Advanced learners’ foreign language-related emotions across the four skills

Individual differences researchers have recently begun to investigate the concept of emotions and their role in language learning (MacIntyre, Gregersen, & Mercer, 2016). Our aim is to report on a project exploring English majors’ feelings related to their use of foreign languages. Using a qualitative research design, participants were asked to write a paragraph in their mother tongue (Hungarian) describing their emotional experiences in connection with foreign languages and one of the four language skills. Our database comprised altogether 166 paragraphs from 31 male and 135 female students, with 43 texts on listening, 35 on speaking, 47 on reading, and 41 on writing. With the help of content analytical techniques, the texts were divided into thematic units and coded by the two authors. A framework of academically-relevant emotions (Pekrun, 2014) was used to guide our initial coding and the categories were modified where it was felt necessary. Results indicate that the two emotions most frequently experienced by English majors are predominantly related to enjoyment and language anxiety, and these emotions vary not only according to the skill involved but also depending on the context of language use (in class or outside class).


Introduction
In individual differences research, the notion of affect was mostly used to refer to phenomena (e.g., attitudes) judged as relevant in connection with motivation (Gardner & MacIntyre, 1993).Also, since early motivational research acknowledged the negative effects of anxiety (Gardner, 1985), language learning anxiety emerged as the most prominent affective variable in studies of second language acquisition (Horwitz, Horwitz, & Cope, 1991).In recent years a more explicit focus on different emotions present in the classroom has emerged, particularly since the turn of the millennium (Dewaele, 2005(Dewaele, , 2010;;MacIntyre, 2002).Around that time, an important change also started in the discipline of psychology: in an influential paper, Seligman and Csíkszentmihályi (2000) argued for introducing the framework of positive psychology.They suggested that, instead of the focus on pathology that characterized psychology for quite a long time, there was a need to concentrate on positive experiences in people's lives.This new approach has also found its way into our field (MacIntyre & Mercer, 2014), which is well reflected by the fact that two books were published on the topic quite recently (i.e., Gabryś-Barker & Gałajda, 2016;MacIntyre, Gregersen, & Mercer, 2016).Since positive emotions are one of the focal points of positive psychology (Seligman & Csíkszentmihályi, 2000), this trend also helped to broaden the scope of affective studies to include positive emotions experienced in the classroom.In the educational literature, there is an increasing interest in emotions experienced in connection with academic success and in different academic subjects (Pekrun & Linnenbrink-Garcia, 2014;Schutz & Pekrun, 2007), while in the field of applied linguistics, language learning related enjoyment has gained attention recently (Dewaele & MacIntyre, 2014, 2016).However, there seems to be limited research to date that aims to provide a comprehensive picture of the range of emotions students experience in connection with foreign language learning and use in different language learning contexts (for an exception, see MacIntyre & Vincze, 2017).In our exploratory study, our aim is to explore English majors' feelings related to their use of foreign languages in and out of the classroom.

What are emotions?
Emotions, similarly to many other concepts of psychology, are notoriously difficult to grasp scientifically.If we asked people around us whether they knew what emotions were, we could safely guess that most of them would answer with a definite "yes," and they would even offer us examples.However, it is much less likely that they could provide a reasonably comprehensive definition or that their definitions would overlap.If we turn to the discipline of psychology, where the study of emotions has quite a long research tradition, we still find some controversy over the definition of emotions.
Arriving at a generally accepted scientific definition of emotions has not been an easy feat.Izard (2010), a renowned researcher of basic emotions, even devoted a small-scale study to the topic.He asked 34 distinguished scientists to define the term emotion, and although he found some agreement on the structures and functions of emotions, he was unable to provide a complete synthesis of the definitions he received.Nevertheless, all the definitions acknowledged the multi-componential nature of emotions, and many of them "gave a definition of emotion that recognized (a) neural circuits and neurobiological processes, (b) phenomenal experience or feeling, and (c) perceptual-cognitive processes as aspects of emotion" (p.368).The definition adopted in our study also acknowledges the multiple aspects that are present in emotions besides drawing attention to the evolutionary role of emotions in adapting to our environment in accordance with our personal goals.Therefore, emotions in our study are defined as "multifaceted responses to events that we see as challenges or opportunities in our inner or outer world, events that are important to our goals" (Keltner, Oatley, & Jenkins, 2014, p. 27).
