\l \ WWWWWWW“HHWIHNH lll\lll\lllllllllllllll\lllllllllllllllllll This is to certify that the thesis entitled ASPECTUAL FORM ~72" [Ill IN JAPANESE presented by Ayako Yamagata has been accepted towards fulfillment of the requirements for .Linguistics degreen 1n WMIJuA—A Major professor Date PHP" I")! 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Mlchlgan State Unlverslty PLACE ll RETURN BOX» monthl- chockoutfrom your record. TOAva FINES Munonorbdmddoduo. DATE DUE DATE DUE DATE DUE MSU lsAn Afflundlvo Action/Emu! Opportunity Institution WW1 ASPECTUAL FORM -TE mu IN JAPANESE By Ayako Yamagata A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Linguistics, Germanic, Slavic, Asian and African languages 1994 ABSTRACT ASPECTUAL FORM -TE IRLI IN JAPANESE By Ayako Yamagata The present thesis examines an aspectual form -te iru in Japanese and defines the core meaning and function of the form, based on the Aristotelian verbal classification. The study proposes that the core meaning of -te iru is determined by a discourse motivated fact, namely, 'heightened temporary relevance' of a described situation with respect to a reference point. It will be shown that all the meanings which have been associated with the form in previous studies, such as progressive, resultative, habitual, repetitive, successive, and experiential are determined based on this core and other situational factors, such as the nature of the described situations, viewpoint setting of the speaker, the use of particles and adverbials, and pragmatic factors such as the speaker's intention. The core meaning/ function of -te iru proposed in this study will explain the occurrences of the form, which have been unexplained in previous analyses. TABLE OF CONTENTS 1. Introduction .............................. 1 2. Aspectual form -te iru .......................... 2 2.1. Aspect ............................... 2 2.1.1. Tense and aspect .......................... 2 2.1.2. Problems regarding aspect ...................... 3 2.1.3. Defining aspect ........................... 6 2.2. Aristotelian aspect ........................... 7 2.2.1. Kenny (1963) ............................ 8 2.2.2. Vendler (1967) ........................... 9 2.2.3. The Aristotelian classification of situations and aspectual category . . . 10 2.3. -Te iru - a grammaticized category of aspect ............... 11 2.3.1. The form and aspectual nature of -te iru ............... 11 2.3.2. -Te iru as a grammaticized category of aspect ............. 12 2.3.3. The meanings expressed by -te iru ................... 14 2.4. Summary ............................... 17 3. Review of previous studies ...... t .................. 17 3.1. Verbal classifications in Japanese .................... 18 3.1.1. Kindaichi (1976) .......................... 18 3.1.2. Pujii (1976) ............................. 22 3.1.3. Machida (1989) ........................... 24 3.1.4. Summary ............................. 25 3.2. Previous work on the meaning of -te iru ................. 26 3.3. Summary ............................... 30 4. Classification of situations in Japanese . . . . . . . ............ 31 5. -Te iru: its core meaning and derived meanings .............. 35 5.1. Problems with states .......................... 35 5.1.1. States in English .......................... 36 5.1.2. States in Japanese .......................... 37 5.2. The core meaning of -te iru as 'heightened temporary relevance of a described situation and further motivations for this View . . . . . . . . 40 5.2.1. Analysis of the -te iru form used with states .............. 40 5.2.2. Discourse function of -te iru ..................... 43 5.3. Types of situations and derived meanings ................ 46 5.3.1. -Te iru with states .......................... 47 5.3.2. -Te iru used with durative actions/ events ............... 50 5.3.3. -Te iru with punctual situations ................... 53 5.3.4. The basic derived meanings ..................... 55 5.4. Other derived meanings and interacting factors .............. 57 5.4.1. Resultative and expriential meanings ................. 57 5.4.2. Progressive and habitual meanings .................. 60 5.4.3. Iterative and successive meanings ................... 62 5.4.4. Plurality of the subject and the successive meaning . . ........ .63 5.4.5. Agentivity and the interpretation of -te iru ............... 66 5.5. Summary ............................... 67 6. Conclusion ............................... 68 1. Introduction The purpose of the present study is to define the core meaning and function of the aspectual form -te iru in Japanese, and explain various meanings associated with the form, based on its core meaning and types of situations classified within the framework of Aristotelian aspect. While the meaning of the -te iru 1 form has been studied quite extensively in the Japanese linguistics literature, it has mainly been at the sentential level. Recently, more attention has been paid to the function of aspect at the discourse level (e.g., Hopper, 1982; Thelin, 1990). The present work advocates an integration of the findings at the sentence level and those at the discourse level, though providing detailed research on the function of this form in natural discourse is beyond its scope. The -te iru form has a core meaning in all occurrences, and it contrasts with other aspectual forms such as —ru (incompletive) and -ta (completive) in terms of its core meaning. I propose that the core meaning of the -te iru form is to show the 'heightened temporary relevance' of a described situation to the reference point in the discourse.2 When ~te iru is used, the described situation is focused and highlighted. This core meaning helps to understand all the occurrences of the -te iru form, including those which haven't received satisfactory account in previous studies on this form. All the different meanings of the form are determined by this core and situational facts such as the types of situations 1 The actual. object of this study is -te-i, rather than -te iru, since -ru in -te iru itself is a morpheme to indicate tense/ aspect. However, since most of the previous work on this form used -te iru, I will follow this tradition to avoid unnecessary confusion. 2 The concept "heightened temporary relevance" was presented in Scheffer (1975), following the research by Van Ek (cited in Scheffer, 1975:38), as a possible central meaning of English progressive form. 2 determined by this core and situational facts such as the types of situations expressed by the predicate, adverbials, and the viewpoint of the speaker. I will also argue that two of the derived meanings, namely, progressive and resultative, are more basic than the others. My observation is based on a classification of various situations and an analysis of the usage of -te iru, rather than on quantitative research. The organization of this thesis is as follows. Following this introductory chapter, Chapter 2 defines aspect and introduces the concept 'Aristotelian aspect' and the form -te iru . Chapter 3 presents a review of the previous studies on verbal classifications in Japanese and the meaning(s) of the -te iru form. Chapter 4 applies Aristotelian aspect to Japanese. Chapter 5 provides an analysis of various derived meanings of the -te iru form based on the classification of the kind of situations given in Chapter 4. In the final chapter, the implications for the future research on the aspectual forms in Japanese will be briefly stated. 2. Aspectual form -te iru This section provides the definition of aspect adopted in this study and introduces Aristotelian aspect, which is a key concept in understanding the form -te iru. Then, a brief explanation on the -te iru form will be given in order to familialize the reader with the aspectual nature of the form and ranges of the meanings expressed by this form. 2.1. Aspect 2.1.1. Tense and aspect "Tense" is a grammatical device for relating an event expressed by the predicate to the reference point, which is usually the present moment or the 3 moment of speech (c.f. Reichenbach, 1947). In this sense, the tense is deictic. If the deictic center does not move in the representation of the events described by the predicates in the sentence or in the discourse, the system is called "absolute tense", while, if there is a shift of the deictic center and the events are described in relation to a reference point other than the present moment, the system is called "relative tense." (Comrie, 1985; Soga, 1983) In contrast, "aspect" is usually considered non-deictic. According to Comrie (1976: 3), "aspects are different ways of viewing the internal temporal constituency of a situation." In other words, aspect is how we view a given event regardless of its location in time. However, this is not the only view on the notion "aspect". Unlike the notion of tense, aspect is quite elusive, and there has been a lot of controversy as to what should be considered as aspect. The next section discusses some of the problems regarding the notion "aspect" and distinguishes aspect from other concepts which are often confused with aspect. 2.1.2. Problems regarding aspect According to Binnick (1991: 136), the term "aspect" is a loan translation from Slavic languages (e.g., Russian vid), which is etymologically related with the English words view and vision. However, despite this meaning of the term itself, the concept which it represents in the linguistic literature is somewhat obscure. Even in the study of Slavic languages where an aspectual contrast is overtly marked by morphology of verbs, exactly what should be considered as aspectual oppositions has been a controversial issue. Among the aspectual oppositions considered are perfect and imperfect, durative and punctual, permanent and transitory, complete and incomplete, habitual] iterative and one-time event. In many European languages other 4 than Slavic, apectual distinctions such as those given above are not morphologically marked uniquely for aspect, but are fused in the forms representing the tense distinctions. For example, in English the present perfect is formed by combining the present tense form of the verb 'have' and the past participle. The expressed event is located in the past, but the reference time is the present moment. Jespersen (1965) and Reichenbach (1947) attempted to explain the present perfect in English in the framework of tense, while Comrie (1976, 1985) treats it as an aspectual form. In some cases, tense and aspect are fused in a single form, and it is hard to distinguish one from the other. For example, the form -ta in Japanese indicates both the past tense and the completive aspect depending on the context (Nakau, 1976; Teramura, 1984). Observe the following examples. (1) A: Kinoo sono bangumi rni-ta. yesterday the program watched 'Did you watch the program yesterday?’ (PAST TENSE) B: Iya, mi -nakat-ta. no watch-NEG-PAST No, I didn't watch it. ‘ (2) A: Moo sono bangumi mi-ta ? already the program watched 'Haveyou already wathced the program?'(PERFECTIVE ASPECT) B: (a) Iya, mada mi -te i-na-i. no yet watch be-NEG-INCOMPLETIVE 'No, I haven't watched it yet.’ (b) Iya, mada mi-na-i. no yet see-NEG-INC No, I haven't watched it yet. (c) *Iya, mada mi-nakat-ta. no yet see-NEG-PAST *No, I didn't see it yet. In (1) the negative answer is given using -ta, while in (2) the negative answer is given in the present tense form of negative morpheme, namely, -na-i. The 5 negative morpheme in Japanese has an adjectival ending. Since adjectives are typically stative, it does not go with the sense of completion. Therefore, -ta in (18) is understood as indicating past tense. On the other hand, since it is ungrammatical to answer (2A) by using a past tense form of the negative morpheme as shown in (2c), -ta in (2A) cannot be indicating the past tense. This kind of -ta is usually considered as indicating the completive aspect. In some other cases the same aspectual form expresses more than one meanings. As will be shown in section 2.3., the -te iru form, when used with different types of verbs, often expresses different aspectual meanings. The idea that the nature of actions or events expressed by the predicates crucially determines the aspectual nature of a described event was proposed by Ryle (1949), Kenny (1963) and Vendler (1967), and is known as "Aristotelian aspect."3 Aristotelian aspect, which deals with verbal aspect, should be separated from the notion of aspect as a grammaticized category.4 I will provide a summary of major work on Aristotelian aspect in section 2.2. Another concept which is often confused with the term aspect is "Aktionsarten". According to Binnick (1991: 458), "Aktionsarten constitutes a classification of expressions for subsituations, phases, and subphases of situations," while aspect concerns "the temporal relationship of a situation to 3 The concept was first discussed in literature by Aristotle (c.f. Dahl, 1981; Binnick, 1991) , and hence is called by this name. 4 The grammaticized category of aspect is realized in the grammar of a language in a systematic way. According to Dahl (1985), the way the grammatical category of TAM (i.e., tense, aspect, modality) is expressed can either be 'by inflection' or 'by derivation'. I adopt Dahl's definition of grammatical category of aspect over the traditional idea that only the morphologically marked opposition of perfect/imperfect as seen in Slavic languages is the grammatical category of aspect. 6 the reference frame against which it is set." We will come back to the issue of distinction between aspect and Aktionsarten when we discuss aspectual categories in Japanese in section 2.3.2. 2.1.3. Defining aspect Binnick's definition of aspect as introduced above is quite close to the one given by Chung and Timberlake (1985: 214): "Aspect characterizes the different relationships of a predicate to the event frame." Timberlake (1985: 153) also states that "aspect can be defined in terms of the relationship of a predicate to time." These definitions are different from Comrie's definition introduced in section 2.1.1.: "aspects are different ways of viewing the internal temporal constituency of a situation." (Comrie, 1976:3) According to Comrie, "aspect is not concerned with relating the time of the situation to any other time-point." (1976: 5) However, closer examination of these definitions would show that they are in fact not so distinct. Comrie defines English perfect as an aspectual form to express "the continuing relevance of a previous situation"(1976: 56), and argues that "it tells us nothing directly about the situation in itself" (1976: 52). This tells us that what he had in mind is the time frame which includes the state after the action or event originally expressed by the root verb. I understand Comrie's point to be that, given the time frame against which the situation is set, the form representing aspect is not particulary concerned with indicating its location in the time with respect to the deictic center or another event. This, of course, does not exclude the possibility that the form may show the location in time as well, especially in languages such as English and Japanese in which the aspectual system is closely interwoven with tense system, as we observed in section 2.1.2. 