.,,.... n", a. ...:.v.‘,-.—rv "Ifl' , , n ;.‘-. “FEMS RARIES \\\\\\\\\\l \lil\ll\\ll\l\l\i\\ ll ll 17 This is to certify that the dissertation entitled A STUDY OF TENSE AND ASPECT IN KOREAN presented by Sangkook Lee has been accepted towards fulfillment of the requirements for Ph- D- degree in _Lingnis_tics Major pro essor Date Jan. 28, 1992 MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State Unlverslty ' PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE 'J Ua‘b $5995 éfiwifiom Ti Ell—'7 MSU Is An Affirmative Action/Equal Opportunity lnditution cmimMpma-pJ A STUDY OF TENSE AND ASPECT IN KOREAN BY Sangkook Lee A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Linguistics and Germanic, Slavic, Asian and African Languages 1991 ABSTRACT A STUDY OF TENSE AND ASPECT IN KOREAN BY Sangkook Lee The two objectives of this study are, first, to identify the full range of meanings of the grammatical categories of £88 and o as tense and aspect markers in Korean, and second, to describe the categories and to present a systematic, if not complete, account of them. This study finds that nonpast tense ¢ contrasts with past tense ESS. The sentences without any overt tense and aspect markers seem to be two-ways ambiguous: one as simple nonpast and the other as nonpast habitual. Also, the sentences with the single ESS have the four apparently incompatible meanings of simple past, present completion, future completion, and past habitual aspect. And there are also sentences with a double form ESS-ESS, which refers to past completion of a situation. It is argued that there are two null markers with the meanings of habitual aspect and nonpast tense. It is also proposed that there is no single ESS when aspect is marked, but, in fact, every instance of the single ESS is a case of double marking on a more abstract level of representation in which the single ESS is sometimes combined with null markers. It is also argued that the marker ESS is, in fact, Sangkook Lee two homophonous but semantically distinct morphemes, i.e., a completive aspect marker E881 and a past tense marker ESSZ. These assumptions will provide a basis to unify the four apparently incompatible meanings of sentences with the marker E85 and without any overt tense and aspect marker. When the habitual (a, co-occurs with past ESSZ, the sentences refer to past habituality. Next, when the nonpast O2 co-occurs with the completive 15:88,, the sentences indicate nonpast completive. If any aspect marker is not marked with the past ESSz, the sentences signify the simple past. Furthermore, when any aspect marker is not marked with the nonpast Oz, the sentences refer to simple nonpast. When the habitual G1 co-occurs with the nonpast $2, the sentences indicate nonpast habituality. Copyright by 0 Sangkook Lee 1992 Dedicated to my parents, Hansul and Jungsook and my wife, Kueyoung and my son, Yunsup ACKNOWLEDGEMENTS The completion of this dissertation leaves me indebted to many people. To a few people the debt I owe is particularly large. I am primarily indebted to my dissertation committee chairperson, Carolyn H. Harford, for guiding the overall research and patient instruction. Without her constant encouragement, advice and help, this dissertation would never have been begun and completed. Profs. Barbara K. Abbott, Grover Hudson, and Ruth M. Brend, members of dissertation committee, read the entire study and made many valuable suggestions. Prof. Seok C. Song, my committee ex-chairperson, has provided often needed guidance and direction. My interest in syntax, especially in tense and aspect, is due, in large part, to him. I heartily thank Prof. Han Sohn at Yonsei University for his constant encouragement. Prof. Sohn has introduced me to this world of linguistics. Without his admonition, this dissertation would not have been possible. I am also thankful to Prof. Chong-Chul Choi at Yonsei University and Prof. Soon-Pong Park at Kunkuk University, who gathered reference materials in Korean and could make me look around them. Perhaps the greatest debt is that which I owe to my wife. Without her love, encouragement, and support, this research would still be incomplete. In addition, I am very grateful for the constant financial and spiritual support of my parents. vi TABLE OF CONTENTS Romanization and List of Abbreviation................ Chapter I. Introduction 1.1 The Goal of the Present Study.. ..... ...... ..... . 1.2 Descriptive Model 1.2.1 Tense...................................... 1.2.2 Aspect..................................... 1.3. Organization of the Present Study.. ............ Chapter II. Review of the Literature 2.1 Null Marker D as Present Tense Marker.................. D as Nonpast-Nonperfective Marker ........ .. 2.1.1 2.1.2 2.2 ['11 U) m NNNN coco oooomoooo .booutdalfildle N O u NNNN O... UUUUMNNNN Chapter III. 3.0 IntrOduCtiODOOOOCOOOOOOOOOOOOOOOOOOO0.0.0.0....O ESS as ESS as ESS as ESS as Past Tense Marker................... Progressive Aspect Marker........... Present Completion.. ...... .......... Present Perfect Marker.............. Cluster..................................... Same Meaning as the Single ESS............. Single Morpheme............................ Doubling of the Same Markers............... Two Different Morphemes.................... Nonpast Time Reference 3.1 Nonpast Tense Marker Dz......................... 3.1.1 Null Nonpast Tense with Stative Verbs...... 3.1.2 Null Nonpast Tense with Nonstative Verbs... 3.1.3 Contrast between D2 and E882 with Aspect Markers.................................... 3.1.4 Conclusion................................. 3.2 Nonpast Incompletive............................ 3.2.1 Habituality and Progressiveness............ 3.2.2 Habituality................................ 3.3 Conclusion...................................... vii Page \l-hN 10 11 13 14 15 16 17 17 19 21 21 26 30 31 37 42 45 45 46 53 57 Chapter IV. Past Time Reference 4.0 Introduction.................................... 4.1 Problems........................................ 4.2 ouble Form ESS-ESS 1 Introduction...... ...... . .................. 2 Same or Different?......................... 3 One or Two? 4.2.3.1 Separable or Inseparable?........... 4 Doubling or Not?........................... 5 Two Different Markers...................... 6 Conclusion................................. n 1 2 D 2. 2. 2 gle ESS...................................... Past Tense................................. Nonpast Completion 4.3.2.1 Present Completion.................. 4.3.2.2 Future Completion................... 4.3.3 Past Habitual.............................. 4.4 Conclusion...................................... Chapter V. Conclusion................................ List of References....................... ..... ....... viii 58 6O 66 67 71 72 74 75 76 78 79 81 82 87 91 95 Romanization The romanization adopted for Korean is Yale romanization. List of Abbreviation AM AUT DEC HON HUM INF-POL INGR. INT MOD NM POL PROG RESULT SFM SUS SUST TOP TRSF VP Accusative marker Authoritative sentence-final marker Declarative sentence-final marker Honorific marker Humble marker Informal-Polite sentence-final marker Ingressive aspect marker Interrogative sentence-final marker Modifier Nominative marker Polite register Progressive aspect marker Resultant aspect marker Sentence-Final marker Suspective sentence-final marker Sustentive aspect marker Topic marker Transferentive marker Verb Phrase Chapter I INTRODUCTION 1.1 The Goal of the Present Study In the present study, an attempt has been made (1) to identify the grammatical categories of E88 and D as tense and aspect markers in Korean, and (2) to present a systematic, if not complete, account about them. The reason for this study is that, in the past, tense and aspect in Korean were studied using various approaches which resulted in considerable confusion on the subject, and which will be discussed in Chapter II. The present study will argue that past tense contrasts with nonpast tense. The past tense is marked by a morpheme E88 and the nonpast tense is not marked by any overt tense marker, i.e., a null marker 9. The sentences without any overt tense marker seem to be two-ways ambiguous: one as simple nonpast and the other as nonpast habitual. Furthermore, the sentences with the single ESS have the four apparently incompatible meanings of simple past, present completion, future completion, and past habitual aspect. And there are also sentences with a double form ESS-ESS, which has the meaning of past completion of a situation. I will argue that there are two null markers, i.e., O, and $2, with the meanings of habitual aspect and nonpast tense, respectively. I also propose that there is no single ESS when aspect is marked, but, in fact, every instance of the single ESS is the case of double marking on a more abstract level of representation in which the single ESS is sometimes combined with null markers. I also postulate that the marker ESS is, in fact, two homophonous but semantically distinct morphemes, i.e., a completive aspect marker and a past tense marker. These assumptions will provide a basis to unify the four apparently incompatible meanings of sentences with the marker E88 and without any overt tense and aspect markers. '1‘; Descriptive Model 11211 Tense The framework which I adopt for the present study of tense and aspect is Comrie's works (1976, 1985). Comrie (1985:9) defines tense as "grammaticalised location in time." For example, this definition makes us say that the difference between JOhn swims and John swam in English is one of tense, while that between John swims and John is swimming is rather one of aspect. Tense may be relevant for more than one time point. For example, Reichenbach (1947) discusses speech time, event time and reference time. The speech time is "the time point 3 of the act of speech" (Reichenbach, 1947:288). In other words, the speech time is the time when the act of speech itself takes place. The perfect "relates some state to a preceding situation" (Comrie, 1976:52). Thus, it expresses "a relation between two time-points" (Ibid.), i.e., the time of the resultant state of a preceding situation and the time of a preceding situation. For example, in Peter had gone by six o'clock yesterday morning, event time is "the time when Peter went" (Reichenbach, 1947:288) and reference point is "a time between event time and speech time," (Reichenbach, 1947:288), i.e., six o'clock yesterday morning. In other words, John's departure is located prior to the reference point. Thus, the speakers can choose any other time point than speech time and can think of the time point as the point that the speaker is referring to. The reference time can be called as a point of reference that functions for a speaker as an alternative to the time of speaking (Johnson, 1981:148). In Korean, speech time is usually the reference point as illustrated in the following: 1. a. ku yeca-ka alumtap-ta that lady NM beautiful DEC She is beautiful. b. ku-ka pap-ul mek-ESS-ta he NM rice AM eat DEC He ate rice. The sentences in (1) do not concern two events. Thus, events in (1) are always determined with respect to speech 4 time. In other words, the reference point is coincided with speech time. However, there can also be another reference point for speakers. For example, in the sentence nay-ka ecey ka-ESS-ta 'I went yesterday', the reference point of time becomes ecey 'yesterday' which is the same as event time. 11212 ASEQQE Comrie (1976:3) takes the view that "aspects are different ways of viewing the internal temporal constituency of a situation." In the example, John was singing when I visited, the second verb visited demonstrates "the totality of the situation referred to," e.g., my visiting, regardless of its "internal temporal constituency" (Comrie, 1976:3). Thus, the whole of the situation is given as "a single unanalysable whole, with beginning, middle, and end rolled into one" (Comrie, 1976:3). The other verb was singing does not describe the situation in the same way. Rather, it concerns itself with internal temporal constituency. It is interpreted as indicating that my visiting is an event that occurred during the time that John was singing. Thus, aspect is not interested in "relating the time of the situation to any other time-point, but rather with the internal temporal constituency of the one situation" (Comrie 1976:5). There is also the contrast between perfective and imperfective. Comrie (1976:18) observes that "the 5 perfective does indeed denote a complete situation, with beginning, middle, and end". But he tries to distinguish 'completed' from 'complete' since 'completed' action emphasizes "termination of the situation," while the perfective aspect of 'complete' situation emphasizes nothing. Thus, the 'complete' situation is different from the 'completed' situation. According to Comrie (1976:19), for example, in Russian, there are derivationally related forms that indicate the completion of a situation. For example, a verb otuZinat' 'finish supper' as in my toIko dto otuznali 'we've just finished supper' is distinct from another verb pouZinat' 'have supper' as in my pouZinali posle polunodi 'we had supper after midnight'. The latter, i.e., pouZinat' 'have supper', refers to "the whole of the meal, not just its end" (Comrie, 1976:19), while the former, i.e., otuZinat' 'finish supper', indicates the completion of the supper. Therefore, in the use of the perfective "all parts of the situation are presented as a single whole" (Comrie, 1976:18). More emphasis on the end of.a situation is at best "only one of the possible meanings of a perfective form, certainly not its defining feature" (Comrie, 1976:19). The imperfective refers to "the internal temporal structure of a situation, viewing a situation from within" (Comrie, 1976:24). This imperfective is subdivided into habitual and progressive aspect. Comrie (1976:26-29) gives a positive definition of habituality as "a situation which is characteristic of an 6 extended period of time, so extended in fact that the situation referred to is viewed....as a characteristic feature of a whole period," while continuousness, including progressiveness, is negatively defined as "imperfectivity that is not habituality" (Comrie, 1976:26). This indicates that progressiveness refers to a nonextended period of time opposite to the habitual meaning. In other words, it suggests that "the basic meaning of the English Progressive is to indicate a contingent situation: this would subsume progressive meaning itself" (Comrie, 1976:38). In the present study, instead of the opposition between perfective and imperfective, I will use the terms 'completive' and 'incompletive' with the following definitions. The completive, which is marked by E851, refers to a completed situation that emphasizes the termination of the situation, rather than a complete situation of the perfective. Thus, it is not appropriate to call the marker ESS1 'perfective' . On the other hand, the incompletive in Korean is negatively defined as aspect that does not refer to completedness of a situation. The incompletive is subdivided into habituality, progressiveness, and resultantness, etc. Habituality indicates a situation which can be "prolonged sufficiently in time, or that can be iterated a sufficient number of times over a long enough period" (Comrie, 1976:30). The progressive indicates a situation that an event continues to the present moment, as S.C. Song (personal 7 communication) points out, over a limited and nonextended period of time, while the resultant indicates a situation where a state continues to the speech time. In this study, the term 'situation' is used, following Comrie (1976:13), as a general cover-term, including 'state', 'event', or 'process'. States are defined as "static, i.e., continue as before unless changed," while events and processes are "dynamic, i.e., require a continual input of energy if they are not to come to an end." Furthermore, events are dynamic situations viewed perfectively, while processes are dynamic situations viewed imperfectively (Comrie, 1976:13). 