MLJ Response Article

How Does AnxietY Affect SecondLanguage Learning? A RePIY toSparks and GanschowPETER D. MACINTIT.EDeparlLnt of PVchoWJUnivetsiry Couzge oI CaP BletonS1dn4, Nwa Scotia, Canitlo BlP 6L2Enail : Wcinq@swc.uccb ns..a

The linguistic coding deficit h)?othesis, introduced by Sparks and Ganschow (1991' 1993a'

t993b1,"postulates thit language aPtitude is the Primary source ofindividual differences in

language achievement. This may be seen to reduce affective variables' such ar language

anx"ietyi to the role of unfortunate side effects, devoid of explanatory Power' This paper

advocates that language anxiety can play a significant causal role in creating individual

differences in both language learning and communication' It reviews evidence from inves-

tigations of anxiety in general and siudies of the role of anxiety in the language learning

pt..rs"" und.ot.tud". that the linguistic coding deficit hypothesis.makes a significant

omission by assignlng mere epiphenomenal status to affective variables in general and

language anxiety in Particular.

FOR MANY STUDENTS, LANCUAGE COLIRSES

are the most anxiety-provoking courses that

they take (Hor{itz, Horwitz & Cope, 1986; Mac-

Intyre & Gardner, 1989, 1991b), and several au-

thors have expressed concern over the amount

of anxiety present in language classes (Cope-

Powell, 1991; Muchnick & Wolfe, 1982) Camp-

bell and Ortiz (1991) considered the levels of

language anxiety amoflg university students to

be "alarming" (p. 159) and estimaled that up to

one halfofall language students experience de-

bilitating levels oflanguage anxiety. ln general,

language learning conlexts apPear to be Parl ic-ularly prone to anxiet,v arousal (Price, 1991).

In rhe qeneral l i lerature on second language

'fhe Mod'n language loundl, 79, i (1995)0026J9o2 /95 /90-99 fi.!'0 /0@'1595 'fhe Mdittn langrye Jounal

learning, there has been some question about

whether anxiety exerts an imPortant influence

on second language learning. Despite the expe-

riences of students, teachers, and administra-

tors, early empitical studies were ". . - unable to

establish a clear picture of how anxiety affects

language learning and performance" (Horwitz

& Young, 1991, p. xiii). In 1978' a review of the

sparse literature concerning anxiety's role in

language learning Pointed to "mixed and con-

fusing" results (Scovel, 1978, p. 132). More re-

cently, Horwitz and Young (1991) noted that"Exactly how anxiety imPedes language learn-

ing has not been resolved" (p. 177). Finally' re-

cent art icles deal ing with Sparks and Gan-

schow's l inguist ic coding defici t hypothesis

(LCDH) (Sparks & Ganscho*;1991, 1993a, 1993b)

have relegated language anxiety to the status of

an unfortunate side effect. The purpose of the

present discussion is to demonstrate that lan-

PeEr D. MacIntJre

guage anxiety can,influence language learningby describing the processes that are believed tounderlie its effects. The second purpose is toshow that the LCDH is incomplete as an expla-nation for individual differences in second lan,guage learning without formally considering af-f€ctive vaiiables, such as language anxiety.

Notwithstanding the difficulties cited in thepreceding paragraph, a good deal of researchhas shown that anxiety is associated with prob-lems in second language learning (Maclntyre &Gardner, 1991c). For example, Gardner, Smythe,Cl€ment and Gliksman (1976) found thar, as thegrade level increased from grade 7 to grade 11,language anxiety became a better predictor ofachievement in second language courses. Sig-nificant negative correlations between lan-guage anxiety and course grades have been re-ported fo. languages such asJapanese (Aida,1994), Spanish (Horwitz, 1986), and French{ Maclnrvre & Gardner, l99lb). Language anxi-ety has also shown significant negative correla-tions with standardised proficiency tests, bothwritten (Gardner, Lalonde, Moorcroft & Evers,1987) and oral (Young, 1990). In addition, stud-ies have shown that the rate ofvocabulary learn-ing is slower for studenrs who reporr experienc-ing language anxiery (Maclntyre & Gardner,1989, 1994a). Finally, factor analytic studies haveshown consistently that language anxiety is as-sociated with a factor defined by self-rated pro-ficiency, actual proficiency, or both with thesecond language (Cl6menr, Gardner & Smythe,1977, 1980; cardnet Smythe & Lalonde, 1984).To explain these effects, it is useful to placelanguage anxiety into the broader context ofthe psychology of social anxieties.

LANGUAGE ANXIETY AS ASOCIAL ANXIETY

Almost everyone has experienced anxiety at$ome time or in some type of situation. A pro-nounced growth in research into the anxietiesof everyday life has taken place over the pastl0-20 years. These common forms ofapprehen-sion among psychologically healthy people canbe broadly classified as social anxieties (Leary,1990t. It has been argued elsewhere rhar lan-guage anxrety stems primarily from the socialand communicative a5pects of language learn-ing and therefore can be considered as one ofthe social anxieties (Maclntyre & Cardner,1989, 1991b). This allows us to draw uDon thisbroader body of work for explanations of theeffects of anxiety in second language learning.

ol

Most theorists agree that socially based anxi-eties have cognitive, affective, and behaviouralcompon€nB (Sarason, 1986; Spielberger, 1966,1983). Social anxiety is defined by "(l) feelingsof tension and discomfort, (2) negative self-evaluarions, and (3) a tendency to withdraw int}re presence of others" (Schwarzer, 1996, D. 1).The arousal of anxiety in any social siruirioncan have these and other affective. cognitive,and behav ioura l consequences (C levenger ,1984; Fischer, 1988). For example, the affectiveexperience of ant(iety includes feelings of ap-prehension, uneasiness, and fear (Whitmore.1987t. Among rhe cognitive effects are increasesin distracting self-related cognition, expecta-tions offailure, and a decrease in cognitive proc-essing ability (Wine, 1980). The behavioural di-mension includes reactions such as rlcrcases lnsympathetic nervous system arousal, inhibitedaction$, and attempts to escape the situation(Levin, 1980).

