Literature Review: Affect In Language Learning (continued)
Previous pages: 1, 2, 3

3.3.4. Anxiety
3.3.4.1. Introduction
Research has confirmed the existence of 'language anxiety' and its effect on second language learning (MacIntyre & Gardner 1991b), pointing to a reciprocity between anxiety and proficiency (MacIntyre et al. 1997:279), such that "even in optimum conditions, students can experience destructive forms of anxiety" (Reid 1999:297). However, this    effect is complex and difficult to measure (Phillips 1992:2), though research (and the experience of teachers) suggests that language learning contexts are especially prone to anxiety arousal (Horwitz et al. 1986; MacIntyre & Gardner 1989; 1991a; Price 1991; MacIntyre 1995:90), with Campbell & Ortiz (1991:159) estimating that up to half of all language students experience debilitating levels of language anxiety, and Horwitz (Horwitz et al. 1986) finding that language anxiety can cause students to postpone language study indefinitely or to change majors. Because of this, language anxiety has been the subject of a good deal of research, on the assumption that an understanding of its causes and investigation into how to reduce language anxiety will improve learner performance and increase learning satisfaction by easing tensions and reducing demands on cognitive processing space (Eysenck 1979). Scovel (1978) provides an early review of anxiety research, which is supplemented by the excellent reviews of MacIntyre & Gardner (1991b), and Gardner & MacIntyre (1993a). More recently, Oxford (1999) has investigated whether language anxiety it is a short-term or lasting trait, whether it is harmful or helpful, which factors correlate with language anxiety, and how anxiety can be identified in the language classroom. In view of these sources of information, the present discussion will confine itself to a brief sketch of the current situation and associated issues.

3.3.4.2. Definitions
Consideration of psychological aspects of learning is important in the study of anxiety, as can be seen in Scovel's reference to an emotional state of "apprehension, a vague fear that is only indirectly associated with an object" (1978:34), and in Horwitz et al's (1986:125) "subjective feeling of tension, apprehension, nervousness, and worry associated with an arousal of the autonomic nervous system." Such psychological definitions most commonly refer to a "transitory emotional state or condition characterised by feelings of tension and apprehension and heightened autonomic nervous system activity" (Spielberger 1972:24), a state which can have both negative and positive effects, and which motivates and facilitates as well as disrupting and inhibiting cognitive actions such as learning. Izard (1972) defines three types of anxiety (trait anxiety, state anxiety, and situation-specific anxiety) identifying reactions of fear, distress, anger and shame (including shyness and guilt) on the negative side, and interest and excitement on the positive side.

From the second-language-learning perspective, and within the context of Horwitz et al's (1986) warning that second language research has neither adequately defined foreign language anxiety nor described its specific effects on foreign language-learning, Gardner & MacIntyre (1993a) see language anxiety as "the apprehension experienced when a situation requires the use of a second language with which the individual is not fully proficient", this apprehension being characterised by "derogatory self-related cognitions ..., feelings of apprehension, and physiological responses such as increased heart rate"(1993a:5). MacIntyre & Gardner (1994:284) also describe the feeling of tension and apprehension specifically associated with second language contexts, in particular second language performance (cf. Horwitz et al. 1986, who identify communication apprehension, social evaluation, and test anxiety, and MacIntyre & Gardner 1989; 1991b). Williams & Burden (1997:92) point out that anxiety is "highly situation specific and itself affected by a number of other factors" (cf. Horwitz & Young 1991).

3.3.4.3. History of research
Research into language anxiety has been characterised by sometimes conflicting evidence from instruments applied in different languages, measuring different types of anxiety, language skills, level of learning, and teaching methodology (Phillips 1992:15), so it is perhaps not surprising that early empirical studies (focusing on trait anxiety and/or state anxiety) showed "mixed and confusing results" (Scovel 1978) and were "unable to establish a clear picture of how anxiety affects language learning and performance" (Horwitz & Young 1991:xiii), a situation still unresolved in 1991 (Horwitz & Young 1991:177). Other early research found a negative correlation between language anxiety and scores on a standardised achievement test (Tucker et al. 1976), results replicated by Horwitz (1991) and others (Steinberg 1982; Tobias 1980), though these effects were moderate (Sarason 1983).

