On the Coligny Calendar and the Neolithic Calendar in Plato’s Critias
An article published in SIS Review similar to the previously unpublished 2006 paper.
Summary
Evidence of a calendar within the Critias of Plato; and extant evidence of the calendar used by the
Druids of Gaul can both be reconstructed according to an eleven-year lunisolar cycle of 5+6 years and
may in fact be the same calendar. This short article summarises the author’s previously published work
on these subjects.
Introduction
The existence of astronomical alignments at Neolithic monuments such as the stone circles, once itself
controversial, is now virtually accepted; somewhat intuitively these may be associated with various
Celtic seasonal festivals and yet without any real understanding of how their calendar worked. Did
they have a system of calendar dates and an era such as we do?
Recent DNA studies have largely overturned the old notion of a wave of Iron-Age Celts invading
Neolithic Britain and Ireland. Neither may we any longer talk vaguely about a ‘Celtic’ calendar as if
continental invaders had introduced their religion and culture, including their calendar, at some
convenient period just prior to the Roman invasions. When we read Julius Caesar, who says that the
Druids and their culture were ‘found existing’ in Britain, then we need to take such a statement by a
contemporary chronicler far more seriously. We may now only use the term ‘Celtic’ as a loose linguistic
grouping.
It is not my intention here to go too deeply into source references in what is intended as an interesting
article for a general audience rather than a scientific paper. I explored these themes in much greater
detail some years ago in my books Atlantis of the West and Under Ancient Skies and in a further
unpublished paper; where all the source references may be found for anyone who wishes to verify my
conclusions.1 There is nothing here that I have not published previously.
The Calendar of Coligny
Clues to the calendar of the Gaulish Druids, and thereby of their astronomy, come from the “Calendar
of Coligny” (Fig. 1), discovered in a French vineyard in 1897. It was part of a wall plaque clearly
designed for display, similar to the annual calendar on your own wall. Its notation displays a five-year
cycle of lunar dates. Conventional academic assessment has been inconclusive because a five-year
period does not work as a lunisolar cycle. This has led to a conclusion that the Celts had only a poor
knowledge of astronomy.
Fig.1. A fragment of the Calendar of Coligny
The structure of the calendar shows 62 months of 29 and 30 days arranged as follows:
30 30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 30 29 30 30* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
Total: 1831 days
The month called Equos, ‘horse’, had nominally 29 days and was variable according to rules that
cannot be determined from the fragment alone. It may sometimes have been 28 or 30 days, so the 5year cycle could have been anything between 1826 and 1835 days.
Other Ancient Calendars
Conservative academic opinion will only allow that the Calendar of Coligny was devised by the IronAge Druids in Gaul just prior to the Roman invasion. A moment’s thought will show the problem here.
The Gregorian calendar on your own wall was printed last year, but the printer did not devise it! It
holds months named after Roman emperors and days named for Germanic gods that are thousands
of years old; the structure is but a revision of the old Latin calendar. Therefore, we may assume that
the devices on the Gaulish calendar similarly evolved from much earlier roots. This would take us back
to the Neolithic and the era of the megalith builders.
The basis of any such calendar is a natural lunisolar cycle, which arranges the months so that they stay
naturally in step with the seasons. Ideally, festivals should repeat at roughly the same season every
year without requiring constant revision by priests and astrologers. In a solar calendar, such as our
familiar Julio-Gregorian calendar, it is the solar year that is held and the lunar phases are allowed to
appear as they fall. However, in a lunar calendar the months must begin strictly with a phase of the
moon and be arranged as either 29 or 30 days long. Consider the following (current astronomical data
used here and subsequently):
One lunar month =
29.530589 days
62 lunar months =
1830.8965 days
One solar year =
365.2422 days
5 solar years =
1826.211 days
Difference:
4.6855 days
As the month and year are not integral, the best arrangement is therefore to use a cycle where a
multiple of lunar months is almost equal to a multiple of solar years. There is no precise equivalence,
but some of the closest available are 8 years, 11 years and 19 years. The Greeks used the 8-year
Octaeteris; the Babylonians used the 19-year Meton Cycle; but no culture is attested to have used an
11-year cycle.
In the Greek Octaeteris, 3 extra intercalary months were added to make the lunar years equate
(roughly) to the solar years; in the Meton Cycle 7 extra months; and in an 11-year cycle 4 extra months
would be needed; so, all lunisolar calendars must employ some lunar years of 13-months in addition
to the 12-month years. If these are evenly spaced, then solstices and equinoxes never slip too far from
their proper season. A good analogy here is Christian Easter, which shows us the opposite case of a
lunar date wandering within a solar calendar, but always held in springtime.
However, the five-year period of the Calendar of Coligny is not a natural lunisolar cycle. 62 lunar
months are 1831 days; 5 solar years are 1826 days. This is an accumulating 5-day slippage. And yet
the notation shows provision for two intercalary months inserted every 2.5 years. This suffices to tell
us that the 5-year cycle is not the entire calendar. The Intercalary months serve no purpose unless
they were intended to hold the months in line with the seasons. So which cycle were they using? Was
it 8 years? 11 years? 19 years? Or perhaps they used some longer cycle.
Another factor to consider is how to make the days add-up. Every calendar must employ a month of
variable length inserted according to established rules. In the Julio-Gregorian calendar we add a day
to February each fourth year and because this is still inexact the Gregorian reform added a century
rule. In the Coligny Calendar we similarly see notation for a variable 29-day month. There is not enough
evidence to say how the rules were applied – but we know that there must have been such rules or
these extra days, like the intercalary months, would serve no purpose.
Other clues are offered by ancient writers who tell us a little about the Gauls; also, about the ancient
Britons; and of their priestly cast known as the Druids. Some other clues may come via myths and
legends. If one may quote Pliny the Elder here: 2
Mistletoe…is gathered with great reverence, above all on the sixth day of the moon (it is the moon that
marks out for them the beginning of months and years and cycles of thirty years) because this day is
already exercising great influence even though the moon is not half-way through its course
(Pliny, Natural History, XVI, 250).