Within the c]u'rrcr.t Zeitgeist of ps/chologt in-terest in the effects of anr(iety is focused on thecognitiye dimension (Schwarzer, 1986). Thiscomponent of anxiety has received considenbleresearch attention and has shor.n the strongestcorrelations with task performance (Holroyd &Appel, 1980). Sarason (1986) argued thar this isnot sufficient reason to ignore the emotionaland behavioural aspects of anxiety altogether,but it does suggest value in stud)'ing the cogni-tive component in its own right. A focus on thecognitive dimension also provides a clear expla-nation for rhe negative eflecrs of anxiety on ian-guage learning,

ANXIETY FROM A COGNITTVEPERSPECTTVE

The cognitive and affective components ofanxiety were identified by Liebert and Morris(1967) as "worry" and "emotionality" respec-tively. Sarason (1986) defined worry as '.. . , dis-tressing preoccupations and concerns aboutimpending events" (p. 2l). This preoccupationofteD takes the form of self-related cognition,which is seldom beneficial for task perfor-mance. The basis for rhe negarive effects ofanx-iety on cognitive activity were summarised byEvsenck (1979) who argued thar:

worry and other task-irrelevanr cognitive actjvitiesassociated with anxiety always impair rhe quatirv ofperformance. The major reason for this is t}la! rhetask-irelevanr information involved in worrv andcognitive self-concern competes with task-relevantinformation for space in the processing system. As a

I-re".rtt,

frighfy anxious subjects are effectiv€ly in a

dual-task or divided atteniion situadon' rn contta3t

io the non_alxious subjects who primarily Processtask-rclevant information (p 364)'

whereas distractions caused by self-related

cosnition can exPlain the negative effects of

"r,ii.ty ott cognitive activity, it has been sug-

cested that some anxiety may actually improve

ierformance (AlPert & Haber' 1960: Scovel '

i9781. To address this possibility, Eysenck (1979)

further suggests that anxious individuals will

.o-p"n"ut" for the increased cognit ive de-

manis by increased effort and that "the ertent

ro which anxiely either faci l i rates or imPairs

Derformance is determined by the extent to

wnich high-anxiety subjects comPensate for re-

duced processing effectiveness by enhanced ef-

fort ' { ; . 365). Thus. rhe arousal of anxiety can

influence both the quality of Performance and

rhe amount of effort invested in it Eysenck

(1979) suggested that much of the research into

the effects of anxiety has assumed that effort

exoenditure is relatively constant, and there-

fore the focus has been on the way in which

anxiety effects the quality of Performance'Eysenck's (1979) theory is able to account for

rhe often cited interaction between the effects

of anxiety and ability on task Performanc€ (for

example, Hunsley, 1985; SPielberger, 1983)' sim-

i lar to the well-known Yerkes-Dodson Law

(Smith, Sarason & Sarason, 1982). This "law"

describes a curvilinear relationship between

anxiety and Performance as a function of task

diff iculty (see Figure l) . To the exrent that a

given task is relatively simPle, anxiety seems to

have little negative effect and may actually im-

prove performance through increased effort'

However, as the demands on the system in-

crease, t}le extra effort may not fully comPen-

sare for the cognit i \e interference, and anxietv

will begin to have a negative effect As demand

further exceeds abi l i ry. the impairment caused

bv anxiety arousal worsens. Thus, those who do

not experience anxiety will be able to Processthe information more quiclJy, more effectively,

or both compared to those who are distracted

by task-irrelevant cognition.whereas anxietv mav cause deficits in cognt-

tive processing and imPair task performance, it

should also be tecognised that task Perfor-mance can provoke anxiety (Maclntyre & Gard-

ner, 1994a). There is no doubt that anxiety can

be provoked by having difficulty in learning th€

language, but that is not to conclude that anxi-

.ry playt no role in contributing to such diffi

culties in the first place These are two different

Thc Mulen Langnge Jotul 79 (1995)

FIGURE Ilnverted "U ' relation between anxiety andperformance.

Atrxi€ty

I

conclusions arising from a focus on two differ-

ent Doints in tlre continuous, cyclical process

invoiving anxiety, cognition, and behaviour'

RELATIONS AMONG ANXIETY,COGNITION, AND BETIAVIOUR

The relations among anxiety, cognition, and

behaviour are best seen as recurrive or cyclical,

where each influences the other (Leary' 1990;

Levitt, 1980). For examPle, a demand to answer

a question in a second language class may cause

a itudent to become anxious: anxiety leads to

worry and rumination. Cognitive Performanceis diminished because of the divided attention

and therefore pelformance suffers' leading to

negative self-evaluations and more self-depre-

catlng cognition which further imPairs Perfor-mance, and so on (see Figure 2). For some stu-

dents, this is a frequent course of events, and

anxiety becomes reliatlly associated with any

situation invohing the second language. Once

established, fiis association leads students to be-

come anxious at tie prospect ofsecond language

learning or communrcatton.At tlli root of this model is a distinction be-

Pct r D. blaclnt re

FIGURE 2Recursive relations among anxiery, cognition,ano oenavlour.

