by Nam-Kil Kim
In: Bernard Comrie (ed.). The World's Major Languages. London: Croom Helm, pp. 881-898
1. Historical Background
For a long time scholars have tried to associate the Korean language to one of the major language families but have not been successful in this venture. There have been many theories proposed on the origin of Korean. Based on the views as to where the Korean language first originated, two prominent views, which are called the Southern theory and the Northern theory, have been advocated by some scholars. According to the Southern theory, the Korean people and language originated in the south, namely the South Pacific region. There are two versions of this theory. One is that the Korean language is related to the Dravidian languages of India. This view is not taken seriously by contemporary linguists, but it was strongly advocated by the British scholar Homer B. Hulbert at the end of nineteenth century. Hulbert's argument was based on the syntactic similarities of Korean and the Dravidian languages. For instance, both languages have the same syntactic characteristics: the word order subject-object-verb, postpositions instead of prepositions, no relative pronouns, modifiers in front of the head noun, copula and existential as two distinct grammatical parts of speech etc.
The other version of the Southern theory is the view that Korean may be related to the Austronesian languages. There are some linguistic as well as anthropological and archeological findings which may support this view. The
linguistic features of Korean which are shared some Polynesian languages include the phonological structure of open syllables, the honorific system, numerals and the names of various body parts. The anthropological and archeological elements shared by Koreans and the people in other regions of the South Pacific are rice cultivation, tattooing, a matrilineal family system, the myth of an egg as the birth place of royalty and other recent discoveries in paleolithic or preceramic cultures. Although this Southern theory has been brought to the attention of many linguists, it is not accepted as convincing by linguists.
The Northern theory is the view that Korean is related to the Altaic family. Although this view is not wholly accepted by the linguistic community, the majority of Korean linguists and some western scholars seem inclined towards believing this view. The major language branches which belong to the Altaic family are Turkic, Mongolian and Tungusic. The area in which the Altaic languages are spoken runs from the Balkans to the Kamchatka Peninsula in the North Pacific. The Northern theory stipulates that the Tungusic branch of Altaic tribesmen migrated towards the south and reached the Korean peninsula. The Tungusic languages would include two mayor languages: Korean and Manchu. The view that Korean is a branch of the Altaic family is supported by anthro-archeological evidence such as comb ceramics (pottery with comb-surface design), bronze-ware, dolmens, menhirs and shamanism. All these findings are similar to those found in Central Asia, Siberia and northern Manchuria. Korean is similar to the Altaic languages with respect to the absence of grammatical elements such as number, genders, articles, fusional morphology, voice, relative pronouns and conjunctions. Vowel harmony and agglutination are also found in Korean as well as in the Altaic languages. Comparing the two theories, it is apparent that the Northern influence in the Korean language is more dominant than the Southern.
It has been discovered in recent archeological excavations that the early race called Paleosiberians lived in the Korean peninsula and Manchuria before the Altaic race migrated to these areas. The Paleosiberians, who include the Chukchi, Koryaks, Kamchadals, Ainu, Eskimos etc., were either driven away to the farther north by the newly arrived race or assimilated by the conquerors when they came to the Korean peninsula. It is believed that the migration of the new race towards the Korean peninsula took place around 4000 BC. Nothing is known about the languages of the earliest settlers. After migration, some ancient Koreans settled down in the regions of Manchuria and northern Korea while others moved farther to the south. Many small tribal states were established in the general region of Manchuria and the Korean peninsula from the first century BC to the first century AD. The ancient Korean language is divided into two dialects: the Puyo language and the Han language. The Puyo language was spoken by the people of tribal states such as Puyo, Kokuryo, Okcho and Yemaek in
Manchuria and northern Korean. The Han language was spoken by the people of the three Han tribal states of Muhan, Chinhan and Byonhan which were created in southern Korea.
Around the fourth century AD the small tribal states were vanquished and three kingdoms with strong central governments appeared in Manchuria and the Korean peninsula. Of these three kingdoms, the biggest kingdom, Kokuryo, occupied the territory of Manchuria and the northern portion of the Korean peninsula. The other two kingdoms, Paekche and Silla, established states in the southwestern and the southeastern regions of the Korean peninsula respectively. It is believed that the Kokuryo people spoke the Puyo language and the Silla people spoke the Han language; however, it is not certain what language the Paekche people spoke because the ruling class of the Paekche kingdom consisted of Puyo tribesmen who spoke the Puyo language. When the Korea peninsula was unified by Silla in the seventh century, the Han language became the dominant dialect paving the way for the emergence of a homogeneous language. The Han language finally became the sole Korean language through the two succeeding dynasties of Koryo (936-1392) and Choson (1392-1910).
Since Silla's unification of the Korean peninsula in the seventh century, it appears that the language spoken in the capital has been the standard dialect. Thus, the Silla capital, Kyongju, dialect was the standard dialect during the unified Silla period from the seventh century to the tenth century. When Silla was succeeded by Koryo in the tenth century, the capital was moved from Kyongju, which was located in the southeastern region of the Korean peninsula, to Kaegyong in the central region of Korea and subsequently the dialect spoken in this new capital became the standard language in Koryo from the tenth century to the end of the fourteenth century. When the Yi (or Choson) Dynasty succeeded Koryo at the end of the fourteenth century, the capital was established at Seoul, the present capital of South Korea, and the language spoken in this area became the standard dialect and has continued as a standard dialect to the present time. Thus, it is obvious that the formation of the standard dialect has been dominated by political decisions. We can find this even in the twentieth century. There are officially two standard dialects existing in Korea; one is the Seoul dialect in South Korea and the other the Phyong'yang dialect in North Korea. Each government has established prescriptive criteria for its own standard dialect and made separate policies on language.
