Language Teachinghttp://journals.cambridge.org/LTA

Additional services for Language Teaching:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

A student's contributions to second-language learning. Part II:Affective variables

R. C. Gardner and P. D. MacIntyre

Language Teaching / Volume 26 / Issue 01 / January 1993, pp 1 - 11DOI: 10.1017/S0261444800000045, Published online: 23 December 2008

Link to this article: http://journals.cambridge.org/abstract_S0261444800000045

How to cite this article:R. C. Gardner and P. D. MacIntyre (1993). A student's contributions to second-language learning. Part II: Affectivevariables. Language Teaching, 26, pp 1-11 doi:10.1017/S0261444800000045

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/LTA, IP address: 137.205.50.42 on 12 Dec 2014

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art articleA student's contributions to second-languagelearning. Part II: Affective variablesR. C. Gardner and P. D. Maclntyre The University of Western Ontario

This is the second part of a two-part article dealingwith individual difference correlates of second-language learning. In this part, attention is directedto a consideration of the role of affective variables insecond-language learning. By affective variables, wemean those emotionally relevant characteristics ofthe individual that influence how she/he willrespond to any situation. In the language learningsituation, many such factors have been identified,but this article will focus on two broad classificationsof affective factors, namely, Language attitudes andmotivation, and Language anxiety and self-con-fidence.

Attitudes and motivationConsiderable research has been conducted on therole of attitudes and motivation in second- orforeign-language learning. This research began withspeculations made by Lambert (1955) that an interestin learning another language often developedbecause of emotional involvement with the otherlanguage community or because of a direct interestin the language itself. In his studies of bilingualdevelopment, Lambert (1955) described twoAmerican university students who had developedparticularly high levels of French/English bilingualskill. One had developed an intense identificationwith France, while the other had devoted much ofher career to teaching French. Lambert speculatedthat their excessively high competence in French(their second language) was quite likely motivatedby their somewhat different involvements with thelanguage. We can see here the seeds of the distinction

Robert C. Gardner is professor of psychology at theUniversity of Western Ontario, London, Canada. Hehas written books and articles on the role of attitudesand motivation on second language learning, co-editeda book on ethnic relations in Canada, and publishedarticles on ethnic stereotypes, psycholinguistics, andstatistics.Peter D. Maclntyre has recently completed his doctoralstudies at the University of Western Ontario. Hegraduated from the University College of Cape Bretonwith an undergraduate degree in psychology andspeech communication. He has published articles in theareas of anxiety and language learning, programevaluation and statistical analysis.

between integrative and instrumental orientationsthat was an integral part of some of the earlierresearch in this area.

The first investigation of the relationship ofattitudes and motivation to achievement in a secondlanguage was published by Gardner and Lambert(1959), even though such relations had beenhypothesised earlier. Arsenian (1945), for example,proposed that attitudes could play a role in second-language acquisition, and in an inaugural article inLanguage Learning, Marckwardt (1948) argued thatthere were five basic motives for learning a secondlanguage. These were described as ' provision of acultural background, the influence of foreign speechislands, the necessity for political and culturalunification, purposes of colonisation and commerce,and the necessity of reading scientific and technicalworks' (pp. 10-11). Note that the integrative andinstrumental orientations are well represented inthese basic motives.

In their initial study, Gardner and Lambert (1959)demonstrated that two independent factors, Lan-guage aptitude and Social motivation, were bothrelated to achievement in French among CanadianEnglish-speaking high-school students. Subsequentresearch also showed relationships between bothlanguage aptitude on the one hand and attitudes andmotivation on the other with achievement in thesecond language, though often more than twofactors were obtained. This would be expected asthe number and complexity of the variables growand as the socio-cultural make-up of the com-munities and level of training and competence in thesecond language varies. The bulk of these studieswere conducted in Montreal, Canada as well as incities in Maine, Louisiana, and Connecticut, and thePhilippines (see Gardner & Lambert, 1972, for areview of much of the earlier research).

Gardner and Smythe and their colleagues wereconcerned with formalising the composition andmeasurement of attitudes and motivation as theyrelate to second-language acquisition. They engagedin a programme of research to develop measureswith high reliability and applicability to differentage levels, levels of second-language training, andsocial contexts. Gardner and Smythe (1981) presentsummary data on the initial development of theAttitude/Motivation Test Battery while Gardner,Smythe, Clement and Gliksman (1976) presentsummary data of the relationships of the various

Lang. Teach. 26, 1-11. Copyright 1993 Cambridge University Press 1

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2measures with achievement in French drawn fromvarious grade levels in seven different cities acrossCanada. Gliksman (1981) investigated similar var-iables with students at the university level.

The Attitude/Motivation Test Battery (AMTB)was developed to measure a number of attributesassociated with second-language learning. Thesecan be grouped into five categories, four of whichreflect concepts that are often used in this research.They are as follows:

(1) Motivation. Three different measures areneeded to assess the various components of mo-tivation. As proposed by Gardner (1985), themotivated individual is one who wants to achievea particular goal, devotes considerable effort toachieve this goal, and experiences satisfaction in theactivities associated with achieving this goal. That is,motivation is defined by three components, desireto achieve a goal, effort extended in this direction,and satisfaction with the task. One would expectthat these three attributes would be correlated withone another, but it is possible that they might notbe, in some circumstances. Consider, for example, aclassroom situation with a stern and severe teacher.It is conceivable that even individuals who are nottruly motivated to learn the material may be shownto display considerable effort in class. If motivationwere defined only in terms of effort, such individualsmay be considered to be motivated, even thoughthey do not have any desire to learn, and even findthe experience distasteful. Other examples couldalso be given to show how one of the other elementsof motivation may be elevated due to situational asopposed to motivational considerations. Given theseconsiderations, motivation is assessed in the AMTBby three measures, (a) Desire to learn the language,(b) Motivational intensity, and (c) Attitudes towardlearning the language.

(2) Integrativeness. Indices of integrativeness arecomprised of attributes that reflect a positive outlooktoward the other language group or out-groups ingeneral. Since the learning of a second languageinvolves acquiring skills associated with anothercultural group, it is proposed that the motivation tolearn the language could involve attitudes towardthat community or more general attitudes towardother groups. In earlier research (see, for example,Gardner & Lambert, 1959, 1972), attention was alsodirected toward a number of measures includingethnocentrism and authoritarianism, but in theAMTB, three measures are used; (a) Attitudestoward the target language group, (b) Interest inforeign languages, and (c) Integrative orientation.

