File PDF .it

Condividi facilmente i tuoi documenti PDF con i tuoi contatti, il Web e i Social network.

Inviare un file File manager Cassetta degli attrezzi Ricerca PDF Assistenza Contattaci



The Lunar Year of the Coligny Calendar a .pdf



Nome del file originale: The_Lunar_Year_of_the_Coligny_Calendar_a.pdf
Titolo: pag 1-6 voorwerk

Questo documento in formato PDF 1.6 è stato generato da Result Export Task Processor, ed è stato inviato su file-pdf.it il 18/05/2016 alle 20:38, dall'indirizzo IP 213.144.x.x. La pagina di download del file è stata vista 1015 volte.
Dimensione del file: 1.3 MB (26 pagine).
Privacy: file pubblico




Scarica il file PDF









Anteprima del documento


TALANTA XXXVIII-XXXIX (2006-2007)

THE LUNAR YEAR OF THE COLIGNY CALENDAR
AS A PRECEDENT FOR THE INSULAR LUNAR YEAR
Brent Davis
In this essay, the origins of the recently-recovered Insular calendar outlawed at
the Synod of Whitby in AD 664 are reviewed, including the calendar’s documented origin in early fifth-century Gaul, its basis on a third-century source from Asia
Minor, and its survival in a nineteenth-century Irish folk-verse that illustrates the
structure of the calendar’s lunar year. It is then demonstrated that a precedent for
this distinctive lunar year can be found in the late-second-century Gaulish brass
artifact called the Coligny Calendar, if the year on the artifact is taken to begin
with March. Finally, the etymologies of the Gaulish month-names are revisited in
this context.
THE INSULAR CALENDAR

In his Historia Ecclesiastica (III.25), Bede describes the calendar disputes that
preoccupied the Insular church in the mid-seventh century. These disputes culminated in 664 in the Synod of Whitby, which declared invalid the 84-year lunisolar calendar of the Insular church, and compelled Christians there to adopt the
ecclesiastical lunisolar calendar of Rome. But none of the primary sources that
mention the conflict describe how this older Insular calendar operated (McCarthy
1993, 205), and so its exact structure remained a mystery.
This situation changed in 1985 with the discovery of the Padua latercus, a table
of Easter dates compiled from the calendar that the Synod of Whitby had rejected (McCarthy and Ó Cróinín 1987-8). As Easter is a lunisolar feast, with a date
depending upon both the sun and the moon, such a table of dates enables the
reconstruction of the lunisolar calendar that produced them; and so from the
Padua latercus, the ‘lost’ lunisolar calendar used by Insular churches before 664
was recovered.
The recovered calendar was presented and analysed by McCarthy (McCarthy
1993). The calendar’s solar year, naturally enough, is the Roman calendar –
specifically, the Julian calendar in its post-Augustan form. The lunar year, however, displays a distinctive characteristic: each lunar month is one day shorter than
the corresponding solar month – for example, lunar May has 30 days, lunar June

9

has 29 days, lunar July has 30 days, and so on (ibid., 212).

Persistence of the Insular calendar in oral tradition
The Padua latercus remains the latest known written source from which the
Insular calendar can be derived. McCarthy (ibid.), however, publicises one later
oral source: a booklet on Irish numerals (O’Leary ca.1920) whose author quotes
a verse that he learned as a child in West Cork in the mid-nineteenth century. The
verse contains a formula for converting a Roman calendar date to a lunar calendar date:
Cómhrimh síos ón Márta mbán,
Go dtí an mí n-a mbeidh tú ann,
Cuir aon fé n-a gceann
Lá an mhí, agus an t-epacht.
Aon nídh fé bhun nó os cionn trí dheich,
Sin agat aois na rae.

Count from the beginning of March,
Down to the month in which you are.
Put one less than
The day of the month, and then the epact.
Then anything below or over thirty,
There you have the age of the moon.

It is worthwhile examining this verse in some detail, as its principles will be of
use later in the essay. To see how the formula works, let us first reword
McCarthy’s translation to make the verse a little easier to use:
From the start of March, count in months how far
It is to the month in which you are.
Add the day of the month less one, and then
Add the age of the moon when March began.
If greater than thirty, take thirty away;
The result is the age of the moon today.

‘The age of the moon’ means ‘the day of the lunar month’. If, for example, this
year’s March 1 fell on the 8th day of a lunar month, then ‘the age of the moon
when March began’ = 8.
We could then also say that this year’s March 1 fell on the 8th day of lunar March
– ‘lunar March’ being simply defined as the lunar month in progress on March 1
(McCarthy 1993, 207). In fact, all the lunar months in this calendar are so defined:
‘lunar April’ is the lunar month in progress on April 1, ‘lunar May’ is the lunar
month in progress on May 1, and so on.
So then, let us suppose that this year’s March 1 did indeed fall on the 8th day of
lunar March, such that ‘the age of the moon when March began’ = 8; and then let
us suppose that today is (say) July 11. What day of lunar July is it today? The
verse provides the answer, as shown in Table 1.

10

TABLE 1: Using the verse to convert July 11 to a lunar date
Operation

Result

Add the day of the month less one ...

Today is July 11, so add 11 - 1 = 10:

4 + 10 = 14

If greater than thirty, take thirty away;

(not applicable)

22 - 0 = 22

Verse

From the start of March, count in months how far
it is to the month in which you are.

... and then, add the age of the moon when
March began.

The result is the age of the moon today.

From March 1 to July 1 there are 4 months:

March 1 was the 8th day of lunar March, so add 8:

4

14 + 8 = 22
22

That is: July 11 falls on the 22nd day of lunar July. We can check the result the
modern way:
1. July 11 always falls 132 days after March 1.
2. A lunar month (according to the moon’s phase) contains 29.531 days,
so 132 days contain 4.4699 lunar months, or 4 lunar months and 13.88
days, or (to the nearest whole day) 4 lunar months and 14 days.
3. March 1 was day 8 of lunar March, so adding 4 lunar months and 14
days will produce day 22 of lunar July.

The verse, then, has correctly converted the Roman calendar date into a lunar one.

Now let us take a simpler case, and suppose that March 1 fell on the 1st day of
lunar March, and that today is July 1. We recalculate, as shown in Table 2.
TABLE 2: Using the verse to convert a July 1 to a lunar date
Verse

From the start of March, count in months how far
it is to the month in which you are.

Add the day of the month less one ...

Operation

Result

Today is July 1, so add 1 - 1 = 0:

4+0=4

From March 1 to July 1 there are 4 months:

4

... and then add the age of the moon when March began. March 1 was the 1st day of lunar March, so add 1: 4 + 1 = 5
If greater than thirty, take thirty away;

The result is the age of the moon today.

(not applicable)

5–0=5
5

That is: July 1 falls on the 5th day of lunar July; while ‘5’ is one greater than the
number of months between March 1 and July 1.
In fact, by substituting any month for July in the preceding table, readers will find
that the final result is always one greater than the number of whole Roman
months since March 1, as shown in Table 3.