+-.+

. 9 3

action. For example, anxious students mayworry about misunderstanding l inguisric struc-tures or worry about inferring meaning fromcontext because of the potential for embarrass-rng errors. In educational settings, Tobias (19g6)has found-rhar anxiery may impiir rhe abil iry totaKe In lntormatton, process it, and lo retrieveit when necessary,

In the literature on language anxiety, a fewstudies have been conducted rhar examine theeffects of language anxiety before the produc-tion stage. Maclntyre and Gardner 0991b) founda significant negative correlation between lan_guage anxiety and the ability to repeat shortstrings of numbers and to recall vocabularvilems. This demonstrates (hal anxiety can limiithe use of borh shorFterm and long-term mem_orv. lr is interesling. and relevant ro Sparks andGanschow's arguments (1991, 1993;, 1993b).that native language versions of the same tasksdid not co.relate with language anxrery.

A more extensive follow-up study (Maclnt Te& Gardner, 1994b) found that anxiery impairedperformance on a wide range of repetition, lis_tenrng, comprehension, reading, and learningtasks. Horwitz et al. (1986) argued rhat re--searchers should examine the .,subtle" effects ofIanguage anxiety. Maclntp-e & Gardner (1994b)concluded thar relatively small effects on specificlanguage learning processes can accurnulite andwill eventually lead to clear differences betweenthe more anxious and the more relaxed students.

In an expe mental study, Maclntyre & Gard-ner (1994a) used a video camera to induce stateanxiety during a computerised vocabularylearn ing lask . Resu l rs showed rha t anx ie rvarousal was associaled wittr performance defi_cits in the learning, recall, and functional use ofthe vocabulary items. When the effects of thevideo camera had dissipated and no longer leadto state anxiety arousal, performance improvedrelative ro those who were experiencing moreanxiety arousal. Based on these findings, we seethat when state anxiety is provoked, perfor_mance on second language tasks suffers, but noperformance deficits are observed when learn-ers are not experien€ing anxiety. Thus, activeinterference seems to arise from state anxietv.and that inrerference can occur at any stage ofthe learning process. This is relatively stiongevidence that anxiety arousal can act as a causalagenr in crealing individual differences in sec_ond language learning.

These srudies contradict the claim thar lan_guage anxiety is a consequence, rather than acause, of problerns in language learning. Fur_

tween state and trait anxiety. On one hand,state anxiety is an imm€diate, transitory erno-t ional experience wirh immediare cognit ive ef_fects. On the odter hand, trair anxiety-is a stablepredisposit ion lo become anxious in a widerange of situations (Spielberger, 19g3). It shouldbe emphasised that state anxiety is the reaction,and trait anxiety represents the tendency to re_act in an anxious manner. The negative effectsof anxiety discussed above, such as coqnitiveinterference, can only be associated with rheimmediate anxiety experience and therefore re-fer to state anxiety arousal.

The majority of larguage anxiety scales, likeHorwitz et al.'s (1986) Foreign Language Class-room Anxiety Scale (FLCAS), are trait-basedscales. The i tems on most language anxteryscales tend to focus on the anxiety experiencedwhile spealdng the second language (Maclnryre& Gardner. 1988, l99lc). These scales are mostuseful in identifying those individuals who haveexperienced state anxiety arousal in the pastand in predict ing tho3e who wil l be most l i lelvto experience state anxiety in the luture. Theavailable measures of language anxiety focusprrmartty on output or production of the sec-ond language (Maclntyre & Cardner, lggg,l99lc) and may be hightighring a general con_cern tor one s publ ic self-presenra t ion (Carver& Scheier, 1986; Leary, 1990).

To make the theo.etical foundation of lan_guage anxlety more complete, we must also al_low for the possibility rhat anxiety influencessecond language activities, such as listening,lea-rning. and comprehension. There is a poren-tial role for anxiety in these processes becausethe anticiparion of second language use in re_ceiving informalion can provoke an anxiety re-

94

th€r, this evidence suggests that the effects of

anxiety are not limited to problems encoun-

tered during sPeaking but per\ade the entire

language learning process. Thus. the LCDH

molel advocated by Sparks and Ganschow(1991, 1993a, 1993b) is at the same time comPat-ible with the anxiety literature and incomPletewithout it. The two explanations for individual

differences in language achievement do not ap-

pear to be antithetical, in sPite of the aPParent

claim to tie contrarY.

.LANGUAGE ANXIETY OR NATIVELANGUAGE APTITUDE DIFFERENCES"

The title ofthis section is taken from the sub-

title ofSparks and Ganschow's (1991) article de-scribing the LCDH. The LCDH is based onwork in the area of reading disabilities and the

work on linguistic aPtitude by Caroll (1981) and

others, It suggests that individuals who exPeri-ence difficulties in the efficient coding of lin-guistic stimuli in the native language' Partic-ularly phonemes, will have similar difficulties in

s".ord hnguag" acquisition. lt is further sug-gested that ". . .the majority of Poor FL leamersare dislinguish€d from good learners by pho-nological coding difficulties . ." (Sparks &Ganschow, f993b, P. 289).

