第一語言對第二語言習得的影響：以華語結果式複合動詞習得研究為例

全文

(1)Chapter One Introduction 1.1 Motivation Languages differ as to what grammatical patterns they use to delineate events consisting of both an action and a result. Mandarin Chinese and English are good examples. In Mandarin, such events are mostly characterized with the resultative verb compounds (RVCs hereafter), e.g., xue-hui in (1a) and da-si in (2a). In English, such the same events are described with different grammatical structures, particularly patterns denoting Accomplishment events, e.g., (1b) and (2b). This point can be clearly illustrated by the following examples. (1) a. Mali Mary. xue-hui-le. zhongwen.. study-understand-LE. Mandarin. Chinese. b. Mary learned Mandarin Chinese. (2) a. Lisi Lisi. da-si-le. Zhangsan.. hit-die-LE. Zhangsan. b. Lisi killed Zhangsan (by hitting him). It should be noted that unlike the Accomplishment in English, which is a kind of situation, 1 the RVC is a morphological construction, often used to denote. 1. The notion of the situation or situation type is different from that of verb or verb class. In Vendler (1957, 1967), verbs are categorized according to the semantic properties of the main verb in a sentence. However, Smith (1997) argues that such constituents as subjects, arguments and adverbials also contribute a lot to the aspectual meaning of a given sentence. Thus, Smith uses the term ‘situation type’ instead of the term ‘verb classes’. See Section 2.2.1 for a discussion of situations proposed by Smith (1997). 1.

(2) Accomplishment or Achievement events in Mandarin Chinese. For a clear understanding of the reason why the action-result event can be expressed with the linguistic structures, it is necessary that we be aware of the semantic meanings of sentences with these structures. As shown in (1a) and (2a), the underlined RVCs (i.e., xue-hui ‘study-understand’ and da-si ‘hit-die’) are two-element verbs, the majority2 of which have the action-result relation between the meanings of their elements (e.g., Li and Thompson 1981; Tai 1984). Because of the semantic properties of their constituents, the RVCs can be used to express events consisting of both an action and a result. Likewise, the English Accomplishment is semantically composed of both a non-detachable process3 and a result; therefore, it can name an event with a natural endpoint (Smith 1997). According to this characteristic, with the Accomplishment learned Mandarin Chinese in (1b) conveys that Mary studied Mandarin Chinese for certain period of time (i.e., process) and as a result, she understood the language (i.e., result). Similarly, killed Zhangsan in (2b) suggests that Lisi hit Zhangsan and Zhangsan died as a result. Although both Mandarin RVCs and English Accomplishments both exhibit the action-result semantic property explicitly or implicitly, they are different in the specification of the completion/result. In Mandarin Chinese, the result of the event is 2. In what follows, we will show that RVCs can be formed with a stative verb as the first constituent, and the meanings of their constituents indicate the state-result relationship. 3 A process is an unbounded action. 2.

(3) specified by the resultative morpheme of the RVC, e.g., hui ‘understand’ and si ‘die’ in (1a) and (2a) respectively, while in English, the specification of completion is inherent in the meaning of the main verb (phrase),4 as in learn Mandarin Chinese in (1b) and kill Lisi in (2b) (Smith 1997). The examples indicate that in Mandarin Chinese the result of an event is ‘overtly’ expressed with the resultative morpheme of the RVC, whereas in English it is ‘covertly’ expressed in the meaning of the main verb in a sentence. In other words, while in Mandarin Chinese the notion of completion is marked with the resultative morpheme that has lexical content, in English it is ‘implied’ in the meaning of the verb. On the basis of the cross-linguistic distinction between the expressions of ‘completion’, Tai (2003) contends that the result of an event would be less salient to English speakers than to Mandarin speakers. This may also suggest that the lack of a clear ‘result-marking system’ in English could make the learning of the meaning and function of the RVC one of the most difficult tasks that an English speaker attempting to master Mandarin Chinese may face. This task is further complicated by the semantic features of the RVC, as illustrated in (3) and (4). (3) a.*Mali Mary. zhengzai imperfective. Asp.. xue-hui. zhongwen.. study-understand. Mandarin Chinese. 4. Another way of saying this is to state that in English whether a closed event is taken as terminated or completed depends to a great extent on the meaning of the main verb in a perfective sentence (Smith 1997). See Section 2.1.1 for more information on this point. 3.

(4) b. Mary is learning Mandarin. (4) a.*Mali Mary. kaishi start. xue-hui study-understand. zhongwen. Mandarin Chinese. b. Mary started learning Mandarin Chinese. (5) a.*Mali Mary. tingzhi stop. xue-hui study-understand. zhongwen. Mandarin Chinese. b. Mary stopped learning Mandarin Chinese. Sentence (3a) shows that Mandarin RVCs are incompatible with the imperfective marker, zhengzai. The reason is that RVCs denote only the result aspect in their aspectual meaning (Tai 1984, 2003), though their first constituent can be either an Activity, as in xue ‘study’ in (3a), or a Semelfactive, as in da ‘hit’ in (2a), both of which can signify a temporally unbounded action. Without the action aspect indicating the feature of duration, RVCs are instantaneous verbs. Accordingly, they cannot appear with durative grammatical markers, such as zhengzai, kaishi and tingzhi in (3a), (4a) and (5a), respectively. Distinct from Mandarin RVCs, which are semantically instantaneous, English Accomplishments are semantically durative and thus can occur with the durative linguistic labels, as shown in (3b), (4b) and (5b). Furthermore, due to the lack of the aspectual meaning of action, the result part is the semantic focus that imparts the central meaning to the RVC sentences. To illustrate, the main idea of Sentence (1a) with the RVC xue-hui ‘study-learn’ is that Mary learned Mandarin Chinese, not that Mary studied Chinese. In contrast to the 4.

(5) Mandarin RVC, the result part of which constitutes the semantic focus, both the action and result are the semantic foci of the English Accomplishment (Smith 1997). The cross-linguistic comparisons above clearly show that we cannot directly translate Mandarin RVCs into English (Accomplishment) verbs. As English does not have a direct counterpart of the RVC, English learners of Mandarin may either have trouble understanding the incompatibility between the RVC and the durative grammatical structure since the first element of an RVC can denote an uncompleted action or have difficulty grasping the essence of RVCs due to the difference in semantic properties5 exhibited in the corresponding English sentences, as in (3b), (4b) and (5b). Motivated by the cross-linguistic distinction between the specifications of completion/result in Mandarin and English and the cross-linguistic variations between Mandarin RVCs and English Accomplishments, this study explores the L2 acquisition of Mandarin RVCs by native English speakers to see whether they have full comprehension of the semantic features of RVCs in general and whether their acquisition varies according to the three RVC types—divided based on the semantic property of the two. constituents—Activity-Result,. Semelfactive-Result. and. State-Result RVCs (e.g., tui-kai ‘push-open’, ci-po ‘poke-broken’ and re-hun ‘hot-faint’, respectively). The theoretical background and the purposes of this study 5. That is, Mandarin RVCs are semantically instantaneous, while English Accomplishments are semantically durative. In addition, while the result part constitutes the semantic focus of the Mandarin RVC, both the action and result are the semantic foci of the English Accomplishment (Tai 1984, Smith 1997). 5.

(6) are presented in the following sections. 1.2 Theoretical Background One of the main purposes of the present study is to examine how the role of the learners’ first language (L1) plays in the second language (L2) acquisition of Mandarin RVCs. Two theoretical perspectives of the L1 influence on L2 acquisition are concerned in this study, namely, language transfer and linguistic relativity. The former, presented in 1.2.1, is concerned with cross-linguistic differences in structure and the latter, introduced in 1.2.2, cross-cultural variations in speakers’ cognition. In the following discussion, we present the theories of language transfer and linguistic relativity. 1.2.1 Theory of Language Transfer The use of the L1 knowledge in the learning of an L2 has been referred to as language transfer. Building on behaviorism, language learning in the theory of language transfer is regarded as a process of habit formation in which learners carry over old verbal habits of the L1, some positive and some negative, to the L2 learning context (Fries 1945). Reinforcing the role of the L1, Lado (1957:57-58) expounds that when a speaker is communicating with another person, it is very unlikely for him/her to think consciously of all grammatical rules. For communicative efficiency, the conversationalist has the operation in the grammatical system reduced to habits.. 6.