Furthermore: for this simple pattern to exist, these lunar months clearly must each
be one day shorter than their solar equivalents – a distinctive characteristic of the
‘lost’ Insular calendar, as mentioned earlier. In short: this nineteenth-century Irish

11

TABLE 3: Day of the lunar month at the start of each Roman month, when lunar
March begins on March 1
Roman date
01-Mar
01-Apr

01-May
01-Jun
01-Jul

01-Aug
01-Sep
01-Oct

Roman months since March 1
0

1

2

3
4

5

6

7

Corresponding day of lunar month
1st day of lunar March
2nd day of lunar April
3rd day of lunar May
4th day of lunar June
5th day of lunar July

6th day of lunar August

7th day of lunar September

8th day of lunar October (etc.)

verse appears to preserve a method for coordinating the Roman solar calendar
with the Insular lunar calendar outlawed at Whitby in 664, despite there being no
known written source later than the tenth-century Padua latercus.
There would have been a practical reason for preserving the Insular calendar orally in this way. The decision at Whitby had been deeply unpopular amongst some
groups, especially in the far north; so it seems that without the sanction to use the
Insular calendar openly, its adherents preserved its basic principles in spoken
form instead. ‘This formula transmitted orally in the nineteenth century,’ notes
McCarthy, ‘is in accord with computistic technicalities written down a thousand
years earlier, eloquent testimony as to the conservative nature of Irish oral tradition’ (ibid., 213).

Origins of the Insular calendar
The question remains, however: where did this calendar originate? In a followup
study (McCarthy 1994), the origins of the Insular calendar are investigated, and
it is demonstrated that Aldhelm of Malmesbury’s late-seventh-century attribution
of this calendar to Sulpicius Severus of Aquitaine was in fact correct. Sulpicius
was a wealthy Gaulish lawyer, historian and chronologer who, converted to
Christianity in 395, published his 84-year lunar cycle in or shortly after 403, and
died in about 420. His calendar was in use in northern Ireland by 438, brought
there most likely by Gaulish Pelagian missionaries who favoured the calendar
because of Sulpicius’ early support of Pelagianism (ibid., 43).
The study then establishes that Sulpicius’ distinctive lunar year, with its months a
day shorter than their solar equivalents, was itself borrowed from Liber Anatolii
(ibid., 25-6), a calendrical work written by Anatolius of Laodicea – now Ladikia,
Syria – sometime before his death in 283. Anatolius was a highly-educated, welltravelled scholar who had served as head of the Aristotelian school at Alexandria,
and his calendrical writings had been praised by Eusebius for their erudition.
Sulpicius, in about 402, shortly before his calendar was published, wrote to
Paulinus of Nola requesting chronological information; Paulinus forwarded this
letter to Rufinus, whose Latin edition of Eusebius would be published in the fol-

12

lowing year. Rufinus’ edition of Eusebius contains a description of Anatolius’
work taken directly from Liber Anatolii, so it seems most likely that Rufinus was
the one who supplied this work to Sulpicius (ibid., 41).
Anatolius was careful to claim authorship of his inventions, but he does not claim
authorship of this lunar year, and so it is thought probable that he borrowed it
from an earlier, unknown source. McCarthy notes that, as to ‘the characteristics
of the lunar calendar on which the latercus was based and which derive from the
lunar table given in Liber Anatolii, it has to be said that these resemble no other
known lunar cycle’ (ibid., 35).
However: a precedent for Anatolius’ lunar year can in fact be found in the secondcentury Gaulish artifact known as the Coligny Calendar, which predates
Anatolius by about a century. McCarthy explicitly mentions the Gaulish lunar
year as unlike that of Anatolius (ibid.); and so my claim will need to be demonstrated, after a brief introduction to the calendar itself.
THE COLIGNY CALENDAR

The artifact
In 1897, a brass hoard was unearthed in a vineyard near Coligny, France. The
hoard consisted of fragments of a 1m x 1.5m brass plate, on which was inscribed
a 62-month calendar in the Gaulish language, using Roman letters (Fig.1).

Fig. 1. The condition of the artifact: shaded areas represent lacunae.

The plate had been deliberately broken. About half the fragments were interred at
the site, together with a partial brass statue of Mars (Duval and Pinault 1986, 357). The paleography indicates that the plate was engraved in the latter half of the
second century; and the style of the statue – though earlier, perhaps AD 100 – supports this late date (ibid., 33-5).

13

Type of calendar
Briefly, calendars are of three types:
1. Solar calendars, such as the Julian, mark solar months, or twelfths of a solar
year.
2. Lunar calendars, such as the Islamic, mark lunar months, as measured by the
moon’s phase. Twelve lunar months are 11 days shorter than twelve solar
months; so a calendar such as the Islamic begins 11 days earlier in each successive solar year, and in about 34 years will drift backward through the entire
solar year.
3. Lunisolar calendars mark both lunar and solar months, and attempt to ensure
that they remain in close alignment (Samuel 1972, 12). This is accomplished
by introducing a 13th lunar month – the intercalary month – once the drift
amounts to a month, thereby restoring the lunar months to their original positions relative to the solar ones. The best lunisolar calendars were able to
ensure that the lunar months consistently ended within the same solar months
every year (McCarthy 1993, 207).
The presence of two intercalary months on the Coligny plate indicates that the
calendar is of the third type, lunisolar, and that its users were concerned with limiting the drift of the lunar months against the solar. The lunar months are explicitly engraved on the plate, while the solar months are implicit; their locations will
be demonstrated.
Explicit lunar months
The plate contains 62 lunar months, and is divided into two 31-month halves,
each half consisting of an intercalary month followed by 30 ordinary months. The
months in each half are arranged into eight columns of four, with each intercalary
month occupying twice the vertical space of an ordinary month (Fig. 2).
ICA

ICB

Fig. 2. The plate consists of two halves and 62 lunar months.

The name of the first intercalary month is missing from the plate, and the name
of the second is only partially extant. In this essay, they are referred to as ICA and

14

ICB. Both contain 30 days. The names and lengths of the 12 ordinary lunar months

are shown in Table 4.

TABLE 4: Names and lengths of ordinary lunar months
Month

Length

Month

Length

Month

Length

Rivros

30 days

Giamon

29 days

Edrin

30 days

Samon

30 days

Duman

Anagant

Ogron

29 days

30 days

Cutios

29 days

Simivison

Equos

30 days

29 or 30 days

Elembiu

30 days

29 days

Cantlos

29 days

These months are arranged on the plate in vertical columns from left to right, as
shown in Fig. 3.
ICA

Riv

Gia

Edr

Riv

Gia

Edr

Riv

Ana

Sim

Can

Ana

Sim

Can

Ana

Sam

Ogr

Equ

Sam

Ogr

Equ

Sam

Ogr

Dum

Cut

Ele

Dum

Cut

Ele

Dum

Cut

ICB

Equ

Sam

Ogr

Equ

Sam

Ogr

Equ

Ele

Dum

Cut

Ele

Dum

Cut

Ele

Gia

Edr

Riv

Gia

Edr

Riv

Gia

Edr

Sim

Can

Ana

Sim

Can

Ana

Sim

Can

Fig. 3. Arrangement of months on the plate; the five years are separated by bold
borders.

As shown by the bold borders, the plate consists of five years. ICB occurs in the
centre of Year III, and ICA occurs before Year I – though ICA is more properly considered the final month of the preceding year, rather than the first month of Year
I (Lainé-Kerjean 1943, 251-2).

Implicit solar months
The calendar’s solar months are not explicitly shown on the plate, but their
boundaries can be deduced from the notation of the intercalary months. As an
illustration, Fig. 4 shows ICB, the second intercalary month.

15

Fig. 4. Intercalary Month B.

