ARE RECASTS EFFECTIVE ACROSS THE BOARD?

In attributing the source of “doubt” (p. 135) regarding recast effectiveness solely to Lyster and colleagues, Goo and Mackey disregard concerns expressed previously by several applied linguists. We trace the first signs of “skepticism” (p. 136) to Corder (Reference Corder1967), who wrote, “simple provision of the correct form may not always be the only, or indeed the most effective form of error correction since it bars the way to the learner testing alternative hypotheses” (p. 168). Similarly, in his seminal review of the literature on error correction, Hendrickson (Reference Hendrickson1978) stated that the procedure whereby teachers provide students with correct forms “is ineffective when helping students learn from their mistakes” (p. 393). On the basis of data from an early classroom observation study, Chaudron (Reference Chaudron1977) concluded that recasting—what he called “repetition with change” (but without emphasis or reduction)—was “especially weak in helping to locate the error” (p. 41).

Goo and Mackey refer to their 2007 meta-analysis as “evidence that recasts work” (p. 134). What they do not mention, however, is that the effect sizes were significantly larger for lab studies than for classroom studies. Nor do they mention having qualified their results as follows: “We cannot say they [recasts] were more or less helpful than other sorts of feedback due to the insufficient number of studies.” Furthermore, they called for “more studies in order to obtain a more clear-cut picture of the effectiveness of different types of feedback” (Mackey & Goo, Reference Mackey, Goo and Mackey2007, p. 442). It would then seem that Goo and Mackey are now retracting their call for comparative studies, and as such we are left to wonder if this is because those that followed did not necessarily provide unwavering support for recasts. For example, Lyster and Saito’s (Reference Lyster and Saito2010) subsequent meta-analysis, which examined the effects of different types of CF specifically in second language (L2) classroom settings, found medium effect sizes for recasts and large effect sizes for prompts, which proved significantly more effective in the within-group contrasts than recasts.

Several reviews of recast studies are available that put recast effectiveness into perspective through reference to a range of linguistic, pragmatic, cognitive, and contextual constraints (e.g., Ellis & Sheen, Reference Ellis and Sheen2006; Nicholas, Lightbown, & Spada, Reference Nicholas, Lightbown and Spada2001; Sato, Reference Sato2011; Sheen, Reference Sheen2011). For example, the positive effects for recasts appear to be reserved for learners in form-oriented settings (Nicholas et al., Reference Nicholas, Lightbown and Spada2001) and for those with high literacy levels (Bigelow, delMas, Hansen, & Tarone, Reference Bigelow, delMas, Hansen and Tarone2006), developmental readiness (Mackey & Philp, Reference Mackey and Philp1998), high working memory capacity (Mackey, Philp, Egi, Fujii, & Tatsumi, Reference Mackey, Philp, Egi, Fujii, Tatsumi and Robinson2002), and high phonological memory, attention control, and analytic ability (Trofimovich, Ammar, & Gatbonton, Reference Trofimovich, Ammar, Gatbonton and Mackey2007). Goo and Mackey acknowledge these and other constraints but then ignore them as they claim instead that research has substantiated “the across-the-board utility of recasts” (p. 135).

For the most part, research demonstrating the effectiveness of recasts has been conducted in laboratory settings where CF can be delivered intensively in consistent ways on specific linguistic targets. These laboratory studies have generally yielded positive results for recasts, but many have methodological problems—such as no control group (e.g., Ishida, Reference Ishida2004) or a control group receiving no CF (e.g., Han, Reference Han2002; McDonough & Mackey, Reference McDonough and Mackey2006; Sagarra Reference Sagarra and Mackey2007)—as well as other limitations, as we will point out, akin to those that Goo and Mackey associate only with the alleged “case against recasts.”