7 Since all the recent attempts to define aspect that I encountered in the literature are not crucially different (though each author stresses the difference in view from previous ones), I will take Binnick's definition of aspect for the rest of the discussion in the present thesis with no further discussion. His definition of aspect is repeated below: Aspect is a distinction having to do with the relationship of a situation to the temporal frame against which it is set. ( Binnick, 1991: 213 ) In other words, aspect is the different way of viewing or perceiving of an event or action with respect to a given time frame. I propose that grammatical category of aspect is a form or some grammatical device to express this difference in view in a systematic manner in the grammar of a language. When I say "systematic", I mean that the form that expresses aspect has broad applicability to the class of words or elements with which the form is used. With this definition of aspect,_ in section 2.3. I will consider whether or not the -te iru form is a grammaticized aspectual category in Japanese. Before we examine this aspectual form, let us review some of the major work on Aristotelian aspect, which is a key concept in understanding the function and the meaning of the form. 2.2. Aristotelian aspect Aristotelian aspect, which is generally taken as representing verbal aspect, deals with the aspect of situations (i.e., a described event, action or state). Aristotelian aspect is crucial to understanding of the -te iru form, since its meanings vary according to the situations. Below I will provide a brief 8 summary of two seminal work on Aristotelian aspect, namely, Kenny (1963) and Vendler (1967). 2.2.1. Kenny (1963) Kenny (1963) divides verbs into three categories: "static verbs", "performance verbs" and "activity verbs". Static verbs expresses states, e.g., know, love, be happy. Performance verbs are those such as discover, learn, find, and verb phrases like build a house, while activity verbs are those such as laugh, talk, listen to. One of the important criteria for distinguishing static verbs from the other two is that they cannot occur in the progressive form. Other two types, i.e., performance verbs and activity verbs, can appear in the progressive form. The crucial distinction between performance verbs and activity verbs is that the former comes to a definite end while the latter may continue indefinitely. This difference corresponds to the distinction made by Carey (1957): telic and atelic.5 Garey (p. 106) explains telic verbs as "a category of verbs expressing an action tending towards a goal" and atelic verbs as "those which do not have to wait for a goal for their realization, but realized as soon as they begin." Telic situation becomes true only when it reaches its terminal point. Among the verbs he classifies as performance verbs, however, there are verbs, in fact, which do not co-occur with progressive form in an ordinary context. For example, take the verb find. It would not be right to say "I am finding it", since there is not a process of finding something. When one says that s/ he finds something, as soon as one goes into that situation, one has 5 This distinction (telic/atelic ) is later called "bounded/nonbounded" situations by Dahl (1981) and the situations in the "open/ closed intervals" by Bennett (1981). 9 found it, or one hasn't reached the point where S/he can say "I have found it". Zeno Vendler's (1967) classification of verbs may provide a solution to this problem regarding performance verb. 2.2.2. Vendler (1967) Like Kenny (1963), Vendler (1967) also uses progressive forms to categorize verbs. Those which can take progressive forms indicate a process and they are divided into two classes: activity and accomplishment. The former is the same as Kenny's activity verbs and the latter corresponds to Kenny's performance verbs. What distinguishes Vendler's classification from that of Kenny's is that Vendler classifies verbs which lack a process into two subcategories: one is state and the other is achievement. The "achievement" verbs are such as recognize, realize, lose, find, reach the summit, win the race. This class of verbs shares a property with activity and accomplishment. The shared property is that they involve the change of state. The subject of the predicate expressing any of these three (i.e., activity, accomplishment, and achievement) experiences some kind of change either physically or mentally. On the other hand, the subject of state verbs does not usually go under any kind of change. Achievement verbs further share another property exclusively with accomplishment. Both accomplishment and achievement are telic, i.e., they have . terminative point when the expressed action or event comes true. 2.2.3. The Aristotelian classification of situations and aspectual category Mourelatos (1978: 423, 1981: 201) present the following model for the Aristotelian verbal aspect, which summarizes the relationships among the above mentioned situations (slightly simplified): situations (state of affairs) F l 1 states non-states (non-dynamic (dynamic sta es of affairs) states of affairs) l l atelic telic (activities) (performances) development punctual occurrences (accomplishment) (achievement) As is clear in the above diagram, any situations that are non-states are dynamic states of affairs, which involve some kind of change of state. The change is either characterized by the indefinite nature as in the case of atelic situations (i.e., activities), or by the definite nature as in the case of telic situation.6 The situations which involve definite change are further subdivided into accomplishments and achievements. As has been discussed in Section 2.2.2., accomplishments are durative, while achievements are punctual. The significance of the Aristotelian verb classification is that it can provide an explanation for the restriction on the occurrence of a given aspectual form in a certain type of situations. For example, states are ‘ 6 Definite/ indefinite changes are the notions introduced by Dowty (1979) in subcategorizing the change-of—state verbs. 10 11 characterized by its lack of dynamic process, and therefore, cannot be expressed in the progressive form which typically expresses an on-going action. The inherent nature of achievements is also incompatible with the meaning of the progressive form, since achievements typically lack the duration as its internal structure. As will be shown in Chapter 5, distinction among different types of situations are important for the interpretation of the -te i ru form as well, since the most prototypical interpretation of -te iru with a given verb is determined by which situational type the verb typically belongs to. 2.3. -Te iru - - an aspectual category 2.3.1. The form and the aspectual nature of -te iru -Te iru in Japanese consists of the te- form (the gerund) of verbs, and an auxiliary verb iru, which originally means 'be, exist'. The te -form is used to combine two predicates together, and is also used with other auxiliaries. According to Kuno (1973: 195), "-te of the gerundive form originates from the perfect tense auxiliary tari of classical Japanese, from which the past tense suffix -ta of the present-day Japanese is also derived." The form -te iru provides us with a different way of viewing the temporal structure of a given event from the one expressed by other aspectual forms. The following example illustrates this point: (3) a. Taroo wa ichi-ji ni hiru-gohan 0 tube -ta. Taroo-TOP one o'clock at lunch. OBJ eat-PAST Taroo ate lunch at one o'clock. b. Taroo wa ichi-ji ni hiru-gohan o tabe-te ita. Taroo-TOP one o'clock at lunch OBJ eating was Taroo was eating lunch at one o'clock. 12 In (3a) Taroo's eating lunch is taken as a whole, and located at the past time regardless of its internal structure, while in (3b) the same event in the past tense of the -te iru form is viewed as having a stretched time frame within which the situation in question is described as an on-going process. Since both expressions of the event is situated in the past, the difference in the meaning is not one of tense. Rather, the difference concerns the internal temporal structure of the given event within its time frame, and therefore, should be ascribed to the difference of aspect according to our definition of aspect given in section 2.1.3. The next subsection examines the distribution of the -te iru form in contrast with the distribution of other syntactic forms, and further clarifies its aspectual nature. 2.3.2. -T e iru as a grammatical category of aspect Teramura (1984) classifies aspectual forms into three categories: the primary, the secondary and the tertiary, based mainly on their distribution.7 While classifying -ru (incompletive) and -ta (completive) as the primary aspectual categories, Teramura regards -te iru as one of the secondary aspect. I propose, however, that the -te iru form should be regarded as one of the primary aspectual categories for the following reasons: First, the -te iru form has a wide distribution over all kinds of verbs except for a very limited number of stative verbs, as will be discussed in Chapter 5. Second, the 7 According to Teramura, because the -te iru form originates from a lexical compound and has restriction on its occurrence with a certain kind of verbs, it should be considered the secondary aspect. Among other forms considered the secondary aspect are: -te aru (be ...ed of purpose), -te simau (do completely] unfortunately), -te iku (to indicate the action away from the speaker). The aspectual categories classified as tertiary by Teramura consist of such auxiliary verbs as -hazimeru (start ...ing), -tuzukeru (continue ...ing), -kiru (completely finish ...ing), which have stronger lexical nature than the secondary aspectual categories. I regard Teramura's tertiary categories as examples of Aktionsarten discussed in Section 2.1.2. 13 occurrences of the «te iru form with various types of verbs can be unified under a single core meaning, just as -ru and -ta have their unique aspectual meanings in all their occurrences respectively. The core meaning of -te iru, as I see it, is to show the temporal relationship of a described event to the entire reference time frame against which it is set. Therefore, it should be separated from the forms which can only indicate the phases and subphases of a described event, such as -hazimeru (start ...ing ), ~tuzukeru (continue ...ing ), -simau (do ...completely). That -te iru is a primary aspectual category is further supported by the following observation. (4) a. Taroo wa tabe-hazime-te iru. Taroo-TOP eat start is Taroo has started eating. b. *Taroo wa tabe-te i-hazimeru. Taroo-TOP eat be start "Taroo will start being eating. (5) a. Taroo wa tabe-tuzuke-te iru. Taroo-TOP eat continue is Taroo continues to eat. . b. ?Taroo wa tabe-te i-tuzukeru. Taroo-TOP eat be continue ?Taroo continues to be eating. (6) a. Taroo wa tabe-te-simatte-iru. Taroo-TOP eat finish is Taroo has finished eating. b. "Taroo wa tabe-te i-te simau. Taroo-TOP eat be finish *Taroo finishes being eating. In (4) through (6) the -te iru form can be attached to the forms indicating Aktionsarten as in (a) sentences, but not in (b) sentences. This leads us to Speculate that the -te iru form is a more general, inclusive aspectual category than Aktionsarten. Furthermore, if we substitute -te iru with -ta in 14 sentences (4) and (5), we obtain the same result for the grammaticality of these sets of sentences. See the sentences in (7) and (8) below3. (7) a. Taroo wa tabe—hazime-ta. Taroo-TOP eat start -COMPLETIVE Taroo has started eating. b. *Taroo wa tabe—te hazimeru. Taroo-TOP eat start *Taroo will start eaten. (8) a. Taroo wa tabe-tuzuke-ta. Taroo-TOP eat continue-COMPLETIVE Taroo continued eating. b. ?Taroo wa tabe-te tuzukeru. Taroo-TOP eat continue *Taroo will continue eaten. This also suggests that the distribution of -te iru is more like that of -ta than the distribution of the forms such as -hazimeru or -tuzukeru in terms of their placement. Based on the above discussions we conclude that the -te iru form is one of the primary aspectual categories just like -ru and -ta. 2.3.3. The meanings expressed by -te iru The form -te iru indicates a range of meanings depending on the context in which it appears. Some of the meanings which have been discussed in the literature are given below (c.f. Soga, 1983). 8 The auxiliary -simau 'do completely] unfortunately' in the sentence (6) requires te-form of verbs before it. Since the continuative form of -ta is -te, substituting ~te iru with -ta in (6) will give us a sentence which seems perfectly grammatical, i.e., Taroo wa tabe—te-simau. Taroo will finish eating (normal reading)‘. However, if we take -te in tabe-te in this sentence to be the continuative form of a tense/aspect marker -ta, this shouldn't be a grammatical sentence (The intended reading for Taroo wa tabe-te simau here is "Taroo will finish eaten/ate.‘ However, since this example might be confusing, I did not provide the substitution test for (6). 15 (9) Taroo wa ima hon o yon-de iru. Taroo-TOP now book OB] reading is (PROGRESSIVE) Taroo is reading a book now. (10) Hanako wakekkonsi-te i ru. (RESULTATIVE) Hanako-TOP married is Hanako is married. (11) Taroo wa mainiti baa ni it-te iru. Taroo-TOP every day bar to going is (HABITUAL) Taoo goes to a bar every day. (12) Hanako wa izen Amerika ni ki-te iru. Hanako-TOP before America to come is (EXPERIENTIAL) Hanako has come to America before. (13) Sakki kara nandomo to o tatai—te iru. a while ago from many timesdoor-OBJknocking is Since a while ago I have been knocking at the door many times. (REPETTI'IVE) (14) Tugitugito hito ga tatemono kara de-te iru. one after another people-SUB building from come—out is People have been coming out of the building one after another. (SUCCESSIVE) There have been a lot of disagreements concerning the meanings expressed by this form and on the relationships between these different meanings and the types of verbs that occur in the construction. One of the common views by previous researchers (Kindaichi, 1976; Fujii, 1976; Kunihiro, 1982; Soga, 1983; Teramura, 1984) is that the form indicates the progressive meaning with durative verbs, and resultative meaning with so- called punctual verbs. Durative verbs are these which express actions or events that continue for a certain duration of time (e.g., yam u 'read' as in (9)), while punctual verbs denotes actions or events that are not perceived as having duration in them (e.g., kekkon suru 'get married' as in (10)). However, this analysis does not explain why two distinctive meanings 16 should be represented by the same form. I would argue that there should be a motivation for having one and the same form instead of two. It also cannot explain the following cases where all the durative verbs in the -te i ru form denote either the resultative or experiential meaning, rather than progressive. (15) Kesa wa asa-gohan o tabe—te iru kara onaka ga this morning-TOP breakfast-OB] eaten is since stomach-SUB suite imasen. empty-NEG Since I've eaten breakfast this morning, I am not hungry. (RESULTATIVE) (16) Sumisu-san wa Nihon de Fuzisan ni nobot-te iru. Smith TOP Japan in Mt. Fuji to climb is Mr. Smith has climbed Mt. Fuji in Japan. (EXPERIENTIAL) (17) Misima no sakuhin wa subete yon-dc iru. Misima-POSS work TOP all read is I have read all the works by Misima. (EXPERIENTIAL) What is responsible for the resultative or experiential interpretation of these sentences is the telicity of the described situation as will be discussed in Chapter 3, and the telicity is signaled by contextual information such as the use of certain particles, adverbials, and pragmatic factors such as the speaker's intention, as will be shown in Chapter 5. The point here is that since all these examples of the use of -te iru are quite common, the generalization regarding the meanings and the use of -te iru, should be able to encompass these as non-exceptional cases. In the present study I will provide a unified analysis that will explain these occurrences of -te iru . 