Current state of foreign language anxiety research
Dating back to traces of foreign language anxiety research in the early 1950s and 1960s with the likes of Castenada, McCandless and Palermo (1956), Alpert and Haber (1960), and Sarason et al. (1960), scholars have long since studied the effects of anxiety on an individual learning a foreign language. This research had conceptualized FLA as being similar to that of psychological anxiety with no discernable differences (see for example, Hilgard, Atkinson and Atkinson 1971). It was not until 1978, however, that FLA studies took a dramatic turn towards redefining this concept with Scovel’s (1978) pivotal research. Scovel examined affective variables of language learning, focusing specifically on anxiety, explaining that:
- the research into the relationship of anxiety to foreign language learning has provided mixed and confusing results, immediately suggesting that anxiety itself is neither a simple nor well-understood psychological construct and that it is perhaps premature to attempt to relate it to the global and comprehensive task of language acquisition. (Scovel, 1978, p. 132).
According to Scovel, FLA’s conceptual basis required a realignment to match the uniqueness of foreign language learning. He argued that “it behooves us to examine other ways in which anxiety can be viewed, not as a simple, unitary construct, but as a cluster of affective states, influenced by factors which are intrinsic and extrinsic to the foreign language learner” (p. 134).
Horwitz et al. revolutionized the study of FLA as a result of Scovel’s assertions in their seminal work “Foreign language classroom anxiety” (1986) by reconceptualising the term to be unique to FL learning. Due to prior mixed and often ambiguous results, Horwitz et al. rationalized that although the affective conditions experienced as a result of anxiety were similar to those felt in other disciplines, such as a general nervousness and fear, the situations which evoked these feelings were not necessarily identical.
Rather than attributing the effects of FL learning anxiety to a general psychological condition, they argued it was defined by three fundamental factors representative of FL learning anxiety, yet similar to a performance anxiety: communication apprehension; test anxiety; and, fear of negative evaluation (Horwitz et al., 1986, p. 127). Yet rather than just applying these factors, the researchers distinguish this specific form of anxiety as “a distinct complex of self-perceptions, beliefs, feelings, and behaviors related to classroom language learning arising from the uniqueness of the language learning process” (p. 128).
Supporting their argument, Horwitz et al. presented an instrument to study this newly conceptualized form of anxiety: the foreign language classroom anxiety scale (FLCAS). This scale, comprised of 33-Likert items, presents a language learner with statements reflecting potentially anxiety inducing situations both in the language classroom and its educational vi-cinity. After completing the questionnaire and tabulating the results, a positive score on the FLCAS indicates that the language learner experiences anxiety and can be labeled as either a low, moderate, or highly anxious language learner.
Since its inception, the FLCAS has been the primary tool of analysis in many of the most prominent works attempting to understand and analyze FL learning anxiety (see Aida 1994; Sparks and Ganschow 1995, 2007; Chen and Chang 2004; Frantzen and Magnan 2005; Yan and Horwitz 2008). These works depict FLA comparably: it is a feeling experienced by certain individuals who are naturally anxious in and as a result of the FL classroom. The FLCAS is therefore administered to determine who experiences FLA, the data from which aids in hypothesizing how this anxiety can be alleviated. The implications and results of these various studies differ greatly, but their common goal of determining how to cope with FLA remains the same.
Although the data collected as a result of these studies has been paramount in estab-lishing a strong scholarly foundation for FLA over the past two decades, it is my contention that this conceptualization of FL is inherently problematic due to its rather antiquated concep-tion of both the learner and views of language. I aim to present a convincing alternative to Horwitz et al.’s approach to FLA; I do not wish to undermine its structural basis or its promi-nence on the second language acquisition community, but rather, I will expand upon it and direct it towards a new understanding of what is vital when attempting to understand the lan-guage learner.
