UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Intransitive verbs with transitive counterparts in Japanese : --with focus on ergative and middle-- Uchida, Miyo 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A8 U34.pdf [ 4.64MB ]
Metadata
JSON: 831-1.0100509.json
JSON-LD: 831-1.0100509-ld.json
RDF/XML (Pretty): 831-1.0100509-rdf.xml
RDF/JSON: 831-1.0100509-rdf.json
Turtle: 831-1.0100509-turtle.txt
N-Triples: 831-1.0100509-rdf-ntriples.txt
Original Record: 831-1.0100509-source.json
Full Text
831-1.0100509-fulltext.txt
Citation
831-1.0100509.ris

Full Text

INTRANSITIVE VERBS WITH TRANSITIVE COUNTERPARTS IN JAPANESE -with focus on ergative and middle-by M I Y O UCHIDA B.A. , The University of British Columbia, 1989 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF ARTS in THE F A C U L T Y OF G R A D U A T E STUDIES (The Department of Asian Studies) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A August 1991 © Miyo Uchida, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date ftuQ 2bf iq<f/ DE-6 (2/88) ii ABSTRACT In this thesis, I have examined intransitive verbs with transitive counterparts in Japanese. In accordance with the differences in their syntactic and semantic characteristics, I have considered that they are categorized into two, so-called ergative and middle verbs. There are various theories concerning "ergative" and "middle" verbs, especially "ergative" verbs since the term ergative is somewhat confusing. For instance, some linguists consider "ergative" verbs as underlyingly transitive verbs while others consider them as lexically intransitive verbs. I have hypothesized that "ergative" verbs are lexically intransitive and "middle" verbs are derived intransitive, and have attempted to support intransitivity or transitivity of those verbs from syntactic and semantic points of view, such as concerning passive possibility, im-plied agent, and so on. For example, "ergative" verbs seem to be passivized whereas "middle" verbs do not seem to be passivized since potentialization, which derives "middle" verbs, has a close relationship with passive. "Ergative" verbs do not necessarily have im-plied agent whereas "middle" verbs always seem to have implied agent and suggest underly-ing agent. My analysis seems to reveal the difference in transitivity between "ergative" and "middle" verbs and supports my hypothesis. iii T A B L E OF CONTENTS ABSTRACT ii LIST OF FIGURES v ACKNOWLEDGEMENTS vi CHAPTER 1: INTRODUCTION 1 CHAPTER 2: CLASSIFICATION OF VERBS 3 2.1. Various Theories 3 2.2. Syntactic Classifications of Verbs 7 CHAPTER 3: INTRANSITIVE VERSUS TRANSITIVE VERBS 20 3.1. Semantic Differences 20 3.2. Case Marking 21 3.3. Further Comments on Syntactic Differences 33 3.3.1. Phrase Structures 33 3.3.2. Passive Construction Possibility 34 3.4. "Intransitive-Transitive Pair" Verbs 40 3.5. "Intransitive-Transitive Pair" Verbs Stems 44 3.6. "Giving and Receiving Pair" Verbs 47 CHAPTER 4: ANALYSIS OF INTRANSITIVE VERBS--ERGATTVE AND MIDDLE 54 4.1. Various Theories on Ergative and Middle 54 4.2. "Ergative" Verbs versus "Middle" Verbs 73 4.2.1. Transitivity 73 4.2.2. Stativity 74 4.2.3. The [NP ni NP ga ] Construction 88 4.3. "Perceptual" Verbs 92 CHAPTER 5: ERGATIVE 104 CHAPTER 6: MIDDLE 110 6.1. Potentiality 110 6.2. Relationship between Potential and Passive 112 6.3. Relationship between "Middle" Verbs and Transitive Verbs 116 CHAPTER 7: CONCLUSION BIBLIOGRAPHY V LIST OF FIGURES CHAPTER 2: Figure 1. Classification of Intransitive Verbs 9 1.1. Volitional Verbs 9 1.2. Non-Volitional Verbs 10 1.2.1. Non-Stative Non-Volitional Verbs 10 Figure 2. Classification of Transitive Verbs 11 2.1. Volitional Verbs 11 2.2. Non-Volitional Verbs 11 Figure 3. Tree Diagrams of Verbal Classification 12 Figure 4. Volitionality of "Intransitive-Transitive Pair" Verbs 18 CHAPTER 3: Figure 1. Phrase Structures 33 1.1. Intransitive Sentence 33 1.2. Transitive Sentence 33 Figure 2. Four Types of Passives 35 Figure 3. Passive Transformations 35 Figure 4. Passive Construction Possibilities 36 CHAPTER 6: Figure 1. Transformation Which Derives "Middle" Verbs 121 vi ACKNOWLEDGEMENTS I would like to thank greatly my supervisor, Dr. Matsuo Soga. He inspired and encouraged me throughout the entire process of writing this thesis. His suggestions were always helpful. Without his support, patience, and time, I could not have completed the thesis in this final form. I would also like to thank my committee members, Dr. Kenichi Takashima and Dr. Bernard Saint-Jacques, for the completion of my thesis. Lastly, I would like to give my special thanks to Mr. Joel Murray for his assistance and encouragement along the way. 1 CHAPTER 1 INTRODUCTION In Japanese, there are intransitive verbs which have transitive counterparts. Al-though those intransitive verbs, on the surface, may be categorized into a single group, they show two distinguishing characteristics. For instance, some intransitive verbs with transi-tive counterparts seem to carry stativity while others do not. Therefore, I will categorize surface intransitive verbs with transitive counterparts into two: "ergative" and "middle." In order to analyze those intransitive verbs, I will first examine and classify Japanese verbs according to transitivity. Then I will focus on those intransitive verbs, which I will later call intransitive verbs of "intransitive-transitive pair" verbs. Prior to the analysis of intransitive verbs of "intransitive-transitive pair" verbs, I will define the terms "ergative" and "middle:" "ergative" verbs are considered to be lexical intransitive verbs and "middle" verbs are considered to be derived intransitive verbs. Having defined those terms, I will observe the differences between "ergative" and "middle" verbs from both syntactic and semantic points of view. Lastly, I will discuss the characteristics of "ergative" and "middle" verbs separately but along with other constructions. Chapter 2 is devoted to the classification of Japanese verbs. First, I will introduce various theories, which are mainly semantic verbal classifications. Then, I will classify Japanese verbs into intransitive and transitive verbs according to their syntactic characteris-tics. My main criteria are transitivity, volitionality, and stativity. Chapter 3 focuses on differences between intransitive and transitive verbs. In this chapter, I will introduce "intransitive-transitive pair" verbs. They are the verbs which have either intransitive or transitive counterparts. In Chapter 4, intransitive verbs of "intransitive-transitive pair" verbs are classified into two: "ergative" and "middle." Since there seems to be some confusion on the term "ergative," I will define the term "ergative" 2 along with the term "middle." In this chapter, I will first introduce various theories on "ergative" and "middle." Then, I will examine differences between "ergative" and "middle" verbs through passivization, stativity, and the [NP ni NP ga ] construction. Along with the discussion on "ergative" and "middle" verbs, so-called perceptual verbs such as kikoeru and mieru are also analyzed. In Chapters 5 and 6, characteristics of "ergative" and "middle" verbs are discussed. Chapter 5 is devoted to discussion on the properties of "ergative" verbs being compared to transitive verbs. Here, the relationship between transitivity and causality is discussed. Then I will attempt to hypothesize that both "ergative" verbs and transitive verbs are de-rived at the lexical level. Chapter 6 is devoted to discussion on the properties of "middle" verbs. They are considered to be derived intransitive verbs. They have underlying transi-tive verbs, and potentialization and intransitivization derive "middle" verbs. In this chapter, the close formal relationship between potential and passive is discussed in order to reason the ungrammaticality of passivized "middle" verbs. In this thesis, I will indicate Japanese sentences in their underlying forms. Since the various morphophonemic rules apply, their surface forms are different from their underly-ing forms. Also to facilitate reading, proper English names will be spelled in English, as in "John" not zyoon. 3 CHAPTER 2 CLASSIFICATION OF VERBS 2.1. Various Theories Classifying Japanese verbs often creates some disagreement among Japanese lin-guists since there has not been uniformity in the classification of verbs. Although linguists usually consider basic verbal criteria, such as stativity, volitionality, or punctuality, those criteria are sometimes hard to define both semantically and syntactically, and definitions are often with many exceptions. The criterion "stativity" is traditionally defined as a verbal characteristic which rep-resents a permanent or near-permanent characteristic of being rather than an action or event. Stative verbs are considered not to co-occur with the auxiliary verb te-iru. which de-scribes a current state or an on-going action or event. They are not to be used in the im-perative construction either, yet some verbs which are considered stative may violate the above stativity criteria. For example, wakaru (understand) is considered to be a stative verb. However, the te-iru form may be attached to wakaru. The sentence John wa wakar-te- iru (John understands) is a grammatical sentence. Another example is ma ((animate thing) exist). Ira is also considered to be a stative verb. However, it can be used in the imperative construction. The imperative sentence, soko ni iro (Be there!) is a grammatical sentence. The criterion "volitionality," which is sometimes used interchangeably with the term "agentivity," is considered to judge whether actions represented by the verbs are carried out with the subject's volition or not. For instance, the sentence ookii piano ga sono heya ni hair-ta (A big piano fitted into that room) is not volitional since "piano" cannot have volition. The sentence gakusei ga takusan sono heya ni hair-ta (Many students entered that room) can be either volitional or non-volitional. If the action of "entering" is considered to have been 4 carried out with "students'" volition, the sentence is volitional. However, if the action of "the students' entering the room" is considered to have been caused by some other force, the sentence is considered to be non-volitional. In conclusion, sentences with an animate subject are mostly volitional. However, sentences with some intransitive verbs such as hairu (enter), kaeru (return), and atumaru (gather) (all of these verbs happen to have transitive counterparts) may be considered volitional or non-volitional although they have an animate subject. Therefore, it is sometimes even confusing to judge whether sentences are voli-tional or non-volitional. Another confusing point is whether volitionality is judged on the lexical level or on the sentential level. For instance, stative verbs such as aru ((inanimate thing) exist), iru ((animate thing) exist), and iru (need) are considered to be non-volitional verbs since these verbs do not represent actions but states. Also, intransitive verbs such as aku (open), kawaku (dry), and moeru (burn) representing events should be considered non-volitional since they do not allow an animate subject. However, intransitive verbs like hairu (enter), kaeru (return), and atumaru (gather) cannot be judged on the lexical level whether they are volitional or non-volitional since these verbs allow both an animate subject and an inani-mate subject and since their volitionality depends on their subject. Therefore, volitionality of these verbs should be judged on the sentential level. In short, although the criterion "volitionality" may be easy to define, the basis for its judgement is sometimes unclear. The criterion "punctuality" is considered to refer to occurrence of time. Punctual verbs, as opposed to continual verbs, represent actions or events which occur instanta-neously. On the other hand, continual verbs represent actions or events which can occur in a durative manner. However, there are some continual-punctual verbs which can be con-sidered some times as punctual and other times as continual~for example, kuru (come) and iku (go), as in John wa roku-zi ni koko ni ku-ta (John came here at six o'clock) (punctual) and as in John wa kesiki o mi-nagara sono miti o ku-ta (while watching the view, John came along 5 the street) (continual). Therefore, although the terms "punctual" and "continual" seem to be opposite, they are not necessarily mutually exclusive. Another problem about the verbal classification is terminological. Different lin-guists, although they consider basic criteria, sometimes invent their own terminology and classify verbs accordingly. One of the well-known postwar Japanese linguists, Kindaichi (1955), in his work on Japanese verbal tense and aspect, classifies verbs into four groups: stative, continual, punc-tual, and the fourth group. His first three groups are within the basic verbal criteria men-tioned in the previous paragraph. However, as he states in his work, the fourth group is based on his own criterion which has never been used by other linguists. Its criterion seems to be similar to his first group in the sense that it is "timeless," but it has never been clari-fied and Kindaichi himself admits that it is open to discussion. Examples in his first group are aru, iru, a potential verb dekiru (can), etc. The te-iru form is considered not to co-occur with the stative verbs. Therefore, deki-te-iru to mean (being able) is ungrammatical. The second group consists of non-stative continual verbs such as yomu (read) and kaku (write) while the third group consists of non-stative punctual verbs such as sinu (die) and tuku (be lighted). The verbs in both the second group and the third groups, unlike stative verbs in the first group, describe an action or event. However, the action or event described by the verbs in the second group can occur for a duration of time whereas the action or event described by the verbs in the third group occurs instanta-neously. Moreover, those second and third group verbs can co-occur with the te-iru form. Non-stative continual verbs plus the te-iru form indicates a progressive action or event whereas non-stative punctual verbs plus the te-iru form refers to resultative state. Further examples in the second group are warau (laugh), oyogu (swim), aruku (walk), etc. Also, some verbs which indicate natural phenomena, such as huru ((rain or snow)fall), and huku ((wind)blow) are classified as the second group verbs. Further examples in the third group are tomaru ((something)stop), kekkon-suru (get married) and so forth. As I 6 mentioned previously, there are verbs which belong to both the second and third groups. Those verbs are iku (go), kuru (come), hairu (enter), deru (get out(of something)) and so forth. These overlapping characteristics of verbs make a uniform verbal classification diffi-cult. The fourth group of verbs indicate that something or someone takes on a certain state1. These verbs always appear with the te-iru form. Verbs in this group are not many. Examples are sobieru (soar), arihureru (be common), omodatu (be principal) and so forth. Makino and Tsutsui (1986) semantically classified Japanese verbs. Their classifica-tion, although basic verbal criteria are used, is somewhat different from Kindaichi's classi-fication. They classified verbs into eight categories according to verbal semantic character-istics. Kindaichi's fourth group verbs are not discussed in their classification. Rather, his stative, continual, and punctual verbs are semantically analyzed in depth. Makino and Tsutsui set up new verbal categories: movement verbs (their terminology), reciprocal verbs, and non-volitional-emotive verbs. Kindaichi classifies kuru (come), iku (go), hairu (enter), and deru (get out(of something)) as both continual and punctual verbs while Makino and Tsutsui classify them as punctual and movement verbs. Kekkon-suru (get mar-ried) is, in Kindaichi's classification, a punctual verb while in Makino and Tsutsui's classification, it is a punctual and reciprocal verb. Makino and Tsutsui's notion of non-volitional-emotive verbs such as yorokobu (be happy), kanasimu (grieve), sinobu (endure) and so forth are identified with another term: psychological verbs. There are arguments on these psychological verbs concerning whether they should be considered to be intransitive verbs or transitive verbs. I will discuss the transitivity of psychological verbs in Chapter 3. kindaichi claims that the fourth group is semantically similar to the first stative-verb group since it refers to the state of being rather than to an action or event. This fourth group is open to discussion. 