(7) It appears that in the behaviorist-based theory of language transfer, the learner’s L1 is a determinative factor in second language acquisition (SLA). The L1-L2 interrelationship is specified in the contrastive analysis hypothesis, which advocates that cross-linguistic differences result in obstacles to L2 learning (i.e. negative transfer), whereas similarities lead to learning facilitation or positive transfer (see Fries 1949 and Lado 1957). Put differently, ease or difficulty in SLA can be attributed to L1 facilitation and L1 interference, respectively. The contrastive analysis hypothesis, renowned for its power of prediction, has been supported by many applied linguists in that it helps account for how a linguistic item in one language is transferred to another (e.g., Catford 1983, Faerch and Kasper 1986, Schachter 1992). Apart from the predicative power, proponents of language transfer hold that in language development, less proficient learners are more likely to depend on L1 transfer than proficient learners (e.g., Odlin 1994, Kellerman 1984, Faerch and Kasper 1986). As Faerch and Kasper (1986) explain, language transfer is a problem-solving strategy, with which the learner deals with L2 production problems when relevant target rules or structures are not available or temporarily inaccessible. In a sense, transfer is used as a strategy for learners with low L2 proficiency to reduce the learning burden and to achieve the communicative efficiency at the same time when they face the demands of communication.. 7.

(8) 1.2.2 Theory of Linguistic Relativity Apart from the cross-linguistic difference in structure, the L1 effect on SLA can be investigated from the perspective of cross-cultural variations in cognition. This area of research, not observable via an examination of the ‘surface’ structure of sentences and lexicons, opens up the possibility that the L1 influence can be explored from a ‘deep’ level of the conceptual structure. The interrelationship between language and human conceptual content is at the heart of discussion in the theory of linguistic relativity. 6 Whorf (1956) suggests that the way we perceive the universe varies with the language we speak. What is more, he proposes that our thinking is shaped by our language. The postulation of the interplay between language and thought in the Whorfian hypothesis is shown below: It was found that background linguistic system (in other words, the grammar) of each language is not merely a reproducing instrument for voicing ideas but rather is itself the shaper of ideas, the program and guide for the individual’s mental activity, for the analysis of impressions…. Whorf (1956: 214). The Whorfian hypothesis can be interpreted in two ways, the strong version and the weak version. While the strong version, claiming that language determines or controls the way we think, has been proven to be untenable and thus criticized frequently, arguments for the weak version, stating that the grammar of a language influences both thought and perception, have been gaining acceptability in linguistics (Connor. 6. This idea was first introduced by Franz Boas and then passed down to Edward Sapir and then to Benjamin Whorf (See M. Aronoff and Janie Rees-Miller (2001)). 8.

(9) 1996). Results of these studies have revealed that structural differences in language sometimes reflect variations in thought patterns (e.g., Bloom 1981, 1984, Liu 1985, Tai 2003). 1.3 Purposes of This Study The purpose of the current study is three-fold. The three specific purposes are given in the following. The first research purpose is to find out whether learners’ performance varies with the three RVC types--State-Result, Activity-Result and Semelfactive-Result RVCs. The second purpose is to discover how much influence learners’ L1 can have on the L2 learning of Mandarin RVCs. In view of the cross-linguistic variations between Mandarin RVCs and English Accomplishments, we examine if English learners would appeal to English Accomplishments (or Achievements) for similar meaning when learning the characteristics of RVCs. The third concern of this research is the interplay between language and thought. Specifically, it intends to find out if English-speaking people would attend more to the action of the event in comparison to Mandarin-speaking people, who pay more attention to the result of the event, claimed by Tai’s (2003). 1.4 Organization of the Thesis. 9.

(10) This thesis is organized into the following chapters. Chapter 1 provides a brief overview of the entire thesis. In Chapter 2, we first examine the characteristics of RVCs. Then, we investigate the similarities and differences between RVCs and English Accomplishments/Achievements in terms of these characteristics. Finally, we look back at previous RVC studies in second language acquisition (SLA). Chapter 3 presents the methodology of this thesis. Chapter 4 is dedicated to the presentation of research findings and the discussion of the results. Chapter 5 is the concluding chapter that offers not only a brief summary of the results, but also the limitations of the study and suggestions for future research.. 10.

(11) Chapter Two Literature Review This chapter reviews the linguistic characteristics and empirical studies of Mandarin resultative verb compounds (RVCs). Prior to the presentation of the linguistic properties of RVCs, we examine the linguistic features Smith (1997) employs to distinguish five types of situations (or types of verbs in Vendler’s (1967) term) in Section 2.1. The background information of Smith’s verb classification sheds much light on our analyses of RVCs and hence is necessary and helpful in reading certain sections. In Section 2.2, we first describe the general characteristics of the RVC in terms of its component parts and function. Then, we introduce a classification of RVCs, made according to the semantic properties of their constituents. Finally, we examine the grammatical characteristics of RVCs, in particular, the incompatibility between the RVC and the three linguistic elements--the imperfective aspect marker zhengzai, the aspectual verbs kaishi ‘begin’ and tingzhi ‘stop’--respectively, and the interpretation of RVC sentences with and without the adverb chayidianr ‘almost’. In Section 2.3, we discuss Tai’s (1984, 2003) and He’s (1992) analyses of RVCs, which bring insightful information to our study. Moreover, one of the major purposes of this study is to investigate the first language (L1) influence on the second language (L2) acquisition of Mandarin RVCs. Therefore, in Section 2.4 the RVCs are. 11.

(12) systematically compared with English Accomplishments, Achievements and States with respect to the grammatical characteristics presented in Section 2.2 to find out if there are similarities and/or differences between the two languages in terms of these situation types. Section 2.5 looks at studies on how Mandarin RVCs are acquired by L2 learners of different L1 backgrounds. A summary of this chapter is given in Section 2.6. 2.1 Situation Types in Smith (1997) Smith (1997) identifies five situation types, 7 namely States, Activities, Accomplishments, Achievements and Semelfactives, with the temporal features, Static, Durative, and Telic. Table 1 summarizes the situation types and their temporal schemata in binary terms. Table 2-1.. Features of the situation types. Situations. Static. Durative. Telic. State Activity Accomplishment Semelfactive Achievement. [+] [—] [—] [—] [—]. [+] [+] [+] [—] [—]. [—8] [—] [+] [—] [+]. 7. The notion of situation type is not the same as that of verb class. In Vendler (1957, 1967), verbs are categorized according to the semantic properties of the main verb in a sentence. However, Smith (1997) argues that such constituents as subjects, arguments and adverbials also have contribution a lot to the aspectual meaning of a given sentence. Thus, Smith uses the term ‘situation type’ instead of the term ‘verb class’. Note also that the distinctions between situation type and verb type are beyond the scope of this thesis. For ease of the cross-linguistic comparison between the Mandarin RVC and the English Accomplishment/Achievement/State, we will use the term verb class instead of situation type in most sections. 8 The feature of telicity is irrelevant to States in that it is used to distinguish event verbs with a final endpoint from those without an endpoint (Smith 1997). 12.

(13) (Smith 1997:20) As Table 1 shows, the three temporal features capture the distinction among the five situation types. The feature Static makes stative verbs distinct from event verbs. The feature Durative separates durative verbs from instantaneous verbs. Telicity, on the other hand, distinguishes verbs denoting events with a natural final endpoint (i.e., telic events) from those with an arbitrary endpoint (i.e., atelic events). In other words, a telic event involves a change-of-state, which constitutes the outcome or goal of the event, while an atelic event has an arbitrary endpoint and thus can stop at any time (Smith 1997:19). These five situation types have their own semantic properties.9 Verbs like know, stand and be sick belong to States. These verbs present stative situations with no dynamicity and no internal structure. Activities denoted by verbs such as du ‘read’, he ‘drink’ and xue ‘study’ are events with an unbounded process. Accomplishments name the events that direct towards a natural endpoint. Examples of Accomplishments are read an article, drink a glass of water and paint a picture. These examples show that English Accomplishments consist of both a process and a result. Achievements, unlike Accomplishments representing a durative telic event, represent an instantaneous telic event. Events denoted by verbs or verb phrases like recognize, win a race, lose and 9. Smith (1997:39) states that the temporal features [Static], [Durative] and [Telic] can be linguistically realized by certain syntactic and semantic properties. We review only the properties that are related to our study. The semantic properties of the five situation types are presented in 2.1 and the syntactic properties in 2.4. For further discussion, see Smith (1997). 13.