(3) Attitudes toward the learning situation. Thisconcept refers to affective reactions toward thelanguage-learning situation. As such, it couldinvolve attitudes toward the instructor, the class,the textbooks, the language laboratory, etc. In theAMTB, attention is directed toward only two

targets, largely because they are more generalisableacross different studies. These are (a) Evaluation ofthe language teacher, and (b) Evaluation of thelanguage course.

(4) Language anxiety. This refers to students'anxiety reactions to situations in which they mightmake use of the target language. Depending on thelanguage-learning context, it could be possible toidentify many possible situations; however, in theAMTB, two general measures are used, (a) Frenchclass anxiety, and (b) French use anxiety. Theformer scale refers to anxiety aroused specifically inthe language class, while the latter refers to feelingsof anxiety that individuals experience in any contextwhere they are called upon to speak the targetlanguage.

(5) Other attributes. In many studies, some attri-butes are included in the AMTB that do not fitinto any of the above categories. In decreasing orderof use, these are (a) Instrumental orientation, (b)Parental encouragement, and (c) Orientation index.

In some studies, scores on the various scales areaggregated to produce scores on these generalconcepts, Motivation, Integrativeness, Attitudestoward the learning situation, and Language anxiety.In other studies, higher order aggregates areobtained. One such measure is the Integrativemotivation score which is the sum of Motivation,Integrativeness and Attitudes toward the learningsituation. Another aggregate score that is used insome studies is the AMI score which is the aggregateof the Integrative motivation score plus Instrumentalorientation minus language anxiety. Lalonde andGardner (1985) compared the relative predictivecapacity of the three basic aggregates, Motivation,Integrativeness and Attitudes toward the learningsituation, and found that Motivation correlatedmore highly with French grades, the behaviouralintention to study French the next year, andobjective measures of French achievement than dideither the measures of Integrativeness or Attitudestoward the learning situation.

Other research has used the AMTB to attempt tounderstand the causal relationships among attitudes,motivation and language achievement, and hassometimes included language anxiety and languageaptitude. Some researchers (Burstall, Jamieson,Cohen & Hargreaves, 1974; Backman, 1976;Strong, 1984) have argued that achievementcauses attitudes and motivation, while the socio-educational model (Gardner, 1985) explicitly pro-poses reciprocal causation. That is, it argues thatmotivation influences language achievement, andthat language achievement as well as experiences informal and informal language contexts influenceattitudes and motivation (which are viewed as someof the many possible non-linguistic outcomes).Some studies have been concerned with testing' causal models' of these relationships using linear

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2structural relations modelling (LISREL) (Joreskog &Sorbom, 1984). Many of these studies are discussedby Gardner (1985, chapter 8). Although the modelschange somewhat from study to study because ofchanges in the socio-cultural context, variablesinvestigated, and the like, all of them indicate that aprime mediator in the language-learning process ismotivation.

Still other research has made use of laboratorytechniques to investigate the causal nature ofattitudes, motivation and language achievement.Such research has the advantage that it controlsextraneous factors that otherwise might cloud causalinterpretations. It has the disadvantage, however,that it is much more artificial than even a classroomlanguage-learning situation (that may be artificialin itself) and is concerned only with the acquisitionof limited vocabulary elements. Nonetheless, suchresearch has demonstrated that the rate of learningFrench/English vocabulary pairs is faster for thosewith high as opposed to low levels of AMI (Gardner,Lalonde & Moorcroft, 1985) as well as integrativemotivation (Gardner & Maclntyre, 1991). Gardnerand Maclntyre also showed that individuals whowere studying the vocabulary items because ofinstrumental motivation (aroused by a monetaryincentive) learned faster than those without thismotivation. Thus, in a controlled laboratory settingat least, both integrative and instrumental mo-tivation were shown to influence second-languageacquisition.

Many studies have used versions of the AMTB incontexts other than those involving English Can-adian students learning French as a second language.Generally speaking, all of these studies foundevidence that motivation or some aspect of languageattitudes correlated significantly with achievementin the second language. However, examination ofthe studies reveals many different forms of thisrelationship. There may be many reasons for this.The social contexts change, the measures are slightlydifferent, the nature of the analyses vary, etc.Nonetheless, in many of them, there is evidence thataffective variables are associated with achievementin the second language. In a series of studies,Clement and his colleagues investigated FrenchCanadian students learning English as a secondlanguage (Clement, Gardner & Smythe, 1977,1980;Clement, Major, Gardner & Smythe, 1977). In all ofthese studies, the highest relationships with achieve-ment in the second language involved the indices ofself-confidence (discussed below in the section onAnxiety), but self-confidence had clear relationshipswith the motivational indices.

Many studies have been conducted outsideCanada. Laine (1977) investigated Finnish studentslearning English and found that indices of self-confidence and motivation were associated withEnglish achievement. In Belize, Gordon (1980)

found that language aptitude was the best singleindicator of English achievement among schoolstudents. An index of attitudes toward the learningsituation was the next best correlate. Muchnickand Wolfe (1982) investigated American studentsstudying Spanish as a second language, and foundthat attitudes toward the Spanish course and Spanishclass anxiety correlated significantly with grades inSpanish. Sison (1991) made use of causal modellingwith a group of American students studying Spanishas a second language. She tested a model that hadlanguage attitudes 'causing' both motivation andlanguage anxiety, motivation and language anxiety'causing' language achievement, and motivation' causing' (negatively) language anxiety. All of thepaths, with the exception of that between languageattitudes and language anxiety, were significant, andthe fit of the model was acceptable.

In a most interesting study of the role of attitudesand motivation in second-language study, Kraemer(1990) investigated Israeli Jewish students studyingeither Arabic or French as a foreign language. Shetoo made use of causal modelling to link languageattitudes, motivation, and indices of proficiency inthe other language. She also included other variables,such as social/political attitudes, political optimism,national security orientation, etc., that were necess-ary to reflect the socio-cultural setting there. Similarcausal models were obtained for both students ofArabic and French. Motivation was found to be acentral mediator in the prediction of languageachievement, but as might be expected in thiscontext, integrative attitudes were not significantcontributors to motivation. This study is particularlyinformative because it shows how it is necessary toconsider carefully the factors that can contribute tothe motivation to learn another language in differentsocio-cultural contexts.

The influence of motivational variables onsecond-language achievement has also been dem-onstrated in a military context. Lett and O'Mara(1990) investigated the predictive power of a numberof variables such as general intellectual ability,language aptitude, demographic variables such assex, etc., attitudes, motivation, learning strategies,personality variables, etc., on the learning of Korean,Russian, German, and Spanish by American militarypersonnel in intensive language-learning program-mes. As might be expected, aptitude tended to beamong the most important predictors of proficiency(especially for Korean and Russian). Attitudes andmotivation assessed during the programme con-tributed significantly to two of the three predictionscomputed for each language (the criteria wereListening, Reading, and Speaking proficiency at theend of the programme). Other variables contributedalso, but the fact remains that even in a militarycontext, attitudes and motivation were clearlyrelated to proficiency in other languages.

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2Although most attention has been directed to the

relation of attitudes and motivation to indices ofsecond-language achievement, some research hasfocused attention on other criteria. One suchcriterion is persistence in language study. Bartley(1969) investigated the relation of language aptitudeand foreign language attitudes to the tendency ofgrade 8 students in Spanish, French, German orLatin to enrol in their language course in thesubsequent term. She found that those who droppedout of language study had significantly loweraptitude scores and less positive attitudes thanstudents who continued with their studies. In afollow-up study, Bartley (1970) assessed studentattitudes at the beginning of the school year andagain at the end. She found that those who droppedout of language study had significantly less favour-able attitudes on both occasions than students whodid not drop out and also showed a significantdecrease in their attitudes from the beginning to theend of the year. In a similar study, Ramage (1990)investigated high school students of French andSpanish and found that continuing students weremore motivated and had more favourable attitudestoward the language class than students who werenot continuing with their language study. Clement,Smythe and Gardner (1978) compared high schoolstudents who continued French study with thosewho dropped out the following year, and foundthat the drop-outs had significantly lower levels ofmotivation, and generally less positive attitudes thanthose who continued language study. They alsotended to have lower levels of language aptitudeand lower levels of French achievement. By andlarge, the index of motivation (comparable to theaggregate measure of motivation described above)was the single best predictor of who would continueas opposed to drop out.

Another variable that is influenced by attitudesand motivation is that of behaviour in the lan-guage classroom. Gliksman (1976) investigated thisrelation in two studies. In the first, he contrastedindividuals characterised as integratively motivatedwith those less integratively motivated (as definedby a median split on an aggregate score based onmeasures of Integrativeness and Motivation asdescribed above). He found, over a series of classsessions, that those classified as integratively moti-vated volunteered more frequently (particularlyamong males), gave more correct answers, andreceived more positive reinforcement from theteacher than students not so motivated. Comparableresults were obtained in a second study (also seeGliksman, Gardner & Smythe, 1982). Naiman,Frohlich, Stern and Todesco (1978) obtained similarfindings though they used a somewhat differenttactic. They observed students who had been selectedby their teachers as being either among the most orthe least proficient in class, and correlated scores on

a series of attitude and motivation measures withtheir participation on 22 student-centred and 17teacher-centred classroom behaviours. Many of thecorrelations were not significant, but those thatwere provided results comparable to those obtainedby Gliksman (1976). In yet another variant on thisapproach, Roger, Bull and Fletcher (1981) comparedstudents identified by their teachers as highlyattentive and enthusiastic with students not sodescribed. The former group had significantly morefavourable attitudes toward learning foreign lan-guages than the latter group. Making the assumptionthat the teacher's characterisations reflected be-haviour in the classroom, such results again indicatean association between attitudes and classroomparticipation.

This review indicates that attitudes and moti-vation play a role in second-language acquisition.These attributes are related to measures of pro-ficiency in the second language, rates of learningvocabulary, persistence in language study andbehaviour in the language classroom. The re-lationships are complex, but taken as a unit, theysuggest that the operative variable in the language-learning process is motivation. Motivation itself,however, is also a complex of factors - proposedhere to encompass desire to achieve a goal, effortexpended in this direction, and reinforcementassociated with the act of learning.

Crookes and Schmidt (1991) provide a similarcharacterisation of motivation as it relates to second-language learning based on their consideration oftheories of motivation proposed in various fieldsof psychology. The important point is that moti-vation itself is dynamic. The old characterisation ofmotivation in terms of integrative vs. instrumentalorientations is too static and restricted. Many studies(see, for example, Dornyei, 1990; Lukmani, 1972;Gardner & Maclntyre, 1991), have shown thatachievement in a second language is facilitated byinstrumental orientations (or motivation) as well asintegrative orientations and attitudes (see, forexample, Gardner, 1985; Gardner & Maclntyre, inpress), or other motivational attributes (Clement,1986; Kraemer, 1990). Whereas motivation appearsto play a primary role in second-language learning,various other attributes, including integrativenessand attitudes toward the learning situation wouldseem to play a role in supporting levels of moti-vation.

Language anxietyA topic of rising importance in the study oflanguage learning is the role of language anxiety.This is a relatively new development largely becausethe role of anxiety in language learning was notrecognised in early research. Studies conducted inthe 1970s were difficult to interpret because of

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2contradictory results (Chastain, 1975; Kleinmann,1977; Scovel, 1978), probably resulting from usinggeneral measures of anxiety. Recent studies, how-ever, have focused on a type of anxiety relatedspecifically to language situations, termed languageanxiety. A recent volume devoted to the subject(Horwitz & Young, 1991) demonstrates its emergingsignificance.

Language anxiety can be defined as the ap-prehension experienced when a situation requiresthe use of a second language with which theindividual is not fully proficient. It is, therefore, seenas a stable personality trait referring to the propensityfor an individual to react in a nervous manner whenspeaking, listening, reading, or writing in the secondlanguage. As with other forms of anxiety, it ischaracterised by derogatory self-related cognitions(e.g. 'I can't do this'), feelings of apprehension, andphysiological responses such as increased heart rate(Endler & Okada, 1975; Eysenck, 1979; Schwarzer,1986).

Horwitz, Horwitz and Cope (1986) describeanxiety in the language classroom as a complexexperience related, in part, to communicationapprehension, social evaluation, and test anxiety.Communication apprehension is a response to thereal or anticipated act of speaking, in the samemanner as one may become anxious when speakingthe native language (Daly, 1991; McCroskey, 1978).Social evaluation apprehension emerges from thesocial nature of language use, because second-language communication involves self-presentationin a language with which only limited competencyhas been attained. The test anxiety component refersto the academic nature of many language-learningenvironments and would be relevant to thosesituations involving formal instruction. Horwitz etal. (1986) further state that language anxiety shouldbe seen as more than the sum of these parts. Theyview language anxiety '...as a distinct complex ofself-perceptions, beliefs, feelings, and behavioursrelated to classroom language learning arising fromthe uniqueness of the language learning process'(p. 128).

