Zompist Bboard Again

The reason for asking this is primarily from a conlanging point of view, but I see no reason not to educate myself in doing so, so I'll take the natlang approach. My question, basically, is what drives innovations in languages when they already have a perfectly good system? For instance why do word orders change? How did PIE's SOV word order end up as VSO in Celtic, SVO in Germanic, etc. Moreover, where do new verb conjugations come from? Or new noun cases? Welsh developed (at some point) new verb conjugations to attach to stems: gwel-, for instance can be gwelaf (I see), gweli (you see), gwelwn (we see), gweld (verb-noun 'see'), etc. but where do (or where can) these things come from?

I mean, language is always changing. The specific reasons why specific changes take place vary. Word order is notoriously unstable and can change as a result of things like focus constructions FWIU. Verb forms often come from incorporating pronouns into the verb, and noun cases I think often come from former nouns or verbs becoming clitics and eventually just case suffixes.

Something comes along which sounds better?

conlanging is the best shared thing...
every group and every member of the group try to find a different way of speaking to differentiate themselves, to get attention, to progress socially...
the result is a permanent linguistic change...

Jonlang wrote: ↑Wed May 12, 2021 10:57 am The reason for asking this is primarily from a conlanging point of view, but I see no reason not to educate myself in doing so, so I'll take the natlang approach. My question, basically, is what drives innovations in languages when they already have a perfectly good system?

The one thing to rule them all seems to be sound changes, notably if you have two things you don't want merging suddenly sounding similar or alike, something else comes along to take its place. Languages can handle some forms being identical, of course, but get a shift of [wiː juː] > [yː], and English is probably going to want to innovate a new pronoun or two.

For instance why do word orders change? How did PIE's SOV word order end up as VSO in Celtic, SVO in Germanic, etc.

I'm not as sure on this point.

Moreover, where do new verb conjugations come from? Or new noun cases? Welsh developed (at some point) new verb conjugations to attach to stems: gwel-, for instance can be gwelaf (I see), gweli (you see), gwelwn (we see), gweld (verb-noun 'see'), etc. but where do (or where can) these things come from?

Contractions, of course! First, you get semantic bleaching of some auxiliary word, then it gets attached to something else, then, there you have it — a new inflection! Where you get particularly messy sound changes, a regular pattern emerges, and tends to take root in less-common words (very common ones can handle being irregular indefinitely, however).

A very general answer would be imperfect learning. Children learn a grammar that roughly produces the language their parents speak but it is not completely the same.
Also language contact of course, both by assimilating and dissimilating to other languages.
And finally, grammaticalization just happens and probably someone has found reasons.

Rounin Ryuuji wrote: ↑Wed May 12, 2021 1:04 pm English is probably going to want to innovate a new pronoun or two.

It already has, hence why most dialects have a form like you guys, y'all, you lot, youse, and so on.

Jonlang wrote: ↑Wed May 12, 2021 10:57 amThe reason for asking this is primarily from a conlanging point of view, but I see no reason not to educate myself in doing so, so I'll take the natlang approach. My question, basically, is what drives innovations in languages when they already have a perfectly good system?

Define "perfectly good". Why don't we use Bell Beaker pottery vessels any more? Those were perfectly good, after all.

Language is a human artifact and is susceptible to all the sorts of change which other human artifacts (such as pottery and music) are. Some changes are driven by utility but others are driven purely by the inexhaustible human need to differentiate ourselves. This is perhaps most obvious in vocabulary, where each new generation creates an innovative youth slang, but it affects all aspects of language. The Great Vowel Shift, for instance, is hypothesised to have been initiated by London merchants who were looking for a way to distinguish their speech from that of mere peasants but without slavishly adopting the habits of the aristocracy. And as for word order, using innovative syntax for dramatic effect--you can see the appeal of that, right? Of course, due to the process of "bleaching", yesterday's bold innovations becomes today's same-old-same-old and so if you want to catch your listener's attention, you need a new trick.

Jonlang wrote:Moreover, where do new verb conjugations come from? Or new noun cases? Welsh developed (at some point) new verb conjugations to attach to stems: gwel-, for instance can be gwelaf (I see), gweli (you see), gwelwn (we see), gweld (verb-noun 'see'), etc. but where do (or where can) these things come from?

Don't these simply continue the Proto-Celtic thematic present forms, which in turn continue the PIE primary thematic active endings?

If you look at other branches of Indo-European, you can clearly seen how new tenses may arise from existing ones. Generally it involves a non-finite form (such as an infinitive--and new non-finite forms get innovated regularly, too) and a light verb, which gets worn down and cliticised and eventually fixed into place like any other affix. The innovative future tense in Romance languages illustrates this process beautifully.

Jonlang wrote: ↑Wed May 12, 2021 10:57 am The reason for asking this is primarily from a conlanging point of view, but I see no reason not to educate myself in doing so, so I'll take the natlang approach. My question, basically, is what drives innovations in languages when they already have a perfectly good system? For instance why do word orders change? How did PIE's SOV word order end up as VSO in Celtic, SVO in Germanic, etc. Moreover, where do new verb conjugations come from? Or new noun cases? Welsh developed (at some point) new verb conjugations to attach to stems: gwel-, for instance can be gwelaf (I see), gweli (you see), gwelwn (we see), gweld (verb-noun 'see'), etc. but where do (or where can) these things come from?