(14) reach the top, which name single-stage events with a final endpoint, are Achievements. Intriguingly, Achievements are semantically instantaneous, but the event of Achievement may have a preliminary stage10 associated with it, as in (1): (1) John won the race. Sentence (1) with the Achievement win a race presents an event as instantaneous, but the event is allowed to occur with some preliminary running process (stage). It is crucial to emphasize that the preliminary process is detachable from the event of Achievement. According to Smith (1997), the preliminary process is not at all a necessary component of the Achievement event. For instance, the Achievement lose characterizes an event that does not require any preliminary losing process. Smith suggests that in the case where the Achievement event occurs with the preliminary stage, the preliminary merely functions to enable the event of Achievement to take place. One important point to be drawn from the review above is that Accomplishments and Achievements can be distinguished on the basis of whether or not they have the non-detachable process in their meaning: while the Accomplishment includes a non-detachable process as part of its meaning and a result, the Achievement may include a detachable process in addition to an outcome. The temporal feature, non-detachability, which associates the process of an event with its outcome, can be 10. Preliminary stages are processes before the attainment of the goal in an Achievement event. 14.

(15) evidenced in the test with the adverb almost, as in the pair of sentences in (2). (2) a. John almost closed the door.. (Accomplishment). b. John almost won the race.. (Achievement) (Smith 1997:44). The Accomplishment sentence (2a) and the Achievement sentence (2b) are interpreted differently with the adverb almost In (2a), the Accomplishment close the door, consisting of both a (non-detachable) process and a result, is ambiguous and can be interpreted in two ways: one is the process interpretation (i.e., John intended to close the door, but somehow changed his mind and did nothing at all) and the other, the result interpretation (i.e., John was pushing the door to close it but he didn’t succeed in closing it.). In contrast, the ambiguity with almost does not appear in the Achievement sentence (2b). The reason is that the preliminary process that occurs with the event of Achievement is detachable and not considered part of the event; therefore, (2b) has only the result reading, that is, John was running to win the race, but he never actually won the race. The last situation type is the Semelfactive. As illustrated in Table 2-1, Semelfactives represent single-stage events, with no final endpoint, as denoted by verbs like blink, cough, and kick. Typical Semelfactives occur very quickly, so they are conceptualized as instantaneous events. In addition, Semelfactives often occur iteratively, as in (3): 15.

(16) (3) Mary knocked for five minutes. (Smith 1997:29) It is clear that Sentence (3) has a multiple-event interpretation, that is, the action of knocking occurs in sequence (Smith 1997). As Smith (1997:30) puts it, this interpretation is triggered by the incompatibility of the instantaneous meaning of the Semelfactive and the durative adverbial for five minutes. It is worthwhile to point out that Semelfactives and Achievements, sharing the feature of instantaneity differ in one important aspect. That is, while the Achievement names an event with a final endpoint, the Semelfactive presents an event without an endpoint. 2.2 General Characterization of Mandarin RVCs Literature on the characteristics of RVCs is abundant in Chinese linguistics. It has been shown that RVCs have several important features. For example, the Mandarin RVCs are productive, suggesting that we can freely create RVCs as long as the semantic relationship between the constituents makes sense to the speaker and the context (e.g., Chao 1968, Li and Thompson 1981, Shi 2003). For example, we can make a handful of RVCs by placing the verb chi ‘eat’, which denotes an action, in the first constituent, as shown in chi-bao ‘eat-full: to become full after eating’, chi-guang ‘eat-empty: to eat up some food’ and chi-ni ‘eat-monotonous: someone eats certain food many times and s/he does not feel like eating the food anymore’. We can also form a large number of RVCs by using the resultative morpheme ni ‘monotonous’ as 16.

(17) V2, as in wan-ni ‘play-monotonous: something is uninteresting to someone because s/he has played it for a long time,’ and kan-ni ‘see-monotonous: someone has seen something many times; as a result, s/he gets bored with seeing the same thing’. The RVC can also be characterized in terms of the form and function of the elements that comprise it. An RVC is made up of two verbs, the second of which serves the function of signifying some result of the action or state signified by the first verb (e.g., Chao 1968, Li and Thompson 1981, Tai 1984 and Liu 1996). In other words, the elements of RVCs can either denote the action-result or state-result causal relationship. For example, the second morpheme si ‘die’ of the RVC da-si ‘hit-die’ indicates the result of the action of ‘hitting’ conveyed by the first morpheme da ‘hit’. On the other hand, the RVC re-yun ‘hot-dizzy’ is formed with a stative verb and a resultative morpheme.11 A state-result relationship can still be found between the states denoted by the two constituents of this RVC. However, not all verbs in Mandarin Chinese can be components of the RVC. Hence, the aims of this section are to find out possible candidates (i.e., verbs) that may serve as the constituents of the RVCs and to explore their basic linguistic characteristics. In Section 2.2.1, we present the verbs that may constitute an RVC and classify RVCs into three categories according to the semantic properties of the two. 11. Because the morpheme in the postverbal position generally has the meaning of the result, we will call this morpheme the resultative morpheme. 17.

(18) constituents-- Activity-Result, Semelfactive-Result and State-Result RVCs. In Section 2.2.2, we explore the grammatical features of RVCs. 2.2.1 Components of Mandarin RVCs and RVC Classification This section discusses the morphemes (i.e., verbs) that may appear as components of Mandarin RVCs based on Smith’s classification of verbs.. As already noted, an RVC consists of two elements, with the second signifying the result of the action or state denoted by the first. The two component morphemes of RVCs can be said to denote a special semantic relationship, as schematized in (4): (4). V1. V2. －. action or state. result. The following examples are illustrations: (5) a. da-si b. da-yun. ‘hit-die’ = to cause someone to die by hitting him/her ‘hit-dizzy’ = to cause someone to feel dizzy by hitting him/her. (6) a. ting-dong. ‘listen-understand’ = to listen to something and understand it as a result. b. ting-fan. ‘listen-annoyed’ = to listen to something many times and get annoyed as a result. (7) a. re-hun b. zuo-ma. ‘hot-faint’ = the hot weather makes someone faint ‘sit-numb’ = one’s leg is numb from sitting too long. (5)-(7) make it clear that the resultative morpheme of RVCs12 functions to signify the. 12. It should be pointed out that the resultative morpheme of an RVC may itself be a compound, as in xi-gan-jing ‘wash-clean’, and xie-qing-chu ‘write-clear’, in which the action is underlined once, and the result is underlined twice. 18.

(19) result of the action or the state conveyed by the verb in the first constituent. In addition, only Semelfactives, Activities and States, which do not include a result (or a final endpoint) as part of their meaning (Smith 1997) can function as the first component of RVCs. As exemplified in (5a) and (6a), the Semelfactive da ‘hit’ and Activity ting ‘listen’, can combine with the resultative morphemes si ‘die’ and dong ‘understand’ to form the RVCs da-si and ting-dong, respectively. Achievements13 like si ‘die’, however, cannot serve as the first component of the RVC perhaps because they already have a result/final endpoint inherently incorporated in their meaning. Based on these, the RVC compounding process of linking an event verb and a resultative morpheme can be said to have the semantic effect of assigning to the atelic event, an Activity or a Semelfactive, a final endpoint. Similarly, the RVC compounding process of linking a stative verb and a resultative morpheme adds a final endpoint to a stative situation. As for the V2 position of the RVC, adjectives are the most frequently seen, e.g., the adjectives hong ‘red’ and ya ‘hoarse’ in the RVCs qi-hong ‘paint-red’ and han-ya ‘roar-hoarse’. This is evident in Ma and Lu’s (1997) study. In the investigation of the morphemes that comprise the RVCs, Ma and Lu discover that 958 out of the 1078 adjectives in a dictionary are morphemes that may appear as the second component 13. The discussion on the possible verbs that may serve as the first constituent of an RVC excludes Accomplishments for the reason that Mandarin does not have the category of accomplishment verbs, as Tai (1984) suggests. Tai’s view that Mandarin lacks Accomplishments will be discussed in detail in Section 2.3.1. 19.

