L 1 Translation as Scaffolding in Tutor Talk : A Case Study of Two Korean Tutors

With the increasing popularity of the communicative approach in the mid-20th century, the use of the L1 was discouraged and stigmatized in foreign language education. Since the emergence of the sociocultural perspective, however, L1 use has been reconsidered as a key mediating tool for second language learning. Utilizing this sociocultural approach, this study examines how L1 translation is used in tutor-tutee talk during writing center appointments by analyzing the talk of two Korean tutors with their seven tutees. The analysis reveals that L1 translation not only helps the tutees clarify their intended meaning in their texts, but also serves as a successful tool for cognitively scaffolding tutee learning—in the forms of pump, hint, and prompt—by helping both the tutor and tutee negotiate meaning and arrive at a solution together. The results bring new insights regarding the use of L1 translation in tutor and tutee interaction by highlighting the usefulness of L1 translation as scaffolding for language learners, particularly when their language proficiency is not advanced. This study implies that the use of L1, in particular, L1 translation may contribute to L2 learners’ learning of English by facilitating their negotiations of meaning with their teachers as to finally arrive at mutual understanding of each other.


INTRODUCTION
Since the adoption of the teaching methods of Audiolingualism and Communicative Language Teaching, which emphasize that second and foreign languages be taught exclusively using the target language, the use of the L1 in second language classrooms has been discouraged and stigmatized because of its strong association with the Grammar Translation Method (Brown, 2007;Larsen-Freeman & Anderson, 2011).For the past two decades, however, L1 use has attracted new attention from scholars taking both cognitive and sociocultural perspectives (Ánton & DiCamilla, 1998;Brooks & Donato, 1994;Cook, 1992;Villamil & Guerrero, 1996).For example, in criticizing teaching approaches that prohibit L1 use in the foreign and second language classroom, Cook (1992) argued that multi-linguals have multi-competence, in which knowledge of multiple languages is viewed as integrated, which means that L2 knowledge should not be viewed as separate from that of the L1.That is, multi-linguals can take advantage of their L1 in learning a second or subsequent language.
From a sociocultural perspective, L1 knowledge can be viewed as a critical tool in mediating L2 learning.Adopting the Vygotskian view that learning originates from an external plane (interpsychological) by means of interactions with others or with cultural artifacts such as signs and tools, and is gradually internalized into an internal plane (intrapsychological), several scholars have emphasized the importance of the L1 in this learning process (Ánton & DiCamilla, 1998;Brooks & Donato, 1994;Villamil & Guerrero, 1996).For example, based on the analysis of talk between American learners of Spanish, Ánton and DiCamilla (1998) showed that the learners' L1 not only helped them to better understand their tasks but also allowed them to provide scaffolded help to each other so that they eventually succeeded in fulfilling the task.
In fact, this Vygotskian perspective has been embraced by many researchers who examined tutor-and tutee-talk at writing centers (Thonus, 1999(Thonus, , 2004;;Weissberg, 2006a;Williams, 2004;Williams & Severino, 2004).They have studied tutor-and tutee-talk at writing centers as a typical model of scaffolded talk between an expert and a novice because tutors can tailor their talk more easily to an individual student's language ability than a teacher can in the classroom.As mentioned in studies that explored how L1 is used as a tool of scaffolding, even in tutor-and tutee-talk, the use of L1 may play an important role, but unfortunately, studies on tutor-and tutee-talk have mainly focused on the talk between a native English tutor and a non-native English tutee.Therefore, no studies to date have investigated how L1 functions in the scaffolded talk of tutors and tutees, who share the same language background on L2 writing at a writing center.Therefore, this study explores the role of L1 in L2 learning, in particular, L1 translation in the talk of two Korean tutors of English with their tutees.

