Greek in Italy

Project Research Blog

Leave a comment

‘Greek in Italy’ in Norfolk


Although a two-week family holiday in North Norfolk was not going to be an epigraphic extravaganza on the same scale as Nick’s visit to Naples, it did provide a reminder to blog about Greek in inscriptions now in Norfolk and some online resources. That said, we happened to see the famous dialect roadsigns that instruct ‘Slow you down!’ (with thanks to the Daily Telegraph).


Some months ago, a friend and colleague asked me about Latin manuscripts at Holkham Hall, of which there are many. At that, I searched Trismegistos, the database of databases for Greek, Latin, and Demotic texts from Egypt (to say the least) for Holkham Hall, as a Trismegistos Collection. The Latin manuscripts fall outside the coverage of Trismegistos, but two Latin inscriptions were reported.

Both are from Rome, both are funerary inscriptions, and both date to early in the second century AD (one may be slightly earlier).


CIL VI 2 14155

CIL VI.3 24008

Thanks to Trismegistos’ own data and its links to Clauss-Slaby (14155 and 24008) and the Electronic Archive of Greek and Latin Epigraphy/the Epigraphic Database Roma (14155 and 24008), full details of the two marbles and the images above can readily be accessed.

Neither inscription involves any Greek sentence, phrases, or script, but there is ‘Greek in Italy’ here nonetheless in the names. 14155.3-5 mentions a Calpurnia Chrysis mater. 24008 reads in full:

D(is) M(anibus) / Petronio / Hedychro / vix(it) an(nos) XXXV m(enses) VI d(ies) VII / Petronia Trophime / conliberto idem / coniugi suo b(ene) m(erenti) fec(it).

‘To the Spirits of the Departed: for Petronius Hedychrus; he lived for 35 years, 6 months, and seven days. Petronia Trophime made <this> for her fellow freedperson and “spouse” alike, who was well-deserving’.

Chrysis is a Greek name (the <ch> and <y> are classic giveaway indications of non-Latin words), as are Hedychrus (<ch> and <y> again – Ἡδύχρους; LGPN omits this bearer) and Trophime (<ph> is one indicator). In Trophime we see also a non-Latin ending. The first-declension nominative singular ends in –a (as in Petronia), but here we have –e as a transliteration of Greek <η>, the equivalent ending for the Greek first declension. One Τροφίμᾱ with -α, the Greek first-declension dialect ending, is also known to the Lexicon of Greek Personal Names.

These three – Goldie, M(aste)r Sweet-Fleshed, and Miss Foster(ed) or Miss Plump – look like slave names, a suggestion corroborated by conliberto ‘fellow freedperson’.

The name Ἡδύχρους (or -χροος) looks like a poetic epithet, as, indeed, it was: GVI 1595.13-14 [Rome, perhaps second century AD]; cf. IG XII, 1 781.4 [Rhodes; second century AD], which has a dative -χροι from a by form in -χρως. However, it was also the name of a perfume. As such, it is also a Greek word known first from a Latin text: Cicero, Tusculan Disputations III 46.

The manuscripts there all have hedyc(r)um (one has aedicrum). Although the <h> and the  <y> were preserved, there is no trace of the <ch>. Whether Cicero spelled it with <c> or with a <ch> that has been lost in transmission, we cannot know. OLD printed the headword as hedycrum with <c>, not <ch>; some texts have the <ch> restored, as LSJ gives it.

Also, –crum indicates that the substantive (like the personal name) was borrowed into Latin from a variety of Greek that has the contraction -χρους, not the uncontracted form -χροος. (Galen has both -χροον and -χρουν).

So much, for now, for Greek in Italy via Norfolk and this experiment with ‘linked data’. I’ll leave it to James to tell the story of the Norfolk clergyperson who argued that the topography of the Iliad was based on his own – and Nelson’s – county.



Hypermobile Mamertines?

At last year’s Laurence Seminar (‘Migration, Mobility, and Language Contact‘), Nick presented a paper about ‘the Mobile Mamertini’, mercenaries from Campania or Samnium who captured Messina circa 288 BCE.

While investigating the earliest appearances of Italians in Egypt, I found that Csaba La’da, one of my papyrology teachers, had reported an instance of the ethnic Μαμερτῖνος in his Foreign Ethnics in Hellenistic Egypt (2002). Thanks to the amicitia papyrologorum, I have obtained a copy of the original edition. I have not yet seen an image or the inscribed surface itself…

The funerary inscription (TM 47467), which remains in situ in Hadra in Alexandria (ad Aegyptum), is still cited from its 1915 (re-)publication as SB I 419f and is dated between 199-30 BCE. It reads:

Μαραῖος Βακείου | Μαμερτῖνος.

‘Maraeus, son of Baceus, a Mamertine’.