I can't say I've yet read the definitive answer on the topic (though I got some promising stuff on my to-read pile. Which is starting to look like a to-read space elevator. But I digress.)

I can run through a few examples though!

Perfect with habeo dates back, at least to Cicero. Except when Cicero said habeo scriptas litteras, well, he literally had written letters. From there it's not a huge step to say the same thing metaphorically. The metaphor was increasingly used, most likely because it brought an added nuance of recency and relevance. (It also helped that Germanic speakers had the same construction.)
In French its use stabilized as a recent past, but of course the recency constraint extented until it replaced the simple past.
(Note: the simple past feels difficult, and I'm tempted to say it fell into disuse because of that. But maybe it feels more difficult because we don't use it as often.)

Sound changes from Latin to French more or less axed the final vowels. This left many verb forms ambiguous as for the subject; gradually subject pronouns (typically used for emphasis) became mandatory. These merged in turn with the following verb or auxiliary:
habeo > ego habeo > j'ai.

Old French speakers were fond of saying things like 'I didn't walk a step.', or 'I didn't drink a drop' for emphasis:
[gloss="I not have step walked']Je n'ai pas marché[/gloss]
[gloss="I not have drop drunk']Je n'ai bu goutte[/gloss]

The construction was so overused that eventually je n'ai pas marché just meant 'I didn't walk a step.'
Old French used to have several little negative words: pas, goutte, point, mie... but as the original meaning was lost, well, speakers started using pas with all verbs: je n'ai pas bu, je n'ai pas écrit.

I don't think this happened purely for the heck of it; the negative morpheme was reduced to ne or n', easily confused with pronouns such on or l' that also occur before the verb.

Having two negative morphemes is kind of overkill, though, so n' disappears in spoken French.

In J'ai pas écrit, though, you're not going to stress j', ai, or pas which are strictly utilitarian.
Interestingly, this construction fossilizes older features, such as SOV order, or nouns ('pas') without articles; more practically this means you can't insert any new material in there.

In other words, it's pretty much a single word.

And that's how you get from non scripsit to [ʒepaekri]

An historical linguistics book of mine has a simplistic, but serviceable explanation. The gist of it, is, people are lazy mumbler who try and keep the effort involved in speaking at a minimum; at some point, though, they have to innovate new forms to replace or supplement the old ones. It looks a bit like a constantly ongoing effort to keep the information content of our speech at a maximum.

There's no "perfect". A system always gets changed, a nudge here and a push there. Even ignoring phonology:

All Chinese languages plus Japanese and Korean have developed a complex system of classifiers, which are usually* mandatory, presumably from historically classifying nouns really often.
(*There are apparent exceptions in Chinese but these employ special "quantifier" forms, which developed from phonological mergers of the cardinal numbers + the general classifier 个/個. They exist in Beijing Mandarin for 1, 2, 3. Some Zhongyuan Mandarin dialects have these for every number. The difference between this system and a system like English's is that "two times three equals six" uses different numeral forms than "two orders each of three books means six books in total".)

They have also developed large numbers of serial verb constructions, presumably from historically using verbs together to describe actions quite often.

Most affixes, adpositions, and markers in Chinese languages are results of historical grammaticalisation of content words. E.g.

• Preposition 从/從...，cóng "from" <- "follow[ing]";
• Postposition ...里/裡~裏 lǐ "in(side)" <- "inside (n.)", which is becoming kind of like a case marker (inessive) in many cases;
• Verbal aspect of experiential ...过/過 guo <- "to pass", which has even got a doublet verbal aspect of excessive ...过/過 guò.
• Preposition 以... yǐ "with" <- "us[ing]";
• Polar question marker ...吗/嗎 ma <- 无/無 "not",
• Postposition ...时/時 shí "when" <- "time"

Reanalysis can also work in unexpected ways for you. Mandarin copula 是 shì derives from Old Chinese "this", because the sentence pattern got reanalysed as Which turned an SOV word order for predication into SVO.

A less surprising case of reanalysis is 把 bǎ "to hold" becoming a direct object marker where there's a notable degree of control, but this effectively turned the word order in such cases into SOV:

"[M]orphology (word‐building) and syntax (word‐arranging) [...] manage the Escheresque trick of getting simpler and simpler until they end up just as complicated as ever."
Justin B. Rye

That reminds me of Japanese. The typical quadrigrade conjugation seems to have its forms descend from */u/ (some sort of verbal marker; alternatively the epenthetic vowel; reconstructions differ here, in my experience), */i/ (a nominaliser), */e/ (a "transitivity switch"), */a/ (the first part of a negational morpheme eclipsing whatever end was there before and being mistaken for a stem?), and, in very old Japanese, possibly smashing */(ə)rə~(ə)jə/ together to get an ending */-je/ (this is not distinguished from */e/ in modern Japanese. You had a pretty four-stem system in Classical Japanese, but now, despite being called godan (quintigrade, but I haven't seen it given an English name), there are often six stems, since the past tense and te-form tend to get some funky innovations.

