There are many significant differences between Latin and its linguistic descendants, the Romance languages. One that stands out from the rest is grammatical gender.
Latin has three genders for its nouns: masculine, feminine and neuter. However, in French, Spanish, Portuguese, Catalan, Italian and all the other many Romance varieties that lack official use and support, there are only two genders.
- For example, the French nouns chat, fromage and garçon (‘cat, cheese, boy’) belong to the masculine gender, while femme, ville and maison (‘woman, city, house’) are feminine.
- Of the Latin nouns dominus, gladiātor, virtūs, porta, caput and argentum (‘master, gladiator, virtue, gate, head, silver’), the first two are masculine, the second two are feminine and the final two are neuter.
It seems that the Latin neuter has been lost over the course of history. The nouns that bore the feature have been reanalysed as nouns of the surviving masculine and feminine genders, thereby carving the neuter up according to where each of its words would fit best.
But why? And how? And when? How can a language lose something as fundamental as a whole gender? How can its absence be so uniform across the Romance family? This article is an introduction to the decline and death of the Latin neuter gender – as well as its many ghosts that continue to haunt its daughter languages.
Pre-Latin and Posh Latin
The neuter gender is very old indeed, arguably older than the other two genders. Through the systematic comparison of ancient languages, we believe that Proto-Indo-European, in its oldest reconstructable state, had only two gender-like categories for its nouns: animate and inanimate (that is, self-moving and not-self-moving things). It was the inanimate that became the neuter gender in its descendant languages, while the animate would later split into two further genders. This later three-gender system is what Latin inherited.
If you’ve studied any Latin (whether it be the Classical Latin of Cicero or Caesar, or the language’s later life in medieval and modern times), the three-gender system seems strong. Roman grammarians recognised this; following Greek practices, they taught that a noun could be masculīnus, feminīnus or neuter (literally ne-uter ‘not either’). Neuter nouns abound across the traditional groupings of Latin nouns, known as its declensions. Students today must learn that there can be neuters of the second declension like bellum ‘war’ or templum ‘temple’, of the third declension like nōmen ‘name’ and tempus ‘time’, and of the fourth declension like cornū ‘horn’ and genū ‘knee’.
However, the Latin that has survived until today reflects only a fraction of the life of the language, and that fraction is for the most part the prestigious Latin of the Roman elite. The power of prestige cannot be underestimated; the norms of standard Latin would have had enormous influence on all of society, acting like a brake on the speed of change – at least in the written language, if not the spoken language too. Our sources for ‘sub-elite’ Latin are rare and offer little more than glimpses into linguistic phenomena outside the villas and senate houses. Graffiti, such as the inscriptions of Pompeii, and quotations in prose texts are therefore treasured for their vulgarity (in all senses of the word) and are subject to much analysis.
To trace the decline of the neuter, we can begin with a theoretical principle: the process must have begun early, at a time when Latin was still a unified language. If we do not accept this as a first premise, it becomes very difficult to explain why the neuter is absent from the whole Romance family – why its decline was so uniform. This is not to say that the neuter had died completely when Romance languages first emerged, only that the seeds had already been sown and the die already cast.
The limited evidence seems to back this up. An oft-mentioned example of the change is found in the Satyricon, a wonderful and quite unique piece of fiction written by Petronius in the first century AD (a date comfortably within the era of Classical Latin). Perhaps the most famous scene in the story is Trimalchio’s dinner, a lavish party hosted by the freedman Trimalchio. Because of the guests’ backgrounds and the idiosyncrasies in their speech, the evening’s conversation is thought to include authentic features typical of sub-elite Latin. For example, Dama, one of the guests, says:
“Et mundum frīgus habuimus. Vix mē balneus calfēcit. Tamen calda pōtiō vestiārius est.“And we’ve been having a real cold spell. My bath has barely warmed me. A hot drink though is as good as a wardrobe. (Satyricon 41)
This piece of small talk includes the word balneus ‘bath’. Here the ending -us identifies the balneus as masculine in gender, yet the word is typically found as balneum, a neuter noun ending in -um. In brief, a neuter noun has become masculine.
