DSpace at VNU: The effects of input enhancement and recasts on the development of second language pragmatic competence

On: 09 April 2015, At: 05:33Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1

On: 09 April 2015, At: 05:33

Publisher: Routledge

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Click for updates

Innovation in Language Learning and Teaching

Publication details, including instructions for authors andsubscription information:

http://www.tandfonline.com/loi/rill20

The effects of input enhancement and recasts on the development of second language pragmatic competence

Minh Thi Thuy Nguyena, Hanh Thi Phamb & Tam Minh Phamba

English Language and Literature, National Institute of Education,Nanyang Technological University, Singapore, Singapore

b Faculty of English Language Teacher Education, University ofLanguages and International Studies, Vietnam National University,Hanoi, Viet Nam

Published online: 07 Apr 2015

To cite this article: Minh Thi Thuy Nguyen, Hanh Thi Pham & Tam Minh Pham (2015): The effects

of input enhancement and recasts on the development of second language pragmatic competence,Innovation in Language Learning and Teaching, DOI: 10.1080/17501229.2015.1026907

To link to this article: http://dx.doi.org/10.1080/17501229.2015.1026907

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the

“Content”) contained in the publications on our platform However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content Any opinionsand views expressed in this publication are the opinions and views of the authors,

and are not the views of or endorsed by Taylor & Francis The accuracy of the Contentshould not be relied upon and should be independently verified with primary sources

of information Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content

This article may be used for research, teaching, and private study purposes Any

substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden Terms &

and-conditions

The effects of input enhancement and recasts on the development

of second language pragmatic competence

Minh Thi Thuy Nguyena*, Hanh Thi Phamband Tam Minh Phamb

aEnglish Language and Literature, National Institute of Education, Nanyang TechnologicalUniversity, Singapore, Singapore;bFaculty of English Language Teacher Education, University ofLanguages and International Studies, Vietnam National University, Hanoi, Viet Nam

(Received 29 October 2014; accepted 28 February 2015)

This study investigates the combined effects of input enhancement and recasts on agroup of Vietnamese EFL learners’ performance of constructive criticism during peerreview activities Particularly, the study attempts to find out whether the instructionworks for different aspects of pragmatic learning, including the learners’ socio-pragmatic and pragmalinguistic knowledge, as well as their frequency of externallyand internally modifying their criticism Over a course lasting approximately seveninstructional hours, the learners received visually enhanced pragmatic input andrecasts of their errors of form and meaning The learners’ pre-to-posttest improvementwas investigated using three production tasks, namely a discourse completion task, arole play, and an oral peer-feedback task The findings show there is potential for inputenhancement and recast in teaching different aspects of second language pragmaticsand are discussed with implications for classroom practices and future research

Keywords: pragmatic competence; input enhancement; recast; speech acts; secondlanguage acquisition; interlanguage pragmatics

1 Introduction

Pragmatic competence is the knowledge of how to express one’s meanings and intentionsappropriately within a particular social and cultural context of communication Thisknowledge involves both having linguistic resources for expressing meanings andintentions (i.e pragmalinguistics) and understanding the sociocultural constraints on theuse of these means (i.e sociopragmatics) (Canale 1983) Pragmatic competence isessential for effective communication and is featured significantly in various models ofcommunicative competence in second language (L2) teaching (see Bachman 1990;Canale and Swain1980; Canale1983) Nonetheless, previous research has shown thatpragmatic knowledge in the target language (TL) is incomplete for many learners,regardless of their proficiency levels (see Kasper and Rose 2002) Unlike grammaticalerrors, pragmatic failure may be treated as offensive, thus likely adversely affecting thelearner’s communication with the native speaker (NS) (Thomas 1983) Earlier studieshave also shown that pragmatic features are learned slowly without the benefits ofinstruction (see for example Bardovi-Harlig and Hartford 1993; Bouton 1994) This isbecause pragmatic functions and their contextual constraints are not salient enough for L2

*Corresponding author Emails:thithuyminh.nguyen@nie.edu.sg;thuyminhnguyen@gmail.com

http://dx.doi.org/10.1080/17501229.2015.1026907

learners to notice despite prolonged exposure (Kasper and Schmidt1996) In other words,mere exposure is insufficient for L2 pragmatic development, and pragmatic acquisitionrequires a certain degree of consciousness of form-function mappings and pertinentcontextual variables (Schmidt1990,1993).