LITERATURE REVIEW SCAFFOLDED TALK
From a sociocultural perspective, talk between native speaker (NS) tutors and non-native speaker (NNS) tutees at writing centers has been viewed as a typical way of scaffolded learning (Ewert, 2009;Thonus, 1999;Williams, 2002Williams, , 2006;;Weissberg, 2006b).Ewert (2009) coded the talk between two NS (referred to as native speakers, hereafter) tutors and their six NNS (referred to as non-native speakers, hereafter) tutees based on six features of scaffolding-recruitment of the learner's interest in the task, reduction of task complexity, maintenance of the goal-oriented activity, markings of discrepancies between learner's production and the solution, control of learner frustration, and demonstration of possible solutions-and found that one of the teachers provided more scaffolding by tailoring her talk to the tutees' proficiency in comparison with the other teacher.In a similar vein, Weissberg (2006b) analyzed the talk between a NS tutor and two NNS graduate students and found that the talk demonstrated various linguistic features of scaffolding mechanisms such as asking questions, repetition of tutee's phrases, or personal affiliations, all of which helped the tutor to create links to the tutee.Based on the findings, Weissberg (2006b), concludes that scaffolding serves as "a highly individualized, negotiated means of delivering oral feedback to L2 students on their writing" (p.247).
While these studies focus on how tutor scaffolding helps tutees to better manage language tasks, it has also been noted that scaffolded talk contributes to higher performance of L2 learners, such as better revisions in the case of second language writing (Aljaafreh & Lantolf, 1994;Goldstein & Conrad, 1990;Williams, 2004).Through the analysis of both teacher conferences with three ESL students and their revised drafts, Goldstein and Conrad (1990) found that the tutor scaffolded the students differently depending on the extent to which they actively participated in the talk and that the tutor's scaffolded talk, when accompanied by the student's negotiation of meaning, resulted in more successful revision.Williams (2004) found similar result regarding the close relationship between tutor talk and the quality of revision after analyzing five writing sessions with five ESL students.She pinpointed several unique features to tutor talk-active writer participation and tutor clarification of critical features, sustained emphasis on goals, organization of tasks, and modeling-that contributed to substantial revision in students' revised drafts.Likewise, while analyzing how tutors differentiated the amount and type of assistance depending on tutee language abilities, Aljaafreh and Lantolf (1994) proved that this assistance was worthwhile by looking at the evidence in student's appropriation of tutor suggestions in their revised drafts.
Although these studies have helped deepen our understanding of tutor talk between NS tutors and NNS students and to clarify how L2 learning can be scaffolded through NS tutor talk, there is a lack of research on how L2 tutors and tutees communicate with each other during appointments and how they negotiate meaning.In particular, between L2 tutors and tutees who have the same L1 background, the use of the L1 may play a key role in scaffolding tutee learning.

USE OF L1 IN SCAFFOLDING
L1 has been shown to provide scaffolding between peers and small group of L2 learners who share the same language background (Ánton & DiCamilla, 1998;Brooks & Donato, 1994;Villamil& Guerrero, 1996).While analyzing the talk between eight pairs of learners of Spanish who were engaged in a problem-solving task, Brooks and Donato (1994) revealed that the L1 talk helped the learners regulate the task and internally reconstruct it so that they could successfully manage it.Villamil and Geurrero (1996) found similar use of L1 among Spanish speaking learners of English during their peer review; that is, L1 was often used to take control of the task and was reported as serving various purposes, such as "making meaning of text, retrieving language from memory, exploring and expanding content, guiding their action through the task, and maintaining dialogue" (p.60).In their investigation of how L1 was used between adult Spanish learners, Ánton and DiCamilla (1998) argued that L1 benefits L2 learning in constructing scaffolded help, building up intersubjectivity, and that the L1 use provides learners with "a social and cognitive space in which learners are able to provide each other and themselves with help throughout the task" (p.338).
In line with the studies on benefits of L1 use among L2 learners, this study raises the question as to whether the benefits observed among students of the same language background can be applied to the talk between L2 tutor and tutee.If such benefits exist, how is L1 utilized between more-expert tutors and less-expert tutees?In particular, in tutor-and tutee-talk, we focus on the use of L1 translation.
L1 translation in second language learning has primarily been studied in the area of vocabulary acquisition and reading, but not in the context of the its application in writing center.Regarding vocabulary acquisition, L1 translation has been found to cognitively facilitate L2 vocabulary acquisition, especially in the early stages of vocabulary acquisition (Daulton, 1999;Kroll & Tokowicz, 2001).For example, Daulton (1999) argued that Japanese students learn English words better and quickly when they are loanwords in Japanese.Regarding reading comprehension, L1 translation also seems to facilitate learners' reading comprehension (Yau, 2010).For example, Yau (2010) examined the role of mental translation, that is, learners' literal or rough translation of a text, among Taiwanese high school students who read classical Chinese and English as a foreign language through both quantitative (reading comprehension tests and translation surveys) and qualitative methods (think-aloud protocols and interviews).Based on the finding that mental translation has a positive effect on learners' reading comprehension of both Classical Chinese and English, Yau (2010) concludes that L1 translation mediates learning by linking the interpsychological to the intrapsychological level.According to Yau (2010), "[translation] is, in a sense, a manifestation of both the internalization and externalization of verbal thought, either within one language or between two languages" (p.385).Drawing on these findings, this study examines how two Korean tutors of English, Ms. Han and Ms. Lee, use Korean translation with their tutees during appointments at the writing center.Our research questions are as follows: 1) What role does L1 translation have in Korean tutor talk with their tutees?2) How does L1 translation affect tutees' understanding of English writing? 3) How does the talk associated with L1 translation contribute to tutee revisions?

