從期望價值動機觀點建構一語言學習模式

全文

(1)CHAPTER ONE INTRODUCTION 1.1 Theoretical Background Motivation, one cognitive and affective aspect of individual differences, has been widely investigated and commonly agreed as a fundamental factor influencing acquisition of a second or foreign language (L2) (Gardner, 1985; Gardner, Tremblay, & Masgoret, 1997). Studies on L2 motivation date back to Gardner and Lambert’s socio-educational model (Gardner & Lambert, 1972), which has dominated the research of L2 motivation for decades. Recent research, however, has suggested that this model may be problematic and limited in its application to understanding L2 motivation in EFL (English as a foreign language) contexts (Belmechri & Hummel, 1998; Clément, Dörnyei, & Noels, 1994; Warden & Lin, 2000). Accordingly, L2 researchers started to make paradigmatic moves from a social approach emphasizing learners’ social integration into the target language culture to mainstream cognitive approaches which look into learners’ mental evaluation and management. Among these cognitively-oriented explorations of learning motivation, a great number of studies (Bandura, 1993; Clément, et al., 1994; Clément & Kruidenier, 1983; Eccles, 1983; Feather, 1988; Noels, 2001; Noels, Clément, & Pelletier, 2001; Noels, Pelletier, Clément, & Vallerand, 2003; Schunk, 1991; Tremblay & Gardner, 1995; Wigfield, 1994; Wigfield & Eccles, 1992) have addressed expectancy and/or value (E-V) constructs encompassed in the E-V framework proposed by Eccles and Wigfield (Eccles, 1983; Eccles & Wigfield, 1995; Wigfield & Eccles, 1992, 2000). Essentially, this framework postulates that motivation to perform a given task is determined by two critical factors: expectancy of success and value. The former refers to the individual’s evaluation of the probability of attaining successful performances, while the latter is defined as the psychological value a learner attaches to success on. 1.

(2) that task or reasons for doing the specific activity. Although E-V related components have been integrated into the investigation of language learning motivation, it is argued that few empirical studies directly examined this E-V framework in EFL contexts and explored its relationships with motivated learning behaviors or outcomes, such as strategy use and learning achievement. Thus, the present study aims to examine language learning motivation of EFL learners in Taiwan from an expectancy-value perspective and to further investigate its association with language learning behaviors and achievement. The link between L2 motivation and L2 achievement, one of the research foci in SLA (second language acquisition), is commonly believed and, in most cases, has been empirically found to be positively significant (Csizer & Dörnyei, 2005; Dörnyei & Csizer, 2002; Ely, 1986; Gardner, 1985; Gardner & MacIntyre, 1991; Gardner et al., 1997; Noels, Clément, & Pelletier, 1999; Oxford & Shearin, 1994; Samimy & Tabuse, 1992; Schmitt & Watanabe, 2001; Tremblay & Gardner, 1995; Wen, 1997). However, contradictory results have also been indicated in either qualitative (Huang & Chang, 1996) or quantitative studies (Chen, Warden, & Chang, 2005; Teweles, 1995; Shaaban & Ghaith, 2000; Vandergrift, 2005). Thus, it appears that high motivation detected at a certain point of time does not guarantee high L2 achievement in the long run and the impact of motivation on language learning is not so linear and straightforward as expected. That is, motivation to learn a foreign language is necessary and fundamental, but not sufficient. Possessing the wish to learn a language well does not necessarily lead to a successful execution of the learning plan. Learners’ self-regulatory capacity, therefore, is needed to protect the initial intention and to further support the demonstration of motivated learning behaviors such as strategy use (Dörnyei, 2005; Tseng & Schmitt, 2008). As Tseng suggested (2006, p. 4), “motivation is like commitment and 2.

(3) self-regulation is like follow-through, i.e., the subsequent processes to protect the initial set-up goal.” Besides, recent research has argued that the initial level of self-regulatory capacity was affected by the magnitude of motivation (Dörnyei, 2001b; Zimmerman, 2000). It implies that motivation and self-regulation are interwoven and both are required in the process of language learning because motivation influences self-regulation and the execution of self-regulatory capacity guides the manifestation of learning behaviors. Although self-regulation has been argued to be significant in language learning and also have been empirically examined together with motivation in some pioneering studies (Garcia, McCann, Turner, & Roska, 1998; Pintrich & De Groot, 1990; Tseng, Dörnyei, & Schmitt, 2006; Tseng & Schmitt, 2008), it is noted that, in the field of SLA research, learners’ self-regulatory capacity is still a fresh concept and its effects on language learning in general have not been systematically investigated. Thus, empirical studies are required to provide a deeper understanding of the role which self-regulation plays in motivated language learning. Besides protecting the initial motivation, self-regulatory capacity also functions as a mediating power to support learning behaviors such as use of learning strategies. Learning strategy use, the most frequently explored type of motivated learning behaviors (MacIntyre & Noels, 1996; Schmidt & Watanabe, 2001; Tseng & Schmitt, 2008; Vandergrft, 2005), has been discussed with regard to its effects on language learning achievement. Research findings regarding the connection between L2 strategy use and L2 achievement, however, were rather mixed. Some admittedly showed a significant positive link (Park, 1997), some a fairly weak correlation (Nisbet, Tindall, & Arroyo, 2005) and others perplexingly a negative association (Gardner et al., 1997). This inconsistency concerning the predictive power of learning strategy use on 3.

(4) L2 achievement, as suggested by SLA researchers (Gardner et al., 1997; Ellis, 1994; Tseng et al., 2006; Tseng & Schmitt, 2008), resulted from the fact that previous research only investigated the quantity dimension of strategy use, i.e. how many language learning strategies were used and how frequently, and addressed little about the quality dimension of strategy use, i.e. how well language leaning strategies were used. In fact, a larger repertoire of learning tactics and a high frequency of strategy use does not guarantee adequacy and effectiveness in the employment of learning strategies. Besides, several researchers (Gardner et al., 1997; Ellis, 1994) have indicated that the quality dimension of strategy use is crucial to L2 achievement and needs to be explored more for a better understanding of L2 learners’ strategy use. Although strategy use has been a popular issue in SLA research, not many studies made an explicit distinction between quantity-quality dimensions and examined their respective roles. Therefore, it seems necessary to systematically evaluate learners' strategic behaviors based on these two distinct aspects, one on learners’ quantity-based overall involvement in using language learning strategies and the other on learners’ quality-based mastery of specific language learning tactics. Grounded on the theoretical backgrounds discussed above, the present study, from an expectancy-value perspective on language learning motivation, aims to examine the causal relationships between the relevant motivational variables in language learning, namely, choice motivation, self-regulation, learning strategy use and achievement. As for the strategy use variable, both the quantity and the quality dimensions are addressed. Moreover, the study intends to construct a model in which L2 motivation as a process is integrated with the motivational variables in language learning.. 4.

(5) 1.2 Purpose of the Study and Research Questions The objective of the present study is to establish a motivated language learning model from an expectancy-value perspective and a process-oriented point of view. It integrates five motivational variables: choice motivation of language learning, self-regulatory capacity in language learning, strategic language learning involvement,. strategic. language. learning mastery, and language. learning. achievement. The study aims to examine the causal links among these motivational related variables in the context of language learning in general and to answer the following research questions: 1. To what extent can motivation as expectancy-value predict the demonstration of self-regulatory capacity? 2. To what extent can self-regulatory capacity influence learners’ strategic language learning mastery? 3. To what extent can strategic language learning mastery predict language learning achievement? 4. To what extent can a motivated language learning model incorporating choice motivation, self-regulatory capacity, strategic involvement, strategic mastery, and language learning achievement, be established?. 1.3 Significance of the Study The study has both theoretical and pedagogical significances: Theoretically, the present study makes the first step to systematically explore language learning motivation of EFL learners based on the expectancy-value framework. In addition, it investigates how motivation as a process involving various variables is integrated with English learning in general and further constructs a motivated language learning model. The model developed in the study may shed light 5.