Recent years have seen an upsurge of interest in the effects of instruction inpromoting pragmatic awareness (see Alcón-Soler 2008; Martínez-Flor, Usó-Juan, andFernández-Guerra2003; Rose and Kasper2001) These studies address three importantquestions: the teachability of L2 pragmatics, the benefits of instruction versus mereexposure, and the relative effects of different teaching approaches Generally, the findings

of these studies have attested to the teachability of pragmatic features and overall benefits

of instruction in developing L2 pragmatic competence (see Jeon and Kaya2006; Kasperand Rose 2002; Rose 2005; Roever 2009; Taguchi 2011; Takahashi 2010, for acomprehensive review) Findings have also suggested that explicit instruction (i.e.instruction that serves to direct learners’ attention to the target forms with the aim ofdiscussing those forms) may work more effectively than implicit instruction (i.e.instruction that allows learners to infer rules without awareness), particularly in teachingsociopragmatics (Jeon and Kaya2006; Takahashi 2010) However, findings in this areamay need to be treated with caution (see Ellis2008; Jeon and Kaya2006) First, becausecurrent research has focused predominantly on relatively‘well defined’ speech acts such

as requests and suggestions (e.g Alcón-Soler 2005; Ghavamnia, Eslami-Rasekha, andVahid-Dastjerdi2014; Eslami, Mirzaei, and Dini2014; Fuykuya and Zhang2002; Koikeand Pearson 2005; Martínez-Flor and Fukuya 2005; Safont 2003; Salazar 2003;Takahashi 2001, 2005; Takimoto 2009), it has remained much less known whetherinstruction works for less clearly defined speech acts or speech act sets such asconstructive criticism that has no predetermined form and thus may cause even moreproblems to L2 learners (see Nguyen 2013; Nguyen, Pham, and Cao 2013; Nguyen,Pham, and Pham 2012) Thus, it is important that further research expand the range oflearning targets under inquiry Second, due to a limited number of studies that haveexamined the impact of implicit instruction on L2 pragmatic development, their findingsmust be corroborated in further research to strengthen the body of evidence (Jeon andKaya2006; Nguyen, Pham, and Pham2012)

Another limitation, as pointed out by Fuykuya and Zhang (2002), lies in theinadequate conceptualization of implicit instruction in many earlier pragmatics studies Incontrast to explicit instruction that is well aligned with a focus on forms approach(involving intentional learning of form via presentation and consolidation of rules inincreasingly communicative practice), implicit instruction seems to be a less developedarea both conceptually and methodologically (Fuykuya and Zhang 2002, 2–3, but seeAlcón-Soler 2005; Ghavamnia, Eslami-Rasekha, and Vahid-Dastjerdi 2014; Eslami,Mirzaei, and Dini 2014; Koike and Pearson 2005; Martínez-Flor and Fukuya 2005;Nguyen, Pham, and Pham2012) In many studies, it is simply defined as mere exposure

to input (Pearson 1998; Tateyama 2001; Takahashi 2001) or the withholding of pragmatic information (e.g House1996) According to Fuykuya and Zhang (2002, 3),this conventional way of conceptualizing implicit instruction ‘leave[s] us with theimpression that the pragmatists have been caught with a fixed notion of simple exposure

meta-to pragmatic examples,’ whereas implicit instruction may, in fact, refer meta-to a wider range

of instructional techniques varying in degrees of implicitness (see Jeon and Kaya2006).1The lack of systematic conceptualization of implicit instruction in earlier studies seems tosuggest that perhaps they are more pedagogically than theoretically motivated (see Ellis2008) Obviously, this calls for more rigorous designs in future interlanguage (IL)

pragmatics research to bring instructional pragmatics research closer to traditions ofmainstream second language acquisition (SLA) research.