PARTICIPANTS
The participants were two Korean tutors of English, Ms. Han and Ms. Lee, 1 and the seven tutees they helped during the week of data collection-S1, S2, and S3 were with Ms. Han and S4, S5, S6, and S7 were with Ms. Lee.They were chosen for this study because their sessions revealed remarkably different patterns of scaffolding in tutor-tutee talk.At the time of the study, Ms. Han and Ms. Lee had been teaching College English for five to six years for a language program coordinated by one of the authors at a university near Seoul, Korea and had been working twice a week as tutors for the writing center run by the university.The tutors were given freedom to choose the medium of communication-whether it is Korean or English.Both of them decided to use Korean with their Korean tutees, based on the assumption that Korean would make communication with the tutees easier and would facilitate the tutees' understanding of their writing problems.All student participants were female sophomores majoring in a variety of subject areas (e.g., Information and Statistics, Child Education, Media Design, and Administration) and were taking College English with instructors other than Ms. Han and Ms. Lee, as a university requirement.Their English proficiency levels ranged from 400 to 500 as measured by the TOEIC, and none had experience in studying in an English-speaking country.Despite a seemingly homogenous group of learners in terms of their language proficiency, as indicated by TOEIC scores, the writing abilities of the seven participants varied greatly, as measured by the modified Weigle (1999) 10-point rating scale (see Appendix).Although Weigle divided the scale into three different areas-content, grammar, and rhetorical control-we simplified the scale, focusing only on content and grammar, given that participants were working on different genres.Table 1 shows the average scores of each participant's writing ability.

DATA COLLECTION
The seven writing tutorials-one tutorial with each student for one assignment-were videorecorded, and copies of the first and the revised drafts were collected.Students visited the writing center as a course requirement and received feedback from the tutors on one of the assignments they completed for class.The participants chose their own topics from the list of topics given by their instructors.Four of them-S2, S3, S5, and S6-wrote descriptive essays about first impressions: S2, S5, and S6 entitled their essays "First Impressions," while S3 used the title "Accuracy of the First Impression."S1 wrote an expository essay entitled "The Natto," introducing the effects of natto on health, or Japanese fermented soybeans, and S4 and S7 wrote descriptive essays entitled "If I won a lottery'' and "The most memorable day," respectively.One week after receiving feedback from their respective tutors, students submitted their revised drafts to the writing center.Each tutorial (one to one) lasted 30 to 40 minutes but was composed of a different numbers of tutor-tutee turns: tutorials with S1 (130 turns), S2 (63 turns), S3 (369 turns), S4 (259 turns), S5 (354 turns), S6 (400 turns), and S7 (205 turns).

DATA ANALYSIS
Scaffolding is usually examined using Guerrero and Villamil's (2000) analysis scheme in second language writing (Ewert, 2009;Weissberg, 2006b).Guerrero and Villamil (2000) adopted the scaffolding mechanism proposed by Wood, Bruner, and Ross (1976), which has six features: "recruiting writer's interest in the task; marking critical aspects of the text requiring correction by the writer; instructing, or giving mini-lessons; providing the writer with problem-solving heuristics; responding contingently, based on the tutor's judgments about the level of help needed by the writer at any given moment; and affective involvement, in which the tutor provided the writer with positive feedback and encouragement to persist in the task."(p.250).Although this mechanism can help illuminate a series of key scaffolding features, it is limited in its ability to clarify the types of scaffolding used and explicate in a detailed manner how each feature of scaffolding is accomplished.For instance, when more knowledgeable partners, such as teachers, tutors, or even peers, respond contingently depending on the writer's current ability, Guerrero and Villamil's (2000) scaffolding mechanism does not allow for a classification of the specific techniques and strategies used by the experts.However, the analysis scheme of Cromley and Azevedo (2005), which the present study has adopted, helps delve into the process of scaffolding in more detail.Although their analysis scheme was originally developed to examine oral interactions between reading tutors and their tutees, the scheme can be adapted to the writing context to obtain an overview of what types of scaffolding are provided and how they are performed by dividing the tutoring process into a series of recognizable steps.Cromley and Azevedo (2005) classified scaffolding into three major types: instruction, cognitive scaffolding, and motivational scaffolding.Additionally, each type is classified into several sub-categories, as seen in Table 2. Instruction includes direct instruction, giving an example, explaining, giving the correct answer, and posing a problem.Cognitive scaffolding includes forced choice, hinting (the tutor suggesting how to solve a problem), prompting (the tutor eliciting the tutee to fill in the last word), and pumping (the tutor asking the tutee to elaborate on a certain point).Motivational scaffolding includes acknowledging difficulty, using humor, giving negative feedback, giving positive feedback, using reinforcement, and using reinforcement with elaboration.
This study adopts Cromley and Azevedo's (2005) analysis scheme with slight modifications: the types that were not found in the current data (i.e., counter example, in which the tutor provides a sample non-application of a rule, and preview, in which the tutor indicates that a new topic is being introduced and describes the topic) were removed and new types (e.g., reinforcement with elaboration, in which a tutor repeats the student's answer with elaboration), which were often utilized by the tutors in the current data, were added.After the video-recorded tutorials were transcribed line by line, each complete statement by the tutor or student was counted and coded, drawing upon this analysis scheme.In particular, we focused on L1 translation2 in order to examine how translation is used in the process of scaffolding.We independently coded the analyzed data, compared our analyses, and finalized the coding by discussing discrepancies.

