Korean and Japanese structural similarity

[Moose-tache]

linguoboy's covered a lot of it already, but once you dive into the Korean verbal complex it becomes obvious that Korean and Japanese aren't that similar. Look at how the two languages handle basic things like relative clauses or negation. And the morphology of questions in Korean is quite a bit more complex than in Japanese, or for that matter most languages I'm familiar with. I'm sure there are entire dissertations about -지 alone.

As for the history, it's certainly not true that we have to wait for Hangeul to have data on historical Korean. Korean has had its own equivalents of manyoogana and hiragana for over a thousand years. Phonetic transcription of Korean grammatical information predates the practice in Japan. The problem is that very little of this data survives because Korea is the Poland of Asia, but there is enough to draw some conclusions. We know, for example, that most of the basic particles and conjunctions, as well as some verbal endings, go back to the Three Kingdoms period. Most likely the "left-most" layer of morphology is very ancient, and (as others have pointed out) things like politeness markers are of more recent origin. There is no evidence of anything like -요 before Middle Korean. Even today, it sometimes gets sprinkled onto random function words to add politeness (maybe linguists treat it as a particle, I don't know, but Koreans tend to think of it as a suffix).

[bradrn]

Does that make Japanese odd for having verby adjectives?

Japanese adjectives are odd in general — I hear it actually has two adjective classes, one nouny and one verby. The universal predicts that the verby adjective class should be closed while the nouny one should be open, which seems to line up with what I can find (Backhouse 2004, ed. Dixon and Aikhenvald).

The weird thing is that it’s not just word order: e.g. OV languages tend to be cased and have nouny adjectives, VO languages tend to be non-cased and have verby adjectives. Why? I don’t have a clue. It’s just one of those inexplicable phenomena you regularly get in linguistics.

The case thing doesn't surprise me at all. I'm guessing that postpositions more easily erode to become case suffixes than prepositions erode to become case prefixes. I can't think of any language right now that has case prefixes (other than, like, arguably French with it's opposition between absolutive l(ə)- / l(a)- / le(z)- and genitive/partitive dy- / dəl(a)- / de(z)-). Can't think of a reason for the differing qualities of adjectives, but yeah, a lot of stuff just seems to have flow-on effects. I know Turkish adjectives are pretty nouny. I thought Japanese adjectives were generally pretty verby though, although I remember there's a few different classes of them from memory.

This is an interesting argument. Could it just be that when people see case prefixes, they tend to call them prepositions instead? (e.g. Hebrew et, which is pretty clearly an accusative case.)

The correlation with adjectives, meanwhile, is generally talked about in terms of tensedness: tensed languages have nouny adjectives and non-tensed languages have verby adjectives, which actually makes quite a bit of sense. (Adjectives are stative, so it’s less natural to apply tense to them.) But this just makes it a question of why tensedness correlates to OV/VO, which if anything is even less obvious.

Czech has had 1000 years of college contact with German (including several hundred years when everyone above a peasant spoke German) and Czech does not look like German grammatically.

Doesn't it? I mean, it's not an active-stative OSV language, is it?

Do active-stative OSV languages even exist, though?

linguoboy's covered a lot of it already, but once you dive into the Korean verbal complex it becomes obvious that Korean and Japanese aren't that similar. Look at how the two languages handle basic things like relative clauses or negation. And the morphology of questions in Korean is quite a bit more complex than in Japanese, or for that matter most languages I'm familiar with. I'm sure there are entire dissertations about -지 alone.

More details please?

[Linguoboy]

Do active-stative OSV languages even exist, though?

You're overthinking the joke.

linguoboy's covered a lot of it already, but once you dive into the Korean verbal complex it becomes obvious that Korean and Japanese aren't that similar. Look at how the two languages handle basic things like relative clauses or negation. And the morphology of questions in Korean is quite a bit more complex than in Japanese, or for that matter most languages I'm familiar with. I'm sure there are entire dissertations about -지 alone.

More details please?

What, on the use of -지 /-ci/ specifically? Martin glosses this as "the uncertain fact whether". Above you have examples of its use in negation. 푸르지 않다 "is not blue", for instance, decomposes into:

푸르-지 아니-하-다
/phulu-ci ani-ha-ta/
blue-uncertain.fact no-[light verb]-declarative

That is, negation is commonly expressed by a negative verb derived from a light verb and "no". A contracted form of 아니 /ani/, 안 /an/, can also be prefixed to verbs in Korean (something which, AFAIK, has no parallel in Japanese) but its distribution is somewhat limited and I don't think ?안푸르다 /an.phuluta/ is used, which is why I didn't include it above.

However, -지 /-ci/ can also be used on its own as a sentence-final ending (with or without the polite particle -요 /-yo/ depending on the speech level) to invite agreement or to express uncertainty:

어렵지? /elyep.ci/ "It's difficult, isn't it?" [Cf. 어려우니? /elyewuni/, which is a more neutral question.]

Of course, given how politeness works, this means that it can also be used as an indirect imperative:

천천히 먹지
/chenchen-hi mek-ci/
slow-ly eat-uncertain.fact
"Eat slowly" (Roughly "And (what) if you ate slowly?")

This is really just scratching the surface. Here's a nice rundown of these and other possible uses: https://ultimatekorean.com/ukr/023/.

