How I see it: An exploratory study on attributions and emotions in L2 learning

Learners’ attributions have received increasing attention in second/foreign language (L2) learning. Studies have shown that how learners attribute their performance influences not only their self-efficacy, motivation, and goal attainment but also their emotions (Hsieh, 2012; Hsieh & Kang, 2010; Hsieh & Shallert, 2008; Weiner, 2000, 2014). This exploratory study investigated how Japanese adult learners of L2 English attributed changes in their L2 learning attitudes and motivation through a 10-week TOEIC preparation program. It also examined emotions expressed in their attributional statements and the differences between learners with lower and higher L2 proficiency. A content analysis of openended questionnaire responses suggested eight attributional categories: perceived L2 improvement, enjoyment, positive feelings, increased L2 exposure, realization of L2 needs and importance, effective L2 instruction, and praise from the teacher for positive changes in attitudes and/or motivation and perceived inefficient L2 skills for negative changes in attitudes and/or motivation. Enjoyment was an emotion the most frequently mentioned by both groups while other emotions, such as joy, happiness, and disappointment, were expressed only by the beginner learners. These results offer important implications for L2 pedagogy and prospects for further research in the area.


Introduction
Attributions refer to the explanations individuals give in a particular situation, generated by "a search for causality represented by a 'why' question" (Weiner, 2014, p. 355). It has been found that attributions that learners make influence their expectation of future success or failure, beliefs about their competence, intensity of work and effort, and in turn, motivation, goal attainment, and emotions (Hsieh, 2012;Hsieh & Kang, 2010;Hsieh & Shallert, 2008;Weiner, 2000Weiner, , 2014. This exploratory study was initially motivated by findings from a needs analysis, which indicated a striking relationship between emotions and attributions among adult learners of English as a foreign/second language (L2) in a 10-week preparation course for the Test of English for International Communication (TOEIC). At the beginning and end of the program, a questionnaire was distributed in order to gain information about the participants as well as their feedback on the program. Two of the items in the post-questionnaire asked students if there had been any changes in their attitudes and motivation regarding L2 learning. It also asked, in open-ended questions, what had brought about those changes. An interesting pattern was observed in the responses. While beginner learners tended to attribute improvements in their attitudes and motivation to emotional aspects of their learning experience, intermediate and advanced learners' accounts were more cognitive in nature.
Emotions and feelings are, in fact, argued to be the results of evaluations which learners carry out in a particular situation while learning (Pekrun, 2000;Méndez López & Aguilar, 2013). In general education studies, learners in academic settings have been found to experience a range of emotions such as enjoyment, hope, pride, frustration, anger, and shame (Pekrun, 2014;Pekrun, Frenzel, Goetz, & Perry, 2007;Pekrun, Goetz, Titz, & Perry, 2002). Various classroom emotions have also been discussed from an attributional theory perspective. They include happiness, admiration, liking, and surprise as success-linked emotions, and unhappiness, humiliation, embarrassment, and hopelessness as failure-linked emotions (Weiner, 2014). Despite this interplay between attributions and emotions, however, research investigating attributions in light of emotions is scarce. Moreover, in second language acquisition (SLA) studies, it is only recently that emotions have been given due attention (Dewaele, 2015;Pavlenko, 2013), with only anxiety (and to a much lesser extent, enjoyment) having been investigated to a degree (Dörnyei & Ryan, 2015).
The aim of the current study is twofold. First, it is to investigate how Japanese adult learners of L2 English attribute changes in their attitudes and motivation toward L2 learning, on completing a 10-week TOEIC preparation program. Second, it is to explore how emotions are expressed in their attributional statements and how these differ between learners with lower and higher L2 proficiency.