ROLE OF L1 TRANSLATION
The analysis of tutor talk revealed that L1 translation was pervasively used in both Ms. Han's and Ms. Lee's tutorials as a tool to instruct and cognitively and motivationally scaffold their tutees.The two tutors translated from English to Korean, but they show different patterns in their interactions with tutees, and in turn, they use L1 translation for different purposes: while L1 translation is used by Ms. Han primarily for instruction, it is used by Ms. Lee for cognitive and motivational scaffolding as well as instruction.

L1 translation as an instructional tool: Ms. Han (T-H).
The typical sequence of teacher feedback in Ms. Han's tutorials with S1, S2, and S3 consists of a series of tools for instructional purpose: 1) tutor poses a problem (PP), explains (EXP), and gives correct answer (GCA), or 2) PP and GCA.In Excerpt 1, L1 translation is usually used by the tutor in either posing a problem or giving a correct answer.T-H

S3
T-H "She did not laugh well," but laugh, you meant jal usji anhda(she did not laugh a lot).You're talking about jaju usji anhda(she did not laugh often), not well.This is not a matter of ability.Umm.(with nod) So here you meant she did not laugh often, very often.You'd better change it.

INS-GCA
In this excerpt, the tutor poses the problem in the sentence "she did not laugh well" and explains that laughing is not an ability she can modify with the adverb "well."Instead, Ms. Han gives the correct answer "she did not laugh often."In this instructional sequence of posing a problem, explanation, and giving a correct answer, Ms. Lee uses L1 to indicate the difference between the tutee's original intention and her English expression.First, she confirms what S3 originally meant by providing its L1 translation jal usji anhda(did not laugh a lot).The adverb jal-literally translated into the English word "well"-can mean both frequency and ability in Korean.Next, in order to reduce the ambiguity of the word jal, Ms. Han replaces it with a better adverb, jaju (often) in the second translation jaju usji anhda (did not laugh often).The use of L1 translation in this case enables the student to become aware of the difference between her intended meaning and her written meaning, as noticed in S3's acknowledgement in turn 106.
Interestingly, during the entire tutorial with S1, Ms. Han intentionally asked S1 to translate her English texts into Korean, sentence by sentence to understand her original meaning, as can be seen in Excerpt 2.

