Differences between Japanese and Korean?

[dɮ the phoneme]

I am quite familiar with Japanese but know almost nothing at all about Korean. Much is made of the similarities between the two languages, but what about the differences? Naively I tend to work under the model that Korean is basically a Japanese relex (...a sentence seemingly fined-tuned to maximally piss off Koreans), but obviously this is not actually true. I know that Japanese and Korean share little vocabulary outside of Chinese loans, but what about syntactically, morphologically, and in terms of lexical typology etc.? What are the major differences between Japanese and Korean?

[Linguoboy]

I am quite familiar with Japanese but know almost nothing at all about Korean. Much is made of the similarities between the two languages, but what about the differences? Naively I tend to work under the model that Korean is basically a Japanese relex (...a sentence seemingly fined-tuned to maximally piss off Koreans), but obviously this is not actually true. I know that Japanese and Korean share little vocabulary outside of Chinese loans, but what about syntactically, morphologically, and in terms of lexical typology etc.? What are the major differences between Japanese and Korean?

Well, I've talked about some differences in the Pronoun thread. Korean pronouns are more of a closed class than Japanese, with the basic first- and second-person pronouns not being derivable from common nouns nor having common nominal uses.

For me one of the most obvious and striking differences is in speech levels. Much is made of the use of polite language and honorifics in Japanese, but where Japanese is generally considered to have three basic speech levels, Korean has seven. These are independent of honorifics (which exalt the subject of verbs and some objects of prepositions) and humble forms (which do the opposite). There are a half dozen suppletive honorific verbs in Korean, but almost any verb can be made honorific with the use of the very productive infix -(으)시- -(u)si-.

Speaking of infixes, I don't recall anything in Japanese parallel to the retrospective infix -던- -ten- but it could be that I just don't know the verbal system well enough. Korean also has three distinct attributive endings used to form relative clauses; IIRC Japanese verbs just used the regular sentence-final endings here, except with descriptive verbs.

Korean is known for having an extensive system of vowel harmony with lexical and syntactic applications. I think its most striking use is in sound symbolism. I know Japanese is also famed for its rich supply of onomatopoeias, but Korean seems to have more variants (what Martin calls "isotopes") based on systematic two- or four-way vowel alternation combined with three-way (plain ~ tense ~ apsirated) consonant alternation. Here's an example from Kong-On Kim's 1977 article "Sound Symbolism in Korean":

(1) Vowel alternation
/piŋkɯl/ '(turn) round and round'
/pɛŋkɯl/ 'round and round (the circle involved is smaller and the movement faster)'
(2) Consonant alternation:
/piŋkɯl/ 'round and round'
/phiŋkɯl/ 'round and round (the movement is more powerful and faster)'
/ppiŋkɯl/ same as /phirjkɯl/

The alternation between /i/ and /ɛ/ in the examples in (i) brings about a
connotation shift in the speed of the movement and also in the size of the moving
object and of the circle made by the circular movement. In the examples in (2),
the alternation between the word initial lenis stop /p/ on the one hand, and its
aspirated or tense counterpart on the other, signals a connotation shift in the
speed and force of the movement, but curiously not in the size of the circle or of
the moving object. These means of changing the connotation of words are highly
productive in the sense that, given a basic word that belongs to the category of
sound symbolic words, native speakers can predict the form and connotation of
the paired member resulting from the phoneme alternation.

This form of sound symbolism even extends to colour terms, with all basic terms exhibiting "light" and "heavy" isotopes, e.g.: 희다 /huyta/ "to be white/grey" (basic term) ~ 하얗다 /hayahta/ "to be pure white" ~ 허옇다 /heyehta/ "to be cloudy white". In addition, some also exhibit consonant alternation, e.g. 검다 /kemta/, 꺼멓다 /kkemehta/, 까맣다 /kkamahta/, 거멓다 /kemehta/, and 가맣다 /kamahta/ are all variants of the basic colour term for "black".

That's just a few examples off the top of my head. I recommend perusing a good grammar of modern Korean, such as Martin's, because I'm sure you'll notice more contrasts that way. (I'm somewhat hampered here by my ignorance of Japanese grammar beyond the basics.)

[Glass Half Baked]

Here are a couple of snippets to give you an idea of the face you should make when people tell you Korean and Japanese have similar grammar:

"I don't want to think."
Japanese: kangaetakunai
Korean: saenggak sipji anda

"For my family"
Japanese: watasi no kazoku no tame ni
Korean: nae kajok wihae

The cases where you can actually translate across a line word-by-word are only slightly more common than for any random pair of SOV languages.

[bradrn]

"I don't want to think."
Japanese: kangaetakunai
Korean: saenggak sipji anda

"For my family"
Japanese: watasi no kazoku no tame ni
Korean: nae kajok wihae

Could we have glosses please?