2.4. Summary Aspect shows the relationship of an event to the temporal frame against which it is set (Section 2.1.2.). Unlike the notion of tense, its basic function is to show the structure of a situation as viewed by the speaker (Section 2.1.1. and Section 2.1.2.). The form -te iru in Japanese is aspectual in this sense. It is a primary aspectual category whose usage contrasts with those of other major aspectual forms such as -ru and -ta (Section 2.3.2.). Various meanings are expressed by the -te iru form. However, the generalization made by previous researchers on the relationship between the kind of verbs and the meaning expressed by -te iru (e.g., the form indicates progressive with durative verbs) does not hold in certain cases (Section 2.3.3.). More problems will be pointed out in Chapter 3, and Aristotelian classification of situations, which was discussed in Section 2.2. can be a key to solving some of the problems. 3. Review of previous studies This Chapter provides a review of previous work on the form -te iru. Section 3.1. reviews verbal classifications in Japanese. Since the traditional classification of verb is problematic for the interpretation of ~te iru, I will propose that it is not verbs by themselves that have to be classified, but the kind of situations expressed by the whole predicates. Section 3.2. reviews previous studies on the central meaning(s) of etc iru and the problems they pose in analyzing the occurrence of -te iru. 17 3.1. Verbal classifications in Japanese 3.1.1. Kindaichi (1976) Kindaichi (1976: 7-12), a seminal work on Japanese aspect, classifies verbs into four categories based on their inherent nature. In classifying verbs, his criteria are whether or not a verb can occur in the —te iru form and what meaning it expresses in the -te iru form. A summary of his classification is given below: 1) Stative verbs: Describe state Cannot be used in the ~te iru form e.g. aru (exist - for inanimate objects), iru (exist - for animate objects), dekiru (can do), wakaru (understand) 2) Durative verbs: Describe an event or an action which continue for a certain duration of time Can be used in the -te iru form in which they indicate action in progress e.g. yo m it (read), kaku (write), wa rau (laugh), utau (sing), taberu (eat), furu (fall - for rain, snow), tiru (fall -- for leaves, petals) 3) Punctual verbs: Describe an event or an action which occurs at a single point in time In the -te iru form they express a state which results from the action/ event expressed by the verb e.g. sinu (die), tuku (belighted), siru (get to know), hazimaru (start), owaru (end), tomaru (stop), kekkonsuru get married) 4) The fourth category of verb: Special kind of stative verbs that are always used in the -te iru form. They express the meanings ofhaving a certain appearance or having certain generic nature. e.g. niru (resemble), sugureru (excel, surpass), omodatu (be prominent), arifureru (be commonplace) 18 19 Kindaichi states that this categorization of verbs is by no means absolute for individual verbs; i.e., the same verb can fall into more than one classes depending on the context. For example, according to Kindaichi (1976: 11-12), yomu 'read' in (1) is used as a punctual verb, and sinu 'die' in (2) is used as a durative verb respectively. (1) Ima yomi-hazime—ta to ommottara moo yon-dc iru. now read ~start-PAST that think-CON] already read is The moment when I thought he started reading, he has already read it. (2) Konogoro wa eeyoossicchoo no tameni tokai no hito ga recently -TOP malnutrition-POSS due to city-POSS people-SUB dondon sin-dc iru. one after another dead is Recently due to malnutrition people in big cities are dying one after another. I basically agree with Kindaichi's view that the verbs change their category depending on the context in which they appear. However, I cannot quite agree with the above observation provided by Kindaichi. I regard yo m u in (1) as a durative verb, since it has a process as is clear from the expression yomi-hazi meta "started reading". I would argue that it is not the property of the verb itself that renders this interpretation, but the property of the whole event hon o yomu 'read a book'. As will be fully discussed in Chapter 5, telicity of the described event together with the fact that the focus is given to the completion of the event are responsible for' the resultative interpretation. As for si nu in (2), it should not be categorized as a durative verb, although the whole situation expressed by the phrase hito ga d ondon shindeiru should be considered as occurring over a period of time rather than at a single moment. We cannot take this sentence to mean that the 20 process of someone's death is going on over a period of time. What the sentence means is repeated occurrences of people's death, each of which is a punctual event. Just as the collection of points constitutes a line, the sequence of punctual events placed on the time line may be taken as a durative situation. This will be discussed more in detail in Chapter 5. Returning to Kindaichi's comment on the variability of the categorization, if the same verb can change the class to which it belongs depending on the context, it is hard to see the significance of categorizing verbs just by themselves. Furthermore, even if we had an exhaustive listing of verbs for each category, if there is too much overlapping of verbs in each category, it would be hard to obtain a generalization on the relationship between the type of verbs and the meaning of an aspectual form. The following pairs of sentences will further illustrate the problems of categorizing a verb by itself. (3) a. Taroo wa Fujisan o nobot-te iru. Taroo-TOP Mt. Fuji-OB] climbing is Taroo is climbing Mt. Fuji. PROGRESSIVE b. Taroo wa Fujisan ni nobot-te iru. Taroo-TOP Mt. Fuji-GOAL climed is Taroo has climbed Mt. Fuji. RESULTATIVE (4) a. Ima Misima no sakuhin o yon-dc iru. now Misima-POSS work-OB] reading is I'm reading Misima's work now. . PROGRESSIVE b. Misima no sakuhin wa subete yon-dc iru. Misima-POSS works-TOP all read is I have read all the works by Misima. EXPERIENTIAL (5) a. Hanako wa uta o utat—te iru. Hanako-TOP song-OBI singing is 21 Hanako is singing a song(s). PROGRESSIVE b. Hanako wa uta o sankyoku utat-te iru. Hanako-TOP song-OB] three-pieces sung is Hanako has sung three songs. RESULTATIVE All the verbs that appear in sentences (3) through (5) belongs to the durative verb in Kindaichi's classification. This is clear if we take out each verb alone and consider its internal structure: noboru 'climb', yomu 'read', utau 'sing'. Each of these indicates certain process and has duration in it. Despite this fact, (b) in all of the above pairs cannot show the process in the -te iru form. The explanation for these phenomena does not lie in the verb itself, but in other elements which forces us to focus on the telicity of the described event] action, as I pointed out in Section 2.3.3. Accroding to Makino 8r Tsutsui (1986: 302), the particle ni in (3b) indicates "a place toward which something or someone moves" (italic mine). McGloin (1989: 29-30) states: "the particle ni . . . puts the focus on the final accomplishment of climbing the mountain." Because of this meaning rendered by the use of the particle ni, the whole situation of yama ni noboru cannot describe a process. Thus, Fujisan ni nobot-te iru in (3b) cannot mean an action in progress.9 Example (4b) cannot express an action in progress because it is impossible to read all the pieces of work written by the author at the same *- 9 In contrast the particle o in (3a) marks a place where the action of climbing takes place. According to Makino & Tsutsui (1986: 349), this kind of 0 "indicates a space in] on] across] through] along which something or someone moves." Therefore, with this Particular particle a verb like noboru necessarily indicates a process, and Pujisan o nobotte iru in (3b) expresses an action in progress. 22 time. Here the adverbial subete 'all' together with the verb is responsible for the resultative reading, but not the verb itself. The use of this adverb makes it impossible for us to conceive of a process in the described event and forces us to focus on the telicity of the event. The same line of argument can explain why we receive the resultative interpretation for (5b). It is hardly imaginable for someone to sing three songs at the same time. We see that the quantifier sankyoku 'three songs' here forces the resultative interpretation and makes it impossible to have the progressive interpretation. There is thus strong evidence for arguing that it is not the verb itself that has to be classified but the kind of the situation described by the predicate. 3.1.2. Fujii (1976) Fujii (1976) offers two categories of verbs in addition to Kindaichi's four categories. Kekka doosi 'result verbs' proposed by Fujii is a kind of verbs that leave a certain observable or perceivable resultant state after the action] event described by the verb has occurred. For example, take a verb like otiru '(something) fall(s)'. When something has fallen on the ground, we can observe a fallen object there. Kekka doosi contrast with hi-kekka doosi 'non- result verbs', which do not necessarily leave an observable / perceivable result. Kekka doosi does not necessarily contrast with already established other types of verbs but can be found among both durative and punctual verbs. For example, among durative verbs, verbs like yam u 'read' and utau 'sing' are categorized as non-result type, while verbs like ti ru '(flowers and leaves) fall' and kiru 'put on (clothes)' are categorized as result type. Punctual verbs also show contrast of result/non-result: verbs like (ziko ga) okoru 'happen' and (soto e) tobidasu 'run out' are categorized as non-result type, while verbs like kekkon suru 'get married' and mi in ka ru 'get found' are categorized as result 23 type by Fujii. By this opposition of result/non-result types Fujii attempted to show that only the result verbs can express the resultant state in the -te iru form. This is obviously not true, since, as we already observed in (4b) and (5b), the verbs like yomu 'read' or utau 'sing' can express the resultant state in the -te iru form. Therefore, Fujii's distinction of result] non-result verbs cannot be very useful for explaining the distribution of different meanings of -te iru, either. Moreover, what is meant by 'perceivable' or 'observable' is not clear in the above examples. For example, both (ziko ga) okoru 'some accident happens' and (soto e) tobidasu 'run out' can leave an observable state. This is clear if we only imagine the scene after a car accident for the former, and the person standing in a yard (after his running out of the house) for the latter, for example. Although Fujii claims that this type of verbs can only indicate the experience, not the resultant state, the -te i ru form of these verbs can clearly express the resultant state in the situations such as those I just described above. Unless what is observable /perceivable is further clarified, Fujii's distinction of result] non-result verbs does not help in the interpretation of -te iru . In fact, whether or not a given action or event described by the verb leaves observable situation is not something that is decided by the nature of the verb itself. Observe the following sentences. (6) a. Taroo ga hon o wasure-te iru. Taroo-SUB book-OB] forget is Taroo has forgot] left his book. b. Taroo wa sando shukudai o wasure-te iru. Taroo-TOP three times homework-OB] forget is Taroo has forgot his homework three times. (6a) is likely to be uttered in a context where we can observe the book which Taroo has left, while (6b) does not usually accompany any observable fact, but 24 rather, is likely to be uttered based on the speakers knowledge about Taroo's past conduct. The former is usually taken as having resultative meaning, while the latter is usually associated with experiential meaning. In the above pairs it is not the meaning of the verb wasureru by itself that gives rise to this difference in our perception of the situation (i.e., whether it leaves observable result or not), but it is the whole event laction described by the predicate together with contextual information that must be considered. 3.1.3. Machida (1989) Machida (1989) proposes that not only verbs alone but the verb phrases must be taken into account. Following Comrie's telic/atelic distinction (Comrie, 1976: 71), Machida subcategorizes durative verbs into two categories: genkaiteki (=telic) and hi-genkaiteki (=atelic). Machida observes that telic verbs or verb phrases can indicate resultative aspect when they occur in the —te iru form. Machida also observes that verbs change their telicity depending on the kind of object they take. For. example, kooen o hasiru 'run in a park' is an example of atelic verb, while hyaku-meetoru hasiru 'run 100 meter' is an example of telic verb. He further analyzes that the adverbial phrase that appear with the form is also crucial to the interpretation of -te iru. (The same observation has been given by other researchers such as Soga (1983), Ando (1986)). However, since he sticks to the Kindaichi's classification of verbs, saying that "the aspectual meaning of the -te iru is basically dependent on the verb types" (1989: 18), his classification does not clarify the relationship between the types of situations and the meanings expressed by -te iru. A major problem in Machida's analysis is the following: based on his belief that fourth category verbs in Kindaichi's classification should be considered as a kind of adjectives rather than verbs, Machida does not 25 include fourth category verbs' in his analysis of the verbal aspect. This is rather an ad hoc solution but is inevitable, given his belief that it is 'verb' types that are crucial for the interpretation of -te iru . Fourth category verbs have been problematic for the analysis of -te iru: they represent generic state but they are mostly used in the -te iru form unlike other stative verbs. They are often regarded as an exceptional, ideosyncratic group of verbs in Japanese. However, since the occurrence of -te i ru with fourth category verbs is systematic, I would argue that it would be more reasonable to think that there is some reason for this systematicity. Moreover, if we are to provide generalization about the aspectual system of Japanese as one which reflects the universal properties of the aspectual systems of natural languages, treating such a phenomenon in Japanese as ideosyncratic or exceptional is to be avoided. Ideally, the behavior of fourth category verbs should also be explained in the same framework in which other types of verbs are treated. 