Day 13, for example, contains the notation +II MD SAMONI. Meanwhile, the 13th
ordinary month in the preceding half of the plate is Samon, as shown in Fig. 5.
ICA

Sam

Riv

Gia

Edr

3

7

11

Ana

Sim

Can

4

8

12

Ogr

Equ

Sam

1

5

9

13

Dum

Cut

Ele

Dum

2

6

10

Riv

Gia

Edr

Riv

Ana

Sim

Can

Ana

Ogr

Equ

Sam

Ogr

Cut

Ele

Dum

Cut

Equ

Sam

Ogr

Equ

Sam

Ogr

Equ

Ele

Dum

Cut

Ele

Dum

Cut

Ele

Gia

Edr

Riv

Gia

Edr

Riv

Gia

Edr

Sim

Can

Ana

Sim

Can

Ana

Sim

Can

ICB

Fig. 5. The 13th month is Samon of Year II.

...and the 13th day of this 13th month contains the notation: +II MD.

In short, the notation at the 13th day of ICB appears to indicate the 13th day of the
13th month on the plate (= Samon 13 of Year II). In the same way, day 14 of ICB
indicates the 14th day of the 14th month (= Duman 14 of Year II); and so on. In
all, the 30 days of ICB indicate the following series of 30 days from the left half
of the plate (Table 5).

16

TABLE 5: Days from the left half of the plate that are indicated by ICB
ICB

Day 1

Indicates:

Year I, Edrin 11

Year I, Anagant 4

Day 14

Year II, Duman 14

Year I, Duman 2

Day 5

Year I, Ogron 5

Day 4

Day 6

Day 7

Day 8
Day 9

Day 10

Indicates:

Day 11

Day 2

Day 3

ICB

Year I, Samon 1
Year I, Rivros 3

Year I, Cutios 6

Year I, Giamon 7

Year I, Simivi 8
Year I, Equos 9

Year I, Elembiu 10

Day 12

Day 13
Day 15

Day 16

Day 17

Day 18
Day 19

Day 20

Year I, Cantlos 12

Year II, Samon 13
Year II, Rivros15

Year II, Anagant 16

Year II, Ogron 17

Year II, Cutios 18

Year II, Giamon 19

Year II, Simivi 20

ICB

Day 21

Day 22

Day 23

Day 24

Day 25

Day 26

Day 27
Day 28

Day 29

Day 30

Indicates:

Year II, Equos 21

Year II, Elembiu 22

Year II, Edrin 23

Year II, Cantlos 24

Year III, Samon 25

Year III, Duman 26

Year III, Rivros 27

Year III, Anagant 28

Year III, Ogron 29

Year III, Cutios 30

The series forms a simple arithmetic progression. The following illustration
shows the pattern that this progression creates on the left half of the plate. ICB is
shown at far right, and the days it indicates are underlined (Fig. 6).

Fig. 6. Days from the left half of the plate that are indicated by ICB.

Meanwhile, ICA, in the upper left corner of the plate, indicates an identical series
of days from the right half of the plate (Table 6).

17

TABLE 6: Days from the right half of the plate that are indicated by ICA
ICA

Day 1

Day 2

Day 3

Day 4

Day 5

Day 6

Day 7
Day 8

Day 9

Day 10

Indicates:

Year III, Giamon 1

Year III, Simivi 2

Year III, Equos 3

Year III, Elembiu 4

Year III, Edrin 5

Year III, Cantlos 6

Year IV, Samon 7

Year IV, Duman 8

Year IV, Rivros 9

Year IV, Anagant 10

ICA

Day 11

Day 12

Day 13

Day 14
Day 15

Day 16

Day 17
Day 18

Day 19

Day 20

Indicates:

Year IV, Ogron 11

Year IV, Cutios 12

Year IV, Giamon 13

Year IV, Simivi 14
Year IV, Equos 15

Year IV, Elembiu 16

ICA

Day 21

Day 22

Day 23

Day 24

Day 25

Day 26

Indicates:

Year V, Rivros 21

Year V, Anagant 22

Year V, Ogron 23

Year V, Cutios 24

Year V, Giamon 25

Year V, Simivi 26

Year IV, Edrin 17

Day 27

Year V, Equos 27

Year V, Duman 20

Day 30

Year V, ‘Cantlos 30’

Year IV, Cantlos 18

Year V, Samon 19

Day 28

Day 29

Year V, Elembiu 28

Year V, Edrin 29

Together, both series form this pattern on the plate (Fig. 7).

Fig. 7. Days indicated by both intercalary months.

As Duval and Pinault note, ‘chaque jour intercalaire, en effet, reçoit la notation
du jour correspondant d’un mois ordinaire précédent, dans l’ordre régulier des
mois’ (1986, 305). Notice that the plate’s 62 explicit lunar months are in this way
divided into 60 implicit solar months, representing the five solar years encompassed by the plate, as shown in Fig. 8.

We do not know what names the Gauls may have given these solar months; but
in other Western lunisolar calendars, they normally bear the names of the lunar
months in which they begin (McCarthy 1993, 207), and so this is the practice followed in this essay. Figure 3 showed the names and locations of the lunar months
on the plate; and Fig. 9 shows the names and locations of the corresponding solar
months.

18

Fig. 8. Boxes indicate the 62 explicit lunar months; shading indicates the 60
implicit solar months.

Fig. 9. In this essay, each solar month is called after the lunar month in which it
begins.

Following ICA in the upper-left corner, solar Samon (shown by shading) begins
on the first day of lunar Samon (shown by a black-bordered rectangle). But the
solar months, being longer, gradually begin later and later within their own lunar
months, until at bottom centre, solar Cutios begins on the last day of lunar Cutios;
and at this point, ICB intervenes to accommodate the remainder of solar Cutios and
thereby restore the pattern.
ICA is used in the same way at the end of the right half, such that with a single
exception, the two intercalary months successfully ensure that each solar month
always begins within its own lunar month. The exception occurs at the end of the
plate: solar Edrin begins on the 29th day of lunar Edrin; but solar Cantlos cannot
then begin on ‘Cantlos 30’, as that lunar month contains only 29 days. Instead,

19

this final solar Cantlos begins on Day 1 of ICA, in the upper-left corner of the
plate; and as a result, the final instance of solar Cantlos becomes the only solar
month not to begin within its own lunar month.
But this exception is necessary. Because the beginnings of the solar months
progress by one day per lunar month, it follows that each solar month is one day
longer than its own lunar month, as shown in Fig. 10.

Fig. 10. Each solar month is one day longer than its own lunar month.

The ‘exception’, then, has in fact ensured the complete regularity of this pattern:
the final solar Edrin and Cantlos, at lower-right and upper-left of Fig. 10, contain
31 and 30 days respectively, as they do in all other years.

Comparison of Gaulish and Roman solar months
Notice that the lengths of these Gaulish solar months are nearly identical to the
lengths of the Roman months, if solar Samon is aligned with March, as shown in
Table 7.

No solar month on the plate is 28 or 29 days long, and so the inequality between
February and Cantlos is to be expected – although here, this disjunction occurs at
the end of the year, where it would cause minimal disturbance. Aside from this,
with solar Equos short, only solar Elembiu differs from its Roman counterpart,
being one day shorter than December.
But whenever solar Equos is long, its 31st day compensates for this shortness of
Elembiu, as shown in Table 8.