Though the dialect distinction of one region from the other is not drastic owing to the relatively small size of the Korean peninsula, each region has its own characteristic dialects. For instance, in the Hamgyong dialect of northern Korea the final p. of verb bases ending in p is pronounced as [b] before suffixed morphemes starting in a vowel, while in the standard Seoul dialect this final p is pronounced as [w] before a vowel; top- 'hot' is pronounced [tabo] in the Hamgyong dialect but [towo] in the standard
dialect. As another example, in the standard dialect palatalisation is normal but in the Phyong'yang dialect palatalisation does not take place: kathi 'together' is pronounced as [kachi] in the standard dialect but as [kathi] in the Phyong'yang dialect. Historically, both Hamgyong and Phyong'yang dialects reflect archaic forms. That is, in the nineteenth-century Yi Dynasty language the words top- and kathi were pronounced as they are pronounced in the Hamgyong and Phyong'yang dialects; and the pronunciation of these words in the standard dialect reflects this historical change.
The Korean language spoken before the fifteenth century is not well known because there are not many records or documents revealing how the language was used before the fifteenth century. It was in the fifteenth century that the alphabetic script (Han'gul) for writing Korean was invented by King Sejong. Before the Korean script was invented, only Chinese characters were used for the purpose of writing. But Chinese characters could not depict the living language spoken by Korean people, since Chinese characters were meaning-based and the grammar of classical Chinese did not have any connection with Korean grammar. Even after the Korean script was invented, Chinese characters were continuously used as the main means of writing until the twentieth century. In traditional Korean society, the learning and study of Chinese characters and classical Chinese were entirely monopolised by a small class of elite aristocrats. For average commoners, the time-consuming learning of Chinese characters was not only a luxury but also useless, because they were busy making a living and knowledge of Chinese characters did not help in improving their lives.
The use of Chinese characters imported a massive quantity of loanwords into the Korean lexicon. More than half of Korean words are Chinese-originated loanwords. Although Chinese loanwords and Korean-originated words have always coexisted, the Chinese loanwords came to dominate the original Korean words and subsequently many native Korean words completely vanished from use. A movement by people who wanted to restore native culture at the end of the nineteenth century tried to stimulate mass interest in the study of the Korean language. When the government proclaimed that the official governmental documents would he written both in Korean script and Chinese characters, the first newspapers and magazines were published in Korean script and the use of the Korean alphabet expanded. In the early twentieth century, more systematic studies on the Korean language were started and a few scholars published Korean grammar books. However, the active study of Korean grammar was discontinued owing to the Japanese colonial policy suppressing the study of Korean.
The study of the Korean language resumed after the end of World War II, but Korea was divided into two countries by the Big Powers. The language policies proposed and implemented by the two governments in the South and the North were different from each other. While both the Korean
alphabet and Chinese characters were used in the South, only the Korean alphabet was used in the North. In the North the policy on the use of Chinese characters has been firm; that is, no instruction in Chinese characters has been given to students and Chinese characters are not used in newspapers, magazines or hooks. This policy has never been changed in the North. Contrary to this, in the South the policy on the instruction of Chinese characters has been inconsistent; whenever a new regime has come to power, both proponents and opponents of the use of Chinese characters have tried to persuade the government to adopt their views. Though the instruction of Chinese characters was abolished a couple of times by the government in the past, this abolition never lasted more than a few years. At the present time in the South, the government has adopted a policy which forces students in secondary schools to learn 1,800 basic Chinese characters. The South and the North also have different policies on the so-called 'purification' of Korean. The purification of Korean means the sole use of native Korean words in everyday life by discontinuing the use of foreign-originated words. The main targets of this campaign are Sino-Korean words. In the North, the government has been actively involved in this campaign, mobilising newspapers and magazines to spread the newly translated or discovered pure Korean words to a wide audience of readers. In the South, some interested scholars and language study organisations have tried to advocate the purification of Korean through the media and academic journals, but the government has never officially participated in this kind of movement. It will be interesting to see what course each of the two governments will take in future with respect to language policy.
The sound system of Korean consists of 21 consonants and ten vowels. The vowels can he classified by the three positions formed by the vocal organs. The first is the height of the tongue, the second is the front or the back of the tongue and the third is the shape of the lips. The vowel systems of Korean can be represented as in table 44.1.
Table 44.1: Korean Vowels
Front Back Unrounded Rounded Unrounded Rounded High i ü u u Mid e ö o o Low æ a
The vowels /ü/ and /ö/ have free variants [wi] and [we] respectively; thus, [kü] 'ear' is pronounced either [kü] or [kwi] and /kömul/ 'strange creature' is
pronounced either [kömul] or [kwemul]. The vowel u is always pronounced [u] after labial sounds; sulphuda is pronounced as [sulphuda] and kamum 'draught' is pronounced as [kamum].