The historian Plutarch also tells us that the Britons held their most important festival each thirty years
as the planet Saturn returned to the constellation Taurus. 3 Consider:
One synodic period of Saturn = 378.09 days
29 oppositions of Saturn = 29 x 378.09 = 10964.61 days = 30.02 solar years
In my earlier reconstruction of the Coligny Calendar, I therefore also considered whether the Druids
tried to base it on a precise thirty-year cycle held to Saturn’s rhythm. So, I investigated how the extant
five-year ‘Coligny’ cycle could be alternated with a six-year cycle. This would allow the short-term
calendar to mesh both with the thirty-year ages and also the 11-year lunisolar cycle.
This brings us on to the 5 + 6-year calendar intimated in Plato’s Critias. To recap his Atlantis narrative,
he tells us that the Egyptian priests remembered an island in the Atlantic that was struck by a
geological catastrophe at a time just before the beginning of the Egyptian state. The ancient dynasty
of kings who ruled this island and the Atlantic coastal regions would gather together “every fifth and
every sixth year alternately” to discuss policy. If they were all to know when to meet, then they must
have used the same calendar; and Plato is giving us a clue how it worked.4
The Coligny calendar gifts us a detailed knowledge of how a real 5-year cycle was constructed. From
this it is possible to work-out how a 6-year cycle must be arranged in order to make use of the 11-year
lunisolar cycle.
In the Coligny fragment the intercalary months are evenly spaced at 2.5-year intervals; one at the start
of the five-year cycle; the other in the middle of the third year. If we reconstruct a six-year cycle using
the same month layout, then the intercalary months should naturally be spaced at the beginning of
the cycle and at the start of the fourth year. Therefore, we may propose:
30 30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
30 30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
Total 2184 days
Here, for simplicity, I have assumed that the variable month always has 29 days, so accuracy is down
to how you wish to organise the number of days that it has in each cycle; i.e. whether the calendar is
to be held strictly lunar; strictly solar; of perhaps to follow Saturn.
These two cycles, alternated, give 4015 days, whereas 11 solar years require approximately 4018 days;
so, the variable month would need an extra 3 days to be added each 11 years. One cannot know for
sure how the calendar rules actually operated, but it doesn’t really matter where you put the extra
days; so long as the number of days in the cycle add-up to your objective and this is fixed by astronomy.
So, the 5+6-year cycle casually mentioned by Plato can produce a very accurate calendar and the
reconstruction of the Calendar of Coligny yields a similar result. Is this a coincidence? Decide for
yourself. Of course, the Neolithic calendar would not look exactly like that of Iron Age Gaul, it’s a
process of evolution, just as our Gregorian calendar has evolved from its original Roman roots.
It was this detail, hidden in plain sight within the Critias, that convinced me of the authenticity of
Plato’s narrative – but you must decide for yourself how much of his catastrophism you wish to accept
alongside the description of ancient geopolitics.
Classical scholars, if they are to remain respectable, may only discuss Plato’s Atlantis narrative as if it
were a classical Greek story. They are not allowed to look beyond Crete, or the volcanic destruction
of Thera, for the inspiration behind Plato’s lost island; yet its own internal content states that it was
based on history told to Solon, as recorded by the Egyptian priests. The story is Egyptian, not Greek;
they recorded the organisation of an island in the Atlantic and of the Atlantic coastal regions. They tell
us that this was contemporary with the earliest period of the Egyptian state (pre-dynastic).
Egyptologists determine this was the late fourth millennium BC.
Archaeology tells us what was going-on along the Atlantic coasts of Europe in the late fourth
millennium BC. This was the middle-Neolithic; a time when the people began to build megalithic
monuments with calendrical alignments; in Britain, in Ireland, in Brittany and as far away as Malta. Is
it therefore so far out to equate the two? Is it so unacceptable to say that if you have a calendar on
your own wall with notations that are thousands of years old, then perhaps the first century Gauls
also used a calendar that was thousands of years old?
My own research has taught me always to believe what the ancient sources actually say, rather than
what academics think they are saying.
Notes and References
Atlantis of the West (2003) and Under Ancient Skies (2005) Both books are now available again in Kindle editions and an
unpublished paper intended for C&C Review is available (see previous), or via www.third-millennium.co.uk
1
The Latin is genuinely ambiguous here; some translators give that the months actually began on the sixth day after new
moon.
2
This shows us that the Druids were observing the Synodic Period; the time required for a planet to return to the same
position relative to the Sun as viewed by an observer on the Earth.
3
If you doubt the importance of using a common calendar, then recall when, in 1805, the Russian Tsar sent his army to join
with the Austrians before the Battle of Ulm, only to find that Napoleon had already defeated them! The Russians were still
following the unreformed Julian calendar which by then was 11 days out of synchronization with the Gregorian calendar used
by the Austrians.
4
Citation:
Chronology & Catastrophism REVIEW 2018:2 pp 50-53
Dunbavin, Paul (2020) On the Coligny Calendar and the Calendar in Plato’s Critias, in Prehistory Papers, pp 23-30 Third
Millennium Publishing, Beverley, ISBN: 978-0-9525029-4-4
The Neolithic Calendar
Paul Dunbavin
Summary: Plato’s Atlantis myth holds a passing reference to a religious festival held by the ancient
kings at alternate 5- and 6-year intervals. This may be evidence of an ancient calendar unrecognised
even by Plato. The author has reconstructed this calendar and compares it with extant evidence of the
calendar of the Druids. The preservation of a practical calendar within the Atlantis myth suggests that
it has come through from a genuine ancient source; and that other aspects of the story, including the
references to a flood catastrophe in the Atlantic, may be historical.
For students of catastrophism in human prehistory, Plato’s story of Atlantis will always be a primary
source text. According to Plato, the myth of an island in the Atlantic Ocean, which disappeared beneath
the waves in a single day and night, was an ancient Egyptian story transmitted to Greece, via Solon,
who visited Egypt at some time around 590 BC.
Classical scholars, however, insist on treating the story as the invention of Plato himself and would
seek to compare it only with other areas of Greek mythology. The possibility that it might truly have
come from Egypt; and that it may contain a memory of a real flood catastrophe that befell the Atlantic
coast of Europe, is left to more unconventional investigators to pursue. If it were possible to prove that
one part of the story is true, then it would strengthen the arguments that other parts of it may be
historical; and that it may hold a memory of a real ancient cataclysm.