Having arrived at a working definition, it might also prove useful to differentiate the term emotion from other related terms like affects, moods, and dispositions.Keltner et al. (2014) used the dimension of time to differentiate these concepts.In their interpretation, self-reported emotions typically last between a few minutes and a few hours, whereas moods last at least for a few hours or even longer, for days or even weeks.Frijda (1993) pointed out another difference between emotions and moods, stating that the former are typically directed at specific people or events, that is, an object, while moods are freefloating and objectless in this sense.The words dispositions, sentiments and traits describe the emotional aspects of someone's personality, which can last a lifetime; in this sense, their temporal range is even longer (Keltner et al., 2014).The terms affective or affect tend to refer to the whole range of phenomena that have been described above (Oatley, 2004).
Besides definitional problems, emotions also pose problems of classification.The words used to describe emotions in different cultures and in different languages tend to vary considerably, and their interpretation might even differ from one person to the other.One way out of the classification problem might be reducing distinct emotions to affective dimensions and viewing them as varying on two dimensions: arousal/activation and valence/pleasantness (Larsen & Fredrickson, 1999).However, considering emotions as dimensions tends to mask the fact that "discrete emotion variables have different predictors and contribute to different behavioral outcomes" (Izard, 2007, p. 267), and these approaches also dismiss evidence of the functionality of discrete emotions outlined by Izard (2007).Moreover, people tend to label their emotions as distinct categories, which seems to underline their psychological reality (Field, 2003).Therefore, another solution to the problem of classification would be to focus on categories of distinct emotions and come up with lists of primary emotions based on biological considerations, as has been attempted by a number of researchers (Reeve, 2009).
One such list of primary, basic or first-order emotions is provided by Izard (2010); in his view, (a) interest, (b) enjoyment/happiness/contentment, (c) sadness, (d) anger, (e) disgust, and (f) fear belong to this group.Although this list is not very extended, the problem is that after early childhood, people do not really experience these first-order emotions in their pure forms.As we grow up, cognition starts to play a decisive role in our emotions; consequently, what we usually refer to as emotions should be more precisely described as emotion schemas.Emotion schemas are defined "in terms of the dynamic interaction of emotion and cognition" (Izard, 2007, p. 265); therefore, besides feelings derived from primary emotions, they also have influential cognitive aspects.Variation in the cognitive aspects can be held accountable for the differences in cognitive schemas, posing potential problems for research.However, since after early childhood emotional schemas can be considered as the most powerful source of human motivation (Izard, 2007), the importance of cognition in forming emotion cannot be disregarded.
The relationship between feelings and cognition is reciprocal.On the one hand, our emotions arise as a result of the process whereby all events are evaluated in line with the individual's concerns, and this evaluation is called appraisal (Lazarus, 1991).The first evaluation process, called primary appraisal, is automatic and reflex-like.It is during the process of secondary appraisal that the person's cognitions have a major role.These cognitions can be labeled attributions (Keltner et al., 2014) or schemas (Izard, 2007), and they will determine the quality of the emotion experienced by the person and eventually the label attached to it.On the other hand, feelings are also known to influence people's cognitions.For example, individuals are more attuned to perceiving emotioncongruent objects and events.Emotions help people prioritize their attentional resources and assist in concentrating on those events and objects that are relevant to whatever they are doing at the moment.When recalling the past, events infused with intense emotions prove to be easier to recall and these events are also recalled more often (Keltner et al., 2014).
Another important question concerns the functions of emotions.Emotion and cognition are intimately connected, and seem to influence each other extensively, but their relationship can be seen in a different light.Although there was a time when emotions were primarily considered as nuisances, disrupting the rational working of the mind (Hebb, 1949;Mandler, 1984), going back even further in time, we find that Darwin (as cited in Reeve, 2009) argued as early as 1872 that emotions help animals adopt to their environment.Today, we also attribute an important role to emotions in coping with fundamental life tasks, such as protection, destruction, reproduction, reunion, affiliation, rejection, exploration, and orientation (Plutchik, 1980).Emotions also serve a social function in the sense that they communicate our feelings to others, they influence how we are interacted with, they invite and facilitate social interaction, and they have a role in establishing, maintaining and dissolving relationships (Manstead, 1991).Consequently, emotions indeed seem to assist adaptation to our physical and social environment and ensure our survival.Therefore, Oxford (2015) seems to be right when arguing that "all learning is a powerful combination of cognition and emotion" (p.371).
