Oct 13, 2015
International Journal of
Educational Research 37 (2002) 285304
Chapter 4
Talking it through: two French immersionlearners response to reformulation
Merrill Swain*, Sharon Lapkin
Ontario Institute for Studies in Education, University of Toronto, 252 Bloor St. West Toronto,
Ont., Canada M5S IV6
Abstract
This article documents the importance of collaborative dialogue as part of the process of
second language learning. The stimulus for the dialogue we discuss in this article was a
reformulation of a story written collaboratively in French by Nina and Dara, two adolescent
French immersion students. A sociocultural theoretical perspective informs the discussion of
three sustained examples of the learners talking it through as they confront and resolve
many language-related problems while comparing their story with its reformulation. A pre-
test/post-test design allows us to demonstrate the combined effects on second language
learning of the reformulation and the students reactions to it.
r 2003 Elsevier Ltd. All rights reserved.
Keywords: Second language learning; Sociocultural theory; Output; Immersion; Reformulation
1. Literature review
1.1. Talking it through
Our program of research has focused on the roles of output (speaking and writing)in second language learning (e.g. Swain, 1995; Swain & Lapkin, 1995). As ourresearch has evolved, so has our theoretical orientation (Swain, 2000). We havemoved from viewing output solely within an information-processing framework oflearning to viewing output within a sociocultural theoretical perspective of learning.An essential difference between these two perspectives is that the latter regardsoutput not only as a message to be conveyed, but as a tool in cognitive activity.
ARTICLE IN PRESS
*Corresponding author.
E-mail addresses: [email protected] (M. Swain), [email protected] (S. Lapkin).
0883-0355/02/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0883-0355(03)00006-5
In this sense, speaking is a cognitive activity (Lantolf & Appel, 1994), the outcomeof which is an utterance. Through speaking, thought is externalized. Externalized asan utterance, it becomes an object. As an object it can be scrutinized, questioned,reected upon, disagreed with, changed, or disregarded. In order to collaborate,learners must speak to each other. Through their dialogue, they engage in makingmeaning, and debate the meaning made. To make their meaning as clear, coherentand precise as possible, learners will debate language form (morphosyntax throughto discourse and pragmatics) and lexical choice (Kinginger, 2000; Kowal & Swain,1997; Swain & Lapkin, 2001). This talk about language (metatalk) mediates secondlanguage learning (Lantolf, 2000; Swain, 2000). Talk supports the process ofinternalization (Lantolf, 2001)the moving inwards of joint (intermental) activityto psychological (intramental) activity (Galperin, 1967; Stetsenko & Arievitch,1997). A detailed (micro-genetic) analysis of the way new linguistic forms andmeanings evolve leads to an understanding of how language learning occurs indialogue, not as a result of it (Swain, 2000).Metatalk is language used to reect consciously on language use. It is one sort of
collaborative dialoguedialogue in which speakers are engaged in problem-solvingand knowledge-building (Swain, 2000). Although collaborative dialogue occurs inmany domains (e.g. mathematics), for our purposes collaborative dialogue relatesonly to problem-solving and knowledge-building in the linguistic domain.Vygotsky (1986) argued that what was needed for learning to occur was the
presence of a more knowledgeable person who would help the learner to move frombeing able do something only with the help of that expert to being able to do itindependently. Because Vygotskys ideas have mostly been applied within develop-mental psychology, the more knowledgeable otherthe expert in an expert/novicepairhas typically been conceived of as an adult (e.g. parent, teacher). However inrecent years, the idea that peerpeer interaction may also foster learning has beenadvanced (e.g. Tudge, 1990; Wells, 2000). This idea has been extended withinsociocultural SLA by suggesting that in peerpeer interaction, peers can beconcurrently experts and novices (Brooks & Swain, 2001; Donato, 1994).A number of recent studies demonstrate the impact of peerpeer dialogue on
second language learning (Swain, Brooks, & Tocalli-Beller, 2002). For example,DiCamilla and Antons (1997) analyses of the discourse of ve dyads of Spanishsecond language learners collaborating on a writing assignment emphasized theimportance of co-constructed scaffolded support and guidance through peerdialogue.Storch (1999a, b, 2000, 2001, this issue) examined the nature of peer assistance in
pairs working together on language tasks. She identied four types of relationshipsamongst pairs: collaborative; expert-novice; dominant-passive; and dominantdominant. There were many differences in the nature of the discourse amongst thefour types of pairs, for example, who initiates questions, frequency of disagreementsand resolutions, use of pronoun we, etc. Adopting a collaborative orientationresulted in co-construction, extension of knowledge, provision of scaffoldedassistance and language development (grammatical accuracy and new lexicalknowledge).
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304286
Other research (e.g. De Guerrero & Villamil, 2000; Ohta, 2001; SpielmanDavidson, 2000; Tang & Tithecott, 1999; Tocalli-Beller, in press; Villamil & deGuerrero, 1998) demonstrates the value of peerpeer dialogue on aspects of secondlanguage learning. The opportunity to talk and discuss language and writing issueswith each other allowed the learners to consolidate and reorganize knowledge ofthe L2 in structural and rhetorical aspects and to make this knowledge explicit foreach others benet (De Guerrero & Villamil, 2000, p. 65).In our research (e.g. Kowal & Swain, 1997; Swain & Lapkin, 1998), students have
worked collaboratively on a number of tasks designed to increase collaborative talk(e.g. dictogloss and jigsaw tasks). In the research reported in this article, studentsrst wrote a story collaboratively and then subsequently they compared what theywrote with a reformulated version of it. Their collaborative efforts, mediated by theirdialogue, reveal what cognitive steps they took to be able later to use theirconstructed knowledge individually.
