The Interface of Error Types , Teacher ’ s Feedback , and Students ’ Uptake

The present study is an attempt to investigate the frequency of different types of errors committed by EFL learners and the most prevalent types of errors, the types of corrective feedback do EFL teachers provide primarily in their classes and the students’ reaction followed by feedback, and the combination of corrective feedback and learner uptake leading to negotiation of form. To perform this study, an observational, analytical and descriptive study was conducted. For collecting data, six classes with 6 different instructors were chosen. The number of participants was 60 female students who were at intermediate level from two subsidiaries of Jahad Language Institutes in Karaj, Albourz Province. Homogeneous groups of language learners were selected. Each class was observed for 5 sessions and the interactions among students and instructors in different classes were recorded. The coding scheme was according to Lyster and Ranta’ (1997) model with some additional parts. Two other types of feedback were added, translation and multiple feedback. Also a combination of errors, multiple errors, was added. The analysis of the database showed that among five types of errors, i.e. phonological, grammatical, lexical, multiple errors and L1, the phonological and grammatical errors were committed primarily by students (43% and 30% respectively). From eight types of feedback given to learners, explicit feedback and recast were the most frequent types of feedback provided by the instructors. Finally, four types of feedbacks including elicitation, clarification request, metalinguistic feedback and repetition of errors led to student uptake: self repair and peer correction.


Definition of Corrective Feedback
Corrective feedback is defined as the case when "negative or positive evidence" to errors are provided to help learners repair the erroneous form based on linguistic correctness and precision (Lyster & Ranta, 1997;Suzuki, 2004).Chaudron (1977) defines it as "any reaction of the teacher which clearly transforms, disapprovingly refers to, or demand improvement of the learner utterance" (p.31).It is described by Lightbown & Spada (2003) "as any indication to a learner that his/ her use of the target language is incorrect" (p.172), it is classified into two categories based on the way they are corrected, explicitly and implicitly.Ellis (1994) maintained that the terms "correction" "repair" and "feedback" are often used to refer to general area of error treatment.Long (1978) distinguished two terms of feedback and correction; he notes that feedback refers to the process of giving students information so that they can tell whether their production or comprehension of the language is correct, while correction refers to the result of feedback or its effect on learning.Long (1996) provided a more comprehensible view of feedback and mentioned that the provided input for the learner can be divided to two broad categories of positive and negative evidence.Positive evidence is defined as providing the model of the target language which is acceptable and grammatical and it is divided to two subcategories of authentic and modified input and model, whereas negative evidence is providing direct and indirect information about what is unacceptable in target language.

Uptake
Uptake is defined by Lyster & Ranta (1997) as "a student's utterance that immediately follows the teacher's feedback and that constitutes a reaction in some way to the teacher's intention to draw attention to some aspects of the student's initial utterance" (p.49).Carroll & Swain (1993) stated that uptake provides an opportunity to learners to practice what they have learned and fill the gap in their interlanguage.Panova & Lyster (2002) believed that the notion of uptake helps the researchers recognize different degrees of the learners' participation while they are corrected.Smith (2005) shortened the definitions of uptake mentioned by Ellis et al. (2001a) as follows: 1. Uptake is student move.
2. The move is optional.
3. The uptake move occurs in episodes where learners have a demonstrated gap in their knowledge.
4. The uptake move occurs as a reaction to some preceding move in which another participant either explicitly or implicitly provides information about a linguistic features.(p. 407-432) 1.2.1 The Benefits of Uptake Uptake has been interpreted to function differently on the part of some researchers.It serves as a predator of general performance of examinee on the test (Loewen, 2005); it could result in a focus on the learners' output (Lightbown, 1998); it contributes to fluency (Swain, 1995); and it provokes the reanalysis and change of non-target form in learners' production while they form new hypothesis and try to test them (Lyster, 1998a).1.2.2Successful and Unsuccessful Uptake Ellis et al. (2001b) made a difference between two uptakes, successful & unsuccessful.He defined successful uptake as a type in which learners try to show their potential to challenge the information offered, for example, by attempting to paraphrase instructor's information or by trying to utilize the information correctly in their production.This is opposite to the unsuccessful uptake in a way that learners try to appreciate the instructor's information or simply repeat what the instructor had mentioned.Some scholars believe that this kind of reaction can be considered as uptake since they are a reaction to the instructors' utterance, but in the perspective of some others, they are unsuccessful uptake due to not contributing to an analysis of information by the learners.

