Blog

read it, tweet it, pin it, love it

Ask a linguist: Why does French have grammatical gender?

As a teacher of French, I’ve heard a lot of grievances about the French language, as though if enough people complained, I (of all people) would suddenly do something about it. One of the most fundamental of these gripes is about French’s grammatical masculine/feminine distinction; the argument goes something like this: “Giving nouns genders is arbitrary (OK, most whiners don’t use the word ‘arbitrary’, but that’s what they mean) and requires effort without serving any apparent purpose.” Actually, that usually just comes out as, “Waaah, learning genders is hard.” Just observing that isn’t particularly admirable, but then some students ask this question: “Why would that even exist in a language?”

It’s interesting enough to be worth a discussion, I think. Why does French make use of an arbitrary distinction between masculine and feminine nouns, when doing so appears to give no benefit to speakers and is cognitively costly for learners?

Before we can look at some possible explanations, let’s talk linguistic typology (please? this is my favorite stuff¹).

First: grammatical gender is different from natural gender. Natural gender describes actual gender distinctions in living beings, like humans or lizards or whatever. English makes some distinctions of natural gender (here are some examples on Wikipedia). Grammatical gender is part of a broader phenomenon of noun classes which exist in many of the world’s languages.

Noun classes divide up a language’s nouns into groups, often with different agreement patterns on related adjectives, pronouns, verbs, and even prepositions. (Example from French: the grammatically feminine Rose (as a name) takes a feminine agreement on its adjectives, so “Rose is mean” would be Rose[+feminine] est méchante[+feminine].) French is pretty tame about agreement; some Bantu languages have up to 16 noun classes and they have to agree their noun classes on pretty much everything else in the sentence (including prepositions). Are you still whining about French?

Sometimes noun classes are logically connected to some idea; for example, in Tshiluba (a Bantu language of the Congo) there’s a noun class that means the noun in question is small. Some nouns are always in this class (like peanuts) and other nouns can be taken from their normal class and made small by marking them with the ‘small’ class. But then again, others classes can be completely arbitrary, and nouns inside the classes might have nothing at all in common.

Alright, I can hear you saying (those of you who have indulged me this far), maybe I can see the logically-connected-to-meaning noun classes being useful to humans, what with our tendency to classify and analyze the world around us. But what about the systems with no apparent motivation (like French)?

Well, there’s one problem we face when trying to figure this out: lazy scientists still haven’t invented time travel.² You see, a lot of these noun class/gender distinctions go back to before the historical record began. Even the historically-reconstructed Proto-Indo-European language (the ancestor of languages as distantly-related as Hindi, Farsi, Greek, Lithuanian, French, and English, to name a few) is believed to have made a noun class distinction. According to some researchers (sources 1 [warning: PDF], 2, 3), this was probably a distinction between animate (usually this is a class including humans and other animals capable of movement) and inanimate nouns (plants and other things). Hey, look at that: it started out as a non-arbitrary, motivated distinction. So what happened? Later, a lot of Indo-European languages for which we have historical records can be seen to have three genders: masculine, feminine, and neuter (many languages still have this, e.g. German). The neuter gender is close to what they think used to be the inanimate class, while the animate class split off into masculine and feminine by analogy to natural gender.

All that basically goes to say that, as time goes on, languages change and split off into multiple languages, and then those languages change, and so on. English used to have grammatical gender, but gave up on it sometime between the 12th and 15th century (good for us, right?). What probably began as a motivated distinction underwent changes that removed the original meaning of the distinction. By this view, grammatical gender in French and related languages is essentially like human vestigial parts such as the appendix: it used to serve a purpose, but no longer does; the organism (/language) keeps it because it gets passed on from parent to child. This may come as a surprise to you, but no one really cares about adult language learners; you just have to deal with all the irregularities and weirdness. Crying about it takes up valuable memorization time, unless you manage to look at your flash cards through the tears.

This brings me to a final point, going back to the idea of cognitive cost. One argument against keeping no-longer-functional noun classes is that it takes effort to learn and use these forms. But here’s the thing: child language learners don’t care. They seriously do not care. English has hundreds of irregular verbs, but by they time they hit puberty, cognitively normal native English speakers have no problem with them. Or how about the fact that something like 80% of Arabic nouns have an irregular plural form? No problem for a human child. They start to figure this stuff out while they’re still pooping in their pants. Given that capacity, it doesn’t seem like such a big deal for a kid to classify every noun in their language by two, three, or 16 different classes. It’s a relatively small task in the grand scheme of acquiring a human language.

Now stop complaining.

 

¹Seriously, if you are actually willing to engage me in a conversation about this stuff, I will love you forever. That can be platonic or romantic; just let me know.

²I’m kidding, scientists, you’re not all lazy. But you guys could probably pick up the pace a bit, don’t you think?

google analytics code: