Socially Awkward Data: Studying Ancient Sociolinguistics

Robin Meyer

“Don’t speak to me like that, I’m not one of your mates!” is, one would imagine, a reply many (British) teenagers have heard from their parents after addressing them in a manner they perhaps thought was perfectly ordinary. Such rebukes are meant, primarily, to remind them of their manners, but also – and more importantly for present purposes – of the fact that languages and language use are not monolithic behemoths but things that can vary greatly. Indeed, languages are ever-changing, their shape depending on a host of factors – time, geography, socio-economics, addressee, context, intention, the list goes on. Speaking to one’s parents is not like speaking to one’s friends – the addressee and context are different; similarly, expressing oneself in 2022 is not like doing so in 1972. We adapt our language to different situations, as Cicero made plain in a letter of 46 BC to his friend Papirius Paetus with regard to Latin:

verum tamen quid tibi ego videor in epistulis? nonne plebeio sermone agere tecum? nec enim semper eodem modo… privatas causas et eas tenuis agimus subtilius, capitis aut famae scilicet ornatius. epistulas vero cottidianis verbis texere solemus. (Cic. Fam. 9.21.1)

But, after all, what do you think of my style in letters? Don’t I talk with you in the vulgar tongue? Why, of course one doesn’t write always in the same style… Private causes and those that are of slight importance we plead in simpler language; those that affect a man’s civil existence or reputation, of course, in a more ornate style. But letters it is our custom to compose in the language of everyday life.

Even if one takes a moment to consider only the varieties of English that spring immediately to mind – British, Canadian, African American, Australian; Old, Middle, Elizabethan, to name only the broadest or perhaps most well-known – it becomes clear quickly enough that there is no such thing as “English” (or indeed “Language X” for that matter), no singular linguistic entity; what we call “English”, “French”, or “German” is just a convenient shortcut referring to the entirety of (or the perceived standard within) a continuum of language varieties.

It cannot surprise us, then, that this very same observation – that languages are really a collection of pluriform, heterogeneous communities of mutually intelligible varieties – should apply also to Latin, Greek, and most other ancient languages. This variability shows itself on various levels, as it does in modern languages. On a macro-level, we can observe quite readily differences of time, genre, and geography: the difference between the language of the Homeric epics, the Iliad and the Odyssey, and Platonic dialogues such as the Symposium is not unlike that between the works of Shakespeare and the plays of Arthur Miller, at least broadly speaking: they are separated by a few hundred years in time, a difference in genre, and a fair few miles.

Such differences can even be used for comic effect, as in Aristophanes’ Lysistrata, where the Spartan Lampito speaks in a (mock) Laconian dialect and thus sounds a little different to Athenian ears: ναὶ τὼ σιώ! The effect of this may be something like, “Yeth, by the Thwin Gods” (Ar. Lys. 90). On a micro-level, that is at smaller temporal, literary, or geographic distances, differences can still be observed, for instance with regard to the lexicon – one might think of the numerous words for “bread roll” in varieties of British English, or for the “core of an apple” in varieties of German.

Next to geographical differences, we can equally see intergenerational variation: what was “groovy” in the 1970s and “cool” or “rad” in the 1990s is perhaps “fab” nowadays. Even in individual speakers, language use can change noticeably over the course of a lifetime, as the vowels of Queen Elizabeth II illustrate. This kind of micro-variation would have been present and observable in Greek and Latin as well. In cursing, for example, female and male characters in Latin comedy tend to behave differently: men swear by Hercules (hercle, mehercle), whilst women prefer to appeal to Castor (ecastor, mecastor). This tendency does not come without complication, however, as we’ll see below. More generally, in trying to discover the extent of sociolinguistic variation in ancient languages, we must face some serious issues.

In order to be able to tell who spoke how to whom in what setting, we essentially need two things: data and context. That is, we need utterances of one sort or another (recordings, interviews, letters, books, diaries, etc.) and the relevant information about the speakers and addressees involved (the time of the utterance, their age, sex, gender, socio-economic status, location, education, relationship to each other). In our quest to figure out what linguistic traits are purely individual and which are linked to these categories, these basic demands already limit what we can know about the sociolinguistics of ancient languages, for we have a limited set of texts and limited information about the speakers or writers involved.

In practice, this means that we cannot set out to find more data, e.g. by recording conversations, asking for letters, looking at new books, or interviewing native speakers – barring the invention of time travel, of course – and that in analysing what we have, we often lack basic information: biographical data about the addressees of a Roman soldier’s letter, for instance. While this means that the amount and degree of detail we are able to gather from analysing ancient Greek, Roman, or other cultures’ texts has its limits, there are still a fair number of things we can seek to find out. Let us have a look at just one example in a little more detail: how Greek-speakers used ‘discourse particles’ – little words that help clarify the tone and flow of speech – and whether or not they are category-dependent.