Advantages and Disadvantages of Recast
Several studies have been done to indicate the merits and effectiveness of recast (Long, 2006;Saxton, 2005): a) recasts appear where the negotiators participate in a "joint intentional focus" in some meaning-based communication(Long2006; p.114); b)recast as a type of corrective feedback contributes to learner's attention and encouragement; c)recast is understood by the learners, therefore it gives more information to the learners, leading to an understanding of "form-function mapping" (Doughty,2001); d) recast does not hinder communication since it has a reactive nature.Hence compared with explicit corrective feedback, it proved to be widely used and more effective.
However, some other researchers notify some demerits in opposition to recast: a) it is believed by some scholars as an ineffective type of corrective feedback and rarely facilitate in target language development; besides it is left unnoticed by the learners (Lyster,1998a;Panova & Lyster, 2002); b) another issue raised against recast is due to its ambiguous nature since it might be considered as reparaphrasing of the learners' utterance (Long, 2006;Lyster & Ranta, 1997;Morris & Tarone, 2003;Nicholas et al, 2001); and finally c) recast rarely leads to repair since in recast, the learners without being given a chance to modify their output, are provided with the correct form of language.(Loewen and Philip, 2006).
1.4 Purpuse And Research Questions Lyster & Ranta (1997) studied the relationship between different corrective feedbacks and learner uptakes.Their studies gave a systematic picture of student-teacher interactional moves, including the types of feedback provided for different types of errors and the types of feedback leading to uptake.Their studies indicated that as far as the learner uptake was concerned, spite the high frequency of recast, it was the least effective type of teacher feedback.Lyster & Ranta's study was carried out with young learners sitting at elementary level in French immersion classroom.Hence, it seems that there is a need to investigate a study in EFL context with English learners sitting at intermediate level to examine if the results confirm the Lyeter & Ranta' study.Furthermore, in the present study the instruction is a mixture of both meaning-based and form-based; whereas, in Lyster & Ranta's was meaning centered.Therefore, the present study aims to shed light on the answers to the following research questions.3: What combination of corrective feedback and learner uptake can lead to negotiation of form?

Participants and Setting
This study was conducted in 6 classes with six different teachers.There were 60 students (about 8-12 students in each class).The participants were female adults aged 23-29 sitting at intermediate level in two branches of Jahad Language Institutes in Karaj.All students did not have any experience of being in target language environments either for a short time or a long time.The learners were studying English for two reasons, to be able to cope with their daily needs at work and to succeed at their university subject matters for their higher education.To assess the participants' level of proficiency, PET was administered to 85 learners at the beginning of the study.Before administering the test to the major group, the test was first piloted in a smaller group of students, consisting of 34 students whose proficiency level was similar to that of the main participants of the study.The reliability of the objective parts of the proficiency test was estimated through KR-21 formula which was 0.87.The teachers were selected based on their willingness to cooperate in this study.All instructors were foreign language learners and their mother language was Persian.All had either BA or MA degrees from state universities in Iran with a score of 7 or upper in IELTS exam.

Instrumentation
In this study, the interaction between teacher and students was recorded by means of a high-quality recorder.Then all interaction was transcribed for the purpose of data analysis.Students in each class were interviewed by the researchers to indicate their attitudes toward the way they were corrected.The researchers also got some information about each student's L1, background knowledge of English, the aim of learning English, and having the experience of spending time in target language environment or not.It is worth mentioning that this intimate interview between the researchers and students was conducted at the end of the term to avoid any impact on students' interaction in the class.

Procedure
Having made sure of the homogeneity of the participants, the researchers observed and recorded about 45 hours of six teachers' classes in two branches of Jahad language Institutes for 6 weeks.First, the interaction between teacher and students was recorded by a high-quality recorder.Second, the recorded voices were transcribed.Third, all students' errors were identified and classified into different types.Fourth, all types of teachers' feedback following learners' errors were identified and their effects on students' learning (uptake) were examined.Finally, the researchers analyzed the data.

Collecting Data
A modified version of Lyster& Ranta's (1997) model was utilized for coding data; in this model (Fig. 1), the process starts when a learner commits at least an error followed by either teacher's corrective feedback or topic continuation.In the case of providing feedback from teacher, it could be either followed by uptake or topic continuation.Learners' non target utterance is either repaired or remained as a needs repair utterance.It is worth mentioning that two categories of feedback types, including, translation and multiple feedback, and one category to error types, namely, multiple error were added to Lyster & Ranta's category.
The Process of Error Management (Adopted from "Corrective Feedback and Learner Uptake: Negotiation of Form in Communicative Classrooms," By Lyster and Ranta, 1997).
Figure 1.This model was modified by the researchers in two parts, Learner Errors and Corrective Feedback

Results
After analyzing the data, five types of error including phonological, grammatical, lexical, multiple errors, and unsolicited use of L1 were recognized.Figure 2 presents the percentage of each type of errors committed by students.Among these 5 types of errors, phonological errors were committed by students mostly and unsolicited use of L1 was the least one (43% and 6% respectively).This graph also gives a percentage of those errors which were not provided feedbacks since instructors did not want to stop their students' speech.As the graph illustrates, 38% of errors received no feedback.Explicit correction and recast were two most frequent feedback types used by the instructors (20 and 16 % respectively).In comparison to other types of feedback, metalinguistic feedback and elicitations are two feedback types that occurred the least (both 2%).