(20) part of the RVC. In addition to adjectives, verbs may also serve as the second element in RVCs. However, the number of verbs that is eligible for the second constituent of an RVC is small. As reported in Chang’s (1999:162) study, only seventeen verbs (e.g. dao ‘arrive’, jian ‘perceive’, dong ‘understand’, wan ‘finish’, si ‘die’, diao ‘fall’, duan ‘break’, etc.) are possible candidates for the second component parts of RVCs. More important, most of the verbs are Achievements. According to the discussion above, the Mandarin RVCs can be made up of an Activity, a Semelfactive, or a State as the first component and a resultative morpheme as the second component. Based on the semantic properties of the two constituents, Mandarin. RVCs. can. be. divided. into. three. types--. Activity-Result,. Semelfactive-Result and State-Result RVCs examined below. (A) Activity-Result (Act-R) RVCs Activity-Result RVCs can be schematized as in (8): (8). V1. V2. －. Activity verb. resultative morpheme. The underlined part of (9) gives an example of Activity-Result RVCs: the Activity verb is underlined once, and the resultative morpheme is underlined twice. (9). Zhangsan. kao-gan-le. natiao. maojien.. Zhangsan. roast-dry-LE. that -CL. towel. ‘Zhangsan dried the towel by roasting it on the fire.’ 20.

(21) As schematized in (8), an Activity and a resultative morpheme comprise the Act-R RVC. RVCs in this class have the action-result semantic relation. For example, in the Act-R RVC kao-gan in (9), kao ‘roast’ is an Activity verb which inherently denotes an uncompleted action, and gan ‘dry’ indicates the result of the roasting action. (B) Semelfactive-Result (Sem-R) RVCs Schematically, we can represent Semelfactive-Result RVCs as follows: (10). V1. V2. －. Semelfactive verb. resultative morpheme. Sem-R RVCs, as illustrated in Schema (10), are formed with a Semelfactive and a resultative morpheme. As reviewed in Section 2.1, Semelfactives, having the temporal features [-Static], [-Durative] and [-Telic], can name either single stage events, or multiple-event activities. When a Semelfactive is part of an RVC, it follows that a variation involving the interpretation of the RVC is possible. The difference between the two interpretations can be nicely illustrated in Example (11): (11) Xiaoming Xiaoming. ti-dao-le. yige. xiangzi.. kick-fall-LE. one CL. box. ‘Xiaoming kicked down one box.’ Depending on the context given for it, the event may either be a single-stage event or a multiple-event activity, both of which are conveyed by the Semelfactive ti ‘kick’. The two possible interpretations are illustrated with the following two speech 21.

(22) contexts. One natural context for this sentence would be one in which the subject Xiaoming is asked to kick down as many empty boxes as he can to win a game of box kicking. The speaker would utter (11) to let the audience know the number of boxes the competitor kicks down at the time of speaking. Because the action of kicking takes a fraction of a second to occur, it is considered a single-stage event. The other appropriate context for (11) would be like this: suppose that Xiaoming is required to kick down many ‘heavy’ boxes in a minute to win the game. The speaker uses Sentence (11) to inform the audience the same information. It is worth noting that the action of kicking, in this case, occurs in sequences. In other words, the Semelfactive ti ‘kick’ presents multiple-event activities instead of a single-stage event. Similar to Act-R RVCs, Sem-R RVCs have the action-result relation between the meanings of their constituents and can thus name a telic event that involves the feature of instantaneity. (C) State-Result (Sta-R) RVCs The State-Result RVCs can be represented by the following schema:. V1. (12). Stative verb. V2. －. resultative morpheme. For example: (13) a.. zuo-ma. ‘sit-numb’. zhan-teng. ‘stand-hurt’ 22.

(23) b.14. re- yun. ‘hot- dizzy’. qi- hun. ‘angry- faint’. Schema (12) illustrates that the Sta-R RVC is made up of two morphemes: the morpheme in the V1 position is a stative verb and that in V2 is a resultative morpheme. The stative verb can be either an ordinary stative verb such as zuo ‘sit’ and zhan ‘stand’ in (13a), or an adjectival stative verb (an adjective) like the underlined morphemes in (13b). No matter whether the first constituent is a stative verb or an adjectival stative verb, it functions to signify a static interval that causes a new, changed state signified by the resultative morpheme to occur. That is, there is a causal relationship between the two states indicated by the elements in the Sta-R RVC, as in: (14) Yuehan John. tang-huai-le. zhezhang. chuang.. lie-broken-LE. this. CL. bed. ‘John lay on this bed, and this bed got broken as a result.’ (15) Hanleng cold. de. tianqi. DE. weather. leng-si-le. nage. laoren.. cold-dead-LE. that. CL. old. man. ‘That old man could not stand the cold weather, and he died as a result.’ From sentences (14) and (15), we can see that the meanings of the elements in the Sta-R RVC indicate a state-result relationship, a causal relationship. Since as He 14. Li and Thompson (1981:66) note an interesting phenomenon that characterizes the use of State-Result RVCs: they do not often convey their literal meanings; rather, their resultative morphemes are used in a metaphorical sense. They further explain that whether the RVCs convey their literal or metaphorical meanings can be determined by the pragmatic context. This implies that, depending on the context, an RVC like qi-hun ‘angry-faint’ in (13b) can mean that someone was so angry that s/he almost fainted, as well as the literal meaning ‘someone made a person so angry that the person fainted as a result.’ Since the metaphorical meaning of the RVC is of little relevance to our topic, throughout this thesis we are referring to their literal meaning. 23.

(24) (1992) and Tai (2003) suggest, the result component constitutes the semantic focus of the RVC, (14) with the Sta-R RVC tang-huai ‘lie-broken’ can be used to announce a new state of affair--this bed is broken. And this new state is related to an earlier state-John lay on this bed previously, expressed by the verb tang ‘lie’. By the same token, the RVC leng-si ‘cold-die’ in (15) is indicative of a new state--the old man is dead now. As we might expect, this newly noticed state is associated with the static interval when the old man felt cold, denoted by the verb leng ‘cold’. 2.2.2 Grammatical Properties of RVCs From the characterization of RVCs presented above, the Act-R and Sem-R RVCs that have the action-result semantic relation between the constituents can be characterized with the temporal features, [-Static], [-Durative] and [+Telic]. Accordingly, these RVCs denote instantaneous, telic events. This implies that the action-result RVCs do not have the feature of duration. The Sta-R RVCs present a state with an endpoint, but they possess the same semantic characteristics as do the other types of RVCs. With these features in mind, we examine the grammatical properties of RVCs in sections 2.2.2.1 and 2.2.2.2. The properties of RVCs that will be discussed in 2.2.2.1 are: their incompatibility with the Mandarin imperfective aspect marker zhengzai, the inceptive verb kaishi ‘begin’ and the terminative verb tingzhi ‘stop’. The interpretation of the RVC sentence and that of the RVC sentence. 24.

(25) that contained the adverb chayidianr ‘almost’ are presented in 2.2.2.2. 2.2.2.1 Incompatibility with Durative Linguistic Markers Zhengzai, Kaishi and Tingzhi The vast majority of RVCs — Act-R and Sem-R RVCs, appear to be like English Accomplishments consisting of a process (or an action 15 ) and a resultant state. However, they do not occur with the imperfective aspect, unlike English Accomplishments, as in (16) and (17): (16). *Ta he. zhengzai. gai-hao. yijian. fangzi.. imperfective Asp.. build-complete task. one CL. house. ‘He is building a house.’ (17). *Ta he. zhengzai. tiao-huai. nazhang. imperfective Asp.. jump-broken. that CL. chuang. bed. ‘He is breaking that bed by jumping on it.’ The reason why the Act-R RVC gai-hao in (16) and Sem-R RVC tiao-huai ‘jump-broken’ in (17) are not compatible with the Mandarin durative marker zhengzai is straightforward: since the RVCs lack the feature of duration and denote instantaneous single-stage events. Thus, Act-R and Sem-R RVCs, Sta-R RVCs are incompatible with the progressive marker zhengzai, either, as shown in (18): (18) *Wangwu Wangwu. zhengzai. lei-bing. shenti.. imperfective Asp.. tire-sick. body. ‘*Wangwu is tiring out.’. 15. It refers to a temporally unbounded action. 25.