The fact that this early and rare example of the change was included in the low-register, colloquial speech of Trimalchio’s guests’ conversation is telling. Who knows how common it was to say balneus at this time? It’s hard to say with any certainty, but Petronius, who wrote this scene with all the necessary reflection and deliberation of an author, must have thought that the choice of word sent the right message.
Other examples of this un-neutering have been proposed and argued over in various sources for Classical Latin. Some even gained popular use in the prestige language; the word opera is seen used not as a neuter plural noun (meaning ‘works’), but as a singular feminine noun in this line by Cicero:
“quīdam nimis magnum studium multamque operam in rēs obscūrās atque difficilēs cōnferunt“Some apply too much study and too much work to obscure and difficult matters. (de Officiis I.19)
This reanalysis of opera is worth noting, because it offers us a chance to consider how effortless and reasonable a change in a word’s gender might be. The ending -a is found not only at the end of neuter nouns, but also in countless feminine nouns in Latin, such as fēmina ‘woman’ or tabula ‘table’. If its meaning also permits it, there is nothing in the shape of the word to stop opera transforming from a neuter third-declension noun into a feminine first-declension noun. It is likely that there was much more variation than we presume, with some alternative forms gaining widespread and long-term success, while others remained limited to smaller areas. After all, Latin was a living language like any other.
Moreover, the loss of the neuter gender coincided with the equally important and slow loss of Latin’s grammatical cases and the different endings that expressed them. As cases and endings merged and faded away, there was even less to distinguish a neuter noun from those of another gender. For example, the Latin noun dōnum ‘gift’ feels strong in its neuter gender as long as the final -m is present and clear enough to mark the word as categorially different from a masculine noun like ventus ‘wind’. However, if the -s, -m and other case endings of these words became muddled and faded away, we would be left with ‘donu‘ and ‘ventu‘, which we would then reasonably group together under the same gender. This is indeed what happened to Latin dōnum and ventus; they became masculine don and vent in French, masculine don and viento in Spanish and masculine dono and vento in Italian.
The re-gendering of neuter nouns had wholesale effect. A fair few plural neuters became singular feminine nouns by means of their common ending -a. We see this in Latin mālum ‘apple’, the plural of which is māla; it was māla that became a feminine noun in its own right, giving Italian its word for ‘apple’, mela. However, the majority became masculine; as we can see in the example of ventus and dōnum, second-declension neuter nouns changed into masculine nouns, thanks to the vowel in their final syllables that they had in common.
This vowel went on to become the defining feature of masculine singular nouns in Romance languages like Spanish; the Spanish words libro ‘book’, hermano ‘brother’, cielo ‘sky’ and negocio ‘deal’ all end in -o and are now all masculine, even though two come from Latin neuter nouns that once ended in -um. Neuter nouns of other shapes and declensions also joined this new and improved o-group of the Romance two-gender system. For example, Latin tempus ‘time’ and corpus ‘body’ (neuter nouns of the third declension) gave way to the new system and became Italian tempo and corpo, both masculine.
Dying… But Not Dead?
So, end of story? Neuter nouns were shared out between the masculine and the feminine, and the neuter gender died out as a grammatical category?
That’s the general idea, but it’s by no means the whole truth. Traces of the neuter gender can still be found across the Romance family today – grammatical ghosts that can go unnoticed or undiscussed. Specifically, the traces to be herein discussed find themselves in Romanian, Italian, Asturian and Romance languages of southern Italy like Neapolitan. With each one, we will see the ways in which the Latin neuter gender still has an effect on the language, but also how these effects do not in fact demonstrate its continued existence.
Lovers of Romance may have noticed, perhaps with some concern, that Romanian has not yet been mentioned. Romanian is also a member of the Romance family and seems at first glance to buck the gender trend. It is traditionally said that Romanian has three genders, like Latin. On one level, this is true, as all Romanian nouns do indeed conform to one of three patterns. Nouns of the ‘neuter’ gender also preserve much of the status and appearance of the original Latin neuter. For example,
- bărbat (‘man’) is masculine. Its plural is bărbați.
- casă (‘house’) is feminine. Its plural is case.