INS-GCA + INS-EXP
In turn 41, S1 gives an L1 translation for her English sentence ("First of all, natto's most important effect is suppressing the aging process").This translation seems to help Ms. Han understand S1's intended meaning.After listening to S1's translation, Ms. Han again uses the same instructional pattern, which was frequently noticed in the case of S3.In turn 42, she poses a problem in her word choice ('suppress') by repeating its Korean translation, eogje (suppress).In turn 43, S1 suggests a different Korean translation, abbag (suppress), which has a very similar meaning to the previous Korean translation.In turn 44, Ms. Han gives the correct answer 'slow down,' along with its Korean translation neujchwojunda (slow down), which she thinks is more suitable for this particular context.Excerpt 2 shows that Ms. Han is giving the correct answer drawing on the tutee's knowledge of Korean vocabulary.
L1 translation as cognitive scaffolding: Ms. Lee (T-L).Ms. Lee also used L1 translation with an instructional function, such as posing a problem and giving a correct answer.However, Ms. Lee's tutorials diverge from Ms. Han's in their interactional pattern.While Ms. Han follows the typical pattern consisting of a series of instructional tools such as PP, EXP, and GCA, Ms. Lee utilizes a series of cognitive and motivational scaffolding strategies, as seen in Excerpt 3.This sequential interaction between Ms. Lee and S6 shows a very typical case of reducing the complexity of a task, which is often quoted as one of the key features of scaffolding (Ewert, 2009;Wood, Bruner, & Ross, 1976).Interestingly, Ms. Lee reduces the task difficulty with the help of L1 translation.In turn 52, Ms. Lee notices a problem with S6's expression 'my thinking has changed because of this,' and asks for the solution from the student.When the student remains silent, Ms. Lee cognitively scaffolds her by providing a hint, that is, L1 translation (iyuga issda).When the student still remains silent, Ms. Lee simplifies the task by cognitively scaffolding her with a prompt and a Korean translation (I have a reason) in turn 56.Faced with continued silence, in turn 58, Ms. Lee gives the correct answer herself, "I have a reason," and repeats the same scaffolding procedure-hint and prompt-using L1 translation from turn 58 to 64.In turn 62, Ms. Lee provides a hint about the phrase that will follow the word "why" by using its Korean translation (saenggageul bakkun iyu).When S6 does not answer in turn 63, Ms. Lee prompts S6 so that she can say each word one-by-one instead of an entire phrase (I changed my mind) by providing its Korean translation (bakkwossda).Finally, arriving at turns 65 and 67, S6 starts to respond to the tutor's hints and prompts, and the tutor confirms the answers with motivational scaffolding and reinforcement with elaboration.
Ms. Lee's usage of L1 translation differs depending on the writing ability of the tutee.For students with lower ability, such as S6, Ms. Lee uses L1 translation more frequently as a tool of cognitive scaffolding.However, in the tutorial with S7, whose writing was evaluated as high, Ms. Lee does not use as much L1 translation because S7 self-corrects many of her own grammatical mistakes, as seen in Excerpt 4. In Excerpt 4, S7 self-corrects her own preposition error by filling in 'to' while reading the sentence, although her attempt is not successful.In turn 43, Ms. Lee complements her attempt by providing positive feedback, a strategy to help motivationally scaffold her.When S7 is not able to answer her pump, "what does the preposition to mean," Ms. Lee gives the answer using L1 translation.When Ms. Lee tries another pump, "what do you want to say here?",S7 delivers her own intention "to home from school."In turn 47, Ms. Lee gives a hint to S7 by translating the target English sentence into Korean (jigeum haggyoeseo dolawasseul ttae).In response to Ms. Lee's hint, in turn 48, S7 successfully provides the correct preposition: 'from' instead of 'to.'Ms. Lee finishes this sequence with motivational scaffolding, reinforcement with elaboration (when I came back from school).In comparison with Excerpt 4 from the tutorial with S6, whose writing ability was evaluated as lower than S7, in the tutorial with S7, Ms. Lee not only uses cognitive scaffolding less frequently-two pumps and one hint-but she also uses less L1 translation.In this case, Ms. Lee does not use any prompt with L1 translation.All this evidence-less cognitive scaffolding and less L1 translation-shows that Ms. Lee does not need to try as hard as in the tutorial with S6 to reduce the difficulty of the task.

EFFECTS OF L1 TRANSLATION AS SCAFFOLDING
During Writing Center Appointments.The ensuing tutor-tutee talk and student revision after the writing center appointments indicate that Ms. Lee's cognitive scaffolding through L1 translation was beneficial at least in terms of students' gaining knowledge of the particular structures discussed during the tutorial.In contrast, as Ms. Han's tutees were not given cognitive scaffolding to collaborate with the tutor in figuring out the solution, Ms. Han's tutees did not correct on their own any of the mistakes pointed out by her during the tutorial.As seen below, in Excerpt 5, Ms. Han's tutorial mainly consists of a series of instructionposing a problem and explanation followed by the tutor giving a correct answer-without much room or many opportunities left for the tutee to contribute to the talk.Hold on.This sentence looks weird.People do not say "some people is disgusted."You should say eotteon salamdeuleun nasttoleul disgustinghadago saenggaghanda (some people think natto is disgusting).A person is disgusted, a be-verb plus past participle means someone gets like this.Some people think natto is disgusting.Here is a spelling mistake."is disgusting because" and then "including me" here.It is too far from the subject, you're one of the people.Including me, another comma, and then makes the sentence start with the subject including me and then put the rest of that.Umm, why did you use "but the" here?It's because I said this before and then this one looks contradictory.It looks like this is my habit.Yes, I think so.You don't have to do that.Hold on.You seem to use but too often.Yes, I think so, too.You don't have to do that.For this kind of essay, you can use but, but in academic writing, people do not use but.You should not start a sentence with but, but you can use a clause but another clause.Next.incorporate the last two pieces of tutor feedback-the spelling mistake and the relocation of the expression, "including me."When a series of PP, EXP, and GCA follow one after another and when the tutee's participation or learning is not stimulated as in Excerpt 6, it is not likely that the tutee will internalize the tutor's suggestion or advice.
Unlike Ms. Han's tutorials, Ms. Lee's tutorials show several instances in which the tutees give the correct answer for themselves as the tutorials proceed through tutor scaffolding, often mediated through a less frequent use of L1 translation, as can be seen in Excerpts 6-8 with S5.