Evidence in favour of Horwitz et a/.'s theory hasbeen accumulating rapidly. Horwitz (1986) foundthat other measures of communication apprehen-sion, social evaluation anxiety and test anxiety eachcorrelated significantly with anxiety in the languageclassroom. Further, language anxiety was sig-nificantly, negatively correlated with expectedand obtained grades in the language course. Alongsimilar lines, Maclntyre and Gardner (1989) usedfactor analysis to identify two anxiety dimensions,General anxiety and Communicative anxiety. Twoof the four scales defining Communicative anxietyreferred specifically to second-language situations.Further analyses revealed that Communicativeanxiety negatively influenced second-language vo-

cabulary learning and production, while Generalanxiety was not associated with performance in thesecond language. A follow-up study (Maclntyre &Gardner, 1991) that employed 22 anxiety measuresidentified a separate dimension of Language anxiety.Scales of communication apprehension, social eva-luation, and test anxiety were not associated withthis dimension, supporting the suggestion thatlanguage anxiety is a specific, relatively unique typeof apprehension. Finally, Young (1990) has shownthat the most anxiety-provoking tasks in languageclassrooms involve public communication and/orevaluation, comprising the three sources of anxietyidentified by Horwitz et al. (1986).

The negative effects of language anxiety onsecond-language acquisition have been well es-tablished. Several studies have shown a negativecorrelation with grades in language courses(Gardner, Smythe, Cle'ment & Gliksman, 1976;Horwitz, 1986; Trylong, 1987; Gardner &Maclntyre, in press). Course grades are one of themost salient measures of language performance;however, correlations with such a global measuretell us little about the manner in which anxietyoperates.

There have been several studies that have usedmore specific indices of performance. Generally, theeffects of language anxiety have been shown to bepervasive over several types of tasks (Maclntyre &Gardner, 1992). Young (1986) reports correlationsbetween scores on an oral proficiency interview andlanguage anxiety. Tucker, Hamayan and Genesee(1976) found that being more adventuresome, lessanxious, and more willing to use the foreignlanguage was associated with achievement in agroup of late immersion students. Gardner, Lalonde,Moorcroft and Evers (1987) found that French classanxiety was significantly correlated with each offour proficiency measures (word production, themetest, listening comprehension, and self-ratings ofproficiency) in grade 12 and grade 13 students.Trylong (1987) found significant correlations ofanxiety with achievement on written tests, oralquizzes, and final grades in a first-year universityFrench course.

Experimental investigations have tried to induceanxiety in a controlled environment in order tostudy its effects. A study by Steinberg and Horwitz(1986) examined two groups. In one, anxiety wasaroused by impersonal treatment and videotapingstudents while they described ambiguous pictures.A second group was treated in a friendly mannerand was not videotaped. The group in whichanxiety was induced was found to be significantlyless interpretative than the more relaxed groupwhen describing the scenes. A later study byGardner, Day and Maclntyre (1992) found thatsimply introducing a video camera is not sufficientto produce anxiety. They suggest that the social

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2interaction in Steinberg & Horwitz's study was akey element in producing differences in the twogroups. Finally, Maclntyre and Gardner (1991)asked beginning adult language students to write abrief essay on either an anxiety-provoking ex-perience or a confidence-building one. The groupswho wrote about anxiety tended to perceivethemselves as less proficient than the group whodescribed a confidence building experience, and themost anxiety-provoking experience reported wasalmost always related to speaking.

One theory of the development of languageanxiety has been offered by Maclntyre and Gardner(1989). They propose that it develops as the result ofrepeated, negative experiences with the secondlanguage. Language anxiety is seen as a learnedemotional response. At the earliest stages, thelanguage learner may experience a form of stateanxiety, a transient apprehensive experience. Afterrepeated occurrences of state anxiety, the studentwill come to reliably associate anxiety with per-formance in the second language. Once it hasdeveloped, language anxiety can have a pervasiveimpact on the performance of language students,even affecting future learning. The negative effectsof language anxiety would be expected to diminishover time as proficiency increases and more positiveexperiences accumulate.

This model has two different components. At theearliest stages of language learning, language anxietyis not very meaningful because negative experienceshave not produced the negative affect or expec-tations of failure that typify anxiety. After severalnegative experiences, however, the relation betweenanxiety and performance should be observed. Astudy by Chapelle and Roberts (1986) found thatthe correlation between English class anxiety andTOEFL scores at the beginning of a semester wasnot significant; however, by the end of the semester,the correlation had increased to a significant level.

The suggestion that anxiety levels are highestearly on in language learning and then decline asproficiency increases also has been supported byresearch. Gardner, Smythe and Brunet (1977)compared beginning, intermediate, and advancedstudents of French on subtests of the AMTB beforeand after an intensive summer-school languagecourse. The highest levels of anxiety were shown bythe beginners and the least by the advanced students.All three groups showed less French class anxiety atthe end of the course than they did at the beginning.Desrochers and Gardner (1981) found that anxietylevels declined significantly among English-speakinggrade-eight students after a four-day excursion to aFrench-speaking community. Among adult stu-dents, Gardner, Smythe and Clement (1979) foundthat proficiency increased while anxiety decreasedafter an intensive French summer-school pro-gramme. The results of these studies indicate that

as experience and proficiency increase, anxietydeclines in a fairly consistent manner.

The theoretical models and related researchdiscussed thus far have considered the negative,sometimes debilitating, effects of language anxiety.The suggestion has been made, however, that someanxiety may be helpful in second-language learning(Scovel, 1978). Kleinmann (1977) reports positivecorrelations between this 'facilitating anxiety' anduse of difficult linguistic structures. On balance,however, this hypothesis has not been well sup-ported. In almost all studies that employ a measurespecifically related to language anxiety (rather thana more general anxiety measure) negative cor-relations with achievement have been obtained.Moreover, the similarities between facilitatinganxiety and motivation are striking, both in howthey are measured and their conceptual definitions.In fact, it has been suggested that the facilitatingeffects of anxiety could be attributable to motivation(Gardner et al., 1992).

The results of these studies of language anxietysuggest that anxious students will have lower levelsof verbal production, will have difficulty in basiclearning and production, will be less likely tovolunteer answers in class, and will be reluctant toexpress personally relevant information in a second-language conversation. Further, it would appearthat language anxiety arises from early negativeexperiences, particularly with speaking. Whilelanguage anxiety may be high initially, it would beexpected to decline as the student gains proficiency,provided that the student continues to study or usethe second language.

