Saturday 10 March 2018

New Year's Day and the Roman calendar

Happy New Year! In early Rome, anyway. Well, maybe ... actually maybe not. Can’t be sure.

‘Date’ and ‘calendar’ are ambiguous words. They can refer to how we label years, but also to how we label days within a year.

But historically, these two things -- year labels and day labels -- haven’t had all that much to do with each other. The date of Julius Caesar’s assassination is the 15th of March, 44 BCE. The Romans called it the 15th of March too (or the equivalent, in Latin) -- but they didn’t call it 44 BCE.

We label years with a number, using the CE/BCE or AD/BC system. This system came into use with 7th-8th century English historians like Bede and Alcuin. It was based on chronological work done by Christian calendrical-liturgical scholars who put a lot of effort into reconciling several different calendar traditions, such as Dionysius Exiguus (‘Dionysius the Puny’) in 525, and going back at least to the 100s.

But we label days within a year using a slightly modified form of the Julian calendar, a system that was introduced under Julius Caesar’s dictatorship in 46 BCE. So these are two independent things.

Just to make things worse: even after both of these systems were in widespread use, from the 700s onwards, 1 January wasn’t necessarily the start of the year. New Year’s Day could be a variety of different days depending on when, where, and whom we’re talking about.

In England before the Norman invasion, the New Year began on 25 March or 25 December -- so the day after 24 March 1050 was 25 March 1051 in the calendar of the time. After 1066 the New Year shifted to 1 January, but was put back on 25 March on the accession of Henry II. It stayed there until switching to 1 January again in 1751. France and Italy also used 25 March or 25 December up until the late 1500s, when they too moved to 1 January. Russia used 1 September up until 1700; that’s still the New Year in the Eastern Orthodox liturgical calendar. In the Catholic church the calendar starts on 1 January for some purposes, but for others (like in lectionaries) it begins on the first Sunday of Advent, at some point in the period 27 November to 3 December. The tax year is out of synch with the calendar year in plenty of countries too. But all of these systems use, or used, the Julian/Gregorian calendar for telling which day it is.

And then we have the Romans. In the past we’ve looked at the myth that the Romans used an ‘AUC’ system for specifying the year (only after the time of Varro, and only rarely). Now let’s talk about the other calendar -- the one for telling which day it is.

The last four months of the year are based on Latin number words. One of the most common complaints about the month-names that we’ve inherited from the Romans is that they’re the wrong numbers:

Month Month number Meaning of name
September 9 septem = 7
October 10 octo = 8
November 11 novem = 9
December 12 decem = 10

A fair number of people have heard a story that these months were originally the 7th, 8th, 9th, and 10th months of the year. That’s what the Romans themselves believed. It may even be true. But there’s quite a lot of misinformation floating around as to how and when it changed (if it did change).

Inscription dating to ‘the 17th day before the Kalends of Germanicus’ (CIL XI 5745 = ILS 6644)

Myth #1: July and August

One myth I’ve seen around is that the Roman calendar had 10 months up until the invention of the Julian calendar, and on that occasion Iulius (July) and Augustus (August) were added, named after Julius Caesar and Augustus respectively, to make a total of 12.

That is where the names July and August come from, but it’s false that they were added. They were just re-named: the republican calendar had 12 months too. July had been called Quintilis or Quinctilis, but was renamed after Caesar’s death, partly to honour Caesar himself (whose birthday was in July) and partly because of the new solar calendar that he had instituted in 46; Augustus himself did the renaming of Sextilis, probably in conjunction with another small calendar reform to correct how the Julian calendar was implemented.