Soarks and Ganschow (1991, 1993a' 1993b)propose the LCDH as an alternative to allectiveexolanations for individual differences in sec-ond language learning. They have claimed re-peatedly that affective variables, and anxiety in

particular, are not likely to be the causes of suchindividual differences. Affective variables areconsidered mere side effects of having difficulties in coding the native language. "ln ourview, low motivation, Poor attitude, or highlevels ofanxietya!e, most likely, a manifestationof deficiencies in t}te efficient control of one'snative language, though they are obviouslv cor-related with difficulty in FL learning" (Sparks

& Ganschow, 1991, p. 10).The existence of a relation between affeciive

variables and individual differences in languageproficiency is not disPuted by SParks and Gan-schow. What is up for debate are the bases onwhich Sparks and Ganschow (1991, 1993a'1993b) criticised affective variables, the Propo-sition that native language aptitude is an alter-native to affective exPlanations and that affectis simply a "manifestation of aptitude." Each ofthese issues will be addressed in tlre discussionbelow.

In their s€ries of articles on the LCDH, Sparks

'fhe M&rn LanguAt Jturid 79 495)

& Ganschow offered a numb€r of criticisms of

affective lariables that require a response. For

example, a recent article (1993b) attacked bothaffective variables and language learning strate-gies as explanations for individual differences in

language achievement. Their criticism of lan-guage learning strategies was that the strategiesextend into other domains of cognitive function-ing and therefore imply general intellectual fail-

ings rather than specific language learning diffi-

culties. First, it should be not€d that some of the

strategi€s studied by researchers (for example,Oxford, 1990) could apply to numerous learning

domains, otlers are highly specific to language

learning, and still others are specific to Particu-lar stages of language learning. Thus, the criti.cism of language leaming strategy research is

overgeneralised. Second, it is somewhat ironic

that SparLs and Ganschow (1993b) themselvespointed out tllat it is not usually the case thatpoor language learners are having trouble in

other areas (p. 29f). This makes it seem moreplausible tlrat langlage students who are notusing appropriate strategies have simPly nottransferred them to the language learning do-main. In fact, the training in phonetic coding

advocated by Sparks and Ganschow (1993b)

might be considered strategy training.As for affective rariables, Sparks and Gan-

schow (1993b, P, 291) noted two criticisms, nei-ther of which is coovincing. The first criticismwas that affective variables are t,?ically meas-

ured using self-rePort questionnaires and thatsuch methods generally have mcasurementproblems, Again, SParks and Ganschow (1993b)

overgeneralise their criticism to all self-rePortmeasures, apparently b€cause of problems with

measures of learning styles and native languagemodality testing. ln fact, highly reliable andvalid measures of language-related affectivevariables are available; Gardner's (1985) Atti-rude/Motivation Test Battery and Horwitz etal.'s (1986) FLCAS are two examples. In anycase, quarrels with self-rePort methodologiesare essentially irrelevant to the Present debatebecause Sparks and Ganschow (1991, 1993a'1993b) argue that affective variables function as

conseouences, not causes, of individual differ-ences. That line of argument leads to the con-clusion thar even perfectly reliable and validmeasures of affective variables would still holdno explanatorv power; a conclusion that was

chatlenged by several studies cited above.

The second crit icism, based on "specula-

tions" by Crain (1989), is that affective variablesare effectively "sealed off" during cognitive

Ptra D. Ma.In\te

processing. The above discussion of the cogni-tive effects of anxiety and the empirical resultscited with it strongly challenge this notion. B€-yond this, however. is the more insidious sugges-rioo that affective lariables are "unrelated tolanguage" and that only variables specific tolanguage learning can be used to explain prob-lems. This was referred to as the "assumption ofspecificity," but with it Sparks and Ganschow(1993b) seemed to be using "language" to referto the purely linSuistic aspects of language.

Their argument warrantg two criticisms, First,measures l ike Gardner's (1985) Attitude,/Motivation Test Battery, Oxford's (1990) Strat-egy Inventory for Language Learning (SILL),Horwitz et al.'s 0986) FLCAS, and similar meas-ures are all specificatly related to languagelearning experiences. They may not measurepurely linguistic variables, but they certainlyarelanguage related, Maclntyre and Gardner0989, 1991b) argued strongly tiat, in the caseof language anxiety, the assessment of anxietymust specifically refer to that arising in languagelearning contexts in order ro obtain consistentcorrelations between anxiety and languagelearning. Essentially, this is t}Ie assumption ofspecificity. Second, rather than posing a protr-lem, the fact that affective variables are notmeasures of pure linguistic processing is a valu-able ass€t. Affect represents a unique source ofrariance impinging on the system. To arbitrarilyrestrict potential explanations for individualdifferences in language learning achievementdiminishes our potential to understand fully thepfocess.

In an earlier paper, Sparks and Ganschow(1991) go to some length to criticise the work ofHorwitz et al. (1986) and their FLCAS. Perhapsthe key criticism is that Horwitz et al. may haveoverlooked native language deficits as the causeof both higher anxiety and lower proficiency. Inthis case, the argument would be that aptitudeinfluences both proficiency and anxiety. Thist)?e of criticism is almost alwats valid for cor-relational studies, where a third variable mightinfluence both of the variables in the correla-tion. In fact, similar criticism might be levelledagainst Sparks and Ganschow's (1991) LCDHbecause it does not take into account the Doten-tial effects of affective variables when consider-ing the relation between aptirude and achieve-ment. One variant of this argument is thatanxiety arousal could influence both scores onlanguage aptitude and language achievementlests. Another possibility is that the arousal ofsocial anxiety hampers coding of authentic

95

communication in either the native or secondlanguage, conceivably through the divided at-tention scenario described above.

