A Review of Grammar Corrective Feedback: The Learning Experience in China’s English Classes

Grammar correction is a common means of instruction in second language classes. However, whether or not to conduct grammar correction is a controversial issue that has triggered researchers’ debates. In China, grammar correction can always be seen in English writing and oral classes. This paper reviews the fundamental theory of grammar correction and discusses the relative merits of grammar correction in the English teaching practice of China by analysing the author’s English learning experience. All in all, grammar correction is recommended as an important teaching method in China’s English classes for learners can benefit from the corrective feedback with proper instructing strategies. The correction with improper strategies will evoke learners’ negative feelings, but they may still extract useful information from it if the feedback itself is effective. Thus, teachers should not overlook or underestimate learners’ ability of self-reflection and autonomic learning during the process of grammar correction. Besides, it is of great concern for teachers to comfort learners’ emotion to ensure the effectiveness of grammar corrective feedback.


Introduction
Feedback is a crucial section belonging to the second language teaching and learning, which will influence the motivation and linguistic accuracy of learners (Ellis, 2009b). During the process of second language acquisition, Chinese learners of English may inevitably make grammar errors or mistakes in their speaking and writing practice. The corrective feedback received from teachers will affect their language learning outcomes. However, grammar correction is a controversial topic among researchers. Truscott, the leader of the opponents, insists (1996,1999) grammar correction is ineffective and even harmful for language learning in speaking and writing. The advocators believe www.scholink.org/ojs/index.php/selt Studies in English Language Teaching Vol. 8, No. 3, 2020 128 Published by SCHOLINK INC. grammar correction is "pragmatically feasible, potentially effective and, in some cases, necessary" (Lyster, Lightbown, & Spada, 1999, p. 457). But Thornbury (1999) points out the current research tends to support the view that ignoring grammar errors and mistakes may put learners' linguistic development at risk. This paper seeks to analyse the effects of grammar correction in the English classes at different stages based on the relevant theories and the author's learning experience of English as a second language. It will introduce the basic concepts and controversies of grammar correction to reflect its advantages and disadvantages in China's English classes from the perspective of second language learners.

Definition of Grammar Correction
Corrective feedback means "the responses to learner utterances containing an error" (Ellis, 2006, p. 28), which is usually involved in two productive language activities: oral speaking and writing. Being opposite to the positive feedback which provides affective supports to learners and motivates their learning, corrective feedback belongs to the negative feedback that reveals the linguistic incorrectness in learners' utterance (Ellis, 2009b). When the error or mistake is related to grammatical issues, corrective feedback is known as grammar correction which refers to teachers' behaviour that provides students with corrections for and comments on their grammar problems (Lee, 2004, p. 286). In this paper, the terms "corrective feedback" and "grammar correction" are interchangeable.

Errors in Grammar Correction
Before conducting grammar correction, the first step is to distinguish the differences between error and mistake. According to the opinion of Corder (1967), errors indicate the evidence of the language system that the learner is using, which can help the teacher know about the learning effect and strategies of the learner and can work as a device that the learner uses to learn. While a mistake cannot make contributions to the language learning process. Ellis believes (2009b) the former results from the lack of knowledge or the gap in competencies; the latter is caused by the shortage of automaticity or the limitation of memory and suggests (1993) the corrective feedback should focus on the salient grammatical features that confuse learners.
There are different ways to categorise errors. Burt (1975) divides them into the global errors, which will "significantly hinder communication" and "affect overall sentence organization", and the local errors, which will "not usually hinder communication significantly" and "affect single elements in a sentence" (pp. 56-57). Thornbury (1999) generally sorts out them into the lexical errors, grammar errors and discourse errors in writing as well as the pronunciation errors in speaking and believes the systematic errors, which "show evidence of a rule being fairly systematically applied" (p. 115), are more suitable for the correction than random errors. Thornbury (1999)  developmental errors. The transfer errors are caused by the influence of learner's first language on their second language. For instance, in written Chinese, the third-person pronoun uses the character "他" to refer to he/him and uses the character "她" to refer to she/her. The different character component, "亻" or "女", represents the different gender. But in spoken Chinese, "他" and "她" share the same pronunciation in the Chinese phonetic system: tā. So, such a mother-tongue habit makes the Chinese learners tend to misuse he/him to refer to both man and woman when they are speaking English, though they understand the principle of using he/him and she/her in English. This example shows the negative transfer from L1 to L2, which is described as the L1 interference. However, the developmental errors derive from "the nature of the second language itself" (pp. 114-115). One typical example is the overgeneralising that may lead to the over-applying of some certain language rules. Some beginning Chinese learners of English will pluralise all nouns by directly putting the letter "s" behind them. Also, Thornbury points the developmental errors will be committed by children when they are learning their mother tongues.