(6) on the elaboration of motivational theories in the SLA field. Pedagogically, the study can provide educators in such EFL contexts as Taiwan with a better understanding of the mediating forces working on the path from learners’ initial language learning motivation to learners’ ultimate language learning achievement. Moreover, the results derived from the study can be used to help learners in EFL contexts become more motivated, self-regulated, and strategically competent in language learning and thus attain higher achievement.. 1.4 Organization of the Thesis This chapter has provided a brief introduction of theoretical backgrounds and stated the research questions of the present study. Chapter 2 presents a review of related research and proposes a hypothesized model of motivated language learning. Chapter 3 describes the methodology used in the study, including participants, instruments, procedures, and data analyses. Chapter 4 reports the results of the analyses and Chapter 5 offers a discussion of the main findings. Finally, Chapter 6 summarizes the study and provides limitations of the study, suggestions for future research, as well as some recommendations for EFL pedagogy.. 6.

(7) CHAPTER TWO LITERATURE REVIEW This chapter discusses previous research concerning the issues addressed in the present study, including cognitive approaches to motivation, motivation as a process, and the factors involved in the path from language learning motivation to language learning achievement. Finally, a hypothesized model of expectancy-value motivated language learning is proposed based on the entire literature review.. 2.1 Cognitive Psychological Approaches to Learning Motivation Motivation is an intricate, multifaceted construct which has been widely investigated from various perspectives and in different approaches. Although disagreements still commonly prevail with regard to interpretations of learning motivation, a great number of recent educational psychological theories on motivation stress the cognitive aspects of motivation to learn (Ames, 1992; Bandura, 1993; Deci & Ryan, 1985; Locke & Latham, 1990; Ryan & Deci, 2000; Weiner, 1992; Wigfield & Eccles, 1992, 2000; Williams & Burden, 1999). Contemporary motivational psychology is characterized by a cognitive approach which places an emphasis on the individual’s thoughts, beliefs, and mental processes that are transformed into action (Dörnyei, 2001b). The fundamental conception underlying the cognitive approach to learning motivation is the belief that human beings are innately active learners with inherent curiosity to know the environment and to meet given challenges with the assistance of mental operations. Learners, not as passive recipients of knowledge, are metacognitively, motivationally, and behaviorally active producers of their learning achievement (Zimmerman, 1990). Accordingly, the main focus in cognitively-based motivational approaches is learners’ mental processes working together to form the existing motivation. 7.

(8) Driven by this belief, educational psychology researchers have expanded the conceptualization of learning motivation by incorporating various concepts, such as self-determination (Deci & Ryan, 1985; Ryan & Deci, 2000), self-efficacy (Bandura, 1993), goal setting and orientation (Ames, 1992; Locke & Latham, 1990), attribution styles (Weiner, 1992; Williams & Burden, 1999), and expectancy/value (Atkinson, 1957; Wigfield & Eccles, 1992, 2000). Among these cognitively-oriented explorations on learning motivation, numerous studies (Bandura, 1993; Eccles, 1983; Eccles et al., 1989; Feather, 1988; Schunk, 1991; Wigfield, 1994; Wigfield & Eccles, 1992) have specifically addressed expectancy and/or value (E-V) constructs. In addition, a substantial body of E-V-directed research have transformed E-V concepts into empirically measurable constructs and examined their effects on learning motivation, cognitive engagement and learning achievement.. 2.1.1 Expectancy-Value Theory Expectancy-value theory, originating from Atkinson’s classic achievement motivation theory (Atkinson, 1957), is one of the most influential and long-standing cognitive motivational theories. Essentially, it postulates that motivation to perform a given task is determined by two critical factors: expectancy of success and value. The former refers to the individual’s evaluation of the probability of attaining successful performances, while the latter is defined as the psychological value a learner attaches to success on that task or reasons for doing the specific activity. Theorists in this tradition argue that learners’ choice, persistence, and performance can be explained by their beliefs about how well they will do on the task and the extent to which they value the task (Wigfield, 1994; Wigfield & Eccles, 1992, 2000). More specifically, the greater the perceived possibility of goal achievement and 8.

(9) the higher the recognized value of the goal, the more positively motivated the learner may be. Conversely, if a learner is convinced that he or she can not achieve the given goal no matter how hard he or she tries, and also fails to perceive the valuable outcomes which the task will contribute to, it is very likely that the learner will lose motivation, and thus invest no effort on a seemingly hopeless and worthless task.. 2.1.1.1 Expectancy of Success In Atkinson’ achievement motivation model (1957, 1964), the expectancy construct was called probability for success, which, together with incentive values, was assumed to determine achievement behaviors. Following Atkinson’s theoretical assumption, Eccles, Wigfield, and their colleagues (Eccles, 1983; Eccles & Wigfield, 1995; Feather, 1988; Wigfield & Eccles, 1992, 2000), holding a more social-cognitive view towards motivation than Atkinson, led the development of modern E-V theories and constructed a comprehensive framework in which achievement performance is predicted by expectancy and task value. In this model, task value and expectancy together are supposed to influence or even determine learners’ choice, persistence, quality of efforts, cognitive engagement and actual performance. Moreover, it is assumed that task value depends on goals while the formation of one’s expectancy is predicted by task-specific self-concept and perceptions of task difficulty. Basically, the expectancy dimension in this influential framework and other related theories is concerned with the question, “Can I do this task?” (Eccles, 1983; Eccles & Wigfield, 2002; Pintrich 1988; Wigfield, 1994; Wigfield & Eccles, 1992) That is, expectancy of success concerns the learner’s perception of confidence in successfully performing an upcoming task. In terms of the antecedents of expectancy, Eccles and Wigfield (2002) proposed that task-specific self-concept beliefs and task difficulty were weighed together to 9.

(10) produce an expectancy judgment. Self-concept represents learners’ evaluation of their current ability to do a task while task difficulty involves their perception of difficulty they may encounter in the process of task completion. Besides indicating the two antecedents, Eccles and Wigfield, based on their extensive related studies, further proposed that both of the determiners were relatively domain-specific and may vary according to subject areas in school contexts (Harter, 1985; Marsh, 1990). For instance, the expectancy in math and in English, to a single learner, may differ dramatically and contribute to different behavioral outcomes. Since expectancy is commonly believed to correlate with learners’ perception of competence in certain domain, it has been frequently discussed within self-efficacy theory (Bandura, 1993, 1997; Bandura & Schunk, 1981), which suggests that the sense of self-efficacy, namely, learners’ assessment of their ability to accomplish a specific task, plays a significant role in learning motivation and performances. Self-efficacy is assumed to be a multidimensional construct which varies in strength, generality, and level (or difficulty). Learners’ efficacy expectations, according to Bandura (1993, 1997), were hypothesized to be a major determinant of goal setting, task choices, effort expenditure and perseverance in the face of difficulties. Moreover, it has been empirically supported to serve as a predictor of learners’ achievement regardless of individual variables such as proficiency level, gender, and age (Condly, 1999). Bandura’s self-efficacy construct is closely related to expectancy for success in Eccles and Wigfield’s E-V framework, especially the self-concept beliefs. However, self-efficacy theory differs from E-V theory in that self-efficacy perceptions are more situation-specific and contextually-changeable (Eccles & Wigfield, 2002). In a series of large-scale correlational studies, Eccles, Wigfield and their colleagues (Eccles, 1983; Eccles et al., 1989; Wigfield, 1994; Wigfield & Eccles, 1992) investigated the effects of expectancy on learning achievement. This 10.