Cognitive and affective components of anxiety were identified by Liebert and Morris (1967) as "worry" and "emotionality", the former being defined by Sarason (1986) as "distressing preoccupations and concerns about impending events" (1986:21), often taking the form of distraction, self-related cognition such as excessive self-evaluation, worry over potential failure, and concern over the opinions of others (Eysenck 1979). Such outcomes often impair task performance, which has itself been the subject of much research into language anxiety (often to the exclusion of the cognitive activity preceding that performance [Eysenck 1979]), results of which suggest that anxiety causes cognitive interference in performing specific tasks (cf. Schwarzer 1986). Horwitz et al. (1986:127) draw attention to three related performance anxieties: i) communication apprehension; ii) test anxiety; and iii) fear of negative evaluation, and Eysenck (1979) offers a reconceptualization of anxiety in terms of this interference, suggesting that the anxious person has his/her attention divided between task-related cognition and self-related cognition:

worry and other task-irrelevant cognitive activities associated with anxiety always impair the quality of performance. The major reason for this is that the task-irrelevant information ... competes with task-relevant information for space in the processing system. (Eysenck 1979:364)

Eysenck (1979) also proposes that anxious students are aware of this interference and attempt to compensate by increased effort. This can facilitate performance if the increased effort outbalances the reduced efficiency of the cognitive processing (Bailey 1983), but it has also been reported that extra study by anxious students does not reflect the effort (Horwitz et al 1986; Price 1991).

Recent qualitative reports suggest that anxiety matters to students of all abilities (Bailey 1983; Horwitz et al. 1986; Price 1991), especially when there is heavy ego-involvement (Bailey 1983; Horwitz et al. 1986; Price 1991; Young 1990), as in oral examinations (Tobias 1980). Horwitz & Sadow (Horwitz & Sadow, submitted) indicate that high language anxiety is related to students' "negative concepts of themselves as language learners, and negative expectations for language learning", begging the question of whether language anxiety is a cause of reduced achievement, a result, or both. Leary (1990) and Levitt (1980) see the relationship between cognition, anxiety and behaviour as interactive, recursive and cyclical, and Gancshow et al. (1994:52) suggest that higher foreign language ability is related to lower anxiety.

Sparks & Ganschow's linguistic coding deficit hypothesis (LCDH) (Sparks & Ganschow 1991; 1993a & b) proposed that language aptitude is the main source of individual differences in language achievement, a view that appears to relegate language anxiety "to the status of an unfortunate side effect." (MacIntyre 1995:90). However, MacIntyre (1995) replies that the effects of anxiety language are more complex than implied by Sparks & Ganschow, and that being a cognitive activity, language learning "relies on encoding, storage, and retrieval processes, and anxiety can interfere with each of these by creating a divided attention scenario for anxious students” (MacIntyre 1995:96).

3.3.4.4. Discussion
MacIntyre & Gardner (1991a) state that anxiety plays little part in the learner's first experience with the foreign language, when language aptitude and motivation are the dominant factors in learning, and that language anxiety develops if subsequent experiences are not positive. Given the tendency of language classrooms to promote anxiety (Horwitz et al. 1986), promoting a low-stress language learning such an environment must be an important priority for the teacher. This involves encouraging realistic expectations about accuracy and errors (Foss & Reitzel 1988), offering training in affective strategies, to help students manage anxiety and improve performance (Oxford & Crookall 1989), reassuring students that they are not alone in their affective reactions and that these feelings are normal (Foss & Reitzel 1988; Campbell & Ortiz 1991), and showing that the teacher/evaluator understands the tension caused by being anxious about appearing anxious Phillips (1992:20).

