The Ecclesiastical Moon

I. The Golden Number

Take the year. Divide it by nineteen. Keep the remainder and add one. The number you are left with is between one and nineteen, and it tells you where you stand in a cycle of nineteen years, after which the phases of the moon return to roughly the same days of the calendar.

Roughly is doing considerable work in that sentence. Everything that follows exists to take up the slack.

Ruth Ansell had done the division on paper since she was eleven, when the previous computist put a pencil in her hand and told her that the machines could do it faster and that she was going to learn it anyway. She was sixty-eight now. The paper came from the mill at Cotter’s, which made it out of stalk fibre and made it badly, and it took ink like blotting cloth, so she wrote small.

For the coming year the remainder was four. Golden Number five.

The name came from medieval calendars, where the numbers were entered in gold. Ruth had never seen gold. There was said to be some in the hold of the Meriden, brought out for contacts and solder, all of it long since used. She had seen a photograph of a page from a calendar of the fifteenth century, in a file that also contained forty thousand other photographs and had not been opened by anyone else in her lifetime, and the gold in the photograph looked yellow-brown, like the light through Cassin’s atmosphere in the second half of the thaw.

She wrote V at the head of the sheet and drew a line under it.

II. The Epact

The Golden Number gives you a lunar cycle. What you want is a moon, so you convert: the epact is the age of the ecclesiastical moon at the beginning of the year, a number from nought to twenty-nine, and from it you can lay out the whole year’s lunations, alternating thirty days and twenty-nine.

Then you correct it, twice, for two different failures.

The first correction is solar. Three times in every four hundred years the Gregorian calendar withholds a leap day from a century year, and each time it does, the epact must come down by one to match. The second correction is lunar, and it exists because nineteen years is not exactly two hundred and thirty-five lunar months. It is about an hour and a half longer, which accumulates to a day every three centuries, so the epact is raised by one, eight times in every two thousand five hundred years, on a schedule fixed in 1582 and running to the end of time.

The corrections are calibrated to the Earth. To the length of the Earth’s year, and the period of the Earth’s moon.

Idris Aban made this point to her in the corridor outside the assembly room, not unkindly, the way you point out to someone that their coat is buttoned wrong.

“You’re correcting for a moon nobody here has ever seen,” he said. “Against a year that has nothing to do with our year.”

“Yes.”

“You say that like it settles it.”

“It’s the second half I’d argue with,” Ruth said. “Not the first.”

III. The Dominical Letter

The days of the year are lettered A to G, over and over, from the first of January, and the letter that lands on Sundays is the Dominical Letter. In a leap year there are two, because the count slips at the end of February.

It is the least interesting part of the calculation and the only part that has never once been broken.

“Not once?” Nell said. She was twenty-two and had the apprentice’s habit of testing a claim by repeating it back with the weight moved.

“In 1582 they took ten days out of October. The fourth was a Thursday. The next day was the fifteenth, and it was a Friday.”

Nell worked it out and then stopped working it out.

“They dropped the dates and kept the weekdays.”

“They kept the weekdays.”

“Why?”

“Because the weeks weren’t wrong,” Ruth said. “The dates had drifted against the sun. The count of sevens hadn’t drifted against anything. There was nothing to fix.” She turned the sheet round so Nell could read it. “You could put it the other way. They were prepared to lose ten days of the calendar and not one day of the week, and I think that tells you what they actually believed they were keeping.”

Nell said, “Idris says you’ll say something like that at the synod.”

“I imagine I will.”

IV. The Paschal Full Moon

Easter is the Sunday after the fourteenth day of the ecclesiastical lunar month whose fourteenth day falls on or after the twenty-first of March. If that fourteenth day is itself a Sunday, you take the Sunday after. The result cannot fall earlier than the twenty-second of March or later than the twenty-fifth of April.

The twenty-first of March is not the equinox. It was the equinox, more or less, at Nicaea. It has not been the equinox for most of the calendar’s life, and in 1582 it was restored to that date by decree rather than by observation, which is to say the men who reformed the calendar chose to make the definition true again rather than admit it had ever been a definition.

The ecclesiastical moon is not the moon. It is a table.

Ruth said all of this to the synod standing up, which took her about four minutes, and then Idris put the question that everybody had come for.

“Easter falls this year on the ninth of April. On Cassin that is the middle of the shut. There is no spring here on the ninth of April. There has not been a spring on the ninth of April in the entire history of this settlement and there will not be one in yours or mine. You are asking us to keep a festival of the rising green in the coldest, deadest, darkest six weeks of the local year, and to do it on the authority of a table about a moon that is not there.”

“Yes,” Ruth said.

“Then give me a reason.”

“Australia.”

Somebody laughed, and then stopped.

“Easter has fallen in the autumn for half the Earth for as long as half the Earth has kept it,” Ruth said. “Four hundred years. Nobody moved it. Nobody proposed moving it. Every argument you have just made was available to a congregation in Melbourne in 1850 and none of them made it, and I don’t think that was stupidity or deference. I think they understood something about what the date is for that we’re at risk of forgetting because we are further away and lonelier.”

“Which is what?”

“That it isn’t a description of the sky. It never was. Nicaea didn’t settle the astronomy, it settled the argument — churches were keeping Easter on different days and could see each other doing it. The whole apparatus, the nineteen years, the epacts, the corrections, all of it is machinery for producing agreement between people who cannot consult one another. That is the entire function. It is designed to be run separately and come out the same.”

Idris said, “There is nobody to agree with. We haven’t had a signal in thirty-one years.”

“I know.”

“Then who are you agreeing with?”

“Anyone running the table,” Ruth said. “It doesn’t require them to be there. It requires them to use the same table if they are.”

She sat down. The vote was taken on paper, because everything on Cassin was, and the reform failed by nine.

V. Easter

Golden Number five. Epact eight, after both corrections. The fourteenth day of the paschal month on the fifth of April. The fifth of April a Tuesday, by a Dominical Letter descending in an unbroken count from a Friday in October 1582, and from a Thursday before it, and back past Nicaea to a week whose beginning nobody recorded because nobody thought it would ever need recording.

Easter, therefore, the tenth of April. Ruth had said the ninth at the synod and been wrong by a day, which she noted on the sheet, and which nobody ever mentioned to her.

They kept it on the tenth. The shutters stayed closed against the cold and the room was lit the way rooms were lit in the shut, which is to say badly. Nell read the date out from the paper because Ruth’s eyes had gone, and put the stress on the wrong syllable of decima, and nobody minded.

Eleven years out, and eleven years back, if there was anybody left to run it, the table gave the same day.

Ruth had never needed to check.