All of these possibilities are consistent withthe type of model shown in Figure 2. As notedaboye, these and similar variables are not in-volved in a strict linear process. Aptitude, cog-nition, anxiety, and language learning behav-iours function in a recursive, cyclical pattern.Thus, aptitude can influence anxiety, anxietycan influence performance, and performancecan influence anxiety. Sparks and Ganschow(1993a) also allowed for the possibility that sru-dents can be trained to increase their phoneticcoding ability; thus aptitude also can be influ-enced. The interesting questions arising fromthis interplay have less to do with uhzther or.e

riable affects the otier and more to do withlr/raa do those effects occur? Sparks and Gan-schow (1991) can be criticised for taking an"either/or" stand.

Additional criticism of Horwitz et al. {1986)was offered in an analysis of FLCAS items(Sparks & Ganschow, 1991). Unfortunatelt rhatdiscussion obscured important differencesanong the items with respect to the source ofthe affective reaction. For example, the items,"feel overwhelmed by the number of rules re-quired to speak language" and "afraid otherstudents will laugh when I speak language," ap-pear to be addressing two very different sourcesof anxiett one cognitive and the other social.Sparks and Ganschow (1991) classified both a5"oral expressive." It is granted that one or bothitems may b€ related to aptitude in some way, butother factors clearly are implicated. An anxiousstudent may feel overwhelmed by the number ofrules because their ability to process informa-tion is temporarily restricted by anxiety-relatedcognition. The fear of being laughed at byother students may occur for a number of rea-sons other than language coding deficits, suchas poor group d).namics and compet rl ireness inthe language classroom (Bailey, 1983; Cl6ment,D6rnyei & Noels, 1994), both of which can beconsidered sources of social anxiety. Clearlv.then, affect is more than a manifestation of ap-titude (for a description of potential sourc$ ofanxiety in the language classroom, see Young,1991).

This demonstrates a broader crit icism ofSparks and Ganschow's model (1991). Their rhe-ory focused exclusively on cognitive ability fac-tors in terms of the coding of linguistic stimuli.They omitted consideration of social factors in-volved in language learning, such as classroom

96

interaction with teachers and other students.the degree of exposure to the language in thecommunity, erhnolinguistic vitality, motivation,attitudes, intergroup relations, and contactwith the target language community. Addi-tional cognitive factors are neglected as well,such as the amount of effort invested in lan-guage study, the student's expectations for suc-cess, and perhaps most importantly, languagelearning strategies that might lessen the impactof native language deficirs. ln shorr. the linguis-tic coding deficit h'?othesis neglects the con-text in which language learning occurs (Cl6m-ent, 1980; Gardner & Maclntyre, 1992, 1993)and ignores the potential for social context toinfluence cognitive processes (Fiske & Taylor,r99r).

Sparks and Ganschow (1993a) provided suchan example. In a series of case studies support-ing the LCDH, the aurhors reported a casestudy of an unmoti ted learner (.Amy") withhigh aptitude who was doing poorly in a Spanishcourse. They did not, however, explicitly inte-gTate this r)?e ofpossibility inro their theoreticalframework. Rather, they made a sweeprng gener-alisation about unsuccessful foreign languagelearners with high aptitude having generallypoor attitudes toward school, and dismissed thecase as being unusual. Perhaps a student withh-igh apritude (l ike "Amy') may perform poorlyif deliberate cognitive processes are preoc-cupied with self-presentation concerns o; otheranxiety-related cogqitive activity, and thereforettre automatic phonetic coding processes arenot accessed. In cases such as these, phoneticcoding impairmenr is not the best place to lookfor the cause of rhe language leirning diff i-culty. This problem is likely to be one of affecr(e.9,, low motivation, poor attitude, high anx!ety) impinging on the cognirive sysrem:

It should be srressed rhar rhe argument hereis not thar rhe LCDH is in.orrect. By definit ion.native language aptitude determines the pa-rarnerers wirhin which language learning wil loccur. It is reasonable to h)?othesise that theselimitations will ertend to second languages aswell. These links should be fully explored. Theposition advocated here is that the propensityto reach one's full potential as a languagelearner is partially determined by affective vaii_ables (Skehan, 1989), such as anxiety. For exam-ple, wil l ingness ro take a language course maybe influenced by altitudes roward rhe secondlanguage group, persistence at study may be in_fluenced by motintion, and the efficiency oflearning may be influenced by anxiety. For sru-

Tlu Mo.la11 Language Jotnnat 79 e9951

dents to maximise their potential in the secondlanguage, they must have a positive contribu-tion from the constellation of affective factors(Brovm, 1991; Rubin & Thompson, 1982). Thepreccding discussion has demonstrated some ofthe wa)s in which the LCDH can be expandedto include affective Eriables.

CONCLUSION AND IMPLICATIONS

As noted above, researchers have had somedifficulty in demonstrating the role of anxietyin second language learning. Even wit-h a numberof empirical studies demonstrating a relationbetween anxiety and achievement, quesl ions re_main about the manner in which anxiery exertsan influence. In proposing tlreir LCDH, Sparksand Ganschow (1991, 1993a, 1993b) have dis-counted anxiety and other affective variables aspossible causes of individual differences in lan_guage achievement.

