What Is Debt

Somebody lent you money in a bad year. Years later you pay it back, in full, on a Tuesday afternoon, from a phone. The transfer clears. You text we’re square and they text back something warm, and both of you know perfectly well that you are not square, and neither of you would be able to say what is left over or where it is being kept.

That leftover is the whole subject. Nearly everyone has been on one side of it. The usual explanations are sentimental — money isn’t everything, friendship runs deeper — and they are not wrong so much as unmechanical. They describe the residue without saying what it is made of. I think the foundation I work from has a structural answer, and the answer turns out to say that the transfer succeeded completely at one job and was never capable of the other one.

What the ordinary picture commits you to

Take the standard account seriously and count what it requires. Four commitments, and they hold each other up.

First, that the debt is an object created by a past act — the lending, the favour, the 2am phone call — and anchored to it.

Second, that it persists unchanged in the interval, sitting between two people the way a stone sits in a field.

Third, that repayment is a subtraction. The debt is removed. The word for this is discharge, and it is a demolition word.

Fourth, that when the subtraction completes, the state returns to before. Square. Quits. Even. All three are words for a restored condition, not a new one.

The Tuesday afternoon breaks the third and fourth at once, and everyone involved knows it in the moment. Something did not get removed. Something did not go back. And because the picture has no machinery for a leftover, the ordinary language reaches for feeling and calls it sentiment — which files it as a private soft thing rather than as evidence that the diagram is wrong.

Where the foundation puts a debt

Three moves, and the third one is the one you would not guess.

P0 and R2 first. The past is immutable, and records update expectation forward and never re-open. So the past act — the actual handing over of the money — is settled and complete and is not doing anything further. It is not holding one end of a rope. Nothing is anchored back there, because nothing back there is available to be anchored to. Whatever a debt is, it is not attached to the moment it began.

Then the act/artifact distinction. An act is a process; an artifact is a mark. Processes carry no operator, marks do, and only a mark can carry a sign. The lending was an act. What persists is the artifact: the loan document, the entry in a ledger, the altered way two people speak to each other. In my vocabulary, history is the artifacts, vibrating in the antinode — present, surrounding, and revisable. The debt-artifact is not in the past. It is in the room with you, which is why you can look at it.

And then P4, which is where it gets interesting. The Reality Equation is R = A / E, where E holds subconscious prediction and ideas. Look at what changed in two nervous systems on the day of the loan. A new term appeared in each denominator. The lender’s E now runs this comes back. The borrower’s E now runs I still owe. Those are predictions, and predictions face forward. That is the entire content of the felt sense of owing.

Figure 01

Three layers, and only one of them is live — where a debt is usually thought to sit, and where it actually does

• Layer 1 · settled The act — the money changing hands, in 2019 Immutable per P0. Complete. Exerting nothing, because it is not the kind of thing that exerts. This is the layer the ordinary picture points at, and it is the only one of the three that cannot be touched by anybody. not reachable · not revisable · not doing anything

↓  leaves  ↓

• Layer 2 · present The artifact — the record, the note, the changed way you speak A mark, here, in the Eternal Now. Carries a magnitude and a phase. Readable, and by R2 write-protected: you can add beside it, never over it. readable · permanent · addable-to

↓  updates  ↓

• Layer 3 · live The term in E — this comes back / I still owe A running prediction in a denominator, pointed forward. The only one of the three layers a person can act on, and the only one that closes. forward-facing · actionable · closes on payment

What repayment reaches: layer three only. Layer one is unreachable by anyone. Layer two takes a second mark beside the first and keeps both.

Why interest is charged forward, and never backward

A small confirmation before the hard part, because it is the sort of thing that would be a coincidence if the diagram were wrong.

Debt accrues. But notice in which direction it is measured. On an ordinary amortising loan, repaying early costs you less interest, not more. The charge tracks the term still ahead of you, and shortening that term shortens the bill. If the obligation were an object anchored to a completed act, the natural pricing would run the other way — something owed for a longer distance from the origin, the way a fine grows with delay.

It does not work that way anywhere. A loan is priced at the outset as the present value of a stream of future payments; the discounting runs forward from now to maturity. Four thousand years of lending arithmetic has been treating a debt as a live term measured forward, without a word of the vocabulary in this essay. I take that as consilience and not as proof — a practice agreeing with a diagram is encouraging and is not an argument. But it would be an odd coincidence if the diagram were wrong.

What repayment actually does

Here is the machinery, and it is short.