Given their fundamental role and their intimate relationship with cognition, it is quite surprising that emotions have largely been ignored in educational contexts until recently.The exception to the rule is anxiety, whose effects on testtaking (Cassady & Johnson, 2002;Spielberger & Vagg, 1995;Zeidner, 1998) and foreign language learning (Gkonou, Daubney, & Dewaele, 2017;Horwitz et al., 1991) have been explored extensively.The recent shift in focus, to other emotions and positive emotions in particular, can be attributed to the emergence of positive psychology (Seligman & Csíkszentmihályi, 2000) and its focus on positive emotions, for example.Fredrickson's (2003Fredrickson's ( , 2008) ) theory hypothesizing the crucial role of positive emotions found its way into research on second and foreign language acquisition (Dewaele & MacIntyre, 2014;MacIntyre & Gregersen, 2012;MacIntyre & Mercer, 2014;MacIntyre & Vincze, 2017, Oxford, 2015).More specifically, Fredrickson's (2003Fredrickson's ( , 2008) ) broaden-and-build theory postulates that negative emotions tend to lead to characteristic thoughts and actions; thus, they result in focusing and narrowing.In contrast, positive emotions "all share the ability to broaden people's momentary thought-action repertoires and build their enduring personal resources, ranging from physical and intellectual resources to social and psychological resources" (Fredrickson, 2003, p. 219).In this sense, positive emotions are broadening rather than narrowing, and they assist building different resources; as such, they are considered quite beneficial in educational contexts.

Emotions in the educational setting
Besides research inspired by the role of positive emotions in general, the effects of distinct emotions have also been investigated within the educational context.Pekrun (2014) argued that emotions, especially those relevant in a school setting, can be grouped into four categories.Achievement emotions relate to feelings connected to success and failure in school; examples include enjoyment of learning, hope and pride related to success, and anxiety and shame related to failure.Another group of emotions relate to cognitive problems encountered while learning; these are labeled epistemic emotions, such as surprise, curiosity, confusion and frustration.The third group of emotions relate to the specific topics that students deal with in the lessons, that is, the content of learning.Topic emotions include instances when students feel empathy towards characters they read about, or when they are disgusted or anxious or feel enjoyment or interest about certain topics.The last group of emotions described are social emotions relating to teachers and peers in the classroom.Examples of such emotions are love, sympathy, compassion, admiration, contempt, envy, anger or social anxiety.
Of these four groups of emotions, the one that has been studied extensively is that of achievement emotions.In their control-value theory of achievement emotions, Pekrun, Frenzel, Goetz and Perry (2007) use a three-dimensional taxonomy of achievement emotions.The first dimension refers to valence (positive/pleasant vs. negative/unpleasant quality), the second to activation (activating vs. deactivating tendency), while the third describes whether the focus of achievement is the activity itself (intrinsic emotion) or the outcome (extrinsic emotion) (Pekrun, Götz, Titz, & Perry, 2002a).In this framework, enjoyment is a positive, activating and activity focused emotion with relaxation as its deactivating counterpart.Anger and frustration are negative and activating emotions, and they are also activity focused, while boredom is their deactivating counterpart.Outcome-focused emotions are more numerous: joy, hope, pride and gratitude are positive and activating, while contentment and relief are positive but deactivating.On the negative side we find anxiety, shame and anger as activating, whereas sadness, disappointment and hopelessness are deactivating emotions.Outcome-focused emotions can be further divided into prospective anticipatory emotions (i.e., hope for success, anxiety over failure) and retrospective emotions (i.e., pride or shame experienced after receiving evaluation) (Pekrun, 2006).In order to measure these achievement emotions in the classroom, Pekrun, Goetz and Perry (2005) developed the Academic Emotions Questionnaire (AEQ), and this instrument has been used in several quantitative studies (e.g., Frenzel, Thrash, Pekrun, & Goetz, 2007;Pekrun, Elliot, & Maier, 2009;Pekrun, Goetz, Frenzel, Barchfeld, & Perry, 2011).
The control-value theory of achievement emotions "stipulates that individuals experience specific achievement emotions when they feel in control of, or out of control of, achievement activities and outcomes that are subjectively important to them, implying that control appraisals and value appraisals are the proximal determinants of these emotions" (Pekrun et al., 2007, p. 16).This means that, on the one hand, achievement emotions experienced by individuals are going to depend on the value, positive (success) or negative (failure), that they attribute to the activity.Values can be intrinsic if they are related to the activity itself, or extrinsic if they are connected to the instrumental utility of the activities to produce the desired outcomes.On the other hand, achievement feelings also depend on the extent to which learners feel in control of the learning situation and their achievements, and the extent to which they are able to attribute success or failure to their own efforts or abilities.