[Glass Half Baked]

No. The point is you don't need glosses. If I showed you a sentence "Dorfsplangmickgwarlbumf" with an English translation "The many yellow trees wave majestically," you don't need to know where the article goes to conclude that the two languages are not the same.

[bradrn]

No. The point is you don't need glosses. If I showed you a sentence "Dorfsplangmickgwarlbumf" with an English translation "The many yellow trees wave majestically," you don't need to know where the article goes to conclude that the two languages are not the same.

Of course. But I’m still curious to know how those sentences are constructed!

(Also… this point falls down a bit when you realise how arbitrary word spacing gets. Zulu and Sotho have very similar structures, yet their word spacing makes them look very different.)

[Travis B.]

(Also… this point falls down a bit when you realise how arbitrary word spacing gets. Zulu and Sotho have very similar structures, yet their word spacing makes them look very different.)

People get confused by this more than some people may think. Take English versus, say, German, Dutch, or Swedish - our NP structures are rather similar (except that English is less likely to insert linking morphemes), but people think they are more different than they are because English normally writes compounds with spaces while German, Dutch, and Swedish typically don't.

[Ketsuban]

"I don't want to think."
Japanese: kangaetakunai
Korean: saenggak sipji anda

"For my family"
Japanese: watasi no kazoku no tame ni
Korean: nae kajok wihae

Could we have glosses please?

I am leaning heavily on Wiktionary for the function of Korean verbal forms, since I do not know the language that well.

kangae-ta-ku-na-i
think-PST-CONT-NEG-ATTR
saenggak sip-ji an-da
thought want-DIS NEG-DECL
The spacing is kind of arbitrary, treating sipji as a form of the verb sipda "want" and anda as a verb in its own right like in Finnish. Hypotheses that the two are related generally assume that Japanese rebracketed panas-an-u speak-NEG-ATTR to panasa-nu.

watasi no kazoku no tame ni
1S GEN family GEN sake DAT
I've seen a fair few Japanese people who would write it as watashino kazokuno tameni, for a sense of how attached to the head word particles are thought of as. tame is a relational noun. I've arbitrarily labelled the particle ni as a dative particle; it also covers locative meanings, e.g. ue ni "above".
nae kajok wiha-e
1S.GEN family serve-ARGH
nae is an abbreviation of na-ui. Note the Sinoxenic term 家族 used for "family" in both Korean and Japanese. -e is an intensely fusional morpheme whose precise function I can't divine from Wiktionary.

[Linguoboy]

watasi no kazoku no tame ni
1S GEN family GEN sake DAT
I've seen a fair few Japanese people who would write it as watashino kazokuno tameni, for a sense of how attached to the head word particles are thought of as. tame is a relational noun. I've arbitrarily labelled the particle ni as a dative particle; it also covers locative meanings, e.g. ue ni "above".
nae kajok wiha-e
1S.GEN family serve-ARGH
nae is an abbreviation of na-ui. Note the Sinoxenic term 家族 used for "family" in both Korean and Japanese. -e is an intensely fusional morpheme whose precise function I can't divine from Wiktionary.

LOL. 에 /ey/ is conventionally glossed "LOC" because most of its meanings involve either location at or motion towards. The main difference between it and Japanese に /ni/ is that 에 /ey/ can't be used with animate nouns. E.g.:

学校に /gakkō ni/ "at/to school"
학교에 /hakkyo ey/ "at/to school"

どようび に /doyōbi no/ "on Saturday"
토요일에 /thoyo.il ey/ "on Saturday"

BUT

友だちに /tomodachi ni/ "to a friend"
친구한테 /chinkwu hantey/

And even in this last case, I suspect that the dative particle 한테 /hantey/ incorporates 에 /ey/, though I don't have Martin's dictionary handy to check the etymology.

[bradrn]

kangae-ta-ku-na-i
think-PST-CONT-NEG-ATTR
saenggak sip-ji an-da
thought want-DIS NEG-DECL

watasi no kazoku no tame ni
1S GEN family GEN sake DAT
nae kajok wiha-e
1S.GEN family serve-ARGH

에 /ey/ is conventionally glossed "LOC" because most of its meanings involve either location at or motion towards. The main difference between it and Japanese に /ni/ is that 에 /ey/ can't be used with animate nouns.

Thanks both!

The spacing is kind of arbitrary, treating sipji as a form of the verb sipda "want" and anda as a verb in its own right like in Finnish. Hypotheses that the two are related generally assume that Japanese rebracketed panas-an-u speak-NEG-ATTR to panasa-nu.

This is the kind of thing I was wondering about in my last post — the word spacing may be different, but it doesn’t necessarily mean that they’re structurally different. And, in this case, it looks like they aren’t particularly structurally different! Especially with the last sentences (clauses?), where it looks like Korean is more fusional but is otherwise very similar.