Korean has a large number of morphophonemic alternations. As major examples of Korean morphophonemic processes involving vowels, we can name the following kinds: vowel harmony, glide formation, vowel contraction and vowel deletion. When non-finite endings starting with a are attached to verbal bases, the initial o of the ending is changed to a after a and o as in nok-oso 'melting' --> [nokaso], and mac-oso 'be hit' --> [macaso]; elsewhere o is not changed as in mok-oso 'eat' --> [mokoso], kiph-oso 'deep' --> [kiphoso], kæ-oso 'clear' --> [kæoso] and so on.
The vowels o, i and a undergo vowel contraction with the vowel i when the vowels in verbal bases and other morphemes such as the causative and passive are combined with each other. Korean has the following kinds of vowel contraction; o + i --> ö: po-i-ta 'be seen' --> [pöta]; o + i --> e: so-iu-ta 'raise' --> [seuta]; a + i --> æ: ca-iu-ta 'make sleep' --> [cæuta]; u + i --> ü: pak'tu-i-ta 'be changed' [pak'üta].
The front vowel i and the back vowels u and o of verbal bases undergo glide formation when they are immediately connected to o or a of suffixes such as -o and -oso; i becomes y and u and o become w: ki-oso 'crawl' --> [kyoso], tu-oso 'leave' --> [twoso] and po-aso --> [pwasa]. As examples of vowel deletion, Korean has two kinds: u-deletion and o-deletion. When verbal bases ending in the vowel u are attached to an ending starting with the vowel o, the vowel u is deleted: s'u-o 'write' [s'o] and k'u-oso 'extinguish' [k'oso].
Finally, o-deletion occurs when endings starting with the vowel o are combined with verbal bases ending with the vowels e, æ, o and a; thus we have the following examples: se-os'ta 'counted' --> [ses'ta], kæ-oso 'clear' --> [kæso], so-o 'stand' --> [so] and ka-oto 'even if he goes' --> [katol. Interestingly, the vowels which force o-deletion are those vowels which do not undergo either glide formation or u-deletion, i.e. i, u and o undergo glide formation; ü and ö have free variant forms [wi] and [we] respectively as in tü-os'-ta 'jumped' --> [twios'ta] and k'ö-os'ta 'lure' --> [kweos'ta]; u is deleted before o. From the above discussion of glide formation and vowel deletion, we can see that all the vowels in the Korean vowel system participate in phonological processes without exception when verbal bases are combined with suffixes starting in o.
Of the 21 consonants, there are 9 stops, 3 affricates, 3 fricatives, 3 nasals, 1 liquid and 2 semi-vowels. The Korean consonants can be illustrated as in table 44.2.
Let us now briefly describe the sound of Korean obstruents (stops, affricates and fricatives). The Korean laxed obstruents are weaker than English voiceless obstruents with respect to the decree of voicelessness. This seems to be due to the fact that Korean obstruents have two other stronger
Manner Point Labial Dental Palatal Velar Glottal Stops voiceless laxed p t k aspirated ph th kh tensed p' t' k' Affricates voiceless laxed c aspirated ch tensed c' Fricatives voiceless laxed s h tensed s' Nasals voiced m n ng() Liquid voiced l Semi-vowels w y
voiceless consonants, the tensed and the aspirated. The laxed obstruents are produced without voice and without aspiration and glottal tension. However, the laxed stops and affricates /p,t,k,c/ are pronounced as the voiced obstruents [b,d,g,j] when they occur between two voiced sounds. Even if some voiceless obstruents have voiceless allophones, Korean speakers are not aware of this change. For instance, the word /aka/ 'baby' is pronounced [aga] because /k/ occurs between two vowels, which are voiced sounds.
The Korean aspirated obstruents are produced with stronger aspiration than English aspirated sounds. The Korean tensed obstruents are one of the most peculiar sounds among Korean consonants. The tensed obstruents are produced with glottal tension, but these sounds are not glottal sounds or ejectives. For instance, the Korean /t'/ is phonetically similar to the sound [t'] in English which is pronounced after [s] in the word stop [stap]; however, the Korean tensed obstruent must be pronounced with more glottal tension.
Liquids and semi-vowels need some explanation. The Korean liquid /l/ has two variants; one is the lateral  and the other the flap [r]. The liquid /l/ is pronounced as the lateral  in word-final position and in front of another consonant, and as the flap [r] in word-initial position and between two vowels. The Korean semi-vowels, /w/ and /y/, occur only as on-glides, never as off-glides.
In the above, the general qualities of Korean consonants were briefly described. Let us now discuss some of the phonological processes affecting Korean consonants. In pronunciation, consonants are always unreleased in word-final position and before another obstruent. Because of this, consonants belonging to a given phonetic group are pronounced identically in word-final position and before other obstruents. For instance, the labial stops /p/, /ph/ and /p'/ are all neutralised into /p/ in word-final position; in the same manner, the velar stops /k/, /kh/ and /k'/ are neutralised to /k/. The largest group of consonants comprises the dental and palatal obstruents
which are pronounced /t/: /t/, /th/, /t'/, /c/, /ch/, /c'/, /s/ and /s'/. When examining consonant clusters, it is found that only single consonants occur in both initial and final position of words. Consonant clusters occur only in medial position in words and only clusters of two consonants are permitted to occur there. Some words have final two-consonant clusters in their base forms, but only one consonant is pronounced and the other consonant is deleted; for instance, the word /talk/ 'chicken' has the lk cluster in its base but it is pronounced [tak], losing l when it is pronounced alone. However the cluster lk occurs in intervocalic positions: /talk/ + /i/ --> [talki]. When the final cluster occurs before a consonant, again one consonant must be deleted as in word-final position to obey the two consonant constraint, e.g. /talk/ 'chicken' + /tali/ 'leg' [taktali] 'chicken leg'.
