Revisiting the Role of Indirect Written Corrective Feedback in the Light of Written Languaging

For years, one of the formidable challenges of EFL/ESL writing teachers has been deciding on the optimal way of responding to learners' linguistic errors. In fact, a heated debate has revolved around the utility of written corrective feedback (WCF) in improving writing proficiency in general and grammatical accuracy in particular. Truscott (1996) raised doubt on the potential and purported effects of grammar correction on improving writing accuracy. Truscott argued that provision of WCF takes up the simplistic stance that language learning is the mere transmission of information from teachers to students; thus, the complex nature of the second language acquisition is oversimplified. He, further, maintained that error correction supports pseudo-learning which does not permit learners to sharpen their understanding, and consequently they grow shallow knowledge and understanding. Nonetheless, over the recent years, research on WCF has yielded some cogent evidence in favor of the effect of WCF on enhancing the learning of a limited number of linguistic structures, what is referred to as focused corrective feedback (e.g., Bitchener, 2008; Bitchener & Knoch, 2009; Sheen, 2007) and on all sorts of errors, what is recognized as unfocused CF, (e.g., Van Beuningen, De Jong, & Kuiken, 2008, 2012). Therefore, the focus of WCF research has shifted towards determining the optimal ways of offering WCF in order to ensure learners' active engagement and further development.

Of the different types of WCF, direct (i.e., indicating the location of errors and giving the correct forms) and indirect (i.e., showing the location of errors by underlining or coding) have attracted SLA researchers' attention. The distinctive feature of direct and indirect WCF is the level of engagement which they elicit from the learners. It is postulated that indirect WCF involves learners in a kind of problem-solving and guided learning activity (Bitchener & Knoch, 2008; Feriss, 2002, 2006). Because direct WCF reduces the amount of processing and engagement on the part of learners, they can easily copy the provided forms rather than pondering and reflecting upon them. Nonetheless, incorporation of direct WCF is supported on the grounds that it provides learners with opportunities through which they are exposed to correct forms; thus, the learners can compare and contrast their own produced forms with the WCF; in essence, learners' initial hypotheses are tested out and possibly reshaped and restructured (Bichener & Knoch, 2010).

One of the theoretical backbones underlying WCF is noticing. Noticing has grabbed overriding attention among SLA researchers as an integral condition for learning language (Ellis, 1995; Schmidt, 1990, 2001; Swain, 1995; Swain & Lapkin, 1995). Schmidt (1990) provoked discussion into the necessity of noticing and awareness for acquiring knowledge. Schmidt (2001) expounds his stance further and notes that "specific attention to linguistic form is the first step toward grammar change" (p. 101). Additionally, the prime function of giving corrective feedback (CF) to learners' output is to make them aware of the nature of the earmarked parts and to orient their heed to notice disparities in their developing interlanguage.

Drawing on the concept of output hypothesis, Swain (1985, 1995) posited that the significance of the output lies in the fact that reflecting upon output augments learners' current knowledge of language further than semantic processing of language. In essence, it initiates the process of awareness of discrepancies between learners' interlanguage system and target language, what Ellis (1995) called cognitive comparison. That is to say, it involves learners in syntactic processing of language (Swain, 1995). Consequently, CF should be offered in an efficient way that learners can optimize their performances by extending the condition to process the offered feedback not only at the level of noticing or perfunctory noticing but also at the level of understanding, that is, substantive noticing (Qi & Lapkin, 2001, p. 291). Qi and Lapkin (2001) postulated that CF implicates engaging learners actively with language to process CF unless it failed to lead to development. Likewise, the level of involvement and contribution has a potential influence on language development (Storch, 2008). In this regard, the role of conscious attention in L2 acquisition is effective in assisting learners to achieve self-regulatory potentials (i.e., control or regulate one's task independently) of the learning process. In line with the same argument, languaging (i.e., learners' verbalizations & reflections) is perceived as one of the ways to extend learners' contributions to perceive feedback in order to assist them to be aware of the reason behind the targeted error (Suzuki, 2012; Swain, 2006). Therefore, languaging brings greater readiness to notice gaps in L2 learners' knowledge so that it possibly helps them to put errors right and carry out this understanding to the subsequent learning context; in other words, languaging aids to gain self-regulation. Rooted in a sociocultural perspective, this study delved into whether written languaging prompted by indirect WCF could contribute to raise learners' attention to refine the degree of understanding about the linguistic errors. Thus, this study set out to explore the level of learners' engagement with offered indirect WCF and written languaging by showing numbers and types of written languaging episodes.  In addition, it made an attempt to investigate the effect of mere indirect WCF and indirect WCF followed by written languaging on language errors in EFL learners’ compositions.