The response to the questions about the roleof anxiety in second language learning, basedon the study ofthe psychological effects of anxi_ety, may be stated as follows: language learningis a cognitive activity that relies on encoding,storage, and retrieval processes. and anxieiycan interfere wirh each of these by creating adivided attention scelario for anxious students_Anxious students are focused on both the taskat hand and their reactions to it. For example,when responding to a question in class. the inx-ious student is focused on answerins theteacher's quesrion and evaluaring rhe social im-plications of the answer uhib gioing t. To theextent that self-related cognitioI1 rncreases,task-related cognition is restricted, and Derfor-mance suffers. Anxious srudenrs rherefore wil lnot learn as quickly as relaxed students. lt isconceivable that witlr increased effort. anxiousstudents could reach the same level of perfor_mance .!s relaxed students (see Maclntyre &Cardner, 1994a).

Considering the abov€ discussion, it wouldappear that the effects of anxiety rnay be morecomplex than has been implied by Sparks andGanschow (t991, 1993a, 1993b). If anxrety ansesduring learning then anxiou5 srudenrs wil l per-form poorlv because rhel hare ls2rns6 i....However, if anything, anxiery may be rnorestrongly aroused by speaking than by learning(Marlntyre & Gardner, 1994a, 1994b). Thus.anxiery may also interfere with the student'sability to demonstrate the amount that she orhe does know The classic example is the stu_d€nt who knows the material but ..freezes uD"

P.ur D. Ma.Itutyre

on a test- Anxious students are caught in thisdouble bind: rhey have learned ler" orr'd ̂"y notbe able to demonsrrare rhe information tharihey have learned. Further, the cyclical relationb€tween anxiety and task performance suggestsrhat as students experience more failure, theiranxiety level may increase even more.

The LCDH advocated by Sparks and Gan-schow (1991) does not recognise the languagelearning context. It is therefore incomDlete.Given rhe exrensive research on the wiys in\^hich the social conrext can influence coqni-rive processes in general (Fiske & Taylor, l99l)and language learning in particular (Cl6ment,1980; Gardner & Maclntyre, 1992, 1993; Skehan,1989), this seems to be a significant omission.

One area in which both the cognitive inter-ference model of anxiety and Sparks and Gan-schowt (1991, 1993a, 1993b) linguistic codingdeficit hypothesis converge is in the remedialaction suggested to address language learningdeficits. In both cases, it is suggested r}tat at-tempts to reduce language anxiety may requiresome skills training as a supplement to anxietyreducrion stralegies in order to compensate fordeficiencies created by anxiety arousal, nativelanguage problems, or both.

91

Bro$o, H. D (l9gll. Bra,ling ttu larLguag. btrizr. !t-mouth, ME: Inlercultura.l prcss.

Campb€ll, C. M., & Ortiz,J. (1991). Helping srudentsover€ome foreign languagc anxiety: A foreignlanguage anxi€ty workshop. In E. K. Horwitz &D. J. Young (Edt.\, Iaiguagt sn&!O: Fon thantand Beach to dattwn in#i(,ltiorr.� lpp. lal-tbli.Englewood CliIIs, Nf: prentice Halt.

Caroll, J. B. (1981). Tw€nry-five years of research onforeign language apritude. In E. K. C. Diller(Ed.), Individual di.fcrar4 and unhssals in lan,guage laning aryiauL (pp. 8g-U7). Rowley, MA:Newbury House.

Carver, C. S,, & Schei€r, M. F. (1986). Functional andd'sfunctional respons€s ro anxiety: The inter_actioo between expectancies and sellfocusedatrention. In R. S€hwarzer (Ed.r, Sq fttald ry-nititnin antbt, and rottuotion (pp. l1l_l42). Hillsdale, NJ: Erlbaum.

Cl6ment, R. (1980). Ethnicity, colrtact, and communi.cative competence in a second language. InH. Giles, W P Robinson & p. M. Smith (Eds.),btnguagc srvial prAabgtnt ptspai*s (pp. 147_154). Oxford: Pergamon press.

Cl€ment, R., D6rnyei, 2., & Noles, K. A. (1994). Mo_tivation and (he for€ign language clarsroom: Astudy of Hungarians learning English. Laagz4g,Lenine 44, 417-448.

Cl6ment, R., Gardner, R. C., & tury,rhe, p. C. 097?). Mo-tiyational i?riables in second language acquisi-tion: A study of Francophones barning English.C6a&lnlo@al { Hwtu"at Sriaw, 2129_133.

Cl6medt, R., Gardner, R. C., & Smythe, p C. (1980).Social ald individuat factors in se€ond lan-guage acquisition. Canadian linfitul oJ fuhalrio\&lScien4 12 293-302.

Clevenger, T. (1984). An analrsis of res€arch on thesocial anxieries. InJ. A. Daly &J. C. Mccroskey(Eds.), Awiding contmun;cat;on: Slry/4\r, ret;unaand cortmunicatiorr drychcrl\ion (pp. 219_236).Beverly Hills, CAj Sage.

Cope-PoweU, J. (1991). Foreign language classroomanxi€ty: lnstitutional responses. In E. K. Hor-wiE[z & DJ. Young (Eds.), tanguag. a'f�ic4, &unthAry end flscarxh tn ct,.tsto.,t t inpltuatio r (pp. t6g_176). Englewood CliITs, NJ: prentice HatL

Crane, S. (1989). Why poor readers misundenta-nd spo.ken sentences. In D. Shankwriler & t. Liberman(EAs.r, hotdig, ond tuldng diebiltq @p.133_165).Ann Arbor MI: University of Michigan press.

Elsenck, M. W (1979). Anxiery, learning and memory:A reconceptualizarion . Jolttrd oI Rlteafth in persotoliD, 13, 3$-3a5.

