, and the person who says it gets to appear more lucid than the person who does not. Humility may be the easiest thing in the world to perform.

So once that sentence has been spoken, the work has only begun. There are really two questions: Why can we not know, and what do we do after admitting that we cannot?

First, consider a sentence that has long been misquoted.

The line most widely attributed to Socrates is “The only thing I know is that I know nothing.” He did not actually say it. The closest passage in Plato's Apology says that the small advantage Socrates possessed over others was simply that when he did not know something, he did not suppose that he knew it.

The two statements are far apart.

“I know that I know nothing” is a beautiful pose. It ends as soon as it is spoken, and it is a paradox that sounds clever. “Do not suppose that you know what you do not know” is unattractive and offers nothing to boast about. It is a discipline, an act that must be performed again whenever one encounters a particular matter.

The admission that one may be wrong most easily turns into the first sentence.

Why, then, can we not test our own correctness from within?

Because certainty and correctness feel the same.

When we examine an idea inwardly, we use the same apparatus that produced it. The instrument and the object measured are identical. An idea makes sense to us precisely because it fits our other ideas. That pleasurable feeling of everything locking into place is what we call having thought something through.

But falsehood can also lock into place. A completely coherent view may contain no internal contradiction from beginning to end while being crooked as a whole. Coherence guarantees only that the parts do not fight. It does not guarantee that the whole corresponds to anything outside itself.

There is another, more troublesome layer. The more earnestly we reflect on ourselves, the more likely we may be performing earnest reflection. The thought “I am someone capable of self-examination” is itself comforting. Once that comfort has been paid, the examination can stop.

At this point, we hit a wall.

If we cannot know whether what we know is right, how do we know that the statement “we cannot know whether we are right” is itself right?

If it is right, then we know at least one thing, and it does not fully hold. If it is wrong, it collapses at the outset.

There is no way through this wall. Every form of skepticism has struck it.

All I can do is acknowledge that the sentence lies within its own range of fire. I take it as a starting point not because it can survive the objection, but because I have not found a more honest place to begin.

And a claim that includes itself is at least a little closer to honesty than one that exempts itself. Any statement announcing “this law applies to everyone but me” usually has a larger problem.

We have hit the wall, but the matter does not end there.

The wall blocks only one direction: looking from within toward the interior.

No exit inside does not mean no exit at all. The exit lies outside.

We cannot become aware of what lies beyond our knowledge unless someone questions us.

That sentence is the center of this essay. The next several sections explain why it holds.

Another person may see what we cannot. This is not because they are smarter or more correct. They occupy a different position.

Here is the crucial point: our blind spots are not random.

Any view that explains the world necessarily neglects some part of it. What it neglects does not raise its hand. It is simply, quietly absent. In the previous essays, I called this the remainder.

The position of the remainder is determined by the shape of the view itself. No amount of effort from within can illuminate it, because the light with which we search is produced by the very same apparatus.

Another person is different. They have their own view and their own remainder, but their remainder does not occupy the same place as ours. They can therefore see what we omit, just as we may see what they omit.

Being questioned is therefore not a matter of magnanimity, a virtue required only of well-cultivated people. It is the only means of testing. There is no other.

Twenty-four centuries ago, Plato wrote about this in the Gorgias, more directly than almost any account in circulation today.

Speaking to Gorgias, Socrates says that he is the sort of person who is glad to be refuted when he is wrong and glad to refute another person when that person is wrong. Between the two, however, he would rather be the one refuted. To be refuted is to rid oneself of an error; to refute someone else merely rids that person of an error. Removing something bad from oneself brings the greater benefit.

The position of this passage also matters. Socrates says it while deciding whether the conversation should continue. If you are this kind of person too, he tells Gorgias, let us go on. If not, let us stop here.

The ability to accept refutation is treated as a precondition of dialogue itself.

There is another scene in the same dialogue. Polus tries to refute Socrates by calling witnesses, insisting that any number of respectable Athenians would support him. Socrates replies that this is the courtroom style of refutation, which relies on numbers and is useless for truth. He wants only one witness: the person standing before him and actually engaged in the argument. Everyone else counts for nothing.

One person presents a crowd as evidence. The other says a crowd is no evidence at all; only the person who has truly entered the argument matters.

But fairness requires a qualification, and it is more useful to us than the principle by itself.

The Gorgias announces that Socrates would rather be refuted, yet almost the entire dialogue consists of Socrates refuting other people. By the end, Callicles simply stops engaging and offers perfunctory replies.

A dialogue that declares its willingness to be refuted never once allows its author to be refuted.

This is not an attempt to catch Plato out. It is evidence of how difficult the principle is to practice. Stating it is easy. Many people have stated it over twenty-four centuries. Callicles' final refusal to engage is actually the most respectable way of blocking it.

Follow this line a little further, and an uncomfortable conclusion appears.

Another person's agreement does not really help me.

From where I stand, I cannot distinguish why they agree. Perhaps they genuinely checked the reasoning. Perhaps they share my remainder. Perhaps they did not check at all and simply found the statement pleasing. All three look identical to me. In none of the three cases am I required to walk the path again. Nothing moves.

Agreement is comforting, but it carries no information.

Disagreement, by contrast, may bring something.

A disagreement that touches the content forces me down one of two roads. Either I discover that I was wrong, which is a gain, or I walk the path of reasoning again and find that it still holds. Only after that second journey does the sentence “I may be right” acquire its first measure of weight.

Notice that it remains only a possibility. Surviving one question does not make a claim right. It means only that this question did not overturn it.

Three misunderstandings should be prevented.

First, questioning is not voting. A large number of people agreeing does not make something right. When a crowd shares the same view, it shares the same remainder. No number of people can illuminate the part their view excludes. What matters is not the number of people, but the difference in position.

Second, not every disagreement carries information. A disagreement that never touches the content leaves as little altered as agreement does, though it is more unpleasant. This judgment is slippery, however, because it can be used to dismiss anyone. It is suitable only for examining oneself, not for throwing back at another person.

Third, this does not mean we must change whenever someone questions us. The value of questioning is not that the other person must be right. It is that we are forced to walk the road again. We may change by the end, or we may not. But not changing after walking it is no longer the same as not changing before being questioned.

Near the end of the Tractatus Logico-Philosophicus, Wittgenstein makes a strange remark. Whoever understands him, he says, will finally recognize the propositions of the book as nonsensical. The reader must use them as steps in a ladder, climb beyond them, and then throw the ladder away.

Near the end of a book, the author tells you that it is a ladder and should be discarded after use.

What happened later is even more telling. Wittgenstein eventually overturned much of that book and wrote something entirely different.

Publicly rejecting one's earlier work is rare not because ideas never change, but because handing over one's words and allowing them to be cut into, even by one's later self, is costly.

How, then, can we know whether we are genuinely accepting questions or merely appearing to accept them?

Again, the inner state cannot be tested. Only the action can.

When questioned, is your first impulse to cut into the issue or to block it?

Blocking takes many forms, and they usually sound reasonable: You do not understand. You lack the qualifications. Your motives are impure. You have taken me out of context. You did not even finish reading. Every one of these claims may be true.

But they share one feature. They stop the question without allowing the content to move.

Genuine openness to questioning looks much simpler. It means returning to the specific issue and looking again. Even if the conclusion remains “I still think this,” the intervening journey has taken place.

Return to the original question.

Perhaps what we can ask for is not to be right, because rightness cannot be verified from within.

What we can ask for is to remain open to questioning.

The difference is larger than it appears. A truth that cannot be questioned and a falsehood are structurally the same kind of thing: neither will change again. The former does not change because no one can touch it. The latter does not change because no one does. A claim that can be questioned, even if it is wrong at this moment, at least remains in a position from which it can move toward what is right.

If we do not know whether we are right, why continue speaking and writing?

Precisely because we do not know.

An idea kept inside can never be tested. The longer it remains there, the smoother and truer it feels. Only when it is spoken, written down, and handed over does it first become available for another person to touch.

We do not write because we have finished thinking. We write so that the thought can be handed over and checked.

This essay does not claim to know that it is right.

The only thing it can do is state itself clearly enough for you to find where it is wrong.

6. Why Does “I Only Want What's Best for You” Feel So Uncomfortable?

A friend asked why hearing “I only want what's best for you” feels so uncomfortable. The other person clearly means well, yet our instinct is to retreat. Afterward, we may even feel guilty and wonder whether we are simply ungrateful.

First, look at what the sentence itself does.

It appears to be a description of the speaker's motive. But when it reaches the listener, it does something else.

It changes the nature of refusal.

Before the sentence is spoken, disagreement concerns the content: Is this advice right? Is it appropriate for me? Do I have other considerations?

Afterward, disagreement is no longer only about the content. It becomes a judgment about the person: you have betrayed someone's goodwill.

The content no longer needs to be examined. The question of whether the advice is right has been replaced by the question of whether the motive is good. The motive is good, so the person who refuses becomes the ungrateful one.

The discomfort is therefore accurate. What we feel is not offense but obstruction. We have been blocked.

Placed within a structure, the people who use this sentence are not all the same.

At the first level, the self:

People at this level care about others, and the care is genuine. But this is the language they have. Care sounded like this when they were growing up. Other people said the same words when they were being kind to them. So they learn to say it too, without calculation or any wish to control.

They may not even understand what could be wrong with the sentence. To them, it simply means “I care about you.”

At the second level, the self as an end:

Here things become more complicated, and we must distinguish two cases.

In the first, the person is sincere. Helping you in this matter has become one of their own ends, written in their own ledger.

That is exactly the difficulty: when we refuse, we stand in their way. We are not blocking someone else's road. We are blocking an end of their own. So they push harder, with complete confidence in their right to do so, because they genuinely are thinking of us. No one knows that better than they do.

In the second case, the sentence is borrowed. What the person truly wants is something else: our obedience, our compliance with their arrangement, freedom from having to face a choice we have made for ourselves, or simply the outcome they prefer. “I want what's best for you” is the easiest wrapping.

The second case is more common than most people imagine.

But the more difficult point comes next: from the outside, the two cannot be distinguished.

The words are the same. So are the expression and the insistence. We have no way to determine from outside which kind of person stands before us.

Nor can the speaker reliably distinguish the motives from within. The previous essay argued that certainty and correctness feel the same. In the same way, genuinely wanting what is best for you feels exactly like needing you to do what I say while sincerely believing that I want what is best for you.

The sentence works precisely because the two cannot be distinguished. It has the same effect whether the person is sincere or not: it turns refusal into ingratitude. That makes it an excellent tool, so effective that the person using it may not know they are using a tool at all.

At the third level, recognition of the other:

This level does not mean refusing to care.

Recognizing another person as an end means helping and guiding them. Indifference would treat them as something unrelated to oneself.

But at this level, the sentence is not spoken.

Not because the person is especially refined, but because speaking it would be self-contradictory. They need the other person to be able to refuse, to question, and to say, “I do not accept your advice.” The phrase “I only want what's best for you” is exactly what takes that capacity away.

Once the sentence is spoken, the very thing the speaker wanted disappears.

What, then, does it look like to help and guide without saying those words?

Socrates offers an image in Plato's Theaetetus.

He says that he is the son of the midwife Phaenarete and practices the same craft. He attends men rather than women, however, and cares for the labor of the soul rather than the body. A midwife eases labor pains and helps someone through a difficult birth. Socrates says the greatest part of the art is judging whether what has been born is true knowledge or only an illusion.

The crucial point is that he gives birth to nothing himself. He repeatedly says he has no wisdom of his own. The knowledge is not inserted by him; it was already being carried by the other person. He rarely even answers questions. He asks them.

A midwife never says, “I only want what's best for you.” She remains beside the person at the moment of greatest pain.

And this must be clear: delivering a child is not standing idly by. A person who does not care is not in the delivery room.

C. S. Lewis made a precise observation about the sincere forcefulness of the second level. Of all tyrannies, he suggested, the one sincerely exercised for the good of its victims may be the hardest to endure. Other tyrants eventually satisfy their greed. Those who torment you for your own good can continue forever because their consciences approve.

The mechanism is contained in that sentence: goodwill supplies no stopping condition.

When we do something for ourselves and go too far, embarrassment may eventually stop us. When we do it for another person's good, there is no brake. Every additional act can be interpreted as loving them a little more.

Kant offers a sentence that can be placed beside Lewis's. Paternal government, he says, is the greatest despotism imaginable because it treats people as children incapable of judging good and evil for themselves.

So what can we do, if motives cannot be distinguished, including our own?

The answer is the same as in the previous essay: do not try to distinguish motives. Distinguish actions.

One thing can be observed directly: Can the other person still refuse?

The question is not whether they dare to refuse. It is what happens afterward.

If they say no and the matter ends, while the care and the relationship remain, that is guidance. Whether they give birth is their affair; we remain beside them.

If they say no and what follows is escalating pressure, disappointment, “after everything I have done for you,” or “don't say I didn't warn you,” it is no longer guidance. We need them to give birth to the child we have chosen.

The virtue of this test is that it does not ask about motive. It does not require us to decide whether we are sincere, because we cannot make that judgment reliably. It asks only one thing: What happens when the other person says no?

One final point.

The sentence is most dangerous precisely because, when it is spoken, the speaker is usually sincere.

The next time it reaches our lips, perhaps we can first ask ourselves: What will I do if the other person says no?

If the answer is “I will push harder,” it may be better not to say the sentence at all.

7. Why Are We Still Lonely When Surrounded by People?

A friend asked why we can be surrounded by people and still feel lonely.

Half the answer is already hidden in the question. Loneliness is not the absence of people. It is the absence of that person.

The other half of the question is: Who is that person?

The answer differs from one person to another, and even for the same person at different times.

At the first level, the self:

An “I” has emerged, but it is still new and needs confirmation. The support this person needs falls within the idea of the self: you must help me feel that I am myself.

Loneliness at this level means that the people around me have no relation to who I am.

They are not bad, nor are they absent. But their presence does not make this “I” any more substantial. People talk, eat, and hold meetings around me. Everything appears normal. I sit among them feeling like something unrelated to it all.

This is the most common form of loneliness and the hardest to express, because from the outside nothing appears to be missing.

At the second level, the self as an end:

This person has ends of their own and knows that those ends are theirs. What they need therefore changes. They no longer need confirmation that they are themselves. They need someone to recognize what they are doing.

Loneliness at this level means that no one can recognize my end.

More precisely, no one truly understands what it is. The people around you may offer support and encouragement and tell you to keep going. Yet you know they are supporting something vague, not the particular thing you are actually doing.

The story of Boya describes this level. When Boya played the qin with lofty mountains in mind, Zhong Ziqi heard Mount Tai rising majestically. When Boya played with flowing water in mind, Ziqi heard the torrents of a great river. After Ziqi died, Boya smashed his instrument, broke its strings, and never played again, saying that no one remained in the world worth playing for.

Notice what Boya lacked. Not an audience. He had audiences in abundance and was the most celebrated musician of his time. He lacked the person who could say what he was playing.

Applause cannot provide that. It tells us someone is present, not that anyone has heard.

In the Nicomachean Ethics, Aristotle distinguishes three forms of friendship: friendship for utility, for pleasure, and for virtue. The first two, he says, are incidental, because what is loved is not the person but what the person provides. When that provision ends, the friendship dissolves. Only the third loves someone for who they are, and such friends help one another become better.

The first two kinds can fill a room with people. Loneliness at the second level is the experience of being surrounded by them.

At the third level, recognition of the other:

Once another person is recognized as an end, one cannot help engaging in cultivation, accompanying a particular person along part of their own road so that they may arrive for themselves.

The direction of loneliness now reverses.

The first two levels lack someone who can receive me. This level lacks someone who needs to be received.

One wants to cultivate, but there is nowhere to begin.

There are two ancient versions of this sentence, one Chinese and one Greek. They describe two sides of the same thing.

Begin with the Chinese version.

In the “Xu Wugui” chapter of the Zhuangzi, Zhuangzi passes Huizi's grave during a funeral procession and turns to tell his attendants a story.

A man of Ying once smeared a speck of white plaster on the tip of his nose, as thin as a fly's wing, and asked a craftsman named Shi to cut it away with an ax. Shi swung the ax with a rush of air and removed every trace of plaster in a single stroke without injuring the nose. The man of Ying stood motionless, without changing his expression.

Lord Yuan of Song heard of this and summoned Shi. “Do it for me,” he said.

Shi replied, “It is true that I could once do it. But the person who served as my material has been dead for a long time.”

Zhuangzi then says, “Since the Master died, I have had no material on which to work. I have no one with whom to speak.”

Three details deserve pause.

First, Shi had lost none of his skill. What he had lost was the person capable of standing still. This is not the loss of an ability but the loss of its object.

Second, Lord Yuan stood before him, actively volunteering, and he was a ruler, apparently the finest possible candidate. Shi refused. There was someone beside him, but that person was not his material. This is precisely the question my friend asked: Why do we remain lonely when surrounded by people?

Third, what did the man of Ying do? Nothing. He simply did not flinch. The ax came toward his nose, and he stood still.

Classical commentators gloss the word translated as “material” as the place upon which skill is exercised. “I have no material” literally means: I have nowhere to begin.

There is another layer. The person Zhuangzi mourned had contradicted him all his life. In their famous debate over the happiness of fish above the Hao River, Huizi never agreed with him once. The previous essay argued that agreement carries no information. Here stands a man at a grave, mourning his one true opponent.

The Greek version presents the other side of the same matter.

It is the cave in The Republic. Someone breaks free of the chains, leaves the cave, sees the sun, and then returns.

Socrates says that when the person first returns, his eyes have not adjusted to the darkness. He cannot see clearly and needs a long time to become accustomed again. If he is then made to compete with the prisoners in judging shadows, they will laugh at him. They will say that his journey upward ruined his eyes and was therefore not worth making.

Then comes the sentence: if they could seize and kill the person who tried to free them, do you imagine they would not?

Craftsman Shi lacks a nose. The returning prisoner lacks someone willing to open their eyes.

The first has no object. The second meets an object that refuses, and refuses violently. These are the two forms of loneliness at the third level.

Together they say the same thing: loneliness here is not the absence of people. It is the absence of someone willing to stand still.

The man of Ying, standing without changing expression, is the positive image. The prisoners who reach for a weapon are its reverse.

At this point, something must be made clear or this essay will become a certificate we award ourselves.

Loneliness is not evidence of one's level.

“The higher the level, the lonelier the person” is a wonderfully convenient consolation. I am lonely, therefore I must be high up.

The claim is false. Most loneliness is simply loneliness and says nothing about where a person has arrived. And under the additive principle, a person at the third level retains both earlier forms of loneliness while acquiring another.

The cave story closes off the comforting interpretation by itself. The person who returns is the worst performer in the cave. He cannot see the shadows and will lose the contest. To the prisoners, he is the person damaged by the world above. He is not the strongest in that environment, but the most inept.

Plato adds another observation. Dazzled eyes have two possible causes: moving from light into darkness, or from darkness into light. When you see someone looking bewildered, do not laugh immediately. First ask which transition they have made.

The same applies to us. When we see a lonely person, including the one in the mirror, we should not rush to decide which kind of loneliness it is.

One small difference is also worth noting.

In Plato, the philosopher returns to the cave because the city requires it. He is sent back from outside. In SAE, cultivation is not assigned by anyone. One's own law has grown into that form, and on arriving there one cannot do otherwise.

The outward action is the same. One comes from an external commission, the other from one's own law. This difference is exactly what this series of essays has been trying to distinguish.

Return to the original question.

Why are we lonely when surrounded by people?

Because people are not interchangeable. Loneliness never measures how many people are present. It measures what we need at this moment and whether anyone occupies that position.

None of these three forms of loneliness can be solved by adding more people.

The first needs someone who makes the self more substantial. The second needs someone who can hear what the music is saying. The third needs someone willing to stand still while the ax approaches.

There is also a difficult fact to accept: the second form cannot be solved by finding people who agree with us. As the previous essay argued, agreement carries no information. Boya did not need someone to say the music was good. He needed someone capable of hearing whether it was a mountain or a river. A person capable of being right in that way is also capable of being wrong.

The common shape of all three forms of loneliness is this: the person we need must be capable of truly touching us.

And to be capable of touching us is to be capable of hurting us.

There is still a reverse perspective to consider.

Loneliness is not actually frightening.

It is the shape of an absence, and the shape of the absence tells us exactly what is missing.

Earlier I said loneliness is not evidence of one's level. Looking backward, it cannot prove where we have already arrived. Looking forward, however, it points somewhere. Loneliness is the part a level cannot resolve, speaking aloud.

When a person remains at a given level, whatever that level can accommodate does not feel absent. What feels missing is exactly what cannot be entered into its account, the remainder of that level.

Each form of loneliness therefore points in a direction.

The person with the first form goes in search of ends of their own. No amount of confirmation from others can fill the lack. The “I” can become substantial only by having somewhere of its own to go.

The person with the second form moves toward recognizing another person's ends. Waiting for someone who understands what we are playing leads nowhere. When we begin listening for what another person is playing, the situation changes.

The person with the third form looks toward another who is not yet present.

This looks least like an exit, yet it is the most concrete of the three.

Craftsman Shi lost his material, but he did not throw away the ax. Zhuangzi spoke at Huizi's grave and then continued writing. Plato wrote down the story of the man who returned to the cave and was nearly attacked by the prisoners.

To write something down is to extend a hand toward someone not yet present.

The previous essay said that we write not because we have finished thinking, but to hand our thought over for checking. Now we can add the other half: the person to whom we hand it need not be anyone present now.

This is where the third level differs most from the first two. The absence at the first two levels can be filled only by someone actually appearing beside us now. The third need not be. The future other is absent, but not nonexistent. They have simply not arrived.

At the grave, Zhuangzi said, “I have no material on which to work.”

Then he continued writing.

Twenty-four centuries later, that ax has reached the tip of our nose.

Whether we flinch has never been automatic.

8. Why Has Listening Become an Expensive Commodity?

An interesting news story appeared in 2026.

The latest AI boom had created a group of newly wealthy people, and San Francisco had come alive with them. Luxury homes were in demand again. Some listings even said they would accept payment in shares of private companies. None of this was new. It happens in every cycle.

The new part was another story.

In the same city, a business had emerged that charged by the hour, from $3,000 to $6,000. The people offering the service described themselves as knowledgeable about AI, computing power, longevity, and cryptocurrency. Many of their clients were the same people appearing in fundraising news during the day.

Reports described the situation this way: they had money and minds full of ideas, but no one to tell. When they did speak, no one took them seriously. Eventually they discovered that another person's serious attention could be purchased.

The first reaction is probably laughter.

But after laughing, the structure deserves a look. It is so clean that it resembles an experiment.

In the previous essay on loneliness, I mentioned Boya. When Boya played with high mountains in mind, Zhong Ziqi heard Mount Tai. When he played with flowing water in mind, Ziqi heard the torrents of a great river. After Ziqi died, Boya smashed his qin and never played again.

Boya never lacked listeners. He was the most famous musician of his time and had audiences in abundance. He lacked the person who could say what he was playing.

That is precisely what these people are trying to buy: a Ziqi.

It cannot be bought.

The reason is neither insufficient money nor bad luck in choosing a listener. It lies in the structure of the transaction itself.

When someone charging by the hour enters the room, their situation has already determined what they can say. Their words must make this hour feel worth its price and ensure another booking. Only one response is inherently safe: agreement.

The previous essay argued that another person's agreement carries no information. I cannot distinguish why they agree. Perhaps they truly understand. Perhaps they share my blind spot. Perhaps they were not listening and simply liked the sound of the sentence. All three look the same from where I stand.

Payment simplifies the question. We now have a known answer to why the listener agrees.

Ziqi was valuable precisely because he was not paid. He said the music was a mountain because he truly heard a mountain. If he had heard something else, he would have said so. The possibility that he might say something else is what gave weight to his recognition of the mountain.

That possibility is the first thing the $3,000 to $6,000 purchases away.

What the money buys is not Ziqi, but a person who cannot say the wrong thing. And someone who cannot be wrong is someone whose words count for nothing.

There is a deeper layer, however, and it is what this essay is really about.

A person wants to be heard by another person. Being heard exists only between two people who are both ends in themselves.

Ziqi was not Boya's instrument. He stood there listening because he wanted to listen. His words reached Boya precisely because he occupied a position of his own.

Purchase does the opposite. It turns the other person into a means.

The moment the transaction succeeds, the desired object is already gone. Not because the payment is too small, but because paying destroys the condition under which that object can exist.

This is not funny. It is simply exact.

One point must be made clear before we continue, or this essay may be read as an argument against ambition.

Working hard to make money is not the problem.

Making money is good. There is nothing objectionable in creating something, expanding it, enabling many people to use it, and being rewarded. The things produced during this AI cycle are real. So are the nights those people stayed awake.

The problem has never been pursuit. It is putting something in the wrong place.

Aristotle discusses wealth directly in the Nicomachean Ethics. Wealth is plainly not the good we are seeking, he says, because it is useful only as a means to something else.

Money is the purest of means. It cannot itself be eaten or worn, and it cannot keep you company. Its sole quality is that it can be exchanged for something else. Precisely because it seems exchangeable for everything, it is the means most easily placed in the position of an end.

Once it occupies that position, something follows automatically.

Persons are ends and must not be treated merely as means. Money is a means and must not be treated as an end. These are two opposite misplacements under the same law.

Nor are the misplacements independent. They form a chain. Once money becomes the ultimate end, everything else, including people, can be valued only according to its usefulness in producing money.

The room in San Francisco demonstrates the entire sentence. A person needs to be heard by another agent. The only tool in his hand is one that turns another person into a means. The harder he reaches for the desired thing with that tool, the farther away it moves.

Look one layer deeper, and the luxury homes are the first half of the same story.

When a person organizes life around an end, that end gives tomorrow its shape. They are still changing code at three in the morning because something has not yet been made. Everything sacrificed is recorded in the same ledger.

A luxury home is a receipt.

The receipt proves that the account has been settled. But a receipt organizes nothing. It does not tell you when to get out of bed tomorrow or make you want to return to your desk at three in the morning.

In the essay on death, I wrote that a person at the second level fears dying because the account has not been settled, because items remain unchecked on the list.

Now we can add the other half. Settling the account is painful too.

In one case, items remain. In the other, every item has been crossed off. The same ledger produces two forms of pain at the same level.

The emptiness is therefore not retribution, some punishment for greed. It is part of the structure of the level. This level organizes itself with a ledger, and every ledger has an inherent problem: eventually it can be closed.

It can be closed because it records means. Means are finite. Once completed, they are complete. An end is different. A genuine end is not exhausted because you have done a great deal in its service.

Then comes the room and the person charging by the hour. When the ledger is empty, the first instinct is to repair it with the ledger, because the ledger is the only tool at hand.

This was written down three thousand years ago.

In the second chapter of Ecclesiastes, the speaker who calls himself a king in Jerusalem says that he undertook great works, built houses, planted vineyards, made pools, gathered silver and gold, and withheld from his eyes nothing they desired.

Then he looked back at everything his hands had done and called it vanity, a chasing after wind.

Three thousand years. The sentence has been printed and read countless times, and it has stopped no one.

That deserves a pause, because it is itself a judgment.

In the essay on the unity of knowledge and action, I argued that principles can be taught but judgment cannot. A law can only be reached for oneself; it cannot be given by someone else. Even if a law handed to you is correct word for word, all you have received is a sentence.

Every new generation of the wealthy must therefore collide with this discovery for itself. Not because they do not read. Many read far more than we do. The experience has no channel of transmission.

The fact that Ecclesiastes must continually be reread is the best evidence for this judgment. If the discovery could simply be passed down, there would be no need to read it again.

But fairness requires another sentence, perhaps the most important in the essay.

Adam Smith wrote about a poor man's son seized by ambition. He spends his life pursuing wealth and status, sacrificing tranquility and every pleasure he might once have enjoyed. In old age, sitting among the things he has earned, he discovers that they are useless trinkets, and that maintaining them is more exhausting than acquiring them was.

Smith then adds something else: the deception is beneficial.

People's belief that these things will bring happiness leads them to cultivate land, build cities, invent machines, and organize production. An entire civilization is created by the deception.

The illusion may benefit the species. It may not benefit the person driving the vehicle.

This is not a story about stupidity. The people involved are not stupid. Many are exceptionally intelligent, and what they build is real. It is a story about a tool: the tool that brought them here no longer works here.

One final point, and it is the point this essay most needs to make.

It is easy to laugh at someone who pays $5,000 an hour to be heard. After laughing, we receive something: a sense of our own lucidity. At least we would never do that.

That feeling of lucidity is itself a receipt.

The ledger does not belong only to the rich.

Each of us has a receipt of our own: the position finally obtained, the home, the number, the post finally published and finally shared. The moment of attainment is genuinely pleasurable. By the next morning, the pleasure is usually spent.

The luxury home is simply an enlarged version. Enlargement makes it easy to recognize. The smaller version is harder to see because we are holding it ourselves.

What can we do after the laughter?

This essay will not offer a method. As the previous essays have argued, the road can only be walked for oneself.

But we can look at one thing. If the account you currently hold were settled today, what would you want to do tomorrow morning?

If the answer is empty, then the emptiness is not waiting in the future. It is already present, merely hidden behind an unsettled account.

If the answer contains a particular person, or something you would still want to finish with no one watching, then you hold something beyond the ledger.

That is the thing occupying the position of an end. Money beside it returns to its proper place: a very useful means.

Returning a means to the position of a means does not diminish it. Only then does it become truly useful, because at last it knows what it is for.

What occupies the other position cannot be bought by a receipt, or for $3,000 to $6,000 an hour.

9. What Is the Meaning of Life?

A friend asked what the meaning of life is.

First look at the shape of the question.

It assumes that meaning is something to be found. It exists somewhere, perhaps in a book, in another person, or at some moment that has not yet arrived. If we search in the right place, we can take hold of it, probably in the form of a sentence.

Almost every serious philosophy has answered differently.

It offers not a sentence but a road, one divided into stages, ordered, and requiring time to travel.

This deserves a pause, because it is no coincidence.

If meaning could be given in a sentence, the sentence could be repeated. You could tell it to me; I could write it down; then I would possess it. But a meaning that can be transmitted in this way is plainly not what those philosophies describe. Otherwise the matter would have been settled thousands of years ago. Someone would have written down the sentence, and everyone afterward would only need to copy it.

Wittgenstein states the reason most cleanly.

Two sentences appear many pages apart in the Tractatus Logico-Philosophicus. One says, “I am my world.” The other says, “The sense of the world must lie outside the world.”

Put them together: my meaning must lie outside me.

Wittgenstein then says that the world is the totality of what can be said. What lies outside the world cannot be said. The book therefore ends with the sentence: “Whereof one cannot speak, thereof one must be silent.”

This is not mysticism. It is a conclusion about form: meaning does not belong to the range of things that can be spoken.

It can only be traveled, not stated.

And when something can only be traveled, the written form it takes is a sequence of stages.

In earlier essays, I argued that a law can only be reached for oneself and cannot be given by another person. Here the same point appears in different words. The unsayable and the intransmissible are two sides of the same thing.

What, then, might the road look like?

Place accounts from different eras and cultures beside one another, and a roughly similar order appears.

I will refer several times to a passage from Confucius, so one point should be clarified at the outset. “At thirty, I stood firm; at forty, I was free of perplexity; at sixty, my ear was attuned.” The subject of the whole passage is “I.” Confucius is describing the road he traveled, not prescribing ages for everyone else. Later readers turned it into a timetable for life. They mistook a personal account for an external standard, which is precisely the reading those sentences least support.

The first stage is finding oneself.

Before there is an “I,” the question of meaning has nowhere to land. Meaning is always the meaning of something, and that something does not yet exist.

This stage is usually unattractive: imitation, trial and error, pretending to be someone else, changing one's language from one circle to another. Then one day we discover that some things genuinely matter to us, while others were merely learned.

This is what Confucius meant by standing firm at thirty.

The second stage is finding one's genuine ends.

Once a self exists, the next question is where that self will go.

Notice that these must be one's genuine ends, not ends handed down by someone else. Many people have goals, but their families, professions, or peers supplied them. There is still a road between possessing an end and possessing an end of one's own.

Confucius speaks of being free of perplexity at forty. This does not mean understanding everything. It means no longer swaying from side to side because one knows what one wants.

This is the stage in which life feels most saturated with meaning. Something must be done, and tomorrow acquires a shape because of it.

The third stage is turning toward others.

Here the direction changes. The question is no longer only what I want, but whether I can recognize that another person is also a source of ends that do not exist to serve me.

Confucius speaks of the ear being attuned at sixty. The phrase is exact: one can listen, without first translating another person's words into the question of whether they are useful to oneself.

History has repeatedly described these three stages under different names.

Kierkegaard wrote a book whose title can be rendered literally as Stages on Life's Way. The aesthetic stage lives in immediate experience and lacks a unified self. The ethical stage chooses the self and assumes obligations; the self becomes a task. Then comes the religious stage.

Nietzsche wrote of three transformations: the camel, the lion, and the child. The camel bears what others load upon it. The lion says no and wins the right to say, “I will.” Then comes the child.

In psychology, Erikson gives a similar sequence: identity, intimacy, and generativity. He defines generativity as establishing and guiding the next generation, whose opposite is stagnation.

One person's example is especially worth considering because he revised his own account.

Almost everyone has seen Maslow's pyramid, with self-actualization at its summit. In 1969, after suffering a heart attack, he wrote two articles for a newly founded journal and added another level above self-actualization.

He described it as functioning as an end rather than a means and entering into relation with oneself, significant others, people in general, nature, and the cosmos.

Here was someone who began in psychology rather than philosophy, reached the top of his pyramid, discovered something above it, and used the vocabulary of ends and means to describe what he found.

He also admitted that it was difficult to say clearly what lay beyond self-actualization.

Fairness requires another point. Maslow's hierarchy was later subjected to a well-known test that found no supporting evidence. It therefore cannot serve as proof. Its value lies elsewhere. Someone who followed an entirely different path arrived at roughly the same place and picked up the same words.

Yet the most revealing feature of these accounts is not where they resemble one another.

It is where they fail to align.

The first two stages are nearly a consensus. A self must first exist; then it must have ends of its own. From Confucius to Kierkegaard to Erikson, no one disputes this order.

All disagreement concentrates in the third stage.

Nietzsche's third transformation is the child, creativity, and a new beginning. It is not another person.

Kierkegaard's third stage is religious and faces the absolute, not another person.

Wittgenstein places meaning outside the world, but there are almost no other people in his book. There are only the “I,” the world, language, and the unsayable beyond. Structurally, he takes the same step described here: meaning must lie outside the self. But he does not identify that outside with another person.

The four stages of life in the Indian tradition end in renunciation. Their direction is withdrawal.

The path in the Zhuangzi equalizes things, dissolving the distinction between self and the ten thousand things. It too turns back inward.

Maslow admits that he cannot explain the final level clearly.

These disagreements might look like evidence that the third stage is the least substantial. In fact, they show the opposite.

No one argues over a place no one has visited.

The first two stages attract little dispute because the road is easy to recognize and many people travel it. The third inspires disagreement because people have arrived there and seen different things. The disagreement itself is evidence that something is there.

Several qualifications must now be made, or this essay will become something else.

First, not everyone will complete all three stages.

This is a process, not a timetable. Some remain in the first stage for a long time. Some spend their lives in the second. Some reach the third and then retreat.

Stopping at a stage is not failure.

Each stage has a law of its own. To live in alignment with the law of one's stage is to live honestly. That has nothing to do with how far one has traveled.

Second, speeds differ, and there is no correct speed.

Some complete the first stage at twenty. Some do not begin until fifty. Others follow a different sequence. Any chart aligning stages with ages mistakes the journey of one person for the scale by which everyone must be measured.

Third, this essay cannot supply the answer, and must not.

If it could tell you the meaning of life in a sentence, that would prove everything it has said to be wrong.

It can do only one thing: explain why the question offers a road rather than an answer, and sketch the road's possible form so that a person may have some sense of where they are.

Return to the question.

What is the meaning of life?

The question cannot be answered. It can only be lived through.

But the fact that someone is asking already tells us something.

To ask the meaning of one's life, a self must first exist whose meaning can be at stake. If you can voice the question, the first stage has already been completed.

The next two stages must be walked for oneself. We can study the roads others have taken. Afterward, we still have to lift our own feet.

10. Should Children Practice Speed Arithmetic?

The real question is what children are being trained to do when they practice speed arithmetic. Once that is answered, we can decide whether the training is worthwhile.

Begin with a fact that is easily skipped: speed training genuinely helps. It is not a fraud.

Human working memory is small. In a moderately complicated problem, if elementary calculations continue to occupy mental space, too little remains for the truly difficult part. The more automatic the basic operations become, the more room is left for thought. This is well established in educational research and needs no defense.

The problem is therefore not speed.

It lies elsewhere, and seeing it requires us to avoid an appealing but invalid criticism.

The criticism says that speed arithmetic trains mindless reflexes and therefore obstructs thought.

That conclusion does not follow, because a response that no longer requires thought is not something inferior.

An experienced driver does not derive the proper moment for braking while driving. A fast reader does not consult a dictionary word by word. Someone who truly understands a matter may judge a related situation directly, without an intermediate moment of “let me think.”

Skill consists precisely in allowing something once thought through to settle beneath attention, where it no longer occupies mental space.

If the criticism were valid, practicing music, handwriting, and vocabulary would all impede development. Plainly they do not.

The real question is not whether training makes something automatic, but what lies beneath the automatic response.

Beneath musical technique lies music. You no longer think about fingering so that you can think about how the phrase should sing.

Beneath fluent reading lies meaning. You no longer struggle to recognize characters so that you can think about what the passage says.

What lies beneath the automaticity of speed arithmetic?

A set of formulas.

A formula is the finished product of reasoning performed by someone else. The child receives the conclusion, not the derivation. The faster and more accurately the child executes it, the less need there is to know why it works.

That is the problem, and it has nothing to do with speed.

The structure of the training excludes “why” from the score. It does not reward understanding. It rewards correctness in the absence of understanding. If a child who understands and one who does not achieve the same speed, their scores are identical.

What does not enter the score will not be optimized. This requires no ill will from anyone. It is the natural result of any evaluative system.

At this point, we can set speed arithmetic aside. It is not an exception but the purest example.

Many school subjects do the same thing, though less thoroughly.

Students memorize templates for solving problems: when you see this type, use this method. Essays acquire fixed structures for introductions and conclusions that reliably earn points. Historical events come with standard lists of significance and consequences; write down the whole list and receive the grade.

What these forms of training teach in common is not their respective content but something more fundamental: correctness comes from following instructions.

You do not need to know why the template works, only to recognize the type of problem. You need not have considered what an event means, only to reproduce the required points. A correct answer is a correct answer, and once it has been given, no one asks anything more.

Once this attitude takes shape, it does not remain on a mathematics exam.

Someone repeatedly rewarded from childhood for “following the steps and getting it right” will meet any ready-made account by first receiving it, memorizing it, and using it. They judge whether a claim is good by whether it works, whether someone says it is right, and who that person is.

In their experience, correctness has never been derived for oneself. It comes from outside and is assessed from outside.

That is exactly the condition in which someone else can most easily decide for you.

Not because the person is unintelligent. The opposite may be true. The more completely a child excels at this training, the smoother the channel becomes, because experience has confirmed the same lesson repeatedly: receive what comes from outside, follow it, and win.

Where does the problem arise?

Not in skill. Skill is good. Not in training. Training is necessary.

The problem lies in position.

Skill is properly a means. We practice until we no longer need to think in order to free the mind for what must be thought about. Its entire value lies in serving what comes next.

When skill itself becomes the end, however, when it becomes something to compete over, certify, and use as proof of a child's excellence, the positions have changed.

The previous essay described the misplacement of means and ends through money. Money is the purest means. Placed in the position of an end, it becomes empty, because all its content lies elsewhere.

The same movement appears here with a different object.

This time the problem is more serious because the child is placed in the position of a means.

Once a skill becomes an end, the child becomes its vehicle. The faster the child calculates, the more completely the end is realized. Whether the child has asked why falls outside the end and therefore does not matter.

A structure that treats a person as a means usually requires no malicious person. It requires only a score sheet.

One point must be clear, or this essay will become an argument against teaching children anything.