1.2. Reformulation
Cohen (1983) dened reformulation as a technique consisting of having a nativewriter of the target language rewrite the learners essay, preserving all the learnersideas, making it sound as nativelike as possible (p. 4). Reformulation, then,encompasses both reconstruction (Levenston, 1978) or the correction of surfaceerrors, and more substantive correction addressing the style and clarity of thought(Cohen, 1982, p. 5). Thornbury (1997) appears to agree, stating that reformulationgoes beyond the surface features of the text.Because they involve changes at all levels of a text (Allwright, Woodley, &
Allwright, 1988), reformulations usually differ in numerous ways from the learnersoriginal texts. Learners must be encouraged to do a thorough comparison betweenthe reformulated version and their own original texts. Cohen (1983) found thatmost students need assistance in comparing their version with the reformulatedoney these comparisons need to be purposely eye-opening and engaging (p. 19).One way of providing such assistance is having learners talk to each other as theywork their way through the texts, identifying or noticing changes and discussing whythese changes might have been made. Brooks and Swain (2001) conducted a pilotstudy involving four adult ESL learners (intermediate prociency) who wrote a storycollaboratively based on a picture prompt. The students then talked through areformulated version of their text, noticing and commenting on many of the changesthe reformulator had made. Next they participated in a stimulated recall session witha researcher in which they viewed a videotape of their noticing activity andcommented further on differences between their original text and the reformulatedtext. Finally, each learner individually re-wrote the original story (the re-write wasreferred to as the post-test) and was interviewed.Brooks and Swain analyzed the post-test responses (i.e. items or structures that the
learners changed as a result of this multi-stage task) in terms of four sources offeedback: peer, reformulation, researcher and self. They demonstrated that the roleof expert is often shared between the learners; where the peer cannot provide
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 287
sufcient expertise, assistance comes from the reformulation together with thedialogue generated by discussing the feedback. Improvements made on the post-teststended to be those linguistic items that had been discussed in both the noticing andstimulated recall tasks.Allwright et al. (1988) found that the whole-class discussion of reformulated texts
was a key component of the technique. They called for studies of whaty learnersthemselves can tell us about the changes they make in their writing (p. 254) as aresult of the opportunity to study reformulations of their texts.Sanaoui (1984) directed her adult students of French as a second language to
notice six different categories of changes: vocabulary selection, syntacticchoices, cohesive devices, discourse functions, overall organization, and stancetoward the reader. In this case, although peer discussion had preceded the writing oftheir rst drafts, the reformulations were dealt with on an individual basis, withinstructions to identify signicant changes, consider how they were made and theeffects of the changes. Piper (1995) guided classroom discussions of reformulatedtexts, and Thornbury (1997) used awareness-raising exercises to enhance hisstudents ability to notice changes between their texts and the reformulated versionsof them.Mantello (1996), working with young adolescent learners of French as a second
language near the beginning of their intensive exposure to the target language,reformulated students texts paying special attention to the compound past tense.Students worked through the reformulated texts on their own, comparing thecorrected past tense verb forms to their original verbs. The stronger students noticedchanges that went beyond the target structure, a result that the researcher had notanticipated.Prociency differences were the focus of a case study by Qi and Lapkin (2001) in
which two adult Mandarin learners of English wrote a text in response to a pictureprompt. Each learner compared his/her text to a reformulated version of it, noticingchanges and sometimes commenting on those changes. One researcher theninterviewed the students, showing each a videotape of the noticing session andconducting a stimulated recall session in which the students were encouraged to sayas much as possible about the changes they had noticed and why they had beenmade. The researchers found that the quality of noticing related directly to theimprovements made when the students were given an opportunity to revise theiroriginal texts one week after the stimulated recall stage.This brief review of studies involving reformulation has identied several ways of
implementing reformulation as a teaching/learning technique to lead to improvedwriting. The effectiveness of the technique depends on learner variables such as theprociency of the participants, and features of its implementation such as theinstructions participants receive, and the opportunity for peer discussion (or lack ofsuch opportunity).In the study presented in the current article, our participants are grade 7 (age 12
years) immersion students in an early French immersion program. They are averageand above average students in their class, and as we will see, they have severalopportunities for talking it through during a multi-stage task.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304288
1.3. French immersion in Canada
The rst French immersion program in Canada was established in 1965 in anEnglish-speaking suburb of Montreal in Canadas only ofcially French-speakingprovince, Quebec (Lambert & Tucker, 1972). The anglophone parents whopressured the school board to set up such an intensive French program wouldprobably not have imagined, that some 35 years later, fully 8% of the Canadianschool population would be enrolled in similar programs in every province andterritory.Immersion programs in Canada have provided a useful context for examining
language learning in progress. Students exposure to French tends to be conned tothe classroom, and target-language models include the teacher (who may or may notbe francophone, but must have near-native prociency in French) and print andaudio materials. As one would expect, the output students hear from their peers ismarked by inaccuracies, and since it also serves as input (Lightbown, 1992; Swain,1985) in the immersion classroom, this contributes to limiting the native-speakingFrench models to which the learners have access. The limited access to native-likemodels of the target language is one explanation for the commonly reported ndingthat after some years of schooling, immersion students communicate uently but notaccurately in French (e.g. Genesee, 1987; Harley, 1992; Lyster, 1995; Swain &Lapkin, 1982).Another factor to bear in mind is the fact that immersion students do not
have many opportunities to produce language, at least according to the fewclassroom observation studies available. For example, Allen, Swain, Harley,and Cummins (1990) found that the overwhelming majority (over 80%)of student utterances in the grade 3 and 6 classes participating in the studywere one clause or shorter in length. Mantello (1996) reported that her grade 8extended French students, who may be considered quite representative,were taken aback at the amount of writing they were asked to do, and that sheherself, as an experienced teacher, had not previously required as muchwriting from her students (in her study the students wrote one compositionper week for 7 weeks). Although it is difcult to generalize from the smallnumber of relevant studies, our experience also suggests that there are fewopportunities for extended output in the upper elementary grades of immersion.The importance of output for second language learning has been rmly establishedboth in theoretical discussions (e.g. de Bot, 1996; Ellis, 1994; Swain, 1985, 2000) andin empirical investigations (e.g. DeKeyser & Sokalski, 1996; Izumi, 2000; Swain &Lapkin, 1995).The two participants in the study presented here come from a typical early French
immersion program. They began their schooling at the Kindergarten level (age 5)entirely in French. Instruction in grades 1 and 2 was also in the medium of the targetlanguage. English language arts was introduced for about an hour a day in grade 3,and thereafter the proportion of the school day in English increased until grade 5,when the instructional day was divided evenly between subjects taught in Englishand French.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 289
2. Methodology
2.1. Research context
The study took place in a middle-class neighborhood in Toronto, a large multi-ethnic city in southern Ontario. The principal of the school gave us access to a smallthough cluttered room where we could set up our equipment and be relatively free ofinterruptions. One research assistant operated a video camera; another two tape-recorders; while a third gave instructions to the students in the task sessions andconducted the interviews. In spite of the presence of the three research assistants andthe equipment, the two student participants, Dara and Nina, appeared quite at easethough highly attentive. When asked during the nal interviews, If we hadnt been inthe room when you did the activity, would you have taken it as seriously?, Dararesponded with I think we would have. I think I would have because it would havebeen as an assignment and I always finish my homework. So I would have taken it
seriously. And independently, Nina replied, Yes, because Im just like that. I dontwant to get in trouble or anything. Both Dara and Nina stated, however, that theyprobably would not have continued speaking in French if the researchers had left theroom.Dara and Nina were in a grade seven French immersion class, having been in a
French immersion program since kindergarten. They had been told that we wereinterested in how students learn French, and that we wanted their opinions about theusefulness of various activities as an aid to learning French. In the nal interviews,they did not hesitate to tell us what they liked and did not like about the activities inwhich they had participated, so our impression is that they were not intimidated bythe situation.