3.1.4. Summary In the preceding sections we have examined the classical categorization of verbs in Japanese and observed evidence for its ineptness in explaining the meanings expressed by -te iru. Since there is no satisfactory solution to the problems within the framework of this verb categorization, I propose that the classification of situations is necessary for providing satisfactory account for the given interpretation of the -te iru form. The following sections further show the problems we face when we depend on the verb types to explain the meaning of -te iru. 3.2. Previous work on the meaning of ~te iru Many of the past researchers who looked into the -te iru form (e.g., Takahashi (1976), Yoshikawa (1976, 1986)) have associated two meanings with the form, i.e., progressive and resultative. Some other researchers have attempted to find a uniform analysis for all the different uses of this form (e.g., Suzuki (1976), Kunihiro (1982), Teramura (1984), Ando (1986)), though none has been quite successful in it. The main problem for all the previous studies is that the kinds of verbs, rather than the kinds of situations are considered for the basis of their analysis. Another problem is that, despite their attempt to provide a uniform view, the use of -te iru with stative verbs was often disregarded. Fujii (1976) lists seven meanings of -te iru: 1) action in progress; 2) continuance; 3) resultant state; 4) experience; 5) simple state; 6) repetition; 7) existence. Fujii does not distinguish the concept of successive from repetitive. In his example, a sentence such as hito ga dondon sindeiru 'people die one after another' is labeled as repetition, though in fact it is not a repetition of the same act by the same person but rather successive events of dying by different people. Furthermore, he lists one use of the -te iru form in the relative clause in his categorization (7), and not all the uses of the form in relative clauses. His list is therefore somewhat ad hoc. Unlike Fujii many researchers are trying to find a central meaning or meanings of -te iru. For example, Takahashi (1976) proposes that the form indicates progressive and resultative, and that the latter includes inherent state expressed by fourth category verbs, such as ni-te iru 'resemble' and sobie-te iru 'tower'. However, fourth category verbs do not necessarily show a resultant state in -te iru. We don't say Yama ga sobie-te iru. 'A mountain is towering.‘ as a result of the mountain's coming into the world. The 26 27 mountain has been there even before the speaker notices it. Therefore, the central meanings proposed by Takahashi are not adequate to explain the occurrence of -te iru with fouth category verbs. Yoshikawa (1976, 1986) also proposes two basic meanings, i.e., progressive and resultative. Following Fujii's (1976) distinction of result] non- result verbs, Yoshikawa proposes that the progressive reading is obtained with durative verbs and the resultative with result verbs. This observation, however, does not always hold as we already saw in 3.1.2. Some other researchers tried to provide a uniform analysis for the occurrences of -te i ru. For example, Suzuki (1976) hypothesizes that -te iru indicates 'continuance' in contrast with the basic form -ru. Okuda (1977, 1978 - cited in Teramura, 1984: 125) and Ando (1986) also defined the meaning of -te iru as "durative" or "continuative." However, if -te iru marks just durativity or continuation, we cannot explain why durative verbs, which already indicate durativity or continuation, are used with the form. It is also puzzling that stative verbs, though they also indicate durativity or continuation, do not appear in -te i ru, while durative verbs do. Majewicz (1985) also analyzes -te iru to mean 'duration in progress' though he does not provide an explanation for how the duration of a given event could be in progress like a dynamic action. Machida (1989) proposes that -te i ru indicates the incompletive aspect while the -ru form expresses completive aspect when we have non-stative verbs. When a described event is stative, however, Machida analyzes that the incompletive aspect is expressed by the -ru form. Thus we do not have any account of the occurrence of -te iru with states in his analysis. Moreover, as already mentioned in the previous section, Machida excludes fourth category 28 verbs from his analysis of verbs. Therefore, we do not have an explanation for the occurrence of -te iru with fourth category verbs in his analysis. Kunihiro (1982) also provides a uniform analysis on -te iru, i.e., 'the continuence of the state resulted from the completion (or the realization ) of some event'. According to Kunihiro, if the lexical aspect (sic.) of the given verb is punctual, what -te iru expresses becomes more like the meaning represented by the present perfect in English, and if the lexical aspect of the given verb is durative or repetitive, the progressive meanng is obtained. Teramura (1984) also presents the same kind of view as Kunihiro's and states that the form indicates 'existence of the result of the completed or realized event'10 Teramura further analyzes that when a given action is durative by nature, the form indicates continued existence of the result of the initiation or realization of the event, while if the event does not have duration in it, the form indicates the continued existence of the result of the completion or the end of the event. Teramura provides the following diagram for these difference. (7) a. yuki ga fut-te iru 'the snow is falling' : 1.......---.- now (01" a reference point) The snow started to fall, and it is still falling. b. saihu ga oti-te iru 'the wallet is on the ground (the wallet has fallen) o----------------- ---------t now (or a reference point) The wallet fell and it is still on the ground. —_.__ 10 Teramura uses the word 'kizen' for representing both completed and realized events. 29 Teramura's observation is almost identical to the one given by Kunihiro. However, Teramura stresses the importance of the context, saying that the above mentioned observation regarding the kind of verbs and the meaning of -te iru typically holds unless there is a specific context which alters the situation. Teramura points out that this typical relation is what the speaker or the hearer are most likely to associate when no specific context is provided. Teramura's implication for the importance of the context is a big step forward toward a unified analysis of the -te iru form. A problem arises when he deals with the occurrence of -te iru with fourth category verbs . Just like other researchers, Teramura also treats this phenomena as somewhat exceptional, and does not provide any analysis within the same framework in which other use of -te iru is analyzed. Kudo (1989) unfolds a rather different view. Kudo proposes that there are two meanings for -te iru. One is what she calls 'durative' or 'continuative' and the other meaning is what she calls "perfect." "Perfect" in Kudo's sense is not the same perfect which is usually translated as kanryoo in Japanese. Kudo argues that her "perfect" should not be considered the same as kanryoo represented by -ta form in Japanese which can only mean present perfect (sic), since "perfect" represented by the -te iru form can mean any of the present, past, and future perfect. Kudo further argues that unlike "continuative", which is basically aspectual, "perfect" subsumes the "anterior" sense as a tense element. This is confusing and misleading, since she is mixing up tense and aspect. The concept she named "perfect" can be explained in the relative tense system without establishing a new concept. I will argue that the well extablished perfective or resultative meaning can indicate any of the past, present, future perfect with the "shift of viewpoint", changing the deictic center taken by the 3O speaker or the writer. If what is intended by Kudo's analysis was that -te iru can represent both tense and aspect (which is a reasonable assumption), tense and aspct should be clearly defined and be applied to the interpretation of the sentence independently from the other. There is, however, an important point in Kudo's analysis of the -te iru form which deserves much attention and further consideration. Following Givon, Hopper, and Thompson (cited in Kudo, 1989: 61), who argued that "the motivation for aspectual system is discourse functional", she analyzed the use of the -te iru form at the discourse level. Although there is a confusion of aspect with the concept of relative tense and many things remain unclear, Kudo's view on the function of this form seems to be on the right track. 3.3. Summary One of the major problems in previous work on -te iru is that the verb types rather than the types of situations were used to capture a generalization on various occurrences of -te i ru.. Another major problem is that the occurrences of -te iru with fourth category verbs, despite their systematicity, are taken as exceptional, and ideosyncratic to Japanese. In order to arrive at a generalization on the use of this aspectual form, however, we need a unified concept which encompasses all the systematic occurrences of the form as non- exceptional cases. The core meaning and function of -te i ru I propose in this thesis will provide such a unified concept. Before we go on to the analysis of various occurrences of the form, we shall classify various situations expressed in Japanese in the framework of Aristotelian aspect. 4. Classification of situations in Japanese Probably the most unfavorable effect of the traditional classification of verbs is that the four classes of verbs are considered to be, in principle, mutually exclusive (although it may not have been Kindaichi's original intention). When four distinct classes are presented, we may not suspect whether any of verbs have some property in common. However, if we consider the kinds of situations as was presented in 2.2.3., we see that punctual situations and durative situations in general share a certain property, i.e., dynarnicity of a situation or a change of state. We also see that both durative and punctual situations can be found under the category of telic situations. The categorization of the verbs alone may well confuse us in differentiating the interpretation of a given situation rather than help us capturing it. Below I will provide the categorization of situations expressed in Japanese. By doing so, I will argue that this way of classifying situations rather than verbs alone is far more informative and advantageous in explaining a given interpretation of the -te iru form. no change of state involved change of state involved atelic telic states duration punctual activity _ accomplishment achievement 1. states group A: stative situations which cannot be expressed with -te iru 1) Situations described by verbs of existence (e.g.,iru 'be, exist', aru 'be, exist') 2) Situations described by i-type adjectives (o to nasi i 'be good tempered', takai “be expensive, high' 3) Situations described by nominals and the copula d a which indicate 31 32 a permanent state or an attribute of the described subject. (e.g., i n u da rbe a dog', hon da 'be a book') 4) Situations described by na-type adjectives with non-agentive subject (e.g., mise wa hima da 'the shop is uncrowded', ryoo wa sizuka da 'the dorm is quiet') group B: stative situations which can be expressed with -te iru 1) Situations described by stative verbs such askotonaru 'differ', h u ku m u 'contain', tekisuru 'be fitted, be suitable' (This group of verbs express generic state in the -ru form, but can take the -te iru form to indicate temporary state.) 2) Situations described by fourth category verbs (e.g., sobieru 'rise, tower',sugureru 'excel, surpass', niru 'resemble', arifureru 'be commonplace', bakageru 'be absurd' (This group of verbs express generic state, and nomally appear in -te iru.) 3) Situations described by nominals and copula da which can optionally indicate temporary state. (John wa ii ko da 'John is a good boy',Taroo wa gakusee da 'Taroo is a student')“ 4) Situations described by na-type adjectives used with agentive subject. (e.g.,Taroo wa ima hima da 'Taroo is free now', Mary wa kiree da. 'Mary is pretty') 2. activities (atelic durative situations): When activities appear in -te iru, they typically indicate a progressive meaning. e.g. kooen o hasiru 'run in a park', hayaku hasiru 'run fast', warau 'laugh', naku 'cry', oyogu 'swim' 3. accomplishments (telic durative situations): When accomplishments appear in -te i ru, they typically express either a progressive or resultative meaning. e.g. hon o yomu 'read abook', san-satu no hon o yomu 'read three (volumes of )books' hyaku-meetoru o hasiru 'run (a distance of ) a hundred meters', yama o noboru 'climb a mountain (process)', 11 Situations such as ii ko da 'be a good boy' or gakusee da 'be a student' by themselves indicate an attribute of a described subject. However, when they appear in -te iru form, the sentences express a temporary state which the described subject are in. In contrast, situations classified as (2) in groupB can never appearin -te iru (e.g., Kore wa kaban da. This is a bag. vs. Kore wa kaban dc iru. This is being a bag.'). This will be fully discussed in Section 5.1.2. 4. achiev St expresse fourth c latter gr situatio predicai in grou] mdo“ in Secti. .1 hasiru State (1. verbs al. Situatior A: may Chdl noboru '( 7 always a PP1 Where they ; 33 e o kaku 'draw a picture', ie 0 tateru 'build a house' 4. achievements (telic punctual situations): When achievements appear in -te iru, they typically indicate a resultative meaning. e.g. sinu 'die', mitukeru 'find', tuku 'arrive', hon o san-satu yomu 'read three books', hyaku-meetoru hasiru 'run a hundred meters', yama ni noboru 'climb a mountain (reach the summit)’ States are classified into two categories: those which cannot be expressed in the -te iru form, and those which can. The situations described by fourth category verbs in Kindaichi's classification of verbs are included in the latter group.12 Since this is not a classification of verbs, but a classification of situations, I also include the stative situations expressed by nominal predicates and adjectives. The contrast between situations classified (3) and (4) in group A on the one hand, and (3) and (4) in group B on the other, has a lot to do with the controllability of the situations. This will be disscussed more in Section 5.1. A careful observation reveals. that the Kindaichi's durative verbs (e.g., hasiru 'run') are actually distributed across three of the situation types except state (i.e., activity, accomplishment, and achievement). Categorizing just verbs alone is useless, since different verb types do not correspond to different situation types. As we observed in 2.3.3., in Japanese the use of particles in a sentence may change the type of the situation described" by the sentence (e.g., yama o noboru 'climb a mountain (process)' vs. yama ni noboru 'climb a mountain 12 It might be confusing to say that they 'can' appear in -te iru, since they actually almost always appear in -te iru as predicate (i.e., at the sentence final position) except for some cases where they appear in proverbs (e.g., [(0 we oya ni niru 'Children resembles their parents). 