July 9. I opened the copy first.
The empty figure on page seventy had been filled. Twenty.
I did not go to count. Instead I wrote twenty on a piece of paper and put the time under it. Nine twelve in the morning. This time I would set the order myself. I had seen what was written and had not yet counted. For as long as I did not count, nobody knew whether the figure was right or wrong. How long I could hold that not-knowing was what I meant to measure.
I held it two hours. I processed seven requisitions in that time. I put two reels on the reader and copied the registration numbers into the ledger. My hand went as it always goes while I copied. After eleven the twenty on the paper kept coming into my eye. I turned the paper over. Turned over, all that was left was knowing which side the figure was on.
Then I ran the search. The same terms as June, in the same order.
Eighteen.
Not twenty. I looked at the screen for a long while. It was the first time. In six months it was the first time the figure written in the copy and the figure I counted had come apart.
Until midday I felt something that could be called relief. Coming apart means the figure was not known in advance. Someone guessed, or miscounted, or inflated it. There are people who inflate everywhere. If six months of matching had been chance, today is where the chance runs out.
At two I searched again.
Twenty.
The two that had been added went up in the afternoon. One at eleven fifty, one at twelve forty. I had counted eighteen at eleven fourteen.
It had not come apart. It was not yet.
I wrote that down and looked at it a long while. There had been a moment when it looked wrong. Only that moment was not wrongness but not-yet. To tell not-yet from wrong I would have to fix how long to wait, and I have no grounds for fixing that term. If it becomes twenty-one this evening then twenty is wrong too, and if it becomes twenty-two tomorrow then twenty-one is wrong as well. Whatever figure is written, something passes it in the end. Until it is passed, it is not yet.
So this table cannot be wrong. A table that cannot be wrong is not a table that is right.
In the afternoon I opened the third of the fourteen again. It is the one whose source I could not confirm in June. Its paragraphs broke in different places from the others and it had one line more.
That line is now in sixteen of the twenty.
It was not in the first thirteen that went up. It went into the third once, and after that it followed into every typed copy. People copying out do not look at the original. They look at the post before theirs. If the line is in the post before theirs, they copy the line too.
The content is nothing much. One sentence has been added to the surveyor's log passage. A sentence saying he left the pit after the last measurement. There is no such sentence in the log in our box. He did not leave.
Six of the sixteen had a short remark under the sentence. The word relief came up three times. One person wrote that it was the first place in the whole thing where they could breathe. With the sentence about leaving, the story closes there. Without it, it does not close. The closing kind is easier to copy out.
The original is in the box under my desk. The box is on no list, and being on no list, nobody looks for it. Of the twenty out there, nobody has seen the original. I am the one who has seen the original.
For six months I have managed the material. The word manage has correcting what is wrong inside it.
To correct it I have to write out the original.
I sat a long while before writing that line. Putting a correction on the board means copying the passage out as it stands and setting it alongside. That the typed kind grows faster than the photographed kind is something I wrote down in June. So a correction is a post that sets the material right and a post that puts one more copy of it into the world.
Leave it and the sentence keeps spreading. Correct it and one more copy of the original spreads.
I set the two side by side and tried to work out which one adds less. The sum would not come. To do the sum I would have to know the rate, and to know the rate I would have to count, and counting puts it up.
At four I made an account.
Since I took the post I have never put anything up outside. Registration on the internal network was where my work ended. The registered copy is entered under an employee number and the number is mine. Boards outside have no employee number. You have to make up a name and put it in.
My hand stopped at the name field. Put in a surname and it becomes a person's name, and a person's name gets attached to one of twenty-one copies. Put in the archive and it amounts to telling people where the box is. Put in nothing and the account is not made.
I put the four digits of the archive code in as the name. Nothing else came to me. After keying the four digits I recognized them as the ones sitting in the recipient field of the predecessor's last log entry. I did not delete them.
I wrote the correction. I copied out the passage from the original and under it wrote that the sentence is not in the copy we hold. I did not say where it is held. Saying so makes people who look for the box.
I read it once more before putting it up. Half of my post was quotation from the original. Half of it is copied out.
I put it up.
At five I searched. Twenty-one.
The one added was mine.
For six months I have been the one who counts. Today I put a line into the side that gets counted. One of twenty-one went out from my hand, and that one has the name I made up on it.
The surveyor measured the same place many times. He kept measuring, knowing that measuring deepens it. I read him three times and on the third I wrote that I understood why he did not stop. The day after writing that I did what he did.
There is one thing that differs. He only measured. Measuring is not laying a hand on the place. Today I laid a hand on it.
Before leaving I opened the copy.
There was one reply under my correction.
The body was three lines. The first line had a page number. Seventy-two. The second line had a date. Not today's date but tomorrow's. The third line had two figures. One was twenty-one.
The other was empty.