The Cleveland Museum of Art has a Buddhist manuscript copied on palm leaves, each leaf about two inches tall and nearly two feet long. It’s the Perfection of Wisdom in Eight Thousand Lines, and the museum’s database splits it into 379 records: one for each side of every surviving leaf, one for each side of the two wooden covers, and one for the whole book. Every one of the 379 gives the same date: “Sunday, September 14, 1119”.

Most museum objects this old are dated to a century, or a span of them. The date comes from the colophon, a note on the last leaf, which the museum has translated: “This is a religious donation of Āryaśrīmittra, an eminent follower of the Mahāyāna, a monk coming from Nepal.” He asks that the merit go first to his teacher, his preceptor, his mother and his father, and then to every living being. Then comes the date: “In the year 239 [Newar Era], on the 8th [tithi] in the bright fortnight of Āśvina, at the great monastery of Vikramaśīla.”

There’s no Sunday in it, and no September. The scribe counted by the moon: the eighth lunar day after the new moon, in the month of Āśvina, in year 239 of the Nepalese era. A lunar day doesn’t start at midnight. It starts whenever the moon reaches a certain distance from the sun, which can happen at any hour. Someone later turned that into a day of the week in the calendar Europe used then. I tried the conversion myself, roughly: the new moon fell around midnight between 6 and 7 September 1119, so the eighth lunar day ran from about Saturday evening to Sunday evening. The usual rule gives a day the lunar day in force at sunrise, so that’s Sunday. The conversion holds. It just isn’t what the scribe wrote.

Then the database copies that one Sunday onto everything. The paintings on the leaves and covers were added in Nepal, after the book came north, the museum says, so not on that day. The last leaf’s back is blank, and it has the date too. At some point between the 1500s and the 1800s, someone pasted yellow Nepalese paper onto the first leaf to stop it falling apart. A corner near the end is patched with the same kind of paper. Folio 2 was lost before the museum bought the book in 1938. None of these days were written down anywhere, so a database with one date field per record files them all under the Sunday.

The colophon wasn’t written to date the book for us. It said who gave it, where, under which moon, and who the merit was for. The museum’s database needs one day, so it has one, and it’s right. Nobody wrote down the other days.