In some ways, the antithesis of the anxiousstudent is the self-confident one (Maclntyre &Gardner, 1991). In multicultural settings, however,self confidence may mean more than a lack ofanxiety. Cle'ment (1980, 1986) considers self con-fidence to be a superordinate construct, encom-passing both a lack of anxiety and positive self-ratings of proficiency in interactions with membersof the target language community. In a series offactor analytic studies of francophones learningEnglish in bilingual communities, Clement and hiscolleagues found evidence for a Self-confidencefactor. Clement, Major, Gardner and Smythe (1977)found a self-confidence factor that was defined bypositive teacher ratings, positive course evaluation,use of the second language outside the classroomand a lack of language anxiety. Clement, Gardnerand Smythe (1977, 1980) found that self-confidencecould be defined by a lack of language anxiety andpositive self-rated proficiency in the second lan-guage. In all three studies, self-confidence was foundto be related to objective measures of proficiency, aswell as indices of motivation.

Gardner, Smythe and Lalonde (1984) investigatedthe factor structure of similar scales in their study of

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2English Canadian students of French across differentregions of Canada and various grade levels. Thefactors were found to replicate fairly consistently,and language anxiety was most often associatedwith a factor related to the self-perception of Frenchcompetence and, less often, a French achievementfactor. These authors conclude that when theopportunity to use the foreign language is presentin the community, language anxiety is negativelycorrelated with second language proficiency.

Clement's model has emerged from a com-bination of these studies with the work of socialpsychologists on the social consequences of languageacquisition and has received empirical support(Clement & Kruidenier, 1985). Clement's theory(1980, 1986) postulates two motivational processespresent in bilingual/muticultural communities. Theprimary motivational process is based on theantagonistic interplay between integrativeness andthe desire to affiliate with the other group, on theone hand, and fear of assimilation and worry overlosing one's cultural identity, on the other. Clement(1980) argues that the result of this process willdetermine the extent to which individuals makecontact with members of the other group. Thesecondary motivational process is based on thefrequency and quality of contact between the studentand members of the target language community.This contact, if positive, can improve self con-fidence with the new language, thus contributing tothe student's motivation. Within monolingualcommunities, self confidence would be less influ-ential because of fewer instances of intergroupcontact.

Labrie and Clement (1986) found support for thismodel in that less frequent, but positive contact wasassociated with improved self confidence. Also, selfconfidence was associated with motivation to usethe second language. In a similar manner, Pak, Dionand Dion (1985) investigated Chinese immigrantslearning English in a large city and found that selfconfidence with English (as defined by self-ratedproficiency only) was associated with linguisticassimilation into the English speaking community.Thus self-confidence, in a multicultural context,appears to have motivational as well as anxiety-related components.

Research on language anxiety and self-confidenceindicates that these variables play an important rolein second language learning. Language anxiety isnegatively related to proficiency whereas self-confidence is positively related. The processesunderlying the two related concepts appear similarin that they both develop as a result of experiencesassociated with learning and using the secondlanguage. One possibility is that a multiculturalsocial milieu enhances correlates of language anxietyand produces a more complex construct combininglanguage anxiety, self-perceptions of proficiency

and attitudinal/motivational components. Anotheris that the underlying construct is the same in bothmulticultural and unicultural contexts, but that thefocus of the different researchers, and the measuresused, emphasise slightly different aspects. Clearly,more research is needed to clarify which possibilityis most appropriate. In any event, indices of lan-guage anxiety are generally associated with self-perceptions of competency and objective measuresof proficiency and often show low but negativecorrelations with motivation. In both researchtraditions, too, language anxiety is shown to beassociated with experiences with the second lan-guage.

Such considerations lead to the generalisation thatlanguage anxiety has a negative effect on second-language acquisition. Its relation to motivationalcharacteristics is less clear cut, and it may well bethat this is mediated by socio-cultural factors.Intuitively, it seems reasonable to propose thathigh levels of anxiety might serve to lessen one'smotivation to learn the language, because theexperience is found to be painful, and that highlevels of motivation result in low levels of anxietybecause the student perceives the experience posi-tively and tends to be successful - both of whichdecrease anxiety. Nonetheless, further research isrequired to determine the precise nature of therelationship, if any, between anxiety and mo-tivation.

A theoretical integrationIn Part I of this review (Gardner & Maclntyre, 1992)attention was directed toward the relation to second-language learning of three classes of cognitiveindividual difference variables, namely, Intelligence,Language aptitude, and Language-learning stra-tegies. In this part, attention focused on attitudes,motivation, language anxiety and self-confidence.This final section is devoted to providing atheoretical overview of how such factors caninfluence, and in some cases be influenced by, thelanguage-learning process. This overview is pre-sented in the form of a revised version of the socio-educational model of second-language learning thatwas presented in Part I (Gardner & Maclntyre,1992). This revised model is presented in Figure 1.In the figure, the socio-cultural milieu is shown asover-riding all aspects of the model. That is, whenconsidering the process of second-language ac-quisition, it is recommended that close attention bedirected to the social context in which the learningis taking place. Although Au (1988) argues that sucha concept 'may serve only to render the theoryimmune to disconfirming evidence, thereby gran-ting infallibility to the theory...' (p. 85), omissionof such a concept would make the theory much toosimplistic. Rather, what is required is research that

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2

Antecedentfactors

Socio-cultural milieu

Individualdifferencevariables

Languageacquisitioncontexs

Outcomes

' >

Biological

Experiential

Intelligence

Languageaptitude

Strategies

Languageattitudes

*

MotivationA

Languageanxiety

- - -

1

1

1

,h

.Formal

\ /Linguistic

/ >-

Informal

Non-linguistic

Figure 1. Schematic representation of the Socio-education model of second-language acquisition.

delineates the significant features of the social milieuthat influences the role of individual differences inlanguage acquisition. Some research based on adifferent theoretical model is already under way inthis regard (see, for example, Clement, 1980, 1986;Labrie & Clement, 1986), while at least one study(Clement & Kruidenier, 1983) demonstrated thateven the reasons for studying a second languagetended to group together to form different orien-tations as a function of the socio-cultural milieu.Moreover, Kraemer (1990) demonstrated that thesocio-educational model could be modified suc-cessfully to take into consideration social forcesthat might operate in a specific language-learningsituation. It is simplistic to assume that a language-learning model should not make explicit recognitionof the very dominant role played by the socialcontext. Would one really expect, for example, thatattitudes toward the other community would berelated positively to the motivation to learn thelanguage of that community by military personnelwho may consider the community a potentialenemy ?

The model posits that there are a number ofAntecedent factors that must also be consideredwhen attempting to study the role of individualdifference variables in the process of learning asecond language. These are described simply as

biological and experiential in the model, with noexemplars or roles shown. Examples of such factors,however, would include gender, age, prior languagetraining, etc. Rather than attempt to describe all thepossible roles that such factors might play, it seemsbest to propose that, at a minimum, they shouldbe considered by an investigator when planning astudy. It seems reasonable to propose, for example,that there is a biological foundation for intelligence,that age might influence what language-learningstrategies are used, or that prior experience with alanguage could influence levels of language attitudes,motivation and/or language anxiety. Clearly, anystudy of individual differences in second-languageacquisition would have to ensure that subjects wererelatively homogeneous with respect to priortraining in that language in order to rule outconfounding effects of prior achievement in thecurrent language-learning process.

