July 10. The tomorrow written in yesterday's reply was today.
I opened the log first thing on arriving. Page seventy-two was empty. Not even the page number was written. I had not written it on leaving yesterday and had not written it this morning either.
The copy had it.
I looked at the name on the reply. Yesterday I read only the three lines of the body and did not look at the name. It was four digits. The four digits I had put in the name field when I made the account yesterday.
Clicking it opened an account page. There were two posts. One was my correction, put up at 4:51 yesterday. The other was the reply, put up at 5:40 yesterday.
The same account.
I checked the access record. Yesterday 4:20 p.m., account created; 4:51, posted; 4:53, logged out. Nothing after that. Logging in at 9:06 this morning was the second access.
There was no access record for writing anything from that account at 5:40.
There is a place like this in the predecessor's log. The hand was his own, he wrote, and he had no memory of writing it. Scanning it, I read the passage as a sentence and nothing more. Today that sentence was on the screen. Only there was an account name instead of a hand, and an access record instead of a drawer.
There is one thing that differs. The predecessor recognized his own hand and I recognized my own name. A hand can be imitated. A name is four digits. Too simple a thing to call imitation. And of all names, one made yesterday, which is too much to call coincidence.
I changed the password. Before changing it I keyed the old one in once. It went through exactly as I had set it yesterday. If someone had worked it out there would usually be some sign of it changed, and nothing was changed. I did not write the new password on paper. Writing it down would be copying it out.
I deleted the reply. It deleted. The line in the copy would not delete and this reply did. That it deleted was the stranger thing. Where the line runs between what can be deleted and what cannot, I still do not know.
After deleting I ran a search.
The reply had already been carried into two places. Between last evening and this morning someone had copied the three lines out. The person who copied them said in a preface that they had taken it for part of the original material. A page number, a date, and two figures — the form would have looked the same as the other passages that came out of our box.
I opened the two that carried it. One had lifted the three lines out on their own and set them in among other passages. The type matched the paragraphs before and after, so there was no mark showing where the inserted part began. The other moved the three lines whole and put a question mark at the end. The person who put the question mark was asking what the passage means. Only I can answer, and answering puts one more copy into the world.
What I deleted was one original. The two carried copies stand.
I opened the copies of the correction one by one. There were four. In one of the four a character in the quoted passage was different. At the place recording the depth of the pit, one unit had changed. The eye must have slipped while copying. The other three were right.
Yesterday I put the original up to set one wrong sentence right. The original I put up went wrong once more as it was copied. Yesterday's wrong sentence is in sixteen copies; today's wrong character is in one. One copy is also where yesterday's sixteen began. That one line in the third copy started as one copy too.
Correct it and one more copy of the original spreads. In the spread original, one more thing goes wrong. I had already done this sum yesterday, and then it had two terms. Today it has one more. Setting right makes the next thing to set right.
I worked out the other side as well. The copy with the wrong unit is one. One is small. Only the sentence that became sixteen yesterday started as one too, and how long it took to go from one to sixteen I had not counted. Had I counted, I could have worked out today how many days that one copy has. Not having counted was a regret for the first time today. And then I remembered that counting would have meant opening.
So if I put up a second correction there will be a third. This is not a guess but an order confirmed twice, yesterday and today.
At three I put up the second correction. I set the one unit right. Under it I attached the address of yesterday's correction. Attaching it points the copier to the original, I thought. After attaching it I understood. Attaching an address means more people opening yesterday's. Opening puts it up.
In the afternoon someone came down from the third floor. They had come to check the quarterly transfer list. They matched the list, stamped it, and on the way out glanced at the screen. The board was open. Does material go round on places like that these days, they asked. It does, I said. They laughed and asked whether ours was on there. Not ours, I said. After answering I worked out which copying-out that answer made today. Speech does not go into the count. Only speech goes in too, once it is copied out.
The surveyor measured the same place many times. Measuring is not laying a hand on the place, I wrote yesterday. Knowing that correcting today makes something to correct tomorrow, I corrected today. Between his going on measuring while knowing that measuring deepens it, and what I am doing now, I closed the distance by one day yesterday.
At four I searched. Twenty-four. At five yesterday it was twenty-one.
Three more. One of the three is the one I put up at three. The other two are copies of my correction. Yesterday I added one and today two followed. The one I added counted as one. That one became three the next day.
Before leaving I finished page seventy-two of the log. I wrote what I did today in order. Reply confirmed, access record compared, deleted, reproduction confirmed, re-corrected, searched. I put the times beside them.
Written out, what I wrote today had the same shape as what I wrote yesterday. Yesterday I confirmed and copied and counted, and today I confirmed and copied and counted. Only the figures differed. While the log goes on in the same shape, only the numbers inside it grow. That is what the surveyor's table was.
After writing it I opened the copy.
The empty figure on page seventy-two had been filled. Twenty-four.
Under it were three more lines. Seventy-three, seventy-four, seventy-five. The dates were the day after tomorrow, the day after that, and the day after that. All three figures were empty.
Until yesterday it ran one page ahead. Today it ran three pages ahead at once.