Emotions in learning
Emotions can be viewed as "short-lived subjective-physiological-functional-expressive phenomena that orchestrate how we react adaptively to the important events in our lives" (Reeve, 2009, p. 9), and feelings are subjective verbal descriptions of emotional experience (Reeve, 2009). In academic settings, emotions that are particularly crucial to learning are conceptualized as academic emotions consisting of four categories: achievement emotions, epistemic emotions, topic emotions, and social emotions (Pekrun, 2014). Achievement emotions are associated with success and failure resulting from achievement activities in academic settings, such as enjoyment, hope, and shame. Epistemic emotions can be provoked by cognitive challenges. The examples of these emotions are curiosity, frustration about obstacles, and delight following the overcoming of a problem. Topic emotions are emotions triggered by learners' interest, whether positive or negative, in learning materials presented in class. Social emotions are linked to teachers and peers in the classroom, such as admiration and social anxiety, and are of particular importance in teacher/student interaction and group work.
From a pedagogical standpoint, emotions are argued to have significant effects on learners' performance and achievement (Dewaele, 2015;Linnenbrink, 2007;MacIntyre & Gregersen, 2012, Pekrun, 2014Schutz & Pekrun, 2007). Dewaele (2015), for instance, posits that "emotions -both positive and negative -are the driving force behind L2 learning" (p. 14) and emphasizes the crucial role that emotions play in the language classroom. Findings from empirical studies in fact lend support to these claims (e.g., Méndez López & Aguilar, 2013;Pekrun, Goetz, Perry, Kramer, & Hochstadt, 2004;Pekrun et al., 2002). Through a series of qualitative case studies, Pekrun et al. (2002) found that students experience a rich variation of both negative and positive emotions in academic settings. They then developed a quantitative self-report instrument and measured the effects of emotions on academic achievement. The results indicated that academic emotions (i.e., enjoyment, hope, pride, relief, anger, anxiety, shame, hopelessness, and boredom) are closely correlated with students' motivation, learning strategies, cognitive resources, self-regulation, and achievement. Méndez López and Aguilar (2013) investigated the effects of emotions of Mexican-speaking students on their motivation to learn English through a qualitative approach (using narratives, journals, and interviews). Their results indicated that both positive and negative emotions contributed to enhancing and diminishing motivation and demonstrated how language learners became responsible based on their reflection on emotional experiences in language instruction.

Attribution theory and L2 learning
A theoretical notion of attribution was first introduced by Heider (1958) within the field of social psychology. The theory was further developed by Rotter (1966), who looked at causal events from internal and external dimensions. Weiner (1976), whose attribution theory has been widely acknowledged and adopted, expanded the notion by formulating a fully-fledged attribution theory model. According to his model, attributions can be characterized by three dimensions: locus, stability, and control. The first dimension, locus, concerns whether causes are perceived to be internal or external to the actor. For instance, when a student attributes his poor grades to his lack of effort, the locus is characterized as internal. On the other hand, if he perceives the poor grades as due to external factors, such as the teacher's poor instruction or learning materials, the locus is external. The second dimension, stability, is concerned with whether a cause is stable or fluctuates over time (i.e., is unstable). Effort is often regarded as unstable while aptitude and ability are typically viewed as stable. The last dimension, control, concerns the extent to which individuals believe they have control over the cause. This dimension therefore relates to judgements of responsibility. Learners' attributions can thus influence their subsequent behaviour at cognitive and affective levels (Hsieh & Schallert, 2008;Weiner, 2000).
Much research has looked into causal attributions and investigated them in relation to various constructs such as achievement (Dweck, 1975), learners' perceptions (Tse, 2000), motivation (Kálmán & Eugenio, 2015), self-efficacy (Hsieh & Kang, 2010;Silver, Mitchell, & Gist, 1995), self-efficacy and motivation (Hsieh & Schallert, 2008), proficiency , age, gender, and target language studied (Williams, Burden, Poulet, & Maun, 2004), as well as culture (Mori, Gobel, Thepsiri, & Pojanapunya, 2010). Many of these studies found the same factors to which learners attributed their success and failure. Williams et al. (2004), for instance, identified 21 attributional categories by surveying 285 secondary students learning foreign languages in the UK. Effort, ability, and interest were the three most cited elements in both positive and negative attributions. The influence of the teacher and environment were other major causes for successful and unsuccessful learning. These factors were also found in other studies in a foreign language learning context (e.g., Gobel & Mori, 2007;Mori et al., 2010;. At the same time, the attributions in each study were uniquely characterized by the context, such as culture and educational traditions, in which the participants were learning the L2. In their study of Japanese and Thai university students, Mori et al. (2010) found that students in both countries attributed their success to teachers and classroom atmosphere while they attributed their failure to lack of ability and effort. In other words, the students focused more on external factors for success and internal factors for failure. They described this tendency as self-critical and concluded that such a tendency is rooted in the non-Western cultures where interdependence is emphasized, and thus modesty and acquiescence are accepted responses. Kálmán and Eugenio (2015) investigated attributions among adult learners of L2 English in a corporate setting in Hungary and found corporate culture as one of the main causes of successful learning. Many of their participants reported that they were grateful to the organization for providing the opportunity and support for their learning. This is another example of attribution studies that highlight the impact of the contexts in which learners are embedded.