Six major individual difference variables areshown in Fig. 1. Intelligence, language aptitude, andlanguage-learning strategies are considered as cog-nitive individual difference variables. In the modelthey are shown as being relatively independent ofone another (they are not linked by any directionalarrows), even though it is recognised that theymight correlate significantly in any study. Themeasures might well share some variance in

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2common, but it seems reasonable to assume that afactor analytic study with sufficient marker variableswould show that the effective variances are in factindependent. This has been shown to be the case, atleast, for intelligence and language aptitude (Gardner& Lambert, 1965).

Language attitudes, motivation and languageanxiety are considered as affective individual dif-ference variables. The concept of language attitudesrefers to any attitudinal variables that might beimplicated in the language-learning context. Earlier,two major classes, Integrativeness and Attitudestoward the learning situation were discussed, butothers have been investigated (see, for exampleGliksman, 1981; Clement & Kruidenier, 1985;Kraemer, 1990). In the model, language attitudesare shown as having a causal influence on motivation(as indicated by the directional arrow). The point is,that motivation needs an affective basis to bemaintained, and it seems reasonable to argue thatattitudes serve this function. In the figure, moti-vation is shown as having a causal influence and asbeing causally influenced by language anxiety (thetwo arrows). These two variables tend to benegatively correlated, and it seems meaningful toargue, given our current state of knowledge, thatnot only might high levels of motivation tend todepress language anxiety but also that high levels ofanxiety might decrease motivation. The model alsoshows causal links from language anxiety andmotivation to language-learning strategies. Theresearch by Oxford and Nyikos (1989) and by Rostand Ross (1991) points to the motivational founda-tion for the use of language-learning strategies whileanxiety reduction techniques could be considered asindirect learning strategies (Oxford, 1990a).Obviously, more research is required, but at thepresent time, it seems meaningful to postulate suchcausal connections.

When attention is directed to the Languageacquisition contexts, all of the individual differencevariables, with the exception of language attitudes,are shown as having a direct effect on learning in theformal language-learning environment (solid di-rectional arrows). This is meant to indicate that inany learning situation where material or skill isbeing transmitted to a learner in some way,individual differences in intelligence, languageaptitude, the use of language-learning strategies,motivation and language anxiety will influence howsuccessful that individual will be in acquiring thatmaterial or skill. Each will play a role in its ownway, some of which were discussed or hypothesisedin earlier sections of this article.

Only motivation is shown to have a direct rolein the informal context (the dotted arrow frommotivation to informal contexts). Because of thevoluntary nature of the informal context, it isexpected that individuals who are not motivated

will simply not take part in the context, while thosewho are motivated will. Once an individual entersan informal context, however, the other individualdifference variables will be implicated and couldinfluence how well material is learned. As above,the role played will be influenced by the relevanceof the task to that particular variable. The indirecteffects of the other variables are shown by thebroken arrows linking them to the informalcontexts. No link is shown between languageattitudes and the informal language learning context.This is because it is assumed that motivation medi-ates any relation between language attitudes andlanguage achievement.

Both formal and informal language acquisitioncontexts are assumed to have direct effects on bothlinguistic and non-linguistic outcomes (the solidarrows). Thus, the model formally recognises theimportance of what takes place in both contexts.Teachers, instructional aids, curricula, and the likeclearly have an effect on what is learned and howstudents react to the experience. Moreover, themodel postulates a causal link from the linguisticoutcomes to the non-linguistic outcomes, a recog-nition that individuals' reactions to the learningexperience will depend to some extent on theirrelative degree of success. Finally, it is proposed thatboth linguistic and non-linguistic outcomes willhave an influence on individual difference variables.Based on the research to date (Oxford & Crookall,1989; Oxford & Nyikos, 1989) it seems clear thatachievement in the language will influence the useof language-learning strategies. For this reason,linguistic outcomes are shown as having an effect onlanguage-learning strategy use. Non-linguistic out-comes are expected in turn to have direct effectson language attitudes, motivation and languageanxiety.

Other variables might be added to this modelwith no loss in generality, but the intent of this two-part article was simply to review the major ones.Thus, for example, some research has demonstratedlinks between personality variables and languageachievement (see, for example, Lalonde, 1982;Ehrman, 1990; Oxford, 1990 b) and between lan-guage achievement and learning style variables suchas field dependence/independence (Oxford, 19906;Chapelle & Green, 1992). These variables were notconsidered in this article, in part because of spacelimitations and in part because of our belief thattheir functional roles in the language-learningprocess are still not clear. Further research couldwell profit from a consideration of their relationshipto variables discussed in this two-part article.

Cook's review of the literature in 1978 dem-onstrated the breadth of variables that can influencelanguage acquisition. Over the past 15 years, theboundaries have been greatly expanded and ourknowledge base deepened. The sheer volume of

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2research currently available points to the complexityof the language-learning process and the need forfurther research to consider the many remainingquestions. As this review attests, there is still muchto be done. Hopefully, the theoretical formulationproposed here will help in this regard.We would like to thank the Social Sciences and HumanitiesResearch Council of Canada for their support by grant no. 410-90-0195 to the first author and Doctoral Fellowship no. 453-91-1277 to the second author.

ReferencesARSENIAN, S. (1945). Bilingualism in the post-war world.

Psychological Bulletin, 42, 65-86.Au, S. Y. (1988). A critical appraisal of Gardner's social-

psychological theory of second-language (L2) learning.Language Learning, 38, 75-100.

BACKMAN, N. (1976). Two measures of affective factors as theyrelate to progress in adult second-language learning. WorkingPapers on Bilingualism, 10, 100-22.

BARTLEY, D. E. (1969). A pilot study of aptitude and attitudefactors in language dropout. California Journal of EducationalResearch, 20, 48-55.

BARTLEY, D. E. (1970). The importance of the attitude factor inlanguage dropout: a preliminary investigation of group andsex differences. Foreign Language Annals, 3, 383-93.

BURSTALL, C , JAMIESON, M , COHEN, S. & HARGREAVES, M.(1974). Primary French in the balance. Windsor, Berks.: NFER.

CHAPELLE, C. & GREEN, P. (1992). Field independence/dependence in second-language acquisition research. LanguageLearning, 42, 47-83.