So, we have another Mamertine, but outside Italy this time and later than expected (the items in Nick’s dossier are all assumed to belong to the third century BCE, the latest being coinage c. 225 BCE).

The content follows the Greek pattern of the deceased’s name in the nominative case, his father’s name in the genitive case, and, then, the deceased’s ‘ethnic’ (or equivalent for the context). The orator Demosthenes illustrates this pattern, but with his Attic demotic rather than this ethnic Ἀθηναῖος: Δημοσθένης Δημοσθένους Παιανιεύς (D.18.187). (The Latin pattern, in full, includes the tria nomina and F(ilius) after the father’s name in the genitive.)

Maraeus, Nick informed me, is a good Oscan name: Marahis. It is rarely attested in Greek texts. The Lexicon of Greek Personal Names (LGPN) knows nine instances, all more-or-less Late Hellenistic in date, several from Asia Minor, and two, comfortingly, from Naples in Campania (e.g.). The nine instances include two father-son pairs. Trismegistos People (TM), as a whole, adds only two more bearers of this name (from two Petrie papyri, both found at Gurob and both written in the last decade of the third century BCE). One of them was a τακτόμισθος (as LSJ puts it, ‘a rank in the army of the Ptolemies‘…). That term is confined to papyri and, on the face of it, might have something to do with soldiers for hire (μισθός). Fischer-Bovet suggested ‘(mercenary) with a fixed wage’ as a gloss.

The father’s name, Βακεῖος is hard to parallel. LGPN’s online search silently corrected it to Βακ<χ>εῖος (the printed edition is transparent on this point), a name that seems unremarkable except that it has only one other attestation: Baccheius of Tanagra, a third-century BCE Hippocratic commentator. Trismegistos People treats the father’s name as a spelling variant of Βάκιος, a name attested on a Roman-period ostracon , whose most recent editors instruct ‘l. Βάκχιος’. (Simplification of <κχ> to <χ> in βακχ- words is illustrated by Gignac (1976:100) and by Threatte (1980: 542f.) and the interchange of <χ> and <κ> between vowels by Gignac (1976: 92) and Threatte (1980: 453ff.). Mayser-Schmoll (1970: 186) provides comparable Βακχ-, Βαχχ-, and βαχ- data, but not βακ-. The simplification of <κχ> to <κ> is not illustrated. <κ> may have been written for <κκ> or Latin .)

Since Greek and Latin texts provide no supporting evidence for the father’s name and the alternative readings proposed in the databases both have weaknesses of their own, what are we to do? Nick drew my attention to the abbreviation BAK() at Pompeii in L BAK() inscribed right to left in Oscan script on a (small) black slip bowl (uasculum) after it had been fired. Whatever name was so abbreviated could give us a point of comparison with the father of our Maraeus and Southern Italy.

However, our knowledge of that inscription is very shaky because it depends entirely on H. Dressel’s drawing and description, which underlie CIL X (1883) 8055.67. On the basis of that drawing, some have read BAD with Oscan <R> for /d/ (Rix, ST Po 88), while others maintain that the final <K> cannot be an incomplete Oscan <R> (Crawford et al., Imagines Italicae, Pompei 86: L(ucius) Bac(culeius) is suggested there, with the Pompeian M. Bacculeius of CIL IV, Supp. 3 (1963) 9256 as a comparison).

Whatever is going on with ΒΑΚΕΙΟΥ, we have a Mamertine with an Oscan name buried in Hadra in Alexandria in the Ptolemaic period.

What about later Μαμερτῖνοι in Egypt? Connections with the Mamertine mercenaries faded and Μαμερτ(ε)ῖνος (LGPN and TM) remained in use only as a cognomen (and hence, from an agnomen, Μαμερτινιανός), Marcus Petronius Mamertinus, the second-century statesman, being its most well-known bearer.

Leave a comment

Gambling in Greek

A new article by James has just been published on ‘The language of a Pompeian tavern: submerged Latin?’.* He discusses the ‘captions’ on a set of four wall-paintings from Pompeii, depicting two men chatting, getting drinks, playing a dice-game, and having a fight and being ejected. These paintings are now quite well-known, having featured prominently in the British Museum’s Pompeii exhibition a few years ago (our colleague Mary Beard has mentioned it in her blog a couple of times, with helpful pictures). James’ focus in the article is what the words tell us about sub-elite Latin, and whether there are similarities to the language of the Roman playwright Plautus, two or three centuries earlier. No spoilers here: you’ll have to read the article to find out the answer.