1.3 Organigarign 9f tne Present Study In Chapter II, the literature will be reviewed to provide a background for discussing tense and aspect in Korean. In Chapter III, I will, first, argue that the null marker O2 for nonpast tense should be set up since the sentences without the overt past tense marker ESS2 contrast with those with it. This contrast is found not only in stative verb constructions but also in nonstative verb constructions. The null marker is also compatible with time adverbs such as cikum 'now', onul 'today', nayil 'tomorrow', etc., not with such time adverbs as ecey 'yesterday', kucekkey 'the day before yesterday', etc. This indicates that the null marker O2 refers to nonpast tense. It will be 8 also argued that sentences refer to simple past or simple nonpast tense when no aspect marker is marked. Third, I will also argue that habitual aspect is different from progressive aspect. The sentences without any overt tense and aspect markers usually indicate simple nonpast tense, marked by the null marker O2, but they also have the meaning of incompletive habitual aspect. This incompletive aspectual meaning is habituality, which is contrasted with the meaning of progressive aspect. This leads us to set up another null marker O5, i.e., the habitual aspect marker, which is different from nonpast tense O2. Section 3.2 will also deal with the contrast between the habitual aspect marker (01 and the progressive aspect marker KO 188. Also, I will argue that the null habitual marker O1 is different from the progressive marker KO 188. Chapter IV will, first, present problems regarding the single ESS. Next, section 4.2 will present some problems about the E88 cluster and then argue that the E85 cluster is composed of two different morphemes, e.g., completive aspect and past tense, so that the E88 cluster has the meaning of past completive aspect. Section 4.3 will argue that the four apparently incompatible meanings of the single ESS come out when the single ESS as one of two homophonous but semantically distinct morphemes, co-occurs with one of the null markers of habitual aspect O1 and nonpast tense O2. With this 9 analysis, the four apparently incompatible meanings of the single ESS can be explained. Chapter V is limited to concluding remarks. The issues in this study allude to only a small part of the Korean tense and aspect system and many of the solutions proposed are tentative and subject to further modifications. However, the present study provides a unified basis for a more fruitful study of the Korean tense and aspect system. Chapter II REVIEW OF THE LITERATURE The purpose of the literature review is to provide a background of related studies regarding tense and aspect in Korean. Particular attention is given to reviewing the studies pertaining specifically to the null marker O, the single E58, and the E88 cluster. 211 Null_uerker_9 21111 Q_e§_£rs§ent_Ten§e_herker Some traditional grammars claim that present tense is indicated by verbs themselves, but without justification (S.-P. Park, 1935:324; I.-S. Chung, 1949:44-45). In other words, the present tense is not realized phonetically. Furthermore, Martin (1954:35) claims that there are two tense morphemes: past E88 and future KEYSS. His analysis does not include a present tense morpheme. Later, this position becomes linked to a claim that present tense does not have any overt tense form (S.C. Song, 1967:22, 240- 241). Song (1976:143) further claims that the unmarked form O refers to present tense, which contrasts with the past tense marker ESS, as in the following: 10 11 1. a. ku-ka pap-ul mek-O-NUN-ka he NM rice AM eat INT Does he eat the rice? b. ku-ka pap-ul mek-ESS-NUN-ka Did he eat the rice? In (1), Song claims that the null marker O in (1a) contrasts with the past tense marker ESS, so that the null marker O refers to present tense in (1a). Recently, H.-J. Chong (1987:26) also claims that the null marker O is the present tense marker. Observe her example, as shown below: 2. Unhi-nun yeppu-O-ta TOP pretty DEC Unhi is pretty. However, she also claims that the marker NUN is also the present tense marker, in the case of (1a). In other words, according to her, there are two present tense markers in Korean. 2.1.; Q as Ngnpnsr-Ngnnerfiegrive Marker H.-M. Sohn (1975:54) claims that an unmarked category O should be set up since this form O centrasts with the marker ESS. According to him (Ibid.), the non-presence of the marker ESS "positively means that an event is going on at present..., will take place in the future, takes place habitually..., or takes place timelessly or eternally.” In other words, the null marker O refers to "non-perfective and non-past" (Ibid.), while its counterpart ESS refers to "perfective (or completive) and past, not just past" (Sohn, 12 1975:51). Observe his examples and translations: 3. a. cikum-kkaci mek-ESS-ta now until eat DEC He has eaten until now. (Perfective) [sic.] b. cikum-kkaci mek-O-NUN-ta He has been eating until now. (Non-perfective) Furthermore, D.-H. An (1980:115) also claims that "the absence of the past tense marker ESS (or the presence of O)...normally represents an event not completed earlier than ST [speech time]." According to him (1980:117-132), the null marker O is used to indicate present time reference in (4a) and future time reference in (4b), semelfactive aspect in (4c), and habitual aspect in (4d). Observe his examples and translations: 4. a. chelswu-uy apeci-nun cikum cangkwun-i-O-ta Chelswu of father TOP now general be DEC Chelswu's father is now a general. b. naynyen khulisumasu-nun ilyoil-i-O-ta next year Christmas TOP Sunday be Next year Christmas falls on Sunday. c. chelswu—ka cikum mutay-wi-eyse nolayhaéO-NUN-ta Chelswu NM now stage on at sing Chelswu is now singing on the stage. (literally, Chelswu sings now on the stage.) d. yenghuy-ka kakkum ppalkan oss-ul ip-O-NUN-ta Younghee NM often red clothes AM wear Younghee sometimes wears red clothes. An (1980:120) claims that (4c) is an example of semelfactive aspect to "refer to a situation that takes place once and once only" (Comrie, 1976:42). According to Comrie (1976:42), in the English verb cough, this verb can refer to 13 one single cough or a series of coughs. The former situation indicates semelfactive aspect and the latter iterative aspect. However, sentence (4c) indicates that Chelswu habitually sings on the stage. This means that his present singing on the stage is one of his singing habit. Therefore, sentence (4c) cannot be semelfactive. He also claims that (4c) with the null marker O is synonymous with the sentences with the progressive aspect marker KO 188, e.g., (5), as in the following:' 5. chelswu-ka cikum mutay-wi-eyse nolayha-KO ISS-O-ta Chelswu is now singing on the stage. To sum up, Sohn (1975) and An (1980) claim that the null marker O has several meanings of nonpast-perfective aspect. They do not claim that the null marker O is either the nonpast marker or nonperfective aspect marker, but that it is a kind of combined category with the meaning of nonpast tense and nonperfective aspect. This position, however, fails to give a unified account of how the null marker interacts with other tense and aspect markers. 212 The_§insle_fi§§ 21211 B§§.§§.2§§£.I§D§§.H§IE§I The most predominant analysis among scholars of the marker ESS has been that it is a past tense marker indicating the situation in question occurred prior to the present moment. Such a traditional linguist as H.-B. Choi (1937 [1980:4491) believes that the marker ESS refers only 14 to past tense. Observe the following sentences: 6. a. ku-ka hakkyo-ey ka-ESS-ta he NM school to go DEC He went to school. b. nay-ka hakkyo-eyse chinkwu-wa nol-ESS-ta I NM school in friend with play I played with friends in school. Thus, the marker ESS refers to a situation located prior to speech time. In other words, the situations of his going to school and my playing with friends in school occurred before the present moment. This line of thinking has been recently confirmed by C.-K. Sim (1980, 1985). It is beyond doubt that the marker ESS has the meaning of past tense, but it is also true that the marker ESS has other meanings than past tense, as discussed in the next sections. 2121; ESS gs Erggressive Aspecr Mgrker C.-S. Na (1963) claims that the marker ESS refers to past progressive. According to him (1963:71-72), the meaning of past progressiveness is expressed in (7), so that it has the same meaning as the progressive aspect marker KO 188. He also claims that sentences with the marker ESS is freely interchangeable with those with the progressive aspect marker KO 188. Observe the following sentences: 7. a. ku-ka ku tangsi pusan-ey sal-ESS-ta he NM that time Pusan in live He used to live in Pusan at the time. b. ku-ka ku tangsi pusan-ey sal-KO ISS-ESS-ta He was living in Pusan at the time. 15 Contrary to Na's claim, (7a) is different in meaning from (7b). .If the marker ESS in (7a) had the same meaning as the markers KO ISS-ESS in (7b), any two sentences with the markers ESS and KO 188 would also be grammatical. However, it is not true as illustrated by the following: 8. a. ku-ka ecey pusan-eyse cwuk-ESS-ta he NM yesterday Pusan in die DEC He died in Pusan yesterday. b. *ku-ka ecey pusan-eyse cwuk-KO ISS-ESS-ta The ungrammaticality of (8b) indicates that the marker ESS does not inherently have the meaning of past progressive aspect. grgr; ESS as Present Completion This aspectual analysis of the marker ESS is credited to C.-S. Na (1963:69-70). He claims that the meaning of present completion is revealed in (9) when such- deictic time adverbs as cikum 'now' are used. Na (1963:70) remarks that the marker ESS refers to a present state resulting from the past situation. Thus, he claims that the sentence (9a) is synonymous with the sentence (9b). Observe the following sentences: 9. a. kkoch-i cikum phi-ESS-ta flower NM now blossom DEC The flower has just blossomed now. b. kkoch-i phi-E ISS-ta The flower is in bloom. 16 It is true that verbs with the marker ESS refer to present completion. The present completion in both (9a) and (9b) indicates that a resultant state of past completion is extended to the present moment (Na, 1963:69). Thus, Na can claim that the marker ESS is synonymous with the resultant aspect marker E 188, which indicates a present state of resulting from a past completed situation, as in (9b). However, the marker ESS in (9a), in fact, indicates that the state of the flower's blossoming is completed at the present moment, but the speaker does not know whether or not the flower will continue to blossom. On the contrary, the marker E ISS in (9b) indicates that the state of the flower's blossoming was completed before speech time and the state of the flower's blossoming is extended up to speech time. 21215. S s ese e ect k Nam (1972 [1978b:8-12]) claims that the marker ESS is a present perfect marker indicating that an action was completed and its result is extended up to the present moment. Observe his examples: 10. a. ney os-ey hukl-i mut-ESS-ta your clothes at soil NM stick Mud has stuck to your clothes. b. ku-ka twis cali-ey anc-ESS-ta he NM back seat in sit He has sat in the back seat. l7 Nam (1972 [1978b:10]) claims that the situation of mud's sticking or his sitting in the back seat was completed before speech time and its resultant state is extended up to the present moment. 2.; ESS Cluster It is well known that there is the double form ESS- ESS, i.e., the ESS cluster in Korean, which appears to consist of two successive past tense markers. This apparent fact results in several problems regarding the nature of the ESS cluster. Until now, most studies also fail to provide a systematic account of the ESS cluster, because of these problems. 2.3.; Sane Meaning gs tne Single ESS Many scholars claim that the ESS cluster is synonymous with the single ESS and that these two markers are freely interchangeable with each other. Among others, C.-K. Gim (1985:46, 49) argues that there is no need to set up two different morphemes for the ESS cluster since it may be regarded as two successive past tense markers. Observe the following sentences: 11. a. i san -ey namu-ka wukeci-ESS-ta this mountain at tree NM grow thick This mountain was overgrown with trees. b. i san-ey namu-ka wukeci-ESS-ESS-ta This mountain was overgrown with trees. 18 According to Gim (1985:85), there is no difference in meaning between in (11a) and in (11b). The only difference is that (11b) is not usually used in Gim's southeastern Kyengnam dialect, but (11b) is used in the central Chwungnam dialect. However, his atypical theory does not appear to be supported by other scholars. Moreover, An (1980:112) claims that ESS cluster is "merely an emphatic expression of -ESS." In other words, the semantic difference between the ESS cluster and the single ESS is, if any, that "the past situation is more emphasized in the latter [the ESS cluster] than in the former [the single marker ESS)" (Ibid.). Observe his examples and translations: 12. a. ku-ka yelsimhi kongpuha-ESS-ta he NM hard study He studied hard. b. ku-ka yelsimhi kongpuha-ESS-ESS-ta He did study hard. An (1980:112) claims that the difference in meaning in (12a- b) is merely "a matter of emphasis" like in English the use of the auxiliary verb do. This claim might be related to adverb reduplications such as ppalli ppalli 'very quickly' in the imperative ppalli ppalli hay! 'Do it very quickly!‘ in Korean, as G. Hudson (personal communication) points out. However, adverb reduplication in Korean is limited to onomatopoeia. In the case of regular adverbs, reduplication is restricted to some adverbs such as ppalli 'quickly' or 19 kkok 'quite' in order to express intensive meaning. If this adverb reduplication were common, sentence (12a) would be grammatical if the adverb yelsimhi 'hard' were duplicated. But it is not true that a sentence *ku-ka yelsimhi yelsimhi kongpuha-ESS-ta is grammatical and acceptable. However, there is no case that any suffix in Korean can be duplicated except for the ESS cluster. For example, other traditionally called tense suffixes such as NUN, TE, and KEYSS, in the verb cannot be reduplicated as follows: 13. a. *ku-ka kongpuha-NUN-NUN-ta he NM study b. *ku-ka kongpuha-TE-TE-la c. *ku-ka kongpuha-KEYSS-KEYSS-ta Furthermore, semantically there are big difference between the single ESS and the ESS cluster. The single ESS refers to simple past or present completion. On the other hand, the ESS cluster indicates that a situation was completed in the past and the situation is no longer extended up to the present moment. For all these reasons, it is unlikely that the ESS cluster, i.e., ESS-ESS, is a reduplication of the single ESS. The semantic difference between the single E88 and the E88 cluster will be discussed in detail in chapter IV. MW Nam (1978a [1978b:140-2]) states that this ESS cluster is attached to all types of verbs, so as to denote 20 discontinuance between the past events or states and the present situation by a speaker's psychological consciousness. This will be illustrated by the following sentences: 14. a. ku-ka paci-lul ip-ESS-ta he NM slacks AM wear DEC He wore slacks. b. ku-ka paci-lul ip-ESS-ESS-ta He has worn slacks. 15. a. kang-i kiph-ESS-ta river NM deep DEC The river was deep. b. kang -i kiph-ESS-ESS-ta The river has been deep. Furthermore, Nam (1978b:11, 143) claims that a single ESS and the ESS cluster can be paraphrased with each other since they are in free variation as in: 16. a. ku-ka sicang-ey ka-ESS-ta he NM market to go He went to a market (and he might still be there). b. ku-ka sicang-ey ka-ESS-ESS-ta He went to a market (but he has already left there). Nam also claims that the marker ESS can be substituted for the ESS cluster (1972 [1978b:10-111), which is not true. He overlooks the semantic difference between (16a) and (16b). As English glosses show, in (16a) the subject ku 'he' went to a market and he might still stay at the market. But in (16b), the subject ku 'he' has already returned from the market. 21 The basic difference in meaning between (16a) and (16b) is that the E88 cluster refers to past completion of a situation and the past situation is not extended up to the present moment. But the single ESS indicates just the location of the situation prior to speech time as the reference point. 2.3.3 poubiing or the Same Markers Martin (1954:35) claims that the situation of PAST- PAST is "more remote or more definitely completed than that of past" probably because past tense refers to "definite, completed action or state (and so usually past)." A On the other hand, W. Huh and C.-H. Park explain this by saying that the ESS cluster is the duplication of the perfect stem supplement ESS. Huh maintains that it refers to "the more remote past” (1969:83), while Park advocates that it indicates "the perfect in the perfect" since "the speaker re-confirms the completion of the action or its description" (C.-H. Park, 1964:82). MW Choi (1937 [1980:4501) claims that the ESS cluster is made up of the past tense and the perfect aspect markers. According to him, the past perfect refers to "an action which was completed and whose result does not hold to the present moment" (Ibid.). Observe the following sentences: 22 17. a. John-i ecey achim khal-ul ilh—ESS-ta John NM yesterday morning knife AM lose John lost a knife yesterday morning (and he still has not found it). b. John-i ecey achim khal-ul ilh-ESS-ESS—ta John lost a knife yesterday morning (but the speaker does not know whether or not John has it now). The pluperfect in English indicates that the past is a reference point, and that the situation in question is located before that reference point. According to Reichenbach (1947:288), the speaker can choose any other time and the reference point indicates the time the speaker is principally referring to. For example, in John had left by five o'clock yesterday morning, the time adverb phrase by five o'clock yesterday morning refers to a reference point in the past, and John's leaving is located prior to this reference point, i.e., by five o'clock yesterday morning. This definition of the pluperfect clashes with Choi's definition of the Korean pluperfect since his definition of the ESS cluster does not allude to two time references, e.g., the past time as the reference point and the time prior to the past. The ESS cluster in (17b) indicates that John's losing a knife is located before speech time but its resultant state is no longer extended up to speech time. Thus, the speaker does not know whether or not John has the knife now. In other words, the situation of John's losing a knife was completed at past time, i.e., before speech time, 23 as the reference point. If the ESS cluster referred to the pluperfect, John's losing a knife would be located before the past reference point, i.e., yesterday morning, but this is not correct. Therefore, there is no reason to use the term 'pluperfect' to refer to the E88 cluster. Furthermore, N.-K. Kim (1975) claims that the ESS cluster is composed of the past tense marker and the experiential-contrastive aspect marker. In other words, one member of the ESS cluster has the meaning of both experience and contrast if the subject of the sentence is animate as in (18a), but only the meaning of contrast if the subject of the sentence is inanimate as in (18b). This claim is illustrated by the following sentences: 18. a. ku-ka pap-ul mek-ESS-ESS-ta he NM rice AM eat DEC He has eaten the rice. b. pi-ka o-ESS-ESS-ta rain NM come It has rained. This position claims that the subject of (18a) ku 'he' has had an experience of eating the rice before or that he ate before but is not eating the rice any longer. In (18b), the situation took place in the past and (18b) implies that raining is no longer taking place. The most vulnerable point of this position is to explain how the subject of sentences can have an experience of death if the ESS cluster is compatible with such verbs as cwuk- 'die', tolaka- 'pass away', and salhayha- 'kill, 24 murder'. The following sentences illustrate this point: 19. a. ku-ka cwuk-ESS-ESS-ta he NM die DEC He has just died. b. ku-pun -i tolaka -SI-ESS-ESS-ta he(POL) NM die(POL) HON DEC He has just passed away. (Polite register) c. John-i Mary-lul salhayha-ESS-ESS-ta John NM Mary AM kill John has killed Mary. Since a person cannot have had "an experience of death or he has died but he is alive now," N.-K. Kim (1975:531) claims that the sentences in (19) are not acceptable. But the sentences in (19) are perfectly acceptable to me, since his death or John's killing Mary in (19) occurred before speech time but the situation no longer extends up to speech time. Thus, these sentences are not semantically anomalous at all. In other words, the sentences in (19) are grammatical because the ESS cluster is not an experiential-contrastive aspect marker. The following sentences will further illustrate this point. 20. a. John-i panccum cwuk-ESS-ta NM halfway die DEC John was on the brink of death. b. John—i panccum cwuk-ESS-ESS-ta John has just been on the brink of death. Sentence (20a) indicates that John was in a dying condition before speech time, but the speaker says nothing about the present condition. On the other hand, sentence (20b) 25 signifies that John was on the verge of death, but he is well again and he is no longer on the brink of death. Thus, the ESS cluster indicates that a situation was completed before speech time and the situation is not extended up to speech time. This will be discussed in detail in chapter IV. To sum up, it has been claimed by the scholars cited herein that the null marker O refers to present tense or nonpast-nonperfective, which contrasts with the marker ESS. It has been also claimed that the marker ESS has the meaning of past tense, past progressive aspect, present completion, or present perfect aspect. Furthermore, it has been claimed that the ESS cluster is synonymous with the single ESS, that it consists of an inseparable single morpheme, that it is the doubling of the past tense marker, and that it is composed of the past tense and the perfect aspect markers. Chapter III NONPAST TIME REFERENCE gig Introducrion In this chapter, I shall discuss two null markers, i.e. , O, and 02, that have the meaning of habitual aspect and nonpast tense, respectively. In Korean, sentences without any overt tense marker indicate not only present and future time reference, but also habitual aspect. When the sentences without any overt tense and aspect markers signify present and future time reference, they contrast with those with the past tense marker ESS. These sentences without any overt tense and aspect markers seem to be also two-ways ambiguous. One of the two readings is that the sentences refer to a situation holding at and right after the present moment without implying that the situation is extended into the past. Another reading is that those sentences without any overt tense and aspect markers also have the meaning of habitual aspect, and so they do not seem to be freely interchangeable with sentences with the progressive aspect marker KO 188. This chapter will conclude that there are two distinct null markers: one with the meaning of habitual aspect O1 and another with the meaning of nonpast tense Oz. 26 27 This leads us to postulate that Korean has a binary tense system, i.e., past and nonpast. Also, it is assumed that aspect marking is optional, but tense and aspect occupy two slots in every verb phrase when aspect is optionally marked. The purposes of this chapter are thus three-fold: (1) to set up two null markers for nonpast tense and habitual aspect, and (2) to present a unified account of why the sentences without any overt tense and aspect markers can have the meanings of both simple nonpast tense and nonpast habitual aspect, and (3) to provide some basis for a systematic account of the single ESS and the ESS cluster. The present study assumes the following representation for verbs and tense and aspect markers. 1. a. Verb Phrase (VP) --> Root + Suffixal Markers b. Suffixal Markers --> (HON) + (Aspect) + Tense + (HUM) + (Modal) + SFM Habitual ----- > 0, ~ Incompletive --> Progressive -> KO 188 Resultant ---> E ISS c. Aspect -> Sustentive --> E TWU Ingressive --> E KA b Completive ----- > ESS1 d. Tense --> Nonpast ------ > O2 Past --------- > ESS2 e. Sentence-Final Marker (SFM) --> Declarative ta/la, Interrogative ka, Authoritative o/so, Informal-Polite e(yo), Suspective ci(yo), Imperative e la, etc. In Korean, suffixal markers are made up of many markers including honorific SI or humble marker SUP and modal 28 markers such as NUN, TE, and KEYSS, as shown in (1b). However, these markers are outside of the scope of the present study. Also, there are many other aspect markers to fill the aspect slot in (1b) such as inchoative e ci 'become to', inceptive ki sicakha 'begin to', and continuative 9 ka 'continue to', etc. But, these markers are also beyond the scope of the present study. Note that, as shown in (1b), two slots for aspect and tense marking are necessary in every verb phrase when aspect is marked. In other words, nonpast tense can occur either with habitual aspect or without aspectual meaning. These two slots for aspect and tense markers give a clue to the question of why the single ESS appears to have the four apparently incompatible meanings of simple past, present completion, future completion, and past habitual. This will be discussed in detail in chapter IV. The resultant aspect marker E 188 and the completive aspect marker ESS1 in (lo) will be also discussed in detail in chapter IV. One reason for giving different numbers to aspect and tense markers is that aspect and tense markers are sometimes homophonous but semantically distinct, so it is necessary to give different numbers to identify them easily. In other words, the null marker 8 represents habitual aspect and nonpast tense. The marker ESS represents both completive aspect and past tense. For the purpose of identification, I will label habitual aspect as O, and nonpast tense as 02. I will also 29 label completive aspect as E881 and past tense as ESSZ. The organization of this chapter will be as follows. In section 3.1, I will argue that a null marker 02 for nonpast tense should be set up since the sentences without the overt past tense marker ESS2 contrast with those with it. This contrast is found not only in stative verb constructions but also in nonstative verb constructions. This null marker is also compatible with time adverbs such as cikum 'now', onul 'today', nayil 'tomorrow', moley 'the day after tomorrow', etc., but not with such time adverbs as ecey 'yesterday', kucekkey 'the day before yesterday', etc. It will be also argued that aspect marking is optional, so that sentences can refer to simple nonpast or simple past tense without having any aspectual meaning. In the section 3.2, I will also argue that habitual aspect is different from progressive aspect. The sentences without the overt tense and aspect markers usually indicate simple nonpast tense, but they may also have the meaning of incompletive habitual aspect. This incompletive-aspectual meaning is habituality, which is contrasted with the meaning of progressive aspect. This leads us to set up another null 'marker 0,, i.e., habitual aspect marker, which is different from the nonpast tense ¢2' Section 3.2 will be also concerned with the contrast between the habitual aspect marker 01 and the progressive aspect marker K0 ISS. 30 gr; Ngnpast Tense narker 92 Matthews (1991:123) points out that an account of zero affixation was "never universally accepted" to answer a question of how plain come and past participle come are formally different. However, in generative morphology, accounts of this zero affixation are still valid (Jensen 1990, Lieber 1981, Spencer 1991). For example, consider the irregularly formed past tense form stood. According to Spencer (1991:184), the vowel alternation stand - stood is governed by a readjustment rule when an abstract affix PAST is added to a verb root marked with such feature as [+ablaut]. 2. V V hide PAST ----> hid [+ablaut] However, a derivation of a regularly formed past tense form grandstanded is different from the derivation of.(2). This is illustrated in the following: 31 3. V l I V I l N I——‘—I A N l v | grand stand 0 O PAST ------ > grandstanded [+ablaut] In the derivation of (2) the morpheme PAST is added directly to a V node which dominates the strong verb root. However, in the derivation of (3), the morpheme PAST is not attached directly to the strong verb but to "two tokens of the zero affix" (Spencer 1991:185) which contain the strong verb. In addition, in generative morphology, word formation by zero derivation is also allowed (Jensen 1990:88). For example, the noun spy 'one who spies' is derived from the verb spy, where an expected overt suffix -er. However, *spier does not exist. These examples indicate that zero morphemes, i.e., phonologically null affixes, are possible in analyzing linguistic data. 3,i.i Null Ngnpasr Tense witn Srntiye yerns Some scholars such as S.C. Song (1967, 1976) and H.-J. Chong (1987) think of the null marker as the present tense marker. Other scholars such as H.-M. Sohn (1975) and D.-H. An (1980) consider the null marker to be the nonpast- nonperfective marker. This section will argue that there is a null marker O2 with the meaning of nonpast tense. 32 Consider, first, the sentences with stative verbs in the following: 4. a. onul kkoch-i cham alumtap-ta . today flower NM very beautiful DEC The flower is very beautiful today. b. *ecey kkoch-i cham alumtap-ta yesterday c. kkoch-i cham alumtap-ESS -ta The flower was very beautiful. 5. a. onul nalssi-ka chwup-ta today weather NM cold DEC It is cold today. b. *kucekkey nalssi-ka chwup-ta the day before yesterday c. kucekkey nalssi-ka chwup-Essitta It was cold the day before yesterday. The sentences (4a-5a), which do not contain any overt tense markers, contrast with the sentences (4c-5c) which do. The sentences with the past tense marker ESS2 have the meaning of past time reference, while the sentences without any overt tense markers in (4a-5a) express present time reference. However, when time adverbs such as ecey 'yesterday', and kucekkey 'the day before yesterday' are used in the sentences without any overt tense marker, the sentences become ungrammatical. This contrast between (a) and (c) of (4-5) leads us to set up a null tense marker Oz, which contrasts with the past tense marker ESSZ, i.e., nonpast tense. Thus, (4a-5a) should be reanalyzed as (6a) and (6b), respectively: 6. a. 33 onul kkoch-i cham alumtap-O -ta The flower is very beautiful today. onul nalssi-ka chwup-Oz-ta It is cold today. One may raise the question of why the null marker Oz is taken to refer to nonpast tense instead of present tense. The answer is that there are many examples in which a null marker 02 is not incompatible with future time reference. First, take a look at the case of stative verbs which indicate future time reference. The sentences without the past tense marker ESSZ, i.e., when they have the null marker Oz,.also contrast with those containing the past tense marker ESSZ. The following examples will show this: 7. a. nayil coh-un kwukyeng-i manh-O-ta tomorrow good MOD attraction NM many There will be many attractions tomorrow. *ecey coh-un kwukyeng-i manh-O-ta yesterday ecey coh-un kwukyeng-i manh-ESS-ta There were many attractions yesterday. kyewul-i o-myen kkoch kaps-i pissarw-ta winter NM come if flower price NM high The flower will be high-priced in winter. *ecey kkoch kaps-i pissa-O-ta yesterday ecey kkoch kaps-i pissa-ESS-ta Yesterday the flower was high-priced. In (7a) and (8a), futurity is here established by time adverbials such as nayil 'tomorrow' and kyewul-i o-myen 'when winter comes', so that the sentences (7a) and (8a) can 34 be used to express future time reference. On the other hand, (7c) and (8c) more clearly indicate past time reference since the past tense marker ESS2 co-occurs with a past time adverb ecey 'yesterday'. Also, the ungrammaticality of the (b) sentences in (7-8) indicates that time adverbs with the meaning of past time reference are incompatible with the sentences without any overt tense marker. The sentences in (7-8) indicate that the null marker Q5 refers to the meaning of future time reference, which is contrasted with the meaning of the past tense marker ESSZ. Thus, the sentences in (7) and (8) should be reanalyzed as (9) and (10), respectively: 9. a. nayil coh-un kwukyeng-i manh-O2 -ta There will be many attractions2 tomorrow. b. *ecey coh-un kwukyeng-i manh-Oz-ta c. ecey coh-un kwukyeng-i manh-ESS'dta There were many attractions yesterday. 