20

TABLE 7: Comparison of Gaulish and Roman solar months when solar Samon =
March (and Equos is short)
Gaulish

Length

Length

Rivros

31

31

Samon

Duman

Anagant
Ogron

Cutios

Giamon
Simivi
Equos

Elembiu

Edrin

Cantlos

31

30

30

31

31

30

31

30
30

31
30

31

30

30

31

31

30

31

30
31

31

28 or 29

Roman
March
April
May
June
July

August

September
October

November
December

January

February

TABLE 8: Comparison of Gaulish and Roman solar months when Samon = March
(and Equos is long)
Gaulish

Length

Length

Rivros

31

31

Samon

Duman

Anagant
Ogron

Cutios

Giamon
Simivi

Equos 1-30
Equos 31

Elembiu
Edrin

Cantlos

31

30

30

31

31

30

31

30
1

30

31

30

31

30

30

31

31

30

31

30
1

30
31

28 or 29

Roman
March
April
May
June
July

August

September
October

November

December 1

December 2-31

January

February

Table 8 shows that when Equos is long, the correspondence between Gaulish and
Roman month-lengths is complete but for a single inequality at the end of each calendar’s variable month – the Gaulish Equos, and the Roman February.

It has always been thought that ‘except for the use of Roman characters and
numeration, the Coligny Calendar shows no trace of Roman influence’ (MacNeill
1926, 3); but Table 8 seems to indicate otherwise. Rome’s irregular series of
month-lengths is distinctively Roman, the product of the Julian calendar reform
(Samuel 1972, 155); and so such a close correspondence seems unlikely not to be

21

the result of Roman influence. Meanwhile, Rome’s practice of permitting cities
and provinces to retain their native calendars after standardising them to Rome’s
is well-attested enough in the various surviving hemerologia – tables of correspondences between Roman and provincial months, prepared as aids for Roman
administrators (ibid., 171-8). The chief manuscripts are medieval, but fragments
survive from as early as the fourth century, and they are in remarkable agreement
with the later copies (ibid., 173 n.3).
Many of the calendars in these hemerologia contain native months whose lengths
have been made to parallel the lengths of the corresponding Roman months (ibid.,
173). The implication is that the calendar on the Coligny artifact represents just
such a compromise – which would only make it easier to understand how this calendar could have been in use at such a late date. Also: with solar Samon (the first
month of the Gaulish year) as the counterpart to mensis Martius, the presence of
the statue of Mars also becomes easier to understand: Mars would then have been
the god of the head of the year.
A PRECEDENT FOR ANATOLIUS’ LUNAR YEAR

Earlier, it was shown how the Irish verse delineates the lunar year used in the calendars of the early Insular church, of Sulpicius, and of Anatolius. In the verse, the
lunar year is counted from March, and consists of months a day shorter than the
corresponding solar months.
Meanwhile, Figure 10 shows that the calendar engraved on the Coligny artifact
also consists of lunar months a day shorter than their solar counterparts; and Table
8 shows that the lengths of these solar months parallel the lengths of the Roman
months when Samon is taken as March. The implication, then, is that the Coligny
calendar contains a precedent for Anatolius’ distinctive lunar year, and therefore
could be the ultimate source from which the lunar calendar of the early Insular
Church was derived.
This implication can be tested quite simply by means of an entirely artificial exercise: if the Coligny plate and the calendar of the early Insular Church do in fact
contain the same distinctive lunar year, then it ought to be possible to use the Irish
verse introduced earlier in this essay to convert Roman dates to Gaulish ones, simply by taking ‘the age of the moon’ to mean ‘the day of the Gaulish lunar month’.
The word ‘month’ is already used in the verse to mean ‘Roman solar month’; so
to keep things clear, let us employ the Gaulish word mid ‘lunar month’ to mean
‘Gaulish lunar month’:
From the start of March, count in months how far

It is to the month in which you are.

Add the day of the month less one, and then

Add the day of the mid when March began.
If greater than thirty, take thirty away;

The result is the day of the mid today.

22

If we repeat the earlier exercise using this version of the verse, we again obtain
Table 3, but in this form:
TABLE 9: The day of the mid on the first day of each Roman month, if Samon
begins on March 1
Roman date
01-Mar

01-Apr

01-May
01-Jun
01-Jul

01-Aug

01-Sep
01-Oct

Roman months since March 1
0

1
2
3
4
5

6
7

Day of mid
Samon 1

Duman 2
Rivros 3

Anagant 4
Ogron 5

Cutios 6

Giamon 7
Simivi 8

Of course, we’ve seen this series of days before; it is the series of solar-month
boundaries indicated by the notation in ICB, as shown earlier in Table 5.
In other words: when solar Samon is equated to March, both lunar years – the one
on the Gaulish plate, and the one in the Irish verse – contain the same distinctive
sequence of lunar months, and therefore create the same simple pattern in operation against the Roman calendar. This demonstrates that the Gaulish artifact,
which predates Anatolius by about a century, does contain a precedent for his distinctive lunar year.

The possibility of direct influence
The existence of this precedent indicates only that this calendar could have been
Anatolius’ source – though there does seem to have been ample scope for direct
influence. Anatolius was a well-travelled, erudite scholar with a keen interest in
chronology; and aside from such Gaulish influence as he may have met with on
his travels, his bishopric was situated on the main route connecting Egypt and
Judaea to nearby Galatia, an area with Gaulish cultural roots. Importantly, he
would have had to hear but a single fact about this Gaulish calendar – that its lunar
year had consisted of months a day shorter than the corresponding Roman months
– for the simple pattern in Table 3 to have become apparent to him.
Then, if Anatolius had found this pattern attractive, its pagan origins would hardly have deterred him from using it. In the third century, there was as yet no calendar called ‘Christian’: all calendars were in effect pagan; various of these pagan
calendars were in use by Christians; and from the elements of these pagan calendars, Christian chronologers created competing systems for determining the date
of Easter. The 112-year Easter cycle of Hippolytus used by Christians in Rome in
the early third century was based on the classical Greek ogdoad (Neugebauer
1975, 620-1; Samuel 1972, 35-42), whose symmetry Hippolytus found attractive;
the 19-year cycle of Meton (Neugebauer 1975, 622-4), preferred by Alexandrian

23

Christians, had originated in Seleucid Babylon (ibid., 354-7); and Easter cycles
everywhere employed the Roman solar calendar, with its origins in the Julian
reform (Samuel 1972, 155). If Anatolius had borrowed this Gaulish lunar year for
its symmetry against the Roman calendar, and pressed it into service in the creation of his Easter cycle, he would only have been following the common practice of Christian chronologers at that time.
ETYMOLOGIES OF MONTH-NAMES

Having presented this hypothesis that solar Samon could coincide with March,
one task remains in this essay: to revisit the etymologies of the month-names in
this context.
Ordinarily, it might not seem germaine to deal with etymology in a study such as
this; but one cannot escape the fact that these twelve etymologies feature prominently in the literature on this calendar, and that any new theory as to the location
of Samon will at once be scrutinised against them. The month-name Ogron, for
example, is almost exclusively described in the literature as having come from the
IE root meaning ‘cold’; yet in this study, we have equated it with July. Scholars
of the calendar deserve an explanation.
Providing an explanation, however, means facing a difficult problem: Gaulish is
still so imperfectly known that only a few month-names have derivations beyond
reasonable doubt. The meanings of most month-names remain obscure; and with
no meaning to guide the philologist, some month-names could credibly be
derived from a dozen different IE roots.
As a result, competing etymologies have emerged in support of various locations
for Samon. Past studies of the calendar, for example, have suggested equating
Samon (which means ‘summer’) to June (Rhys 1905, 73; Rhys/Fotheringham
1910, 210), October (Duval/Pinault 1986, 403), and even December (Olmsted
1992, 89-90); yet most studies present credible month-name etymologies to support their theories. All these etymologies, though they support different theories,
are the product of considerable experience in Celtic philology, and cannot simply
be dismissed.
Faced with numerous possible derivations for some month-names, and competing etymologies for others, it seemed that perhaps a new approach might be in
order. To this end, all twelve etymologies were instead produced by means of an
experiment whose results occupy the remainder of the essay.