(11) considerable body of research adopted cross-sectional as well as longitudinal research designs in which upper elementary and junior high students were given self-report measures directed to self-conception of capabilities and expectancy of success in math and English. The studies have shown that the strongest predictors of achievement, measured with standardized exams or course grades, were self-perception of competence (self-efficacy). This finding exhibited the mediating role of expectancy between environmental contexts and actual achievement and further demonstrated the importance of learners’ beliefs and the cognitive aspect of learning motivation. Besides examining the relations between expectancy and achievement, researchers were also interested in the interplay of expectancy and cognitive engagement. For instance, Pintrich and his associates (Pintrich & De Groot, 1990; Pintrich & Schrauben, 1992) examined how expectancy beliefs corresponded to learners’ use of cognitive strategies such as elaboration and metacognitive strategies. They have consistently found that the reported use of learning strategies was positively related to the level of expectancy and perceptions of competence. That is, the more competent the participants felt, the more learning strategies they reported using and the more cognitively engaged they were. In summary, expectancy of success concerns learners’ cognitive evaluation of their competence and their appraisal of the contextual factors within a given task. Empirically, it has been supported to be a motivational mediator between cognition and behavior.. 2.1.1.2 Value The other component in the E-V theory, value, also labeled as ‘valence’, ‘incentive value’, ‘task value’ by different researchers, refers to the subjective value a learner associates with the success in a given task. Compared with expectancy, as 11.

(12) Wigfield and Eccles (1992) indicated, the value component received less attention and fewer theoretical discussions because most theorists using E-V models have recently focused on expectancy. Due to this inadequacy, Eccles, Wigfield and their associates (Eccles, 1983; Eccles, Adler, & Meece, 1984; Wigfield, 1994; Wigfield & Eccles, 1992, 2000) developed a thorough framework in which value was elaborately discussed. Rooted in traditional E-V models such as Atkinson’s, their model incorporates social cognitive research on intrinsic and extrinsic motivation inspired by Deci and Ryan’s research (1985), and assumes that both expectancy and value are cognitive beliefs related to the decisions learners consciously make about their achievement. According to Eccles and Wigfield (1995), achievement task value consists of four components: attainment value, intrinsic value, utility value, and cost. Attainment value refers to the importance of doing well on a task. Intrinsic value is the enjoyment a learner gains from doing the task or the aesthetic appreciation of the task itself, while utility value is defined as the usefulness of the task in terms of learners’ current or future goals such as fulfilling a requirement for a master degree or getting higher grades for entering a prominent university. Intrinsic value and utility value in the value framework echo with the distinction Deci and Ryan (1985) made in their influential intrinsic/extrinsic motivational model. According to Deci and Ryan, intrinsic motivation is the most self-determined form of motivation, which can be defined as doing an activity for its inherent satisfaction rather than for external reasons whereas extrinsic motivation concerns engagement in a task for obtaining external rewards such as grades or for avoiding possible punishment (Ryan & Deci, 2000). The last component of value is cost, which refers to the perceived negative consequences of engaging in the task, which may include not only the perceived amount of efforts required as well as time invested, but also anticipated emotional 12.

(13) states such as anxiety and fear. As a whole, these four elements are assumed to operate together to determine the task value a learner perceive and thereby influence learners’ intensity of motivated learning behaviors. With a comprehensive value pattern, in their empirical work, Eccles et al. (Eccles, 1983; Wigfield, 1994; Wigfield & Eccles, 1992, 2000) have focused on how the value components interplay with expectancy beliefs to affect learners’ actual behaviors in authentic school settings. Longitudinal and correlational in design, the studies used self-report surveys with large samples of students learning English and math. Several consistent findings have emerged in the large-scale body of serial research. First, the first three components of value, namely, attainment, intrinsic, and utility values, were empirically differentiable constructs. That is, learners did make distinct judgments about importance, interest, and usefulness of a learning task. Furthermore, the research has shown that these value components positively correlated with expectancy beliefs. For instance, the students tended to highly value the tasks they thought they could do well (Eccles & Wigfield, 1995). This finding strongly supported the assumption that expectancy and value are positively related, not inversely associated as originally proposed by Atkinson (1964). Third, value was positively correlated with achievement. That is, the extent to which the learner valued the domain-specific task predicted their actual performances (Eccles, 1983; Eccles & Wigfield, 1995). However, when both expectancy beliefs and task values were used to predict achievement, expectancy beliefs were significant determinants while values were not (Eccles et al., 1983; Meece, Wigfield, & Eccles, 1990). In contrast, when it comes to choice behaviors, such as intentions to take courses or to keep enrollment, values, including attainment value, intrinsic value, and 13.

(14) utility value, had more predictive power than expectancy beliefs (Eccles et al., 1983; Meece, et al., 1990). It seems, from an E-V perspective, expectancy is more directly related to learning achievement while value is more closely tied to decision-making process which provides further opportunity for future achievement. This finding suggests that once a learner has started the learning process in a school setting, in order to improve their classroom achievement, it is more crucial for instructors to enhance learners’ expectancy and self-efficacy beliefs, rather than worry about increasing their value judgment and interest in the course material. However, enhancing learners’ value beliefs is very likely to have a beneficial impact on their enrollment as well as willingness to participate in the given learning activities, which may indirectly determine the degree of learners’ achievement.. 2.1.1.3 Assessing Expectancy-Value Learning Motivation Like most empirical studies on motivation, the vast majority of the research based on a cognitive approach to learning motivation, especially from the expectancy-value perspective, have been using self-report instruments to ask about participants’ judgments on their expectancy of success and value beliefs toward a domain-specific task. In general, these questionnaire items are rated on a Likert scale and analyzed with a variety of statistic tools. Among the cognitively-based motivational instruments, the Motivated Strategies for Learning Questionnaire (MSLQ) (Pintrich, Smith, Garcia, & McKeachie, 1993) is one of the most reliable and valid measures. It has been translated into multiple languages and conducted by hundreds of researchers for different research purposes (Rao & Sachs, 1999; Sachs, Law, & Chan, 2001). As its name indicates, the instrument items cover two broad areas, 31 of which on motivation and 60 on learning strategies ─here only the motivation section will be 14.

(15) elaborately discussed. In essence, the MSLQ was developed based on a social-cognitive view of motivation, assuming that learners are active processors of information whose cognitive beliefs mediate important instructional input and task characteristics. Besides, behind the MSLQ stands a notion that motivation is dynamic and contextually bound (Garcia & McKeachie, 2005). In other words, learners’ motivation varies according to different domains such as physics, mathematics, and English. Therefore, the MSLQ was designed to focus on the course level, situated between the very general level of all learning situations and the impractical level of every specific situation within one course. The motivation section of the MSLQ is grounded on three constructs, expectancy, value and affect, and consists of six subscales: control of learning beliefs (expectancy), self-efficacy for learning and performance (expectancy), intrinsic goal orientation (value), extrinsic goal orientation (value), task value (value), and test anxiety (affect). Expectancy components refer to learners’ beliefs of the extent to which they can control learning outcomes and to which they expect themselves to accomplish a task. Value components focus on the reasons why learners engage in an academic task, such as learning for learning itself and mastery (intrinsic goal orientation), learning for grades or approval from others (extrinsic goal orientation), and judgments on the usefulness, enjoyment, and importance of the learning course. The third motivational construct is affect and has been operationalized into the test anxiety scale, which assesses learners’ worry and uneasiness over taking exams. Although MSLQ has been extensively applied in a substantial body of educational research and also has been proven to be highly reliable and valid (Garcia & McKeachie, 2005), few studies adopted it to examine L2 motivation. It is reasonable to argue that such an academic-motivation directed and instructional 15.