A new paradigm
In establishing a new method by which FLA can be understood and analyzed, the theoretical foundation of Horwitz et al.’s conceptualization merits reconsideration. Funda-mentally, distinguishing FLA from a generalized psychological form of anxiety is a worth-while endeavour and has aided in establishing the theoretical basis of this field. By focusing on the role of FLA in three categories - communication apprehension, test anxiety and fear of negative evaluation - researchers have been able to study the experiences felt by learners which directly correlate to these characteristics of FLA. Research has undeniably demonstrat-ed too that many language learners experience anxiety to some degree in these categories, all of which are contingent upon the unique environment of the language learning classroom where communication exists between the instructor, learners, and their fellow classmates. This cannot be understated and remains prominent in the conceptualization of FLA that I propose.
Absent from this conceptualization of anxiety however is a thorough investigation and interpretation of language learners themselves. Horwitz et al. theorize learners as being a facet of the language learning classroom, thereby having an identity that is distinguishable by its language acquiring nature from other aspects of their identity; as they argue, “the ‘true’ self as known to the language learner and the more limited self as can be presented at any given mo-ment in the foreign language would seem to distinguish FLA from other academic anxieties” (Horwitz et al., 1986, p. 128). Arguing that a “true” self exists that is unique to the language learning environment would imply that the learner’s identity is positioned as being solely structured within the classroom; learners adopt an identity as “learner” that persists in the language learning classroom exclusively – once they leave the classroom, they are no longer afforded the opportunity to be language learners and should no longer experience foreign language classroom anxiety.
It is upon these theoretical grounds that I refute this conceptualization of anxiety as underdeveloped and unable to take into consideration modern identity theory in second lan-guage acquisition research. Furthermore, an application of such theory may lead to a better understanding of how anxiety occurs and interacts with the nature of the learner. I will not, however, attempt to formulate ways by which learners can solve problems of anxiety, nor will I explain why certain situations are more anxiety inducing than others. These goals are arguably unattainable if the proposed reconceptualization of anxiety is applied, due to the dy-namic nature of the language learner.
Currently, FLA is conceptualized as being a state-specific anxiety; it occurs in specific situations such as the language learning classroom. In this sense, it differs from trait-specific anxiety in that anxious language learners are not necessarily anxious in other aspects of their lives; language anxiety is argued to be very much situated specifically within the FL classroom. Scovel (1978) formulated his call for a reconceptualization of FLA based upon trait-specific approaches to FLA, which Horwitz et al. (1986) took up as they conceptualized an anxiety that is unique to the FL language classroom and to this specific state. This is to say that learners who experience anxiety in the FL classroom are anxious language learners be-cause certain situations are inherently anxiety-provoking. It may take into consideration facets of the classroom such as the instructor, peers, or the myriad activities that occur within the classroom as contributing to this defined state of being a FL learner in the FL classroom. The goal of the researcher is then to determine what these situations or factors are and how they can be resolved.
Yet this classification of the language learner, as being potentially susceptible to an-xiety if the situation is relevant assumes that certain language learners will always feel anxious in specific situations or states. The identity of the language learner is therefore rationalized as rather static in this context. The trend in SLA however is suggesting a transition toward post-structuralist theories of identity, whereby the language learner is considered to have a dy-namic, fluid identity that is always shifting. Tabouret-Keller (1997) contends that “at any given time a person’s identity is a heterogeneous set made up of all the names or identities, given to and taken up by her” (Tabouret-Keller, 1997, p. 316). If the language learner is to be understood by these means, then he or she cannot be argued to have a “true” self that is only found within the confines of language learning, as Horwitz et al. would argue. Furthermore, if these identities are myriad and dynamic, an understanding of FLA as state-specific cannot take into effect the fluid nature of the learner’s identity – what is on one day anxiety-provoking may cease to be if the learner’s identity has altered accordingly.
