Gender Diversity in Greek and Latin Grammar: Ten Ancient Discussions

Tom Hendrickson
Jan 30, 2020 · 9 min read

Latin obsessively genders its speakers. From pronouns to choice of name and use of adjectives. Grammatical gender is actually complex and shifting, a fact that Greek and Roman grammarians were highly conscious of.

As a Latin teacher, I’ve been thinking a lot about how I present grammatical gender to my students. Latin obsessively genders its speakers. From pronouns to choice of name and use of adjectives, it’s one thing if you identify as masculine or feminine, but Latin presents real questions for non-binary speakers. In addition, there is a chauvinism inherent in Latin grammar, like in the way that mixed-gender groups are referred to as masculine. There is a similar chauvinism in the Latin pedagogical tradition: the very ordering of adjectival forms as masculine/feminine/neuter (-us/-a/-um) presents a hierarchy of being masquerading as a lexicographical convenience.

On top of that, I worry that our modern pedagogical practices implicitly present students a highly essentialized vision of gender: every noun has one gender exclusively, rigid and unchanging.

Yet grammatical gender is actually complex and shifting, a fact that Greek and Roman grammarians were highly conscious of. They debated whether grammatical gender was inherent in the things themselves or a product of human convention. They devised a technical vocabulary to articulate the difference between grammatical gender identity and grammatical gender expression. They argued about whether their language had three genders — as we teach — or perhaps only two, or maybe five, or even as many as seven, if you count “fluid” and “uncertain” and as distinct categories (and they did!).

Grammatical gender, at its most basic, is a system of noun classes whose forms agree with other inflected words, like adjectives. French nouns have two genders, German three, Chechen six, and Ganda ten. Wikipedia has a handy list of languages by type of grammatical gender.

The fact that we think of these classes as “genders” at all is a product of historical circumstance rather than linguistic necessity. There is nothing inherently feminine about aqua (“water”) or masculine about rivus (“stream”). From a purely morphological perspective, we could dump the very concept of grammatical gender and call these groups “noun classes,” which is apparently the usual term for them in Bantu languages.

Yet there is more to a language than its morphology. Greeks and Romans believed that their words had gender, and Anthony Corbeill, author of Sexing the World, has shown that this belief affected the way that they thought about their words, and about their world. The gender diversity of the Greek and Latin languages is not something I encountered in my own training in Greek and Latin. I suspect that the erasure of this gender diversity is not a matter of pure chance, but that’s a topic for another day.

As a resource for other interested teachers, I’ve put together the below list of passages in which ancient Greek and Roman grammarians discuss gender. I’ve also included links to open-access or public domain versions of the texts, just in case you don’t happen to have your copy of Ammonius Hermiae with you at the moment.

Let me stipulate that I don’t mean to claim any particular expertise or authority on matters of gender, whether ancient or modern. I’m just a Latin teacher trying to foster an inclusive classroom environment. I also don’t mean to suggest we should be uncritically emulating the views of ancient Greeks and Romans on gender (or really on anything). Rather, I put together this list together simply to bring attention to the gender diversity of the Greek and Latin languages, and to the ways that Greeks and Romans conceptualized that diversity. The passages on this list come largely from Marcin Kilarski’s book Nominal Classification, which I highly recommend.

1. A Fragment of Protagoras, Earliest Reference to the Three Genders:

The names for the genders go back at least to Protagoras, the 5th century BCE philosopher. Aristotle (Rhetoric 3.5.5) notes that Protagoras divided words into three types/genders (γένη): males (ἄρρενα), females (θήλεα), and things (σκεύη). Greek text available here and English translation here.

2. Aristotle on the Third Gender as “In Between,” and Whether Grammatical Gender is Formal or Natural

Aristotle (fourth century BCE) had a different name for the three genders (Poetics 21): masculine, feminine, and “in between” (μεταξύ). Aristotle also differed from Protagoras in seeing gender as a formal characteristic of the noun (that is, of its endings). Whereas Protagoras’s terms, in particular the neuter as “things,” suggested that the category was based on the nature of the object in question, Aristotle saw it rather as a linguistic convention. Aristotle expands on his disagreement with Protagoras in Sophistical Refutations 14 and 32. The text of Poetics 21 (1458a) is available here in Greek and here in English; Sophistical Refutations 14 (173b17) and 32 (182a7) are available in facing Greek/English pages here and here.

3. Dionysius Thrax on the Third Gender as “Neither,” and on the Possibility of Five Genders

Dionysius Thrax, writing in the second century BCE, differed from Aristotle in calling the third gender “neither” (οὐδέτερον) masculine nor feminine. Romans later translated this term as neutrum, which we have bizarrely borrowed into English as “neuter” rather than translating it as “neither.” Dionysius might have taken this term from earlier writers in the Alexandrian scholarly tradition, like Aristophanes of Byzantium. Dionysius also writes that some people add two more genders to the traditional three: a “common” gender (κοινόν, later taken into Latin as commune) and a “communal” gender (ἐπίκοινον, Latinized as promiscuum but sometimes just transliterated as epicoenum). The “common” gender is for words like “horse” (ἵππος), which can take modifiers of either gender. The “communal” gender is available for any creature to use, no matter whether it is male or female. So a “swallow” (χελιδών), for instance, will always take feminine modifiers, even if the swallow in question is male. This discussion is in his Ars Grammatica 12; Greek text is here and English translation here. (Note: this English translation differs from the Greek in chapter numbering, and in the text itself.)