One of the most interesting characteristics of Korean phonology is its rich consonant assimilation. Korean consonant assimilation comprises nasalisation, labialisation, dentalisation, velarisation, palatalisatioll and liquid assimilation. Of these, nasalisation is the most productive; for instance, the stops k, t and p (including the neutralised stops) become ng, n and m respectively before nasals: /kukmul/ 'soup' --> [kungmul], /patnunta/ 'receive' --> [pannunda], and /capnunta/ 'catch' --> [camnunda]. As another example of nasalisation, the liquid l becomes n after the nasals m and ng and the stops k, t, p: e.g. /kamlo/ 'sweet' --> [kamno], /pækli/ 'one hundred li' --> [pækni], /matlyangpan/ 'first son' --> [matnyangpan] and /aplyok/ 'press' --> [apnyok]. Interestingly, in the last three examples, the stop sounds which caused l to nasalise assimilate to the following new nasals and become nasals themselves: [pækni] --> [pængni], [matnyangpan] --> [mannyangpan] and [apnyok] --> [amnyok]; thus these three examples undergo two nasalisation processes.
The consonant h behaves interestingly in medial positions; when h occurs in intervocalic position, it is deleted: /cohun/ 'good' --> [coun]; and when h occurs before laxed stops and affricates, metathesis takes place as in the following example: /hayah-ta/ 'white' --> [hayatha] --> [hayatha]. As another example of consonant deletion, we can name l-deletion: when the consonant l occurs in the initial position, it is deleted: /lyangpan/ 'aristocrat' --> [yangban]. However, l-deletion in the initial position is not absolute, because l is changed to n in the same position depending on the following vowels, as in /lokuk/ 'Russia' --> [noguk]. Thus, the right way of explaining the l phenomena would be to say that the consonant l does not occur in initial position.
Thus far, we have seen the Korean phonemic system and some of its I phonological processes. In the remaining portion of this section, the Korean I writting system will be briefly presented.
As can be seen from table 44.3, the Korean alphabet, which is called Han'gul, consists of 40 letters: 10 pure vowels, 11 compound vowels, 14 basic consonants and 5 double consonants. The Korean writing system is
Table 44.3: Korean Alphabet
based on the 'one letter per phoneme' principle. However, comparing the number of phonemes with the number of letters, it is found that the writing system has nine more letters. This is because the diphthongs are also represented by their own letters. Thus, the semi-vowels y and w do not have their own independent letters. They are always represented together with other vowels occurring with them. For instance, the letter is a combination of y and a, and the letter is a combination of w and a.
As a general rule, in writing, Korean letters are formed with strokes from top to bottom and from left to right. The letters forming a syllable have a sequence of CV(C)(C) and they are arranged as a rebus: e.g. ka 'go' --> ; kak 'each' --> ; talk 'chicken' --> . One interesting thing about the Korean writing system is that a vowel cannot be written alone: for instance, a cannot
be written as and i cannot be written as . In the Korean writing system, the absence of a consonant is represented by a Ø consonant, which is shown by the symbol and written to the left of the vowel. Thus, a is written as and i is written as . For instance, a word ai 'child' which consists of the two vowels, a and i, is written as .
Korean words can be divided into two classes: inflected and uninflected. The uninflected words are nouns, particles, adverbs and interjections. Inflected words are classed as action verbs, descriptive verbs, copula and existential. The distinction between action and descriptive verbs can be shown by the way in which paradigmatic forms such as propositive and processive are combined with verbal forms. For instance, a descriptive verb lacks propositive and processive forms. Thus, whereas the action verb plus the propositive ca or the processive nun is grammatical, the combinations of descriptive verbs with the same endings are not: mok-ca 'let's eat' and mok-nun-da 'is eating' but *alumtap-ca 'let's be beautiful' and *alumtap-nun-da 'is being beautiful'. While the copula behaves like a descriptive verb, the existential behaves like an action verb with respect to conjugation; thus, *i-ca 'let's bet and *i-n-ta 'is beeing' are ungrammatical but is'-ca 'let's stay' and is'-nun-ta 'is staying' are grammatical.