Attributions and emotions
Attribution theory highlights an important link between emotions and achievements. Weiner (2014) contends that cause-emotion relations are a crucial element in the attributional approach and that "feelings are determined by thoughts, and specifically by beliefs about causality . . . and emotions depend on 'how it seems to me'" (p. 355). Based on a large body of empirical research in the domain, Weiner (2014) discusses success-linked emotions and failurelinked emotions. The former include happiness, pride, envy, admiration, liking, disliking, gratitude, and surprise. The latter involve unhappiness, shame (humiliation and embarrassment), guilt (regret), sympathy (pity), scorn (contempt), anger, helplessness, hopelessness, hope, and schadenfreude. For example, happiness often arises following success, regardless of its cause. Then, happiness functions as a reward (pleasure) and increases achievement motivation. Gratitude is a positive emotion toward others who have volitionally contributed to the success. The public expression of gratitude elicits modesty and appreciation. As failure-linked emotions, shame is typically aroused by an attribution to low aptitude whereas guilt and regret are typically aroused by an attribution to a lack of effort. Schadenfreude is a positive self-directed affect aroused by the damage befalling another person. For example, if a student is accepted to a high rank class because of cheating and then fails an exam, fellow students could experience schadenfreude (see Weiner, 2007, for detailed descriptions of these emotions).
An examination of qualitative data in attribution studies in fact reveals various emotions expressed by L2 learners. In the study by Williams et al. (2004), for example, they reported that "interest was cited 77 times (8.0%) and included wanting to learn, liking, enjoying" (p. 22). Their participants also referred to the teacher by expressing liking and disliking. Attributions for not doing well at school included comments such as "it's boring," "I hate languages," and "I don't enjoy languages." In the study by Tse (2000), her participants expressed appreciation for having opportunities to speak the language and to participate in discussions. They also felt that teacher attention and sympathy contributed to maintaining their interest in learning the language. These learners also expressed enjoyment of taking their courses. Negative emotions associated with unsuccessful learning experience were also found. A student said in the interview, "I do not want to feel that alienated again" (p. 78).
The above literature review depicts an important relationship among attributions, emotions, and other affective constructs. Depending on the ways in which learners perceive their performance, different types of emotions are triggered. Those emotions also have an influence on other affective constructs such as motivation as well as cognitive aspects of learning and achievement. The current study seeks to further explore this interrelation. Specifically, the study investigates Japanese learners' attributions for the changes in their L2 learning attitudes and motivation, taking into account emotions articulated in their attributional statements. To that end, the following research questions (RQs) were posed: RQ1. What changes do Japanese learners report in their attitudes toward L2 learning and motivation to improve their L2 skills, over the 10 weeks of their TOEIC preparation program? RQ2. To what aspects of L2 learning experience do the learners attribute the changes? RQ3. How are emotions expressed in the attributions? RQ4. How do learners with lower and higher L2 proficiency differ in their attributions?
It should be noted that the first research question is closely related to the second research question.