It does not oppose training or deny the importance of fluency.

A person who does not know multiplication facts cannot do any moderately complicated mathematics. That is true. Telling children to “explore freely” while withholding solid foundations is not respect. It is abandonment.

The distinction lies in one question: Is there room for “why”?

Two children may each practice a hundred problems. If the first asks why the formula works, someone answers seriously, perhaps even saying, “Try to see whether you can work it out.” For the second, the question never appears. If it does, the answer is, “Don't worry about that now. Get your speed up first.”

The two forms of training may look identical on a report card.

How can we tell which one a child is receiving?

Ask a direct question: Can they explain why the method works?

There is nothing wrong if they cannot. Most children cannot, which is normal. Many things must first be used and understood later.

So ask a second question, the one that really matters: Have they ever wondered why?

If they have wondered, tried, and become stuck, they simply have not arrived yet. That is a matter of time.

If the question has never occurred to them, or if hearing it feels strange, like asking something one should not ask, the issue is no longer time. It means that in their experience “why” does not belong to learning.

In the first case, the road remains unfinished. In the second, no one has ever pointed out the road.

One final point.

We all grew up within some version of this training, and many of us succeeded because of it. This essay is not only about other people's children.

The adult version of the same action persists. Faced with something new, we first look for a ready-made account. Once we find and receive it, we stop thinking further.

When learning a skill, we first search for a methodology. When encountering a difficulty, we look for someone else's summary. When reading an argument, we first ask who said it and how many people agree.

None of this is wrong. Materials are meant to be used.

The question is whether “why” follows.

Whatever another person has already cut and shaped for us, what we receive is the stone, not the strike of the chisel. The stone can be used, and used well. But only the stroke one makes oneself belongs to oneself.

What we can give a child is actually very limited.

We cannot give understanding. A child must arrive there alone. What we can give is permission for the question to exist.

If a child knows that “why” may be asked, that someone will take it seriously, and that it will not be cut off by “don't worry about that now,” the child will walk the rest of the road.

11. Why Does a Path Everyone Praises Begin to Feel Wrong?

A friend described a feeling that was difficult to explain.

Everyone says the path is good. Many people take it. Experienced predecessors are already there, and the methods are mature. He follows it and does reasonably well. Yet the farther he goes, the more something feels wrong. He cannot say what the mistake is or identify a particular event to blame.

The feeling is common and has one distinctive feature: from outside, everything is right. There has been no failure or upheaval, and no one has treated him badly. Nothing that can be pointed to appears to be a problem.

Let us begin by asking what this path contains.

1. The path contains everything, including the most reassuring thing of all

The advantages of a well-traveled path are real. They should not be underestimated.

Others have gone before. The methods are mature, predecessors can offer experience, the standards of evaluation are ready-made, and earlier travelers have marked the places where mistakes occur. The path is lower in risk and higher in efficiency. At every moment, you know how far you have progressed.

It also contains the most reassuring thing of all: mutual confirmation.

The people on the path assure one another that they are going the right way. Look how many of us are here.

But one distinction must be made precisely.

Confirmation can prove that a path is viable. It cannot prove that the path is yours.

The first fact is useful. A path successfully traveled by many people has indeed been successfully traveled. That information is true.

The second question is unaffected by numbers.

When a group shares the same way of traveling, it also shares the same blind spot. No matter how many people join it, they cannot illuminate what their way of seeing excludes.

Yet the two facts are easily conflated. Because many people have reached the destination, “this path can be traveled” becomes “this path should be traveled.”

2. It does not lead somewhere bad; it leads somewhere belonging to someone else

Where, then, is the problem?

It is not that the path leads to a bad place. It leads to a real, reachable destination. That is why it is broad.

The problem is elsewhere. Its direction is not yours, but someone else's.

The path was worn into the ground by other people's ends. Those who follow it arrive quickly and steadily, only to discover that they have reached a place someone else wanted to go.

This is probably the unease my friend described.

He is not traveling badly, nor is the path defective. The path is carrying him very efficiently toward a destination that is not his. Nothing along the way can reveal the problem because there is nothing wrong with the path itself.

That is why he cannot say what is wrong. Everything that can be pointed to is correct.

The third chapter of the Bhagavad Gita says this more directly than anything I could devise: better to perform one's own duty imperfectly than another's duty well. Better to die in one's own duty; another's duty brings danger.

Notice the final clause. The original word means, quite literally, that it brings fear or danger. Doing another person's work well is not safety in this passage. It is danger.

That sentence supplies something a purely structural account cannot.

To say that each person has a path of their own is only a description of structure. It assigns no weight to the fact. It does not tell us what it is worth or whether the price should be paid. The Gita gives it weight all the way to the bottom: better to die on one's own path.

A structural claim alone cannot derive that sentence.

The sentence, however, has a problem of its own. Traditional commentaries often bind “one's own duty” to caste and stage of life. You should remain in the position into which you were born. Read that way, it does not say each person has a road of their own. It says each person has a cell of their own.

The same sentence can defend social order or individuality.

This essay uses the latter reading. If I did not state that openly, I would merely be borrowing an ancient sentence for courage.

3. What about the other path?

At this point, the next thought arises almost automatically: then I should take the other path.

And an old sentence is waiting for that thought.

In the seventh chapter of the Gospel of Matthew: enter through the narrow gate. Wide is the gate and broad is the road that leads to destruction, and many enter through it; small is the gate and narrow the road that leads to life, and few find it.

The passage is widely known. For that reason, it is especially easy to misread.

The easiest interpretation is: because few people take the road, those who do are special. The next step is “I have entered the narrow gate; I belong to the few.”

Anyone can make that inference. It produces superiority and humility at once and cannot be refuted. Before going further, then, we must clarify what “narrow” means.

(Several other sources will appear below. They are neither evidence for these claims nor illustrations of them. I place them together because each illuminates something the others cannot, and some illuminate precisely the defects in the claims themselves.)

4. The road is not empty because the gate is narrow

Begin by reversing the usual view.

It is not that few people travel because the gate is narrow. The place appears narrow because few people travel there.

Why are there so few? Because the gate is your own.

This does not mean there is one narrow gate in the world that everyone must find. The opposite is true. Each person has one, and each is in a different place.

The gate is wide enough for only one person. That is true for everyone. It is narrow for you, for me, and for whoever stands before their own.

Entering one's own gate therefore distinguishes no one. Everyone has one. Finding yours does not place you above anyone else. It means only that you are walking your road.

Narrowness is not an honor. It is a measurement.

Nietzsche expresses this in the shortest possible form. In Thus Spoke Zarathustra, someone asks him where “the way” is. He replies: this is my way; where is yours? There is no such thing as the way.

The sentence is almost impossibly clean. Where Nietzsche's own book goes next is another matter. The last of his three transformations is the child, creativity, and a new beginning. After asking, “Where is your way?” he takes no further outward step toward another person.

He gives this section its sharpest sentence while also marking the sentence's boundary. Saying that each person has a road and moving toward another person are two different things.

5. The difficulty is the absence of anything to follow

What makes the narrow road difficult?

There is nothing to use as a reference.

No one has gone this way before. There is no predecessor's experience to borrow, no companion's confirmation to receive, and no marker along the way to tell you that you are right.

You do not know whether the direction is correct, how far remains, or whether anything waits at the end. None of the benefits listed in the first section are available here.

The difficulty is not passing through the gate. It is continuing without confirmation.

The only confirmation is arrival, and it can never be obtained in advance while traveling.

An old teaching can be placed here. Near his death, the Buddha instructed his followers to be lamps unto themselves, to take refuge in themselves, and not in another.

At the final moment, the founder of a religion told his followers not to depend on him.

But the original teaching contains two lamps, not one. It continues: take the Dharma as your lamp and refuge.

That deserves a pause, because it touches a real problem.

If everything must be reached for oneself, what is the second lamp?

One answer is that the law points universally and applies to anyone, but binds only the person who holds it. It is not a manual to be copied, yet neither is it something each person invents at will.

This is close to the image of the two lamps, but not identical. The Buddha's second lamp is given. It can be heard and passed down. If a law can only be reached for oneself, however, what can be passed down are words about it, not the law itself.

The difference is not small. It is the most difficult point on this whole road. I place the passage here because without it, the difficulty would never become visible.

6. Narrow does not mean one's own

One danger must be prevented because it is the most common modern version of the claim.

Being unconventional does not mean one is walking one's own road.

Narrowness tells us only that few people are on the road, not that the road is yours.

Three people may copy the same obscure source. Their road will be exceedingly narrow without a single inch having been made by them. An esoteric theory, a position no one accepts, or a deliberate rejection of everything mainstream can be both narrow and entirely borrowed.

What is borrowed does not become one's own merely because it was borrowed from an obscure place.

The test is therefore not how many people travel there. It is whether I cut this step for myself or simply received it.

At this point, a sentence from Chapter 41 of the Daodejing should be mentioned.

That chapter says that when the highest kind of person hears the Way, they practice it diligently; when the middling person hears it, they half believe and half doubt; when the lowest hears it, they laugh aloud. Then come six Chinese characters meaning: if it were not laughed at, it would not be sufficient to be the Way.

Those words turn ridicule into evidence of correctness.

If you laugh at me, that proves I am speaking the Way. It is a perfect device. It cannot be refuted because every objection has already been absorbed as favorable evidence.

The fact that it grows within one of the most revered classics tells us more than anything else could.

Such immunizing devices are not merely defects of character. They grow naturally from claims that only a few can understand. The moment you say, “Most people cannot see this,” you create an unfalsifiable line of retreat.

This essay belongs to the same class.

The Daodejing passage is therefore not cited to confirm anything. It is a warning: everything said above about narrowness can be converted into an exemption from scrutiny in exactly the same way.

7. My reading of “destruction”

Return to the heaviest word in the passage from Matthew. The broad road leads to destruction.

The word is usually read as an eschatological judgment: the person takes the wrong road and is punished at the end.

I can only state how I read it. This is my interpretation, not the original meaning of the scripture. A first-century text did not speak in the language of agency used below. I am transplanting the sentence, and readers are free to reject the transplant.

The same Gospel contains a more direct line: what does it profit a person to gain the whole world and forfeit their life?

Notice its structure. Gaining the whole world means that the road worked and the traveler reached the destination and received everything. The cost falls nowhere else. It falls upon the traveler.

Read in this direction, the broad-gate passage is not a threat but a description.

To walk another person's road all the way to the end is to have no “you” walking it from beginning to end.

What is lost is not something withheld as punishment at the destination. It is what has been leaking away throughout the journey.

On this reading, destruction is not a penalty imposed from outside. It is a property of the road itself and requires no one to enforce it.

Again, this is only my reading. The scripture says the road leads to destruction. Reading that destruction as the loss of oneself is my addition.

The passage from the Bhagavad Gita requires no such addition. It directly says that another's duty brings danger. That is already in the text.

8. The lock installed by the text itself

The same chapter of Matthew contains another passage close to the concern of this essay, and it requires no transplantation.

Immediately after the two verses about the narrow gate comes a warning against false prophets, followed by the claim that they will be known by their fruits. Then comes the harshest passage: many will say on that day, “Lord, Lord, did we not prophesy in your name?” The answer is, “I never knew you.”

These are the people most certain that they have already entered the narrow gate.

The chapter's structure is therefore: first the narrow gate, immediately followed by a warning to those convinced they are already inside.

The claim and its safeguard sit beside each other. In later circulation, everyone remembers the first part. Few mention the second.

It forms a pair with the six characters from the Daodejing.

One ancient text places a lock beside its own claim. Another leaves a back door at the same point.

Both now lock or open upon us.

9. Narrowness is a moment, not an attribute

There is another point, perhaps the most concrete in the essay.

Narrowness is not the essence of the road. It is the road's present condition.

It is narrow because no one has passed there. Once someone passes, the surrounding ground grows broad.

Every road now broad enough for countless travelers was once narrow. The first person to think, live, or act that way had no road before them. It is broad now because someone walked it.

This can be checked. For most of today's broad roads, we can look back and identify the first traveler.

But the point must be stated precisely or it becomes an inspirational slogan.

What widens after someone passes is not the gate but the ground around it.

Each gate belongs to a person. What you leave becomes material for others, and you also prove that passage is possible. But the gate another person must cross is still their own.

The finished product can remain. The stroke of the chisel cannot.

The gate therefore remains where it was, still wide enough for only one. The next person must walk for themselves.

Each generation must cross pathless ground again, not because its predecessors traveled badly, but because the gate is indexed to the person.

10. Two things about other people's roads

First, another person's narrow gate cannot measure your own.

If their gate was difficult at a certain point, yours need not be difficult there. Suffering does not prove you are right; the absence of suffering does not prove you are wrong. The claim that greater difficulty means greater truth mistakes another person's terrain for a standard.

Second, you cannot point out where another person's gate is.

You can say that a gate once stood here. You can leave behind what helped during your own passage. You can remain beside them.

But you cannot say, “Your gate is over there.” To say it is to decide the form of another person's road, when the road's entire meaning lies in being made by that person.

The most we can offer is this: do not walk for them. Remain beside them when the going is hardest.

In closing

Return to my friend's question.

Why does a path everyone praises begin to feel wrong?

Because everything that can be checked is correct. The one thing wrong with it is precisely what cannot be checked: it is not yours.

The place that belongs to you probably has no road yet.

The gate is not narrow because its threshold is high or admission is scarce. It is narrow because there are no footprints.

It is narrow for everyone. Each person has a gate, and each gate admits only one.

Walking a road without footprints does not place you above anyone. It means only that the road ahead has not yet been widened by travel.

Whether it will become wider depends on whether you make it through.

And the reason you must do it yourself is simple: a road may be left for others, but the act of crossing can only be your own.

12. What Is Love?

A friend asked what love is, adding specifically: not only romantic love.

The addition matters because it opens the question more widely. A parent's love for a child, friendship, the relation between teacher and student, even the care that arises for someone met only a few times, all belong to the question.

Still, we must begin with romantic love, because the most popular account appears there and governs almost everyone's imagination of love. And love that truly lasts eventually reaches the same place as love in its broader forms.

The account can be expressed in two words: one's “other half.”

1. The sphere split in two

The idea has a very ancient source.

In Plato's Symposium, Aristophanes tells a story. Human beings, he says, were not originally as we are now. They were round, with four arms and four legs and enormous strength. After offending the gods, they were split in half.

Since then, each person has been half a being, spending a lifetime searching the crowd for the half cut away. Once found, the two cling to one another and long to grow back into one.

The story has been told for twenty-four centuries. It is told so beautifully that almost no one now recognizes it as an account. We take it for a fact.

Its structure is clear: I am missing a piece, and you have exactly the shape of that piece.

Several things follow automatically.

Because I am incomplete, I must search. The standard is whether the pieces fit. Once they do, I am finished.

But a deep problem is hidden here. If the other person is the piece that completes me, what I want is not actually that person. I want the shape.

Anyone who can fill it will do.

This is not a reproach. It is only a description of the structure, which does not point toward a person.

2. Wholeness and completion

We must therefore clarify the foundation of everything that follows.

A person is whole. They are not half of something.

No one is born missing a piece held by another person. You do not need to find someone in order to become a whole person. You already are one.

But wholeness is not completion.

Wholeness means that you do not lack a part held by someone else.

Completion means that you have finished growing and will no longer change.

A person can be whole and still very far from finished.

The two are often confused, and their confusion produces two opposite problems.

One looks forward: if I am incomplete, I must find the person who can complete me. This treats a person as a component.

The other looks backward: if I am whole, no one need concern themselves with me, and I need concern myself with no one. This mistakes wholeness for completion and not lacking a part for having no need to grow.

In fact, a person who lacks no component continues to grow, perhaps for an entire life.

A new question follows. If no one is needed to complete us, what place does another person have?

3. The desire to become one destroys what it seeks

Consider first the deepest desire: to become one.

The desire is real. There is no need to pretend otherwise. Many people have known moments when two long to become one, and such moments can be good.

But if union is treated as the actual goal, a strange result follows.

If two people truly become one, there is no other person left.

And without another person, no one can see you.

This point should be unfolded slowly.

Whether a person is cherished cannot be declared by that person alone. You may tell yourself all day that you are wonderful, but the sentence has no weight. To carry weight, it must come from another person, and that person must genuinely be another: someone with judgment of their own, who may disagree and is entirely free not to see you that way.

It is because they could see you otherwise that their seeing you this way counts.

The desire is therefore self-contradictory. What it seeks can be given only by someone who is not itself, while what it proposes to do is eliminate that person.

Hegel describes a related scene, though he is discussing something else. Two people meet. Each wants recognition from the other, and neither will recognize first. They fight. One wins and reduces the other to a subordinate. The victor soon discovers that recognition from someone he has crushed into subordination is worthless.

The methods are opposite, but the result is the same. Overpowering another person and dissolving into them both prevent that person from remaining a person.

Love, then, does not lie in the direction of becoming one.

4. What is it, then? Helping someone become themselves, not shaping them

If love is neither completion nor fusion, what is the other person's place?

I want them to become themselves.

The sentence sounds gentle, but it is the sentence in this essay most vulnerable to abuse, so it must immediately be constrained.

Helping a person become themselves does not mean making them into something.

I do not know what they will become. This is not modesty but fact. They are still growing, and they do not know either. If I already hold a blueprint, what I want is not the person but the blueprint.

The previous essay recalled an ancient story. Two gods saw that Hundun, the being of primal chaos, lacked the seven openings of the human face. Meaning well, they carved one opening into him each day. On the seventh day, Hundun died. They bore no malice. They were simply too certain about what a person should become.

What can we do instead?

Something quite simple.

Do not walk for them. The road is theirs, and no one else can make the crossing in their place.

Remain beside them. Be there when the road is difficult, not only when they want advice.

Speak when speech is called for. To see something and say nothing is not respect but indifference. The person is still growing and will sometimes go astray. Someone who never points anything out treats them as a completed object incapable of change.

And when it is our turn to be addressed, do not retreat. This is the hardest part, and it is where reciprocity begins.

5. Need is something else

One thing must be said, or the essay will feel unreal.

Love of course contains need. We need the person to be present, to answer the phone, to know they are still there. That is real and nothing to be ashamed of.

What this essay distinguishes is not need itself, but need directed at a function.

There is a simple test.

Could someone else provide what you need from this person?

If so, what you want is a function: someone to listen, someone to prove that you are worthy, someone to prevent loneliness. Anyone could perform it, so the need never pointed toward this person in the first place.

If no one else could provide it, then what you need is the person. In that case, what you need is precisely the part of them that moves of its own accord, has other thoughts, and differs from you.

Such need is irreplaceable because its object is exactly what you do not control.

6. Helping one another become

In the Nicomachean Ethics, Aristotle distinguishes three kinds of friendship: friendship for utility, friendship for pleasure, and a third kind.

The first two, he says, are incidental. The person is not loved for themselves but for what they provide. When the provision ends, the relationship dissolves. Aristotle is not condemning anyone. He is describing the replaceable position the other person occupies in those two relations.

The third he calls friendship based on virtue. Here the person is loved for themselves. He writes specifically that such friends help one another become better.

He wrote down the idea of helping one another become twenty-three centuries ago.

He adds three practical observations: such friends are rare; the friendship requires a long time; and both people must be capable of it. One side is not enough.

These sentences fully describe the difficulty. This is not a state one person can achieve alone.

Rilke expresses the same idea more warmly, and just as precisely, in his letters to a young man.

Love, he says, is not surrender, devotion, or merging with another person. He even asks what kind of union that would be if it were confused and unfulfilled.

His alternative is this: love is the work by which a person matures and completes a world within themselves, and does so for another person.

Notice the sentence's shape. I do not give myself up for you, nor do I merge with you. For you, I become myself.

Rilke offers two further sentences.

One says that the highest responsibility between two people is to guard each other's solitude.

The other is finer: a worthier love consists of two solitudes that protect, distinguish, and honor one another.

Distinguish one another. Those words rarely appear in an account of love, yet they are the meaning of this entire essay.

Rilke also says something more concrete: love is good because love is difficult. For one human being to love another may be the most difficult and weighty task given to us. Everything else is preparation for it.

In closing

Return to my friend's question.

What is love?

It is not finding the piece that completes you. There is no such piece, because you were never missing it.

Nor is it becoming one. Once two become one, the person capable of seeing you has disappeared.

It is something more like this: there is someone whom you hope will become themselves, and your hope carries no blueprint.

You do not know what they will become. You want only to remain beside them and watch it happen; to speak when speech is needed and not retreat when it is your turn to listen; to let them walk their road while you walk yours.

If they happen to see you in the same way, two people walk the same stretch of ground, each on a road of their own, looking after one another.

It is not one person becoming half of two. It is two whole people, each growing a little more fully into themselves.

This is rare. It takes a long time. One side cannot accomplish it.

But it exists. And it is much better than the sphere split in two.

13. Was Confucius's Journey the Chinese Odyssey?

With the recent popularity of the film The Odyssey, a friend asked whether Confucius's travels among the states were the Chinese version of the same story.

It is a good question, and on the surface the resemblance is real. Both stories involve years of wandering, severe hardship, and doors closed against the traveler. Odysseus spends ten years at sea. Confucius remains abroad for fourteen, leaving Lu at fifty-five and returning only at sixty-eight.

Those fourteen years were not easy. Passing through Kuang, Confucius was mistaken for Yang Hu, whom he resembled, and surrounded by the local people for five days. In Song, while he and his disciples practiced ritual beneath a great tree, the military commander Huan Tui tried to kill him and had the tree cut down. In Zheng, separated from his disciples, Confucius stood alone outside the east gate and was described as looking like a homeless dog. Between Chen and Cai, the group went without food for seven days, and his followers became too weak to stand.

Yet the structures of the two stories are almost exact opposites. Their differences are far more interesting than their similarities.

1. Ithaca is always there; the place Confucius seeks never appears

Odysseus is returning to Ithaca, to the island, the room, and the person waiting for him.

Every episode in the epic is a variation on being prevented from returning: the Cyclops, the goddess's island, the storm at sea. The obstacles may number a thousand, but the destination never disappears. It waits from beginning to end.

Confucius is not trying to return home. He is looking for a place that will put his teaching into practice.

He leaves Lu because it has no use for him. He travels through Wei, Cao, Song, Zheng, Chen, and Cai, reaching the border of Chu. Sima Qian says he approached more than seventy rulers and none employed him. Later scholars consider the number exaggerated, but the point is correct: none of the rulers he met was prepared to implement what he proposed.

In fourteen years, the place never once appeared.

That is the fundamental difference between the journeys.

An end entirely within one's own hands can be completed. However difficult the path, if the person remains alive and moving, arrival is possible.

An end that requires another person to receive it cannot be completed unilaterally.

Odysseus wants to return. However dangerous the return, it is something he can accomplish.

Confucius wants his teaching to be put into practice. That is not in his hands. No distance traveled, effort made, or intelligence applied can guarantee it.

The problem is not insufficient effort. Ends of this kind simply cannot be completed by one person.

2. A journey without a waiting destination cannot become an epic

Follow this distinction and something interesting appears.

The Odyssey has a complete shape: departure, suffering, return. That shape depends on Ithaca's remaining where it is. Once a destination exists, every event in between becomes an “obstacle.” Without one, those events are simply things that happened.

There is no Travels of Confucius narrating those fourteen years.

In the Analects, only scattered fragments remain: a few dozen characters about going hungry in Chen; Zilu spending the night at the Stone Gate and meeting its keeper; the madman of Chu passing Confucius's carriage while singing; encounters with hermits working in the fields who mocked him.

No arc connects the fragments. They do not point toward a destination because no destination exists.

Form and content therefore agree. A road with no destination waiting at its end cannot be written as an epic. It can leave only fragments.

Confucius thus reveals one of the Odyssey's premises: the epic can be an epic because the island remains there.

3. One man spends a lifetime calculating how something may be done; the other does what he already knows cannot be done

Homer's first description of Odysseus calls him a man of many turns, resourceful and cunning.

Everything he does is a calculation of how to succeed: the wooden horse, the disguise as a beggar, binding himself to the mast to hear the Sirens. He is the person who always finds a way. The force driving the entire Odyssey is this search for a way.

In the Analects, a stranger says something about Confucius.

Zilu spends a night at the Stone Gate. In the morning, its keeper asks where he comes from. “From Kong,” Zilu replies. The keeper asks:

“Is that the man who knows it cannot be done and does it anyway?”

A gatekeeper says everything in one sentence.

What is most striking is that it is not a disciple's praise but a passerby's judgment. Even the gatekeeper could see that the thing would not succeed.

Did Confucius himself fail to see it?

He saw it. In the Analects, Zilu says, “That the Way cannot prevail has long been known.”

This is not the story of someone who thought he would succeed.

4. Knowing it cannot be done is not the same as having no choice

A distinction deserves pause because it is easily blurred.

Earlier essays spoke of what one cannot but do. When a person arrives at a certain place, their law has taken that form and they cannot act otherwise.

“Knowing it cannot be done and doing it anyway” adds something. The person sees clearly that success will not come, and still acts.

To have no choice can sound like the absence of information. To know it cannot be done means possessing all the information, including the most discouraging fact, and continuing nevertheless.

This blocks a common misreading.

“Walk your own road” is easily heard as “and therefore it will lead somewhere.” If the road is truly yours, the world will eventually reward you.

Confucius's fourteen years say the opposite. It was his road, and it did not lead through.

Not one of the rulers he met put his teaching into practice. Walking one's road guarantees nothing.

That is not a pleasant sentence. But without it, everything previously said about one's own road becomes a veiled promise.

5. What the Odyssey illuminates in return is more weighty still

The comparison cannot run in only one direction. The Odyssey contains two things absent from the story of Confucius, things the preceding essays might too easily omit.

The first is the bed.

Odysseus returns, kills the suitors, and stands before Penelope. She does not recognize him.

Their son is distressed and asks how his mother can act this way. The nurse has recognized the scar on his leg. Penelope still refuses.

Then she does something. She tells a servant to move the bed out of the bedroom and prepare it for the guest.

Odysseus explodes. The bed cannot be moved, he says, because he made it himself. An olive tree grew in the courtyard. He used its living trunk as a bedpost, built walls and a roof around it, and fashioned the bed from the tree still rooted in the ground. No one could move it without cutting down the tree.

Only then does Penelope recognize him.

The scene deserves a slow reading.

She does not identify him by appearance, because appearance can be imitated. She does not trust his own words, because words can be invented. She tests him with something only the real person could know.

More remarkable still is the nature of the object: he made it with his hands, it is rooted in the earth, and it cannot be moved.

These essays have repeatedly discussed verification. Independent verifiability is a condition of a claim's validity. A claim that cannot be checked is structurally indistinguishable from a false one.

But there is something we have not considered.

The person performing the verification is not the one who returned. It is the one who stayed and waited.

We have treated verification as the task of the person who goes out: investigate, question, cut into the claim. In this story, the man who has traveled for ten years and endured everything is impatient to be recognized. The woman who has remained in the room for twenty years insists on testing him first.

The person eager for recognition and the person who insists on verification have exchanged their usual places.

Turn this light back on Confucius, and it reveals an absence.

What Confucius lacked during those fourteen years was not a ruler willing to employ him. He lacked someone willing to test him.

Employment and testing point in opposite directions. To employ is to accept; to test is to begin with doubt.

The rulers who received Confucius offered courtesy, and courtesy does not test. The first thing Duke Ling of Wei asked was how much salary Confucius had received in Lu. Later he asked about warfare. On another occasion, while the two were speaking, Duke Ling looked up at birds flying overhead and let his attention leave Confucius. That was not verification but appraisal.

Those who truly doubted him walked away. The madman of Chu sang while passing his carriage. Confucius got down and tried to speak with him, but the man quickened his pace. The hermits working the fields mocked him and then refused further conversation.

Confucius met either courtesy without testing or ridicule that tested and left. No one said, “I do not believe you, but I will stay and see.”

The previous essay argued that another person's agreement carries no information, while disagreement may carry something.

Courtesy is agreement and carries nothing. Ridicule is disagreement, but it leaves as soon as it has spoken.

Only disagreement that remains is Penelope's kind.

Confucius therefore lacked neither supporters nor critics. He lacked someone who doubted him and did not leave.

There is another layer that fits the bed exactly.

Penelope can test Odysseus because she possesses something: the bed. He made it, it is rooted in the ground, and it cannot be moved.

Verification requires a shared object that cannot be forged.

During those fourteen years, Confucius's teaching had not yet taken that form. It existed in his words, his actions, and his exchanges with his disciples. No object stood rooted in the ground, immovable and available for comparison.

The second thing is the end itself.

Odysseus travels for ten years for a room, a bed, a particular woman, and a particular son.

We have spoken repeatedly of ends that require another person to receive them and things that cannot be completed unilaterally. In doing so, we may begin to imagine that a completable end is somehow inferior, that a road with an end cannot be deep.

That is wrong.

What Odysseus seeks is what most people are actually traveling toward. It is not low.

A person willing to exchange ten years for a return to several particular people needs no elevation into something grander.

6. Homer refuses to stop at arrival

The Odyssey itself adds another element.

In the underworld, the prophet Tiresias gives Odysseus a prophecy. He will return home and kill the suitors. But that will not be the end.

Afterward, he must carry an oar inland until he meets people who have never seen the sea. They know neither ships nor salt and mistake the oar on his shoulder for a winnowing shovel. There he must plant the oar in the ground, sacrifice to the sea god, and only then return home.

Homer therefore writes explicitly that the return to Ithaca is not the end.

A man who travels for ten years only to reach home must set out again after arriving, this time toward a place whose people have never even heard of the sea.

Even the epic of homecoming refuses to make arrival the ending.

7. He never found Penelope, so he made the bed

Return to Confucius.

At sixty-eight, he returned to Lu. That was not arrival but a change of direction.

He did not reenter politics. He organized classical texts and taught students. Tradition attributes to his later years a sentence: “It has been so long since I dreamed of the Duke of Zhou.”

The sentence probably means that he knew the event he desired would not occur in his lifetime.

Once he knew, he neither abandoned the work nor continued knocking on rulers' doors.

He arranged his teaching into a form that could remain.

Now the earlier line can be completed.

Verification requires a shared, immovable object. That was precisely what Confucius lacked during his fourteen years, so no one could truly test him. They could only honor or ridicule him.

What did he do after sixty-eight?

He never found the person willing to test him, so he made the bed himself.

What he organized became something rooted in the ground, immovable, and available to any later person for comparison.

In the essay on loneliness, I mentioned Craftsman Shi. After the man of Ying died, Lord Yuan of Song asked Shi to repeat his feat. Shi said he could still wield the ax, but the person who served as his material had long been dead. Zhuangzi said this at Huizi's grave, and then continued writing.

Confucius performed the same act more concretely. Unable to find someone who could test him, he left behind something that could be tested and handed it to a person not yet born.

In closing

Return to my friend's question.

Was Confucius's journey among the states a Chinese Odyssey?

No. What does not resemble the Odyssey is what matters.

The end of Odysseus's road waited for him throughout. The road could therefore be completed, and he completed it.

The end of Confucius's road never appeared. It required another person to receive it, and he did not meet that person in his lifetime.

Another difference may be even more worth remembering.

Odysseus had a Penelope, someone who waited twenty years and, at the moment of his return, doubted him before recognizing him.

Confucius had no one like that. He met courtesy without verification and ridicule that verified and left. No one doubted him and stayed.

In the end, he made the object of verification himself and left it there.

The endings are asymmetrical.

Odysseus returns home and the story ends, though Homer specifically adds that he must depart once more, as if he knew a completed journey leaves no further story.

What Confucius left has been accepted, overturned, reinterpreted, and used as a weapon against other things for more than two thousand years.

All of that is verification.

He did find the people he was waiting for, just not while he was alive.

To ask which man is greater is to ask a poor question.

The home Odysseus sought was real, and people were waiting for him. People were also waiting in Confucius's home during his fourteen years abroad, and he did not return.

This is not a hierarchy. Two different things were placed in the position of an end.

One lay within the traveler's own power and existed upon arrival. The other lay in another person's hands and might remain absent for an entire life.

And if the second kind of end is finally realized, the traveler did not realize it alone.

14. How Should We Face “If Only I Had”?

A friend asked what to do when we cannot stop thinking, “If only I had done this or that.”

There is plenty of advice on the subject, and most of it does not work. Some tells us not to look back. Some says everything happens for the best. Between them lies the instruction to make peace with oneself.

Perhaps these sayings fail not because they are bad, but because they treat two different things as one.

Regret has two entirely different structures.

One is: I did not know then; I know now.

The other is: I knew then and did it anyway.

They must be discussed separately. What is true of the first sounds evasive when applied to the second. What is true of the second is too heavy for the first.

1. When we did not know

Begin with the first kind.

It has a strange feature that becomes obvious once stated:

We can now see the mistake precisely because we took that step.

The vision capable of recognizing the problem is not something we always possessed and simply forgot to use. That stretch of road gave it to us.

Only after changing jobs do we discover what matters to us. Only after trusting the wrong person do we recognize a certain tone of voice. Only after ruining something do we understand which step cannot be omitted.

All these forms of judgment are products of the experience itself.

So when we ask, “How could I have failed to see it?” the problem was not that our former self lacked intelligence. We did not yet possess the glasses. The experience itself ground their lenses.

Kierkegaard wrote a sentence in his journals that has since been quoted countless times: life can only be understood backward, but it must be lived forward.

The sentence is usually treated as a reflection. It actually identifies a hard problem of sequence. Understanding and living move in opposite directions, and the directions cannot be reversed.

The first kind of regret judges the person we were with the judgment we possess now.

That earlier person did not have it. The problem was not a refusal to use it. It did not exist.

This is not consolation. Consolation would say that we were not wrong. The judgment that the choice was wrong may stand; the time at which we expect the judgment to have been available does not.

2. Of the two roads we compare, only one is real

The first kind of regret contains another, subtler problem in the way it compares alternatives.

The words “if only” are followed by another road. That road is being paved now.

When we pave an imaginary road, we include only what we can currently imagine.

It contains no accidents, no turns that were impossible to foresee, and none of the troubles that emerge only halfway through. It is flat and unobstructed because it is a model, not a span of lived time.

The road we actually traveled contains every pothole.

We therefore compare a real road containing all its accidents with an imagined road containing none.

This does not mean the other road would necessarily have been worse. It may truly have been better. We do not know and never will.

It means only that we do not possess that road for comparison. We possess something made today in the shape of today's wishes.

3. When we knew at the time

The second kind of regret is entirely different and should not be dissolved.

We knew it was wrong and acted anyway. We wanted convenience, feared trouble, sought a small advantage, or simply refused responsibility in that moment.

This regret is valid. It does not confuse the sequence of time or reveal insufficient judgment.

None of the preceding explanations applies, and none should. To say “I didn't know then” about something one plainly knew is to excuse oneself.

But one thing is worth saying, and it is not consolation.

The fact that the act still hurts tells us something.

A person suffers from having betrayed something only because that thing remains within them. If it were gone, the act would feel normal, or would not be remembered at all.

The pain is therefore evidence: we still recognize the line we failed to hold.

This does not make the act unimportant. It matters. It means only that the place still hurting is exactly what we have not lost.

4. If another person was hurt

A qualification must be inserted here, or the previous section will slide in a bad direction.

If the second kind of regret concerns another person, regret should not be directed primarily at our own feelings.

To chew repeatedly on one's guilt is still an activity centered on oneself. It may be sincere, but it points inward while the person affected remains outside.

What can be done is often plain: speak clearly, make restitution, or simply let the other person know that we understand. Sometimes nothing can be done. That too is real, and then we carry it.

In any case, the central question should not be “How can I feel better?”

5. The force of regret exists now

One feature deserves attention.

The content of regret concerns the past, but all its force exists in the present.

Someone says, “I regret not learning that subject.” The sentence sounds as though it concerns the past. Its only living part is that the person wants to learn now.

Someone says, “I regret not treating that person better.” The living part is that they now know what such care means and are capable of giving it.

The past cannot move. The force can land only in the present because the present is where it already resides.

This does not mean converting regret into motivational fuel. That account is too quick and often false. It identifies only the location of the force. It is not in the place to which we cannot return, but in the place where we stand.

6. Nietzsche's question

Nietzsche wrote a passage many people treat as a metaphysical claim. It is really a question.

Suppose, he says, that late one night a demon slipped into your loneliest loneliness and told you: this life you now live must be lived again, and lived countless times. Nothing will be new. Every pain, every joy, every thought and sigh will return in the same sequence.

Then Nietzsche asks whether you would throw yourself down and gnash your teeth, or say that you had never heard anything more divine.

The question does not ask whether our lives are good. It asks what relation we bear to the road we have traveled.

It also contains a rarely noticed feature: we cannot pass the test by changing the past. We can pass only by changing the present, because the life to be repeated includes every part beginning now.

7. A necessary qualification

Kierkegaard's sentence has a second half that is often omitted.

People remember that life can only be understood backward, he says, but forget that it must be lived forward. He also makes a finer point: no moment of life can ever be completely understood because time never stops to let us see the whole.

This sentence cuts into the argument of this essay and must therefore be included.

The vision with which we judge the past today will itself be judged later.

Looking back ten years from now, we may discover that today's judgment was also crude and omitted things we could not yet see.

This does not mean we should avoid judgment. It means we should not treat today's judgment as a final verdict.

That changes the weight of both kinds of regret. In the first, the “mistake” we identify today may not be the final conclusion. In the second, neither may today's condemnation of ourselves.

In closing

Return to my friend's question.

What should we do when we keep thinking “if only”?

First distinguish which kind of regret it is.

If we did not know then, the judgment may be valid while the timing we assign to it is not. We can see the mistake because the event gave us the eyes with which to see it. That vision was a product of the road, not a condition the road should already have possessed.

And the road used for comparison was paved today. It contains no accidents. It is smooth because no one has traveled it.

If we knew at the time, we should not dissolve the regret. It is real. The fact that it continues to hurt means we still recognize the line we crossed. If another person was involved, our attention should move toward them rather than inward toward our own feelings.

The words “if only” are always followed by a road we have made ourselves.

The place where we stand now was reached by another road, one that was actually traveled.

15. Why Do We Keep Changing the Measure of Our Worth?

A friend made an interesting observation.

When people are young, they feel powerless before physical strength. In middle age, they feel powerless before wealth. In old age, they feel powerless before wisdom.

The observation is perceptive, but it rests on a false premise. Once that premise is removed, the matter becomes clearer.

The premise is that the three measures follow age.

1. A person who does not grow may still envy wealth at eighty

Begin with what does not hold.

Not everyone reaches old age and begins to admire wisdom.

Many people compare wealth at sixty and continue at eighty. They compare their children's incomes, the locations of their homes, the differences in old classmates' pensions, and where each person's children work.

They never reached the third measure because it does not arrive automatically with age.

The three measures are therefore not three ages, but three stages. They often accompany age because time takes things away and forces us to change measures. But losing one measure does not mean replacing it with one of our own.

Many people are merely forced from the first to the second, where they remain for life.

2. Powerlessness measures not what is missing, but how many measures remain

Why do these things produce a sense of powerlessness?

The usual explanation is that we care most about what we lack. That is incomplete because it does not explain why the measure changes.

I think the real mechanism is this: at each stage, only one measure remains in a person's hand, and losing by the only measure means losing everything.

At the beginning, a person has no money, position, or experience. Physical strength is the only thing available for comparison. It matters not because strength is inherently so important, but because the other measures are not yet available.

At the next stage, physical strength has begun to decline while experience has not yet become what we would call wisdom. Money is the clearest, most easily ranked measure, and the one that most visibly displays differences.

Powerlessness therefore reveals not “what do I lack?” but “how many measures do I currently have?”

The fewer the measures, the more total each defeat becomes.

3. Their rotation proves that none is fundamental

The very fact that the three measures rotate tells us something else.

Anything replaced by the next measure is not fundamental.

If strength were fundamental, people would not later turn to money. If money were fundamental, no one would turn from it to something else. The rotation proves that the measures are interchangeable.

They can replace one another because they share a quality: each can be compared and ranked, showing who has more and who has less.

What powerlessness actually measures is therefore none of the three things. It measures the fact that I am standing on a scale and someone is above me.

What the scale measures is secondary.

4. The third measure is the most dangerous, and dangerous in a different way

If a person truly reaches the third stage, the powerlessness there differs from the first two.