(16) learning situated measure is potentially useful to explore the nature of L2 motivational dispositions EFL learners hold because most of the learners receive formal English instruction as a major learning medium and generally take English as a school subject for examinations rather than a tool for interpersonal communication.. 2.1.2 Expectancy-Value in Studies on L2 Motivation Before 1990, the investigation of L2 motivation was dominated by a social psychological approach, especially guided by the socio-educational model by Gardner and Lambert (Gardner, 1985; Gardner & Lambert, 1972). However, realizing the practical but limited applicability of this social approach, after 1990, L2 researchers (Clément et al., 1994; Mori, 2002; Mori & Gobel, 2006; Noels et al., 1999; Noels et al., 2003) started to make paradigmatic moves from a social approach to mainstream psychological cognitive approaches, aiming at the expectancy-value related constructs in particular. No concrete E-V model has been offered in L2 motivation research, but some E-V related components have been incorporated into several L2 motivation studies (see Dörnyei, 2001b, for a comprehensive review) For instance, Tremblay and Gardner (1995) included ‘valence’ and ‘self-efficacy’ variables in their prominent socio-educational model to signal the positive role which learners’ cognitive judgments may play in L2 motivation. They indicated that valence and self-efficacy, together with goal salience, will affect motivational behaviors, including attention, motivational intensity, and persistence, and in turn work upon language learning achievement. In terms of the aspects of expectancy, Clément and his associates (1983, 1994) explored L2 learners’ linguistic self-confidence and have identified its significance in intercultural communication. Basically, linguistic self-confidence is similar to 16.

(17) self-efficacy, since both refer to a L2 learner’s perception of his or her ability to perform tasks competently. However, linguistic self-competence is more generalized in nature while self-efficacy is highly task-specific (Dörnyei, 2001b). Moreover, in contrast to the cognitive dimension of self-efficacy, self-competence seems to be socially-oriented because it involves the quality and quantity of the direct or indirect contact between language learners and native speakers (Dörnyei, 2001b). With regard to value components, Noels and his associates (Noels, 2001; Noels et al., 2001, 2003) investigated how intrinsic motivation (intrinsic value) and extrinsic motivation (utility value) were related to integrative orientation and the motivational model proposed by Clément and Kruidernier (1983), which is composed of travel, friendship, knowledge and instrumental orientation. It was found that extrinsic regulation (the least self-determined form of extrinsic motivation) correlated strongly with instrumental orientation and the others were more linked with intrinsic types of L2 motivation. Although these aforementioned L2 studies have incorporated some E-V related components into their investigation, few empirical studies in the field of EFL (English as a foreign language) directly examined the expectancy-value framework. That is to say, the application of E-V theory in the L2 field is a hybrid combination in which social dimensions of L2 motivation is emphasized with the cognitive aspect as a minor supplement. The only E-V based studies which could be found in the literature were those conducted by Mori et al. (Mori, 2002; Mori & Gobel, 2006). These researchers, from an E-V perspective, explored what constituted English learning motivation in a specific EFL setting of Japanese universities. By a principle components analysis, it was shown that expectancy for success, attainment value, intrinsic value, and utility value were not only conceptually distinct but also empirically distinguishable in the 17.

(18) contexts under investigation. It suggested that the general forms of motivation identified in the expectancy-value theory could adequately reflect foreign language motivation in Asian EFL contexts. Though Mori et al. have taken a preliminary step to directly apply the E-V theory to the investigation of EFL learning motivation, they examined solely the constructs encompassed in the E-V framework and did not explore their relationships with motivated behaviors or learning outcomes, such as strategy use or achievement. Therefore, for a more complete picture of L2 motivation, the links between E-V constructs, learners’ actual motivated behaviors, and language learning achievement should be simultaneously and closely explored.. 2.2 Motivation as a Process Although most of the aforementioned motivational theories imply that motivation is relatively stable and can be precisely measured by tapping into it at a specific point of time, learning motivation has been gradually believed to be dynamic, contextually-bound and temporally changing because a great body of research has shown that learners’ motivation to learn a language does not remain constant and often weakens over time (Dörnyei & Csizer, 2002; Inbar, Donitsa-Schmidt, & Shohamy, 2001; Koizumi & Kai, 1992; Tachibana, Matsukawa, & Zhong, 1996). In addition, L2 learners often experience a change of motivational types with the school years increase (Tachibana et al., 1996; Wen, 1997). That is, instead of an unchanging state, L2 motivation is fluctuating in nature. It may involve a process where a sequence of events and variables may interplay to influence motivated behaviors and thereby learning achievement. In fact, it is certainly reasonable to hold the belief that L2 motivation fluctuates more or less according to external contextual changes or internal mental 18.

(19) transformation, because L2 learning, whether in EFL or ESL contexts, often takes a considerable period of time to accomplish. As Schumann (1998) indicated, L2 learning is sustained deep learning, which displays distinct motivational features from short-term activities. Therefore, how motivation evolves and how motivational strength is maintained for a long period is supposed to be a significant issue in L2 studies. To go deep into the nature of L2 motivation, SLA (second language acquisition) researchers should take the temporal and dynamic dimension of learning motivation into consideration and examine what factors mediate the motivational process. Although this dynamic and process-oriented dimension is seldom doubted, few of the available motivational theories, especially in the SLA field, provided a detailed, systematic framework to display motivational transformation over time. However, with the increasing awareness of motivation as a process, this aspect of motivation have been recently addressed by a rising number of studies in psychology, especially those inspired by Heckhausen and Kuhl’s action control theory (1985). Moreover, some preliminary endeavors have also been made in L2 research such as a process-oriented model of L2 motivation proposed by Dörnyei and Ottó (1998).. 2.2.1 Heckhausen and Kuhl’s Action Control Theory One step to conceptualize motivation as a process which involves distinct phases was taken by Heinz Heckhausen and Julius Kuhl (Heckhausen & Kuhl, 1985; Kuhl, 1985, 1987), two prominent motivational psychologists in recent decades. They proposed a motivational theory, referred to as action control theory, and presented a “temporal perspective that begins with the awakening of a person’s wishes prior to goal setting and continues through evaluative thoughts entertained after goal striving has ended” (Gollwitzer, 1990, p. 55). In other words, motivation, according to the 19.

(20) action control theory, entails different stages and evolves from initial formation of a goal, striving for the goal, and eventually to an evaluation of the whole process. This theory attempts to explain the common observation that people do not always behave in accordance with their intentions which have been formed based on an expectancy-value assessment. That is, there seems to be an obvious gap between one’s intention and one’s behavior. Besides, it is prevalently observed that even though the planned goal or intention remains unaltered, people’s motivational strength wax and wane over time. It is also a common phenomenon that, even though more attractive alternative activities exist, people still persist in pursuing the goal and engage in given tasks. For an adequate account for these ordinary phenomena, Kuhl, Heckhausen and their associates (Gollwitzer, 1990; Heckhausen & Kuhl, 1985; Kuhl, 1985, 1987; Kuhl & Beckmann, 1994) highlighted the dynamic nature of motivational processes by making a distinction between two separate, temporally-ordered phases, and most importantly, introduced the mediating role of action control in the discrepancy between intention and behavior.. 2.2.1.1 Distinction of Two Motivational Phases The action control theory proposes that a goal-directed activity should be divided into two distinct major phases, the predecisional and the postdecisional phases: Predecisional phase: a decision–making or intention-formation stage at which the intention to learn is formed based on the assessment of efficacy and value. It concerns a pre-actional process in which choice motivation exercises to select a goal and reach a resultant intention. Postdecisional phase: an implementation stage at which an established intention is executed and maintained by means of volitional management. 20.

