Korean grammar

This article is a description of the morphology, syntax, and semantics of Korean. For phonetics and phonology, see Korean phonology. See also Korean honorifics, which play a large role in the grammar.

Note on romanization

This article uses a form of Yale romanization to illustrate the morphology of Korean words. The Yale system is different from the Revised Romanization of Korean seen with place names.

Under the version of Yale used here, morphemes are written according to their underlying form rather than their spelling in the Korean writing system or pronunciation. Under this system, for example, the syllable which is written in Korean as 었 is analyzed as ess even though the the ss would be pronounced t before another consonant, and the vowel e ㅓ is pronounced low and somewhat rounded, closer to o. To avoid confusion, bold type will represent the morphology (in Yale), and italics will represent Revised Romanization.

Classification of words

Korean grammar
Hangul	9품사
Hanja	9品詞
Revised Romanization	9 pumsa
McCune–Reischauer	9 p'umsa

Korean grammarians have been classifying Korean words to parts of speech for centuries, but the modern standard is the one taught in public schools, chosen by South Korea's 1963 Committee on Education. This is the 9 품사 pumsa system, which divides words into nine categories called pumsa.

The pumsa are themselves grouped together according to the following chart.

• Content words
  • Substantives
    • 명사 myeongsa (nouns)
    • 대명사 daemyeongsa (pronouns)
    • 수사 susa (number words)
  • Verbs (broadly speaking)
    • 동사 dongsa (action verbs)
    • 형용사 hyeongyongsa (descriptive verbs or adjectives)
  • Modifiers
    • 관형사 gwanhyeongsa (determiners, prenouns, or indeclinable adjectives)
    • 부사 busa (adverbs)
  • Other content words
    • 감탄사 gamtansa (interjections or exclamations)
• Function words
  • 조사 josa (particles or postpositions)

Both cardinal and ordinal numbers are grouped into their own part of speech. Descriptive verbs and action verbs are classified separately despite sharing essentially the same conjugation. Verb endings constitute a large and rich class of morphemes, indicating such things in a sentence as tense, mood, aspect, speech level (of which there are 7 in Korean), and honorifics. Prefixes and suffixes are numerous, partly because Korean is an agglutinative language.

There are also various other important classes of words and morphemes that are not generally classified among the pumsa. 5 other major classes of words or morphemes are:

• 어미 eomi (verb endings)
• 지시어 jisieo (demonstratives)
• 접속어 jeopsogeo (conjunctions)
• 접두어 jeopdueo (prefixes)
• 접미어 jeommieo (suffixes)

Substantives

Main article: Korean nouns and pronouns

Nouns

명사(名詞) Myeongsa, "nouns," do not have grammatical gender and though they can be made plural by adding the suffix 들 deul to the end of the word, in general the suffix is not used when the plurality of the noun is clear from context. For example, while the English sentence "there are three apples" would use the plural "apples" instead of the singular "apple", the Korean sentence 사과가 세 개 있습니다 sagwaga se gae isssumnida "apple three(things) exist" keeps the word 사과 sagwa "apple" in its unmarked form, as the numeral makes the plural marker redundant.

The most basic, fundamental Korean vocabulary is native to the Korean language, e.g. 나라 (nara, country), 날 (nal, day). However, a large body of Korean nouns stem from the Korean pronunciation of Chinese characters e.g. 산(山) san, "mountain," 역(驛) yeok, "station," 문화(文化) munhwa, "culture", etc. Many Sino-Korean words have native Korean equivalents and vice versa, but not always. The choice of whether to use a Sino-Korean noun or a native Korean word is a delicate one, with the Sino-Korean alternative often sounding more profound or refined.

For a list of Korean nouns, see wikt:Category:Korean nouns. ]

Pronouns

Korean pronouns (대명사, daemyeongsa, 代名詞) are highly influenced by the honorifics in the language. Pronouns change forms depending on the social status of the person or persons spoken to, e.g. the pronoun for "I" there is both the informal 나 (na) and the honorific/humble 저 (jeo). In general second person singular pronouns are avoided, especially when using honorific forms. For a larger list of Korean pronouns, see wikt:Category:Korean pronouns.

Numbers

Main article: Korean numerals

Korean numbers or numerals (수사, susa, 數詞) constitute two regularly used sets: a native Korean set and a Sino-Korean set. The Sino-Korean system is nearly entirely based on the Chinese numerals. The distinction between the two numeral systems is very important. Everything that can be counted will use one of the two systems, but seldom both. The grouping of large numbers in Korean follow the Chinese tradition of myriads (10000) rather than thousands (1000) as is common in Europe and North America.