Whereas the lab studies to which Goo and Mackey appeal for support fall short of substantiating their claim for across-the-board benefits of recasts, the classroom studies to which they refer (i.e., Doughty & Varela, Reference Doughty, Varela, Doughty and Williams1998; Goo, Reference Goo2012; Loewen & Nabei, Reference Loewen, Nabei and Mackey2007) provide even less support, as we explain next. First, in Doughty and Varela’s classroom study comparing CF to no CF, recasts were used solely as secondary moves in the event that the primary move, which was a prompt that repeated verbatim the learner’s error, failed to elicit self-repair. Thus, the fact that Goo and Mackey refer to this study as evidence of recast effectiveness is misleading. Also worrying is their lack of criticism of this study for using multiple feedback types and not controlling for modified output, a criticism they seem to apply only to studies that appear unsupportive of recast effectiveness. Second, Goo and Mackey classify Loewen and Nabei (Reference Loewen, Nabei and Mackey2007) as a classroom study, but it was not conducted in intact classrooms. Instead, it was conducted outside of the classroom, with a researcher who interacted with small groups of four learners, and so it may be better classified as a lab study. As with other lab studies comparing different types of CF (e.g., Lyster & Izquierdo, Reference Lyster and Izquierdo2009; McDonough, Reference McDonough and Mackey2007), Loewen and Nabei found no difference across CF types.

Third, the treatment in Goo’s (Reference Goo2012) study was a form-oriented task requiring learners to ask the teacher a question in order to complete two fill-in-the-blank exercises, each with 15 blanks. Whereas the 14 participants in the recast group repeatedly heard positive exemplars of the target form each time an error was made, the 32 participants in the metalinguistic group heard a rule, such as, “Don’t use the conjunction ‘that’ when you ask about the subject of the subordinate clause” (p. 455), without opportunities for self-repair. We have never heard of rule repetition as a potentially effective type of CF and find it surprising that the recast group was unable to outperform the metalinguistic group. Moreover, that both groups outperformed the control group does not provide any support for recast effectiveness because the control group did not participate in the same exercises and so was unable to benefit from similar exposure to the target form. A control group in a CF comparison study needs to participate in the same treatment tasks, but without CF, as was the case in Loewen and Nabei (Reference Loewen, Nabei and Mackey2007), Lyster (Reference Lyster2004), and Yang and Lyster (Reference Yang and Lyster2010), so that all groups have similar exposure to the target forms, with and without CF (Lyster, Saito, & Sato, Reference Lyster, Saito and Sato2013).

INSTANCES OF UPTAKE ARE NOT INSTANCES OF LEARNING

Goo and Mackey portray our earlier work on uptake in a particularly misleading way and fail to acknowledge that Lyster’s program of research has continued to evolve since the publication of Lyster and Ranta (Reference Lyster and Ranta1997). In that seminal work, we stated upfront that, because we examined only immediate uptake, “claims related to language learning remain speculative and subject to further empirical investigation” (p. 57), and that our results needed to “be validated by means of coding sets of classroom data from other contexts” (p. 56). This call for further classroom research led to several studies that indeed revealed discrepancies in rates of uptake following recasts across instructional settings, most notably Ellis, Basturkmen, and Loewen (Reference Ellis, Basturkmen and Loewen2001), Sheen (Reference Sheen2004), and Lyster and Mori (Reference Lyster and Mori2006), as well as other studies that also revealed low rates of repair: in English immersion classrooms in Korea (Lee, Reference Lee2007) and English as a foreign language (EFL) classrooms in Hong Kong secondary schools (Tsang, Reference Tsang2004). Goo and Mackey’s reference to studies of dyadic interaction (Braidi, Reference Braidi2002; Oliver, Reference Oliver1995) as counterevidence to our findings is unconvincing, especially when we consider Braidi’s (Reference Braidi2002) conclusion that, even in a laboratory setting, it was impossible to determine whether learner responses to recasts addressed form or meaning, as in, “‘Yes, I recognize that that is the correct form’ versus ‘Yes, that is what I meant to say,’” (p. 31) and Oliver’s (Reference Oliver2000) subsequent observation that, in contrast to dyadic interaction, “the nature of whole class interactions diminishes the opportunity for students to respond to the feedback” (p. 126). The reference to Ohta’s (Reference Ohta, Hall and Verplaeste2000) classroom study as evidence “that L2 learners do in fact produce uptake” (Goo & Mackey, this issue, p. 137) also seems underwhelming, given that it showed only that four adult learners, who produced private speech with moderate to high frequency while wearing lapel microphones in a form-oriented foreign language classroom, tended to repeat recasts. These references—taken as evidence to suggest that the rates of uptake observed by Lyster and Ranta in French immersion classrooms were somehow wrong—completely miss the point that we now know that rates of uptake vary from context to context.