But the inscriptions also show how commonplace Greek words were in Latin, even (or especially) at social levels for whose language we have less literary evidence. The first panel has one man saying to another ‘nolo cum Murtale uasu’, which probably means ‘I don’t want a drink with Myrtale’. Myrtale is a Greek name and the final -e is probably meant to represent the Greek dative singular ending -ηι (the final -ι would have been lost in speech by this time). In the third panel, one speaker says to the other ‘non tria, duas est’, ‘that’s not a three, it’s a two’, using the Greek word duas ‘a two, deuce’. This word doesn’t appear again for centuries, and there are no other examples of it being used in the context of numbers on a dice. If it weren’t for this inscription, we’d have no idea that it had this meaning in Latin, or that it had been borrowed so early.

This inscription is also particularly interesting for me, because I’ve been thinking lately about Roman spelling, and in particularly how features that tend to be described as ‘old-fashioned’ or ‘archaic’ actually had a much longer after-life than we probably give them credit for – perhaps especially in the sub-elite educational tradition. These scenes can be quite accurately dated to between 62AD, when an earthquake hit Pompeii, and 79AD, since that is when Vesuvius exploded and destroyed/preserved Pompeii. Latin inscriptions had started using the letter y to represent Greek υ by the early first century BC, but here they are still using u 150 years later! And that’s not even the record; the captions spell the word ego ‘I’ eco, a mere three hundred years after the letter g was invented. But maybe we shouldn’t be so surprised: spelling can be a remarkably conservative practice –  after all, English spelling largely carries on as it was established several centuries ago, often reflecting pronunciation from even before that, despite an almost-complete failure to match how we actually speak nowadays.

*The article is published in Early and Late Latin: Continuity or Change, edited by Jim Adams and Nigel Vincent, published by Cambridge University Press

Leave a comment

From Teano to Bacoli

After Geoff, Livia and Katherine went home, James and I stayed in Naples for another week, to attend the 22nd International Conference of Historical Linguistics. Apart from going to lots of interesting papers, we also took the chance to do a bit more exploring, this time of places closer to Naples. Our first stop was Teano, which is a lovely hillside town: while I suspect Sunday midday is not the time to see it at its bustling best, the museum, much to our relief, was open. It’s just one floor, but it crams an amazing amount of fascinating material into the space. There’s some nice Greek-looking statuary and grave goods, but for me of course the highlight was the large number of Oscan inscriptions. Funerary inscriptions are relatively uncommon in Oscan, but there are quite a number of examples in Teano, including one of the few cases of a woman’s name (for more info on these see my post Name Games), pictured below. The inscription says Ep(pio?) Loukio daughter of Min(is).

Ep. Loukio

Teano in the late 4th century BC was also home to a family firm making plates and bowls: helpfully they were in the habit of writing variations on the information you can see in the picture below on their pots. It says: Minis Berriis. They (the pots) were made at the workshop (?) in Teanum. Note that Minis was clearly a popular name in Teano.

Minis Beriis

Next stop was the – really large – museum in Capua, which has an entire room devoted to epigraphy (bliss!) as well as lots of inscriptions scattered elsewhere about the place. For those who are slightly less one-track-minded than us, probably the most striking thing is the amazing collection of Matres Matutae: these are scores of – presumably religious – statues of seated women holding babies. The early ones (6th century BC) start off being only vaguely female and holding a couple of babies in swaddling: by the end (2nd century BC) they are distinctly beefy and unimpressed-looking matrons clutching up to twelve of the things. Very weird, especially in large numbers. Finally, via a detour to the lovely hill-side church at Sant’ Angelo in Formis, formerly a Roman temple, to Santa Maria Capua Vetere, for a few more inscriptions, and a great tomb-painting of a Samnite warrior.

Our next adventure, later in the week, was to Bacoli (ancent Baiae) where we saw the Piscina Mirabilis, which is aptly named, since it is indeed amazing. It is not a cathedral, as you might think from the picture, nor a swimming pool, but an underground reservoir which was used to collect and distribute water from the aqueducts going into Baiae. It’s a bit difficult to get into, since you have to pre-arrange an appointment with the woman who holds the keys. Fortunately we had the services of Ludovica, our superlative guide, who organised everything. Pictured are me and James pretending to say wise things about Roman brickwork, while Ludovica looks sceptical.

Piscina Mirabilis 1

Piscina Mirabilis 2

From there we went to the Museo Archeologico dei Campi Flegrei. Like so many of the museums we’ve been to in Italy it’s brilliantly presented and full of fantastic stuff – and in an old Aragonese castle with a wonderful view over the Bay of Naples. Cumae, just around the corner, was the earliest mainland Greek colony in Italy, and the museum has a number of early Greek inscriptions, several with features which are more or less unique – it’s amazing that even in a language which we know as well as Greek, a single inscription can make a big difference to what we think we know. The inscription below is a nice example of boustrophedon writing, where the lines are written alternately left-to-right and right-to-left.