When researching mythology, the present author’s method is to look for ‘mythological fossils’.
These are pieces of detail that are not strictly needed to tell the main story. They may be references that
can be independently checked, or which may have survived unchanged through generations of oral
repetition of the myth.
One such vital piece of evidence is hidden within Plato’s Critias.1 Here he tells us that, at a remote
period, before the catastrophic sinking of the island, the kings of the various parts of their empire
gathered together at the temple of Poseidon:
…every fifth and every sixth year alternately, thus giving equal honour to the odd and to the even number.2
I first investigated this statement in my book The Atlantis Researches in 1995.3 Why did Plato need
to introduce these alternated five and six-year periods into his narrative? A fictional storyteller could
simply have said that the ancient kings gathered together every few years to discuss policy. Why bother
to be so precise?
It is common sense that in order for kings from the outlying provinces to gather on a prescribed feast
day they would all need to be using a similar calendar and be able to measure the year to sufficient
accuracy. We have here a piece of numerate evidence that can be checked by science. We can
investigate whether it is possible to construct a real calendar that works on the principle of alternated
five-year and six-year periods.
There is no reason to believe that such a calendar is either Greek or Egyptian, The Greeks of Solon’s
day, used a lunar calendar based on an eight-year cycle, the Octaeteris. Egyptian calendars, as
understood by Egyptologists, offer no evidence of either five- or six-year devices. A cycle of five-plussix years implies a lunisolar calendar based on an eleven-year intercalation cycle, requiring four
intercalary months. It is important to appreciate that 11-years is a real intercalation cycle; indeed, it is
slightly better for the purpose than the Octaeteris.
The importance of intercalation cycles is that a precise number of months and years will repeat, with
the smallest cumulative error, to reconcile the lunar year of 354 days (12 x 29.5-day lunar months) with
the solar year. The insertion of intercalary months requires that some calendar years must have 13
months. The half-day in the lunar cycle requires that all calendars must combine both 29-day and 30day months; and since even this is not exact, a variable month is always needed so that sun and moon
can be regulated according to easily understood rules.
The accuracy of various intercalation cycles
No calendar is known to have been based upon the 11-year intercalation cycle. The 19-year cycle was
introduced to Greece from Babylon by Meton in 432 and forms the basis of modern Jewish and Muslim
calendars. A 19-year cycle is out by only about two hours, whereas an 11-year cycle would be out by
more than one and a half days.
So, if the Atlantis myth were merely Plato’s fiction, then why would he go to so much trouble to
invent a completely new calendar based upon a real eleven-year intercalation cycle? Moreover, why
does Plato not mention the five-plus-six-year calendar in any of his other astronomical works? I believe
that Plato never recognised this detail for what it is. It is evidence of a real ancient European calendar,
carried through as a ‘mythological fossil’, surviving intact through language changes to Egyptian,
Greek, and on into modern English. So far as I know, no researcher has investigated this until I did so
in The Atlantis Researches.4
The same mythological fossil may be found within one of the songs of the Finnish epic Kalevala.5
Here we find a story about the maid Marjatta who laments the passage of a long period of years as, “for
five or six of summers”. The use of this five- and six-year formula in Finno-Ugrian oral tradition gives
us a strong clue that the 11-year cycle has its origin in prehistoric Europe.6
Another place where we encounter a five-year calendar cycle is in Celtic Gaul. In Julius Caesar's
commentaries of his Gallic wars, between the years 58 and 51 BC, we find a brief account of the
activities of the Gaulish Druids. Caesar apart, we find only a few passing references to their science in
the works of other classical writers such as Pliny. Caesar tells us that the Druids were capable
astronomers who debated the size of the Earth and the movements of the planets.7
The Romans ruthlessly suppressed the Druids in Gaul, especially during the reigns of Tiberius and
Claudius and we must presume that they also persecuted them within their British province. Indeed,
the Druids’ continuing influence in Gaul may have been a principal motive behind the Claudian
conquest of AD 43. Caesar again, tells us that the order originated in Britain; and that a Druids
College existed there.8
Evidence of how the Gaulish calendar operated is preserved in the Coligny calendar, which was
discovered in a vineyard near Bourg-en-Bresse in 1897. The extant calendar consists of several
fragments, constituting some three-fifths of a bronze tablet, which appears to have been deliberately
broken-up and buried. The reconstituted calendar shows annotation in the extinct Gaulish language,
using the Roman alphabet; and this has been used to date it approximately to the reign of Augustus. The
mechanism of the calendar, however, shows no Roman influence and must therefore belong to the
final period of native Gallic culture. A less well-preserved fragment found near Villards d’Héria is
thought to date from the second century AD.9
Although most commentators on the Coligny Calendar have treated it as a purely Gaulish artefact
no older than the Roman era, it is important to appreciate that all calendars evolve gradually from
ancient roots. The first century Gaulish Druids did not invent the Coligny Calendar. Take as an example
the Gregorian calendar that hangs on your own wall; the days and weeks are pagan Germanic; the
months are named from Roman months that predate Caesar’s reforms and must be at least 2500 years
old! The Coligny fragment may similarly hold evidence of astronomy that was performed in Western
Europe thousands of years before Roman times. Caesar’s comment that Druidism was ‘found existing
in Britain’ would suggest that we should look to the British Isles for the origin of the calendar.
The Coligny tablet holds notation for five years; set out in sixteen columns. Each column holds the
notation for four named months; the exception being the first and the ninth columns, which hold an
intercalary month, followed by two normal months. The intercalations are therefore positioned at the
beginning of the five-year cycle and in the middle of the third year. Hence, it may be seen that the
intercalations were spaced at intervals of two-and-a-half years.
With a single exception, the length of each month is always either 29 or 30 days. Each month
is marked-out in days, listing their various festivals; and with a peg-hole to mark each date.