Verbs (broadly speaking)

Main article: Korean verbs

Processual verbs

Korean 동사(動詞) dongsa, which include 쓰다 (sseuda, "to use") and 가다 (kada, "to go"), are usually called, simply, "verbs." However, they can also be called "action verbs" or "dynamic verbs," because they describe an action, process, or movement. This distinguishes them from 형용사(形容詞) hyeongyongsa.

Korean verb conjugation depends upon the tense, aspect, mood, and the social relation between the speaker, the subject(s), and the listener(s). Different endings are used depending on the speaker's relation with their subject or audience. Politeness is a critical part of Korean language and Korean culture; the correct verb ending must be chosen to indicate the proper degree of respect or familiarity for the situation.

Descriptive verbs

형용사(形容詞) Hyeongyongsa, sometimes translated as "adjectives" but also known as "descriptive verbs" or "stative verbs," are verbs such as 예쁘다 yeppeuda, "to be pretty" or 붉다 bukda, "to be red." English does not have an identical grammatical category, and the English translation of a Korean hyeongyongsa is usually a linking verb + an English adjective. However, some Korean words which do not match that formula, such as 아쉽다 aswipda, a transitive verb which means to "to lack" or "to want for", are still considered hyeongyongsa in Korean because they don't involve an action. For a larger list, see wikt:Category:Korean adjectives.

Copula

The copula clitic i 이 may be historically related to the nominative case clitic i 이. Regardless, nouns do not take the case clitic i/ka -ga when followed by the copula. The copula inflects like any verb, except that it has a special honorific form.

The copula takes the negative prefix an 안, but the result is written as if it were a single morpheme: 아니 ani. Nouns do take the nominative clitic i/ka 이/가 -ga before the negative copula. The derived form ani yo 아니요 is the word for "no" when answering a question. (In the case of a negative question, ani yo is equivalent to "yes" in English.)

The copula is only for "to be" in the sense of "A is B". For existence, Korean uses the verbs iss- 있 iss-/it- "there is" and eps- 없 eobs- "there isn't." The honorific existential verb for iss- is kyesi- 계시 gyesi-.

Attributive verbs

Korean does not have relative pronouns. Instead, attributive verbs modify nouns, as adjectives do in English. Where in English one would say "I saw the man who walks the dog", the structure of Korean is more like "The dog-walking man I saw".

The structure is ROOT + valence + attributive suffix, with little of the complexity of finite verbs above.

Attributive verb template

Prefix	0	I	II	III
negative	ROOT	valency	tense	attributive (tense)

Active verbs use the attributive suffix -un 은 -eun after a consonant, or -n ㄴ after a vowel, for the past tense. For descriptive or stative verbs, often equivalent to adjectives in English, this form is used for generic (gnomic) descriptions; effectively, "eaten food" is food which once was eaten (past), whereas "a pretty flower" is a flower which has become pretty, and still is (present/timeless). To specify the on-going action for an active verb, the invariable suffix -nun 는 -neun is used instead. This is not found on descriptive verbs, as it makes no sense to say that *"a flower is being pretty". For the future, the suffix -(u)lq 을/ㄹ (-(eu)l with reinforcement of the following consonant) is used, and in the imperfective/retrospective (recalling what once was) it is -ten 던 -deon.

For example, from the verb mek 먹 "to eat", the adjective yeppu 예쁘' "pretty", and the nouns pap 밥 bap "food" and kkoch 꽃 kkot "flower", we get:

Attributive forms

	Active verb			Descriptive verb
Present progressive	먹는 밥	meongneun bab	"food which is being eaten"
Perfective	먹은 밥	meogeun bab	"eaten food (food which was eaten)"	예쁜 꽃	yeppeun kkot	"a pretty flower"
Imperfective	먹던 밥	meokdeon bab	"food which one used to eat"	예쁘던 꽃	yeppeudeon kkot	"a flower which was once pretty"
Future	먹을 밥	meogeul ppap	"food to be eaten"	예쁠 꽃	yeppeul kkot	"a flower which will be pretty"

The perfective suffix -e-ss 었 -eoss- is sometimes used as well, on active verbs. It precedes the attributive suffix:

• 먹었던 밥 meogeotdeon bap "food which had been eaten"

For action verbs, -e-ss -었 is used for completed actions or processes that result in a present state. The individual verb’s meaning can help determine which interpretation is appropriate. Hence kyel.hon-hay-ssta 결혼했다 gyeorhon haetta can mean ‘got married,’ focusing on the past event, or ‘is married,’ focusing on the present state resulting from the past event. But kong ul cha-ssta 공을 찼다 gongeul chatta ‘kicked the ball’ can only denote a past action and cal sayngkye-ssta 잘 생겼다 jal saenggyeotta ‘is handsome’ can only denote the present state. (sayngkita 생기다 saenggida is an action verb, meaning ‘get formed/ created.’)