Selection of Errors
Errors are diverse. Teachers are supposed to concentrate on some error types rather than all errors learners make because teachers can design suitable assessing instruments to check and correct the errors according to their predetermined error types (Ellis, 2009b;Harmer, 2007;Penny, 1996). Thus, it is necessary to determine the correction priority of errors. Some errors will cause serious misunderstandings and ambiguity. Others just like a slip of the tongue that will not influence learner's aggregate performance. Thornbury (1999) considers intelligibility as one of the criteria for selecting the prior errors which will check "to what extent does the error interfere with or distort the speaker's or writer's message" (p. 115). Meanwhile, teachers also need to respond to the errors which are out of the range of the predetermined errors in an intuitive way.

Strategies of Grammar Correction
There are various strategies of grammar correction that can be employed in the English classes: (1) Self-correction: teachers are suggested to guide students to self-correct themselves (Hedge, 2000).
For students, they can conduct self-correction with their existing linguistic knowledge. In other words, they can correct their mistakes rather than errors by themselves (Ellis, 2009b). Ferris (2004) asserts the result of L2 writing research shows it is more likely for students obtaining corrective feedbacks to self-correct themselves than those without feedback, which demonstrates the correction uptake can promote students' linguistic competence in the long term.
(2) Explicit correction: if the self-correction fails, teachers can offer a clear correction to the errors caused by their students (Ellis, 2009b), which means teachers can offer the correction to students by pointing out the errors directly without preparation. This is a common strategy employed in the ELT classes of China. Students can immediately receive and understand the feedback about their errors, but they may feel frustrated and embarrassed because some teachers' inappropriate expressions in the correction will be regarded as criticism. (3) Peer-correction: when the self-correction cannot work, teachers can also ask other students to provide suggestions so that students can correct each other (Thornbury, 1999). From the perspective of my learning experience, the support from peers can release the stress of the student who is expected to respond to his or her errors in the oral speaking activities and can help the students strengthen the impression on varieties of writing errors in the pair and group works.
(4) Repetition: teachers can underline the errors committed by students by repeating their errors in the utterance with emphatic stress (Ellis, 2009). Compared with the explicit correction, it may bring less stress to students. But students may misunderstand the teacher's intention of doing so (Thornbury, 1999). The uncertainty will make students feel hesitated to say something in oral activities.
(5) Elicitation: teachers can repeat the partial utterance of students and hint students to complete the erroneous part that has been left with a rising intonation (Ellis, 2009b). This strategy may be more intelligible than repetition for students to figure out what they are going to do in line with the teacher's prompts.
(6) Clarification request: teachers can suggest students that their messages are unclear or distorted.
This comparably friendlier strategy will evoke students to re-cast their messages and improve their performance (Thornbury, 1999). Teachers' suggestion will lead students to correct and reorganise their utterance with clear motivation.
(7) Paralinguistic signal: teachers can indicate there is an error or mistake by using a gesture or facial expression to prompt students (Ellis, 2009b). Body or facial languages can offer vivid clues to guide students to notice the errors and enlighten them to come up with ideas to correct themselves.
(8) Impromptu reactive teaching: teachers can make use of student's errors to deliver some impromptu teaching materials. The instructions are not predetermined and will respond to the errors to be corrected in a student-centred environment. But students may feel reluctant to express themselves (Thornbury, 1999). Thus, teachers are supposed to explain the impromptu instruction as clear as possible.
(9) Recast/reformulation: teachers can expand or re-formulate student's incorrect utterance. This strategy will foster students' improvement by providing a temporary scaffold to them, though young students may be unable to realise the difference between teacher's reformulation and their utterances (Ellis, 2009b;Thornbury, 1999). This strategy is popular among Chinese English language teachers.
Teachers will change and correct students' utterance, but excessive correction should be avoided to prevent teachers from distorting the meaning that students intend to convey. A typical example will be illustrated a little bit later.