Fischer, W. F. 11988). Thaniat of anrictJ (Znd ed.). Lan_ham, MD: University Press of Amcrrca.

Fiske, S. T., & Taylor, S. E. (1991). ,%cdl.olnlt,lrz. NewYork: M(Crah, Hi l l .

Cardner, R. C. 11985'). bcial ?syhotog, and saond tan-guo,g, acqtirition: Th. nL of attitu.d.s and 7.at;vstion.London: Edward Arnold.

Gardner, R. C., Lalo[de, R. N., Moorcroft. R.. &

ACKNOWLEDGEMENTS

This research was facilitated by a post Doctoral Fel-lowship from the Social Science and Humanities Re-search Council of Canada. Thc author would like tothank R. Cl6menr, K. Noels, and the enonymous re-\ierlers for their comments on drafts of thrs rnanuscrrpr.

REFERENCES

Aida, Y (1994). Examination of Horwitz. Horwitz. &Cope s consrruct of foreign Ianguage anxiety:The case of students ofJapanese. Modern Lan-guage Joumal, 78, t55-768.

Alperr, R., & Haber, R. N. (1960). Anxrery rn aca-demic achievement sit]d tiol.,s, Joufial of Abnr-,vt and &,cial Prchatig, Iq 207-215.

Bail€y, K. M. (1983). Compeririv€ness and anxiery inadult second language learning: Looking a,and ,rroryi the diary srudies. In H. W: S€liger &M. H. Long (Eds.t, CJasyooa oruna,atttnl insetonn langry aquindaa (pp. 67-102). Rowley,MA: Newbury House.

98

Evers. F. T. (1987). Second language attrition:

The role of doti iion a d rLse. Jow'nal of l/tn'

guag, anl Soeial Pryholag' 6' 29-47

Gardner. R. C., & Maclntyte, P D. (1992) A student's

contribution to second language learning. PartI: Cognitive Evariables. I'anguge Tinching, 25'211'220.

Cardne!. R. C., & Maclntyr€, P. D. (1993). A studenfscontribution to second laryuage learninS. Part lI:Affective variables. language Tii&ching, 26 l-11,

cardner, R. C., Smythe, P C., Cl€flent, R., & Gliks-man, L. (1976). Sccond language acqutuition: Asocial psychological persPecti\re, Candidn Md'cm Languagc Raiat, 32, $a-2l3.

Gardner. R. C., Smythe, P C., & Lalonde, R N.(1984). Th, NMte and RtfliaabilitJ of Fa.urs in S.l'ond Language Aequitilion. (Research BulletinNo. 605). Londofi, Ontario: University of West-ern ontario.

Holroyd, K. A., & ApPel, M A 0980). Test anxietyand physiological responding. In I. G. Sarason

lBd.), Tett ansicr: Trwry, ns&fth and aPplicotions(pp. 129-151). Hillsdale, NJ: Erlbaum.

Horwitz, E. K. (1986). Preliminary evidence for thereliability and validiry of a foreign languageanxiety scale. IESOL EtMafJ, 2A,559-562.

Horwitz, E, K., Hoi'itz, M. 8., & Cope, J. (1986).

Foreign language classroom amiety. Mod?rnLanguage hlrnd" 70, 125-112.

Horwitz, E. K., & Young, D.J. 0991). keface. In E. K.Horwitz & D. J. Young (Eds.\, Languagr auiztl:Frorfl thior) and reszanh to datfi@m imPlicat;ons(pp. xii-xiv). Englewood Cliffs, NJ: PrenticeHdl.

Hunsley, J. (1985). Test arxiety, academic Perfor-. mance, and cognili'te apprzisals. Jolnnal oJ fulx-

cdimal baholq, 77, 678-682.Leary, M. R. (1990). Anxiery cognition alld behaviori

ln search of a broader peFpective. In M. Booth-Butterfield (Ed.r, Con tuntalion, aBnititm, annaflxteq (pp. 39-44). Newbury Park, CA: Sage.

Levit!, E. E. (1580\. fhe Prchdtg ofAnrlrD. Hillsdale,NJ: Erlbaum.

Lieb€rt, R. M., & Morris, L. W. (1967). Cognitive andemotional components of test anxiety: A dis-linction and some initial data. Pqcfutogical R2-

M5 2q 975-918.Maclntyre, P D., & Gardner, R. C. (1988). Tfu nesnm-

flent of aMury and apwittttiotr: b sctcnd lawuagcbaming: An anndad. Uthogaplry. ( Research Bul-letin No. 672). London, Ontario: University ofWesrern Ontario.

Maclnryre, P D.. & Cardner, R. C. (1989)- Anxiety andsecond language learning: Toward a theoreticalclarifi,cation. Lengu.agr lzanin& 32, 251-275.

Maclntyre, P D., & Cardner R. C. (1991a). lnvestigat-ing language class anxiety using the focusedessay technique. Modzrn Language JounaL 75,296-304.

Maclntyre, P D, & Gardner, R. C. (1991b). Languageanxiety: lts relation to orher anxieties and top-

Tlv MdEn Languagc Journal 79 (19955

rocessing in Dalive and secood languages. l-an-gulrg. barning, 41, 513-534.

Maclntyre, P D., & Gardner, R. C. (l99lc). Me$odsand results in $e study of foreign languageanxieiy: A revie* of the lite.^t\re. I'angryeI'ening, 41, 85-117.