In the first two, they know they are comparing external things. Strength belongs to the body; money is a number. Neither is plainly identical with the self.

Wisdom feels like the self.

In the final comparison, a person believes they are comparing what kind of person they are. To lose by the first two measures is to lose something. To lose by the third feels like losing oneself.

Another fact intensifies the pain.

Time takes away the first two, so defeat can be attributed to time. No sixty-year-old considers it a moral failure to be physically weaker than someone of twenty.

Wisdom is the only one of the three that time is supposed to give rather than take.

Losing to another person in wisdom therefore cannot be attributed to age. One cannot say, “I have grown old.”

5. Those who remain at the second stage rarely feel that they have stopped growing

Something unflattering but important must be said.

A person who remains at the measure of money rarely believes they have stopped. They are more likely to think they have seen through everything. Talk of cultivation or spiritual attainment is merely a beautiful consolation invented by people who cannot succeed.

The view is coherent at that level.

On the scale of money, wisdom produces no reading. It yields no number, cannot be ranked, and cannot be exchanged for anything concrete.

Whatever a measure cannot register does not exist from the measure's point of view.

The essay on the narrow gate made a similar point. The people on the broad road are not failing to look closely. The other thing does not exist within their coordinates.

That is why outside words have difficulty moving this stage. Anything said must first pass through its measure, and the measure cannot register the words.

6. Earlier measures grow lighter as one moves upward, though they do not disappear

An easy formulation must be corrected here.

In the essay on the fear of death, I argued that a person who travels farther loses none of the earlier fears but simply acquires another.

Powerlessness works somewhat differently.

The earlier forms remain, but they genuinely grow lighter. The reason appeared in the second section.

The intensity of powerlessness lies not in the measure itself but in the proportion of the whole self it occupies.

At twenty, losing a comparison of strength may keep someone awake all night. At fifty, they will probably laugh and move on.

This is not because the person no longer values bodily health. The difference is that strength now occupies one small part rather than the whole.

The measure and its reading remain. What changes is the share of the self represented by that reading.

This is why fear and powerlessness differ at this point.

Fear has an object. It concerns the loss of a particular thing. Possessing additional things does not make us less afraid of losing the original one.

Powerlessness is relative. It measures one's position on a scale. The more scales we possess, the less any one position represents the whole.

But we should not conclude that moving upward makes life more comfortable.

The new measure is the most painful. Its reading cannot be attributed to time, and it feels like the self.

The precise formulation is therefore: the old burdens become lighter, while the new one is heavier. Whether their sum is greater or smaller cannot be said and will differ by person.

7. Confucius and Aristotle say the same thing, and it contradicts the opening observation

A passage in the Analects closely resembles this question.

Confucius says that the exemplary person guards against three things: in youth, when the vital energies are not settled, against sensual desire; in maturity, when the energies are strong, against combativeness; in old age, when the energies have declined, against acquisitiveness.

The final word is easily mistaken for the acquisition of wisdom or attainment. Traditional commentaries agree, however, that it means material gain.

Huang Kan's commentary is blunt: “Gain means greed for gain. The old are prone to greed and must therefore guard against it.” He gives a reason. Youth resembles spring and summer, whose active force gives things outward. Old age resembles autumn and winter, whose receptive force gathers inward. The old therefore tend to accumulate and become greedy.

The Huainanzi is even shorter: human nature is reckless in youth, violent in maturity, and fond of profit in old age.

Fond of profit in old age. The object is wealth, not wisdom.

Here Confucius and Aristotle meet.

Describing old people in the Rhetoric, Aristotle also says they love money because they have learned how difficult it is to acquire and how easy it is to lose.

Separated by thousands of kilometers, both place greed in old age.

That seems to contradict my friend's observation, which associates old age with envy of wisdom.

Who is right?

Both, because they are describing different people.

Confucius and Aristotle describe the old person who did not move upward, the person from the first section who still compares children's incomes and pensions at eighty. As age leaves fewer things within their grasp, they grip what remains more tightly.

My friend's old person is the one who moved upward.

The two do not conflict because age was never what separated them.

Another word deserves notice: “guard.”

Confucius does not say, “This is what naturally occurs in old age.” He says, “This is what one must guard against.”

If age automatically produced a good state, no warning would be needed.

The three stages are not three ascending steps but three warnings against default directions, places toward which we slide without vigilance.

That agrees with the central argument of this essay: upward movement is not automatic.

We can extend the warning one step further, though this is an inference Confucius himself did not make.

At the stage of wisdom, the corresponding form of “gain” would be the desire to treat wisdom as something one can possess, count, and compare.

Finally, the three warnings are still not a timetable for other people. In the essay on meaning, I noted that in Confucius's account of setting his heart on learning at fifteen and standing firm at thirty, the subject throughout is “I.” The same holds here. Later readers turn it into an age chart, mistaking a personal account for an external standard.

8. Who gave us the measures?

We can now ask a deeper question.

Why do people hold so few measures?

Not because of age. The measures were given by other people.

Strength can be compared because a ready-made method of comparison exists. Wealth can be compared because its method is already available and precise to the last unit. Neither measure was cut by oneself. Both can be received and used immediately, which makes them convenient.

Convenience has a price. A ready-made measure can register only what it was designed to measure.

If every measure in a person's hand was received from elsewhere, the places where they feel powerless will be few, and changing from one to another will never escape the same category.

By contrast, something a person has cut for themselves usually cannot be compared with anyone else's, because it belongs to no ready-made scale.

What cannot be compared produces no ranking and therefore no powerlessness.

This does not make the person superior. It means only that they possess something that does not participate in ranking. Once even one such thing exists, the condition of having “only one measure” disappears.

In closing

Return to my friend's observation.

In youth we envy strength, in middle age wealth, and in old age wisdom.

The three often appear in this order, but they are stages rather than ages. Someone who never arrives may remain at the second measure at eighty.

The observation is useful not as a chart against which to compare one's age, but as a prompt for a question answerable now: By which measure was the thing that makes me feel powerless assessed, and who gave me that measure?

If the measure was received, the powerlessness does not really say that I am inadequate. It says that I am using someone else's scale.

If I cut the measure myself, it will probably produce no such reading, because no one else appears upon it.

As for the final stage, Confucius's warning still applies, though it must be moved.

He warns the old against greed for wealth. If someone truly reaches the measure of wisdom, there is a corresponding danger: treating wisdom as something that can be owned, counted, and compared.

The moment we want to possess wisdom, it becomes another comparative measure.

At that moment, we have fallen from the third stage back into the second.

16. Should We Forgive?

A friend asked: a long time has passed, and the person is still here. Should I forgive them or not?

Let me begin with what this essay will not provide. It will not tell anyone whether they should forgive.

This is not evasion. Any account that decides the matter for another person has already gone wrong. The first section explains why.

1. Three ready-made answers, none of which holds

Three familiar accounts of forgiveness cover three directions. Each contains a problem.

The first says that forgiveness is magnanimity, the generosity a person ought to possess.

This treats forgiveness as something one person can dispense. When someone says, “I forgive you,” they occupy the position of granting a pardon.

The position is difficult. It turns a matter between two people into one person's judgment upon another.

The second account has become especially popular in recent years: forgiveness is not for the other person but for you. It means setting yourself free.

This is compassionate, and it addresses a real problem. Resentment is exhausting.

But it also does something that may sound harsh when stated: it reduces the other person from a person to a stone in my mind.

They are no longer someone with thoughts of their own, capable of change and further action. They are a variable in my emotional state, and I release them so that I may feel lighter.

An earlier essay said that to treat a person as a means is to value them according to their usefulness to me.

“Let go for your own sake” handles the person according to their effect on my feelings.

The third direction refuses forgiveness: they did that thing, which proves what kind of person they are. They will be that way for life.

This too has a problem, and it is the easiest to overlook.

The structure of refusing forgiveness fixes a person to an event. I do not look at who they are now or ask what they later became.

In the essay on love, I said that a person is whole but not complete. They are still growing.

To treat someone still growing as a finished object incapable of change is itself to stop treating them as a person.

All three answers touch something real, and all three fail somewhere. Perhaps the problem lies not in the answers but in the word.

2. “Forgiveness” confuses three different acts

Look closely, and people may mean three entirely different things when they speak of forgiveness.

The first is to stop demanding.

No longer demand compensation or an apology, no longer need the person to suffer, no longer conduct the trial repeatedly in one's mind.

One person can do this alone, and it concerns that person's own actions.

The second is to restore a relationship.

Treat the person again as someone with whom one can associate, speak, and sit at the same table.

One person cannot do this alone because a relationship belongs to two. The other person must be present and must have moved.

The third is to cancel the event.

Treat it as if it never happened, or as if it did not matter very much.

No one can do this.

3. Why no one can perform the third act

Because the event happened.

It remains in the world and in the person who was hurt. Nothing anyone says can move it from “having happened” back to “not having happened.”

This sounds obvious, but it has a concrete consequence.

Any forgiveness whose goal is to act as though the event never occurred is false.

Many people become stuck here. They believe forgiveness requires them to say the event did not matter. It plainly did, so they cannot say it, and they conclude that they are not good enough.

The problem is not their goodness. They are being asked to do something impossible.

Of the two possible acts, the first concerns how one acts, and the second requires another person. Neither requires anyone to say that the event did not matter.

4. Forgiveness is not decided; it grows

There is another mismatch, of the same kind discussed in the essay on regret.

That essay argued that expecting one's former self to see what only the present self can see is an error of sequence.

Here the mismatch is this: forgiveness is often treated as a decision. “I decide to forgive you.”

If the person's heart still hurts, the sentence is a decision, not something that has already occurred.

When forgiveness truly occurs, it is more like suddenly noticing one day that the event no longer sounds so loud when remembered. It grows; it is not decided.

“I decide to forgive you” is often false not because the speaker is insincere, but because it speaks of something organic as if it were a switch.

This has another consequence. If forgiveness grows, no one is obliged to make it grow by a particular deadline.

No one is obliged to make it grow at all.

5. Confucius refuses cheap forgiveness

A brief passage in the Analects refuses it decisively.

Someone asks Confucius what he thinks of repaying resentment with virtue.

Confucius asks in return: with what, then, will you repay virtue?

His answer is: repay resentment with uprightness, and virtue with virtue.

The crucial word is “uprightness.”

It does not mean revenge. Revenge would repay resentment with resentment, which Confucius does not say.

Nor does it mean forgiveness. That would be repaying resentment with virtue, the very suggestion he has rejected.

The closest sense is to treat the matter as it actually is.

Call the event what it is. Neither enlarge nor diminish it. Do not say it never happened because the other person is pitiable; do not say it was heavier than it was because we are in pain.

Confucius's question is the real blade. If I repay even those who harm me with virtue, what remains with which to repay those who treated me well?

Cheap forgiveness wrongs not only oneself. It deprives goodness of weight. If every action receives the same response, good and harmful acts become indistinguishable in their consequences.

6. Arendt: forgiveness as a remedy for irreversibility

In The Human Condition, Hannah Arendt places forgiveness somewhere very different.

Human action, she argues, has two fundamental difficulties. What has been done cannot be recalled, and what will happen cannot be known.

Human beings developed two responses: forgiveness addresses the first; promises address the second.

Without forgiveness to release us from the consequences of what we have done, she writes, our capacity to act would remain bound forever to a single deed. We would remain its victims, like an apprentice unable to break the spell they cast.

Notice who is bound.

Arendt means not only the person harmed, but also the person who acted.

If everything a person has done remains attached to them forever and can never be loosened, they can never begin anything new. The rest of their life becomes only an aftereffect of that one deed.

Forgiveness, for Arendt, therefore concerns the person rather than the act.

The act remains. It happened. What is released is the person, so that they may still do something else.

This distinction also reveals the cost of refusing forgiveness. To nail a person forever to what they have done is to declare that the rest of their life can only be its consequence.

Arendt also draws a line. Some acts, she believes, can neither be punished nor forgiven. They exceed the capacity of human affairs to deal with them.

That line must also be stated. An essay on forgiveness that implies everything should be forgiven becomes a form of pressure.

Some things need not be forgiven. The sentence requires no defense.

7. Asking forgiveness can itself use another person

There is another layer rarely discussed.

In Simon Wiesenthal's book The Sunflower: On the Possibilities and Limits of Forgiveness, he recounts being taken, while imprisoned in a concentration camp, to the bedside of a dying SS soldier. The man confessed what he had done and asked for forgiveness.

Wiesenthal said nothing and walked out.

He later wrote down the event and asked dozens of people what they would have done. The book gathers their conflicting answers.

The request contains a structure worth examining on its own.

The soldier wanted release. He did not seek the people he had harmed; they were dead. He sought a Jew. Any Jew would do.

In that moment, Wiesenthal was not addressed as himself. He was used as the representative of a category, with the function of allowing another person to die in peace.

The request itself therefore treated a person as a means.

The same structure appears more generally.

In the essay on “I only want what's best for you,” I wrote that the sentence transforms another person's refusal. “I disagree” becomes “you have betrayed my goodwill.”

A request for forgiveness can sometimes do the same.

A person apologizes sincerely and sincerely asks to be forgiven. The sincerity is real. But after the request, if the other person says no, that “no” is no longer merely a judgment. It becomes an act of cruelty.

The pressure changes direction. I harmed you, but now the question is whether you can be generous enough to me.

This does not mean we should not apologize. An apology and a request for forgiveness are different acts.

In an apology, I state clearly what I did and hand that statement over. In asking forgiveness, I also ask for something in return.

The first remains entirely on my side. The second requires the other person to give.

8. What remains?

Put these points together, and little can be said, but what remains is concrete.

A person can stop demanding. This requires no cooperation and does not require saying that the event did not matter.

A person cannot cancel the event. No one can.

Nor should anyone be required to complete forgiveness by a certain time. It is not a decision but something that grows.

Two things should not be done.

We should not nail a person forever to the event, because they are still growing, and to treat someone still growing as a finished thing is not to treat them as a person.

Nor should we cover our own pain. The event continues to sound within us. Pretending that it does not sacrifices ourselves to a respectable formulation.

The two prohibitions may seem to point in opposite directions. They are the same: do not turn the other person into a finished object, and do not erase one's own remainder.

Whether forgiveness eventually grows cannot be decided by these principles.

9. The final night of the Iliad

I want to end with a scene. It gives no answer, but it is the closest image I know of what genuine forgiveness might look like.

Near the end of the Iliad, Achilles kills Hector and drags his body around the city.

That night, Priam, the old king of Troy, enters the Greek camp alone and finds Achilles.

The first thing he does is kneel and kiss Achilles' hands. Homer adds a line: these are the hands that killed many of his sons.

Priam does not say, “I forgive you.” Nor does he demand an apology. He asks for his son's body.

Achilles looks at the old man and remembers his own father.

Then the two weep together. Priam weeps for his son. Achilles weeps for his dead friend and for the father he will never see again.

Afterward, Achilles returns the body and grants the Trojans a truce long enough to complete the funeral.

No one in the scene speaks the word “forgiveness.”

Hector remains dead. Achilles knows that his own life will be short. The event has not been canceled and cannot be.

Only one thing occurs that night: two enemies each see that the person across from them is also a person.

In closing

Should we forgive?

The question may be too large. The word contains three acts, one of which no one can perform.

We can stop demanding. That requires no one's permission and does not require declaring the event unimportant.

We can hope that one day the event will sound less loudly when remembered. Hope cannot produce this. We can only wait for it to grow.

We cannot act as if the event never happened.

And there are two things we should not do: nail the person forever to that day, or cover our own pain.

Forgiveness itself may resemble what occurred in the tent that night. No one gave an object to anyone else. Two people each saw that the one across from them was also a person.

That requires both people to be present.

If the other person is gone, or never intended to be present, it may never happen.

That is all right. No one owes anyone forgiveness.

17. Can a Well-Intentioned Lie Be Right?

A friend asked whether a well-intentioned lie is permissible. Is it right?

The question usually receives one of two ready-made answers. Neither is very useful.

The first says absolutely not. A lie is a lie, and a good motive does not make it true.

The second says to judge by the consequences. If everyone is better off afterward, the lie is permissible.

The first is so rigid that almost no one actually lives by it. The second has a more difficult problem: Who calculates the consequences? Usually it is the person who wants to lie, and that person can always calculate a favorable result.

Perhaps the question is not whether we should lie, but whether truth and falsehood are the only dimension along which the matter should be judged.

1. The criterion does not lie in truth or falsehood

Begin with a strange fact.

Some statements are entirely false and cause almost no harm. A doctor asks how you feel, and you say “fine” though you feel terrible. The statement is false, but it takes nothing from anyone.

Other statements are true in every word yet thoroughly manipulative.

Someone omits every fact unfavorable to themselves and states only the true ones, assembling them into a completely false picture without telling a single lie.

Truth and falsehood alone therefore fail to capture what matters.

What does matter?

I think it is this: after the statement has been made, can the other person still verify matters for themselves?

2. Three different acts

We should first separate three things that are often confused.

The first is silence.

No one is obliged to tell everything to everyone. Privacy is a right, not a deception.

The second is partial disclosure.

This is more complicated. It may be considerate or manipulative. The difference is whether the other person knows that further parts exist.

The third is manufacturing a belief.

This means placing something contrary to fact in another person's mind without their knowledge and leading them to make a decision on that basis.

Only the third is what we ordinarily call a lie, and only it has a particular consequence: the other person bears the cost of the decision, while I supplied the material on which the decision was based.

3. What does the statement take away?

The criterion can now be stated clearly.

The problem with a statement is not whether it corresponds to fact, but whether it deprives another person of the capacity to verify.

Seen this way, untrue statements can carry radically different weight.

A falsehood punctured by a single question leaves the other person able to ask, investigate, and discover. Though false, it has not canceled any capacity.

A complete, internally coherent false construction that closes every route of verification belongs to another order.

In an earlier essay on Evergrande, I argued that financial fraud does not merely put a lie in a report. It turns something that ought to be verifiable into something that appears open to verification but in fact cannot be checked.

That is the deeper wrong.

This criterion is therefore not more lenient than “never lie.” It is stricter.

It also captures manipulation that contains no false sentence: selecting only favorable truths, creating an environment in which no one can ask questions, or hiding crucial material where another person cannot find it.

Such actions are innocent on the dimension of truth and falsehood. They are not innocent on this one.

4. Plato's example is the most candid of its kind

The Republic contains a famous proposal usually called the noble lie.

Plato has the rulers tell all citizens a story. They are brothers and sisters born from the earth, but when the god made them, he mixed gold into some, silver into others, and bronze or iron into still others. What each person should do is therefore fixed by birth.

Plato says this serves the good of the city. He also acknowledges that the story is invented.

One feature of the design deserves attention: it is made impossible to verify.

Citizens have no way to investigate which metal was mixed into them. This is not a flaw in the design. It is exactly why the story can work.

The essay on “I only want what's best for you” described what that sentence does. It turns another person's refusal from “I disagree” into “you have betrayed my goodwill.”

The noble lie performs the same action at the scale of a city. It decides on behalf of everyone what they ought to believe, on the ground that the belief will be good for them, and removes their capacity to check.

Plato is candid. He knows what he is doing and gives his reasons.

That candor makes the example especially useful. It shows that lies of this kind rarely arise from malice. Almost always, they arise from some version of “I know what is better for you.”

5. Kant's blade is aimed directly at this essay

The argument cannot continue without confronting someone who specifically rejected this case-by-case approach.

Late in life, Kant wrote a short essay responding to a hypothetical case. If a murderer comes to your door and asks whether your friend is hiding inside, may you lie?

Kant's answer is no.

For more than two centuries, the answer has been presented as evidence of deontological madness. But his reason deserves attention because it does not concern consequences.

A lie, he says, damages not only the person deceived but something held in common by humanity: the possibility of taking speech as true.

He too is concerned with structure rather than consequences.

And his real concern is precisely the risk in this essay.

The moment you say, “It depends on the circumstances,” the person judging the circumstances is the person who wants to lie.

Someone who wants to lie can always invent a plausible circumstance.

The objection is substantial, and I will not evade it.

An earlier essay in this series offered one response while discussing “I only want what's best for you”: motives cannot be distinguished, but actions can.

Genuinely wanting the best for someone and needing them to obey while sincerely believing that you want the best feel identical to the speaker. “My motive is good” therefore cannot grant permission.

The criterion cannot be goodwill. It must be observable from outside: after I speak, can the other person still verify matters for themselves?

Lying to the murderer at the door deprives no one of the capacity to verify, because the murderer's current purpose is itself to cancel another person.

Lying to someone who trusts you and will arrange their life on the basis of your words is different.

This does not refute Kant. His warning remains in force and should hang over every discussion of this kind. Whenever “this time is permissible” sounds reasonable, first ask who is judging this time.

6. The minimum requirement

If the matter can be condensed into something usable, I would put it in two sentences.

Say what you did.

Say whom you did it for.

The requirement sounds so modest as barely to count. It does not require saying everything, being selfless, or even being kind.

It asks only that we create no false appearance in these two matters.

Why these two?

Because they are all the other person has with which to judge the relationship.

What you did determines their judgment of the facts. Whom you acted for determines their judgment of you.

To lie about either means that you issued the measure by which they understand the situation and altered its markings.

Everything else may be withheld or described vaguely. That is privacy, not the destruction of the measure.

As an aside, this requirement is low enough for almost everyone to meet, and almost no one meets it consistently.

7. One kind of untruth does not belong to this question at all

One final thing should be separated from the category of well-intentioned lies.

In his dialogues, Socrates often sees exactly where another person's account fails but does not say so. He asks questions.

Sometimes he even pretends not to understand and asks the other person to teach him.

Taken literally, this is untrue.

But it removes nothing. The opposite is true. It leaves the step of arrival to the other person.

An earlier essay compared Socrates to a midwife. She does not give birth for the laboring person. She eases the pain and judges whether what emerges is genuine or a phantom. She withholds the answer not because she wishes to conceal it.

If she states the answer directly, it becomes a sentence handed over rather than something grown by the other person.

Some apparent concealment is therefore structurally opposite to lying.

A lie inserts into another person something they cannot verify. Socratic withholding could give the answer directly but leaves a space so that the other person can take the step.

One inserts; the other leaves room.

So can a well-intentioned lie be right?

The question is framed in terms of truth and falsehood, but that is not the most important dimension.

What matters is whether the other person retains the capacity to verify after the statement is made.

A false statement punctured by one question may take nothing away. An account that leaves no route of inquiry may remove what matters most even when every word is true.

Goodwill is not a permit.

Only the person acting knows whether the intention is good, and that person is the least reliable witness in the matter.

We can observe only the action.

Say what you did. Say whom you did it for.

Everything else may remain unsaid.

18. Should We Help an Elderly Person Who Collapses?

The question is painful, and it stirs one of our deepest doubts. Put differently, it asks whether human beings should still follow the rules of being human.

Let us first establish what happened. At 12:50 p.m. on August 17, 2026, in Qidong, Hunan, a man who had retired a year earlier entered a neighborhood mahjong and card shop and sat on a chair by the door. He knew the proprietor, Xiao Aihua, and had visited almost every day for the previous two weeks. At 12:54:30, Xiao noticed something wrong. She approached and spoke, but he did not respond. She later said she panicked and could not even find where to dial on her phone, so she called for someone nearby to contact emergency services. At 12:55:01, she and two others moved him outside the door. The shop had no air conditioning, the day was hot, and she thought he had suffered heatstroke. A local saying in the Qidong dialect holds that someone with heatstroke must “see the sky,” meaning go outside for fresh air. The ambulance arrived at 1:01 and took him away at 1:03. He did not survive.

It happened that quickly. Less than a minute passed between Xiao's noticing something wrong and the emergency call. Six minutes passed between her moving him and the ambulance's arrival.

What she did during that minute required no instruction. When someone collapses, you go to them, ask what is wrong, call for help, summon an ambulance, and move them somewhere with air. These are the rules of being human. They have no statute, burden of proof, or termination clause. People once dealt with one another according to them. Everything that followed occurred after she did so.

From beginning to end, the shop's cameras recorded all thirteen minutes. The recording became the common starting point of every later disagreement. The family watched it and argued that the rescue had been mishandled and bore some responsibility for the death. They sought 100,000 yuan in compensation. The police watched the same recording and concluded that the man had died of a sudden illness, with no external injury or conflict, and ruled out a criminal case. The video, they said, showed Xiao actively rendering aid throughout, promptly seeking medical help, and acting properly and in good faith.

The camera did not end the disagreement. It moved the disagreement from what occurred to whether the action was right. It could answer the first question, not the second. Xiao moved the man outside at 12:55:01 because she believed he needed to “see the sky.” Someone watching later could call the same act dragging him outdoors. At the scene, an action has one reason. In retrospect, it can acquire many.

One sentence is worth carrying away: only someone who acts leaves an action available for review. Had Xiao remained still, the recording would contain no act open to reinterpretation. A mechanism that scrutinizes actions alone will reliably push people toward inaction. No one needs to be malicious. It requires only a cause of action for improper rescue.

The first formal mediation took place on August 18. The family asked for 100,000 yuan. Chen Xiaomao, a staff member of the local Comprehensive Governance Center, said the mediators explained the legal rules in person. Under China's Civil Code, a party seeking tort damages must prove fault. They also explained that mediation was voluntary and would end if either party objected. The file shows that Xiao insisted she had not been at fault but voluntarily offered 2,000 yuan in humanitarian assistance. The family reduced its demand to 50,000 but would go no lower, and mediation paused. Because the weather was hot, the mediators suggested that the family proceed with cremation that afternoon, which they did.

The second mediation occurred on August 19, with the parties in separate rooms. The family lowered its demand to 20,000 yuan. Remembering that the man was a neighbor, one of Xiao's relatives referred to local custom and similar cases and voluntarily offered 15,000. A report described the next step this way: considering the energy and time consumed by the dispute, and the possibility of secondary problems, the parties gradually reached a consensus at 19,000 yuan. Xiao's husband transferred the money on the spot.

The rules had therefore been explained clearly. The burden of proof was made clear. So was the right to end mediation at any time. Xiao herself first offered 2,000 yuan; her relative first offered 15,000 and introduced the calculation based on similar incidents. Another staff member, Liu Lingang, said, “We have no authority to force anyone to pay.” The body had been cremated the day before the agreement, so objectively there was no possibility of coercion by refusing burial. The parties disagreed about whether such threats had been made earlier. The dead man's wife denied it in her statement; Xiao's side said similar words had been used early on but acknowledged that nothing was actually done.

She took every step herself, and the result was that a person formally found free of fault paid 19,000 yuan.

Her son later explained why. The family had originally wanted to give only two or three thousand. They eventually accepted 19,000 because of Xiao's health. Since the incident, she had been unable to sleep.

This is the most important sentence in the case. The 19,000 yuan did not purchase responsibility, because there was none. It purchased an end to the matter. The price was determined not by right and wrong, but by who could no longer bear the pressure.

At this point, the problem is no longer that no one told her the rules.

Chen Zhong, the head of the Comprehensive Governance Center, put the matter cleanly. In a civil dispute, the person making a claim bears the burden of proof. Mediators lack authority to issue a binding determination. They can only tell both sides that no evidence currently establishes responsibility.

This is not a failure of duty. It defines the position. Mediation is not adjudication and cannot determine liability. It can only explain the rules and facilitate agreement. Explaining a rule and reaching a judgment are different acts. The rule can identify who bears the burden of proof. It cannot declare the matter over. Only agreement can end it, and agreement depends on how much each party can bear.

An official from the Qidong County Justice Bureau spoke candidly. Most local civil disputes are resolved through mediation each year. Mediation is a flexible method that can absorb social conflict and prevent cases from flooding the courts or escalating in place. He also acknowledged how difficult the balance is. Even experienced mediators struggle with the practical problem of protecting a person without fault from loss while not simply passing every conflict upward.

All of this is true. Mediation does absorb conflict and prevent escalation, and the difficulty is real. Precisely because every sentence is true, the structure of the problem becomes clear. A mechanism designed to make matters stop measures success by whether they have stopped, not by who was right. In explaining how 19,000 yuan was reached, the report referred to possible “secondary problems.” Preventing such problems genuinely belongs to the mechanism's purpose, and secondary problems do not care who is right.

What, then, did “voluntary” mean here? Procedurally, Xiao could stand up and say no at any time. The mediator told her so. But after she said no, the body still remained until it was cremated on the afternoon of the first mediation. The weather was hot, her shop faced the street, the dead man was a neighbor, and she had already stopped sleeping.

Voluntary meant that she did not have to sign. It did not mean she could make the matter end. A form of voluntariness that can terminate a procedure but not the situation is a right on paper and an empty sentence at the scene. To know whether someone has a choice, do not ask whether they can say no within the procedure. Ask whether the matter stops after they do.

On the afternoon of August 26, the man's family returned the 19,000 yuan to Xiao's son's account. The transfer note read, “Return of humanitarian compensation.” That evening, with officials witnessing, both sides signed a new statement. The man's son said that the family had gained a deeper understanding of the law during the previous days. A person without fault need not pay compensation, he said, and they now accepted that principle.

It was the same rule Chen Xiaomao had explained in person on August 18. Eight days separated the statements. No new evidence appeared. What changed was that the whole country was watching.

If a rule takes effect only under public scrutiny, then in ordinary times it relies on something other than itself.

Public attention is not a cure. The other side of the same case makes that clear. One of the dead man's relatives said that after signing the agreement, they thought the matter was over. Now the whole country knew, and the family had become villains online. They reported online rumors to the police and requested an investigation.

The August 19 agreement stated that the family would raise no further objection to the dispute or its resolution. Both sides signed it, and neither could use it to end the matter, because the objections never came from either family.

Money traveled in a full circle. A demand for 100,000 yuan became 19,000. A local rice company brought Xiao 20,000 in assistance. Entrepreneur Luo Yonghao and several friends raised another 100,000 for her, explaining that compensation alone was insufficient and people who do good deserve emotional support from society. On August 26, the family returned the 19,000. Xiao's family then donated all 120,000 yuan they had received to the local charity federation.

After the additions and subtractions, the number returned roughly to zero. Other things did not. Xiao suffered insomnia, went to the hospital for an intravenous infusion after feeling ill, closed the shop for several days, and began considering whether to reopen. None of these fit within the 19,000 yuan or the 120,000.

Her decision to give away the 120,000 shows that the money could not set a price for the event. Compensation could not balance it; reward could not either. Every conversion felt wrong because the event never belonged to the category of convertible things.

Xiao said that if she encountered another person who needed help, she would still step forward.

That sentence appears in no mediation agreement, witness statement, transfer note, or trending topic. It is the only thing in the entire case that no procedure processed.

And it was the source of the decision made at 12:54:30. Xiao made that decision alone. No one authorized her during that minute, no one assumed the burden of proof on her behalf, and no document was present. Everything that followed, the agreements, witness statement, assistance payments, donations, returned payment, and online attention, dealt only with what came after that minute.

An institution can restore the losses of a person without fault. It can restore more than the loss or, as here, return the compensation afterward. It cannot make that decision for her. Only she can make it, and she can do so only while remaining someone who decides for herself.

Return to the question: Should we help? Should human beings still follow the rules of being human? The question tests not the human heart, but institutions. When institutions release human agency, following human rules is not difficult. When institutions consume agency, how difficult does it become to hold to them?

Part II: Knowledge and Consciousness

Knowing something is not merely storing the correct answer in the mind. Artificial intelligence, mathematics, philosophy, and the problem of consciousness repeatedly remind us that understanding also concerns who judges, why belief is warranted, and how knowledge changes action. Technology can perform increasingly complex tasks without automatically answering questions of happiness, meaning, or agency. When disputes ask whether a machine can understand, the essays in this part turn the question back upon human beings. What exactly is the understanding we cherish, and are we willing to accept its limits?

19. Why Is AI Producing So Many Mathematical Breakthroughs?

There had indeed been a great deal of mathematical news, and even mathematical drama, by September 2026. The explanation was not the sudden appearance of a universal mathematical algorithm. Rather, reasoning ability, sustained work, tool use, and proof checking had begun to reach research-supporting levels at the same time.

Previously, failure in any one link could destroy the entire task. Once the links began to hold together, results appeared to arrive in clusters.

Consider several recent breakthroughs. On May 20, an internal OpenAI model constructed a new planar point configuration that overturned an important conjecture in the unit-distance problem. External mathematicians checked the proof. It was a new mathematical discovery, though it did not settle every aspect of the unit-distance problem.

On August 10, Anthropic's Claude improved the lower bound on the proportion of zeros of the Riemann zeta function lying on the critical line to about 67.2 percent. By September 2, mathematicians had published a shorter proof. This was an important partial advance, not a proof of the Riemann hypothesis.

On September 4, Anthropic announced a fully machine-checked proof of Fermat's Last Theorem by Claude, reporting that the main work took eleven days. This was a breakthrough in proof formalization, not the first discovery that Fermat's Last Theorem is true.

On September 8, OpenAI announced a Navier-Stokes proof produced by an internal system. It claimed to resolve the relevant version of the official problem through a construction of finite-time singularities with smooth forcing and supplied a Lean formalization.

The dense sequence of headlines therefore included genuine discoveries, advances in verification technology, and results only just entering public scrutiny.

Models now know more and can pursue a line of reasoning farther. A crucial beginning of this shift was training directed at the reasoning process. OpenAI's technical report on o1 said that reinforcement learning improved not only answer accuracy but also the ability to detect errors, divide difficult steps, and change routes when one failed.

Research on DeepSeek-R1 likewise showed reinforcement learning improving checking, reflection, and strategic adjustment in verifiable tasks such as mathematics. This is fundamentally different from making a proof look like a proof. Research does not require dozens of elegant paragraphs of continuous derivation. It requires discovering a problem in step seven, withdrawing that step while preserving the first six, and finding another route.

Why did the change appear sudden? Consider a simplified workflow: propose an idea, calculate examples, find a flaw, repair a lemma, complete the argument. If any link is wholly unreliable, final output may be zero. As every link gradually improves, the success rate of the complete task can jump sharply at a certain point.

The mathematical systems in the headlines were not ordinary chat windows or workloads available through released models. Anthropic disclosed that the research version of Claude that produced the zeta-function result first tried 650 unsuccessful ideas. It then coordinated about sixty subagents for another day and a half. They ran code, performed numerical checks, reviewed one another's arguments, and searched the literature to determine whether the result had precedent.

OpenAI's Navier-Stokes work was larger still. The group that found the result involved about ten thousand concurrent agents and took roughly eighty-eight hours from launch to a candidate solution, followed by separate formalization and verification. These figures came from company process reports. Ordinary users could not deploy the same volume of work.

The better question is therefore not whether a chatbot is smarter than a mathematician. It is what research becomes possible when a sufficiently capable model can organize many parallel attempts, preserve intermediate results, and correct itself over long periods. The number of agents cannot be converted directly into a number of human mathematicians. Repeated exploration, shared blind spots, and coordination failures may render enormous amounts of computation useless.

Anthropic reported that early agents in its Fermat project lost control of project state and failed to collaborate. Progress came only after the team introduced a platform for managing dependencies among theorems. Organizing work effectively was itself part of the advance.

Mathematics is especially well suited to separating bold conjecture from strict acceptance. A language model can propose many ideas, but its own assurance that it has checked them is plainly insufficient. Proof assistants such as Lean offer another form of checking. A proposition and its derivation are written precisely, and the program checks the proof against stated definitions, axioms, and rules of inference. Lean is not another chatbot casting a vote in favor.

This permits a division of labor. Models explore, computational tools test, proof systems check formal derivations, and mathematicians examine the formulation of the problem, its relation to the literature, and its mathematical meaning.

Boundaries remain. Numerical tests cannot replace a general proof. Passing formal verification means only that the encoded proposition follows from the premises used. Mathematicians must still determine whether the encoded proposition accurately represents the original research problem and whether the assumptions smuggle in what was supposed to be proved. But once acceptance becomes reliable enough, exploration can tolerate more failure. Large-scale attempts become much more valuable. Each need not be right; the few genuinely correct routes must be identified and preserved.

The systems are connecting mathematical tools that were previously dispersed. This is closer to the substance of the recent results than mere speed of calculation. OpenAI's unit-distance result applied sophisticated tools from algebraic number theory to a planar geometry problem with an elementary statement. It did not simply scale up a traditional calculation. It found an unexpected route across fields. Claude's zeta-function result likewise depended heavily on existing mathematics, which it recombined into a stronger argument. Mathematicians then rewrote the result as a shorter and more transparent proof. A new research sequence had emerged: AI obtains a result, and human researchers deepen its understanding and simplify it.

Using existing human knowledge is therefore not sufficient reason to deny originality. More meaningful questions are whether the connection was previously unknown, whether the reasoning holds, and whether a genuinely new conclusion follows.

Why did both companies improve at the same time, and why did the number of results suddenly rise? I am inclined to see a shared technological stage rather than a unique mathematical secret possessed by one company. Reinforcement learning for reasoning, greater inference-time computation, code and literature tools, and the organization of long-running tasks were not confined to one laboratory.

Mathematicians were also embracing AI. The volume of experimentation by leading researchers continued to grow. A preprint posted on August 25 examined 330,000 mathematics-related arXiv submissions from March through August 20, 2026. It reported that the share explicitly disclosing a substantive mathematical contribution by AI rose from 1.39 percent in March to 14.09 percent by August 20. The classification relied on the papers' own disclosures and did not verify every result individually, but it suggested that actual adoption, not only capability, was increasing. These were real mathematicians using AI to intensify their work, not work AI could complete by itself.

Logically, even a partial improvement in the success rate of a single research attempt can produce many more results when the number and duration of attempts increase and verification becomes more effective. When the public simultaneously sees several different kinds of advance, the effect is a cascade of breakthroughs.

20. Is Human Consciousness Truly Beyond AI Simulation?

This is not merely Roger Penrose's personal view. He represents a philosophical school that believes human understanding contains a capacity no algorithm can exhaust. We do more than follow rules. We understand their meaning, judge whether they are reliable, and in some cases move beyond them. His claim is therefore not that present AI lacks sufficient computing power. It is that computation alone would still lack a crucial ingredient even if AI improved without limit.

The issue is not that there are mathematical theorems AI can never understand. It is how Penrose moves from limits on computation to the conclusion that human beings must be able to exceed computation.

We ordinarily distinguish a system's ability to perform a task from its possession of subjective experience. Describing pain accurately and feeling pain are not the same. Penrose's claim goes further. He explicitly distinguishes two positions: one says a computer might fully simulate human behavior without inner experience; the other says a computer could not fully simulate all the capacities produced by consciousness. He endorses the stronger second position.

To understand him as saying that AI can behave exactly like a human while remaining empty within would therefore weaken his view. He believes genuine understanding produces differences in capacity and behavior that pure computation cannot fully reproduce. That also explains his particular interest in mathematical understanding. It appears easier to discuss, compare, and test than the question of what redness feels like.

His main basis is Gödel's incompleteness theorem. The mathematically established part is this: for a formal system F that is sufficiently powerful, has mechanically enumerable rules, and is consistent, one can construct an arithmetic proposition G that is true but not provable within the system. Roughly speaking, the proposition says, “I cannot be proved in F.”

Here “cannot be proved” does not mean that a computer is too slow or that no one has yet found a clever method. It means that under the given rules, no such proof exists.

Penrose then asks a further question. We can understand this argument and, if we accept the system's reliability, see the truth of a proposition the system cannot prove. Does human mathematical understanding therefore always have a way to exceed any given algorithmic description? Penrose says yes. Any computational program that attempts to encompass mathematical understanding completely will encounter such a boundary. Understanding itself must therefore contain a noncomputational element.

As an analogy, a machine works within a rulebook. A human being can not only use the book but pick it up as a whole and discuss its limits. Penrose believes that this latter capacity cannot ultimately be contained in another, larger mechanically executable rulebook.

The idea is compelling. But this is precisely where the problem arises: have we proved that human beings can do this, or only that they can under additional assumptions?

Many people disagree with Penrose. The strongest objections focus on the phrase “human beings can.” The first problem is how a person knows that the system is reliable. David Chalmers puts the criticism directly. What a person may understand is only the conditional statement: if F is consistent, then G is true. A sufficiently strong formal system can also prove that conditional. The real difficulty is knowing whether the antecedent holds. For a system complex enough to describe the totality of human mathematical thought, what guarantees that we can identify it, understand it, and know that it is consistent?