CHAPELLE, C. & ROBERTS, C. (1986). Ambiguity tolerance andfield independence as predictors of proficiency in English asa second language. Language Learning, 36, 27-45.

CHASTAIN, K. (1975). Affective and ability factors in second-language acquisition. Language Learning, 25, 153-61.

CLEMENT, R. (1980). Ethnicity, contact and communicativecompetence in a second language. In H. Giles, W. P.Robinson & P. M. Smith (eds.), Language: social psychologicalperspectives. Oxford: Pergamon.

CLEMENT, R. (1986). Second language proficiency and accul-turation : an investigation of the effects of language statusand individual characteristics. Journal of Language and SocialPsychology, 5, 271-90.

CLEMENT, R., GARDNER, R. C. & SMYTHE, P. C. (1977).Motivational variables in second language acquisition: astudy of francophones learning English. Canadian Journal ofBehavioural Science, 9, 123-33.

CLE'MENT, R., GARDNER, R. C. & SMYTHE, P. C. (1980). Socialand individual factors in second language acquisition.Canadian Journal of Behavioural Science, 12, 293-302.

CLEMENT, R. & KRUIDENIER, B. G. (1983). Orientations insecond language acquisition. I. The effects of ethnicity,milieu, and target language on their emergence. LanguageLearning, 33, 273-91.

CLEMENT, R. & KRUIDENIER, B. G. (1985). Aptitude, attitudeand motivation in second language proficiency: a test ofClement's model. Journal of Language and Social Psychology, 4,21-37.

CLEMENT, R., MAJOR, L. J., GARDNER, R. C. & SMYTHE, P. C.(1977). Attitudes and motivation in second language ac-quisition: an investigation of Ontario francophones. WorkingPapers on Bilingualism, 12, 1-20.

CLEMENT, R., SMYTHE, P. C. & GARDNER, R. C. (1978).Persistence in second language study: motivational con-siderations. Canadian Modern Language Review, 34, 688-94.

COOK, V.J. (1978). Second-language learning: a psycho-10

linguistic perspective. Language Teaching and Linguistics:Abstracts, II, 73-89.

CROOKES, G. & SCHMIDT, R. W. (1991). Motivation: reopeningthe research agenda. Language Learning, 41, 469-512.

DALY,J. (1991). Understanding communication apprehension:an introduction for language educators. In E. K. Horwitz, &D. J. Young (eds.), Language anxiety: from theory and researchto classroom implications. Englewood Cliffs, NJ: Prentice Hall.

DESROCHERS, A. & GARDNER, R. C. (1981). Second languageacquisition: an investigation of a bicultural excursion experience.Quebec: International Centre for Research on Bilingualism.

DORNYEI, Z. (1990). Conceptualising motivation in foreignlanguage learning. Language Learning, 40, 45-78.

ERHMAN, M. (1990). The role of personality type in adultlanguage learning: an ongoing investigation. In T. S. Parry,& C. W. Stansfield (eds.), Language aptitude reconsidered,pp. 126-78. Englewood Cliffs, NJ: Prentice Hall Regents.

ENDLER, N. S. & OKADA, M. (1975). A multidimensionalmeasure of trait anxiety: the S-R Inventory of general traitanxiousness. Journal of Consulting and Clinical Psychology, 43,319-29.

EYSENCK, M. W. (1979). Anxiety, learning and memory: areconceptualisation. Journal of Research in Personality, 13,363-85.

GARDNER, R. C. (1985). Social psychology and second languagelearning: the role of attitudes and motivation. London: EdwardArnold.

GARDNER, R. C , DAY, B. & MACINTYRE, P. D. (1992).Integrative motivation, induced anxiety, and languagelearning in a controlled environment. Studies in SecondLanguage Acquisition, 14, 197-214.

GARDNER, R. C., LALONDE, R. N. & MOORCROFT, R. (1985).The role of attitudes and motivation in second languagelearning: correlational and experimental considerations.Language Learning, 35, 207-27.

GARDNER, R. C , LALONDE, R. N., MOORCROFT, R. & EVERS, F.T. (1987). Second language attrition: the role of motivationand use. Journal of Language and Social Psychology, 6, 29-47.

GARDNER, R. C. & LAMBERT, W. E. (1959). Motivationalvariables in second language acquisition. Canadian Journal ofPsychology, 13, 266-72.

GARDNER, R. C. & LAMBERT, W. E. (1965). Language aptitude,intelligence, and second-language achievement. Journal ofEducational Psychology, 56, 191-9.

GARDNER, R. C. & LAMBERT, W. E. (1972). Attitudes andmotivation in second language learning. Rowley, Mass.: New-bury House.

GARDNER, R. C. & MACINTYRE, P. D. (1991). An instrumentalmotivation in language study: who says it isn't effective?Studies in Second Language Acquisition, 13, 57-72.

GARDNER, R. C. & MACINTYRE, P. D. (1992). A student'scontributions to second language learning: Part I. Cognitivevariables. Language Teaching, 25, 4, 211-20.

GARDNER, R. C. & MACINTYRE, P. D. (in press). On themeasurement of affective variables in second languagelearning. Language Learning.

GARDNER, R. C. & SMYTHE, P. C. (1981). On the developmentof the Attitude/Motivation Test Battery. Canadian ModernLanguage Review, 37, 510-25.

GARDNER, R. C , SMYTHE, P. C. & BRUNET, G. R. (1977).Intensive second language study: effects on attitudes,motivation and French achievement. Language Learning, 27,243-61.

GARDNER, R. C , SMYTHE, P. C. & CLE'MENT, R. (1979). Intensivesecond language study in a bicultural milieu: an investigationof attitudes, motivation and language proficiency. LanguageLearning, 29, 305-20.

GARDNER, R. C , SMYTHE, P. C , CLEMENT, R. & GLIKSMAN, L.(1976). Second language acquisition: a social psychologicalperspective. Canadian Modem Language Review, 32, 198-213.

http://journals.cambridge.org Downloaded: 12 Dec 2014 IP address: 137.205.50.42

State of the art: Second language learning. Part 2GARDNER, R. C. & SMYTHE, P. C. & LALONDE, R. N. (1984).

The nature and replicability of factors in second languageacquisition. Research Bulletin No. 605, Department ofPsychology, The University of Western Ontario, London,Canada.

GLIKSMAN, L. (1976). Second language acquisition: the effects ofstudent attitudes on classroom behaviour. Unpublished master'sthesis, University of Western Ontario, London, Canada.

GLIKSMAN, L. (1981). Improving the prediction of behavioursassociated with second language acquisition. UnpublishedDoctoral dissertation, University of Western Ontario,London, Canada.