7 2.2. Syntactic Classifications of Verbs Kindaichi and Makino and Tsutsui's classifications are based on semantic characteristics of verbs. It seems to be that such classification tends to become subjective since it does not depend upon "formal" criteria. I believe that any verb classification must be based upon syntactic criteria also if the classification is to be more objective. Therefore, in what follows, I will attempt to classify Japanese verbs both semantically and syntacti-cally, with the focus mainly on syntactic characteristics of verbs. By so doing, the verbs can be classified systematically and the lexical verbs which are original and the derived verbs which are intransitivized or transitivized can be distinguished more clearly2. First, I assume that all verbs belong to either intransitive or transitive groups. Therefore, in order to classify Japanese verbs syntactically, I will consider those two basic syntactic criteria along with some semantic basic criteria such as stativity, and punctuality which Kindaichi and Makino and Tsutsui use in their semantic classification. Moreover, I would like to treat the notion of volitionality, which is sometimes identified with agentivity3 and sometimes with the [self-controllable] feature4, as a crucial criterion for my syntactic classification. Thus, all Japanese verbs are considered to be either intransitive or transitive, first. Furthermore, I will consider that each group has two semantic sub-groups: volitional and non-volitional. As a result, four major syntactic-semantic groups appear: the volitional intransitive group, the non-volitional intransitive group, the volitional transitive group, and the non-volitional transitive group. All these four groups are further sub-categorized. The non-volitional intransitive group can be divided into stative and non-stative groups. The 2Derived intransitive verbs are derived from underlying transitive verbs through the transformation "intransitivization" and derived transitive verbs are derived from underlying intransitive verbs through the transformation "transitivization." 3Inoue (1990) assigns the theta role Agent to the volitional subject. Therefore, she uses the term "agentive" rather than the term "volitional" in her work. 4Kuno (1973) uses this feature. 8 other three groups are considered to be all non-stative. Furthermore, both the volitional and the non-volitional non-stative intransitive groups are subcategorized into two: verbs with transitive counterparts and verbs without transitive counterparts. The non-volitional non-stative verbs without transitive counterparts are further divided into three: "natural phenomenal" verbs, other unexpected event verbs, and others. The volitional non-stative transitive group is, like the non-stative intransitive groups, subcategorized into two: verbs with intransitive counterparts and verbs without intransitive counterparts. The non-voli-tional non-stative transitive group is also subcategorized into two: psychological verbs and other unexpected event verbs. Stative verbs, as has been mentioned previously, are verbs which represent states, such as "existing" and "needing." They are all non-volitional intransitive. Traditionally mieru (be visible) and kikoeru (be audible) are said to be stative verbs. However, it is ob-served that despite their stativity, they are somewhat different from other stative verbs. I will argue in Chapter 4 that they are derived stative verbs which have gone through two transformations, namely the potential transformation, which adds stativity to their underly-ing non-stative verbs, and the intransitivization transformation, which transitivizes their underlying transitive verbs to intransitive verbs. In other words, mieru and ki1cop.rn are considered to be underlyingly non-stative transitive verbs. "Natural phenomenal" verbs are verbs which represent natural phenomena, such as "raining," "snowing," and "wind blowing." These verbs are non-volitional intransitive. Other unexpected event verbs are verbs such as "meeting someone unexpectedly" and "losing consciousness." They may be either intransitive or transitive, but they are all non-volitional, since the events represented by those verbs are unexpected. Other non-voli-tional non-stative intransitive verbs without transitive counterparts, such as iku (go), kuru (come), hasiru (run), and so forth, are formally identical to volitional counterparts. This difference in volitionality is due to the [animate] or [inanimate] feature of a noun phrase in the subject position. For instance, iku (go) with an animate subject is considered to be voli-9 tional whereas iku with an inanimate subject, such as basu (bus), kuruma (car), and so forth, is considered to be non-volitional. Despite Makino and Tsutsui's claim, which is that psychological verbs are intransi-tive, I consider psychological verbs to be transitive verbs since these verbs behave in the same way as other transitive verbs. The transitivity of psychological verbs will be discussed in Chapter 3. Here are the classifications and examples: Figure 1. Classification of Intransitive Verbs I N T R A N S I T I V E Volitional Non-volitional Non-Stative Stative Non-Stative Figure 1.1. Volitional Verbs I N T R A N S I T I V E Volitional Non-Stative Verbs with transitive counterparts  Verbs without transitive counterparts  10 Figure 1.2. Non-Volitional Verbs INTRANSrnVE Non-Volitional Stative Non-Stative Verbs with transitive counterparts Verbs with-out transitive counterparts Figure 1.2.1. Non-Stative Non-Volitional Verbs INTRANSITIVE Non-Volitional Non-Stative Verbs without transitive counterparts "Natural phenomenal" verbs Other unex-pected event verbs Others 11 Figure 2. Classification of Transitive Verbs TRANSITIVE Volitional Non-volitional Non-Stative Non-Stative Figure 2 .1. Volitional Verbs TRANSITIVE Volitional Non-Stative Verbs with intransitive counterparts  Verbs without intransitive counterparts  Figure 2.2. Non-Volitional Verbs TRANSITIVE Non-Volitional Non-Stative Psychological verbs Other unexpected event verbs The above classification may be more clearly represented in the diagram on the following page. 12 Figure 3. Tree Diagrams of Verbal Classification Intransitive Volitional Non-Stative Non-Volitional Stative Non-Stative Verbs with Vt Verbs without Vt Verbs with Vt Verbs without Vt "Natural Phenomenal" Verbs Others Other unexpected event verbs Transitive Volitional Non-Stative Verbs with Vi Verbs without Vi Non-Volitional Non-Stative Psychological Verbs Other unexpected event verbs Vt: transitive verbs Vi: intransitive verbs 13 (1). Volitional intransitive A. Non-Stative a. Verbs with transitive counterparts agaru "(someone) go up" atumaru "(people) gather" kaeru "(someone) return" korogaru "(someone) roll" taoreru "collapse" tooru "(someone) pass through" b. Verbs without transitive counterparts aruku "walk" hasiru "run" iku "go" kuru "come" oyogu "swim" (2). Non-volitional intransitive A. Stative am "(inanimate thing) exist" iru "(animate thing) exist" iru "need" wakaru "understand" 14 B. Non-Stative a. Verbs with transitive counterparts agaru "(something) go up" atumaru "(things) gather" kaeru "(something) return" korogaru "(something) roll" taoreru "(something) fall down" tooru "(something) pass through' aku "(something) open" hazimaru "(something) start" kawaku "(something) dry" moeru "(something) bum" nam "(something) ring" okoru "(something) happen" b. Verbs without transitive counterparts • "Natural ohenomenal" verbs huku "(wind) blow" hum "(rain, snow) fall" • Other unexoected event verbs deau "come across" dekuwasu "come across" humeru "tremble" souguu-suru "encounter" 15 • Others aruku "walk" hasiru "run" iku "go" kuru "come" oyogu "swim" (3). Volitional transitive A. Non-Stative a. Verbs with intransitive counterparts ageru "raise" atumeru "gather" kaesu "return" korogeru "roll" taosu "knock down" toosu "pass (something or someone) through" akeru "open" hazimeru "start" kawakasu "dry" moyasu "burn" narasu "ring" okosu "cause" 16 b. Verbs without intransitive counterparts kaku "write" korosu "kill" nomu "drink" taberu "eat" tataku "hit" yomu "read" (4). Non-volitional transitive A. Non-Stative a. Psychological verbs itamu "mourn" kanasimu "grieve" sinobu "endure" tanosimu "enjoy" yorokobu "be happy" b. Other unexpected event verbs nakusu "lose" ushinau "lose" Notice that some intransitive verbs, such as agaru (go up), atumaru (gather), kaeru (return), iku (go), and kuru (come) can be volitional or non-volitional. If intransitive verbs have an animate subject, they are considered to be volitional. The animate subject is considered to be an instigator of the action or event represented by the verbs. If intransitive verbs have 17 an inanimate subject, they are considered to be non-volitional. Naturally, the inanimate subject is not considered to be an instigator of the action represented by the verbs. There are some cases that intransitive verbs are considered to be non-volitional even though they have an animate subject. In those cases, the animate subject is considered to have no voli-tion with respect to the stated action or event. Observe (5): (5)a. John wa yuki ga hur-ta node basu de kaer-ta. Top snow Nom fall-Past because bus Abl return-past "Because it snowed, John returned (home) by bus." (volitional) b. Horyo kookan no tame konsuizyootai no horyo POW exchange Gen for coma Gen POW made mo kyuukyuusya de kaer-ta. even also ambulance Abl return-Past "Because of the prisoners of war (POW) exchange, even POW's in coma also returned by ambulance." (non-volitional) Sentence (5a) is volitional. John is considered to have volition to do the action of "returning home." However, sentence (5b) should be considered to be non-volitional although it has an animate subject. The fact that the soldiers returned is outside of their volition. There-fore, the verb in sentence (5b) should be considered non-volitional. Some intransitive verbs with transitive counterparts, such as aku (open), hajimaro (start), and kawaku (dry) are always non-volitional and always have an inanimate subject. 18 I refer to verbs which have counterparts of opposite transitivity as "intransitive-transitive pair" verbs. Since clear understanding of the relation between intransitive and transitive verbs in "intransitive-transitive pair" verbs is crucial to the understanding of my thesis, I will include another figure representing only "intransitive-transitive pair" verbs and their relation. B in the following diagram refers to the abstract verb base (Chew's terminol-ogy), R to the intransitive morpheme, and S to the transitive morpheme. Intransitive verbs are represented as B+R+ru and transitive verbs as B+S+ru (intransitive and transitive stems will be discussed in chapter 3). Figure 4. Volitionality of "Intransitive-Transitive Pair" Verbs INTRANSITIVE TRANSITIVE Volitional Non-volitional Volitional Non-volitional B + R + ru B+R + ru B + S+ru B + R + ru B + S+ru 19 Observe (6): (6) "Intransitive-transitive pair" verbs INTRANSITIVE TRANSITIVE Volitional Non-volitional Volitional Non-volitional kaeru — kaeru kaesu "(someone) re- "(something) re- "return" turn" turn" moeru moyasu "(something) "burn' burn" I will discuss characteristics of "intransitive-transitive pair" verbs in depth in Chap-ter 3. 20 CHAPTER 3 INTRANSITIVE VERSUS TRANSITIVE VERBS 3.1. Semantic Differences The major semantic differences between intransitive and transitive verbs are the di-rection of an action or event represented by the verbs and the number of participants re-quired1. The action or event represented by intransitive verbs is inwardly oriented whereas the action or event represented by transitive verbs is outwardly oriented. Kuru (come), iku (go), aruku (walk), hasiru (run), and oyogu (swim) are the examples of intransitive verbs. For instance, examine sentence (1): (1). John ga hasir-ta. Nom run-Past "John ran." In sentence (1), the action of running is restricted within the agent, John, who is an initiator of the action, rather than going beyond the agent and affecting the outside world. It is also observed that there is only one participant in this sentence. Taberu (eat), kaku (write), tataku (hit), korosu (kill), etc. are the examples of transi-tive verbs. Observe sentence (2): 1Jacobsen (1982) examines the number of participants required in intransitive verbs and transitive verbs. 21 (2). John wa Mary o koros-ta. Nom Acc kill-Past "John killed Mary." In sentence (2), it is observed that there are two participants, and the action initiated by the agent John goes outward and affects the other participant. As (2) shows, transitive verbs require at least two participants. 3.2. Case Marking Since Japanese uses case forms quite extensively, I will first attempt to clarify the term "case," which is somewhat confusing. In fact, there is a great deal of confusion in the term "case." Sometimes the term "case" is used to refer to the "traditional" case system adopted from the case system of Latin. These cases are nominative, vocative, accusative, genitive, dative, and ablative. Other times, the term "case" is used to refer to the case system which expresses the seman-tic relation of nominals to the verb. These cases are, for example, agent, experiencer, in-strument, patient, and so forth. More confusingly, these two different definitions of the case system are frequently used without being properly defined. Evidently, these two dis-tinguished case systems are not mutually exclusive. In fact, the case system which expresses the semantic relation between a noun phrase and a verb may be assigned to the case system adopted from Latin. For example, a noun phrase to which the nominative case is assigned may be assigned the case of agent, experiencer, or something else. In this thesis, I will use the term "case marker" to refer to the "traditional" case as well as the semantic case. Also, I will make some additions to the normally used cases. For example, I will consider "topic" as a case. I will not attempt to find a unified correspon-22 dence between the traditional and semantic cases for each case marker. Nor will I attempt to find one-to-one correspondence between a case and a case marker. For example, the nominative marker ga may be associated with the case of agent or with the case of experi-encer. Also, a case particle (or a seemingly one and the same particle) may be used for more than one case. Observe the following: (3)a. John ga Mary o tatak-ta. Nom Acc hit-Past Agent "John hit Mary." b. John ga Mary no si o kanasim-ta Nom Gen death Acc grieve-Past Experiencer "John grieved at Mary's death." Notice that in (3), John is nominative in both (a) and (b) but the particle ga indicates agen-tive for (3a) and experiencer for (3b). Another example is one of the many ablative case markers, de. De may be associated with the case of location or with the case of instrument. Observe the following: 23 (4) a. John wa kooen de ringo o tabe-ta. Top park Abl apple Acc eat-Past Location "John ate an apple at the park." b. John wa hasami de Top scissors Abl Instrument "John cut the paper with scissors." Verbal transitivity can be examined through the number of participants since the number of participants differs between intransitive verbs and transitive verbs. Intransitive verbs have one participant which is marked in the nominative case. Transitive verbs have two or three participants: the leftmost noun phrase is in the nominative case, the noun phrase immediately followed by a verb is in the accusative case, and the third one (if any) is in the dative case. Therefore, the basic structures of intransitive and transitive sentences can be represented as in (5): (5) a. intransitive verb kami o kir-ta. paper Acc cut-Past rNP ga L Nom 1 J b. transitive verb TNP ga NP o 1 L Nom Acc J [NP ga NP ni NP o L Nom Dat Acc Examples are: (6)a. intransitive verb John ga ku-ta. Nom come-Past "John came." b. transitive verb John ga keeki o tabe-ta. Nom cake Acc eat-Past "John ate the cake." 25 c. John ga Mary ni hana o okur-ta. Nom Dat flower Acc send-Past tt 'John sent flowers to Mary. it Both intransitive verbs and transitive verbs have a unique characteristic to sub-categorize the number of noun phrases. [NP ga] is associated with intransitive verbs and marker. O is the accusative marker. M is the dative marker. It is considered to be at the surface structure level where grammatical notions such as subject, direct object, and indirect object are most easily observable. Typically, the nominative case marker marks subject, the accusative case marker marks direct object, and the dative case marker marks indirect object2. On the surface, it may appear that Japanese, unlike English, does not have a one-to-one correspondence between accusative case marking and the grammatical notion of the direct object. It may appear that the accusative case marker o marks not only the goal to which the actions represented by the verbs are directed but also something else such as space covering by the motion of verbs, which is often translated as "along," "through," and so forth, and origin of location of the motion of verbs, which is often translated as "from." Observe the following: 2Kuno (1973) claims that the particle ga can mark object. For example ga in John wa eigo ga wakaru (John understands English) marks object. Therefore, eigo is the object of the verb wakaru. He argues that eigo in the given sentence does not show the grammatical characteristics of a subject as we see in honorifics, for example. Thus, he claims that it is the object. However, he does not give a convincing argument that it is the object. It even seems that his analysis is semantically influenced by English. In English, it is true that English is the object of the verb understand. However, in Japanese, eigo can be the subject if the verb wakaru is interpreted as "be understandable" instead of "understand." Although Kuno claims that ga can mark object, his claim cannot be considered as a counter-example for the general understanding that the accusative case marker o marks direct object. [NPga(NPni)NPo ] is associated with transitive verbs. Ga is the nominative 26 (7)a. John ga sonohito o tatak-ta. Nom that person Acc hit-Past "John hit that person." b. John ga sonomichi o ik-ta. Nom