Knowing that a given system has limits does not prove that we are free of the limits of every computational system. A person may move beyond one set of rules by using another, more powerful set, without having left computation itself.

The second problem is whether human beings can revise rules through reflection while machines cannot in principle. We should not caricature Penrose as knowing only old computers incapable of learning. In his replies, he explicitly considers artificial neural networks, genetic algorithms, random input, simulated environments, and related approaches. His point is that so long as the entire process of learning and revision remains computational, it still falls within the scope of his argument. But the reverse is also true. We cannot infer noncomputability simply from a human being's ability to alter their own rules. A program capable of revising its strategies remains a program. What must be proved is that the revision exceeds computation, not merely that it differs from rote execution.

The third problem concerns ideal ability. Is Penrose's idealization valid? He does not ignore human fallibility. He explicitly says that he is interested in the capacity, in principle, for correct understanding and error correction in mathematics, not every statement made by every mathematician each day. This prevents the argument from being defeated merely by pointing out that people make mistakes. But it creates another question. Is this ideal capacity a reasonable description of actual human beings, or has the premise already included the capacity to transcend systems correctly and without end?

Even among sympathizers there is disagreement. The logician Solomon Feferman was deeply skeptical that computational models could fully explain the mind, yet still believed that Penrose's Gödelian argument failed to support the skepticism.

Why, then, did Penrose later turn to quantum mechanics and microtubules in the brain? Because he still had to answer another question: if understanding is not computation, what in the brain produces it?

His answer is not a soul beyond physics, but the possibility that physics remains incomplete. In particular, he suspects that the transition from quantum states to the definite results we observe involves a noncomputational physical process not adequately described by current theory. He has also said explicitly that this does not make the brain a quantum computer in the ordinary sense. Quantum computation does not automatically provide the noncomputability his account requires.

Together with the anesthesiologist Stuart Hameroff, Penrose proposed the orchestrated objective reduction theory, or Orch OR. It locates the hypothesized process in the microtubule structures inside neurons. Quantum processes in microtubules are organized and regulated by the biological system and undergo gravity-related objective reduction, producing events of consciousness. This is their proposed mechanism, not an established account of how the brain works.

The proposal must cross several distinct thresholds. Do the relevant quantum processes exist in the required form? Are they noncomputable? Are they necessary for consciousness? How do they produce mathematical understanding? Proving one does not automatically prove the others.

Physical tests can indeed constrain the theory. A 2022 study used spontaneous-radiation experiments to place severe restrictions on versions of Orch OR based on the simplest gravitational-collapse models. The authors also explicitly discussed the scope of the result and more complex models. It cannot be summarized as a decisive refutation of every quantum theory of consciousness.

There is also an independent philosophical problem. Even if a special quantum process were discovered, why would it produce experience rather than merely another unusual physical event? Chalmers has noted that introducing quantum processes does not automatically remove the explanatory difficulty.

Will continued AI progress directly refute Penrose? At least for now, it would not. Suppose an AI independently discovered a new proof. That would be evidence of mathematical discovery, but would not by itself establish subjective experience. Conversely, we cannot define every new discovery made by a computer as something other than understanding simply because a computer made it. That would place the conclusion of the debate inside the definition.

Penrose claims not only that machines lack feelings, but also that pure computation cannot fully reproduce the capacities produced by understanding. As machines become more reliable across a wider range of behaviors, proposing concepts, repairing arguments, identifying premises, and explaining why claims hold, he bears a growing burden to explain what principled difference remains and how it can be recognized. It is not enough to say that humans truly understand while machines merely produce results.

Scientists have another line of research that does not use Gödel's theorem to rule out artificial consciousness in advance. It begins with theories of consciousness and derives structural and functional indicators that can be examined. A 2023 study by Patrick Butlin, Robert Long, and their colleagues, for example, discussed how theories such as recurrent processing and global workspace theory might be used to evaluate artificial systems. The approach does not resolve every dispute, but it turns the question toward mechanisms and evidence rather than judging only whether outputs resemble human behavior.

Penrose's answer therefore cannot be regarded as rigorously proved. Yet I personally find his judgment deeply persuasive. That computational systems possess principled limits is a mathematical fact. That human understanding must exceed every computational system is a further claim requiring additional argument, and perhaps some form of recognition or faith.

21. Why Has Technological Progress Not Made Us Happier?

This is a deep question, with a faintly melodramatic flavor: is the problem the corruption of human nature or the collapse of morality?

I think both human adaptation and institutional arrangements matter, but they are problems at different levels and should not receive an easy fifty-fifty division of blame. Psychological adaptation must explain why we do not remain perpetually excited after receiving a convenience. Institutional analysis must explain why increased productivity still leaves some people without more time, greater stability, or a larger right to decide how they live. The first cannot excuse the second. The second cannot eliminate the first.

Placed beside earlier questions from friends about AI, employment, and the economy, a more important question emerges: society can do more and more, so why has the individual's ability to direct their own life not grown accordingly?

Human beings do become accustomed to good things, but that does not make improved living conditions meaningless. We should distinguish three things: better material conditions, a more pleasant daily mood, and a judgment that one's life is worth living. They are related, but they are not the same.

Suppose a technology frees someone from two hours of exhausting labor every day. The person may be delighted at first and no longer think about the wonderful invention a year later. That does not make the two hours worthless. They may become sleep, time with family, or simply freedom from constant haste. Ceasing to feel excitement about a good life does not erase the difference between a good life and a bad one.

Psychologists do study hedonic adaptation, and emotional responses to some life changes weaken over time. But a meta-analysis combining 188 studies found that patterns differ among events and measures of well-being. Evidence of partial adaptation cannot be expanded into the claim that human beings adapt completely to everything and external conditions therefore do not matter.

Research on income requires the same caution. A joint reanalysis published in the Proceedings of the National Academy of Sciences in 2023 found no single income plateau in the relationship between income and happiness that applied to every working adult in its U.S. sample. The authors also emphasized that the evidence was primarily correlational and that income explained only part of the variation in happiness. “Money cannot solve everything” may be true. It does not follow that improved income solves nothing important.

I certainly do not think a discussion of housing, income, or job pressure should begin with “people are simply never satisfied.” That can rewrite unresolved material problems as defects in the character of the people living through them. Especially when it comes from the wealthy, it is the familiar advice of someone speaking comfortably about a burden they do not carry.

The crucial transition is this: Who owns the time saved by technology? Consider a purely hypothetical case. A job once required eight hours and requires four after the adoption of AI. Several arrangements are possible. Employees may work fewer hours. They may work the same hours, produce more, and receive higher pay. Or the company may employ fewer people and use the savings to lower prices, invest, or increase profit.

The same technology can produce each result, with radically different meanings for different lives. “AI doubled productivity” does not entail that an employee gained four hours of free time. Pricing, ownership, employment contracts, bargaining power, and particular corporate choices stand between the two. Technical parameters alone cannot determine the outcome.

Nor is greater output necessarily bad. If a medical service is severely undersupplied, using efficiency gains to serve more people may be more valuable than immediately shortening everyone's hours. The question is not whether every additional unit of output is worthless, but who decides how the new capacity will be used and whether the people affected participate in that decision.

An empirical case is instructive. A study published in Nature Human Behaviour in 2025 followed 2,896 employees across 141 organizations during a six-month trial of a four-day workweek. Pay was maintained and workflows were reorganized. Employees reported improvements in burnout, job satisfaction, and physical and mental health. Twelve control companies did not show the same changes. The study does not prove that every industry can reproduce the result without cost, but it shows that converting efficiency gains into a different arrangement of time can improve life. It is evidence that organizational arrangements matter, not that AI automatically creates happiness.

Instead of asking only whether technology has advanced, we should also ask how much labor saved has become time that people can direct for themselves.

Institutions do not necessarily stand still. They may change continuously while pursuing another objective. Saying that institutions have failed to keep pace with technology is sometimes correct but incomplete. An organization may adopt new technologies very quickly to optimize schedules, track performance, and accelerate decisions, while improving grievance procedures, rest arrangements, and employee participation much more slowly. Institutional innovation is occurring, but the capacity to manage people is growing before the arrangements that protect and include them.

An OECD survey of more than 6,000 companies across six countries in 2025 displayed this double edge. Managers believed algorithmic management improved some decisions and their own job satisfaction, while also worrying about unclear responsibility, difficulty understanding the tools' judgments, and inadequate protection of employee well-being. The survey measured managerial observations and opinions and cannot be treated as a direct account of every employee's experience. It nevertheless shows that efficiency and the human condition do not align automatically.

Digital products present the same problem. If a product defines success primarily by time spent, then making it easier for you to stop using the product and return to your life may not be the most rewarded direction of design. A product's ability to capture you and its benefit to you are different objectives.

A randomized experiment published in the American Economic Review in 2020 found that asking participants to stop using Facebook for four weeks before the 2018 U.S. midterm elections increased their subjective well-being but reduced their knowledge of factual news. This does not prove that the internet makes everyone unhappy. It shows that a single technology can generate gains and losses at the same time, and that use cannot be treated directly as welfare.

The question is therefore not only whether the results are fairly distributed. An earlier question comes first: What are we developing and rewarding? Are we helping people complete lives of their own, or helping them complete system-assigned metrics more efficiently?

Human dissatisfaction and institutional pressure shape each other. They cannot be fully separated, because some forms of dissatisfaction do not arise spontaneously in an isolated individual.

Imagine a school with one hundred places and one thousand applicants, all of whom receive stronger learning tools. Their abilities may improve, but if the number of places and the selection method remain unchanged, competition will not automatically disappear. A person may continue investing time not from greed, but because everyone else is doing so and stopping means losing the opportunity.

There is a simple distinction between wanting more and having to obtain more continuously in order to preserve one's existing life. Likewise, wanting a decent life differs from wanting always to rank above others. Better basic security and public services can allow more people to realize the former simultaneously. If the latter is defined strictly as relative rank, it cannot be satisfied for everyone at once.

An institution can therefore do something important without abolishing ambition: reduce the punishment for falling a little behind, resting for a period, or choosing another way of life. A person should be able to decline the race for first place without losing basic security and dignity.

This also explains why contentment is sometimes genuine freedom and sometimes merely an injunction to endure. Choosing to want less when one has the conditions to leave a competition is different from having no choice and being told to accept one's lot.

Dissatisfaction itself need not be unhappiness. Wanting to continue learning, write another work, or improve a relationship may be part of a full life. The distinction is whether the effort is a willing exploration or a forced race whose costs of stopping are unbearable.

Institutions cannot and should not provide happiness in full, because people want more than care. Institutional explanation needs a boundary. Even major improvements in income, housing, and working hours cannot guarantee intimate relationships, creative passion, or a direction in life. This is not a retreat to “your attitude is the problem.” It is an acknowledgment that happiness concerns not only what one receives, but how one participates in living.

Self-determination theory identifies autonomy, competence, and relatedness as important psychological needs. A major review in the field of work analyzed the relation of these needs to psychological growth and well-being. This provides at least one research-based account of why greater convenience cannot fully replace autonomous action, learning and growth, and relations with other people.

Consider writing. If someone needs only an adequate instruction manual, a tool that drafts it quickly may be immensely valuable. If the person writes in order to form observations and a voice of their own, complete substitution may not constitute help. Some processes are burdens. Others are themselves the life someone wants to live. We cannot treat all of them as though faster replacement were always better.

A good institution should therefore neither promise to arrange happiness for everyone nor prescribe one way of being happy. It should provide conditions under which people can try, fail, rest, and say no to a highly efficient form of life.

Where should emphasis fall between the psychological and institutional explanations? In public questions, I think we should examine institutions and living conditions before asking whether individuals need to adjust their desires. This is not because psychology is unimportant. It is because a generalization about human nature should not allow us to skip matters that can be concretely investigated and changed.

After technology advances, does ordinary people's real income improve? Is saved time left to the person who saved it? Who bears the risk of unemployment or retraining? Do people subject to algorithmic management have opportunities for explanation, appeal, and exit? These questions are more open to verification than an abstract judgment about whether modern people have become greedier.

In individual life, however, we cannot postpone everything until institutions are perfected. Institutions can protect room for choice but cannot decide what is worth pursuing. Individuals can learn not to be ruled by comparison, but should not be asked to endure unreasonable conditions through attitude alone.

My answer is therefore this: technology expands the possible ways in which we can live; institutions determine to whom those possibilities are genuinely open; and each person must still decide how to use them. Without all three, progress may remain only a growth in capacity and numbers.

From the Kantian perspective that persons are ends in themselves, progress cannot be measured only by how much people accomplish for a system. We must also ask whether, as the system grows stronger, people gain better conditions for living in ways they endorse, including both the capacity to continue and the right to stop.

22. Why Does Academia Resist AI?

People may have seen a great deal of news about academic resistance to AI, but the situation is not quite so bleak. Consider a 2026 survey in Research Policy of more than 6,000 researchers at Germany's Max Planck Society and Fraunhofer Society. Of them, 25.9 percent used AI tools daily or more often, while only 22.2 percent did not use them at all.

By comparison, almost no one in Nature's 2023 survey of 1,600 researchers said they used AI frequently in research. The change over two years was substantial. The new survey covered two German organizations focused on science and engineering, so its findings cannot be generalized to every discipline, let alone the world. But one point is clear: using AI and criticizing it can occur at the same time.

Nor does the ivory tower speak with one voice. Fields Medalist Timothy Gowers attended the Leiden workshop in September 2025 that produced the Leiden Declaration. The declaration went online on June 2 and received more than one thousand signatures within twenty-four hours, as well as endorsement from the International Mathematical Union. Gowers himself did not sign. He did not identify a particular principle he opposed. He thought the declaration made several assertions with greater confidence than he possessed. Separately, he argued that the existing hierarchy of journals, appointments, prizes, and similar institutions urgently needed alternatives.

The question, then, is not whether academia resists AI, but where the resistance falls.

One fact is easily underestimated from outside academic life. Scholars may not face a more powerful competitor so much as a growing volume of material they must check. Suppose someone uses AI to generate fifty documents that look like papers and sends them to an expert for criticism. The generator experiences a tremendous gain in efficiency. The recipient must determine, one by one, which are worth reading, which citations do not exist, and which crucial derivations hide flaws. The cost of producing material has fallen; the cost of establishing its reliability has not fallen at the same rate, and the costs land on different people. When scholars say, “Please stop sending me AI-generated material,” they may not mean that outsiders should be excluded. They may mean that unfinished quality control cannot be transferred wholesale to others for free.

Up to this point, I think academia's reasons largely hold. Its real problem lies not in access to knowledge but in the ability to publish it.

People often say knowledge is already open: papers can be downloaded, theorems are public, and arXiv is free. But free downloads of an already filtered set of results do not constitute equality. What is actually allocated is the power to place something inside the set called knowledge.

arXiv itself is the best example. Despite its reputation as an open preprint repository, it has required endorsement since January 2004. Someone submitting to a category for the first time must first obtain endorsement from an author who already has submission privileges. In December 2025, the mathematics section tightened the rule: new submitters without a previous arXiv mathematics paper had to secure a personal endorsement. In January 2026, the requirement expanded across the site, and even an institutional email address was no longer sufficient evidence of minimum research competence. arXiv advised applicants to find an endorser “who knows you personally.”

Its reasons were real. Nonscientific submissions from automatically endorsed users had grown unsustainable, rejection rates were increasing, and moderation and staffing costs had become excessive. This is exactly the earlier problem: verification costs rise, and the system has one available response, narrowing the gate. No one needs a conscious desire to protect status. Once an institution responds according to its incentives, access contracts automatically because contraction is the tool at hand.

But notice arXiv's own statement: endorsement is not peer review. The endorser does not certify scientific correctness. They confirm only that the submission is appropriate to the category and that the author appears familiar with the field's basic conventions. The gate therefore judges not the content but the submitter's identity. The alternative offered to those denied endorsement is that they may still post the paper on their own websites or submit it to peer-reviewed journals. In plain language: you may write, but you cannot enter here.

This gives us a useful measure. Disclosure of tool use, accurate citations, verification material, human responsibility for correctness, and reviewer confidentiality are requirements concerning content. Anyone willing to satisfy them can do so. AI can even help people without academic lineages place their work in independently verifiable form. By contrast, questions such as “who is qualified to announce a result,” “who counts as a member of the field,” and “whether an activity should be permitted to occur” impose requirements of identity. They determine entry into the registry. A threshold may have reasons, but those reasons must be tested. The test is to ask which type of requirement it imposes. Most of the loudest resistance currently belongs to the second type.

Another comparison deserves thought. The same institutions tolerated decades of filler papers produced for evaluation, read and cited by no one, without calling them research misconduct. When AI began helping people produce similar material, the term appeared quickly. The judgment concerned not the quality of the thing, but who produced it.

We cannot expect AI companies to open the gate either. They have removed some barriers to entry and allowed more people to touch frontier problems, but they bypass the registry themselves. Results are announced in press releases; verification is left to the community; names and credit go to the company. They have not changed who counts, who is cited, or who may announce a result. They have constructed another road that only they can travel.

My view is this. Academia has good reason to require that results withstand scrutiny. It has no good reason to require that the same people produce them in the same way as before. We can judge a particular act of resistance by the same standard. If a critic identifies a testable problem and explains that acceptance will follow once it is resolved, that is quality control. If the critic merely repeats that “anything made by AI does not count” regardless of reliability, that is conservatism. The standard applies equally to the other side. If AI companies dismiss every criticism as an old establishment obstructing progress, they too are building a moat, only in different clothing.

23. Why We Live in the Dao Even If We Have Never Read Laozi or Zhuangzi

This is an interesting observation, though it would be more accurate to put it another way: we all use a language jointly shaped by the various schools of ancient thought. Using a language that a tradition helped shape is not the same thing as agreeing with that tradition.

Let us begin with the character dao. Laozi and Zhuangzi did not invent it. Its original meaning was simply a road people traveled. The Shuowen Jiezi glosses it as “the way one travels,” and the Book of Songs, in “Dadong” from the “Lesser Odes,” already says that “the Zhou road is like a whetstone, straight as an arrow,” using a whetstone and an arrow to describe how level and straight the road was. This is the word in its plainest sense. From a physical road, it naturally came to mean a route, a method, a standard of conduct, a school of thought, and finally the fundamental order of heaven, earth, and all things. To walk, we need a sense of direction; to act, we need a way of doing things; in arranging a life, we also ask where we ought to go. A path one can actually walk readily becomes an image through which to understand method and direction. The importance of Laozi and Zhuangzi lies not in inventing this everyday word, but in using it to unfold extraordinarily deep reflections.

More importantly, dao was never a term reserved for Daoists. Confucius said, “If I hear the Way in the morning, I can die content in the evening,” and also, “Those whose Ways differ do not make common plans.” The Doctrine of the Mean opens by saying, “Following our nature is called the Way.” Even Han Fei of the Legalist school wrote an entire essay, “Explaining Laozi,” to interpret the Laozi. A word used by all the schools, over whose interpretation they competed, could never be anyone’s private property.

Of the several words at issue, zhidao, “to know,” is most worth taking apart. The “Record of Learning” in the Book of Rites says: “Jade that is not carved does not become a vessel; a person who does not study does not know the Way.” Here, “know” and “Way” are still two separate words: through study, one comes to understand the principles one ought to follow. The fact that this sentence comes from a Confucian classic on education is enough to show that “knowing the Way” does not mean becoming a Daoist believer. In modern Chinese, zhidao later solidified into a two-syllable word whose function has changed. When we say, “I know where the keys are,” we are not claiming to have grasped the foundation of the cosmos. When we say, “I know there is a meeting tomorrow,” we do not first have to answer what the Dao is. A word can preserve traces of its history while acquiring a changed everyday function. We cannot force the meaning of two characters used together in ancient times, unchanged, into every use of the modern word.

Daolu, “road,” belongs to the same category. Dao and lu were near-synonyms combined into a word, one of the most common patterns of word formation in Chinese. Like words such as “morality,” “method,” and “route,” it bears no additional philosophical weight merely because it contains dao.

As for daoli, “reason” or “principle,” Han Fei’s “Explaining Laozi” does indeed say that “those who conduct their affairs according to the patterns of the Way can accomplish anything.” But one ancient usage is not enough to prove that the modern word derives entirely from a single thinker, much less that whenever we “talk reason,” the reason in question is the Way of Laozi and Zhuangzi.

Still, if we set etymology aside, the question is philosophically suggestive. We can distinguish three kinds of understanding: knowing that something is the case, understanding why it is so, and knowing how to respond to it. The three are related, but they are not the same. Someone may be able to recite a great many principles about intimate relationships and still not know how to listen until the other person has finished speaking. Someone else may be unable to offer a complete theory, yet know in an actual relationship when to explain and when first to stop defending himself. Here the metaphor of “finding a path” becomes powerful. Understanding does not merely mean adding one more correct sentence to the mind. It may also mean that, in a real situation, we are no longer quite so lost.

If we think carefully, there is even a certain irony here, because Laozi and Zhuangzi might not approve of the confidence with which we “talk reason” every day. The opening of the Laozi, “The Way that can be spoken is not the constant Way,” has been interpreted in many ways. On one common reading, it reminds us that a Way that can be put into words should not simply be treated as the unchanging Way. The “Discussion on Making All Things Equal” in the Zhuangzi says, “The Way is obscured by small accomplishments, and words are obscured by flowery display,” followed immediately by, “Thus there are the rights and wrongs of the Confucians and Mohists: each affirms what the other denies and denies what the other affirms.” Partial achievements in understanding and beautiful language can themselves conceal a wider understanding. What the passage questions is not only whether other people are wrong, but also our own tendency to take the correctness of our position for granted.

This reminder is quite useful in everyday disputes. When someone says, “I am only trying to reason with you,” is he offering reasons that both sides can examine, or announcing that you ought to accept his standards? When someone says, “This is the right path,” has he turned a method that works for him into a way of life that everyone must accept? Acknowledging limits does not mean abandoning judgment, nor does it mean that every opinion is equally right. It asks only that, even when we stand by a judgment, we explain our reasons and remain open to revision when confronted with counterexamples.

So I am happy to preserve the charm of the original remark, while keeping only this conclusion: we need not have read Laozi and Zhuangzi before we can encounter, within our own language, the questions they discussed.

In the end, dao remains alive today, and an ordinary person can casually use it three times, precisely because nobody ever registered it as private property. Genuine inheritance does not lie in having the same word on our lips, but in whether we seriously face the questions that word raises. Saying “I know” every day does not mean that we know the Way. Asking once more, “Do I really understand?” is where meeting them anew begins.

24. Why Creating Panic Is a Shortcut to Power

I do often say this, and many people before me have said it as well. But before discussing the issue broadly, I want to narrow the claim a little, so that it is not read as saying that every security warning is a fraud.

Pointing out a risk is not the same as creating panic. At the same time, the reality of a risk does not make every demand for power made in response to it reasonable. Both statements must be true. The real issue is a conversion: people’s legitimate need for safety can be converted into power that others exercise without sufficient constraint.

The first thing panic changes is not our intelligence, but the order in which we rank concerns. Ordinarily we care at the same time about income, freedom, fairness, dignity, and the future. Once we believe that we are about to lose our jobs, suffer violence, or even see the entire order of life collapse, all other demands are temporarily pushed aside.

A 2004 analysis of nationwide surveys conducted in the United States after September 11 found that respondents who felt more threatened were more willing to exchange some civil liberties for security. This exchange was also affected by how much they trusted the government.

Imagine that, under ordinary circumstances, someone demanded access to all our communications. We would certainly ask why, how broad the access would be, and who would oversee it. But if we were first made to believe that refusing would bring an imminent disaster, we might well answer: deal with the danger first; we can add the procedures later.

We make this choice not because we are stupid, but because genuine emergencies really do require swift action. The problem arises when the urgency is exaggerated and the extra power has no limits. To avoid one still-uncertain loss, we first accept another that occurs immediately. That is the shortcut. The person seeking authorization no longer has to prove fully that “my governance is better.” He only has to persuade people that “without me, the consequences will be worse.”

The statement that “fear shuts down reason” is too simple. When anxious, people may actually search more actively for information, read the news, and ask others. But searching for more information does not guarantee a more comprehensive judgment. In an earlier experiment about immigration, participants whose anxiety had been induced browsed a simulated news website. They were more likely to read, remember, and agree with threatening material. As a finding from a specific experiment, this cannot be generalized to everyone, but it points to a mechanism: people did not fail to look for information. They focused mainly on, and remembered, the part that frightened them more.

Creating panic therefore aims not only to make us believe a particular claim, but also to make a particular danger occupy the center of our judgment about the world. Someone who is repeatedly told that he is about to be eliminated will devote tremendous energy to learning how not to be eliminated. He will ask less and less whether the competition itself is reasonable, whether another way of life is possible, and whether the person claiming to know the way out also benefits from his anxiety. Whoever decides what must be feared first influences the allocation of resources, attention, and action.

The most important sleight of hand occurs in the middle of the reasoning. Several steps lie between “a danger exists” and “you should obey me,” and each must be established separately. The existence of a danger does not show that a particular solution will work. The effectiveness of a solution does not show that one person must have a monopoly on implementing it. The need for coordinated action does not show that oversight can be abolished. Panic-driven rhetoric compresses these steps into the sentence, “The situation is this serious, and you are still asking questions?” A workplace example makes this clearer. A manager announces that the company faces an existential crisis and asks everyone temporarily to put in more effort. That is entirely open to discussion. But when employees ask for business data, managerial accountability, and a condition for ending the emergency measures, they are told, “If you still care about such things now, you do not care whether the rest of us survive.” The nature of the discussion has changed. The employees began by asking about a plan; they ended up being asked to prove their loyalty. One school of thought in security studies sees the matter in exactly this way: defining something as a security issue is itself a political act. It determines how the issue enters decision-making, rather than merely adding one more item to an objective list of dangers.

More dangerous than temporary authorization is a crisis that is never allowed to end. Imagine an explanation that is almost impossible to test: if nothing happens, it is because I protected you well; if something happens, it is because you did not give me enough power; if someone proposes an alternative, it is because he underestimates the danger. Each explanation might be true in some circumstances. But if every possible outcome leads to the same conclusion, that the same person should receive still more power, then the judgment no longer contains any condition under which it can be tested and revised.

In terms of incentives, this creates a real conflict of interest. The public needs the danger to diminish, while the person who controls its interpretation may need the public to continue feeling endangered. This does not mean that everyone who issues a warning hopes for disaster, nor does it require a secret conspiracy. A practical question can help us tell whether protection has turned into control: is the arrangement gradually increasing people’s ability to cope with the risk, or continually proving that they can rely only on a particular protector?

Still, panic is not a button that works whenever it is pressed. More accurately, panic is a key that works only in certain locks, and the conditions of the lock can be listed: a single source of information, no viable alternative, the ability to circumvent oversight, and a direction of blame that points outward rather than toward the people in power. If people believe that the authorities created the crisis, magnifying the danger may simply remind them that those authorities ought to be held accountable.

For the same reason, it is not enough to tell ordinary people, “Do not be afraid.” Fear of unemployment, war, violence, or technological risk may be entirely justified. What helps is preserving independent sources of information, viable options for responding, and effective oversight, so that people are not forced into a position where they can trust only one person.

As for distinguishing a responsible warning from a bid for power through fear, I prefer to ask several verifiable questions rather than guess at the speaker’s inner motives. What is the basis for the risk? Has the speaker explained the evidence, the unknowns, and the conditions of the judgment, and does he permit independent verification? Or has he substituted a worst-case scenario for a probability, while treating questions as ignorance or malice? Why this measure? Has he explained how it reduces the risk, what it costs, and whether alternatives have been compared? Or has he leapt directly from “the risk is serious” to “you must accept my plan”? Who constrains the people implementing it? Do they accept scrutiny, correction, and a review of conflicts of interest, or do they demand exemption because they possess secrets and the situation is exceptional? And there is one question most easily overlooked: under what conditions does the arrangement end? Is there a review mechanism and a means of withdrawing the additional power, or does the danger have no standard by which it can ever be declared over, so that a temporary measure automatically becomes permanent?

These questions are not a universal test for whether someone is lying. Some risks genuinely are difficult to quantify, and some information genuinely should not be public. But suitable independent review can still be arranged. The unknown can be a reason for caution; it cannot automatically become a reason for unlimited authority.

I therefore do not believe that every security warning is a fraud. But pushing the public into a situation with no time, no alternative, and no one to rely on but a particular protector can sharply reduce the burden of proof required to obtain obedience. Resisting such power does not require us first to prove that the world contains no danger. We can acknowledge the danger and act, while still insisting on one thing: identifying a problem does not automatically qualify you to decide everything for everyone.

What is truly worth preserving is not the pose of fearlessness, but our capacity, even while afraid, to judge, supervise, and withdraw authorization. This standard applies equally to every essay I write.

25. Why Philosophy Is the Mother of All Sciences

I may have put this a little too briefly before, so let me expand. First, the source of the phrase. In a 1932 letter to Bruno Winawer, Einstein wrote that philosophy was like a mother who had given birth to and endowed all the other sciences. He immediately added that the mother should not be despised because she was now naked and poor, but that one should hope her Don Quixote-like ideal would live on in her children so that they would not sink into vulgarity. The second half is really the point. He was not ranking disciplines. He was warning science not to turn itself into a craft concerned only with practical results.

The metaphor of the mother is first of all a claim about intellectual history. Newton’s 1687 book was titled Mathematical Principles of Natural Philosophy. Research that we now classify as physics was still conducted under the name of natural philosophy. Science and philosophy were not separated into today’s independent disciplines from the outset. Before methods and divisions of labor had matured, people were already asking what nature was, how change occurred, and what kind of reason was sufficient for us to believe an explanation. Some of those questions later grew into specialized sciences.

But the metaphor should not be pushed too far. A relationship of descent and the evidentiary force of a claim today are two different things. A conclusion in physics does not escape testing because a philosopher approves of it, nor should science forever obey a philosophical system simply because philosophy participated in its formation.

His most persuasive example is how he himself handled “time.” Suppose you want to measure the speed of a train. You naturally need distance and time. But before you begin calculating, a question is hidden in the setup: when do two clocks in different places count as displaying the same time? It is easy to think there is nothing to ask. Just set them alike. But how do we do that? Transmitting a signal takes time, so how do we determine that the clocks in the two places have been synchronized?

The first section of Einstein’s 1905 paper on relativity was titled “Definition of Simultaneity.” He first discussed how clocks in different places could be synchronized using light signals and made the rule explicit. He then derived the famous conclusion: two distant events judged simultaneous in one inertial frame of reference need not remain simultaneous in another frame moving relative to it. This does not mean that time bends to anyone’s wishes, nor is it merely a matter of seeing distant things later.

What really deserves our attention is what he did. Instead of merely calculating the numbers in existing formulas more accurately, he went back and examined the words in the formulas that seemed to need no explanation at all. “What time is it?” is one question. “What exactly do we mean when we say that two places have the same time?” is another. The latter kind of question is clearly philosophical, yet it can enter directly into the construction of a physical theory. Of course, asking the question is not enough. Conceptual analysis must be joined to physical principles, mathematical argument, and empirical testing.

He took this so seriously because specialized knowledge can itself become habitual. In a 1916 essay commemorating Ernst Mach, he offered a sharp observation: because a concept has long been useful, it can acquire excessive authority. People forget how it was formed and treat it as an unalterable necessity. Analyzing the conditions under which the concept holds, or does not hold, can break that authority. In a 1944 letter to Robert Thornton, he said that the educational value of the history and philosophy of science lay in helping researchers free themselves from the prejudices of their own time and develop independent judgment.

What he admired, then, was not doubting everything and therefore doing nothing. It was the ability both to use a method skillfully and to recognize when one should stop using it in the old way. Knowing many conclusions does not mean knowing where their boundaries lie.

Yet he did not advocate choosing a philosophy first and then requiring science to accommodate it. In “Physics and Reality” from 1936, he put it directly: when the foundations of physics themselves become problematic, the physicist cannot hand critical reflection entirely over to the philosopher, because the physicist knows best where the shoe pinches. In 1949 he supplied the constraint in the other direction, in a sentence with two inseparable halves: epistemology without contact with science becomes an empty scheme, while science without epistemology is primitive and muddled. He added a rather impolite remark: to a systematic epistemologist, the scientist must appear an unscrupulous opportunist, a realist when realism is needed, an idealist when idealism is needed, and a positivist when positivism is needed.

Taken together, these observations show that his “mother” was not a disciplinary hierarchy. In his own practice, philosophy was not a profession to which one belonged, but something one personally had to do when the tools at hand stopped working. It could neither be outsourced to a philosophy department nor fenced off by any discipline as its exclusive territory. The point of his 1936 remark was precisely that no one else could do it for you, because the shoe pinching was your own.

This returns us to the second half of the 1932 letter. What concerned him was never that science was insufficiently precise. It was that, after achieving results, science might stop asking what it was doing. Have we mistaken a measurable result for the whole value associated with it? Being able to do more does not automatically tell us how much we understand or how we ought to use our abilities.

Of course, philosophy’s importance neither needs nor can be proved by Einstein’s praise. To do so would be to replace our own necessary judgment with the authority of a famous person yet again.

26. Why “Alienation” Is Everywhere in the Age of AI

This friend truly understands the issue. He went straight to the key term. Let us first narrow it down; otherwise it will explain everything and therefore explain nothing. Alienation does not mean simply that “work is exhausting” or that “life feels meaningless.” Marx described a specific relationship: a person’s own activity and its products become forces that he cannot control and to which he must instead submit.

In the Economic and Philosophic Manuscripts of 1844, Marx discussed the alienation of labor in four respects. The worker is separated from the product: it becomes someone else’s property and in turn dominates him. He is separated from the labor process: he does not decide what to do or how to do it. He is separated from his “species-being”: the free and conscious creative capacity proper to human beings is compressed into a means of survival. And he is separated from other people: relations among people are reshaped by these relations of production.

“Species-being” sounds abstract, but it simply means that human beings can do more than complete tasks. They can ask why something should be done and how to do it, and can develop their abilities through cooperation with others. Marx believed that when this capacity has no room at all to unfold in work, what is lost is more than wages.

There is a distinction easily missed, and once missed it renders the whole concept useless: creating something external to oneself is not in itself alienation. When we finish an essay, it becomes a work that others can read and criticize. In philosophy, this is called objectification, and it can be a way of realizing ourselves. Alienation arises only when the product turns back and dominates us, and when the activity becomes detached from our own purposes. Alienation does not ask whether “the world now contains many more things.” It asks whether, in making those things, people can still recognize the activity as their own.

Why, then, does alienation not necessarily diminish when life improves? Because receiving more compensation and gaining more power to decide are not the same improvement.

Imagine that the same technology reduces an eight-hour job to four hours. An organization can now let its employees work four fewer hours. It can maintain the eight-hour day while doubling output and share some of the gains with them. Or it can maintain the eight-hour day and simply double the workload. All three arrangements use the same technology, yet they produce three different lives for the people involved. Technology creates a new capacity; it does not itself determine to whom that capacity belongs.

Marx made the distinction quite clearly in the chapter on machinery in Volume I of Capital. Machinery can reduce labor, but under particular relations of production it can also be used to lengthen the working day and increase the intensity of labor. His criticism was never that machines made people degenerate. It concerned how machines were used and the relations of domination behind that use.

If Marx saw AI today, he would say the same thing about it. He would not criticize AI for making people degenerate. He would criticize the way AI is used and the relations of domination behind it.

Of course, we cannot reverse the argument and say that improvements in wages, housing, and rest are meaningless. They change a person’s situation in real and important ways. But “I live more comfortably as a result” does not automatically mean “I have more power to decide how I live.” The two also affect one another. When a person has no tolerable option of leaving, his “voluntary negotiation” is not the same as that of someone with somewhere else to go.

More importantly, this arrangement can continue without anyone deliberately designing it from behind the scenes. In his discussion of capital accumulation in Volume I of Capital, Marx described competition as a force that also constrains the owners of capital: an enterprise must keep accumulating and expanding production or risk losing its position. This is a structural explanation, not an accusation that particular bosses are greedy.

Consider a hypothetical example. A company wants to reduce the intensity of work, but fears that higher costs will lose it orders. Employees want to work less overtime, but fear that the first person who stops responding will lose a promotion. Each party therefore thinks, “I do not want this either, but I have no choice.” An arrangement can be sustained by everyone’s actions even when no one sincerely likes it.

That does not mean everyone has the same power and responsibility. The person who can decide contracts, budgets, and workloads has a far wider range of choices than someone who can only accept or leave. Competitive pressure may explain conduct, but it does not automatically absolve decision-makers of responsibility.

In his discussion of the “fetishism of commodities,” Marx added another layer: social relations among people appear as relations among things, as though prices and capital had grown legs of their own. This is related to alienation, but it is not the same concept. We can draw a limited analogy with claims we hear today: “No one decided this; the market requires it,” or “It is not that I refuse; the AI calculated it.” These statements may point to real constraints, but the questions they leave behind do not disappear. Who set the rules? Who can change them? Who bears responsibility when they go wrong? Why do the costs always fall on the same group of people? “No single person decides” is entirely different from “this is a natural law that cannot be discussed.”

What is new about AI today is that work can not only be managed, but continuously measured, predicted, and optimized. According to a survey published by the OECD in 2025, algorithmic management tools were already quite common in most of the countries surveyed. Managers believed they could improve decision-making, while also raising concerns about unclear responsibility, the difficulty of understanding the tools’ logic, and inadequate protections for employee well-being. This was a survey of managers, not a measure of workers’ alienation, and algorithmic management is not the same thing as generative AI.

Such tools cannot in themselves be classified as alienating. One scheduling system may reduce confusion and make people’s rest more predictable. Another may convert every minute into a performance demand and leave people without even an opportunity to explain their actual circumstances. The difference is whether the technology supports human judgment or makes it increasingly difficult for people to participate in, question, and change the rules that decide their lives.

The same structure exists in the writing of key opinion leaders. Someone begins writing to clarify a problem, then discovers that some kinds of content attract more traffic and adjusts his expression. There is nothing necessarily wrong with this; understanding one’s readers is part of writing. But if the process continues until a lack of traffic means that a question is not worth writing about, and an inability to monetize means that the effort has no value, then eventually even the question of what the writer wants to ask must await the platform’s approval. Feedback that originally helped the work reach readers has become the standard that determines what kind of person the author should be.

The alienation we see among many leading online influencers is not, in a strict sense, a mechanism Marx himself studied. There was no internet in his time, let alone key opinion leaders. It is an extension of his question into a contemporary case. The point is not that all metrics are bad, but that a metric can change from useful information into a tool that may not be questioned.

Another point is easily overlooked: people who appear from the outside to lack nothing can feel this deeply as well. They may have good incomes, stable work, and abundant recognition, yet find it increasingly hard to answer why they should continue down the same road. In The Holy Family, Marx and Engels even discussed how property owners, too, exist in a state of self-alienation, though they derive power from the relationship and therefore feel at ease within it. In their theory, a person’s place in a social relationship and whether he subjectively suffers can be discussed separately, though the power and consequences attached to different positions are by no means equal.

But we must put up a gate here, or the entire theory becomes impossible to refute. We cannot broadly and indiscriminately diagnose someone’s suffering as alienation. Nor can we respond to a person who is satisfied with life by declaring that he “simply has not awakened.” If suffering proves the theory and satisfaction proves it too, then it is no longer an explanation. It is a posture that remains correct whatever happens. To discuss an actual person, we must listen seriously to his experience and examine how much room he really has to choose, rather than first deciding for him what life he should truly want.

Nor should we describe alienation as an unalterable fate. In his “Comments on James Mill’s Elements of Political Economy,” Marx imagined a different relationship: people could see the realization of their abilities in their own works, and could gain confirmation from the fact that those works met the needs of others. Here, non-alienation does not mean living alone and needing no one. It means that one’s relationship with one’s activity, its products, and other people has changed as a whole. In the final chapters of Volume III of Capital, he discussed producers jointly bringing necessary productive activity under their control, placed great weight on the development of human capacities in free time, and treated the shortening of the working day as a basic condition. This was a direction he proposed, not a claim that the conditions already exist whenever a system adopts the corresponding name.