2. Literature Review

2.1 Languaging and Written languaging

The contribution of output to second language learning has been recognized by Swain (1995, 1998, 2000, 2005, & 2006). She spelled out three functions of output: the hypothesis testing function, the noticing function, and the metalinguistic (reflective) function. Firstly, output provides opportunities for learners to test hypotheses on the target language grammar. With respect to corrective feedback, learners should be able to evaluate their production with the received feedback and modify what they are going to convey. Secondly, output serves to help learners to notice gaps and holes in their interlanguage system. Finally, the third role of output which contributes to language learning is metalinguistic (reflective) function, that is, "reflecting on the language produced by others or the self-mediates second language learning" (Swain, 2005, p. 478). While learners are speaking and writing, they process their output production for making links between the forms and the functions; therefore, this function of output enables them to become aware of language they are using. Moreover, Swain and Lapkin (1995) postulated that the noticing function can lead to conscious awareness if learners are prompted to engage in linguistic problems and processing them. Swain (2006) underlined the importance of the metalinguistic output in speaking and writing and termed it as languaging. Languaging is defined as "producing language in an attempt to understand – to problem-solve – to make meaning" which can effectively influence the second language learning outcomes (p. 96). In effect, languaging considerably boosts the noticing and metalinguistic functions.

Languaging originates from the mindset of Vygotsky's sociocultural theory of mind which is defined as "an action- a dynamic, never-ending process of using language to make meaning" (Swain, 2006, p. 96). The sociocultural theory explains that learning, first, occurs with the help of the surrounding environment and capable individuals to deal with difficulties; afterward, tasks can be accomplished independently (Lantolf & Thorne, 2006). The progression from the object or other regulation to self-regulation is called internalization (Lantolf & Thorne, 2006). One of the most significant tools in making meaning and internalization is language (Swain, 2006). Languaging or using language actively involves EFL learners in perceiving and producing linguistic entities. 

From a Vygotskian sociocultural theory of mind, L2 learners can employ language as a cognitive tool to regulate their thinking and to mediate their mental activities (Lantolf & Thorne, 2006; Negueruela, 2008; Swain, 2006, 2011). With regard to this functioning of language, languaging can assist L2 learners to gain and develop an insight towards a phenomenon through the language they are using. Languaging is a broad construct which takes the forms of collaborative dialogue and private speech/writing. On the one hand, collaborative dialogue is a dialogue with others in an attempt to solve a problem or to acquire and enhance knowledge about a contradiction which speakers are engaging in (Donato, 1994;
Swain, 2000). In fact, it shows the social mediation. On the other hand, private speech/writing is self-directed talking with/writing to the self (Dicamilla & Lantolf, 1994; Ohta, 2001) which paints a picture of self-mediation. Languaging was put forward by Swain (2006) in order to cast more light on the role of using language not only as a means of communication but also as an aid to mediate and shape cognition. Following Swain's series of studies (2006, 2010, 2011), Suzuki (2012) proposed that written languaging as an external memory can assist L2 learners to demonstrate an understanding of a linguistic entity they are dealing with by expressing their own deliberations and considerations; hence, they can broaden their insights towards this linguistic phenomenon.