Since the average period of the lunar month is 29.53 days, any lunar calendar must incorporate
alternating 29 and 30-day months. Furthermore, one of the months must also be variable to regulate the
correspondence with the real Moon. The 30-day months are each suffixed as 'good', whereas the 29day months are styled 'not-good'. As the month Equos ('horse-month') was similarly 'not-good', this
indicates that it was not considered an ordinary thirty-day month and scholars have concluded that it
was the variable month. It may have alternated in length between 28 and 30 days.
The extant fragment contains notation for only sixty-two months and covers a period of
approximately five solar years. Sixty-two lunar months contain 1831 days, whereas five solar years
require only 1826 days. The sixty-two months exhibit notation for a possible 1835 days. Therefore, if
the Coligny fragment were to constitute the entire calendar, then, by strict lunar reckoning, it would
suffer a cumulative error of four days in every five years: and an error of nine days by solar reckoning.
A five-year intercalation cycle simply does not work! Such a huge error defeats the purpose of the
intercalary months. There must therefore have been a correction mechanism or another, missing, cycle
alternated with this one, in order to pull back the divergence.
For a calendar to be of use to the general population, it must be organised according to simple rules
that everyone will remember. The best way to do this would be to use one of the self-correcting
intercalation cycles. A theory that the Celtic calendar was based upon the nineteen-year Meton cycle
was proposed by Fotheringham and Rhys, some of the first scholars to investigate the Coligny
calendar.10 However, there is no historical evidence that such a cycle was ever used by the ancient Celts,
either in Gaul or in the British Isles. Later commentators such as Olmsted have proposed a theoretical
25-year cycle, but he suggests that this replaced an earlier 30-year cycle. However, while there is
evidence for a five-year festival, there is no textual evidence for the existence of a 25-year cycle in
ancient Gaul.11
The reconstruction given here is based upon the principle of alternated five- and six-year periods
and preserves all the known rules of the Coligny fragment. Within the six-year cycle the intercalary
months are spaced every 3 years, to give the required 4 intercalations in 11 years. An extra day has been
added to the variable month in the first eleven-year cycle and two days to the second eleven-year cycle.
One consequence of this configuration is that the start of a month falls behind the real Moon by 3 days
every 22 years. After 44 years it would therefore give the situation described by Pliny, who tells us that
the Druids’ month began on the sixth day after new moon.12
If this 3-day slippage is designed into the calendar then the lunar cycle is held naturally in balance
with the solar cycle over twenty-two years, with a discrepancy of only about thirteen minutes. This
compares most favourably with the two-hour discrepancy of the Meton cycle over nineteen years. The
arrangement shown in the reconstruction is at least an improvement on the Greek octaeteris.13 Certainly,
it is greatly superior to the unreformed Roman calendar. A further advantage is that a simple cycle of 5
+ 6 years repeats indefinitely, with no complex rules to remember.
However, if the eleven-year cycle is required to repeat with the 30-year Druid ages then a long-cycle
of 330 years (30 x 11) must elapse before the two cycles will mesh. In my earlier work I suggested that
a double Druid cycle of 60-years may have been employed – and indeed many such reconstructions are
possible without violating the rules of the extant 5-year cycle. However, I am now satisfied that the
simple 5+6-year formula given by Plato, repeated indefinitely, is all that is required.14
The present author’s hypothesis would suggest that, in addition to the extant five-year cycle of the
Coligny calendar, there should also be a ‘lost’ cycle of six years. However, until hard evidence of a
missing six-year cycle comes to light from somewhere in the Celtic regions, there can be no conclusive
proof that the Coligny Calendar actually operated this way.
The mechanism proposed is therefore an eleven-year lunisolar cycle of 4017/4018 days that was
allowed to wander alongside a thirty-year cycle based on the observation of Saturn. This cycle actually
requires less intervention than the Metonic cycle over two eleven-year cycles. The lunisolar difference
over eleven years is almost exactly a day and a half, so doubling this:
Sun:
Moon:
2 x 4017.6642 =
2 x 4016.1601 =
Difference =
8035.3284 days
8032.3202 days
3.0082 days
This slippage might explain why the Celtic months began on the sixth day of the moon at the era when
Pliny recorded it.
Thus, we see revealed the remarkable accuracy hidden within Plato’s Atlantis calendar. The calendar
as reconstructed here is beautifully simple. Given the realities of astronomy, constrained by the extant
Coligny fragment, we can be confident that any real calendar based upon alternating five- and six-year
periods must work in a similar way. I would invite any scholar to put this mechanism into their own
spreadsheets and satisfy themselves that this so.
The mention of this 5+6-year formula within Plato’s narrative is too great a coincidence for it to be
mere fiction. This passage has always convinced the present author of the authenticity of the Atlantis
myth. The preservation of the 5-year cycle by the Druids, and its mention independently in a FinnoUgrian myth, suggests that a real calendar of this kind may have been used by the kings of ancient
Europe just as is recorded in the Critias. This further raises the question of whether other aspects of
Plato’s Atlantis narrative deserve more respect as a record of history.
Tags: Coligny calendar, Celtic calendar, Gaulish calendar, Druids, Druid calendar, Pliny, Plato, Plato
Critias. Lunisolar calendar. 11-year cycle
Next Page:
A reconstruction of the 5+6-year cycle over 22 years as originally published in Under Ancient Skies in
2005. Other arrangements for the days per month are possible as long as the result conforms to the
lunisolar cycle and the known arrangement of the Coligny calendar.
Notes and References
1
Plato, Critias, 119 (translation by Benjamin Jowett)
Within the Timaeus, Plato says that the empire extended to the whole island and also to the
continent: Europe as far as Tyrrhenia and Libya. However, he refers to only one specific sub-king
who ruled over a peninsula of the island. By analogy with the later empires of Rome and Napoleon’s
Europe, we may posit that tributary kings also ruled over the conquered territories of Europe.
3
Dunbavin, P. The Atlantis Researches – the Earth’s Rotation in Mythology and Prehistory, Third
Millennium Publishing, Nottingham (1995)
4
Dunbavin, P. Atlantis of the West – The Case for Britain’s Drowned Megalithic Civilisation.
Constable & Robinson, London (2003)
5
W.F. Kirby (London 1907) Kalevala: the Land of Heroes, (republished by the Athlone Press, 1985);
see Runo L, p 634.
