Preserving Purity: Atticism and the Power of Linguistic Norms in Antiquity
From the Oxford comma to split infinitives, debates over language usage are anything but new. We are often told that certain forms are wrong—don’t say ain’t, never end a sentence with a preposition, always use whom in formal writing—even when it is not entirely clear why. These are classic examples of prescriptivism: the belief that language should conform to rules established by authoritative sources, and that deviation from these norms somehow results in improper or incorrect language.
In this post, I reflect on the power of prescriptivism and standardisation in antiquity, exploring how Roman and Greek grammarians dealt with linguistic norms and favoured the language of the classical period over the contemporary one. When does standardisation of language feel empowering, and when does it seem limiting? But where do you stand—do you follow the rules or challenge them?
Prescriptive grammars have long made it their mission to establish a normative standard for language: to protect it from fleeting trends and to provide a reliable guide to its fundamental grammatical structure. Traces of prescriptivism—and its cousin, linguistic purism—are still very much alive today. In a recent public statement, a group of English authors protested the inclusion of popular words into the English dictionary. Similarly, the French Academy has voiced its strong opposition to a government proposal aimed at preserving regional dialects.
But prescriptivism isn’t just about rules—it is about power. It is about deciding whose way of speaking and writing counts as the standard, and whose does not. All too often, the ‘standard’ is defined and enforced by a small, influential minority–typically male, and even more often, elderly. What starts as a grammar guide can become a cultural gatekeeper, reinforcing social hierarchies and shaping how people think, write, and even govern. Needless to say, prescriptivism and standardisation are closely intertwined, perhaps even just two sides of the same coin.
Take the story of L’Académie Française, or the French Academy, for example. Originally founded to promote the production of French literature, it hosted the famous salons of the 1620s and 1630s. During these gatherings, writers and literary connoisseurs discussed French literature, fought over the art of writing, and debated what role literature should play in society. But here is the twist: as the Academy came to represent the most authoritative voice on matters related to language, its objective gradually shifted from describing language to prescribing it. In fact, this became its sole purpose after becoming an official part of l’Institute de France (the French Institute). In this capacity, it established the first editions of the official French dictionary, wrote the first authoritative, prescriptive French grammar and became (perhaps excessively) protective of outside influences on the French language—captured in several laws that followed, most notably La Loi Toubon.
Fig. 2: Pietro Antonio Martini (1738-1797), ‘Exposition au Salon du Louvre en 1787’. Metropolitan Museum of Art, New York. Public Domain.
Of course, things were completely different in Antiquity. Neither Athens nor Rome had a system for language regulation or standardisation, at least not in the way we think about it today. It’s also important to remember that what scholars today consider to be ‘standard Latin’ or ‘standard Greek’ generally covers the language of only a limited set of authors: for Latin, this is typically Cicero for prose and Virgil for poetry; for Ancient Greek, scholars often look to Lysias, Plato and Demosthenes for prose and to Euripides, Sophocles and Aristophanes for poetry—though this can vary among scholars. Still, both ancient societies were faced with similar issues as the French Academy many centuries later: bilingualism and language contact influencing the language, the spread of the language as it was superimposed on nearby territories, the increasing use of the language as a lingua franca, and especially the lack of clear standardisation as far as the rules of the language.
While Latin grammarians and rhetoricians such as Cicero and Quintilian certainly discuss what makes ‘good Latin’ (~latinitas)[1], their focus tends to be more on rhetorical style—figures of speech, the structure of persuasive arguments, and the art of oratory— rather than strictly on grammar. Varro’s De Lingua Latina offers a more detailed linguistic look at Latin grammar, but even then, his work is mostly descriptive and explanatory, rather than prescriptive. This raises a key question: to what extent was Latin standardised during the Classical (1st c. BCE–1st c. CE) and Post-Classical periods (1st c. CE–ca. 200 CE)? And more specifically, what did ‘linguistic norms’ look like for Classical authors such as Catullus, who were routinely influenced by the language of their Greek predecessors?
