What Is Proof

Contents

1. Proof, stated fairly
2. Three cases the picture cannot hold
3. The neighbours
4. What the word is the name of
5. What a proof reaches
6. Why one record feels like a window
7. “Pics or it didn’t happen”
8. What it does not mean
9. What it costs
10. An ordinary Tuesday

Proof, as people ordinarily use the word, is what shows that something happened. You keep the receipt so you can prove you paid. You screenshot the message so you can prove what was said. You take the photo, and if you don’t, somebody at the table says pics or it didn’t happen, half joking. “I have proof” ends arguments — “the receipts” has become a word for exactly that — and an argument without proof tends to go to whoever sounds surest.

The picture underneath is a picture of looking back. The event is over, but it left a trace, and a good trace lets you see the event through it the way clean glass lets you see the street. A blurry photo is a smeared window; a forged receipt is a painted one; a clear, timestamped, unedited record is as close as anyone can get to having been there. The stronger the proof, the clearer the view. And where there is no proof at all, the event itself seems somehow to be in doubt.

This is not a foolish picture. It is right about the thing that matters most in practice: some records are much better than others, and it is worth knowing which. The question is what, exactly, a good record is better at.

Three cases the picture cannot hold

The seam is the same in all three. The picture puts the strength of a proof on the far side of the record, in the event. And in all three the strength of the proof moved — it fell when the till failed, it vanished when the clock was found out, it was absent from the start by the river — while the event stayed exactly where it was. Whatever a proof’s strength is a measure of, it is on this side. All three cases are composites, not reported events.

The neighbours

Several earlier pieces stand close to this one, and the resemblances should be named rather than left for the reader to find.

The piece on history took the same result into the archive room and found that a shelf of records looks like access because there is so much of it. That is a collection. This piece stays with a single record — the receipt, the photo, the screenshot, the one called decisive — and asks a narrower question: if abundance is what produces the appearance, how does one mark on its own produce it so strongly?

The pieces on memory and on a witness both land on the same sentence in different words: a memory is not a window; a witness is not a window. Both are marks in a person. Proof is usually what we bring in to check those marks — the external, impersonal record — and it is natural to think that it, at least, is a window. This piece says it is not, and asks the question those pieces did not need to ask: why it feels exactly like one.

The piece on agreement counted readings — heads in a room — and asked what their agreement is worth. Here there is one reader and one record, and the agreement that matters is among marks, most of which nobody in the scene is looking at. The piece on an ancestor owns the life that left no record at all, and what a missing mark does and does not mean; this piece’s third case comes close to that and stops short, because it needs only what pics or it didn’t happen says about the people at the table. And Proof of Person, an August essay outside this series, used the word in a different sense — proof that a particular person made a message, proof of work. This piece is about proof that something happened.

What the word is the name of

I work from a small set of numbered primitives and the results derived from them. The one doing the work today has two clauses: the Past is inaccessible, and the abundance of present records makes it appear otherwise (R9). The foundation derives it from the axiom and from the result that what cannot change cannot perceive or stand in relation to anything. The first clause says nothing reaches the Past. The second says why it seems otherwise, and it names a culprit: records — here, now, plenty of them.

Three more lines are needed. The foundation’s locked definitions make an artifact an imperfect historical mark, and history the artifacts themselves — present, surrounding, revisable. A receipt is a mark. So is a photograph, a screenshot, a log file, a signature on a delivery slip. They are here. Whatever they are marks of is over. The axiom, in its precise form, says the past is immutable as a constraint on records: the records admit exactly one consistent assignment, and no present operation can produce a different one (P0). And the quotient that a living reader actually meets is the actual over the expected, with the expectation term belonging to the reader (P4); records update that expectation forward and do not reopen anything (R2).

What a proof reaches

Put the first clause against the receipt. The receipt cannot reach the purchase, because nothing reaches the Past. So when the clerk checks it, what does she check it against? The till’s record of the sale — another present mark. If that is missing, the card statement, the tag still on the sleeve, the date printed at the bottom. Each check she can run is a comparison between things on her side of the counter. When enough of them agree, she stops expecting a problem, and the refund goes through.

Something was settled at that counter, and it was hers: her expectation about the sale. That is what proof does. It is a great deal. It is how money is returned, how accounts close, how people stop arguing at dinner. It is not a view.

The word’s older sense was already closer to this. Dictionaries trace proof to a Latin verb for testing, and the sense survives in a printer’s proof: a trial impression pulled today and read, line by line, against the manuscript beside it. Two present marks, compared now, and a correction made in the margin. That is an illustration, not a ground — the argument above stands without it — and the etymology is reported from standard dictionaries rather than from a source cited here.

Why one record feels like a window

Now the second clause. If a proof reaches nothing, why does a good one feel exactly like seeing? The foundation names the cause: the abundance of present records. In an archive room the abundance is on the shelves. But a single receipt has no shelf. Where is the abundance?