2.2. Research questions
This study investigates two main questions:
1. What do students notice while comparing a story they wrote in their L2 to anative-speaker reformulation of it? Do they accept or reject the changes made totheir story? For what reasons?
2. Do the participants make revisions to their original stories (an indication thatlearning occurred) based on the reformulation and their collaborative dialogues?
2.3. Design
The data presented in this paper come from one pair of learners, Nina and Dara(pseudonyms), and constitute a subset of data collected from 12 grade 7 Frenchimmersion students (aged 12) during the spring of 2000. The participants were drawnfrom a single class in a middle school in the greater Toronto area where moststudents come from middle-income families. From among the students who returnedsigned permission forms, the teacher constituted four pairs, two composed of
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304290
average students, and two composed of a relatively strong and a relatively weakerstudent. The teacher assigned a rating to each student in the class on a 7-point scale,where 7 indicates a high level of prociency in French. Nina got a rating of 6 onthe seven-point scale, while Dara got an average rating of 4.The data collection unfolded over a period of two school weeks in the following
stages/sessions (the language of each session and the approximate length of time ittook Dara and Nina to do each session appear in parentheses for each stage):
Stage 1: Writing (in French; 30min). At the beginning of the session we showedthe students a 5-min videotaped lesson focusing on pronominal verbs in French (lesverbes r!efl!echis); the video included a short segment where two students modelled ajigsaw task (for details, see Lapkin & Swain, 2000; Lapkin, Swain, & Smith, 2002).Dara and Nina then did a similar jigsaw task (described below) collaboratively, rstorally and then in writing. A native speaker reformulated their text, in preparationfor stage 2. The reformulator had been asked to revise the students text to reecttarget-language usage while preserving the students original meaning.1
Stage 2: Noticing (in French; 10min). The students were asked to noticedifferences between their stage 1 text (which we had typewritten prior to the session)and the reformulated version of it (also typewritten). Dara and Nina used ahighlighter to mark the differences that they noticed, usually verbalizing the item orstructure from their own text and from the reformulated text, and occasionallycommenting on it.
Stage 3: Stimulated recall (in English; 40min). We showed the videotape of stage 2to the learners, stopping the tape at each feature they had noticed, and asking themto comment on the differences. To prepare for this stage, the research team watchedthe video of stage 2, locating where Dara and Nina had verbalized differencesbetween the two texts. In stage 3, then, one research assistant operated the video,stopping at each relevant location. A second researcher operated tape recorders anda third identied the changes and asked the participants to comment on what theywere thinking as they compared the two texts.
Stage 4: Post-test (15min). We gave each learner a typewritten copy of theiroriginal story and asked her to write the story again, making any changes shewanted. The two students worked independently.
Stage 5: Interview (in English; 15min for Nina and 20min for Dara). One researchassistant interviewed Dara and Nina individually to elicit their perceptions of allstages in the task.These sessions took place over two school weeks as shown in Table 1.
2.4. Task
The jigsaw task was an information-gap activity that we have used in previousresearch (e.g. Swain & Lapkin, 1998; Swain & Lapkin, 2001), based on a series ofeight pictures (Appendix A). One student held pictures numbered 1, 3, 5 and 7, and
ARTICLE IN PRESS
1The reformulation thus included both reconstruction and reformulation as Cohen (1983) discusses
there.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 291
the other pictures 2, 4, 6 and 8. They took turns to narrate the story told by thepictures. Then they laid all the pictures out in sequence on the table, and wrote outthe story they had told. In this case, Dara began the writing, and Nina took overafter a short while.
2.5. Data analysis
We transcribed the talk of the learners as they worked together to write the story(stage 1), to notice differences between their written narrative and the reformulatedversion of it (stage 2), and to reect on those differences in the stimulated recallsession (stage 3). We also transcribed their individual interviews (stage 5).We then coded the transcribed data from stages 13 for language-related episodes
(LREs), originally dened as any part of the dialogue where learners talk about thelanguage they are producing, question their language use, or correct themselves orothers (Swain & Lapkin, 1995). The LRE has been used to analyze the dialogue thatoccurs during collaborative writing (here, stage 1, see Swain & Lapkin, 1998). Foranalyses of stages 2 and 3, we extended the denition to include any part of thedialogue where learners talk about the language they produced, and reect on theirlanguage use; this usually includes an indication of whether they accepted or rejectedthe way in which their original text had been reformulated. Depending on thelinguistic focus of the LRE, we coded it in one of three main categories:
* Lexicalincludes adverbs, nouns, adjectives and verbs.* Formincludes article gender, possessive pronoun/article, preposition, preposi-
tion+article, pronoun reference, sentence structure, spelling, pronominal verb,verb form.
* Discourseincludes discourse marker, logical sequencing, stylistics, tensesequencing, temporal sequencing, text structure.