An experiment in semantics
This experiment is based upon the strong dualism evident in the calendar, not
least of all in the structure of the plate itself. This dualism is also present in those
month-names with the best-known etymologies: Samon and Giamon, for example, are derived from the roots for ‘summer’ and ‘winter’, and Ogron and Edrin
from the roots for ‘cold’ and ‘to burn, to glow’. In both cases, the pairs are six
months apart; and so the other eight month-names are also suspected of forming

24

pairs of this sort, month-names whose meanings are in some way complementary.
It has been mentioned that some month-names with obscure meanings could
credibly be derived from many roots. But if the twelve etymologies are arrived at
in pairs so as to preserve some notion of complementarity between opposite
month-names, then the number of credible derivations ought greatly to be
reduced; and the application of this additional constraint might then prove a sieve
through which something interesting could fall.

The parameters of the experiment
This experiment was executed in three stages:
1. All month-name etymologies in the literature were reassessed and catalogued.
2. Through a close study of the corpus of known IE roots, as many additional derivations as possible were posited for each month-name. (In this process, IE
roots with no attested words in Celtic languages were excluded from consideration.)
3. The resulting long lists of derivations were studied in opposite pairs – Samon
with Giamon, Duman with Simivisonn, and so on. Newly-posited derivations
in one list with no clear semantic complement in the other were eliminated,
and the lists were gradually shortened in this way. (Etymologies already in the
literature, however, were retained even if they appeared to fail this test.)

At the end of this process, the lists consisted largely of etymologies already in the
literature, as well as a small number of newly-posited derivations not eliminated
in Step 3. At this point, each of the six pairs of lists was found to contain a single
pair of semantic complements that stood out as particularly apt; and these are the
six pairs of derivations presented in the following sections.

Results of the experiment
Because the published etymologies were treated so conservatively, eight of the
twelve month-names have retained previously-suggested derivations, despite the
novel approach – though the old derivations for Ogron and Edrin emerge with new
shades of meaning. Of the remaining four derivations, those of Rivros and
Simivisonn appear in part in the literature; only those of Cutios and Equos are
entirely new.
Each of the sections below is headed by the canonical form of the month name as
proposed by Duval (1964-5, 16), followed by a statement of the proposed modern equivalents for both the solar and lunar versions of the month. The equivalents for the solar months were given before, in Table 8; while the locations of the
lunar months are arrived at in a straightforward fashion, as can be illustrated by
taking the case of Samon.
Table 8 equates solar Samon to March. We have already seen that the Insular calendar contains a lunar year like the one on the plate; and so lunar Samon will
largely correspond to the Insular calendar’s lunar March – that is, the lunar month
in progress on March 1. This means that March 1 could coincide with any of the

25

30 days of lunar Samon: if it coincides with the last day, then lunar Samon will
have largely corresponded to February, while if it coincides with the first day, then
lunar Samon will largely coincide with March. The range of locations for lunar
Samon is therefore given as ‘February/March’– and so for the other lunar months
as well: the range of any particular lunar month is always roughly equivalent to
its corresponding solar month plus the preceding solar month.

SamonEquivalents. Lunar month = February/March; solar month = March.
Derivation. IE *sem- ‘summer’1 > *semo- > Old Irish sam-, Gaulish month-name
Samon- ‘SUMMER’.
Comments. This month-name indicates the start of the summer half of the calendar year, perhaps as marked by the first signs of spring.

DumannEquivalents. Lunar month = March/April; solar month = April.
Derivation. IE *dheu- ‘to steam, to smoke, to blow about, to give off (vapour or
odour)’, and the like, as well as ‘to be in violent, surging motion, also in emotional and psychological senses’2 > IE *dheum- ‘cloud, mist, vapour, smoke’ >
*dheumo-, *dheumә-: Latin fūmāre ‘to give off smoke, vapour or odour’
(*dheumә-mi); Middle Irish dumacha ‘cloud’; Greek θύμος ‘pertaining to soul,
spirit, mind, temper, passion’ = ‘thymian’, Greek θῡμός ‘soul, life; spirit, passion,
etc.’ (*dheumo-s); Gaulish monthname *Dūmannos (*dheumә-snos), with a substantivising suffix, so meaning something like ‘THYMIAN-NESS’, genitive
*Dūmanni, in accordance with the plate’s nominative DVMAN-, genitive DVMANNI, variant DVMAN[N]I.
Comments. This month-name could be a reference to the reawakening of nature
– that is, to the return of life or ‘spirit’ to the natural world.

Rivros
Equivalents. Lunar month = April/May; solar month = May.
Derivation. IE *per- ‘over, before, first’3 > dative of direction *prei-, *pri-: *prei> Old Latin prῑ ‘prae- = fore-’; *pri > CC (Common Celtic) *[p]ri- > Gaulish re‘fore-’, as in re-bellias ‘fore-beautiful’, not superlative but intensive, in the sense
of ‘(a) most beautiful ...’ (Lambert 2002, 159).
Then, IE *per- ‘to sell’4 > reduplicated form *pepor- ‘bestow, grant’ > CC
*[p]e[p]or- ‘consecrate, dedicate’ > early Gaulish *eior- ‘dicāre = consecrate, dedicate’ (> Gallo-Greek ειωρ-ου > Gallo-Roman IEVR-V /ieur-u/, Old Irish (ro)-ír

Pokorny 1959, 905; Thurneysen 1899, 532; Loth 1904, 130; Rhys 1905, 73; MacNeill
1926, 41; Lainè-Kerjean 1943, 258; Duval 1964-5, 9; Duval/Pinault 1986, 426.
2
Pokorny 1959, 261-7; Duval/Pinault 1986, 423.
3
Pokorny 1959, 810-816; Thurneysen 1899, 533; MacNeill 1926, 41.
4
Pokorny 1959, 817.
1

26

‘dicavit = he consecrated, he dedicated’ (Lejeune 1968-9, 36-7; Lambert 1979,
208-9).
Finally, in combination: early Gaulish re- ‘fore-’ + *eior- ‘dicāre = consecrate,
dedicate’ = *re-eior- > *rēior- literally ‘fore-consecrate, fore-dedicate’ (cf. Latin
prae-dicāre ‘praise’) > Gallo-Greek *ρειωρ > Gallo-Roman *rēor- > *rēurthrough assimilation to the inventory of licensed diphthongs in Gaulish > late
Gallo-Roman *rīur-,5 hence month-name Rīuros ‘FORE-DEDICATION’, perhaps
‘PRAISE’, genitive Rīuri, in accordance with the plate’s nominative RIVROS, genitive RIVRI.
Comments. May 1 will almost always fall within lunar Rivros. This month-name
could be a reference to the celebrations of the return of greenery that were held
not just in Celtic countries but throughout the Continent at this time of year
(Hutton 1991, 272).

