The Gaulish (Coligny) Calendar


Today's date in the Coligny calendar:


This is the calendar which was used in Gaul in the latter part of the BCE period and, possibly, into the beginning of the CE period. The workings of this calendar were preserved in a bronze tablet that was discovered buried near Coligny in France. This was, presumably, because the calendar was banned by the Romans as it was an indication of druidic practices and, more importantly, flouted the rule that the Julian calendar should be the official calendar in all parts of the Roman Empire.

The calendar has been reconstructed from the tablet and started anew by various people, e.g. Ray White, who begins the current epoch (NCC) on 8th October 1999. This date was chosen because it "corresponds with a new moon on the date of the beginning of European winter. (when the sun is at 15 Deg Scorpio).". According to my information, however, the new moon in October 1999 actually took place on 9th October. Since lunar calendars usually begin a new month on the day after a new moon*, I have changed the epoch to the 10th October 1999 for my reckoning, with the first year of the epoch being NCC 1.

* This is debatable as there is conjecture that the Gaulish month began at 1st quarter, or even 3rd quarter, and not with the new moon. However, I have chosen to stick to modern convention and to take the month as beginning the day following the new moon.

How the calendar works

The workings of the Gaulish calendar would probably be a mystery to us if the remains of a bronze tablet had not been discovered near Coligny in France. It is thought that the Gauls buried this to preserve the details of their calendar, which had been banned by the ruling Romans. A section of the tablet is shown to the right.

On a yearly basis the calendar works like most lunar calendars, i.e. it has 12 months of 29 or 30 days, keeping in tune with the phases of the moon. This gives a "year" of 354 days and means that the months would soon start drifting through the seasons, getting earlier and earlier each year (as in the Islamic calendar). To counteract this, an extra month (Ciallos Bis) is inserted every 2 and a half years to keep in synchronisation with the seasons, which works in the short term, but over longer periods is very inaccurate.

The Gaulish calendar was therefore run on a cycles of thirty years at a a time, split into smaller periods of five years, each of which had 62 months including the two extra months that are inserted. In the sixth and final period of the 30 years, however, only one extra month was inserted, so that five-year period only had 61 months.

This was meant to make the calendar more accurate, but my own attempt at recreating this system produced an error between the calendar and the moon's phases of five days (but see My Hypothesis below). It would seem that there would have to be other, longer term adjustments to keep the calendar synchronised. A better system is one that follows the the natural 19-year cycle of moon phases, like the Jewish calendar, of which I have created a version on my Celtic calendar page.

Gaulish months

According to the Coligny tablet, the Gaulish months, along with their respective lengths, were as follows:

Name Days
Samoni (Summer's End) 30
Dumannos (World Darkness) 29
Rivros (Cold & Ice) 30
Anagantios (Staying Home & Storing) 29
Ogroni (Cold's End) 30
Cutios (Wind) 30
{Ciallos Bis (Extra Moon)} {30}
Giamoni (Winter's End) 29
Semivisonna (Midsummer Brightness) 30
Equos (Horse) 29 {30}
Elembivos (Claim) 29
Edrinios (Arbitration) 30
Cantlos (Song) 29
{Ciallos Bis (Extra Moon)} {30}

The structure of the calendar over a five year period is shown in the diagram of the tablet displayed below. The tablet was split into 128 squares in 16 columns and 8 rows. The extra months, for one reason or another, cover four squares each, whereas the other months only cover two squares each.

Good and Bad

It is thought that the Gauls divided all time periods into two halves, called MAT (good) or ANMAT (bad), and that these referred to either light and dark parts or long and short parts. For instance, the months were MAT if they had 30 days, and ANMAT if they had 29 days. A year had six months that were MAT (the summer months) and six that were ANMAT (the winter months). Each day began at sunset and the period from then until sunrise was ANMAT and from sunrise until sunset was MAT.

This was also thought to be how the months were split as well. If so, the likelihood is that each month actually began at first quarter instead of new moon, and that the period from then, through full moon, until third quarter was the MAT, or light, half of the month. The second half of the month from third quarter through new moon and back to first quarter again was ANMAT, or dark.

Each square on the tablet covers one half month, starting with the light (MAT) half then the dark (ANMAT) half, which was headed in each case by the word ATENOUX, which is thought to have meant "return of night".

My hypothesis

To have the calendar work accurately over a period of centuries, if not millennia, is very difficult, but I have worked out a relatively simple format which produces an average month length of 29.5309973, which differs from the average lunation by 0.0004085, i.e. less than a minute. This may be noticeable over thousands of years, but over a few hundred it would be accurate enough.

 
Year 1
Year 2
Year 3
Year 4
Year 5
Ciallos Bis 30        
Samoni 30 30 30 30 30
Dumannos 29 29 29 29 29
Rivros 30 30 30 30 30
Anagantios 29 29 29 29 29
Ogroni 30 30 30 30 30
Cutios 30 30 30 30 30
Ciallos Bis     30    
Giamoni 29 29 29 29 29
Semivisonna 30 30 30 30 30
Equos 29 29 29 29 30
Elembivos 29 29 29 29 29
Edrinios 30 30 30 30 30
Cantlos 29 29 29 29 29
TOTAL 384 354 384 354 355

The five year model shown above would be used over the first 25 years of each 30 year period, and the following points should be noted:

  1. There are 12 months in a standard year, six having 30 days and six with 29 days.
  2. At the beginning of the first year, and halfway through the third year, an extra month is inserted, with 30 days.
  3. In the fifth year, the ninth month (Ekwos) has an extra day.

In the last five year period of the overall 30 year cycle, there is only one extra month, i.e. the one in the first year. Consequently, the third year only has 354 days in this period instead of the usual 384. Using this simple and easy to remember system, the calendar remains very accurate over long periods of time.


Main Page