As predictable from the above discussion, each inflected form consists of a base plus an ending. Bases and endings can be classed into groups according to the ways in which alternant shapes of bases are combined with endings. There are two kinds of ending: one-shape endings such as -ko, -ta, -ci and -kes' and two-shape endings such as -supnita/-upnita, -una/-na and -un/-n. Two-shape endings are phonologically conditioned alternants; thus, for instance, the formal form -supnita occurs only with base forms ending in a consonant, but the altemant form of the formal form /-upnita occurs only with base forms ending in a vowel. Based on these classes of endings, verb bases can be divided into two groups: consonant bases (i.e. bases ending in a consonant) and vowel bases (i.e. bases ending in a vowel). There are, however, some classes of bases whose final sounds are changed when attached to endings. Thus, in addition to regular bases which do not alter when combined with the ending, there are about five classes of consonant bases which alter with the ending: bases ending in t, bases ending in w, bases ending in h, bases ending in sonorants and s-dropping bases. Vowel bases have three classes in addition to the regular vowel bases: l-extending vowel bases, l-doubling bases and l-inserting vowel bases.
In order to see how the base form is changed when it is attached to the endings, the partial conjugation of regular and irregular bases ending in t is illustrated:
Irregular Regular Base /mut-/ 'ask' /tat-/ 'close' Gerund [muk-ko] [tak-ko] Suspective [muc-ci] [tac-ci] Formal Statement [mus-sumnita] [tas-sumnita] Infinitive [mul-o] [tat-o] Adversative [mul-una] [tat-una]
When comparing the two base forms, mut 'ask' and tat 'close', ending in t, it is found that both forms undergo morphophonemic changes when combined with endings starting with a consonant. These morphophonemic changes are phonologically conditioned; t is changed to k before k; t is changed to s before s and so on. However, t is changed to l before vowels only in the base mut, but not in the base tat.
Below, the partial conjugation of an l-inserting vowel base is illustrated together with the conjugation of an ordinary vowel base.
Irregular Regular Base /phulu-/ 'be blue' /t'alu-/ 'close' Gerund [phulu-ko] [t'alu-ko] Suspective [phulu-ci] [t'alu-ci] Formal Statement [phulu-mnita] [t'alu-mnita] Infinitive [phulul-o] [t'al-o] Adversative [phulul-una] [t'al-una]
In regular vowel bases such as t'alu-, the final vowel is deleted when attached to the endings starting with a vowel, as shown in the conjugation of the infinitive and the adversative. However, in the case of irregular vowel bases such as phulu-,l is inserted before the same endings.
The number of endings which can he attached to the base is said to be over 400. In finite verb forms, there are seven sequence positions where different endings can occur: honorific, tense, aspect, modal, formal, aspect and mood. The honorific marker si (or usi) is attached to the base to show the speaker's intention or behaviour honouring the social status of the subject of the sentence. Tense has marked and unmarked forms; the marked form is past and the unmarked form present. The past marker os'/s' has the meaning of a definite, completed action or state.
Aspect occurs in two different positions because there are two different aspects: experiential-contrastive and retrospective, which are mutually exclusive, i.e. if one occurs, then the other cannot. The experiential-contrastive os'/s', which has the same form as the past tense marker and only occurs after the past tense, has been called 'the double past'. The two sentences, John i hakkyo e ka-s'-ta and John i hakkyo e ka-s'-os'-ta, are usually translated in the same way as 'John went to school'. However. this does not mean that they have the same meaning. To translate them more precisely, the first sentence merely indicates the fact that the subject has
gone to school and is there now. But the second sentence has the meaning that the subject has had the experience of being in school or that he has been in school before but has come back to the place where he is now. Thus, the two sentences have quite different meanings. Only the second sentence has an aspectual meaning of experiential-contrastive.
The retrospective to indicates that the speaker recollects what he observed in the past and reports it in the present situation. The sentence John i cip e ka-to la has roughly the meaning 'I observed that John was going home and now I report to you what I observed'.
The modal kes' has the meaning indicating the speaker's volition or supposition and is used both for a definite future and a probable present or past. When the modal kes' is attached to a verb whose subject is first person the sentence only has the volitional meaning and is used only with reference to the future: næ ka næil ka-kes'-ta 'I will go tomorrow' but * næ ka næil ka-s'-kes'-ta. When the modal kes' occurs in a sentence whose subject is second or third person, the sentence has only the suppositional meaning and is used for both a definite future and a probable present or past: Mary ka næil kakes'-ta 'I suppose that Mary will go tomorrow' and Mary ka oce ka-s'-kes'-ta 'I suppose that Mary left yesterday'.
The formal form supni/pni is used for the speaker to express politeness or respect to the hearer: onul top-supni-ta 'it is hot today' and onul top-ta 'it is hot today'. The only difference between the two sentences is the presence or absence of the polite form supni in the verbal form. The first sentence could he used for addressing those whose social status is superior to the speaker's but the second sentence would be used for addressing one who is inferior or equal to the speaker in social status (here, social status includes social position, age, sex, job etc.).
Among a large number of mood morphemes, the most typical moods are declarative, interrogative, imperative and propositive. In Korean, sentence types such as declarative, interrogative, imperative and propositive sentences are identified by the mood morphemes: ta, k'a, la and ca. These mood morphemes occur in the final position of finite verbal forms, e.g. declarative: ka-pni-ta 'he is going'; interrogative: ka-pni-k'a? 'is he going?'; imperative: ka-la 'go'; propositive: ka-ca 'let's go'.