We also coded the LREs occurring in stage 3 (stimulated recall) in terms ofwhether Nina, Dara, or both students accepted or rejected the way in which the itemor structure in their original text had been reformulated. These coding decisions werebased on information found in stages 2 and 3.
ARTICLE IN PRESS
Table 1
Data collection timetable for Dara and Nina
Week 1 Week 2
Tuesday Thursday Monday Thursday Friday
Stage 1 Stage 2 Stage 3 Stage 4 Stage 5
Writing (pre-test) Noticing Stimulated recall Post-tests Interviews
30min, in French 10min, in French 40min, in English 15min for each
student
1520min for
each student
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304292
Four members of the research team worked together to develop and code aportion of one stage 3 stimulated recall transcript. Then we worked individually on afurther portion and met again to ensure that we agreed both on the boundaries ofindividual LREs and on their categorization. Two research assistants then coded theremainder of the transcripts (stages 13) and identied any coding difculties; thelatter were discussed and resolved at a subsequent project meeting.
3. Findings
3.1. Overview of LREs
In the writing session, Nina and Dara produced 47 LREs.2 In the noticing session,they produced 21 LREs. The reformulator, a native speaker of French, made a totalof 29 changes to their text; of these, the students failed to comment on eight duringthe noticing stage; most of these unnoticed items were accents or punctuationmarks such as commas. Finally, in the stimulated recall session, Dara and Ninaproduced 23 LREs. In general, in each of the three sessions of writing, noticing, andstimulated recall, Nina and Dara paid more attention to form than to lexis ordiscourse.3
3.2. Qualitative analyses
We turn now to three examples in which we look across all sessions, taking note ofthe original form the students writing took, its reformulation, the relevant LREsfrom all stages, and the post-test results. The data presented in these examplesaddress our research questions, that is what the students notice as they compare theirtext to a reformulation, whether they accept the changes or not and why, and whatthey learn from the reformulations and the collaborative dialogues as evidenced bythe post-test results.Based on the noticing and stimulated recall protocols, we categorized the LREs
into two broad types: those where the learners (one or both) accepted thereformulation and those where the learners (one or both) rejected the reformulation.First we look at two examples where Dara and Nina do not accept thereformulation. We were particularly interested in why they would reject such expertfeedback, and our rst two examples suggest two main reasons for their decision to
ARTICLE IN PRESS
2 In a previous study using the same stimulus material, but with grade 8 French immersion students, we
found an average of 8.8 LREs per pair of students. (The range was from 1 to 26 LREs in the 12 dyads in
that class: Swain & Lapkin, 1998). Clearly the context in which the data were collected had an inuence:
pairs were withdrawn from the classroom and worked together in a quiet room with 3 researchers present
in the current study, whereas pairs worked together in a classroom with all the attendant noise and
distractions in the earlier study.3When averaged across the three sessions, 52% of the LREs were form-based; 28% lexis-based; and
20% discourse-based. We feel that the LREs discussed in this article are representative of the entire set of
LREs that Nina and Dara produced.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 293
reject the authority of the reformulated text.4 These are followed by one furtherexample representing a case where Nina and Dara accepted the reformulation.Overall, Nina and/or Dara accepted 65% of the reformulations and rejected 35% ofthem.The examples that follow, then, provide excerpts from the talk that took
place during the writing, noticing and stimulated recall stages: these are sets of LREsin tabular form, setting out stages 13 in sequence. The heading of each tableincludes the item or structure as written by the dyad, its reformulation, and itsrealization in each students post-test. Italicized headings represent our codingdecisions. To put these LREs in context, refer to Appendix B which includesthe original story written by Nina and Dara, the reformulation, and each studentspost-test.Table 2 presents an example of a set of LREs relating to the prepositional phrase
de la maison (one LRE in the writing stage, a corresponding LRE in the noticingstage and another in the stimulated recall stage). Here the learners reject thereformulated phrase initially, in part because of an incorrect rule one of them (Nina)has internalized. In turn 464 of the writing stage, Nina questions whether the correctphrase is de la maison or du maison, and Dara (turn 465) is equally hesitant. In turn466, Nina articulates the incorrect rule that de+la equals du, and Dara (turn 467)accepts this.In the stage 2 noticing session, in turn 109, Nina explicitly notices and
verbalizes the difference between their text and the reformulated one, and Daralaughingly says I told you so! In the stimulated recall, however, Dara (turn 214)states that either alternative would have been acceptable. Nina reminds her (turns215 and 217) that they had hesitated during the writing stage and had agreed on du.Nina reiterates her rule in turn 219 (is not it de la equal du?) and acknowledges Idont know; and later (turn 226) Dara asserts that De la does make more sense.Nina (turn 227) still holds, albeit with some uncertainty, to her rule as the exchangecomes to an end.This set of LREs illustrates the power of an internalized rule to prevail, as
Nina sticks with du maison in her post-test; Dara, on the other hand, holds to herassertion that De la does make more sense and uses the correct de la maison in herpost-test.Another reason for rejecting a reformulated item or structure is illustrated
in Table 3. Here the learners reject the adverbial phrase as it was reformulated(dans le silence) and use the acceptable and grammatically correct alternatives(en silenceNina, and silencieusementDara) in order to preserve theiroriginal meaning. In their story, the students had written Il est maintenant6:01 et elle sendore sans bruit, using the phrase sans bruit (without a sound[literally, without noise]), and the reformulator decided to change the phrase todans le silence (in the silence), slightly altering the meaning in so doing. Ninaand Dara put the emphasis on how Marie-France falls asleep, i.e. without a
ARTICLE IN PRESS
4The students mistakenly thought that the research assistants had reformulated their original story;
their challenging some of the changes to their story therefore took a lot of courage!
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304294
sound, and the reformulation highlights the state of the room, which is now silent. Inthe brief noticing segment (turns 5962), Nina and Dara notice the difference, andare able to articulate it in turn 109 of the stimulated recall: I think sans bruit is more,she, she fell asleep and she didnt make any noise. But silence is like everything around
her is silent. Here the determination to preserve their own meaning is a powerfulforce, although the students do so using the lexical item (silence) that thereformulator proposed.