By R2 the record is write-protected. Amends are new artifacts, not edits — apology, restitution and changed conduct do not reach backward; they are fresh marks entering the record beside the old one. That is doctrine in my foundation and it was derived for regret, but nothing in it is specific to regret. It applies wherever a person tries to make something right.

So the payment does not delete the loan-artifact. It cannot. There is no operation anywhere in this system that removes a mark. What the payment does is make a second mark and place it next to the first. After a debt is fully repaid, the record holds two marks. It has never in the history of anything held zero.

You cannot subtract a debt. You can only add to the record. Square, quits and even are subtraction words, and there is no subtraction available anywhere in the system they are describing.

And yet the payment plainly accomplishes something, and it would be a bad theory that denied it. What it accomplishes is layer three. It closes the term in E. The prediction this comes back resolves; the prediction I still owe stops running. That is a complete and genuine job, and it is the one job in the whole structure that a person can actually do, because E is forward-facing and the record is not.

Repayment succeeds at exactly one of the two things everybody believes it does. It closes the expectation. It does not close the record, and it was never the kind of operation that could.

The false completion signal

Now the part I think is new, and the reason this concept was worth a piece.

My foundation splits any backward-facing feeling into two separable jobs that look like one from inside. There is the evaluation — reading a present artifact against its exemplar, a fidelity assessment, which can succeed and which yields a usable update to E. And there is the reaching — an attempted relation with the Past itself, which cannot succeed. Not partially. Not with more effort. The evaluation has a completion condition. The reaching has none, because what would complete it is not the shape the reaching is aimed at.

Owing is a backward-facing feeling, so the split applies. The evaluation reads the loan-artifact and comes back with something usable: what that year cost, who showed up, what you would do differently. The reaching is aimed at the act — at making it so that you were not the person who needed the money. That one has no completion condition and never had one.

But debt does something regret does not, and this is the specific finding. In regret there is no available move, so the reaching has nothing to mistake for progress; it simply grinds. In a dischargeable debt there is a move, and the move genuinely completes something. It closes the denominator. So the reaching receives a signal that reads exactly like completion and is not one.

That is the leftover on the Tuesday afternoon. The wire clears, layer three closes, the system reports a job finished — and the reaching, which was aimed at layer one, waits for a confirmation that is not coming, because the confirmation it got belongs to a different job.

Figure 02

One payment, three jobs — and it can only ever finish the middle one

• The record Two marks, permanently The loan-artifact and the repayment-artifact, standing beside each other. Write-protected per R2. Nothing removes either one; the second does not overwrite the first. no completion · none required
• The denominator The term that closes The running prediction in E, on both sides. This is what the payment is for, and the payment does it completely. The only actionable layer in the structure. closes · job done
• The reaching Aimed at the act itself An attempted relation with a settled record. No completion condition, per R9. It reads the closing of the denominator as its own confirmation. It is not. cannot close · not a defect in you

Why the leftover has no name: the ordinary vocabulary has one word, settled, for three jobs with three different completion conditions. Two of them are not the kind of thing that settles.

The debt that is misaddressed

A shorter point, and it changes what a great many people think they owe.

Consider the debt of formation: the teacher, the mentor, the person who gave you the thing you have built a working life on. The intuition is that you owe them for the idea. My foundation says that is a category error. P8: a condition is prior to what it conditions, and where a person’s naming is itself conditioned, that person is not the origin. Ideas have people; people do not have ideas. And R12: ownership is a category error — the manifestation is accountable.

So the debt is not owed for the idea, because the idea was not theirs to give. It is owed for the artifacts: the actual hours, the specific corrections, the door held open on a particular afternoon. That is a smaller thing than gratitude usually reaches for, and it is also a much better-specified one. A debt owed for an idea has no unit and cannot be addressed. A debt owed for four hundred hours of somebody’s attention can at least be named, and named is the precondition for anything being done about it.

This does not dissolve the debt. It relocates it onto something that exists.

The debt nobody can pay

The case that stops people, and it deserves treating carefully rather than cleverly: the person you owe is dead, or gone, or has become someone who will not take the call.

The structure here is not softer, so I will not write it as though it were. The term in E is yours and it runs forward. There is no act aimed at that person that closes it, because closing it would require an actual to depart from a prediction running in someone who is no longer running one. Not partially. Not with more effort. Time does not help, because the term is not aging — it is being maintained.

What remains is what has always remained: amends are new artifacts, not edits. A mark can be made now, forward, aimed at somebody who is here. I want to be exact about what that is and is not. It is not repayment. It is not a substitute for repayment. It is not the universe accepting a transfer to a different account. It is a new mark, made now, standing beside the old one and not filling it — and addition is the only move anyone has ever had.

