Iversen wrote:I see two opposite pressures in linguistics: the most important is probably the laziness of the speakers, but in some cases it does indeed seem that the very same speakers add as many complications as they can deal with. Or at least they accept that certain sound changes result in a mess, and there is no collective push to find ways to deal with the problems - except when the language users simply give up en bloc and discard something they find too complicated.
Except that's the thing. It has nothing to do with the laziness of the speakers. It has everything to do with sound changes causing some forms to merge, other forms becoming preferred, paradigms being regularized (my dialect of English, for instance, has regularized negative "do" to "don't", and past tense "be" to "was", for all persons (among other things) -- "he don't need nothing", "we was (often "we's") just leavin'"), or grammaticalization happening making a preposition into a case ending, for instance. It really has nothing to do with laziness though, and language changes slowly shift through a population. Apart from taboo instances, I'm not aware of a single instance of users giving up "en bloc" to discard something they find too complicated. A lot of times they didn't even realize it was going until it was already gone; it certainly wasn't a conscious decision of the speakers of my dialect to regularize their past tense paradigm, for instance.
For instance it would be much easier for everybody including native speakers of Latin if all Latin case endings were unique and easy to recognize, but for instance in the 1. conjugation of nouns and adjectives we find the ending "ae" in several places. Who benefitted from that? The "ae"'s may have developed from different forms in an even earlier proto version of Latin (I haven't studied the prehistory of that language), but why didn't the ancient Romans invent ways to keep them separate? They actually did something, but much later, and that was simply to drop a large part of the morphology (and much of the syntax too), and the result is known as the Romance languages. I see this as some kind of Darwinian selection, where something dies out.
Except it doesn't necessarily work as Darwinian selection, as it has nothing to do with what's better suited to an environment or what's simpler. It's just random sound changes that then spread through a population for various reasons (conquest, prestige, etc.) The case endings in Latin just eventually collapsed together, through various sound changes and other syntactical features taking their place. The "dropping" of the morphology and syntax didn't happen overnight. It was a process that took centuries and occurred slowly. And other, just as complicated, structures arose in their place to compensate for it.
Languages develops because certain errors against all expectations become so popular that they replace existing words or patterns. And nobody promised you that errors would be neither systematic nor logical - and least of all if the error in question consists in a contrived way of mispronouncing certain sounds.
This is just a normal sound change. It's only an error in-so-far as you assume there's one version of the language that is "correct". Sounds change one generation to the next and it spreads (again, through conquest, prestige, etc.). It's not a "contrived way of mispronouncing" it -- it's quite literally the correct way in that group's dialect.
As for redundancy: it may help people to understand whispering and badly recorded speech, but I see that as a side effect that doesn't explain the nature of the redundancies. Let's take an example, namely the sentence "M..cus stult.. est". Oh yes, I know that the stupid person has a name from the 2. declension, but I don't know who it is. On the other hand, if the endings disappeared ("Marc.. stult.. est") I could be fairly that the poor fellow was named Marcus. Apart from that, the argument that the endings help you to understand sentence structures is true, and to get heads and tails out of Latin poetry and artsy writing like the speeches and letters of mister Cicero it may be necessary to study the endings. But it is very unlikely that people in the streets of Rome used such complicated sentence structures when they discussed among themselves, and when I look at modern Russian or German texts most sentences use a fairly strict and welldefined word order which makes the information contained in the endings less crucial. There are of course counter examples (some of which have been constructed for the purpose), but I would get more information from a Russian text without the endings than from one without the roots.
How many possible names are there that would fill that ending? And, assuming that the ending didn't exist and you misheard the same part you'd have "M... stult ... est", which gives you
less information than the part where you do hear the ending. And it has more to do with just whispering an recorded speech. Spanish has, for instance, several words that differ only in gender. Let's say you miss the referent, or the article the first time around (for instance, you came in late to a conversation). But you pick it up when the speaker uses the article later and you can infer what they're talking about. For instance, say you walk in on two of your friends talking about bands, and you hear "... batería ... la". You'd know one of two things (a) they weren't talking about bands but about batteries or (b) they were talking about a band with a female drummer. It clarifies a lot, and that's precisely the redundancy at work.
As to word order, it still is flexible, and, even if there is a preferred one, it can be moved and is by speakers at times.
And finally: how did biological gender end up as grammatical gender, which in many cases goes squarely against the biological facts (like German Mädchens who are neutrum beings)? I just say that the idea to have a distinction must have occurred because women and men looked at each other and noted that there was a difference, and that it was important. And then they found a way to express this in language - let's just for fun say that all female names got an -a at the end and all male names got -us (and yes, I know that this is Latin and not whatever markers the first speakers used). But then somebody looked at a baobab tree or a stone and wondered what gender that thing might have, and from there everything went wrong.
This is a conflation of two words, to be honest. Gender was originally used similar to "genre" to refer to a class, and originally referred to just two "kinds", an didn't really take on its current role until after the eroticization of the word "sex" in the 20th century... well after "gender" had been used as a name for noun classes. It's best to completely delink the two concepts when you discuss noun classes, as the two/three gender system is fairly uncommon (it's widespread simply because of Indo-European languages). Bantu languages, for instance, are known to have upwards of 20 genders. But, as to how words came to be grouped that way: usually because of similar sounds. So the word ended the same, which caused an article to take on the same allophone, etc. Eventually the allophones split into grammaticalized words associated with that gender. Then semantic broadening could've happened, or a collapse of groups of genders (see: PIE's 3 merging to 2, as mentioned by reineke).
And that also precisely explains your example of Mädchen being a neuter noun -- it's the diminutive, and, in German, diminutives are all neuter nouns. They're not grouped by biological gender, but based on sound and original grammatical roles. It's the same with cailín, girl, being masculine in Irish. the diminutive -ín ending makes a word masculine, so therefore cailín is. Which should show that gender is more governed by
sounds that by any perceived connection to natural gender (doubly so since you'd still use the feminine pronoun when talking about a cailín in Irish).