As we were leaving the museum the exhaust fell off our car, so that was the end of the ancient part of our adventures that day!

1 Comment

Name games

I’m writing a book at the moment on how Oscan-speakers spelled their language when it was written in the Greek alphabet, and one of the side-effects of this is that I needed to know how women’s names worked in Oscan. Men’s names are easy, and work much like in Latin (and a bit like in English): everyone has a praenomen (which is like our first name), and a gentilicium (like our surname, passed down in families); usually we’re also told what their father’s praenomen was too, in the genitive (in Latin this is followed by f. for filius ‘son’, but there’s no equivalent abbreviation in Oscan). In Classical times, Roman men also had a cognomen, which started off as a kind of nickname, but later became part of the official part of the name. Sometimes Oscan men have cognomina too, but it seems to have been quite rare. So, a typical Oscan man’s name would be Vibis Púpidiis Vibieis “Vibis Popidiis, son of Vibis”.

So what are women called? You might assume that women’s names would work much the same, with a praenomen and gentilicium. But that’s not the case in Latin, at least by Classical times, where women just get one official name, the gentilicium. Clearly, if there was more than one daughter in the family, there must have been some way for people to distinguish them, and indeed once the cognomen began to take off for men, we also find women with cognomina. So, women have one less name than men: either just a gentilicium (e.g. Quinctilia) to a man’s praenomen + gentilicium , or a gentilicium + cognomen (e.g. Appuleia Varilla) to a man’s praenomen + gentilicium + cognomen. Just as the men’s names can be followed by the father’s names in the genitive, so the women’s names can be followed by their father’s names, or by their husband’s names. As I say, this is the way that things work in the Classical period, but there’s a small number of examples of female names from Praeneste, near Rome, in the third century BC, which seem to imply that at this stage women could have a praenomen, like men: was this later dropped, or did they just have different naming habits in Praeneste?

Since the system of men’s names is so similar in Oscan and Latin (and, with minor differences, in Umbrian and Etruscan), we might expect that the same is true for women’s names. But is this in fact the case, and even if so, did Oscan women’s names work like those in Praeneste or in the same way as in Classical times? It turns out that it’s very difficult to know for sure.

One reason for this is that, unlike men’s names, which we find in inscriptions all over the place, there are hardly any women’s names written down, so we don’t have much evidence. The other, is that it’s often quite hard to know how to analyse the evidence. Some of our examples are just single names. Should we take these as praenomina or gentilicia? One way to tell is to look at the name and see whether the male equivalent is used as a praenomen or a gentilicium. But this isn’t as easy as it sounds. Most gentilicia are originally derived from praenomina (that is, they probably started off as a patronymic, before becoming fixed as a family name, just as the English surname Jackson originally meant Jack’s son). This is done by adding is to the name (so next to the praenomen Heirens there is a gentilicium Heirennis). But some praenomina already end in -is, and then the result is a gentilicium in -iis (for example, praenomen Dekis, gentilicium Dekiis). The equivalent for women is praenomina or gentilicia in –, gentilicia in –iiú.

This means it’s not always easy to tell whether a name ending in –is or –iú is a praenomen or a gentilicium unless you have definite cases of that name being used in a male name written out in the full sequence praenomen + gentilicium + father’s name, where you can tell by the position of the names. There’s also a further complication: normally, at least if you find a spelling -iiú, you can be sure that you’re dealing with a gentilicium. But at some point there was a rule in Oscan that changed a single –i– after a –v– into a double -ii- (we can see this in words like menereviius ‘belonging to Minerva’, which we know for etymological reasons should be menerevius). So in the case of the single name Úviiú ‘Ovia’, we’ve no way of knowing whether it’s a praenomen or a gentilicium. And things are even more difficult in the Greek alphabet, where no distinction is made in writing between -iú and –iiú: they’re both spelt -ιο. Unfortunately, by far the majority of our examples of female names are written in the Greek alphabet…

We may not be able to tell from the single names. But when we have two names together, it’s a fair bet that thεy’re a praenomen + gentilicium sequence, right? Well, not necessarily. Some scholars have suggested that in cases like siviiú magiú, what we have are two gentilicia, one that of the woman’s father, one that of the husband. I think that this is wrong, and that Oscan women did have a praenomen of their own, as well as a gentilicium. But in the end, this comes down to a single piece of evidence, the name ahvdiú ni(umsieís) ‘Audia daughter/wife of Numsis’. We’ve got an example of the male equivalent of this name as a gentilicium, and it’s spelt ahvdiis, so the female gentilicium ought to be ahvdiiú, and ahvdiú ought to be a praenomen.

So it looks as though Oscan women, like the women of Praeneste, were better off than most of their Roman equivalents in the way of names, having two, just like their menfolk. But it would be nice to have more evidence, so if you find any, let me know!