The present study attempts to add a variation in the way of conceptualizing implicitinstruction by examining the potential of implementing focus on form instruction inteaching L2 pragmatics In contrast to the intentional learning of form afforded by focus

on forms instruction, in focus on form instruction, learners’ attention to form issubconsciously accomplished while they are primarily engaged in comprehending orproducing meaning (see Ellis2001,2008; Long1991; Long and Norris2009; Long andRobinson1998) A focus on form might be achieved by means of input-based activitieswhere input is manipulated in a way that causes attention to form to take placeincidentally (e.g textual input enhancement), or implicit feedback (e.g recasts) thatbriefly addresses learners’ difficulty in using the target form (see Ellis2008) Because ofits unobtrusiveness, this instructional approach may be a suitable option for teaching form

in meaning-based contexts where the primary focus is maintained on content ofcommunication In the framework of this paradigm, the present study, therefore, examinesthe combined impact of two instructional techniques, i.e provision of textually enhancedinput and recasts, given during a writing class to draw students’ attention to languageform needed for giving pragmatically appropriate constructive criticism on a peer’swritten work Since much focus-on-form literature addresses L2 grammatical develop-ment, more research is needed in the area of L2 pragmatic development (Fuykuya andZhang2002; Koike and Pearson2005; Martínez-Flor and Fukuya2005)

The study reported in this article is part of a larger-scale project that sets out toinvestigate the relative effectiveness of explicit and implicit form-focused instruction onthe acquisition of the speech act of constructive criticism by a group of Vietnamese EFLlearners (see Nguyen, Pham, and Pham2012) While the comparative effects of the abovetwo instructional approaches on the learners’ overall pragmatic performance werediscussed in Nguyen, Pham, and Pham (2012), the present article focuses on examininghow the implicit approach works for different aspects of pragmatic learning that has notyet been explored in the earlier article (i.e Nguyen, Pham, and Pham2012) It is hopedthat by offering an in-depth analysis of the different aspects of the learners’ performance,the present article would be able to provide insights into questions such as what aspects

of pragmatics are amenable or resistant to instruction and therefore would provide usefulinformation to guide classroom practices In particular, one of the questions that thepresent article seeks to answer is whether implicit instruction improves learners’performance in both areas of sociopragmatic appropriateness (conformity to socioculturalcode of language use) and pragmalinguistic accuracy (accurate usage of linguisticresources for expressing the intended meaning) The reason for an examination of thisquestion is that despite that grammar is distinctive from pragmatics (see Bachman1990;Canale 1983; Canale and Swain 1980), grammar ‘serves as a resource for encodingdifferent kinds of meaning,’ including pragmatic meaning (Ellis2008, 195) Learners’limited grammatical competence may restrict their capacity to produce linguistic actionseffectively (see Nguyen2008) This is particularly the case in a foreign language contextwhere learners lack sustained exposure to L2 pragmatic input while their classrooms mayplace a greater emphasis on developing linguistic than pragmatic competence (see Ellis2008) In a previous study, it was found that while explicit instruction generally producedpositive impact on L2 learners’ performance of constructive criticism, this type ofinstruction seemed to benefit sociopragmatic appropriateness to a greater extent thanpragmalinguistic accuracy (Nguyen, Pham, and Cao 2013) Thus, it would be useful toexamine how implicit instruction impacts the two above aspects in the learners’ speech

production Another question of interest to the present study is whether implicitinstruction enables learners to notice both external and internal illocutionary forcemodifying devices (politeness devices that serve to tone down the force of the utterance).