The Categorisation of Grammar Correction
According to the categorisation of the corrective feedback strategies neatened by Ellis (2009), two pairs of standards can be used to classify the above strategic items: (Ⅰ) the explicit and implicit correction feedback proposed by Carroll and Swain (1993) and (Ⅱ) the input-providing and output-prompting correction feedback admitted by Lyster (2004) and Ellis (2006). According to the first standard, the implicit correction feedback consists of the strategy of self-correction, peer-correction, repetition and clarification request, while the rest strategic items belong to the explicit correction feedback. Ellis, Basturkmen and Loewen (2001) conclude from the research that the implicit and explicit correction feedback can assist language acquisition and the latter is generally more effective than the former. Concerning the second standard, except for the recast and explicit correction, the rest items belong to the output-prompting correction feedback, which is preferable because "they enabled learners to increase control over linguistic forms that they had partially acquired" (Ellis, 2009b, pp. 33-34) and are more effective than the input-prompting strategies in the oral field (Lyster, 2004). As far as I am concerned, the impromptu reactive teaching can be put into any above-mentioned taxonomy and its nature will be finally determined in line with the specific needs of teachers and students. For instance, if the teacher wants to exemplify the difference between hoodie and cardigan, he or she can take the student who wears the corresponding clothes as an example to show the natures of these two nouns to the class and guide the students to correct the inappropriate words they used. In this case, such impromptu teaching can be classified into implicit correction and input-based instruction.

Differences between Written and Oral Correction
Timing is another problem that needs to be taken into consideration by grammar correction. There are two kinds of corrections that are conducted at a different time: (Ⅰ) immediate correction and (Ⅱ) delayed correction. Ellis (2009b) notes the general timing agreement of the oral corrective feedback advises that the correction should be left at the end of the fluency activity but errors in the accuracy activity should be corrected immediately. Besides, the writing corrective feedback will always be provided after teachers have collected and checked all scripts.
Apart from timing, there also exist some other differences between oral corrective and writing feedback. Sheen (2010) indicates "written correction imposes less of a cognitive load on memory than oral CF (corrective feedback)" (p. 176) because learners have to make use of their short-term memory as much as possible during the oral correction. Additionally, the written correction needs to process more information, such as linguistic competence, discourse competence, coherence, mechanics and so forth, than the oral correction that simply focuses on learners' errors in communicative activities (Polio, 2001).

Attitude and Controversy
As previously mentioned, different researchers hold divergent perspectives on grammar correction that has sparked continuing controversy. Typically, in the writing section, Truscott (1996) claims that grammar correction in L2 writing classes should be abandoned because plentiful studies have manifested it is ineffective, useless and even harmful. In the oral section, Truscott (1999) asserts maintaining the practice of grammar correction is not beneficial because the oral correction is unable to promote learners to speak grammatically. Some of his essential viewpoints are concluded as followed: (1) Grammar correction cannot promote the development of learners' writing and will disrupt communicative activities in speaking, for it will impose pressure on students and drain their motivation in writing and provoke students' negative emotions in communication, such as embarrassment, fear, anger, discouragement and so forth.
(2) Grammar problems are too difficult for teachers to recognise, identify and explain the errors if they do not have a good understanding of the correct use and grammar theories. Students may have difficulty noticing, understanding, remembering, generalising and applying grammar correction in writing and oral speaking.
(3) It is harmful to both teachers and students to pay attention to corrective feedback because grammar correction will divert class resources from more appropriate tasks. Thus, rejecting grammar correction can help teachers and students transfer their focus to other activities that are more productive and pleasant.
(4) The lack of concern with grammar correction will not lead to fossilisation that makes students stuck at a low level of grammar skills. Teachers don't need to provide corrective feedback to students, though students superficially think grammar correction is helpful and want to be corrected.
Based on the above opinions, Truscott concludes that grammar correction should be abandoned to avoid its negative influence on students' language learning in writing and speaking. Such an extreme statement successfully triggered other researchers' rebuttal. Ferris (1996) summarises the main problems of Truscott's research review, such as the incomparable subjects, the wide variation of paradigms and teaching strategies, the overstatement of negative pieces of evidence of grammar correction as well as the dismissal of positive research results, which means Truscott cannot completely prove the grammar correction is ineffective and harmful. Meanwhile, Ferris points out the necessity of continuing grammar correction. For example, surveys (e.g., Ferris, 1995;Hedgcock & Lefkowitz, 1994) have manifested students attach importance to grammar correction from their teachers, so students' needs of corrective feedback cannot be ignored. The absence of grammar correction will influence students' motivation, confidence and even their goal and achievement. It will also make students lose consciousness and the ability to edit their skills. Therefore, Ferris asserts students can benefit from the effective grammar correction, which is selective, prioritised and clear. Lyster et al. (1999) respond to Truscott's essay and overturn his conclusion from the aspects of feasibility and effectiveness. They believe grammar correction is "pragmatically feasible, potentially effective, and, in some cases, necessary" (p. 457), because studies (e.g., Lyster, 1998;Lyster & Ranta, 1997) have proven grammar correction is a feasible discourse move and an intrinsic part of the classroom process that neither damages students' morale nor prevents communication; students can notice and will expect corrective feedbacks; and experienced teachers can treat with the grammatical errors, performance and needs of students suitably. As to the effectiveness, Lyster et al. (1999)