ARTICLE IN PRESS
Table 2
LREs relating to de la maison
Original story (pre-test): Elle sort du maison
Reformulation: de la maison
Post-test Nina: du maison
Post-test Dara: de la maison
Stage 1 writing Stage 2 noticing Stage 3 stimulated recall
Form LRE
(preposition+article)
Nina rejects (and Dara accepts)
reformulation
464. N. OK, elleyvingtminutes plus tard elle sort
du maison. Elle sort de la
maison? Ou du maison?
107. N. Elle sort du maison 213. R. Ok so de la maison and du
maison
465. D. De, non, du
108. D. en courant.
214. D. uh, it could have been any
way.
466. N. Non, parce que de
la cest du
109. N. du mai?ydeyIlsont mis. Quest-ce qu on a
mis? On a mis du maison,
ils ont fait de la maison.
215. N. it has to do with du, de la. I
remember when we were rst writing
it.
467. D. Oui
110. D. Je tai dit! Ou tu as
dity[laughs] 216 D. Yeah.468. N. sort du
maisonysortyen courant217. N. and we went like is it either
du or de la? And we agreed that it
was du but I dont know [laughs]469. D. sort du maisony218. D. I dont think I can remember
it exactly, but
219. N. Yeah, sometimes its like de
la? Isnt it de la equal du? I dont
know.
220. D. Nooo! Shoot, uh, ok
221. R. So you know theres a rule in
there, but
222. D. um-hum, like I cannot
remember things that I learned in
grade ve. They are a complete blur
[laughter], so
223. R. Many things are blurry to me.
226. D. de la does make more sense.
Well, I dont know. De la
227. N. equalsy228. D. Really?
229. N. Yeah.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 295
Table 4 presents an LRE where the learners accepted the reformulated item andappeared to learn it during stages 2 and 3. In stage 1 Nina hesitates between thepronominal verb se souvient and the (incorrect) souvient (turns 335 and 347), askingin turn 351 which form is correct. Dara is equally uncertain, repeating bothalternatives twice in turn 352, and settling on the incorrect form in turn 354. Ninaappears to accept this (turn 357), and Daras private speech5 in turn 358 (where sheagain produces the form se souvient, but then rejects it) reveals their decision aboutthe correctness of souvient. In turn 361 Dara makes explicit her rejection of sesouvient and the two learners agree on souvient (turns 362 and 363) and use it in theirstory (turn 366).In the noticing stage, Dara reads from their own text, elle souvient (turn 79), Nina
queries the reformulated se souvient? (turn 80), and then the students play a game ofI told you so! Ninas je tai dit in turn 82 probably reects the fact that she wasthe one to question which form was correct in stage 1 (turn 335). Dara highlights thecorrect se souvient and they move on.
ARTICLE IN PRESS
Table 3
LREs relating to dans le silence
Original story
(pre-test):
sans bruit
Reformulation: dans le silence
Post-test Nina: en silence
Post-test Dara: silencieusement
Stage 1 writing Stage 2 noticing Stage 3 stimulated recall
N=A Lexical LRE(adverb)
Reject reformulation
59. D. Okay. Se
rendort en silence.
101. R.ythe part that we wanted to ask youabout was sans bruit [shows relevant part of stage
2 video]yDid you get that part?60. N. Quest-ce quon a mis? 102. N. Yes
61. D. sans bruit 103. R. When
62. N. okay. 104. N. sans bruit. Yeah
105. R. Ok
106. D. What did you change to?
107. R. Dans le silence.
108. D. Ah, ok
109. N. I think sans bruit is more, she, she fell
asleep and she didnt make any noise. But silence
is like everything around her is silent.
110. D. is silent
5Dara is speaking very softly in turn 358, leading us to label that utterance as private speech.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304296
Prompted by the researcher in the stimulated recall, Nina states the rule forreexive verbs in turns 174 and 176.6 In the post-tests, both learners get the item
ARTICLE IN PRESS
Table 4
LREs relating to se souvient
Original story (pre-test): elle souvient
Reformulation: se souvient
Post-test Nina: se souvient
Post-test Dara: se souvient
Stage 1 writing Stage 2 noticing Stage 3 stimulated recall
Form LRE (Pronominal verb) Accept reformulation Accept reformulation
334. D. Elle. 79. D. Tout "a coup, elle
souvient
173. R. so why did you need se
in any case?335. Elle se souvient, non, Elle
souvient quelle a uny unepratique de chorale, alors elle se
l"eve
80. N. Elle se souvient? 174. N. Because, because its
like something that is
happening to her.
336. D. Tout a coup
81. D. Ahh.
175. D. its sort of like suddenly
I think.347. N. Elle ssse
82. N. Oh! Je tai dit!
176. N. No, its not like that.
Something is happening to her,
like the action is happening to
her. So it has to be a verbe
r!efl!echi.
350. D. Elle souvient
83. D. non, tu nas pas!
351. N. se souvient ou souvient?
84. N. Oui!
352. D. Elle souvientyahh, elle sesouvientyElle souvientyElle sesouvient, non
85. D. Oh. Tu !ecris.
353. N. pasy
86. N. Ok, um. Elle se
souvient.
354. D. Elle souvient quelle doit
aller au bandy?