(21) It involves executive motivation which guides how intentions are realized and how action control is manipulated to protect the desirable intention against other competing activities and thus maintain motivational intensity.. The boundary separating the predecisional phase (the planning stage) and the postdecisional phase (the implementation stage), is the commitment (or Rubicon as a metaphorical analogy) the learner makes to an intention. At the first stage, a goal turns into an intention, or an ‘activated plan to which the actor has committed herself or himself’ (Kuhl, 1987, p. 282), while at the second phase, this committed intention is activated into actual behavior with a mechanism dealing with the task itself (e.g. the use of cognitive learning strategies) and with another mechanism employed to maintain or protect the intention (e.g. the use of self-regulatory strategies). The basic assumption behind this distinction is that how intentions are formed is one thing and how intentions are implemented is another (Heckhausen, 1991). In fact, for a more in-depth understanding of motivational processes, these two phases should receive equal attention and even be explored simultaneously to investigate their connections. However, most mainstream psychological theories have only focused on the predecisional phase with an exhaustive analysis of goal-setting processes and factors which may influence intention formation. This line of research inquiry has directly guided the direction of L2 motivational studies. To put it specifically, most of the previous studies on L2 motivation (Belmechri & Hummel, 1998; Clément & Kruidenier, 1983; Gardner, 1985; Gardner & Lambert, 1972; Mori, 2002; Noels et al., 2003; Warden & Lin, 2000) focused on the motivational antecedents which influenced learners’ motivational dispositions, rather than examine how these motivational tendencies were protected, scaffolded, and enhanced. 21.

(22) However, in an EFL environment such as Taiwan where English is a compulsory school subject of secondary education and to learn English is usually an non-self-determined intention of learners, we may argue, it is more meaningful and pedagogically valuable to explore the ongoing motivational process learners experience when implementing their learning tasks and the interplay between motivation and behavior than to narrow the research focus down on what forms L2 motivation and the reasons for doing a specific learning activity.. 2.2.1.2 Action Control: the Mediator between Intention and Behavior Besides clearly differentiating two temporally-ordered motivational phases, the action control theory emphasizes the crucial impacts of action control, also termed as volitional control or self-regulatory control, on learning behaviors. It is assumed to dominate the processes “that mediate intention-behavior consistency by protecting a current intention from being replaced by a competing action tendency” (Kuhl, 1985, p. 102). According to Kuhl (1987) and Corno (1993), motivation and volition should be distinguished because the former drives the decisions to engage in the task and form an intention but the latter guides the actual behaviors to pursue a decided intention. In other words, volition is the mediator bridging the gap between intention (choice motivation) and behaviors (performances). Kuhl argued that many motivational theorists have neglected volitional processes by assuming that motivation directly, linearly leads to outcomes or achievements. He proposed a contrary concept that the path from motivation to achievement is dynamic and concerns certain mediating power because motivation only contributes to the intention to learn and can not completely dominate the execution of this intention. He further suggested that once a learner engages in a learning activity, volitional processes, governed by action control, 22.

(23) will take over. Moreover, the efficiency of self-regulatory control will influence intention execution and task completion. In a broader sense, action control can be defined as “those thoughts and/or behaviors that are directed toward maintaining one’s intention to attain a specific goal in the face of both internal and external distractions” (Garcia et al., 1998, p. 393). It is exercised in the post-decisional phase, conceptualized as executive motivation, to help a committed intention preserved until the current intention is accomplished or terminated. In a learning context, action control is assumed to mediate the intention to learn (choice motivation) and the use of learning strategies (cognitive engagement) (Corno, 1993; Kuhl, 1985). Based on Kuhl’s more detailed theory of action control, learners with low level of action control can be characterized as state-oriented, while learners with high level of action control are recognized as action-oriented. Research has indicated that action-oriented learners, even encountering failures, persisted in using effective strategies and maintained their sense of competence (Brunstein & Olbrich, 1985). A series of qualitative studies conducted by Reed, et al. (Reed, Hagan, Wicker, & Schallert, 1996; Reed & Schallert, 1993; Reed, Schallert, & Deithloff, 2002), in addition, has suggested that volitional control was critical in helping learners enter into involvement. Also, action control has been found to lead to a better performance on achievement tests (Menec & Schonwetter, 1994; Menec, Schonwetter, Struthers, & Perry, 1993). In general, action control has been operationalized as a repertoire of self-regulatory strategies (Kuhl & Beckmann, 1994). Based on his preliminary studies on the techniques which the learners reported using to strengthen their learning focus and avoid distractions, Kuhl concluded the following taxonomy of self-regulatory strategies, which was believed to facilitate the protection of a current intention against 23.

(24) the attractive power exerted by other competing action tendencies (summarized from the work by Kuhl (1985, 1987) and Dörnyei (2001a)): 1. Active Attentional Selectivity refers to selectively attending to relevant information and intentionally neglecting attractive alternatives. 2. Encoding control refers to selectively encoding those features of a stimulus related to the current intention. 3. Emotion control refers to the inhibition of the emotional states which might hinder the implementation of the intention. Emotion control also involves the generation of positive emotions that are beneficial for intention execution. 4. Motivation control refers to the strengthening of the current intention’s motivational basis, aiming at a change of the present hierarchy of tendency strength. Frequently-used motivation control strategies are to think about what will happen if the original intention fails, and to remind oneself of the favorable expectancies or positive incentives. 5. Environment control refers to the manipulation of the environment to make the abandonment of the intention more difficult and less convenient. Typical environment control strategies are to make social commitments and to create a setting facilitative for intention implementation. 6. Parsimony of information-processing refers to a ‘let’s not think about it any more but get down to doing it’ strategy.. In response to Kuhl’s taxonomy of volitional control strategies, Corno and Kanfer (1993) presented a scheme composed of five distinct categories of self-regulatory actions, namely, metacognitive control, environment control, attentional control, emotion control, and motivation control. For an assessment of learners’ ability to exercise volitional control in academic situations, the Academic Volitional Strategy Inventory (AVSI) was developed (McCann & Garcia, 1999; McCann & Turner, 2004) and have recently adopted in several psychological studies (Bembenutty, 2004; McCann & Garcia, 1999; McCann & Turner, 2004). Although empirically applicable and adequate in terms of its psychometric properties, however, this self-report instrument only addresses students’ management in Corno’s categories 24.

(25) of emotion and motivation and neglects the other three types of self-regulatory strategies which have been shown to be effective for learners’ self-regulation (see Zimmerman & Schunk, 2001, for review). Thus, the development of a more exhaustive measure tapping into self-regulatory capacity is critically required for an entire picture of learners’ exerted effort to protect their learning intentions. Several empirical studies (Bembenutty, 2004; Garcia et al., 1998; McCann & Garcia, 1999; McCann & Turner, 2004) have been overtaken to examine the mediating role of action control in the relationships between motivation and learning strategy use. These studies have consistently indicated that motivation and volition were not only conceptually distinct but also empirically distinguishable, which was in accordance with Kuhl’s assumption about the distinction of choice motivation and executive motivation. In addition, it has been shown that the use of volitional control strategies, although having no significant effects on course performance measured by final course grades, positively affected learners’ cognitive engagement measured by the use of learning strategies (Garcia et al., 1998; Pintrich & De Groot, 1990). It is also noticeable that, according to Garcia et al.’s study (1998), the indirect link of motivational factors such as task value, self-efficacy and intrinsic goal orientation, to learning strategy use through action control was much stronger than the direct effect of choice motivation on the use of learning strategies. That is, volitional control did protect one’s intention to learn and maintain goal-directed activities by amplifying the effects of initial motivational level on the use of learning strategies. To sum up, all of these related studies have provided sound evidence to support the assumption that volitional control or self-regulatory capacity is a mediator in the motivation-cognitive engagement-performance equation.. 25.