10. a. kyewul-i o-myen kkoch kaps-i pissa-Oz -ta The flower will be high-priced in winter. b. *ecey kkoch kaps—i pissa-Oz-ta c. ecey kkoch kaps-i pissa--ESS2 -ta Yesterday the flower was high-priced. The contrast between the (a) sentence and the (c) sentence of (9-10) shows that the null marker 02 in the (a) sentence refers to future time reference since future time reference is established by time adverb such as nayil 'tomorrow'. Thus, it also contrasts here with the past tense marker ESSZ. 35 This contrast between the null nonpast marker a5 and the past tense marker ESS2 can also be illustrated in the constructions containing the copula verb i- 'be', as shown below: 11. a. onul kim taywi-ka tangcik-i-Oz-ta today Kim captain NM on duty be DEC Today Captain Kim is on duty. b. nayil kim taywi-ka tangcik-i-O -ta Tomorrow Captain Kim will be on suty. c. *ecey kim taywi-ka tangcik-i-Oz-ta Yesterday d. ecey kim taywi-ka tangcik-i-ESSzvta Yesterday Captain Kim was on duty. In ( 11) , the null marker 02 is compatible with time adverbs indicating present and future time reference. This supports the view that the null marker 02 indicates nonpast time reference and so it contrasts with the past tense marker ESSZ. Also, the contrast between the null nonpast marker O2 and the past tense marker ESS2 can be found in the constructions with existential verbs iss 'exist, be' and eps 'not exist' as follows: 12. a. onul isahoy-ka iss-Oz-ta today council meeting NM exist DEC Today there is a council meeting. b. nayil isahoy-ka iss-Oz-ta tomorrow Tomorrow there will be a council meeting. c. *ecey isahoy-ka iss-Oz-ta yesterday 36 d. ecey isahoy-ka iss-ESSzdta Yesterday there was a council meeting. 13. a. onul—to gasollin-i eps-(bz-ta today also gasoline NM not exist Today we have run out of gas, too. b. nayil-to gasollin-i eps-Oz-ta tomorrow Tomorrow we will have run out of gas, too. c. *ecey-to gasollin-i eps-Oz-ta yesterday d. ecey-to gasollin-i eps-ESSz-ta Yesterday we ran out of gas, too. The contrast between the null marker 02 and the past tense marker ESS2 in (12-13) is the same as in (4-11). The ungrammaticality of the sentences (12c-13c) is also due to the fact that the null nonpast marker 02 cannot be compatible with time adverbs indicating past time reference such as ecey 'yesterday'. To sum up, in stative verb constructions the sentences without the past tense marker ESS2 contrast with those with it as in (4-13). Furthermore, the sentences without any overt tense and aspect markers co-occur with time adverbials indicating present and future time reference. This indicates that the sentences without any overt tense and aspect markers have the meaning of nonpast tense. Thus, this meaning suggests the position that the null nonpast tense marker 02 should be postulated. This will be represented as in the following: 37 14. Tense Past Nonpast I 8-__-_ N ass2 3.i.g Nuil Nonpast Tense with Nonstative Verbs In the previous section, it was argued that the null nonpast tense marker 02 contrasts with the past tense marker ESS2 in stative verb constructions. This section will argue that the null nonpast tense marker 02 also contrasts with the past tense marker ESS2 in nonstative verb constructions. The following sentences will illustrate this point: 15. a. onul TV-ey sin ceyphum kwangko-lul nay—ta today TV on new product advertisement AM submit I advertise a new product on TV today. b. nayil TV-ey sin ceyphum kwangko-lul nay-ta tomorrow I will advertise a new product on TV tomorrow. c. *ecey TV-ey sin ceyphum kwangko-lul nay-ta yesterday d. ecey TV-ey sin ceyphum kwangko-lul nay-Essivta yesterday I advertised a new product on TV yesterday. The sentences without the past tense marker ESS2 in (15a-b) co-occur with time adverbs such as onul 'today' and nayil 'tomorrow'. In this way, the sentences without the past tense marker ESSz are contrasted with those with it of (15d). The ungrammaticality of (15c) lies in that the null nonpast tense marker 02 is incompatible with such time 38 adverbs as ecey 'yesterday', with the meaning of past time reference, as discussed in the previous section. Thus, the same argument may be made that the null nonpast tense marker Q5 can be also set up in nonstative verb constructions, as in stative verb constructions. Therefore, (15a) and (15b) should be reanalyzed as (16a) and (16b), respectively: 16. a. onul TV-ey sin ceyphum kwangko-lul nay-Oz-ta I advertise a new product on TV today. b. nayil TV-ey sin ceyphum kwangko-lul nay-OZ-ta I will advertise a new product on TV tomorrow. This contrast between the null nonpast tense marker O2 and the past tense marker ESS2 can be also found in constructions with other nonstative verbs, as shown by: 17. a. onul so-lul tul-ey naynoh-Oz-ta today cow AM field to put out Today, I turn cows out to the field. b. nayil so-lul tul-ey naynoh-Oz-ta tomorrow Tomorrow, I will turn cows out to the field. c. *ecey so-lul tul-ey naynoh-Oz-ta yesterday d. ecey so-lul tul-ey naynoh-Essidta yesterday Yesterday, I turned cows out to the field. Just as in (15-16), nonpastness is here established by time adverbs such as onul 'today' and nayil 'tomorrow', so that the sentences (17a) and (17b) are contrasted with sentence (17d), which refers to past time reference. Sentence (17c) is ungrammatical since the null nonpast tense marker O2 is 39 incompatible with the time adverb ecey 'yesterday'. This contrast also confirms that the null nonpast tense marker OZ contrasts with the past tense marker ESSZ. The sentences in (15-17) may sound odd to Korean native speakers, but they are still grammatical and acceptable. The (a-b) sentences in (15-17) are usually used in very limited situations such as writing a diary, newspaper headlines, or slogans, etc. In the (a-b) sentences of (15-17), the marker NUN is usually inserted between the verb roots such as nay 'submit' and naynoh 'put out', etc., and the declarative sentence-final marker -ta, e.g., nay-NUN-ta or naynoh-NUN-ta . However, this present study is concerned only with the contrast between the null nonpast tense marker and the past tense marker, so that I do not give examples with the marker NUN in order to avoid confusion since the marker NUN is also the outside of the present study. Concerning the contrast between the nonpast Oz and the past ESSZ, we can find other evidence if we consider sentence-final markers other than the declarative sentence-final marker -ta. It is well known that there are several levels of speech in Korean which are decided by many complex factors in actual speech environments. Speech levels are indicated by different sentence-final markers, which are obligatory like gender marking in French nouns or case marking in German nouns. The contrast between the null nonpast tense marker 02 and the past tense marker ESS2 is still maintained 40 even when sentences contain different sentence-final markers. The following illustrations show this: 18. a. ku-ka onul hoysa-lul ttena-Offla he NM today company AM leave AUT He quits the company today. b. ku-ka nayil hoysa-lul ttena-Ofxa tomorrow He will quit the company tomorrow. c. *ku-ka ecey hoysa-lul ttena-OZ-so yesterday d. ku-ka ecey hoysa—lul ttena-ESSZ-so yesterday He quit the company yesterday. 19. a. ku-ka onul pap-ul manhi mek-tDz-so he NM today rice AM much eat AUT He eats a lot of rice today. b. ku-ka nayil pap-ul manhi mek-Oz-so tomorrow He will eat a lot of rice tomorrow. c. *ku-ka ecey pap-ul manhi mek-Oz-so yesterday d. ku-ka ecey pap-ul manhi mek-ESSZ-so yesterday He ate a lot of rice yesterday. In (18-19) , the null nonpast tense marker OZ also contrasts with the past tense marker ESS2 as it does in plain declarative constructions of (15-17) ending with the sentence-final marker -ta. When sentences end with the suspective sentence- final marker ci(yo) 'I suppose' (hereafter SUS), the past tense marker ESS is also contrasted with the null nonpast tense marker (Dz. The following sentences will illustrate this point: 41 20. a. ku-ka onul pap-ul mek-OZ-ci(yo) he NM today rice AM eat SUS I suppose he eats the rice now. b. ku-ka nayil pap-ul mek-Oz-ci(yo) tomorrow I suppose he will eat the rice tomorrow. 0. ku-ka ecey pap-ul mek-Oz-ci (yo) yesterday d. ku-ka ecey pap-ul mek-ESSz-ci(yo) yesterday I suppose he ate the rice yesterday. This suspective sentence-final marker ci(yo) 'I suppose', according to Martin, et a1. (1968:1522), refers to casual statements, questions, suggestions, or commands (often inviting confirmation or agreement) depending on four different intonation. No matter what intonation is used, the tense contrast among (20a), (20b), and (20d) is still maintained. Sentence (20a) refers to present time and sentence (20b) has the meaning of future time reference, while sentence (20d) indicates past time reference. Sentence (20c) is ungrammatical since the null nonpast tense marker Oz is not compatible with the time adverb .ecey 'yesterday', as discussed in earlier. Thus, the null tense marker 02 in (20a-b) can refer to nonpast tense, so that it contrasts with the past tense marker ESS2 in (20d). It will be discussed in the next section why the null nonpast tense marker 02 appears to have the meaning of progressive aspect. To sum up, just as in stative verb constructions, the sentences with the past tense marker ESSztalso contrast 42 with those without it in nonstative verb constructions. I propose to explain this contrast by setting up a null marker Oz, 'nonpast tense', contrasting with the past tense marker ESSZ. 3.1.3 Contrast between Q2 and ESS2 witn Aspegt Markers In the previous section, it was argued that the null nonpast tense marker 02 contrasts with the past tense marker ESSZ. The contrast between the nonpast 02 and the past ESS2 is also maintained when aspect is marked, just as when aspect is not marked, as shown in (4-20). The purpose of this section is to motivate the aspect slot in the verb. This is illustrated in the following sentences: 21. a. onul ku chayk-‘ul ilk-E TWU-Oz-e(yo) today that book AM read SUST INF-POL We read the book today (for future use). b. nayil ku chayk-ul ilk-E TWU-Oz-e(yo) tomorrow that book AM read SUST INF-POL We will read the book tomorrow (for future use). c. *ecey ku chayk-ul ilk-E TWU-Oz-e(yo) yesterday d. ecey ku chayk-ul ilk-E TWU-ESSZ-e(yo) We read the book yesterday (for future use). The marker E TWU refers to sustentive aspect (hereafter SUST), which indicates that a situation is kept for later use. In (21), the sustentive aspect marker E TWU is immediately followed by the verb root ilk 'to read'. The sentences in (21) indicate that the contrast between the null nonpast tense marker 02 and the past tense marker ESS2 43 is still maintained despite the fact that the sustentive aspect marker E TWU is inserted between the verb root ilk 'to read' and the tense markers. This contrast between the null nonpast O2 and the past ESS2 in (21) is the same as the case of the ingressive aspect marker E KA (hereafter INGR). The ingressive aspect marker E RA in Korean indicates the beginning of a new situation. The following sentences illustrate this point: 22. a. John-i onul phyenci-lul ssu-E KA-OZ-ci(yo) NM today letter AM write INGR SUS I suppose that John begins to write a letter today. b. John-i nayil phyenci-lul ssu-E KA-Oz-ci (yo) NM tomorrow letter AM write INGR SUS I suppose that John will begin to write a letter tomorrow. c. *John-i ecey phyenci-lul ssu-E KA-OZ-ci(yo) yesterday d. John-i ecey phyenci-lul ssu-E KA-ESSZ-ci (yo) NM yesterday letter AM write SUS I suppose that he began to write a letter yesterday. In (22), the ingressive aspect marker E KA also comes between the verb root ssu 'to write' and tense markers. Sentence (22a) indicates that the speaker supposes that John begins to write a letter with respect to speech time, while the sentences in (22b-d) indicates that the speaker suppose that John will begin or began to write a letter, respectively. However, sentence (22c) is ungrammatical since a time adverb ecey 'yesterday' is not compatible with the nonpast tense marker 02. Thus, the sentences in (22) 44 contrast with those without the ingressive aspect marker E KA as in the following: 23. a. John-i onul phyenci-lul ssu-Oz-ci(yo) NM today letter AM write SUS I suppose that John writes a letter today. b. John-i nayil phyenci-lul ssu-Oz-ci (yo) NM tomorrow letter AM write SUS I suppose that John will write a letter tomorrow. c. *John-i ecey phyenci-lul ssu-E KA-Oz-ci(yo) yesterday d. John-i ecey phyenci-lul ssu-ESSz-ci(yo) NM yesterday letter AM write SUS I suppose that he wrote a letter yesterday. In (22-23) , the nonpast 02 is still contrasted with the past ESSZ, regardless of the existence of the ingressive aspect marker E KA. Thus, the null nonpast tense marker 02 contrasts with the past tense marker ESS2 even when aspect markers such as the sustentive aspect E TWU and the ingressive aspect markers E KA co-occur with these tense markers. This analysis also suggests that aspect marking is followed by tense marking, as assumed in (1) in the section 3.0. This analysis indicates that there should be two slots for aspect and tense marking when aspect is optionally marked, as shown below: 45 24. a. VP --> Root + (Aspect) + Tense b. Aspect Tense | Sustentive Ingressive Nonpast Past I l l u l . E TWU E KA 02 13532 3.1.4 Conclusion With regard to tense, the sentences with the past tense marker ESS2 are contrasted with those without any overt tense marker in nonstative verb constructions and stative verb constructions. Therefore, I propose to set up a null tense marker 02 in order to account for this contrast. Thus, this null tense marker Oz contrasts with the past tense marker ESSZ. Furthermore, this null tense marker 02 can co-occur with time adverbs indicating present and future time reference. This indicates that the null marker Oz refers not only to present time reference but also future time reference, i.e., it has nonpast time reference. Moreover, the null nonpast tense marker 02 also contrasts with the past tense marker ESS2 when aspect markers co-occur with tense markers, just as when they are not marked. 112W In the previous sections, it was argued that the null marker 02 refers to nonpast tense regardless of whether 46 or not aspect is marked, so that it contrasts with the past tense marker ESSz. However, the sentences without any overt tense and aspect markers also refer to nonpast incompletive habitual aspect. So such verbs seem to be two-ways ambiguous: one with the meaning of simple nonpast tense and the other with the meaning of incompletive habitual aspect. Section 3.2.2 will argue that the null nonpast tense marker Oz does not have the two-way ambiguity. Instead, there is a second null marker, 8,, with the meaning of incompletive habitual aspect whose meaning interacts with the null nonpast tense marker 02. One consequence of this analysis is that habitual aspect marker O1 is not freely interchangeable with the progressive aspect marker KO 188, as will be discussed in the section 3.2.2. In Korean, sentences without any overt tense and aspect markers have been claimed to have incompletive aspectual meaning (Sohn, 1975 [1986:160-1621). These have been regarded as having progressive aspect, so that the following sentences have been claimed to have the same meaning as the English progressive. Observe the following sentences: 25. a. tangsin-un olaystongan pap-ul mek-O-so you TOP for a while rice AM eat AUT i) {You have been eating rice for a long time. ii) You have the habit of eating rice for a long time. 47 b. ku-ka olaystongan pap-ul mek-O-ci(yo) he NM SUS i) #I suppose that he is eating rice for a long time. ii) I suppose that he has the habit of eating rice for a long time. However, the examples in (25) do not carry any progressive meaning, as seen in the contrast between the translations. The notation # in (25a.i) and (25b.i) does not mean that the English sentences are ungrammatical, but it indicates that they do not have the intended sense of (25a.ii) and (25b.ii). Moreover, if the sentences in (25) really had the meaning of progressive aspect, they would be freely interchangeable with the sentences with the progressive marker KO ISS. However, they are not. There is an aspect marker KO ISS, which indicates progressiveness, as shown by: 26. a. ku-ka onul swuyeng-ul ha-KO ISS-e(yo) he NM today swimming AM do PROG INF-POL He is swimming today. b. ku-ka ecey swuyeng-ul ha-KO ISS-ESS-e(yo) yesterday He was swimming yesterday. Sentence (26a) refers to present progressive aspect, whereas sentence (26b) indicates past progressive aspect. With these meanings, the sentences in (26) should be reanalyzed as: 27. a. ku-ka onul swuyeng-ul ha-KO ISS-Oz-e (yo) He is swimming today. b. ku-ka ecey swuyeng-ul ha-KO ISS-ESSz-e(yo) He was swimming yesterday. 