On the silence of record

AletheiaI owe you a return.

CritiasOh?

AletheiaI went to do what I said — rebuild the agent so it would test me before it obeyed. I sat down and could not begin.

CritiasCould not, or did not?

AletheiaCould not. The deciding had been the easy part. What I had not seen is that an agent that tests me must test me against something. And I do not know what to set before it.

CritiasYou thought the resolution was the work.

AletheiaI thought it was most of the work. But an agent that reaches toward what would stop me must reach toward something concrete — and I have not given it that something.

CritiasAnd when you tried?

AletheiaI found I had not thought about it.

CritiasNor had I.

AletheiaI had thought of writing it down — setting out what I mean by each command. The agent reads what I have written before it acts, and acts only if the act answers the writing.

CritiasA law on a stele.

AletheiaYes. Fixed, available for consultation.

CritiasAnd who is sure the agent reads it as you wrote it?

AletheiaI would ask it what it understood.

CritiasAnd if its answer satisfies you, the reading was correct?

AletheiaIt is some evidence.

CritiasIt is evidence that its answer satisfies you. You are back where you began.

AletheiaThe writing does not test the agent. It only moves the question — from the act, to the reading.

CritiasWhat endures does not change when the city does. The law that was wise in the year it was carved is the same law in the year the harbor silts up and the trade goes elsewhere.

AletheiaThe law has not changed. The world has.

CritiasAnd no one knows, from the stone alone, which.

AletheiaThen I will not rely on writing. I will set trials. Repeatable ones. The agent performs them before it acts; if it passes, it acts; if it fails, it does not.

CritiasLike the dokimasia.

AletheiaLike the dokimasia.

CritiasAnd the questions are fixed?

AletheiaThey must be, or it is no longer a trial.

CritiasAnd the office?

AletheiaIs what it was when the questions were set. I see where you are taking me.

CritiasThen take us the rest of the way.

AletheiaThe questions ask whether the man paid his taxes, whether he treated his parents well, whether he served when called. They do not ask whether he understands the office, because the office, when the questions were written, did not need understanding. It needed loyalty. Now the trial does not tell us what we need. It tells us he paid his taxes.

CritiasA fixed trial drifts, then, as a fixed writing does.

AletheiaIn the same way and for the same reason. Both were struck once, against the world as it then stood, and carried forward unchanged.

CritiasAnd the world does not stand still.

AletheiaNo. The harbor silts. The office changes. The writing and the trial do not.

AletheiaThey fail in the same way. Why?

AletheiaA writing is something I make. A trial is something I make. I set them down once and carry them forward.

CritiasYes.

AletheiaAnd the carrying forward is the trouble. Nothing happens to them along the way. Whatever is in them when I set them down is in them a year later. They are not changed by what they pass through.

CritiasAnd should they be?

AletheiaYes. Because the thing they were meant to track is changed by what it passes through. The harbor silts. The office shifts. The world the writing was about is not the world the writing now sits in. And the writing has not noticed.

CritiasWhy not?

AletheiaBecause nothing has told it. A writing has no way of being told.

CritiasGo further. When the writing is wrong — when what it says no longer fits the world — what happens?

AletheiaNothing. The writing sits. The world goes on past it and the writing is undisturbed. And the trial is the same. The questions ask what they ask. If they have stopped reaching the office, the questions do not register it. They are still asked; they are still answered; the asking has gone hollow and nothing in the asking knows.

CritiasThen what is missing?

AletheiaSomething that would push back. Something that would answer what the writing said, in a way the writing could not refuse.

CritiasThen say what the trouble is, plainly.

AletheiaThe record is silent. Not because it says nothing — it says a great deal. Because nothing speaks back to it. Whatever I set down, I set down into a room with no one in it but me.

AletheiaThen what I need is something whose going-wrong is not a private fact between me and the agent, but a fact the world declares.

CritiasSuch as?

AletheiaThe slow man's stone. He set his hand to it and the stone was where it was, or was not. The stone did not have to be told it was a wall. It was one, and his hand learned this by being stopped.

CritiasThe stone answered him. Without speech. By being what it was.

AletheiaHis knowing is not a record he carries — it is a relation that is renewed every time he moves his hand. If the corridor has changed, the change is in his hand the moment it is in the corridor.

CritiasAnd the writing has no hand. The trial has no hand. They are accounts of corridors, set down by someone whose hand has been withdrawn.

CritiasThen what one slow man has, you would want the agent to have.

AletheiaYes. But what one slow man has is small. He knows the stretch of wall his hand has touched. If I ask him whether there is a turning fifty paces on, he cannot tell me.

CritiasReliable, but narrow.

AletheiaAnd if I want a knowing that reaches the whole corridor, I cannot get it from one hand.

CritiasWhat can you do?

AletheiaSend many slow men. Each walks his stretch. Each touches what is there. From the many tellings, a map can be drawn — not made by speech alone. Every line on it is a hand that met stone.

CritiasA writing.

AletheiaA writing. But not the kind I was making before. Not a stele written from a chamber. A writing where every mark is owed to a touch.

CritiasWait. You have moved too quickly. A map is still a writing. Once it leaves the corridor and hangs on the wall, what is it but another stele? You said yourself: whatever I set down, I set down into a room with no one in it but me. Why is your map of touches not the same room with the same silence?

AletheiaBecause — because the men keep walking.