If you know your Latin, you’ll spot that these follow a similar pattern to the other ‘number’ months, just with ordinal numbers instead of cardinals:

Month Month number Meaning of name
Quintilis 7 quintus = 5th
Sextilis 8 sextus = 6th

Some later rulers in the principate tried to rename months after themselves too. Thankfully, they never stuck. Gaius (Caligula) renamed September Germanicus, after one of his surnames; Nero renamed April, May, and June all after himself, as Neroneus, Claudius, and Germanicus; Domitian renamed September and October as Germanicus and Domitianus; Commodus renamed all twelve months after his various adopted names. Shudder. (Still, if any of these had stuck, there’d be a bright side: imagine celebrating Nero Fool’s Day on the 1st of Neroneus.)

A passage from one of Cicero’s letters to his brother Quintus, written in 56 BCE, mentioning a few dates: highlighted is K. Quintilis (1 July).

Myth #2: January and February

January and February were put at the start of the year at some point. But we don’t have any real idea how, why, or when, except that it was pretty early.

Roman tradition held that the mythical king Numa, the second of Rome’s legendary seven kings, added them onto a pre-existing 10-month calendar. But even if Numa was ever a real person, which is vanishingly unlikely, the Romans certainly didn’t have any records of any kind from that period.

They didn’t have records from anywhere near that period. There’s no authentic Roman history at all before the reign of Tarquinius Priscus, the fifth king. Even after that point, it’s more myth than history up until the 300s BCE. (As for Priscus, the only reason we suspect he’s real is because of the criterion of embarrassment: we assume that people don’t like inventing embarrassing stories about themselves, and Priscus was a potential embarrassment because he was Etruscan, not Roman.)

Livy, one of Rome’s best known historians, tells us himself that no written records survived from before the 300s. He guessed that the records were all destroyed in the Gaulish sack of Rome, traditionally dated to 390 BCE. More likely, no records ever existed. If Numa had existed, he’d have been in the 700s BCE: four hundred years before any actual written records. For stories set in that period our default assumption has to be that they’re all completely false.

That includes the stories about Numa reforming the Roman calendar. And it also includes the idea that there were originally 10 months in the calendar.

Now, the story that January and February got added to the start of the year looks intrinsically reasonable. First, there ought to be an explanation for the number-names of Quintilis to December; the January-February story explains it nicely. Second, the idea that they were originally intercalary winter months -- adjusted each year to make the dates come out in synch with the seasons -- fits well with the fact that intercalation in historical times was done in February (and still is today). So it looks very plausible that they have a separate origin.

When we see Roman writers coming up with the same story, it’s not because they had access to some kind of secret knowledge from the 700s BCE: it’s because it’s a good theory and they thought of it too. It doesn’t follow that anything else they say about the early calendar is true.

So while it’s completely plausible that January and February were added on, we know nothing at all about how they were added, and nothing about how the early Romans compensated for their absence beforehand. Roman writers tell us stories of extra intercalary months in winter; stories that January and February were originally at the end of the year, after December; stories that the early Romans just lived with having only 304 days in the calendar and as a result it was sometimes summer in December. We can’t draw any conclusions from those, because they’re all just guesses.

Myth #3: the names of the other months

So months 5 to 10 had names based on Latin numbers: ‘fifth-ilis’, ‘sixth-ilis’, ‘seven-ber’, ‘eight-ber’, and so on. What about the other months?

Month Meaning of name
Ianuarius (January) ‘month of the door’ (ianua ‘door’)
Februarius (February) ‘month of the februa’ (related to Lupercalia)
Martius (March) ‘month of Mars’
Aprilis (April) probably Etruscan: ‘month of Fortune’ (from Etruscan afr, apru(n))
Maius (May) ‘month of the elder’ (mai- ‘older, greater’)
Iunius (June) ‘month of the younger’ (iuni- ‘younger’)

January. Popularly thought to be named for the god Janus, but there’s no real basis for that. There was a very minor festival called Agonalia on the 9th of January that some sources claim was in honour of the god Janus -- but Agonalia also took place on the 20th of May and the 10th of December. It looks pretty obvious that the link to Janus was invented in hindsight.