Anagantio(?-)
Equivalents. Lunar month = May/June; solar month = June.
Derivation. IE negative particle *nĕ6 > privative prefix *ṇ > Gaulish, Old Irish,
Welsh, Cornish, Breton an- ‘no-, not-’ , as in Gaulish an-mat ‘not-good’.
Then, IE *aĝ- ‘to drive, to go with swinging arms’7 > *ag-ōn ‘a driving, a going
with arms swinging’ (> Greek ἀγω
́ ν ‘effort, exertion, struggle’, CC *ag-ān), genitive *ag-nes ‘of driving, of going with arms swinging’ (> Gallo-Roman *agnis
‘a stride’, Old Irish áin ‘a bustle, a game, a match’); CC *ag-ān- > present participle *agān-tios > Gaulish *agāntios ‘driving, striding, exerting, struggling
(person)’, plural *agāntiōs, genitive plural *agāntiōm.
Finally, in combination: Gaulish an- ‘not’ + *agāntiōs (plural) ‘driving, striding,
exerting, struggling (people)’, genitive plural *agāntiōm > month-name
*Anagāntiōs (plural) ‘no-driving, no-striding, no-exerting, NO-STRUGGLING (people)’, genitive (plural) *Anagāntiōm, in accordance with the plate’s nominative
ANAGANTIO-, genitive ANAGANTIO-, with the variant ANAG[AN]TIOS as an abbreviated nominative, as Duval suggested it might be (Duval 1964-5, 10).
Comments. Hesiod’s Works and Days, though it is a Greek document, contains a
statement that illustrates the gloss: ‘But when the artichoke is in flower [in June]...
one might have the shadow under the rock, and the wine of Biblis, a curd cake,
and all the milk that the goats can give you, the meat of a heifer... and of baby
kids also. Then, too, one can sit in the shadow and drink the bright-shining wine,
his heart satiated with eating...’ (Lattimore 1959, 87-9).

OgronEquivalents. Lunar month = June/July; solar month = July.

Lejeune 1968-9, 36-8; Lambert 1979, 209.
Pokorny 1959, 756-8; Duval 1964-5, 10; Duval/Pinault 1986, 422.
7
Pokorny 1959, 4-6, 756-8; Rhys 1910, 220 & n.1; MacNeill 1926, 34; Duval 1964-5, 10;
Duval/Pinault 1986, 422.
5
6

27

Derivation. IE *oug- ‘cold’8 > *ougro- ‘cold’ > Old Irish úar, Irish, Gaelic fuar,
Welsh oer, Cornish oir ‘cold, chilly’; past participle *ougrәnós, *ougrónos
‘chilled (thing)’ > Old Irish úarán ‘well, cold-spring’ (*ougrәnós), Gaulish
month-name *Ōgrónos (*ougrónos) ‘WELL, COLD-SPRING’, genitive *Ōgróni,
dative *Ōgrónu, in accordance with the plate’s nominative OGRON-, genitive
OGRONI, with the variant OGRONV as a possible dative.
Comments. Perhaps this was a month associated with wells and springs.
Extensive deposits of artifacts have been found at such sites throughout Celtic
Europe, indicating that they were ritually visited; and such visits must necessarily have involved travel, often to higher altitudes, where springs tend to originate.
July seems a practical time of year to do this, while the weather is warm, and the
crops are still ripening.

Cutios
Equivalents. Lunar month = July/August; solar month = August.
Derivation. IE *keu- ‘to hear, observe, notice, pay attention’9 > Greek κοέω ‘to
hear, to mark, to note’; *kowo-s ‘audient, observant’ > *kowo-kantos ‘fully audient, quite observant’ > Old Breton (int) cou-cant, Middle Welsh (yn) geu-gant
‘very experienced’ (ibid., 526-7), Cornish cow-gans ‘certain’; *keu-tā ‘(something having the quality of) audience, observance’, hence perhaps ‘(an) audience,
(a) hearing’, like *teu-tā ‘(something having the quality of) swollenness,
strength’, hence ‘(a) folk, (a) land’ (ibid., 1080-5); Gallo-Greek *κοουτιος >
Gallo-Roman *coutios, later *cūtios, Gaulish personal-name Cūtius ‘audient,
observant (person)’ like Gallo-Greek τοουτιος > Gallo-Roman toutios > later
*tūtios, Gaulish personal-name Tūtius ‘folk (person)’, hence ‘tribesman, (fellow-)citizen’ (ibid.; Lejeune 45-6); late Gaulish *cūtios ‘audient, observant (person)’, plural *cūtiōs, genitive plural *cūtiōm > month-name *Cūtiōs (plural)
‘audient, observant (people)’, hence ‘(an) AUDIENCE, (a group of) hearers’, genitive (plural) *Cūtiōm, in accordance with the plate’s nominative CVTIOS, genitive
CVTIO-; the variant QVTIO- then represents an attempt to Latinise the syllable ku-.
Comments. August 1, associated with the god Lugh, will nearly always fall within lunar Cutios. The month-name could be a reference to a mass-gathering that
took place at this time of year.

GiamoniEquivalents. Lunar month = August/September; solar month = September.
Derivation. IE *ĝhei- ‘winter’10 > *ghi-iōm-, *ĝhi-iemo- > Middle Irish gem-, Old
Welsh gaem, Gaulish month-name Giamoni- ‘WINTER’.

8
Pokorny 1959, 783; Thurneysen 1899, 534; Loth 1904, 130; Rhys 1905, 73; MacNeill
1926, 41; Lainè-Kerjean 1943, 259; Duval 1964-5, 11; Duval/Pinault 1986, 425.
9
Pokorny 1959, 587-8.
10
Pokorny 1959, 425-6; Thurneysen 1899, 534; Loth 1904, 130; Rhys 1905, 73; Rhys
1910, 218 n.1; MacNeill 1926, 38; Duval 1964-5, 12; Duval/Pinault 1986, 424.

28

Comments. This month-name indicates the start of the winter half of the calendar
year, perhaps as marked by the first signs of autumn.

SimiuisonnaEquivalents. Lunar month = September/October; solar month = October.
Derivation. IE *sem- ‘one; together as one; with, together’11 + IE *ei- ‘to go’12 >
IE *sem-ei- ‘go together’, 3rd-sing. *sem-ei-ti ‘it goes together’ > pre-Latin
*sem-eitus ‘(a) going-together’ > Latin adv. simītu ‘together; simultaneously’;
while the same *sem-ei-ti ‘it goes together’ > CC *sem-i-ti- > Old Irish adv. emi-th ‘tanquam, quasi- = like, just like, just as if, as it were, a sort of, almost as
much as, nearly’, Welsh adv. hef-y-d ‘also, as well’; *sem-ei- ‘go together’ > CC
*sem-i- ‘just like, as it were, etc.’ > Gaulish *sem-i-, with a later variant *sim-isimply as a result of Latin influence (cf. Latin sim-ilis, sim-ītu, sim-ul, etc.).
Then, IE *weis- ‘to flow (out or away)’13, attested in Brythonic nouns and rivernames: for example, CC *weis-aro- > Welsh gwy-ar ‘blood’, literally ‘flowingout (stuff)’, Gaulish river-name Vis-era > French Vesère, literally ‘flowing-away
(thing)’; CC river-name *Weis-ā > Welsh Gwy ‘Wye’; Old British river-name
*Wis-uriā > English Wear; Gaulish river-name Vis-uvia > French la Vezouse; and
so on (ibid.).
Then, IE *pen- ‘water’14 > Middle Irish en ‘water’, on-chū ‘otter’ (literally ‘waterhound’), river-name On, Gaulish river-name *Ona, as well as the suffix -onnā in
Gaulish river-names, such as Bebr-onnā ‘beaver-stream’, literally ‘beaver-water’
(ibid., 136-7).
Finally, in combination: Gaulish *sem-i-, later *sim-i- ‘just like, as it were, etc.’
+ vis- ‘flowing(-away)’ + -onnā ‘water’ = *semivisonnā, later *simivisonnā ‘LIKE
FLOWING(-AWAY) WATER’, ‘flowing(-away) water, as it were’.
Comments. The gloss suggests a taboo-formation of a type common in IE languages: a metaphor for something whose own name is being avoided. The month
is October; perhaps the underlying notion is one of ‘dying’ or ‘dying away’, as in
the departure of life from nature, or the flowing of ‘spirit’ out of the world; and
in that sense, it stands in contrast to THYMIAN-NESS, the gloss for the opposite
month.