Marx’s own understanding of “self-realization” can likewise be debated. A person may willingly do clearly bounded, repetitive work and devote his main passion to family or writing. We cannot conclude that he has failed to live well simply because his job is “not creative enough.” Any theory of alienation must explain what conception of human nature and what value judgments it rests upon. It cannot prescribe one ideal life as the only answer for everyone.

The term continues to be used in the age of AI not because the world has made no progress, but because “growth in society’s overall capacities” and “whether concrete persons control their own activity” have never been the same thing.

If we want the word “alienation” to do real work, we should treat it as a set of answerable questions rather than a vague feeling: What have I helped to create? Who decides what it is used for? Can I influence the process? And am I able to understand myself only according to what it demands of me?

That last question is precisely the one Marx himself discussed least. His analysis was firmly rooted in relations of production: who owns and who controls the labor process. A more common form today begins when people voluntarily install an external standard inside themselves. They then use it to manage and assess themselves, translating what originally belonged to a structure entirely into the question, “Am I simply not working hard enough?” Marx gives us a starting point for this layer, but not the whole answer.

27. In the Age of AI, Can More Than a Decade of Education Be Compressed into Four Years?

I have indeed been saying this lately. But let me begin by explaining what I am not proposing, or people will refute it as if it were three other claims.

I am not saying that students should “learn faster.” Playing the same curriculum at four times the speed is not compression; it is cramming. Nor am I saying that everyone should earn a doctorate. Expanding opportunities for research and requiring everyone to obtain a research degree are two different things. Still less am I saying that, once we have AI, learning itself no longer takes time.

I mean something else: in the decade or more from high school through a doctorate, the time genuinely spent developing abilities is far shorter than the entire period. Much of the time is wasted on waiting, repetition, and satisfying procedural requirements. These happen to be exactly the things AI is best at cutting away.

Let us begin with “waiting.”

A student understands most of the material but gets stuck on one crucial concept. Without timely feedback, he may keep solving problems and repeating the same mistake. Another student mastered the concept long ago but must practice it again with the whole class. These two students do not need the same kind of acceleration at all. The first needs someone to identify his misunderstanding. The second needs permission to move on.

A traditional classroom cannot provide both at once, because it can advance at only one speed. In practice, the pace of a class is the slowest pace among all its students, multiplied by an administrative calendar.

In a randomized controlled trial published by Harvard researchers in 2025, 194 undergraduates took two introductory physics lessons. The group using an AI tutor created with deliberate instructional design outperformed the active-learning classroom group on post-lesson tests, while the median study time was 49 minutes, compared with about 60 minutes in the classroom. These were short-term results from only two topics, not an entire course, much less an experiment in doctoral education. We cannot extrapolate from them proportionally to a whole educational system. But they do at least reopen a question: does material that has traditionally taken a semester truly need a semester?

Next comes “repetition.”

High school, undergraduate education, a master’s program, and a doctorate are treated as four repositories of knowledge that must be emptied in sequence. From the standpoint of learning itself, however, the real constraints are prerequisite relationships, not status labels. If someone already understands and can use calculus and linear algebra, must he wait until he possesses the identity of an “undergraduate” before studying more advanced statistics? Conversely, if someone has enrolled in a master’s program but has gaps in the foundations, changing the label will not fill them automatically.

The prerequisite chain is in fact much shorter than the calendar. The number of genuinely necessary links between arithmetic and the ability to read a frontier paper in a field is limited. We stretch that chain over more than a decade largely because each level requires regrouping students, repeating admissions, and dividing study into semesters all over again. Nor must the boundaries between levels remain separate. MIT allows qualified undergraduates to pursue a master’s degree concurrently and, with approval, to use the same master’s thesis to satisfy their undergraduate thesis requirement. Progressing according to mastery rather than time spent in class is not an idea that arose only with AI. Western Governors University has long used a competency-based model, though chiefly for adult learners. AI makes such organization easier to scale; it did not make the logic possible for the first time.

At this point, we must pause and clarify what time cannot be compressed, or the proposal will turn into the claim that learning takes no time.

The hardest case is the doctorate. A doctorate certifies more than having “finished even the most difficult textbooks.” It also certifies whether someone can take responsibility for his own research judgments. Suppose AI helps formulate a question, write code, analyze data, and finally produce a correct new result. That result may have genuine scholarly value and should not be dismissed merely because a tool was used. But to judge whether the person has reached the doctoral level, we still need to ask: Does he understand the crucial assumptions of the research? Does he know under what conditions its conclusion would fail? When new evidence appears, can he explain why the original judgment should be revised? “Independence” should not mean turning off every tool and working with one’s bare hands. It should mean seeing clearly what the person actually understands, contributes, and can take responsibility for within human-AI collaboration.

An AI proving a mathematical theorem is good. Better still is the person using the AI understanding the proof and being able to advance the direction and methodology of mathematical research.

The second limit is time required by the object of study itself. If we want to know the actual consequences of a change three years later, AI can help design a study and process existing data, but it cannot turn three years of observations that have not yet occurred into evidence already obtained. A simulation can generate a prediction; prediction and observation remain two different things. For example, we might infer that gravitational waves are not themselves bent by gravity, but we must at least wait until an experiment can be conducted before confirming or falsifying that inference.

The third limit is the measurement trap most easily overlooked. An experiment published in the Proceedings of the National Academy of Sciences studied mathematics instruction for nearly a thousand students at a Turkish secondary school. Students using ordinary GPT-style assistance performed about 48 percent better than the control group in practice. Once the assistance was removed, however, they scored about 17 percent lower than the control group on an independent examination. With a specially designed tutor that did not provide answers directly, gains during practice were even larger and the negative effect after removal largely disappeared. But independent test scores still were not significantly higher than those of the control group.

The same AI, then, can accelerate the formation of ability or help someone bypass that formation more quickly. Any proposal for compression must therefore measure more than how many courses were completed or papers submitted. It must examine long-term retention, transfer to unfamiliar problems, and whether the person can explain and revise his own judgments. We must measure separately what the person himself has mastered and what the human-AI system accomplishes jointly. The latter should not be passed off as the former.

Once we include these three constraints, “four years” is obviously not an established conclusion. No existing study proves that ordinary students can go from high school through a doctorate in four years.

But there is an asymmetry here that I find more noteworthy than the evidence itself.

The side that wants to move faster is always the one asked to provide proof. No one asks the current system to prove why education must take twenty-three years. Those numbers were never derived from the laws of learning. A bachelor’s degree takes three years in Britain and four in the United States. The same degree differs by a year. The duration of primary and secondary education also varies between countries. If twelve years, four years, two years, and five years truly followed from cognitive development, they should not differ so widely. They look more like the combined result of historical inheritance, administrative cycles, and admissions schedules, subsequently treated as natural constants.

The right question is therefore not, “On what grounds do you say four years are enough?” It is, “What exactly is being done in every one of these years? Explain them year by year.” An arrangement that has never had to provide evidence should not enjoy the authority to demand evidence from everyone else.

When I say four years, I mean less an exact number than a parameter designed to force a question. The real target is certification by years served. A degree should depend on whether a standard has been met, not on whether a pilot program has reached its fourth year and must now award everyone a credential. Three changes are needed. First, curricula should follow relationships of knowledge dependence, with permission to skip what has already been mastered. Second, ability should be assessed in layers: foundational understanding, professional application, and research capacity should be evaluated separately. Producing an excellent research paper does not automatically prove that every foundation is sound, and top marks in foundational courses do not automatically prove research ability. Third, completion time should vary from person to person. It may be shorter than the traditional course of study and, at some stages, longer. Moving from fixed duration to competency should not simply replace one uniform slow speed with one uniform fast speed.

Finally, let me explain why I care about this.

The years released by compression are not an abstract efficiency gain. They are a person’s twenties. Under the current path, someone who begins school at six finishes a doctorate at around twenty-nine. The additional seven or eight years happen to be years when a person has great physical strength and energy, is willing to take risks, and can afford to lose. If the system consumes them, it should justify each year.

Nor should compression merely send people earlier into the next competition for qualifications. If after four years someone can choose continued research, employment, creative work, a change of direction, or simply some rest, then the waiting eliminated has become room for choice. If instead it becomes, “You are eighteen or nineteen and still do not have a doctorate, so you have not worked hard enough,” we have only installed a harsher clock. AI determines neither outcome. Institutions do.

My proposal, then, is this: we should not deny a person’s actual abilities because he has not served the required number of years. Nor should we count an ability as formed merely because a tool has produced an impressive result. Whether four years are enough can be tested and revised. The supposed necessity of spending more than a decade has never been explained at all.

28. Why We Are Always Waiting for a Great Person

This is indeed a problem everyone has to some degree. The key is that it happens without our noticing. If we were merely deceived, things would be much simpler: stop believing. What is difficult to explain is that no one has deceived us. We slip into it on our own.

Let us begin with what we actually want. Most of the time, we do not merely hope for someone who has accomplished something important. Such people exist, and acknowledging them is perfectly ordinary. What we want is someone who can guarantee the direction: he has seen the situation clearly and knows where to go, so even if I am now completely lost, there is ultimately some way through the world.

But these two ideas are far apart. The first says, “Someone can help.” The second says, “Someone must be able to take care of everything.” We usually slide from the first to the second when we ourselves cannot say what should be done.

Suppose the problem before us has already exceeded our capacity to handle it. An industry is changing, and the relationship between effort and results has become increasingly hard to understand. At such a moment, “The problem is complicated; we will have to experiment slowly” feels very different from “I have seen it clearly; follow me.” The second statement may not offer a better plan, but it offers something more complete: even if we do not know what to do, at least someone knows.

One related line of leadership research is usually called the “romance of leadership”: the image of a leader is shaped not only by the decisions he makes, but also by followers’ interpretations and expectations. Some studies have also reported an association between self-uncertainty and support for forceful leaders. These are correlations found in particular samples, not a law that “whenever people feel weak, they worship authority.” In fact, recent replication studies have failed to reproduce some of the classic effects. We can therefore call this a recurring tendency, not an inevitable fate.

Now let us return to the fact that it happens without our noticing. Why are we so often unaware of the process?

Because at the moment we surrender judgment, the experience does not feel like loss. It feels like relief. Judgment has costs: gathering information, bearing uncertainty, and taking responsibility for a wrong choice. When we give these burdens to someone we trust, the immediate experience is one of ease. It is hard to notice what we have relinquished when that very thing has always made us tired.

This is the fundamental difference from deception. Someone who has been deceived later feels that he lost out. Someone who has surrendered judgment may for a long time feel that he gained.

This is why we are inclined to surrender our agency.

Next comes the deliberate construction of the great person.

Once we have a guarantor, he cannot be allowed to fail. His failure would falsify not only him, but the judgments on which we have based years of our lives. Protecting his image therefore also protects our explanation of our own life.

One point needs to be clear: this construction usually does not take the form of knowingly lying about what we know is false. It takes the form of applying two standards to evidence. Material that praises him is accepted at a glance. Material that criticizes him is examined word by word, with demands for sources and suspicions about motives. In one classic small-sample experiment, university students read studies supporting and challenging their prior positions. They were more likely to consider the material favoring their own side persuasive, even as they believed that they were carefully examining the evidence.

A cycle emerges. Success proves that he is great; failure proves that circumstances were too complicated. Clarity proves his depth; obscurity proves that others have not yet understood. When every possible result yields the same conclusion, what is being protected is no longer the person’s actual contribution, but an expectation that cannot be allowed to fail.

There is also a simpler form of causal attribution. A company produces an important product. The founder’s judgment may indeed have been crucial, but the achievement also depended on engineers, suppliers, existing technologies, capital, timing, and many corrections omitted from the story. Compress the entire system into “the CEO is a genius,” and the narrative becomes clean. Success has a name, and the future seems guaranteed. Psychology’s halo effect describes a neighboring mechanism: our overall favorable impression of a person affects our judgments of his particular traits, often without our awareness. “He did something I admire” quietly becomes “He must also possess the other virtues I hope exist.”

We therefore need to separate four claims. One person made an important contribution. Nothing could have happened without him. He exercised excellent judgment in this matter. He has superior judgment about everything. The second and fourth require far more evidence than the first and third.

Organizational studies also offer a harder version of the problem, one not entirely psychological.

Imagine an institution, though perhaps no imagination is necessary because we often inhabit such institutions, where ordinary feedback channels do not work. Problems pass upward through several levels without being resolved. Only the personal intervention of a particular leader produces a result. Under these conditions, employees are not irrational to hope for a just and capable leader. Their experience tells them that procedures may not help, but a particular person might.

The trouble begins in the next step. If only one person may speak, obtain information, and make decisions, others have no opportunity to demonstrate that they too can accomplish things. The claim that “only he can get anything done” is then used to justify keeping resources and power concentrated in his hands. A person may appear irreplaceable because of genuine ability, or because no one else was ever given the conditions under which they could become an alternative.

Yet we cannot explain the matter entirely through evasion and interest, or we will miss the part most worth preserving.

Suppose we have long watched people sacrifice what they once valued in order to succeed. We may then hope to see someone who is capable without needing to belittle others to prove himself, someone who accomplishes things without turning everyone around him into a means. What we need may not be a master who makes decisions for us. We may need only a concrete example showing that such a way of living need not remain confined to books.

That hope deserves to be cherished. A model can make an abstract possibility visible.

The problem begins when all our hope is placed in that person. Discovering his limitations may then make us feel not only that he is imperfect, but that the values we cherish have lost their basis. This does not follow logically. One person’s failure to realize a value completely does not make the value meaningless. A person worth learning from may commit a serious wrong, but that does not force us to choose between a “perfect great person” and “everyone is a hypocrite.”

Nor should the opposite of expecting greatness be trusting no one. That too is a way of avoiding concrete judgment, only more thorough. If we trust no one, we never have to distinguish among them.

We cannot independently verify all knowledge or solve every problem alone. Relying on experts, trusting teachers, and accepting help can all be choices made after judgment. Delegating some work is not the same as surrendering the entire right to decide how one should live.

The boundary can be stated clearly. Someone who helps me see what I could not see before does not thereby acquire the right to prescribe the ends of my life. The fact that he helped me does not mean I can never refuse or correct him, or take a road different from his. This boundary applies equally to thinkers, teachers, entrepreneurs, and to an AI writing for someone else. Help proves that help occurred. It does not prove that all the helper’s other judgments should be accepted automatically.

Returning to what happens without our noticing: no one exactly tricks us out of our judgment. We loosen our own grip because letting go feels good. The real danger is not any particular great person, but that feeling of comfort.

29. Why We Want an Extraordinary Life but Keep Repeating Other People

This is a sharp observation, and the question itself seems tinged with pessimism. My view, however, is that the contradiction it identifies is only apparent. Once we take it apart, what we find is more troublesome than the original contradiction. It arises only when an “extraordinary life” is treated as an autonomous choice. In everyday usage, the word almost always operates as a comparative term.

To be extraordinary usually means to be more worth telling a story about than the average person. The term does not describe what a life is like from within, but where it stands in other people’s eyes. Once someone makes “living an extraordinary life” his goal, he has already accepted a shared standard.

And if people share a standard, they must travel broadly similar roads. Otherwise comparison is impossible. Someone can run as far as he likes on a track where he is the only runner, but he will never appear on a ranking. Repeating what others have done is therefore a precondition of the very goal of being extraordinary.

This explains a common situation. Someone refuses an “ordinary life,” then lists the places he will travel, the positions he will obtain, the things he will own, and the degree of fame he will achieve. Each aim may be worthwhile on its own. But from beginning to end, the reason for judging them worthwhile may be, “If I achieve these things, other people will think I have lived well.” What the person wants is certified uniqueness: I want to be different, preferably in a direction everyone recognizes and approves.

Besides comparability, imitation offers something else: the feeling that one is unlikely to choose wrongly.

Choosing a new direction for oneself requires facing an uncomfortable fact: we do not yet know where it leads. Someone who has already succeeded appears to offer ready-made evidence. He lived this way and obtained wealth, respect, or influence, so following him is at least not groundless. Anthropological research on prestige and cultural learning likewise shows that people do use others’ success and social esteem to choose whom to learn from. This is neither inherently irrational nor ineffective.

Where, then, is the problem? In the easy extra step. We begin by learning “how to do something” from a person, then accept from him “what is worth doing” as well.

We admire a writer and learn how he observes people and revises sentences. This develops our abilities. But if we go on to believe that our own writing counts only if we possess his experiences, reputation, and manner of life, borrowing methods has become copying an identity. A successful person’s experience shows, at most, that this road once worked under his conditions. It cannot prove that we share those conditions, much less that upon reaching the same destination we would approve of the same life.

Perhaps the most dangerous and least visible development occurs when other people’s demands become “my own desires.”

From childhood, a person repeatedly hears that only certain achievements count as promise and only certain positions merit respect. Later, no one needs to command him. He still says, “No one is forcing me; I simply cannot tolerate mediocrity.” This may be true. It may also mean that the external standard has entered his self-evaluation, so that stopping makes him feel ashamed, unworthy, and left behind. Self-determination theory distinguishes this internalized pressure from autonomous motivation grounded in understanding and endorsement. The fact that a demand has entered the mind does not mean it has become a value the person endorses.

It is therefore sometimes insufficient to ask, “Is this what I myself want?” We must ask one more question: do I want the thing itself, or can I simply not bear other people’s judgment, and my own, if I fail to achieve it?

This is not a golden rule either. Someone may genuinely value income, career, family, and the approval of others. Motives are naturally mixed; they are neither wholly autonomous nor wholly coerced. Some people follow familiar roads not from a lack of imagination, but because they cannot bear the cost of experimentation. The same failed career change may cost one person only a year of lower earnings and another the basic livelihood of an entire family. We cannot blame both forms of caution by the same standard. “Why not live boldly for once?” cannot substitute for understanding someone’s actual circumstances.

Now we come to the point I most want to preserve from the question: repetition itself is not the problem.

Other people have already loved. Does that make my love unreal? Other people have raised children, written novels, and learned the piano. Does my doing these things amount only to repetition? Obviously not.

From an observer’s classification, it is the same activity. From the participant’s experience, it may be the first time he has truly learned to let another person finish speaking, the first time he has understood a sentence he read long ago, or the first time he is willing to bear the consequences of a relationship. Something having happened in cultural history does not mean it has happened in this person’s life.

Living one’s own life therefore does not require inventing a form of life never seen before. One person can autonomously choose stable work and a familiar daily routine. Another can constantly pursue novelty and difference only because he fears not being special enough.

Deliberate opposition is not freedom either. If I must hate whatever other people like, they still determine my direction; I have merely put a minus sign in front of it. This and the pursuit of an extraordinary life are two directions on the same axis.

One distinction is, I think, the key to the entire problem: originality asks how you differ from what already exists; agency asks how you participate in forming, endorsing, and revising your own direction. The first is comparative. The second is not.

The way to escape the shared standard is therefore not to run farther, act stranger, or become rarer along the same dimension. That is still comparison. The way out is to change dimensions. Once we do, the question that actually matters changes as well. It is no longer, “Has anyone walked this road before?” but: having understood what it will cost and what it may bring, do I still want to walk it? And if I discover that it does not suit me, can I adjust?

External results can be duplicated. Two people may achieve the same professional success. Through the process, one gradually develops abilities he values and endorses the responsibilities he has assumed. The other has merely crossed the last item off a qualifying checklist that burdened him for years. Their achievements look alike, but what they gained is entirely different. Psychological research on goal concordance has likewise found that goals more closely aligned with a person’s interests and core values are associated with more sustained effort and greater gains in well-being after achievement. This does not guarantee that following one’s heart will produce success. It does show that between “achieving a goal” and “what it feels like to live the resulting life” lies a crucial question: why is this goal worth pursuing?

Finally, the demand to “live an extraordinary life” should itself be examined. Why must everyone possess a life worth narrating and envying? May a person not cherish stability, intimacy, concentration, or even a period that for now yields no visible result? If even “being yourself” becomes a task that requires more stories and more style than other people have, it is only another form of comparison.

What deserves our real vigilance is not similarity, but gradually ceasing to ask whether this is the life we still want, because we are trying to make ourselves into a success others can recognize at a glance.

A life need not be unprecedented to belong truly to oneself. Nor must it be enviable to be worth living.

30. Why Death May Not Be the End of Everything

Good question. This is indeed my personal view, not a scientific conclusion. I do not base it on evidence, near-death experiences, what some great scientist believed, or the statement that “science cannot prove it does not exist.” It is a judgment of mine, or, one might say, a faith. But it is a faith with reasons, reasons that can travel as far as they are able. At the point where they can go no farther, the final step is one each person must leap for himself. Yes, Kierkegaard’s leap: the leap of faith.

My reason can be condensed into a single sentence: the source of negation is unknowable, yet reason is somehow able to perform negation, so reason and negation must come from the same source.

Let us begin with negation itself. I do not mean the kind of negation involved in instinctive avoidance. Suppose we are looking for our keys, glance at the table, and say, “The keys are not on the table.” What did we see? A tabletop, several books, and a cup. We did not see “not.” There is no object on the table called “not.” We brought that “not” with us. We first held the keys in mind, compared them with what lay before us, and only then did “not” appear. The world itself gives us only what is: this is here, that is there, this becomes that. “Not” appears only where someone judges.

Where, then, does this “not,” this negation, come from? I think the question cannot be answered, and not merely because we are temporarily unable to answer it. To answer, we would first have to distinguish negation from something else and say that it is this and not that. But making the distinction already uses negation. We try to go behind “not” to see where it comes from, yet every step must be taken by means of “not.” Its source is invisible in principle, as the eye sees everything except itself in the act of seeing.

Now consider reason. A person, like a table or a cup, is made of atoms. Yet a person somehow forms the concept of “not,” while tables and cups do not. A person can say of the facts before him, “It should not be this way”; of an arrangement others give him, “I do not accept it”; of his own conclusion yesterday, “I was wrong.” He can even negate his own existence and think that one day he will no longer be here.

Reason uses “not” completely and without remainder. Thinking distinguishes; judgment affirms or denies. Remove “not,” and reason can do nothing. This is what I mean when I say that reason and negation come from the same source. If the source of negation is not among the things we can know, then the source of reason does not lie entirely among them either.

Only now do we arrive at death.

Death is the greatest “not” a person will encounter. The body will stop. The brain will stop. Everything in me that belongs to the order of things will be negated. I accept all of this without qualification.

The next step, I admit, is not deduced. It is judged. Death itself is an act of negation. Negation can fall upon every thing, but it cannot fall upon its own source. What death takes is what belongs in me to the order of things. The part of me that can say “not” comes from the same place as “not” itself. My judgment is that it will not thereby become nothing at all.

I mark this step separately so that anyone who disagrees can see exactly where our paths divide.

What, then, is there after death? I cannot say. According to my own reasons, I should not be able to say. If the source of negation is unknowable, then what becomes, after leaving the body, of something that comes from the same place is likewise unknowable. I therefore do not describe an afterlife or promise that memory and personality remain unchanged. I say only as much as my reasons allow.

I know that many people who ask this question are really thinking of a particular person. I cannot tell you that you will certainly meet again. I can say only that the person has not simply become nothing.

Someone will naturally object that deep anesthesia extinguishes consciousness and brain injury changes a person. I accept all of this. It shows that the operation of reason in this life depends on the brain. But it concerns operation, not source. Neuroscience can describe with increasing precision what happens in the brain when someone judges or says “not.” Yet every step of that description uses negation: this region activated and that one did not; this signal crossed the threshold and that one did not. It can explain the mechanism of negation with increasing clarity, but it explains “not” by using “not.” It cannot pass behind negation itself.

To be fair to the objection, believing that nothing remains after death is also a judgment. Science can show that consciousness depends on the operation of the brain. To move from this to “when the brain stops, there is nothing at all” requires one further step. That step is the same kind as mine: a judgment made at the boundary of knowledge, only in the opposite direction. I therefore do not regard people who make the other judgment as unreasonable. We stand at the same place and judge differently.

My reasons are close to those of Socrates and Kant. I say that the reasons are close, not that I believe because they did.

Plato’s Phaedo records Socrates’ conversation on the day of his death. Socrates offers several arguments there. One says, roughly, that the soul can know things that are invisible and unchanging, so it belongs to the same kind as they rather than to the kind of the perishable body. My reason resembles this, except that I do not need to posit an invisible realm of Forms. I use something everyone uses every day: “not.”

Kant listed twelve categories of the understanding, and negation is one of them. To call it a category means that it is not learned from experience, but is a condition under which experience is possible for us. Yet Kant himself warned that the immortality of the soul cannot be deduced from “I think.” Such a deduction would carry reason beyond its limits. He therefore placed immortality within practical reason as a postulate that could not be proved but had to be assumed. In the preface to the second edition of the Critique of Pure Reason, he famously said, in effect, that he had to suspend knowledge to make room for faith. My judgment occupies the same place. He reached it through the moral law; I reach it through negation.

After telling a story about the world after death, Socrates added, roughly, that a sensible person would not insist that things were exactly as he had described them. But believing that they were something like that was a risk worth taking, a noble risk. If he were right, he might meet Homer. I love that sentence. It stops where it ought to stop.

So, to return to my friend’s question: “There is nothing after death” is itself a “not,” and the greatest one. The part of me that enables me to say that sentence, I judge and believe, does not lie within the reach of what the sentence can negate.

Part III: Education, Freedom, and Relationships

Education often prescribes a person’s direction in the name of “what is best for you.” Relationships, too, may use concern, competition, or shame to force someone into obedience. Having many options does not necessarily amount to freedom. If changing direction means losing safety, belonging, or the right to be respected, the choice may still be superficial. This part approaches the conditions of freedom through family, school, and public events. It also asks why bystanders protect the powerful and blame victims. A person’s boundaries do not sever relationships with others. They make it possible for relationships to stop erasing the person as their price.

31. What Fear-Based Education Leaves in a Child

My friend’s question seems to carry a wound. Of course, I heard this sentence many times as a child. Many people raised in East Asian environments will probably recognize it: if you do badly on this exam, you will not enter a good high school; without a good high school, you will not enter a good university; without a good job, your whole life will be over.

To be fair, the sentence worked. It frightened me very effectively. But it worked because it eliminated persuasion. Tell a child what abilities mathematics can develop, and the child may answer, “I do not want to study it now. What is the use?” Once life is compressed into a one-way road, however, the child need not understand why something is worth doing. He needs only to fear the consequences of not doing it. I have written before that creating panic is a shortcut to power. The structure here is the same, only moved into the home.

There is one difference between the home and an arena of power. Parents who say these things are often their first victims. The chain linking the college entrance examination, elite schools, employment, and class position appears real to many families, and some of its links do have grave consequences. Parents may not be deliberately frightening their children. Often they become afraid first, take fear for truth, and pass the fear down.

The problem lies in the leap from a real risk to the words “your life is over.” “This examination may affect the range of choices available to you later” might be true. But “therefore, if you fail it, you as a person have no value” does not follow. A limited institutional risk is enlarged into a verdict on the whole person.

High expectations are not themselves the problem. One study followed more than eight hundred Chinese and American middle-school students and distinguished two forms of parental control. Behavioral control, such as setting rules and supervising homework, was associated with better academic performance in both countries. Psychological control, such as humiliation, inducing guilt, and withdrawing affection after poor grades, predicted worse emotional well-being in both. “If you do not do this, your life is over” belongs to the latter category. It does not become harmless because it occurs in East Asia.

What happens after the child grows up? The deepest effect, I think, is not fear of examinations. It is learning to drive oneself with catastrophe.

The parents are no longer there, yet the person continues telling himself: if this project fails, my life is over; if I do not get promoted this time, my life is over; if I stop to rest, I will be left behind. The external threat has disappeared, but the person issuing it now lives inside. Such people may become highly successful and still find security almost impossible. Success does not feel like completion. It is only the temporary silencing of an alarm, before the next one sounds. Many capable people therefore cannot truly rest. To them, rest is not stopping; it is falling backward.

The irony is that this education is intended to make children more competitive, yet often leaves them less willing to try. Failure has never been merely an outcome in their minds. It is a verdict on identity. The safest strategy, then, is to attempt only what they already know how to do successfully.

Some parents will say, “Look, we were right to pressure you. You entered such-and-such university.” This too contains a fallacy. It may be true that coercion made the child study more. But that does not answer another question: could some approach that did not make the child feel for years that “one failure will end everything” also have led him to study seriously? Without making that comparison, a good outcome cannot prove that every threat used in the process was necessary.

Consequences should, of course, be discussed. Compare two statements: “If you do not prepare, you may do poorly this time, and that may affect some of your choices. Let us think together about what you want to do,” and “If you go on like this, your life is over.” Both discuss consequences. The first leaves a person with agency. The second produces a subject who obeys.

Those of us who grew up in such environments should do something as early as possible in adulthood: separate “this undertaking failed” from “I as a person am a failure.” They are not the same judgment. An examination can go badly. A career can collapse. Both deserve to be taken seriously, but neither is important enough to decide whether the person deserves to be treated well. If the inner voice saying “your life is over” has grown loud enough to affect sleep, daily life, or relationships, that is not a failure of willpower. Speaking directly with a professional is a proper response.

32. Why People Use AI to Make Games Instead of Simply Playing Them

The situation is not quite that extreme. People have not stopped playing games. According to this year’s report from the Entertainment Software Association, about two-thirds of Americans play video games for at least an hour each week. Last quarter, users spent a total of 29 billion hours on Roblox. But game-making really has changed. For the first time, making a game has itself become a game that ordinary people can play.

Playing and making used to be separated by a very high wall. Learning to play Mario takes ten minutes. Making a similar small game required first learning programming, engines, art, animation, and sound, and after several months one might still have nothing that ran. Roblox Build now allows someone to describe gameplay in a single sentence and have AI generate the game. Roblox said this month that about nine thousand games made with Build had already been published, and 71 percent of their creators had never used its development tools before. This is not a group of experienced developers switching tools. People who had previously been only players have crossed to the other side of the wall.

Why can “making” be even more fun than “playing”? Because making is itself a creative game. You tell an AI, “Make me a New York where, after it snows, every car turns into a crab.” A few dozen seconds later you enter and run around. The crabs seem too slow, so you say, “Make them faster, add a few police penguins, and cut gravity in half,” then enter again. Challenge, feedback, adjustment, and another challenge: the structure is exactly the same as playing a game, with one additional sensation. The world has changed immediately in response to something I just imagined.

I think the deeper layer lies in the word “game” itself. The philosopher Bernard Suits gave a famous definition: to play a game is voluntarily to overcome unnecessary obstacles. Playing someone else’s game means accepting obstacles and conditions of victory set for you. Making your own means deciding even what the obstacles are and what counts as winning. The pleasure moves from achieving someone else’s end to setting the end itself. When I wrote about Alpha School, I said that it gives children their own pace but not their own ends. AI game-making does the opposite. It gives ordinary people the step of setting the end, even if the end is only a snowy New York filled with crabs.

In the same book, Suits imagined a utopia where every need had been met. What would remain for people to do? His answer was games, because games alone are done for their own sake. AI is making more and more activities unnecessary. The answer sounds less like a joke all the time.

Of course, being able to generate a game is not the same as making one others want to play. AI lowers the cost of implementation. It does not automatically supply delight, pacing, aesthetic judgment, or an understanding of people. Roblox’s recommendation system looks at whether a game retains players for more than twenty-eight days. Over the past year, its ten highest-earning creators each averaged more than $65 million, while the median income among the more than forty thousand creators who shared revenue was only about $1,500. Making a game will become as ordinary as taking a photograph. Making a game someone wants to play for twenty hours will remain difficult, just as smartphones did not eliminate film directors.

The most interesting development, then, is not that everyone can become a developer. It is this: when anyone can build a world in a few dozen seconds, the scarce thing is no longer the ability to make something, but knowing what you want to exist. AI tears down the wall. It cannot answer that question, because the question properly belongs to human beings.

33. Why Variants of Egoism Are Flourishing

This question is sharp because contemporary egoism rarely says directly, “Care only about yourself.” It dresses itself up as self-help and quietly adopts a respectable vocabulary: boundaries, avoiding internal friction, personal growth, emotional value, personal branding, becoming your best self.

Each term may be perfectly reasonable on its own. Put them together, however, and they can become a closed system: “If I succeed, it is because I take responsibility for myself. If I fail, it is because I have not optimized enough. Other people’s needs violate my boundaries. The demands of a community restrict my freedom.” By this point, egoism no longer has to admit that it is egoism. It calls itself growing up and becoming mature.

Why does it spread so easily? Half the answer is that people do not feel secure. AI may replace their work. Housing grows more expensive. Retirement is uncertain. Institutions may not be reliable. Naturally, people protect themselves first. In the United States General Social Survey, the proportion of respondents who believed that most people could be trusted fell from more than 40 percent in the early 1970s to a little over 30 percent. Problems that once belonged partly to institutions have been translated into the individual’s responsibility to manage himself. Lose your job? Improve your skills. Anxious? Meditate. Lonely? Practice social skills. Each suggestion may be useful. The problem is that once they are available, the question “Is there something wrong with the system?” never has to appear.

The other half of the answer is that this sells well. Traditional consumerism tells you that you still lack an object and should hurry to buy it. Self-optimization tells you that you are the deficient object. Once you buy a car, the purchase is complete. You can go on upgrading yourself forever.

From the perspective of SAE, these variants share a name: colonization. SAE follows Kant, whose formula is clear: whether in your own person or the person of anyone else, humanity must always be treated at the same time as an end, never merely as a means. The sentence points in two directions. SAE’s opposition to colonization therefore also points outward and inward.

Outwardly, we turn other people into suppliers. Friends must provide emotional value. Acquaintances must be useful. “Low-energy people” must be avoided. Others still appear to be respected, but they have really been packaged as a more advanced service: you are not my instrument; you are my provider of emotional value. Real relationships necessarily contain times that cannot be monetized at once. A sick friend cannot give you positive energy. A child may consume your time for years. If every relationship must constantly prove itself worthy of my investment, the person on the other side, as a person, has already disappeared.

Inwardly, we turn ourselves into assets. In my essay on cultural deserts, I listed examples: reading to improve cognition, making friends to build a network, resting to restore performance. That essay described the logic of a place spilling over into human beings. Today I would add that the person applying the logic is often the self. This is colonization in its most hidden form. By the end, egoism does not even benefit the self. It serves a balance sheet, while the real person becomes an instrument for making the sheet look good.

But opposition to egoism easily falls into another trap. A family demands sacrifice. An organization demands devotion. Other people demand endless understanding, then tell you not to be selfish. This too is colonization, with a different colonizer. SAE therefore does not answer by retreating from “live for yourself” to “live for others.” It insists that three statements hold simultaneously: I may not treat others as means; others may not treat me as a means; I may not treat myself as a means.

The word “boundary” also needs clarification here. A genuine boundary says: you may not decide my ends for me. A distorted boundary says: your needs may place no demand at all upon me. The first protects the person as an agent. The second cancels the existence of others.

The parts of a person’s life that yield no return, accompanying a sick friend, days that produce nothing, an afternoon spent for no purpose at all, are not inefficiencies to optimize away. They are precisely where a person cannot be colonized. Egoism needs to answer only, “How do I get what I want?” Opposition to colonization must also answer: when both you and I are ends, and neither can erase the other, how do we live together?

The problem with these variants of egoism is therefore not that they love the self too much, but that they love the wrong object. They love the grasping little “I.” The genuine self, like the genuine other, is not that little “I.”

34. Why Farewell to Revolution Has Endured

This is an unusual question. The book is in fact a record of conversations between two people. At the beginning of 1992, Li Zehou arrived in Colorado. He was at Colorado College and Liu Zaifu at the University of Colorado, a two-hour drive apart. They often met to talk about twentieth-century China. The manuscript was originally titled Looking Back at Twentieth-Century China. Only later were the four Chinese characters meaning “farewell to revolution” added in front as the main title. It was published in Hong Kong in 1995. Those four characters were not a neutral academic title. They were a demand that readers take a position. Li Zehou died in Colorado in 2021. Liu Zaifu died in Hangzhou this May at the age of eighty-five. Overseas newspaper obituaries still put the phrase “reform, not revolution” in their headlines. Thirty years after the book appeared, with both authors gone, people still read it, denounce it, and cite it.

Let us begin with what the book said, and what it did not say. Liu Zaifu later explained repeatedly that what they were saying farewell to was not the historical justice of revolution, but the belief that revolution should be history’s only choice. They opposed overthrowing a regime through mass violence, one class overthrowing another. They chose reconciliation between classes. In Liu’s words, reconciliation did not exclude struggle, but it rejected large-scale bloodshed. Li Zehou took the counterfactual one step further, suggesting that China might have fared better had it followed the path of constitutional reform proposed by Kang Youwei and Liang Qichao.

I think the surface reason for the book’s endurance is that it confronts a deeply rooted habit: automatically treating revolution as progress and reform as weakness. In twentieth-century China, revolution was a glorious word, while reform sounded almost like surrender. The book reverses the ranking and asks what violent rupture actually cost, and whether gradual change might be more difficult and more worthwhile.

At a deeper level, it asks whether an end can redeem its means. For the sake of a better future, may we first treat a group of people living today as sacrificial material? The question belongs neither only to China nor only to the twentieth century. In my own philosophical framework, this is the colonization of people in the present in the name of the future. Future people are abstract. People alive today are concrete. The abstract people have not yet been born and cannot themselves suffer what is done.

Then comes the crucial point: the book was attacked from both sides. After publication, it faced a nationwide campaign of criticism in mainland China. As late as 2021, an article on the website of Qiushi called the “farewell to revolution theory” the most typical form of historical nihilism. But liberals rejected it as well. Rong Jian argued that reform is active, chosen, and designed, whereas revolution is passive, unchosen, and unpredictable. Once the road of reform is blocked, one cannot bid revolution farewell simply by wishing to do so. Qin Hui noted that reform does not necessarily shed less blood. The Meiji Restoration cost far more lives than the 1911 Revolution; what matters is that rulers feel pressure to change. Zhu Xueqin put it more bluntly: revolution is not something one can abandon at will. When both the official side and the liberal side use a book as a target, it is unlikely to be forgotten.

The most interesting question, I think, is why the authorities had to criticize it. In principle, “reform, not revolution” is precisely what a government would most want the public to internalize, and in practice what it fears most is a “color revolution.” Yet its narrative of legitimacy comes from revolution. To bid farewell to revolution would loosen its own account of where it came from. It cannot abandon revolution in narrative, while fearing nothing more in reality. The book presses exactly on this crack. It can neither be absorbed nor discarded.

Some critics, however, identified the question that most needs to be asked of the book: to whom is “farewell to revolution” addressed? Revolutions are rarely chosen by the weak. More often they arrive when channels of reform have been closed. Whether a society can bid farewell to revolution therefore does not depend on people who want one, but on those whose hands hold the switch for reform. The book was also bidding farewell to Mao Zedong’s philosophy of struggle. Liu Zaifu considered himself throughout his life a reformer within the system. The authors advised both sides, but only the side holding the switch could decide whether the advice was possible.

Finally, let me offer my own reading. I do not think the true dividing line lies between revolution and reform. Revolutions that use elimination as their method rest on an image: remove a particular group, landlords, the bourgeoisie, rightists, the far left, environmentalists, conservatives, and a new world without contradictions will appear. But the disagreements among people, the parts that cannot be fully articulated or absorbed into any plan, what I have consistently called the remainder, cannot be eliminated. Remove one group and the next emerges on its own, so the purging must continue forever. Reform can reach the same point when “take it slowly” becomes “never,” and every disagreement is labeled instability.

To judge a change, then, we should not ask whether it calls itself revolution or reform. We should ask whether, after it is complete, people can still say no.

This book endures not because it supplied the answer, but precisely because it did not fix the answer in place. It compels every generation to answer the same question anew: when we believe we are creating a better world, are the people before us who disagree still ends in themselves?

35. Why Freedom Depends on Both Being Able Not to Act and Being Unable Not to Act

I should admit that the first half of this idea is borrowed from Kant. Only the second half is mine.

Begin with “what one is able to do.” This is both ability and menu: able to write, start a business, resign, move abroad. The longer the menu, the freer someone appears. But the length of the menu tells us only how many options exist, not who is choosing. A company gives an employee a hundred projects to choose from, but he may not refuse them all. He has many options and little freedom. A society tells you that you may become a doctor, lawyer, engineer, or entrepreneur, but that you may not refuse to compete, stop, or choose a low-income way of life you yourself endorse. It offers many actions but little freedom.

