On Haccents in the Roman World

…sit quaedam certa uox Romani generis urbisque propria…

…there is a certain voice (=accent) peculiar to the Roman people and city…

Cicero, De Oratore 3.44

What can we say about accents in the Roman world? We know, can see around us, that eventually Latin would diversify into the modern Romance languages. Are we then to imagine senators from Gaul twirling their moustachios and swapping out hon hon hon for Plautus’ hae hae hae? What about the inevitable interference from languages elsewhere in the Empire (Greek, Armenian, Syriac, Etruscan etc etc etc)?

You can throw a pin at the Roman map and find something interesting to talk about linguistics wise. Even within Rome itself social stratification would have rendered a few different accents, the same as any city at any time. This is an interesting topic (both in terms of subject matter and in trying to avoid being too technical) and I think it might be one I return to again and again.

We need to heavily narrow the terms of our enquiry and so I’d like to posit we examine aspiration as a loose nexus. In doing so, we can look at the phenomenon within Latin and as affected by non-native speakers. Aspiration seldom occurs alone and so, naturally, we’re going to look a little at aspirated consonants and a few vowels via both literary and inscriptional evidence and attitudes towards this linguistic diversity. This is hardly an essay, but a hodgepodge of connected musings.

Let’s select a passage from Catullus to get us started:

Chommoda dicebat, si quando commoda vellet

dicere, et insidias Arrius hinsidias,

et tum mirifice sperabat se esse locutum,

cum quantum poterat dixerat hinsidias.

credo, sic mater, sic liber avunculus eius.

sic maternus auus dixerat atque avia.

Hadvantages, Arrius would say whenever he meant to say

advantages. Ambushes, too, [he called] hambushes.

Then he was hoping he had spoken wonderfully

when he said hambushes as much as he could

Thus I believe his mother spoke, his free uncle

his maternal grandfather and grandmother.

Catullus 84 1-6

Even without its famous punchline, the humour of the poem is the product of more than slapstick over poor pronunciation (but seriously go read it). The use of imperfect tense suggest a repetitive action, deliberately taken over and over. In colloquial English we might say that Arrius is ‘putting it on’ and are meant to laugh at this parvenu incapable of aspirating correctly.

But what’s the linguistic implication? Either (as seems most likely) Arrius is a native speaker of a Latin dialect which has lost the /h/ sound or he’s a second language speaker unable to replicate the Roman sound. I must say, the latter was my initial reaction. After all, why else the reference to his family speaking the same way? The reference to a ‘free uncle’ clearly is meant to contrast with former slave status. Certain textual critics have even tried (untenable) to correct liber into a non-Roman name. It’s easy to imagine Catullus’ disgust at the product of the recently free making his way around high society and failing to blend in due to his poor speech.

Against that, however, is the fact that this Arrius (Harrius, surely? … sorry) is most likely the Q. Arrius we know from elsewhere, famed for his ostentatious failed bid at the consulship. Under Sulla various ex-slaves did well (cf. Chrysogonus) but it would have been unthinkable to aim for the consulship.

Either way, it seems that his dialect( /interference from first language) gave him trouble with aspirating hence his over compensating. This is a process known in linguistics as hyper-correction (see the link below) and we can safely conclude that there were Latin dialects where the aspirate was inconsistently applied.

In the context of philology, processes such as hyper-correction and analogy are studied as matter of course, but in a sociolinguistic context we might say something about the attitude present. In other words, we’ve established that this variation exists – but why is Arrius so keen to falsify his mode of speech?

The answer, I think, is obvious. He’s trying to fit in with the other upper-class Romans and so needs something approaching Cicero’s vox Romana. Poor aspiration, however, seemed to be a particular signifier of poor speech as a fragment of the grammarian Nigidius Figulus, who was active around the same time, suggests: rusticus fit sermo, inquit, si adspires perperam (speech becomes rustic, he says, if you aspirate wrongly).

It’s notable there’s a kind of urban bias here. M. Clodius Pulcher the “patrician tribune”, famously affected a plebian pronunciation hence his name being Clodius and not, as would be proper, Claudius. Someone of his station could get away with this idiosyncrasy and, besides, he spoke with an accent of the city and not the hinterland. Conversely, in the Imperial period, both Hadrian and Septimius Severus (Hispania and Africa respectively) could be castigated for their speech despite belonging to the upper echelons of society. They were, after all, provincials.