As we have always maintained, instances of uptake are not instances of learning (Lyster, Reference Lyster1998, Reference Lyster2002, Reference Lyster2007; Lyster & Ranta, Reference Lyster and Ranta1997). Instead, uptake refers to a range of possible responses made by students following CF. Goo and Mackey follow Long (Reference Long2007) in misrepresenting Lyster’s program of research by attributing to it an “emphasis on uptake” and the “implication that uptake is a measure of learning” (p. 137). For example, they refer disapprovingly to Panova and Lyster’s (Reference Panova and Lyster2002) suggestion that “if recasts and translations are essentially corrective in purpose, there is little evidence that L2 learners in the present study processed them as such” (p. 591), which was based on Panova and Lyster’s observation of minimal uptake after recasts and translations. Yet the observational data in Panova and Lyster, similar to those in Lyster and Ranta, indeed suggested that learners might not be processing the corrective function of recasts. That responses to CF may be a sign of noticing was claimed by others as well, including Mackey herself, who suggested that “uptake may be related to learners’ perceptions about feedback at the time of feedback” (Mackey, Gass, & McDonough, Reference Mackey, Gass and McDonough2000, p. 492). Concerning the value of uptake following recasts, however, Panova and Lyster noted that “uptake consisting of a repetition may not have much to contribute to L2 development, because of its redundancy in an error treatment sequence where the repair is both initiated and completed by the teacher within a single move” (p. 579).

Goo and Mackey claim in their article to have “considered the issue of uptake and its meaning” (p. 146), yet they do not explain uptake as a discourse phenomenon, and they fail to make an important distinction between (a) utterances still in need of repair or (b) utterances with repair. Utterances in need of repair include simple acknowledgements such as yes, hesitations, off-target responses, partial repair, and occurrences of either the same or a different error, whereas utterances with repair entail the correct reformulation of an error and thus differ from modified output, which may or may not be a correct reformulation. Learner repair can either be a repetition or self-repair, and these different types of immediate repair are specific to one type of CF or another: Recasts and explicit correction can lead only to repetition of correct forms by students, whereas prompts can lead either to self-repair or peer repair, not to repetition. Lyster and Ranta (Reference Lyster and Ranta1997) suggested that “student-generated repairs in the error treatment sequence may be important in L2 learning” (p. 57). Thus, it was this line of research that Lyster pursued, predicting that different types CF would have differential effects on learning insofar as different types of repair (e.g., self-repair after prompts vs. repetition after recasts) entail different processing mechanisms (see Lyster, Reference Lyster1998, Reference Lyster2002, Reference Lyster2004).

This research has shown benefits for both recasts and prompts, with some benefits for prompts over recasts, but not to the extent predicted. For example, in the case of young immersion students, recasts were as effective as prompts in oral production measures and less effective than prompts only in written production measures (Lyster, Reference Lyster2004). Recasts were as effective as prompts for young English as a second language learners with high pretest scores but less effective than prompts for learners with low pretest scores (Ammar & Spada, Reference Ammar and Spada2006). Adult EFL students in China benefited equally from recasts and prompts in improving the accuracy of irregular past tense forms but benefited more from prompts than recasts in improving their accurate use of regular forms (Yang & Lyster, Reference Yang and Lyster2010). With adult Japanese learners of English, recasts were more effective than no CF for improving pronunciation of familiar lexical items but not for unfamiliar items (Saito & Lyster, Reference Saito and Lyster2012). In the case of Lyster and Izquierdo’s (Reference Lyster and Izquierdo2009) study, which Goo and Mackey group with other classroom studies even though the feedback treatments were delivered individually by a researcher in the lab, all participants benefited equally from recasts and prompts. To construe these results as categorically against recasts appears reductionist, and this dichotomous argumentation has led them to gloss over important nuances that have emerged from comparative CF studies in classrooms, which we consider worthy of further pursuit.