Participants
The participants were 42 adult Japanese learners of L2 English, aged from their early twenties to their late fifties, working in various departments, such as human resources, accounting, and education, at the university where the program was offered. Based on their previous TOEIC scores, other English proficiency tests (e.g., Step Eiken), and self-reported proficiency levels, 28 (12 males and 16 females) were placed in a beginner class, and 14 (6 males and 8 females) in an intermediate/advanced class. The average TOEIC score of the beginner class was 271.67 (SD = 72.15), and that of the high intermediate/advanced class was 649.38 (SD = 206.19), which respectively correspond to low A2 and between B1 and B2 on the Common European Framework of Reference (Educational Testing Service, 2015). The majority of the participants had similar previous L2 learning experiences through studying English as a compulsory subject at junior high school (i.e., 3 years) and senior high school (i.e., 3 years). One student in the beginner class homestayed in Australia for 1 month, and two students in the intermediate/advanced class homestayed in Canada for 3 weeks and 1 year, respectively. One student in the latter class participated in a study aboard program for 2 years in the United States when she was in college.
Regarding their current L2 use, 18 students (64.3%) in the beginner class and six students (42.9%) in the intermediate/advanced class reported at the beginning of the program that they had no opportunities or needs to use English at all in their daily lives. Six students (21.4%) in the former group and four students (28.6%) in the latter group reported that they would use English for up to 1 hour per week, and four students in each group (14.3% for beginner and 28.6% for intermediate/advanced) would use English from 1 to 5 hours per week, mainly at work through tasks such as writing and sending emails, and answering questions from international students studying at the university.

Course description
The program was offered at a university in Southern Japan for 10 weeks. It was designed to improve employees' TOEIC scores as part of the university's globalization agenda. Each group met once a week for a 90-minute class (i.e., a total of 10 classes each). Both courses were taught by the author, using officially authorized TOEIC exercise books, TOEIC Test Koushiki Purakutisu Risuningu Hen (Educational Testing Service, 2011) and TOEIC Test Koushiki Purakutisu Riidingu Hen (Educational Testing Service, 2014a) for the beginner class, and TOEIC Test Shin Koushiki Mondaishuu Vol. 6 (Educational Testing Service, 2014b) for the intermediate/advanced class. While those exercise books were the main classroom materials with which students worked on solving TOEIC listening and reading practice questions, various communicative tasks and activities were provided as well.

Questionnaire
The data for the current study were collected by means of a questionnaire administered at the end of the course. The first question asked learners to indicate changes in their attitude toward learning English from 1 (it became very negative) to 7 (it became very positive). The next question, in an open-ended format, asked them to write about what caused the changes. Questions about motivation also followed the same pattern. The first question asked learners to indicate changes in their motivation to improve their L2 skills from 1 (I became very demotivated) to 7 (I became very motivated), and the next question asked them to write about what caused the changes. The original questionnaire was written in Japanese, and the learners wrote their response in Japanese as well. Table 1 includes the English translation of the questionnaire items. Have there been any changes in your motivation to improve your English skills through the course?
(1) I became very demotivated (2) I became demotivated (3) I became a little demotivated (4) There is no change (5) I became a little motivated (6) I became motivated (7) I became very motivated QB-2 What caused the changes?

List of emotions
All the emotions discussed as academic emotions by Pekrun (2014) and the ones elaborated on by Weiner (2014) in his attribution theory were put on a list, which is included in Table 2. This list was aimed to help discern emotions expressed in the attributional statements.

Data analysis
The learners' rating scores in QA-1 and QB-1 were imported into an Excel spreadsheet, and their means, standard deviations, skewness, and kurtosis were calculated in order to answer RQ1. Based on the responses to QA-2 and QB-2, a total of 79 (53 from beginner, and 26 from intermediate/advanced learners) statements explaining reasons for either positive or negative changes in L2 learning attitudes and/or motivation were collected. Using the numerical data from QA-1 and QB-1, those statements were first divided into two types (negative changes, positive changes) for each element (attitudes, motivation) for each group (beginner, intermediate/advanced). More specifically, statements provided by the students who chose answers 1, 2, and 3 were categorized as attributions for negative changes and those by the students who chose answers 5, 6, and 7 as attributions for positive changes. Next, the statements were content analyzed, a method of analysis in which the qualitative categories used are not predetermined but are derived inductively from the data analyzed (Dörnyei, 2007, p. 245). This analysis was aimed to answer RQ2 and RQ4. In order to explore RQ3, attributional statements were examined based on the emotion list (Table 2) to see whether and how any of the emotions on the list were articulated.