Passive and causative verbal forms can be derived by adding suffixes to bases. There are a number of passive and causative suffixes such as i, hi and li which have common shapes. Generally, causative suffixes can be divided into three groups according to the vowel in the suffix: i-theme causatives, u-theme causatives and æ-theme causatives. Passive suffixes can be grouped with the i-theme causative because their theme vowel is only i. Because both causative and passive suffixes have identical shapes, homonymous causative and passive verbal forms are frequently produced from the same base: k'ak'-i 'cause to cut' and k'ak'-i 'be cut' from the base k'ak' 'cut'; anc-hi 'seat' and anc-hi 'be seated' from anc 'sit'. Besides the causative morphemes -i-
and -hi-, there are -ki-, -ukhi-, -ikh'-, -li-, -liu- and -iu- morphemes in i-theme causatives.
In addition to lexical causatives and passives which are derived from the combination of verb bases with the causative or passive suffixes, Korean has periphrastic causatives and passives. The periphrastic causative is formed by the combination of verb base with the adverbial ending -ke followed by the verb ha 'do', e.g. ip-ke-ha-n-ta 'make (someone) put on'. Some verbs take both lexical and periphrastic causatives, but some other verbs take only periphrastic causatives. Comparing the two types of causative, periphrastic causatives are more productive than lexical causatives in Korean.
In Korean, passives are not so commonly used as in some other languages, such as English or Japanese. There are many transitive verbs which do not undergo passivisation; for instance, the verb cu 'give' does not undergo either lexical or periphrastic passivisation. Thus, the number of transitive verbs which undergo passive formation with the passive suffix is limited to a certain group of verbs. There are two kinds of verbs which undergo periphrastic passivisation: one is a group of verbs which take an inchoative verb ci and the other a group of verbs which take an inchoative verb tö in their passive formation. The passive of the first group is formed by adding the infinitive ending o to the base followed by the inchoative verb ci: pusu-o-ci-da 'be broken'. All the transitive verbs which take the inchoative verb tö in passive formation are derived from Chinese-originated loan verbs plus the verbaliser ha. In the passive formation of these verbs, the verbaliser ha is changed to the inchoative verb tö; thus, the passive of sængkakha-ta 'think' is sængkaktö-ta 'be thought'.
Finally, there are a great number of nouns which are derived from verbs by adding the nominalising morphemes to verbal bases. There are three nomilialisers ki, um/m and iwhich can he added to the base. As examples of derived nouns, we have the following: ki-derived nouns: talliki 'running', næki 'bet', chaki 'kicking' and poki 'example'; um/m-derived nouns: olum 'ice', cam 'sleep', k'um 'dream' and chum 'dance'; i-derived nouns: koli 'hanger', noli 'game', kili 'length' and nolpi 'width'. Though there is no general rule deciding which nominaliser is attached to which base, more nouns are derived from verbal bases by adding the nominalisers um/m and i than the nominaliser ki.
In this brief sketch of Korean syntax, the discussion will concentrate on representative examples which make Korean different from many Indo-European languages, especially English. One of the most frequently cited features of Korean syntax is the word order. Korean is a SOV language, meaning that the basic word order of transitive sentences is subject-object-verb. Korean has a relatively free word order compared to
English; here, the phrase 'relatively free' means that Korean a completely free word order language. The Korean language obeys a strict grammatical constraint requiring that the sentence end with a verb. As long as the sentence obeys this constraint, a permutation of the major constituents in a sentence is permissible; thus, the sentence John i Mary eke chæl ul cu-os'-ta 'John gave a book to Mary' can be said in the following ways: John i chæl ul Mary eke cu-os'-ta; Mary eke John i chæl ul cu-os'-ta; chæl ul John i Mary eke cu-os'-ta; chæl ul Mary eke John i cu-os'-ta. However, the following sentences are ungrammatical: * John i Mary eke cu-os'-ta chæl ul; * John i cu-os'-ta Mary eke chæl ul. The ungrammaticality of the last two sentences is due to the violation of the verb-final constraint.
In the above examples of Korean sentences, the grammatical elements i, eke and ul are postpositional particles corresponding to the cases nominative, dative and accusative. There are other kinds of postpositional particles such as e 'to/at', eso 'at/in', to 'also', nun 'topic', putho 'from' and k'aci 'to/till'. All these particles must occur after nouns, but some of them can occur after other particles; ice putho to ha-l-su is'-ta 'we can do it from now, too'; uli tosokwan e nun chæk i manh-ta 'in our library, there are many books'.
Comparing the Korean example with its English translation, it is found that chæk 'book' in Korean does not have any number marker, singular or plural, whereas books in the English translation has a plural marker s. This does not mean that Korean does not have a plural marker. In Korean, the plural marker attachment is not so obligatory as in English. Especially in cases where quantifiers or numerals appear in sentences as in the above example, the plural marker is usually not attached to the noun. Another characteristic of number in Korean is that the plural marker can be attached to adverbs, e.g. p'alli-tul il ul ha-n-ta 'they do work fast'. In the example, the plural marker tul attached to the adverb p'alli 'fast'. Usually, in this kind of sentence, the subject is deleted, but it is understood that the subject of the sentence is plural instead of singular owing to the presence of the plural marker on the adverb.