(26) From Sentence (18), we can see that the Sta-R RVCs, which present a state with an end point, are odd with the imperfective aspect marker zhengzai. Because the Sta-R RVC lack the feature of duration, they do not occur with the imperfective marker zhengzai, which is only compatible with durative verbs. Here are other examples that demonstrate this property of RVCs very clearly: (19) *Lisi Lishi. kaishi. tui-kai. nashan. men.. start. push-open. that. CL. door. ‘Lisi started to push that door open.’ (20) *Xiaoming Xiaoming. kaishi. reng-po. tade. sujiao. bei.. start. throw-broken. his. plastic. cup. ‘*Xiaoming started to break his plastic cup by throwing it onto the ground.’ (21) * Zhangsan Zhangsan. kaishi. zuo-huai. nazhang. yizi.. start. sit-broken. that. CL. chair. ‘*Zhangsan started to sit on that chair and as a result, his leg became numb.’ As indicated in (19)-(21), an RVC is not allowed to go with the aspectual verb kaishi ‘begin’, which signals the inception of an event. Kaishi, according to Ma (2005), selects a process or a state as its complement. The Act-R RVC tui-kai ‘push-open’ in (19) and Sem-R RVC reng-po ‘throw-broken’ in (20) are verbs involving the feature of instantaneity, not verbs with a process; hence, they cannot appear with the aspectual verb kaishi. Neither can the Sta-R RVC zuo-huai ‘sit-broken’ in (21) occur with the verb kaishi in that this class of RVCs does not. 26.

(27) represent a state. Rather, it names a different kind of state, a state with a final endpoint. Similar to the inceptive verb kaishi ‘begin’, the terminative verb tingzhi ‘stop’ selects a process as its complement (Ma 2005). Now that RVCs exhibit the feature of instantaneity, they are not compatible with the verb tingzhi ‘stop’. The following examples are illustrations of this point: (22). *Meimei sister. tingzhi. ku-shi. zhentou.. stop. cry-wet. pillow. (Act-R RVC). ‘My younger sister stopped wetting the pillow by crying.’ (23). * Ta. tingzhi. qiao-po. stop. knock-broken. he. huaping.. (Sem-R RVC). vase. ‘*He stopped breaking the vase by knocking on it.’ (24). *Tamen. tingzhi. zhu-guan. xiao. fangjian. (Sta-R RVC). they. stop. live-accustomed. little. room. ‘* They stopped getting accustomed to living in a small room.’ The above examples clearly show that RVCs of the three types do not occur with the terminative verb tingzhi ‘stop’ in that they are not process verbs. 2.2.2.2 The Result Interpretation of the Adverbial Chayidianr ‘almost’ in RVC Sentences We have considered one of the most important grammatical characteristics of RVCs; that is, the RVC presents instantaneous events, so it is not allowed to go with durative linguistic labels. In this respect, Mandarin RVCs are like English 27.

(28) Achievements, not Accomplishments. As noted in Section 2.1, Accomplishments and Achievements can be distinguished on the basis of whether or not they have the non-detachable process in their meaning, and this non-detachability can revealed in the test with the adverb ‘almost’. This section investigates the interpretation of the RVC sentences with the adverb chayidianr ‘almost’ to see whether the process (or temporally unbounded action) denoted by the first constituent of the RVC is a non-detachable process, or detachable process. Recall that Accomplishment sentences with almost have both the action and result interpretations, while such ambiguity does not appear in Achievement sentences with almost. If the Act-R and Sem-R RVCs are syntactically similar to Achievements, then, the ambiguity in interpretation will not appear. For example: (25). Yisheng doctor. chayidianr almost. jiu-huo. Mali.. save-alive. Mary. (Act-R RVC). ‘The doctor almost saved Mary.’ (26). Huairen. chayidianr. kan-duan. tade. jiao.. hooligan. almost. cut-broken. her. foot. (Sem-R RVC). ‘The hooligan almost cut off her foot.’ An examination of the above sentences yields the observation that RVC sentences with chayidianr ‘almost’ only have the result reading. To illustrate, (25) with the Act-R RVC jiu-huo ‘save-alive’ has only the result reading — the doctor was trying to save Mary, but he did not succeed. By the same token, (26) with the Sem-R 28.

(29) RVC kan-duan ‘cut-broken’ only has the result interpretation — the hooligan was cutting her foot, but he didn’t cut it off. In addition to Act-R and Sem-R RVCs, the Sta-R RVCs exhibit the same semantic property, as in: (27) Ta he. chayidianr. zuo-ma. almost. sit-numb. tade. tui.. his. leg. (Sta-R RVC). ‘His leg almost became numb resulting from sitting on a chair for a long period of time.’ In Sentence (27), a sentence with the Sta-R RVC zuo-ma ‘sit-numb’ and the adverb chayidianr ‘almost’, the ambiguity in interpretation does not appear in the. Hence, (27) has the result interpretation that he sat for a long time, but his leg never actually became numb. In Section 2.2, we have presented the internal structure of the RVC with regard to its form, function and semantic relation between the constituents. We have also examined the grammatical properties of RVCs. In the next section, we will review Tai’s (1984, 2003) and He’s (1992) linguistic analyses of the RVCs to gain a more complete understanding of this particular construction. 2.3 Linguistic Studies of RVCs Literature on the linguistic analyses of RVCs is abundant in Chinese linguistics (e.g., Chao 1968, Li and Thompson 1981, Tai 1984, 2003, He 1992, Liu 1996 among others). Among them, some focus on the classification of RVCs, made according to. 29.

(30) the type of result characterized by either their two component parts, or the second part (e.g., Li and Thompson 1981, and Liu 1996) and some pay attention to the semantic characteristics of RVCs such as Tai (1984, 2003) and He (1992). In this section, we will review Tai’s (1984, 2003) and He’s (1992) analyses in turn since the current research puts more emphasis on the semantic properties of RVCs. 2.3.1 Tai (1984) Tai (1984), in his study on the semantic meaning of verbs in Mandarin Chinese, expounds that Mandarin RVCs and English Accomplishments have the action-result semantic relation between their semantic components, but they are different in one important aspect. That is, unlike English Accomplishments, which encode the aspectual meaning of both the action and result, the RVCs, however, include only the result aspect 16 in their aspectual meaning. To illustrate this point, consider the sentences in (28): (28) a. * Ta he. zhengzai. hua-wan. nazhang. hua.. imperfective Asp.. paint-finish. that CL. picture. b. He is painting that picture.’ As can be seen from Example (28a), the RVC hua-wan ‘paint-finish’ is not allowed to go with the imperfective marker zhengzai. The reason is simply that RVCs do not encode the aspectual meaning of the action, though they may have the 16. The terms ‘action aspect’ and ‘result aspect’ are used by Tai (1984). In the RVC construction, the action aspect confers the aspectual meaning of the action verb, and the result aspect, that of the result component. 30.

(31) action-result semantic relation (Tai 1984). Without the aspectual meaning of action, RVCs lack the feature of duration. Hence, they are incompatible with durative linguistic labels such as the imperfective marker zhengzai. Contrarily, the aspectual meaning of the English Accomplishment includes both the action aspect and result aspect, and hence the Accomplishment is compatible with the English progressive, as in (28b). On the basis of the contrast between RVCs and Accomplishments, it is not appropriate to treat RVCs as Accomplishments. There are three points about Tai’s (1984) analysis of Mandarin RVCs that are worth noting. First, RVCs denote instantaneous events on account of the notion that the time schema for RVCs does not have continuous tenses (Tai 1984). This supports our earlier analysis that RVCs denote instantaneous events. Second, in contrast to Tai’s study, in some work attempting to investigate the semantic structure of Mandarin verbs, RVCs were taken as accomplishment verbs by mistake (i.e., Teng 1985 and Smith 1997). Tai’s analysis helps us pinpoint where those linguists go wrong in their research: they fail to recognize that the aspectual meaning of RVCs excludes the action aspect. In other words, they do not realize that Mandarin RVCs and English Accomplishments are different in a fundamental manner: while Accomplishments are semantically durative, RVCs are semantically instantaneous. Third, the cross-linguistic variations between Mandarin RVCs and English. 31.