A number of studies have suggested that oral languaging is a source of L2 leaning (e.g., Negueruela, 2008; Qi & Lapkin, 2001; Storch & Wigglesworth, 2010; Swain, Lapkin, Knouzi, Suzuki, & Brooks, 2009); however, very few studies have targeted the efficiency of private writing or written languaging in the terrain of L2 learning (Dicamilla & Lantolf, 1994; Moradian, Miri, & HosseinNasab, 2017; Suzuki & Itagaki, 2009; Suzuki, 2009, 2012). Therefore, this study aimed at enriching our understanding of the merits of written languaging in the L2 learning context. This investigation attempted to explore the utility of written languaging in reply to indirect WCF in development of grammatical accuracy of EFL learners' compositions from a sociocultural perspective.

2.2 Sociocultural Theory

Sociocultural theory (SCT), which dwells on Vygotsky's proposal, brings a more robust mindset on second language development than other second language approaches. Vygotsky's writing (1978) provides a fresh and broader perspective towards the function of language, learning and development, and the role of learners. Grounded in a Vygotskian SCT of the mind, all mental activities are mediated and directed through physical and symbolic tools; moreover, among other mediational tools, language functions as a symbolic tool which can usher learners to regulate their thoughts in order to internalize the activities scaffolded. Studies within this groundwork differ from L2 research venues, and a host of studies are grounded in the qualitative nature which provides an insight into the learning process and a rich description of learners' involvement in the process of learning. Moreover, it highlights the importance of how L2 learners blossom and progress through offering feedback phases when they are scaffolded. The pioneering investigation in SCT is Aljaafreh and Lantolf's study (1994). Their participants received one-to-one feedback from their tutor on their writings. They read and checked through their own writings for the committed errors; however, the tutor did not intervene this process unless they could not rectify their errors. Aljaafreh and Lantolf noted that learners' actual level can be unfolded through negotiation between the tutor and learners, and the tutor underwent a series of prompts from the most implicit to the most explicit level until learnersʼ appropriate level was determined. The assistance was graduated, firstly, by indicating the location of errors and, eventually, by giving explanation about the nature of the error, and moving towards the implicit strategy indicated learning and achieving self-regulation. Aljaafreh and Lantolf's study showed that learners' developmental trajectories are different and the tutor’s gradual help provided an opportunity for learners to acquire their self-regulatory potential in order to accomplish the similar subsequent task autonomously.

Following Aljaafreh and Lantolf's study, Nassaji and Swain (2000) compared the usefulness of feedback on article use within a learner's ZPD and a learner who received random feedback. They concluded that negotiated assistance in the learner's ZPD is more effective than non-negotiated error correction. In another study, Storch and Wigglesworth (2010) addressed two types of WCF: reformulation (direct WCF) and error coding (indirect WCF). Their learners collaboratively wrote a text based on a graphic prompt. In Session Two, they received their texts with the pertinent feedback types and rewrote the texts in pairs. All learners' interaction during the review and rewriting phase was audio-recorded. In Session Three, learners wrote a text individually on the same prompt in Session One. They found that learners in the indirect group produced more language-related episodes, and the level of engagement appeared more extensive for the indirect CF group; additionally, learners in the direct group generated more lexicon than form-based language-related episodes. Concerning writing accuracy, both groups significantly revealed gains in their writing accuracy in the revised texts. Indeed, Storch and Wigglesworth (2010) confirmed that:

learners (particularly adult learners) [are] intentional agents in their language learning   activity who assign relevance and significance to certain events and whose behavior is guided by their own goals (Lantolf & Pavlenko, 2001; Lantolf & Thorne, 2006). These beliefs and goals may affect what learners notice, whether they accept or reject the feedback provided, and how much of the feedback they retain. (p. 306)         

More specifically relevant to the present study, Suzuki (2009, 2012) and Moradian et al., (2017) conducted particular investigations of more relevance to WCF within this outlook. In these studies, L2 learners were scaffolded by writing down their own deliberations and reflections upon the highlighted parts by the instructors. The results of their studies showed that involving EFL learners in written languaging could assist them to get deep insights towards the highlighted parts and mediate their next performance on their writings.  In essence, research within SCT can unfold the processes learners undergo when they perceive certain types of feedback.