6
It should be recorded that when the author attempted to publish a similar article on the 5+6-year
calendar in a journal of astronomical history, the Kalevala reference was dismissed as ‘a patently
solar reference’ having no place in a discussion of a lunisolar calendar. Indeed, the very mention of
Plato’s Critias as the primary source for the calendar evidence produced a dismissive negative
reaction. Hence the article was revised for publication elsewhere.
7
Caesar, The Gallic War, VI.xiii-xiv; VI.xviii.
8
Caesar, The Gallic War, VI,Xiii
9
Duval, P.-M & Pinault, G Recueil des Inscriptions Gauloises, 3: Les calendriers (Coligny, Villards
d’Heria). Paris: Supplement a Gallia 45 (1986)
10
Mac Neill, E. On the Notation and Calligraphy of the Calendar of Coligny, Eriu, X, pp1-67 (1928)
11
Olmsted, G. The Gaulish Calendar, Dr Rudolph Habelt GMBH, Bonn (1992)
12
Pliny, Natural History, XXX. xiii
13
Geminus, Elementa Astronomiae, C.8.
14
Dunbavin, P. Under Ancient Skies: Ancient Astronomy and Terrestrial Catastrophism, Third
Millennium Publishing, Nottingham (2005).
2
Citation footnote added 2021
The above text is unchanged from the form that the article was left in abeyance in 2006, other than
formatting for publication in Prehistory Papers in 2020:
https://www.academia.edu/50962364/The_Neolithic_Calendar_in_Platos_Critias
Dunbavin, Paul (2020) The Neolithic Calendar, in Prehistory Papers, pp 13-22, Third Millennium
Publishing, Beverley, ISBN: 978-0-9525029-4-4 (a previously unpublished 2006 paper)
…and the publication of a related article in the C&C Review in 2018.
Dunbavin, Paul (2020) On the Coligny Calendar and the Calendar in Plato’s Critias, in Prehistory
Papers, pp 23-30 Third Millennium Publishing, Beverley, ISBN: 978-0-9525029-4-4
Dunbavin, Paul (2018) On the Coligny Calendar and the Calendar in Plato’s Critias, in Chronology &
Catastrophism REVIEW, 2018:2 pp 50-53.
On the Coligny Calendar and the Neolithic Calendar in Plato’s Critias
An article published in SIS Review similar to the previously unpublished 2006 paper.
Summary
Evidence of a calendar within the Critias of Plato; and extant evidence of the calendar used by the
Druids of Gaul can both be reconstructed according to an eleven-year lunisolar cycle of 5+6 years and
may in fact be the same calendar. This short article summarises the author’s previously published work
on these subjects.
Introduction
The existence of astronomical alignments at Neolithic monuments such as the stone circles, once itself
controversial, is now virtually accepted; somewhat intuitively these may be associated with various
Celtic seasonal festivals and yet without any real understanding of how their calendar worked. Did
they have a system of calendar dates and an era such as we do?
Recent DNA studies have largely overturned the old notion of a wave of Iron-Age Celts invading
Neolithic Britain and Ireland. Neither may we any longer talk vaguely about a ‘Celtic’ calendar as if
continental invaders had introduced their religion and culture, including their calendar, at some
convenient period just prior to the Roman invasions. When we read Julius Caesar, who says that the
Druids and their culture were ‘found existing’ in Britain, then we need to take such a statement by a
contemporary chronicler far more seriously. We may now only use the term ‘Celtic’ as a loose linguistic
grouping.
It is not my intention here to go too deeply into source references in what is intended as an interesting
article for a general audience rather than a scientific paper. I explored these themes in much greater
detail some years ago in my books Atlantis of the West and Under Ancient Skies and in a further
unpublished paper; where all the source references may be found for anyone who wishes to verify my
conclusions.1 There is nothing here that I have not published previously.
The Calendar of Coligny
Clues to the calendar of the Gaulish Druids, and thereby of their astronomy, come from the “Calendar
of Coligny” (Fig. 1), discovered in a French vineyard in 1897. It was part of a wall plaque clearly
designed for display, similar to the annual calendar on your own wall. Its notation displays a five-year
cycle of lunar dates. Conventional academic assessment has been inconclusive because a five-year
period does not work as a lunisolar cycle. This has led to a conclusion that the Celts had only a poor
knowledge of astronomy.
Fig.1. A fragment of the Calendar of Coligny
The structure of the calendar shows 62 months of 29 and 30 days arranged as follows:
30 30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 30 29 30 30* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
.
30 29 30 29 30 30 .
29 30 29* 29 30 30
Total: 1831 days
The month called Equos, ‘horse’, had nominally 29 days and was variable according to rules that
cannot be determined from the fragment alone. It may sometimes have been 28 or 30 days, so the 5year cycle could have been anything between 1826 and 1835 days.
Other Ancient Calendars
Conservative academic opinion will only allow that the Calendar of Coligny was devised by the IronAge Druids in Gaul just prior to the Roman invasion. A moment’s thought will show the problem here.
The Gregorian calendar on your own wall was printed last year, but the printer did not devise it! It
holds months named after Roman emperors and days named for Germanic gods that are thousands
of years old; the structure is but a revision of the old Latin calendar. Therefore, we may assume that
the devices on the Gaulish calendar similarly evolved from much earlier roots. This would take us back
to the Neolithic and the era of the megalith builders.
The basis of any such calendar is a natural lunisolar cycle, which arranges the months so that they stay
naturally in step with the seasons. Ideally, festivals should repeat at roughly the same season every
year without requiring constant revision by priests and astrologers. In a solar calendar, such as our
familiar Julio-Gregorian calendar, it is the solar year that is held and the lunar phases are allowed to
appear as they fall. However, in a lunar calendar the months must begin strictly with a phase of the
moon and be arranged as either 29 or 30 days long. Consider the following (current astronomical data
used here and subsequently):
One lunar month =
29.530589 days
62 lunar months =
1830.8965 days
One solar year =
365.2422 days
5 solar years =
1826.211 days
Difference:
4.6855 days
As the month and year are not integral, the best arrangement is therefore to use a cycle where a
multiple of lunar months is almost equal to a multiple of solar years. There is no precise equivalence,
but some of the closest available are 8 years, 11 years and 19 years. The Greeks used the 8-year
Octaeteris; the Babylonians used the 19-year Meton Cycle; but no culture is attested to have used an
11-year cycle.
