But only some numbers are represented. Why don’t we see Romans named ‘Primus’, ‘Secundus’, ‘Tertius’, or ‘Quartus’? Or for that matter ‘Septimus’, ‘Octavus’, or ‘Nonus’?
Today’s post isn’t really a debunking of a popular myth. An unpopular myth, maybe. I’m posting it because it’s something I just learned this morning, and it shocked me. It’s one of those things that’s staring you in the face all the time when you’re reading about the Greco-Roman world. So when I found out the true explanation, I felt a little bit betrayed -- as though it was something I ought to have known all along.
I guess it’s hard to find the time to get around to thinking about why Romans had names that meant ‘fifth’, ‘sixth’, and ‘tenth’. If you do think about it, you’re likely to make the same assumption that I did: that children were named for the order in which they were born. The 1st son would be Primus, the 2nd Secundus, the 3rd Tertius, and so on.
But that isn’t the case. If it were, we’d see corresponding names for the first to fourth children. And they just don’t exist. We do find Primus, Secundus, Tertius, and Quartus as regular cognomina -- official nicknames -- but not as personal names, and not at an early period. They start to pop up in the imperial period, and they’re not in Rome: they appear in Celtic contexts, which tends to suggest contamination from Celtic naming customs (Petersen 1962: 349 n. 6).
What’s the solution, then? Should we assume that the first four sons would get ‘real’ names, like Marcus, Titus, Publius, and so on? And when the parents got to their fifth child, suddenly they’d be all like ‘Hey let’s start using numbers now.’
Nope. The scholarly consensus is that these names originally came from the names of the months in which they were born.
If that surprises you, you’re not alone. I was startled too. It was originally suggested by the 1st century BCE scholar Varro, and it appears he was dead right. The standard modern discussion is 56 years old, Petersen 1962, but Petersen’s argument hasn’t been superceded. If anything the argument has been strengthened by parallels that have been found in other ancient Italian languages.
There are a couple of complications. First: Quintus and Sextus don’t sound like month names. However, prior to the lives of Julius Caesar and the emperor Augustus, the Romans had different names for the 7th and 8th months of the year: Quintilis (‘fifth-ilis’) and Sextilis (‘sixth-ilis’). (see this previous post from March for more details, and for discussion of why the months’ names don’t match up to their position in the year.)
Second: it’s not just the number-months. March, May and June aren’t named for Latin numbers, but children born in those months got names related to the months anyway: not Martus, but Marcus; Maius is a rare name, but it exists; and June gave Iunius both as a praenomen and as a gentilician (family) name.
And third, there’s a semi-regular pattern of praenomina ending in -us with corresponding family names ending in -ius: hence Marcus ≈ Marcius. This helps fill in the gaps with some of the numbers. We don’t see Octavus as a praenomen, but we do see Octavius; Iunius is rare as a praenomen, but common as a family name; Maius appears both as a praenomen and a family name.
|Iunius (June)||Iunius (very rare)||Iunius|
|September||Septimus (rare, archaic)||Septimius|
Varro, who first came up with this explanation, was bothered a bit by the lack of any names corresponding to April. But not enough to prevent him from proposing it anyway, and not enough to put off modern proponents. (The fact that the name ‘April’ appears to come from Etruscan may have a lot to do with this gap in the table. Maybe one of the other traditional Roman praenomina, like Gaius and Publius and Titus, was related to an older Latin name for April? Who knows.)
What are the arguments in favour of Varro’s theory? Well, first is the fact that the number-based praenomina start with Quintus and end with Decimus, and this constraint corresponds tidily to the fact that month-names also start at ‘fifth’ (Quintilis) and end at ‘ten’ (December).
Second is the fact that we find some related names in two ancient Italic languages related to Latin, Oscan and Faliscan, in ways that suggest they’re also related to month names.
Oscan was spoken by the ancient Samnites, in the Appenine mountains south of Rome, and we have a Samnite family named Decimius attested in Roman sources. In Rome itself the corresponding name was ‘Decius’. In a study of the ancient Samnites, E. T. Salmon (1967: 53) cites the Oscan names Mamerkis ≈ Marcus, Sepis ≈ Septimus, and Dekis ≈ Decimus, and states that Mamerkis is actually formed from the Oscan name for the month of March. I haven’t been able to confirm the last point, but it’s certainly true that Mamerkis is related to the god Mars, who was called Mamertis in Oscan.
Petersen cites some parallels from Faliscan and Oscan too. Marcius only appears as a family name in Rome, but Petersen points to an example of Marcius as a praenomen in a Faliscan inscription (1962: 352 n. 16).
And most strikingly of all, he cites an Oscan family name Sehsimbriis. This is transparently based on an alternate formation for the name of the month of August. In republican-era Latin, August was called Sextilis (‘sixth-ilis’); Sehsimbriis must reflect an Oscan formation which would correspond to Latin *Seximber.
|Note, a day later: I’m no longer too sure of this. If Sehsimbriis did correspond to *Seximber, it’d be an artificial formation, based on analogy with September, November, etc.: the Latin for ‘six’ is sex, not *sexem/sexim. A strict formation ought to be just *Sexiber. (Or in Oscan maybe something like *Sehsikbri-: sehsik- appears to be the Oscan for ‘six’.)|
So there we have it: some of the most common praenomina in Latin, Marcus, Quintus, Sextus, and Decimus, come from month names. If your name is Mark, you’re named after the month of March. The idea must surely be that in the earliest times, the names would have corresponded to the month in which the boy was born.
There are some caveats and provisos, mind. First, customs changed over time. The practice of naming children directly for months was long gone by the historical period. And in the imperial era, some number-based names were formed by analogy with the traditional names: as a result we start to see some innovations like Decimius as a praenomen, as in the name of the poet Decimius Magnus Ausonius.
Second, women’s names are different. Men’s names followed a regular pattern of praenomen plus gentilician name; women’s names were much less regulated. And among women we do see Prima, Secunda, Tertia, and Quarta as personal names alongside Quinta. So however exactly the month-name custom worked originally, it didn’t work the same way for women’s names. It could be that for women these names did originally indicate order of birth.
And third: we do see some other number-names popping up as gentilician names which do not correspond to month names. These names seem to come from non-Roman contexts. For example, the Roman name Pomponius is based on a non-Roman word for ‘five’. ‘Five’ in proto-Italic was *kwenkwe. Latin preserved /kw/ sounds relatively faithfully, and so ended up with the form quinque ‘five’. But in many languages, /kw/ transformed into /p/: so in Oscan the word for ‘five’ was pumperias or pompe. This or a related language must have provided the gentilician name Pomponius, basically an Oscan equivalent of Quinctius. In the same way proto-Italic *kwetwōr ‘four’ ended up as quattuor in Latin, but pettiur or pitora in Oscan: a form related to these must be the origin of the Roman gentilician name Petronius.
- Petersen, Hans 1962. ‘The numeral praenomina of the Romans.’ Transactions and Proceedings of the American Philological Association 93: 347-354.
- Salmon, E. T. 1967. Samnium and the Samnites. Cambridge: Cambridge University Press.