It is behind the receipt, and mostly out of view. The receipt is decisive because the clerk expects that the till, the card statement, the tag and anything else she could check would agree with it. She does not check them all, and she does not need to. That is what decisive means: the reader no longer expects another mark to disagree. The slip of paper does not carry the weight alone. It carries the expected agreement of every record it would have to agree with. The abundance the second clause names is really there — present records, plenty of them — but behind a single proof most of it sits in the reader’s expectation of what the other records say.

The slip of paper does not carry the weight alone. It carries the expected agreement of every record it would have to agree with.

And that is the window feeling. Looking at a street through clean glass is, among other things, the experience of not expecting a second look to show you something different. A decisive proof produces the same state by another route: expectation stops moving, because no further record is expected to move it. From inside, a settled expectation and a clear view are very hard to tell apart.

Case two shows the mechanism running backwards. The photograph was decisive for exactly as long as everybody expected the camera’s clock to agree with it. The moment one more present mark turned out to disagree, the photograph stopped being proof — at once, without anybody touching it, and without anything happening to the weekend.

Figure 01

What carries the weight in five ordinary proofs

The proof	What it is read against — all present marks	What it settles
A receipt at a returns desk	The till’s record of the sale; the card statement; the tag on the sleeve	The clerk’s expectation that the sale took place
A screenshot of a message	The thread on the other phone; the platform’s copy; what the sender says now	The group’s expectation about what was said
A dated photograph	The camera’s clock; other photos from that day; who else remembers the weekend	The family’s expectation about who was where
A signed delivery slip	The courier’s scan; the parcel on the step; the buyer’s account of it	The sender’s expectation that the parcel was delivered
“Pics or it didn’t happen”	No mark was made, so there is nothing to read against anything	Nothing about the hawk. The table’s expectation stays where it was

“Pics or it didn’t happen”

Now the hawk. Taken at its word, the sentence says that the record decides whether the event took place. The axiom rules that out in its precise form: the records admit exactly one consistent assignment, and no present operation can produce a different one. A photo produced at lunch, a photo not produced, a photo faked — each of these is something done or not done at a lunch table. None of them reaches the river.

So the sentence cannot mean what it says. What does it mean? Run the first construction. What a photograph would have settled is the friends’ expectation. Pics or it didn’t happen is a statement about the table: we will not update without a mark. It is a standard for changing your mind, dressed in grammar about the event.

As a standard it is often reasonable. Friends who believed every hawk story would be wrong a good deal. As grammar it is wrong, and the wrongness leaks. Said often enough, it teaches people to hear the unrecorded as the unreal — to treat I can’t show you as if it meant it may not have been. Her having no photograph leaves the catch exactly as it was, and leaves her account no more and no less accurate than it was when she told it.

What it does not mean

It does not mean proof is worthless, or that receipts are theatre. Keep receipts. A good record is the best instrument there is for settling expectation, and settling expectation is most of what anybody needs from the past.

It is not a claim about any standard of proof that a court, an employer or an institution should use. The foundation says nothing about how much agreement ought to be enough, and this piece takes no side in any case of any kind.

It is not about a theorem’s proof, the strength of a spirit, or dough left to rise. A proof in arithmetic is not a record of an event, and nothing here applies to it.

It does not tell you to doubt each decisive record you hold. Most stay decisive. The claim is about what their decisiveness is, not about how often it fails.

And it does not settle what true means. The foundation has no locked definition of truth. Where this piece says an account is accurate, it uses the word in its ordinary sense.

What it costs

The cost is plain, and I will not soften it. Nothing you have proved put you in touch with what happened. Each certainty about the past that rested on a record rested on records agreeing with one another, and on your expecting the ones you did not check to agree as well. A proof is decisive until the next mark, and nothing guarantees that the next mark will agree. That includes the proofs you were most relieved to have.

It also takes away a comfort people do not notice they are using: that a good enough record would let the past speak for itself in an argument. It will not. The past does not speak, and a record speaks only to somebody reading it now, against other things they can read now. Every dispute about what happened is a dispute among present marks and the people reading them.

And the cost falls hardest on whoever has no record.

An ordinary Tuesday

1. When you say “I have proof”, say what it is read against. A receipt and a till. A photo and a clock. Knowing the second item tells you how your proof could fail, which is most of what is worth knowing about it.
2. Hear “pics or it didn’t happen” as a sentence about the speaker. It says what they will need before they change their mind. You can decline to supply a record without conceding the event.
3. Keep records for the reader, not for the event. A receipt is a mark you make now for somebody who will read it later — often you. It keeps nothing of the purchase. It keeps the next argument short.
4. When a proof falls, say “my reading moved”, not “so it didn’t happen”. The overturned photograph stays in the record, beside the clock that overturned it. Nothing was edited; something was added.
5. Notice the window feeling and name it. When a record seems to show you the event, what you are feeling is that you do not expect another record to disagree. That is usually a good bet. It is still a bet.