As indicated in previous studies, internal modifiers (politeness devices occurring withinand making up an integral part of the head act) tend to carry less transparent pragmaticmeaning than external modifiers (semantic moves to support the head act) and thereforeare less likely noticeable to learners (Hassal2001; Nguyen2008) Although it would bepedagogically important to examine whether internal modifiers are amenable tointervention, previous studies on the effects of instruction of this type of modifiers arerather scarce and yet produce inconclusive evidence (e.g Fukuya 1998; Fukuya andClark 2001; Martínez-Flor 2008; Nguyen 2013; Safont 2003) Hence, this question isworthy of further investigation

2 The role of enhanced input and recasts in SLA

Input enhancement is a technique for making the target form more prominent eitherphonologically (by means of oral repetition) or typographically (e.g color-coding orbold-facing) with an attempt to induce learners’ noticing in an implicit and unobtrusivemanner (Doughty and Williams1998) This idea is based on the assumption that noticing

is a prerequisite for intake (see Gass1988; Schmidt1990,1993; Sharwood Smith1981)and that without being prompted to attend to target forms in language input learnersgenerally lack sensitivity to them, particularly if these forms lack perceptual salience(Sharwood Smith 1993) A number of empirical studies have been conducted toinvestigate the effects of input enhancement on L2 learning (e.g Doughty 1991; Izumi2002; Leeman et al 1995; White 1998; Wong2003) Their outcomes are inconclusive,however For example, while some studies found significant results (e.g Alcón-Soler2005; Doughty1991; Leeman et al.1995; Martínez-Flor2006; Martínez-Flor and Alcón-Soler 2007; Martínez-Flor and Fukuya 2005; White 1998), others found limited or noeffects (e.g Izumi2002; Fukuya and Clark 2001; Wong2003) Han, Park, and Combs(2008) attribute this lack of congruence to methodological idiosyncrasies of the earlierstudies, for example, different choices of target structures (meaningful vs nonmeaning-ful) and learners (with vs without prior knowledge of the target form), nature and lengths

of treatment, and types of enhancement (used with vs without other attention-gettingdevices) Han, Park, and Combs (2008) also point out that despite the mixed results,previous studies, nonetheless, have shown that although input enhancement may inducethe noticing of form, particularly if the form is meaningful, whether or not it leads toacquisition depends largely on whether or not the learner has prior knowledge of the form(also see Sharwood Smith 1981) Further, input enhancement combined with otherattention-getting devices such as corrective feedback may be most effective in causing thelearner to process the form at a deeper level

Recasts as a type of corrective feedback have been an area of inquiry of great interest

to SLA Recasts are generally defined as the reformulations of erroneous utterances intomore target-like versions while preserving the original meaning (see Long 1996; Lyster2004) From the interactionist perspective, recasts are considered useful because theyoccur during interaction and the process of negotiation of meaning, thus causing learners

to attend to linguistic form while maintaining their primary focus on meaning (Long1996) Saxton (1997) also hypothesizes that recasts from caregivers to children learningtheir first language enable the latter to make cognitive comparison of their own outputwith the adults’ version, thus providing them opportunities to learn the correct form

Despite strong advocating theories, however, there is controversial empirical evidenceregarding the facilitative impact of recasts (e.g Carroll and Swain 1993; Doughty andVarela 1998; Lyster and Ranta 1997; Mackey and Philp 1998; Ortega and Long 1997;Panova and Lyster 2002; Y Sheen2004) Observational classroom studies have shownthat recasts may not always be noticed by the learner, particularly when used withincontent-based, meaning-focused (as opposed to form-focused) instructional contexts (e.g.Lyster1998,2004; Y Sheen2004) Compared to other types of corrective feedback, forexample, elicitations or clarification requests, recasts are less likely to lead to uptake andimmediate repair (Lyster and Ranta1997; Panova and Lyster2002) On the other hand,many experimental laboratory studies have found that learners who receive recastsgenerally improve in postexperimental observations (e.g Carroll and Swain 1993;Doughty and Varela 1998; Mackey and Philp1998; Ortega and Long1997).

