The Roman Calendar – Time for a City and an Empire
The Western calendar tradition is dominated by the Roman antecedents in many respects. This holds true for the names of the months and their length, for the mechanisms of intercalation and last but not least its graphic representation. The very word ‘calendar’ has a Roman ancestry, calendarium denoted the notebook recording loans and settlement dates.
Like most circum-Mediterranean cultures, the Roman calendar that we can trace back to the fifth, maybe sixth century BCE, was a lunisolar calendar. As such, it featured empirical lunar months, starting from around the New Moon. In order to deal with the difficulties of observation and the resulting changing length of the apparent lunation from New Moon to New Moon, a variable period was built into the first phase of the month. The observation was entrusted to a priesthood, the pontiffs (pontifices), probably originally entrusted with the construction of the bridge in voting procedures of the people, that is all the family clans, as a whole – hence their name ‘bridge-makers’. They were close to the political leadership, but not actually political authorities themselves. On the day of observation they estimated the distance to the full moon and ‘called out’ (in Latin kalare, hence kalendae as the name of the first day of every month) the number of days until a day called nonae, that is the ‘ninth’ (Latin nonus/a) day before the Full Moon). Thus, they enabled the stable sequence of three ‘weeks’ of eight days, one from the Nones to the Ides (the day of the Full Moon, Latin idus or archaic eidus) and two from the Ides to the new Kalends. Counting was backwards from these three fixed points of the month. As the Romans counted their days including the starting as well as the ending day, their week of eight days was called ‘a period of nine days (nundinae)’, two days before the Ides was ‘before the third day of the Ides (ante diem tertiam Idus)’. The head days of these weeks, also called nundinae, were market days, but also had the character of a weekly feast day, off school, the barber’s day, better food.
An Etruscan festival list found on a brick at Capua, the so-called tabula Capuana, dating from the fifth century BCE, indicated a system of lunar intervals parallel to the Roman Full and New Moons. Still by the end of the second century BCE the Etruscans seemed to have stuck to the same type of empirical or conventionalised lunar months, fixing them to the length of 29 and 30 days The liber linteus, an Etruscan sacrificial calendar of this period, indicated several dates of a calendar which numbered the dates of the month from 1 to 30, thus most probably still a lunisolar calendar. The Roman calendar was very different by this time.
The lunar months of the early Roman calendar were brought in alignment with the solar year by occasionally adding – or as the Roman would say ‘calling inbetween (inter-calare)’ – a thirteenth month to the year, a mensis intercalaris. We do not hear of any rationalised scheme of intercalations. Presumably, the situation was similar to Greece. Intercalation was done at the whim of local authorities, even when more elaborate schemes were already available.
The names of the Roman months sound familiar. The year opened by a month called (mensis) Ianuarius (‘January’) after the name of the Roman god of openings and thresholds, Janus. Evidently, the winter solstice, just days before, was a point of orientation. Intercalation, if necessary, followed only after the second month of the year called Februarius. If the first of January was ‘New Year’ and a festival throughout the Roman world, one of the most successful exports of the Roman calendar to the Mediterranean world and beyond, February nevertheless was a month of endings, of actual and spiritual cleaning. Hence, the month following February (and perhaps an intercalatory month), March (in Latin Martius after the god Mars, the supposed ancestor of the Romans), was also a sort of start of the year. This is reflected in the names of some later month. Martius was followed by Aprilis (the Romans never achieved consensus whether this was derived from Aphrodite or ‘opening (of spring)’), Maius and Iunius. Throughout the republican period, down to the death of Julius Caesar (44 BCE), the other months of the year were just numbered: Quinctilis, Sextilis, September, October, November, December, starting to count from March. At the beginning of the Imperial period, Caesar was honoured by renaming the ‘fifth month’ into Iulius and his successor Augustus by renaming the ‘sixth’ into Augustus. Several times successive emperors or members of the imperial family were flattered by renaming later months (Germanicus, Domitianus etc.), but such attempts hardly survived the death of the respective emperor, neither at Rome nor in the provinces. In the end, the time-honoured sequence survived even the end of Antiquity and Paganism.
If the names of the months suggest utter stability, this was certainly not true for many technical details below that surface. As in many other cultures, the alignemnt of sun and moon, of months and solar year, and the social and political problems of intercalation for complex and imperialistically growing societies triggered reforms. Rome was unique among the ancient circum-Mediterranean societies, in using writing and graphic representations of the calendar to push ever more radical solutions, thus building something by the end of the first century BCE that basically survived until today.
Late republican authors imagined some sort of fixing already by the early republican period as part of a fundamental codification of law in the middle of the fifth century BCE. Indeed, the use of writing and the radical character of even the first step of the reforms were interdependent, but must be dated down to the end of the fourth century. Evidently, the vision was to have a system that allowed to replace the eight-days-weeks by an uninterrupted, recurrent sequence (like the Jewish week of seven days, certainly unknown to Rome of the time) of such weeks and to have months of stable length that made precise countings and appointments by day possible even for the months, if not years to come. Or vice versa, fixing in writing necessitated to stabilise the number of days throughout the month and enabled to introduce both innovation. Without shifting the surface of the monthly structure, the months were given 31 (January, March, May, Quinctilis, October), 28 (February) or 29 days (all the others), 355 in total. In the longer months, the Ides were fixed to the 15th of the month, in the shorter ones to the 13th. Thus, the terminology of a lunar month and even the rituals of moon observation was kept. But to call out five or seven times (for Nones on the 5th or 7th day of the month) was regulated by the calendar, not by the actual size of the moon, which might already be in its waning phase.