February. Not ‘month of fevers’ (febres), as I vaguely remember being told when I was younger, but month of the februa. These were purification offerings for the festival of Lupercalia, on the 15th of February. For the same reason we also occasionally see Lupercalia referred to as ‘the februated day’ (dies februatus).

March. Mars was always absolutely central to Roman state religion. He was one of the Big Three along with Jupiter and Ceres, and intimately tied to Roman foundation myths. His place here doesn’t need much explanation.

April. Ancient writers liked to think that Aprilis came from the Greek goddess Aphrodite (originally pronounced Ap‘roditē), and this story is still in circulation. But it’s obviously guesswork, and a late idea: Romans in the time of the principate may have liked to think of Latin as a Greek dialect (wrongly), but the early Romans certainly didn’t use Greek like that. Another possibility (De Vaan 2008: 48) is suggested by the fact that Aprilis shares a suffix with two number-named-months, Quintilis and Sextilis. Aprilis could in principle be another one: it would come from an early compound of Latin ab/ap- ‘away from, off’, as in the verb aperio ‘to open’: *ap(e)rus could be then be an ordinal, ‘the following, second in sequence’, with Aprilis as a month-name based on that. But most probably the name is a borrowing from Etruscan. We have the names of some of the months in the Etruscan calendar, and April happens to correspond to Etruscan apru, aprun, or apira. The Etruscan name is based on the word afr or apher ‘fortune’, indicating the meaning ‘month of Fortune’, where ‘Fortune’ was an Etruscan divinity.

May. Maius is sometimes linked to the Roman goddess Maia (not to be confused with the Greek Maia, who was one of the Pleiades and mother of Hermes -- though the Romans eventually came to identify them with each other). That’s the reason we sometimes see Maia equated with the bona dea (‘good goddess’), whose festival was on the 1st of May. But more probably May and June are a pair: May ‘elder’, and June ‘junior’. This too is an ancient theory -- it’s Varro’s idea -- but unlike many ancient attempts at etymology, this one looks pretty likely.

June. This one is often linked to the goddess Juno, but as we just saw it’s pretty likely to mean ‘younger’, paired with May ‘elder’. If the name had come from Juno it would have to be Iunonius, not Iunius. Some sources (like the Oxford Classical Dictionary) suggest that the name is Etruscan -- Juno’s Etruscan name was Uni, and it’s a lot easier to see Iunius coming from that -- but that theory doesn’t hold water. First, Juno is an Indo-European name: the Etruscans borrowed Uni from Iuno, not the other way round. Second, we know what the Etruscans called the month of June, and it ain’t related to Uni: they called it acale or acle.

May and June are still connected to Maia and Juno, just indirectly. The names are linguistic cousins, not linguistic parent and child. That is to say: Maia doesn’t come from Maius, but they do both come from the same origin, mai-, meaning ‘elder’ or ‘greater’. And it’s the same story with Iunius and Iuno. Presumably Maia would originally have meant ‘elder (goddess)’ or ‘greater (goddess)’, and Juno would have been ‘goddess of youth’ at some point.

Just to finish off, here’s Ovid, with a nice mixture of myths and accurate etymologies:
These were the things that Quirinus [= Romulus] paid attention to
    when he gave his laws for the year to the rustic people.
The first month belonged to Mars, the second to Venus [= Aphrodite];
    she was the author of the race, he its father.
The third got its name from old people, the fourth from the young,
    and the crowd that followed were known by number.
-- Ovid, Fasti 1.37-42
Ovid knew his Varro. So he’s got March, May, and June right; but he’s wrong about April, and we just don’t know if he’s right or wrong about March being the first month.


  • De Vaan, Michel 2008. Etymological dictionary of Latin and the other Italic languages. Leiden/Boston: Brill.


  1. I remember reading that Mars was originally an agricultural god, so having a month named after him in spring would make a certain sense.

    1. That could certainly fit. I'd also love to hear an etruscologist's view on the history of April!