(26) 2.2.2 Dörnyei and Ottó’s Process-oriented Model of L2 Motivation In response to the dynamic and temporally unstable nature of L2 motivation, although few, some SLA researchers have started to address L2 motivational processes and discuss respective elements involved at each motivational phase. One of the latest and most comprehensive frameworks constructed with this effort was a process model of L2 motivation proposed by Dörnyei and Ottó (1998). Inspired by Heckhausen and Kuhl’s action control theory, Dörnyei and Ottó tried to introduce a process-oriented perspective of L2 motivation and synthesized an ample body of L2 motivation research into a unified and systematic framework. This model, as Dörnyei summarized (2003, p. 18), is a process model of L2 motivation that “breaks down the overall motivational process into several discrete temporal segments organized along the progression that describes how initial wishes and desires are first transformed into goals and then into operationalized intentions, and how these intentions are enacted, leading (hopefully) to the accomplishment of the goal and concluded by the final evaluation of the process.” The following section gives a detailed description of Dörnyei’s model and provides related empirical findings.. 2.2.2.1 The Sequential Pattern of L2 Motivational Process Similar to the action control theory, Dörnyei and Ottó’s model demonstrates an action sequence which is basically composed of three discrete phases: preactional (concerning choice motivation preceding the implementation of action, dealing with how motivation is formed), actional (referring to executive motivation which regulates the motivation generated at the first stage and maintains the current intention when it is implemented) and postactional (regarding an overall appraisal after the action has been accomplished or terminated). Preactional phase includes goal setting, 26.

(27) intention formation, and initiation of intention enactment. Actional phase concerns action while post-actional phase addresses a retrospective evaluation after the action has been completed. This model, furthermore, specifies the corresponding motivational effects on each phase. More specifically, the model details the motivational forces or motives that facilitate or hinder the behavioral process. It is worth noting that the motives indicated to affect each motivational stage differ greatly as learners move from one to another. For example, the executive motivational influences associated with the actional phase are not directly related to the motives working upon the earlier or later stages of the motivational process. This tenet echoes with the assumption of the action control theory that the predesicional phase (choice motivation) and the postdecisional phase (executive motivation) encompass essentially distinct variables. A recent study by MacIntyre, MacMaste, and Baker (2001) has provided empirical evidence to indicate the process aspect of L2 motivation. These researchers conducted a factor analysis to examine the convergence of 23 motivational scales from multiple L2 motivational frameworks. They found that these scales neatly fell into three separate factors. The first two, labeled as attitudinal motivation and self-confidence, were associated with established components of L2 motivation, chiefly concerned with the preactional phase. However, the third factor, termed as action motivation, was closely tied with the actional stage of L2 motivation. The emergence of action motivation as an independent variable explicitly indicates the process dimension of L2 motivation. As a whole, Dörnyei and Ottó’s process-oriented model highlights the dynamic character and temporal variation of L2 motivation, which is a potentially fruitful approach to interpreting and integrating motivational factors that affect language learning behaviors in classroom settings (Dörnyei, 2003; Dörnyei & Skehan, 2003). 27.

(28) Two recent empirical studies (Chen, Warden, & Chang, 2005; Tseng & Schmitt, 2008) have made initial steps to approach L2 motivation based on this model and have both indicated that this process-oriented perspective on L2 motivation led to a fuller understanding of the whole motivational learning process. Chen et al. (2005) examined the interrelations among variables functioning in preactional (instrumental, integrative, and required orientation), actional (actual language use measured by effort and success), and postactional phases (self-evaluation). It was shown that the preactional phase did not highly correlate with the variables in the postactional phase, but the actional phase was significantly related to postactional phase factors. This finding supported the mediating effect of the actional phase and implied that high levels of initial motivation may not transform directly to skill achievement. The other study from a process-oriented motivational view is a causal one by Tseng and Schmitt (2008). Tseng and Schmitt, by constructing structural equation modeling, validated a proposed model of motivated English vocabulary learning. This model, drawing on the work by Dörnyei (2001a, 2001b) about the motivational stages, consisted of six variables, one of which (initial appraisal of vocabulary learning experience) represented the outcomes of the preactional phase, three (self-regulating capacity, strategic involvement, and mastery of learning tactics) the actional and two (vocabulary knowledge and post-appraisal) the postactional phrase. A systematic cycle from the initial motivational level (the preactional stage) to the retrospective evaluation (the postactional stage) was found, which demonstrated that, instead of an initial state, motivation was an integral segment of the whole vocabulary learning process. It also showed that the predictive power of choice motivation on effective strategy use was positively significant only via the demonstration of self-regulatory capacity. To put it differently, in an EFL setting, volitional control was also confirmed 28.

(29) to mediate between motivational factors and strategic learning behaviors. Although Tseng and Schmitt (2008) examined the role of motivation as a process and made the first endeavor to explore learners’ self-regulatory capacity in language learning, their research target solely fell on the vocabulary learning process, which is merely one particular domain of language learning. Accordingly, further studies are needed to examine L2 motivation in general from a process-oriented perspective and to investigate the potential impacts of volitional control by applying Tseng and Schmitt’s soundly-developed vocabulary learning model to general language learning processes.. 2.2.2.2 Self-Regulatory Strategies Following the main assumption of the action control theory, Dörnyei and Ottó’s (1998) process model of L2 motivation also marked the influence of action control strategies during the actional stage. As mentioned above, action control strategies are self-regulatory mechanisms activated to enhance, scaffold or protect learning-specific actions (Dörnyei, 2001a). They are particularly important, as Dörnyei and Ottó suggested, for two basic reasons. First, it has been found that academic schoolwork were generally considered by teenage learners the least rewarding, motivating, interesting activity (Wong & Csikszentmihalyi, 1991; Schneider, Csikszentmihalyi, & Knauth, 1995). This negative belief widely held by learners establishes numerous opportunities for a wide range of distractions to interfere with learning. Self-regulatory strategies are therefore needed for the sake of learning effectiveness. Second, in school contexts, learners were frequently imposed with tasks which they had little room to choose during the preactional phrase. Since the motivational force stimulated at the choice motivation stage is relatively fragile, the volitional control mechanism is required to maintain and 29.

(30) even enhance learners’ motivation to reach the ultimate goal of acquisition. Based on Kuhl’s (1987), Corno and Kanfer’s (1993) taxonomy of action control strategies, Dörnyei (2001a) developed a scheme of language learning self-regulatory or self-motivating strategies as follows (cf. Dörnyei, 2001a, p. 110-115): 1. Commitment control, techniques that help to preserve or enhance the learners’ original goal commitment (e.g. keeping in mind favorable expectancies or positive incentives, or focusing on what would happen if the original intention failed). 2. Metacognitive control, techniques used to monitor and control concentration and to stop procrastination (e.g. giving oneself regular self-reminders to concentrate, giving oneself regular self-reminders of the deadline, or intentionally ignoring attractive alternatives or irrelevant aspects). 3. Satiation control, technique intended to add extra attraction to the task (e.g. adding a twist to the task, or using one’s fantasy to live up the task). 4. Emotion control, techniques used to manage obtrusive states and generate emotions which are conducive to implementing the intentions (e.g. generating useful diversions, self-affirmation, constructing positive narratives of events, or finding humorous elements). 5. Environment control, techniques used to eliminate negative environmental influences and exploit positive ones (e.g. removing environmental sources of interference and temptations, asking friends to restrict you, or making a public commitment). Dörnyei’s (2001a) taxonomy of self-motivating strategies, mainly derived from the psychological field, has not been extensively transferred to a concrete instrument assessing language learners’ self-regulatory capacity. The only measure of self-regulation in L2 learning we can find from the literature is a self-report instrument conceptualized, developed, and validated by Tseng, Dörnyei, and Schmitt (2006), which is named ‘Self-Regulating Capacity in Vocabulary Learning Scale (SRCvoc)’. With the notion that the most crucial aspect of strategic learning dose not lie in the specific techniques learners actually employ but rely on the fact that learners. 30.