The graphic representation invented must have looked quite similar to calendar overviews popular today, featuring thirteen instead of twelve months, however. The listing of the days was organized by the recurring letters A to H (for a continuous week of eight days) at the beginning of every entry. None of these letters were tied to the market day, but used to trace the latter by changing the letter at the turn of the year (or after intercalation) in order to keep up the recurrent length of eight days. The main information concerned the public character of the day, especially indicating its juridical status, i.e., whether it was fas (“right”) to perform certain important legal steps or not (nefas). As a consequence the graphic representation was called fasti, a name replaced by calendarium only after the ancient period. The comitia, the Roman legislative body of the people, could not meet on such dies nefasti . and had also to avoid some of the dies fasti. Shortly after the first reform, a further category of dies comitiales was included and displayed in the calendar. Religious traditions, too, were integrated: feriae, a special class of days given to the gods as property (and hence free from every mundane activity) were marked out as ‘dies nefasti whose violation made piacular sacrifices necessary’ (marked by the letters N(efas) P(iaculum) and abbreviations of the festival names). Once established, more and more information was accrued and added in much smaller letters, in order to not disturb the readability of the calendar as a whole. By the early second century the anniversaries of urban temple foundations were marked by indicating the divine owner and the location. From the end of the first century BCE the graphs were flooded by holidays for imperial victories, birthdays and the like, typically subject to revision after the death of an emperor and the more so after the end of a dynasty.
How could intercalation be made regular on the basis of such stable months? The additional month of 27 days was not fully added, but replaced the last five or four days of February in turn every second year. This resulted in a net intercalation of 22.5 days every second year, adding up to an average of 366.25 days. Evidently, the reception of Greek mathematics and astronomy was faulty, but that was not the main problem in the centuries to come. Intercalation, even when moved from political authorities to the hand of the pontiffs, remained arbitary. In principle, however, the Roman calendar had turned into a purely solar one, not any longer tracking actual lunations by its civic, standardised months.
In fact, the handling of the calendar proved to be a catastrophe. Calendar dates moved to discrepancies up to four months with regard to their supposed seasonal position, traditional knowledge tied to calendar dates about astronomical periods or agricultural seasons was annihilated. Thus, it was not so much the calculation and astronomical mistake of a surplus of one day over 365.25 days, but the practice of intercalation that needed to be addressed.
The period to do this were the chaotic years of civil strife and parallel attempts at rationalisation in many areas of the final years of the Republic. Gaius Julius Caesar used his unique power position as a dictator (and also supreme pontiff) to reduce the intercalatory period to just one day, to be intercalated every fourth year. This was achieved by distributing the days of the intercalatory months to the months of 29 days to 30 (January, April, June, September, November) or 31 (Sextilis, December). The fasti, the graphic representation, shrank from thirteen to twelve columns. Intercalation stayed at the traditional position on the fifth last day of February (in Roman counting: dies sextus before the Kalends of March) and by simply doubling this day. In an intercalary year, the 24th February was doubled (bi-sextilis). As before (and in contrast to Greek practices) this additional period was conceptualised as ‘void’, no interest was to be paid for it (less relevant for an intercalatory day than an intercalary month, though), any birthday occuring on such a day was to be shifted to the previous, the real sextilis. A singular lengthening of the year 46 BCE to 445 days brought the revised calendar in an intended alignment with the solar year, rendering 25 December winter solstice. Here a precedent for the Gregorian calendar reform with its dropping of eleven days was set.
A misunderstanding of the rule for intercalation (every three instead of four years) was corrected by Augustus in 8 BCE, clarifying the misunderstanding of ‘every four years’ into ‘every fifth year’. As pointed out before, translation of Roman counting in calendar matters was and is a difficulty. As might be imagined, Caesar’s reform caused immediate criticism, but not any successful resistance.
Rome did not consciously export this calendar, but it was the calendar referred to in Roman law, a set of norms defining (if necessary by analogous reasoning) juridical procedures more and more even on lower levels throughout the empire. The Roman fasti were the calendar also used by the Roman army, even in garrisons in far distant Mesopotamia. And it was the calendar used by the Roman administration throughout the empire. Many colonies and towns in the Western part of the Empire adopted it directly, many Eastern calendars renamed months, refixed their New Year’s day, or stabilized the length of months in order to achieve close parallels, while at the same time keeping the appearance of traditional local calendars. For an empire, not homogeneisation, but translatability of calendars was paramount.
This ‘Iulian’ calendar, stabilizing the solar year not only in theory, but finally also in practice, to 365.25 days was used in Europe and European colonies far into the early modern period and beyond. Its slow deviation from its astronomical and seasonal equivalents was not realised in Europe until the high Middle Ages. The reformed ‘Gregorian calendar’ of 1582 added a minor detail and a deviance from the Iulian calendar three times in four hundred years. It did not change the shape of the Roman calendar, originally designed for a single city, in any visible way.

Jörg Rüpke is a scholar of religion, focusing on the ancient world. He is the co-spokesperson of the KFG “Religion and Urbanity: Reciprocal Formations”.
Rüpke, Jörg. “The Roman Calendar – Time for a City and an Empire”. Religion and Urbanity Blog. KFG “Religion and Urbanity: Reciprocal Formations” (FOR 2779). 29/06/2020. https://urbrel.hypotheses.org/1077
OpenEdition suggests that you cite this post as follows:
Jörg Rüpke (June 29, 2020). The Roman Calendar – Time for a City and an Empire. Religion and Urbanity: Reciprocal Formations. Retrieved February 7, 2025 from https://doi.org/10.58079/v3cq
1 Response
[…] need or even defied a written fixation. Graphical representations of the type developed in Rome (see this blog post for details) were unknown in Greece. The so-called ‘sacrificial calendars’ known for example from Athens […]