five-year calendar, and hence the effective loss of lunar synchronism, would
have been deemed a worthwhile sacrifice for the sake of achieving more
accurate solstice predictions. This explanation, however, is untenable on a
number of counts. It amounts to arguing that the Coligny calendar was
essentially an astronomical tool for the prediction of solstices, and that what
counted most to its users was the accuracy of these predictions, whereas the
calendar itself and its relation to the moon were only a subordinate concern.
Yet, the form and contents of the Coligny inscription suggest on the contrary
that its primary function was to establish the dates of festivals and‘good’and
‘not good’days in a calendar that was clearly meant to be lunar, whereas the
three-lettered and other solar notations were only additional and ancillary to
it.^24 In addition, it is very unclear whether there would have been anything to
gain, for purposes of solstice predictions, from switching from a 30-year to a
25-year cycle. Olmsted demonstrates that with a 25-year cycle the predictions
would have been more accurate, but in practice, this higher accuracy would
hardly have been noticeable, and would certainly not have justified the loss of
synchronism with the lunar month.^25 It is most unlikely, therefore, that this
25-year cycle was ever used as a calendar in practice.
A more plausible suggestion has therefore been made by McCluskey (1998)
58 – 60, who refrains from conjecturing the existence of cycles longer thanfive
years (for which there is no explicit evidence in the Coligny inscription), and
suggests instead that thefive-year Coligny calendar was used with irregular,ad
hocadjustments so as to keep it in line with the sun and moon. It may be
(^24) Olmsted is, remarkably, not bothered by the one-day discrepancy of hisfive-year cycle
from the lunar phases, dismissing it as‘not really a problem’(1992: 15; his lengthy demonstra-
tion, ibid. 126–30, of the remarkable accuracy of his reconstruction of the Coligny calendar
obfuscates the fact that it is only accurate with regard to its solar, predictive notations, but not
with regard to the lunar month). Olmsted argues that as long as this discrepancy was known,
calendar users could easily have predicted the true dates of the lunar phases (for which, indeed,
the three-lettered signs could also have been used). But this argument, again, is based on the
implausible assumption that the function of the Coligny calendar was essentially predictive (i.e.
of solstices and, as here argued, lunar phases), and that all that counted to people was to make
accurate astronomical predictions—whereas the lunar accuracy of the calendar itself would have
been of no concern.
(^25) As Olmsted (1992) 126–30 shows, the three-letter scheme applied to the 25-year cycle falls
behind the solar year by one day in 454 years; I calculate that a similar scheme applied to the 30-
year cycle yields the same discrepancy in only 113 years. This greater accuracy of the 25-year
cycle is only noticeable, however, after more than a century, and is unlikely to have been
important to the calendar users. It is also questionable whether thesefigures, which are based
on present-day values of the mean tropical year, would have been known to the authors of the
calendar (more on this below, n. 33). In relation to the solstice dates of the Julian year (which
they certainly could have known; see ibid.), the scheme applied to the 25-year cycle is short of
one day in 100 years (which may explain, incidentally, why in Olmsted’s interpretation the three-
letter scheme avoided going beyond three long cycles or 75 years), whereas when applied to the
30-year cycle it is short of one day in 60 years. Again, the difference between the two is unlikely to
have been significant enough for the 25-year cycle to have been adopted at the expense of lunar
synchronism.
Dissidence and Subversion 307