Modifiers

Determinatives

Korean 관형사(冠形詞) gwanhyeongsa are known in English as "determiners," "determinatives," "pre-nouns," "adnouns," "attributives," "unconjugated adjectives," and "indeclinable adjectives." Gwanhyeongsa come before and modify or specify nouns, much like attributive adjectives or articles in English. Examples include 각(各) kak, "each." For a larger list, see wikt:Category:Korean determiners.

Adverbs

Korean adverbs (부사, busa, 副詞) include 또 (tto, "also") and 가득 (gadeuk, "fully"). Busa, like adverbs in English, modify verbs. For a larger list, see wikt:Category:Korean adverbs.

Other content words

Exclamations

Korean interjections (감탄사, gamtansa, 感歎詞) are also known in English as "exclamations". Examples include 아니 (ani, "no"). For a larger list, see wikt:Category:Korean interjections.

Function words

Particles

Main article: Korean particles

Korean particles (조사, josa, 助詞) are also known in English as "postpositions" or case markers. Examples include 는 (neun, topic marker) and 를 (reul, object marker). Particles come after nouns and are used to indicate the role (subject, object, complement, or topic) of a noun in a sentence or clause. For a larger list, see wikt:Category:Korean particles.

A special particle is ida (이다), a "predicative particle" (seosulgyeok josa; 서술격 조사) that behaves much like the English copula "be" (in joining subjects to their complements) and is therefore called a verb in English sources, though it is not regarded as a dongsa in Korean. 이다 cannot be negated by adding the prefix 안; instead it has a special negative form 아니다.

Cases

Both nouns and pronouns take case clitics. Pronouns are somewhat irregular, and are covered above. As with many clitics and suffixes in Korean, for many case clitics different forms are used with nouns ending in consonants and nouns ending in vowels. The most extreme example of this is in the nominative (subject), where the historical clitic i 이 is now restricted to appearing after consonants, and a completely unrelated (suppletive) form -ka (pronounced -ga) appears after vowels.

Case clitics

Case	After V	After C
Nominative	ka 가 -ga	-i 이
Accusative	lul 를 -reul	ul 을 -eul
Genitive	-uy 의 -e1
Dative (also destination)	-ey 에 -e(inanimate) -uy key2 에게 -ege (animate)
Locative (place of event, also source)	-ey se에서 -eseo (inanimate) -uy key se2 에게서 -egeseo (animate)
Instrumental	-lo 로 -ro3	-ulo 으로 -euro
Comitative (also and)	-hako 하고 -hago
	-wa 와	kwa 과 -gwa
	lang 랑 -rang	-i lang 이랑 -irang

1 -uy 의 is a morphophonemic spelling, which is pronounced the same as 에, -e.

2 -uy key and -uy key se are compounds. Key is a noun meaning "neighborhood, behalf." Standard Korean spelling treat them as opaque, but here it is broken down etymologically.

3 -lo also occurs with stems ending in ㄹ l.

Information clitics

Type	After V	After C
Topic*	nun 는 neun	un 은 -eun
Also*	to 도 -do
And (and so on)	na 나 -na	-i na 이나 -ina

* The topic marker and the also marker mark the noun phrase with case markers. They override the nominative and accusative case markers rather than being attached after those case markers.

Syntax

Korean is typical of languages with a verb-final word order, such as Japanese and Turkish, in that most affixes are suffixes and clitics are enclitics, modifiers precede the words they modify, and most elements of a phrase or clause are optional.

Number

(It is against grammatical rules for the plural marker '-deul' to occur at some of the alternative positions given in the following examples, and it is highly uncommon or at least somewhat unnatural for the most of them. Please review the grammatical consistency of this subsection.)

Korean has general number. That is, a noun on its own is neither singular nor plural. It also has an optional plural marker tul 들 -deul, which is most likely to be used for definite and highly animate nouns (primarily first- and second-person pronouns, to a lesser extent nouns and third-person pronouns referring to humans, etc.) This is similar to several other languages with optional number, such as Japanese.

However, Korean tul may also be found on the predicate, on the verb, object of the verb, or modifier of the object, in which case it forces a distributive plural reading (as opposed to a collective reading) and indicates that the word it is attached to expresses new information.