As pointed out by Ellis and Sheen (2006) and Sheen (2008), these mixed resultsreflect differences in the design of the above studies, including research designs(classroom observation vs experiments), interaction settings where the recasts occur(form-focused vs meaning-focused), linguistic forms targeted for recasting (e.g morpho-syntax vs phonology), types of recasts (e.g explicit vs implicit)2, learner factors (e.g.developmental readiness, cognitive orientation), and outcome measurements Hence,Sheen (2008) calls for further research that allows for a more systematic control of thesevariables in order to fully understand the efficacy of recasts Importantly, Ammar andSpada (2006) have pointed out that it is problematic to associate the effectiveness ofrecasts with the rate of uptake and repair following this technique (as did observationalstudies) This is because uptake or the absence of it may not always serve as the bestevidence of learning or lack of learning For example, sometimes opportunities for uptake

to occur are impossible In other cases, repairs might simply be a sign of‘mimicking.’Therefore, it has been argued that the evidence of learning should be measured morereliably by means of carefully designed experimental research

Unlike in mainstream SLA research, the role and effects of textually enhanced inputand recasts have been scarcely investigated in L2 pragmatic research (e.g Alcón-Soler2005; Fukuya et al.1998; Fukuya and Clark2001; Fuykuya and Zhang2002; Martínez-Flor and Fukuya2005) Findings of these studies are also relatively mixed For example,whereas Fukuya et al (1988) found no effects for recasts in teaching sociopragmaticaspects of L2 requests, Fuykuya and Zhang (2002) reported relatively large impact ofrecasts on improving learners’ performance of requests in terms of both sociopragmaticappropriateness and pragmalinguistic accuracy Fukuya and Clark (2001) failed to findpositive effects for textual input enhancement in teaching request modifiers On the otherhand, Alcón-Soler (2005) and Martínez-Flor and Fukuya (2005) found significantcombined effects of textual input enhancement and recasts on the acquisition of L2requests and suggestions, respectively See also the positive impact of combining inputenhancement and input flooding in Ghavamnia, Eslami-Rasekha, and Vahid-Dastjerdi’s(2014) investigation of suggestions and input enhancement and recasts in Eslami,Mirzaei, and Dini’s (2014) investigation of requests

Despite their contributions in terms of offering a more developed definition ofimplicit instruction compared to earlier works (for a review, see Jeon and Kaya2006), theabove studies nevertheless have exclusively focused on requests and suggestions, thusleaving unanswered the question of how implicit teaching works for other speech acts.Further, most of the above studies employed only a single outcome measure task, thuslimiting the validity of the data (e.g Alcón-Soler2005; Fukuya et al.1988; Fukuya andClark2001; Fuykuya and Zhang2002) Finally, due to the absence of a delayed posttest

in the reviewed studies, it remains unknown if their findings are stable over an extendedperiod of time.

3 The present study

3.1 Research question

Due to a number of outstanding issues, there is a need for continuing the above line ofinquiry to further our understanding of the applicability of implicit instruction in thepragmatic realm The present study therefore attempts to find out whether implicitinstruction that is based on enhanced input and recasts works for different aspects ofpragmatic learning, including learners’ sociopragmatic and pragmalinguistic knowledgethat is needed for performing constructive criticism during peer review activities, as well

as their frequency of externally and internally modifying their criticism to reduce thepotential offense to their peers The present study is hoped to address the current gaps inthree ways First, it focuses on a less commonly researched speech act, i.e givingconstructive criticism to a peer’s performance, hence contributing to expanding the range

of learning targets for investigation (see Nguyen2013; Nguyen, Pham, and Pham2012;Nguyen, Pham, and Cao2013) Second, the study employs a mix of written and oral dataelicitation methods to measure the learning outcomes with a view to enhancing thevalidity of the data Third, the study aims to examine both immediate and durable effects

of input enhancement and recasts, thus hoping to be better capable of informing L2classroom practices To achieve the specified aims, the current study seeks to answer thefollowing question:

Does the combination of input enhancement and recasts benefit the different aspects of thelearners’ performance of constructive criticism, as outlined above, in both short and longerterms?