COG-HI
In the first draft, S5 wrote, "first impressions can be very different than their characters or personality," and in turn 77, Ms. Lee poses a problem with the phrase "first impressions."In turn 79, the tutor gives a hint to S5 using L1 translation, nwukwunkaey tayhan chesinsang (first impressions of someone), but from turn 81, the tutor gives the correct answer, "first impression of someone," after giving its literal translation again, salamtuley tayhan (about people), nwukwunkaey tayhan chesinsang (first impressions of someone).In turns 83, 85, and 87, Ms. Lee uses L1 translation several times when she gives the correct answer "my first impressions of someone."About sixty turns later, the same issue is noticed by the tutor.Although it takes 12 turns in Excerpt 7, it takes only 7 turns for the tutor and the tutee to arrive at the same solution: "my first impressions of someone."In fact, at Ms. Lee's mere use of cognitive scaffolding, in the form of a hint, which is provided in the format of L1 translation, nwukwunkaey tayhan chesinsang (first impressions of someone), S5 successfully gives the correct answer: "first impressions of someone."After about 40 turns, the same problem is identified by Ms. Lee again (i.e., "if you remember your first impression").This time, Ms. Lee cognitively scaffolds the tutee using a hint without any L1 translation.In turn 192, Ms. Lee only reminds the tutee of what they discussed in the previous turns, the tutee gives the correct answer, "of someone," immediately.S5's writing ability was identified as the lowest, but these three consecutive excerpts show that despite S5's low writing ability, she seems to gain grammatical knowledge of a particular structure through the cognitive scaffolding employed by Ms. Lee, and, in particular, the use of L1 translation.In addition to S5, Ms. Lee's tutorials with S4, S6, and S7 contain similar incidences, showing that they are becoming more independent as the tutorials proceed with the help of tutor scaffolding.For example, before Excerpt 9 (below), Ms. Lee had given S6 the correct expression "those who" as well as its Korean translation by saying "those who mean etteettehan salamtul (those who)."The same phrase was given again by the tutor with its Korean translation later.Finally, arriving at turn 376 in Excerpt 9, the tutee, not the tutor, provides the correct answer.
Excerpt 9 (Ms.Lee) Turn Speaker Utterance Scaffolding Type Good first impression?You can write like the previous one.How can you say it?(no response) On the contrary, and then do as you did before.people people people who give us Right.
In Excerpt 9, the tutor does not use L1 translation, but simply stimulates the tutee by urging her to do as she did in turn 373.In turns 374 and 376, S6 gives the correct expression, "people who give us good impressions," correcting her original expression in the sentence "good first impression on people have a good feeling, but may be cheaters or a bad person."Excerpts 6-9 reveal that Ms. Lee's use of L1 translation helps the tutees to become aware of the problem and also to provide correct answers for themselves as the tutorials proceed.
Student Revision.As mentioned previously, student revisions reveal that compared to Ms. Han's tutees, Ms. Lee's tutees were more likely to incorporate tutor feedback into their revisions and were less likely to delete the parts where problems were indicated during the tutorials.Despite the idiosyncratic variation among the participants-Ms.Han's S2 incorporated most of her feedback into the revision, while Ms. Lee's S6 incorporated less feedback into her revision in comparison with Ms. Lee's other tutees-93% of tutor feedback was reflected in the revisions of Ms. Lee's tutees, while only 67% of tutor feedback was incorporated by Ms. Han's tutees.Of course, we should be cautious in simply assuming that higher percentages of incorporation of tutor feedback necessarily indicate improvement in tutee writing quality or in their general writing ability.However, the large discrepancy in the amount of deletion as well as in the lack of incorporation of feedback is worth paying attention to.In the case of Ms. Han's tutees, 12% of her feedback was not incorporated into her students' revisions, and 21% of the trouble sources she pointed out were simply deleted in the revisions.Excerpts 10 and 11 demonstrated cases in which the tutor's failure to meet the tutee's needs leads to the deletion of the sentence that includes the troubled source.