2.3 Research Questions

1. What types of written languaging episodes (correctly resolved, incorrectly resolved, unresolved) prompted in response to indirect written corrective feedback do L2 learners produce more from the first to the last writing task?
2.  To what extent does the provision of indirect WCF accompanied by written languaging and mere indirect WCF affect the grammatical accuracy of the compositions of the Iranian EFL learners at the intermediate level of proficiency?

3. Methodology

3.1 Participants

The participants were EFL learners at the intermediate level of language proficiency in the context of a private language institute in Khoramabad, Iran. They were aged between 15 and 17. They were homogenous based on their formative and summative scores registered in the preceding semester when the study was conducted. Furthermore, the results of the Quick Oxford Placement Test carried out by the institute confirmed the homogeneity of the EFL learners. They attended the class three sessions a week, each one lasted 90 minutes. All female participants had received two years of instruction at this institute. The institute syllabus was aimed at developing the EFL learners' communicative competence with instruction being offered in listening, speaking, reading, and writing skills. It is worthy to point out that the intact classes conducted by one of the researchers. The teacher researcher had more than ten years of experience in Teaching English as a Foreign Language (TEFL).

3.2 Procedure

All data were collected over a period of ten sessions. Five episodes from Intermediate Steps to Understanding (Hill, 1981) were singled out by the teacher. In the first session, participants listened to a passage read by the teachers two times. The teacher read the passage at a normal speed and the learners listened without taking any notes. Subsequently, the learners were invited to take note to its content as they were listening to the passage read by the teacher with a slight pause between the sentences. Finally, they were requested to reconstruct the passage individually based on their notes. They were given 20 minutes to regenerate the text including at least 100 words. In the following session, the learners received their compositions in which the teacher researcher had offered indirect WCF, underlining, to all linguistic errors. Then, learners in the written languaging group were asked to language about their errors by writing down their own explanations and deliberations on the underlined parts earmarked by the teacher, which were dubbed as written languaging episodes. Consequently, these written languaging episodes were classified as correctly resolved, incorrectly resolved, and don't know.It took the learners on average 16 minutes to perform the written languaging task. Nonetheless, the indirect WCF group were asked to review and number their grammatical errors in their writings. At the end of the session, learners' compositions and written languaging sheets were collected for later analysis. These steps were followed for four next writings. It is worthy to mention that offering indirect WCF and coding the written languaging sheets were done by the teacher researcher.

3.3 Writing Tasks

Five dictogloss tasks were designed for the study. The writing tasks were singled out from Intermediate Steps to Understanding (Hill, 1981). These writing tasks were selected for three reasons. First, the tasks were based on the level of EFL learners' language proficiency. Then, all dictogloss tasks were related to the same genre of writing. Third, participants were assisted through listening to the same episode so as to write a composition. 

3.4 Scoring Procedure

All linguistic errors in EFL learners writings were coded and categorized by the teacher researcher according to numbers of errors and types of written languaging episodes (correctly resolved, incorrectly resolved, unresolved). Furthermore, following Suzuki (2012), a normalized error score was calculated for each composition. In fact, the normalized score was calculated by dividing the number of errors by the number of the words in a composition multiplied by the average number of the words in each group. The normalized error scores were calculated for the first and the last writings to be compared. However, for the types of the written languaging episodes, all EFL learners written languaging sheets were analyzed to classify them based on correctly resolved, incorrectly resolved, and do not know written languaging episodes. It is worthy of mentioning that the indirect WCF and assessing all WLEs sheets were provided by the teacher researcher. 