In conjunction with this notion of a dynamic identity is the fact that if these various sources of influence are taken into consideration, then research needs to look external to the FL language classroom for possible sources of anxiety which may breach the arbitrary physi-cal boundaries of the classroom and permeate the language learner and his environment. Sources of anxiety should not be confined exclusively within the physical realm of the class-room; rather, if identity is dynamic and influenced by external sources to the classroom, then sources of anxiety should too be susceptible to these influences. Also associated with the identity of the learner to a large degree are the views of language which the learner holds. By this I refer to not only the views of foreign language currently being studied, but views of one’s native language and other foreign languages that may have been studied concurrently or in the past. Horwitz et al. reflect momentarily on this notion, arguing that certain beliefs about the function of language may contribute to feelings of anxiety in the classroom (Horwitz et al., 1986, p. 127). They cite examples such as the belief that the language should only be spoken when it can be spoken correctly, thereby inhibiting attempts to speak the language in the FL classroom. Furthermore, these beliefs “must produce anxiety since students are ex-pected to communicate in the second tongue before fluency is attained and even excellent language students make mistakes or forget words and need to guess more than occasionally” (p. 127). While arguably valid, two aspects of this assertion deserve further attention.
Firstly, to state that language beliefs must produce anxiety presumes that language learners react to state-specific anxiety in identical ways. While certain individuals may feel anxious to speak under such conditions, others may find this motivating and strive to perform even better in order to ensure that their language skills will someday rival those of native speakers, supporting the re-emergence of facilitating anxieties (Alpert & Haber, 1960). Faci-litating anxiety is said to “motivate the learner to ‘fight’ the new learning task; it gears the learner emotionally for approach behavior” (Scovel, 1978, p. 139). FLA, understood in this context, need not necessarily inhibit language learning, but can rather stimulate the learner to perform even more effectively. While perhaps not a common method by which to deal with anxiety, it can and does occur, yet is given no attention in Horwitz et al.’s conceptualization of FLA.
Secondly, the language beliefs which are referenced concern only the target language being studied. This would indicate an opinion that language views in the context of the FL classroom and in accordance with anxious feelings can only pertain to the present studied language. This view of language reaffirms Horwitz et al.’s understanding of the language learner as having a “true” self, an identity which is situated within the confines of the FL classroom. However, views of language do not necessarily exclusively exist concerning the target language. Past experience studying languages or the role of the learner’s native lan-guage in his or her life may greatly influence how the language learner perceives the foreign language.
With the distinctions I have made in terms of learner identity and language views, I would like to present a reconceptualised paradigm for understanding FLA. Rather than em-phasizing the common situations which FL learners may find anxiety-provoking, I emphasize the dynamic roles of language learners themselves and their views of language as being of ut-most importance in theorizing and understanding FLA. Furthermore, I contend that not only are these facets important, but that they are intrinsically related to FLA and neither can be understood in this context without the other (Figure 1).
Figure 1: Foreign language anxiety paradigm
This paradigm details three primary facets which conceptualize a revised understand-ing of how FLA functions. Learner identity, language views, and anxious feelings all play equally important and reciprocal roles in this reconceptualization. The lines connecting each aspect signify a bilateral connection; therefore, views of language may influence the identity construction of the language learner and his feelings of FLA. Yet reciprocally, feelings of FLA may themselves impact the language beliefs of learners and how they construct their identity. Consequently, I place significant emphasis on the role of all three aspects as consti-tuting and impacting one another. I will outline the theoretical framework and implications of each facet in order to construct a better understanding of how this paradigm can function and fundamentally transform the understanding of FLA in this discourse.
The first facet of the proposed paradigm implements current identity theory into FLA research. Horwitz et al. (1986) define language learner identity through two factors. Firstly, the language learner is said to have both a “true” self, that which is known to the language learner, and a limited self that is a more typical portrayal of the language learner to the sur-rounding society (Horwitz et al., 1986, p. 128). Secondly, the identity expressed in the FL language classroom is definitive; “students with debilitating anxiety in the foreign language classroom setting can be identified and… they share a number of characteristics in common” (p. 129). This statement assumes that language learners are alike in the FL classroom, as it is the setting, rather than the individual learners, which define them. This further validates the state-specific understanding of anxiety that is so prominent, yet it remains problematic in light of advances in SLA discourse.