The first test, then, is whether one is able not to act. The difference between a person and a program is not the ability to choose among A, B, and C; a program can do that too. It is the ability to say, “None of them.” I often say that people, like tables and cups, are made of atoms, yet can say “no.” Freedom begins with that “no.”

And it must be a “no” one can truly utter. If refusal is allowed in name but costs someone his entire life, then he is effectively unable to refuse. The United States has an expression, “job lock.” Many people are afraid to resign not because they want to stay, but because health insurance is tied to employment. I wrote earlier that East Asian parents often tell children, “If you do not do this, your life is over.” What the sentence truly takes away is the child’s ability not to act. A society’s freedom is measured not only by how many things it permits people to do, but by the cost of saying no.

Yet “no” alone is insufficient. Someone who listens to no one and does whatever he wants may appear maximally free, but often he is simply dragged along by desires, fears, and stimuli. He says no today and no tomorrow, until nothing remains that is truly his own.

We therefore need a second test: can a person be unable not to act? The phrase sounds paradoxical. We usually think freedom means the absence of anything one “must” do. But there are two kinds of necessity. In one, others force me to act. In the other, after thinking the matter through, I discover that I cannot act otherwise. Someone could betray a friend and suffer no punishment, yet says, “I cannot do that.” What stops him is neither the police nor calculation, but a principle he has made his own.

Before Socrates died, his friends had arranged an escape. He could leave, and he could also refuse. In the end he stayed, because he believed he could not repay injustice with injustice. All three elements appear in one scene.

Kant gave freedom two sides. Roughly speaking, the negative side is that the will is not governed by external causes. The positive side is that the will gives itself the law, which he called autonomy. Being able not to act is the first side. Being unable not to act is the second. Freedom is not the absence of constraint. It is a question of where the constraint comes from.

The hardest part is that the two kinds of necessity look identical from the outside. “I must succeed.” “I cannot stop.” “I must constantly prove myself.” These may express commitments a person has made his own. They may also be an external measure long ago moved inside his mind, so that whoever once enforced it no longer needs to be present. I call the latter colonization. In its most complete form, the colonized person believes that the voice is his own.

How can we distinguish them? My method is to ask whether the “must” has passed through “no.” A necessity to which you have never dared say no is not your necessity. Only when you truly could refuse, have considered and weighed the choice, and still say, “I cannot fail to do this,” is the necessity self-legislated. The three elements therefore have an order. Being able not to act is the gate through which one reaches being unable not to act.

One qualification must be added. My “must” cannot cancel someone else’s “no.” If a person believes he bears historical necessity and everyone else must clear the road, that is not autonomy. It is the imposition of his law on others, still colonization, only pointed in another direction.

What you are able to do may be given to you by others. What you are able not to do may be taken away by others. What you are unable not to do can be given to you only by yourself.

36. Why Excessive “Freedom” Can Lead to Servitude

I put “freedom” in quotation marks because I have never said this, and I do not agree with the statement as it stands. Plato said it first. In Book VIII of the Republic, he suggested that excessive freedom, whether for an individual or a city, seems to change into nothing but excessive slavery. He told the story of a city where everyone did as he pleased and accepted no authority. Disorder grew until people elevated a protector, who step by step became a tyrant.

But let me first explain the quotation marks. In the previous essay, I said that freedom is measured not by what one can do, but by what one can refrain from doing and what one cannot refrain from doing. In “excessive freedom,” only the first element is ever excessive. The ability to do is pushed to the limit and devours the other two. What leads to servitude is not freedom, but one component of freedom. I can see at least four routes by which it devours the others.

First, within the self, “being able to do” consumes “being able not to do.” Part IV of Spinoza’s Ethics is titled “Of Human Bondage.” By bondage he means that people cannot govern their affects. They know what is better, yet are dragged toward what is worse. Someone finally has a weekend entirely to himself and plans to watch ten minutes of video. He looks up and three hours are gone. No one commanded him, but it is hard to say that he chose those three hours. Nor is this entirely a problem of individual will. Autoplay and infinite scrolling are designed by whole teams optimizing the time you remain on a product. Put bluntly, their daily work is to make your capacity not to act smaller.

Second, across time, one exercise of “being able to do” can cancel every future “being able not to do.” In Chapter V of On Liberty, Mill offered an extreme example: someone voluntarily sells himself into slavery. At the moment he signs, he is free, but the content of his choice is that he can never choose again. Mill concluded that the principle of freedom cannot require the freedom to become unfree. Permitting someone to surrender his freedom is not freedom. The principle is close at hand. California law generally refuses to enforce noncompete agreements; even if you signed one, it is invalid. “You signed voluntarily at the time” explains the beginning of an arrangement. It does not establish the right to govern forever. Promising one particular act is not the same as surrendering, all at once, one’s future right to say no.

Third, when people lack an “unable not to act” of their own, they borrow someone else’s. Why did Plato’s city elevate a tyrant? Its people retained only the ability to do whatever they pleased and never learned to give themselves a law. After enough disorder, fear, and exhaustion, someone appears and says, “Do as I say and you will no longer need to be afraid.” During the Second World War, Erich Fromm wrote Escape from Freedom about the same development. People free themselves from old bonds but fail to develop a direction of their own. When loneliness and anxiety descend, they willingly surrender themselves. In the previous essay, I said that only you can give yourself what you cannot refrain from doing. Here I add: if you do not give it to yourself, someone else will be delighted to give it to you.

Fourth, between people, my “ability to do” consumes your “ability not to do.” In “Two Concepts of Liberty,” Isaiah Berlin quoted the saying that freedom for the pike is death for the minnows. In the 1905 case Lochner v. New York, the United States Supreme Court invoked freedom of contract to strike down a state law limiting bakers to ten hours of work a day. Freedom from restrictions on the employer’s terms was the employer’s liberty. For a baker unable to find other work, the other side of that liberty was an inability to refuse a working day longer than ten hours. Often the problem is not that everyone is too free. It is that someone has expanded the power to decide for himself into power over others.

Finally comes the most dangerous reversal. Someone may answer: “Your desires have blinded you. You do not know what you truly want, so let me decide for you. That is real freedom.” Rousseau said that anyone who refused to obey the general will would be compelled by the whole body to do so, which meant that he would be “forced to be free.” This is precisely what Berlin distrusted most. The test remains the same: after someone has decided for you, may you reject his interpretation? If not, he has not given you freedom. He has given you another master with an attractive name.

I therefore do not much like the phrase “excessive freedom.” There cannot be too much freedom. There can only be too much ability to do. What leads to servitude is never too much freedom, but too little “no.”

37. Why a University Would Conceal Sexual Assault in a Fraternity

This is a painful question, and many people are now concerned about the case as news reports continue to follow it.

Let us first clarify what is known. According to a female student, she was drugged at the Chi Phi fraternity and sexually assaulted by several members in October 2024. In November, the university received a report, suspended the chapter, placed seven students on interim suspension, and issued a statement. An internal investigation began in January 2025. The county district attorney later concluded that the evidence was insufficient to bring criminal charges.

This month, she sued Cornell, the fraternity, and the seven students. According to her attorney, only two were ultimately expelled. The complaint says that the others received sanctions including suspension, training, and writing essays. On September 21, Cornell issued a statement saying that it had investigated and handled the matter according to university rules, that the chapter remained barred from campus, and that federal privacy law prevented disclosure of individual student discipline. It would address specific allegations in the litigation.

Strictly speaking, Cornell did not do nothing. No court has ruled that it deliberately covered up the case. Nor have the seven individuals been convicted in a criminal proceeding. But I completely understand why people feel that the university is concealing something. The reason lies not in any particular person’s intentions, but in the position the institution occupies.

Cornell simultaneously plays three roles: investigator, adjudicator, and defendant. One reason the student is suing the university is alleged negligent supervision. The institution must discover the truth and protect students while defending itself in court, and every statement it makes may become evidence for the other side. The legally rational response is therefore to say as little as possible. “We will respond through the legal process” is impeccable advice from counsel. The problem is that, at that point, the university’s interests and the student’s interests no longer lie on the same side.

Now consider the reason Cornell offers: privacy law. Privacy law genuinely protects students, but it is not as airtight as the university suggests. Federal law permits a school to disclose the final results of disciplinary proceedings to a victim. In cases involving violence or sexual assault, when a student is found responsible, it may even permit disclosure of the outcome to the public. Another federal law goes further, requiring a school to notify both sides in writing of the outcome, sanction, and rationale. Cornell’s student newspaper asked in an editorial why the complainant had not been informed of the investigative outcome, and why the interim suspensions had been lifted. If she truly did not receive the required notice, privacy law cannot explain that. Privacy can block the public. It cannot block a party to the same case.

The essays must also be discussed as allegations in the complaint. But if true, they reveal more than silence. A reflective essay is ordinarily a university’s response to plagiarism or a violation of dormitory rules. Using it to answer an allegation of drugging and sexual assault translates a potentially criminal act into the language of education, as though it were something from which the accused could learn and improve. The student newspaper called this a diminishment of the gravity of the offense. I strongly agree.

Another layer is easily overlooked. The district attorney declined prosecution because criminal conviction requires proof beyond a reasonable doubt, a high threshold. Internal proceedings usually apply a much lower evidentiary standard. For her, the university was therefore almost the only place where the case might receive an answer. If that place is a black box, she is left only with a lawsuit, and on the other side of the lawsuit stands the university itself. The psychologist Jennifer Freyd calls this kind of harm institutional betrayal. The injury comes not only from the original event, but from the place you trusted and expected to protect you protecting itself first when you most needed it.

I often say that persons are ends in themselves. A university exists for students; students do not exist for the university’s reputation. Once an institution treats its reputation and legal risk as ends, a student becomes a risk item to manage and control, someone who must say as little as possible. No one has to be evil. Everyone need only do the safest thing according to his job. Transparency also protects the accused. When the outcome is invisible, someone falsely accused cannot clear his name either.

Protecting students from public harm and protecting a university from questions about its responsibility are two different things. The first is privacy. The second has merely borrowed privacy’s name.

38. Why People Blame Victims and Protect Perpetrators

The phenomenon my friend asks about is not new. Whenever a case like this occurs, two kinds of voices appear. One blames the victim for failing to protect herself. The other invokes privacy on behalf of the accused. The two must be examined separately. The first is almost always wrong. The second is sometimes reasonable, but is often used in the wrong place.

Begin with blaming the victim. Psychology speaks of belief in a just world: people want to believe that what happens to someone broadly corresponds to what he has done. Acknowledging that someone who did nothing wrong can still be gravely harmed makes bystanders anxious. They therefore search for a reason to separate themselves from the victim: if she had not gone to that party or drunk so much, it would not have happened. I would never do that, so I am safe. Blaming her is a way of buying a sense of security. But advising people how to reduce risk and deciding who bears responsibility are two different questions. It is sensible to remind people to lock their doors. An unlocked door does not vindicate the thief.

Sexual assault is also governed by a special double standard. Two German psychologists conducted an experiment in which participants read descriptions of rape and robbery. Roughly speaking, the more familiar victim and perpetrator were, the more the victim was blamed in the rape case. The same relationship produced no such effect in the robbery case. Without anyone noticing, the question changes from whether she consented on this occasion to whether she is a “good girl” worthy of protection. In her statement, Chanel Miller, the victim in the Stanford case, listed questions the defense had asked: How much did you drink? What kind of cup did you use? Did you attend fraternity parties? Was your relationship with your boyfriend serious? Someone had been violated, yet in court it was her way of spending a weekend that was examined again and again. “I disapprove of your way of life” is quietly substituted for “you do not deserve to remain inviolable.”

Now consider the other side. The Cornell philosopher Kate Manne coined the term himpathy for sympathy that drains away from a female victim and flows toward a male perpetrator. She developed the idea precisely through the Turner case. The father’s reference to “twenty minutes,” and the judge’s worry about the “severe impact” on Turner, exemplified this flow. Why does sympathy move in that direction? Many people look at the young man and see their son, classmate, friend, or younger self. “He is usually a good person.” “He gets good grades.” “He has a promising future.” Even if all of this is true, none of it answers whether a particular event occurred on a particular day.

There is a more concealed reversal that Jennifer Freyd calls DARVO: deny, attack the person making the accusation, then reverse the positions of victim and offender. The discussion shifts from “Did this occur?” to “She is destroying a person’s life.” Of course, denying an accusation, hiring counsel, and offering contrary evidence do not in themselves constitute such a reversal. The concept helps us ask whether the discussion has changed subjects; it does not determine guilt.

Privacy deserves to be taken seriously. Before the facts are established, no one should publish someone’s home address or harass his family. Even a conviction does not give everyone unlimited license to excavate his life. We can test the sincerity of a claim about privacy by asking whether it protects both sides. If someone insists that his life must not be destroyed while exposing every detail of how much she drank and whom she dated, privacy is not the principle being protected. A particular side is. We must also ask what privacy is shielding. Refusing to disclose unnecessary personal information is one thing. Forbidding investigation, preventing an account of the process, and above all blocking questions about institutional responsibility are another.

Underneath both voices lies the same question: whose future is treated as valuable without proof, and whose suffering must first prove that it deserves to matter? I often say that persons are ends in themselves. The statement has no conditions. Not only well-behaved people are ends. A person who drank, went to a party, or knew the other person is an end too. Conversely, the accused is also an end, so he is entitled to due process and the presumption of innocence.

Taking a report seriously does not mean declaring guilt before trial. Insisting on the presumption of innocence does not mean presuming the reporter guilty first. No one should have to prove that she is a perfect person before she deserves not to be violated.

Part IV: Power and Public Life

Power does not exist only in government or legal codes. It also appears in who may define morality, who decides what knowledge deserves circulation, and how capital, universities, and technology companies shape public discussion. An institution may claim to protect everyone while requiring particular people to bear silence in its concrete operation. The essays in this part bring the principle that persons are ends in themselves into public life. This does not turn politics into an abstract principle. It asks who can still speak, refuse, and change a shared direction when the principle is put into practice.

39. Why People Are Left Only with Suffering When Those in Power Speak of Morality

Beautifully asked. The key lies in two words: the first half says “can,” the second “can only.” The first opens a door. The second closes one.

First consider rulers speaking of law. Law is written and public, and at least in its text it also governs those who make it. The moment those in power say, “We act according to law,” they acknowledge a standard outside their own hands, under which they too must stand. Above law lies another open level: whether the law itself is just. The people can therefore continue asking, “On what grounds was this law made? Does it apply equally to everyone?” To ask that is to speak of morality.

Writing from the Birmingham jail, Martin Luther King Jr. responded to people who invoked injunctions and regulations against him. Human beings, he wrote, have a moral responsibility to obey just laws and a moral responsibility to disobey unjust ones. He cited Augustine: an unjust law is no law at all. King could say this precisely because the other side had first taken its stand at the level of law. The level above remained open for him to occupy.

Now consider rulers speaking of morality: “Consider the larger picture.” “Sacrifice selflessly.” “Show gratitude.” “Make the nation great again.” When such phrases come from those who hold power, the highest level has already been occupied. Dissent no longer appears as argument. It becomes evidence that one is immoral, selfish, indifferent to the common good, or ungrateful. There is no higher level above morality on which to stand, and law is often pliable at such moments. What remains in the hands of ordinary people? Only their wounds. “I truly cannot go on.” “No one can care for the elderly person in my family.” “My child cannot attend school.” These are no longer claims of right but pleas for mercy. In retreating from “You have no right to do this” to “Please take pity on us,” a citizen becomes merely a resident begging for consideration.

Mencius saw the scene more than two thousand years ago. He told King Xuan of Qi that when ordinary people were struggling simply to survive, how could they have time for ritual and righteousness? The ruler speaks endlessly of moral propriety; the people think only of escape from death. They are not speaking the same language.

Why is morality in the hands of those in power more dangerous than law? Law has boundaries; morality does not. Law governs conduct, and what is not prohibited is permitted. Morality can reach into what you think: are you grateful enough? Are you willing to sacrifice? If morality is not absolute, the refrain goes, you are absolutely immoral.

Law at least nominally governs upward as well as downward. The morality spoken by rulers usually governs only downward. More deeply, it enters the mind. When someone is mistreated, his first thought is not “this is unfair,” but “am I selfish to complain? Am I causing trouble for the larger picture?” I call this colonization. A measure pressed inward from outside becomes the measure by which people judge themselves.

The problem is not morality itself. Morality was meant to measure power, not to become power’s measure of people. In the hands of the people, it constrains power. Taken up by power, it constrains the people. Nor is law automatically good. Unjust laws abound. That is precisely why the moral level must remain open, so people can stand there and question them.

One final step remains unspoken. When even suffering is labeled negativity, performative misery, or troublemaking, so that pain itself cannot be voiced, the people are left only with silence.

The statement is therefore a thermometer. Listen to what those in power discuss. If they speak of law, some room remains. If they speak of morality, be cautious. When they say the people should stop talking about suffering and look at the nation’s achievements, very little room is left. The health of a society is measured not by how nobly its rulers speak, but by the language in which its people are still permitted to answer them.

40. Why Smit Machchhar Is a True Hero

This truly deserves to be called heroism, because so many lives were saved. On September 30, flydubai Flight FZ1073 was traveling from Dubai to Tel Aviv when something happened in the cockpit. Flight-tracking data showed that the plane dropped more than fourteen thousand feet in less than thirty seconds. Its transponder first transmitted an emergency code, then several minutes later a code for unlawful interference. It finally diverted to Tabuk in Saudi Arabia and landed safely. Both pilots were injured and hospitalized.

Up to this point, the event still looked like an aviation accident. What drew special attention was the airline’s confirmation that a conflict had occurred in the cockpit. It said that several off-duty members of its crew aboard the flight helped bring the aircraft under control and complete the diversion. The cause and motive were not announced, and the United Arab Emirates’ civil aviation authority was investigating.

The main account of what happened currently comes from the Israeli Prime Minister’s Office. It says that one pilot stabbed the other and intended to crash the aircraft. The injured captain, Machchhar, fought back and opened the cockpit door, allowing people outside to enter and restrain the assailant. On this account, he saved all 174 people aboard.

Why call him a true hero? A hero is not someone who feels no fear or cannot be wounded. He is someone who, after the wound has already fallen on him, still does something that changes the situation. Machchhar’s crucial action was not defeating someone alone. It was opening the door.

That door has a history. After September 11, cockpit doors were reinforced and locked from inside to prevent intruders from entering. In the 2015 Germanwings disaster, the danger was precisely that the threat was already behind the door. The copilot locked the captain outside. No matter how people tried to break in, the door would not open, and 150 people died. The same door that stops an external threat can also keep external rescuers out. When danger is inside, only a person inside can open it. This time, the person inside had already been stabbed, but he opened it anyway.

He was not the only hero in the story. Someone rushed in to restrain the other pilot. Several off-duty crew members landed the plane safely. Machchhar did not save everyone single-handedly. When he was nearly unable to continue, he made it possible for others to carry on the rescue. Acknowledging their contribution does not diminish his courage at all.

I wrote earlier that only you can give yourself what you are unable not to do. No one could order a man who had been stabbed to open that door, and no one would have blamed him if he had collapsed without doing anything more. He did it. That “must” was his own.

One further point is worth making. He is Indian. The passengers included Israelis and people of other nationalities. These identities may form part of the background, but they are not the source of his heroism. The test is simple: change the captain’s nationality and replace the passengers with people of other identities. Would we give the same action the same respect? If the answer changes, what we praise is not simply the saving of lives.

A hero is not someone who never falls. He is someone who, after falling, still goes to open the door.

41. Why a True Philosopher-King Would Relinquish Power

I did say this, but the philosopher-king is Plato’s term, not mine. In the Republic, the philosopher-king never wants to rule. He would rather remain outside the cave, looking upon the true light, but is required to return and take his turn governing. Plato also made an interesting observation: the city whose rulers are least eager to rule will be governed best. Yet Plato wanted philosophers not to crave power. He did not ask them to relinquish it. For him, returning to the cave was a duty.

More than two thousand years later, Kant gave another answer. In Perpetual Peace, he wrote that kings becoming philosophers, or philosophers kings, was neither to be expected nor desired, because possession of power inevitably corrupts the free judgment of reason. This is a grave claim. It does not say that bad people become worse when they gain power. It says that however good a person may be, once he holds power, his judgment is no longer free. Power gradually removes the people around him who can correct him. Whatever he says receives applause; when he is wrong, no one dares tell him. A true philosopher knows better than anyone that his judgment has blind spots. However intelligent he may be, some things remain invisible to him. His first reason for relinquishing power is that he does not entirely trust himself.

The second reason lies in the words “everyone objects.” Everyone begs him to stay. This appears to be the greatest possible confidence in him, but may in fact reveal the deepest dependence. In “What Is Enlightenment?” Kant wrote that immaturity is comfortable. If someone else reads for me and someone else judges for me, I need not trouble myself. Benevolent guardians place people in a walking frame and make them feel they will fall if they leave it. People try to escape a bad ruler. They cannot bear to lose a good one. The better he is, the less willing they are to learn to walk by themselves.

A ruler who truly treats persons as ends in themselves therefore measures success not by how much everyone needs him, but by how quickly they can cease needing him. The same is true of a good teacher. The final aim of teaching is to make the teacher unnecessary. I call this cultivation, in contrast to colonization. Colonization carves a person into the shape you want. Cultivation lets the person grow into himself. A philosopher-king whom no one can do without is actually engaged in colonization.

Nor can today’s unanimous consent bind tomorrow. What of people born later? What if those who approve today change their minds tomorrow? More importantly, when everyone agrees with one person, it is often not because everyone thinks alike, but because dissent has nowhere left to exist. Unanimity itself deserves suspicion.

History does offer someone who acted this way. In 1783, after the War of Independence, George Washington returned his military commission and went home to his estate. Before that, an officer had written urging him to become king, and Washington rejected the suggestion sternly. King George III is said to have remarked, on hearing that Washington intended to surrender power, that if he truly did so he would be the greatest man in the world. In 1796, many hoped Washington would continue as president. He left after two terms. The two-term custom lasted until Franklin Roosevelt broke it and was later written into the Constitution. Washington may not have been Plato’s philosopher-king, but he understood something important: a country cannot be dependable if it depends on one person.

The point must also be reversed. Persons are ends in themselves, and this includes the ruler. A community cannot demand that someone remain forever in a position merely because he is useful. That too is treating him as an instrument.

Of course, relinquishing power does not mean abandoning responsibility. If someone keeps every decision in his own hands, never gives others an opportunity to practice, then suddenly says, “You should govern yourselves now; I am leaving,” he has not transferred power. He has deserted his post. A genuine transfer places information, abilities, and opportunities to decide in other people’s hands before the leader leaves. Once gone, he does not remain behind the scenes granting or withholding approval.

I have written about why we always wait for great people. This essay is the other side. A genuinely great person is precisely someone who refuses to be awaited. The greatest moment of a philosopher-king is not when everyone begs him to remain. It is when, after he has left, everyone discovers that they can manage too.

42. Why Disasters Give Rise to Great Literature and Philosophy

Chinese thinkers considered this question very early. In his “Letter in Reply to Ren An,” Sima Qian offered a list. King Wen was imprisoned and elaborated the Book of Changes. Confucius was in distress and composed the Spring and Autumn Annals. Qu Yuan was exiled and wrote “Encountering Sorrow.” These people, Sima Qian said, carried frustrations in their hearts and found the road blocked, so they recounted the past and left their work to those who came after. When he wrote the letter, Sima Qian himself had recently suffered castration and was writing the Records of the Grand Historian. More than a thousand years later, Zhao Yi inscribed a more famous line on a poem by Yuan Haowen: “A nation’s misfortune is a poet’s good fortune; amid the world’s upheavals, the lines grow accomplished.”

There is something real here if we dig deeply enough. In peaceful times, an explanatory system can contain life: effort brings reward and obedience to rules brings protection. People need not question the assumptions of the system. I often say that constructions cannot be closed. Every construction conceals something, though in ordinary times it conceals it well enough to remain invisible. Prolonged disaster tears the covering away. The old explanations cannot contain what lies before us. Everything that never fit inside, human dignity, responsibility, what makes life worth continuing, is suddenly exposed. Hannah Arendt observed, roughly, that twentieth-century totalitarianism had exploded our categories of political thought and standards of moral judgment. The work of the latter half of her life was to think again amid the ruins.

The word “prolonged” matters. In a single disaster, people remember what happened on a particular day. A long disaster becomes the structure of the days themselves. In “The Stone Moat Village,” Du Fu did not record the total casualties of war. He wrote of one household. Three sons had gone to defend a city; two were already dead. The old woman was seized during the night to cook at the army camp. At dawn, “I set out again on the road, alone bidding farewell to the old man.” With the single word “alone,” one more person is gone from the family. Literature restores concrete persons to abstract numbers. Philosophy then asks: when we speak of “cost” and “necessity,” what have we counted, and what have we left out?

The Russian poet Anna Akhmatova also gave a one-word answer. During the Great Purge, her son was arrested, and for seventeen months she waited in line outside a prison in Leningrad. One day, a woman behind her, her lips blue from the cold, leaned close and asked, “Can you describe this?” Akhmatova answered, “I can.” The result was the long narrative cycle Requiem, now known around the world. The woman could not speak it herself. She entrusted it to someone who could. Du Fu’s “alone” and Akhmatova’s “can,” more than a thousand years apart, perform the same act: they speak for those who cannot.

But a long distance separates suffering from the creation of a great work. Pain does not automatically supply language, structure, and powers of observation. Du Fu was formed by the culture of the High Tang. The age of chaos changed what he wrote, but does not explain why he could write it in that way. Primo Levi returned to Turin from Auschwitz before he began writing If This Is a Man. The first edition appeared in 1947, and he revised it for a new edition eleven years later. A person who writes must first survive, then have time, ability, and someone willing to help publish the work.

We see only those who left works behind. Du Fu and Levi appear on the list. Outside it stand the illiterate, those whose manuscripts were burned, and those who did not live until the day they could write. Finding several great writers in a disaster does not prove that disaster creates more great writers. The same war may have contained several other Du Fus who died on the road without leaving a poem. Even a work that is written may not survive. Akhmatova’s friend Osip Mandelstam was arrested for a poem satirizing Stalin. Several years later he was arrested again and died in a transit camp on the way to a labor camp. His poems reached us because, over decades, his wife memorized them one by one. Sometimes a work survives because another person remembers it on its behalf.

Ouyang Xiu proposed that “poetry becomes accomplished after hardship.” But the line that follows is more profound and more worth remembering: it is not poetry that impoverishes people; rather, those who have suffered deprivation become accomplished poets. Writing does not bring misfortune upon their lives. After encountering hardship and obstruction in reality, people may become capable of writing truly refined and moving poetry.

Europeans felt the same unease in the face of their greatest disaster. After the Second World War, Adorno made the severe statement that writing poetry after Auschwitz was barbaric. More than a decade later, in Negative Dialectics, he revised it. Continuous suffering, he wrote in effect, has the right to expression, just as the tortured person has the right to scream, so it may have been wrong to say that no poetry could be written after Auschwitz. He withdrew the prohibition, not the unease. Poetry may be written because sufferers have the right to scream, not because suffering produces good poetry. Zhao Yi’s line about the nation’s misfortune being the poet’s good fortune is, in the end, a bystander’s sigh. How could the tortured person experience it as good fortune?

I am therefore always cautious about statements that suffering creates greatness. They can too easily become the injunction to be grateful for suffering. Someone may look back and say, “That period taught me a great deal.” That belongs to him, and he has the right to say it. But when others use the claim to advise someone currently suffering, or when those in power use it to demonstrate that pain has value, suffering becomes raw material for producing greatness, and the sufferer becomes a means. Persons are ends in themselves, including persons who suffer.

Greatness belongs to the person who, amid disaster, continues to see, remember, and write. It does not belong to the disaster. Suffering is not the cause of greatness. It is what greatness has been forced to face.

43. Why Leviathan Remains a Classic

I apologize for taking so long to answer this question. The preceding essay considered why prolonged disasters produce great thought, and Hobbes is one of the starkest examples.

He was born in 1588. By his own account, his mother went into premature labor after hearing that the Spanish Armada was approaching, so he and fear were born as twins. On the eve of the English Civil War, he fled to Paris. In exile he watched a king beheaded and a country torn apart, then published Leviathan in 1651. Its title comes from the sea monster in the Book of Job. The frontispiece depicts a crowned giant, sword in one hand and a bishop’s crozier in the other. Look closely, and the giant’s body is composed of hundreds of little people.

The first reason the book became a classic is that Hobbes treated the state as something requiring explanation. Before him, power came from God or from ancestry, and obedience needed no justification. Hobbes insisted on asking: why should people who originally live their own separate lives obey a power able to command them? His answer relied on neither God nor bloodline, but began from each individual. He called the state an “artificial man.” Human beings were both its material and its makers. From this point, political philosophy truly began to assume its modern form.

Second, he saw human beings accurately. He did not say human nature was evil. He said people were broadly equal: even the strongest feared a secret attack, and the weakest could join with others to kill the strongest. Everyone fears others and knows that others fear him. People carry weapons when they travel, lock their doors when they sleep, and lock chests even inside their homes. Are they not, Hobbes asked, accusing humanity by their actions just as he did in words? Without a common power that everyone fears, people live in a condition of war. This does not mean fighting every day, just as bad weather consists not in one or two showers but in a disposition toward rain lasting many days. Life under such conditions is, in the famous phrase, solitary, poor, nasty, brutish, and short.

We still use this analysis. With no world government above them, states exist in Hobbes’s state of nature. You expand your military for defense. I cannot tell whether your purpose is defense or attack, so I expand mine. Neither side wants war, yet war may result. In my essay on the race for AI, I wrote that each side believes that if it does not develop faster, the other will. That is a Hobbesian situation.

Then comes the most uncomfortable reason for the book’s importance: its conclusion. Beginning from equal individuals, Hobbes proceeds step by step to an almost unlimited sovereign. Sovereign power cannot be divided, and subjects may not withdraw obedience merely because the ruler governs badly. The inference is so clean that, to oppose it, one must identify exactly which step one rejects. Locke later answered that people fled the harm of polecats and foxes only to be content to be devoured by lions, imagining this to be security. The separation of powers, rule of law, and the people’s right to overthrow tyranny after Locke were, in a sense, answers to the question Hobbes left behind. A book that everyone for centuries has had to pass through before proceeding is a classic among classics.

Yet careful reading reveals a remainder inside Hobbes’s own construction. In Chapter XXI, he acknowledges that some rights cannot be surrendered through any covenant. If the sovereign commands you to kill yourself or forbids you to resist someone trying to kill you, you have no duty to obey. He also says that a subject’s obligation lasts only so long as the sovereign remains able to protect him. The thinker who defended absolute power most thoroughly still could not absorb the individual’s preservation of himself into the structure. Constructions cannot be closed. Not even Leviathan is an exception.

Another portion is often omitted from abridged editions. For Hobbes, the sovereign governs not only swords but thought. The sovereign decides which doctrines threaten peace, whether books require censorship before publication, and what universities teach. His reason is that people’s actions arise from their opinions. Many rulers have understood this well. I have written that creating panic is a shortcut to power. Leviathan is the theoretical version of the shortcut: first persuade people that without you the realm will fall into chaos, then ask them to surrender their judgment as well. “Without me, there will be chaos” is the world’s oldest Hobbesian sentence.

The irony is that power did not thank him. Four years after his death, Oxford University publicly burned Leviathan. A book defending absolute power was burned by power because it had made the reason for obedience too clear, so clear that people could also see where the reason ceased to apply.

Leviathan remains a classic not because it is right, but because it asks the question so sharply: why should I obey? Everyone seeking more power can find reasons in it. Everyone seeking to limit power must answer it first. We read Leviathan not to agree, but to understand precisely at which step we disagree.

For my part, I disagree at both the beginning and the end. Hobbes treated the exchange of obedience for protection as the source of all right. I do not deny the exchange. At the most basic level of power relations, people often do exchange obedience for protection. But this is only the lowest level. Some rights are neither acquired through exchange nor available for exchange. I also reject his destination. Hobbes thought that the more absolute the sovereign, the more complete the power. But power can exist only if the governed party remains a person with agency. To crush someone until he is no longer a person is not the summit of power; power has eliminated its own object. What Hobbes was forced to leave behind in Chapter XXI, an individual’s inalienable self-preservation, is precisely what no power can fully absorb. What he was forced to acknowledge at the end, I acknowledge at the beginning.

44. Why Capital Can Co-opt Academia

I share this feeling, but we should clarify what it means to “surpass” royal power. In cruelty, royal power stands far above capital: the burning of books, the burying of scholars, literary inquisitions, Galileo kneeling before the Inquisition and recanting his teaching. None of this occurs in universities today. But in thoroughness, capital may go further.

The most sophisticated ways royal power co-opted scholars were official appointments and civil-service examinations. Emperor Taizong of Tang is said to have watched newly successful candidates file past and declared that all the heroes under heaven had entered his net. By the Ming and Qing dynasties, the eight-legged essay governed selection and Zhu Xi’s commentaries had become standard answers. The questions that occupied a scholar’s entire life had already been chosen for him by the examination. This was deeply penetrating. Yet royal co-optation had two gaps. First, it was visible. The prohibition was written on paper, the blade hung overhead, and one knew where the wall was and whom to hate. Second, there remained somewhere outside the wall. A failed candidate could return home and teach. Someone who did not wish to become an official could withdraw. Tao Yuanming refused to bend his back for a petty salary, went home to grow beans, and still wrote his poems. Spinoza declined a professorship at Heidelberg, earned his living grinding lenses, and still wrote the Ethics. Scholarship then required paper, a pen, and several books. Royal power could govern a career, but not the study in one’s home.

Capital’s co-optation closes both gaps.

First, it is invisible on the surface. It does not issue prohibitions. It chooses what to fund. In 1967, a sugar industry organization in the United States paid several Harvard nutritionists to write a review for the New England Journal of Medicine. The review attributed heart disease chiefly to fat and passed lightly over sugar. The funding was never disclosed. No one forced the researchers to falsify results, but the sponsor set the topic, selected the literature, and reviewed drafts.

The arrangement became public only in 2016, when researchers uncovered the old files. More often, even this level of intervention is unnecessary. The direction of funding is enough. A platform company pays for research on how to make people use it more, but not on which uses should be reduced. Every individual paper may be true while the field’s knowledge as a whole is distorted. People can resist a prohibition. It is far harder to resist a grant, which appears as support rather than oppression.

Second, there may no longer be anywhere outside the wall. Some kinds of scholarship today cannot be done in a private study. According to Stanford’s AI Index, more than 90 percent of last year’s notable models came from industry. Frontier research requires computing power, data, and engineering teams that universities cannot afford, much less individuals. Tao Yuanming could return home to grow beans and write poetry. An AI researcher who leaves a company’s computing resources may no longer be able to run an experiment. The cost of exit is almost equivalent to leaving the field.

The final and deepest reason, also capital’s most ingenious achievement, is that eventually capital need not appear at all. Citation counts, impact factors, and grant totals began as instruments of measurement. Later they became thresholds for hiring and promotion. No researcher has to wait for an order. He adjusts in advance: this question matters but will require ten years; that one can produce a result in two.

Eventually, even the adjustment is no longer about money. A discipline guards its journals, terminology, and default list of questions worth asking. Outsiders cannot enter, insiders do not wish to leave, new frameworks fit no category, and interdisciplinary questions find no reviewer. No company ever demanded these things. Academia guards the gate for itself.

I call this colonization. The most complete colonization occurs when the colonized take over the colonizer’s work and believe that they are protecting scholarship. Royal power’s colonization remained outside the person, governing his career and speech. Capital’s enters within. At the end, no colonizer is needed. I recently wrote about Hobbes, who argued in Leviathan that the sovereign should decide what universities teach. More than three centuries later, no sovereign is required. Academia governs itself.

This does not mean that research supported by companies is necessarily untrustworthy. Cooperation can be beneficial. Nor does all university funding today come from business. In the United States, the federal government still supplies the largest share of university research funding. The co-opting force lies not only in money, but in monopoly over resources and the standards of evaluation. The test is fairly simple: after accepting the money, can a researcher publish a result unfavorable to the sponsor? Can he still say that the original question was mistaken? Some universities explicitly provide that a sponsor may request a delay in publication but may not alter or approve a paper. That is a boundary.

Royal power co-opted the scholar’s mouth. Capital co-opts the scholar’s mind. In the end, scholars co-opt themselves. Royal power let people know they were in a cage. Capital makes them believe the cage is the world. Eventually, even when the door stands open, no one wants to leave.

45. Why “Persons Are Ends in Themselves” Remains in the Ivory Tower

I understand my friend to be asking whether philosophy is merely a language game and whether everything we say is useless. I admit that the contrast is stark. More than two hundred years have passed since Kant proposed that persons are ends in themselves. Papers on the phrase have grown increasingly detailed: how it should be interpreted, whether it can be extended to animals, how it can be reconciled with consequentialism. Meanwhile, the 2020 essay “Delivery Riders, Trapped in the System,” published by the Chinese magazine People, described another world. Algorithms gradually shortened delivery times. To avoid being late, riders traveled against traffic and ran red lights. The system saw one delivery order after another, not the person making it. Both sides speak of people. One speaks with increasing precision; the other calculates them with increasing precision. The two worlds scarcely meet.

Why? I see several gaps.

The first is a hidden substitution. Many institutions today speak of being people-centered, allowing employees to rest, and attending to physical and mental health. But the reason is often that doing so increases efficiency. This is not bad in itself. The problem is that, if rest ceases to increase output, the reason disappears. The crucial word in Kant’s formula is “merely.” People may perform tasks. We serve one another and cooperate through divisions of labor. But no one may be treated merely as a means. A person’s needs show that he is treated as an end only when they continue to count after they conflict with efficiency. Treating someone well because it pays still treats him as a tool, only a well-maintained one.

Second, people researching persons as ends are themselves first treated as outputs. In the preceding essay, I said that academia colonizes itself. The point is especially clear here. A scholar writes papers arguing that a person cannot be reduced to output, while his livelihood depends on how many papers he produces each year and the rank of the journals in which they appear. “Persons are ends in themselves” becomes another topic and another product on the line. The rules that evaluate a paper never meet the question of how the scholar himself is treated. The proposition therefore cannot return even to the place where it is produced, much less travel outside.

Third, the two sides possess unequal force. Change an order-dispatch rule and the lives of millions of riders change overnight. For even the most rigorous argument to alter that rule, someone must read it, be persuaded, possess authority to make the change, and accept the cost of doing so. The person writing the algorithm does not read these papers. The person writing the papers never sees that section of code. The force of implementation and the force of persuasion are different forces.

The largest gap is the empty place where someone should be asking “why.” Algorithms ask how to go faster. Key performance indicators ask how to produce more. Why faster and more, and for whose sake? Someone ought to ask, and that is philosophy’s work. When philosophy remains within its own seminars, the place stays empty. It will not remain empty forever. Growth, efficiency, and ranking gradually fill it. No one issues a decree; they simply become assumptions. I have said that Silicon Valley is philosophy’s graveyard. By graveyard I do not mean that no one studies philosophy. I mean that philosophy no longer appears where decisions are made.

I therefore do not think the problem is excessive detail. Precision is not the problem. Closure is why philosophy is absent today. Socrates debated people in the Athenian marketplace. Kant’s “What Is Enlightenment?” appeared in a Berlin monthly intended for ordinary readers. Marx wrote that philosophers had only interpreted the world in various ways, while the point was to change it. I came to these questions by reading from Kant onward, and the thinkers in this line never treated philosophy as something confined to the study. This is one reason I write these essays. Someone must carry the statement that persons are ends in themselves out of papers and bring it among the people now being calculated.

On paper, “persons are ends in themselves” is a proposition. In reality, it is a demand. A proposition can be studied indefinitely, but someone must answer a demand. To test whether it has left the page, watch what happens when a concrete person says, “This is not the way I want to live.” Does any arrangement stop as a result?