In studies linking academic achievement to achievement emotions, fairly consistent findings have emerged.Positive activating emotions such as activityrelated enjoyment and outcome-related hope and pride have been found to be related positively to students' academic achievement, while negative deactivating emotions such as hopelessness and boredom are linked to negative achievements (Pekrun, Götz, Titz, & Perry, 2002b).The effects of deactivating positive emotions, such as relaxation, and of negative activating emotions, such as anger, anxiety and shame are hypothesized to be more complex.For example, although Pekrun et al. (2002b) found that anger, anxiety and shame are negatively correlated with achievement, in a diary study investigating individual students' achievement emotions before and after their final exams, Pekrun and Hoffmann (as cited in Pekrun et al., 2007) found both negative and positive correlations.
Recently, in the field of second and foreign language learning, there have also been some studies which attempted to offer a more comprehensive view of emotions and move away from anxiety as their exclusive focus.In a largescale questionnaire study involving 1746 participants, Dewaele and MacIntyre (2014) found that while learning languages, learners experienced foreign language enjoyment more often than foreign language anxiety and argued that these two emotions should be treated not as endpoints of a single continuum but as two different dimensions.This means that enjoyment is more than simply the lack of anxiety and that these two emotions can even be experienced together.Ross and Stracke (2016) analyzed learners' perceptions of pride in an interview study and found that, even within this single emotion, different dimensions can be identified which can also be related to the different learning contexts.
Instead of concentrating on just one or two emotions, MacIntyre and Vincze (2017) attempted to link 10 positive and 9 negative distinct emotions to models of motivation (which included language anxiety as a motivational construct) on a sample of Italian language learners studying German in a high-contact context.They found that the correlations between positive emotions and motivation-related variables are positive and strong, except for the correlations with anxiety (which not surprisingly are negative in sign).At the same time, correlations involving negative emotions are weaker and show a less consistent pattern, although they are mostly negatively correlated with motivation-related variables, with the exception of anxiety where the correlations are positive.Their regression analyses, which attempted to predict motivation with the help of emotions, showed that motivation-related variables can be best predicted by a combination of positive and negative emotions, thereby drawing attention to the dynamic interplay of the two.They found that Fredrickson's (2013) positivity ratio, which instead of looking at positive or negative emotions alone examines if positive emotions are prevalent in language learners' experiences, is a measure that correlates well with language learning motivation.
Another questionnaire study conducted by Pishghadam, Zabetipour and Aminzadeh (2016) attempted to link the achievement emotions explored by the AEQ to the four skills in an Iranian context.Although it is quite interesting that different language skills tend to be linked to different groups of emotions (e.g., listening is linked to all of the examined negative emotions while speaking seems to trigger mostly positive emotions), the closed format of the questionnaire raises the question whether these are indeed the only or even the most relevant emotions that students experience in connection with the four skills.Since studies exploring the range of possible emotions that could be experienced in connection with language learning and language use are clearly missing, we attempted to fill this research gap with our study.Pekrun (2014) acknowledged that there are other potentially important emotions influencing learning besides the achievement emotions he explored, so we felt the need to have an extended scope for our research and to investigate additional categories of academically-relevant emotions (Pekrun, 2014).Language learning is different in nature from learning other academic subjects, since classes do not necessarily have a set content as language is primarily the medium of communication (Coyle, Hood, & Marsh, 2010;Stryker & Leaver, 1997) and language learning also involves aspects of culture and personal identity (Dewaele, 2016;Dörnyei, 2005;Gardner & Lambert, 1959) in ways other school subjects do not; therefore, there is a possibility to adjust the content of classes to learners' interests.Also, social factors play a more dominant role in language learning and language use than in other subjects (Gardner & Lambert, 1959), as the primary aim of language instruction is to enable students to communicate in the foreign language and interact in social situations.Therefore, we set out to explore the full spectrum of emotions experienced by learners in and outside the classroom by asking them to describe their feelings in various language-related learning contexts.