(32) Accomplishments prompt Tai (1984) to contend that Mandarin exhibits three types of verbs, namely States, Activities and Results. The verb category of Results, according to Tai (1984:295), is mostly expressed with the Resultative Verb Compounds (RVCs). The author further suggests that among the four categories of verbs Vendler (1967) identifies-- States, Activities, Accomplishments and Achievements, Achievements and Accomplishments are realized in Mandarin in the form of RVC. For example, the Mandarin counterparts of the Achievement verbs17 find and hear are RVCs, namely ‘zhao-dao seek-reach’ and ‘ting-jian listen-perceive’, respectively. Apart from the Achievement, Tai (1984:290-291) reports that the English Accomplishment verb may be expressed with the RVC in Mandarin, such learn, corresponding to as in xue-hui ‘study-understand’. However, Tai (1984) claims that RVCs are not comparable to English Accomplishments since there is one notable difference between the semantic focus of the Mandarin RVC and that of the English Accomplishment: while the result part constitutes the semantic focus of the RVC, both the action and the result parts are the semantic foci of the Accomplishment. Or in Tai’s (1984:292) term, the RVC has only the result aspect. Tai thus concludes that Mandarin lacks the category of Accomplishment verbs. One important insight from Tai’s (1984) study is that English lacks a direct. 17. There are Achievements taking the form of a simple verb in Mandarin, as in si ‘die’ (Tai 1984:294). 32.

(33) counterpart of Mandarin RVCs. Although it seems that the English Accomplishment is the closet translation of the Mandarin RVC in that both of them exhibit the action-result semantic property explicitly or implicitly, there is however one crucial cross-linguistic distinction between the RVC and Accomplishment. To illustrate this point, consider the sentences in (29): (29) a.. Ta. sha-si-le. Lisi.. he. kill-dead-LE. Lisi. b. He killed Lisi. (29b) is the translation equivalent of (29a); however, there is a subtle difference between (29a) and (29b). The difference lies in how the completion/result of a closed event is expressed in Mandarin and English. In Mandarin, the notion of completion is explicitly characterized with the resultative morpheme of the RVC, while in English the specification of completion is inherent in the meaning of the main verb (i.e., kill) in a perfective sentence. Our analysis conforms to Tai’s (2003) that the result of the event is covertly expressed in English, but overtly expressed in Mandarin. In other words, unlike Mandarin, English does not have a clear ‘result-marking system’ which characterizes the result of the event with the resultative morpheme like that of the RVC. 2.3.2 He (1992) and Tai (2003) Based on Tai’s (1984) analysis that RVCs encode only the result aspect in their 33.

(34) aspectual meaning, He (1992), in his investigation on the semantic characteristics of RVCs, maintains that the result component is the semantic prime that confers the central meaning to the RVC. For example: (30) a. Wo I. du-dong-le. neiben. shu.. read-understand-LE. that CL. book. ‘I understood that book through reading.’ b. Wo. du-le. neiben. shu.. I. read-LE. that CL. book. ‘I read that book.’ c. Wo I. dong-le. neiben. shu.. understand-LE. that CL. book. ‘I understood that book.’ (He 1992:122) He (1992) observes that the result predication in (30c) expresses the main idea of (30a) with the RVC du-dong ‘read-understand’. Tai (2003), similar to He (1992), holds that the result part constitutes the semantic focus of the RVC. Tai also points out a cross-linguistic difference between Mandarin and English in structuring events consisting of both an action and a result. That is, Tai suggests that in Mandarin, such events are unequivocally expressed with the compound verbs of RVCs; English, however, unlike Mandarin, which has the consistent action-result schema18 in the. 18. Tai (2003:305) states that like Mandarin, in which the action-result schema are linguistically expressed by the RVC, English has the resultative construction that realizes the action-result schema, but the English resultative construction is structurally different from Mandarin RVCs, as in: (i) a. He kicked the door open. 34.

(35) event description, uses various grammatical patterns to characterize the same events. We can demonstrate this cross-linguistic difference by citing one of Tai’s examples, as in: (31) a.. Ta. jia-cuo-le. laogong. she. marry-wrong-LE. husband. b. She has married the wrong husband. (Tai 2003:304) In (31a) with the RVC jia-cuo ‘marry-wrong’, the mistake expressed with the resultative morpheme cuo ‘wrong’ signifies the result of the action expressed with jia ‘marry’. In contrast, (31b) shows that in the English event description, the mistake signified by wrong modifies the object noun husband. In light of the cross-linguistic difference in structuring the action-result event, Tai states that the result of an event is overtly expressed with the resultative morpheme of RVCs in Mandarin, whereas in English the result is covertly expressed or implied in the meaning of the perfective sentence. On the basis of the cross-linguistic difference between Mandarin and English regarding the expression of the result of an event, Tai claims that the result part of an event would be more salient to Mandarin-speaking people than to. b. He painted the house red. It is also important to point out that although the action-result schema could be directly triggered by the resultative construction in English, English is much less liberal in the use of the resultative construction than Mandarin in the use of RVCs. There are many constraints on the formation of the English resultative. For example, Carrier and Randall (1992) argue that the result predicate is fairly free in terms of category (i.e., an AP, a PP, or an NP), but it does not take every potential result phrase within these categories. For further discussions of the English resultative construction, see Goldberg (1991) and Carrier and Randall (1992). 35.

(36) English-speaking people. In other words, English speakers would attend more to the action part of the event than to the result. Tai’s view that the result is not emphasized in English event description is supported by Hoekstra (1988), who in his study on the English resultative construction expounds that the result predication is a ‘shadow interpretation’ in that it is a cancelable predication, as in (32): (32) They cooked [the chicken dry].19 (Hoekstra 1988:117) Hoekstra’s explanation for the ‘shadow interpretation’ is twofold. First, he observes that there is not a sensible semantic relationship between the verb cook and the postverbal NP the chicken. Rather, the postverbal NP has a sensible semantic relationship with the following predicative expression. Second, Hoekstra points out that the main verb in the English resultative construction does not usually take an object,20 which means that the result predication is merely an adjunct to the verb. This implies that the result predication and the main verb are completely independent of each other. Based on the two accounts, Hoekstra concludes that the focus of (32) is on the main verb, not on the result interpretation. Therefore, the cancellation of the result predication will not change the meaning of the sentence.. 19. The two constituents in the result predicate form a small clause (Hoekstra1988). According to Hoekstra (1988), the main verb in the English resultative construction is usually an intransitive verb, as in: (i) He painted [the door green].. 20. 36.

(37) 2.4 Cross-linguistic Comparisons between Mandarin RVCs and English Accomplishments, Achievements and States As noted in the previous review, Mandarin action-result RVCs21 are similar to English Accomplishments. This study aims to investigate if the L2 learners who are native speakers of English would appeal to English verbs (i.e., Accomplishments, Achievements and States) for similar meaning when learning the characteristics of RVCs. To further explore where the learners’ difficulty may occur, the sections that follow will provide a systematic comparison between RVCs and the above-mentioned English verbs with respect to the three properties--(1) RVCs are incompatibile with the durative linguistic labels of the imperfective aspect marker zhengzai, the aspectual verbs kaishi ‘begin’ and tingzhi ‘stop’; (2) the result part of the RVC functions as the center of predication; and (3) the RVC sentence with the adverb chayidianr ‘almost’ has only the result interpretation. Therefore, Section 2.4.122 discusses the similarities and differences between Mandarin Activity-Result and Semelfactive-Result RVCs and English. Accomplishments. and. Achievements.. Section. 2.4.2. presents. the. cross-linguistic comparisons between Mandarin State-Result RVCs and English States. Section 2.4.3 predicts the learners’ performance of RVCs on the basis of the. 21. By action-result RVCs, we mean the Activity-Result and Semelfactive-Result RVCs, both of which have the action-result semantic relation between the meanings of their constituents. 22 We include only the Activity-Result and Semelfactive-Result RVCs in the cross-linguistic comparison in Section 2.4.1 since both of the RVCs represent events. The State-Result RVCs, which name a state with a final endpoint, are compared with the English States, naming a stative situation, in Section 2.4.2. 37.