Results
First, the changes in L2 learning attitudes and motivation will be presented. Next, the categories of attributions emerging from the content analysis will be reported followed by the emotions found in the attributions. Lastly, the differences in attributions between beginner and intermediate/advanced learners will be reported. Table 3 shows descriptive statistics for each group's scores for attitudes and motivation obtained from QA-1 and QB-1 in the questionnaire, in which learners were instructed to indicate changes in their attitudes toward L2 learning and motivation to improve their L2 skills, based on the seven possible response choices for each item. For both attitudes and motivation, 1 indicated that they had changed very negatively, and 7 indicated that they had greatly improved, with 4 indicating no change. As shown, the majority of the students in both groups (75% of the beginner and 85% of the intermediate/advanced learners) reported their L2 learning attitudes had improved. While five students in the beginner group indicated their L2 attitudes had become negative, none of the students in the intermediate/advanced group reported such negative changes. Two students in each group reported no changes in their attitudes. Similarly, the majority of the students in both groups (89% of the beginner and 84% of the intermediate/advanced learners) reported their increased motivation, and none of the students in the two groups reported decreased motivation. Three beginner learners and two more advanced learners indicated no change in their motivation level.  Table 4 presents the number of attributional statements categorized by two directions of the changes (i.e., positive or negative) in attitude and motivation by each group, based on the numerical data obtained from QA-1 and QB-1 in the questionnaire (Table 3). The content analysis of those statements suggested eight attributional categories: perceived L2 improvement, enjoyment, positive feelings, increased L2 exposure, realization of L2 needs and importance, effective L2 instruction, and praise from the teacher for positive changes in attitude and/or motivation, and perceived inefficient L2 skills for negative changes in attitude and/or motivation. Table 5 includes the categories along with representative extracts translated from Japanese into English. Of the 78 attributional statements, improved L2 skills that the learners perceived during the program was the most commonly cited element to explain the positive changes. For example, one student wrote, "I feel more motivated to study English because I noticed that my listening skill got better." Many learners displayed enjoyment that they experienced through learning as the reason for the positive changes. Learners also attributed their improved attitudes and motivation to various positive feelings. For instance, one student wrote, "I felt happy when I knew the answer, so I feel more positive about studying English." In the attributions termed increased L2 exposure, learners explained that whereas they hardly had any chance of using English in their daily life previously, the program gave them such opportunities, and that had a positive influence on their attitudes or motivation. Students referred to their better understanding of learning needs and goals as the reason for their improved attitudes or motivation. Effective L2 instruction they received in the class was also mentioned by several students. Other students referred to the positive feedback they received from the teacher, and this component was typically mentioned with gratitude. In the statements attributing the negative changes in attitudes and motivation, the most commonly cited element was inefficient L2 skills that learners perceived while studying. Due to the small number of attributional statements given for the negative changes, this was the only category that emerged from the analysis.

Types of emotions
As illustrated in the categories above, enjoyment was the type of emotion most frequently cited in learners' responses to the open-ended questions. However, there was a variation regarding which aspects of the class students reported to have enjoyed. While many referred to their classes in broad terms, as in "I enjoyed the classes" or "the classes were so much fun," some discussed more specific elements such as writing and conversation activities. A few wrote that they felt enjoyment from learning new things or being able to understand English better. In attributional statements, which were mostly categorized as positive feelings, various affective concepts were expressed with the Japanese term kimochi 'feeling.' Some described a feeling of being refreshed by studying for the first time in a long time, and some described an uplifting feeling evoked by participating in class after long hours of working. Three participants indicated their negative feelings toward English had decreased. One used the Japanese term kennokan, which is best translated as hatred, saying that her old hatred toward English had softened. One participant referred to happiness: "I felt happy when I knew the answer." A few noted that the program had raised their curiosity and made them feel like studying more. Several students expressed hope for their further L2 improvement. Gratitude was another emotion expressed in statements about receiving praise from the teacher.
Unlike the statements relating to positive changes, few emotions were expressed in those relating to negative changes. One participant wrote that she was shocked to realize how poor her English was, using the Japanese term gakuzentosuru. The word, describing a strong negative sense of surprise, depicts a feeling of shock, disappointment, and even shame. Another student described his low L2 proficiency with the Japanese word tsuukan, which captures a feeling of pain, possible sadness and disappointment.