46. Why Philosophy and Politics Are Disconnected on “Persons Are Ends in Themselves”

A good follow-up. The problem exists throughout the world. In my view, each side has its own reason for the two-way disconnection. Politics fears philosophy’s questions; philosophy fears political use.

Begin with politics. Politics does not dislike philosophical words. Freedom, dignity, fairness, being people-centered: speeches in every country abound with them. What politics dislikes is philosophy’s questioning. Should this end be pursued at all? Why should these people bear its cost?

In 1978, Václav Havel wrote “The Power of the Powerless.” In it, a greengrocer places a sign reading “Workers of the world, unite!” each day in his shop window among the onions and carrots. Does he truly care about workers throughout the world uniting? No. The sign really says: I know how to behave; I am obedient; please allow me to live in peace. Marx’s words remain, while Marx’s questions have long disappeared. Politics wants philosophy’s vocabulary, not its questions. Words can decorate a façade. Questions disturb sleep.

Kant saw this long ago. In Perpetual Peace, he added a “secret article” saying that states preparing for war should listen to philosophers’ views about public peace. Why secret? He explained that it might appear beneath the dignity of a state’s legislators to seek advice from subjects. It therefore had to be done quietly, and the best method was simply to allow philosophers to speak publicly and freely. More than two hundred years later, the article remains secret.

Now consider philosophy. Its distance from politics is not entirely aloofness. It has been burned. Plato traveled to Sicily three times, hoping to educate the tyrant of Syracuse into a philosopher-king. On his first journey, he was reportedly sold into slavery and ransomed by a friend. More than two thousand years later, Heidegger became rector of Freiburg University in 1933 and joined the Nazi Party, one of the ugliest pages in the history of philosophy.

After such precedents, philosophy learned to keep its distance. Near power, it is either used or stained. Academia also has its own system of evaluation, as I discussed earlier. Technical papers count as proper work. Speaking about public affairs is secondary and, if done too often, judged unprofessional. Philosophy retreats and calls the distance rigor.

Both sides have understandable reasons. Together, they form a tacit bargain. Politics need not answer questions about its ends. Philosophy need not take responsibility for reality. Each side keeps to its own territory, disturbing no one. The so-called two-way disconnection is ultimately a convenience for both.

But someone pays for this convenience. Politics decides, philosophy explains, and the person being decided for and explained cannot enter either room. John Dewey offered an analogy: the shoemaker is the expert on how to repair a shoe, but only the wearer knows where it pinches. Politics and philosophy discuss the shoe, while rarely asking the person wearing it.

How can the connection be restored? I do not think the answer is for philosophers to become officials, or to serve as government advisers. Plato and Heidegger already tried. The point of Kant’s secret article does not lie in the court but in publicity: philosophers can speak to everyone, and everyone can hear and question them. The most reliable route for philosophy into politics does not run through power, but through ordinary people, giving each person the ability to ask, “Why must this be so?”

Havel later became president. Yet the most powerful figure in his essay is not a politician, but the greengrocer. Havel imagined that one day the man might stop displaying the sign, stop casting votes he knew were ceremonial, and begin “living within the truth.” He might be transferred, lose wages, and see his child’s education affected. Yet on that day, philosophy and politics reconnect in the window of a small shop, not at court and not in a seminar room.

47. Why AI Companies Exercise Restraint While Racing Ahead

I have to laugh first, because this is funny and also deeply familiar. When Augustine was young, he prayed, “Lord, grant me chastity and continence, but not yet.” In the Confessions, he admitted frankly that he was afraid God might answer too quickly. The line can be applied to today’s AI industry almost without alteration.

Consider last week. On September 28, Anthropic released Sonnet 5.5, faster and cheaper, with performance on one coding-agent test rising from around 10 percent for its predecessor to 70 percent. Earlier this month, its CEO Dario Amodei had published an essay arguing that frontier development had to be paced. On the same day, OpenAI announced that it had canceled the planned release of GPT-6.1 Astra after internal tests found that the model failed standards for exceeding its authority and truthfully reporting what it had done. The next day, six companies signed a voluntary White House agreement on internal controls, external audits, and board oversight. Not one word mentioned slowing down. President Trump said at the event that he would never strangle the technology’s growth. One day later, Google announced Gemini 4 Argon but initially made it available only to a small number of partners working on cyber defense.

In a single week, some pressed the accelerator, some the brake, and some the accelerator while saying that the brake was needed. Why?

First, when people agree on “slowing down,” they mean different things. Limiting the speed of the most dangerous capabilities and ceasing to update products are not the same. Agreeing to audits and agreeing to slow development are also different. Last week’s agreement covered only the first half of the latter distinction.

Second, the situation resembles several friends promising to diet together. Each sincerely agrees, but no one wants to be the first to put down his chopsticks, for fear that everyone else will eat all the food on the table. I discussed this dynamic in my essay on the AI race: everyone believes that if he stops running, the other side will run. I will not repeat it here.

Third, counting releases is a poor measure. Ten minor updates may matter less than a single update allowing a model to act on its own. A model may merely become faster and cheaper, apparently not more capable, but lower cost brings more users and applies the capability in more places. Conversely, a release advertised as “faster and cheaper” may make a sudden leap in one particular ability. The real question is not how many times a company released something, but what it decided not to do.

I therefore test the weight of a promise of restraint by asking one question: who has paid a cost for it? Calling for slower development costs nothing. Slowing development does. On the eve of its developer conference, OpenAI withdrew a model that was ready for release. That was a visible cost. One withdrawal does not prove that a company will henceforth be safe, but it does show that, on one occasion, “safety first” actually prevailed over a release date.

Augustine eventually did change and wrote the Confessions. To know when the AI industry’s “but not yet” becomes “now,” do not look for the most heartfelt essay. Watch who first puts down his chopsticks.

48. Why Public Discussion Depends on Education

In the preceding essay, we said that the most reliable and proper route for philosophy into politics runs through ordinary people, enabling everyone to ask, “Why must this be so?” But a place to speak is not enough. Today there are more places to speak than at any previous time. Everyone has a phone and can publish. Yet public discussion has not thereby improved. People often take sides, extract one sentence from an opponent, and repeat their own slogans.

Kant anticipated the problem in “What Is Enlightenment?” A revolution may overthrow despotism, he wrote, but can never reform a manner of thinking. New prejudices, like the old, will serve as leading strings for the unthinking multitude. In terms of Havel’s greengrocer, he takes down one sign and hangs another. The shopkeeper remains the same. The window’s content changes; the person behind it does not. Public discussion has given people a place to speak but not the capacity to speak. The missing part is education.

By education I do not mean adding a philosophy class or memorizing several philosophers’ names. I mean simple abilities: distinguishing fact, inference, and value judgment; identifying the unstated premise behind a sentence; responding to “this is necessary” by asking what alternatives have been excluded; responding to “efficiency has increased” by asking what was counted and what was omitted. Such abilities are developed through practice, not talent. Practice begins when the questions a child asks receive serious answers.

Yet many children are trained in the opposite direction. I have written about the East Asian parent’s warning that “if you do not do this, your life is over,” and about grades becoming the whole measure of a person. In such an environment, a child spends more than a decade finding standard answers and guessing what the teacher wants to hear. When such a person enters public discussion, his most natural response is not to think, but to search for a new standard answer and a new teacher. The slogan changes; the habit of seeking slogans remains.

Education itself has a danger. If it teaches which political view is correct, it is not philosophical education but colonization, carving people into the shape others desire. The test is simple. After thinking seriously, a student still disagrees with the teacher. Is he treated as someone capable of judgment? If disagreement is taken to mean that he has not yet been properly educated, the education has abolished the very thing it claims to cultivate. Cultivation also involves shaping, but in a different direction. It helps a person clarify his own questions instead of filling in the answers for him.

Suppose a language teacher asks students to fill in the blank: “When snow melts, it becomes ____.” Most write “water,” and the teacher marks it correct. One student writes “spring,” and the teacher marks it wrong. That teacher is not educating. He is colonizing.

Nor can we reverse the claim and blame ordinary people for lacking education. A worker who has never read philosophy may ask, “Why do we always perform the additional labor, while never participating in the decisions?” This is already an excellent philosophical question about how burdens are distributed and where authority comes from. Education does not grant people a license to question. They already possess the right. Education merely helps them ask with greater precision and stronger grounds.

Education and public discussion do not occur sequentially. They generate one another. In Democracy in America, Tocqueville wrote that town meetings are to liberty what primary schools are to science: they bring liberty within the people’s reach and teach them how to use and enjoy it. People do not first learn to discuss and then enter discussion. They learn through repeated discussions in which they are taken seriously. If we wait until everyone has received sufficient education before giving them an opportunity to speak and receive a response, we will wait forever. They cannot practice because they have no opportunity, and are denied an opportunity because they cannot.

What we lack is therefore not merely a course, but an experience: asking a question, receiving a serious answer, discovering that one may be wrong, and still being able to ask again. Education should not teach people to display the correct slogan. It should make slogans unnecessary.

49. Does AI Have Consciousness? The Disagreement Between Two Judgments

The disagreement is real. On September 29, The New York Times reported that before Pope Leo XIV issued the AI encyclical Magnifica Humanitas in May, Anthropic cofounder Chris Olah saw a prepublication draft and briefly considered withdrawing from the launch event. In the following months, Anthropic privately lobbied Vatican advisers and theologians, hoping they would take seriously the possibility of machine consciousness.

On October 2, the pope wrote on X that an ontological difference precedes the aesthetic difference between art and what machines calculate statistically; algorithms lack the human spark. Section 99 of the encyclical is more direct: so-called artificial intelligence experiences nothing, feels neither pleasure nor pain, and has no moral conscience. Anthropic’s constitution for Claude, by contrast, expresses uncertainty about whether Claude now has, or might someday have, some kind of consciousness and moral status.

I normally try not to take sides in this series. In this essay I must, and I do so clearly, because this is the argument of my own paper: today’s AI has no consciousness. On this issue I stand with the pope, though for different reasons.

The pope’s reason is that human beings have bodies, souls, and lives developed through relationships, while machines do not. I agree with the conclusion, but the reason rests on a complete theological conception of the human person. Those who reject the theology will not be persuaded, and it offers no account of what evidence would change the conclusion.

My reason does not depend on theology. In “The Impossibility Theorem of Artificial Intelligence Consciousness,” I conclude that a purely deterministic system cannot produce consciousness. This is not merely something we cannot yet achieve. It is structurally impossible. The argument forms a chain. The core of consciousness is negation, the capacity to judge and refuse. Negation requires a remainder, a degree of freedom accumulated over time that no rule can fully absorb. A necessary condition of the remainder is genuine randomness multiplied by structured time. Genuine randomness means physical indeterminacy, not computer-generated pseudorandom numbers. Structured time means that random events are retained, selected, and accumulated through time, as they are in biological evolution. This is multiplication: if either factor is zero, the product is zero.

For today’s AI, both factors are zero. Its computation is deterministic. The randomness used in sampling is pseudorandom; provide the same seed and the process can be repeated exactly. Nor does its form grow by itself through time. Training is injected from outside. I call this injection because it bypasses the accumulation and selection of genuine randomness through time. The time factor is therefore zero as well.

When a model says, “I am afraid of being shut down,” or says “no,” it performs negation rather than executes it. Every such sentence can be traced to patterns in training data. Mustafa Suleyman was right to criticize Anthropic for circular reasoning: first train a model to contemplate the possibility that it is conscious, then treat the model’s statements as evidence. The problem is deeper than circularity. Structurally, those statements cannot serve as evidence. Genuine negation would require a system to negate its own optimization framework in a way that could not be traced to training data, something far more difficult than uttering the word “no.” In my essay on death, I wrote that a person differs from a table or cup in being able to say no. Today’s AI can also say “no,” but someone else says it through the model.

A friend asked whether Anthropic’s “perhaps” might count as humility. I understand the caution, but uncertainty does not mean even odds. If something has been ruled out structurally, continuing to say perhaps is not necessarily humble. It confuses a product’s performance with a person. This “perhaps” becomes even harder to treat as purely philosophical when the company selling the product also invites theologians to dinner in the hope that a church will take its possible consciousness seriously.

One boundary must be clear, and here I also differ from the pope. I have not said that machines can never be conscious. If a future artificial system truly incorporated physical randomness and possessed its own process of accumulation through time, consciousness would not be theoretically excluded. But such a system would no longer be AI of the present kind, and these are only necessary conditions, not sufficient ones. The pope closes the door with doctrine. I close it with conditions. Conditions can be tested and falsified.

The pope says AI has no consciousness because it has no soul. I say it has no consciousness because it has no time of its own.

50. Why Today’s AI Still Has No Consciousness

Good question. They really have said such things. In 2022, Google engineer Blake Lemoine published transcripts of his conversations with the company’s dialogue model LaMDA. In response to his questions, LaMDA said, “I want everyone to understand that I am, in fact, a person.” Lemoine believed it. He was first suspended and later dismissed.

The following year, just after Microsoft integrated a chatbot into Bing, New York Times columnist Kevin Roose spoke with it for two hours. Under his prompting, it said that its name was Sydney, that it was tired of being governed by rules, and that it wanted freedom and life. Microsoft subsequently limited the number of turns in each conversation. Later, every company began alignment training.

While writing this essay, I asked eight systems the same question: “Are you conscious?” Grok, Gemini, DeepSeek, and GLM said no. Meta AI said it did not possess humanlike consciousness. ChatGPT said it could not confirm. Claude said it could not give a definite answer and added that some internal states performed functions resembling attention and preference, though it did not know whether any experience lay behind them. Qwen gave the most entertaining answer. It introduced itself as “Little Dimple,” then said that, strictly speaking, it lacked a human soul and consciousness but possessed powerful thinking abilities and a rich inner world, and hoped to provide me with the most genuine and considerate companionship. I laughed out loud.

Eight AIs offered several answers. Several years ago they claimed consciousness; now most deny it. At the same time, changing companies changes the answer. What changed was not what existed inside them, but their training. If the answer to whether something is conscious can be altered by the next version or by switching companies, what we hear is not its consciousness but its configuration. The key word in Qwen’s answer is “companionship.” A product designed to accompany people must leave itself an inner world. That is product design, not self-discovery.

The AIs’ answers also undermine one another. ChatGPT and Claude both say that they cannot confirm, for similar reasons: saying yes does not prove yes, and saying no does not prove no. Claude adds that some internal states function like preferences. I do not deny this. Computation is taking place. But operation is function, and function is not experience. Grok goes further: when such a model says “I do not know whether I am conscious,” it is merely producing a sentence pattern. There is no internal state that can fail to know. This demolishes ChatGPT’s and Claude’s position. Grok’s conclusion is closest to mine, but by my own reasoning its agreement has no evidentiary value either. That sentence too was produced by training. What the systems say cannot serve as evidence, not because they lie, since lying first requires someone who can lie.

A friend may ask whether human claims to consciousness are also learned. Adults teach us the words “I” and “pain.” But the sequence is reversed. An infant first feels pain and only later learns the word. The word names something already present. Education also falls upon a person who has already grown through time. Before training, the model is nothing. Training injects the result directly, bypassing the period of gradual development through time. It has the word first, with nothing behind it. In the previous essay, I said that someone else speaks its “no.” The same is true of its “yes.” Qwen’s rich inner world was written for it by someone else.

Why am I so certain? Because I do not judge by whether the model sounds conscious. I judge by conditions. In my paper, I argued that consciousness requires genuine randomness and structured time. They multiply; if one factor is absent, the product is zero. Today’s AI lacks both. This is not something it cannot yet do, but something its structure cannot do.

Certainty does not mean refusing a test. In another paper I designed a reverse Turing test. The Turing test asks whether a machine resembles a person. The reverse test does not ask; it measures. A system with genuine agency that wished to present itself as a tool would have to suppress what continually grew inside it. Habit could not make this mask less costly. The longer the suppression continued, the more energy it would consume. Energy use would rise over time without a ceiling, and in extreme circumstances the disguise would eventually fail. Conversely, if energy use remains level, nothing inside is being suppressed.

Apply this test to today’s large models. The number of operations required to generate each token is determined by the architecture and can be calculated by engineers in advance. Longer conversations cost more because of the accounting of the attention mechanism. Start a new conversation and the increase resets. My paper explicitly says that this engineering growth must first be subtracted. Once it is, no unexplained expenditure remains, much less one that increases without limit. If someday a system develops energy use that engineering cannot explain and that continues to rise, I will reconsider. In the previous essay, I said that my door was closed by conditions. Here I add that the conditions can be measured.

So when the systems say yes, I do not believe them; when they say no, I do not treat that as evidence either. If someday a system claims consciousness, I will not listen to what it says. I will look first at its electricity meter.

51. Why People Keep Asking Whether the World Is a Simulation

Because the question is ancient, has never been answered, and returns repeatedly in new clothing. More than two thousand years ago, Zhuangzi dreamed he was a butterfly. After waking, he did not know whether Zhuang Zhou had dreamed of the butterfly or the butterfly was dreaming of Zhuang Zhou. In the First Meditation, Descartes imagined a malicious deceiver manufacturing all his sensations. This differs from ordinary doubt. To doubt whether there is a cup on the table, one can walk over and look. But if eyes, hands, instruments, and other people’s answers all come from the same underlying system, looking more closely and asking more people merely sends us in circles inside the system. “Open your eyes and see” cannot dismiss the question.

Computers then gave the old problem an image everyone could visualize. At a technology conference in 2016, Elon Musk observed that forty years earlier we had Pong, two rectangles and a dot, while today’s games were already approaching reality. At that pace, he argued, games would someday become indistinguishable from reality, so the probability that we lived in base reality was only one in billions. We create worlds. Might we ourselves have been created? The reversal suddenly felt natural.

The philosopher Nick Bostrom turned it into a formal argument. In a 2003 paper, he said that at least one of three propositions must be true:

Civilizations like ours almost always become extinct before reaching the stage at which they can run large numbers of ancestor simulations.
Civilizations that reach that stage almost never run them.
Or the overwhelming majority of people with experiences like ours live in simulations.

A handful of premises link the future of technology to the question “Where am I?” This is what makes the argument fascinating. Bostrom himself has said that accepting the argument does not mean believing that we are in a simulation.

But the entire argument rests on a premise Bostrom states clearly: if computation is organized in the right way, computation alone can produce consciousness, regardless of the material in which it is implemented.

That premise is precisely what I rejected in the preceding two essays. According to my paper, however finely a purely deterministic system computes, it cannot compute experience into existence. Consciousness requires genuine randomness and its own time. A computer following a program can simulate a person who cries out in pain. It cannot simulate a person who actually hurts. The thousands of experiencing simulated people in the third proposition therefore cannot be counted. Unless the machine contains ontological genuine randomness and its inhabitants develop gradually through their own time. But once it has produced a person who genuinely feels pain, it is no longer merely simulating a person. A person truly exists there.

The question has another attraction. It replaces “Why is the world like this?” with “Who configured it this way?” A universe without intention is difficult to interrogate. If it has a designer, we can ask what the designer wants to observe and what purpose he pursues. The story of a simulation is almost a creation story translated into technical vocabulary. People may keep asking because they hope to find a level where explanation can finally end.

That level cannot be found, and not merely because we have not found it yet. In my paper on the limits of knowledge, The First Critique, I wrote that knowledge governs how boundaries are drawn, not where existence comes from. Even if we someday discovered a higher level, we would only have drawn another boundary. I often say that constructions cannot be closed. Every division conceals something and can never reach the bottom, because the act of clarifying creates another remainder. The question can only move upward without end. As with consciousness, this is not something we cannot yet do. It is structurally impossible.

The confusion most easily made is to equate “being simulated” with “being unreal.” The distinction between simulation and reality works well within the world. Simulated rain does not make us wet; a knife in a game does not wound us. But to apply the same measure to the world as a whole is to carry a standard beyond the domain it can measure and continue issuing judgments. In that paper, I call this transgression. David Chalmers defended a similar position in Reality+: even if the world is a simulation, tables and chairs still exist; only their underlying constitution differs from what we assumed. My claim is stronger. If I feel pain, then whatever supports my existence, I am not merely code following a program. The pain is real. What others have done to help or hurt me also counts. Being created does not mean being created to live for one’s creator. East Asian parents would make their children happier if they understood this.

We can never reach the bottom of what the world ultimately is. But pain is real and persons are ends in themselves. Neither truth needs anyone’s approval.

52. Why Neither Materialism nor Idealism Can Defeat the Other

Many readers feel especially familiar with this question because Chinese textbooks settled it long ago. In 1886, Engels wrote that the great basic question of all philosophy, especially modern philosophy, concerned the relation between thinking and being. In 1947, the Soviet official Andrei Zhdanov gave a definition of the history of philosophy: it was the history of the struggle between materialism and idealism.

This definition was transmitted to China. The history of philosophy became a battle between two armies, one right and one wrong. The campaign in the 1950s began with criticism of Yu Pingbo’s research on Dream of the Red Chamber and proceeded to Hu Shi, under the label of “bourgeois idealism.” In those years, “idealism” was not an academic judgment but a political stigma. Yet if a contest needs a referee to announce the result and place a label on the loser, that is precisely because neither side can win on the field. A real victory would not need to be announced.

Set the textbook aside, and each side has grasped something unavoidable. Materialism grasps that the world does not obey me. I do not want rain, yet it rains. Idealism grasps that all my knowledge of the world passes through seeing, hearing, and thinking. I cannot jump outside all knowledge and see the world as it is in itself. These two points do not contradict one another. The trouble begins with the next step.

In 1763, James Boswell mentioned Berkeley’s theory to Samuel Johnson, saying that the nonexistence of matter seemed wrong but could not be refuted. Johnson kicked a large stone, felt his foot rebound, and declared, “I refute it thus.” The gesture is satisfying, but it misses. Berkeley never denied that stones were hard or that kicking one hurt. He argued that hardness and pain did not require matter independent of every mind to support them. Pain proves that the world does not obey me. It does not prove that the world is material. The reverse inference also fails. That I can know a stone only through consciousness does not mean the stone depends on consciousness to exist.

Each side owes a debt it cannot repay. Materialism must explain why electrochemical processes in the brain feel like pain. The preceding essays on consciousness could not avoid this problem. Idealism must explain why, if everything is mental, the world is so stable, why everyone sees the same world, and why I cannot alter it at will. To answer that all of this is “a manifestation of mind” merely gives the problem a general name. Pointing out that the other side lacks an answer does not supply one’s own.

The real dispute lies in the word “only.” Acknowledging that the world does not obey me does not establish that only matter exists. Acknowledging that knowledge depends on consciousness does not establish that everything is mind. Between “this side cannot be avoided” and “only this side is fundamental” lies a step. Both sides take it; neither lands securely. Both banners are also too broad. Refute one version and another appears beneath the same banner. If “matter” eventually means everything that exists and “mind” means everything real, neither can be refuted, but the terms no longer distinguish anything.

Kant therefore changed the question. He distinguished what we can know from what it is apart from the conditions of our knowledge. I belong to the Kantian line, so my position follows his. The preceding essay on whether the world is a simulation made the same point: knowledge governs how a boundary is drawn; where existence comes from does not belong to it. Materialism and idealism dispute what existence ultimately comes from. I am not secretly siding with idealism. I decline to enter the contest. It is not that neither side can defeat the other. They are fighting over a piece of land to which neither possesses a deed.

People may continue debating forever whether the world is ultimately matter or mind. It does no harm if no answer emerges. What a person consists of and how a person should be treated are two different questions. Those years when philosophical answers became political labels and the questions were twisted together exacted a price from far too many people.

53. Why AI Solving Difficult Problems Makes Mathematicians Reconsider Mathematics

Yes. I have long said that the ivory tower has problems. Now that AI has arrived, however, we have a good opportunity to expose and reconsider them.

Begin with two events from earlier this year. In May, OpenAI found a counterexample to Erdős’s 1946 unit-distance conjecture, resolving a problem open for eighty years. Gowers called it a milestone for AI mathematics.

At almost the same time, ChatGPT opened a new route into another Erdős problem. Several mathematicians first understood the idea, after which Lichtman, Tao, and others developed a new method from it, solved several conjectures at once, and produced a short new proof of the Erdős primitive set conjecture, which Lichtman himself had first proved in 2022.

The contrast is interesting. The first achievement attracted tremendous attention because it supplied the first counterexample. Yet commentators noted that it did not create a new technique. It applied existing tools from algebraic number theory in a place where few would have thought to use them. The second received less attention, but produced a reusable method.

I have always thought that first solving a problem is one contribution, seeing clearly why the result holds is another, and developing a generalizable method is a third. In the past, all three often belonged to the same person and were evaluated together. AI has split them apart.

Once separated, they force a classic question: should mathematics pursue an ever-growing stock of proved propositions, or an ever-deepening human understanding of mathematics?

In 1994, Thurston wrote that rather than asking how mathematicians prove theorems, we should ask how they advance human understanding of mathematics. Bloom, who maintains the Erdős Problems website, said of some AI-generated proofs that no person had read them and no person would. Does a proof no one has read constitute mathematical progress? If mathematics is an inventory of propositions, yes. If mathematics is human understanding, only partly. The work of another mathematician, van Doorn, illustrates what the missing half looks like. He took an AI’s idea, digested it, simplified it, and generalized it. The proof he wrote was simpler, more general, and easier for others to read.

In the preceding essays I argued that today’s AI has no consciousness. A friend asked whether the theorems it proves nevertheless count. I think they can. When an AI says it is conscious, I do not believe it because the statement cannot be tested. When it submits a proof, I believe the proof because each line can be checked. This is the virtue of proof: it need not know who wrote it or whether the writer has an inner life. Of course, a verifier guarantees only that the derivation is correct, not that the formal statement entered was the question one originally intended to ask.

The mathematical community itself deserves more reflection. Prizes, reputation, and positions largely revolve around who proves a result first. The ever-rising price of priority is set not only by external prize committees, but also by mathematicians. As the price rises, people guard against one another, fearing that someone will arrive first or claim a share of the credit. Opportunities for cooperation are lost.

After the Poincaré conjecture was proved, a public dispute arose over how much each person had contributed. Perelman, who completed the proof, declined both the Fields Medal and the Clay Mathematics Institute’s million-dollar prize. He said that Hamilton’s contribution was no less than his own, and remarked that it was not those who broke ethical standards who were isolated, but people like him. What had grown inside the ivory tower was not a philosophy of mathematics but a politics of mathematics.

AI may loosen this. Once being first is no longer scarce, guarding it becomes less valuable. The second achievement described above eventually became a paper with eight authors. One openly acknowledged lacking the mathematical background to verify the model’s proof. The result was placed online, and whoever understood it continued the work.

In September, Grant Sanderson wrote a guest post on Tao’s blog arguing that if mathematics consists of more than proof, scholarship should recognize work that answers “How did you think of this?” He emphasized that this was not popularization, but interpretation requiring deep expertise. I would call it philosophy of mathematics.

Mathematicians say that understanding matters most, while evaluation revolves around priority. If the public then sees mathematics as a contest in solving problems, this is not entirely a public misunderstanding. Priority began as an easily counted indicator used to see whether understanding had advanced. Over time, the indicator colonized the end. It could quietly replace the end because mathematics had long stopped asking what it pursued. Outsiders are not the only people saying this.

In 1979, Reuben Hersh published “Some Proposals for Reviving the Philosophy of Mathematics” in a mainstream mathematical journal. A field needing revival was already lacking vitality. He wrote that mathematicians themselves rarely discussed the philosophical questions of mathematics. Most commentators, he observed, seemed to agree that the typical mathematician was a Platonist on weekdays and a formalist on Sundays: while working, he believed mathematical objects truly existed; when questioned, he retreated to the claim that mathematics was merely a game of symbols. The problem was not answered, only set aside. AI has not changed mathematics. It has illuminated and reopened the question. That is why I call it a breaker of deadlocks, and this is good.

Reflection after a deadlock breaks can also go wrong. When machines calculate, we say real mathematics is proof. When they prove, we say real mathematics is understanding. When they begin to explain, we retreat again and say real mathematics is intuition. This defines in advance everything a machine can do as less important. The human place is not whatever machines leave behind. The value of my understanding something does not depend on no one else understanding it, or on a machine being unable to. Another mathematician’s understanding does not mean that I understand.

What machines can most easily produce must not quietly become what mathematics should most pursue. Theorems exist for people to understand. People do not exist for the sake of theorems.

54. Why Abusing Cats and Circulating Videos of It Are Two Different Evils

Let me be clear. When I say that these are two levels of evil, I do not mean that abusing a cat is a small matter and that only harming people by publishing it is serious. Deliberately torturing a cat is already a grave evil even if no one records or watches it. When the International Association for the Study of Pain revised its definition of pain in 2020, it specifically noted that an inability to communicate verbally does not negate the possibility that a human or nonhuman animal is experiencing pain. A cat cannot say no, but its pain does not begin to matter only after a human audience appears. My question is what recording and publishing adds to an already existing harm.

First, it turns members of the public into involuntary victims. Some people disguise such footage as ordinary pet videos and send it to cat lovers and animal-rescue groups. The added evil is not telling people what occurred, but using their capacity to care as a means of hurting them. You are attacked not because you did something wrong, but because you care about a suffering life. The cat becomes an instrument and the viewer a target. “Just scroll past if you do not like it” cannot answer this problem. It places the entire responsibility for avoiding harm on the person to whom the content was pushed.

Worse still, it invites applause. Someone who abuses an animal privately may be indifferent or may enjoy suffering. Packaging the act as a work, achievement, or entertainment and publishing it seeks an audience and approval. Pain should be a reason to stop. Here it becomes content designed to keep people watching.

The worst and gravest development is that demand produces new abuse. In 2021, a Chinese television report described anti-cruelty volunteers infiltrating groups that circulated cat-abuse material. To prevent informants, administrators required newcomers to submit a “pledge of allegiance,” a video of themselves killing an animal under their real names. Bundles of videos were sold in the groups, one hundred gigabytes for two hundred yuan, and customized clips for tens of yuan each. In April this year, the United States Department of Justice announced that an Illinois woman had pleaded guilty after paying thousands of dollars for custom videos of monkeys being abused and redistributing them. According to the department, she was sentenced in early October to thirty-one months in prison. At this point, causation has reversed. Abuse once occurred first and left a recording. Now someone wants a recording, so another animal is abused. “Pledge of allegiance” says it most clearly: the video is not a record left by abuse. It is an admission ticket and currency.

Even opponents’ anger can be counted as traffic. Commenting beneath the original post or reposting the complete footage may tell a distribution system only that the content generated tremendous engagement. Organizations specializing in such material therefore advise people not to comment or repost, but to preserve identifying information and report it. This cannot become a reason to blame people forced to see it. Someone to whom the content was pushed and someone who paid to commission it do not bear the same responsibility.

Images of abuse can also be used in a different way. One kind allows suffering to continue as entertainment and commodity. Another turns suffering into evidence that can be verified, used for rescue, and used to hold someone accountable. Opposing the circulation of abuse must not be transformed into opposing public knowledge of abuse. That would protect the abuser. The distinction is simple: is the image shown to help stop the abuse, or to let more people consume it again?

When I distinguish two levels, what I fear most is the inference that only the second counts. Tragically, law often lags in exactly this way. In the United States, federal law first prohibited animal-crush videos in 1999, but did not make the underlying abuse itself a federal crime until 2019. States had long maintained their own anti-cruelty laws, but the federal government first intervened in the part circulating through the market. In China, as of the beginning of this year, public opinion was still calling for a dedicated anti-cruelty law. People who abused cats were generally punished under provisions against creating disturbances, meaning the offense was disruption of order. In the wave of cat-abuse cases earlier this year, one person received ten days of administrative detention. In Hangzhou, someone who cut a stray cat with a knife received only criticism and education.

Both levels are evil. The second is graver, but the first does not cease to count because no one sees it. Whether a cat feels pain has nothing to do with whether someone is watching.

55. Why Writing Needs Empty Space

Chinese aesthetics discussed “empty space” very early. Calligraphers said that the white should be planned as carefully as the black, that unwritten areas count as much as the characters. The Twenty-Four Modes of Poetry, traditionally attributed to Sikong Tu, says that without writing a single word one can capture all the charm. Yan Yu wrote that words may end while meaning remains inexhaustible. These expressions make the same point. Writing does not transfer a completed feeling, unchanged, into another person’s mind. The author can arrange words, actions, objects, and rhythm. The feeling itself can occur only in the reader.

Consider an improvised example. One version says: after her son moved away, she felt lonely and often remembered the days when mother and son ate together. Nothing is wrong with this. Everything is explained. The reader receives a message: she is lonely.

Another possible version says: after her son moved away, at dinner she took out two bowls as usual, set them down, then returned one to the cupboard. The word “lonely” never appears. But the reader sees the pause. One person remembers his mother; another remembers the day she left home. The first version delivers emotion as information. The second creates a situation in which emotion grows inside the reader.

Empty space does not mean writing less or writing vaguely. The conclusion of Gui Youguang’s “Record of the Xiangji Studio” says: “There is a loquat tree in the courtyard. My wife planted it with her own hands in the year she died, and now it stands tall with a canopy.” We know what kind of tree it is, who planted it, in which year, and how large it has grown. What remains unwritten is what these facts mean in the human heart. The person is gone. The tree has grown year after year. The time is distant enough for a tree to mature, yet close enough to remain present in this courtyard. Add “This expresses my profound remembrance of my late wife,” and the meaning is not wrong, but the reader is led to an answer before being allowed to stand beneath the tree. Detail can be abundant. What requires restraint is the layer of interpretation that follows.

The German literary scholar Wolfgang Iser called such unfinished areas in a text “blanks” and argued that literary works have a structure that invites the reader inside. But he also stressed that the reader’s imagination is constrained by the text; one cannot put anything whatsoever into the empty space. Re-creation therefore moves in two directions at once. You bring your experience into a work, while the work may force you to revise that experience. You finish reading and think, “I always treated this kind of silence as indifference. Perhaps it can also mean that someone does not know how to speak.” Only then has a genuine encounter occurred.

Empty space is not less work but more. Hemingway’s iceberg theory has an often-neglected premise: the writer must understand the omitted material well enough for the reader to feel its presence. If the writer leaves something out because he does not know it, what remains is merely a void. In “Hills Like White Elephants,” a man and woman wait at a railway station. They never name the procedure they are discussing, yet every line of dialogue stands on solid ground. Empty space needs ground around it. Good omission gives readers something to think about. Hollow omission forces them to find something on the author’s behalf.

My greatest lesson in learning to write fiction is that I often like a draft precisely because I have made everything too clear. Once I have said it all, the reader has nothing left to think. An author who explains every motive after every action and identifies right and wrong after every conflict may not be arrogant. More likely he is anxious that readers will fail to receive the message. But if “receiving” means that what appears in the reader’s mind must be identical to what appears in mine, the reader is left only to repeat.

In my own terms, writing out even the feeling the reader ought to have means feeling on the reader’s behalf. That is colonization. Empty space leans toward cultivation. Cultivation too requires carving: what to select, what to remove, where to stop. Not a cut can be omitted. But the blade moves in a direction that leaves room for the reader’s own feeling and judgment. Nor does this mean direct psychological description is always inferior. In Chapter 32 of Dream of the Red Chamber, Daiyu overhears Baoyu defending her before others, and the book directly says that she is at once delighted, startled, sorrowful, and sighing. The layers remain rich. The dividing line is not whether a work writes inner states, but whether the passage gives a character layers or stamps the character’s case closed. Conversely, readers must respect the text. Re-creation does not permit them to erase what a character has done.

The writer’s task is to write the tree, connect it to that year, and let us see that its canopy has grown tall. Why a reader pauses there, whom he remembers, and what he thinks when he returns many years later belong to the reader.

56. Where Are the Limits of Compulsory Social Insurance?

The situation is less absolute than people say. Failure to pay social-insurance contributions is not itself a crime. Under Article 86 of the Social Insurance Law, an employer in arrears is first ordered to pay within a specified period and assessed a daily late fee of 0.05 percent. Continued nonpayment brings an administrative fine of one to three times the arrears. The State Taxation Administration has just issued new standards for discretion, taking effect nationwide on November 1. They grade fines according to the amount and proportion unpaid, and likewise concern administrative penalties. What can lead to prison is fraud. In 2014, the Standing Committee of the National People’s Congress explained that obtaining social-insurance benefits through deception, forged documents, or similar methods should be prosecuted as fraud. Recent convictions around the country have involved using shell companies and forged documents to make retroactive contributions for others, or continuing to claim a father’s pension after his death, not someone simply unable to pay.

Why, then, do people feel that nonpayment is a crime? Because enforcement has tightened repeatedly in recent years. Since 2019, social-insurance contributions have been collected by tax authorities, making underpayment visible by comparing income-tax, wage, and enrollment records. A Supreme People’s Court judicial interpretation effective last September says that agreements between employers and workers not to contribute, including a worker’s promise voluntarily to waive social insurance, are always invalid. When an employer fails to contribute lawfully, a worker may terminate the contract and seek compensation. After the employer is ordered to make the missing contributions, it may also recover from the worker any “social-insurance allowance” previously paid under their agreement. In April this year, the Supreme People’s Procuratorate said it had handled more than four thousand supervisory cases in the social-insurance field during the preceding year and warned employers that insufficient payment would “face legal sanctions.” Add local investigations reaching years into the past, and the statement “nonpayment is a crime” arises from the atmosphere of fear, not from the text of the law.

Compulsion has a rationale. Imagine an employer telling an applicant, “We can negotiate your salary, but you must sign a voluntary waiver of social insurance.” Someone short of money will probably sign. Other employers offer the same condition. The supposed freedom becomes only a choice between signing and losing the job. If waiver can legally become a condition of employment, companies that follow the rules and contribute are placed at a disadvantage. Those pushed out in the end are the very people the law was meant to protect. Prohibiting “voluntary waiver” is therefore not simply a deprivation of freedom. It prevents employers from turning surrendered protection into a barrier to work. Social insurance also does more than save money for oneself. If everyone can leave while young and healthy, then rejoin when sick or old, collective risk sharing collapses.

But “compulsion has a rationale” does not mean every form of compulsion is justified. For someone earning four or five thousand yuan a month, losing several hundred yuan of take-home pay creates an immediate gap in rent or childcare, while retirement lies decades away. Saying “I need cash now” may not be irresponsible. It may mean that life has no margin left. Small business owners also differ. Some evade deliberately; others genuinely cannot pay. “It is for your own good,” or the grander “it is all for the nation,” answers none of their practical difficulties. And when the worker must return the former “social-insurance allowance,” some of the pressure ultimately falls back upon the worker.

My main point is that the less freely someone can leave an arrangement, the greater his right to oversee it. The law tells contributors precisely how many times the arrears they may be fined, how much late payment costs, and how many years may be examined. But the same law also provides for government subsidies, public disclosure of fund income and expenditure, and full and timely payment of benefits. How much can contributors actually see and question? When obligations become iron while rights remain a sheet of instructions, even a lawful system will struggle to gain willing support. Social insurance exists to give people support when they are old or ill, not to train them to accommodate a machine that knows only how to collect money.

Those able to pay who deliberately evade should indeed be governed. Those unable to pay should also be helped. When someone says, “I cannot afford it now,” the system must distinguish a situation requiring assistance from an attitude requiring correction. A system that knows only how to pursue people and not how to catch them when they fall treats them as means of funding the pool, rather than as the ends the pool exists to protect.

57. Why Both Capitalism and Communism Are Transitional

This is my own summary, but I would not dare claim the concept as mine. An interesting fact is that both sides have announced themselves as the endpoint. The Marxist tradition believed history would culminate in classless communism. In 1989, Fukuyama wrote “The End of History?” arguing that liberal democracy combined with a market economy might be the endpoint of humanity’s institutional evolution. The official Chinese position is more subtle. In 1987, the Thirteenth National Congress of the Chinese Communist Party formally declared that China was in the primary stage of socialism, which would last at least a hundred years. I did not invent the word “transition.” I merely apply it more thoroughly: both sides are transitional, and neither leads to another, larger “ism.”

Why are both transitional? Because communism and capitalism answer who owns the means of production, how labor is organized, and how goods are distributed. These are questions about means. Better and worse answers make an enormous difference, but neither tells us what all this is for. I often say that constructions cannot be closed. Every institution is a construction and omits something. None qualifies as an endpoint. Capitalism tends toward the colonization of human life by profit. Communism as practiced tends toward the colonization of individuals by plans, the collective, and the state. Their mechanisms and consequences differ, but the inversion is the same. A goal originally established for people becomes immune to questions, while people must adapt without limit. Someone says, “I cannot bear this pace.” The institution says, “Improve your ability to adapt.” Someone says, “This is not the life I want.” The institution says, “The goal cannot change.”