that person Abl go-Past "John went along that street." c. John ga sonotonneru o toorinuke-ta. Nom that tunnel Abl pass through-Past "John passed through that tunnel." d. John ga sono eki o de-ta. Nom that station Abl leave-Past "John left that station." Sentences in (7) all appear to have the same construction: [NP ga NP o Verb]. Sentence (7a) has the particle o which is the accusative case marker and which also marks the direct object. The sentences (7b), (7c), and (7d) also have the particle o. However, the particle o in these sentences does not indicate the goal to which actions represented by the verbs are directed. Rather, it indicates space along, through, and from, which the motions represented by verbs are semantically associated with. What this particle indicates, in fact, is similar to the so-called case of location which is indicated by one of the ablative case 2 7 markers3. Therefore, the particle o in those sentences should be considered as indicating the ablative case rather than the accusative case. In fact, the verbs in them are all intransitive verbs, which are believed not to take the direct object. By assigning the ablative case marker to the particle o in sentences (7b), (7c), and (7d), I will be able to distinguish between intransitive verbs and transitive verbs from the point of view of case marking, especially accusative case marking. Intransitive verbs can be defined as verbs without the accusative case marker whereas transitive verbs can be defined as verbs with the accusative case marker. There is an argument that psychological verbs such as yorokobu (be happy) and kanasimu (grieve) should be considered as intransitive verbs, rather than transitive verbs4, which take the ablative case marker rather than the accusative case marker for their "direct object" (I will come back in the following pages to the argument which claims those verbs to be intransitive). I consider that psychological verbs are transitive verbs since those verbs behave syntactically in the same way as other transitive verbs. I consider the particle o in the sentences of psychological verbs as the accusative case marker like in other transitive sentences. By considering the particle o preceding psychological verbs as the accusative case marker and by considering the particle o in intransitive sentences as the ablative case marker, the definition of intransitive and transitive verbs from the point of view of case marking may hold without exceptions. Observe the following intransitive and transitive sentences with the particle o. The ablative case marker o and the accusative case marker o happen to be homophonous. (8a) is an example of an intransitive verb in which the particle 2 marks the ablative case indicating the location of origin; (8b) is an example of a transitive 3As I have mentioned, in Japanese there are many particles marked in the ablative case. Examples are de to indicate location or means (instrument), Q to indicate space which the motion verbs take place, and so forth. Notice that the ablative case marker o and the accusative case marker o happen to be homophonous. 4See Makino and Tsutsui. They consider psychological verbs intransitive. 28 verb in which o marks the accusative case; (8c) is an example of a psychological verb in which o again marks the accusative case as in (8b). (8)a. John wa ie o de-ta. Top house Abl leave-Past (intransitive) Location "John left home." b. John wa ie o kaw-ta. Top house Acc buy-Past (transitive) Patient "John bought a house." c. John wa Mary no seikoo o yorokob-ta. Top Gen success Acc be happy-Past Human-Emotion (Transitive: psychological verb) "John was happy for Mary's success." In the previous paragraph, I simply stated that psychological verbs are transitive verbs. Contrary to my observations, Makino and Tsutsui (1986) claim that psychological verbs are intransitive verbs since those verbs represent inwardly-oriented events, a typical semantic characteristic of intransitive verbs. Furthermore, they claim that the particle o 29 preceding psychological verbs is not the "ordinary" direct object (their term) and that the particle o indicates the cause of human emotions. Makino and Tsutsui's claim has many flaws. First, in their claim, they do not seem to have a clear distinction between the grammatical notion, "ordinary" direct object, and the semantic notion, the cause of human emotions. Rather, they use those two different types of notions as if they were the same notion in order to argue for the intransitivity of psychological verbs. Second, Makino and Tsutsui's term "ordinary" direct object is confusing. They fail to specify the differences between "ordinary" direct object and "non-ordinary" direct object preceding psychological verbs. I believe that psychological verbs should be considered to have a direct object which is marked in the accusative case. Semantically, this accusative case marker is associated with the cause of human emotions while the accusative case marker in other transitive verbs is often associated with the case of patient. Third, Makino and Tsutsui attempt to classify psychological verbs only on a semantic basis. In fact, as they claim, there is an obvious semantic difference between psychological verbs and other transitive verbs; psychological verbs may represent, in their view, "inwardly-oriented" event whereas other transitive verbs, "outwardly-oriented" event. However, I consider that this semantic difference may still be weak since X o yorokobu does not necessarily have to be "inwardly" oriented. That is, someone's joy may be "directed toward X," so to speak. It is true that psychological verbs have the case of experiencer associated with the nominative case marker, whereas other transitive verbs have the case of agent associated with the nominative case marker. Furthermore, for the former the human emotion is associated with the accusative case whereas for the latter, the patient is associated with the accusative case. Yet, for both, it is evident that the accusative case remains, and so the verbs are transitive. Let us consider syntactic argument, against Makino and Tsutsui's claim. Although psychological verbs and other transitive verbs seem to be different semantically, psychological verbs behave syntactically almost in the same way as other transitive verbs. 30 For instance, psychological verbs, like other transitive verbs, may be passivized with the result of a simple passive instead of an affective passive5. Observe (9) and (10): (9)a. John ga hon o kak-ta. Nom book Acc write-Past "John wrote the book." b. Hon ga John ni-yotte6 kak-rare-ta. Nom Abl Pass-Past "The book was written by John." (10)a. John ga Mary no si o taihen kanasim-ta. Nom Gen death Acc much grieve-Past "John grieved much at Mary's death." 5There are two types of passive in Japanese. One is called "simple" passive (or "pure" passive) and the other is called "affective" passive (or "adversity" passive). The "pure" passive is similar to the English passive and the "affective" passive expresses that the speaker or the person indicated in the subject position is affected by the action represented by the verb. Passivized intransitive verbs are all "affective" passives. 6Ni(-yotte) consists of the ablative case ni, yor-u (be due to), and the tei form. Ni(-yotte) appears in the simple passive construction indicating the initiator of the actions or events. It is translated as "by" in English. In this thesis, the meaning of "by" is indicated in the simple passive by ni(-votte) rather than by the underlying form ni(-yor-te). 31 b. Mary no si ga John ni-yotte taihen kanasim-rare-ta. Nom by Pass-Past "Mary's death was much grieved at by John." Notice that in (10) with a psychological verb, there is the active-passive correspondence just like the one in (9) with a typical transitive verb. Let me consider another syntactic argument which might be useful for transitive-intransitive distinction. In general, the auxiliary verb te-aru. which indicates that something has been done, may be attached to most of the transitive verbs. Since psychological verbs, as the term suggests, represent human emotions rather than some action's being done, psychological verbs, unlike other transitive verbs, do not allow the attachment of the auxiliary te-aru form. The fact that the te-aru form cannot be attached to psychological verbs is simply due to the semantic characteristics of psychological verbs and does not disprove the transitivity of psychological verbs. Observe (11): (ll)a. Tegami ga kak-te-aru. letter Nom write have been done-Non-Past "The letter has been written." b. *Mary no si ga kanasim-te-aru. Gen death Nom grieve have been done-Non-Past Lit: "Mary's death has been grieved at." 32 In fact, there are other transitive verbs which do not occur in the te-aru construction. For example, sira (know), wara(w)u (laugh), okasu (commit (a crime)), and so forth. Observe the following ungrammatical sentences: (12)a. *Sore ga sir-te-aru. that Nom know-have been done-Non-Past Lit: "That has been known." b. *Kare no matigai ga waraw-te-aru. he Gen mistake Nom laugh Lit: "His mistake has been laughed at." c. *Tumi ga okas-te-aru. crime Nom commit Lit: "The crime has been committed." The above examples indicate that te-aru is not a very firm criterion for transitivity after all. From the syntactic evidence above, I must conclude that psychological verbs are, against Makino and Tsutsui's claim, transitive verbs. All intransitive verbs seem to have the NP pattern [NP ga] whereas all transitive verbs seem to have the NP pattern [NP ga (NP ni) NP o]. Ga always marks the nominative case, o the accusative or ablative case, and m (if any) the dative case. Intransitive verbs do not have the accusative case marker whereas transitive verbs have the accusative case marker. In conclusion, in Japanese, intransitive and transitive verbs may be distinguished from the point of view of case marking with the understanding that there is no necessary surface one-to-one correspondence between a marker and a case. 3.3. Further Comments on Syntactic Differences 3.3.1 Phrase Structures Intransitive and transitive verbs have distinct structures. Observe the following basic tree structures of intransitive and transitive sentences: Figure 1. Phrase Structures Figure 1.1. Intransitive Sentence S NPga VP V Figure 1.2. Transitive Sentence S NPga VP (NPni) NPo V 34 The basic structure of intransitive sentences involve a single noun phrase and the basic structure of transitive sentences involve two or three noun phrases. Transitive verbs, such as taberu (eat), tataku (hit), korosu (kill), and so forth, which do not take the indirect object, involve two noun phrases. Transitive verbs, such as watasu (hand), osieru (teach), and so forth, which take the indirect object involve three noun phrases. A major syntactic difference between intransitive and transitive verbs is clearly observed in the phrase structures indicated above. 3.3.2. Passive Construction Possibility As mentioned previously, a crucial difference between intransitive and transitive verbs may be observed in the passive construction possibility. In Japanese, there are two basic kinds of passive: the simple passive and the affective passive. The simple passive is similar to the passive English. The affective passive, which implies that some action or event takes place and the third person (mostly the speaker) is affected by the action or event. Each of these two passives is further sub-categorized in two groups. The simple passive is sub-categorized into the direct-object passive and the indirect-object passive. The affective passive is sub-categorized into the indirect-object passive and the "third-person" passive. The direct-object passive is the passive whose subject is derived from the underlying direct object. The indirect-object passive is the passive whose subject is derived from the underlying indirect object. The "third person" passive is the passive whose subject is the speaker himself or the third person. For example, the sentence ame ga hur-ta (it rained) may be passivized into the affect "third-person" passive with the speaker himself as an affectee, ame ni hur-rare-ta (I got rained on), or with the third person as an affectee, John wa ame ni hur-rare-ta (John got rained on). 35 Figure 2 represents four types of passives and Figure 3 represents passive transformations. Figure 2. Four Types of Passives SIMPLE AFFECTIVE Direct-Object Indirect-Object Indirect-Object "Third-Person" Figure 3. Passive Transformations NPga (NPni) (NPo) V 1 2 3 4 Simple Direct-Object => 3 ga 2 1 ni(-yotte) 4 -(r)are Indirect-Object => 2 ga 3 1 ni(-yotte) 4 -(r)are Affective Indirect-Object => 2 ga 3 1 ni 4 -(r)are "Third-Person" =>lni 2 3 4-(r)are The passive auxiliary verb is -(r)are-. The underlying subject in the affective passives is always indicated by ni (the ablative case) whereas the underlying subject in the simple passives is indicated by either ni or ni-yotte. [NP ga NP ni NP o ] may be passivized into all four as in (15). Transitive verbs which involve the phrase structure [NP ga NP o ] may be passivized into two simple direct-object passive and affective "third-person" passive as in (14). Since they do not have indirect object, they lack both simple and affective indirect-object passives. 36 Unlike English intransitive verbs, Japanese intransitive verbs can be passivized. However, passivization of intransitive verbs results only in the affective "third-person" passive. Observe Figure 4, which represents passive construction possibilities of intransitive and transitive verbs, and following examples: Figure 4. Passive Construction Possibilities INTRANSmVE TRANSITIVE [NPgaNPo 1 NPgaNPniNPo 1 • Affective "Third-person" • Simple direct-object • Affective direct-object • Simple direct-object • Simple indirect-object • Affective indirect-object • Affective "third-person" (13)a. John ga Mary no uti ni ku-ta Nom Gen house Dat come-Past "John came to Mary's house." b. John ni *(-yotte) Mary no uti ni ku-rare-ta. Abl Gen Dat Pass-Past "(Someone) was affected by the fact that John came to Mary's house." (Affective "third-person" passive) 37 (14)a. John ga Mary no tegami o yom-ta. Nom Gen letter Acc read-Past "John read Mary's letter." b. Mary no tegami ga John ni(-yotte) yom-rare-ta. Nom Abl Pass-Past "Mary's letter was read by John." (Simple direct-object passive) c. John ni ?(-yotte) Mary no tegami o yom-rare-ta. Abl Acc Pass "(Someone) was affected by the fact that Mary's letter was read by John." (Affective "third-person" passive) (15)a. John ga Mary ni tegami o watas-ta. Nom Dat letter Acc hand-Past "John handed the letter to Mary." 38 b. Tegami ga Mary ni John ni(-yotte) watas-rare-ta. Nom Dat Abl hand-Pass "The letter was handed to Mary by John." (Simple direct-object passive) c. Mary ga tegami o John ni(-yotte) watas-rare-ta. Nom Abl hand-Pass "Mary was handed the letter by John." (Simple indirect-object passive) d. Mary ga tegami o John ni?(-yotte) watas-rare-ta. Nom Abl Pass "Mary was affected by the fact that John handed her the letter. (Affective indirect-object passive) e. John ni?(-yotte) Mary ni tegami o watas-rare-ta. Abl Pass "(Someone) was affected by the fact that John handed the letter to Mary." (Affective "third-person" passive) Notice that the above examples show that intransitive verbs result in only affective passive and transitive verbs result in both simple and affective passives. Furthermore, these 39 examples show that the ni-yotte never appears in affective passives of the intransitive verbs although the affective passives of the transitive verb might take ni-yotte among some native speakers. The te-aru form, which I discussed in the previous pages, can occur with transitive verbs excluding psychological and several other verbs but not with an intransitive verb. Transitive verbs with the te-aru form represent the perfective meaning of an action which has been done. When the te-aru form is attached to transitive verbs, the subject-object or nominative-accusative shift occurs as in (16), (16)a. John wa tegami o kak-ta. Top letter Acc write-Past "John wrote a letter." b. Tegami ga kak-te-aru. Nom has been done-Non-Past "The letter has been written." Since intransitive verbs do not occur with an accusative, the subject-object shift cannot take place. Hence, the structure is impossible for intransitive. Clearly, then, passive may be considered useful in Japanese to distinguish intransitive verbs from transitive verbs, and the structure -te aru may also be considered useful to some extent since no intransitive verbs can occur in that structure. 