4. Sergius on the Difference between the “Common” and “Communal” Genders

The fifth-century CE grammarian Sergius expanded on the difference between the common and communal genders in his commentary on Donatus. A word like canis (“dog”) is common gender because you can say “good dog” as canis bonus or canis bona, whereas a word like aquila (“eagle”) is communal because you can only say aquila bona, even if the eagle happens to be male. Sergius hazards an explanatory theory: the name for an animal will be common if you can tell its sex by looking at it (like a dog) and communal if you can’t (like an eagle). Yet he has to immediately admit his theory doesn’t work, because “camel” and “elephant” are communal even though you can definitely tell from looking whether they are male or female. I provide here a link to the Latin in Keil’s Grammatici Latini, vol. 4 (start at the very bottom of page 493). The link is just Latin; I’m not aware of an English translation.

5. A Fragment of Varro on Gender and Genus

The word genus in this context is certainly not just “type,” but it also doesn’t quite match with our modern concept of gender. (Though we shouldn’t expect any ancient concept to line up perfectly with modern equivalents.) Still, it’s clear that Greeks and Romans saw grammatical gender as an amorphous conglomeration of characteristics and tendencies related to, but not wholly defined by, biological sex. The sole attempted definition of grammatical gender is Varro’s etymological reasoning genus a generando (“gender from generating” — fragment 245 Funaioli). The comment only survives as quoted in other authors, which leaves open the question of whether Varro is suggesting that only animate objects can have gender, or that gender is a function of heteronormative male/female procreation. The version of this quotation preserved in Sergius (quoted above in #4) seems to suggest both interpretations. In favor of the first interpretation, Sergius adds that only humans and animals have a gender “by nature” (a natura), while inanimate objects have a gender rather “by tradition” (ab auctoritate). In favor of the second, he concludes that there are really only two genders: masculine and feminine. Yet even the two-gender paradigm allows for multiple permutations, since he adds that the neuter gender is neither masculine nor feminine, the common gender is both, and the communal gender can be either. The link here is the same as above, but start on the bottom of page 492.

6. Varro on Gender Identity and Gender Expression

Varro’s (principally) two-gender paradigm also allows for variability within the masculine and feminine genders. Take first-declension masculine nouns like agricola (“farmer”): the inflections look feminine, but the word agrees with masculine adjectives. Varro writes that in cases like these, the figura is feminine and the materia masculine. I think that we can most accurately translate figura here as “gender expression” and materia as “gender identity,” since Varro explains the distinction as being like a man who wears feminine clothing, or a woman in masculine clothing. Every translation risks anachronism, and particularly so with social and cultural constructions. Yet in rendering figura and materia as “gender expression” and “gender identity,” I don’t think we’re imposing modern ideas on ancient Rome. In fact, this seems to be a case where English terminology has only recently caught up with Roman notions. The distinction between materia and figura are discussed in De Lingua Latina 10.11 (Latin and English available here); the examples of a man in feminine dress and a woman in masculine dress are in De Lingua Latina 10.27 (Latin and English available here).

7. Varro on Gender Change

Varro believed that grammatical gender was a matter of social convention and experience rather than nature, which means that gender can also change as our conventions and experiences change. As an example, Varro writes that the word “dove” (columba) was once communal, but that after doves were domesticated the males and females began to take on new roles, and the word’s gender then changed to reflect that differentiation (columba and columbus). The passage is De Lingua Latina 9.56, Latin and English available here.

8. Priscian on the Seven Genders of Latin

The sixth-century CE grammarian Priscian provides the most extensive discussion on grammatical gender, covering thirty pages of Keil’s massive Grammatici Latini (vol. 2, pp. 141–171). Priscian writes that while there are “two principal genders” (genera … principalia sunt duo), there are also five additional genders: neuter, common, and communal, as well as two categories of nouns whose gender can vary. The first group comprises nouns whose gender isn’t quite certain, like finis (“limit”) and margo (“edge”). Priscian calls this group dubia genera, which we might translate as “uncertain” (“questioning”?) genders. The second group is composed of nouns that can happily be more than one gender, like filius/filia (“son”/“daughter”). He calls this group mobilia genera, which we might translate as “fluid” genders. Priscian sees each category not as a monolith but as a group of genders, since nouns can be “uncertain” or “fluid” in more than one way. (For instance, he outlines five distinct types of fluid genders, and he cites an entire work De Dubiis Generibus by Probus.) The Latin text is here; I do not know of any English translation.

9. Ammonius Hermiae on the Rationality of Grammatical Gender:

Ammonius Hermiae, a commentator on Aristotle from the fifth century CE, articulates the most extreme version of the perspective that grammatical gender springs from the nature of things themselves. In making the case, Ammonius tries to rationalize the gender of all words, and in doing so he puts forward a pervasive heteronormativity as an explanatory tool. As an example, Ammonius writes that seas (θαλάσσας) and bays (λίμνας) are feminine while rivers (ποταμούς) are masculine, because whereas seas and bays are penetrated (ὑποδοχάς) by rivers, the rivers are doing the penetrating (ἐμβάλλοντας). The Greek text can be found here in the Commentaria in Aristotelem Graeca (vol. 4, part 5, pp. 35–36). I’m not aware of an English translation.

10. Sextus Empiricus on the Arbitrariness of Grammatical Gender

While Ammonius presents the extreme version of the “natural” view of gender, Sextus Empiricus (second century CE) presents the most extreme version of the view that grammatical gender is a product of human convention. After all, he writes, if grammatical gender corresponded to something in the nature of things themselves, then we should expect the words for each thing to be the same gender in each language, when in fact the gender of words can vary even among the different Greek dialects. The relevant passage is Against Grammarians 142–153; Greek text with English translation here. (The first page of the Greek text is upside-down, but it’s the only free online version of this text that I know of.)

Tom Hendrickson teaches Latin and English at Stanford Online High School.


Towards a more equitable Classics