Horwitz et al. argue that “no other field of study implicates self-concept and self-expression to the degree that language study does” (p. 128), and although this is decidedly true, their approach to this assertion, when analyzed in the current context of SLA discourse and identity research, is rather out-dated. At the time of publication, identity research in SLA had not yet explored the possibility of a dynamic and fluid identity, hypothesized in poststructuralist definitions of identity. It was not until 1995 that SLA began to theorize the incorporation of socially dependent variables into its construction and understanding of iden-tity. This progress was the result of Norton’s (1995) influential work, in which she wrote that “SLA theorists have not developed a comprehensive theory of social identity that integrates the language learner and the language learning context” (Norton, 1995, p. 12). She argues that language learners do not live in “idealized, homogenous communities” (p. 12) and for this reason, cannot be studied under overarching terms such as ‘language learners’, sharing common characteristics and a shared identity. Interestingly, she too presents one of the only accounts of establishing a link between poststructuralist theories of identity and FLA, writing that “anxiety is not an inherent trait of a language learner, but one that is socially constructed within and by the lived experiences of language learners” (Norton, 2000, p. 123). This argu-ment is not expanded upon in great detail, but the initial theorization did indeed exist.
Contrasted with Horwitz et al.’s suggestion that language learners share common cha-racteristics and are identified in terms of their FLA as a result of the FLCAS and its limited approach to anxiety, we find that their assertion is entirely problematic. Language learners need to be given more accountability as unique individuals who can mould their experiences through their dynamic identities. Consequently, it proves difficult to argue that language learners in the context of the FL classroom will behave similarly and have anxious feelings in similar anxiety-inducing situations. The FLCAS is also accountable, due in part to its reliance on situational anxiety that quantifiably identifies the learner and is able to distinguish the ex-tent to which anxiety influences the learner based purely upon the results of a questionnaire; it is problematic to utilize the FLCAS in this way and merits reconceptualization.
To incorporate poststructuralist theory of identity, I incorporate and apply Block’s (2007) theorization to the proposed paradigm. Block examines a variety of theories that he sees as vital to an understanding of poststructuralist identity - taking into account concepts such as communities of practice, performativity, and power – and describes these identities as “socially constructed, self-conscious, ongoing narratives that individuals perform, interpret and project in dress, bodily movements, actions and language” (Block, 2007, p. 27). He lists seven identity types, all of which co-construct an individual’s identity: ethnic, racial, national, migrant, gender, social class, and language (p. 43). Of specific interest for FL learning is his classification of language identity, that which is the “assumed and/or attributed relationship between one’s sense of self and a means of communication which might be known as a lan-guage (e.g. English) a dialect (Geordie) or a sociolect (e.g. football-speak)” (p. 40). Block states however that “it is difficult to discuss one type of identity without mentioning others” (p. 42); although researchers may like to focus on language identity and situate the learner’s identity solely within the FL classroom, other aspects of identity, such as social class or ethnic identity may prove influential as well and fundamentally alter the language identity of a learner.
In an effort to expand the contextual area examined when analyzing the relationship between the cause and effect of FLA, the second category of this proposed reconceptualiza-tion pertains to the views of languages and language learning that each individual consciously or subconsciously holds. This has been explored briefly by Horwitz et al. as previously men-tioned, but their position pertains solely to concerns of how the language currently being stu-died should be used (for example, whether native language skills are required in order to speak confidently). While this can certainly be argued to be an aspect of what constructs a learner’s view of language, the field is more expansive than just this. Not only should the views of the current studied language be taken into consideration, but so too should views of other languages that may have significant meaning to the learner.
The native language (L1), perhaps not typically considered a facet of FLA, may have implications that transcend the context of the L1 and impede upon the FL classroom. Sparks and Ganschow have focused ample research on the possibility of the L1 affecting FLA, con-structing one of the few critical oppositions to Horwitz et al.’s generally adopted definition of FLA (see Sparks & Ganschow 1991; 2007). They argue that the skills used to learn a foreign language are similar to those used to learn one’s native language; they therefore questioned “Which came first?”; do students develop affective learning disabilities, such as anxiety, or do the native language problems innately cause difficulties?” (Sparks & Ganschow, 1991, p. 5). Although not considered under Horwitz et al.’s conceptualization of anxiety, the research by Sparks and Ganschow could indeed be enlightening in helping understand the role of the L1 in FLA construction as being a factor.