Differences between beginner and intermediate/advanced learners
Since the study did not obtain enough data on attributions for negative changes, discussions regarding differences between the two groups will focus on the attributions for positive changes only. Table 6 presents the number of attributional statements in each category. Positive changes in attitudes and motivation for the two groups are further illustrated in Figures 1 and 2, respectively. The most notable difference between beginner learners and intermediate/advanced learners was that certain categories appeared only in either one of the groups. Increased L2 exposure, which was most cited by the intermediate/advanced learners as the reason for positive changes in attitudes (45%) and motivation (25%), was not mentioned by the beginner learners at all. Positive feelings constituted nearly one fourth of the attributions for positive changes in attitudes (20%) and motivation (25%) among the beginner learners but were not expressed among the intermediate/advanced learners at all. The same applied to praise from the teacher.
Another striking difference between the two groups was the proportion of affective characteristics in the responses. As discussed above, various emotions and feelings were identified in the attributional statements categorized as enjoyment, positive feelings, and praise from the teacher. While more than half (56%) of the attributional statements in the beginner group contained affective characteristics, such statements constituted only 17% in the intermediate/advanced group.  Figure 2 The proportion of attributional statements in each category for motivation

Discussion
The first research question concerned the changes adult Japanese learners of L2 English reported in their attitudes toward L2 learning and motivation to improve their L2 skills over the 10 weeks of their TOEIC preparation program. The majority of the students in both the beginner and intermediate/advanced groups reported that their attitudes had become more positive while several students in both groups indicated no improvement, and three students in the beginner group reported that their attitude had become slightly negative. With regard to motivation, all of the beginner learners and most of the intermediate/advanced learners reported that they felt more motivated while a few in the latter group indicated no change. Reasons for those changes were provided in the openended questions, which were used to investigate the second research question. The second research question looked into the aspects of L2 learning experience to which the participants attributed changes in their L2 learning attitude and motivation. The content analysis of learners' responses to the questionnaire suggested eight attributional categories: perceived L2 improvement, enjoyment, positive feelings, increased L2 exposure, realization of L2 needs and importance, effective L2 instruction, and praise from the teacher for positive changes, and perceived inefficient L2 skills for negative changes in attitudes and/or motivation. These categories share some similarities with those found in other attribution studies. As discussed, the role of teachers was one of the most frequently cited factors influencing success (Mori et al., 2010;Tse, 2000;Williams et al., 2004). Similarly, the learners in the current study referred to the teacher's effective L2 instruction and praise as the reasons for their improved learning attitudes and motivation. At the same time, these attributions uniquely reflect characteristics of the learners in this study, as demonstrated by other studies (e.g., Kálmán & Eugenio, 2015;Mori et al., 2010). The two factors, increased L2 exposure and realization of L2 needs and importance, appear to have had an impact because of the context in which the learners were embedded. As discussed in the course description, most of the participants were enrolled in the program as the request of their departments. At the beginning of the program, many of them reported that they did not use English for work, let alone in their daily lives. They probably did not possess a strong need for their L2 improvement, either. However, once the program started, the opportunity to use the L2 notably increased. This increased exposure by itself was influential enough for the participants to change their learning attitudes and motivation. During the program, the participants were also introduced into various ways in which L2 skills can enrich their lives. This could have led them to further realize the importance and value of possessing good L2 skills. These elements can also be characterized as unstable based on Weiner's attributional dimensions. Given that the program was offered only for the limited time, the L2 exposure can easily decrease together with the needs of L2 skill improvement unless learners maintain their motivation and continuously seek out such L2 opportunities.
Another interesting characteristic can be found in perceived L2 improvement, the most cited element to explain positive changes in learning attitudes and motivation. In attribution theory and research, ability is typically characterized as uncontrollable and stable. At the same time, ability can be viewed as controllable and unstable based on implicit theories (Dweck, 1999). The theories propose two types of mindsets and posit that learners with a growth mindset believe that their abilities can be developed through effort and hard work, as opposed to those with a fixed mindset who perceive ability as a fixed trait. It has been argued that when learners are equipped with an incremental view of ability (i.e., a growth mindset), they become motivated to seek out challenges and opportunities to learn (see Dweck, 1999 for the motivational model of achievement). Based on this notion, the result that the learners in this study reported their L2 improvement indicates that they most likely thought their ability can be enhanced even by a short period of time of practice, leading them to feel more motivated to improve their L2 skills. This finding also lends support to the result of a study on adult learners' attributions by Kálmán and Eugenio (2015), which found that ability significantly contributed to motivation.