Equos
Equivalents. Lunar month = October/November; solar month = November.
Most past scholars have chosen ‘horse’ as the gloss for this month-name; but
many have done so rather reluctantly. Some (e.g. Duval/Pinault 1986, 424) note
that a derivation from IE *[p]eku- ‘cattle’ is also possible, though it would pose
Pokorny 1959, 902-5; Rhys 1910, 257 n.1.
Pokorny 1959, 293-7.
13
ibid., 1134.
14
ibid., 807-8.
11

12

29

some phonetic difficulties; while others (e.g. Duval 1964-5, 14) state simply that
the gloss might be ‘horse’, but that Equos might also be derived from IE *[p]eku‘cattle’, or even be a cognate of Latin aequus ‘equal’.
The reason for this reluctance is that Gaulish is a member of the branch of Celtic
languages in which IE kw was inherited as p. As a result, the Gaulish word for
‘horse’ is epos, not ekwos (Pokorny 1959, 301-2). Tellingly, the Gaulish horsegoddess was Epona, a well-attested name that appears nowhere as *Equona;
while the month-name Equos appears nowhere on the calendar as *Epos. In fact,
no form of equos is attested in any other Gaulish inscription.
As a result, I share this reluctance to gloss Equos as ‘horse’, particularly in light
of the ubiquity of the form Epona. A more straightforward explanation, it seems
to me, is that the qu of Equos, like the qu of Qutios (a variant spelling of Cutios),
represents an attempt to Latinise the spelling of the syllable cu, and that Equos is
essentially a three-syllable word: E-cu-os.
Derivation. IE *epi ‘at, by’15 > CC *e[p]i- = *ei- > Gaulish *ē-16, Old Irish éi-, é-,
as in Old Irish éi-thech, é-thech ‘perjury, forswearing’, literally ‘around-swearing’ (cf. Greek ἐπι-ορκέω ‘I perjure’); or as in *epi-bhelo- ‘over-burning (thing),
on-burning (thing)’ > Irish éi-bheall ‘ember’; or as in *epi-weriō ‘protected-round
(place)’, hence ‘bounded land, hill, island’ > Old Irish É-riu ‘Ireland’, Old Welsh
Ī-werδon ‘Ireland’ (Pokorny 1959, 1160-2).
Then, IE *keu- ‘to hear, observe, notice, pay attention’17 > early Gaulish *cou- >
late Gaulish cū-, as described in the section on Cutios.
Finally, in combination: IE *epi-keu- > CC *e[p]i-kou- (cf. Greek ἐπ-ἀκούω
‘heed, obey’) > early Gaulish *ē-cou- > late Gaulish *ē-cū-, hence month-name
*Ēcūos, written Ēqūos, perhaps ‘OBEDIENCE’, genitive *Ēcūi, written Ēqūi, in
accordance with the plate’s nominative EQVOS, genitive EQVI; the QV- then represents an attempt to Latinise ku-, as was the case with QVTIO- for CVTIO-.
Comments. FORE-DEDICATION, the gloss for the opposite month, is an active notion,
in that praise or offerings are bestowed; whereas OBEDIENCE is a passive notion, in
that judgments or laws are received. It is worth noting that November 1 will nearly always fall within lunar Equos; and at least in Ireland, November was the month
in which tribal assemblies were held, and laws were made (Hutton 1991, 177).

ElembiuEquivalents. Lunar month = November/December; solar month = December.
Derivation. IE *el- ‘red, brown’18 > *el-en- ‘deer’ > Welsh elain ‘hind’ (*elәnī),
perhaps Middle Irish ell ‘herd’ (*elnā).
ibid., 323-5.
Lejeune 1968-9, 36 n.30.
17
ibid., 587-8.
18
Pokorny 1959, 302-4; Thurneysen 1899, 535; Rhys 1905, 101; Lainè-Kerjean 1943, 260;
Duval/Pinault 1986, 424.
15
16

30

Then, IE *bhei(ә)- ‘to hit, strike’19 > *bio- ‘(a) strike, striker’, as in *vailu-bios
‘wolf-striker’ > Middle Irish personal-name Faíl-be ‘Wolf-slayer’; or as in *vidubion ‘wood-strike’ > Gallo-Roman vidubium ‘wood-axe’, Old Welsh uiidimm,
Welsh gwyddyf ‘pruning-knife’, Middle Irish fidba ‘bill-hook’.
Finally, in combination: IE *elen-bion > Gallo-Roman *elembium, perhaps ‘DEERSTRIKE’ or ‘HERD-STRIKE’, hence month-name *Elembium, genitive *Elembī, in
accordance with the plate’s nominative ELEMBIV-, genitive ELEMBI.
Comments. The struggle and the hunt for food implicit in DEER-STRIKE stand in
contrast to the ease and plenty implied by NO-STRUGGLING, the gloss for the opposite month. At its earliest, lunar Elembiu- coincides with November, which in
Welsh is Tachwedd ‘Slaughter’, as this is the month in which herd-animals unfit
for the coming winter were slaughtered for meat. Also, this is an excellent time
of year for hunting elk and deer, after the autumn rutting-season is complete, but
before deep winter, while the animals are still fat from summer feeding. A light
snow-cover is actually a benefit, as it assists in tracking; while the lower temperatures lengthen the time during which the meat can be butchered and preserved
safely.

(A)edrini-s(?-)
Equivalents. Lunar month = December/January; solar month = January.
Derivation. IE *aidh- ‘to burn, to glow’20 > u-stem *aidhu- > Gaulish tribe-name
Ædui, Old Irish áed (< *aidhus) ‘fire’; r-stem *aidh-(e)r ‘glowing thing’ = ‘(clear
daytime) sky, upper sky, upper air’ > Greek αἰθήρ ‘clear sky, upper air’ = ‘ether’;
genitive *aidhr-es ‘of the ether’, oblique stem *aidhr- ‘ethereal’ > n-class noun
*aidhr-ōn ‘ethereal (thing)’ = ‘brightness of the sky; etherealness; upper sky;
uppermost heaven’, genitive *aidhr-enes > Gaulish month-name Ædrinis
‘(month of) brightness of the sky; etherealness; upper SKY; uppermost heaven’,
genitive Ædrinis; later spelling Ēdrinis, genitive Ēdrinis, in accordance with the
plate’s nominative EDRINI-, genitive EDRINI-; the variant ÆDRINI- then displays the
earlier spelling.
Comments. The warmth, glow and height of the upper SKY stand in contrast to the
coolness, darkness and depth of the WELL, the proposed gloss for the opposite
month.
Interestingly, the Greek adjective αἰθρη-γενής ‘ether-born; clearing up, chilling’
indicates an association between ‘ethereal skies’, clearness, and coldness that will
be familiar to anyone who has experienced a cold winter. When temperatures are
very low, the coldest days are the clearest ones, when there is no cloud-cover to
hold in the sun’s warmth. With moisture frozen from the atmosphere, such a sky
is very clear and blue, unlike the softer blue of a clear sky in warm weather, when
atmospheric humidity helps diffuse the sunlight; and skywatchers know that cold,
Pokorny 1959, 117-8; Thurneysen 1899, 535.
Pokorny 1959, 11-12; Rhys 1905, 101; Lainè-Kerjean 1943, 259; Duval 1964-5, 14;
Duval/Pinault 1986, 424.
19
20