Consider a student who enjoys writing and finds inspiration in works of translated German literature. She enjoys using her L1 to write and finds it a useful tool for communica-tion. Although her knowledge of the German language may be rudimentary, her passion for the language and its applicability allows her to perceive and view the language differently and perhaps more positively than a learner who is taking a language course for a variety of other reasons. The effect of motivation and learner attitude must be considered to some degree, and although the field of discourse on this is admittedly wide, I perceive great benefit by examin-ing Gardner’s (1985) socio-educational model, proposing that:
the beliefs in the community concerning the importance and meaningfulness of learn-ing the language, the nature of skill development expected, and the particular role of various individual differences in the language learning process will influence second language acquisition (Gardner, 1985, p. 146).
Emphasis deserves to be placed on the connection between motivation and language beliefs, as “attitudes involving other ethnic groups and the language learning situation underlie moti-vation” (p. 149). Although this model considers the effect of situational anxiety as well, I see this being less problematic as it is associated specifically with the cultural context and is not the basis for an entire FLA model.
Within the confines of the classroom, this learner’s language views could potentially impact her anxiety in a multitude of ways. The typical debilitating anxiety caused by aspects such as communication apprehension may be negligible due to her passion to learn German and her prior, albeit limited, experience employing the language. Yet so too could her anxiety be solely debilitating - her desire to learn the language and her prior knowledge of it supports her view that it is a difficult language to learn. She may feel her achievement is marginal in her limited exposure to the language, thus causing any combination of anxious feelings.
Facilitating anxiety may also emerge in such a situation. Consider the motivations of this language learner: the desire to understand her favourite book in German may be the spe-cific motivational support required when facing adversity in the classroom, before it manifests into the anxious feelings described previously. Rather than succumbing to these feelings, she could endure and manipulate them as motivation to perform even better. Thus, this language learner is unique - her language views differ from others’ and she may approach her anxious feelings in a multitude of ways. Consequently, it remains difficult to structure FLA solely on the current conceptualization of anxiety that exists: state-specific and able to be diagnosed through thirty-three Likert-items depicting various circumstances that may induce anxiety. Yet as each language learner experiences not only anxiety in different ways, but also holds unique language views, then the FLCAS and its accompanying conceptualization of FLA cannot accurately account for all possible factors.
The concluding aspect of this paradigm examines the actual feelings experienced by the learner as an integral aspect of this proposed conceptualization, and as stated previously, I generally accept how Horwitz et al. (1986) have conceptualized FLA as distinct from general, psychological manifestations and deserving of its own theoretical foundation. The three com-ponents outlined by Horwitz et al. are primarily found within the FL classroom and are in-deed distinguishable from other anxieties; rather than understanding FLA as simply being the conglomeration of these three facets, however, I propose that these facets are but one aspect of what constitutes FLA and instead classify them as anxious feelings instead of FLA in its entirety.
As much of the classroom dynamic is situated around speaking the foreign language and practicing it in a variety of forms, communication apprehension is quite poignant - whether in speaking, listening, writing or reading, all methods of communication are prevalent in the FL classroom and can lead to anxiety. Although the majority of research focuses on anxiety in oral communication, written forms of anxiety can be just as anxiety inducing in many instances (see Cheng 2002). The fear of negative evaluation too can be unique to the FL classroom considering the typical structure of these classes. Dependent largely on com-municating and learning the fundamentals of the language, evaluation opportunities are plen-tiful and require the learner to be consistently alert and prepared to perform effectively. Al-though evaluation occurs in every class, it is the sheer prevalence in the FL classroom that distinguishes it from other anxieties and further supports it being recognized as a FL specific anxiety.