For this purpose, let us think about Athenian comedy, which gives us as close an approximation to everyday speech as we are likely to get in ancient sources. Menander, an Athenian comic poet of the late 4^th century BC, is the most well-known representative of New Comedy, a genre in which quotidian situations and characters were the basis of plot and humour. As in Roman comedies – some of which were based on Menander’s work – we find characters from all walks of life: men as well as women, old and young, free and enslaved. Their speech is peppered with these discourse particles – little words like English “well”, “quite”, “I mean”, “you know”, and (in)famously “like”. The usage of such words could be completely up to the individual, or could show certain tendencies, for instance being more typical of one or another kind of speaker – and it is exactly that question we are trying to answer.

One instance of such a word in Greek is γε (ge), “indeed” or “at any rate”. The particle γε often occurs to emphasise a word or phrase (e.g. “Sprouts I cannot stand!”) or to specify a reference (“I personally cannot stand sprouts… but you might like them.”). To figure out whether this word is used more or less equally by everyone or whether there are sociolinguistic differences, we need to count how often it occurs and in what circumstances (speaker, context, etc.). In the surviving comedies of Menander, we find 162 occurrences in contexts that we can evaluate – that is, excluding fragments and broken lines. The category that we can distinguish most clearly in comedy is gender/sex, since characters’ ages are referred to only inconsistently and their social standing or indeed liberty is not always clear either.

What we find at first glance is that women seem to use the particle γε almost twice as frequently as men, suggesting it might be a marker of female speech. But, as tempting as this finding may be, it is somewhat problematic: the number of lines of female speech we have is less than 10% of Menander’s work, meaning that we know much less about it than about male speech. To complicate matters further, a large number of instances where γε is used fall into specific contextual patterns, e.g. particularly emotional situations. The crassest example might be that of Chrysis, a young mother and mistress of Demeas, who owing to a set of misunderstandings is evicted by the latter and now fears for her and her child’s future, exclaiming: τάλαιν᾽ ἔγωγε τῆς ἐμῆς τύχης (“Oh dear me, how am I suffering from my fate!”, Men. Sam. 398).

In a differently dramatic situation, Simiche (an elderly slave) is in despair over a lost bucket, and shows this as her master Knemon approaches in anger: περιτρέχων ταύτην πάλαι ζητεῖ βοᾷ τε – καὶ ψοφεῖ γε τῆν θύραν (“now he runs around looking for this thing under curses and – now he’s bloody well rattling the door”, Men. Dys. 586–7). This prevalence in emotional contexts applies not only to female speech, it turns out, but also to male speech.

Finally, we need to keep in mind that Menander was a comic poet, an entertainer, who painted a literary picture of society and its more or less comical members. He used language for a purpose and employed particular characters or character types in stereotypical situations. The high incidence of the particle γε in female speech is therefore not necessarily a sign that women used this particle more frequently. But it does indicate that Menander portrays women more commonly in particular situations, specifically emotional ones, in which they might use this particle. Rather than a marker of gender/sex, γε seems to mark discourse situations. This chimes well with other observations we can make concerning the use of this word, e.g. its prevalence in answers to questions and its frequent co-occurrence with personal pronouns.

A similar problem of distribution arises when we turn to cursing once more. In the early comedians Plautus (254–184 BC) and Terence (185–159 BC), the invocation edepol, “by Pollux!”, seems to be used quite differently – more often by men in Plautus, more often by women in Terence. Was there a change in the use of this word? Did it skip from one gender to another? As in the case of Menander above, frequency can be misleading. Edepol has a greater absolute frequency in Terence’s female characters, but they also curse more than four times as frequently as male characters, bringing their relative frequencies almost on par. While this suggests that edepol has not undergone any changes, we can induce that, for Terence, cursing in general was a marker of female language. Whether that is a true reflection of reality or a comic stereotype is a different question.

What then do we learn from these small snippets of engagement with ancient sociolinguistics? For one, that the data we have available need to be looked at with great care and in their context: literary uses may skew them and other explanations may be available. Second, that we are limited by our sources: we use comedy as an approximation of everyday language, even though it is still a literary genre. Third, and perhaps most importantly, that ancient languages are subject to the very same variation of style, register, and context as modern languages are – it’s just a little harder to figure out the details that matter.

---

Robin Meyer is assistant professor of historical linguistics at the Université de Lausanne and works on Latin, Greek, Armenian and Iranian syntax and language contact. He can be very particular about ancient language use.

---

Further Reading

Jerome Moran, “Standard and Non-standard Latin,” Journal of Classics Teaching 19 (2018), available here.

Eleanor Dickey, “Forms of address and conversational language in Aristophanes and Menander,” Mnemosyne 48 (1995) 257–71.

David Bain, “Female Speech in Menander,” Antichthon 18 (1984) 24–41, available here.

J.N. Adams, “Female Speech in Latin Comedy,” Antichthon 18 (1984) 350–73, available here.