In the Greek Octaeteris, 3 extra intercalary months were added to make the lunar years equate
(roughly) to the solar years; in the Meton Cycle 7 extra months; and in an 11-year cycle 4 extra months
would be needed; so, all lunisolar calendars must employ some lunar years of 13-months in addition
to the 12-month years. If these are evenly spaced, then solstices and equinoxes never slip too far from
their proper season. A good analogy here is Christian Easter, which shows us the opposite case of a
lunar date wandering within a solar calendar, but always held in springtime.
However, the five-year period of the Calendar of Coligny is not a natural lunisolar cycle. 62 lunar
months are 1831 days; 5 solar years are 1826 days. This is an accumulating 5-day slippage. And yet
the notation shows provision for two intercalary months inserted every 2.5 years. This suffices to tell
us that the 5-year cycle is not the entire calendar. The Intercalary months serve no purpose unless
they were intended to hold the months in line with the seasons. So which cycle were they using? Was
it 8 years? 11 years? 19 years? Or perhaps they used some longer cycle.
Another factor to consider is how to make the days add-up. Every calendar must employ a month of
variable length inserted according to established rules. In the Julio-Gregorian calendar we add a day
to February each fourth year and because this is still inexact the Gregorian reform added a century
rule. In the Coligny Calendar we similarly see notation for a variable 29-day month. There is not enough
evidence to say how the rules were applied – but we know that there must have been such rules or
these extra days, like the intercalary months, would serve no purpose.
Other clues are offered by ancient writers who tell us a little about the Gauls; also, about the ancient
Britons; and of their priestly cast known as the Druids. Some other clues may come via myths and
legends. If one may quote Pliny the Elder here: 2
Mistletoe…is gathered with great reverence, above all on the sixth day of the moon (it is the moon that
marks out for them the beginning of months and years and cycles of thirty years) because this day is
already exercising great influence even though the moon is not half-way through its course
(Pliny, Natural History, XVI, 250).
The historian Plutarch also tells us that the Britons held their most important festival each thirty years
as the planet Saturn returned to the constellation Taurus. 3 Consider:
One synodic period of Saturn = 378.09 days
29 oppositions of Saturn = 29 x 378.09 = 10964.61 days = 30.02 solar years
In my earlier reconstruction of the Coligny Calendar, I therefore also considered whether the Druids
tried to base it on a precise thirty-year cycle held to Saturn’s rhythm. So, I investigated how the extant
five-year ‘Coligny’ cycle could be alternated with a six-year cycle. This would allow the short-term
calendar to mesh both with the thirty-year ages and also the 11-year lunisolar cycle.
This brings us on to the 5 + 6-year calendar intimated in Plato’s Critias. To recap his Atlantis narrative,
he tells us that the Egyptian priests remembered an island in the Atlantic that was struck by a
geological catastrophe at a time just before the beginning of the Egyptian state. The ancient dynasty
of kings who ruled this island and the Atlantic coastal regions would gather together “every fifth and
every sixth year alternately” to discuss policy. If they were all to know when to meet, then they must
have used the same calendar; and Plato is giving us a clue how it worked.4
The Coligny calendar gifts us a detailed knowledge of how a real 5-year cycle was constructed. From
this it is possible to work-out how a 6-year cycle must be arranged in order to make use of the 11-year
lunisolar cycle.
In the Coligny fragment the intercalary months are evenly spaced at 2.5-year intervals; one at the start
of the five-year cycle; the other in the middle of the third year. If we reconstruct a six-year cycle using
the same month layout, then the intercalary months should naturally be spaced at the beginning of
the cycle and at the start of the fourth year. Therefore, we may propose:
30 30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
30 30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
.
30 29 30 29 30 30 .
29 30 29* 29 30 29
Total 2184 days
Here, for simplicity, I have assumed that the variable month always has 29 days, so accuracy is down
to how you wish to organise the number of days that it has in each cycle; i.e. whether the calendar is
to be held strictly lunar; strictly solar; of perhaps to follow Saturn.
These two cycles, alternated, give 4015 days, whereas 11 solar years require approximately 4018 days;
so, the variable month would need an extra 3 days to be added each 11 years. One cannot know for
sure how the calendar rules actually operated, but it doesn’t really matter where you put the extra
days; so long as the number of days in the cycle add-up to your objective and this is fixed by astronomy.
So, the 5+6-year cycle casually mentioned by Plato can produce a very accurate calendar and the
reconstruction of the Calendar of Coligny yields a similar result. Is this a coincidence? Decide for
yourself. Of course, the Neolithic calendar would not look exactly like that of Iron Age Gaul, it’s a
process of evolution, just as our Gregorian calendar has evolved from its original Roman roots.
It was this detail, hidden in plain sight within the Critias, that convinced me of the authenticity of
Plato’s narrative – but you must decide for yourself how much of his catastrophism you wish to accept
alongside the description of ancient geopolitics.
Classical scholars, if they are to remain respectable, may only discuss Plato’s Atlantis narrative as if it
were a classical Greek story. They are not allowed to look beyond Crete, or the volcanic destruction
of Thera, for the inspiration behind Plato’s lost island; yet its own internal content states that it was
based on history told to Solon, as recorded by the Egyptian priests. The story is Egyptian, not Greek;
they recorded the organisation of an island in the Atlantic and of the Atlantic coastal regions. They tell
us that this was contemporary with the earliest period of the Egyptian state (pre-dynastic).
Egyptologists determine this was the late fourth millennium BC.
Archaeology tells us what was going-on along the Atlantic coasts of Europe in the late fourth
millennium BC. This was the middle-Neolithic; a time when the people began to build megalithic
monuments with calendrical alignments; in Britain, in Ireland, in Brittany and as far away as Malta. Is
it therefore so far out to equate the two? Is it so unacceptable to say that if you have a calendar on
your own wall with notations that are thousands of years old, then perhaps the first century Gauls
also used a calendar that was thousands of years old?
