gradually shifted towards the day offirst visibility of the new moon, which was
surely not the original intention of the creators of this calendar.^74
In support of his argument, Parker cites a total of seventeen epigraphic and
documentary sources with a double date, civil and lunar, dating from the
Ptolemaic and early Roman periods (between 237BCEand 190CE). According
to Parker, all the lunar dates conform to the 25-year cycle, which confirms that
this cycle was in standard use throughout this period. He argues further that
these dates could not have been reckoned in any other way, e.g. empirically,
because the range of these dates, from the mid-third centuryBCEto the second
centuryCE,reflects accurately the gradual shift of the 25-year cycle from day of
invisibility of the old moon to day of visibility of the new moon; this would not
have arisen if these dates had been consistently reckoned through empirical
observation of invisibility of the old moon.^75
But Parker’s interpretation of pap. Carlsberg 9, and his claim that it was
used as the standard lunar calendar in the Ptolemaic and Roman periods, can
be criticized in several ways. According to Parker, pap. Carlsberg 9 only
supplies the dates of alternate lunar months, i.e. months beginning in even-
numbered months of the civil calendar; the dates of lunar months beginning in
odd-numbered civil months must be reconstructed, and this can only be
achieved on the basis of conjecture. Of the seventeen cases where the double
dates corroborate (according to Parker) the 25-year cycle, nine correspond to
the dates of odd-numbered civil months, i.e. to dates that are not in pap.
Carlsberg 9 but only in Parker’s reconstruction. These nine cases, therefore,
cannot serve as evidence that the 25-year cycle was used. More importantly,
Parker’s assumption that pap. Carlsberg 9 lists only alternate lunar months is
strange and intrinsically unlikely.^76 A completely different reading has been
suggested by Depuydt (1998), according to which, as is more likely, the dates
ofallthe lunar months in the 25-year cycle are supplied.^77 According to this
reading, however, pap. Carlsberg 9 no longer matches the double-dated
documents and inscriptions that Parker invoked in support of his theory.^78
(^74) Parker (1950) 16–17. Note also that in spite of referring only to Roman regnal years, this
document makes use of the ancient Egyptian civil calendar (i.e. as it was before the Alexandrian
reform): A. Jones (1997) 161. But this does not necessarily prove that the cycle must have been
composed before the beginning of the Roman period.
(^75) Parker (1950) 17–23; Clagett (1989–99) ii. 25, 137 n. 30; Jones (1997) 161–2.
(^76) Neugebauer (1975: i. 563 n. 4) suggests that the alternate months omitted in the document
were deliberately intended to be left indeterminate andflexible; he disagrees, therefore, with
Parker’s assumption that these dates must be somehow reconstructed. It seems to me, however,
that the notion offlexible months in this document would contradict the spirit and purpose of a
schematic calendar.
(^77) The 25-cycle as reconstructed by Depuydt also has the advantage of beginning neatly on a
I Akhet day 1 (the civil New Year), whereas Parker’s cycle begins on II Akhet day 1. See Depuydt
(1998) for further arguments (e.g. linguistic) in favour of his reading.
(^78) I have found that according to Depuydt’s reading (1998), the cycle differs from the
following double dates (as numbered by Parker 1950: 19–22): nos. 5 (46BCE), 7 (142BCE),
The Egyptian Calendar 151