If there were classroom researchers who were actually against recasts, they would be inclined to conceal the benefits of recasts revealed by their research and to proscribe recasts in their conclusions. In contrast, Lyster and Ranta (Reference Lyster and Ranta1997) concluded simply that “teachers might want to consider the whole range of techniques they have at their disposal rather than relying so extensively on recasts” (p. 56). Lyster and colleagues have followed suit throughout the years and have encouraged teachers to use a variety of CF types (Lyster, Reference Lyster2002, Reference Lyster2004, Reference Lyster2007; Lyster & Mori, Reference Lyster and Mori2006; Lyster & Saito, Reference Lyster and Saito2010; Lyster et al., Reference Lyster, Saito and Sato2013; Panova & Lyster, Reference Panova and Lyster2002). This recommendation is based on the premise that recasting what students already know is unlikely to be the most effective strategy to ensure continued L2 development and that prompting learners to draw on what they have not yet acquired is equally ineffective.

To depict Lyster as responsible for the alleged case against recasts, Goo and Mackey rely on the following quote: “The analysis of recasts as used by the four immersion teachers in the present study leads to the suggestion that, in studies investigating the effects of negative evidence on classroom SLA, recasts themselves may be red herrings” (Lyster, Reference Lyster1998, p. 74). Yet, by “red herrings,” Lyster was specifically referring to the ways in which the then-existing classroom research had operationalized recasts as very explicit forms of CF. For example, Doughty and Varela (Reference Doughty, Varela, Doughty and Williams1998) operationalized implicit negative feedback as a repetition of the error with rising intonation and additional stress to emphasize the error, followed by a recast only if the learner failed to self-repair. Additionally, the teacher provided learners in the recast condition with written feedback on the target forms, and, as students viewed their videorecorded presentations, she paused when errors occurred and asked the class to repeat the correct forms simultaneously. Similarly, in Tomasello and Herron’s (Reference Tomasello and Herron1988, Reference Tomasello and Herron1989) studies, teachers provided recasts that were, as Long (Reference Long, Ritchie and Bhatia1996) argued, “delivered in the form of explicit correction” (p. 441), wherein a teacher first wrote the incorrect form on the chalkboard, crossed it out, and then wrote the correct form while saying it out loud (i.e., the recast), before providing a brief explanation. Lyster’s concern was that classroom research, claiming at that time to investigate recasts, was really investigating something other than implicit negative evidence. Goo and Mackey appear to have dissociated this concern from its original context to suggest that a case against recasts exists.

MODIFIED OUTPUT OPPORTUNITIES

Goo and Mackey argue that, in comparative studies of different types of CF, learners in prompt conditions “could be seen to be in a relatively advantageous position, all other things being equal” because they have opportunities for modified output, whereas those in recast conditions “receive feedback that rarely leads to modified output” (p. 148). First, this assertion contradicts their argument disputing “the [alleged] claim that recasts do not lead to uptake” (p. 137). Second, if recasts are as effective as they are described to be by Goo and Mackey (owing to their semantic transparency, temporal immediacy, saliency, and opportunities for comparison: see p. 130), then one may argue that learners receiving recasts are, in fact, at an advantage because they consistently receive positive evidence, whereas learners in the prompt groups do not. In any case, the purpose of comparisons is indeed to ascertain which treatment is more advantageous than others. Comparing identical rather than different types of CF would lead us back to nebulous comparisons of recasts and negotiation (see Mackey & Philp, Reference Mackey and Philp1998). Prompts and recasts can be compared precisely because they are different.

On the one hand, Goo and Mackey question the role of uptake in L2 development, yet they acknowledge, on the other hand, that modified output (learner repair in our terms) can “lead learners to focus on the linguistic form in which the feedback was provided” (p. 148). According to this view, learner repair trumps all the advantages of recasts, and, therefore, recasts and prompts should not be compared because learners receiving prompts would have an advantage, enabling them to perform better than those receiving recasts. Thus, although Goo and Mackey extoll recasts for making “relatively few participatory demands” (p. 148) on learners, they also appear to believe that the greater participatory demands placed on learners by prompts are more beneficial. We would argue that good teaching should indeed place increased participatory demands on students, and one way to do so is by using CF types that push them to modify their output.