The rationales for implementing both of the instructional techniques are as follows First,the provision of corrective feedback in tandem with instruction is generally supported asthey are deemed complementary in pedagogical terms (see Lyster, Saito, and Sato2013).Further, research has shown that employing multiple instructional strategies may producemore positive learning effects than the adoption of a single teaching strategy (seeIzumi 2002)

Regarding the learning target under inquiry, the present study defines constructivecriticism as a negative assessment of an individual’s work in progress with the aim ofimproving current or future performance This speech act usually involves theidentification of a problematic action, choice, or product as well as advice on how tochange or correct the problem (Nguyen 2005) The use of constructive criticism as ameans of improving L2 learners’ writing skills has been generally supported in theliterature for ‘its social, cognitive, affective and methodological benefits’ (Rollinson

2005, 23) However, due to its face-damaging nature, providing constructive criticism to

a peer can become a daunting experience if learners lack training in it Research hasdocumented that students from certain cultures may feel uncomfortable criticizing theirpeers’ work or worse yet face considerable difficulty in conveying their messageappropriately For example, unlike the NS, Vietnamese learners of English tend to softentheir constructive criticism far less frequently while aggravating it far more often, usingmodal verbs of obligation inappropriately (see Nguyen 2005) Given that constructivecriticism may pose a challenge to learners who are not familiar with the way it isperformed in the TL, this speech act deserves due pedagogical attention

3.2 Participants

The present study was conducted at a teacher training institution in Vietnam Two highintermediate EFL intact classes (N = 41) were invited to participate in the study Beforebeing recruited, the learners were explained about the research3and signed consent forms

to indicate their agreement to participate The learners (2 males and 39 females) werepreservice EFL teachers doing their Year 3 English major The two classes were randomlyassigned to control (N = 22) and treatment (N = 19) conditions There were no notabledifferences between the two groups in terms of age range, lengths of English study, andexposure to English outside the classroom In both groups, learners’ ages ranged between

20 and 22 Their lengths of English study ranged between 6 and 9 years None of themhad ever resided in an English-speaking country They had had limited opportunities tohear and speak English outside the formal instructional time

3.3 Choice of target forms

The treatment was integrated into a writing syllabus that focused on paragraph and essaywriting The syllabus specified that students participate in peer-feedback activities for atleast four writing assignments That is, they had to read and give critical comments on atleast four pieces of written work by a peer The writing classes met for three class hoursevery week Instruction on constructive criticism was implemented for the treatmentgroup for one class hour (i.e 45 minutes) every week over a period of 10 weeks, resulting

in an approximate total of seven instructional hours The third author taught the controlgroup and the second author taught the treatment group

The two classes followed the same writing syllabus and schedule, that is, they bothengaged in peer-feedback activities for at least four writing assignments (see above) Theonly difference was the additional exposure to the language for expressing constructivecriticism via input enhancement and recast activities by the treatment group It is notedthat the control group may have received some exposure to these linguistic itemsincidentally through teachers’ feedback in their writing class and other courses However,

it remains uncertain if the items had been explicitly taught to them in previous courses.Given that the students had reached a high intermediate level, it is plausible to assume apositive answer Nonetheless, since English instruction in Vietnam is heavily grammar-based, it could be the case that although the students may have learned these grammaticalforms, they may not have learned all corresponding pragmalinguistic functions As aresult, they did not always put the forms to the correct pragmatic use, as can be seen intheir pretest performance

The target features selected for instruction included two major criticism realizationstrategies: (1) identification of problem and (2) giving advice The pragmalinguisticconventions for realizing identification of problem included three structures: (1) NP wasADJ; (2) You V (past tense); and (3) You had (a/an) (ADJ) NP The pragmalinguisticconventions for realizing advice included 12 structures: (1) You can + V; (2) You could +V; (3) You could have + V (past participle); (4) You may + V; (5) You might + V; (6) Youmight want to + V; (7) (If I were you) I would; (8) It would be better if you + V; (9) Itwould be better + V (infinitive); (10) If you + V, it may; (11) NP may be + V (pastparticiple); and (12) Why don’t you?