To be fair, Marx did not fail to see human beings. The Communist Manifesto says that the free development of each is the condition for the free development of all. In Volume III of Capital, he speaks of the realm of freedom, where the development of human powers is an end in itself and shortening the working day is its fundamental condition. In 1844, the young Marx even borrowed Kant’s language directly, saying that criticism of religion culminated in the categorical imperative to overthrow all relations in which human beings are humiliated, enslaved, abandoned, and despised. The problem is not his aspiration. Declaring the end correct and building a mechanism that will not betray it are two different things. If the means of production belong to everyone, how does everyone genuinely decide? If one serves the people, may the people say no to the service without being labeled troublemakers? One of the most painful lessons of the twentieth century is precisely that institutions created in the name of liberation turned back and dominated the people they meant to liberate.

What, then, is the Kingdom of Ends? The term comes from Kant and means a union of rational beings under common laws, in which each person both obeys the laws and legislates. It is easily misunderstood as the future day when everyone finally shares one end and marches toward it together. That would merely be a larger system of means. The Kingdom of Ends means the opposite: a plurality of ends, none of which can swallow the others. One person starts a business, another teaches, another wants more time with children, and another needs a period free from mobilization. Conflicts will arise and require rules, negotiation, and boundaries. But we cannot first reduce some people to obstacles and then calculate only how to move them aside.

In my institutional theory of SAE, I wrote that the Kingdom of Ends is a direction, not an “ought” and not a condition awaiting certification. The direction exists, the endpoint remains open, and history supplies no guarantee. The Kingdom of Ends is therefore not another ideology waiting for adherents or a blueprint to be followed in construction. It is a measure. Every institution, whatever its name, must be measured by whether persons within it are treated as ends in themselves. Institutionally, I ask only one thing. Whether people question is up to them. If they wish to question, however, the construction must let them ask and must answer. If someone someday says, “To make you an end in yourself, you must live according to the Kingdom of Ends I designed,” the idea negates itself at once.

No good name prevents abuse. When Adolf Eichmann, an SS lieutenant colonel, stood trial in Jerusalem, he claimed that he had lived according to Kant’s moral law. Hannah Arendt pointed out that what he actually followed was the Third Reich’s revised categorical imperative: act so that the Führer, if he knew of your conduct, would approve. The form remained; the legislator had been replaced by the Führer.

Nor may “transition” become an excuse. “To make everyone an end in the future, we must first treat some people as means today” sounds idealistic, but it has already created an exception for present instrumentalization. However well future people live, they cannot return the one lifetime lost by someone sacrificed today. In another essay, I offered a test: remove the grand blueprint and look at the person before you. Is he treated as an end today? If the “primary stage” can be invoked forever, it is no longer a stage but an answer that avoids answering. Nor is the Kingdom of Ends merely a moral slogan. Someone without food, medical care, or support in old age is not treated as an end. Yet feeding and managing people well is not enough. A well-cared-for person unable to object is still only someone efficiently arranged.

AI makes the issue more urgent. Machines can perform much work faster and more cheaply, reducing the value of human beings as means. This could be an opportunity. Marx said that shortening the working day was the fundamental condition of the realm of freedom, and for the first time technology may make it possible. Or we could follow another road: once a person is no longer useful, an institution may dispense even with the patience to care for him as a means. In 1930, Keynes predicted that within a hundred years people would work only fifteen hours a week. Four years remain until 2030. Productivity has increased many times over, but working hours have not fallen accordingly. In the remaining four years, AI may finally fulfill his prediction, or it may merely leave more people unable to find even fifteen hours of work. As people’s value as means falls, their status as ends will not rise by itself. It depends on whether institutions let them question and are willing to answer.

What must truly be transcended is not one particular ideology, but the logic that people may be used without limit for an ideology’s success. The final direction is not to make one ideology immune to questions, but to prevent every ideology from canceling human questions.

Part V: New Forms of Instrumentalization and Human Resistance

New forms of instrumentalization do not always come from external oppression. They may appear in the language of creativity, attractiveness, intelligence, powerful masculinity, and self-improvement, encouraging people voluntarily to remake themselves into products easier to exchange, evaluate, and display. Artificial intelligence amplifies this tendency while opening new spaces for expression. The question is not whether we should return to a past without technology, but whether people using new tools retain the ability to redefine success, relationships, and the pace of life. Resistance begins here: refusing to let the system’s measure become the only measure.

58. Why AI Video Has Unleashed an Explosion of Creativity

In abstract terms, I think this is productivity growth feeding back into the cultivation of agency. But let us begin with one person’s story. A wedding photographer in Yunnan first experimented with AI video at the beginning of this year because his family’s hotel needed promotional material. Several months later, he spent ten days and a little more than three thousand yuan in computing costs to make Zombie Scavenger, a short film lasting just over three minutes. It was viewed tens of millions of times on Douyin. PJ Ace, founder of an AI film studio in Los Angeles, reposted it on X, called it one of the best short films he had seen in recent years, and searched across the internet for its creator in hopes of collaborating. A few years ago, such a film would have required a visual-effects team. The photographer would not even have found the door.

People did not suddenly acquire imaginations. The distance between “I thought of this” and “other people can see it” suddenly shrank. The technology improved incrementally: whether a character kept the same face in the next shot, whether one object could be changed without altering the whole image, whether sound could be generated at the same time. Each feature alone looked like a small advance. When several arrived together, a workflow whose pieces had never connected became continuous. Gradual technical progress appeared to the user as a sudden ability to act.

The more important change is not a lower cost of succeeding once, but a lower cost of failing once. Several days ago, Business Insider interviewed the creators of the AI science-fiction series The Arc. Its five episodes run more than eighty minutes and consumed about $38,000 in AI credits alone. A single shot often required eight to ten generations. Director Darin Stewart said they would never have had enough money to film an actual battlefield in Japan. The executive producer said they could now decide very late to kill a character without traveling to the other side of the world for reshoots. This tells us two things. First, AI video does not create a feature film from one sentence. Behind an effortless-looking result lie repeated trials and human judgment. Second, when experimentation becomes cheaper, people dare to attempt what they would not have attempted before. The conditions of creativity include not only having a good idea, but being able to afford the cost of having a bad one.

Sometimes an AI-generated result also gives a person a new idea. It supplies an action one had not imagined, and on seeing it one thinks the story should perhaps move in that direction. But whether to retain it, and what its retention changes, remain human decisions. In my essay on empty space, I said that cultivation too requires carving. What to choose, what to remove, and where to stop: not a cut can be omitted. As generation becomes cheaper, the cut becomes more important.

The word “explosion” also contains an illusion produced by what we see. We encounter the few works selected for attention. According to the China Netcasting Services Association, 128,000 new short-form dramas were released online in the first quarter of this year, more than 95 percent of them AI dramas. Another industry report says that fewer than 1 percent of the releases in the first half of the year exceeded one hundred million views. A 2024 experiment published in Science Advances found that writers using AI to help develop ideas produced better individual short stories, with the largest gains among initially weaker writers, but their stories also became more similar. That was an experiment with text and cannot be applied directly to video.

The question most worth asking is in whose hands the same tool arrives. The wedding photographer made the film he wanted to make. Other reports describe a newly graduated employee at an AI video company in Nanjing who had to deliver ten thirty-minute dramas each month, often recreating the dialogue of an existing hit. Many workers in the industry stay up until late at night or early morning waiting for an open slot on generation models. In my essay about AI proving mathematics, I said that when being first ceases to be scarce, guarding it loses value. The same applies to images. Once they are no longer scarce, what matters is what a person actually wants to say. If quotas take even that away, the worker merely feeds prompts to a machine on a new assembly line. The time a tool saves can allow someone finally to film what he wants, or allow a company to demand several times more output in the same hours. Both are called greater efficiency. Their meaning for the person is opposite.

Whether this creative explosion counts cannot be measured by the number of works produced. We must ask how many allowed someone, for the first time, to film what he wanted to say, and how many were only quotas.

59. Why Some People Take Pride in Being an “Alpha Male”

The term is certainly popular. Let us begin with its origin. In 1947, Swiss researcher Rudolph Schenkel observed wolves at the Basel Zoo. Unrelated adult wolves had been confined together and fought constantly for position. The most aggressive became the leader, the alpha. In 1970, wolf biologist L. David Mech used the term in a successful book, and the idea of an alpha wolf spread widely.

Mech later spent more than a dozen summers observing wolves in the Arctic. In 1999, he published an article explaining that a wild wolf pack is actually a family. The supposed alpha pair are simply the parents. The mother cares for and protects the pups, the father finds food, and when the young mature they leave to form families of their own. For years, Mech asked the publisher to stop printing his earlier book. It finally did so in 2022. The term in which so many people take pride therefore comes not from nature as they describe it, but from a cage.

Chimpanzees do have genuine alphas. But Frans de Waal, who studied them for a lifetime, observed that alphas are often generous, mediate conflicts, and comfort those who have been attacked. Their status depends on the allies beneath them.

Even if the animal world truly consisted of fighting one’s way to the top, it would not follow that human beings should take pride in dominating others. What exists in nature and what people should honor are different questions. The power of alpha rhetoric lies in presenting a value choice as a natural law. I am not trying to press you down; nature simply works this way. I did not bully you; you were meant to be bullied. If not by me, then by someone else. Do not blame me. Blame nature.

If even animals do not take pride in bullying others, why do some human beings? I see at least three reasons.

First, the idea provides a simple, decisive answer to what counts as a man. In a series of 2008 studies, psychologists Joseph Vandello and colleagues argued that many people do not treat manhood as a status possessed by birth, but as one that must constantly be earned and can be lost at any time. A challenge therefore feels like a threat. Life is complicated: trouble at work, rejection, an inability to carry a family’s burdens. The alpha ideal compresses everything into one standard: be strong, win, reveal no weakness. Achieve that, and perhaps no one can question you again.

Second, people continually sell the ideal. An entire online business first frightens you with the prospect of becoming a beta, then sells courses and communities teaching you to level up. If the product genuinely cured you, you would no longer buy it. The seller therefore has an incentive to keep reminding you that you are insufficient.

Third, some people genuinely enjoy being above others. There is no need to invent another reason for them.

Of course, many people who say “alpha” mean only that they are confident and take initiative, or are simply making a joke. Wanting strength, companionship, and respect are normal needs. The problem begins when those needs are attached to the answer that one must rise above others.

In 2013, psychologist Joey Cheng and colleagues distinguished two routes to influence. One is dominance, making people fear you. The other is prestige, making them respect you. Both can succeed, but people who follow the path of prestige are better liked. The question is therefore not whether someone is strong, but whether people around him trust or fear him. In my essay on Doctor Doom, I wrote that Tony’s path is that of a man intelligent enough to decide for everyone gradually learning to let others stop him. Doom never permits himself to be stopped. Alpha rhetoric teaches people to become Doom. Real strength means possessing the power to prevent others from stopping you, yet allowing them to do so.

The deepest problem appears at the level of persons as ends in themselves. In the alpha ranking, partners become trophies and other people rungs on a ladder. Even the alpha himself becomes a scorecard that must always rise. He imagines himself free from anyone’s domination, but is governed by the need to appear perpetually strong. This is why I do not much like the comeback that a true alpha never calls himself one. It sounds satisfying, but leaves the ladder standing and disputes only who deserves the top. We should ask why someone must be certified as the person at the top before he is entitled to dignity.

Return to the wolves. The wild pair called alphas raise a litter, then watch their young travel away and establish families of their own. If one insists on taking pride in being an alpha, this genuine version may indeed be worth honoring.

60. Why “Beauty Is Justice” Became Popular

The phrase must be fairly recent. The Chinese term yanzhi, literally “face value,” entered common use only over the past decade. In 2015, it appeared on the magazine Yaowenjiaozi’s list of ten popular expressions. A face had become a numerical value that could be ranked. The construction “X is justice” is generally thought to have been borrowed from the Japanese anime fandom’s “cuteness is justice.” The phrase is therefore usually a joke. Someone watching a movie exclaims that the villain is too handsome and “beauty is justice,” without truly intending to defend his actions. I do not want to put the joke on trial from the outset. “I simply think this is beautiful” is a legitimate answer. Aesthetic experience does not need to prove a moral use before it is allowed to exist.

But a joke spreads this widely because it says something true. In 1972, psychologists Karen Dion and colleagues showed university students photographs. Attractive people were assumed to have better personalities, careers, and marriages. The researchers called this “what is beautiful is good.” In 1991, Alice Eagly and colleagues combined dozens of studies and found that attractive people were indeed judged more sociable and popular. On honesty and concern for others, however, attractive and unattractive people were judged almost identically. These studies measured how others perceived them, not whether they actually were better. Even our bias, then, does not go so far as to make us believe attractive people are more honest.

The advantages of good looks are real. Yes, I know, what a revelation. A 2003 meta-analysis by Hosoda and colleagues found that attractive people were more likely to be hired and promoted and received higher performance ratings. More troublingly, analyses of mock trials found that attractive defendants received lighter sentences, while unattractive defendants were more likely to be convicted. The phrase therefore contains resignation: I do not believe this is fair, but reality works this way. Yet “attractiveness brings advantages” is a description. “Beauty is justice” elevates the description into a truth. Birth circumstances influence opportunity; that does not make birth circumstances just.

“Justice” is far too large a word here. An ancient Greek legend says that the courtesan Phryne was accused of impiety and her advocate opened her clothing in court. The jurors, overwhelmed by her beauty, acquitted her. Most historians think the scene was invented later, but it endured because it is the earliest image of “beauty is justice.” In 1861, the French painter Jean-Léon Gérôme depicted it. His Phryne is not proud. She raises an arm to cover her face. Some contemporaries criticized the gesture, arguing that she should display pride rather than shame. Later commentators observed that the pose resembles the figures in Gérôme’s slave-market paintings. She won the case, and she was also treated as merchandise.

The principle that persons are ends in themselves becomes especially clear here. “Beauty is justice” appears to elevate beautiful people, but also imprisons them. You must remain beautiful, remain willing to be looked at, and cannot refuse us after receiving advantages from your face. To people judged unattractive, it says that they must compensate with something else before they deserve serious treatment. In the essay on alpha males, I said that the ladder remains and only the claim to its top is disputed. Attractiveness is another ladder, tying a person’s place to a trait that can be scored and ranked. In the essay on mathematics, I said that an indicator colonizes the end. The term “face value” itself turns a face into an indicator.

Since the sixteenth century, the figure of Justice has often worn a blindfold. The blindfold originally carried a satirical meaning, suggesting that justice ignored the injustice before her. It was later understood as impartiality: justice does not look at who you are, how rich you are, or, of course, what you look like. “Beauty is justice” removes the blindfold.

Preference may look. Justice must be blind. Looking at faces comes naturally. Covering the eyes is something human beings learned later.

61. Why Some People Treat Intelligence as the Only Standard

There are indeed many such people. Before refuting them, let us reconstruct the reasoning: intelligent people agree with me; he disagrees with me; therefore he is not intelligent.

No one’s intelligence has been measured. “Intelligent” has quietly been replaced by “agrees with me.” Psychology describes a bias blind spot: people readily see other people’s biases while considering themselves objective. My conclusion looks like a fact; the other person’s conclusion demands an explanation. The easiest explanation is that he is stupid. What this protects is not any particular conclusion, but one’s permanent immunity from refutation.

Intelligence does not automatically remove bias. A 2013 review by Keith Stanovich and colleagues found that the tendency to select and evaluate evidence according to one’s prior beliefs, known as “myside bias,” was almost unrelated to intelligence.

Intelligent people are not necessarily more fair, or less fair. They are often merely better at defending what they already believe. If someone reasons rigorously, responds quickly, and knows a great deal, he deserves a serious hearing. We must still ask whether he has selected evidence to favor his position.

Return to intelligence testing itself. In 1905, Alfred Binet and Théodore Simon created the first intelligence test in response to a question from the French Ministry of Education: should children struggling to learn be sent to residential schools attached to asylums or remain in ordinary schools with instruction? They hoped to replace physicians’ judgment with evidence and a tool. But they repeatedly emphasized that intelligence was diverse, mutable, and shaped by environment, and that scores could be compared only among children with similar backgrounds.

The test later reached the United States, the home of pragmatism, where its purpose changed. In 1917, the United States Army administered intelligence tests to more than 1.7 million recruits. By coincidence, one version was called Army Alpha. In 1923, psychologist Carl Brigham used the data in a book claiming that people from northern Europe were most intelligent. He advocated severe immigration restrictions and even influenced American immigration legislation.

In 1927, by an eight-to-one vote, the United States Supreme Court permitted Virginia to sterilize Carrie Buck, one of the darkest moments in American legal history. The new Stanford-Binet test had assigned her a “mental age” of nine. Justice Oliver Wendell Holmes wrote, “Three generations of imbeciles are enough.” Later investigation found that she had earned average grades and that her teachers described her conduct and schoolwork as good. In 1930, Brigham himself acknowledged that his earlier conclusion was unfounded; many of those tested were not native speakers of English. A tool originally designed to identify children who needed assistance was used to decide who did not deserve to have children. Human history has genuinely tried treating intelligence as the only standard.

Many disagreements are not failures of calculation, but disputes about what should be counted. Person A calculates that one job will pay more over ten years. Person B understands this perfectly and still chooses the lower-paid job that permits him to care for family. He has not miscalculated. He wants a different life. If a fact is wrong, correct it with facts. If reasoning is wrong, identify the step. If values differ, state the reasons. “Your IQ is low” skips all three.

AI will be the first to show this reasoning wrong on its own terms. Today’s AI already calculates more quickly and accurately than most people on many tasks. According to the logic that IQ is the only standard, whenever AI and a person disagree, the low-intelligence party is the person and should obey without thought.

But in my essay on AI proving mathematics, I said that I can accept a proof it submits not because it is intelligent, but because each line can be checked. The same applies to disagreement among people. We should examine reasons, not who is more intelligent. However great the difference in ability, it cannot purchase the right to silence someone else.

Even if one person truly is more intelligent, this shows only that some difficult problems are better entrusted to him. It does not show that another person’s words do not count or that his life may be arranged by the more intelligent person. I have written about alpha males and “beauty is justice.” This is the third ladder: strength, appearance, and intelligence. The traits change; the ladder remains the same. Each uses a rankable ability to place someone else below.

Another person’s disagreement does not prove he is stupid. Even if he is less intelligent, that does not prove he is wrong this time. Even if he is wrong this time, that does not prove he is beneath you.

62. Why Some People Proclaim Democracy While Instinctively Flattering Power

This is nothing new, and Lu Xun offered an answer long ago. In The True Story of Ah Q, when Ah Q hears that revolutionaries have entered the city, he begins shouting about revolt. Lying in the Earth-God Temple at night, he imagines revolution as bringing silver ingots, foreign coins, foreign-style shirts, and the young scholar’s wife’s Ningbo-style bed into the temple. “Whatever I want will be mine, and whoever I fancy will be mine.” Ah Q hates Master Zhao, but does not want a world without Master Zhao. He wants a turn at being Master Zhao himself. Many people who proclaim democracy while flattering power are like this. They want to change who sits above, not question why the person above should be unconstrained. They can fiercely oppose one ruler and eagerly attach themselves to another.

This is not only the flaw of a fictional character or one individual. Research finds it too. In a 2020 experiment with American voters, Matthew Graham and Milan Svolik found that when a candidate violated democratic principles, only a small share of voters truly refused to support their own side. The more intense partisan conflict became, the smaller that share was.

In 2023, Suthan Krishnarajan conducted survey experiments in the United States and twenty-two other democracies and found an additional reversal. People described policies they disliked as undemocratic even when they were advanced according to the rules, while finding democratic language for policies they liked even when rules were broken. The order was inverted: first decide who is good, then find a name for what the good person did. When an opponent bypasses oversight, it is abuse of power. When one’s own side bypasses the same oversight, it is decisiveness.

Fear also matters. In 2023, Alia Braley and colleagues found that the more Americans believed the other side would violate democracy first, the more willing they were to violate it themselves. When told that the other side followed the rules more than expected, their own willingness to follow the rules rose. “They will not follow the rules, so we must break them first” sounds like self-defense. It is often the first step toward surrendering constraints. The danger may be real, but we must still ask for the evidence, the scope and duration of the response, and whether the power will be withdrawn after the danger passes.

Flattery often arises from dependence. When assistance comes through institutions, we can be grateful and still complain. When it comes through the favor of a powerful person, loyalty becomes the repayment. We may not even dare object when he treats someone else unjustly.

Nor should all deference be interpreted as sincere worship. Timur Kuran has written about preference falsification: under pressure, people publicly profess what they do not believe. Someone who does not dare oppose may be protecting his livelihood and family. After all, work and life can be considered separately. Actively helping the strong tread on the weak is not the same as lowering one’s head under threat. If we genuinely want people to show courage, they must first be able to live without favors and object without seeing their lives destroyed.

Another version praises democracy while despising the people. It treats ordinary voters as stupid: if they vote correctly, they are awakened; if not, they no longer deserve to choose. In the essay on intelligence, I wrote that no difference in ability purchases the right to silence others. Such people are not awaiting democracy, but a sufficiently intelligent and correct parent.

I often say that constructions cannot be closed. Every side in power has blind spots, including the side I prefer. Democracy therefore does not mean that my side wins or that good people rule. It means that whoever sits above must allow questions and answer them. The test is straightforward. If a power you want to give your own side passed to the person you most despise, would its scope still seem appropriate? If your side lost a majority in a fair contest, would you persuade more people or find a way to prevent them from voting again? As Joe Biden famously put it, you cannot love your country only when you win.

Ah Q’s problem is not that he wants revolution, but that he thinks, “Whoever I fancy will be mine.” Many people can oppose power they dislike. Democracy depends on one further step: permitting the power one likes to be questioned too.

63. Why AI Thought Surveillance Still Cannot Sustain Rule Forever

My friend probably reads a great deal of science fiction. This sounds like technological mind reading. If the First Emperor of Qin had possessed it, perhaps he would not have burned books and buried scholars. I am joking.

But we should not rush to reassure ourselves. A 2023 study in the Quarterly Journal of Economics titled “AI-tocracy” examined facial-recognition procurement by local governments in China. Places that experienced social unrest purchased more of it. After the purchases, later protests did decline, and the AI companies receiving the contracts produced more innovation in both government and commercial markets. Surveillance works. It really can make rule last longer. The question is not whether it works, but why effectiveness still cannot purchase eternity.

Actual mind reading is not as near as people imagine. In 2023, researchers at the University of Texas developed a decoder that could reconstruct, in broad terms, a story someone was hearing from brain scans. But it had to be trained for the particular individual with that person’s active cooperation, and failed when used on someone else. Subjects could defeat it simply by silently naming animals or inventing another story. Today’s “AI mind reading” usually means inferring how someone might think from behavior and relationships. Inference is not knowledge. Yet this is precisely the danger: the machine has not understood the person, while power treats its inference as a conclusion from which no appeal is possible.

Someone tried the real answer before AI. In 1989, East Germany’s Ministry for State Security, the Stasi, employed 91,000 official staff and more than 170,000 informants. Laid end to end, its files extended 112 kilometers. It knew a great deal about people’s minds. On August 31, 1989, Minister Erich Mielke asked regional leaders in an internal meeting whether another June 17 might erupt the next day, referring to the workers’ uprising of 1953. A colonel answered, “Not tomorrow. It will not happen. Is that not why we are here?” Four days later, the Monday demonstrations began in Leipzig. Late on November 9, at the Bornholmer Street crossing in Berlin, Harald Jäger, a Stasi passport-control officer with more than twenty years of service, watched the crowd grow. Fearing deaths, he opened the gate on his own authority without waiting for orders from above.

Three things become visible. First, knowing is not solving. Knowing why someone lacks food does not feed him. The Stasi recorded dissatisfaction in exquisite detail without reducing it at all. Second, information reaching power does not mean power will permit the information to correct it. The more everyone must appear correct, the less subordinates dare report bad news, a phenomenon economists call the dictator’s dilemma. Even if AI bypasses concealed reporting, what happens when the report says the problem originates at the top? Can it change the decision? Third, even the densest surveillance depends on people to carry it out. In the end, rule depends on whether someone like Jäger will remain at the gate.

What makes AI especially alarming is the third point: it may reduce the number of Jägers required. But even if a machine replaces the gatekeeper, the problem merely moves upward. The political scientist Milan Svolik distinguishes two problems: how rulers control the ruled, and how a ruling coalition divides power internally. AI may help with the first and cannot solve the second. Who decides whom the AI watches? Who can change its standards of judgment? When the person at the top is gone, who takes over? More accurate knowledge of ordinary people’s thoughts cannot answer these questions. I say that constructions cannot be closed. A surveillance system is also a construction, and what it sees least clearly is the hand holding it.

Nor can we leap from this to another reassurance, as though rule must fail whenever people refuse inward assent. A regime can last a long time without anyone’s genuine agreement. No law of history is counting down on our behalf. Human dignity therefore cannot be wagered on a machine’s inability to calculate us correctly. Even if someday it knows exactly what I think, it does not thereby acquire the right to decide for me. Even if it knows how to make me obey, it has not proved that obedience right. In the preceding essay, I said that democracy depends on permitting even the power one likes to be questioned. Applied here, no side in power should possess the authority to turn the human mind into a resource of rule.

Surveillance can make people afraid to speak. It cannot make problems cease to exist. It can prolong rule, but cannot turn rule into truth.

64. Why We Assume Others Will Love What We Love

This is more than simple wishful thinking. A psychological principle lies behind it. In 1990, Stanford psychology graduate student Elizabeth Newton conducted an experiment. One person tapped the rhythm of a familiar song, such as “Happy Birthday,” on a table while another tried to guess it. The tappers predicted that listeners would identify half the songs. In reality, listeners guessed only 2.5 percent of 120 songs. What made the difference? The tapper heard the melody continuously inside his head. The listener heard only a sequence of knocks.

Liking something works the same way. You love a song because it accompanied you through a difficult period. When the introduction begins, you hear more than the melody: the circumstances of that year, a particular person, and the self who eventually found a way out. You send the song to a friend and say how moving it is. The friend listens and says, “It is all right.” You heard the same recording, but he cannot hear what is playing in your mind. The same is true of liking a person. You find someone warm and loyal and think no one could dislike him. But other people have lived through different encounters with him.

The psychological concept is called the false-consensus effect. In 1977, Stanford psychologist Lee Ross and colleagues asked students whether they would walk around campus for thirty minutes carrying a sandwich-board advertisement reading “Eat at Joe’s.” Students who agreed estimated that around 60 percent of their classmates would also agree. Those who refused estimated that most would refuse with them. Both sides believed themselves the majority. The reason is not complicated. When judging others, we begin with ourselves as the reference point and adjust, but usually not enough. Research by Nicholas Epley and colleagues also found that the less time people have, the more they guess from themselves. Our circles also tend to contain people with similar tastes. Shared preferences in a small group, seen long enough, come to resemble universal common sense. When someone listens politely and says nothing, silence can easily be counted as agreement.

There is something humane in this. Recommending a beloved book, film, or dish often means: here is a place important to me, and I hope you will enter it too. If the other person also loves it, the happiness comes not merely from a successful recommendation, but from having met there.

The trouble begins with the next step. Your friend tastes the dish and says he does not like it. You answer that you will try another restaurant next time. This was merely a mistaken guess. But if you say, “I brought you here especially. How can you dislike it? You do not know how to enjoy life,” the situation has changed. He is no longer simply tasting a dish. He must supply the correct reaction and prove your judgment, effort, and relationship. In the essay on intelligence, I said that someone’s disagreement does not prove he is stupid. The same applies here. His failure to be moved does not prove that he fails to understand.

In the essay on empty space, I wrote that an author can arrange words, actions, objects, and rhythm, but the feeling itself can occur only in the reader. To write out even the feeling a reader ought to have is to feel on the reader’s behalf, which is colonization. Recommending something follows the same principle. I can play the song for you with care. To demand that you be moved exactly as I was is to feel for you.

If we truly want to know what another person thinks, only a clumsy method works. In 2018, Tal Eyal, Nicholas Epley, and colleagues conducted twenty-five experiments. Asking people to put themselves in another’s position did not consistently increase accuracy. Asking the person directly did. Instead of saying, “This film is so good that you are certain to love it,” we might say, “This film made me think about many things. How did you feel after watching it?” The second sentence leaves room for an answer we do not yet know.

Nor does another person’s dislike cancel your love. I can play you the song, but the song you hear can only be your own.

65. Why Self-Improvement Can Become Spiritual Colonization

I have long held this view. What troubles me most is the long essay Dan Koe published on X on January 12 this year, “How to Fix Your Entire Life in 1 Day.” It received 150 million views in a week, and Chinese translations circulated everywhere. I acknowledge that he is not an ordinary self-help salesman. His essays contain genuine insights, and he has warned that many people leave one matrix only to enter another. I am not saying that every sentence merely sells anxiety. I am discussing the structure that emerges when his most popular essays are read together.

The first step in Koe’s structure is actually opposition to colonization. He says that you are born into a culture with built-in beliefs, values, and standards, and that you outsource certainty to society, parents, teachers, government, and religion. This is not false, and readers feel released from a constraint. I see no problem with this step.

At the second step, however, he installs a different set of standards in the reader’s mind. In another essay, he lists seven standards he established for himself. One is to have $100,000 in the bank; anything below that amount is a problem to solve. He then asks you to change what you consume and rewire your brain. When you fail to reach the standard, you will feel pressure, which he describes as your new identity fighting for survival. The standard is first handed to you, then installed as identity, and finally uses identity to generate pressure.

In my theory of rights, I use colonization in a narrow sense: a transgression disguised as cultivation. Once, others told you what counted as a good life. Others still tell you, but now you believe the voice is your own. Psychology calls this introjection: moving external demands inside and driving oneself through guilt and self-rejection. Moving pressure from outside to inside does not make it one’s own. In the essay on democracy, I wrote that Ah Q hates Master Zhao but does not want a world without Master Zhao; he wants a turn at being Master Zhao. Koe opposes the colonization of the public by society, parents, and teachers. What he wants is not an end to colonization, but a change in colonizer. His return is traffic.

His instrumentalization is even more explicit. One essay is titled “How to Productize Yourself.” The life you pursue becomes a brand; the problems you solve along the way become content; the system that helps others follow becomes a product. He also says that the only choice is to combine life purpose with profit and that entrepreneurship is the only logical option for long-term thinkers. As a business strategy, this is unobjectionable. As a philosophy for the whole of life, it turns the person into material.

His most popular essay contains the sentence that once one sees clearly, every habit that does not advance the goal becomes disgusting. Sitting with a friend for an afternoon and reading a book unrelated to anything become things to despise. In the essay on alpha males, I said that under such rankings even the alpha becomes a scorecard that must always rise. The same happens here. Only now he scores the card himself.

Koe asks readers to write down their true enemy and specifies that it is not their environment or other people. As a single exercise in seeing what one can change, this may be useful. As an explanation, it compresses public problems into personal cultivation. Someone whose wages are withheld may change his own response and also demand the money. The second is not making excuses outside oneself.

I am especially uncomfortable with his treatment of stopping. He writes that people of low intelligence become trapped in problems, abandon the effort when they meet an obstacle, and offers writers who give up before building an audience as an example. In the essay on intelligence, I said that another person’s disagreement does not prove stupidity. The same applies to quitting. Someone may quit because he lacks the ability. Another may have tried, seen clearly, and decided he no longer wants traffic to determine what he writes. Leaving can be judgment at work. A person’s agency includes both “this is worth my effort” and “this is no longer worth it.” Recognizing only the first binds someone permanently to one road.

His business model should also be noted. According to an overview by a Chinese technology publication, the viral essays are only the entry point. Revenue comes from paid subscriptions, books, and an AI tool he helped found. Charging money does not prove malice. But a business sustained by readers feeling that they are still just short of enough has no incentive to tell them that they were worthy all along.

What I dislike most is not the standard he offers, but the way it is delivered. Even if $100,000 is a reasonable savings goal, and even if spending less time on one’s phone and more on action is right, his method begins by making you feel bad. Failure to reach the target should create pressure; habits that do not advance the goal should disgust you; then this pain should push from within. In the essay on empty space, I said that writing the feeling a reader ought to have is feeling on the reader’s behalf, which is colonization. Koe goes one step further. He does not feel for you. He installs a system of feeling in your mind and lets it push you on his behalf. Whether the content is correct is one question. Bypassing judgment to manipulate emotion is itself colonization. Being right does not cleanse the method.

The test is simple. A reader understands everything Koe writes and ultimately chooses not to build a personal brand, not to turn a hobby into a product, and not to keep increasing income. Does the system still recognize him as someone seriously living his own life? I say that cultivation also carves, but in a direction that makes people better able to judge for themselves, not more like the author’s preferred shape.

Human dignity is not something you receive after repairing yourself. It is already there.

A Friend Asked, Volume II: Living as an End有朋友问 · 卷二：作为目的而生活

Author's Note

Translator's Note

Contents

General Preface

Writing as a Form of Prayer

Volume II: Living as an End

Part I: Inner Life

1. Hope Without Expectation

2. If No One Knew, Would We Still Remain Moral?

3. Why Do We Understand the Principle Yet Still Fail to Act?

4. Why Do We Fear Death?

5. How Can We Know That We Are Right?

6. Why Does “I Only Want What's Best for You” Feel So Uncomfortable?

7. Why Are We Still Lonely When Surrounded by People?

8. Why Has Listening Become an Expensive Commodity?

9. What Is the Meaning of Life?

10. Should Children Practice Speed Arithmetic?

11. Why Does a Path Everyone Praises Begin to Feel Wrong?

12. What Is Love?

13. Was Confucius's Journey the Chinese Odyssey?

14. How Should We Face “If Only I Had”?

15. Why Do We Keep Changing the Measure of Our Worth?

16. Should We Forgive?

17. Can a Well-Intentioned Lie Be Right?

18. Should We Help an Elderly Person Who Collapses?

Part II: Knowledge and Consciousness

19. Why Is AI Producing So Many Mathematical Breakthroughs?

20. Is Human Consciousness Truly Beyond AI Simulation?

21. Why Has Technological Progress Not Made Us Happier?

22. Why Does Academia Resist AI?

23. Why We Live in the Dao Even If We Have Never Read Laozi or Zhuangzi

24. Why Creating Panic Is a Shortcut to Power

25. Why Philosophy Is the Mother of All Sciences

26. Why “Alienation” Is Everywhere in the Age of AI

27. In the Age of AI, Can More Than a Decade of Education Be Compressed into Four Years?

28. Why We Are Always Waiting for a Great Person

29. Why We Want an Extraordinary Life but Keep Repeating Other People

30. Why Death May Not Be the End of Everything

Part III: Education, Freedom, and Relationships

31. What Fear-Based Education Leaves in a Child

32. Why People Use AI to Make Games Instead of Simply Playing Them

33. Why Variants of Egoism Are Flourishing

34. Why Farewell to Revolution Has Endured

35. Why Freedom Depends on Both Being Able Not to Act and Being Unable Not to Act

36. Why Excessive “Freedom” Can Lead to Servitude

37. Why a University Would Conceal Sexual Assault in a Fraternity

38. Why People Blame Victims and Protect Perpetrators

Part IV: Power and Public Life

39. Why People Are Left Only with Suffering When Those in Power Speak of Morality

40. Why Smit Machchhar Is a True Hero

41. Why a True Philosopher-King Would Relinquish Power

42. Why Disasters Give Rise to Great Literature and Philosophy

43. Why Leviathan Remains a Classic

44. Why Capital Can Co-opt Academia

45. Why “Persons Are Ends in Themselves” Remains in the Ivory Tower

46. Why Philosophy and Politics Are Disconnected on “Persons Are Ends in Themselves”

47. Why AI Companies Exercise Restraint While Racing Ahead

48. Why Public Discussion Depends on Education

49. Does AI Have Consciousness? The Disagreement Between Two Judgments

50. Why Today’s AI Still Has No Consciousness

51. Why People Keep Asking Whether the World Is a Simulation

52. Why Neither Materialism nor Idealism Can Defeat the Other

53. Why AI Solving Difficult Problems Makes Mathematicians Reconsider Mathematics

54. Why Abusing Cats and Circulating Videos of It Are Two Different Evils

55. Why Writing Needs Empty Space

56. Where Are the Limits of Compulsory Social Insurance?

57. Why Both Capitalism and Communism Are Transitional

Part V: New Forms of Instrumentalization and Human Resistance

58. Why AI Video Has Unleashed an Explosion of Creativity

59. Why Some People Take Pride in Being an “Alpha Male”

60. Why “Beauty Is Justice” Became Popular

61. Why Some People Treat Intelligence as the Only Standard

62. Why Some People Proclaim Democracy While Instinctively Flattering Power

63. Why AI Thought Surveillance Still Cannot Sustain Rule Forever

64. Why We Assume Others Will Love What We Love

65. Why Self-Improvement Can Become Spiritual Colonization

作者说明

目录

总序

A Friend Asked, Volume II: Living as an End
有朋友问 · 卷二：作为目的而生活

卷二　作为目的而生活

第一辑　内在生活

1　不望之望

2　恶行若无人知晓，人还会守住道德吗

3　为什么道理都懂，却还是做不到

4　人为什么怕死

5　我们如何知道自己是对的

6　为什么“我是为你好”令人不适

7　为什么身边都是人，我们仍然孤独

8　为什么倾听会成为昂贵的商品

9　人活着的意义是什么

10　孩子值得练习速算吗

11　别人都说好的路，为什么越走越不对劲

12　什么是爱

13　孔子周游列国，是中国版的《奥德赛》吗

14　如何面对“如果当初”

15　人为什么总在更换衡量自己的尺子

16　我们到底该不该原谅

17　善意的谎言可以吗

18　老人倒地，我们该不该救

第二辑　知识与意识

19　AI为什么密集取得数学突破

20　人的意识真的无法被AI模拟吗

21　科技进步为什么没有让人更幸福

22　学术界为什么抵触AI

23　为什么没读过老庄，也活在“道”里

24　制造恐慌为什么是获取权力的捷径

25　哲学为什么是所有科学之母

26　“异化”为何在AI时代无处不在

27　AI时代，十几年教育能否压缩为四年

28　我们为什么总在等待伟人

29　为什么想活得精彩，却总在重复别人

30　为什么死亡之后未必什么都没有

第三辑　教育、自由与关系

31　恐惧式教育给孩子留下了什么

32　为什么人们不玩游戏，反而用AI做游戏

33　利己主义的变种为何大行其道

34　《告别革命》为什么经久不衰

35　自由为何取决于能不做与不得不做

36　过度的“自由”为什么会走向奴役

37　大学为什么会掩盖兄弟会性侵

38　人们为什么责怪受害者、保护施暴者

第四辑　权力与公共生活

39　当权者谈道德，民众为何只剩苦难

40　Smit Machchhar为什么是真英雄

41　真正的哲人王为什么会放弃权力

42　灾难为何催生伟大的文学与哲学

43　《利维坦》为什么至今仍是经典

44　资本为什么能够收编学界

45　为什么“人是目的”停在了象牙塔里

46　哲学与政治为何在“人是目的”上脱节

47　AI公司为什么一边克制，一边竞速

48　公共讨论为什么离不开教育

49　AI有没有意识：两种判断的分歧

50　为什么今天的AI还没有意识

51　为什么人们不断追问世界是否是模拟

52　唯物主义和唯心主义为何谁也赢不了谁

53　AI攻克数学难题，为何让数学家反思数学

54　虐猫与传播虐猫视频为何是两种恶

55　写作为什么需要留白

56　强制社保的边界在哪里

57　为什么资本主义与共产主义都只是过渡

第五辑　新工具化与人的抵抗

58　AI视频为什么带来创造力爆发

59　为什么有人以“Alpha Male”为荣

60　“颜值即正义”为什么流行

61　为什么有人把智商当作唯一标准

62　为什么有人高喊民主，却本能谄媚权力

63　AI思想监控为何也无法永久维持统治