INS-PP MOT-NF + INS-DI
In Excerpt 10, the tutor poses a problem with S3's choice of the vocabulary 'assess' in the sentence "I had no choice but to assessing all friends," and she also rejects S3's alternative, 'evaluate,' judging both of them as being too strong in the given context.After about 50 turns, when the tutor reads a similar kind of sentence "the reasons made evaluate her badly," containing the very word, 'evaluate,' which S3 suggested but the tutor rejected in turn 68, the tutee initiates the discussion by posing a problem in the sentence.28 look up the word, make, the phrase of make an object and an infinitive means to make the object do something.So this is right.Because of that, the word evaluate is right.This sentence is perfect.S3's comment "this one, too" with rising intonation indicates that she is questioning her choice of the word 'evaluate,' used in a similar context as in Excerpt 10.To S3's surprise, however, the tutor responds by positively evaluating her word choice.Looking at Ms. Han's explanations, it becomes evident that she focuses on the usage of the word 'make + object + infinitive' instead of the word choice, missing S3's implied question.At the end of turn 127, Ms. Han evaluates the target sentence as perfect.Ms. Han's conflicting evaluations of the word 'evaluate' in similar contexts may have confused S3 and deprived her of the chance to gain knowledge of the word.In her revised draft, S3 removed the sentence, in spite of her tutor's positive evaluation: "this sentence is perfect." In comparison with Ms. Han, Ms. Lee's tutees rarely removed or deleted sentences where she initially posed problems, most likely due to the fact that she urged them to participate in solving the problems through consistent use of cognitive scaffolding.Although mere acceptance of tutor feedback can also demonstrate lack of student agency, the high percentage of incorporated feedback and the low percentage of deletion witnessed in Ms. Lee's tutees suggest that they made efforts to solve the problems discussed during the tutorials, given that deletion or removal of trouble sources would not lead to a solution.Furthermore, it is highly probable that Ms. Lee's thoughtful use of cognitive scaffolding and L1 translation, used with varying degrees depending on student writing ability, contributed to this high rate of incorporation and the low rate of deletion.

DISCUSSION
The analysis of these two Korean tutors' interactions with their tutees during writing center appointments shows that even though both of them used L1 translation as a tool for scaffolding, they used this tool for different purposes.While Ms. Han used L1 translation mostly for instructional purposes, such as posing a problem in a student essay or giving the correct answer, Ms. Lee also used L1 translation to cognitively and motivationally scaffold her tutees, which is evidenced by the fact that she not only reduced the task difficulty using L1 translation, but also varied the amount of L1 translation depending on writing ability of the tutee.This finding both confirms and expands on past research on L1 use in L2 learning (Cumming & So, 1996;Kim & Yoon, 2014;Polio & Duff, 1994).Although these scholars acknowledged the usefulness of the L1 in second language classrooms, including writing center tutorials, they also showed concern about L1 use.Through the comparison of L1 and L2 use in tutor-and-tutee talk, Cumming and So (1996) stated that L1 use may provide "deeper and profound opportunities to make logical links between the two languages" (p.216), but they also warned that exclusive use of the L1 may deprive learners of opportunities to learn to negotiate in the L2.Polio and Duff (1994) witnessed frequent L1 use across foreign language classrooms that served a variety of functions, such as giving grammar instruction, taking a stance of empathy and solidarity, and compensating for students' lack of comprehension.Drawing upon this finding, they recommended that foreign language teachers should be aware of their frequent use of the L1 and attempt to use more of the target language.
The findings of this study, however, demonstrate that L1 use varies depending on the tutor and that it has the potential to scaffold tutee learning.Ms. Lee's tutorials in this study demonstrate that a certain kind of L1 use-in particular, a certain kind of L1 translation as we witnessed in Ms. Lee's case-has the potential to both cognitively and motivationally scaffold students.Her typical pattern of cognitive scaffolding consists of pump, hint, and prompt, followed by reinforcement with elaboration.At every stage of this sequence, Ms. Lee used L1 translation to facilitate tutee understanding and involve them in the negotiation of meaning.Specifically, Ms. Lee prompted tutee answers by providing them with Korean translations for the target English word.Ms. Lee also modified her use of L1 translation as cognitive scaffolding depending on tutee English ability.For tutees with lower writing ability, Ms. Lee was more likely to go through all the stages (i.e., pump, hint, and prompt) and repeat this sequence several times until the tutee gave a correct answer, relying on L1 translation more heavily.On the other hand, for more advanced writers, Ms. Lee used pumps and hints without using prompts or relying on L1 translation.
Ms. Lee's modified use of L1 translation and cognitive scaffolding depending on her tutee's English ability seems beneficial in facilitating learning, whereas Ms. Han's uniform use of L1 translation as a mere instructional tool across all tutees is limited.In fact, even later in the same tutorial, Ms. Lee's four tutees showed improved understanding of their problem sources when the same problem was pointed out by Ms. Lee, demonstrating increased independence by giving correct answers themselves.The amount of tutor feedback incorporated into revisions also differed between Ms. Han's and Ms. Lee's tutees.In the case of Ms. Han, while two-thirds of the trouble sources she pointed out during the tutorials were corrected in their revised drafts, one-third was either left uncorrected or simply deleted.On the other hand, Ms. Lee's tutees showed a higher percentage of incorporation, with few instances (9%) of uncorrected or deleted trouble sources.Although this higher acceptance of tutor feedback does not necessarily correlate with a higher quality of writing, the fact that Ms. Han's tutees did not even attempt to correct the trouble sources and simply deleted certain sentences may imply that the type of scaffolding provided by Ms. Lee may have led to a higher probability of feedback incorporation.