87. D. souvient.
355. N. Chorale.
356. D. Chorale.
357. N. xxxTout "a coup elle
souvient quelleydoit aller a lachorale
358. D. [very softly] elle se
souvientynon.359. D. Alors elle
360. N. non, waitytout "a coupelleyse souvient?361. D. Je pense pas que cest se
souvient
362. N. oh, souvientysouvient363. D. Elle souvient quelle a le
chorale
364. N. Quelle doit se pr!eparer
365. D. Oui
366. N. pour le choraleynon,tout "a coup elle souvient quil y a
une pratique de chorale
6 In turn 175 Dara is misled in her response by her tout "a coup in turn 366 of stage 1.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 297
right, suggesting their noticing in stage 2 and their collaborative dialogue in stage 3played a role in their ability to produce the correct pronominal form of the verb.Their producing se souvient in the post-test suggests that they were ready for thefeedback provided by the reformulated text.7
3.3. Post-test results
As stated above, there were a total of 47 LREs in the writing session. Of those,37% or 78% were resolved correctly. In essence, then, this means that Ninaand Daras written story can be considered as a pre-test. That is to say, Ninaand Daras story consisted of a text which incorporated their solutions to thelanguage problems they had identied while composing. But in some cases, theywere unable to resolve them correctly. Their text, then, was left with thosecases where they were unable to correctly resolve the problem, and those caseswhere they did not even notice that there was one. These were, in general, theitems or structures reformulated by the native speaker. Thus the changesthat Nina and Dara each made to their original written story when they individuallyrewrote it towards the end of week 2 represent what each learned fromnoticing the feedback (in the reformulated text) and from their dialogueabout it during the stimulated recall in which they reected on the changes madeto their writing. It is in this sense that we consider each students rewritten text as apost-test.We coded the post-test (stage 4) data by noting, for each reformulated item or
structure, whether Nina and/or Dara got the item right or wrong. Items wereconsidered correct if (a) they corresponded to the reformulated version of (b) theywere an acceptable (correct) alternative. Eighty percent of Daras and Ninaschanges were correct. Of those, approximately two-thirds corresponded exactly tothe reformulation, and one third were an acceptable alternative. The majority (three-quarters) of Nina and Daras responses were correct when they accepted thereformulation. Interestingly, approximately three-quarters of their responses werealso correct when they rejected the reformulation. This apparent anomaly can beexplained by reference to the three sets of LREs discussed above. Both acceptanceand rejection of the reformulators changes led to talking it through. Thisintermental process of talking it through mediated internalization such that eachlearner could individually draw on the knowledge they had previously jointlyconstructed.
4. Conclusions and discussion
In summary, we found that reformulation of learners writing, as implemented inour study, is an effective technique for stimulating noticing and reection onlanguage. The various stages of the task proved to engage these learners and provide
ARTICLE IN PRESS
7That is, the feedback was in their ZPD.
M. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304298
numerous opportunities for collaborative dialogue. During their collaborativedialogues, Nina and Dara noticed most of the differences between their story and thereformulated text, accepting or rejecting what they noticed. In just over a third of thecases, they rejected the reformulation. Rejection, however, does not necessarily meanthat no learning occurred.There were two main reasons why Nina and/or Dara rejected a reformulation.
First, the student had a rule that the reformulation violated, and she sawno reason to give up the rule. In Ninas words, I thought some of thecorrections were not necessary. Second, the students felt that the reformulationchanged their intended meaning; it did not respect the meaning they hadcreated and so they were unwilling to incorporate it into the existing text.As Nina said during her interview, Some of them [the reformulations], they seemedlike they changed the story sort of and it wasnt really ours. In her interview Daraexplained that: it would have been [more] helpful to have someone at about the samelevel or a little higher, like in possibly grade nine or something, go over it [theirstory]cause thats like closer to us and we would understand more the words and stuff.In other words, in some instances the reformulation did not serve as a successfulscaffold.Nina and Dara each got approximately 78% of the post-test items or structures
correct, a fact that attests to the power of the multi-stage writing, noticing andstimulated recall processes. Multiple opportunities to talk it through meant thatthe learners could reect on the language point in question and come to a deeperunderstanding of the proposed change.We conclude with Daras comment during her interview which indicates her view
on the value of her dialogue with Nina: When youre doing something in pairs youhave different ideas and you can incorporate them to make it better. She believed thatfor feedback to have an impact on her learning she needed to be able to talk about it.OK, so they changed something. But you dont know why they changed it. So thatswhy its better to have the person like, talking to you about the corrections that theyve
made. In a real classroom context, of course, one would hope that the teacherwould play this role.
Acknowledgements
This research was made possible through a grant (No. 410-99-0269) from theSocial Sciences and Humanities Research Council of Canada to Merrill Swainand Sharon Lapkin. We wish to thank the project team, Carole Bracco, LindsayBrooks, Agustina Tocalli-Beller, Linda Schmeichel, and Monika Smith. We alsowish to thank Andrew Cohen, Alister Cumming and Miles Turnbull who providedvaluable feedback on an earlier draft of this paper. We could not have collected thedata without the cooperation and assistance of the principal, teachers and studentsof the school; although we cannot identify them, we thank them for theircollaboration.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 299
Appendix A. Pictures used in Jigsaw task
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304300
Appendix B
Nina+Dara
Pre-test (collaboratively written)
Dormir o "u ne pas dormir?C!etait un jeudi matin et le soleil se l"eve. Marie-France est encore endormi. Cest 6:00du matin et soudainement, elle entend le sonnet de son alarme. Elle essaye dedormir malgr!e le bruit. Finalement, elle en a assez. Elle decide darr#eter lalarmeavec son orteil. Il est maintenant 6:01 et elle sendore sans bruit. Quand elle ne se l"evepas, larme devint frustr!e. Un bras y sort et commence de la chatouiller avec uneplume. Tout "a coup, elle souvien quil y a une pratique de chorale. Elle se force "asortir du lit et elle se prepare pour l!ecole en brossant ses dents et cheveux. 20minplus tard elle sort du maison en courant et essaye dattraper son ami Mike pour serendre "a l!ecole.
Reformulation
Dormir o "u ne pas dormir?Cest un jeudi matin et le soleil se l"eve. Marie-France dort encore. Il est 6:00 du matinet soudain, elle entend la sonnerie de son r!eveil. Elle essaie de dormir malgr!e le bruit.Finalement, elle en a assez et d!ecide darr#eter la sonnerie avec son orteil. Il estmaintenant 6:01, et elle se rendort dans le silence. Voyant quelle ne se l"eve pas, ler!eveil d!ecide dagir. Un bras en sort et commence "a la chatouiller avec une plume.Tout "a coup, elle se souvient quil y a une pratique de chorale. Elle se force "a sortir dulit, et elle se pr!epare pour l!ecole: elle se lave les dents et se brosse les cheveux. 20minplus tard, elle sort de la maison en courant et essaie de rattrapper son ami Mike pourse rendre "a l !ecole avec lui.
