contrast is evident, for example, in 522BCEwhen the same month is accurately
given 29 days in the Astronomical Diaries, and a 30th day in an administrative
text.^47
The most simple explanation is that dates in administrative/economic texts,
particularly at the end of the month, were less accurately recorded than in the
Astronomical Diaries.^48 The reason for this can be explained,first by analysing
Huber’s own data. His sample of administrative and economic texts comprises
a total of 155 months, of which 151 are of 30 days, and four of 29 days. The
very much higher proportion of 30-day months is surprising—it is certainly
not reflected in astronomical sources—and suggests that there is something
wrong with this sample. A partial explanation may be that in the context of
administrative/economic documents, 30-day months are easier to identify (on
the basis, simply, of‘day-30’dates), whereas months of only 29 days are more
difficult to identify and positively prove. However, I suspect that the very high
proportion of 30-day months in Huber’s sample also reflects, to a significant
extent, a scribal tendency to date documents to the‘30th day’even if the
calendar month only counted 29 days. I would question, therefore, Huber’s
fundamental assumption (1982: 51) that documents with day-30 dates neces-
sarily attest 30-day calendar months. As I shall now argue, the preponderance
of day-30 dates in administrative/economic documents teaches us something
about scribal dating practices, but not necessarily about the Babylonian
calendar or its accuracy.
A scribal tendency to date documents to‘day 30’even in 29-day calendar
months is only to be expected. In most cases, indeed, scribes writing docu-
ments on the day following the 29th of the month would not have known
whether the new month had already begun, i.e. whether it was the 30th of the
old month or the 1st of the new month. This question depended on the king
and his astrologers, whose decision would normally have been taken on the
day itself, and hence could not have been instantly known even in the capital
city; it would have taken a few days to reach the other cities of Mesopotamia
(ibid. 55). Indeed, the new moon should have been visible on 20 May and 18 June 93BCE, thus
making Aiaru a 29-day month; the date in this text suggests that the month either began one day
early, or ended one day late (thus yielding a 30-day month). The latter is assumed (for no clear or
cogent reason) byWallenfels (1992), who takes this as evidence that the month was still based in
this period on actual sightings of the new moon (which in this case would have been postponed,
on 18 June, because of bad weather).Whatever the merits of his argument, the date of 30 Aiaru in
this document may be considered suspect, for reasons that I shall explain below.
(^47) Year 7 of Cambyses, month XI (=c. February 522BCE) is given in Sachs and Hunger (1988–
2006) v. 168–9 (no. 55, col. iii l. 12) as 29 days long, as indeed it should have been according to
new moon visibility, but the document in Strassmaier (1890) 228 no. 395 is dated to the 30th of
that same month (Huber 1982: 54). I have identified another ten months that are attested both in
astronomical and in administrative/economic sources (from Huber’s list), but there, the dates are
in agreement. 48
For a rudimentary and not quite accurate suggestion on these lines see Cohen (1993) 4–5.
The Babylonian Calendar 85