(31) exert purposive effort to improve their learning, this measure taps into general tendencies and inclinations of language learners’ self-regulation rather than focuses on specific behavioral descriptions. By means of confirmatory factor analysis, Tseng et al. demonstrated that this psychometrically-based measure of L2 learners’ self-regulatory capacity reached satisfactory reliability and validity, and thus suggested that Dörnyei’s (2001a) self-regulatory framework is empirically valid. However, this measure, to date, has solely been used to assess learners’ self-regulatory capability in a specific domain, i.e. English vocabulary acquisition. The self-regulation of L2 learning in general has not been explored, which can be effectively probed into through Dörnyei’s self-regulatory scheme (2001a) and Tseng et al.’s proposed measurement (2006).. 2.3 From L2 Motivation to L2 Achievement The link between L2 motivation and L2 achievement has been a popular issue under SLA investigation for decades. Researchers have indicated that motivation to learn a second or foreign language (L2) is a crucial predictor of success in language learning (Csizer & Dörnyei, 2005; Dörnyei & Csizer, 2002; Ely, 1986; Gardner, 1985; Gardner & MacIntyre, 1991; Gardner et al., 1997; Noels et al., 1999; Oxford & Shearin, 1994; Samimy & Tabuse, 1992; Schmitt & Watanabe, 2001; Tremblay & Gardner, 1995; Wen, 1997). That is, L2 learners with higher degree of learning motivation are widely believed and, in most cases, have been empirically found to achieve better proficiency in the target language. Although a wide range of L2 studies have supported the strong predictive power of motivation on language learning achievement, contradictory results have been shown in either qualitative (Huang & Chang, 1996) or quantitative studies (Chen et al., 2005; Teweles, 1995; Shaaban & Ghaith, 2000; Vandergrift, 2005). For example, 31.

(32) Huang and Chang (1996), in an ESL (English as a second language) setting, found that the participants’ motivational beliefs such as self-efficacy level did not correlate with their learning achievements correspondingly but their achievements positively reflected their motivational level. It suggested that L2 motivation did not necessarily lead to successful performance but it served as a precondition for better achievement. In a comparative study of two EFL settings, i.e. Japan and China, Teweles (1995) demonstrated that the level of motivation did not highly correlate with proficiency regardless of test-type. Vandergrift (2005) also indicated that L2 motivation seemed not to be a reliable predictor of proficiency in L2 listening comprehension. All of the aforementioned studies signaled the insignificant link between motivation and learning outcomes. It strongly suggested that the path from language learning motivation to language learning achievement is not linearly connected but indirectly correlated via mediators. One mediator which has been found to be significant in several pioneering studies (Garcia et al., 1998; Pintrich & De Groot, 1990; Tseng et al., 2006; Tseng & Schmitt, 2008) was learners’ self-regulatory capacity, which was argued to help protect the initial intention and to further support the demonstration of motivated learning behaviors such as strategy use (Dörnyei, 2005; Tseng & Schmitt, 2008). In the field of L2 research, in general, the use of language learning strategies has been the most frequently explored type of actual learning behaviors which SLA researchers targeted to contribute to a deeper understanding of the indirect association between L2 motivation and L2 achievement (Garcia et al., 1998; MacIntyre & Noels, 1996; Printrich & De Groot, 1990; Schmidt & Watanabe, 2001; Tseng & Schmitt, 2008; Vandergrft, 2005). The following section provides a short review of the empirical studies on language learning strategy use and specifies the inherent problems in the L2 research on strategy use. 32.

(33) 2.3.1 L2 Motivation and Language Learning Strategy Use For a link of language learners’ motivational disposition with their actual motivated behaviors, a considerable body of research has been conducted to explore the interrelations between choice motivation and learners’ cognitive engagement or strategic involvement in learning tasks (Garcia et al., 1998; MacIntyre & Noels, 1996; Printrich & De Groot, 1990; Schmidt & Watanabe, 2001; Tseng & Schmitt, 2008; Vandergrft, 2005). More specifically, these empirical studies aimed to uncover how motivational tendencies reflect learners’ reported use of learning strategies, which are generally defined as actions, behaviors, steps or techniques used to enhance language learning (Oxford, 1990). With regard to the correlation between choice motivation and learning strategy use, previous research has demonstrated that the level of motivation directly affected learners’ use of learning strategies, not only the frequency (Gardner et al., 1997) but also the types of strategy use (Garcia et al., 1998; Schmidt & Watanabe, 2001). However, conflicting results also existed. For example, Tseng and Schmitt (2008) found that choice motivation did not directly and positively influence strategy use. It was found that learners’ self-regulatory capacity was a determinant mediator functioning on the path from choice motivation to cognitive engagement. Accordingly, it is reasonable to hypothesize that the relation between choice motivation and strategy use can be established by means of learners’ capacity in regulating their learning behaviors as a mediator.. 2.3.2 Language Learning Strategy Use and L2 Achievement Besides the relation between L2 motivation and language learning strategy use, how strategy use correlates with language learning achievement has also been an intriguing research issue drawing a wide range of attention. Still, like the empirical 33.

(34) findings concerning motivation and learning strategy use, the results of L2 studies which examined the role of learning strategy use in language learning success were also mixed. Some admittedly showed a significant positive link (Park, 1997), some a fairly weak correlation (Nisbet, Tindall, & Arroyo, 2005), some no significant link (Mori, 2007), and others perplexingly a negative association (Gardner at al., 1997). This inconsistency and discrepancy of empirical findings concerning the predictive power of learning strategy use on L2 achievement may result from two fundamental problems: one inherent in the frequency-based measurement of strategy use, and the other deriving from the conceptual ambiguity of language learning strategies. In general, studies on strategy use assessed learners’ strategic behaviors by means of self-report questionnaires. In the L2 field, the most frequently adopted measure of language learning strategies from 1990s has been the ‘Strategy Inventory for Language Learning’ (SILL), developed by Rebecca Oxford (1990). This instrument, composed of six dimensions, focuses on specific strategic behaviors and asks about L2 learners’ frequency of usage of strategies which ranges from ‘always’ to ‘never’. Scale scores are attained by computing the average of the item scores within a sub-dimension. Essentially, the psychometric assumption behind this measure is that the more usage, the better. However, it has been argued that the frequency-based instrument only measures the quantity of strategy use but fails to reflect another important dimension of strategic behavior, namely, the quality of strategy use, i.e. how well these strategies are used (Tseng et al., 2006; Tseng & Schmitt, 2008). A high score on SILL reflects only a high frequency and a wide range of strategy use. It does not indicate the effectiveness of strategy use and learners’ adaptive manipulation of available learning tactics. 34.