My own research has taught me always to believe what the ancient sources actually say, rather than
what academics think they are saying.
Notes and References
Atlantis of the West (2003) and Under Ancient Skies (2005) Both books are now available again in Kindle editions and an
unpublished paper intended for C&C Review is available (see previous), or via www.third-millennium.co.uk
1
The Latin is genuinely ambiguous here; some translators give that the months actually began on the sixth day after new
moon.
2
This shows us that the Druids were observing the Synodic Period; the time required for a planet to return to the same
position relative to the Sun as viewed by an observer on the Earth.
3
If you doubt the importance of using a common calendar, then recall when, in 1805, the Russian Tsar sent his army to join
with the Austrians before the Battle of Ulm, only to find that Napoleon had already defeated them! The Russians were still
following the unreformed Julian calendar which by then was 11 days out of synchronization with the Gregorian calendar used
by the Austrians.
4
Citation:
Chronology & Catastrophism REVIEW 2018:2 pp 50-53
Dunbavin, Paul (2020) On the Coligny Calendar and the Calendar in Plato’s Critias, in Prehistory Papers, pp 23-30 Third
Millennium Publishing, Beverley, ISBN: 978-0-9525029-4-4
On the Coligny Calendar and the Neolithic Calendar in Plato’s Critias
Paul Dunbavin
Citation: Chronology & Catastrophism REVIEW 2018:2 pp 50-53
Summary
Evidence of a calendar within the Critias of Plato; and extant evidence of the calendar used by the Druids of
Gaul can both be reconstructed according to an eleven-year lunisolar cycle of 5+6 years and may in fact be the
same calendar. This short article summarises the author’s previously published work on these subjects.
Introduction
The existence of astronomical alignments at Neolithic monuments such as the stone circles, once itself
controversial, is now virtually accepted; somewhat intuitively these may be associated with various Celtic
seasonal festivals and yet without any real understanding of to how their calendar worked. Did they have a
system of calendar dates and an era such as we do?
Recent DNA studies have largely overturned the old notion of a wave of Iron-Age Celts invading Neolithic
Britain and Ireland. Neither may we any longer talk vaguely about a ‘Celtic’ calendar as if continental invaders
had introduced their religion and culture, including their calendar, at some convenient period just prior to the
Roman invasions. When we read Julius Caesar, who says that the Druids and their culture were ‘found existing’
in Britain, then we need to take such a statement by a contemporary chronicler far more seriously. We may now
only use the term ‘Celtic’ as a loose linguistic grouping.
It is not my intention here to go too deeply into source references in what is intended as an interesting article for
a general audience rather than a scientific paper. I explored these themes in much greater detail some years ago
in my books Atlantis of the West and Under Ancient Skies and in a further unpublished paper; where all the
source references may be found for anyone who wishes to verify my conclusions.1 There is nothing here that I
have not published previously.
The Calendar of Coligny
Clues to the calendar of the Gaulish Druids, and thereby of their astronomy, come from the “Calendar of
Coligny” (Fig. 1), discovered in a French vineyard in 1897. It was part of a wall plaque clearly designed for
display, similar to the annual calendar on your own wall. Its notation displays a five-year cycle of lunar dates.
Conventional academic assessment has been inconclusive because a five-year period does not work as a
lunisolar cycle. This has led to a conclusion that the Celts had only a poor knowledge of astronomy.
Fig.1. The Calendar of Coligny
The structure of the calendar shows 62 months of 29 and 30 days arranged as follows:
30
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
30
29
30
30* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
Total 1831 days
The month called Equos, ‘horse’, had nominally 29 days and was variable according to rules that cannot be
determined from the fragment alone. It may sometimes have been 28 or 30 days, so the 5-year cycle could have
been anything between 1826 and 1835 days.
Other Ancient Calendars
Conservative academic opinion will only allow that the Calendar of Coligny was devised by the Iron-Age
Druids in Gaul just prior to the Roman invasion. A moment’s thought will show the problem here. The
Gregorian calendar on your own wall was printed last year, but the printer did not devise it! It holds months
named after Roman emperors and days named for Germanic gods that are thousands of years old; the structure
is but a revision of the old Latin calendar. Therefore, we may assume that the devices on the Gaulish calendar
similarly evolved from much earlier roots. This would take us back to the Neolithic and the era of the megalith
builders.
The basis of any such calendar is a natural lunisolar cycle, which arranges the months so that they stay naturally
in step with the seasons. Ideally, festivals should repeat at roughly the same season every year without requiring
constant revision by priests and astrologers. In a solar calendar, such as our familiar Julio-Gregorian calendar, it
is the solar year that is held and the lunar phases are allowed to appear as they fall. However, in a lunar calendar
the months must begin strictly with a phase of the moon and be arranged as either 29 or 30 days long. Consider
the following (current astronomical data used here and subsequently):
One lunar month = 29.530589 days
62 lunar months = 1830.8965 days
One solar year = 365.2422 days
5 solar years = 1826.211 days
Difference: 4.6855 days
As the month and year are not integral, the best arrangement is therefore to use a cycle where a multiple of lunar
months is almost equal to a multiple of solar years. There is no precise equivalence, but some of the closest
available are 8 years, 11 years and 19 years. The Greeks used the 8-year Octaeteris; the Babylonians used the
19-year Meton Cycle; but no culture is attested to have used an 11-year cycle.
In the Greek Octaeteris, 3 extra intercalary months were added to make the lunar years equate (roughly) to the
solar years; in the Meton Cycle 7 extra months; and in an 11-year cycle 4 extra months would be needed; so all
lunisolar calendars must employ some lunar years of 13-months in addition to the 12-month years. If these are
evenly spaced then solstices and equinoxes never slip too far from their proper season. A good analogy here is
Christian Easter, which shows us the opposite case of a lunar date wandering within a solar calendar, but always
held in springtime.