Post-testDara
Dormir o "u ne pas dormir?Cest un jeudi matin et le soleil se l"eve. Marie-France est encore endormi. Cest 6:00 hdu matin, soudainement, elle entend la sonnerrie de son alarme. Elle essaie dedormir, malgr!e le bruit. Finalement, elle en a assez, et d!ecide darr#eter la sonnerrieavec son orteil. Il est maintenant 6:01 h, et elle sendore silencieusement. Voyonsquelle ne l"eve pas, lalarme devient frustr!e. Un bras en sort et commence de lachatouiller avec une plume. Tout "a coup, elle se souvient quil y a une pratique dechorale. Elle se force "a sortir du lit et se pr!epare pour l!ecole: en brossant les dents etles cheveux. 20min plus tard, elle sort de la maison en courant et essaie dattraper sonami Mike pour se rendre "a l!ecole.
Post-testNina
Dormir o "u ne pas dormir?Cest un jeudi matin et le soleil se l"eve. Marie-France dors. Cest 6:00 du matin, etsoudain, elle entend son sonnerie. Elle essaie de dormir malgr!e le bruit. Finalement,elle en a assez et decide darr#eter la sonnerie avec son orteil. Il est maintenant 6:01 etelle se rendore en silence. Quand elle ne se l"eve pas, la sonnerie decide dagir. Un bras
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 301
en sort et commence "a la chatouiller avec une plume. Tout "a coup, elle se souvientquil y a une pratique de chorale. Elle se force "a sortir du lit et en preparant pourl!ecole, elle brosse ses dents et cheveux. 20min plus tard, elle sort du maison encourrant et essaie de rattraper son ami Mike pour se rendre "a l!ecole.
References
Allen, P., Swain, M., Harley, B., & Cummins, J. (1990). Aspects of classroom treatment: towards a more
comprehensive view of second language education. In B. Harley, P. Allen, J. Cummins, & M. Swain
(Eds.), The development of second language proficiency (pp. 5781). Cambridge: Cambridge University
Press.
Allwright, R. L., Woodley, M.-P., & Allwright, J. M. (1988). Investigating reformulation as a practical
strategy for the teaching of academic writing. Applied Linguistics, 9, 236256.
Brooks, L., & Swain, M. (2001). Collaborative writing and expert feedback: How they support second
language learning. Toronto: OISE/UT.
Cohen, A. (1982). Writing like a native: The process of reformulation. ERIC ED 224 338. Retrieved, from
the World Wide Web.
Cohen, A. (1983). Reformulating second-language compositions: A potential source of input for the learner.
Retrieved, from the World Wide Web.
de Bot, K. (1996). The psycholinguistics of the output hypothesis. Language Learning, 46, 529555.
De Guerrero, M. C. M., & Villamil, O. (2000). Activating the ZPD: Mutual scaffolding in L2 peer
revision. The Modern Language Journal, 84(1), 5168.
DeKeyser, R. M., & Sokalski, K. J. (1996). The differential role of comprehension and production
practice. Language Learning, 46, 613642.
DiCamilla, F. J., & Anton, M. (1997). Repetition in the collaborative discourse of L2 learners: A
Vygotskian perspective. The Canadian Modern Language Review, 53(4), 609633.
Donato, R. (1994). Collective scaffolding in second language learning. In J. P. Lantolf, & G. Appel (Eds.),
Vygotskian approaches to second language research. Norwood, NJ: Ablex Press.
Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University Press.
Galperin, P. Y. (1967). On the notion of internalization. Soviet Psychology, 5, 2833.
Genesee, F. (1987). Learning through two languages: Studies of immersion and bilingual education.
Cambridge, MA: Newbury House.
Harley, B. (1992). Patterns of second language development in French immersion. Journal of French
Language Studies, 2, 159183.
Izumi, S. (2000). Promoting noticing and SLA: An empirical study of the effects of output and input
enhancement on ESL relativization. Unpublished Ph.D. thesis, Georgetown University, Washington,
DC.
Kinginger, C. (2000). Learning the pragmatics of solidarity in the networked foreign language classroom.
In J. K. Hall, & L. S. Verplaetse (Eds.), Second and foreign language learning through classroom
interaction (pp. 2346). Mahwah, NJ: Lawrence Erlbaum.
Kowal, M., & Swain, M. (1997). From semantic to syntactic processing: How can we promote
metalinguistic awareness in the French immersion classroom? In R. K. Johnson, & M. Swain (Eds.),
Immersion education: International perspectives (pp. 284309). Cambridge: Cambridge University
Press.
Lambert, W. E., & Tucker, G. R. (1972). Bilingual education of children: The St. Lambert experiment.
Rowley, MA: Newbury House.
Lantolf, J. P. (2000). Second language learning as a mediated process. Language Abstracts, 33, 7996.
Lantolf, J. P. (2001). Private speech and internalization in second language learning: A
theoretical argument with some empirical evidence. Paper presented at the eighth annual
meeting of the sociocultural and second language learning research working group, Toronto:
OISE/UT.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304302
Lantolf, J. P., & Appel, G. (1994). Theoretical framework: An introduction to Vygotskian approaches to
second language research. In J. P. Lantolf, & G. Appel (Eds.), Vygotskian approaches to second
language research (pp. 132). Norwood, NJ: Ablex Publishing Corporation.
Lapkin, S., & Swain, M. (2000). Task outcomes: A focus on immersion students use of pronominal verbs
in their writing. Canadian Journal of Applied Linguistics, 3, 10011017.
Lapkin, S., Swain, M., & Smith, M. (2002). Reformulation and the learning of French pronominal verbs in
a Canadian French immersion context. Modern Language Journal, 86, 485507.
Levenston, E. A. (1978). Error analysis of free comparison: The theory and the practice. Indian Journal of
Applied Linguistics, 4, 111.
Lightbown, P. (1992). Getting quality input in the second/foreign language classroom. In C. Kramsch, &
S. McConnell-Ginet (Eds.), Text and context: Cross-disciplinary perspectives on language study
(pp. 197198). New York, NY: D.C. Heath.
Lyster, R. (1995). Instructional strategies in French immersion: An annotated bibliography. Canada:
Canadian Association of Immersion Teachers.