(35) As for the nature of learning strategy use, researchers (Gardner et al., 1997; Ellis, 1994) have suggested that the quality dimension of strategy use is crucial to learning achievement and should be distinguished from the quantity dimension of strategy use because a larger repertoire of learning tactics and frequent use of learning strategies does not guarantee adequate and resourceful use of language learning strategies which will probably lead to better learning achievement. Therefore, the measurement of learners’ strategic competence should touch upon two distinct parts, one on learners’ quantity-based, overall involvement in using strategies for language learning and the other on learners’ quality-based mastery of specific language learning tactics. In addition to the inherent instrumental problem, research on language learning strategies has been confronting with a theoretical problem, that is, the conceptual ambiguity of learning strategy (Dörnyei & Skehan, 2003). Disagreements prevail in terms of the conceptualizations of language learning strategy and the discriminating criteria for strategic learning from ordinary learning (see Dörnyei, 2005, for a comprehensive review). In fact, during the past decades, educational psychology has abandoned the confusing term ‘learning strategy’ and discusses learners’ strategic contribution to their learning in a more versatile concept of ‘self-regulation’. This theoretical transition implies that it is not what learners do that makes them strategic learners, but rather the fact that they make purposive efforts, get consciously involved, and regulate their own learning process (Pintrich, 2000; Skehan & Dörnyei, 2003; Tseng et al., 2006). The focus, thus, shifts from the actual techniques employed for dealing with the learning task itself to the self-regulatory process whereby learners’ underlying self-regulatory capacity works. Again, it appears justifiable to conclude that on the way from L2 motivation to L2 achievement, in addition to strategic learning behaviors, 35.

(36) learners’ self-regulatory capacity may also take an important part.. 2.3.3 Motivation-Achievement as a Reciprocal Cycle Although commonly believed to be a determinant of success in language learning, L2 motivation has been gradually regarded also as a result of language learning achievement. As the ‘Resultative Hypothesis’ claims, “learners who do well are more likely to develop motivational intensity and to be active in the classroom” (Ellis, 1994, p. 515). This proposition, suggesting the positive impact of achievement on motivation, has been supported in several empirical studies (Berwick & Ross, 1989; Gardner & Tremblay, 1997; Strong, 1984; Ushioda, 1996; Tian, 2005). All of the studies indicated that L2 learners’ motivational intensity was influenced by their perceived achievement, and this perception of success in language learning did promote their language learning motivation. It appears that the relation between motivation and achievement is interactive, reciprocal and mutual. In other words, it involves a bi-directional causality. A high level. of. motivation. indeed. facilitates. language. learning,. but. perceived. accomplishment in achieving L2 goals can also help learners maintain their existing motivation and even strengthen learners’ motivational dispositions (Ellis, 1994).. 2.4 The Hypothesized Model of Motivated Language Learning Based on the literature review of previous studies, mainly drawing on the work by Tseng et al. (2006, 2008), the present study, from a expectancy-value-directed, process-oriented perspective on L2 motivation, aims to construct a model which incorporates five motivational related variables: choice motivation of language learning, self-regulatory capacity in language learning, strategic language learning involvement, strategic language learning mastery, and language achievement. The 36.

(37) hypothesized model is presented in Figure 1 to demonstrate the hypothesized causal paths among these five factors. All the paths are assigned a number to signal how the processes of the model may develop and ‘+’ signs are used to indicate an assumed positive impact. Phase 1(Predecisional). 2. +. Choice Motivation. Phase 2 (Postdecisional) Strategic Involvement 3. 1. +. + 5. Outcome. 6. Strategic Mastery. +. Self-regulatory Capacity. +. + + 7. Language Achievement. 4. + 8. Figure 1. The Hypothesized Model with Five Variables The following eight hypotheses about the relationships between the variables are further constructed based on the theoretical discussions in this chapter: Hypothesis 1. Choice Motivation of Language Learning influences Self-Regulatory Capacity in Language Learning. Hypothesis 2. Choice Motivation of Language Learning influences Strategic Language Learning Involvement. Hypothesis 3. Self-Regulatory Capacity in Language Learning influences Strategic Language Learning Involvement. Hypothesis 4. Self-Regulatory Capacity in Language Learning influences Strategic Language Learning Mastery. Hypothesis 5. Strategic Language Learning Involvement influences Strategic Language Learning Mastery. 37.

(38) Hypothesis 6. Strategic Language Learning Involvement influences Language Learning Achievement. Hypothesis 7. Strategic Language Learning Mastery influences Language Learning Achievement. Hypothesis 8. Language Learning Achievement influences Choice Motivation of Language Learning. As proposed in Chapter 1, the aim of the study is to construct a model of expectancy-value motivated language learning and the research questions are raised as follows: 1. To what extent can motivation as expectancy-value predict the demonstration of self-regulatory capacity? 2. To what extent can self-regulatory capacity influence learners’ strategic language learning mastery? 3. To what extent can strategic language learning mastery predict language learning achievement? 4. To what extent can a motivated language learning model incorporating choice motivation, self-regulatory capacity, strategic involvement, strategic mastery, and language learning achievement, be established? The eight hypotheses are thus raised to answer the questions. Hypothesis 1 is proposed to answer question 1, Hypothesis 4 question 2, and Hypothesis 7 question 3. The rest of the research hypotheses are proposed to answer the primary question 4. To sum up, the proposed model signifies eight specific hypothesized relationships between the five variables under investigation to demonstrate the possible structure of L2 motivation in EFL contexts. All of the links are evaluated by means of multiple regression and path analysis for the strength of each hypothesized path and the overall fitness of the hypothesized model to the empirical data. 38.

(39) CHAPTER THREE METHODOLOGY The study, from an expectancy-value perspective on L2 motivation, was designed to examine a hypothesized model of motivated language learning, which addresses L2 learners’ choice motivation, self-regulatory capacity, strategic language learning involvement, strategic language learning mastery, and L2 achievement. This chapter offers the research methodology of the present study, including participants, instruments, procedures and data analysis.. 3.1 Participants This section provides an introduction of the participants involved in the pilot study and the formal study.. 3.1.1 Pilot Study A total of 93 students at the second year of one public boys’ senior high school in Taipei participated in the pilot study. These participants, coming from three separate classes of natural science, used one identical version of English textbooks in Taiwan, i.e. Long-Tung textbooks, as the major English learning material and had the same English teacher, who was also the researcher of the present study. Since English is a required course generally starting from the fifth year of elementary schools in Taiwan, most of the participants have learned English at least for seven years. According to the background information provided by the learners, 34.4% of the students have learned English for seven years, 39.8% for eight to nine years and 23.5 % for more than nine years. As for the English instruction these participants received in general, as in other senior high schools in Taiwan, English, a compulsory subject of secondary education, 39.

(40) was taught with a chief focus on learners’ literacy competence, particularly in reading. Besides, learners’ achievement in English learning within a semester was principally assessed by three monthly pencil-and-paper tests which were mostly directed to learners’ reading and writing abilities.. 3.1.2 Formal Study A total of 190 students at the third year of three public senior high schools were invited to participate in the formal study. The participants, including 92 male students and 98 females, consisted of two classes of the students from a Taipei county senior high school, two from a Taipei city senior high school, and two from a Taipei city boys’ senior high school. Five of the classes belong to ‘social science’ group and the other ‘natural science’ group. In Taiwan, senior high students, depending on their willingness and dispositional preferences, are mainly grouped into three types of classes, social science, natural science, and medical science. This classification is aimed at students’ preparation for future academic majors in colleges. As for the length of English learning, the participants have learned English for more than seven years since the fifth year of elementary schools. Table 1 provides a summary of the participants involved in the formal study. Table 1. Summary of the Participants’ Background Information School Class A B C. 1 2 1 2 1 2. M 4 12 11 24 21 20. Number English English Learning (years) F Total Textbook 36 40 8.55 Far-East Series 26 38 7.74 25 36 Long-Tung 7.86 Series 11 35 8.22 0 21 Long-Tung 8.19 Series 0 20 9.10. 40.