Phillips (1992:21) also points out that evaluations involving partner and small-group work, interviews, problem-solving, and role-plays are usually enjoyed by students (Phillips 1990; Young 1990) and can reduce anxiety-raising competitiveness (Bailey 1983) and apprehension (Foss & Reitzel 1988). Familiar tasks also create less anxiety (Bailey 1983), so practice of test-tasks will encourage confidence:

By providing students ample oral practice and the opportunity for group evaluation, the language teacher may meet the challenge of encouraging the development of communicative competence without creating an anxiety-ridden atmosphere that can result in negative attitudes toward language learning and testing. (Bailey 1983)

Oller (1979b) mentions that measurement of affective variables is "necessarily inferential and indirect" (1979b:10), a situation that has led to the use of self-reports in research (cf. Horwitz et al's [1986] self-report instrument, the Foreign Language Classroom Anxiety Scale (FLCAS) the 33 items of which include questions about test anxiety, communication apprehension, and fear of negative evaluation in the classroom). Results of such reports can be problematical if the subjects "tend to give answers that are associated with the respondents' perceptions of the predispositions of the researcher" (Oller 1979b:17), but reflective self-reports can also inform those beliefs, affective reactions and attitudes about language anxiety which may be more important to students than "any external reality" (Bailey 1983:86), and which determine whether they continue studying a foreign language.

Chastain (1975) found conflicting results when investigating the correlation between language test scores and anxiety, which led him to conclude that mild anxiety could be beneficial (cf. Kleinmann 1977). This issue of facilitating versus debilitating anxiety (Alpert & Haber 1960; Bailey 1983) may be central to research on anxiety in SLA (Scovel 1978:132) and was examined by Bailey (1983:93) through analysis of diary studies, in which the relationship between competitiveness and anxiety appeared to result in either an unsuccessful or successful self-image. In this model (figure 12, below), the successful self-image and the unsuccessful self-image subjected to facilitating anxiety enter a cycle of enhanced learning and positive (mostly internal) rewards. The unsuccessful self-image subjected to debilitative anxiety however, enters a different cycle of perceived failure, which can be broken by future facilitating anxiety (if the learner continues in the language course):  

FIGURE 12: COMPETITIVENESS AND THE SECOND LANGUAGE LEARNER (BAILEY 1983:97)

The competitive self-image is identified by Bailey as originating from a number of factors:

1.  Overt self-comparison of language learning.
2. Emotive responses to the comparisons in 1.
3. A desire to out-perform other language learners.
4. Emphasis on or concern with tests and grades.
5. A desire to gain the teacher's approval.
6. Anxiety experienced during the language lesson.
7. Withdrawal from the language-learning experience (Bailey 1983:93).

3.3.4.5. Conclusion
Widely used measures of language achievement such as course grades and standardised proficiency tests, consistently show significant negative correlations between anxiety and performance in research (Clément et al. 1980; Gardner & MacIntyre 1993b; Gardner et al. 1984;  Horwitz 1986; Phillips 1992; Trylong 1987). Horwitz et al. (1986) acknowledge this, but suggest that researchers should study the more subtle effects of language anxiety on specific processes involved in language acquisition and communication. Research suggests that development of oral skills can increase levels of foreign language anxiety experienced by some learners (Horwitz et al. 1986; Scott 1986; Phillips 1989; 1990; Price 1991), although students show most interest in developing their ability to communicate verbally in the target language (Lindenau 1987). This performance anxiety can be intensified by communicative oral testing since anxiety increases in evaluative situations (Tobias 1980), a situation that Phillips (1992:14) argues should cause concern to language educators because of its potential impact on oral tests, on students' affective reactions, and hence on their attitude to language learning (Phillips 1992:22).

MacIntyre & Gardner see the "inconsistencies" of previous research as "attributable to an inappropriate level of instrument specificity" (1989:272), and claim that "a clear relationship exists between foreign-language anxiety and foreign-language proficiency" (1989:272). They propose that language anxiety develops as the result of repeated, negative experiences with the second language, and that anxious students are capable of showing high levels of achievement, given sufficient time to study (MacIntyre & Gardner 1994:298; cf. Tobias 1986). Learner-perceptions are important in the matter of reducing language anxiety, and studies of students' negative correlations between anxiety and output quality indicate that anxious students have more difficulty expressing themselves and tend to underestimate their level of ability ("self-derogation") compared with more relaxed students ("self-enhancement") (MacIntyre & Gardner 1994). It is therefore important for the teacher to identify anxious students (Horwitz et al. 1986:128) and any elements of the classroom environment which produce this reaction.

Questions remain regarding the influence of language anxiety on oral exam performance, how it interacts with other personality variables such as learning styles, motivation, and personality types, and what techniques are effective in controlling it.

Continue reading this literature review ("Learning Styles")