48 Now, observe the following sentences: 28. a. ku-ka mayil pap seykki-lul mek-OZ-so he NM every day rice three meals AM eat AUT He eats three meals every day. b. *ku-ka mayil pap seykki-lul mek-KO ISS-Oz-so 29. a. ku-nun hangsang hoysa-ey nuckey o-Oz-ci(yo) he TOP always company to late come SUS As for him, I think he always comes to the company late. b. *ku-nun hangsang hoysa-ey nuckey o-KO ISS-Of- ci(yo) In (28-29), the progressive marker KO ISS cannot co- occur with adverbials of habituality, while the nonpast Oz can. This indicates that the nonpast 02 is not freely interchangeable with the progressive aspect marker KO ISS. This also indicates that the nonpast Oz refers to habituality, not to progressiveness. One may raise the question of whether or not the nonpast tense marker Oz is two-ways ambiguous. The sentences in (28-29) indicate that one meaning of the null marker 02 in Korean is simple nonpast tense, whereas another meaning is habitual aspect. How can this apparent ambiguity of the null marker D be explained? I propose that there is another null marker with the meaning of habitual aspect, i.e., 61. 'Therefore, the sentences in (29a) and (29a) should be reanalyzed as: 30. a. ku-ka mayil pap seykki-lul mek-¢,-¢2-so He eats three meals every day. 49 b. ku-nun hangsang hoysa-ey nuckey o-O1-Oz-ci(yo) As for him, I think he always comes to the company late. Why is another null aspect marker 8, necessary? There are other examples indicating habituality without overt habitual adverbs such as mayil 'every day', hangsang 'always', and enceyna 'always' as in (28-29). This will be illustrated by the following sentences: 31. a. ku-ka kongpu sikan-ey cal col-e(yo) he NM study hour in well doze INF-POL He has the habit of dozing off well in class. b. kong-ul cha-l ttay-nun wonpal-lo cha-e(yo) ball AM kick MOD when TOP left foot by kick He has the habit of kicking a ball with the left foot when he tries to kick. In (31), the informal polite marker e(yo) (hereafter INF- POL) is used in polite statements, questions, commands, or suggestions when accompanied by different intonation. As the glosses in (31) show, the sentences without overt habitual adverbs also refer to habituality. Therefore, we cannot say that the sentences in (28-29) have the meaning of habituality because they have habitual adverbs. If habitual adverbs co-occur with the sentences, the meaning of habituality becomes more obvious and specific as in (28-29). Note that there are also sentences with the meaning of habitual aspect when the past tense marker ESS2 is present even when overt habitual adverbs do not co-occur with the past tense marker. The following sentences will illustrate this point: 50 32. a. ku-nun TV-lul po-myense pap-ul mek-ESSZ-SUP-NUN-ta he TOP AM see while rice AM eat HUM DEC He used to eat rice while watching TV. b. celm-ESS-ul ttay achim-ey sanpo-lul ha-Esszwta young MOD when morning in walk AM do I used to take a walk in the morning when I was young. I have argued that the null nonpast OZ and the past ESS2 are contrasted with each other, i.e., they are in complementary distribution. However, if the nonpast 02 were to contribute the meaning of habituality in (31), then it cannot be in complementary distribution with the past ESSZ. IMoreover, if the nonpast O2 were two-ways ambiguous, it would combine tense and aspect meanings, which we have already proposed are assigned to separate verbal slots, as shown in table (1) of the section 3.0. Furthermore, as shown in (31-32), habituality is established without overt habitual adverbs, unlike (28-30). Therefore, it is necessary to set up a second null habitual aspect marker 8,. Moreover, if the sentences without any overt tense marker refer only to the situation holding at and after speech time, how can they be analyzed to distinguish them from other constructions with habitual, progressive, and resultant aspect? Observe the following sentences: 33. a. ku-ka onul thoyscca-lul mac-ci(yo) he NM today rejection AM meet SUS He gets rejected today. b. ku-ka nayil thoyscca-lul mac-ci(yo) . tomorrow He will get rejected tomorrow. 51 c. ku-ka ecey thoyscca-lul mac-ESS: -ci (yo) He got rejected yesterday. When we examine the meaning of these sentences closely, we find that the sentences in (33) do not indicate that the situation of his getting rejected occurs habitually or is going on in (33a-b), or occurred habitually or was going on in (33c). This indicates that the sentences in (33) do not have such aspectual meanings as habituality, progressiveness, and resultantness, etc. In the sentences of (33), however, the contrast between the null nonpast OZ and the past E882 is still preserved, as argued earlier. Therefore, it is plausible to reanalyze the sentences in (33) as (34) having without any aspect marking, respectively: 34. a. ku-ka onul thoyscca-lul mac-(D2 -ci (yo) he NM today rejection AM meet SUS He gets rejected today. b. ku-ka nayil thoyscca-lul mac-Oz-ci (yo) tomorrow He will get rejected tomorrow. c. ku-ka ecey thoyscca-lul mac-ESS 2-ci(yo) He got rejected yesterday. No aspect marker is followed by the nonpast Oz and the past ESS2 since, as shown in the previous section and in (22) and (27-28), aspect is optionally marked. Then, the question arises of why it is not necessary to set up any aspect marker. The sentences in (33-34) do not have the meaning of progressive aspect in (26-27), ingressive aspect in (22), 52 and habitual aspect in (25) and (28-32). Therefore, no aspect marker should be set up in order to distinguish it from other aspectual meanings, especially habitual aspect O1 and the nonpast tense marker 02. This analysis will be represented as in the following: 35. a. VP = Root + (Aspect) + Tense b. Aspect TeTse Completive Incompletive Nonpast Past I l . : I Habitual Non-Habitual PROG SUST INGR I I I I I I : ESS1 O, 1K0 188 E TWU E KA Oz ESS2 (where PROG = Progressive, SUST = Sustentive, INGR = Ingressive) To sum up, it has been argued that the null nonpast tense marker OZ itself does not inherently have the meaning of habitual aspect when the sentences with the null nonpast tense marker 02 have the meaning of habitual aspect. The meaning of nonpast habitual aspect comes from the fact that the null nonpast tense marker 02 is combined with another null habitual aspect marker 8,. Therefore, when the sentences without any overt tense and aspect markers refer to simple nonpast tense, they should be reanalyzed as having only the nonpast tense marker 02 since aspect is marked 53 optionally. The analysis will be summarized as: 36. a. Nonpast Habitual ----> O1-O b. Simple Nonpast ------ > (32 grgrg Habitngiity In the previous section, it was argued that the sentences with the null aspect marker 0, have the meaning of habitual aspect. This section will discuss how habitual aspect O1 differs from progressive aspect in Korean. First, as discussed in chapter I, Comrie (1976:26- 29) positively defines habitual aspect as a situation indiCating an extended period of time. On the other hand, progressive aspect indicates a more or less temporary or contingent situation (Comrie, 1976:38), whose imperfectivity is not habitual aspect (Comrie, 1976:34). This distinction between the extended and the nonextended period of time can be applied to Korean: 37. a. ku-ka sewul-ey sal-01-Oz-e(yo) he NM Seoul in live INF—POL He lives in Seoul. b. ku-ka sewul-ey sal-KO ISS-Oz-e(yo) He is living in Seoul. An extended period of time is indicated in (37a), while (37b) refers to a relatively non-extended or temporary period of time. Thus, (37a) implies that the speaker's normal residence is Seoul, whereas (37b) implies that he temporarily resides in Seoul. This distinction becomes more obvious when adverbials with the meaning of habituality are 54 used. The following sentences illustrate this point: 38. a. ku-ka taytaylo sewul-ey sal-O1-O2-e(yo) he NM for generations Seoul in live INF-POL He (habitually) lives in Seoul for generations. b. *ku-ka taytaylo sewul-ey sal-KO ISS-Oz-e(yo) The contrast between the habitual aspect marker O1 and the progressive aspect marker KO 188 can be illustrated by using habitual adverbs such as maycwu kumyoil 'every Friday'. Observe the following: 39. a. maycwu kumyoil-ey tosekwan-ey ka-O1-OZ-e(yo) every Friday at library to go INF-POL I go to the library every Friday. b. *maycwu kumyoil-ey tosekwan-ey ka-KO ISS-e(yo) In (39), habituality is established by the adverbial maycwu kumyoil 'every Friday', so that (39a) is grammatical since the nonpast OZ with the null aspect O1 refers to nonpast habituality, but (39b) is ungrammatical since the progressive form KO 188 is incompatible with habituality. The distinction between the extended period of time and the nonextended period of time is also clearly illustrated with adverbs with the meaning of limited period of time, as in the following: 40. a. *han sikan tongan phyenci-lul ssu-O1-Oz-e(yo) one hour during letter AM write INF-PLO *I write a letter for one hour. b. han sikan tongan phyenci-lul ssu-KO ISS-e(yo) I have been writing a letter for one hour. 55 In (40), the phrase han sikan tongan 'for one hour' sets a limit on the speaker's writing period of time. In (40a), the phrase phyenci-lul ssu-e(yo) 'write a letter' implies a habitual action, so that the null aspect marker in this phrase is the habitual aspect marker. Also, the phrase of this limited period of time makes (40a) ungrammatical, since sentence (40a) with the habitual 0, refers to the habituality of the extended period of time. However, (40b) is grammatical since the progressive form KO ISS indicates a limited and nonextended period of time by the phrase han sikan tongan 'for one hour'. S.C. Song (personal communication) points out that the null habitual aspect marker 0, can sometimes co-occur with an adverbial phrase indicating a limited and nonextended period of time. This is illustrated by the following sentence: 41. ilnyen tongan supheyin mal-ul paywu-O,-¢2-ci (yo) one year during Spain language AM learn SUS He has studied Spanish for one year. In (41), the verb paywu- 'learn' implies a habitual action, so that the null aspect marker in the phrase ‘paywu-ci(yo) is the habitual aspect marker O5. Sentence (41) indicates that the speaker has taken a course in > Spanish for a year. Song supposes that with adverbials of this type, the null habitual aspect marker 01 takes on a less habitual, but more continuous, feature of nonpast meaning. The acceptability of (41) comes from the fact that 56 taking a Spanish course implies more habitual or repeated study even during the limited period of time, but long enough to form a habit. Song (personal communication) further points out that the nonpast tense marker 02 may co-occur with time adverbs such as camkan 'a little while, a minute'. Observe the following sentence: 42. ku kos-ey camkan tul-le(yo) that place to a minute stop by INF-POL Let's stop by there for a minute. However, the grammaticality of sentence (42) lies in that the nonpast Oz occurs without aspect marking, not with the habitual 01 since the situation of stopping by there does not often imply habitual actions. Therefore, sentence (42) should be reanalyzed as having without any aspect marking: 43. ku kos-ey camkan tul-Oz-le(yo) Let's stop by there for a minute. To sum up, the null habitual aspect marker O1 refers to habituality usually over an extended period of time, but sometimes to that over a limited period of time as in (41). The progressive marker KO ISS indicates the progressive meaning of a limited and nonextended period of time. This difference in meaning prevents the two markers from being freely interchangeable with each other. 57 gr; Co c s' n The past tense marker ESS2 contrasts with the null nonpast tense marker Oz in stative and nonstative verb constructions, regardless of whether or not aspect is marked. It was also argued that sentences without the past tense marker ESSZ, i.e., with the null nonpast tense marker Oz, signify simple nonpast time reference when aspect is not marked. Moreover, the sentences with the null nonpast tense marker OZ may have the meaning of habitual aspect. However, the null nonpast tense marker 02 itself does not have the meaning of habitual aspect and some sentences, in fact, contain a combination of habitual aspect O1 and nonpast tense 02. In other words, when aspect is not marked with nonpast tense OZ, the sentences refer to simple nonpast tense. However, when habitual aspect O1 is marked with nonpast tense OZ, the sentences have the meaning of nonpast habitual aspect. Furthermore, it was also argued that the null habitual aspect marker 0, usually refers to the habituality over an extended period of time, but sometimes to that over a limited period of time. The progressive marker KO ISS indicates the progressive meaning of a limited and nonextended period of time. This difference in meaning prevents two markers from freely being interchangeable with each other. Chapter IV PAST TIME REFERENCE 4.0 Introducrion It is well known that there is a morpheme ESS in Korean. However, the syntactic behavior of the morpheme ESS and of the constructions containing it is not well understood by Korean linguists. This is because they fail to explicate the relationship that exists between the null nonpast O2 and the past ESSZ. Furthermore, the different semantic interpretations of the single form ESS by Korean scholars add confusion. Because of this, some further problems arise in dealing with the double form ESS-ESS. This chapter will discuss the morpheme ESS with the four apparently incompatible meanings of past tense, present and future completion of a situation, and past habitual aspect. It is also well known that there is a double form ESS-ESS, i.e., the ESS cluster, in Korean. The ESS cluster has the meaning of past completion of a situation. An analysis of the E88 cluster presents a crucial clue for resolving the question of how the single form ESS bears the four apparently incompatible meanings of simple past tense, present and future completion, and past habitual aspect. 58 59 In this chapter, I will conclude that there are two homophonous but semantically distinct morphemes: one with the meaning of completive aspect ESS,,iand the other with the meaning of past tense ESSZ. .Also, there is no single ESS when aspect is marked, but, in fact, every instance of ESS is a case of double marking on the underlying representation in which the single ESS is sometimes combined with the null markers, e.g. , habitual aspect O1 and nonpast tense Oz. The purpose of this chapter is: (1) to show the semantic difference between the single ESS and the ESS cluster, (2) to present an analysis of the ESS cluster as composed of a sequence of two different morphemes, i.e., completive aspect ESS1 and past tense ESSZ, (3) to confirm the analysis in the previous chapter that there are two slots for aspect and tense marking on the verb and aspect markers are always followed by tense markers when aspect is marked, and (4), thus, to provide a systematic account of how the single ESS can refer to past tense, present and future completion, and past habitual. The organization of this chapter will be as follows: In section 4.1, I will present problems regarding the single ESS. But, as remarked earlier, it will be necessary to provide some answers to problems of the ESS cluster before an analysis of the single ESS is possible. Thus, section 4.2 will be concerned with the discussion of the ESS cluster. I will, first, present some problems concerning 60 the ESS cluster. Next, this section will argue that the ESS cluster is made up of two different morphemes, i.e., past tense and completive aspect, so that the ESS cluster has the meaning of past completive aspect. Section 4.3 will argue that the four apparently incompatible meanings of the single ESS come from the fact that the single form ESS, in fact, is two homophonous but semantically distinct morphemes which occurs in combination with one of the null markers of habitual aspect O1 and nonpast tense OZ. With this analysis, the four apparently incompatible meanings of the single ESS can be explained. 