Context of our study
The context of our study was a prominent university in Budapest, Hungary, where one of the most populous majors is English studies.In the 2015/16 academic year, 348 students commenced their studies in the English BA program.During their three years of study, students are obliged to take language development classes, courses specifically aimed at developing their academic reading and writing skills, courses involving extensive listening to lectures, and extensive reading in preparation for content classes in literature, culture, and history.Generally, we can say that the complete spectrum of English language skills is practiced at some point in the curriculum, with different skills being emphasized in different courses.Since academic emotions tend to be examined in domain-specific contexts (Goetz, Frenzel, Hall, & Pekrun, 2008), it is quite possible that different language emotions will emerge in association with different skills.
The students enrolled in English studies tend to have a foreign language proficiency level of around B2 level according to the Common European framework of reference (Council of Europe, 2001).They generally also have had experience of learning a second foreign language in high school, as it is compulsory according to the Hungarian national core curriculum (National Curriculum, 2012).Having gone through the system and having experienced both compulsory learning of languages as well as optional opportunities (choosing a language major) led us to believe that exploring the emotions related to language learning of this target group would help us gain insight into the role affect plays in the course of language learning.

Research questions
Based on the above, the research questions guiding our study were the following: a. What kind of emotions do English majors experience in connection with their use of a foreign language and the four language skills?b.Are there any differences in the types of emotions language learners tend to experience with regard to the four skills?c.Are there any differences in the types of emotions language learners tend to experience with regard to their year of study at university? d.Are there any differences in the types of emotions language learners tend to experience in-and outside the foreign language classroom contexts?

Method
In order to investigate the research questions formulated above, we took a primarily qualitative approach in our study.We hoped that this type of design would yield rich, descriptive and insightful results.Through purposive and convenience sampling, 166 English majors from a large Hungarian university (31 male and 135 female students; 93 first-year, 17 second-year, and 53 third-year students -data for three participants was missing) were selected to take part in our study.The mean age of the participants was 20.9.These students were purposefully selected because they were guaranteed to have ample foreign language learning experience in order to provide us with information concerning their emotions while learning languages.The sample was also conveniently chosen as they comprised learner groups where the teachers offered to assist us in our project.
Data was collected during different university seminars, where participants were asked to write a paragraph in their mother tongue (Hungarian) describing their emotional experiences in connection with foreign languages and one of the four language skills.The instructions read as follows: "In a paragraph, please describe how you feel when speaking/writing/listening/reading in a foreign language in and outside the classroom."Altogether our final pool of data consisted of 43 texts on listening, 35 on speaking, 47 on reading, and 41 on writing.
In order to analyze the narratives, we used the Atlas.tiv. 7.5.17software.We adopted a deductive qualitative content analysis approach (Elo, Kääriäinen, Kanste, Pölkki, Utriainen, & Kyngäs, 2014) with Pekrun's (2014) theoretical framework in mind.In our analysis we also opted for Strauss and Corbin's (1990) approach to qualitative data analysis with three phases of analysis: open coding, axial coding and selective coding.First, through open coding (Strauss & Corbin, 1990), we compiled a coding scheme that consisted of 75 codes.These codes were emotion words (in English) adapted from Pekrun's (2014) list as well as English translations of emotion words appearing in the Hungarian texts as participants used them.Thus, some of the codes emerged from the texts themselves (see Strauss & Crobid, 1990).After analyzing the Hungarian texts and reaching a point of saturation, we ended up with 1352 coded segments.The intercoder reliability coefficient of Cohen's kappa = .91provided an indication of the conformability (Elo et al., 2014) and trustworthiness of our data analysis.Next, through axial coding (Strauss & Corbin, 1990), we looked at the context of the 75 codes related to emotions and tried to find links between them and group them into larger categories accordingly.Finally, through selective coding, we looked at the main categories of emotions that seemed to account for most of the variation in our dataset (see Strauss & Corbin, 1990).Specifically, we counted the frequencies with which the main types of emotions were mentioned across the four skills, across university years, and linked to in-and outside classroom contexts.Finally, we investigated these differences using non-parametric techniques with the help of the Statistical Package for the Social Sciences (SPSS) 16.0 for Windows software.

Results and discussion
In presenting our findings, we will focus on the results of the axial and selective coding processes, namely, we will discuss the most salient categories of emotions and try to establish their link to contextual factors on an abstract level.For the sake of anonymity, the letters and numbers in brackets after the quotations depict the participants' identification numbers.