34 (reach the summit)'). The nature of the object of a verb may also change the type of the described situation. For example, kooen o hasiru 'run in a park' is an activity, while hyaku meetoru o hasiru 'run (a distance of) 100 meters' is an accomplishment. Since it is an accomplishment,Hyaku meetoru o hasitte iru can mean either an action in progress:'(S/ he) is running (a distance of) 100 meters.‘ or the resultant state:'(S/he) has run (a distance of) 100 meters.‘ This is due to the definite nature of the object hyaku meetoru '100 meters' vs. indefinite nature of the object kooen 'a park'. In other words, one can continue to run in a park forever unless s/he wants to stop running, or someone/something stops her] him from running, while one's running of 100 meters comes to an end when 5] he reaches the finishing line. Now, if we take out the particle o from hyaku meetoru o hasiru 'run (a distance of) 100 meters', it becomes an achievement. Since achievements lack a process, hyaku meetoru hasitte iru can only mean the resultant state, namely, '(S/he) has run 100 meters.‘ At this moment, I do not have an explanation for this last phenomenon. In order to see in what environment a given verb express an activity, an accomplishment, and an achievement respectively, the function of particles in Japanese has to be further clarified. The importance of the Aristotelian classification of situations lies not only in the fact that it can explain the restriction on the occurrence of -te iru with certain types of situations, but also in that it will tell us the tendancy for a given verb in the -te iru form to have a certain interpretation. For example, durative verbs like yo m a 'read', when they appear in -te i ru, are likely to be interpreted to mean an action in progress, unless its telicity is focused. Punctual verbs like otiru 'fall' in -te iru may be most likely to be associated with a resultant state meaning, unless it is taken as having duration. In fact, some of the things Kindaichi (1976) pointed out on the relationship between 35 verb types and the meaning of -te iru are correct, if only they are taken as tendencies. In the following sections, I will analyze various use of -te iru based on the classification of situations given in this chapter, and show how all the different interpretations can be determined based on the possible core meaning of this form. 5. -Te iru : its core meaning and derived meanings This section presents my analysis of the core meaning of the -te iru form. As will be shown in 5.2., the core meaning of this aspectual form is the heightened temporary relevance of a described situation at the reference point in the discourse. The situation described by -te iru is often a focused event in the discourse. As will be clear in the following discussions, this analysis of the core meaning and function of -te iru can explain its occurrence in various situations in a uniform manner and is superior to any of the past explanations which left many phenomena unexplained. I will specifically argue that all the meanings associated with -te iru as has been proposed earlier are derived meanings, and that the contextual information, including the type of the described situations, crucially determines the interpretation of the form. 5.1. Problems with states In this section , I will analyze the relationship between the -te iru form and 'states'. Many of the phenomena we will observe below provide a motivation for positing a core meaning and function of the -te iru form. As has been pointed out by Kenny (1963), Vendler (1967), and others who studied Aristotelian aspect, states typically cannot be expressed by the 36 progressive form except in certain cases. Since one of the meanings expressed by-te iru is action in progress, we expect the same kind of restrictions on the occurrence of -te iru when a given situation indicates state. Before we examine Japanese data, let us look into the relation between the kind of states and the progressive form treated in the study of aspect in English. 5.1.1. States in English Comrie (1976: 36) points out that many verbs are "treated sometimes as stative, sometimes as non-stative, depending on the particular meaning they have in the given sentence." If we take Comrie's view, we can simply exclude those situations expressed by the non-stative use of stative verbs from the situation labeled as 'state'. Dowty (1979), on the other hand, leaves those instances of stative verbs in the situation type labeled as 'states', but subcategorizes states into those that can appear in the progressive form and those that cannot. I will take Dowty's position, since I assume that the non- stative use of stative verbs pointed- out by Comrie (1976) indicates a state.13 According to Dowty, agentive adjectives and predicate nominals as well as some group of stative verbs with human subject can appear in the progressive form, while non-agentive stative verbals cannot. Examples are given below1 4: (1) Tom is being polite now. AGENTIVE (2) John is standing by the door. (with animate subject) AGENTIVE 13 For example, Tom is being a good boy indicates a state of Tom's being a good boy temporarily. It is not an activity, accomplishment, or achievement. 14 Dowty's use of agentive/non-agentive opposition does not necessarily correspond to the opposition of the subject being animate/inanimate. Non-agentive situations also includes the situations expressed with animate subjects but are non-controllable. Dowty (1979:185) points out that his agentive] non-agentive opposition is the same as that of controllable] non-controllable. 37 (3) a. Tom is in the garden. NON-AGENTIVE b. "Tom is being in the garden. (4) a. The house stands on the hill. NON-AGENTIVE b. "The house is standing on the hill. (5) a. Iknow him. NON-AGENTIVE b. *I am knowing him. Dowty's generalization about the relationship between the agentivity of states and the possibility of verbs to occur in the progressive form is also relevant in the analysis of Japanese-te iru. Next section examines agentive] non-agentive states described with -te iru. 5.1.2. States in Japanese Observe the following instances of the use of the -te iru form. (6) Taroo wa itu made gakusei de iru ki da. Taroo-TOP when until student being is mind is Until when he remains to be a student? (When does he intend to start working? ) (7) Hanako wa kyoo wa ii ko de iru. Hanako-TOP today-CONTRAST good girl being is Hanako is being a good girl today. (8) Itu made mo kirei-de iru no wa muzukasii. forever pretty being is-NOM-TOP hard It is hard to stay pretty forever. In the above examples, nominals with the copula and adjectives all appear in the -te iru form. The situations described in (6) - (8) are agentive, or controllable situations, and their grammaticality is explained by Dowty's observation. The point becomes clear if we look into the use of adjectives and nominals in non-agentive constructions such as those given below: 38 (9) a. Asoko wa mada resutoran da. that place-TOP still restaurant is That place is still a restaurant. b. *Asoko wa mada resutoran de iru. that place-TOP still restaurant being is That place still remains to be a restaurant. (10) a. Kyoo wa ryoo wa sizuka da. today-CONTRAST dorm-TOP quiet is The dorm is quiet today. b. *Kyoo wa ryoo wa sizuka de iru. today-CONTRAST dorm-TOP quiet being is The dorm is being quiet today. The inability of the copula and adjectives to appear in the -te iru form in the above examples (9) - (10) is explained in terms of the non-controllability of the discribed situations. Here we see that Dowty's observation on the relationship between the agentivity, or non-controllability of states and the progressive form of English is also applicable to the relationship between the agentivity, non-controllability of states and -te iru in Japanese. However, this does not apply to all the situations expressed by the stative predicates in Japanese. First, the situations expressed by fourth category verbs (e.g. sobieru 'rise, tower', sugureru 'excel, surpass', niru 'resemble', arifureru 'be commonplace', omodatu 'be prominent'), which do not indicate agentive or controllable situation, usually appear only in the -te iru form as illustrated below: (11) a. *Takai tatemono ga kawa no mukoo ni sobieru. tall building -SUB river - Poss across on tower b. Takai tatemono ga kawa no mukoo ni sobie-te iru. tall building - SUB river - POSS across on towering is The tall building is towering on the other side of the river. (12) a. *Kono modemu wa sohuto ga sugureru. this modem - TOP software-SUB excel 39 b. Kono modemu wa sohuto ga sugure-te iru. this modern - TOP software-SUB excellent is The software of this modem is excellent. Second, some stative verbs indicating relations that hold between the subject and another argument (e.g. kotonaru 'be different', hukumu 'contain', taioosuru 'correspond, match', tekisuru 'be fitted, be suitable') can optionally be expressed by the -te iru form even though they are typically non-agentive. See the following examples: (13) a. Kono kinzoku wa tetu o hukumu. this metal -TOP iron - OBJ contain b. Kono kinzoku wa tetu o hukun-de iru. this metal -TOP iron -OBJ containing is This metal contains iron. (14) a. Kono kutu wa yamanobori ni tekisuru. this shoes-TOP mountain-climbing to suit b. Kono kutu wa yamanobori ni tekisi-te iru. this shoes-TOP mountain-climbing to suitable are This shoes are suitable to mountain-climbing. If we assume that the occurrences of the -te iru form with states are exceptional cases where the situations described are agentive, we cannot explain the (b) sentences in (11) - (14), where the use of the form with non- agentive states is grammatical. None of the previous analyses of the possible core meaning of -te iru can capture the phenomena we have just observed. There is some other factor involved in addition to the opposition of agentive] non-agentive that constrains the co-occurrence of the -te i ru form and states. It may be reasonable to speculate that this factor is something that has general applicability rather than language-specific. In the next section I will propose the possible core meaning and function of -te iru, and will further seek the motivation for positing this core meaning. 5.2. The core meaning of -te iru as 'heightened temporary relevance' of a described situation and further motivations for this view In the following sections, I will argue that the core meaning of the -te i ru form is the 'heightened temporary relevance' of a described situation at the reference point in the discourse. The reference point usually corresponds to the speech moment in conversational text, though it could often be some other point in the narrative discourse. The situation described by -te iru may be a focused event in the discourse, and its relevance at the reference point is heightened by the use of this form. 5.2.1. Analysis of the -te iru form used with states One strong motivation for postulating that the -te iru form indicates 'heightened temporary relevance' of a described situation is that this notion can explain the occurrences of the form which was problematic to the earlier analyses. Observe the following sentences: (15) a. Shinjuku wa takai biru ga sobieru mati da. Shinjuku-TOP tall building-SUB tower town is Shinjuku is a town where tall buildings tower. b. Me no mae ni takai biru ga sobie-te iru. eye-POSS front at tall building -SUB towering is There is a tall building towering in front us. (16) a. ?Kono suponzi wa mizu o takusan hukumu. this sponge -TOP water-OB] a lot contain This sponge contains a lot of water. b. Kono suponzi wa mizu o takusan hukun-de iru. this sponge -TOP water-OB] a lot containing is (Lit.) This sponge is containing a lot of water. (17) a. Taroo wa kanji ga kireeni kak-e-ru. Taroo-TOP Chinese character-OB] neatly write-POT Taroo can write Chinese character neatly. b. Taroo wa kono kanji ga kireeni kak-e-te iru. Taroo-TOP this Chinese character-OB] neatly write-POT is 40 41 (Lit.) Taroo is being able to write this Chinese character neatly. (18) a. Ko wa oya ni niru. child-TOP parent(s) to resemble Children resemble their parents. (proverb) b. Sono ko wa oya ni ni-te iru. the child-TOP parent to resembling is The child resembles his parents. Generic situations such as those expressed by fourth category verbs, some stative verbs, and potential verbs in Japanese transcends the concept of time. In other words, they do not have to be assessed its truth value with respect to the reference point. If we anchor such situations to the present moment, however, it would restrict their truth value as being relevant only at the present moment, and thus, highlights their relevancy at the present moment. This is exactly the case in examples (15) though (17). The relevance of sobieru 'tower' in (15b) is restricted to the speech moment by the use of -te iru , and contrasts the description of the attribute of the town in (15a). Incidentally, in English the same contrast is obtained with the verb 'tower'. If one says 'tall buildings are towering', it means that the described situation is relevant only within his ]her current frame of reference (i.e., 'within the speaker's glance'). It contrasts with the expression 'tall buildings tower', which also refers to the state outside the speaker's current reference time frame. This further supports the validity of our analysis of -te iru. In (16a) the statement 'this sponge contains a lot of water' describes the property of the sponge. It is not something that changes depending on the context. (17a) also express a state which does not change over time, i.e., Taro can write Chinese character neatly. In contrast, (16b) and (17b) represent specific observed situations. They express their relevancy at the speech mement, but it may not be true at some other point in time. The use of -te iru is thus relevant in 42 these expressions. The contrast of (18a) and (18b) may be slightly different from the above cases, but it also supports our core meaning of -te iru. Since (18a) is a saying, niru 'resemble' in (18a) is supposed to apply to any child, and therefore, is used as more like a generic statement. On the other hand, ni- te i ru 'resemble' in (18b) applies only to the particular child described here, and its relevance is limited within our current reference frame. The temporary relevance of the described situation within this reference frame surrounding the moment of speech is heightened by the use of -te iru. Thus we have strong support for the view that the aspectual form -te iru indicates 'heightened temporary relevance' of a described situation. A problem arises when we consider inpossibility of the occurrence of -te iru with certain situations such as those shown in (b) of the following sets: (19) a. Ima Taroo wa niwa ni iru. now Taroo-TOP garden in is b. *Ima Taroo wa niwa ni i-te iru. now Taroo-TOP garden in being is Taroo is in the garden now. (20) a. Heya wa ima kirei da. room-TOP now clean COP b. *Heya wa ima kirei-de iru. room-TOP now clean is The room is clean now. (21) a. Sono mise wa ima mo kissaten da the shop—TOP now even coffee shop COP b. *Sono mise wa imamo kissaten - de iru . the shop-TOP even now coffee shop being is That shop is still a coffee shop. These are the situations which were labeled as non-agentive or non- controllable by Dowty as introduced in section 5.1.1. Our new definition of the -te iru form is also relevant for explaining the above cases where the use 43 of the form with given situations is ungrammatical. Unlike fourth category verbs and some stative verbs, which indicate generic states or inherent properties of the described object, verbs of existence such as iru 'be, exist' and aru 'be, exist', adjectives and nominals with copula ending which I classified in group A of states in Chapter 4 can indicate relevancy of a described situation at the present moment. Therefore, the latter group of predicates do not have to be in the -te iru form in order to highten its relevancy at the speech moment. The meaning provided by the -te iru form is redundant information for the description of a given situation for this group of predicates. Thus, we can account for the ungrammaticality of (b) sentences of (19) through (21) with our core meaning of -te iru . 