GLIKSMAN, L., GARDNER, R. C. & SMYTHE, P. C. (1982). Therole of the integrative motive on students' participation inthe French classroom. Canadian Modern Language Review, 38,625^7.

GORDON, M. E. (1980). Attitudes and motivation in second languageachievement: a study of primary school students learning English inBelize, Central America. Unpublished Doctoral dissertation,University of Toronto.

HoRwrrz, E. K. (1986). Preliminary evidence for the reliabilityand validity of a foreign language anxiety scale. TESOLQuarterly, 20, 559-62.

HoRwrrz, E. K., HORWITZ, M. B. & COPE, J. (1986). Foreignlanguage classroom anxiety. Modern Language Journal, 70,125-32.

HORWITZ, E. K. & YOUNG, D.J. (1991). Language anxiety.Englewood Cliffs, N.J.: Prentice Hall.

JORESKOG, K. G. & SORBOM, D. (1984). LISREL VI: Analysis oflinear structural relationships by the method of maximum likelihood.Mooresville, Indiana: Scientific Software Inc.

KLEINMANN, H. H. (1977). Avoidance behavior in adult secondlanguage acquisition. Language Learning, 27, 93-107.

KRAEMER, R. (1990). Social psychological factors related to the studyof Arabic among Israeli Jewish high school students. UnpublishedDoctoral dissertation, Tel-Aviv University.

LABRIE, N. & CLEMENT, R. (1986). Ethnolinguistic vitality, self-confidence and second language proficiency: an investigation.Journal of Multilingual and Multicultural Development, 7,269-82.

LAINE, E. (1977). Foreign language learning motivation in FinlandI. Turku: Turku University Press.

LALONDE, R. (1982). Second language acquisition: a causal analysis.Unpublished Master's thesis, University of Western Ontario,London, Canada.

LALONDE, R. N. & GARDNER, R. C. (1985). On the predictivevalidity of the Attitude/Motivation Test Battery. Journal ofMultilingual and Multicultural Development, 6, 403-12.

LAMBERT, W. E. (1955). Measurement of the linguistic domi-nance of bilinguals._/oMiW of Abnormal and Social Psychology,50, 197-200.

LETT, J. & O'MARA, F. E. (1990). Predictors of success in anintensive language learning context: correlates of languagelearning at the Defense Language Institute Foreign LanguageCenter. In T. S. Parry & C. W. Stansfield (eds.), Languageaptitude reconsidered, 222-260. Englewood Cliffs, N.J.: Pren-tice Hall Regents.

LUKMANI, Y. M. (1972). Motivation to learn and learningproficiency. Language Learning, 22, 261-73.

MACINTYRE, P. D. & GARDNER, R. C. (1989). Anxiety andsecond language learning: toward a theoretical clarification.Language Learning, 39, 251-75.

MACINTYRE, P. D. & GARDNER, R. C. (1991). Language anxiety:its relation to other anxieties and to processing in native andsecond languages. Language Learning, 41, 513-34.

MACINTYRE, P. D. & GARDNER, R. C. (1992). The pervasive'

effects of language anxiety on second language performance.Canadian Psychology, 33, 276 (Abstract C#ll) .

MARCKWARDT, A. H. (1948). Motives for the study of modernlanguages. Language Learning, 1, 3-99. (Reprinted in LanguageLearning (1988), 38, 161-9.)

MCCROSKEY, J. C. (1978). Validity of the PRCA as an index oforal communication apprehension. Communication Mono-graphs, 45, 192-203.

MUCHNICK, A. G. & WOLFE, D. E. (1982). Attitudes andmotivations of American students of Spanish. The CanadianModern Language Review, 38, 262-81.

NAIMAN, N., FROHLICH, M., STERN, H. H. & TODESCO, A.(1978). The good language learner. Research in Education ScriesNo. 7, Ontario Institute for Studies in Education, Toronto,Ontario.

OXFORD, R. L. (1990a). Language learning strategies: what everyteacher should know. New York: Newbury House.

OXFORD, R. L. (1990fe). Styles, strategies, and aptitude:connections for language learning. In T. S. Parry & C. W.Stansfield (eds.), Language aptitude reconsidered, 67-125.Englewood Cliffs, N.J.: Prentice Hall Regents.

OXFORD, R. L. & CROOKALL, D. (1989). Research on languagelearning strategies: methods, findings, and instructional issues.Modern Language Journal, 73, 404-19.

OXFORD, R. L. & NYIKOS, M. (1989). Variables affecting choiceof language learning strategies by university students. ModernLanguage Journal, 75, 292-300.

PAK, A. W.-P., DION, K. L. & DION, K. K. (1985). Correlatesof self-confidence with English among Chinese students inToronto. Canadian Journal of Behavioural Science, 17, 369-78.

RAM AGE, K. (1990). Motivational factors and persistence inforeign language study. Language Learning, 40, 189-219.

ROGER, D., BULL, P. & FLETCHER, R. (1981). The constructionand validation of a questionnaire for measuring attitudestowards learning foreign languages. Educational Review, 33,223-30.

ROST, M. & Ross, S. (1991). Learner use of strategies ininteraction: typology and teachability. Language Learning, 41,235-73.

SCHWARZER, R. (1986). Self-related cognition in anxiety andmotivation. Hillsdale, N.J.: Lawrence Erlbaum Associates.

SCOVEL, T. (1978). The effect of affect on foreign languagelearning: a review of the anxiety research. Language Learning,28, 129-42.

SISON, M. P. A. (1991). The relationship of college student attitudes,motivation, and anxiety to second language achievement: a test ofGardner's model. Unpublished Doctoral dissertation, Uni-versity of Southern California.

STEINBERG, F. S. & HORWITZ, E. K. (1986). The effect ofinduced anxiety on the denotative and interpretive content ofsecond language speech. TESOL Quarterly, 20, 131-6.

STRONG, M. H. (1984). Integrative motivation: cause or resultof successful second-language acquisition? Language Learning,34, 1-14.

TRYLONG, V. L. (1987). Aptitude, attitudes, and anxiety: a study oftheir relationships to achievement in the foreign language classroom.Unpublished doctoral dissertation, Purdue University.

TUCKER, G. R., HAMAYAN, E. & GENESEE, F. H. (1976).Affective, cognitive and social factors in second-languageacquisition. Canadian Modern Language Review, 32, 214-26.

YOUNG, D.J. (1986). The relationship between anxiety andforeign language oral proficiency ratings. Foreign LanguageAnnals, 12, 439-48.

YOUNG, D.J. (1990). An investigation of students' perspectiveson anxiety and speaking. Foreign Language Annals, 23,539-53.

11