(These verbs may or may not have originally had a stem in */a/ or */ə/, depending on your reconstruction of Old Japanese; I find the */ə/ one most compelling because of a few vestigial forms in the kuru paradigm in ko-. Little irregularities in very common verbs can also be preserved for a while, note that the Classical Japanese copula was ari rather than aru in the predicative.)

Classical Japanese had four (formally five, but two of them were virtually identical) regular conjugational patterns rather than two (formally three, but again, two are more-or-less identical), but it feels on some level like the modern language gets more complicated with all the proliferated contractions that are now normative rather than spoken variants of a more regular form.

The complex system of number classifiers seems to have been borrowed from Middle Chinese; Japanese also has its own native counter つ that can be used with native numerals to count anything, if I understand right, demonstrating that weird complicated innovations can be borrowed and maintained, especially if there's a pressure of cultural prestige going round.

(Obligatory noting to keep your salt handy, since I'm not an expert, and my resources on the subject are rather limited.)

Seirios wrote: ↑Thu May 13, 2021 6:25 am Reanalysis can also work in unexpected ways for you. Mandarin copula 是 shì derives from Old Chinese "this", because the sentence pattern

xx

,

是

this

yy

,

也

COPULA

"xx, this is yy"

got reanalysed as

xx

,

是

COPULA

yy

,

也

INTERJECT

"xx is yy"

Which turned an SOV word order for predication into SVO.

The first stage of that sounds suspiciously like the usage of Russian это, sans the copula (examples shamelessy taken from Wiktionary):
Э́то — брат и сестра́.
this brother and sister
They are brother and sister.

Со́евое молоко́ — э́то напи́ток, полу́ченный из зёрен со́и.
soy-ADJ milk this beverage, received from grain-GEN.PL soy-GEN.SG
Soy milk is a beverage produced from soybeans.

I wonder if это will eventually come to be the copula. Though somebody in wiktionary had something else to say about it:

One of the uses of this form is to connect subject and predicate as kind of topic or subject marker. (However, as opposed to actual topic markers such as は (wa) in Japanese, ‘это’ may be omitted without changing the meaning or breaking the sentence structure.) Such usage does not get translated into English.

I might be wrong but from some recent research into Classical Chinese, I'd think the transition of 是 from 'this' to a copula would in part be affected by the rest of the language being predominantly SVO, and so bringing the construction more in line with the rest of the language. But that's only my guess and there do seem to have been constructions where certain objects moved before the verb in certain situations.

Does anyone know if there's any evidence whether older forms of Chinese might have been SOV?

linguistcat wrote: ↑Thu May 13, 2021 10:09 pm I might be wrong but from some recent research into Classical Chinese, I'd think the transition of 是 from 'this' to a copula would in part be affected by the rest of the language being predominantly SVO, and so bringing the construction more in line with the rest of the language. But that's only my guess and there do seem to have been constructions where certain objects moved before the verb in certain situations.

I think the OC constructions are easier to understand if we accept that 也 is a particle, not a verb. It appears where a particle does, it can't be negated normally or take adverbs, it can co-occur with a verb. It may simply be a declarative particle, or have an aspectual meaning (an ongoing state).

Thanks for the replies, everyone! Some good stuff for me to think about, but too much for me to thank people individually. Much appreciated.

Rounin Ryuuji wrote: ↑Thu May 13, 2021 1:24 pm The complex system of number classifiers seems to have been borrowed from Middle Chinese; Japanese also has its own native counter つ that can be used with native numerals to count anything, if I understand right, demonstrating that weird complicated innovations can be borrowed and maintained, especially if there's a pressure of cultural prestige going round.

"Attorney General", "letters patent", "battle royal", "president-elect", "professor emerita/us", "God Almighty", "devil incarnate"

The "God Almighty" one demonstrates it spreading to native words.

Seirios wrote: ↑Sat May 15, 2021 12:06 am

Rounin Ryuuji wrote: ↑Thu May 13, 2021 1:24 pm The complex system of number classifiers seems to have been borrowed from Middle Chinese; Japanese also has its own native counter つ that can be used with native numerals to count anything, if I understand right, demonstrating that weird complicated innovations can be borrowed and maintained, especially if there's a pressure of cultural prestige going round.

"Attorney General", "letters patent", "battle royal", "president-elect", "professor emerita/us", "God Almighty", "devil incarnate"

There is a tendency to reanalyse these to make them conform better to English's native patterns; compare Shakespeare's atturneyes generall to modern attorney-generals and his letters patents to modern letters patent.

Rounin Ryuuji wrote: ↑Sat May 15, 2021 10:00 am The "God Almighty" one demonstrates it spreading to native words.

Alternatively, it could be a survival of indigenous NA word order; compare line 3070 of Beowulf: þeodnas mære "famous lords". However, you're probably right after all, as OE and Early ME exclusively have forms of the type ælmihtig god; Romance influence handily explain why the type god almighty only appears later on.