40 3.4. "Intransitive-Transitive Pair" Verbs So far, I have discussed differences between intransitive and transitive verbs. In Japanese, there are, in fact, numerous verbs which appear in an intransitive-transitive pair. As I mentioned in chapter 2, some intransitive verbs of "intransitive-transitive pair" verbs are considered as both volitional and non-volitional whereas some are considered only as non-volitional verbs. Transitive counterparts of all intransitive verbs in "intransitive-transitive pair" verbs are all considered as volitional. Observe the following volitional and non-volitional intransitive verbs of "intransitive-transitive pair" verbs and their transitive counterparts. (17)a. "Intransitive-transitive pair" verbs INTRANSITIVE Volit ional Non-volitional agaru "(someone) go up" agaru "(something) go up" atumaru "(people) gather atumaru "(things) gather" kaeru "(someone) return" kaeru "(something) return" korogaru "(someone) roll" korogaru "(something) roll" taoreru "collapse" taoreru "(something) fall down" tooru "(someone) pass through" tooru "(something) pass through" aku "(something) open" hazimaru "(something) start" kawaku "(something) dry" 41 moeru "(something) bum" nam "(something) ring" okom "(something) happen b. "Intransitive-transitive pair" verbs TRANSITIVE Volitional Non-volitional agem "raise" atumeru "gather" kaesu "return" korogasu "roll" taosu "knock down" toosu "pass (something or someone) through" akeru "open" hazimeru "start" kawakasu "dry" moyasu "bum" narasu "ring" okosu "cause" Transitive verbs of "intransitive-transitive pair" verbs such as kaesu (return (something)), moyasu (burn (something)), and okosu (cause), if compared to transitive 42 verbs which do not have intransitive counterparts such as nomu (drink), kaku (write), and tataku (hit), seem to involve causality. For example, causality in kaesu can be seen easily if kaesu is interpreted as in (18b). Sase is the causative auxiliary verb. (18) a. John wa hon o kaes-ta. Top book Acc return-Past "John returned the book." b. John wa [hon ga kaer-] sase-ta. Top book Nom return cause-Past Transitive verbs without intransitive counterparts cannot be paraphrased as in (18b) with the sase form since they do not have causality. It seems that transitive verbs of "intransitive-transitive pair" verbs are "causative" forms of their intransitive counterparts. Causality of transitive verbs will be further discusses in Chapter 5. Interestingly, sometimes stativity seems to be observed in sentences with non-volitional intransitive verbs of "intransitive-transitive pair" verbs. Observe (19) and (20): (19) a. Sono huku wa kawak-ta. that clothes Top dry-Past "Those clothes dried." 43 b. Sono huku wa sugu kawak-ru. soon dry-Non-Past "Those clothes dry soon." (20)a. Sono huku wa kawak-te-iru. "Those clothes are dry." b. *Sono huku wa sugu kawak-te-iru. Sentences in (19) are all non-volitional. Sentence (19a) represents the event of "clothes' drying" whereas sentence (19b) seems to represent quality or a permanent characteristic of "clothes" which are "drying soon." In this sense, sentence (19b) seems to be a stative sentence. In fact, as the sentences in (20) show, the te-iru form can be attached to sentence (19a). However, the te-iru form cannot be attached to sentence (19b) since sentence (19b) is already stative. Stativity is not observed in volitional intransitive sentences since volitional intransitive sentences represent actions or events. In fact, stativity and volitionality are mutually exclusive terms. I will discuss that there are non-volitional intransitive verbs which are derived form transitive forms and which appear on the surface to be in the same form as intransitive verbs of "intransitive-transitive pair" verbs. I consider that those derived intransitive verbs are syntactically derived through transformations, namely, potential and intransitivization. When the potential transformation applies to verbs, the stativity, which is carried in the potential auxiliary verb -(rar)e-. is added to the verbs. As a result, derived intransitive 44 sentence show stativity. I will refer to derived intransitive verbs as "middle" verbs and to non-volitional intransitive verbs of "intransitive-transitive pair" verbs as "ergative" verbs. "Middle" and "ergative" verbs, which is the major focus of this thesis, will be discussed in Chapters 4,5, and 6. 3.5. "Intransitive-Transitive Pair" Verb Stems Inoue (1990) points out that verbs have either an intransitive, a transitive, or a neutral root. The intransitive suffix r{el is attached to either a transitive or a neutral root, and the transitive suffix s£e) is attached to either an intransitive or a neutral root. "Intransitive-transitive pair" verbs, according to her theory, share the same root. The root could be a transitive, intransitive, or neutral root. Observe (21): (21)a. transitive root + r[e) intransitive ex. tor- "take off (something)" (transitive) tor-re- "(something) come off" (intransitive) b. intransitive root + s(e) transitive ex. ak- "(something) open" (intransitive) ak-se- "open (something)" (i.e. ake on the surface) (transitive) c. neutral root + r(e} => intransitive neutral root + s(e) => transitive too- "passing through" (neutral) too-r- "(something or someone) pass through" (intransitive) too-s- "pass (something or someone) though" (transitive) 45 Different from Inoue's claim, I argue that all "intransitive-transitive pair" verbs are composed of the verbal stem, which is similar to Inoue's neutral root, plus either the intransitive suffix r and the transitive suffix s attached to the base verbal stem rather than Inoue's proposition of verbs having fixed intransitive or transitive roots. (22) base stem + r => intransitive base stem + s => transitive Those intransitive and transitive morphemes have allomorphs7. (23) is the representation of intransitive and transitive allomorphs. The capital letters R and S represent intransitive and transitive morphemes, and the small letters represent their allomorphs: (23)a. intransitive allomorphs R: r, re, ri, ar, er b. transitive allomorphs S: s, se, as, os "Intransitive-transitive pair" verb stems can be represented in the following manner: (24) kae + R => kae-ru "(someone) return" (intransitive) "(something return)" kae + S => kaes-ru "return" (transitive) 7I am indebted to Chew (1973). korog + R => korogar-ru "(someone) roll" "(something) roll" korog + S => korogas-ru "roll" (intransitive) (transitive) tao + R => taore-ru "collapse" (intransitive) "fall down" tao + S => taos-ru "knock down" (transitive) ak + R ak-ru "(something) open" (intransitive) ak + S => ake-ru "open" (transitive) nar + R => nar-ru "(something) ring" (intransitive) nar + S => naras-ru "ring" (transitive) ok + R => oki-ru "(something) happen" (intransitive) ok + S okos-ru "cause" (transitive) Phonological changes occur as Chew (1973) points out in his phonological work: ...where the first phoneme of the basic or underlying form of the suffix is of the same type as the last phoneme of the verb stem, the first phoneme of the suffix is deleted according to the formula (2a) V iV 2 V i 8Chew (1973), pp. 26-27. VjV2 => does not seem to be observed in the process of intransitive and transitive suf f ixiation. It is not clear where he can apply the rule, except perhaps in the te-iru construction where i in iru may be deleted: mat-te-iru => matte-ru (be waiting). However, the deletion of i in the te-iru form is optional and does not occur in formal writing. What I am interested in here is the second rule of consonants: C1C2 => C\. 47 Observe (25): (25). ak + se "open" => ake-ok + ri "rise" => oki-nig + re "escape => nige-C4C2 => C\ can be observed in (25). 3.6. "Giving and Receiving Pair" Verbs Giving and receiving verbs such as ageru (give) and morau (receive), which are both transitive verbs, exhibit the so-called active-passive deictic difference. In general, there are two voices, which represent viewpoints for the actions or events to be looked at. One is referred to as the active voice and represents actions or events from the instigator's point of view. The other is referred to as the passive voice and represents actions or events from the affectee's point of view. Thus, sentences may be represented in either the active or the passive voice. The difference between active and passive sentences is deictic. The following sentences in (26) show the deictic difference. (26)a. John ga Mary o tatak-ta. Nom Acc hit-Past "John hit Mary." 48 b. Mary ga John ni(-yotte) tatak-rare-ta. Nom Abl Pass-Past "Mary was hit by John." Sentence (26a) is active and represents the action "hitting" from John's point of view. On the other hand, sentence (26b) is passive and represents the action from Mary's point of view. The deictic difference which is observed between active and passive sentences seems to be observed between giving and receiving verbs as well. Observe (27): (27)a. John ga Mary ni kane o age-ta. Nom Dat money Acc give-Past "John gave Mary money." b. Mary ga John ni kane o moraw-ta. Nom Abl Acc receive-Past "Mary received money from John." Sentence (27a) involves the giving verb ageru and represents the action of giving from John's point of view with John as a giver. Sentence (27b) involves the receiving verb morau and represents the action from Mary's point of view with Mary as a receiver. Notice that the particle ni is marked by the dative case in (27a) and the ablative case in (27b). Ni in (27a) indicates the goal of the action and is never replaced with kara (from) whereas ni in 49 (27b) indicates the source of the action and can be replaced with another ablative case, kara (from). In fact, in Japanese there are some other kinds of giving and receiving pairs of verbs. I will refer to these verbs as "giving-receiving pair" verbs. These verbs, such as osieru (teach) and osowaru (be taught), azukeru (deposit) and azukaru (keep/be deposited), sazukeru (give) and sazukaru (be given), and so forth, exhibit the same type of deictic difference observed between active and passive sentences. Observe (28): (28)a. John ga Mary ni eigo o osie-ta. Nom Dat Acc teach-Past "John taught Mary English." b. Mary ga John ni eigo o osowar-ta. Nom Abl Acc be taught-Past "Mary was taught English by John." Sentence (28a) involves the "giving" verb osieru and represents the action of teaching from John's point of view with John as the giver. Sentence (28b) involves the "receiving" verb osowaru and represents the action from Mary's point of view with Mary as the receiver. As shown in (28), "giving-receiving pair" verbs, like the pair of ageru and morau. take the accusative case indicating direct object. Furthermore, the "giving" verb as in (28a) takes the ablative case ni which can be replaced with kara (from). Both giving and receiving verbs, such as ageru and morau. may be passivized into both simple and affective passives. However, most "receiving" verbs of "giving-receiving pair" verbs, such as osowaru (be taught), azukaru (keep/be deposited), sazakaru (be given), 50 and so forth, cannot be passivized although their "giving" counterparts such as osieru (teach), azukeru (deposit), sazukeru (give), and so forth, may be passivized into both simple and affective passives. The fact that most "receiving" verbs cannot be passivized may lead to the assumption that "receiving" verbs underlyingly carry passive quality. In fact, sentence of "receiving" verbs exhibit strong similarities to affective indirect object passive sentences. Observe (29): (29)a. Mary ga John ni hon o watas-rare-ta. Nom Abl book Acc hand-Pass-Past "Mary was affected by the fact that John handed the book to her." b. Mary ga John ni eigo o osowar-ta. Nom Abl Acc be taught-Past "Mary was taught English by John." c. *Eigo ga osowar-ta. Nom be taught-Past Lit: "English was taught." Sentence (29a) is an affective indirect object passive sentence and sentence (29b) is a sentence of a "receiving" verb. In both sentences (29a) and (29b), the particle ni is marked by the ablative case indicating the initiator of the actions. Sentence (29c) implies that although a sentence of a "receiving" verb is like the passive, it is not like the simple passive. In fact, (29c), which is similar to the simple passive construction, is ungrammatical. 51 In conclusion, sentences of "receiving" verbs of "giving-receiving pair" verbs are both semantically and syntactically similar to the passive. They are syntactically and semantically quite similar to the affective indirect object passive. Thus, I consider that "receiving" verbs are derived from their "giving" counterparts via affective passivization. Observe the following: (30). sazuk-e-ru "give" + -rare- "Pass" U sazuk-e-rare-ru II sazuk-ar-ru "be given" (31)a. John ga syoohin o sazuke-rare-ta. Nom prize Acc give-Pass-Past "John was given a prize." (affective passive) John ga syoohin o sazukar-ta. be given-Past b. John ga eigo o osie-rare-ta. Nom English Acc teach-Pass-Past "John was taught English." II (affective passive) 52 John ga eigo o osowar-ta. be taught-Past (30) indicates that a passivized "giving" verb sazuk-e-ru (give) is lexicalized into a "receiving" verb sazuk-ar-ru (be given). The first sentences in both (31a) and (31b) are affective passive with "giving" verbs, sazuke-ru (give) for (31a) and osie-ru (teach) for (31b). (31) again shows that the passivized "giving" verbs are lexicalized into "receiving" verbs sazukar-ru (be given) and osowar-ru (be taught). Notice that sentences in (31) do not indicate any change in the affective passive construction but lexicalization of the passivized "giving" verbs. I will now include the list of "giving" verbs and "receiving" verbs to illustrate the relationship between those. (32). "giving" verbs receiving" verbs azuk + S => azuke- azuke-rare- (azukar-) 'deposit' 'keep/be deposited' osie + S => osie- osie-rare- (osowar-) 'teach' 'be taught' sazuk + S => sazuke- sazuke-rare- (sazukar-) give 'be given' tamaw + S => tamaw-"give" tamaw-rare- (tamawar-) "be given" 53 As shown in (32), "giving" verbs involve the transitive morpheme S. "Receiving" verbs underlyingly have the structure of "giving" verbs plus the passive morpheme rare and are recognized as in the parentheses. In this chapter, the focus has been placed on the differences between intransitive and transitive verbs. In the following chapter, I will discuss mainly surface intransitive verbs which are lexical and derived. 54 CHAPTER 4 ANALYSIS OF DmiANSITIVE VERBS - ERGATIVE AND MIDDLE -4.1. Various Theories on Ergative and Middle There have been various theories proposed on "ergative" and "middle". The term ergative, as Fagan (1988) mentions in her work, is confusing. Here, I will mention the theories proposed by Dixson, Keyser and Roeper, Fagan, Inoue, and Miyagawa. Then, I will define the terms "ergative" and "middle" for the purpose of this thesis. Dixson (1979), in his ergative work, studies the Australian language Dyirbal and refers to the term ergative associated with a type of grammar. The "ergative" grammar, according to his theory, marks transitive subject (A) by the marked ergative case,while it marks both intransitive subject (S) and transitive direct object (O) by the unmarked absolutive case. Thus, S and O, differently from A, are treated in the same way. A language which has the "ergative" grammar is called an "ergative" language. As opposed to the "ergative" grammar, there is another type of grammar called the "accusative" grammar. The "accusative" grammar, unlike the "ergative" grammar, marks A and S by the unmarked nominative case and marks O by the marked accusative case. Thus, S and A are identified in the "accusative" grammar. A language which has the "accusative" grammar is called an "accusative" language. Observe the following "ergative" and "accusative" grammar patterns. (1) S V (intransitive) A O V (transitive) 55 t! ergative" grammar s=o (unmarked absolutive case) A (marked ergative case) II accusative" grammar S=A (unmarked nominative case) O (marked accusative case) Dixson also points out that "ergativity" can be observed at two grammatical levels which are case-marking at the morphological level (mentioned in the previous paragraph) and various syntactic constructions at the shallow structure1. Keyser and Roeper (1984) use the term "ergative" in a different sense. They studied ergative and middle in English and referred to the terms ergative and middle as intransitive verbs of "intransitive-transitive pair" verbs. "Ergative" verbs and "middle" verbs appear in pairs with transitive counterparts. Observe the following pairs: 1Dixson claims that languages do not have ergative/accusative distinction in deep structure. The distinction is made at two grammatical levels, shallow structure and surface morphology. The ergative case-marking appears in the morphological level. Various ergative syntactic operations such as coordination and subordination are conducted at the shallow structure level. The syntactic level, according to Dixson, is divided into three levels: deep structure, shallow structure, and surface structure. Passive, antipassive (antipassive, which changes the transitive AOV structure to the intransitive SV structure), and reflexive derive shallow structure, and coordination and subordination derive surface structure: deep structure => shallow structure => surface structure. (2)a. John broke the toy. b. The toy broke. 56 (3)a. Someone broke the toy. b. The toy breaks easily. The pair in (2) is referred to as an "ergative" pair whereas the pair in (3) is referred to as a "middle" pair. Both sentences (2a) and (3a) involve the transitive verb break and both sentences (2b) and (3b) involve its intransitive counterpart break. However, break in sentence (2b) is considered to be an "ergative" verb and break in sentence (3b) is considered to be a "middle" verb. In general, "middle" verbs are like stative verbs and represent state of being while "ergative" verbs represent an action or event. Thus, the "ergative" verb break in (2b) represents the "breaking" action and the "middle" verb break in (3b) represents the generic quality of the "toy." Furthermore, Keyser and Roeper claim that "middle" verbs, unlike "ergative" verbs, can occur neither in imperative nor in progressive constructions since they are stative. Keyser and Roeper also point out that although "ergative" verbs and "middle" verbs seem to be derived from the same underlying transitive structure, "ergative" verbs are lexically derived and "middle" verbs are syntactically derived. In other words, "ergative" verbs exist in lexicon whereas "middle" verbs do not. Fagan (1988) elaborates on Keyser and Roeper's work, focussing mainly on the English middle. Her terms ergative and middle coincide with those of Keyser and Roeper. However, against their theory, she argues that both "ergative" and "middle" verbs are lexically derived. She agrees with Keyser and Roeper that the crucial contrast between "ergative" and "middle" verbs is between events and states. However, she considers that that contrast is due to the difference in the semantic nature between "ergative" and "middle" verbs. 