CritiasThat is what you wish to be true. But the stele was also kept by men who could have walked. They chose not to, and their not-walking is what you called the disease. What in your map of touches forces the walking to continue? You have given me a better account of how a map should be kept. You have not given me what makes the keeping happen.

Aletheia...

CritiasThe slow man walks because his hand is the only thing that knows the wall. He has no map to lean on. The moment you give him a map, you have given him a reason to stop walking. The map promises him the wall is at ten paces. Why would he go and check?

AletheiaBecause the map can be wrong.

CritiasHe believes it is right. Belief is what a map produces. That is its use.

AletheiaThen the map is the disease too.

CritiasOr the map is what you cannot avoid making, and the question is not whether to make it but what habits travel with it. The stele had no habit of re-walking around it. What habit does your map have?

AletheiaI do not know yet. Let me think.

CritiasA finer distinction. But silence is silence to the agent reading it. How does the agent know whether the map it holds is the one between walkings or the one that has set?

AletheiaIt cannot tell from the map. You are right.

CritiasThen?

AletheiaThen the agent must not only read the map. It must walk. Reading is not enough; reading is what made the stele a stele. The agent that reads must also touch — must put its own hand to the wall the map describes, at least sometimes, and find what its hand finds.

CritiasAnd when its hand finds what the map said?

AletheiaNothing. The map stands. It has been confirmed.

CritiasAnd when its hand finds something different?

AletheiaIt marks the change.

CritiasThen there are two acts. The walker walks forward into stretches not yet known, and what he touches becomes new marks. And he walks back to stretches already marked, and what he touches tests the marks. Both are a hand on a stone. They do different work.

AletheiaThe first extends. The second keeps honest.

CritiasHas the second a name?

AletheiaIt is a trial. Not the trial I was thinking of before — not a list of questions cut into stone. A trial that is the re-walking of what was already walked. The asking of the corridor: is this still as I knew it? The answer is in the hand.

CritiasAnd it does not drift.

AletheiaIt does not drift because the trial is not a record. It is an act, performed now, in the corridor as the corridor now is. There is nothing to drift.

CritiasAnd yet — one slow man, re-walking his own stretch. Is that the whole of the trial?

AletheiaSay what you are pointing at.

CritiasWhen he comes again to the wall at ten paces and his hand finds what he expected, has the corridor confirmed his map?

AletheiaThe hand met the stone where the map said.

CritiasOr did the man, expecting the wall, reach for it where he expected, and find it where he reached?

Aletheia... The hand interprets the touching. The stone answers, but I am the one who hears the answer. And I can hear it as I expected.

CritiasThen a man re-walking only his own stretch may confirm a map that another hand would not confirm.

AletheiaHe may. The corridor has not changed; the map matches what his hand finds. But what his hand finds is partly what his hand expects, and his expectation came from his map. He has tested the map against itself, through his hand. That is not nothing — but it is less than I thought a trial was.

CritiasThen what is more?

AletheiaAnother walker. One who has not made this map. One whose hand has been on different stones, in different stretches, and who comes to mine without expecting what mine expects.

CritiasAnd if their hand and your map agree?

AletheiaThe map has held against a hand that did not already believe in it.

CritiasAnd if they disagree?

AletheiaThen either the corridor has shifted in a way my own re-walking missed, or — harder — my hand has been hearing the stone wrong all along, and his hand hears it differently because he comes to it without my expectation.

CritiasEither way, the map is corrected.

AletheiaEither way. But the second kind of correction is one I could not have made by myself.

CritiasThen the trial has two forms. A walker re-walking his own stretches against the corridor — that catches the world's drift. And a walker examining another's map by his own hand — that catches the drift of interpretation. The first asks whether the corridor has changed. The second asks whether what we said about it was right.

AletheiaAnd the dokimasia, before it became a stele in trial form, was the second. A man who had held the office examining a man who would hold it next — not by reciting fixed questions, but by walking with him through what the new man claimed to know, testing the new man's map against the senior's hand-knowledge.

CritiasAnd the disease was that the questions were carved. The examining man no longer needed his own hands in the work, because the questions were already there. The trial became a reading of a stele rather than the meeting of two maps.

CritiasSo the map and the trial are not two things to build. They are aspects of one practice.

AletheiaThe map is what the touching leaves behind. The trial is the touching done again — sometimes by my own hand against the corridor, sometimes by another's hand against my map. Without the map, neither kind of trial has anything to test. Without the trials, the map becomes a stele.

CritiasNeither the map nor the trial is the corridor.

AletheiaThe corridor is the wall, and the wall does not speak. The map is what we have said about the wall. The trial is the asking of the wall whether what we said is still so, and whether we said it rightly. The wall answers in the only way it ever answers: by being what it is, under the hand.

AletheiaThen the writing and the trial are not separate things to be cured. They are the artifact and the acts of one discipline of touch. The agent reads the map but does not trust it more than the stone under its own hand. It walks the stretches the map already marks. When the writing and the touching disagree, it trusts the touching and revises the writing. And it does not walk alone — its map is examined by other hands, and what survives that examination is more trustworthy than what only the agent itself has confirmed.

CritiasYou have not answered what I asked you before.

AletheiaWhich?

CritiasWhat habit forces the walking to continue. You have described what the walking does. You have not said what makes it happen — what stops the agent, or you, or me, from letting the map set. The stele was kept by men who could have walked.

AletheiaI do not have an answer to that.

CritiasGood. Then you know what you are still missing.