Mantello, M. A. (1996). Selective error correction in intermediate extended French writing programs: A
comparative study of reformulation and coded feedback. Unpublished Ph.D. thesis, University of
Toronto, Toronto.
Ohta, A. (2001). Second language acquisition processes in the classroom: Learning Japanese. Mahwah, NJ:
Lawrence Erlbaum.
Piper, A. (1995). Reformulation, evaluation and revision: Do they help academic writers? In M. L. Tickoo
(Ed.), Reading and writing. Theory into practice, Vol. 35 (pp. 2038). Singapore: SEAMEO Regional
Language Centre.
Qi, D., & Lapkin, S. (2001). An inquiry into the effects of noticing in a three-stage second language writing
task. Toronto: OISE/UT Modern Language Centre (mimeo).
Sanaoui, R. (1984). The use of reformulation in teaching writing to FSL students. Carleton Papers in
Applied Language Studies, 1, 139146.
Spielman Davidson, S. (2000). Collaborative dialogues in the zone of proximal development: Grade eight
French immersion students learning the conditional tense. Unpublished Ph.D. thesis, University of
Toronto, Toronto.
Stetsenko, A., & Arievitch, I. (1997). Constructing and deconstructing the self: Comparing post-
Vygotskian and discourse-based versions of social constructivism. Mind, Culture, and Activity, An
International Journal(4), 159172.
Storch, N. (1999a). Are two heads better than one? Pair work and grammatical accuracy. System, 27(3),
363374.
Storch, N. (1999b). How collaborative is pair work? ESL tertiary students composing in pairs. Language
Teaching Research, 5(1), 2953.
Storch, N. (2000). Is pair work conducive to language learning? The nature of assistance in adult ESL pair
work and its effect on language development. Paper presented at the conference: scaffolding and
language learning in educational contexts: Socio-cultural approaches to theory and practice, at the Centre
for language and literacy, University of Technology, Sydney, December 68, 2000.
Storch, N. (2001). An investigation into the nature of pair work in an ESL classroom and its
effect on grammatical development. Unpublished Ph.D. thesis, University of Melbourne,
Melbourne, Australia.
Swain, M. (1985). Communicative competence: Some roles of comprehensible input and comprehensible
output in its development. In S. M. Gass, & C. G. Madden (Eds.), Input in second language acquisition
(pp. 235253). Rowley, MA: Newbury House.
Swain, M. (1995). Three functions of output in second language learning. In G. Cook, & B. Seidlhofer
(Eds.), Principle and practice in applied linguistics. Oxford: Oxford University Press.
Swain, M. (2000). The output hypothesis and beyond: Mediating acquisition through collaborative
dialogue. In J. P. Lantolf (Ed.), Sociocultural theory and second language learning. Oxford: Oxford
University Press.
Swain, M., Brooks, L., & Tocalli-Beller, A. (2002). Peerpeer dialogue as a means of second language
learning. Annual Review of Applied Linguistics, 22, 171185.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304 303
Swain, M., & Lapkin, S. (1982). Evaluating bilingual education: A Canadian case study. Clevedon:
Multilingual Matters Ltd.
Swain, M., & Lapkin, S. (1995). Problems in output and the cognitive processes they generate: A step
towards second language learning. Applied Linguistics, 16, 371391.
Swain, M., & Lapkin, S. (1998). Interaction and second language learning: Two adolescent French
immersion students working together. Modern Language Journal, 82, 320337.
Swain, M., & Lapkin, S. (2001). Focus on form through collaborative dialogue: Exploring task effects. In
M. Bygate, P. Skehan, & M. Swain (Eds.), Researching pedagogic tasks: Second language learning,
teaching and testing (pp. 99118). Harlow, England: Longman.
Tang, G. M., & Tithecott, J. (1999). Peer response in ESL writing. TESL Canada Journal, 16(2), 2038.
Thornbury, S. (1997). Reformulation and reconstruction: Tasks that promote noticing. ELT Journal, 51,
326335.
Tocalli-Beller, A. (in press). Cognitive conict, disagreement and repetition in collaborative groups:
Insights into the affective and social dimensions. Canadian Modern Language Review.
Tudge, J. (1990). Vygotsky, the zone of proximal development and peer collaboration: implications for
classroom practice. In L. C. Moll (Ed.), Vygotsky and education (pp. 155172). Cambridge: Cambridge
University Press.
Villamil, O., & de Guerrero, M. C. M. (1998). Assessing the impact of peer revision on L2 writing. Applied
Linguistics, 19(4), 491514.
Vygotsky, L. S. (1986). The collected works of L.S. Vygotsky. Volume I: Thinking and speaking. New York,
NY: Plenum.
Wells, G. (2000). Dialogic inquiry in education: Building on the legacy of Vygotsky. In C. D. Lee, & P.
Smagorinsky (Eds.), Vygotskian perspectives on literacy research. New York, NY: Cambridge
University Press.
Dr. Merrill Swain is a Professor in the Department of Curriculum, Teaching and Learning at the Ontario
Institute for Studies in Education of the University of Toronto. Her interests include bilingual education
(particularly French immersion education) and communicative second language learning, teaching and
testing. Her present research focuses on the role of collaborative dialogue in second language learning. She
was President of the American Association for Applied Linguistics in 199899, and is currently a VP of the
International Association of Applied Linguistics (AILA).
Dr. Sharon Lapkin is Professor in the Modern Language Centre and Second Language Education Program
of the Ontario Institute for Studies in Education/University of Toronto. She has published widely in the
area of French second language education. Professor Lapkin is co-Editor of the Canadian Modern
Language Review/La Revue canadienne des langues vivantes.
ARTICLE IN PRESSM. Swain, S. Lapkin / Int. J. Educ. Res. 37 (2002) 285304304
Talking it through: two French immersion learners response to reformulationLiterature reviewTalking it throughReformulationFrench immersion in Canada
MethodologyResearch contextResearch questionsDesignTaskData analysis
FindingsOverview of LREsQualitative analysesPost-test results
Conclusions and discussionAcknowledgementsPictures used in Jigsaw taskReferences