The target features also comprised two types of criticism modifiers: (1) externalmodifiers, including compliments (e.g It was an interesting paper), disarmers (e.g Youhad a few spelling mistakes here and there but I think that’s because you’re writing prettyquick, nothing too major), and grounders (e.g I think‘it’ is better than ‘are’ there because

‘traffic’ is single) and (2) internal modifiers, including past tense (e.g I thought it wouldmake more sense that way), modal verbs (e.g may, might, could, would), modal adverbs(e.g maybe, perhaps, probably), expression of uncertainty (e.g I wasn’t sure that was thebest phrase you could’ve used), hedges (e.g sort of, kind of, seem), and understaters (e.g.

a bit, a little bit, quite, rather) (see details in Nguyen, Pham, and Pham2012) The abovestrategies, modifiers, and pragmalinguistic conventions were selected for teachingbecause they were found to occur most frequently in NS criticism in equal powersituations but to present considerable difficulty to many learners of English (seeNguyen2005)

3.4 Instructional procedures

Over the 10 weeks, the treatment group was engaged in the following activities:(1) Exposure to visually enhance input in the first three sessions: In the first session,learners read samples of NS peer-feedback conversations, answered compre-hension questions, and evaluated the effectiveness of the NS criticism using alist of criteria such as whether the criticism is specific, well-grounded, soundspositive, includes suggestion for improvement, and does not provoke negativeresponses from the hearer (see Tracy, van Dusen, and Robinson 1987) Theymight add more criteria to the list if they wished to In the second session, thelearners read and wrote feedback on a sample essay Then they read the sample

NS feedback on this essay and compared the feedback with their own withreference to the above criteria.4In the third session, the learners read differentsample essays, identified the problems with them, and matched them withcorresponding NS feedback samples Note that all the NS feedback samplescontained bold-faced target structures, and the learners were instructed to payattention to these highlighted parts when reading the samples in order tocomplete the tasks (seeAppendixfor samples of the materials)

(2) Communicative practice comprising performance on a discourse completiontask (DCT) and oral peer-feedback tasks (OPFs) in the remaining sevensessions

(3) Performance evaluation: For this activity, learners recorded their peer-feedbackconversation, listened to the recording, and evaluated their own performance,using a set of guiding questions given to them (seeAppendix)

(4) Receiving recasts of both pragmatic and grammatical errors that arose out ofcommunicative tasks (see below) The recasts were also written on theblackboard at the end of the lesson for students to note down

Recasts were provided in the form of confirmation checks as these were assumed topresent a clearer corrective intention than the reformulation of errors alone (see Gass andAlison 2007) Recasts were done as follows First, the teacher, also the second author,repeated the erroneous part of the utterance in a rising tone to attract students’ attention.Then, the teacher said the appropriate utterance, preceded by‘You mean’, also using arising tone The corrected part was also stressed, as seen in the example below:

Student: You must pay attention to grammar

Teacher: Must?↑ You mean ‘Perhaps you could pay more attention to grammar?’↑

In order to decide what and how to recast, the framework proposed by Fuykuya andZhang (2002) was adopted In particular, if an utterance is pragmatically appropriate butgrammatically inaccurate, the teacher recasts only the linguistic form (type 1) If anutterance is pragmatically inappropriate but grammatically accurate, the teacher recasts itsillocutionary force by using one of the pragmalinguistic conventions for expressingconstructive criticism (type 2) Finally, if an utterance is neither pragmatically appropriatenor grammatically accurate, the teacher recasts both of its form and illocutionary force

by using one of the pragmalinguistic conventions for expressing constructive criticism(type 3) This procedure together with illustrative examples is presented inFigure 1.5

