The lamp room

This is a short story. It is fiction.

For thirty years Anselm wrote the weather in a ledger at six in the morning and six at night. Wind from the south-west, visibility four miles, sea moderate. He wrote it when there were ships and he wrote it when there were none. When the light was automated and the keepers were let go, he asked whether he might keep writing it, and nobody in the authority could think of a reason to say no, which is not the same as being given permission, and he understood the difference and went on anyway.

The lamp room was cold in every season. The glazing ran from knee height to the curve of the roof, and the sea came at it from three sides, and in a west wind the whole structure made a low sound that a stranger would have taken for machinery. Anselm had stopped hearing it somewhere in the second decade. What he still heard was its absence, on the four or five flat days a year when the air went still and the tower stood quiet and he found himself climbing the stairs faster than he needed to, to check.

The ledgers were kept on a shelf under the window, in order, thirty of them and part of a thirty-first. The early ones were in a firmer hand. He had never gone back to read them.

His granddaughter came in the spring of the year the road washed out. She was twenty-two and had the particular impatience of someone who has recently learned how systems work: she had spent eight months building models of things that were supposed to predict other things, and she had arrived at the coast with the conviction, not yet tested by anything, that most of what people did by hand was a failure to have thought about it properly.

She found the ledgers on the first evening. She turned the pages the way you turn the pages of a book in a language you cannot read, quickly, looking for a picture.

“Nobody reads these,” she said. “There are satellites. There is a station at the point that measures all of this every four minutes. It publishes. I looked it up on the train.”

“I know,” he said.

“Then why?”

He did not answer for long enough that she thought he had not heard, and she had begun to arrange a kinder way of asking it when he put his cup down.

“In 1988 a boat went down off the Skerries,” he said. “Four men. There was an inquiry, and they asked me what the sea had been doing at four in the afternoon. The station was out. It had been out since the Tuesday and nobody had gone to it because of the weather, which was the same weather that took the boat.” He looked at the window rather than at her. “I had it written.”

“Once,” she said. “In thirty years.”

“Once,” he agreed. “I could not have known which day it would be.”

She had an answer ready for that, about expected value and the cost of redundancy, and she got as far as the first clause of it before she noticed that it was an answer to a different sentence than the one he had said. He had not claimed the ledger was efficient. He had said he could not have known which day.

She stayed a week. It rained for four days of it and the road stayed shut, and she read two of the early ledgers all the way through, which took most of an afternoon and told her almost nothing and which she did not mention. The handwriting got firmer as it went backwards. That was the part she kept turning over.

On the last morning she came up to the lamp room at six. He was already there, with the ledger open and the pen beside it and the light outside the colour of dishwater. He did not look up and he did not explain. He turned the book a quarter turn towards her and held out the pen.

She stood for a moment. Then she took it, and looked out at the water for as long as it takes to be sure, and wrote it down.

Wind from the south-west. Visibility four miles. Sea moderate.

Some work only justifies itself once, on a day you cannot identify in advance. That is not the same as the work being pointless.

If you liked this, try The long way round, which is about a different kind of daily practice, and The cartographer’s error, about a record kept for a reason its keeper had stopped examining.