As Bloomer (2017) points out, this issue is a tricky one. Ancient linguistic prescriptions were often anything but systematic. While Roman authors certainly had ideas about what counted as ‘correct’ language, they didn’t follow a formal, codified system. In short, they didn’t establish a true standard. Bloomer (2017: 68) notes:
A recurrent difficulty in trying to understand how the ancients understood, or applied principles of linguistic and stylistic correctness is their lack of systematicity. Principles were advanced, but from the first extant censor of Latin usage, the satirist Lucilius, through the literary figures of the Antonine age, the masters of Latin style will allege that they employ taste and judgment and not mere system. They will admire the old stylists and search the old literature for words but will not undertake a program to purify language by returning it to some pristine era. They will follow usage and yet admit the old expression. They will follow Cicero but avoid his fullness or Cato but avoid his aridity. The metaphorical terms employed likewise are not of much help. Further, there is a strong self-serving element in claims that the present author is the true judge or exemplar of Latinity. (…) And finally, the Roman writers, despite recycling a definition of Latinity from Varro and ultimately from Stoic theory of language, in fact proceed by making individual observations. (see also Clackson 2015)
Greek grammarians faced similar challenges and, at first, it seemed like they too lacked any sort of systematic approach to language. Yet, this changed drastically during the Second Sophistic (1st c. CE – early 3rd c. CE), a period when contemporary Greek writers such as Aelius Aristides, Dio Chrysostom and Plutarch suddenly, and quite profusely, mimicked the language of Classical Greek prose authors, and to a lesser degree, Classical poets. This movement, known as Atticism, saw a shift towards using the language of Classical Greek as a model for ‘proper’ usage. The trend was structurally supported by the lexica of Phrynichus (Selection of Attic Words; Sophistic Toolbox), Moeris (Atticist), Pollux (Onomasticon), Philemon (About Attic analogy/contradiction in words), the so-called Antiatticist, and the Pseudo-Herodianic Philetaerus.[2] These works systematically compared a contemporary word/construction with its counterpart in Classical Greek authors. Though the degree of explicitness and attitudes toward linguistic correctness varied among these lexica, one thing is clear: they preferred the Classical Greek forms over the more contemporary equivalents . Interestingly, the difference in usage wasn’t just about grammar—it also carried social weight. The lexica often used short phrases that conveyed judgments about the cultural sophistication or social status of speakers. For example, in the lexicon of Moeris we find:[3]
(1) ε 60: εἰ γάρ Ἀττικοί· εἴθε γάρ κοινόν.
‘εἰ γάρ ('if only') [is] Attic: εἴθε γάρ ('if only') [is] common.’
(2) η 6: ἤρεσέ με Ἀττικοί· ἤρεσέ μοι κοινόν
‘ἤρεσέ με ('it pleased me (acc.)') [is] Attic: ἤρεσέ μοι ('it pleased me (dat.)') [is] common.’
Both examples highlight the tension between the use of Attic writers (Ἀττικοί) and that of the κοινόν. A closer look at the corpus, though, reveals something interesting. In (1), the structure εἴθε γάρ, labelled κοινόν, appears only in high-register texts or passages where the register is intentionally elevated (cf. la Roi 2022: 218–219). On the other hand, for (2), the picture gets a bit more complicated. Attic writers used both the accusative and the dative as an argument with the verb ἀρέσκω. However, the accusative quickly fell out of use after the Classical period, making the dative the ‘standard’ case. So, in this instance , κοινόν refers to the more general use of the dative, regardless of the register of the text (see also Cluyse forthcoming).[4] The goal of these linguistic sources, however, remains clear: to elevate the Classical form above the contemporary one.
The linguistic impact of this Atticism movement on the grammar of the Post-Classical Greek period cannot be overlooked. In practice, this meant that words or constructions, which had fallen out of use for centuries, were revived in contemporary texts. A good example is the aforementioned verb ἀρέσκω, which had stopped being used with the accusative after the Classical period. Yet, during the Second Sophistic, the accusative construction reappears in authors like Plutarch, Galen, Dio Chrysostom, Aelius Aristides, Lucian, and Cassius Dio—often alongside the dative. The only exception to this comes from Hellenistic poetry, especially from Apollonius of Rhodes, thereby closely mimicking Homer’s language, who only used the accusative.
In essence, by advocating for the use of anachronistic or archaic constructions, the Atticists actively (and perhaps unconsciously) pushed back against the natural linguistic variation and change happening in the Greek language. Take the grammarian Aelius Dionysius, for instance. He notes that the Classical Greek aorist εἶπον ‘I said’ is preferred over the contemporary εἶπα, where the ending had undergone analogical levelling with other aorist forms ending in -α (Ael. Dion. ε 22). By pleading in favour of the aorist in -ον, he would thus reintroduce into the contemporary language a verbal form which in fact counters ongoing morphological changes.
Interestingly, non-literary texts seem less influenced by such Atticist trends, while the major works of key authors like the ones mentioned earlier show a distinctively high number of ‘Classical’ forms—often alongside contemporary forms. What’s fascinating, though, is that while these prescriptive lexica may have been trying to impose a standard, they also offer invaluable insights into the evolution of the Greek language. They tell us a lot about changes in e.g., nominal and verbal morphology, semantics, phonology, and more.