5.2.2. Discourse function of -te iru This section provides further support for our analysis on the possible core meaning of the -te iru form. As was pointed out by Kudo (1989), many linguists have been speculating that the motivation for the existence of the aspectual oppositions is to provide the cohesion in the text. Hopper and Thompson (1980) argue for the role of aspect to mark the distinction of foregrounding and backgrounding of the events in the discourse. Givon (1984) also points out that temporal coherence, or event sequentiality, is an important factor for maintaining coherent discourse, and that aspectual category of 'perfect' in many languages is marked for 'counter-sequentiality' and 'current relevance'. Recently, as more attention is paid to the function of the aspect system in the discourse level, an integration of the findings at the sentence level and those at the discourse level is becoming an important subject in the study of aspect. If the existence of an aspectual system is motivated by its function in the discourse, 44 analysis of an aspectual form which does not reflect its discourse functional meaning cannot be useful even at the sentential level. As Thelin (1990: 22) puts it, temporal perspectivization cannot be understood properly unless related to a broader pragmatic concept of situation structure and aspect (the semantic system of perceptual-conceptual distinctions structuring this temporal perspective), correspondingly, cannot be fully understood unless treated as a function of discourse organisation assigned only secondarily to individual propositions or sentences. In this respect our definition of -te iru as 'heightened temporary relevance of a described situation with respect to a reference point in discourse' may be further supported by its function at the discourse level. Before we examine how this core meaning helps our understanding of all the occurrances of the form, let us briefly go over Kudo's (1989) analyses of -te iru , which, to my knowledge, is the first serious attempt to find its discourse function. Following Okuda (cited in Kudo, 1989), who pointed out the importance of the relationship among sequence of events in discourse in Japanese aspectual system, Kudo (1989) puts forth the view that aspects performs a text-constructing function. Based on this View, she proposes two functional meanings for the -te iru form: the continuance (or duration) and the perfect. Both the continuence and the perfect in Kudo's sense show a contrast with the completiveness of the situation represented by -ru form. As I pointed out in Section 3.2., her analysis with regard to the perfect meaning of -te iru confuses tense and aspect. However, the following observation made by Kudo (1989: 79, 81) is insightful: ...site i ta ( perfect ) introduces preceding events without disturbing the story line given by the sequence of events expressed by si ta. 45 This seems to mean that -ru/- ta forms introduce events in time sequence and construct the basic skeleton of the text, while -te iru and -te ita (the past tense form of -te iru) show the simultaneity or the precedence of a given event to the event introduced by -ru and -ta. Kudo (1989: 78, 80) provides the diagrams such as (22) (alphabetical letters are used rather than actual verb phrases used in Kudo's original diagrams). Let us suppose that A, B, C, D indicate events introduced by -ta , and that X and Y indicate events introduced by -te i ta . The broken line indicates continuance of the event X, and the arrow shows completion of the event Y. (22) A E E: i) -ta -ta -ta -ta X Y -te ita -te ita Ibasically agree with Kudo regarding the function of -te iru and -te ita in discourse. What distinguishes my view from Kudo's is that I unify the two uses of -te iru proposed by Kudo and posit that the difference is obtained by other factors (i.e., not an inherent property of the form itself). In my analysis whatever the interpretation of -te iru in a given text, the function of the -te iru form is to focus on the described event, and heighten its relevancy at the reference time point. In a simple sentence the reference point is usually the moment of speech. In a complex sentence or a narrative discourse, however, an event described by -te iru and -te ita may have a reference point other 46 than the speech moment.15 Events indicated by -ta , namely A, B, C and D in the above diagram are introduced into the text in their totality regardless of their internal structure, and construct the main story line. The use of -te ita in the description of the events X and Y draws listener's (or reader's) attention out of the main story line for a second, and pulls it onto the situation expressed by the -te ita form. I will argue that the form makes our eyes focus on the internal structure of the time frame against which the described situation is set. Thus, a situation described by -te ita is a highlighted, or focused event in the discourse, and its temporal relevancy at the reference point gets heightened, showing contrast with the same situation being described by -ta. Based on the above analysis of the central meaning and function of the -te iru form, in the next section , I will provide an analysis of the occurrances of -te i ru in various situations, and will explain how a given interpretation of -te iru is obtained in each case. 5.3. Types of situations and derived meanings This section examines various sentences where -te iru is used, based on its core meaning presented in the previous section and the classification of the types of situations given in Chapter 4. By doing so, I will try to provide the reasons for a given interpretation of the form in each situation type. As will be shown in the course of discussion, two of the derived meanings, 1 5 In the above diagram B serves as a reference point for X, and D serves as a reference point for Y. The two events have different reference points in this text, and neither of the reference points are the speech moment. 47 namely, progressive and resultative, are more common than others, and thus regarded as somewhat prototypical with the form. 5.3.1. -Te iru used with states It is commonly held that situations classified as states basically cannot be expressed with the -te iru form (e.g., i ru 'exist', gakusee da 'be a student', isogasii 'be busy'). However, as we observed in section 5.1.2., some states such as the one described by the fourth category verbs are normally expressed in -te i ru . We also observed that adjectives and nominals with copula that indicate controllable stative situations also take -te iru . The phenomena such as these can be easily explained by the nature of the situations and our definition of the core meaning of -te iru . See the following diagram. (23) situations described by the sentences such as yama ga sobie-te iru 'the mountain is towering' or kono sofuto wa sugure-te iru 'this software is excellent‘ sobieru sugureru sobie-te iru te iru sobiiru sugureru @ As was discussed in Section 5.1.2., the situation expressed with -ru of the fourth category verbs express generic state or inherent property of the described object. Therefore, the relevancy of the expression like *yama ga sobieru 'the mountain towers' or *kono sofuto wa sugureru 'this software excels' do not indicate their relevance at the moment of speech. However, when -te iru is attached to these expressions, they become relevant at the moment of speech or within the time frame perceived as the present, and receive the interpretation in the real world. One may argue that there are 48 cases in which the expression with the fourth category verbs do not specifically show the relevancy at the present moment. (24) a. Ano gakusee no sakubun wa itumo sugure-te iru. that student-POSS composition -TOP always excellent is The compositions written by that student are always excellent. b. Taroo ga iu koto wa tokidoki bakage-te iru Taroo-SUB say -NOM- TOP sometimes nonsense is What Taro says is sometimes nonsense. However, our assumption on the core meaning of -te iru still holds. I argue that although the use of adverbs such as i tumo 'always' or tokidoki 'sometimes' makes the sentences sound as if they do not specifically show a relevancy of a described situation at the present time, they do, in fact, only over a wider scope of time frame. The temporal frame against which these situations are set is spread along the time line to include all the recent time instances which are perceived as present time, and the relevancy of the given situations are assessed within this broader frame. The diagrams for sentences (24a) and (24b) are shown in (25a) and (25b) respectively. (25) a. Ano gakusee no sakuhin wa itumo sugure-te iru. itumo sugure-te iru t5! 49 b. Taroo ga iu koto wa tokidoki bakage-te i ru. \ tokidoki bakage-te iru / The adverb i tumo 'always' in (24a) expand our reference frame to include all the instances where the student's composition being excellent as well as the speaker's expectation towards the future as shown in (25a), while the adverb tokidoki 'sometimes' in (24b) indicates that Taroo's comment being foolish is peceived as less frequent within the given time frame as is shown in (25b). In both cases the function of -te iru is to anchor the generic state onto the time frame in the real world and to heighten its relevancy within this time frame which centers around the moment of speech. Now let us examine the use of -te iru with nominal predicate. In (26) -(27) -te iru indicates the temporality of the described situation. With its temporality being emphasized, the relevancy of the described state at the speech moment is heightened. (26) Taroo wa ima no tokoro ii ko de iru Taroo-TOP for the time being good child be is For the time being Taroo is being a good boy. (27) Taroo wa mada gakusee de iru. Taroo-TOP still student be is Taroo is still being a student. 50 (28) situations described by the sentences such as i iko de iru, gakusee de iru iiko de iru gakusei de iru realization of states Unlike the situation expressed by phrases such as iiko da rbe a good boy' or gakusee da 'be a student' which normally indicate an attribute of the described subject, the -te i ru counterparts of these indicate temporal state which the subject is in. In (26) the boy became a good boy at a certain point in time and is still observed as being a good boy at the moment of speech. His being a good boy within the time frame, therefore, is highlighted against more usual occasion of his not being a good boy. Here again, our definition of core meaning of -te iru can explain the occurrences or -te iru, which had been problematic in previous analyses. 5.3.2. -Te iru used with durative actions/events In this subsection I will examine the occurrence of -te iru with durative actions] events. According to the classification of situations given in Chapter 4, situations which have duration include activity and accomplishment, i.e., both atelic and telic situations. Although activity by nature is atelic and does not have to have a definite terminal point to it, it is usually perceived as coming to an end eventually due to the physical limitation on the part of the performer of the activity. Accomplishment on the other hand, has to have a definite terminal point for it to represent what it means. Based on the above observations, we see that both activity and 51 accomplishment as a durative situation have the following internal structure.16 (29) durative situations A M realization completion The above diagram can account for at least two variations for the interpretation of durative situations expressed with -te iru. If -te iru indicates the existence of the realized or completed event as Teramura (1984: 127) suggests, we would expect the ambiguity in its interpretation with durative situations such as that shown in the diagram (29), given the fact that there are both realization and completion for a described situation. The following example shows this point. (30) Sonobangumi o mi-teiiru. that program-OB] watch is a. (5] he) is watching the (TV) program. b. (5] he) has seen the (TV) program. Observe the following diagrams for (30a) and (30b) given below as (31a) and (31b) respectively. 16 This diagram is drawn based on the diagram provided by Soga (1983:29) in representing different stages of the process of an action or event. 52 (31) Sono bangumi o mi-te iru. a. (S / he) is watching (TV) program. b. (5] he) has seen the (TV) program. finish watching the program completion \ As is clear in (31a) and (31b), the difference in the meaning can be explained in terms of the difference in viewpoint set by the speaker or the bearer. If we set our viewpoint onto the initiation of the situation, with its internal structure being focused, we obtain the progressive reading as in (30a), while if we set our viewpoint onto the completion of the situation, with the state following the completion of the event in our reference frame, we obtain the resultative reading as in (30b). Although Kunihiro (1982) and Teramura (1984) pointed out that -te iru indicates either the continuance of the initiated or completed event, their claim was not intended to account for the ambiguity such as shown in (31). The claim was basically to associate durative situations with progressive 53 meaning and punctual situations with resultative reading. As we saw in the above diagram, however, durative situations can be associated not only with progressive, but also with resultative reading. This is because we can set our viewpoint to either of the two points which signals the change of state into or out of the durative situation. In both cases -te iru shows the heightened relevancy of the described situation to the reference point, which is regarded as the moment of speech unless the context set its reference point otherwise. Since our viewpoint setting is crucially determined by each environment in which the expression with -te iru appears, the context in which the form appears is often a decisive factor in its interpretation. However, verbs like yomu 'read' and hasiru 'run', even though they could indicate any of activity, accomplishment and achievement, depending on the environment in which they appear, may be more often associated with progressive meaning in -te iru due to their durative nature. I would, therefore, regard progressive as a default reading of the -te iru expressions with durative verbs. 