3.5 Assessment tools

Previous research has shown that the type of outcome measure may affect the observedmagnitude of instructional effects (Jeon and Kaya2006) Studies employing elicited dataonly tend to produce smaller effect sizes than those employing both elicited and naturaldata Thus, to maximize the possibility to track postexperimental changes, both elicitedand naturalistic data were employed in this study Specifically, the current study made use

of an OPF on actual written works, an 8-item DCT, and a 6-item role play (RP) Details

of these instruments were reported in Nguyen, Pham, and Pham (2012) Also seeNguyen, Pham, and Pham (2012) for samples of these instruments

The treatment and control groups’ performance was compared on a pretest and animmediate posttest, using the three production tasks mentioned above The pretest wasconducted at the onset of the study while the immediate posttest at the end of thetreatment period To examine whether the treatment group’s gains (if any) were durable

Type 2Inappropriate but accurate

(e.g You must

pay attention to grammar)

Type 3Inappropriate and inaccurate

(e.g Your

introduction are

too long)

Learner made an error

Type 1Appropriate but inaccurate

(e.g If I were

you, I will

revise it)

Recasting linguistic form

(e.g If I were

you I would

revise it)

Recasting illocutionary force

(e.g Your

introduction

was probably a bit long)

Figure 1 Types of errors and corresponding recasts

beyond the treatment, a delayed posttest consisting of the above three production taskswas also conducted for this group five weeks after Due to limited resources, however, thedelayed posttest was not conducted for the control group To guard against thepossibilities of the learners’ memorizing responses from the pretest, some adjustmentswere made as follows For the OPF task, the learners were asked to critique a differentessay in each test Similarly, the order of DCT and RP scenarios was also reshuffled eachtime Finally, to keep the variable of social distance constant, the learners were allowed tochoose their own pairs for the RP and OPF, and this pairing was maintained throughoutall the three tests.

3.6 Data analysis

The data were coded into different categories of criticizing strategies and modifiers, using

a coding scheme devised and validated by Nguyen (2005) (also see Section 3.3) The datawere first coded independently by each researcher in the team and then were carefullycross-checked Cases of discrepancies were discussed until an agreement was reached Inorder to analyze the frequency with which the learners in each group externally andinternally modified their constructive criticism, means and standard deviations of eachcategory of modifiers were computed This was done by first calculating the number ofexternal and internal modifiers produced per criticism in each production task byindividual learners in each group and then averaging the outcomes of the three tasks foreach learner Next, the average outcomes for individual learners in each group wereentered into an SPSS spreadsheet to compute the means and standard deviations for thegroups

In order to assess the sociopragmatic and pragmalinguistic aspects of the learners’performance of criticism, two rating scales, which were adapted from Martínez-Flor andFukuya (2005), were employed Each scale consisted of five points, with 0 representingthe lowest and five the highest possible score Sociopragmatic appropriateness (hence-forth‘appropriacy’) was assessed in terms of knowledge of what to say to a particularinterlocutor in a particular context of situation In this study, appropriacy was determined

by the choice of realization strategies and politeness devices from the list that was taught

to the students (see Section 3.3) These items were derived from an NS database ofconstructive criticism between peers in the institutional context that was collected byNguyen (2005) Pragmalinguistic accuracy (henceforth‘accuracy’) was assessed in terms

of knowledge of various expressions for conveying intentions and determined by thecorrect usage of relevant linguistic structures, also from the aforementioned list Note thataccuracy was to be scored only when appropriacy had been achieved This is because theultimate goal of pragmatic instruction is to enable students to find socially appropriatelanguage for the situations that they encounter (Bardovi-Harlig and Mahan-Taylor2003);thus, only when learners can connect newly learned grammatical forms to thecorresponding pragmatic functions should the instruction be considered successful.Also note that scores were given only when learners were able to use one of the targetforms which had been taught to them in the treatment (see Section 3.3)

The scoring procedure for appropriacy is as follows If the student was able to employthe realization strategy and at the same time modify its illocutionary force appropriatelyaccording to the context, he or she would be awarded the full mark of 5 He or she,however, would be awarded only 2.5 points if his or her utterance was lacking inmodality or was accompanied by an inappropriate modifier (e.g making use of anintensifier instead of a downgrader) This is because modality constitutes an essential