(38) cross-linguistic comparisons in 2.4.1 and 2.4.2. 2.4.1 Similarities and Differences between Action-Result RVCs and English Accomplishments/Achievements The action-result RVCs (Actity- and Sem-Result RVCs), having the temporal features of [-Static], [-Durative] and [+Telic], exhibit the following characteristics: (A) They are not compatible with durative linguistic structures. (B) The semantic focus of the RVC is the result component, which implies that the result predication is the center predication of the RVC sentence. (C) In the test with the adverb chayidianr ‘almost’, a sentence with the RVC has only the result interpretation. In the following, we will discus the similarities/differences between Mandarin RVCs and English Accomplishments and Achievements with regard to the three characteristics above. A1. The English progressive The imperfective aspect like the English progressive prototypically spans an interval that is internal to a situation (Smith 1997:73). As mentioned earlier, English Accomplishments are semantically durative, which means that they have internal stages. Hence, they are entirely compatible with the English progressive, as shown in (33). In contrast to Accomplishments, Achievements are semantically instantaneous and therefore, do not normally co-occur with the progressive, as in (34a): 38.

(39) (33) Mary was running to school.. (Accomplishment). (34) a. *John was noticing Tina in a party.(Achievement without the preliminary) b. Mary was reaching the top.. (Achievement with the preliminary). However, some Accomplishment are compatible with the English imperfective aspect, e.g. (34b), running counter to our expectation. The reason why the imperfective is available for some Achievements in English is simply that the imperfective focuses on the preliminary stage (i.e., the interval before the event takes place) of the Achievement. On account of that, (34b) does not indicate that the Achievement event actually takes place. Rather, it presents the preliminary stage of the instantaneous event—the stage before Mary reached the top. A2. The inceptive verb ‘begin’ and the terminative verb ‘stop’ Smith (1997) states that the inceptive and terminative verbs begin and stop, respectively are compatible with durative events, but they cannot go with events that involve the feature of instantaneity, as shown in (35a) vs. (36b) and (35b) vs. (36a): (35) a. Mary began walking to school. b. *The balloon started (began) to burst. (36) a. Sam stopped walking to school. b. * The bomb stopped exploding.. (Accomplishment) (Achievement) (Accomplishment) (Achievement) (Smith 1997:44-47). B. The semantic focus of the Accomplishments and Achievements It is clear from the forgoing discussion in Section 2.1 that the Accomplishment is. 39.

(40) composed of both the action and result. Because the semantic meaning of Accomplishments includes an action and a result, both the action and the result components are the semantic foci of the Accomplishments. As indicated in Section 2.1, Achievements consist of a result only, and thus the result component constitutes the semantic focus of the achievement verb. C. The test with the adverb ‘almost’ It has been shown in Section 2.1 that English Accomplishments, composed of both an action (process) and a result, are ambiguous in the test with the adverb almost and can be interpreted in two ways: one is the action interpretation and the other, the result interpretation. On the contrary, the same ambiguity does not appear in Achievements, having only the result interpretation. The. similarities. and. differences. between. the. action-result. RVCs,. Accomplishments and Achievements are summarized in Table 2-2. Table. 2-2.. Similarities. and. differences. between. action-result. RVCs,. Accomplishments and Achievements Verbs Features of RVCs. 1. presenting an event consisting. RVCs. Accomplishments. Achievements. Yes. Yes. No. Yes. No. Yes. Yes. No. Yes. of both an action and a result 2. incompatible with durative linguistic labels 3. the result part is the semantic. 40.

(41) focus 4. having only the result. Yes. No. Yes. interpretation while co-occurring with almost 2.4.2 Similarities and Differences between State-Result RVCs and English States In Section 2.1, the semantic meaning and temporal features of English States have been presented. A further comparison shows that Mandarin State-Result RVCs and English State verbs are different in both meaning and function; however, the two types of verbs exhibit the grammatical property that they cannot occur with durative linguistic labels. Semantically, English States consist of only a state, and thus function to name a stative situation. As shown in Section 2.2.1, State-Result RVCs have the state-result semantic relation between their comportment parts; therefore, this type of RVCs can serve the function of denoting a stative situation which causes a change of state—i.e., a result. Interestingly, the Sta-R RVCs have the same grammatical properties as do the Act-R and Sem-R RVCs. With respect to the grammatical characteristics, States in English, according to Smith (1997), cannot occur with durative linguistic labels like the progressive, the verb begin and the verb stop, as in (37), (38) and (39), respectively: (37) *John is knowing the answer. (38) * Sam began to be angry.. 41.

(42) (39) *Mary stopped being sick. (Smith 1997 74, 47) In respect to feature of the semantic focus, the English States consist of only a state, so the state is the focus in their meaning. Notice that the test with the adverb almost does not co-occur with stative verbs in English, as shown in (40): (40) *Mary almost loved John. Table 2-3 displays similarities and differences between the Sta-R RVCs and English States. Table 2-3. Similarities and differences between state-result RVCs and English States Verbs Features of RVCs. 1. describing a stative situation. RVCs. States. Yes. No. Yes. Yes. Yes. No. Yes. N.A.. which causes a change of state 2. incompatible with durative linguistic labels 3. the result part is the semantic focus 4. having the result interpretation while co-occurring with almost 2.4.3 Predictions of the Learners’ Performance Based on the cross-linguistic comparison between Mandarin RVCs and English verbs, we make predictions on the type(s) of RVCs and the linguistic features of 42.

(43) RVCs that may cause confusion among English learners. As already mentioned, Sta-R RVCs differ from the English States in meaning. Act-R and Sem-R RVCs, which have the. action-result. semantic. relation,. are. semantically. similar. to. English. Accomplishments in that the two kinds of verbs select both an action and a result as their semantic components. However, these RVCs resemble Achievements in their grammatical properties-- (1) they are incompatible with durative linguistic labels; (2) the result part functions as the semantic prime; and (3) an RVC/Achievement sentence with the adverb chayidianr ‘almost’ has only the result interpretation. Thus, we predict that the learners’ performance on the Sta-R RVCs would be different from those on the Act-R and Sem-R RVCs. Besides, in light of the similarities and differences between Action-Result RVCs 23 and Achievements/Accomplishments presented above, we predict that the English learners are liable to transfer what they have perceived in the English verbs to the new forms: there will be a negative transfer if they appeal to Accomplishments when learning the grammatical properties of the action-result RVCs; a positive transfer will take place if they treat such RVCs as Achievements. Furthermore, because of the possible interference from the learners’ L1, we predict that the learners’ performance on the Sta-R RVCs will be better than that on. 23. Activity-Result and Semelfactive-Result RVCs are Action-Result RVCs. 43.

(44) the Act-R RVC/Sem-R RVC. As noted already, the L1-L2 structural similarities and differences may lead to positive/negative L1 transfer, and the L2 learners are likely to carry over their L1 knowledge into the L2 learning of Mandarin RVCs. On the basis of the theory of language transfer, we predict that the L1 effect will be mitigated by the learners’ L2 proficiency. 2.5 Empirical Studies on Mandarin RVCs There has been considerable research devoted to analyzing linguistic characteristics of RVCs. However, only a few studies have been done on RVCs in the field of SLA. Two most recent studies are Guo (2003) and Chen (2004), reviewed below. Guo (2003), in the investigation of how Japanese learners of Mandarin Chinese would acquire RVCs, finds out that the lack of structural correspondence between Japanese and Mandarin may pose some problems for Japanese-speaking learners in learning Mandarin RVCs. The learning difficulty is exemplified in (41): (41) * Guo laoshi. de. Guo teacher. DE. Li laoshi. de. Li teacher. De. zuoye. wan-le,. homework zuoye homework. finish-LE, haimei. wan.. negation. finish. ‘I am finished with Miss Guo’s homework, but I haven’t finished Miss Li’s.’ (Guo 2003:75). 44.

