Greek in Italy

Project Research Blog

The inscriptions of Santa Maria in Trastevere


Santa Maria in Trastevere (the name of the neighbourhood means “across the Tiber”) stands on the site of one of Rome’s earliest churches. While not as big as San Pietro in the Vatican or the Basilica of St John Lateran – two of the other Christian sites of a similar age in Rome – Santa Maria is a beautiful church with stunning gold mosaics and mediaeval icons. But for me, the best part of the church is the wall outside, where hundreds of ancient inscriptions and fragments take pride of place.

DSC02508 DSC02502

Almost all of the inscriptions are gravestones from around the third century AD. Some were found in or near the church itself, while others were brought there from Appian Way outside Rome. Some include phrases and images that we associate with early Christian gravestones, but some could be either Christian or pagan. A few are carefully drafted and carved, and were clearly expensive monuments. But most are brief inscriptions that look like they were produced quickly and (comparatively) cheaply – and that means they are full of spelling mistakes.

Most epigraphers and linguists would horrify a high school Latin teacher, because we love spelling mistakes. Many non-standard spellings tell us what the language sounded like to native speakers (just as the jokey spelling “skool” tells you that school starts with a SK sound and not SH), which is especially important for studying languages which are no longer spoken. If we have enough examples from different periods, they can also show us change in pronunciation over time.

The inscriptions at Santa Maria show a fantastic range of telling mistakes. For example, it’s very common to miss out the letter N in particular places, like it the word coniunx (wife) which shows us that this was no longer pronounced. “To his wife” should be coniugi, but is often spelled coiugi, showing us that this was now how the word was pronounced.

My absolute favourite, though, is an inscription that combines losing the N in coniugi and merenti with spelling bene as baene. This is particular kind of mistake called “hypercorrection”. In the later Roman Empire, the sound represented by the earlier spelling <AE>, which sounded like the vowel sound in English bike, or the /ai/ sound in Italian hai, started to be pronounced exactly the same as one of the sounds represented by the letter <E>, a long /e:/ sound a little like in English there or Italian bene. The most educated writers had the opportunity to learn the ‘correct’ spelling of each word individually, but many writers spelled words as they pronounced them, using <E> whenever a long /e:/ sound came up, whether the word had originally been spelled with <AE> or with <E>.

The writer of this particular inscription had clearly been told off for getting <AE> and <E> confused. He knew that what sounded like <E> should sometimes be spelled <AE>, but rather like Eliza Doolittle and her Hs, he didn’t know exactly when this should happen.


What I love about this mistake is that bene merenti is an incredibly common set phrase, meaning “well-deserving”, i.e. deserving of a tombstone. It appears, correctly spelled, in almost every inscription on display at Santa Maria. It’s hard to shake the impression that had the writer just glanced a metre or two to his left, he would have seen a correct example to copy from.

The wall of the church also gives us a nice impression of the use of Greek in Rome during the third century. There are a few scattered inscriptions in Greek, and one that seems to use both Latin and Greek. And like all collections of inscriptions from the Roman Empire, there are many Greek names mixed in with the Latin ones – lots of slave and freedmen came from Greek-speaking areas of the Empire, and so names made up of both Latin and Greek elements were very common.

A few examples suggest that the use of Greek might have influenced how Latin was spelled. In two examples, carissima (dearest) is spelled karissima. The Latin alphabet did include the letter K, but it was used only in a tiny handful of words. The letter K was much more common in Greek, which had no letter C – so one possibility is that the spelling here shows a connection to the Greek alphabet. It could also be an attempt to make the inscription look “archaic” in some way.


PS – If you go to Trastevere, I recommend having a pizza in Dar Poeta. I also recommend bringing a Roman with you to translate the Romanesco (Roman dialect) poetry on the walls!

2 thoughts on “The inscriptions of Santa Maria in Trastevere

  1. Great post.

    “In the later Roman Empire, the sound represented by the earlier spelling , which sounded like the vowel sound in English bike, or the /ai/ sound in Italian hai”

    That is the consensus, but I’m not that sure that was the case. Ι’m of the impression that the distinction between an classical /ai/ diphthong and a late /e:/ long vowel takes spelling as a transparent representation of how Latin was pronounced… that one day, suddenly was abandoned in the whole of the Roman empire…

    Transparency, however, is always graded, and most languages don’t have a transparent spelling system at all (English, ehem, ehem, but not only). Combination of letters are also usually used to distinguish in writing what you don’t distinguish in speech (as Latin American Spanish does with s, c, and z).

    The earliest inscriptions have AI for /ai/, so I don’t think they needed a new combination for that sound. Instead, you have cases of EI vs I distinguishing long /i/ (open?) from short /i/ (closed?) sounds (like FILIEIS (dat. pl.) in CIL I, 583.). So, I would venture that AE was never truly pronounced as /ai/ but always as either a long /e:/ sound or any other kind of long vowel. Otherwise, I can’t see how you explain the need for hiatus in poetry (if you need to make clear that the two vowels are to be pronounced separately, it means that they’re not as a default position).

    I know, it’s my personal crusade, but when I hear dies irae pronounced as /irai/ instead of /ire/ it makes me cringe…

    • Thanks for reading, Juan! Nice to hear from you.

      You’re right of course, that most languages don’t have a transparent spelling system – but in most cases this reflects an earlier historical spelling. (English KNIFE shows that there used to be a K in that word, and an E sound on the end, for example). That’s exactly what has happened in this case. In later Latin AE and E both spell an /e/ sound – but the two spelling reflect an earlier difference of pronunciation.

      The use of hiatus in poetry happens because a diphthong /ai/ is a single vowel sound (one syllable), while two vowels /a/ and /i/ in hiatus would be two syllables. The ability to slightly manipulate the number of syllables can be useful in poetry, but it doesn’t mean Latin isn’t a diphthong.

      As for everyone changing their minds one day – actually I was not very detailed about when this sound change happened, but you can see evidence of it taking root even in the late Republic. It’s a very widespread language change, that Romans themselves noticed and commented on.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s