Learning Experience Review and Discussion
On the whole, I am the advocator of grammar correction. The necessity and benefits of grammar correction discussed in the studies of Ferris (1995Ferris ( , 1996 and Lyster et al. (1998;1997) are reflected in my learning experience of English as a second language. But it doesn't mean Truscott's (1996Truscott's ( , 1999 views are inadvisable because the negative influence of grammar correction can reveal the problems existing in the English teaching and learning environment of China. Here, I will extract some impressive examples of grammar correction from my writing and oral speaking learning experience to analyse its advantages and disadvantages.

Example of Positive Written Correction
When I was in high school, freewriting was one of the most common English writing tasks that would ask the students to finish exercises as "writing paragraphs, essays, notices, reports, letters, diaries etc on the given topics where they are required to express their own opinions with careful planning" (Khatri, 2015, p. 20). After submitting the scripts, the students would obtain grammatical feedback from the teacher, which was following the delayed timing of written correction.
Usually, my teacher would mark all the errors I committed in the script, offer explicit corrections to each error and write comments to point out the most frequent and salient problems. The direct corrective feedback, as well as the focused and unfocused feedback, were employed in the correcting process. Ellis (2009a) indexes the directive corrective feedback can provide learners with clear guidance to correct errors, especially when they are not sure about the correct form. And several recent studies (e.g., Ellis, Sheen, Murakami, & Takashima, 2008) have proven that the more explicit the feedback is, the greater the benefit for the students to enhance their linguistic accuracy (Sheen, 2010).
The focused feedback will concentrate on several typical types of errors and is deemed to be more effective than the unfocused feedback as the former can lead learners to intensively handle a single error with multiple correction forms. On the contrary, the unfocused feedback requires learners to get touch with various errors and brings difficulty to them to reflect on every error at a time. But it can support learners in the long run (Ellis, 2009a). Maybe, the combination of focused and unfocused feedback can deal with the key errors and general errors harmoniously.
My teacher preferred to underline my errors, such as using inappropriate conjunctions and punctuations, in the scripts. And I tended to collect all the similar errors according to the teacher's correction and to summarise the rules of how to use them accurately. Such a post-writing review left a profound impression in my mind and helped me to understand the related grammatical knowledge in depth. Thus, I valued teachers' corrective feedback highly. Until now, when I need to write something, I will consciously check whether my writing complies with the proper English grammar of using conjunctions and punctuations which are common errors in a variety of writing tasks. This case shows it is necessary to conduct grammar correction in L2 writing teaching and learning because its function can promote learners' acquisition of the second language. At the meantime, learners' needs of grammar correction can never be ignored and their capacity of comprehending, self-reflecting and applying should not be underrated.

Example of Negative Written Correction
When I was in secondary school, the main English writing task used to be guided writing which would provide the students with some freedom of selecting lexical items and structural patterns based on the given writing prompts or suggestions (Khatri, 2015). Compared with freewriting, however, there was less freedom for the students to express themselves in guided writing. Such a writing condition might influence the teacher's judgement on the scripts and the standard of conducting grammar correction.
Here, I will call the involved teacher as teacher 2.
Once, the theme of our writing task was to depict the natures of four seasons with the rhetoric skill of metaphor. After we obtained the feedback from teacher 2, I found part of my classmate's writing content was ruled out with red marks by teacher 2 as my classmate likened "Autumn" to "sorrowful separation". Teacher 2 left a comment at the margin of the paper and recast my classmate's idea into "Autumn should be the joy from the harvest", in which the words "joy" and "harvest" seemed to be more consistent with the traditional lexical choices about Autumn that could meet people's common expectation.
I tend to regard such a correction as the misapplication of the reformulation due to the teacher's correction distorted the meaning that the student wanted to express and looked unnecessary and even ridiculous. According to the definition of reformulation given by Ellis (2009b), what teachers need to do are incorporating the content words of students' incorrect utterance, correcting their errors and reconstructing the utterance from the aspects of phonological, syntactic, morphological, lexical and so forth. But teacher 2 did not obey another criterion proposed by Ellis (2009a) that reformulation is supposed to keep the content of the original intact. In this case, there is no technical mistake or error in the student's writing, but teacher 2's correction is too subjective to entirely change the original meaning of the script. Similar situations may make students may feel unacceptable and angry about the results.