- măr (‘apple’) is neuter. Its plural is mere, and it comes from the Latin neuter noun mālum.
However, although three categories of noun can be identified in Romanian and labelled with the same three terms used for Latin, the status of the Romanian neuter is disputed. The issue is that these nouns do not behave like a fully-fledged gender. Importantly, there are no dedicated neuter endings for adjectives; instead, when a Romanian ‘neuter’ noun combines with an adjective, that adjective will be either masculine or feminine.
- un bărbat frumos (‘a beautiful man’) – the adjective has a masculine ending to match the noun
- o casă frumoasă (‘a beautiful house’) – the adjective is feminine
- un măr frumos (‘a beautiful apple’) – the noun is neuter, but the adjective and indefinite article associated with it are masculine, as with bărbat
To be precise, Romanian neuters take masculine adjectives when they are singular, but feminine adjectives when they are plural. Compare case and mere here:
- doi bărbați frumoși (‘two beautiful men’)
- două case frumoase (‘two beautiful houses’)
- două mere frumoase (‘two beautiful apples’)
The fact that these nouns must use masculine and feminine adjectives goes against the independence of the neuter gender. As Hockett (1958) says, “genders are classes of nouns reflected in the behaviour of associated words”; without dependent words that are similarly neuter, Romanian ‘neuter’ nouns present only the illusion of a separate category. Because of their use of both feminine and masculine adjectives, linguists refer to these nouns as ‘ambigenous’ – that is, they belong to both of the two genders. This view may surprise some people, seeming contradictory to what has traditionally been taught about Romanian. So, allow me to put it another way: if Romanian neuter nouns do not effect the gender of their associated words, and are only identified as neuter because of their own inflected forms, surely this is only a matter of noun declension, not of a separate grammatical gender?
Thus, in Romanian, yet again, the neuter seems to have died out.
Ambigenous nouns can also be spotted across the Romance languages of Italy, including a handful in Standard Italian. As in Romanian, most of these nouns inherit their special status from the Latin neuter gender and again show a division according to whether they are singular or plural. For example, the Italian nouns uovo ‘egg’, labbro ‘lip’ and braccio ‘arm’ do look and indeed are masculine – when they are singular. While the majority of masculine nouns have plural forms ending in -i (e.g. orso ‘bear’, orsi ‘bears’), uovo, labbro and braccio have the plural forms uova ‘eggs’, labbra ‘lips’ and braccia ‘arms’. These are now treated as feminine by Italian, but really it was the neuter gender that provided their plural -a ending. Compare the modern words with their Latin ancestors: ōvum, lābrum and bracchium (in the plural: ōva, lābra and bracchia). For example, ‘I eat a delicious egg’ in Italian is:
- mangio un uovo delizioso
In this, uovo and all its associated words are masculine. If you want to express that you eat a multitude of tasty eggs, all those words become feminine in gender:
- mangio molte uova deliziose
The third and final challenge to the death of the Latin neuter comes from the Romance languages of the south of Italy and north of Spain. Distinct but unofficial Romance languages like Neapolitan (spoken in and around Naples) and Asturian (spoken in Asturias, a part of Spain) seem to have definite articles (equivalent to English the) for three genders, not two. Asturian has el for masculine nouns, la for feminine and lo for those nouns termed ‘neuter’. In Neapolitan and similar varieties of that part of Italy, we can find neuter definite articles like ‘o, lo, u and ru that stand alongside the articles for masculine and feminine nouns.
As an aside, it’s important and interesting to qualify that in modern-day Neapolitan, while there are seemingly only two definite articles, namely ‘o and ‘a, we find that ‘o sometimes triggers raddoppiamento sintattico on the noun that follows it. This is a fascinating phenomenon, a hallmark of south Italian Romance. Essentially, ‘raddoppiamento sintattico‘ means that the first consonant of the noun is doubled for some grammatical reason. For example, in some contexts, Roma ‘Rome’ will be pronounced Rroma in Neapolitan; to go ‘to Rome’ is to go a Rroma, with a doubled or ‘geminated’ consonant. If you’re interested, this wonderful thing can in fact be found in what is probably the most famous example of the Neapolitan language: the song ‘O sole mio (‘My sunshine’), specifically in the words “‘e llastre d”a fenesta toia” (which rather prosaically translates to ‘your window panes’).
