and beyond.^49 Thus an early Achaemenid-period letter from a temple official
in Larsa to a colleague in Uruk mentions receiving a report that the last month
had been of 29 days—suggesting that this information did not reach everyone
immediately. The same is implicit in another letter, referred to above, of some
temple official to a governor at Sippar asking for a report on whether the last
month was 29 or 30 days.^50 It is no surprise that the administrative/economic
document from 522BCEthat assumes a 30-day month, whereas the Astronom-
ical Diaries have only 29 days for this month (see above), was written at
Sippar—at some distance from Babylon where our Astronomical Diaries were
compiled and where the calendar, in this period, is most likely to have been set.
All this suggests that on the 30th day itself, if not also later, most people would
have had no idea when the new month had begun.
If the beginning of the new month was widely unknown, the most sensible
policy for scribes would have been to assume a 30-day month and to write
‘30th’(on the day after the 29th) as a provisional or tentative date. This
practice would have been justified in a number of ways.^51 Firstly, this option
was least likely to lead to errors: for if it later turned out that there had only
been 29 days in the outgoing month, the date of‘30th’could easily be
corrected, or rather re-interpreted, as meaning the 1st; whereas if on the
contrary the scribe had written‘1st’(assuming a 29-day month), but then
the month turned out to be of 30 days, it might have become unclear, later on,
which day the scribe had really meant.^52 Secondly, in astrological tradition
30-day months were considered normal and auspicious, and 29-day months
were inauspicious.^53 This may explain a preference for day 30 (Beaulieu 1993:
(^49) Beaulieu (1993) 71 and n. 18. A similar problem affected the dissemination of decisions to
intercalate the year, at least in the neo-Assyrian period: thus an order from Esarhaddon in 670
BCEto intercalate Ululu and postpone the cult ceremonies to the next month (i.e. to the second
Ululu) only arrived from Assyria to Babylon on or after the 6th of the month, when the
ceremonies of the 1st–6th had already been performed (Parpola 1970–83: i, no. 287, with
commentary ii. 284–5; 1993: no. 357). In Stern (2000a), I argue that in much more remote
locations such as Elephantine in southern Egypt, where reports on the new month could never
have arrived within the same month, the inhabitants had no option but to improvise—often
rather erratically—the dates of the Babylonian calendar. For similar evidence from Idumaea and
Bactria, see below, nn. 75 50 – 6.
Above, n. 43.
(^51) For a similar practice in ancient Greece, see Ch. 1, near n. 69.
(^52) The preponderance of day-30 dates in extant administrative/economic documents suggests
that documents were not corrected or rewritten, but at the most, reinterpreted. The same is likely
to apply to documents dated to thefirst days of the month, when the length of the outgoing
month was yet widely known; in this context, however, it would have been difficult to avoid
uncertainty, later on, as to which day the scribe had really meant.
(^53) Hunger (1992) e.g. nos. 7, 9–13 (a new moon on the 30th is a bad omen, on the 1st (= 31st)
a good omen); nos. 106, 119, 290–1 (the‘normal’length of a month is 30 days); see Brown (2000)
106 – 21, 249.
86 Calendars in Antiquity