The first is that by having different pronouns for each gender, when you must use more pronouns in the same sentence, there are good chances that you will be able to use different pronouns so there will be no ambiguity about what the pronouns are referring to.
The grammatical gender is a hashing function, with 2 or 3 buckets in the Indo-European languages but with many more buckets in some languages.
When you are lucky and there are no hash collisions, you use only different pronouns in a sentence, when you are unlucky and there are hash collisions, i.e. nouns with the same gender, you use the same pronoun twice, or even more times, and the sentence becomes ambiguous.
Because this role of the genders is hard to replace, even English has retained the gender for pronouns.
The second role for the genders was for adjectives. By making the adjectives repeat the gender of the noun to which they applied, when there are no hash collisions it is clear to which nouns the adjectives are connected, otherwise, like for pronouns, the sentences remain ambiguous.
This was important in the old languages, with free word order, where the adjectives were not necessarily placed besides their nouns.
English has lost the agreement in adjectives, but this simplification has been paid by the loss of flexibility in word order. English has a much more rigid word order than most other languages.
I think a better analogy is error correction code. Some redundancy in a language can be a good thing. Gender information may be redundant, but the fact that adjectives agree in gender with their nouns, helps better distinguish what adjective refers to what noun. Same with pronouns.
For example, the famous Jespersen's cycle[0]: in many languages, "I not know" eventually becomes "I not know a thing" (like in French "je ne sais pas"). "A thing" is completely redundant here, but it helps better understand that it's a negation, in case the first "not" was not heard.
> Because this role of the genders is hard to replace, even English has retained the gender for pronouns.
Only vestigially. He is a single human man/boy (or maybe a pet), she is a single human woman/girl (or maybe a pet or a boat), everything else is it/they. This is very different from what I would consider truly gendered languages, where every single noun is independently gendered (like the knife, fork, spoon example in the article: das Messer, die Gabel, der Löffel)
> The second role for the genders was for adjectives. By making the adjectives repeat the gender of the noun to which they applied, when there are no hash collisions it is clear to which nouns the adjectives are connected, otherwise, like for pronouns, the sentences remain ambiguous.
> This was important in the old languages, with free word order, where the adjectives were not necessarily placed besides their nouns.
Is this really the case? I don't know Latin, but AFAIU, adjectives have to go right next to their nouns. In any case it only helps by 'luck'. If you happen to have multiple nouns with the same gender in your sentence, you get no power to disambiguate. My understanding had always been that it's declension based on case that allows Latin to have a free word order. I can see the value of declension, it's harder for me to understand the value of gender.
In Medieval Latin the adjectives were more frequently close to their nouns, because the writers were influenced by the word order from their native languages.
In Classical Latin, e.g. in Caesar or Pliny the Elder, the adjectives may be located either before or after the nouns and frequently with several other words interposed.
In poetry, having adjacent nouns and adjectives is more an exception than the norm.
In ancient languages, poetry was highly valued, being important both for entertainment and as a mnemonic aid for utilitarian texts.
Unrestricted word order was most useful for poetry, allowing any rearrangements of a sentence that were needed to match a desired rhythm.
The transition to a constrained word order, accompanied by the loss of many of the grammatical means used to mark the roles of the words, which became redundant with a fixed word order, went in parallel with the diminished importance of poetry.
>Is this really the case? I don't know Latin, but AFAIU, adjectives have to go right next to their nouns.
I don't know about Latin, but here's examples from Russian (also true for Acient Greek) where adjectives can be found far away from their nouns (word-by-word translations in their original order, just found on Lenta.ru):
"Lost his family in fire Russian critizes verdict"
("lost" refers to "Russian")
"Found in car minister's corpse moose's seized investigators"
("found" refers to "corpse")
"Recreated existing inside planet giants dark ice"
("existing" refer to "ice")
Gender and declension help tell what refers to what.
It is hard to give a general answer. It wasn't invented, as such, but in the Indo-European linguistic family, a distinction was drawn in the mists of time between an 'animate' and an 'inanimate' class (which is perhaps the most important category distinction that there was to be drawn), and it is from that binary division that a three-fold gender system evolved.
The Afroasiatic languages (including the Semitic languages) already had a grammatical distinction between masculine and feminine nouns many thousands of years before the animate gender has split into masculine and feminine genders in most Indo-European languages.
Because this distinction between masculine and feminine nouns is quite rare among the known language families, it is possible that the appearance of masculine and feminine genders in the Indo-European languages was influenced by the contact with the speakers of Semitic languages (e.g. Akkadian).
It is known that such an ancient linguistic contact has existed, because there are a number of very old reciprocal loanwords between the Indo-European languages and the Semitic languages, dating to about the same time.
Gender is a specific case of the more general structure of classifiers. I think this is mentioned in the article but so far I have only skimmed it. For example Arabic has “sun” and “moon” words. Japanese has a rich set. I have always thought of them as a kind of disambiguator to reduce semantic errors in listening. But really the question of “why” makes as little sense in linguistics as it does in evolution, as any structure is a combination of happenstance, just so, environment, and such.
I have read that affix discriminators (e.g. declining a noun) started off as standalone words. Could be true; the distinction of “word” feels to me like an artifact of writing.
Instead of Latin, can I please have a logical language intended for thinking about scientific and engineering processes invented that's similar to the idea of the Ancients in Star Gate (the TV series)?
Why "instead of Latin"? Your requirement seems to be orthogonal to the existence of Latin.
Latin's a dead language; we can't submit pull-requests.
English (etc.) are living languages; people are committing changes to them all the time, with no maintainer to oversee the project.
Sure, I'd be happy not having had to spend schooldays learning French and Latin irregular verbs. But these languages weren't developed for dealing with science and engineering.
My guess would be animism giving genders to objects along with spirits, and from there on the newer, more abstract nouns either developed from older gendered nouns or got a default gender.
And it gets continued usefulness as an error detection and correction mechanism.
tl;dr Originally Proto-Indo-European had no genders. The concept was added piecemeal. Subjects were marked with ·s. Results of an action or grammatical objects were marked with ·m, this was generalised and transferred onto actual objects/things. This was the birth of the first gender distinction, neutrum, and the ·s words assumed the role of the default gender. Collective words (before the invention of plural) and abstract words ended in ·a. The animate/inanimate distinction mentioned by thread neighbours is a red herring, disregard that theory.
The conflation of genus and sexus and hence the unfortunate misnaming of default gender as masculine and abstract gender as feminine which is still in use today is due to bad Roman grammarians; the HN submission article mentions that in contrast Varro understood it better.
As a native English speaker, I've never understood why it was invented in the first place.