4.; Problems The most prominent analysis among traditional Korean scholars of the single ESS has been that it is a past tense marker. It is indisputable that the single form ESS may refer to simple past tense, i.e., location of a situation before speech time, without implying such aspectual meanings as habituality, progressiveness, and resultantness, etc. The following sentences illustrate this point: 1. a. wuli cosang-i yekise sal-ESS-ta our ancestors NM here live DEC Our ancestors lived here. b. ku-ka ecey nongkwu-lul ha-ESS-ta he NM yesterday basketball AM do He played basketball yesterday. In (1), the situations of our ancestors' living here and his playing basketball are located prior to speech time but they 61 do not imply whether or not the situations occurred over an extended period of time prior to the speech time. However, the single form ESS seems to have more than the meaning of past tense, e.g., location of a situation before speech time. First, the single form ESS may also refer to present completion of a situation. The following examples will show this: 2. a. ku-ka uyca-ey anc-ESS-ta he NM chair in sit DEC He has just sat in the chair. b. ku-ka swukcey-lul kkuthnay-ESS-ta he NM homework AM finish He has just finished his homework. As shown in (2), the single form ESS is used to express a present completion of a situation. In other words, the situation occurs and is just completed at speech time. However, the single ESS in (2) does not refer to a present state of the past completion of a situation. For example, it is well known that there is a resultant aspect marker E ISS. This resultant aspect marker E ISS refers to the present state resulting from a past situation. In other words, it indicates that states in the past continues to the present moment, while progressive aspect signifies that events in the past continue to the speech time (S.C. Song, personal communication). This is illustrated by the following sentences: 62 3. a. kkoch-i phi-ESS-ta flower NM blossom DEC The flower has just blossomed. b. kkoch-i phi-E ISS-ta The flower is in bloom. In (3), sentence (a) indicates that the flower's blossoming has just occurred, so that (3a) can be true even if the flower had folded. On the other hand, sentence (b) signifies that the flower's blossoming was already completed before speech time but its state is still going on. Thus, the resultant aspect marker E ISS refers to the present condition of the past completion of the situation. The single ESS in (3a), in fact, means that the flower's blossoming has been just completed at speech time, but the speaker does not know whether or not the flower will continue to blossom. On the other hand, the resultant E ISS indicates that a blossom already came out and its state of blooming is still going on. The semantic difference between the single ESS and the resultant E ISS can be illustrated more clearly in the following: 4. a. kkoch-i pangkum phi-ESS-ta flower NM just now blossom DEC The flower has just now blossomed. b. *kkoch-i pangkum phi-E ISS-ta The flower is in bloom just now. In (4), sentence (b) is ungrammatical since the meaning of the adverb pangkum 'just now' implies that a situation occurs right now at speech time. The meaning of the adverb 63 pangkum 'just now' contradicts with the meaning of the resultant aspect marker E ISS and the nonpast tense marker Oz. 'The reason is that resultant aspect indicates the state resultant from the completion of a situation. Therefore, the resultant aspect marker E ISS should not be compatible with the adverb pangkum 'just now' since the resultant E ISS assumes that a situation occurred before speech time and is going on at speech time. Because of the semantic difference between (3a) and (3b), the marker ESS and the marker E ISS are not interchangeable with each other, as is also shown in the following sentences: 5. a. *John-i olaystongan nwu-ESS-ta NM for a long time lie down DEC John lay down for a long time. b. John-i olaystongan nwuwe-E ISS-ta John is lying down for a long time (= John is in a state of resulting from lying down). In (5), sentence (a) is ungrammatical since the meaning of the adverb olaystongan 'for a long time' implies that a situation is holding from the past up to speech time. The meaning of the adverb olaystongan 'for a long time' contradicts with the meaning of the completive aspect marker ESS2 in (5a). The incompatibility of the adverb olaystongan 'for a long time' with the completive aspect marker ESS2 makes (5a) ungrammatical. On the other hand, sentence (5b) is grammatical since the resultant aspect marker E ISS is compatible with 64 the adverb olaystongan 'for a long time'. The resultant aspect marker E ISS assumes that a situation occurred before speech time and is going on at speech time. Therefore, sentence (5b) indicates that John lay down on the bed or something before speech time and his state continues up to speech time. Furthermore, the single ESS may also refer to future completion of a situation. Observe the following sentences: 6. a. wuli-ka iki-ESS-ta! we NM win We will win! b. cokum man te ha-myen ta kkuthnay-ESS-ta. a little just more do if all finish If I do it just a little more, I will be all done. The single form ESS in (6) refers to a situation that will be completed immediately after the speech time, i.e., in the immediate future. In other words, the situations of our winning and my doing in (6) have been partially completed as of speech time and will be finally completed after speech time. The marker ESS in these contexts implies that the future completion is imminent. Sentence (6a) may be used when the game is almost over and a victory of the speaker's team is virtually predictable. Moreover, sentence (6b) may be used when the speaker has almost done and is finishing up his work. Therefore, the sentences become ungrammatical when adverbs with the meaning of relatively distant future time co-occur with the marker ESS. This will be illustrated by the 65 following sentences: 7. a. *naynyen-ey wuli-ka iki-ESS-ta! next year in we NM win We will win next year! b. *cokumman te ha-myen naynyen-ey ta kkuthnay-ESS-ta. little just do if next year in all finish If I do it just a little more, I will be all done next year. Moreover, the single form ESS may also indicate habituality over an extended period of time before speech time. This will be illustrated by the following sentences: 8. a. ku-ka olaystongan pusan-ey sal-ESS-ta he NM for a long time Pusan in live DEC He used to live in Pusan for a long time. b. ku-ka maypen simpulum-ul ka-ESS-ta he NM each time errand AM go DEC He used to go on an errand each time. In (8a), the single form ESS is used to express that his living in Pusan occupied an extended period of time before speech time. The form ESS in (8b) also indicates that the situation of his going on an errand was repeated a number of times over a long period before speech time. These uses of the single form ESS appears to indicate that sentences with it refer to habituality. When we compare the uses of the single ESS in (1-8), we see that it can have the four apparently incompatible meanings: past tense in (1), present completion in (2-4), future completion in (6), and past habitual in (8). The observations by many Korean scholars such as Choi (1937 66 [1980]), Na (1963), Nam (1978b), etc., are essentially correct as far as they go, but they fail to present a systematic account of these apparently incompatible meanings that the single form ESS has. Before I present the analysis of the single form ESS, it is necessary to analyze the ESS cluster as background and to give an explanation of the four meanings of the single ESS. The analysis of the ESS cluster will also give evidence for the idea that two slots for aspect and tense markers are occupied in every verb phrase when aspect is marked. 4.2 qugle Fgrm ESS-ESS 4,2,1 anroduction It is well known that the Korean language has the double form ESS-ESS, i.e., the ESS cluster, which appears to consist of two successive past tense markers (Martin, 1954c35) as follows: 9. a. ku-ka pap-ul mek-ESS-ESS-ta he NM rice AM eat DEC He has eaten the rice. b. ku-ka sewul-ey o-ESS-ESS-ta Seoul to come He has come to Seoul. In (9), the ESS cluster indicates that his eating the rice or his coming to Seoul was completed before speech time but its resultant state is not extended up to speech time. There has been controversy over the differences between the 67 single ESS and the double ESS-ESS. One disagreement is concerned with whether or not the E88 cluster is different from the single form ESS in meaning. A second disagreement is concerned with whether the ESS cluster is composed of a single marker or two different markers. A third disagreement is related to the question of whether the ESS cluster is a doubling of the same morpheme or whether it consists of two different morphemes. A fourth disagreement is related to the following question: if the ESS cluster is made up of two different morphemes, what are they? Each of these questions will be taken up in turn. 4.2.3 ame i e ent? We now examine the question of whether or not the ESS cluster has the same meaning as the single form ESS. First, it has been claimed that the single ESS is a variant of the ESS cluster so both markers are synonymous, so that they are always interchangeable with each other (Gim, 1985; Nam, 1978b; An, 1980). Observe the following: 10. a. kang-i kiph-ESS-ta The river was deep (and the river might or might not be deep). b. kang -i kiph-ESS-ESS-ta river NM deep DEC The river has been deep (but the river is no longer deep now). However, as the glosses in (10) indicate, the speaker knows that the river was deep but he does not know whether or not 68 the river is still deep in (10a). In (10b), the speaker knows that the river was deep in (10b), but also knows that the river is no longer deep. This semantic difference indicates that (10a) and (10b) are not used interchangeably and are not synonymous with each other. The following examples further illustrate this point: 11. a. i san -ey namu-ka wukeci-ESS-ta this mountain at tree NM grow thick This mountain was overgrown with trees. b. i san-ey namu-ka wukeci-ESS-ESS-ta This mountain has been overgrown with trees. However, there is also a difference in meaning between (11a) and (11b). The single form ESS in (11a) indicates that the situation of wukeci 'being overgrown' is located prior to the speech time. On the other hand, the ESS cluster in (11b) alludes to the termination of a past situation of being overgrown, implying that the situation is no longer extended to the speech time. Moreover, the semantic difference between the single ESS and the ESS cluster can be also found in nonstative verb constructions. The following sentences will illustrate this point: 12. a. ku-ka paci-lul ip-ESS-ta he NM slacks AM wear DEC He wore slacks. b. ku-ka paci -lul ip-ESS-ESS-ta He has worn slacks. 69 There is no question, as shown in (12), about the semantic difference between the single form ESS and the ESS cluster. Sentence (12a) refers to simple past tense. In other words, the situation of his wearing slacks occurred before speech time, but the speaker does not know whether or not the situation of his wearing slacks was completed. On the other hand, sentence (12b) indicates that the situation of his wearing slacks occurred and was terminated or completed prior to speech time. Furthermore, this semantic difference is more evident in the case of telic verbs, such as ka- 'go', 0- 'come,' etc. The following sentences illustrate the semantic difference, as shown here: 13. a. ku-ka sewul-ey ka-ESS-ta he NM Seoul to go He went to Seoul. b. ku-ka sewul-ey ka-ESS-ESS-ta He has been to Seoul. Sentence (13a) means that his going to Seoul occurred before speech time and so he has not returned, but he might or might not be in Seoul. Note that sentence (13a) carries no implication of past completion of the situation. Thus, sentence (13a) expresses only that the situation of his going to Seoul occurred before speech time. On the other hand, sentence (13b) signifies that his going to Seoul occurred in the past, and that he is no longer in Seoul. Therefore, the meaning of the ESS cluster 70 is different from that of the single past tense marker ESS. There is another piece of evidence that the ESS cluster is different from the single ESS. The following sentences will illustrate this point: 14. a. b. 15. a. b. 16. a. b. *nay-ka pap-ul mek-ESS-ESS-ko ku-nun I NM meal AM eat and he TOP kwukswu-lul mek-ESS-ESS-ta noodle AM DEC I have eaten the meal and he the noodles. nay-ka pap-ul mek-ESS-ko ku-nun kwukswu-lul mek-ESS-ESS-ta I have eaten the meal and he the noodles. *nay-ka ceki-ey ka-ESS-ESS-taka o-ESS-ESS-ta there to go TRSF come I returned after I went there. nay-ka ceki-ey ka-ESS-taka o-ESS-ESS-ta I returned after I went there. *nay-ka ka-ESS-ESS-umyen ton-ul pel-ESS-ESS-ta I NM go if money AM earn If I had been there, I would have earned money. nay-ka ka-ESS-umyen ton-ul pel-ESS-ESS-ta If I had been there, I would have earned money. In (14-16), the sentences with the ESS cluster is not freely interchangeable with those with the single form ESS. There is no reason that the sentences in (14-16) could not be both used as in (14) if it were correct that the ESS cluster is the same thing semantically or syntactically as the single ESS. Thus, such conjunctions as coordinative ko 'and', transferentive taka 'while', and conditional umyen 'if' do not allow the ESS cluster to co-occur with them, while they allow the single ESS to do. 71 To sum up, it has been claimed by some scholars that the single ESS has the same meaning as the ESS cluster does. Because of this alleged synonymity, it has been also claimed that the ESS cluster can be substituted for the single ESS. However, there is a semantic difference between the ESS cluster and the single form ESS as in (9-16): the single ESS may refer to simple past tense, i.e., location of the situation before speech time. However, the speaker does not know whether or not the situation was completed. On the other hand, the ESS cluster indicates that a situation occurred and was completed or terminated before speech time, and so the situation is no longer extended up to speech time. Therefore, this semantic difference prevents the single form ESS from being freely interchangeable with the ESS cluster. Furthermore, there are also syntactic differences so that there are many conjunctive constructions such as coordinative, transferentive, and conditional, which are compatible with sentences with the ESS cluster, but not those with the single ESS. 4,2,; one gr Tug? 44.24.14 W The present subsection is concerned with the question of whether the ESS cluster consists of one or two morphemes. One piece of syntactic evidence against the idea that the E88 cluster is an inseparable single morpheme is 72 that the single ESS can co-occur with the other aspect markers, while the ESS cluster cannot. The following sentences will illustrate this point: 17. a. ku-ka pap-ul mek-KO ISS-ESS-ta he NM rice AM eat DEC He was eating the rice. b. *ku-ka pap-ul mek-KO ISS-ESS-ESS-ta 18. a. kkoch-i pi-E ISS-ESS-ta flower NM blossom DEC The flower was in bloom. b. *kkoch-i pi-E ISS-ESS-ESS-ta We suggest that the observation that sentences (17b) and (18b) are ungrammatical follows from two assumptions: (1) there is a single slot for an aspect marker on the verb, as described earlier; and (2) the ESS cluster consists of two morphemes, not one. Therefore, the ungrammaticality of (17b) and (18b) results from the fact that two different aspect markers cannot co-occur with each other in the same