The third research question concerned how emotions were expressed in the attributional statements. So many emotions were articulated that several categories emerged (enjoyment, positive feelings, and praise from the teacher). Among various emotions and feelings articulated in the statements, enjoyment was the most notable emotion referred to by many learners. Among the beginner learners it was the most and second most cited reason for improved L2 learning attitudes and motivation, respectively. This result corroborates the findings from previous studies (Pekrun et al., 2002;Pekrun et al., 2004), in which enjoyment correlated positively with learning motivation, and is in line with the notion that positive emotions such as enjoyment of learning most likely increase interest and strengthen motivation . Enjoyment is also one of the academic emotions discussed by Pekrun (2014) as an example of achievement emotions (see the literature review in Section 2).
Other emotions found in the attributional statements, such as happiness, curiosity, hope, hopelessness, disliking, and disappointment, are also among those referred to within the attributional perspective on emotions and/or academic emotions. This finding shows the powerful role that emotions play during the L2 learning process, aligning with the discussion of significant effects of emotions on learning voiced by many scholars (e.g., Dewaele, 2015;Linnenbrink, 2007;MacIntyre & Gregersen, 2012;Pekrun, 2014;Schutz & Pekrun, 2007).
The fourth research question concerned the differences in attributional statements between learners with lower and higher L2 proficiency. It was found that certain categories only emerged in one of the groups. Increased L2 exposure, which was the most cited reason for the improvement of both L2 learning attitudes and motivation among the intermediate/advanced students, was not mentioned by the beginner students. This finding may be accounted for in light of L2 metacognition and autonomy. In essence, metacognition refers to "individuals' awareness and management of their learning process" and consists of two components: metacognitive awareness and metacognitive strategies (Raoofi, Chan, Mukundan, & Rashid, 2013, p. 37). The former refers to information which learners possess about their own learning, while the latter are skills with which learners manage and guide their own learning (Rahimi & Katal, 2012). Previous research has indicated that possessing and utilizing metacognitive awareness and knowledge helps learners to be successful in L2 learning (Cotterall & Murray, 2009;Raoofi et al., 2013). It has also been argued that developing metacognitive awareness of one's own language learning ultimately facilitates autonomous learning (for further discussion of autonomous learning, see, e.g., Reinders, 2011). Taking into consideration these views, it can be speculated that the intermediate/advanced learners in this study were more metacognitively developed and able to make effective use of the resources available to them. The beginner learners, on the contrary, may not have had such knowledge or skills so that they did not benefit as much from mere exposure and increased access to the L2, which, in turn, may have resulted in little influence on their learning attitudes or motivation.
In a similar vein, praise from the teacher, which only appeared in attributions by the beginner learners, may not have had such an impact on the intermediate/advanced learners, who may have been used to being acknowledged or praised for their good L2 skills.
Another remarkable difference found between the two groups was the proportion of affective attributions. In explaining reasons for the improvements of their L2 learning attitudes and motivation, the beginner learners expressed a number of emotions such as joy about learning new things, happiness about getting their answers correct on the TOEIC questions, gratitude for the teacher's praise, and feeling refreshed from studying English for the first time in a long time, which added up to 56 % of their attributional statements. On the other hand, enjoyment was the only emotion expressed by the intermediate/advanced students, and it only constituted 17% of their attributional statements. Several explanations can be offered. In Weiner's (2014) attribution perspective on emotions, happiness is experienced when a learner is successful, regardless of the cause of their success (p. 357). As illustrated in Figures 1 and 2, perceived L2 improvement was one of the most cited attributional categories among the beginner learners. When one considers this improvement as success (or successful L2 learning), it may be natural for those who recognized their achievement to express positive emotions. However, this account may not be sufficient to explain why enjoyment constituted a smaller proportion of the intermediate/advanced learners' attributional statements, and why other emotions were not identified. It is possible that the number of participants in this group was too small to obtain any other emotions. Another possible interpretation is that it may be simply natural for beginners to have emotional reactions first. As such individuals make progress and expand their knowledge and skills, they become more aware of cognitive aspects of their learning. In other words, those who are more advanced are more conscious of and thus likely to be affected by cognitive elements in their learning, whereas beginner learners, not yet having built metacognitive awareness or knowledge, may be inclined towards affective aspects of L2 learning, which results in changing their perceptions based on feelings and emotions.