5.3.3. -Te iru with punctual situations Punctual situations, or achievement in Vendler's (1967) term is a kind of situation which does not have internal duration as the initiation of the situation coincides with its completion. Because of this nature of punctual situations, no ambiguity is expected to arise with the use of -te iru in punctual situations in nomal cases, and it is perceived as expressing a resultant state. See the following diagram for an example sentence describing a punctual situation. (32) Saihu ga oti-te iru. 'A wallet has been fallen. wallet-SUB fall is (and it is on the ground)‘ saihu ga otiru © 'the wallet falls' realization=completion A sentence like (32) is normally understood as expressing a resultant state. In (32) the state resulted from the preceding event is focused (rather than the internal structure of the event itself). A progressive reading is impossible for this sentence, as it has no internal duration. If we add the phrase yuka ni 'on the floor' to this sentence, resultative interpretation is fostered. This may be because it is hardly imaginable for one to observe and describe the process of dropping of purse from someone's pocket onto the floor, which is a momentary event. However, there are occasions when punctual situations are perceived as having internal duration. If a given punctual situation is magnified and stretched along the time line, it becomes possible to perceive the same situation as having duration in it. For example, see the following sentence with a diagram. (33) Ano biru kara nanika ga oti-te iru. that building from something-SUB fall is Something is falling from that building. nanika ga otiru 'something falls’ realization @ tcompletion 55 The sentence (33) describes a process of something falling from a building. The witnessed situation is perceived as having an internal structure just like a durative situation, with the adverbial phrase ano biru kara 'from the building' singnaling a ralatively long path passed through by the falling object. The resultative reading is difficult to obtain with (33), since it may be hard to decide where the fallen object came from when we focus on the object already fallen on the ground. The adverbial phrase 'from the building' draws our eyes onto the origianl location of the falling object rather than the place it dropped onto. Our viewpoint is set to the initiation of the event, and the process of 'something falling' gets focused. However, unless the context forces us to have this kind of reading, punctual verbs like otiru 'fall', tiru '(petals, leaves) fall', kiru 'put on' are usually associated with a resultant state in -te iru due to its nature which lacks duration. Therefore, I would regard resultative as a default reading for the -te iru expressions with punctual verbs. 5.3.4. Basic derived meanings In the previous section we observed that achievements or punctual situations expressed with -te iru are typically interpreted as resultative. With the realization of an event being simultaneous with its completion, we set our viewpoint onto the completion of the event, and accordingly, the state resulted from the completion of the event rather than its internal structure is focused. What is crucial to the resultative reading is the telicity of a described situation. According to the classification of situations given in Chapter 4, two classes, namely, accomplishments and achievements are characterized by their telicity. As I noted in section 5.3.2., an activity, too, is most likely perceived as having an end point due to either the physical limitation of the human being or the time constraints. Thus, three out of four types of 56 classified situations (i.e., activity, accomplishment, achievement) satisfy the condition for the resultative reading. It is no wonder that the resultant state is viewed as somewhat basic or typical among various meanings reperesented by the -te iru form. In the same way, we can also account for the reason why progressive is commonly associated with -te iru. As we saw in Section 2.2.3. and Chapter 4, Aristotelian categorization of situations tells us that a decisive factor for the progressive reading is whether or not we can perceive dynamicity in the described situation. Both activities and accomplishments have dynamic internal structure and therefore, can be associated with on-going action unless their telicity is focused on. Even achievements which typically lack internal duration could represent progressive meaning if a given context makes us look into the internal structure of the momentary event (Section 5.3.3., (33)). Moreover, states can be perceived as on-going event if a change of state is implied for some contextual reason (Section 5.3.1., (26)-(28)). Accordingly, all the types of situations we classified can possibly express progressive meaning. Thus we see that most of the situations can represent either progressive or resultative meaning with the -te iru form. It is reasonable to speculate that these two meanings are probably most prototypical. Nevertheless, I would argue that they are not the core meanings expressed by the form itself. Both progressive and resultative are the meanings naturally obtained from the nature of the described events and the speaker's (or hearer's) viewpoint in a given specific context. We should not forget that neither the progressive or resultative meaning is obtained when -te iru is used with the states described by the fourth category verbs. Therefore, progressive and resultative are 57 simply two basic meanings of the form which are derived from its core meaning. 5.4. Other derived meanings and interacting factors In this section I will Show how other interpretations of the form such as habitual, experiential, iterative and successive are determined. It will be shown that all these meanings are actually derived from the two basic meanings, namely, progressive and resultative. These peripheral meanings are determined by a number of situational factors. Thus, individual context in which the form appears becomes crucial to the interpretation of these meanings. 5.4.1. Resultative and experiential meanings Observe the difference between (a) and (b) in the following pairs. (34) a. Tanaka-san wa moo koko ni ki-te iru. Mr. Tanaka-TOP already here to come is Mr. Tanaka has already come here (and he is here now). b. Tanaka-san wa maeni koko ni ki-te iru. Mr. Tanaka-TOP before here to come is Mr. Tabaka has been here before. (35) a. Hanako-san wa ima kaimono ni it-te iru, Hanako -TOP now shopping to going is Hanako has gone shopping (and is not here now). b. Hanako wa sannen mae ni nihon ni it-te iru. Hanako-TOP three years before _ Japan to go is Hanako has been to Japan three years ago. (36) a. Dareka ga boku no nikki o yon-de iru. somebody-SUB I-POSS diary-OB] read is Somebody has read my diary (as there is some trace). b. Boku wa mae ni kono hon o yon-de iru. I -TOP before this book-OB] read is I have read this book before. 58 In (34) through (36) above (a) sentences indicate a resultant state while (b) sentences express an experienial meaning. The difference is that (a) sentences indicate some kind of observable situation while (b) sentences just express speaker's recollection. The Situations described by (34a) and (34b) may have the structures (37a) and (37b) respectively. (37) a. Tanaka is here Tanaka came here. © Tanaka has been here / The sentence (34a) is likely to be uttered in a context when Tanaka's arrival is Tanaka came here. important information. The diagram (37a) shows that the result of Tanaka's arrival, i.e., his existance, is relevant at the moment of speech. On the other hand, (34b) expresses Tanaka's experience of having come here before. We don't know whether or not he is still here at the moment of speechgThus there is no focus on the completion of the described event. In (b) sentences of (34) through (36) the time adverbials indicated in bold letters play a role for getting experiential interpretations. These adverbials set our viewpoint to the past, and. forces us to see the described events as not directly related to the situation at the present moment. These events are related to the present moment only as the speaker's recollection of a past event. 59 The role of time adverbials is Singificant when we have activities such as (36). Since nikki o yomu 'read a diary' is a durative situation, if it were not for the time expressions like mae ni 'before' that draw our viewpoint to the past, we would focus on the internal structure of these situations and typically associate progressive reading for these expressions as a default reading. In some cases, it is hard to distinguish the experiential from the resultative. See the following sentences. (38) Sono sakuhin wa moo yon-de iru. the piece of work-TOP already read is I have read the piece of work already. (39) Surnisu-san wa moo susi o tabe-te iru. Mr. Smith -TOP already susi—OB] eat is Mr. Smith has already eaten susi. In (38) and (39) it is difficult to decide whether the -te iru indicates a resultant state or experiential without context. In (38), for example, if the speaker's intention of uttering this sentence is to express that 3] he has finished reading the book (so that someone else can go ahead and read it), it may be taken as showing a resultant state. If this utterrance is made as a reply to a question like 'Have you ever read this book?', (38) is likely to be taken as showing an experience of the speaker. In the same way, (39) can mean a resultant state if it is utterred when somebody asked whether Mr. Smith has already had dinner or not, while it can also be a statement to express Mr. Smith's experience of having eaten sushi before. Thus, the interpretations of these sentences are heavily dependent on the context. In conclusion, experiential is a special case of resultant state, and its interpretation is fostered by other factors such as time adverbials or contextual information that interact with types of situations. The use of certain time 60 adverbials draws our attention to the past and let us view the incident in its totality rather than letting us focus on its completion and the result. However, if there is no crucial information found at a sentence level, we must resort to other information provided at a discourse level, such as what is focused, for what purpose the utterance is made. 5.4.2. Progressive and habitual meanings I propose that habitual reading is a special case of progressive in which regular occurrences of the same action] event are placed on the time line, and the collection of these separate incidents is perceived as a single durative situation. The following diagram clarifies this point. (40) Taroo wa maiasa kooen o hasit-te iru. 'Taroo is runs in a park every morning.’ maiasa kooen o hasitte iru kooen o hasiru © The key for the habitual interpretation is again the adverbial expressions which signal the regularity of the described event. Examine the following sentences. (41) a. Tanaka-san wa mainiti kohii o non-de iru. Mr. Tanaka-TOP every day coffee-OB] drink is Mr. Tanaka drinks coffee every day. b. Hanako wa saikin maishuu paatii ni it-te iru. Hanako—TOP recently every week party to go is Recently Hanako is going to the parties every week. c. Konogoro taitee siti-ji ni oki-te iru. these days mostly six o'clock at wake up is I have mostly been waking up at six o'clock these days. 61 In all the sentences in (41) the adverbial expressions indicated by bold letters broaden the time frame against which the described situation is set. Therefore, the relevancy of the described situation is assessed within this broader temporal frame which surrounds the present moment. The sequence of events repeated many times within this time frame may be perceived as a continued on-going action. Although there is no particular evidence for this analysis, the following examples might be helpful for bridging the gap between these two separate concepts (i.e., progressive and habitual). (42) a. Taroo wa kono ni nen kan zutto benkyoo si-te iru. Taroo-TOP this two year for continuously study is Taroo has been studying continuously for two years. b. Taroo wa kono ni nen kan zutto mainiti Taroo-TOP this two year for continuously every day benkyoo si-te iru. study is Taroo has been studying every day continuously for two years. It is difficult to determine whether (42a) expresses progressive or habitual. It might be the case that (42a) is more likely to be associated with progressive reading, but if we add the adverb maini ti 'every day' to this sentence as shown in (42b), the sentence is more likely to be associated with habitual reading. With the addition of mainiti 'every day' we are more sure that it is a repetition of the same action, not a single continuous action. However, even without this adverb it is clear that in reality what Taroo is engaged in is not an uninterrupted single action of studying. It is just perceived as such. Thus, sentences (42a) and (42b) basically express the same thing, except that habituality is emphasized by the use of the adverbial that signals the repetition of the described event in the latter case. In order to show the 62 validity of this analysis, however, some kind of quantitative research on the native speakers will be needed. 5.4.3. Iterative and successive meanings The iterative and successive meanings can also be perceived as a kind of progressive in the same sense as habitual meaning is. Both iterative and successive meanings are a collection of the same action or event repeated many times. When an action is repeated by a single individual, we have iterative meaning, and when different individuals are performing the same action, we get successive interpretation. See the following examples for both iterative and successive with the diagram for each. (43) Nando mo to o tatai-te iru ga daremo detekonai. many times door-OBJ knock is but nobody come out - NEG I have been knocking on the door many times but nobody comes out. ITERATIVE nando mo to o tatai-te iru to o tataku 'knock on the door' © (44) Gakusee ga tugitugi to kyoositu ni hait-te iru. student-SUB one after another classroom to enter is Students are entering the classroom one after another. SUCCESSIVE 63 In both (43) and (44) the use of adverbials indicated in bold letters foster the intended readings. Soga's (1983: 121) subcategorization of the meaning of continuity given below supports our View that iterative and successive are a kind of progressive. (45) Continuous Interrupted Uninterrupted /\ Repetitive Successive Continuous uninterrupted situations in the above diagram correspond to the simple progressive. 