With regard to Pekrun's (2014) framework, a clear dividing line appeared between the positive and negative type of achievement, epistemic, topic and social emotions.Participants mentioned positive achievement emotions concerning foreign language learning, such as pride, contentment, comfort, relaxation (cf.Pekrun et al., 2007), generally positive feelings related to success, feeling the lack of anxiety, and the feeling of enjoyment linked to academic accomplishments.An example of such emotions is the following: "When I communicate with native speakers in writing, I feel proud that I am able to use this language" (P83).
Although not a primary emotion (see Izard, 2007), based on our data, it made sense to include the feeling of freedom in all four groups of positive academic emotions (achievement, epistemic, topic, and social emotions), respectively, since participants mentioned it relatively frequently (11 times in total) in connection with a variety of language use situations.Based on the above theorizations, the feeling of freedom can be viewed as an emotion schema (Izard, 2007), which is developed through time by way of a dynamic interplay between emotion and cognition.It suggests a high level of control appraisal, meaning that individuals have a high sense of control over the activity they are engaged in, which in turn leads to the positive emotional experience of freedom.The following quotation from the data illustrates the feeling of freedom as a positive achievement emotion: "Outside class, I like to listen to people speaking in a foreign language, or even to join in a conversation.I feel free, I'm not worried, if I don't understand something, I just ask them to repeat what they said, or ask them to simplify it" (P166).
Alongside positive feelings, negative achievement emotions also emerged from the data.These included expressions of anger, disappointment, impatience, anxiety, and remorse.Participants recounted such feelings as the following: "I might be a bit more impatient and nervous when reading in Hungarian because I cannot proceed as quickly and smoothly" (P12).
The second main category comprised epistemic emotions primarily linked to the cognitive aspects of foreign language learning.Such positive emotions included feelings that showed that learning a language poses a cognitive challenge and instigates enjoyment; feelings related to self-efficacy, the belief of possessing the cognitive ability to complete language related tasks; feelings of excitement; happiness; freedom; and curiosity.Excerpts such as the following were grouped under these categories: "A foreign language text always presents challenges" (P4).The negative counterparts of these emotions included general negative feelings, being frustrated, confused, tired, bored, depressed, and not self-efficacious with regard to the cognitive demands of learning a language.
Participants described this type of feeling as follows: "I feel uncertain even when I write very simple texts" (P50).
The next large category of emotions is related to topic in the academic context.Here we identified feelings of interest, freedom, liking, enjoyment and feeling energetic in connection with particular issues that language learners discuss, write about, listen to, or read about: "If I find the text interesting, I enjoy it" (P29), "I also like reading novels in foreign languages for pleasure" (P31).Negative topic emotions included anxiety, feelings of dislike and disinterest related to the object of language use.Feeling unenthusiastic about a topic also emerged as an issue categorized as a negative topic emotion: "I am usually less enthusiastic in class, but it also depends on the task we are given" (P99).
The last group of academically relevant emotions in Pekrun's (2014) framework are social emotions.Here again, based on our data, we distinguished between positive and negative emotions related to the social aspects of language learning.Expressions of liking, enjoyment, freedom, the lack of anxiety, being brave in connection with socializing in the foreign language and having positive attitudes towards others were grouped under the umbrella term positive social emotions (e.g., "I communicate in English with my foreign friends, and it feels really good that we are able to understand each other even though neither of us is a native speaker" (P72)).In contrast, feeling scared, anxious, stressful, not brave, and expressing dislike along with possessing negative attitudes towards others with respect to the social aspect of language learning were identified as negative social emotions.Examples here included the following: "In class, I am much more critical towards others, I pay more attention to their grammar mistakes" (P151); "I feel under stress to understand everything the teacher says, and I try to avoid embarrassment because of the level of my knowledge" (P160).
Upon first looking at the data from our sample, Pekrun's (2014) framework of academic emotions seems to provide an adequate theoretical background of emotions experienced in a university context.The most prominent emotions in all categories seem to be those related to enjoyment and liking along with anxiety (cf.Dewaele & MacIntyre, 2014, 2016).This also means that enjoyment, anxiety, and liking appeared in more than one emotion category.This is in line with the multi-dimensionality described by Dewaele and Mac-Intyre (2014), specifically with regard to these emotions.The authors explain that these emotions can be linked to both the private or the social context of an individual.In the present study, this is underpinned by the fact that participants mentioned these three emotions in all categories, including social academic emotions (using a foreign language in a social context or to communicate with a specific person); the more private topic emotions (liking a specific topic or not) and achievement emotions (notions related to their own successes and failures); and epistemic emotions (linked to the emotions evoked by the cognitive aspects of language learning).