Since the definite article ‘o causes raddoppiamento on only a limited group of nouns, this is taken to be the continuation of a distinct neuter definite article and gender. For example,
- ‘o libro (‘the book’ ) – masculine in gender
- ‘a casa (‘the house’) – feminine
- ‘o ffierro (‘the iron’) – neuter, with raddoppiamento, from the Latin neuter noun ferrum.
What I find most interesting is that this doubling of consonants seems to be the way that the language compensated for the loss of another consonant. This idea goes back to the Latin word ille, which meant ‘that’ and is the origin of definite articles like French le, Spanish el and Neapolitan ‘o. The neuter form of ille was illud, and when the final -d of illud disappeared, this gave rise to the double consonants in Neapolitan neuters like ‘o ffierro.
I digress. The key question is whether these supposed neuter definite articles demonstrate the existence of a neuter gender. Again, sadly, the answer is no. These definite articles and the nouns they are used with (no matter what their origins) seem not to constitute a neuter gender, but rather to signify that the noun is something abstract or a physical mass – that is, an indivisible substance, like water or gold. It’s not uncommon for mass nouns to have their own specific grammatical features; for an English example, it’s unnatural to combine them with the indefinite article and say things like a water or a gold. Because Neapolitan ‘o ffierro ‘the iron’ and Asturian lo lleche ‘the milk’ refer to a metallic and a liquid substance, they are marked as such by the ‘neuter’ definite article.
Why does this contradict the existence of a neuter gender? Because we find this “mass gender” (Harmon 2007) used with nouns that were not neuter in Latin, and because nouns in these languages can appear both with and without the feature, according to whether they have a mass meaning or not. This is therefore not a grammatical feature on the same level as the masculine and feminine genders; instead, the mass gender is a feature of nouns determined by semantics – to what sort of thing the noun refers.
In Neapolitan, we can find both ‘o pane ‘the loaf’ (masculine) and ‘o ppane ‘the bread’ (neuter, with raddoppiamento), in which the neuter noun refers to a general substance and the masculine to an individual thing. The noun itself comes from Latin pānis, originally a masculine word, and so has joined this group of Neapolitan mass nouns through the new semantic function of the neuter. Likewise, the difference between Asturian el pelu (masculine) and lo pelo (neuter) is that one concerns a single hair, while the other, the neuter, refers to hair in general.
It does seem that the neuter gender of Latin is indeed dead. As the Romance family emerged in the early medieval period, some of the new languages thoroughly divided up the neuter between the masculine and the feminine. This is the case with French, Spanish, Portuguese and Italian. Other languages, most notably Italian and Romanian, preserved the illusion of a third category for some of their nouns, but really these are ambigenous – caught between two genders and changing according to quantity. Lastly, languages like Neapolitan and Asturian have transformed the Latin neuter into something new. For them, it has come to work alongside grammatical gender, distinguishing countable nouns from uncountable.
There is more that could be mentioned, but I think this is enough to demonstrate the variety of the Romance family and the many phenomena that can emerge from the dissolution of something as big as a grammatical gender. The Latin neuter may be dead, but it didn’t go out without a fight.
- Carretero Garcia, P. 2017. Agreement in Asturian. In: Butt, M., & King, T. H. (eds.): Proceedings of the LFG’17 Conference. University of Konstanz. 188–208. Stanford, CA: CSLI Publications.
- Colacino, C. 1999. Neapolitan: An Introductory Course. Available here: http://www.duesicilie.org/OLDSITE/Neapolitan5.html
- Harmon, S. E. 2007. Gender in the Romance Languages: An Evolutionary Approach. The University of Texas.
- Hockett, C. F. 1958. A course in modern linguistics. New York: Macmillan
Many thanks to Arnold Platon (@Arnold_Platon) for his help with the Romanian examples.
Distribution or reproduction of this post is free for non-commercial use with attribution.