Goo and Mackey suggest that the ideal way to compare the effects of different types of CF would be “to carefully control production of modified output following corrective feedback—regardless of the type of feedback—such that learners produce it or are prompted to move on without producing it” (p. 148). They refer to Goo’s (Reference Goo2012) classroom study to illustrate a context in which students were prevented from modifying their output after recasts and metalinguistic feedback alike. This was pragmatically appropriate in his study because there was no purposeful goal in the form-focused treatment task, which required university-level education students to orally produce a question to fill in a blank. Rather than fill-in-the-blank exercises, other classroom studies have tended to use more communicatively oriented tasks designed to create contexts of interaction. In such contexts, we do not recommend that students be either prevented from modifying their output or required to produce it, but we do suggest that instances of modified output be better documented than has been the case in experimental classroom studies. This would enable classroom researchers to include modified output as a variable in their interpretations (rather than terminating all comparison studies, as Goo and Mackey suggest). The extent to which learners modify their output after different types of CF in classrooms is an empirical question that can be answered by analyses of video recordings of treatment sessions (e.g., Yang & Lyster, Reference Yang and Lyster2010) or audio recordings of language-related episodes (e.g., Loewen, Reference Loewen2004).

In their lab study, Lyster and Izquierdo (Reference Lyster and Izquierdo2009) found recasts and prompts to be equally effective, even though modified output was allowed after prompts but not after recasts. The alleged advantage of prompts over recasts—owing to greater output opportunities—has not yet been substantiated by studies predicting such an advantage. Lyster and Izquierdo concluded that, although both types of CF proved equally effective, the types of processing that led to their respective effectiveness were very different; that is, learners receiving recasts can benefit from repeated exposure to positive exemplars as well as from opportunities to infer negative evidence, whereas learners receiving prompts can benefit from repeated exposure to negative evidence as well as from opportunities to practice using the target form as they modify their output. As Lyster et al. (Reference Lyster, Saito and Sato2013) and others have pointed out (e.g., Loewen & Philp, Reference Loewen and Philp2006), positive learning outcomes can be associated with self-repair following prompts but not necessarily with learner repetition following recasts, the effectiveness of which may instead be associated with specific input features such as length and prosody. These are important avenues to continue to explore in classroom contexts because, contrary to Goo and Mackey’s claims, they have not yet been settled insofar as the benefits of modified output associated with prompts do not necessarily trump the benefits associated with recasts.

SINGLE-VERSUS-MULTIPLE COMPARISONS

Goo and Mackey’s argument against comparing single CF moves with multiple CF moves is an important one but is again misleading when applied to classroom studies, such as those by Ammar and Spada (Reference Ammar and Spada2006), Lyster (Reference Lyster2004), and Yang and Lyster (Reference Yang and Lyster2010), in which learners in the prompt groups received a greater variety of CF types but not a greater amount of CF. In these studies, prompts were considered a single CF strategy, which withholds correct reformulations but has multiple manifestations (e.g., clarification requests, repetition of learner error, elicitation, metalinguistic clues). Because prompts include more than one way of providing negative evidence while withholding positive evidence, we agree that it may be their variety (but not greater frequency) that adds to their effectiveness. If it were feasible and desirable for a teacher to provide only one type of prompt repeatedly, then this refinement may add more rigor to the comparisons and would help tease apart the effects of prompts considered implicit (e.g., clarification requests) from those considered more explicit. We hasten to add, however, that we consider Goo’s (Reference Goo2012) operationalization of prompts as a single rule explanation repeated after every target error to be questionable in terms of pedagogical appropriateness. Variety in the use of prompts may be considered more pedagogically sound and also methodologically valid in comparative studies when we consider that recasts, too, come in many shapes and sizes (e.g., Loewen & Philp, Reference Loewen and Philp2006; Sheen, Reference Sheen2006). Mackey and Goo (Reference Mackey, Goo and Mackey2007) themselves acknowledged that recasts are “increasingly recognized as being elastic in nature,” although they considered them, as many researchers do, “as a monolithic construct” (p. 413). Of interest to pursue in this regard is that variety of CF—rather than quantity—may play a major role in its effectiveness (see Lyster et al., Reference Lyster, Saito and Sato2013).