CONCLUSION
Several scholars have reported that non-native speaker tutees feel uncertain about how to perform their roles appropriately in writing center tutorials partly because of their lack of L2 proficiency (Conrad & Goldstein, 1990;Harris, 1997;Thonus, 2004;Williams, 2004).This study reveals that when tutors and tutees share the same mother tongue, the use of the L1 not only lowers the language barrier that might have existed between them if they interacted with each other in the L2, but it also can be used as a key tool to scaffold the understanding of the tutees.Of course several shortcomings in NNS tutor-NNS-tutee talk exist in comparison with NS tutor-NNS tutee talk, such as NNS tutors' imperfect knowledge of the L2 as witnessed in Ms. Han's rather prescriptive explanations of the coordinating conjunction 'but' and NNS tutees' fewer opportunities to practice the L2 during the tutorials.Despite these shortcomings, however, the use of the L1 (and L1 translation in particular) can be beneficial, especially to NNS tutees.If bilinguals have different cognitive processes and are multi-competent (Cook, 1992;Cook & Singleton, 2014), there is no reason why only the target language should be used in foreign language learning.Specifically, for low proficiency L2 learners, L1 translation can be used to effectively and efficiently mediate their learning.The use of L1 translation as scaffolding to L2 learning, as this study shows, can open up new pedagogical possibilities.
Due to the small number of cases involved, this study is limited in its generalizability to different populations and different language learning contexts.Further studies are needed to extend the focus to a larger population in a variety of learning contexts.Additionally, most tutor feedback in this study is focused on grammatical features such as syntactic features and lexical choices rather than issues of content, organization, logical flow, coherence, or rhetorical style.Despite these limitations, however, the findings of this study are insightful, taking into account the fact that formalistic features such as lexical choices, morphology, and syntax are more emphasized in teaching writing in EFL contexts.While writing centers have been considered important venues for 'learning to write' (LW) in the North American context (Ortega, 2011), this study shows that a writing center may serve different purposes depending on the context in which it is situated and that it can also play an important role in promoting "writing to learn a language' (WLL) in EFL contexts like Korea.Ortega (2011) once asked for further research that examines WLL, advocating that it become "a more central concern in L2 writing scholarship" and that "SLA theory and research grapple with L2 writing as a site of language learning" (p.249).In the same way, WLL in EFL contexts should be studied in more depth and not be disparaged as an extraordinary form of the writing center that is dominant in the North American context.
you intended to say here.What kind of fact?my, my first impressions are different from.First impressions, right?So you mean my first impressions of someone, my first impressions can be wrong.Yes.How can we write?We talked about before.How do you say nwukwunkaey tayhan chesinsang(first impressions of someoneif you remember your first impressions about people you can improve your understanding of people."Mmmm.There's also something wrong here.Can you catch it this time?If you remember your first impressions, what expressions did we use before?Of someone Right.You can also say of people.

TABLE 1
1 All these names are pseudonyms.

TABLE 2 .
Analysis Scheme of Scaffolding What expression shall I use?You meant ileon peuloseseuleul neujchwojunda(slow down this process).You meant slow down.Then I think a word other than suppress will be used in that context.You meant anti-aging, and then consult relevant materials about aging and anti-aging.

TABLE 3 .
Ms. Han's Tutees' Incorporation Rate of Tutor Feedback

TABLE 4 .
Ms. Lee's Tutees' Incorporation Rate of Tutor Feedback