57 Inoue (1990)2 studies ergative and middle in Japanese using Keyser and Roeper's framework. Compare English and Japanese sentences in (4) and (5)3: (4) a. The boy over there broke my toy. (agentive transitive) b. My toy broke. (ergative) c. My toy breaks easily. (middle) (5) a. Asokoni iru otokonoko ga boku no omotya over there exist boy Nom I Gen toy o kowas-ta. Acc break-Past "The boy over there broke my toy." (agentive transitive) 2I believe that her work has opened up the area in which the Japanese verb system may be examined from the viewpoint of "ergativity" which brings many other related problems such as causativization, passivization, potentialization, etc. into discussion. In this sense, Inoue's work is truly significant. (The present thesis has been inspired by her paper). 3Inoue (1990), p. 2, example sentence in (2) and (3). 58 b. Boku no omotya ga koware-ta. Nom break-Past tt" 'My toy broke. it (ergative) c. Boku no omotya wa kantan ni koware-ru. Top easily break-Non-Past ir 'My toy breaks easily. (middle) Sentences (4b) and (5b) are "ergative" sentences with "ergative" verbs break/koware-ru. Sentences (4c) and (5c) are "middle" sentences with "middle" verbs break/koware-ru. Inoue examines both ergative and middle using the semantic feature called the theta role, which is similar to the case I have mentioned previously, such as the theta role Agent or Experiencer or Instrument, and so forth. She observes mostly the theta role of the subject and concludes that neither "ergative" nor "middle" sentences have the theta role Agent. Both "ergative" and "middle" sentences, as in (4b), (4c), (5b), and (5c), are non-agentive4. However, Inoue claims that "middle" sentences retain the agentivity of the underlying transitive subject, the so-called implied agentivity whereas "ergative" sentences lack the implication of agentivity completely. Observe the following sentences5: 4Inoue's term "agentive" is the same as the term "volitional" I have mentioned previously. 5Inoue (1990), p. 16, example sentences in (26). 59 (6)a. Kono to ga ak-ta. this door Nom open-Past "This door opened." b. Kono to ga kantan ni ak-ru. easily open-Non-Past "This door opens easily. Sentence (6a) is an "ergative" sentence and sentences (6b) is a "middle" sentence. Both sen-tences are non-agentive. However, Inoue claims that the "middle" sentence (6b), unlike the "ergative" sentence (6a), has the implied agent and may be translated as "someone can open this door easily." Notice that the implied agent is always indefinite and is translated as "one," "someone," "people," and so forth. Different from Inoue's view, I consider that "ergative" sentences, like "middle" sentences, also have the implication of agentivity and have the cause-effect relationship with their transitive-sentence counterparts. I will discuss the implied agent in "ergative" and "middle" sentences later in this chapter. Furthermore, Inoue points out that "middle" verbs are derived by suffixiation of the potential morpheme re. Since the potential morpheme re is stative, verbs to which re is attached receive its stativity in the process of "middle" verb formation. Thus, "middle" verbs are considered to be stative. "Ergative" verbs, on the other hand, do not involve the potential morpheme, but the intransitive morpheme. Since the intransitive morpheme, unlike the potential morpheme, is not stative, "ergative" verbs are considered to be non-stative. She also mentions that in Japanese, the potential morpheme and the intransitive morpheme, both of which are re, happen to homonymous. Therefore, on the surface level, 60 "ergative" verbs and "middle" verbs appear to be identical. Observe the following Inoue examples6: (7)a. Hanako ni pai ga yak-re-ru. Dat pie Nom bake-Pot-Non-Past "Hanako can bake pie." (potential/middle) b. Pai ga yak-re-ta. Nom bake-Intr-Past "Some pie was baked." (ergative) Yak-re-ru in sentence (7a) involves the potential morpheme re whereas yak-re-ru in sentence (7b) involves the intransitive morpheme re. Sentence (7a) is interpreted as both potential and middle whereas sentence (7b) is interpreted as ergative. Inoue's claim that "middle" verbs are derived by suffixiation of the potential morpheme might lead to a wrong assumption that all the verbs with the potential morpheme such as kak+re-ru (kak+pot re-ru), tabe+re-ru (tabe+pot re-ru), and so forth, are "middle" verbs. Like Keyser and Roeper, Inoue states that "ergative" and "middle" verbs are intransitive. However, different from Inoue, I consider that "middle" verbs are derived intransitive verbs which have gone through the potential transformation before the intransitivization. Thus, on the surface, "ergative" verbs and "middle" verbs, both of which 6Ibid., p. 22, example sentences (38b) and (39a). 61 are intransitive, appear to be identical, not because the intransitive morpheme re for "ergative" and the potential morpheme re for "middle," as Inoue claims, but because of the different derivations, which I will subsequently discuss. Inoue claims that there is a zero potential suffix. Observe (8)7: (8)a. Kono torakku ni piano ga sandai nor-ta. this truck on Nom three get on-Past "Three pianos got on this truck." (ergative) b. Kono torakku ni piano ga san dai nor-ru. this truck on Nom three get on-Non-Past "One can put three pianos on this truck." (middle) According to Inoue's analysis, nor- (get on) has an intransitive root. Nor- in sentence (8a) is an "ergative" verbs which is intransitive whereas nor- in sentence (8b) is a "middle" verb which has a zero potential suffix. Since the middle verb nor- in sentence (8b) has a zero potential suffix, no more suffixiation is possible. It is true that the attachment of the potential morpheme re to sentence (8b) results in the ungrammatical sentence *kono torakku ni piano ga san-dai nor-re-ru (Lit: One can put three pianos on this truck). Also observe the fact that the attachment of the potential morpheme re to sentence (8a) results in the ungrammatical sentence *kono torakku ni piano ga san-dai nor-re-ta (Lit: Three pianos 7Ibid., p. 23, example sentence (24a). 62 could get on this truck). According to Inoue's analysis, nor- (get on) in sentence (8a), unlike the one in sentence (8b), should be able to permit suffixiation of the potential morpheme since it does not have any potential morpheme present. However, contrary to her analysis, it does not permit the suffixiation. Here, I propose that on the surface level, nor- in both (8a) and (8b) are non-volitional intransitive and suffixiation of the potential morpheme to non-volitional intransitive verbs results in ungrammaticality. I will come back to this issue later. As I have mentioned so far, Inoue considers the potential morpheme as re. However, I believe that rare instead of re should be considered as the potential morpheme. As a result, the potential morpheme rare is "formally" quite similar to the passive morpheme rare. In fact, the potential and the passive exhibit many similarities both semantically and syntactically. I will discuss the relationship between the potential and the passive in Chapter 6 along with the discussion on the middle. Miyagawa (1980) has also dealt with "ergativity" in Japanese. His concept of "ergativity" is quite different from that of Inoue's. It seems that his is only another "term" for stative verbs. Yet, since he uses the term, let us briefly examine his concept of "ergativity." Miyagawa defines "ergative" verbs as those which take two or three noun phrases and have the so-called ergative case array; the nominative case marker ga instead of the accusative case marker o is assigned to the rightmost noun phrase. He points out that the case marking pattern must be included in the lexical entry of verbs. Thus, "ergative" verbs and transitive verbs differ in the case marking pattern in the lexical level. Observe the following Miyagawa's patterns: 63 (9)a. [NP(ni) NPga Dat Nom (ergative) b. [NPga NPo Nom Acc (transitive) (9a) is the "ergative" sentence pattern and (9b) is the transitive sentence pattern. Sometimes, "ergative" verbs leave the first noun phrase unmarked rather than assigning the dative case. The unmarked noun phrase is later assigned the nominative case by the redundancy rule8. His examples for "ergative" verbs are iru (need) and aru (have). Observe the following examples9: 8Miyagawa (1980) suggests a "case redundancy rule:" if there is one unmarked noun phrase, the rule assigns the nominative case ga to it; if there are two noun phrases, the rule assigns ga to the leftmost noun phrase and the accusative case o to the rightmost noun phrase; and if there are more than two unmarked noun phrases, the rule assigns ga to the leftmost noun phrase, o to the rightmost noun phrase, and the dative case to any other noun phrase. Here is his rule: Case Redundancy Rule [NP ]=*[NPga ] [NP NP ] => [NP ga NP o ] [NP .. . NP ... NP ] [NP ga ... NP ni... NP o ] 9Miyagawa (1980), p. 74, example sentences (31a) and (32a). 64 (10)a. John ni/ga kane ga ir-ru. Dat/Nom Nom need-Non-Past "John needs money." b. John ni/ga kodomo ga ar-ru. Dat/Nom child Nom have-Non-Past "John has a child." The nominative case marker ga for the first noun phrase both in sentences (10a) and (10b) is assigned by the redundancy rule. According to Miyagawa's analysis, "ergative" verbs can be represented in the following way with the "ergative" feature [ +ergative] and "ergative" case marking pattern. Observe (ll) 1 0: (11). T ir-ru / ar-ru ] I I I +verb I I + ergative I I [NPNPga ] I L [NPniNPga ] J Miyagawa, furthermore, claims that the potential morpheme rare is [ +ergative] and the attachment of the potential morpheme to verbs results in the "ergative" potential verbs. 10Ibid., example (65). 65 According to his analysis, so-called potential verbs are lexically derived verbs11 and may be represented in the following way in lexicon12: (12). T [X] 1 T [X]rare 1 I +verb I I +verb I I -ergative I I + ergative I L [(NP)* ] J L [ (NP) n ___] J [(NP)n ] in (12) indicates that the attachment of the potential morpheme (rar)e does not affect the number of noun phrases in the sentences13. Contrary to Miyagawa's claim on the ergative case array, I consider that the particle ni in his "ergative" sentences is in the ablative case instead of in the dative case since it functions differently from the dative case marker ni assigned to the indirect object. Semantically the dative case ni can be replaced by ni taisite. but the ablative case ni cannot be replaced by ni taisite but can be replaced by ni totte. Observe (13): n A s Inoue (1990) mentions in her ergative and middle work, in Japanese it is hard to draw a clear line between lexical and syntactic derivations. Some linguists consider the so-called potential verbs as syntactically derived verbs while others consider them as lexically derived verbs. 12Miyagawa (1980), p. 76, example (69). 13Sometimes, the attachment of some suffixes change the number of noun phrases. For instance, Miyagawa (1980) mentions that the attachment of the causative morpheme increases the number of noun phrases by one: F [X] 1 f [X]sase I +verb | => | +verb I -ergative I I -ergative L [(NP)n ] J L [(NP)n + l ] For example, the causative sentence of John ga ringo o tabe-ta (John ate an apple) would be Mary wa John ni  ringo o tabe-sase-ta (Mary made John eat an apple). The underlying sentence has two noun phrases and its causative sentence has three. The number of noun phrases increased in the causative sentence. 66 (13) a. John ga Mary ni / ni taisite / *ni totte. uso o iw-ta. Nom Dat lie Acc say-Past "John told Mary a he." b. John ni / *ni taisite / ni totte. sonnani ookii kuruma wa Abl such big car Top ir-nai. need-Neg-Non-Past "John does not need such a big car." Sentence (13a) has the dative case marker ni whereas sentence (13b) has the ablative case marker ni. Also, observe (14) and (15): (14) a. John ga Mary ni hana o okur-ta. Nom Dat flower Acc send-Past "John sent flowers to Mary." b. John ni sigoto ga ir-ta. Abl job Nom need-Past "John needed a job." 67 c. John ni san-nin kodomo ga ar-ru. Abl people child Nom exist-Non-Past "John has three children." d. John ni eigo ga hanas-rare-ru. Abl English Nom speak-Pot-Non-Past "John can speak English." (15)a. *John ga Mary ga nana o okur-ta. Nom Nom b. John ga sigoto ga ir-ta. Nom Nom c. John ga san-nin kodomo ga ar-ru. Nom Nom d. John ga eigo ga hanas-rare-ru. Nom Nom Sentences in (15) are intended to be equivalent to sentences in (14), except for the fact that the particle ni in (15) is replaced with the nominative case marker ga. Notice that sentences (14b), (14c), and (14d) may replace the particle ni with the nominative case marker ga and permit the multiple-subject construction, as sentences (15b), (15c), and (15d) show. However, sentence (14a) does not allow the particle ni, which indicates indirect object, to 68 be replaced by the particle ga as in (15a). Thus, contrary to Miyagawa's claim, I suggest that the particle ni in sentences (15b), (15c), and (15d) (Miyagawa's "ergative" sentences) is in the ablative case which is different from ni for indicating the dative case. The particle ni in the dative case and the particle ni in the ablative case happen to be homophonous. Although Miyagawa considers "potential" verbs as "ergative" verbs, "potential" verbs seem to be quite different from his other "ergative" verbs ira (need) and aru (have). Ira and am always have the rightmost noun phrase in the nominative case whereas "potential" verbs have the rightmost noun phrase sometimes in the nominative case but sometimes in the accusative case. This irregularity that "potential" verbs have in the case marker of the rightmost noun phrase might be due to the fact that they are underlyingly transitive verbs. Observe (16) and (17): (16)a. John ga eigo o hanas-ru. Nom English Acc speak-Non-Past "John speaks English." b. John ga eigo o hanas-rare-ru. Nom Acc Pot-Non-Past "John can speak English." c. John ni eigo ga hanas-rare-ru. Abl Nom Pot-Non-Past "John can speak English." 69 (17)a. John ni kane ir-ru. Abl money Nom need-Non-Past t i 'John needs money. t t b. *John ga kane o ir-ru. Nom Acc Non-Past Lit: "John needs money. II Sentence (16b) and (16c) have a "potential" verbs, and sentence (16a) which is the underlying sentence of (16b) and (16c) has a transitive verb. Notice that the rightmost noun is in the accusative in sentence (16b) and is in the nominative case in sentence (16c). Both sentences are grammatical. However, sentence (17b) which has the rightmost noun phrase in the accusative is ungrammatical. I believe that although "potential" verbs and other "ergative" verbs, iru (need) and aru (have), happen to have the same case marking pattern (ergative case array in Miyagawa's terminology), those verbs should not be grouped together as "ergative" verbs because "potential" verbs, and iru and aru, are both syntactically and semantically quite different from one another. In conclusion, although there has been no consensus on the definition of the term "ergative," Dixson, Inoue, and Miyagawa seem to employ some basic common concepts of the meaning of "ergativity" in their works. Dixson's ergative case marking identifies transitive direct object (O) and intransitive subject (S). Inoue's ergative pair identifies O and S. Also, Miyagawa's potential verbs seem to identify O and S although his "ergative" verbs ara (have) and iru (need) do not seem to identify O and S. Observe (18): 70 (18)a. Dixson's ergative case marking A O V (transitive) S V (intransitive) o=s b. Inoue's ergative pair [Otoko no ko ga] [omotyao] [kowas-ta] (transitive) A O V "The boy broke the toy." [Omotya ga] [koware-ta] (ergative) S V "The toy broke." o=s 71 c. Miyagawa's potential verbs14 [John ga] [eigo o] [hanas-ru] (transitive) A O V "John speaks English." [John ni] [eigo ga] [hanas-rare-ru] (potential) S Pot V "John can speak English." o=s Observing Dixson, Inoue, and Miyagawa's analyses in (18), I may safely conclude that the common concept concerning the meaning of "ergativity" between analysis of those three linguists is the identification of transitive direct object and intransitive subject. I have so far mentioned various theories on ergative and middle (mostly ergative) proposed by different linguists and identified some common grounds. Before I proceed with my discussion on ergative and middle, I must now clearly define the terms ergative and middle. Like Inoue, I will employ Keyser and Roeper's framework with some modifications in my thesis. I will refer to the terms ergative and middle as intransitive verbs of 14Miyagawa does not necessarily claim that potential verbs and transitive verbs are related. Rather, he simply states that potential verbs assign the nominative case marker ga to the rightmost noun phrase. However, it seems to be true that potential sentences and transitive sentences are syntactically related. Therefore, I have presented these example sentences. 