For instance, in

아이들이	김에게	빵을	많이	주었어요
ai tuli	kimuy key-0	ppang ul-0	manhi-0	cwue-ss-e-0
aideuri	gimege	bbangeul	manhi	jueosseo
child-pl-NOM	Kim-to	bread-ACC	a_lot	give-PRET-INT,

Tul could also occur at any of the 0's. If it did, it would not be redundant with the plural marking on the subject. For example, in,

학생들이	풍선	하나를	샀어요
haksayng tuli	phungsen	hana lul	sa-ss-e-yo
haksaengdeuri	pungseon	hanareul	sasseoyo
student-pl-NOM	balloon	one-ACC	buy-PRET-INT-POL

"The students bought a balloon",

it's not clear if they bought one balloon together, or one each. However, without the ACC on "one",

학생들이	풍선을	하나	샀어요
haksayng tuli	phungsen ul	hana	sa-ss-e-yo
haksaengdeuri	pungseoneul	hana	sasseoyo
student-pl-NOM	balloon-ACC	one	buy-PRET-INT-POL

"The students bought a ballon together"

and, with ssik ("each") on "one"

학생들이	풍선을	하나씩	샀어요
haksayng tuli	phungsen ul	hana ssik	sa-ss-e-yo
haksaengdeuri	pungseoneul	hanassik	sasseoyo
student-pl-NOM	balloon-ACC	one-each	buy-PRET-INT-POL

"The students bought a balloon each",

"balloon" is specified as a distributive plural.

Subject-verb agreement

While it is usually stated that Korean doesn't have subject-verb agreement, the conjugated verbs do, in fact, show agreement with the logical subject (not necessarily the grammatical subject) in several ways. However, subject-verb agreement in Korean usually only narrows down the range of subjects. If the logical subject of a sentence is narrowed down enough to be obvious from verb endings, then it is ungrammatical to redundantly state it explicitly. Personal agreement is shown partly on the verb stem before the tense-aspect-mood suffixes, and partly on the sentence-final endings.

Korean distinguishes:

• Honorific subjects from non-honorific subjects in the second or third person via a verb suffix.
• Korean distinguishes first person from non-first in emotion verbs, which take the verb -hada as a suffix in the third person.
• first person from third person, partially, in the future and the past tense.
• inclusive first person from exclusive first person, and first person from third person, in the jussive mood

Korean does not distinguish:

• singular from plural on the verb (though this is systematically marked on pronouns)
• second person from third person in statements
• second person from first person in questions

The following table is meant to indicate how the verb stem and/or the sentence ending can vary depending on the logical subject. The row labeled "1st" contains the form used for the 1st person statements (also used for the 2nd person in questions). The row labeled 3rd contains the forms used for the other persons. The column labeled "emotion verbs" illustrates the paradigm of muesi silh-ta 무엇이 싫다 mueosi silta, a typical emotion verb phrase meaning "to dislike something". The column labeled "jussive ending" contains the various jussive sentences endings in the plain style. The difference between -(u)m a 마 and ca 자 ja is clusivity; -(u)ma is exclusive of the listener, -ca is inclusive.

Logical person		Person agreement on stem Emotion verbs	Person agreement on final ending Jussive ending
1st Person		muesi silh- 무엇이 싫	-(u)m a 마 -ca 자
3rd Person	Non-honorific	muesul silhe ha- 무엇을 싫어하	-e/a la 아라/어라
	Honorific	muesul silhe hasi- 무엇을 싫어하시	-(u)sye la 셔라

Subordinate Clauses

Verbs can take conjunctive suffixes. These suffixes make subordinate clauses.

One very common suffix -ko 고 -go, can be interpreted as a gerund if used by itself, or, with a subject of its own, as a subordinating conjunction. That is, mek.ko 먹고 meokgo means approximately "eating," koki lul mek.ko 고기를 먹고 gogireul meokgo means "eating meat," and nay ka koki lul mek.ko 내가 고기를 먹고 nae-ga gogi-rul meog-go means "I eat meat and..." or "My eating meat."

Another suffix, somewhat similar in meaning, is se 서 -seo which is, however, attached to long stem of a verb. The long stem of a verb is the one that is formed by attaching e/a 어/아 -eo/-a after a consonant.

Both sometimes called gerunds, the verb form that ends in se and the one that ends in -ko juxtapose two actions, the action in the subclause and the action in the main clause. The difference between them is that with se the action in the subclause necessarily came first, while -ko conveys more of an unordered juxtaposition. Se is frequently used to imply causation, and is used in many common expressions like manna se pan.gapsupnita 만나서 반갑습니다 Manna-seo bangapseumnida (literally, "Since I met you, I'm happy" -or- "Having met you, I'm happy"). If -ko was used instead, the meaning would be closer to "I meet you and I'm happy," that is, without any implied logical connection.

These are both subordinating conjunctive suffixes and can't (in the more formal registers, at least) derive complete sentences of their own without the addition of a main verb, by default the verb is 있. Nay ka koki lul mek.ko issta 내가 고기를 먹고 있다 naega gogireul meokko itta therefore means "I am eating meat." The difference between this and the simple sentence nay ka koki lul meknun ta 내가 고기를 먹는다 is similar to the difference in Spanish between "Estoy almorzando" and "Almuerzo," in that the compound form emphasizes the continuity of the action. The -se 서 form is used with the existential verb iss 있 for the perfect.

See also

• Korean particles