Night shift, day fast

A fast that is named by sunrise can begin while the person keeping it is still at work, and the day it belongs to is then a day spent asleep. Put that in a room: three in the morning, the break room behind a ward, a nurse on a 12-hour night working out when her fast starts. In the reckoning her family keeps, it starts at sunrise, and sunrise this week falls inside the last hour of her shift.

She is a composite, assembled out of the sort of question that arrives from people whose working day is somebody else's night. Nurses on 12-hour rotations, drivers on long hauls, the office crews in Indian cities who keep American hours. The calendar in these households was made for a body that sleeps when it is dark. The rule at the middle of the calendar was made for a sky. Only one of those two has moved, and it is not the sky.

So it is worth being exact about what the rule asks for, because it is smaller than people assume.

A vrat (vrata, व्रत) is usually pinned to a tithi (तिथि), a lunar day: one step in the widening gap between the moon and the sun. A tithi runs a little under one of our days, so it will not sit tidily inside a date. It begins on one afternoon and finishes the next evening. The almanac, the panchang (pañcāṅga, पञ्चाङ्ग), therefore needs a way to decide which of our dates carries which tithi's name, and the way it does that is one minute long. Whichever tithi is running at sunrise names the whole day. That is the udaya tithi (उदय तिथि), the risen tithi, and which day is your fast, actually works the rule through on real sheets.

An instant, then. Not a stretch of morning, not a waking hour, not a meal. The rule wants to know what the sky was doing at one moment, and it never asks who was there to see it.

The minute you are usually awake for

This part surprises people, including the people it happens to. Take a 12-hour night, seven in the evening to seven in the morning, worked in Delhi. In August the sun comes up there at about a quarter to six. In the last week of December it comes up at about ten past seven. Both of those fall inside the shift. Somebody working that pattern is awake for the naming minute in every month of the year, and asleep for most of the day the minute has just named.

Shorten the shift and the year splits in two. Nine at night to six in the morning: in August the sun clears the horizon a quarter of an hour before you hand over, and in December it comes up while you are at home with the curtains shut. Same person, same rule, same fast. The sheet prints the same date either way, because the rule was never counting attendance.

That is the first thing to hold on to, and it is a relief rather than a problem. The naming of the day is settled by the sky and it stays settled. Sleeping through a sunrise does not make a household's date wrong, any more than a cloudy morning does. The rule only runs out where the sky stops handing the minute over at all, which is sunrise at sixty degrees north.

What moves is everything else. Every part of an observance that has a length rather than a moment has to be fitted to a body, and the body is on a rota.

Two hours and thirteen minutes

Take ekadashi (ekādaśī, एकादशी), the eleventh lunar day, which a great many households fast on. Kamika ekadashi this August is a clean case. The sheet computed for New Delhi prints it like this: the eleventh begins at 1.59 on Saturday afternoon, 8 August, and ends at 11.04 on Sunday morning. At Sunday's sunrise the eleventh is running, so Sunday is the fast day.

Then comes the part that catches shift workers, because it is the part nobody mentions when they tell you the date. The fast is not simply over at bedtime. It is broken the next morning, inside a window the sheet prints for you: parana (pāraṇa, पारण), from 5.47 to 8.00 on Monday 10 August. Two hours and thirteen minutes.

Each edge of that window is a rule. It opens at sunrise, which on that Monday in Delhi is 5.47. It closes when dwadashi (dvādaśī, द्वादशी), the twelfth, runs out, because the breaking belongs inside the twelfth. A third condition sits behind those two. Hari vasara (hari vāsara, हरि वासर), the first quarter of the twelfth, is meant to pass before anybody eats, and here the twelfth began on Sunday morning, so that quarter was finished by late Sunday afternoon and the Monday window is clear of it.

Now put a shift against that. You come off at seven, you are on the bus at twenty past, you are through your own door at ten past eight, and you would like to sit down first. The window shut at eight. You kept the whole fast and missed the morning it was meant to end in, and nothing careless happened: you were at work.

The almanac has met this before, though not for this reason, and it already carries a clause. Break the fast within the twelfth, the rule runs, unless the twelfth is over before sunrise. Watch what that clause does in another city. Those two instants, 11.04 on Sunday morning and 8.00 on Monday morning, are events in the solar system, and they reach everybody at once. Delhi clocks run five and a half hours ahead of Greenwich, and London in August runs one hour ahead of it. So the twelfth ends at 3.30 on Monday morning in London, and the London sun comes up at 5.38. In London the window does not shut too soon after opening. It is over before it can open at all, and the clause takes over: break the fast after sunrise. The same fast, the same two instants, and a difficulty that is sharp in one city and absent in the other.

Convention, and the rest

Now the question the reader came with: what a household is allowed to do about any of this. It is worth separating what the tradition has written down from what people have worked out themselves. Both are in use, and they are not the same kind of thing.

Written down, and printed on the sheets: the three-day shape, a single light meal the day before, the fast itself, the breaking the next morning. The parana window with its two edges. The clause for the morning when the window has gone. And, plainly, two dates. Almanac publishers print separate ekadashi lists for Smarta and Vaishnava households, and the Vaishnava day can fall the day after the Smarta one. Neither list is the correction of the other.

Behind those lists is a shelf of books about nothing but this. Deciding which day an observance falls on is its own branch of Sanskrit writing, and the two volumes the almanac-makers still lean on, the Nirnaya Sindhu of about 1612 and the Dharma Sindhu of 1790, spend their length on questions of exactly this shape. A household weighing two defensible answers is not falling short of the tradition. It is doing what the tradition has been doing in writing for four hundred years.

Then there is the rest, which has no text behind it and is honest about that. A person on nights keeps the sheet's date and eats at the first chance after the window, which is most people, most months, without anyone calling it a decision. A family holds its meal until nine at night, when the one who has been asleep all day is up and the others are still awake, so the fast at least ends at a table with people at it. Someone rings a priest in India from a night break, when it is the middle of his working day and the middle of her shift, and takes the answer he gives for her line and her household. Someone keeps their own waking hours instead of the sheet's, knowing quite well what they have swapped.

I cannot tell you how many households do each of these, and I would not trust anybody who could. What I would say is that the second list is not a lesser version of the first. A convention is an improvisation that enough people kept for long enough, and the printed sheet in front of you is the survivor of a great many arguments in rooms like the one you are sitting in.

One thing does not belong to the calendar at all. Whether a fast is a good idea on a 12-hour night belongs to the person keeping it and to whoever looks after their health. The sheet tells you which day. It has never told anybody what their body can do on that day, and it was not written to.

The households that find this manageable have usually done one thing once, on a day off, in daylight, with the roster and the sheet on the same table. Four numbers, written down. The minute the sun comes up here. The minute the tithi ends. The minute the window shuts. The hour the shift finishes. Those four say quickly enough whether a given month is easy or whether it needs a decision.

Three in the morning is a poor hour for settling any of it, and it is the hour at which the question always seems to arrive.