72 "intransitive-transitive pair" verbs. "Ergative" verbs and "middle" verbs appear in pairs with their transitive counterparts. Observe the following sentences: (19)a. John ga sono to o ake-ta. Nom that door Acc open-Past "John opened that door." b. Sono to ga ak-ta. Nom open-Past "That door opened." c. Sono to ga kantan ni ak-ru. Nom easily open-Non-Past "That door opens easily." Sentences (19a) and (19b) form an "ergative" pair and aki ((something) open) in sentence (19b) is an "ergative" verb. Sentences (19a) and (19c) form a "middle" pair and ak-((something) open) in sentence (19c) is a "middle" verb. In the following pages, I will discuss the differences between "ergative" and "middle" verbs in depth. 73 4.2. "Ergative" Verbs versus "Middle" Verbs 4.2.1. Transitivity I consider that both "ergative" and "middle" verbs are non-volitional intransitive verbs and have transitive counterparts. Thus, they may appear in pairs with their transitive counterparts: an "ergative" pair and a "middle" pair. However, "ergative" verbs are lexical intransitive verbs whereas "middle" verbs are derived intransitive verbs. In other words, "ergative" verbs are intransitive verbs of "intransitive-transitive pair" verbs in lexicon while "middle" verbs are underlyingly transitive verbs and have gone through transformations, which are potential and intransitivization. As I have mentioned previously, in Japanese, intransitive verbs may be passivized and the passivization of intransitive verbs results in the affective passive as in ame ni hur-rare-ta (I got rained on). Passivization of "ergative" verbs seems grammatical and results in the affective passive. However, passivization of "middle" verbs seems ungrammatical. Observe (20): (20)a. Kinoo wa hasi no mannake de kuruma yesterday Top bridge Gen middle at car m koware-rare-te taihen-datta. Abl break-Pass-because terrible was "It was terrible because I was affected by the fact that my car broke down in the middle of the bridge yesterday." (ergative) 7 4 b. *Watasi no kuruma ni kantanni koware-rare-ru I Gen car A b l easily break-Pass-Non-Past Li t : "I am affected by the fact that my car breaks down easily." (middle) Sentence (20a) is grammatical and involves a passivized "ergative" verb. Sentence (20b) is ungrammatical and involves a passivized "middle" verb. I consider that potentialization, which derives a "middle" verb, has a close relationship with passivization. Therefore, the form of passive is closely associated. (I will discuss the close relationship between potential and passive in Chapter 6). The observation above may suggest that "ergative" verbs are lexical intransitive verbs whereas "middle" verbs are derived intransitive verbs. 4.2.2. Stativity "Middle" verbs seem to acquire stativity from the potential morpheme rare, which is stative. In fact, "middle" sentences show similarities to potential sentences while "ergative" sentences do not. One may consider that both "middle" and potential sentences are stative. Thus, "middle" verbs, like "potential" verbs and other stative verbs, represent state of being rather than an action or event whereas "ergative" verbs are non-stative and represent an action or event. For example, the suffix which indicates a point in time, -tokoro (Lit: at the point where), can occur with "ergative" verbs. However, -tokoro cannot occur with "middle" verbs since "middle" verbs do not specify particular time, but rather express state, which is above specific time point. Observe (21): 75 (21)a. John no kuruma ga koware-ta-tokoro-da. Gen car Nom break-Past-at the point where-be "John's car has just broken down." b. *John no kuruma wa kantanni Top easily koware-ru-tokoro-da. break-Non-Past-at the point where-be Lit: "It is at the point where John's car breaks down easily." Sentence (21a) involves an "ergative" verb koware-ru (break) and expresses that the event has just occurred with the suffix -tokoro. Sentence (21b) involves a "middle" verb koware-ru (break). Since sentence (21b) does not indicate any action or event, it results in an ungrammatical sentence after the suffixiation of -tokoro. The difference in stativity between "ergative" verbs and "middle" verbs can be observed in many other syntactic characteristics, such as possibility of the attachment of the te-iru form, reference of tense, meaning difference of the nagara construction, possibility of other suffix attachment, and so forth. Stativity of "middle" verbs prevents the attachment of the te-iru form, which indicates the progressive action or the perfective action. The attachment of the te-iru form to the "ergative" verbs results in the resultative meaning rather than the progressive meaning. Observe the following example sentences: 76 (22) a. Sono kuruma wa kantan ni koware-ru. car Top easily break-Non-Past "That car breaks easily." (middle) b. *Sono kuruma wa kantan ni koware-te-iru. Lit: "That car is being in the state of breaking easily." (23) a. Sono kuruma wa koware-ta. break-Past "That car broke down." (ergative) b. Sono kuruma wa koware-te-ita. "That car was broken." (resultative) *"That car was breaking down." (progressive) Sentence (22a) is "middle" and a "middle" sentence with the te-iru form like sentence (22b) with the meaning specified is ungrammatical. Of course, (22b) can have the meaning such as "That car has broken down easily," but then it indicates event and the underlying verb 77 used here is "ergative." Sentence (23a) is "ergative" and an "ergative" sentence with the te^  iru form like sentence (23b) is perfectly grammatical and indicates the resultative meaning. "Ergative" sentences seem to hold the "cause-effect" semantic relationship with their transitive sentences. For example, an "ergative" sentence to ga ak-ta (the door opened) may be seen as the result of a transitive sentence John ga to o ake-ta (John opened the door). The transitive sentence John ga to o ake-ta may be seen as a "causative" sentence. Observe (24): Semantically, a transitive sentence John ga to o ake-ta (John opened the door) can be seen as (24). Si in (24) indicates the transitive sentence and S2 indicates an "ergative" sentence to ga ak- (the door opens). (I will discuss the relationship between "ergative" sentences and "causative" sentences in Chapter 5). Since "ergative" sentences seem to represent effect or result of "causative" sentences, they are often in past tense. However, there are some occasions that "ergative" sentences are in non-past tense. Non-past tense in "ergative" sentences, like in other sentences of actions and events, refers to future time or to some habitual actions. "Middle" sentences, unlike "ergative" sentences, are often in non-past tense and represent the quality of being. Non-past tense in "middle" sentences, unlike in "ergative" sentences, seems to refer to present time. Observe the following ambiguous sentence: 78 (25). Kono huku wa sugu kawak-ru. this clothes Top quickly dry-Non-Past "These clothes will dry quickly." (ergative) "These clothes dry quickly." (middle) Sentence (25) is ambiguous. It could be an "ergative" sentence representing the future effect or could be a "middle" sentence representing the quality of "the clothes." However, if specific time adverbial phrases are added, the ambiguity disappears resulting in only the ergative interpretation. Observe (26): (26)a. Kono huku wa asita made-ni kawak-ru. this clothes Top tomorrow by dry-Non-Past "These clothes will dry by tomorrow." (ergative) b. Kono huku wa asita made-ni kawak-ru. tomorrow by *"These clothes dry by tomorrow." (middle) 79 The middle interpretation as in (26b) is ungrammatical since it includes a specific time reference: asita made-ni (by tomorrow). The nagara construction which expresses two simultaneous actions or states has two distinct meanings depending on stativity of sentences. If nagara occurs with a non-stative verb, it expresses the meaning either "while" or "although." If a non-stative verb is a durative verb, nagara expresses either "while" or "although," but if it is a punctual verb, nagara expresses "although" only and has a resultative meaning. If the nagara construction occurs with a stative verb, nagara expresses only the meaning "although." However, nagara with a stative verb, unlike nagara with a non-stative punctual verb, has no resultative meaning. Observe the following sentences with the nagara construction: (27)a. John wa tabe-nagara benkyoo-si-ta. Top eat while study-Past "John studied while eating." b. John wa takusan tabe-nagara mattaku hutor-nai. Top a lot although at all get f at-Neg-Non-Past "Although John eats a lot, he does not get fat at all. c. John no otoo-san wa zyuu-nen mae ni Gen father Top ten year before at nakunar-nagara ima demo minna ni pass away although now even everyone Abl 80 oboe-rare-te-iru. remember-Pass-Non-Past "Although John's father passed away ten years ago, he is still remembered by everyone." d. John wa gakusei-nagara benkyoo-su-nai-de Top student although study Neg instead of asob-de bakari i-ru. play only do-Non-Past "Although John is a student, he only plays around instead of studying." Sentences (27a), (27b) and (27c) involve a non-stative sentence with the nagara construction. Sentences (27a) and (27b) involve a non-stative durative verb tabe-ru (eat). Therefore, nagara is translated as either "while" or "although" depending on the matrix sentences. Sentence (27c) involves a non-stative punctual verb nakunar-ru (pass away). Therefore, nagara is translated as "although" and has a resultative meaning. However, sentence (27d) involves a stative verb phrase gakusei da in nagara construction. Therefore, nagara expresses only the meaning of "although." It has no resultative meaning. "Ergative" verbs are considered to be non-stative. Nagara with "ergative" verbs, like nagara with punctual verbs, is translated as "although" only and expresses a resultative meaning. "Middle" verbs are considered to be stative. Nagara with "middle" verbs is translated as "although" but, unlike nagara with "ergative" verbs, has no resultative meaning. Observe (28) and (29): 81 (28)a. Subete no mado ga ak-ta. all Gen window Nom open-Past "All the windows opened." (ergative) b. Subete no mado ga ak-nagara kono all Gen window Nom open although this heya wa mada atui. room Top still hot-Non-Past "Although all the windows are opened, this room is still hot. (ergative) (29)a. Sono to wa kantan-ni ak-ru. that door Top easily open-Non-Past "That door opens easily." (middle) 82 b. Sono to wa itsumo kantan-ni ak-nagara kyoo that door Top always easily open although today dake wa doomo aki ga warui. only Top really opening Nom bad "Although that door always opens easily, only today the opening is really bad." (middle) Sentences (28a) and (28b) are "ergative" and the latter expresses a meaning similar to a resultative sentence subete no mado ga ak-te-i-nagara (although all the windows are open). Sentences (29a) and (29b) are "middle" and the latter has no resultative meaning. In general, stative sentences are so-called timeless and cannot co-occur with auxiliary verbs which imply time, such as -hazimeru (start to), -kakeru (begin to), and so forth. I will refer to these auxiliary verbs as "time" auxiliary verbs. (30)a. Kono heya ni gakusei ga go-nin this room in student Nom five-people i-ru. exist-Non-Past "There are five students in this room." 83 b. *Kono heya ni gakusei ga go-nin this room in student Nom five-people i-hazime-ta. exist-start to-Past Lit: "There started to be five students in this room." c. *Kono heya ni gakusei ga go-nin this room in student Nom five-people i-kake-ta. exist-be bound to-Past Lit: "There was bound to be five students in this room." Sentence (30a) is a stative sentence with a stative verb iru ((animate thing) exist). Because of stativity of sentence (30a), the attachment of the auxiliary verbs to indicate the inceptive point results in ungrammatical sentences as in (30b) and (30c). "Ergative" verbs seem to permit the attachment of the auxiliary verbs -hazimeru and -kakeru whereas "middle" verbs seem to reject it. I consider that the fact that "middle" verbs do not seem to allow the attachment of those auxiliary verbs is due to stativity of "middle" verbs. Observe the following sentences in (31) and (32): 84 (31)a. John no kuruma ga koware-ta. Gen car Nom break-Past "John's car broke down." (ergative) b. John no kuruma ga koware-hazime-ta. break-start to-Past "John's car started to break down." c. John no kuruma ga koware-kake-ta. break-begin to-Past "John's car began to break." (32)a. John no kuruma wa kantan-ni koware-ru. Gen car Top easily break-Non-Past "John's car breaks easily." (middle) 85 b. *John no kuruma wa kantan-ni koware-hazime-ru. break-start to-Non-Past Lit: "John's car starts to be breakable easily." c. *John no kuruma wa kantan-ni koware-kake-ru. break-begin to-Non-Past Lit: "John's car begins to be breakable easily." Sentences in (31) involve an "ergative" verb koware-ru (break) and are all grammatical after the attachment of the "time" auxiliary verbs. Sentences in (32) involve a "middle" verb koware-ru (break/breakable) and are all ungrammatical after the attachment of those auxiliary verbs. When "middle" sentences in (32) and stative sentences in (30) are compared, there seems to be a great similarity between "middle" and stative sentences. Thus, it seems to be true that one of the major differences between "ergative" and "middle" verbs is their stativity. "Middle" verbs, which seem to adopt stativity from the underlying stative potential morpheme rare, may be less stative than "potential" verbs and other stative verbs such as aru ((inanimate thing) exist), iru (need), and so forth. For example, adjectives -yasui (be easy to) and -nikui (be difficult to) can be attached to "middle" verbs whereas they cannot be attached to "potential" verbs nor to other stative verbs. Observe (33), (34), and (35): 86 (33) a. Kono tako wa karui node yoku agar-ru. this kite Top light because well go up-Non-Past "This kite flies well because it is light." (middle) b. Kono tako wa agar-yasui. be easy to-Non-Past "This kite is easy to fly." c. Kono tako wa agar-nikui. be difficult to-Non-Past "This kite is difficult to fly." (34) a. John ni kono tako ga age-rare-ru. Abl this kite Nom raise-Pot-Non-Past "John can fly this kite." b. *John ni kono tako ga age-rare-yasui. Pot-be easy to-Non-Past Lit: "To John it is easy to be able to fly this kite." 87 c. *John ni kono tako ga age-rare-nikui. Pot-be difficult to-Non-Past Lit: 'To John it is difficult to be able to fly this kite." (35 )a. John wa kuruma ga san-dai ar-ru. Top car Nom three have-Non-Past "John has three cars." b. *John wa kuruma ga san-dai ar-yasui. be easy to-Non-Past Lit: "It is easy for John to have three cars." c. *John wa kuruma ga san-dai ar-nikui. be difficult to-Non-Past Lit: "It is difficult for John to have three cars." Sentences in (33) have a "middle" verb aga-ru (go up), sentences in (34) have a "potential" verb age-rare-ru (transitive verb age (raise) + the potential morpheme -rare), and sentences in (35) have a stative verb ar-ru (exist). Since the degree of stativity in a potential verb as in (34) and a stative verb as in (35) is high, the suffixiation of -yasui and -nikui results in ungrammatical sentences as in (34) and (35). However, the degree of stativity in "middle" verbs does not seem to be very high, since the suffixiation of -yasui and -nikui does not result in ungrammatical sentences. This low stativity for "middle" verbs leads me to the 88 assumption that "middle" verbs themselves do not have full stativity but simply have a stative underlying structure, which is the potential construction; that is, stativity is adopted into "middle" verbs in the process of "middle" verb formation. Thus, I must state that "middle" verbs have underlying low degree stativity. I will discuss the relationship between "middle" verbs and the potential construction in the following pages and in Chapter 6. 4.2.3. The [NP ni NP ga ] Construction Both "ergative" and "middle" verbs seem to permit the [NP ni NP ga ] construction. However, ni in "ergative" sentences seems to be mostly in the dative case expressing the goal whereas ni in "middle" sentences is in the ablative case expressing the agent. Observe (36) and (37): (36)a. Kane ga takusan atumar-ta. money Nom a lot gather-Past "A lot of money gathered." (ergative) b. John ni / ni taisite kane ga takusan atumar-ta. Dat "A lot of money gathered to John." *"John gathered a lot of money." 