However, the five-year period of the Calendar of Coligny is not a natural lunisolar cycle. 62 lunar months are
1831 days; 5 solar years are 1826 days. This is an accumulating 5-day slippage. And yet the notation shows
provision for two intercalary months inserted every 2.5 years. This suffices to tell us that the 5-year cycle is not
the entire calendar. The Intercalary months serve no purpose unless they were intended to hold the months in
line with the seasons. So which cycle were they using? Was it 8 years? 11 years? 19 years? Or perhaps they
used some longer cycle.
Another factor to consider is how to make the days add-up. Every calendar must employ a month of variable
length inserted according to established rules. In the Julio-Gregorian calendar we add a day to February each
fourth year and because this is still inexact the Gregorian reform added a century rule. In the Coligny Calendar
we similarly see notation for a variable 29 day month. There is not enough evidence to say how the rules were
applied – but we know that there must have been such rules or these extra days, like the intercalary months,
would serve no purpose.
Other clues are offered by ancient writers who tell us a little about the Gauls; also about the ancient Britons; and
of their priestly cast known as the Druids. Some other clues may come via myths and legends. If one may quote
Pliny the Elder here: 2
Mistletoe…is gathered with great reverence, above all on the sixth day of the moon (it is the moon that
marks out for them the beginning of months and years and cycles of thirty years) because this day is already
exercising great influence even though the moon is not half-way through its course. (Pliny, Natural
History, XVI, 250)
The historian Plutarch also tells us that the Britons held their most important festival each thirty years as the
planet Saturn returned to the constellation Taurus. 3 Consider:
One synodic period of Saturn = 378.09 days
29 oppositions of Saturn = 29 x 378.09 = 10964.61 days = 30.02 solar years
In my earlier reconstruction of the Coligny Calendar, I therefore also considered whether the Druids tried to
base it on a precise thirty-year cycle held to Saturn’s rhythm. So I investigated how the extant five-year
‘Coligny’ cycle could be alternated with a six-year cycle. This would allow the short-term calendar to mesh
both with the thirty-year ages and also the 11-year lunisolar cycle.
This brings us on to the 5 + 6 year calendar intimated in Plato’s Critias. To recap his Atlantis narrative, he tells
us that the Egyptian priests remembered an island in the Atlantic that was struck by a geological catastrophe at a
time just before the beginning of the Egyptian state. The ancient dynasty of kings who ruled this island and the
Atlantic coastal regions would gather together “every fifth and every sixth year alternately” to discuss policy. If
they were all to know when to meet then they must have used the same calendar; and Plato is giving us a clue
how it worked.4
The Coligny calendar gifts us a detailed knowledge of how a real 5-year cycle was constructed. From this it is
possible to work-out how a 6-year cycle must be arranged in order to make use of the 11-year lunisolar cycle.
In the Coligny fragment the intercalary months are evenly spaced at 2.5 year intervals; one at the start of the five
year cycle; the other in the middle of the third year. If we reconstruct a six-year cycle using the same month
layout, then the intercalary months should naturally be spaced at the beginning of the cycle and at the start of the
fourth year. Therefore we may propose:
30
30
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
30
29
30
29
30
30
.
29
30
29* 29
30
29
Total 2184 days
Here, for simplicity, I have assumed that the variable month always has 29 days, so accuracy is down to how
you wish to organise the number of days that it has in each cycle; i.e. whether the calendar is to be held strictly
lunar; strictly solar; of perhaps to follow Saturn.
These two cycles, alternated, give 4015 days, whereas 11 solar years require approximately 4018 days; so the
variable month would need an extra 3 days to be added to the variable month each 11 years. One cannot know
for sure how the calendar rules actually operated, but it doesn’t really matter where you put the extra days; so
long as the number of days in the cycle add-up to your objective and this is fixed by astronomy.
So the 5+6 year cycle casually mentioned by Plato can produce a very accurate calendar and the reconstruction
of the Calendar of Coligny yields a similar result. Is this a coincidence? Decide for yourself. Of course the
Neolithic calendar would not look exactly like that of Iron Age Gaul, it’s a process of evolution, just as our
Gregorian calendar has evolved from its original Roman roots.
It was this detail, hidden in plain sight within the Critias, that convinced me of the authenticity of Plato’s
narrative – but you must decide for yourself how much of his catastrophism you wish to accept alongside the
description of ancient geopolitics.
Classical scholars, if they are to remain respectable, may only discuss Plato’s Atlantis narrative as if it were a
classical Greek story. They are not allowed to look beyond Crete, or the volcanic destruction of Thera, for the
inspiration behind Plato’s lost island; yet its own internal content states that it was based on history told to
Solon, as recorded by the Egyptian priests. The story is Egyptian, not Greek; they recorded the organisation of
an island in the Atlantic and of the Atlantic coastal regions. They tell us that this was contemporary with the
earliest period of the Egyptian state (pre-dynastic). Egyptologists determine this was the late fourth millennium
BC.
Archaeology tells us what was going-on along the Atlantic coasts of Europe in the late fourth millennium BC.
This was the middle-Neolithic; a time when the people began to build megalithic monuments with calendrical
alignments; in Britain, in Ireland, in Brittany and as far away as Malta. Is it therefore so far out to equate the
two? Is it so unacceptable to say that if you have a calendar on your own wall with notations that are thousands
of years old, then perhaps the first century Gauls also used a calendar that was thousands of years old?
My own research has taught me always to believe what the ancient sources actually say, rather than what
academics think they are saying.
Notes and References
1
Atlantis of the West (2003) and Under Ancient Skies (2005). Both books are now available again in Kindle
editions and an unpublished paper intended for C&C Review is available upon request to the editorial address, or
via www.third-millennium.co.uk
2
The Latin is genuinely ambiguous here; some translators give that the months actually began on the sixth day
after new moon.
3
This shows us that the Druids were observing the Synodic period; the time required for a planet to return to the
same position relative to the Sun as viewed by an observer on the Earth.
4
If you doubt the importance of using a common calendar, then recall when, in 1805, the Russian Tsar sent his
army to join with the Austrians before the Battle of Ulm, only to find that Napoleon had already defeated them!
The Russians were still following the unreformed Julian calendar which by then was 11 days out of
synchronization with the Gregorian calendar used by the Austrians.
Citation: Chronology & Catastrophism REVIEW 2018:2 pp 50-53