Gender marking

A. In European and many other languages, nouns are divided into gender classes. Spanish has masculine and feminine, marked with an article and often with the final vowel: el sombrero, la casa. German has three: “the spoon,” “the fork,” and “the knife” are der Löffel, die Gabel, and das Messer. This is not necessary in a language: it is an accident of history.

B. Stage one. In many languages, we can see how this marking begins. In Dyirbal, spoken in Australia, all nouns must be preceded by a separate word. Which word a noun takes depends on which of four categories it fits into. One is for males and animals, another for female things, another for food that is not flesh, and another is the grab bag.

Dyirbal gender classifiers:

C. Stage two. Over time, separate words such as these erode and become prefixes or suffixes—grammaticalization again. At first, the new prefixes or suffixes still correspond fairly well to categories. Swahili is at this stage. Swahili has seven “genders” (although because sex is not one of the categories marked, linguists call them noun classes). The one with an m- prefix contains people: mtu, “man”; mtoto, “child.” The one with an n- prefix contains animals: ndege, “bird”; nzige, “locust.”

D. Stage three. But as time goes on, sound change, cultural changes, eccentric semantic switches (such as the one that made the word for sister-in-law masculine in Proto-Indo-European), and other processes make the correspondence between marker and category increasingly vague. European languages are an example of this stage, where only marking actual male beings masculine and actual female beings feminine makes any immediate sense anymore.

Thus, a great deal of what a language’s grammar pays attention to is technically a kind of window dressing. Keep in mind that there are actually some languages that do not mark tense at all, and some where I and we are the same word, he and they are the same word, and so on, because pronouns mark person but not number! This shows that it is inherent to human language to overelaborate.