Biases regarding pronunciation are rational in so much as we are able to discern that reasons exist (e.g urban vs rural) but they rarely are logical or reasonable. I want to illustrate this with a brief interlude from Eddie Izzard, which also concerns aspiration.


Not to detract from Eddie Izzard’s joke, but our (British) pronunciation of /h/ here is in fact an ‘error’. The Americans have it right in so much as the traditional English pronunciation should eschew h in the same way we do for honour or when (ostensibly, I don’t personally) pronouncing the name of the letter itself. Yet somehow Izzard’s pronunciation would be seen as more high status. Attitudes trump reality.

The case in Latin is similar. In many cases consistent aspiration amongst even the learned classes was a fairly recent phenomenon. Ennius had pulcer not pulcher, triumphus was originally triumpus (of Greek origin, via Etruscan) and so on. Actually, philology can triangulate interferences between these three languages (Greek, Etruscan, Latin) to note that there is a shift from Greek to Etruscan that involves devoicing (b > p, g > k etc) and therefore conclude that since the Etruscan ear was less sensitive to voiced consonants, they would have likewise struggled with these in Latin.

H seems more or less always to have been weakening in Latin (although even by St. Augustine’s time people are still insisting on it). Many of the word initial h’s had a fuller sound in the parent language: homo, man, was  *ǵʰmṓ in PIE, reduced to *hemo in proto-Italic (nemo, nobody, < *nehemo, it was commonly lost between letters) and of course the modern romance languages have continued this trend and reduced h further: omo (Italian), homme (French, etymological h), hombre (Spanish, etymological h) etc.

Even Cicero (Orator 160) can’t be consistent and ends up following popular usage. Learned speech likewise permits stray h’s via hypercorrection: humidus was properly umidus for example, but anyone pronouncing it was such would have been labelled by men like Nigidius Figulus as a bumpkin.

Poor Arrius! Who can keep up?!

We could go on piece by piece focusing on different sounds in different Latin accents, but let’s keep with the h theme, only this time we’ll look briefly at a foreign accent – that of Greek. Unsurprising, Greek gives us our greatest amount of evidence for second language interference in Latin both in textual and inscriptional terms. Admittedly, most discussions of foreign interference tend to focus on either a) lexical borrowing or b) morpho-syntactical peculiarities (properly termed solecisms) e.g incorrect case usage. That’s just the nature of the evidence, but it is possible to talk a little about accent.

This inscription, or graffito rather, from Pompeii is a good jump off point. I’m…eh…not going to translate it for obvious reasons given what it says:

Tiopilus, canis,

cunnu lingere no-

li puellis in muro

CIL IV 8898

Tiopilus here stands for the Greek name Theophilos; the differences in rendering here are elucidating and we can take them section by section. Given this post’s focus on aspiration, I’m sure we can all see why this was chosen.

Aspiration in ancient Greek wasn’t like its modern counterpart, where θ is pronounced the same as English th. In fact, though I don’t have the figures to hand, this is a rare phoneme hence when jokes are made about German accents in English the definite article is always rendered ‘ze’. Originally, φ, χ and θ were simply aspirated stops: imagine a subtle expulsion of breath after each consonant. Later, these became the fricatives we all know and love. This is why the Romans spelt philosophy philosophia and not *filosofia despite widely employing f: because the Greek sound was different.

So, we can conclude that this is an area where a Greek speaker might have trouble. In fact Quintilian tells us of a case where a Greek witness was unable to pronounce the name Fundianus and came out with Hundianus. This makes sense, f was missing from Greek’s phonemic inventory and his native accent would have permitted poor speakers to utter either *Pundianus or *Hundianus.

The other letters are equally interesting even if not really germane to our h theme. What in Latin script has been rendered as ‘i’ are actually two difference sounds in the Greek (ε and ι). How big a difference existed between these letters in Latin? Learned speech kept them distinct, but even so we know of many instances of ‘e’ standing in for ‘i’ (the historian Livy’s accent would have merged these sounds), the distinction must have been difficult for Greeks.

As for os vs -us for endings well this is a typical equivalent. Even now, everyone knows Latin names end in us and Greek in os e.g Constantinus to Κωνσταντῖνος but this confusion isn’t always learned and occurs elsewhere e.g in the so called Colloquia, remnants of bilingual texts for teaching Latin or Greek, secunda is rendered as σεκονδα or even σεκοντα (ντ for δ is a whole different kettle of fish).

Even in Latin, the distinction wasn’t always obvious. Yes, we all know the myriad ways in which U > O became in Romance but one even comes across inscriptions with spellings like apud loco. Loco? I hear you ask, how can this be? Is this evidence for the later collapse of the case system in Latin?

Well, yes, but more importantly it’s evidence for the phonological underpinnings of these changes. Apud takes the accusative and therefore one would expect apud locum. Interestingly final m is often left off in more casual inscriptions (as in the 2nd line of the inscription above) because m was nasalised. Because the remaining sound wasn’t quite either a u or an o (try it out for yourself) it made sense for less learned writers to shift the spelling this way. This isn’t an error in case usage but an accurate rendering of spoken habit.

Given this propensity to change this vowel sound depending on what follows by native speakers, It’s easy to imagine Greeks not necessarily accurately mapping their u’s and o’s to their Latin equivalents either. After all, subtleties like this are the most common stumbling blocks in language acquisition.

In fact, we can go further. Modern Greek pronunciations of Latin retains some elements of the accents Romans would have heard. Emphasis on some, Greek has experienced its own significant changes, as we’ve had cause to note above. Vowel quantity has been lost, but there is still a sense of breadth to remaining vowels and, yes, including confusion in articulating ‘e’ vs ‘i’ in speech. Most Greek dialects (Cypriot a sometime exception) no longer distinguish between geminated and singular consonants (e.g ss vs s) and we know that Roman writers from the Imperial period commented on this. Ok, two examples are not a lot, it’s easy to overstate the case, but you must admit it’s cool to realise that you can hear something close to what the Romans did just by observing a Latin lesson in Greece.

Can we offer a quick conclusion? We haven’t, obviously, come anywhere near the point in this post where we could posit the full Greek accent of Latin (by what speakers? At which period? etc) or gone into great depth about regional Latin pronunciation (same questions) but I hope I’ve given a reasonably satisfying glimpse of what’s out there. So, when thinking about the Latin heard in Roman streets perhaps now you’ll contrast the ‘correct’ pronunciation we’re taught in school with the unaspirating rustic hawking his fish or the smooth consonant-broad vowelled Greek visitor. Perhaps something for us to come back to.

Further Reading


Any introduction will do but I enjoyed:

Wardhaugh, R (2009) An Introduction to Sociolinguistics. Oxford

Perhaps the best way into seeing these applied to the ancient world are:

McDonald, K (2015) Oscan in Southern Italy. Cambridge. Especially chapter 2

Latin specifically

Literally anything by J Adams, even his grocery list will probably teach you some hitherto unknown fact about the accentuation of the plu perfect passive by Sarmatian horsemen stationed in Eboracum but my thinking on this subject is specifically indebted to:

Adams, J. N (2007) The Regional Diversification of Latin. Cambridge

Linguistic terms

http://www.odlt.org/ this is a place holder until I find better, specific, links. Until then just ask or consult the brilliant Oxford Linguistics Dictionary.

Ancient Texts

Quintilian and Cicero (especially De Oratore, Orator, Brutus) are brilliant first stops for thinking about how language was used and consequently thought about in the Roman world, besides them there is:

Keil, G (1855-80) Grammatici Latin. Leipzig (8 volumes) = once an absolute obsession of mine, brilliant collection but if you’re new to this start here:

Lord, F. E (1894) The Roman Pronunciation of Latin. Boston, MA. Yes, it’s old and methodologically out of date etc etc κτλ κτλ but the point is, it’s an easy to read collation of sources. Remember! nullum esse librum tam malum ut non aliqua parte prodesset!



2 thoughts on “On Haccents in the Roman World

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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