Although Pekrun's (2014) framework proved to be fruitful for our analysis, accounts of certain emotions emerging from our data did not fit into its boundaries such as not feeling anything or feeling special.The lack of feeling was listed eight times (e.g., "When I read in English, it feels totally neutral, it does not feel special in any way" (P34)), which should be interpreted with caution, as it might simply be an artefact caused by the phrasing of the questions which explicitly enquired about students' feeling and emotions.Alternatively, it may be attributed to the disembodied nature of the second language, a common theme in emotion and multilingualism (cf.Pavlenko, 2012).At the same time, in five instances, participants included in their accounts feeling special in connection with foreign language learning, which signals a difficulty in the precise labelling of their emotions although the value attached to the feeling clearly appears positive.An example from a language learner reads as follows: "I feel privileged that I have access to such a great universe" (P4).These feelings do not seem to appear in Pekrun's academic emotions framework, and their interpretation and possible inclusion requires further research.
After categorizing the type of emotions language learners recounted in connection with foreign language learning, we looked at overarching categories and tried to establish abstract links between them and contextual characteristics.With the results of these analyses we sought answers to research questions b, c, d, as listed above.In other words, we examined the overall distribution of the main categories of emotions, their distribution across the four skill areas (see Table 1), across years (see Table 2), and in different contexts (in and outside the classroom) (see Table 3).
For our participants, overall the most frequently mentioned feelings were positive topic emotions (N = 94), followed very closely by positive achievement emotions (N = 93).More precisely, they appeared in connection with reading (N = 42) and listening (N = 33) in a foreign language (see Table 1), respectively.Generally speaking, among the receptive skills, reading is the one that can be practiced with the least amount of time pressure and where students have the largest space to choose topics they are interested in.While listening to foreign speech is still less time-bound than speaking and misunderstandings are also less obvious, being able to understand what the other person is saying seems to bring about positive emotions linked to success.The frequency of positive topic and achievement emotions linked to the receptive skills can be explained by the fact that our sample consisted of university students majoring in English, who tend to read a considerable amount of fiction and non-fiction in the foreign language and attend lectures held in English.However, it must be noted that our analysis was not restricted to English, but any foreign language the English major students chose to share their feelings about.Thus, positive emotions linked to the receptive skills also point beyond the foreign language major and perhaps indicate that, generally, our participants have positive emotional experiences tied to the topics they have a chance to read about in a foreign language and linked to successfully understanding others' speech in a foreign language.This finding contradicts the one in Pishghadam et al. (2016), where listening was primarily associated with negative emotions.Differences in the findings can be explained by differences in the sample, in the educational context, and also in the research methodology used.Perhaps taking a processbased approach in future studies will also be helpful in clarifying the situation.
To see whether differences in the frequency of the four emotions can be found across the skills, a Kruskal-Wallis non-parametric test was performed.We found a statistically significant difference in terms of social (χ 2 = 41.003,p < .001)and topic emotions (χ 2 = 15.039,p < .002)related to the four skills; more specifically, participants mentioned social emotions most often in the context of speaking skills and topic emotions in the context of reading skills.Such differences echo the findings of studies on one particular emotion, language anxiety, which have shown that learners tend to experience different levels of anxiety tied to distinct foreign language skills (Pae, 2013;Xiao & Wong, 2014).Our results are straightforward in that they reflect the nature of these skills: speaking involves oral communication, which by definition is a social activity; hence the importance of social emotions.Particularly, Dewaele, Witney, Saito, and Dewaele (2017) found enjoyment to be linked to the proportion of class-time spent on speaking.On the other hand, as our learners tended to focus more on the theme or topic of texts, topic emotions played a crucial role.Furthermore, given that emotions might change as skill and achievement progress, we compared the distribution of emotions across the years in the BA English studies program (see Table 2).We decided to group together secondand the third-year students for these analyses because the distribution across the three years was not equal, and it made sense to separate freshmen from the students who were more advanced in their studies.