Conclusion
The current study sought to investigate how adult L2 learners attributed changes in their L2 learning attitudes and motivation during a TOEIC preparation program. It was found that the beginner learners tended to attribute improvements in their attitudes and motivation to emotional aspects of their learning experience while the intermediate and advanced learners' accounts were more cognitive in nature. This finding highlights the important role that emotions play in L2 learning, particularly among beginner learners.
Several limitations of this study must be noted. The first has to do with its small sample size, which produced a relatively small number of attributional statements to analyze, particularly those from learners with higher L2 proficiency as well as those relating to negative changes. The exploratory nature of this investigation may also face criticism, especially in light of attribution research, which by and large studies attributions within a success-failure framework. Another issue is concerned with the absence of another researcher to assist with the content analysis. By investigating emotions expressed in attributions, the current study treated emotions as a result of evaluations, that is, cognitive appraisal as the antecedent of emotions. Yet, the author acknowledges that the causal relationship between appraisals and emotions is not always linear and uni-directional and that there can be a reverse causal relationship, where emotions and/or mood influence information-processing, as suggested by the mood-congruence effect (e.g., Johnson & Tversky, 1983) and the affect-infusion model (AIM; e.g., Forgas, 1995).
Based on the limitations discussed above, together with the findings from the current study, several suggestions can be made for further research. A larger sample will provide richer accounts of the differences observed between learners with lower and higher L2 proficiency. It will also enable the collection of more data to form other attributional categories for negative changes in L2 learning attitudes and motivation. Alternative theoretical perspectives, such as communicative or transactional approaches (Parkinson, 2009), may offer insight into the roles and functions of emotions in L2 learning. Another promising approach in the emerging field of positive psychology is to consider emotions through the lens of human strengths, such as resilience, courage, and gratitude (e.g., Seligman & Csikszentmihalyi, 2014;Snyder, Lopez, & Pedrotti, 2010).
The results of the present study, though limited, suggest implications for L2 pedagogy, particularly in test-related EFL programs. Administrators and instructors in such programs by and large pay greater attention to cognitive aspects of L2 learning, for example, what language features are to be taught, in what order, and in what ways. While these are all important, clearly affective factors, such as emotions, do play a critical role in L2 learning, especially among learners with lower L2 proficiency. As suggested by the current study, providing learning experiences whereby students feel joy and enjoyment is as important as equipping them with new L2 knowledge or test-taking strategies in increasing motivation and improving attitudes, and, in turn, assisting their further L2 development.