When nouns occur with numerals, classifiers are attached to numerals almost obligatorily. Korean has a rich system of classifiers. Each classifier is related to a class of nouns. In other words, a certain classifier occurs only with a certain class of nouns, e.g. chæk han-kwon 'one volume of a book'; mækcu tu-pyong 'two bottles of beer'; namu han-kulu 'one tree'; congi han-cang 'one piece of paper'. Another interesting thing with respect to numerals is that there is an alternative word order. Thus, the sequence of numeral + classifier, which occurs after nouns in the above examples, can also occur before nouns. When this floating takes place, the genitive particle ui is inserted between numeral + classifier and the noun: han-kwon ui chaek 'one volume of a book'; tu-pyong ui mækcu "two bottles of beer'; han-kulu ui namu 'one tree'; han-cang ui congi 'one piece of paper'.
As may have been noticed in some of the examples, deletion of subjects is allowable as long as subjects are recoverable from linguistic or non-linguistic context. Deletion of the first person and second person in Korean is especially free, as in chæk ul sa tuli-kes'-upni-ta 'I will buy you a book'; once t'ona-seyo? 'when do you leave?' In the first sentence, the first person subject is deleted and in the second, the second person subject is deleted because these subjects are recoverable in a discourse context. Although deletion of the third person subject is not so common as deletion of first and second person subjects, it is also possible: Mary ka cip e kass-ult'æ Ø uphonpætalpu lul manna-s'-ta 'when Mary went home, she met the mailman'. The zero indicates the position where the third person subject is deleted. In the last example, we discover another difference between Korean and English. In the English translation of the last Korean example, the noun mailman is preceded by the definite article the. This same noun could he preceded by the indefinite article a. This means that English has distinct definite and indefinite articles. But Korean does not have articles indicating definiteness or indefiniteness. Although definiteness is indicated by demonstratives in some cases, the distinction between definite and indefinite, in general, is not made in Korean.
Modifiers such as demonstratives, genitives, adjectives and relative clauses precede head nouns in Korean, e.g. i chæk un cæmiis'ta 'this book is interesting'; John ui aboci nun uisa-ta 'John's father is a doctor'; yep'un k'ochi is'-ta 'there is a pretty flower'; hakkyo e ka-ko is'-nun haksæng un na ui chinku-ta 'the student who is going to school is my friend'. All constituents in bold print are located to the left of the head noun. These modifying constituents make Korean a left-branching language. The notion of left-branching becomes clear in the following sentence containing three relative clauses [[[[næ ka a-nun] haksæng i tani-nun] hakkyo ka is'-nun] tosi nun khu-ta] '[the city [where the school is [where my friend goes [who I know is big]]]]'. One of the characteristics of the relative clause in Korean is that it lacks relative pronouns. Demonstratives can also be classified as one class of modifiers. Korean demonstratives have two distinct characteristics which differ from English demonstratives. First, Korean demonstratives cannot occur independently, i.e. they must occur with nouns. The second difference is that Korean demonstratives have a triple system, unlike that of English. In addition to the demonstratives 'thist and 'that', Korean has a demonstrative which has the meaning 'that over there': i 'this', ku 'that' and co 'that over there'. The same triple system is found in demonstrative locative nouns, e.g. yoki 'here', koki 'there' and coki 'yonder'.
Korean predicates do not agree in number, person or gender with their subjects. However, predicates show agreement with honorificness and politeness in different styles of speech. Three main levels of speech are distinguished with respect to politeness: plain, polite and deferential. Many other speech levels can also be represented among these three basic speech
levels by different endings. She main speech levels of declarative sentences have the following ending forms: plain: ta; polite: yo; deferential: (su)pnita. Thus, when the speaker expresses his politeness toward the hearer, either the polite or the deferential speech level is used, e.g. sonsængnim i cip e ka-yo 'the teacher is going home'; sonsængnim i cip e ka-pnita. In contrast to this, when the speaker does not express any particular politeness toward the hearer, the plain speech level is used; e.g sonsængnim i cip e ka-n-ta.
If the speaker wants to express his respect toward the referent of the subject, the honorific marker si is inserted between verbal bases and endings: e.g. sonsængnim i cip e ka-si-oyo; sonsængnim i cip e ka-si-pnita. In the last example, the insertion of the honorific marker si is possible in the predicate of a sentence ending in the plain speech level, since the honorificness is expressed to the subject, but not to the hearer. In the above example, if the subject is a student instead of a teacher, then unacceptable sentences are produced: *haksæng i hakkyo e kasi-oyo; * haksæng i hakkyo e ka-si-pnita; *haksæng i hakkyo e ka-si-nta. The ungrammaticality of the last examples is due to the violation of agreement "between the subject and the predicate with respect to honorificncss. In other words, the subject haksæng 'student' cannot occur with the predicate containing the honorific marker si, because haksæng belongs to the class of nouns which cannot be referred to with the honorific marker si.