Goo and Mackey’s criticism of single-versus-multiple comparisons seems only to apply, however, if the comparison did not come out in favor of recasts. For example, they refer to Mackey and Philp’s (Reference Mackey and Philp1998) study to support recast effectiveness, but this study compared recasts with negotiated interaction, which contains both clarification requests and confirmation checks; negotiated interaction thus combined input-providing and output-pushing types of CF and also overlapped with recasts, many of which, in their own data, proved to “function as confirmation checks” and thus to occur as “part of negotiation sequences” (Mackey & Philp, Reference Mackey and Philp1998, p. 342; see also Loewen & Philp, Reference Loewen and Philp2006; Lyster, Reference Lyster1998; Sheen & Ellis, Reference Sheen, Ellis and Hinkel2011).

Similar support for recast effectiveness is attributed to the Long, Inagaki, and Ortega (Reference Long, Inagaki and Ortega1998) lab study that compared recasts with models. Learners in both the recast and model treatments were equipped with headphones as they played a communication game in dyads with a researcher. In the recast condition, they were told in their first language (L1) to perform a specific task. As they performed the task, they had to translate the directive to describe their action in their L2 to the researcher, who then recast their utterance. In the model condition, learners heard a prerecorded utterance in the L2 and then repeated the utterance for the researcher, who responded by performing the action. Comparing the effects on learners of receiving either (a) a recast after translating from the L1 to the L2 or (b) no CF after repeating a model does not seem to us to be a particularly robust and valid way of assessing recast effectiveness, given the additional processing involved in the recast condition (i.e., listen to the L1 utterance, translate into the L2, and receive recast) versus the model condition (i.e., listen to the L2 utterance and repeat it). Nonetheless, this study, too, was exempt from Goo and Mackey’s criticism of design issues that affect comparability.

FORM-FOCUSED INSTRUCTION

Although theoretical justifications abound for isolating the effects of CF from instruction, Goo and Mackey’s suggestion that CF must be provided in the absence of instruction even in classroom research shows a lack of understanding of instructional practices and also research designs. Without equivalent instructional activities across all groups, including a control group, studies cannot tease apart CF effects from task effects, as in Goo’s (Reference Goo2012) study comparing two CF groups, who completed an instructional task while receiving either recasts or rules, with a control group, who did not engage in the same task. Learners in the control group missed equivalent exposure to positive exemplars of the target form, which, Goo acknowledged, was a “rare” and “limited” form unless targeted by instruction (p. 453).

In classroom settings, we believe that CF is best provided strategically with form-focused instruction and that there are ways to experimentally tease apart the effects of CF from those of form-focused instruction. For example, in the classroom studies by Ammar and Spada (Reference Ammar and Spada2006), Lyster (Reference Lyster2004), and Yang and Lyster (Reference Yang and Lyster2010), the CF and no-CF groups alike were all exposed to the same instructional activities, which included focused production tasks that provided practice opportunities in all groups and contexts for CF provision in the treatment groups. To integrate these treatment tasks into immersion students’ regular curriculum, Lyster’s (Reference Lyster2004) study also included various noticing and awareness tasks. Engaging all groups in these tasks served to level the playing field for comparisons between CF and no-CF groups and also to ensure that those receiving CF were at a similar level in their knowledge of the target feature. As Li (Reference Li2010) argued, because the no-CF group, which engaged in the same instructional activities, served as a control group, any differences detected among the recast, prompt, and no-CF groups “must be due to the presence or absence of feedback” (p. 319).