89 (37)a. Kane ga kantanni atumar-ru. money Nom easily gather-Non-Past "Money gathers easily." (middle) b. John ni kane ga kantanni atumar-ru. Abl "John can easily gather money." or "Money gathers easily for John." c. John ni totte kane ga kantanni atumar-ru. "John can easily gather money." Sentences in (36) are "ergative." The particle ni is in the dative case, and expresses the goal. It can be replaced by ni taisite as in sentence (36b). Sentences in (37) are "middle." The particle ni is in the ablative case, and expresses agent. It can be replaced by ni totte as in sentence (37c). Potential sentences involve the [NP ni NP ga ] construction; ni, which indicates agent, is in the ablative case, and ga, which indicates underlying object, is in the nominative case. Thus, "middle" verbs in the [NP ni NP ga ] construction share the same structure as that of potential sentences. Observe (38): 90 (38)a. Kodomo ni sono to ga kantan ni ake-rare-ta. child Abl that door Nom easily open-Pot-Past "Children could open that door easily." (potential) b. Kodomo ni sae sono to ga kantan ni ak-ru. Abl Nom open-Non-Past "Even for children, that door opens easily. / Even children could open that door easily." (middle) Both potential sentence (38a) and "middle" sentence (38b) permit the presence of agent marked by the ablative case marker ni. Notice that the English translation for the "middle" sentence (38b) may be similar to that of the potential sentence (38a). The evidence above again may imply that "middle" verbs are closely related to the potential construction. As I have mentioned previously, Inoue claims that "middle" verbs have the so-called implied agent while "ergative" sentences do not. It can be proved that "middle" verbs have the "implied" agent from the fact that the agent may appear in "middle" sentences as in the example middle sentence (38b). Contrary to Inoue's claim, "ergative" verbs also seem to have the so-called implied agent although on the surface, the agent never seems to appear in "ergative" sentences. I will compare "ergative" verbs and "natural phenomenal" verbs, which are also non-volitional. Observe (39): 91 (39)a. Ame ga hur-ta. rain Nom fall-Past ir It rained. t t b. To ak-ta. door Nom open-Past ft' 'The door opened. tt (ergative) "Natural phenomenal" sentence (39a) does not seem to have the "implied" agent whereas "ergative" sentence (39b) may have the "implied" agent since it can be viewed as a sentence in a "cause-effect" pair with a transitive counterpart ake-ru (open). In a sentence such as "watasi ga to o issyookenmei ni os-tara, to ga ak-ta" (when I pushed the door hard, the door opened), the presence of the agent in the underlined sentence, which is (39b), is very clear. In conclusion, although "ergative" and "middle" verbs are considered to be non-volitional, it may be true that they have the so-called implied agent. "Middle" verbs seem to have the "implied" agent due to the process of the "middle" verb formation, which involves the potential construction. "Ergative" verbs also seem to have the "implied" agent as the event represented by an "ergative" verb is considered to be caused by someone. However,they may not have the "implied" agent when the event is considered to be initiated by itself. For instance, the event opening of the door in the sentence zidoo doa ga  ak-ta is not considered to be caused by anyone but to be initiated by itself. The sentence does not have the "implied" agent. Thus, I will conclude that the "implied" agent always exists in "middle" verbs but does not necessarily exist in "ergative" verbs. 92 4.3. "Perceptual" Verbs Transitive verbs which express perception mini (see) and kiku (hear) have intransitive counterparts mieru (be visible) and kikoeru (be audible)15, and appear in intransitive-transitive pairs. Like other intransitive verbs of "intransitive-transitive pair" verbs, mieru and kikoeru are considered to be non-volitional. They do not appear in the imperative construction. For example, the imperative sentences Mie-ro (Be visible!) and Kikoe-ro (Be audible!) are ungrammatical. However, unlike many of the other intransitive verbs of "intransitive-transitive pair" verbs, they are generally considered to be stative. Contrary to the general analysis, mieru and kikoeru seem to behave like non-stative verbs also. For example, there are some occasions that those verbs permit the attachment of the te-iru form, and of other suffixes which indicate the point in time such as -tokoro-da (be at the point where), -hazimeru (start to), and -kakeru (begin to). Thus, it seems that mieru and kikoeru are in some cases stative and in other cases non-stative. From the evidence above, I will consider that mieru and kikoeru are both "ergative" and "middle" verbs. In fact, the pairs mini and mieru. and kiku and kikoeru. seem to exhibit characteristics of both an "ergative" pair and a "middle" pair. Observe (40): 15In English, ggg and hear do not describe actions but perceptions. They usually contrast with look at and listen to, which more clearly involve the idea of volition and effort on the part of the observer. Japanese does not have the distinguished pairs of verbs which show the contrast mentioned above. Mini is translated as both see and look at and kiku as both hear and listen to. Since mieru (be visible) and kikoeru (be audible) describe perception, I will consider that those verbs are paired with mini (see) and kiku (hear) rather than with mini (look at) and kiku (listen to). 93 (40)a. John wa tonari no monooto o kik-ta. Top neighbor Gen noise Acc hear-Past "John heard the neighbor's noise." (transitive) b. Tonari no monooto ga kikoe-ta. Nom be audible-Past "The neighbor's noise was heard." c. Tonari no monooto ga yoku kikoe-ru. Nom well be audible-Past "The neighbor's noise is well audible." Sentence (40a) is a transitive sentence with a transitive verb kiku (hear) and sentences (40b) and (40c) are intransitive sentences with an intransitive verb kikoeru (be audible). Sentence (40b) expresses an event whereas sentence (40c) expresses state of being. Thus, it is considered that the pair of sentences (40a) and (40b) is similar to an "ergative" pair, and the pair of sentences (40a) and (40c) is similar to a "middle" pair. In this thesis, I will consider that both mieru (be visible) and kikoeru (be audible) are either lexical intransitive verbs or derived intransitive verbs. Lexical intransitive verbs mieru and kikoeru are considered to be "ergative" verbs, and derived intransitive verbs mieru and kikoeru are considered to be "middle" verbs, which are derived from "potential" verbs. 94 In what follows, I will further examine the characteristics of mieru and kikoeru. and attempt to differentiate "ergative" verbs mieru and kikoeru from "middle" verbs mieru and kikoeru. Example sentences (40a) and (40b) are considered to be in an "ergative" pair. However, this "ergative" pair does not seem to exhibit a "cause-effect" relationship since mini (see) and kiku (hear), unlike other transitives with intransitive counterparts, are not "causative" verbs. In fact, mini and kiku express perception rather than an action with a causative meaning. For example, a sentence John ga mado o ake-ta (John opened the window) may be interpreted with a causative meaning as John Imado ga ak-1 cause (John caused the window to open). However, a sentence with mini or kiku cannot be interpreted as a "causative" sentence. A sentence John ga hen-na monooto o kik-ta (John heard a strange noise) cannot be interpreted as a sentence * John [hen-na monooto ga kikoe-1 cause (John caused a strange noise to be audible). Observe (41) and (42): (41) John ga mado o ake-ta. Nom window Acc open-Past Agent "John opened the window." 95 (42) John ga hen-na monooto o kik-ta. Nom strange noise Acc hear-Past Experiencer "John heard a strange noise." S 2 John hen-na monooto kikoe- cause (41) shows a transitive sentence and its causative interpretation exhibited in a tree structure. Notice that the particle ga in (41) is associated with the case of agent. Thus, the sentence in (41) is an agentive transitive sentence. However, (42) involves a non-agentive 96 transitive sentence. The sentence in (42) has no causative meaning. Thus, the causative interpretation exhibited in the tree in (42) is ungrammatical. Notice that the particle ga in sentence (42), unlike in sentence (41), is associated with the case of Experiencer. Although an "ergative" pair with mieru (be visible) or kikoeru (be audible) is somewhat different from other "ergative" pairs, it has contrastive characteristics when it is compared to a "middle" pair with mieru or kikoeru. "Ergative" verbs mieru and kikoeru are like non-stative verbs since they permit the attachment of the te-iru form and of other suffixes which imply time, such as -tokoro da (be at the point where), -hazimeru (start to), and -kakeru (begin to). "Middle" verbs mieru and kikoeru. on the other hand, are like stative verbs since they do not seem to permit the attachment of those suffixes. Observe (43) and (44): (43)a. Fuzi-san ga mie-ta. Mt. Fuji Nom be visible-Past "Mt. Fuji appeared." (ergative) b. Fuzi-san ga mie-te-iru. Non-Past "Mt. Fuji has appeared (and is in sight)." c. Fuzi-san ga mie-ru-tokoro-da. be at the point where-be-Non-Past "Mt. Fuji is about to appear." 97 d. Fuzi-san ga mie-hazime-ta. start to-Past "Mt. Fuji started to appear." e. Fuzi-san ga mie-kake-ta. begin to-Past "Mt. Fuji began to appear." (44)a. Megane o kake-ta node sono zi ga glasses Acc wear-Past because that character Nom yoku mie-ru. well be visible-Non-Past "Because I put on my glasses, that character is well visible." (middle) b. ?Sono zi ga yoku mie-te-iru. Progressive Lit: "That character is being well visible (temporarily)." 98 c. *Sono zi ga yoku mie-ru-tokoro-da. be at the point where-be-Non-Past Lit: "That character is about to be well visible." d. *Sono zi ga yoku mie-hazime-ru. start to-Non-Past Lit: "That character starts to be well visible." e. *Sono zi ga yoku mie-kake-ru. begin to-Non-Past Lit: "That character begins to be well visible." Sentence (43a) is "ergative" and like other "ergative" verbs permits the attachment of the suffixes as shown in sentences (43b), (43c), (43d), and (44e). Sentence (44a) is "middle" and does not permit the attachment of the suffixes as shown in sentences (44b), (44c), (44d) and (44e). From the evidence above, it is clear that "ergative" sentence (43a) is non-stative whereas "middle" sentence(44a) is stative. "Ergative" verbs mieru and kikoeru in the present tense, like other non-stative verbs in the present tense, seem to imply either future time or reoccurring events. On the other hand, "middle" verbs mieru and kikoeru in the present tense, like other stative verbs, refer to the present time. Observe the following example: 99 (45). Fuzi-san ga mie-ru. Nom be visible-Non-Past "Mt. Fuji will appear." "Mt. Fuji is visible." (ergative) (middle) Sentence (45) is ambiguous. It could be an "ergative" sentence representing an event in the future or reoccurring event, or a "middle" sentence implying some potentiality. However, if specific time adverbial phrases are added, the ambiguity disappears resulting in only the ergative interpretation. Observe (46): (46)a. Fuzi-san ga asita mie-ru. tomorrow "Mt. Fuji will appear tomorrow." (ergative) 100 b. Fuzi-san ga asita mie-ru16. tomorrow *"Mt. Fuji is visible tomorrow." (middle) The middle interpretation as in (46b) is ungrammatical due to the specific time adverb asita (tomorrow). Nagara with "ergative" mieru and kikoeru. like nagara with other "ergative" verbs, is translated as "although" only and expresses a resultative meaning. Nagara with "middle" verbs mieru and kikoeru. like nagara with other "middle" verbs, is translated as "although" and expresses no resultative meaning. Observe (47) and (48): (47)a. Taiyoo ga mie-ta. sun Nom be visible-Past "The sun appeared." (ergative) b. Taiyoo ga higasini mie-nagara soto wa in the east although outside Top 16There is a verb mieru with the meaning of "show up" as in Tanaka-san ga asita mieru (Mr. Tanaka will show up tomorrow/*Mr. Tanaka will be visible tomorrow). There is no question that it is derived from mieru that we are discussing here. Notice that mieru with the meaning of "show up" is strictly "ergative" with non-stative characteristics. 101 mada usu-gurai. still slightly dark-Non-Past "Although the sun has appeared in the east, the outside is still slightly dark." (Note the resultative meaning) (ergative) (48)a. Sono zi ga yoku mie-ru. that character Nom well be visible-Non-Past "That character is well visible." (middle) b. Sono zi ga yoku mie-nagara John wa although Top mie-nai to uso o iw-ta. be visible-Neg Comp lie Acc say-Past "Although that character is well visible, John told a lie that it is not visible." (middle) Sentences (47a) and (47b) are "ergative" and the latter expresses a resultative meaning similar to mie-te-i nagara. Sentences (48a) and (48b) are "middle" and the latter has no resultative meaning. 102 I consider that "ergative" mieru and kikoeru are lexical intransitive verbs whereas "middle" mieru and kikoeru are underlyingly transitive verbs and have gone through transformations, which are potential and intransitivization. Since potential and passive have a close relationship, which I will discuss in Chapter 6, passivization of "middle" mieru and kikoeru is ungrammatical. However, passivization of "ergative" verbs seems grammatical and results in the affective passive. Observe the following example sentences: (49)a. Huuhu-genka no koe ni made married couple quarrel Gen voice Abl even kikoe-rare-te sono huuhu wa be audible-Pass-because that married couple Top hazukasii omoi o si-ta. embarrassment feeling Acc do-Past "The married couple was embarrassed by the fact that even their voice of quarrel was heard." (ergative) 103 b. *Hotyooki de tiisai oto ni made hearing aid Abl small sound Abl even yoku kikoe-rare-ta. well be audible-Pass-Past Lit: "He was affected by the fact that even small sounds were well audible with a hearing aid." (middle) Sentence (49a) is grammatical and involves "ergative" kikoeru. Sentence (49b) is ungrammatical and involves "middle" kikoeru. In conclusion, mini (see) and kiku (hear), like other transitive verbs of "intransitive-transitive pair" verbs, have intransitive counterparts mieru (be visible) and kikoeru although they do not have causative meaning. Those intransitive counterparts mieru and kikoeru are at some times considered to be lexical intransitive verbs and at other times to be underlyingly transitive verbs17. In fact, they show characteristics of both "ergative" verbs (lexical intransitive verbs) and "middle" verbs (derived intransitive verbs). Thus, I have examined those verbs from the point of view of "ergative" and "middle" verbs. In the next chapter, I will focus on the relationship between "ergative" verbs and causality. 17Makino and Tsutsui (1986) classify mieru and kikoeru as lexical intransitive verbs. Soga (1978) considers those verbs as irregular non-volitional potential forms of transitive verbs mini (see/look at) and kiku (hear) as opposed to regular volitional potential forms mi-rare-ru (can be seen/can be looked at) and kik-rare-ru (can be heard). Soga (1986) recognizes their non-stative as well as stative characteristics. In this thesis, those contradictory characteristics are considered to be due to their "ergative" and "middle" differentiations. 104 CHAPTER 5 ERGATIVE Lakoff (1965) studied the relationship between transitive verbs and their intransitive counterparts in English and concludes that transitive verbs and their intransitive counter-parts are related by way of causativization. Observe his analysis: (l)a. Intransitive Sentences II [ +causative] Ii Transitive Sentences b. The sauce thickened, (intransitive) Ii John brought it about that the sauce thickened. ([ +causative]) Ii John thickened the sauce, (transitive) (la) shows the semantic relationship between intransitive verbs and transitive verbs via causativization. Lakoff's example sentences as in (lb) illustrate the relationship. In fact, transitivity and causality seem to be closely related to one another. The following statement by Waterman (1963) also reflects the close relationship between causative and transitive: 105 Unfortunately, the Greek terminology did not always fare so well in translation. The accusative case, as an example, was called by the Greeks aitiatike, that is, "the thing caused by the verb." But aitia means both "cause" and "accusation," and certain of the Latin grammarians later called it the "accusing" case rather than the "causing." Today we should really refer to the "causative case," but tradition has frozen the mistranslation into our terminology, and it will probably stay there forever1. In Japanese, like Lakoff's analysis in English, intransitive verbs and their transitive counterparts might be analyzed as related to one another via causativization as in the following example sentences: (2)a. Huku ga kawai-ta. Clothes Nom dry-Past "The clothes dried." (ergative) Ii John wa [huku ga kawak-] sase-ta Caus Ii John wa huku o kawakas-ta. Top clothes Acc dry-Past "John dried the clothes." (transitive) Waterman (1963), p. 8. Kane ga nar-ta. bell Nom ring-Past "The bell rang." (ergative) II John wa [kane ga nar-] sase-ta Caus II John wa kane o naras-ta. Top bell Acc ring-Past "John rang the bell." (transitive) Ziko ga oki-ta. accident Nom happen-Past "An accident happened." (ergative) II John wa [ziko ga oki-] sase-ta Caus II 107 John wa ziko o okos-ta. Top accident Acc cause-Past "John caused an accident." (transitive) However, different from Lakoff s analysis, I consider that both intransitive verbs ("ergative" verbs) and their transitive counterparts are derived at the lexical level. The former have the intransitive morpheme R and the latter have the transitive morpheme S. Observe the following examples: (3). "ergative" verbs transitive verbs ak + R ak-ru "(something) open" ak + S ake-ru "open" nar + R => nar-ru "(something) ring" nar + S => naras-ru "ring" ok + R => oki-ru "(something) happen" ok + S => okos-ru "cause" Like Waterman's claim, the terminology transitive and causative may be confusing in Japanese also. Numerous transitive verbs such as akeru (open), atumeru (gather), hazimeru (start), and so forth might be analyzed to have causality especially when they appear in a pair with their intransitive counterparts. For example, watasi wa mado o ake-ta 108 (I opened the window) may have an implied causative meaning and suggest the event mado  ga ak-ta (the window opened). However, unlike Lakoff s analysis, I consider that causality happens to be one of the characteristics that those transitive verbs have rather than suggesting that they are actually derived from their intransitive counterparts via causativization. In fact, Shibatani (1976), who studied Japanese verbs on the basis of their causality, classified those transitive verbs as lexical causative verbs as opposed to productive causative verbs with the causative morpheme sase2. According to his analysis, lexical causative verbs need to be present in the lexicon whereas productive causative verbs do not need to be present in the lexicon. Also, Inoue (1990) claims that transitive verbs of "intransitive-transitive pair" verbs carry a causative meaning and refers to these verbs as causative verbs. Nevertheless, like Shibatani, she seems to treat them as lexical verbs rather than derived verbs. In conclusion, I, like Shibatani and Inoue, acknowledge a causative meaning in transitive verbs of "intransitive-transitive pair" verbs, and consider it as the characteristic of those transitive verbs. When the relationship between "ergative" verbs and their transitive counterparts is examined, there is a possibility that they might be considered to be related by way of causativization, like Lakoff's analysis. That is, transitive counterparts might be considered to be derived from "ergative" verbs via causativization. However, I consider at this stage of my analysis that both "ergative" verbs and their transitive counterparts are at the lexicon. I believe that this treatment avoids confusion between transitive and causative. Seemingly, there are many different theories on the relationship between "ergative" verbs and transitive counterparts, and I believe that mine can be seen as one of them. In this chapter, the relationship between "ergative" verbs and transitive counterparts has been discussed along with causality that transitive counterparts have. In the next 2The allomorphs for the causative morpheme sase are sase and ase; sase for vowel verbs and ase for consonant verbs. chapter, I will concentrate on "middle" verbs and other constructions such as potential, passive, and transitive. 110 CHAPTER 6 MIDDLE 6.1. Potentiality Since potentialization is considered to derive "middle" verbs, "middle" and "potential" verbs share great similarities both semantically and syntactically. For instance, both the "middle" sentence kono to ga kantan-ni ak-u and the potential sentence kono to ga  kantan-ni ake-rare-ru may be translated as "one can open this door easily." Furthermore, both "middle" and potential sentences have ni in the ablative case in the [NP ni NP ga ] construction. Ni expresses the agent and ga expresses underlying object. Inoue (1990) claims that the intransitive morpheme and potential morpheme, both of which are recognized as re, are homonymous in Japanese. Observe her example sentences: (l)a. Taroo ga kono hako no futa o this box Gen lid Acc tor-ta. remove-Perfect "Taro removed the lid of this box." (agentive-transitive) I l l b. kono hako no futa ga tor-re-ta. remove-Intr-Perfect "The lid of this box came off." (ergative) c. kono hako no futa wa rakuni Top easily tor-re-ru. remove-Pot-Non-Perfect "One can remove the lid of this box easily." (middle)1 Notice that the intransitive morpheme re in (lb) is identical to the potential morpheme re in (lc). Thus, it is very likely that "middle" verbs are derived intransitive verbs via potentialization. Although Inoue's claim that "middle" verbs have the potential morpheme may be justifiable, her claim that the intransitive morpheme and the potential morpheme are homonymous may not be. Unlike Inoue, it seems better to consider that the potential morpheme is rare rather than re, and that Inoue's potential morpheme re is simply its allomorph. For instance, a potential form of taberu (eat) is either tabe-rare-ru (can eat/can 1Inoue (1990), example sentences (35 a, b, c), p. 20. 112 be eaten) with rare or tabe-re-ru with re. In fact, vowel verbs form two types of "potential" verbs either with rare or with re. Considering the potential morpheme as rare. I will discuss a close relationship between potentialization and passivization in the following pages. In doing so, I will claim that passivization of "middle" verbs is ungrammatical. 6.2. Relationship between Potential and Passive Potentialization has a close relationship with passivization. there are a few linguists who have studied this relationship. For instance, Chapin (1967) examined English potential and passive. The potential suffix is recognized as able in English. Observe his example sentences: (2)a. One can redeem coupons for cash. (transitive) b. Coupons can be redeemed for cash. (passive) c. Coupons are redeemable for cash. (potential) d. Coupons are able to be redeemed for cash2. (potential and passive) 2Chapin (1967), p. 54. 113 Sentence (2b) is passive and sentence (2c) is potential with the potential suffix able. They both involve passive transformation, which is characterized by the subject-object interchange. Sentence (2d), which involves both passive and potential constructions, is ungrammatical. Chapin (1990) claims that in English, potential constructions are closely related to passive constructions. In fact, passivization, along with other transformations, namely agent deletion transformation and Able transformation (potentialization), derives potential constructions. For instance, the potential sentence Glass is breakable has an underlying structure [glass be able for/to \A break glassl SilS2- First passive transformation applies to S^  and then agent deletion transformation applies. At last, Able transformation (potential transformation) derives the surface potential sentence Glass is breakable. In Japanese, like in English, potential and passive seem to be closely related. Matsushita (1961), for instance, categorizes potential as "formal" passive as opposed to "true" passive which includes simple and affect passives. In fact, syntactically simple passive and potential sentences share great similarities. They both permit the [NP ni NP ga ] construction and involve both the subject-object shift and the morpheme rare, which seems to be identical. On the surface, both in passive and potential sentences, agent is marked by the ablative case marker ni and underlying object is marked by the nominative case marker ga. There are differences between the two, however. Interestingly, potential constructions sometimes retain underlying structure (agent in the nominative case and object in the accusative case) rather than going through the subject-object shift. In such a case, it is considered that potential constructions do not involve the passive transformation. Observe (3): 114 (3)a. John ga susi o tabe-ru. Nom sushi Acc eat-Non-Past "John eats sushi." b. John ga susi o tabe-rare-ru. Nom Acc eat-Pot-Non-Past "John can eat sushi." c. John ni susi ga tabe-rare-ru. Abl Nom eat-Pot-Non-Past "To John, sushi can be eaten/can be edible." Sentence (3a) is a transitive agentive sentence. Sentences (3b) and (3c) are both potential sentences which are derived from sentence (3a). Sentence (3b) does not involve the passive transformation and retains the underlying structure [NP ga NP o ]. On the other hand, sentence (3c) involves the potential transformation and shows the [NP ni NP ga ] construction. Since passive morphemes and potential morphemes share the same form rare, sometimes ambiguity occurs. For instance, the sentence nihon de oisii kome ga tabe-rare-ru is interpreted as both a passive sentence (In Japan, tasty rice is eaten) and as a potential sentence (In Japan, tasty rice can be eaten). Semantically, passive and potential are certainly dissimilar. Passive sentences express some events or actions from the patient point of view whereas potential sentences 115 do not express any actions or event but simply state the generic quality. Furthermore, passive sentences are non-stative whereas potential sentences are stative. Observe (4): (4)a. Senkyo no kane ga atume-rare-ru. election Gen money Nom gather-Pass/Pot-Non-Past "Money for the election will be collected." (passive) "Money for the election can be collected." (potential) b. Senkyo no kane ga atume-rare-te-iru. election Gen money Nom Pass-Prog-Non-Past "Money for the election has been/is being collected." (passive) Sentence (4a) is ambiguous. It can be passive or potential. Since the potential interpretation of sentence (4a) is considered to be stative, the attachment of the te-iru form to the sentence results in an ungrammatical sentence whereas the te-iru form can be attached to the passive interpretation of sentence (4a) as in sentence (4b). Notice that the present tense in the passive sentence (4a) refers to future time (or habitual event), while the present tense in the potential sentence (4a) refers to present time. The evidence above may also suggest that potential sentences are stative. In conclusion, potential and passive sentences have many formal similarities. However, the potential morpheme rare carries stativity whereas the passive morpheme rare does not. Furthermore, potential does not necessarily require the subject-object shift 116 whereas passive (simple passive) does. Thus, I will consider that the potential morpheme and the passive morpheme happen to be homonymous and that it is not the case that passive must underlie in potential. 6.3. Relationship between "Middle" Verbs and Transitive Verbs I consider that "middle" verbs are derived from transitive verbs via potentialization and intransitivization. I have previously discussed similarities between "middle" verbs and potentialized transitive counterparts through various ways, such as stativity, the [NP ni NP ga ] construction, and so forth, so that it seems to be clear that "middle" verbs and potentialized transitive verbs are syntactically related. I consider that "potential" verbs are potentialized transitive verbs whereas "middle" verbs are derived intransitive verbs from potential forms. For instance, the "potential" verb ake-rare-ru (can open/can be opened) has a transitive verb stem ake- whereas the "middle" verb ak-u (open) has the intransitive verb stem ak^ . It can be considered that intransitivization of "potential" verbs derives "middle" verbs and their underlying transitive verbs in the following way: (5). John ga kata te de sono to o Nom one hand with that door Acc ake-ru. open-Non-Past "John opens that door with one hand." (transitive) •11 (Potentialization) 117 John ga kata te de sono to o ake-rare-ru. open-Pot-Non-Past "John can open that door with one hand." (potential) Ii (Intransitivization) Kata te de sono to ga ak-ru. Nom open-Non-Past "That door opens with one hand." (middle) (6). Haha ga kantan-ni sono hari ni mother Nom easily that needle Abl ito o toos-ru. thread Acc put through-Non-Past "The mother easily put thread through that needle." (transitive) Ii (Potentialization) 118 Haha ga kantan-ni sono hari ni ito o toos-rare-ru. put through-Pot-Non-Past "The mother can easily put thread through that needle" (potential) •li (Intransitivization) Kantan-ni sono hari ni ito ga Nom toor-ru. go through-Non-Past "The thread easily goes through that needle." (middle) Sentences in (5) and (6) illustrate the derivation of the "middle" verb ak-u (open) and toor-u. respectively. Middle "perceptual" verbs, mieru (be visible) and kikoeru (be audible), which I have mentioned previously, also seem to exhibit the relationship with their underlying transitive verbs mini (see) and kiku (hear) via potentialization. Observe the following sentences: (7). Gakusha ga seikoo-na kenbikyoo de saikin scholar Nom elaborate microscope with bacteria o mi-ru. Acc see-Non-Past "Scholars see bacteria with an elaborate microscope." (transitive) •11 (Potentialization) Gakusha ga seikoo-na kenbikyoo de saikin o mi-rare-ru. Acc see-Pot-Non-Past "Scholars can see bacteria with an elaborate microscope." (potential) (Intransitivization) Seikoo-na kenbikyoo de saikin ga Nom mie-ru. be visible-Non-Past "Bacteria are visible/can be seen with an elaborate microscope." (middle) (8). Ninpu ga pregnant Nom woman kikai o toosite machine Abl through taizi no fetus Gen 120 sindoo o kik-u. heartbeat Acc hear-Non-Past "Pregnant women hear the heartbeat of a fetus through a machine." (transitive) U- (Potentialization) Ninpu ga kikai o toosite taizi no sindoo o kik-rare-ru. hear-Pot-Non-Past "Pregnant women can hear the heartbeat of a fetus through a machine." (potential) U- (Intransitivization) Kikai o toosite taizi no sindoo ga kikoe-ru. Nom be audible-Non-Past "The heartbeat of a fetus is audible/can be heard through a machine." (middle) Sentences in (7) and (8)also illustrate the derivation of middle "perceptual" verbs mie-ru (be visible) and kikoe-ru (be audible). As the sentences show, mie-ru and kikoe-ru are derived from their potentialized transitive counterparts mi-rare-ru (can see) and kik-rare-ru (can hear). 121 I consider that there is in general a transformation which transforms specifically transitive verbs of "intransitive-transitive pair" verbs into "middle" verbs. So far, I have referred to this transformation as intransitivization. The application of this transformation is quite restricted; the transformation applies only to potentialized transitive verbs of "intransitive-transitive pair" verbs. This transformation may be stated as follows: Figure 1. Transformation Which Derives "Middle" Verbs NP ga stem + S - rare-=> NP ga stem + R -In Figure 1, S indicates a transitive morpheme and R indicates an intransitive morpheme. Figure 1 illustrates that the transformation forces the first NP ni to drop and intransitivizes a verb as the flip-flop of the S and R shows. 122 CHAPTER 7 CONCLUSION Any discussion on "ergative" and "middle" verbs in Japanese, as I have mentioned earlier, is still debatable since the definitions of "ergative" and "middle" slightly differ according to various linguists. In my thesis, I first attempted to define clearly the terms "ergative" and "middle" and then moved onto the examination of those verbs supporting my definitions. Having analyzed both syntactic and semantic properties of "ergative" and "middle" verbs, I may conclude that "ergative" verbs are lexically intransitive verbs and "middle" verbs are derived intransitive verbs, although they formally appear to be identical. 123 BIBLIOGRAPHY Chapin, Paul G. 1967. On the Syntax of Word-Derivation in English. Cambridge, Massachusetts: The MITRE Corporation. Chew, John J. 1973. A Transformational Analysis of Modern Colloquial Japanese. The Hague: Mouton & Co. N. V. Dixson, R.M.W. 1979. Ergativity. Language. 55.1,59-138. Fagan, Sarah M.B. 1988. The English Middle. Linguistic Inquiry. 19.2,181-203. Farmer, Ann K. 1980. On the Interaction of Morphology and Syntax. Ph.D. Dissertation, MIT. Inoue, Kazuko. 1976. Henkei bunpoo to nihongo (I). Tokyo: Taishukan Shoten. . 1990. On Nonagentive Subjects-English Japanese Comparative Syntax. Papers from the Middlebury Conference on Japanese Linguistics and Japanese Language  Teaching. 3-38. Jacobsen, Wesley. 1982. Transitivity in the Japanese Verbal System, reproduced by the Indiana University Linguistics Club, Bloomington, Indiana. Keyser, Samuel J., and Thomas Roeper. 1984. On the Middle and Ergative Constructions in English. Linguistics Inquiry. 15.3,381-416. Kindaichi, Haruhiko. 1955. Nihongo dooshi no tensu to asupekuto. Tokyo: Mugi Shobokan. Koizumi, Tamotsu et al. 1989. Nihongo kihon dooshi yoohoo jiten. Tokyo: Taishukan Shoten. Kuno, Susumu. 1973. The Structure of the Japanese Language. Cambridge, Massachusetts: The MIT Press. Lakoff, George. 1965. On the Nature of Syntactic Irregularity. Mathematical Linguistics and Automatic Translation, report No. NSF-16 to the National Science Foundation, Cambridge, Massachusetts. 124 Makino, Seiichi, and Michiko Tsutsui. 1986. A Dictionary of Basic Japanese Grammar. Tokyo: The Japan Times. Matsumoto, Makoto. 1976. Kiso nihongo bunpoo gaisetsu. Tokyo: Hakuteisha. Matsushita, Daizaburo. 1961. Hyoojun nihongo koogohoo. Tokyo: Hakuteisha. Miyagawa, Shigeru. 1980. Complex Verbs and the Lexicon. 1. Working Papers in Linguistics from A - > Z. Tucson, Arizona: University of Arizona. Shibatani, Masayoshi. 1976. The Grammar of Causative Constructions: A Conspectus. Syntax and Semantics 6. New York: Academic Press, 1-39. Soga, Matsuo. 1968. Derivational Morphology and the Lexicon in a Generative Grammar. Eibungaku kenkvuu [Studies in English Literature]. XLIV, 2,233-242. Soga, Matsuo, and Noriko Matsumoto. 1978. Foundation of Japanese Language. Tokyo: Taishukan Publishing Company. . 1986. On Stativity in Japanese. Nitobe-Ohira Memorial Conference on Japanese Studies. 174-222. Yamada, Takao. 1936. Nihon bunpoogaku gairon. Tokyo: Hoobunkan. Waterman, John T. 1963. Perspectives in Linguistics. Chicago: The University of Chicago Press. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100509/manifest

Comment

Related Items