31

clear winter nights provide the best skies for viewing. In temperate climates, then,
‘ethereal skies’ are a feature of cold months rather than warm ones; and solar
Ædrinis is here equated to January.
Cantlos
Equivalents. Lunar month = January/February; solar month = February.
Derivation. IE *kan- ‘to sing, to sound’21 > *kan-tlo- > Middle Irish cētal, Welsh
cathl ‘song’, Breton kentel ‘(chanted or sung) lesson’, Gaulish *cantlos ‘SONG’ or
perhaps ‘CHANT’ or ‘CANTICLE’, hence month-name *Cantlos, genitive *Cantli, in
accordance with the plate’s nominative CANTLOS, genitive CANTLI.
Comments. The sound and activity of SONG stand in contrast to the silence and
passivity of AUDIENCE, the gloss for the opposite month. It is worth noting that
February 1, associated with the goddess Brigid, will almost always fall within
lunar Cantlos.
Proposed glosses of month-names
At the end of this experiment in semantics, we are left with the following array of
glosses shown in Table 10.
These results depend heavily on the existing literature; yet all six pairs now form
semantic complements like Samon and Giamon, with glosses appropriate to a
year beginning in March.

CONCLUSIONS
If solar Samon is taken as March, then the Coligny calendar contains a distinctive
lunar year attested elsewhere only in the Insular calendar devised by Sulpicius of
Gaul in about 403, and in a mid-third-century work by Anatolius of Laodicea. The
Coligny artifact predates Anatolius by a century or less, such that the calendar
engraved upon it constitutes a precedent for the lunar year he describes in his work.
A set of etymologies can be derived in which opposite month-names have complementary meanings, after the model of Samon/Giamon ‘summer/winter’; and
the resulting glosses also seem apt for a solar year beginning in March. At the
same time, most of the month-names remain derived from roots that have already
been suggested in the literature.
Finally, the close correspondence shown in Tables 7 and 8 implies something that
Fotheringham suspected nearly a century ago: that ‘the Coligny calendar is, like
our Easter calendar, a calendar accommodated to the Julian calendar’
(Rhys/Fotheringham 1910, 285), and that the plate expresses a standardised relationship between the two. However, based upon the fact that Samon means ‘sum21
Pokorny 1959, 525-6; Thurneysen 1899, 536; Rhys 1905, 101; Rhys 1910, 259; MacNeill
1926, 35; Lainè-Kerjean 1943, 259; Duval/Pinault 1986, 423.

32

Solar

Aug

Jul

Jun

May

Apr

Mar

month

Lunar

Jul/Aug

Jun/Jul

May/Jun

Apr/May

Mar/Apr

Feb/Mar

month

Proposed glosses

SUMMER

THYMIAN-NESS

(listening, silence, ?reverence)

AUDIENCE, HEARING

(cold, dark, deep)

WELL, COLD-SPRING

(ease, having, thriving)

NO-STRUGGLING

(active, giving, gifts, gods)

FORE-DEDICATION, PRAISE

(infilling, regeneration, being born)

(beginning of summer half of year)

(and related senses)

TABLE 10: Proposed glosses of month-names
Proposed

form

SamonDūmannos

Rīuros

Anagāntiōs
Ōgronos

Cūtiōs

Proposed glosses

WINTER

(and related senses)

LIKE-FLOWING(-AWAY)-WATER

(beginning of winter half of year)

OBEDIENCE, HEED

(outflowing, degeneration, dying away)

DEER-STRIKE, HERD-STRIKE

(passive, receiving, obligations, leaders)
SKY, ETHER

(struggle, hunting, slaughter)

(fiery, glowing, high)

SONG, CHANT, CANTICLE

(singing, sound, ?jubilation)

Lunar

Solar

Proposed

form

Giamoni-

month
Sep

month

Aug/Sep

Simivisonnā
Ēcūos

Elembium

Oct

Nov
Dec

Sep/Oct

Oct/Nov

Nov/Dec

Ædrinis

Cantlos

Jan

Feb

Dec/Jan

Jan/Feb

33

mer’, Rhys placed Samon in June, near the summer solstice (ibid., 210; Rhys
1905, 73); and so it would be worthwhile to reassess Fotheringham’s idea with
Samon as March. Such a reassessment is the subject of the second in this pair of
essays.
BIBLIOGRAPHY

Duval, P.M. 1964-5: Observations sur le calendrier de Coligny IV, Études Celtiques 11, 7-45.
Duval, P.M./G. Pinault, 1986: Recueil des inscriptions Gauloises, Volume III: Les calendriers,
Paris.
Hutton, R. 1991: The Pagan Religions of the Ancient British Isles: Their Nature and Legacy,
Oxford.
Lainè-Kerjean, C. 1943: Le calendrier celtique. Zeitschrift für keltische Philologie 23, 249-284.
Lambert, P.Y. 1979: Gaulois IEVRV: irlandais (ro)-ír ‘dicauit’, Zeitschrift für keltische
Philologie 37, 207-213.
Lambert, P.Y. 2002: Recueil des inscriptions Gauloises, Volume II.2: Textes Gallo-Latins sur
Instrumentum, Paris.
Lattimore, R. 1959: Hesiod: the Works and Days; Theogony; the Shield of Herakles, Ann
Arbor.
Lejeune, M. 1968-9: Note sur les emplois de ΕΙ, Η, Ω dans les textes lapidaires gallo-grecs,
Études Celtiques 12, 36-48.
Loth, J. 1904: L’annèe celtique d’après les textes irlandais, gallois, bretons et le calendrier de
Coligny. Revue Celtique 25, 113-162.
MacNeill, E. 1926: On the Notation and Chronography of the Calendar of Coligny. Eriu 10, 167.
McCarthy, D./Ó Cróinín, D. 1987-8: The ‘Lost’ Irish 84-Year Easter Table Rediscovered,
Peritia 6-7, 227-742.
McCarthy, D. 1993: Easter Principles and a Fifth-Century Lunar Cycle used in the British Isles,
Journal of the History of Astronomy 24, 204-224.
McCarthy, D. 1994: The Origin of the Latercus Paschal Cycle of the Insular Celtic Churches,
Cambrian Medieval Celtic Studies 28, 25-49.
Neugebauer, O. 1975: A History of Ancient Mathematical Astronomy, Berlin.
O’Leary, P. C. 1920: Irish Numerals and How to Use Them, Dublin.
Olmsted, G. 1992: The Gaulish calendar, Bonn.
Pokorny, J. 1959: Indogermanisches Etymologisches Wörterbuch, Bern.
Rhys, J. 1905-6: Celtae and Galli, Proceedings of the British Academy 2, 71-107.
Rhys, J./J.K. Fotheringham, 1909-10: The Coligny Calendar, Proceedings of the British
Academy 4, 207-89.
Samuel, A.E. 1972: Greek and Roman Chronology: Calendars and Years in Classical
Antiquity, Munich.
Thurneysen, R. 1899: Der Kalendar von Coligny, Zeitschrift für keltische Philologie 2, 523-544.

Brent Davis
Centre for Classics and Archaeology
School of Historical Studies
University of Melbourne, Australia
b.davis2@pgrad.unimelb.edu.au

34



Parole chiave correlate