verb phrase. This indicates that the ESS cluster is not an inseparable morpheme, but it consists of two different morphemes, one of which is an aspect marker. £1211 DQ!211DQ.Q£_N2£Z If, as argued in the previous section, the ESS cluster consists of two different morphemes, what are they? It has been claimed that the ESS cluster can be analyzed into a combination of two past tense markers. According to this analysis, the situation of PAST-PAST is "more remote or 73 more definitely completed than that of past" (Martin, 1954:35). However, consider the following examples: 19. a. secong-i 1443 nyen-ey hankul-ul mantul-ESS-ta Sejong NM 1443 year in Hankul AM make DEC Sejong the Great invented Hankul (Korean alphabet) in 1443. b. John-i ecey achim khal-ul ilhe-ESS-ESS-ta John NM yesterday morning knife AM lose John lost a knife yesterday morning (but the speaker does not know whether or not John has it now). If the ESS cluster refers to a more remote past than the single ESS does, as S.C. Song (personal communication) points out, sentence (19b) should make reference to a more remote past than (19a). But it does not. When (19a) is compared to (19b), the grammaticality of (19b) indicates that the E88 cluster does not indicate a more remote past than the single form ESS in (19a). Therefore, the ESS cluster does not indicate a past more remote than that indicated by the single ESS. Some scholars such as Martin (1954), C. Lee (1987), and Shin (1988), etc., claim that the ESS cluster can refer to the pluperfect, arguing that the ESS cluster is made up of a combination of two past tense markers. In other words, the event should occur before the reference point, i.e., past, which is located before speech time. This definition, however, does not apply to the ESS cluster in Korean. Observe the following sentence: 74 20. John-i caknyen-ey ton-ul manhi pel-ESS-ESS-ta John NM last year in money AM much earn DEC John has earned a lot of money in the last year. If the verb in sentence (20) were the pluperfect, the reference point would be caknyen 'last year' and John's getting money would be also located before that time point. However, sentence (20) indicates that John's getting money was completed in the last year, but the speaker does not know whether or not John is still earning a lot of money now. Thus, the ESS cluster does not refer to the pluperfect. To sum up, it has been claimed that the ESS cluster is composed of two past tense markers and it refers to the remote past or the pluperfect. However, neither of these claims is satisfactory. In the following section, we will investigate the consequences of assuming that the ESS cluster is composed of two different morphemes. 41.2.2 Wham . We propose that the basic solution for this problem lies in that the ESS cluster is made up of a sequence of two successive different markers: one is the completive aspect marker and the past tense marker. Thus, we will analyze the ESS cluster as completive aspect marker ESS1 and the past tense marker ESS2 i.e., ESS1-ESSZ. Observe the following sentences: 75 21. a. ku-ka hoyuy-ey chamsekha-ESS1-ESSZ-ta he NM meeting at attend DEC He has attended the meeting. b. ku-ka ne-eykey cenhwa-lul ha-ESS1-ESS2-ta he NM you to telephone AM do He has called you. Sentence (21a) indicates that a situation of his attending the meeting occurred and was terminated prior to speech time. This meaning comes from the ESS cluster since one member of the ESS cluster has the meaning of completive aspect and the other of it has the meaning of past tense. Specifically, the meaning of completive aspect comes from the first ESS of the ESS cluster, while the meaning of past tense comes from the second ESS of the ESS cluster. The reason that aspect marking precedes tense marking is, as discussed in the previous chapter, that other aspect marking such as progressiveness or resultantness is also followed by tense marking. This analysis also applies to (21b). As discussed for (21a), sentence (21b) signifies that the situation of his calling you occurred and was completed before speech time. Therefore, it suggests that the ESS cluster has the meaning of completive aspect and past tense. $1215 Conclusion The meaning of the ESS cluster is different from that of the single form ESS. The single form ESS has the meaning of simple past tense when such aspect markers as habitual, progressive, and resultant markers, are not 76 marked. This indicates that a situation is located prior to speech time. The single form ESS does not say anything about whether or not the speaker knows that the situation was terminated or completed before speech time. In addition to this, the single form ESS also says nothing about whether the situation occupies an extended period of time before speech time or it is extended up to speech time. The single ESS will be discussed in more detail in the next section. The ESS cluster has the meaning of completive aspect and past tense. This signifies that a situation occurred and was terminated or completed prior to speech time and so the situation is not extended up to speech time. This is because the ESS cluster is made up of a sequence of two successive different morphemes, i.e., the completive aspect marker E881 and the past tense marker ESSZ. This meaning of past tense and completive aspect makes the ESS cluster different from the single form ESS which has the meaning of simple past tense. Furthermore, the two meanings of completive aspect and past tense also indicate that the ESS cluster is neither simply the doubling of the same morpheme such as the past tense marker, nor is it a single inseparable morpheme. 51;. §iDQl§_E§§ In the previous section, it was proposed that the ESS cluster is composed of two different morphemes, i.e., the completive aspect marker ESS1 and the past tense marker 77 E882. This analysis contributes to a unified account of the question of how the single form ESS can have the four apparently incompatible meanings of past tense, present completion, future completion, and past habitual as shown in the previous section 4.0. For convenience, it is worth repeating here each one of these examples that have been mentioned in (1-7) of section 4.0. Observe the following sentences: 22. a. wuli cosang-i yekise sal-ESS-ta our ancestors NM here live DEC Our ancestors lived here. b. ku-ka uyca-ey anc-ESS-ta he NM chair in sit He has just sat in the chair. c. cokum man te ha-myen ta kkuthnay-ESS-ta. a little just more do if all finish If I do it just a little more, I will be all done. d. ku-ka olaystongan pusan-ey sal-ESS-ta he NM for a long time Pusan in live He used to live in Pusan for a long time. The single form ESS in (22a) refers to simple past tense, while it expresses the meaning of present completion in (22b). Moreover, the single form ESS in (22c) indicates the future completion of a situation, whereas it means past incompletive habitual aspect in (22d). I propose that, in (22), there is no single ESS on a more abstract level of representation when aspect is marked since, in fact, the completive aspect marker ESS1Ior past tense marker ESS2 is combined with habitual O, or nonpast tense OZ, as discussed in the previous chapter. Recall that 78 aspect marking is followed by tense marking when aspect is optionally marked and that the null marker O represents incompletive habitual aspect and nonpast tense. This hypothesis can give a systematic explanation how the single form ESS has the four apparently incompatible meanings of simple past, present and future completion, and incompletive habituality, as shown in (22), and how the single form ESS is systematically related to the ESS cluster. £1111 Past Tense As described earlier, repeated here for convenience, the single form ESS may refer to simple past tense, i.e., location of a situation before speech time. The following sentences will illustrate this point: 23.a. nay-ka pap-ul mek-ESS-ta I NM rice AM eat DEC I ate the rice. b. ku-ka chwukkwu-lul ha-ESS-ta he NM soccer AM do He played soccer. In (23), the events of my eating the rice and his playing soccer are located prior to speech time. But the examples do not say anything about whether or not this situation occupies some extended period of time prior to speech time. In other words, the sentences in (23) do not mean that the situations of my eating the rice and his playing soccer occurred habitually or they were still going on. Thus, in (23) aspect such as habituality or progressiveness, etc., is 79 not marked at all. The sentences in (23) should be reanalyzed as: 24. a. nay-ka ecey pap-ul mek-Esszita I ate the rice yesterday. b. kucekkey chwukkwu-lul ha-Esszdta He played soccer the day before yesterday. In (23-24) , when the past tense marker ESS2 occurs without aspect marking, the sentences refer to simple past tense. This does not differ from the assumption that there are two slots for aspect and tense marking since aspect marking is optionally marked as assumed in chapter III. 4.3,; Nonpasr Compietion MW The sentences with the single ESS may also refer to present completion of a situation at the present moment. On the other hand, some scholars such as Na (1963) claim that the marker ESS has the meaning of present completion, but the marker ESS is synonymous with the resultant aspect marker E ISS. This is incorrect since the resultant marker E ISS refers to the present state of past completion of a situation, as discussed in the previous chapter. The meaning of present completion is quite different from past tense. This meaning of the single ESS is aspectual and it is not related to past tense. The following sentences illustrate this point: 80 25. a. ku-ka twis cali-ey anc-ESS-ta he NM back seat in sit He has just sat in the back seat. b. kkoch-i phi-ESS-ta flower blossom The flower has just blossomed now. In (25), the situations of his sitting in the back seat and the flower's blossoming do not refer to the past situations before speech time. Rather, they indicate that those situations have just been completed at the present moment. The question here is how the would-be past tense marker can refer to present completion. Under our account, the meaning of completion of the situation comes from the single ESS, while the meaning of present tense results from the null nonpast tense marker O2. 'Therefore, the single ESS in (25) is not the past tense marker ESSZ, but the completive aspect marker ESS1. 'This conclusion leads us to reanalyze the sentences in (25), as in the following: 26. a. ku-ka twis cali-ey anc-ESS-2-ta He has sat in the back seat. b. kkoch-i phi-ESS-2-ta The flower has hlossomed now. As shown in (26), the meaning of present completion comes from the composition of completive aspect ESS1 and nonpast tense Oz in the verb phrase anc-ESS1-O2-ta 'to have just sat'. To sum up, the single ESS in (25—26) indicates present completion of the situation, not the present 81 condition of past completion of the situation. In other words, the situation occurs and is just completed at speech time. This meaning of present completion comes from the composition of completive aspect ESS1 and nonpast tense OZ. 4.3.2.2 Puture Completion The sentences with the single ESS may also refer to future completion of a situation. As discussed in an earlier section, one problem is concerned with how the apparent past tense marker ESS can refer to future completion. The following illustration will show this meaning: 27. cokumman te ha-myen ta kkuthnay-ESS-ta. a little just more do if all finish If I do it just a little more, I will be all done. We propose that the single ESS in (27) is in fact the completive aspect marker, i.e., ESS1, in combination with the nonpast tense 02, whose meaning is further narrowed by the adverb cokum man te 'a little more'. Therefore, sentence (27) should be reanalyzed as: 28. cokum man te ha-myen ta kkuthnay-ESS -OZ-ta. If I do it just a little more, I will be all done. This combination of nonpast and completive aspect, i.e., ESS1-OZ, can be also illustrated as in the followings: 29. a. nayil John-un honna-ESS-— -ta 2 tomorrow John TOP be in trouble DEC John will be in trouble tomorrow. 82 b. kicha-ka o-ESS1-OZ-tal train NM come The train will have come! In (29), the situations of John's being scolded or the train's coming will be completed subsequent to speech time. Therefore, the marker ESS1 refers to future completion of the situation. In (27), sentence (a) may be used when the speaker knows that a teacher or John's parents will give John a sharp scolding after the present moment. Sentence (27b) may be used when the speaker knows that the train almost is coming to the station and will arrive at the platform soon. To sum up, when the marker ESS1 co-occurs with nonpast tense 02, this combination of ESS1-O2 can indicate present completion of the situation as in (25-26) of the previous section, and future completion of the situation as in (27-29), namely, nonpast completion of the situation. 4 .3 t 'tua The sentences with the single ESS may also have the meaning of the past habitual. Observe the following sentences: 30. a. ku-ka olaystongan pusan-ey sal-ESS-ta he NM for a long time Pusan in live DEC He used to live in Pusan for a long time. b. ku-ka maypen sang-ul ta-ESS-ta he NM each time prize AM receive He used to win a prize each time. 83 It is correct that the sentences with the single ESS can have the meaning of past habitual aspect, but this meaning of past habitual aspect does not inherently come from the single form ESS itself. Rather, as we assumed earlier, the verb phrase in (30) contains in fact the combination of the past tense ESS2 and the null incompletive habitual aspect O1. This combination makes the sentences in (30) carry the meaning of the past habitual. Thus, the sentences in (30) should be reanalyzed as: 31. a. ku-ka olaystongan pusan-ey sal- -O -ESSZ -ta He used to live in Pusan for a long time. b. ku-ka maypen sang-ul ta- -O -ESS2 -ta He used to win a prize each time. Habituality refers to a more permanent state of an extended period of time, as discussed in the previous chapter II. This more permanent state of the situation makes the past tense marker ESS2 compatible with such time adverbs as taytaylo 'for generations', samsipnyen-tongan 'for thirty years', olaystongan 'for a long time', etc., so as to clearly indicate past habitual aspect. The following sentences can illustrate this: 32. a. wuli cosang-i taytaylo yekise sal-ESS-ta Our ancestor NM for generations here live Our ancestors used to live here for a long time. b. samsipnyen-tongan nongkwu-lul ha-ESS-ta thirty years during basketball AM do I used to play basketball for thirty years. 84 The sentences in (32) indicate that the situations of our ancestors' living and my playing basketball occupy a relatively extended period of time before speech time. So, the sentences in (32) do not have the meaning of progressive aspect since progressive aspect refers to a limited and nonextended period of time, as discussed in the previous chapter. Therefore, the sentences in (32) do not co-occur with the progressive aspect marker KO ISS. This will be illustrated by the following sentences: 33. a. *wuli cosang-i taytaylo yekise sal-KO ISS-ESS-ta b. *samsipnyen-tongan nongkwu-lul ha-KO ISS-ESS-ta The sentences in (32) need the habitual aspect marker O1zh1 order to indicate the meaning of habituality. Thus, the sentences in (32) should be also reanalyzed as, respectively: 34. a. wuli cosang-un olaystongan yekise sal-O -ESS -ta Our ancestors used to live here for a long t1me. b. sipnyen-tongan nongkwu-lul ha-O,-ESSz-ta I used to play basketball for ten years. This distinction between extended and nonextended period of time becomes more obvious when adverbials of habituality are used. The following sentences can illustrate this distinction: 35. a. ku-ka mayil swuyeng-ul ha-O1-ESSZ-ta he NM every day swimming AM do He used to swim every day. 85 b. *ku-ka mayil swuyeng-ul ha-KO ISS-ESSzwta 36. a. ku-nun nul achim-ey khephi-lul masi-O1-ESSZ-ta TOP always morning to coffee AM drink As for him, he used always to have coffee in the morning. b. *ku-nun nul achim-ey khephi-lul masi-KO ISS-Essi- ta 37. a. maycwu ilyoil-ey sengtang-ey ka-O1-ESSZ-ta every Sunday at church to go I used to go to church every Sunday. b. *maycwu ilyoil-ey sengtang-ey ka-KO ISS-Esszwta In (30-37), habituality is established by such adverbials as maypen 'each time', olaystongan 'for a long time', taytaylo 'for generations', mayil 'every day', and nul 'always', maycwu ilyoil 'every Sunday', etc. The (a) sentences in (35-37) are grammatical since the aspect marker O1 refers to habituality over an extended period of time, but the (b) sentences in (35-37) are ungrammatical since the progressive form KO ISS is incompatible with adverbs of habituality. However, the (b) sentences in (35-37) may be acceptable if the time adverbs indicating a limited period of time co-occur with adverbs of habituality. The following sentences will illustrate this point: 38. cinan kyewul mayil swuyeng-ul ha-KO ISS-Essivta last winter every day swimming AM do He used to be swimming every day in last winter. 