On top of this, there's another nonfinal verbal ending 지 /ci/ which is transparently a bound noun preceded by relative endings and expresses the meaning of "since".

담배를 끊은지 한 달 됐다.
/tampay lul kkunh.un ci han tal toayss-ta/
cigarettes ACC stop-PST.REL time one month become-PST-DECL
It's been one month since I stopped smoking.

Perhaps there's a historical relationship between the two morphemes ("time, occasion" > "uncertain fact" doesn't seem a huge leap to me) but their syntax is quite distinct now.

[Moose-tache]

안 is used far more frequently than -지 아니다, but you're right 안푸르다 doesn't sound right to me. My mind wants to find a way to just make 아니다 the main verb: 파란색 아니다 or something. Also, maybe it's in your link, but -지 is used with other common constructions, most notably -지 마. So far as I'm aware, nothing remotely like this suffix exists in Japanese. The -te ending patterns more like Korean -고, and nothing else comes close.

Other differences include: Korean is more permissive of juxtaposing bare nouns, while Japanese is more permissive of unmarked verbs modifying nouns. Japanese "honorifics" are an archaic, closed class of verbs, while Korean "honorifics" are morphological and add an addition layer to the already Byzantine politeness gradient. Korean has some vestigial vowel harmony in its verbal endings, which is unknown in Japanese. This one is subjective, but I've always found Japanese word boundaries easy to distinguish in speech but impossible in writing, while the reverse is true in Korean.

[Linguoboy]

Korean has some vestigial vowel harmony in its verbal endings, which is unknown in Japanese.

There is evidence for vowel harmony in Old Japanese, but you're right that this has left basically no traces on the modern language. And Korean vowel harmony is anything but vestigial when it comes to its sound symbolism, which is more developed than for any other language I've ever studied. There are hundreds of terms--not just onomatopoeias (which Japanese is also known for) but ordinary adverbs and adjectives, including colours and moods--which come in multiple "isotopes" (to use Martin's terminology) based on these historical alternations. Add in "intensive" and "para-intensive" (again, Martin's terms) based on consonant alternations and you have literally thousands of attested variants. Consider for instance:

통통 /thongthong/
퉁퉁 /thungthung/
똥똥 /ttongttong/
뚱뚱 /ttungttung/

all expressing varying degrees of plumpness.

[Ketsuban]

There is evidence for vowel harmony in Old Japanese, but you're right that this has left basically no traces on the modern language.

It's worth noting that as much as Altaicists love Arisaka's laws for much the same reason they love Yonaguni reflexes of words like "intestines" and "mountain", the system of vowel harmony they can be taken to imply doesn't really resemble the vowel harmony systems of Korean or Tungusic languages. Francis-Ratte doesn't even reconstruct vowel harmony for Proto-Korean-Japanese - he takes Arisaka's laws to be the result of sound change where schwa was lowered by a following /a o u/.

[Rounin Ryuuji]

As far as Arisaka's Law goes, I've encountered conflicting accounts on it, including one where Frellesvig mentions that */a/ might have been some sort of plural formant at an earlier stage, an idea that strikes me as not implausible — note the pluraliser ra persisting in modern words like warera and bokura (chiefly only for pronouns, but it seems to be fossilised in sakura, "blooming things", and maybe also mura, but I have some doubts on that last one) — and the apophony of *pitə (modern hito) "one", *yə (modern yo) "four, but *puta (modern futa) "two", *ya (modern ya) "eight"; he also notes OJ *tanasoko (palm), *manakwo (pupil), *minamətə (fountain, spring; note this apophony occurs in a word that has */ə/ already, unless I'm mistaken about the Old Japanese vowels in the second morpheme). The alternation of */i~u/ in *pitə~*futa may be somehow connected to that in *mi "three", *mu "six". The vowel inventory of Proto-Japonic was probably /i (j)e u ə ɨ (w)o a/, and I think it would be odd to suppose that harmony only existed between */ə a/; */ɨ/ turns up mostly in monosyllables, where it mostly merged with */ə/, but leaves traces in words like kesa "this morning" (not *kosa or *kasa), kyō "today" (archaic kefu rather than *kofu), but kore "this (pn.)", kono "this (det.)").

As far as Yonaguni and Yaeyama go, the initial */b d/ are more-or-less certainly innovations (that they're fortition of inherited */j w/, rather than inherited, leaving an unsatisfying gap for */g/ in Proto-Japonic); from what I've managed to gather, Proto-Japonic probably had a consonant inventory of */m n p ^mb t ⁿd k ^ŋg s ⁿz r w j/, and I don't really see anything in this that would support Pre-Proto-Japonic */b d (g)/ > */w j Ø/; if — if — Pre-Proto-Japonic did have initial */b d g/, I suspect they instead merged into */p t k/ and left a low-pitched initial sound (a similar change happened in Thai, I think it was), but there isn't any evidence to suggest such a change happened, and nothing about the Proto-Japonic we can reconstruct, to my knowledge at least, needs to have initial voiced stops at all.

It might, on this note, be a fun mental exercise to reconstruct Pre-Proto-Japonic as having "dim" vowels */ɯ ə uo ɒ/ and bright vowels */i æ ie u/, and initial voiced stops, but that would veer almost entirely into the realm of pure fantasy.