Example of Negative Oral Correction
There indeed exist many advantages of oral grammar correction that help learners realise their errors and foster their improvement, which has been proven by many studies (e.g., Ellis, Sheen, Murakami, & Takashima, 2008;Sheen, Wright, & Moldawa, 2009). However, if teachers do not concern about learners' feelings, oral corrective feedback may frustrate them and dampen their motivation to express more.
When I was in college, public speaking was a popular event in the English classes where students needed to prepare their speeches in line with the given topics. Since the speech paid more attention to assessing the fluency of the students' utterance, during the speech, the teacher would not interfere with the students' performance so that the completeness could be assured. The students would obtain their feedbacks after finishing the speaking tasks. The timing of correction accorded with the general agreement of putting the feedback at the end of the fluency activities (Hedge, 2000). After I finished my speech, teacher 3 ruthlessly pointed out my errors by exaggeratively imitating my mispronunciation of the word "diamond" in front of the whole class. Then, the negative emotional experience depicted by Truscott (1999) happened to me, which made me feel embarrassed, humiliated and frustrated during the process of grammar correction. However, the terrible experience strengthened my impression on the authentic pronunciation of "diamond" and helped me to avoid similar oral errors after the speech.
Therefore, I think learners' feeling of discouragement is not derived from grammar correction itself but from teachers' inappropriate attitude of offering instructions to their learners. Lyster et al. (1999) note that experienced teachers not only need to have a good sense of students' language performance and needs but also need to pay attention to individual students' personalities. In this way, students can get rid of the sense of frustration caused by grammar correction and deepen their understanding of the related grammatical principles to develop their oral expression.
In general, from my point of view as an English language learner, the positive functions do exist in grammar correction that can promote learners' acquisition of English. Thus, grammar correction should not be uncritically abandoned as Truscott (1996Truscott ( , 1999 claims. Lyster et al. (1999) indicate what the teacher can do is to provide corrective feedback to the learner, but it is the learner who makes the final decision of whether to correct the error or not. In other words, when learners think the corrective feedback is unacceptable, they have the autonomy of refusing to correct the errors or mistakes underlined by their teachers. When the corrective feedback is considered reasonable, learners may positively response to teachers' instruction, though they will feel frustrated by teachers' irritating or unpleasant expression. Or, we can say, when the corrective feedback is effective, it is the improper way of conveying the corrective feedback that will bring harmful effects, such as learners' negative emotions and an attitude of resistance, that impede their language learning rather than the corrective feedback itself.

Conclusion
This paper introduces grammar correction from the aspects of its basic concepts and controversies. It exemplifies the advantages and disadvantages of grammar correction in China's context by reviewing Chinese learners' experience of learning English as a second language in writing and speaking. Many existing studies have proven that learners can notice they are being corrected and are willing to obtain corrective feedbacks from teachers (Ellis, 2009b). Based on the analysis of the English learning experience, we can conclude that learners can benefit from grammar correction when it is effective.
Thus, it is inadvisable for teachers to arbitrarily abandon grammar correction in the language teaching process. If teachers want to obtain satisfactory teaching results from grammar correction, they need to select proper strategies according to learners' specific performance and avoid provoking learners' negative emotional response. Besides, learners' needs of grammar correction should not be ignored for they are also important participants with autonomy in the correction activities who may exert their subjective initiative to reading teachers' feedback, reflecting on their errors or mistakes and drawing lessons from the past. It is worth mentioning that, although grammar correction will sometimes make learners feel burdensome or anxious, they can still acquire useful information from teachers' effective feedback which is delivered with bad subjective attitudes. Therefore, I strongly agree with Ellis' opinion (2009b) that "teachers should monitor the extent to which corrective feedback causes anxiety in learners and should adopt the strategies they use to ensure that anxiety facilitates rather than debilitates" (p. 14). This shows teachers' subjective attitudes and means of expression will influence the effect of grammar correction, too. Future research may detect to what extent teachers' different subjective methods of conveying corrective feedbacks will influence the effectiveness of grammar correction and students' learning outcome.