Maclntyre, P D., & Gardner, R. c. (1994a). The ef-fec!s of induced anxiety on three stages of cog-nitive processing in computeriscd vocabularylear'ritng. Studi.s in 9end languzg, Aqitition, 161-t7

Maclntlre, P D., & Gardner, R. C. (1994b). The subtleeffects of language anxiety on cognitive proc-essing in the second l^ngnage. lang4e I'avn-ing 44,283-305.

Muchnick, A. G., & \4'olfe, D. E. (1982). Attitudes andmotivations of American students of SPanilh.Canadian Modan Langage tuvia1r, i8,262-28t.

Oxford, R. L. (1990r. langnge banring stralqi.s: Wha!eL,ed tzaahzt thatd hnou. N€w York: Heinle.

Price, M. L. (1991). The subjective exp€rieoc€ of for-eign language anxielv: Interviews with highlyanxious students. In E. K. Homitz & D. J.Young (Eds.), Langua4e an*bE: Ftun hary annrcsearch to classtoom inplielttions (pp. 101-108).Edgtewood Cliffs, NJ: Prentice Hall.

Rubin, J., & Thompson, I. (1982). Hot! to It o ,wr2errc6$n languag, len6 Boston: Heinle.

Sarason, l. G. (1986). Test anxiety, worry, and cogni-tive interference. ln R. Schwarz€r (Ed.), gtr-

ftIattn cqnirion in anxizt, and ndilnlrm (PP. 19-34). Hillsdale, NJ: Erlttaum.

Schwarzer R. (1986). self-related cognition in anxietyand motivation: An introduction. In R. Schwar-zer (Ed.), W-ftlatd cognition in a'|;wq and noliln-a''n (pp. l-17). Hillsdale, NJ: Erlbaum.

Scovel, T. (1978). The effect of affect on foreign lan-guage learninS: A review of the anxiety re-search. Languag, IatrrLing, 28' 129-142

Skehan, P (1989). lnltuiduzl dilJemcd in s.@nn bnguag'baning. London: Edward Arnold

Smith, R. E., Sanson, I. G., & Sanson, B. R. (1982). Al-chdog: Thz fftrn,ias ol lphsltiir. New'lbrk: HarP€r &Row

Sparks, R. L., & Ganschow, L. (1991). Foreign lan-

Suage learning diffetences: Affectl!€ or natrv€language aptitude differetces? Mdan lLzngwgc

Journal, 75, 3-16.Sparks, R. L., & Ganschol^, L. (1993a). The impatt of

native language learning problems on foreignlanguage learning: Case studv illusrrations oflhe linguisric coding deficit h)?othesis. ,VodtrtLanguage Jrurnol" 7 7,'. 8 -7 4.

Sparkq R. L., & Canschon, L. (1993b). Searching fo!the cognitile locus of foreign language learningdifficulties: Linkjng first and second languagelearning. Mulen Iaflguage lamal" 77, 289-302.

Spielb€rger, C. D. (1966). Complex learning and aca-demic achievemenr. In C. D. SPielberger (Ed.),AttxizE and Mux'ior (pp. 361-398). Ne\t York: Ac-ademi€ Press.

Petet D. Madnryte

Spiefb€rger, C. D. (79-89). Manualf.,,tu surt -ttait auictlin atut, (font y). palo Alto, CA: Consulring psy_chologists press.

Tobia5. S. {1986). Anxieryand cognir ive pro(essing ofrnsfructron. In R. Schwarzer \Ed.t gtf-rett t t lco-gniton in anxiaD and ̂ ot;""uo" tpp. 't_SC).Hillsdale, NJ: Erlbaum.

Whitmore, R. Q9ATI Li.ving u;th stress and aruial. Man-chester, England: Manchester University press.

Wine,J. D. (198i). Cognirive-attentional rheori of testanxiety. In I. c. Sarason (Ed.), Iist anrbr: Thz,or, nearci and o@cadr"r (pp. 549_385). HiIs-dale, NJ: Erlbaum.

Young, DJ. (1990)..The relationship berween arxieryand foreign language oral pioficiency ratings.fofagn Language Anrurk, D, 539_b59.

Young, D. J. (lggl). Creating a low_anxiery classroomenvlronment: What does the language anxietyresearch suggesr? Modan Langu;g Jounta!, 7i,426-431.

99

Sparks and Ganschow respond innext issue, 79, 2 (f995), oi the

in theML].

CaIl for Articles: AAUSC Volume in LanguageProgram Direction

THE AMERICAN ASSOCIATION OT' UNIVERSITY SUPERVISORS, COORDINATORS, AND DI-:"..-,:.Il

t-:g" langulqe

programs announces the seventh volume in their series, Issuzs in Lan_guag? L'togfttm Dh?rtion, pubiished by Heinle and Heinle.

Patutns and polioes: Th" Ch4ng,ng. D.nogrophw of Fmeign Languag? Instr..rlctiolledited by Judirh E. Liskin-Gasoar-ro

This volume will address ,n" ,-ii:l-:I,:nlging demographics on loreign tanguage programs:student. populal ions. curriculum, marerrats, speclal programs.Deadline for submission is February r, r99b. tarlier s-ubmission is encouraged. rhe series is re-fereed by the AAUsc Editorial Board ana "rrr.. t roru.. r.o* il;;;..-". papers are to tre sent to:

Rofessor Judith E. Liskin-GasparroDepartment of Spadish & portugueseUniversity of lowall4 PhilliF Halllowa City,IA i2242E nai I : jud.ith-li*i n -gosparm@ uioua_ du(319) 335-2244: Fax: (Sl9) 335-2270