Figure The Percentage of Different Types of Feedback
Figure 4 summaries the findings of this study in terms of students' total number of errors, teachers' provision of feedback, uptake moves-repair and needs repair.In this study, 1064 error episodes were found by the researchers in which 401 (38%) error cases were left without feedback.663 (62%) students' errors were provided by 7 types of feedback and a combination of feedbacks (multiple feedback).After the students were provided different types of feedback, they either paid attention to teachers' feedback (uptake) or they did not (no uptake).From 663 teachers' feedback, 167 (25%) of feedbacks remained without uptake.This graph shows that approximately 40% of students' errors did not receive feedback by the instructors.The reasons for this ignorance or not giving feedback may be due to some factors such as topic continuation, not interrupting students' flow of speech, and motivating students to continue talking.From among 663 numbers of feedbacks offered, 496 had uptake, whereas 167 of teachers' feedbacks were not paid attention to by students.The following Graph (Figure 5) presents a better picture of uptake.The whole percentage is shown in terms of no uptake and uptake, repair and needs repair.The first column graph indicates those feedbacks remained with no uptake (25%).The second and third column graphs revealed the percentage of uptake divided into two categories, repair and needs repair.75% of teachers' feedback led to uptake, 40% repair and 35% needs repair.Table1, illustrates the frequency of repair, needs repair, and no repair.As it was mentioned before, 25% of the teachers' feedbacks led to no feedback which was different from Lyster & Ranta's (1997) study in that 0.45% of feedback provided by teachers in their study was left without uptake.This showed that in this study, students were more motivated to respond to teachers' feedback either in repair or needs repair form.(32%).The least feedback provided for lexical were multiple feedback and metalinguistic.(0 and 2.5% respectively).Multiple error and unsolicited use of L1 received mostly recast.Both error types received no repetition feedback (0%).