'Repetitive' in Soga's classification includes both habitual and iterative. According to Soga, all of them "share the semantic feature of continuation or on-going, regardless of whether they are interrupted or not" (1983: 121). This generalization by Soga will further support our View that ‘ habitual, iterative, and successive are classified or perceived as 'on-going' just like real progressive is. It is no wonder that the progressive meaning is strongly felt as prototypical among the meanings represented by -te iru. 5.4.4. Plurality of the subject and the successive reading In the previous section we observed that the adverbials such as tugitugi to 'one after another' strengthen the successive interpretation of the event] action described by -te iru. There are still other factors interacting with the successive reading. According to Teramura (1984), in order to obtain the interpretation which makes it possible for the collection of separate events as 64 being pereceived as a single continued situation, all we need is plural subjects. This observation seems to be true with the successive situation, though it does not necessarily hold with habituals and iteratives. Observe the following successive situations expressed by -te iru . (46) a. Ahurika de wa mainiti takusan no hito ga shin-dc Africa in -TOP every day many-POSS people-SUB die iru is In Africa many people are dying every day. b. Kono iti nen kan ni kinjo de akatyan ga san nin these one year for in neighborhood at baby - SUB three umare-te iru born is Within one year three babies were born in our neighborhood. c. Mankai no sakura kara hanabira ga in full bloom-POSS cherry blossom from petals - SUB harahara to tit-te iru. flutteringly fall is Petals are fluttering from the cherry blossom in full bloom. In all the sentences in (46), the situations described by -te iru , which usually indicate the resultant state, receive successive interpretation with the plural subject or the expression that implies the plurarity of the subject (in case of (46c)). A decisive factor for the successive interpretation actually is not purely the plurality of the subject of the described event by itself. The point becomes clear in the following sentences in which the event with the plural subject does not lead to successive interpretation. (47) a. Kyonen no daizisin de takusan no hito, ga sin-de last year-POSS earthquake at many-POSS people-SUB dead iru is Many people has been dead in the earthquake which happened last year. b. Sengetu Tanaka-san no uti ni san-nin no akatyan ga last month Mr. Tanaka-POSS home at three -POSS babies -SUB umare-te iru. born is Last month three babies have been born at Mr. Tanaka's house. c. Kinoo no ame de niwa no sakura no yesterday-POSS rain by garden-POSS cherry blossom-POSS hanabira ga minna tit-te iru. petals -SUB all fallen is All the petals of the cherry blossom in the garden have been fallen due to yesterday's rain. Despite the fact that the subject of -te i ru are all plural, the above sentence are usually taken not to express the successive aspect. (47a) and (47b) are most likely to be interpreted as recollection, and (47c) expresses a resultant state. We may infer that what is responsible for the successive reading for (46a) through (46c) was not only the plurarity of the subject itself, but probably also the adverbials that appeared in the sentence. In (46a) the adverb mai ni ti 'every day' signals that the described event did not happen just once. 1 ti nenkan 'for a duration of one year' in (46b) also implies that the three babies were not born at the same time, but at different times over the one year period. In addition, the place of the quantifier plays an important role in deciding whether three babies are a set or not. With the quantifier placed in front of the noun as in (47b), we take this as a unit, while if it is placed after the noun as in (46b) it is not necessarily the case. Moreover, kinzyo de 'in neighborhood' in (46b) as opposed to Tanaka-sun no uti de 'at Mr. Tanaka's home' (47b) also makes it difficult for the event to be taken as a one-time incident. Finally, in (46c) the adverbial harahara to 'flutteringly' forces the successive reading rather than the plurarity of the subject. In conclusion, the plurarity of the subject of -te iru may not necessarily lead to successive reading without the help of other factors interacting with it. As interacting factors being so diverse, it is hard to come to a generalization at 66 this moment. Further research is needed, which includes a variety of occurrences of -te iru with successive reading from natural discourse (both spoken and written) in order to identify what is actually responsible for a given interpretation of the form. 5.4.5. Agentivity and the interpretation of -te iru Another factor that affects the interpretation of the -te iru form is the agentivity of the described situation. As we already observed in section 5.1.2. agency or controllability of a described situation is a key factor for determining whether stative situations can be expressed in -te iru or not. In the following pairs we find that not only agentivity determines the occurrence of -te iru with states, but also affects the interpretation of the form with other situations (48) a. Mado ga ai-te iru. window-SUB open is The window is open. (i.e., It has been opened and is still open.) b. Mado o ake-te iru. window-OB] opening is He] S he is opening the window. (49) a. Roosoku no hi ga kie-te iru. candle -POSS light-SUB gone out is The candle light has been gone out. b. Roosoku no hi 0 kesi-te iru. candle -POSS light-OB] put out is He / She is putting out the candle light. (50) a. To ga koware-te iru. door-SUB broken is The door is broken. b. To 0 kowasi-te iru. door-SUB breaking is They are breaking the door. 67 In the above pairs (a) sentences indicate non-agentive situations with intransitive verbs while (b) sentences show agentive situations with transitive verbs. As is clear from the translations, all the instances of -te iru in (a) sentences are most likely to be understood as expressing resultant state, while those in (b) sentences are taken to express the progressive meaning. This is due to the fact that agentivity of situations is closely tied with their dynamicity. As was pointed out in Section 2.2.3., agentive or controllable situations which implies the volitional involvement of the subject with the described event makes it easier for us to conceive of its developmental process. In case of non-agentive situations it might be more difficult for us to imagine the dynamicity of the described event. We see, thereofere, that not only the punctuality or durativity of the described situation, but also the agentivity of the situation is an important factor to determine whether -te iru indicates a progressive or a resultative meaning. 5.5. Summary The core meaning of -te iru proposed in this study, namely, 'the heightened temporary relevance of a described situation to the reference time point', is supported by the fact that it can account for the occurrence of the form with fourth category verbs, which was unexplained in previous studies (Section 5.1.2.). This core has its motivation from the discourse function of ’the form observed by Kudo (1989) (Section 5.2.2.). Positing this core meaning helps us understand the occurrence of the form in various situations expressing different meanings (Section 5.3.). Two of these meanings, namely, progressive and resultative are more basic than the others (Section 5.3.4.). Crucial factors to determine these two basic meanings are the shared properties of the described situations: durativity and dynamicity of a described 68 event in case of progressive, and telicity in case of resultative. Other situational factors such as the viewpoint of the speaker are also crucial (Section 5.3.2. & Section 5.3.3.). All the other meanings (i.e., habitual, iterative, successive, and experiential) are derived from these basic meanings, and they are determined by contextual information such as the use of the adverbials, plurarity of the subject, and agentivity of the described situation (Section 5.4.1. - 5.4.5.). 6. Conclusion The present study shows that various meanings associated with the -te iru form which have long been recongized as progressive, resultative, habitual, iterative, successive, experiential are all encompassed by the core meaning] function of the form, namely, the heightened temporary relevance of a described situation to the reference point. One problem of previous studies is that they based their analyses of the form on the types of the verbs that occur with -te iru (Chapter 3). This problem is overcome by applying the Aristotelian classification of situations to Japanese (Chapter 4). Another problem in previous studies is that the occurrence of -te iru with fourth category verbs of Kindaichi's (1976) classification has not received proper treatment, despite their systematicity. The heightened temporary relevance of a described situation as a core meaning of the -te i ru form can embrace the use of this form with fourth category verbs as well as other states which were problematic in previous analyses (Section 5.2.1. 8: 5.3.1.). As was clarified in Section 5.3.4., two of the derived meanings, namely, progressive and resultative, turns out to be more typically associated with -te iru than the others, due to the shared properties among the types of situations (namely, durativity and dynamicity for the progressive reading and 69 telicity for the resultative reading). All the other meanings are derived from these two basic meanings, and are determined by situational factors such as the use of the adverbials, viewpoint setting of the speaker, the speakers purpose of uttering the sentence in a given context (Section 5.4.) Given that our definition of -te iru is based on its discourse function (Section 5.2.), it may be supported by the actual occurrences of the form in various environments in natural discourse. The present study did not aim at accounting for -te iru appearing in natural discourse in contrast with other aspectual forms, and was limited to the sentential level. However, if we posit that the main motivation for the existence of the aspectual categories is their discourse function as discussed in Hopper (1979, 1982), Li and Thompson (1982), Givon (1984), Thelin (1990), among others, the use of -te iru must also be clarified in discourse. Further research on the use of this aspectual form in natural discourse both written and spoken is needed in order to provide more convincing supports for the present hypothesis on the core meaning and the function of -te iru . BIBLIOGRAPHY BIBLIOGRAPHY Ando, S. 1986. Eigo no Ronri, Nihongo no Ronri - Taishoo-gengogaku-teki Kenkyuu (The Logic of Japanese Language and the Logic of English Language) Tokyo: Taishuukan. Bennett, M. 1981. "Of Tense and Aspect: One Analysis." Tedeschi and Zaenen. 1981.Syntax and Semantics vol.14 - Tense and Aspect. 13-29. New York: Academic Press. Binnick, R. 1991. Time and the Verb - A Guide to Tense and Aspect. Oxford: Oxford University Press. Chung, S. and Timberlake, A. 1985. "Tense, Aspect and Mood." Shopen. ed. Language Typology and Syntactic Description. vol. 3. Cambridge: Cambridge University Press. Comrie, B. 1976. Aspect. Cambridge: Cambridge University Press. Comrie, B. 1985. Tense. Cambridge: Cambridge University Press. Dahl, O. 1981. "On the Definition of the Telic-Atelic (Bounded-Nonbounded) Distiction." Tedeschi and Zaenen. 1981. Syntax and Semantics vol.14 - Tense and Aspect. 79-90. New York: Academic Press. Dahl, O. 1985. Tense and Aspect Systems. New York: Basil Blackwell. Dowty, D. 1979. Word Meaning and Montague Grammar. Holland: D. Reidel Publishing Company. Fujii, T. 1976. "'Doosi + te iru' no imi ( The Meaning of 'Verb + te iru' )." Kindaichi ed. Nihongo Doosi no Asupekuto (Aspect of Japanese Verbs ). 97- 116. Tokyo: Mugi Shoboo. Gareygfl. 1957. "Verbal Aspect in French." Language 33. 91-110. Givon, T. 1990. Syntax — A Functional Typological Introduction. vol.II. Amsterdam] Philadelphia: John Benjamins Publishing Company. Hopper, P. and Thompson, S. 1980. 'Transitivity Grammar and Discourse." Language 56. 251-99. Hopper, P. 1979. "Aspect and Foregrounding in Discourse" T. Givon. ed. Syntax and Semantics: Discourse and Syntacs. 213-241. New York: Academic Press. 70 71 ‘ Hopper, P. ed., 1982. Tense-Aspect: Between Semantics and Pragmatics, Containing the Contributions to a Symposium on Tense and Aspect, held at UCLA, May 1979. Amsterdam] Philadelphia: John Benjamins Publishing Company. Jespersen, O. 1965. The Philosophy of Grammar. New York. The Norton Library. Originally published in 1924. Johnson, M. 1981. "A Unified Temporal Theory of Tense and Aspect." Tedeschi and Zaenen. 1981. Syntax and Semantics vol.14 - Tense and Aspect. 145-75. New Yorlc Academic Press. Kenny, A. 1963. Action, Emotion, and Will. London: Routledge & Kegan Paul. Kindaichi, H. 1976. "Kokugo Doosi no Iti-Bunrui (A Proposal for Japanese Verb Classification). Kindaichi. ed. Nihongo Doosi no Asupekuto (Aspect of Japanese Verbs ). 7-26. Tokyo: Mugi Shoboo. Repr. - Originally published in Gengo Kenkyuu 15. (1950). 48-63. Kudo, M. 1989. "Gendai Nihongo no Paafekuto o Megutte (On the Perfect of Modern Japanese)" Gengo-gaku Kenkyuukai (Research Goup for Linguistics). ed. Kotoba no Kagaku (Science of language). Tokyo: Mugi Shoboo. Kunihiro, T. 1982. "Nihongo, Eigo.( Japanese and English )." Teramura. ed. Kooza Nihongo-gaku - Gaikokugo to no Taishoo 11 ( Lectures on Japanese Language - Contrastive Study ). 2-18. Tokyo: Meiji Shoin. Machida, K. 1989. Nihongo no Jisei to Asupekuto (Japanese Tense and Aspect ). Tokyo. Aruku. Majewicz, A. 1985. Grammatical Category of Aspect In Japanese and Polish in a Comparative Perspective. Poznan: Universytet im. Adama Mickiewicza. Mourelatos, A. 1981. "Events, Processes, and States." Tedeschi and Zaenen. 1981. Syntax and Semantics 001.14 - Tense and Aspect. 415-34. New York: Academic Press. Nakau, M. 1976. "Tense, Aspect, Modality." Shibatani, M. 1976. Syntax and Semantics vol.5 421-82. New York: Academic Press. Reichenbach, H. 1947. Elements of Symbolic Logic. New York: The Macmillan. Ryle, G. 1949. The Concept of Mind. London: Hutchinson. Scheffer, J. 1975. The Progressive in English. Amsterdam: North-Holland Publishing Company. Soga, M. 1983. Tense and Aspect in Modern Colloquial Japanese. Vancouver: University of British Columbia Press. 72 Suzuki, S. 1976. "Nihongo Doosi no Toki to Sugata (Tense and Aspect of Japanese Verbs)" Kindaichi. ed. Nihongo Doosi no Asupekuto ( Aspect of Japanese Verbs). 63-81. Tokyo: Mugi Shoboo. Teramura, H. 1984. Nihongo no Sintakuru to Imi (Japanese Syntax and Meanings). vol.ll. Tokyo: Kuroshio Shuppan. Timberlake, A. 1984. Reichenbach and Russian Aspect. Filier and Timberlake. ed. The Scope of Slavic Aspect. 153-67. Columbus, Ohio: Slavica Publishers, Inc. Thelin, N. 1990. Verbal Aspect In Discourse. Amsterdam: John Benjamins Publishing Company. Vendler, Z. 1967. "Verbs and Times." Linguistics and Philosophy. 97-121. Ithaca, N.Y.: Cornell University Press. Reprint of 1957. "Verbs and Times." Philosophical Review 66. 143-160. Yoshikawa, T. 1976. "Gendai Nihongo Doosi no Asupekuto Kenkyuu (The Study of Aspect of Modern Japanese)" Kindaichi. ed. Nihongo Doosi no Asupekuto Aspect of Japanese Verbs ). 155-327. Tokyo: Mugi Shoboo. "Illllillililillil'5