Let us now turn to negation. Korean has three different negative morphemes: an, ma and mos. The morpheme an occurs in declarative and interrogative sentences and the morpheme ma occurs in propositive and imperative sentences, e.g. declarative: cip e an ka-n-ta 'I do not go home'; interrogative: cip e an ka-ni? 'don't you go home?'; propositive: cip e ka-ci mal-ca 'let's not go home'; imperative: cip e ka-ci ma-la 'don't go home'. The remaining negative morpheme mos has the meaning 'cannot', e.g. cip e mos ka-n-ta 'I cannot go home'. There are three types of negation in Korean. In the first type, the negative morphemes an and mos occur immediately before the main verb, as in the declarative and interrogative, as in the last example. The other two types involve more complicated operations. In the second type, the negative behaves like the main predicate and the complementiser ci is incorporated, as in the propositive and inoperative. The third type of negation involves the main predicate ha 'do' in addition to ci complementation; cip e ka-ci ani ha-n-ta 'I don't go home'; cip e ka-ci mos ha-n-ta 'I cannot go home'. From these three types of negation, we can observe different occurrences of negative morphemes. That is, while the negative morpheme an appears in all three types of negation, the morpheme mos appears in the first and third types of negation. The remaining negative morpheme ma appears only in the second type of negation.
As a final example of Korean syntactic characteristics, Korean sentential complements will be briefly discussed. Sentential complements are marked
with the nominalisers kos, ki, um and ci and with the complementiser ko. Several differences exist between nominalisers and complementisers: first, case particles occur after nominalisers but cannot occur after compiementisers: e.g. in the sentence na nun i chæk i cæmiis'-nun kos ul a-n-ta 'I know that this book is interesting', the accusative particle ul occurs right after the nominaliser kos, but in the sentence *na nun i chæk i cæmiis'ta ko lul sængkakha-n-ta 'I think that this book is interesting', the variant accusative particle lul cannot occur after the complementiser ko. Secondly, while the nominaliser is preceded by non-finite modifier forms -nun- and -n/un-, the complementiser is preceded by the finite verbal ending form -ta. Thirdly, the nominaliser occurs in both the subject and object positions, but the complementiser occurs only in object position. Sentential complements containing the nomimaliser have different syntactic behaviour from sentential complements containing the complementiser. Sentential complements containing the nominaliser behave like regular noun phrases. Thus, whereas sentential complements with the nominaliser undergo syntactic processes such as topicalisation, pseudo-cleft formation, passivisation, noun phrase deletion and pronominalisation, sentential complements with the complementiser do not undergo the same syntactic processes. Of the above nominalisers and complementisers, ci is used as a question nominaliser and ko is used as quotative complementiser: na nun John i once o-nun ci molu-n-ta 'I do not know when John will come'; na nun John i næil o-n-ta ko malha-yos'-ta 'I said that John would come tomorrow'.
Sentential complements containing ki can be differentiated from sentential complements containing um/m by syntactic and semantic characteristics. In the majority of cases, um/m is used for factive complements (i.e. complements whose truth is presupposed), but ki is used for non-factive complen1ents. A given predicate will take only one of these two nominalisers, e.g. na nun John i cip e ka-l-kos ul wonha-n-ta 'I want John to go home.'; *na nun John i cip e ka m ul wonha-n-ta; na nun John i cip e ka m ul al-as'-ta 'I knew that John was going home'. *na nun John i cip e ka ki lul al-as'-ta. The exampies show that the non-factive predicate wonha 'want' occurs only with ki and the factive predicate al 'know' occurs with um/m. The nominaliser kos occurs with both factive and non-factive complements: na nun John i cip e ka-nun kos ul wonha-n-ta 'I want John to go home'; na nun John i cip e ka-nun kos ul al-as'-ta 'I knew that John was going home'.
For discussion of the origins and history of Korean, reference may be made to Chin-Wu Kim (1974) and Ki-Moon Lee (1967). the latter also available in a German translation. Ho (1965), a monograph treatment of Korean phonology, is available only in Korean, but two studies are available in English: Martin (1954), a classic treatment of Korean morphophonemics, and B.K. Lee (1977), from the generative viewpoint. Among descriptive grammars, Choi (1954) is available only in Korean and in the absence of a comparable descriptive grammar in English, the most useful sources of general information on Korean grammar are the pedagogical texts by Martin (1969) and Lukoff (1982). Nam-Kil Kim (1984) is a study of Korean sentence complementation from a generative viewpoint.
Choi, Hyon Bae. 1954. Uli Malbon (Jongumsa, Seoul)
Ho, Woong. 1965. Kuko Umunhak (Jongumsa, Seoul)
Kim, Chin-Wu. 1974. The Making of to Korean Language (Center for Korean Studies, University of Hawaii, Honolulu)
Kim, Nam-Kil. 1984. The Grammar of Korean Complementation (Center for Korean Studies, University of Hawaii, Honolulu)
Lee, B.K. 1977. Korean Generative Phonology (Iljisa, Seoul)
Lee, Ki-Moon. 1967. 'Hankuko Hyongsongs', in Hankuk Munhwasa Taekye, vol. 5 (Korea University Press, Seoul; also available in German translation, Geschichte der koreanischen Sprache, translated by B. Lewis, Dr Ludwig Reichert Verlag, Wiesbaden, 1977.)
Lukoff, F. 1982. An Introductory Course in Korean (Yonsei University Press, Seoul)
Martin, S. 1954. Korean Morphophonemics (Linguistic Society of America, Baltimore)
---------- 1969. Beginning Korean (Yale University Press, New Haven)
Nam-Kil Kim, "KOREAN, In: Bernard Comrie (ed.). The World's Major Languages. London: Croom Helm, pp. 881-898