39. hantal kan ku-ka nul achim-ey talliki-lul one month during he NM always morning to running AM ha-KO ISS-ESSz-ta do He used always to be running in the morning for a month. 86 The sentences in (38-39) indicate that the progressive aspect marker KO ISS may co-occur with adverbs of habituality only if a limited period of time is established by adverbs such as cinan kyewul 'in last winter', or hantal kan 'for one month', etc. Thus, the sentences in (38-39) indicate that the situations of his swimming and his running may be repeated over a limited period of time, not over a extended period of time. To sum up, the sentences containing the single ESS have the meaning of past incompletive aspect. However, incompletedness does not come from the ESS itself but from the null habitual O, of the combination of O,-ESSZ. The combination of habitual aspect and past tense, O,-ESSZ. usually refers to the habituality of the extended period of time. The progressive form KO ISS is incompatible with adverbs of habituality. This indicates that the progressive aspect marker KO ISS has the different meaning from the habitual aspect. To conclude, the analysis of the ESS cluster into the combination of the completive aspect marker ESS,Iand the past. tense marker ESS2 provides a clue for a unified explanation of the four apparently incompatible meanings of the single form ESS. The analysis of the ESS cluster suggests that there is no single ESS when aspect is marked and, in fact, every instance of ESS is a case of double marking on a more abstract level of representation in which the single ESS is sometimes combined with the null markers, 87 i.e., the habitual O1 and the nonpast 02. When the single ESS is marked but aspect markers with the meaning of habituality, progressiveness, and resultantness, etc., are not marked at all in simple sentences, the sentences refer to simple past tense. However, when incompletive habitual aspect O, co-occurs with the single ESS, i.e., the past ESSZ, in simple sentences, the sentences indicate past incompletive habituality. Moreover, when nonpast tense Oz occurs and the single ESS, i.e., the completive ESS1, is marked in simple sentences, the sentences signify nonpast completive aspect. Thus, a systematic analysis of the single form ESS is possible when it is postulated that the single ESS is combined with different null markers, i.e., the habitual O1 and the nonpast OZ. 4_._4. 95222114111211 , The ESS cluster is identified as a sequence of two successive separate morphemes, that is, the completive aspect marker ESS1 and the past tense marker ESSZ. This analysis proposes that two slots are necessary for aspect and tense marking in verb phrases in Korean when aspect is marked. This typology provides a unified explanation of how the past and the nonpast markers have the meaning of the simple nonpast or simple past tense. It also presents a systematic account of how the sentences have the meaning of habitual aspect. 88 The analysis provides a reasonable account of how the single ESS has the four apparently incompatible meanings of past tense, present completion, future completion, and past habitual aspect. These apparently incompatible meanings do not always inherently come from the single ESS. The single ESS is, in fact, two homophonous but semantically distinct morphemes, e.g., the completive ESS1 and the past ESSZ. When these two homophonous morphemes co-occur with the 'null markers, e.g., habitual aspect O1 and nonpast tense 25, the single ESS has the four apparently incompatible meanings of past tense, present completion, future completion, and past habitual aspect. Thus, the analysis of the ESS cluster suggests a clue that there is no single ESS. Rather, every instance of single ESS is in fact the case of double markers on a more abstract level of representation in which the single ESS is sometimes combined with the null markers 0 when aspect is marked. The null marker O is really two homophonous but semantically distinct markers: habitual aspect O1 and nonpast tense Oz. The habitual O1 is different from the progressive KO ISS. The habitual O, refers to more or less permanent state of the situation or iterativity over the extended period of time, whereas the progressive KO ISS does not. The nonpast 02 contrasts with the past ESSZ. This analysis may be represented as the following subdivision of tense and aspect in Korean. 89 40. a. VP = Root + (Aspect) + Tense b. Aspect Tense Completive Incompletive Nonpast Past I I | : : I I . I I ' Habitual Non-Habitual I I I I I I I l I I I I I I I I : : PROG SUST INGR : : I I I I I I : : : : : : ESS, 2, KO ISS E TWU E KA (02 13332 (where PROG = Progressive, SUST = Sustentive, INGR = Ingressive) The table in (40) shows that there are two slots for tense and aspect markings when aspect is optionally marked. Incompletive aspect has the subsystem of habituality, progressiveness, and resultantness, etc., which are marked by 8,, KO ISS, and E ISS, respectively. The combination of the markers in the table in (40) provides a unified explanation of the four apparently incompatible meanings of the single form ESS and the relationship between the single ESS and the ESS cluster. It also presents a systematic account of how sentences without overt temporal expressions have the meaning of simple nonpast and nonpast habitual aspect. This will be illustrated by the following: 9O 41. VP = Root + (Aspect) + Tense Nonpast OZ Simple Nonpast O, Oz Nonpast Habitual KO ISS OZ Nonpast Progressive ESS1 Oz Nonpast Completive Past ESS2 Simple Past O1 ESS2 Past Habitual KO ISS ESS2 Past Progressive ESS1 ESS2 Past Completive For instance, when the nonpast OZ occurs without aspect marking, the sentences carry the meaning of simple nonpast. If the completive ESS1 co-occurs with the nonpast OZ, the sentences have the meaning of nonpast completive, i.e., ESS1-OZ. When the incompletive habitual ID, co-occurs with the nonpast Oz, the sentences signify the nonpast habitual, i.e. , 01-82.. If the nonpast Oz co-occurs with the progressive KO ISS, the sentences refer to nonpast progressive, i.e., KO ISS-Oz. Furthermore, when aspect is not marked with the past tense marker ESSZ, the sentences refer to the simple past. Further, when the completive aspect ESS1 and the past tense ESS2 are marked, the sentences mean the past completive, i.e., ESS1-ESS2. If the incompletive habitual O1 and the past tense ESS2 are marked, the sentences express the past habitual, i.e, 0,-ESSZ. When the past tense ESS2 co-occurs with the progressive KO ISS, the sentences indicate the past progressive, i.e., KO ISS-ESSZ. Chapter V CONCLUSION Throughout this study, it has been shown that there are two slots for aspect and tense marking in the verb in Korean when aspect is optionally marked. Also, aspect marking always precedes tense marking if aspect is marked. The past is marked by the marker E882 and indicates the location of a situation before speech time. The nonpast is represented without any overt tense marker, i.e., by the null marker OZ, and indicates the location of a situation at and right after the present moment. The marker ESSzIalways contrasts with the null marker Oz in nonstative verb constructions as well as in stative verb constructions. The sentences with the single ESS have the four apparently incompatible meanings of simple past, present completion, future completion, and past habituality. Also, the sentences with the double ESS, i.e., the ESS cluster, have the meaning of past completion. On the other hand, the sentences with no overt tense and aspect markers, i.e., the null marker (62, have the meanings of simple nonpast and nonpast habituality. 91 92 The analysis of the ESS cluster provides a clue for resolving the question of how a small part of tense and aspect system works in Korean. First of all, it is argued that the ESS cluster is composed of two homophonous but semantically distinct morphemes: one with the meaning of the completive aspect marker ESS1 and the past tense marker ESSZ. This indicates that there are two slots for aspect and tense marking in the verb in Korean. Therefore, it is proposed that there is no single ESS when aspect is marked, but, in fact, every instance of the single ESS is the case of double marking including the single ESS in which it is sometimes combined with one of two null markers, i.e., habitual aspect O1 and nonpast tense Oz. Moreover, the reason that the marker ESS has the four apparently incompatible meanings is that the marker ESS emerges as the completive aspect marker ESS1 or the past tense marker ESSZ. ‘When the single ESS appears in sentences, null markers are necessary to complete verb phrases in Korean since there are two slots for aspect and tense marking in verb phrases when aspect is marked. Thus, two null markers are set up to indicate habitual aspect O, and nonpast Oz. When each of these null markers co-occurs with two homophonous but semantically distinct markers ESS1 and E882, the question of how the single ESS has the four apparently incompatible meanings can be resolved. First, when the null habitual aspect O1 co-occurs ‘with past tense ESSZ, the sentences refer to past 93 habituality. Next, when the null nonpast O2 co-occurs with the completive aspect marker ESS1, the sentences indicate present and future completion, i.e., nonpast completive aspect. If no aspect is marked with the past ESSZ, the sentences signify the simple past. Moreover, note that the sentences without any overt tense and aspect markers have the meaning of the simple nonpast and nonpast habituality. These two meanings of the simple nonpast and nonpast habituality come out when two null markers co-occurs with one another. In other words, when the null nonpast tense O2 occurs without aspect marking, the sentences refer to the simple nonpast. When the null habitual aspect marker O1 co-occurs with the null nonpast tense marker O2, the sentences have the meaning of nonpast habituality. This analysis will be represented as in the following: Nonpast O2 Simple Nonpast O1 O Nonpast Habitual ESS1 O22 Nonpast Completive Past ESS2 Simple Past O1 ESS2 Past Habitual ESS1 ESSz Past Completive Habituality in Korean generally refers to prolongation or iterativity over an extended period of time. On the other hand, progressiveness indicates that events continue to the present moment over a limited and nonextended period of time. 94 The present study provides a systematic account of a small part of the tense and aspect system in Korean. I have discussed only simple sentences in this dissertation. I did not deal with tense and aspect phenomena in sentences with subordinate clauses. This remaining area awaits for future study and, I hope, I will be able to carry on my research beyond the present state into this more complicated phenomena in complex sentences. But I hope it will be a springboard for future studies of the Korean tense and aspect system. LIST OF REFERENCES LIST OF REFERENCES An, Dong-Hwan. 1980. Semantics of Korean tense markers. Ph. D. dissertation, Georgetown University. Choi, Hyun-Bae. 1937. Wuli Malpon (Our Grammar). Seoul: Cengumsa. [8th edition, 1980] Chong, Hi-Ja. 1987. A study of the function of tense and aspect in Korean narrative discourse. Ph. D. dissertation, Ball State University. Chung, In-Sung. 1949. Pyocwun cwungtung Munpep (Standard Secondary Grammar). Seoul: Amunkak Co. Comrie, Bernard. 1976. Aspect. Cambridge and London: Cambridge University Press. . 1985. Tense. Cambridge: Cambridge Univeréity Press. Gim, Cha-Kyun. 1980. "Kwuke sicey hyengthayso-uy uymi (A study on the meaning of Korean tense morpheme)," Hankul (Korean Language) 169:45-116. . 1985. "{ASS}-kwa {ESS}-uy uymi-wa sang (A study of the meaning and aspect of {ASS} and {ESS})," Hankul (Korean Language) 188:3-64. Huh, Woong. 1969. Pyocwun Munpep (Standard Grammar). Seoul: Sinkwu Munhwasa. Jensen, John T. 1990. Morphology. Amsterdam: John Benjamins Publishing Co. Johnson, Marion R. 1981. "A unified temporal theory of tense and aspect," Syntax and Semantics 14:145-175. Kim, Nam-Kil. 1975. "The double past in Korean," Feundations of Language 12.4:529-536. Lee, Chungmin. 1987. "Temporal Expression in Korean," in Verschueren, Jef., and Marcella Bertuccelli-Papi, '(eds.) The Pragmatic Perspective: Selected Papers from the 1985 International Pragmatics Conference. 95 96 Amsterdam: John Benjamin Publishing Co. pp. 435-453. Lee, Hansol Hyun-Bok. 1989. Korean Grammar. Oxford: Oxford University Press. A Lee, Sangkook. 1986. On the Korean tense and aspect marker ESS. Ms. Michigan State University. . 1988. "Korean tense and aspect marker ESS reconsidered," Papers from the Sixth International Conference on Korean Linguistics. Edited by Eung-Jin Baek. International Circle of Korean Linguistics and Department of East Asian Studies, University of _Toronto. pp. 416-431. . in press. "The marker NUN revisited," Proceedifigs from the Fourth International Symposium on Korean Language Education. August 22-24, 1991. Edited by Han Sohn. The Institute of Language Research and Education, Yonsei University. Lieber, R. 1981. "Morphological conversion within a restricted theory of the lexicon," in Moortgat, M., H. van der Hulst, and T. Hoekstra. eds., The Scope of Lexical Rules. Dordrecht: Foris. Martin, Samuel E. 1954. Kerean Morphophonemics. Baltimore: Linguistic Society of America. ., Yang-Ha Lee., and Sung-Un Chang. 1968. New Korean-English Dictionary. Seoul: Minjungseogwan. Matthews, P.H. 1991. Morphology. 2nd ed. Cambridge: Cambridge University press. Na, Chin-Suk. 1963. "Kwuke wumcikssi-uy ttaymaykim yenkwu "(Studies on Korean verbal tense)," Hankul (Korean Language) 134:50-147. Nam, Ki-Sim. 1972. "Hyentay kwuke sicey-ey kwanhan muncey (Problems concerning tense in modern Korean)," Kwuke Kwukmunhak (Korean Language and Literature) 55-57:213- 238. Reprinted in K.-S. Nam (1978:1-34). . 1978a. ”ESS-ESS-uy ssuim-ey tayhaye (On the use of ESS-ESS)," Hankul (Korean Language) 162. Reprinted in K.-S. Nam (1978:133-147). . 1978b. Kwuke MUnpep-uy Sicey-ey kwanhan Yenkwu (Studies on Kerean Tense). Seoul: Tower Press. Park, Chang-Hai. 1964. Hankwuke KWucolon Yenkwu: Hyengthaysolon mich Hyengthayso Payhaplon (Studies on 97 Korean Syntax: Morphology and Morpheme Combination). ,Seoul: Korean Language Institute, Yonsei University Park, Seung-Pin. 1935. Cbsen Ehak (Kerean Linguistics). Photo reprinted with index in 1972. Seoul: Tongmunkwan. Reichenbach, Hans. 1947. Elements of Symbolic Logic. London and New York: The Free Press. Shin, Sung-Ock. 1988. Tense and aspect in Kerean. Ph. D. dissertation, University of Hawaii. Sohn, Ho-Min. 1975. ”Tense in Korean," in H.-M. Sohn (ed.) The Korean Language: Its Structure and Social Projection. Occasional Papers of the Center for Korean Studies. Honolulu: Center for Korean Studies, University of Hawaii. pp. 47-61. Reprinted in H.-M. Sohn (1986:155-165). . 1986. Linguistic Expeditions. Seoul: Hanshin Publishing Co. Song, Seok Choong. 1967. Some transformational rules in Korean. Ph. D. dissertation, Indiana University. . 1976. "Review of Seung-Bin Park: Chosen Ehak (Korean Linguistics)," Ehak Yenkwu (Language Research) 12.1:133-146. . 1988. Explorations in KOrean Syntax and Semantics. Berkeley, CA: Center for Korean Studies, University of California at Berkeley. Spencer, Andrew. 1991. Merphological Theory. Oxford: Basil Blackwell.