(45) As Guo (2003) puts it, in Japanese the result part of an event is more important than the action part, so the action part is often left unexpressed. For instance, in Japanese, [終おつた] ‘finish’ can be used to express the notion of completion for many actions. This may imply that in the description of an event composed of both an action and a result, the Japanese speaker only needs to describe the result part. In contrast, in Mandarin Chinese, the action-result event is expressed with the RVC, formed with an action verb or a stative verb and a resultative morpheme. Due to the difference between Chinese and Japanese in the event expression, the author suggests that to an L2 learner whose L1 lacks a similar compound verb which can be used to describe the action-result event, like Japanese, the learning task of the Mandarin RVC can be overwhelming. Gao’s research, however, is considered insufficient in that the results simply reveal an observation that Japanese learners usually have difficulty producing the action verb in the V1 position in Act-R and Sem-R RVCs. She does not show whether learners’ performance on Sta-R RVC reflects such a tendency. What is more, the author fails to show if the grammatical properties of RVCs, presented in Section 2.2, would pose difficulty in acquisition to the Japanese learners since as our analyses of RVCs indicate, RVCs can be best characterized with these properties. Chen’s (2004) study is a corpus-based. In the study, she concludes that Mandarin. 45.

(46) RVCs that do not appear in the textbooks pose great problems for the L2 English-speaking learners. As an illustration, consider examples (42) and (43): (42) * nage that CL. erzi. nian. quanbu. naben. shu,. son. read. whole. that CL. book. keshi. heishi. kan. bu. dong.. but. still. read. negation. understand. ‘Although that son read the whole book, he still could not understand it.’ (43) * nage. kelian. that CL poor. de. nanshen. zuotian. bei. DE. boy. yesterday. passive. tade. didi. she –le.. he GEN. brother. shoot –LE. ‘That poor boy was shot by his little brother yesterday.’ (Chen 2004:93) As the underlined parts in (42) and (43) suggest, the learners have difficulty learning the RVCs nian-wan ‘read-finish’ and she-si ‘shoot-die’. Chen’s explanation for the learning problem is simply that the learners have not learned the RVCs at the time of testing, so the production such RVCs are difficult for these learners. It is also important to point out that sentences (42) and (43) seem to provide evidence for our analysis on the cross-linguistic difference between the specification of completion in Mandarin and English. As mentioned in 2.3.2, in English, the completion of a closed event is covertly implied in the meaning of the main verb in a perfective sentence; but in Mandarin the notion of completion, in general, is overtly 46.

(47) expressed with the resultative morpheme of the RVC. This has the implication that English L2 learners may have problem learning Mandarin RVCs. In other words, this difference in the specification of completion between Mandarin and English may account for why these English learners produce awkward sentences like (42) and (43). 2.6 Summary In this chapter, the basic properties and previous studies of the RVC construction have been examined. We first investigate the possible candidates that may serve as the components in RVCs and find out that the RVC can be formed with an Activity, a Semelfactive, or a State verb and a resultative morpheme. Then, we divide the RVCs into three distinct categories — Activity-Result, Semelfactive-Result and State-Result RVCs — according to the semantic properties of the constituents. We also present the essential features of RVCs and make cross-linguistic comparisons between RVCs and English verbs with the goal of discovering some similarities and differences between RVCs and English Accomplishment, Achievement and Stative verbs. It is found that the Act-R and Sem-R RVCs are similar to the English Accomplishments in that they both consist of an action and a result. In addition, the comparison reveals that the temporal features of RVCs are the same as those of English Achievements. Specifically,. the (action-result). RVCs and. Achievements are semantically. instantaneous. Because RVCs do not include the action aspect denoting the feature of. 47.

(48) duration in their aspectual meaning, they are incompatible with durative grammatical labels. Moreover, like Achievements, the result component constitutes the semantic focus of the RVCs. Similar to Achievements, which do not include a non-detachable process (or an action) as part of their meaning, the ambiguity with almost does not appear in the RVC/Achievement sentence. As for Sta-R RVCs, they have the same linguistic features as do Act-R and Sem-R RVCs except that the meanings between their constituents indicate a state-result relation. With regard to the empirical research of the RVC construction, Guo (2003) found that Japanese learners of Chinese have difficulty producing RVCs. According to Guo, owing to the fact that in Japanese the action part of a telic event is often left unexpressed, the production of the RVC, consisting of both an action verb and a resultative morpheme, pose a serious problem to the Japanese L2 learners. Moreover, Chen (2004) reports that English learners of Chinese are weak at producing RVCs. She also observes that the learners are less familiar with the RVCs whose component morphemes are new to them than with those whose component morphemes they have learned already. Based on the linguistic properties of RVCs and the similarities and differences between RVCs and English verbs, this study investigates the L2 acquisition of Mandarin RVCs by English L2 learners to see whether they have full understanding of. 48.

(49) the semantic properties of RVCs in general and whether their acquisition varies according to the three RVC types. Moreover, we are interesting in knowing if the learners will carry over their L1 knowledge into the L2 learning of RVCs.. 49.

(50) Chapter Three Methodology This chapter is dedicated to the experimental design of the present study. Section 3.1 describes the subjects in the experiment. The methodology and instrument are introduced in sections 3.2. Section 3.3 presents the procedure of the experiment, including the pilot study, and the type of statistics employed in our data analysis. A summary is given in Section 3.4. 3.1 Subjects A total of 40 English-speaking adults learning Mandarin at the Mandarin Training Center (MTC) of National Taiwan Normal University in Taipei participated in this experiment. The selection of the subjects from the MTC was based on the consideration that they attended Mandarin classes regularly--two hours a day, 10 hours per week with Mandarin-speaking teachers. On the basis of their performance on the MTC placement test, the 40 subjects were evenly (i.e., n=20) divided into two proficiency groups and labeled as Mid and High. Besides, the current study included 20 native speakers of Mandarin who were undergraduates at one of the universities in Taipei. The purpose of recruiting the Chinese speakers was to establish a base line of information or norm which could then be compared with the performance of the two English-speaking groups. Tables 3-1 and 3-2 present the personal information of the. 50.

(51) subjects and language background of the L2 learners, respectively: Table 3-1. Personal information of the subjects Subject. Mid. High. Control. Male. 14. 15. 8. Female. 6. 5. 12. Total. 20. 20. 20. Gender. Table 3-2. L2 learners’ language background Subject. Mid (n=20). High (n=20). Mean. SD. Mean. SD. 1.73. 0.50. 3.5. 1.09. Language background Total length of Chinese learning. As can be seen in Table 3-1, the ratios of gender in the experimental group and the control group were not equal: while the mid group had 14 male and 6 female and the high group 15 male and 5 female, the control group included 8 male and 12 female. The discrepancy was not significant since the focus of this study was on the performance of learners with a difference in their L2 proficiency. In addition, as Table 3-2 indicates, the number of years of exposure to L2 is 1.73 for the mid group and 3.5, for the high group. Though the standard deviation for the L2 learning experience of the high group was larger than that of the mid group, it was considered insignificant because the L2 subjects were chosen on the basis of their overall proficiency in. 51.

(52) Chinese, evaluated by the MTC.24 According to the admission and class offerings of the MTC, our subjects in the mid group had an average of 1 year Chinese learning experience at the MTC and those in the high group 2 years. Further, the L2 learners in the mid group were Beginner-High, Intermediate-Low and Intermediate learners who were studying one of the following books at the time of testing: Practical Audio-Visual Chinese II, Practical Audio-Visual Chinese III, Taiwan Today and Chinese Folk Tales. The subjects in the high group included Intermediate-High, Advanced-Low and Advanced-High learners, and they were studying these books: Mini Radio Plays, News Readings II, Thought and Society: an Advanced Spoken Level Text or Various Chinese Newspapers and/or Periodicals. It is also important to note that the target structure -- the Mandarin RVC -- is introduced late, in Lesson 21 of the learners’ first textbook Practical Audio-Visual Chinese I. To assure the participants’ awareness of the target structure, only those who had finished studying Lesson One in Practical Audio-Visual Chinese II at the time of testing were selected to be the subjects of the present study. 3.2 Methodology and Instrument This study takes the quantitative methodology through the use of objective tasks. 24. Learners at the MTC are usually given a language test on entering the school system. After the placement test, students will be evaluated on their overall proficiency in Chinese and suitable classes will be suggested and arranged. Notice also that the students at the MTC can be classified as Beginner, Beginner-Mid, Beginner-High, Intermediate-Low, Intermediate, Intermediate-High, Advanced-Low and Advanced-High learners. 52.