Discussion
Analyzing the first research question and data obtained from figure 2, five types of errors were observed including phonological (43%), Grammatical (30%), lexical (13%), multiple errors (8%) and unsolicited use of L1 (6%).The phonological and grammatical errors were observed to be the two most prevalent types of errors.The frequency of occurrence of L1 and gender errors was low in this study.The low rate of gender error might be due to proficiency level of student.Another possible reason might be attributed to the feedback type the participants received since the errors were mostly corrected through peers and self repair and not through teacher feedback.
Analyzing the frequency and percentage of eight types of feedback showed that explicit feedback was the most frequent type of feedback (32%), and metalinguistic feedback as the least frequent feedback type (3%).The findings of the study were not in parallel with the results of Lyster and Ranta' (1997) study in that recast was recognized as the most frequent type of feedback (55%) and the least was repetition(5%).However in this study, recast was the second most frequent type of feedback (26%).Also it is worth mentioning that the result of this study was inconsistent with Sheen's (2004) study since he reported recast as the most frequent type of feedback.The findings of this study showed that recasts led to the lowest rate of uptake which is in parallel with Sheen's (2004) result.The researcher believed that the proportion of recasts contributing to uptake and repair could be influenced by the context in which they are used.In other words, recasts may lead to uptake in contexts where the focus is primarily on recast and through the use of reduced or partial recasts students' attention are directed towards linguistic form rather than meaning.Moreover, the classroom observation showed that some teachers were reluctant to use clarification, elicitation, and other types since they thought these feedback types were time consuming and needed more patience.In the current study and Lyster and Ranta's study, elicitation, metalinguistic feedback, clarification request led to the greatest amount of uptake by providing the students the opportunity to self correct.Recast, despite its high frequency, led to the lowest amount of uptake (16%).Probably this could be attributed to the ambiguity of recast (Lyster, 1998b).In other words, recast might be confusing to the learners and they might be confused whether the instructor was correcting the error or repeating the correct form or rephrasing their utterance.Another reason could be their proficiency level since some researchers emphasized that the effectiveness of the recast would be increased at advance levels (Doughty & Varela, 1998).In some cases, it was observed that the students understood the intended aim of the teacher's provision of recast but not utter anything.It is worth mentioning that in some cases, instructors took the opportunity from students for reaction to recast.
Regarding the third research question, the result indicated that four types of feedback including elicitation, clarification request, metalinguistic feedback and repetition of error led to student-generated repair (self-repair and peer correction).The findings were consistent with that of Lyster and Ranta's (1997) study.In deed, similar to their study, the current study's data revealed that the feedback-uptake sequence helped the learners engage more actively when the teachers do not provide the correct form to the learners, as recasts and explicit correction do, students attempted to reformulate their erroneous utterance more actively.This might mean that these four types of feedback are potentially more useful in helping learners notice their linguistic inadequacies.Due to the importance of giving feedback and the little information provided to instructors on the issue of error correction, the result of this study could be helpful for instructors since most of them are unaware of the beneficial effects of different types of feedback.
Since in this study the frequency of repair (self and peer repair) was low in spite of their high effectiveness in learning a language, some teachers took this valuable opportunity from their students to correct their errors by themselves or their classmates, so it is highly recommended to instructors before correcting students' errors, give a chance to correct themselves or their classmates.Even after the individual student and class have failed to provide self-correction, it is still not recommended to give student the correct form.The instructor can repeat the incorrect utterance and, by pausing immediately before or after the error, highlight it in the hope that there will be sufficient help to encourage a student to produce the correct answer.
Although recast is considered as the second most frequent type of feedback in this research, the effectiveness of this feedback type in leading to uptake is low, so instructors are recommended to use this type of feedback less than others or to combine it with other feedbacks; for example with elicitation: S: I am agree with Sarah about this problem of society.(Error-grammar) T: Really, do you agree with Sara about cultural problems of family?Sarah I am agree or I agree?(Feedback-Recast and Elicitation) S: Sorry sir.I agree.(Uptake-repair-self) In the above example, the instructor mixed two implicit and explicit feedback types to raise the effectiveness of feedback and inform the student of her error.The current study was carried out at intermediate level; therefore the next study could be replicated at elementary or advance level.
The present study could be done by the instructors teaching to children to explore which type of feedback has the most effect on children in different ages, gender, and language proficiency.In this study, the classifications of errors was based on Lyster & Ranta's category.The other studies could be done with other categories of errors e.g.errors related to stress, intonation, register, omissions and appropriacy.Classroom observation indicated that some instructors used other way(s) of correcting such as facial expression, delay error correction, …., so the future studies could consider more types of feedback.Since this study was carried out among male and female instructor, it seems that female students were more comfortable, motivated, and confident to negotiate with female instructors; they had more tendency to receive feedback especially explicit feedback from the instructors with the same gender.The next studies could be done to examine the effect of gender on receiving feedback in EFL context as a moderator variable.
The current study was carried out to investigate the immediate effect of 8 feedback types; therefore the long term effect of each feedback type could be some decent topics for further research.And finally, further studies could be carried out to investigate the relationship between different types of corrective feedbacks given by instructors and the learners' level of competency and proficiency.
1: what are the different types of errors committed by EFL learners and what types of errors are the most prevalent?2: What types of corrective feedback do EFL teachers use mostly in their classes and what is the students' reaction to them?

Figure 2 .
Figure 2. The Percentage of each Error TypeFigure3, gives us a general view about the percentages of different feedback types given to students while committing errors, including recast, explicit correction, clarification request, metalinguistic feedback, elicitation, repetition, translation and multiple feedback.This graph also gives a percentage of those errors which were not provided feedbacks since instructors did not want to stop their students' speech.As the graph illustrates, 38% of errors received no feedback.Explicit correction and recast were two most frequent feedback types used by the instructors (20 and 16 % respectively).In comparison to other types of feedback, metalinguistic feedback and elicitations are two feedback types that occurred the least (both 2%).

Figure 4 .
Figure 4.The Frequency of Occurrence of Errors, Feedbacks, and Uptakes

Figure 5 .
Figure 5.The Percentage of No uptake, Repair and Needs repair

Table 2
, shows that in Lyster and Ranta' study, the total percentage of repair was 0.27% while in the current study it was 0.40% showing the higher rate of learning.Based om Lyster and Ranta's findings, the participants were not interested in replying to 69% of recasts and 50% of explicit feedback and for other types of corrective feedback including repetition (22%), metalinguistic feedback (14%), and clarification request (12%).It is worth mentioning that Lyster & Ranta did not consider two types of feedback; namely, translation and multiple feedback-due to the scarcity of these two types of feedback.

Table 1 .
The frequency and percentage of uptake in relation to corrective feedback type

Table 2 .
Lyster and Ranta (1997)tive feedback inLyster and Ranta (1997) 3.2 Analyzing Error Types Receiving FeedbackTable3presents the percentages of error types receiving feedback types.The data revealed that phonological and grammatical errors were mostly provided by explicit and recast.Lexical errors received mostly explicit feedback

Table 3 .
The frequency and percentage of error types leading to feedback