The final aspect of Horwitz et al.’s theorization of FLA is that of test anxiety. It too is distinctively prominent in the language learning classroom, yet I would contend that it is too similar to other forms of test anxiety and cannot necessarily be considered unique to the theo-rization of FLA, as MacIntyre and Gardner have argued (MacIntyre & Gardner, 1989, p. 268). I do not propose however that test anxiety be removed from the currently accepted con-ceptualization of FLA, but simply that it be understood as not unique to this context. Accord-ing to the previously accepted state-specific conceptualization of FLA this may prove prob-lematic, but if anxiety is understood as neither state- nor trait-specific, understanding test an-xiety to be of a more general conceptualization is not detrimental to my proposed reconceptu-alization of FLA.
Implications of the paradigm
This proposed paradigm entails several implications if adopted. Of primary importance is the reliance upon the FLCAS that has dominated FLA discourse. The goal of the FLCAS is to detect anxiety in language learners by posing situations that may induce anxiety, and asking the learners to reflect on them. The implications of such a questionnaire positions learners as being equivalent and having similar views on language, while considering only events within the context of the classroom as being potentially influential in determining an-xiety – the FLCAS should thus not be used as the primary tool of analysis if the proposed paradigm is considered. By positing that language learners can be studied with a generalized, quantitative questionnaire, these learners are immediately relegated as non-dynamic learners and their poststructuralist identities are not considered; it is therefore difficult to adapt such a conceptualization in light of current SLA discourse. Furthermore, the influence of language views on the learner extends far beyond the limitations of the classroom, yet these are not taken into consideration either by exclusively utilizing the FLCAS.
For these reasons, I propose that the FLCAS not be the sole research method in stu-dies concerning FLA (see studies by Frantzen & Magnan, 2005; Yan & Horwitz, 2008; De-waele et al., 2008); rather, I support its implementation as a preliminary tool to educate the learner instead of diagnosing him. Its strength is in concisely presenting a language learner with a selection of situations that have been found to be anxiety-provoking by various lan-guage learners. This is not to say that every language learner will react to these situations identically, or that by reacting to certain situations, the learner is inherently anxious, but that these situations have the possibility of affecting a learner and should therefore be recognized. By examining the FLCAS on a metalinguistic level, the learner can gain a relatively quick glimpse into many situations that he or she has perhaps experienced in the FL classroom, gaining an understanding that while some of these situations may apply, many others may not; what is imperative is the basic understanding that anxiety can exist in such a large variety of situations.
If the FLCAS continues to be implemented as the sole tool of analysis, the results will be unable to accurately define the new definition of the learner in SLA discourse. It cannot account for the dynamic and fluid identity of the language learner as it only considers the language learner within a state-specific model of anxiety, formulated exclusively as a result of what occurs in the classroom or as a direct result of the classroom. Yet there exists much more that can affect feelings of FLA for a language learner that deserves further study.
To do so, a focus on qualitative study should direct future research in FLA. This trend has already gathered traction as can be seen in studies such as Yan and Horwitz’s (2008) and Dewaele, Petrides and Furnham’s (2008), both of which utilize the FLCAS in conjunction with qualitative methods. I consider this a necessary development in this field of research, as the response of the learner is the only way to understand how exactly FLA is constructed. The issue that I observe in these studies, however, is that they persist in their attempts to de-termine which situations specifically cause anxiety and strive to modify the situations in an effort to alleviate anxiety. This approach assumes again that language learners are similar and experience FLA in similar situations, which cannot be the case if learners adopt poststructu-ralist identities. I therefore place emphasis on using this proposed paradigm to discover how learners experience and perceive their anxious feelings. It is perhaps too difficult to present solutions as to how to best remedy FLA – a remedy may in some instances not even be de-sired, as evidenced by the facilitating aspects of FLA in previous studies – but this paradigm can help both researchers and language learners better recognize why it exists.