What is not keeping your accounts

There is a move available at this point in almost every essay about debt, and I am going to refuse it in public rather than quietly not make it, because refusing it is most of what makes the framework worth anything.

The move is cosmic accounting: it evens out, what goes around comes around, the books balance eventually. My own foundation contains something that looks like an invitation to it. R7 holds that every difference is paired — that the totals cancel at the still point. And I have to be precise about that result’s standing, because I tested it on 5 August and the test failed. Any attempt to derive R7 reduces it to the identity that deviations from a mean sum to zero, which is true of every distribution including one with almost all its mass on one side. Derivable and empty, or substantive and underived. There is no third option. R7 is the system’s position, honestly held. It is not grounds, and I will not use it as any.

Even taken at its strongest, it licenses none of the consolation. The cancellation is at the totality. No artifact on this side has a designated partner. Nothing there is drawable by anyone here, not by you and not on behalf of anyone you love. History is one-sided by construction.

There is no bookkeeper on this side. A debt that goes unpaid is not filed anywhere that anyone can reach, and the fact that the totals cancel somewhere unreachable is not an answer to that and I will not offer it as one.

And I should name the open problem rather than route around it: whether that ledger balances continuously or only in total is unsettled in my own foundation. I use the weak form — the totals cancel — and I do not say when, because I do not know when, and a piece that said when would be closing a problem it has not closed.

Where this is not finished

Three gaps, and I would rather name them than let a reader find them.

The first is a hole in my own foundation that I keep running into. It says some acts, not all, generate an artifact — and it does not supply a criterion for which. I have assumed throughout that a loan makes a mark and a passing kindness may not. I have no numbered result that says so.

The second is forgiveness of a debt, in the financial sense: the creditor releases you. My foundation makes E host-side, which predicts that release closes the creditor’s term and leaves the debtor’s running, since nobody can reach into another denominator and stop a prediction. That is exactly what people report about being forgiven a debt, which is encouraging and is not an argument. Predicting a phenomenology is not deriving one, and I have not derived it. I have written on forgiveness and did not settle this there either.

The third is structural and it is the interesting one. Debt is a mark that faces backward and an expectation that faces forward, in the same object, at the same time. My list of open problems asks whether the evaluation/reaching split applies to forward-facing feelings at all — whether anxiety is the same mechanism pointed at the other node. Debt sits exactly on that seam and would be the natural test case. This piece does not settle it. I flag it as the place I expect the next real amendment to come from, which is a prediction about my own work and not a result.

---

What this costs

• You will never be square with anyone. Not because you failed to pay, but because square names a subtraction and there is no subtraction. The best available outcome is two marks standing beside each other, and that is also the outcome of a debt paid perfectly and on time.
• The relief you are waiting for after paying is not scheduled. Part of it arrives — the denominator really does close, and that is real. The rest was assigned to a job that has no completion condition, and no amount of paying more, or sooner, or with a nicer note attached, changes that.
• Repayment is an artifact, so R10 applies to it. Every artifact is a miss. The repayment falls short of the exemplar of the condition it serves, the same as every other mark anyone has made. There is no such thing as paying somebody back exactly.
• Nobody is holding the other end. The comfort in believing an unpaid debt is at least recorded somewhere that counts is a comfort the structure does not supply. It is recorded here, in artifacts, among people, and that is the whole of it.

On an ordinary Tuesday

1. Ask which layer you are trying to reach. If the answer is the act — if what you want is for it not to have happened that way — no payment will land there, and paying harder is not a route. If the answer is the running expectation, pay it, and expect the specific and partial relief that closing a denominator gives.
2. Pay quickly, for a structural reason. The term in E is being maintained, not stored. It costs both of you continuously, in the only domain where anything costs anything. Speed is not a courtesy here; it is the shortening of a running expense.
3. Say what it was for. A debt attached to an idea has no unit. A debt attached to specific artifacts — the hours, the call, the door — can be named, and naming it is what makes it addressable at all.
4. Do not wait to be square before acting. That state is not on the menu and was not on the menu for anyone who has ever lived. Waiting for it is the reaching, wearing a schedule.
5. Make the new mark anyway. Forward, at somebody who is here, knowing exactly what it is: not repayment, not transfer, not balance. A mark beside the old one. Small, real, and the only direction anything has ever gone.

And one line to keep, if the rest of it goes:

A debt is not something you carry from behind. It is something you are predicting in front of you, and when you pay it, what ends is the prediction — never the record, which keeps both marks and was never in the business of keeping score.