In the end, the Atticist movement of the Second Sophistic offers a fascinating counterpart to the more fragmented prescriptive tendencies seen in Latin grammars. Unlike for Greek, where the Atticists actively promoted the use of classical forms as a standard, Latin grammarians such as Cicero and Quintilian focused more on rhetorical style than enforcing a codified ‘correct’ form of the language. Still, in both cases, we see a desire to elevate certain forms of the language—whether it be the Classical Greek of Lysias and Demosthenes or the refined oratory of Cicero—as a way of asserting cultural and social ideals. What sets Greek apart, however, is the establishment of a more systematic, prescriptive tradition through the lexica and grammars of the Second Sophistic. These efforts were not just about language; they were about creating a sense of linguistic authority, with the power to shape social status and cultural identity. While the situation for the Latin language remains largely unsystematic (cf. Bloom 2017), the Greek counterpart shows just how powerful language standardisation can be, and how attempts to preserve the past can clash with the natural evolution of a language. Ultimately, both Latin and Greek sources offer us insight into the delicate balance between linguistic purity, cultural identity, and the constant push and pull of language change.
Main image credit: Detail: MS. Barocci 95: f. 1r (Digital Bodleian, Bodleian Library, Oxford, CC-BY-NC 4.0).
Notes:
[1] The concept of latinitas is complex and covers many distinct features. I refer the interested reader to the works of Smiley (1906); Díaz y Díaz (1951); Siebenborn (1976); Chahoud (2007); Clackson (2015), and Bloomer (2017).
[2] More on the fundamental features of Atticism can be found in Monaco (2021); Monaco, Machado & Bozia (2024) and Tribulato, Favi & Prauscello (2024). Early traces of this movement can already be detected in the first century BCE, particularly in Dionysius of Halicarnassus. In much the same way, one can also find Atticist lexica after the Second Sophistic e.g., the works of Hesychius, Stephanus of Byzantium, and Pseudo-Cyril (all 6c. CE).
[3] The Ca’ Foscari University of Venice recently launched an online database (DEA) to collect all citations from Atticist sources, arranged as individual entries. Each entry contains extensive quotations from prescriptive sources, alongside detailed discussions of the linguistic reality of the Atticists’ statements. The database can be accessed here: https://atticism.eu
[4] I am currently working on a more in-depth study on this argument structure alternation for ἀρέσκω in diachrony with Jóhanna Barðdal (Ghent University), likely to be published in 2026.
Further reading:
Bloomer, Martin. 2017. Latinitas. In Daniel L. Richter & William A. Johnson (eds), The Oxford Handbook of the Second Sophistic, 67–80. Oxford: Oxford University Press.
Chahoud, Anna. 2007. Alterità linguistica, latinitas e ideologia tra Lucilio e Cicerone. In Renato Oniga & Sergio Vatteroni (eds), Plurilinguismo letterario, 41–58. Soveria Mannelli (Catanzaro): Rubbettino.
Clackson, James. 2015. Latinitas, Ἑλληνισμός and Standard Languages. Studi e Saggi Linguistici 53(2): 309–330.
Cluyse, Brian. (forthcoming). ἀρέσκω, ἀπαρέσκω (Moer. η 6; [Hdn.] Philet. 29, 88). In Olga Tribulato (ed), Digital Encyclopedia of Atticism.
Díaz y Díaz, Manuel Cecilio. 1951. Latinitas. Sobre la evolución de su concepto. Emerita 19: 35–50.
la Roi, Ezra. 2022. The Atticist lexica as metalinguistic resource for morphosyntactic change in Post-Classical Greek. Journal of Greek Linguistics 22(2): 199–231.
Monaco, Chiara. 2021. The Origins and Development of Linguistic Atticism. Cambridge University Unpublished PhD Dissertation.
Monaco, Chiara, Robert Machado & Eleni Bozia (eds). 2024. Redefining the Standards in Attic, Koine, and Atticism. Leiden: Brill.
Siebenborn, Elmar. 1976. Die Lehre von der Sprachrichitigkeit und ihren Kriterien. Studien zur antiken normativen Grammatik. Amsterdam: John Benjamins.
Smiley, Charles Newton. 1906. Latinitas and ΕΛΛΗΝΙΣΜΟΣ. The Influence of the Stoic Theory of Style as Shown in the Writings of Dionysius, Quintilian, Pliny the Younger, Tacitus, Fronto, Aulus Gellius, and Sextus Empiricus. Bulletin of the University of Wisconsin, Philology and Literature Series 3(3): 205–272.
Tribulato, Olga, Federico Favi & Lucia Prauscello. 2024. Ancient Greek Purism. The Roots of Atticism. Berlin: De Gruyter.
Brian Cluyse is a doctoral researcher in the Department of Classics, Trinity College Dublin.
His thesis topic is ‘Beyond the break: unravelling the linguistic dimensions of the caesura in Latin elegiac poetry’ (co-supervised by Anna Chahoud and Boris Kayachev). His work is funded by Taighde Éireann – Research Ireland as part of the Enjambement in Latin poetry: prosody, pragmatics and word order project led by Boris Kayachev (grant number 22/PATH-A/10638).