While acknowledging the lasting impact of the seminal research by Horwitz et al. (1986), with this paper I wish to reopen the discussion on FLA and incorporate advancements in SLA discourse that I feel are necessary to our continued understanding of this complex feeling. I would also suggest that the terms facilitating and debilitating anxiety be reintroduced into the discourse’s vocabulary. The majority of studies have focused entirely on a debilitating anxiety in the FL classroom, yet if the new paradigm and the uniqueness of each individual learner are considered, then we need also to consider the positive and facilitating effects that FLA may have. FLA remains a complex term to conceptualize, but if it is re-evaluated, perhaps research can begin to better understand its effect and reinvigorate this exciting discussion in SLA discourse.
Aida, Y. (1994). Examination of Horwitz, Horwitz, and Cope's construct of foreign language anxiety: The case of students of Japanese. Modern Language Journal, 78(2), 155-168.
Alpert, R., & Haber, R. N. (1960). Anxiety in academic achievement situations. The Journal of Abnormal and Social Psychology, 61(2), 207-215.
Block, D. (2007). Second language identities. London: Continuum.
Castenada, A., McCandless, B. R., & Palermo, D. S. (1956). The children's form of the manifest anxiety scale. Child Development, 27, 317-326.
Chen, T., & Chang, G. (2004). The relationship between foreign language anxiety and learning difficulties. Foreign Language Annals, 37(2), 279-289.
Cheng, Y. (2002). Factors associated with foreign language writing anxiety. Foreign Language Annals, 35(6), 647-656.
Dewaele, J., Petrides, K.V. & Furnham, A. (2008). Effects of trait emotional intelligence and sociobiographical variables on communicative anxiety and foreign language anxiety among adult multilinguals: A review and empirical investigation. Language Learning, 58(4), 911-960.
Frantzen, D., & Magnan, S. S. (2005). Anxiety and the true beginner-false beginner dynamic in beginning French and Spanish classes. Foreign Language Annals, 38(2), 171-190.
Gardner, R.C. (1985). Social psychology and second language learning: The role of attitudes and motivation. London, UK: Edward Arnold.
Hilgard, E. R., Atkinson, R. C., & Atkinson, R. L. (1971). Introduction to psychology (5th ed.). New York: Harcourt Brace Jovanovich.
Horwitz, E. K., Horwitz, M. B., & Cope, J. A. (1986). Foreign language classroom anxiety. Modern Language Journal, 70(2), 125-132.
MacIntyre, P. D., & Gardner, R. C. (1989). Anxiety and second language learning: Toward a theoretical clarification. Language Learning, 39(2), 251-275.
Norton, B. (2000). Identity and language learning: Gender, ethnicity and educational change. Harlow, England: Longman, 2000.
Norton, B. (1995). Social identity, investment, and language learning. TESOL Quarterly, 29(1), 9-31.
Sarason, S. B., Davidson, K. S., Lighthall, F. F., Waite, R. R., & Ruebush, B. K. (1960). Anxiety in elementary school children. Oxford, England: John Wiley.
Scovel, T. (1978). The effect of affect on foreign language learning: A review of the anxiety research. Language Learning: A Journal of Applied Linguistics, 28, 129-142.
Sparks, R. L., & Ganschow, L. (1991). Foreign language learning differences: Affective or native language aptitude differences? Modern Language Journal, 75(1), 3-16.
Sparks, R.L. & Ganschow, L. (1995). Effects of direct instruction in Spanish phonology on the native-language skills and foreign-language aptitude of at-risk foreign-language learners. Journal of Learning Disabilities, 28(2), 107-120.
Sparks, R. L., & Ganschow, L. (2007). Is the foreign language classroom anxiety scale mea-suring anxiety or language skills? Foreign Language Annals, 40(2), 260-287.
Tabouret-Keller, A. (1997). Language and identity. In Florian Coulmas’ (Ed.), The handbook of sociolinguistics (pp. 315-326). Oxford, UK: Blackledge.
Yan, J. X., & Horwitz, E. K. (2008). Learners' perceptions of how anxiety interacts with per-sonal and instructional factors to influence their achievement in English: A qualitative analysis of EFL learners in China. Language Learning, 58(1), 151-183.