Using the Mann-Whitney test, which is non-parametric test, we investigated the differences between the two groups.Overall, first-year students reported significantly more emotions (U = 2771, z = -1.99,p = .046)than secondand third-year students in connection with foreign language learning.The differences were found to be significant in terms of negative emotions (U = 2756.50,z = -2.137,p = .033),particularly in the case of negative achievement emotions (U = 2789.50,z = -2.60,p = .009).Interestingly, upper-year students reported more feelings tied to failure or unsuccessful experiences, although overall they tended to report experiencing fewer emotions related to foreign language learning.These results can be explained by the notion that probably the main source of language learning experience stems from the university context for our sample.In the first year, students are eager to learn about the subject they have chosen to pursue a career in; they have entered a new level of education, a new institution where there is still plenty to explore; hence the frequent mention of positive achievement, epistemic, and topic emotions.However, in later years, students tend to focus on obtaining good grades for their courses and on getting their degrees; so, their emotions are mainly linked to achievement in connection with both success and failure.Another issue we investigated concerned emotions experienced inside and outside the language classroom (see Table 3).According to our data, the educational context seems to evoke relatively more negative emotions (especially in the form of anxiety), whereas outside the classroom, enjoyment, general liking of language-related activities, and freedom in using the language were most frequently mentioned.While there is abundant literature on foreign language anxiety in the language classroom (Gkonou et al., 2017), relatively few studies have been published on the experiences of other emotions, albeit this situation is rapidly changing (e.g., Galmiche, 2017;MacIntyre et al., 2016).Based on our data, we can assume that outside the classroom, students have more control over the foreign language-related activities they are engaged in, their social aspects, the topic, and the challenge that some tasks present.As argued earlier, such positive emotional experiences can stem from such control appraisals (Pekrun et al., 2007).This line of reasoning also parallels studies on self-regulation and autonomy where it has been shown that a greater sense of control leads to higher levels of language learning motivation (Ushioda, 2011).Likewise, positive emotions such as freedom also appear in the pyramid model of willingness to communicate (MacIntyre, Clément, Dörnyei, & Noels,1998), where they play a role in interpersonal motivation and the language learners' desire to communicate with a particular person.

Conclusion
In the present study, we set out to explore the type of emotions English majors experienced in connection with their use of a foreign language and the four language skills.We chose a primarily qualitative approach to explore the variety and frequency of learners' feelings.In our data analysis, we embraced a deductive approach and interpreted the data within Pekrun's (2014) framework of academic emotions.According to the findings, language learners experience both positive and negative feelings linked to achievement, epistemic, topic and social emotions.Overall, in line with previous studies that have focused on language learners' affective states, the two most frequent emotions were enjoyment and language anxiety.Emotions were linked to particular language skills, and we found that participants recounted significant differences in experiences of topic and social emotions.Reading skills were mainly associated with topic emotions, whereas speaking skills were primarily related to social emotions.Language learners' feelings varied not only according to the skills involved but also depending on their year of study, with freshmen mentioning emotional accounts more frequently, while upper-year learners providing evidence of a higher frequency of negative achievement emotions.Finally, our data suggest that the context of language use (i.e., inside or outside class) can also be characterized by particular emotional experiences.Language learners tended to list more positive feelings in connection with outside classroom language learning situations.Our findings also have pedagogical relevance in shedding light on emotions in academic contexts.It seems that emotions are not negligible when it comes to language learning.Students experience a wide array of emotional states and teachers need to be aware of these.To maximize language learning experiences, it would be crucial for teachers to reduce negative emotions and encourage positive emotions in the classroom.Also, teachers should recognize that the use of particular language skills will invoke different emotions even in the same learner.By promoting enjoyment, especially with topics learners like and feel positive about in the classroom, language learning experiences in the instructional setting could also be enhanced.
Our study also has limitations.Although it was English language majors we investigated, they could recount emotional experiences concerning any foreign language they had learnt.It may have been worth making a distinction between such accounts of English as a foreign language and other languages.Also, as the data comprised retrospective narratives, they may have differed from how the learners felt on the spot at the time these events they were recalling actually took place.Last but not least, it must be noted that the data were collected in Hungarian (L1 of the participants), while the language of data analysis and interpretation was English.Thus, translation issues may have allowed room for misinterpretation in spite of the fact that two researchers coded the data instead of just one.
In terms of further research, it may be worth investigating the roles of the type of languages being studied, the proficiency level and the language teaching methods that are used when investigating emotional experiences in the process of language learning.It may also be important to look at emotions as language learning events as they unfold to gain deeper insights into what language learners feel in such situations (see Boudreau, MacIntyre, & Dewaele, this special issue).

Table 1
The frequency of the different emotions across the four skills

Table 2
The frequency of the different emotions across the years of study

Table 3
The distribution of emotions in and outside the foreign language classroom