We believe that the design of these CF comparison studies is a strength because we are interested in how CF fits into the bigger picture of classroom instruction. In contrast, Goo and Mackey are less interested in the role of CF in instructional practices and more interested in isolating the effects of CF during interaction in lab settings to control for prior knowledge. Advocating a range of research approaches to understanding the complex role and variable effects of different types of CF would seem more fruitful than advocating one program of research associated with one specific theoretical orientation. In our more recent writing on CF (Ranta & Lyster, Reference Ranta, Lyster and DeKeyser2007), we have drawn more from skill acquisition theory than from the interaction hypothesis to explain CF effectiveness because skill acquisition theory explicitly acknowledges a role for CF within an instructional sequence that includes language practice (see also Lyster & Sato, in press). The purpose of CF in educational settings is for learners not only to notice target exemplars in the input but also to consolidate emergent L2 knowledge and skills through practice.

PRIOR KNOWLEDGE

Goo and Mackey criticize a set of classroom studies for not controlling for prior knowledge. In so doing, they imply that research in support of the positive effects for recasts has consistently controlled for prior knowledge, which is not the case. They appear to have been influenced by Long (Reference Long2007), who claimed that, with regard to the purpose of CF, “acquisition of new knowledge is the major goal, not ‘automatizing’ the retrieval of existing knowledge” (p. 102). In our view, the ultimate goal of instruction is not to continuously present only new knowledge to students without sufficiently providing subsequent opportunities for assimilation and consolidation of that knowledge; rather, students need repeated opportunities to retrieve and restructure their knowledge to become fluent and accurate users of the target language. Goo and Mackey miss this point as they criticize the Yang and Lyster (Reference Yang and Lyster2010) study for its interactional activities designed to elicit target features because they “might have evoked learners’ prior knowledge of the targets and have provided practice opportunities for learners in the control group even without corrective feedback” (p. 155). The study was with Chinese EFL learners at the postsecondary level (i.e., English majors), who, of course, had had plenty of prior knowledge of past tense forms—as Yang and Lyster acknowledged—but who might have lacked communicative practice in using them. That learners had prior knowledge of the target forms does nothing to detract from the results showing (a) positive effects for learners engaging in practice activities without CF, (b) more positive results for those receiving recasts, and (c) even better results for those receiving prompts (depending on the linguistic structure).

Goo and Mackey’s recommendation that CF studies should “examine L2 targets to which learners have never been exposed” but “without providing instruction” (p. 153) is inadvisable not only from an educational perspective but also from a research perspective. We see little point in providing CF on forms to which learners have not yet been exposed, mainly because it remains unclear how a learner with zero knowledge of a form could attempt to produce it during meaningful interaction in a way that would elicit CF from the teacher or interlocutor (see Lyster et al., Reference Lyster, Saito and Sato2013).

HOW SHOULD CF RESEARCH BE CONDUCTED?

CF RESEARCH AND L2 TEACHING

To push the apples-and-oranges metaphor a bit further, we would argue that teachers are concerned with the “nutritional needs of learners” (Schachter, Reference Schachter, Clarke and Handscombe1983) and thus need to know when apples are a better choice over oranges. According to Schachter, one nutritional need of learners is negative input (i.e., evidence). For language teachers, the framing of this issue is in terms of how they should respond to learners’ errors in speaking. Surveys of teachers in different contexts have shown that teachers feel that CF is important (Schulz, Reference Schulz2001) and that teachers prefer to recast learners’ errors (Bell, Reference Bell2005). This is possibly due to the influence of teacher-education literature and courses in which the potential negative affective impact of explicit forms of CF is often stressed (e.g., Harmer, Reference Harmer2007). In our own pre- and in-service teacher-education courses and professional development workshops, we emphasize the need for the use of a variety of CF moves (Lyster, Reference Lyster2007; Ranta, in press), as do Ellis (Reference Ellis2009) and Goo and Mackey (this issue, p. 132). Importantly, our advice to teachers is not based on an opinion or belief but, rather, on empirical evidence from CF comparison studies carried out in classrooms whose overall characteristics we are able to describe. There is reason to believe that teachers are likely to be influenced by evidence that comes from teaching and learning settings comparable to their own. Hence it is necessary for classroom research to be conducted in a wide variety of settings rather than considering that the issue has been settled.