4. Results

The first research question was addressed to examine the type and the number of written languaging episodes. Participants in the written Languaging group were requested to write down their explanations for the indirect WCF which were provided by the teacher. As shown in Table 1, in the first writing, participants could correctly recognize the reasons behind their errors on 85 (47.80%); additionally, the number and percentage of incorrectly resolved WLEs and do not know WLEs were 32(17.58%) and 63 (34.61%), respectively. In the second writing, participants managed to give correctly resolved WLEs on 55 (47.82%), incorrectly resolved WLEs on 21(18.26%), and do not know WLEs on 39(33.91%). Furthermore, in the third writing, the number of correctly resolved WLEs was 48(50%); moreover, 19 (19.79%) and 29 (30.20%) were produced for incorrectly resolved WLEs and do not know WLEs, respectively. In the fourth writing, the number and percentage of correctly resolved WLEs were 38 (48.71%), and these values for incorrectly resolved WLEs and don't know WLEs were 15 (19.23%) and 25 (32.05%), respectively. in the last writing, participants produced 30 (50%), 13 (21.66%), and 20(28.34%) for correctly resolved WLEs, incorrectly resolved WLEs, and don’t know WLEs, respectively. As depicted in Table 1, the number of errors which were demanded to generate WLEs was decreased and the percentage for the types of WLEs was the same from the first writing to the last writing. In effect, participants could manage to produce, firstly, a higher number of correctly resolved WLEs. Then, they could generate a higher percentage for unresolved WLEs than incorrectly resolved WLEs.

Table 1

The Frequency (The Percentage) and the Type of Written Languaging Episodes (WLEs)

The second research question aimed at investigating the effect of indirect WCF accompanied by written languaging and indirect WCF on improving the grammatical accuracy of dictogloss composition tasks. The normalized scores of post-tests of these two groups were compared. As depicted in Table 2, the Mann-Whitney test showed that there was a statistically significant difference between the scores of the indirect WCF followed by written languaging and the mere indirect WCF in fostering the grammatical accuracy of their compositions on the post-test, z = -2.76, p = .016, with a large effect size r = .40.The significance level was smaller than 0.05 so, it was found that the indirect WCF accompanied by the written languaging group outperformed more significantly than the mere indirect WCF group on the posttest. Therefore, the difference between the two sets of scores was significant.

Table 2

Inferential Statistics for Comparing Post-tests of Indirect WCF+ Indirect WCF plus Written Languaging

5. Discussion and Conclusions

The study explored the effect of mere indirect WCF and indirect WCF followed by written languaging on language errors of EFL learners' compositions. Moreover, it sought to examine numbers and types of WLEs produced by the EFL learners. The findings of the study showed that the participants in the indirect WCF followed by written languaging produced the highest percent of correctly resolved WLEs during all five writing tasks. Furthermore, EFL participants managed to generate a higher percent of do not know WLEs than incorrectly WLEs. The findings also corroborated that indirect WCF followed by written languaging was more effective than mere indirect WCF in the accuracy of EFL learners' compositions.

Results of the study indicated that written languaging could involve EFL learners in a deeper level of understanding, so the level of involvement for participants in the languaging group was more extensive than those in the indirect WCF group. The results were in line with Storch's (2008) and Storch and Wigglesworth’s (2010) studies which revealed that the extensive level of involvement with WCF extended the learners' contributions to process WCF. It was also found that limited engagement with feedback like the mere indirect WCF in the current study could not scaffold EEL participants enough to gain understanding towards their language errors in their compositions. Qi and Lapkin (2001) stated that when learners are engaged in perfunctory noticing, they could not attend substantively to reasons behind the feedback so as to be aware of the rationale behind it. Therefore, the improvement in participants’ writings through written languaging can attribute to the opportunities which were open to them to deliberate and then written language on language errors in their compositions. Indeed, languaging could assist EFL learners to mediate and regulate their learning. Hence, languaging can be considered as a means to extensively and attentively broaden EFL learners' contributions to process feedback(e.g., Negueruela, 2008;  Moradian et al., 2017; Storch & Wigglesworth, 2010; Suzuki, 2009; 2012; Swain et al., 2009). In general the level of noticing and awareness for both groups in this study was different, so the written languaging group gained an edge over the mere indirect WCF.

Several implications for L2 pedagogy should be considered based on the findings of the study. In order to promote EFL learners’ active and mindful engagement in their language production, they should be invited to reflect and language about their output. Likewise, teachers would invite EFL learners to keep diaries, journals, and written records to deliberate about their linguistic problems. Thus, learners may be provided with more time for reflection and deliberation to glean insights into the language and the language learning process they are involved in. These written records may be of critical and valuable points for both EFL teachers and learners.