Continuing with our discussion of High-Leverage Teaching Practice #6 (Providing Corrective Oral Feedback to Improve Learner Performance) in Enacting the Work of Language Instruction (2017) …
Today we will take a look at Glisan and Donato’s Research and Theory Supporting the Practice.
The authors state, “A body of research has confirmed the effectiveness of corrective feedback in supporting language learning and development …” (2017, p. 142)
I do not question that a body of research has confirmed the effectiveness of corrective feedback in supporting what was tested. My questions revolve around what was tested. At the moment I have no answers, nor do Glisan and Donato provide any.
Does corrective feedback support acquisition, i.e. the ability to use the language in spontaneous use for communication (the expression, interpretation, and negotiation of meaning)? Or does it support Monitor use, i.e. the ability to edit utterances by thinking of and applying the “rules of grammar”? What was tested? Discrete grammar items? How was it tested? Was spontaneous communication tested, or was the test conducted under conditions conducive to Monitor use?
Krashen rejects error correction as useful for acquisition and provides case study evidence of acquisition without error correction or many of the other things traditionally practiced by language teachers.. Krashen does grant error correction a small place for Monitor use. I wonder, though, if he defines error correction the same way that Glisan and Donato define it. All too often, writers use terms without clearly defining them.
VanPatten opines that corrective feedback may lead to acquisition if it is part of negotiating meaning. (We will see in a moment that he probably still disagrees with Glisan and Donato about the place of error correction.)
This raises the question of research on the different kinds of oral corrective feedback. Has it been done? What are the results? Are all types of corrective feedback equally effective? Based on what standards?
Glisan and Donato note that “… different types of feedback can be more or less useful in leading to what is called uptake, how learners use the oral feedback offered by the teacher to repair their error or not …” (2017, p. 142; emphasis in original) Does this mean that different types of feedback are inherently more or less useful? Does the usefulness of any specific type of feedback vary depending on the context and other factors?
The authors predicate the usefulness / effectiveness of the feedback on students’ noticing its corrective nature. This takes us to the Noticing Hypothesis, which both Krashen and VanPatten question or reject. John Truscott, in his article “Noticing in Second Language Acquisition: A Critical Review“, notes the extremely limiting constraints of noticing and the weakness of any support for its effectiveness.
Based on my own case studies, I note that a couple of types of corrective feedback enumerated by Glisan and Donato seem to be helpful in negotiating meaning, i.e. in actual communication, and thus potentially helpful for acquisition.
However, Glisan and Donato distinguish between certain strategies employed as oral corrective feedback and those same strategies used as conversational feedback. They write, “It bears mentioning that many of these types of CF [Corrective Feedback] can be used to provide conversational feedback instead of to call attention to linguistic errors.” (2017, p. 143)
I have at least two questions from this:
- Is this a distinction without a difference? If I give feedback that shows I have not understood the utterance, my interlocutor still has to adapt the utterance so that it becomes understandable. This may well involve changing certain elements of the utterance to conform to standard language use.
- Are Glisan and Donato themselves clear on their own distinction? In the paragraph that contains the statement above, the authors give examples of how certain strategies for providing oral corrective feedback may be used as conversational feedback. However, in the same paragraph they state, “Further, instead of using a linguistic utterance, the teach could also use a paralinguistic signal [emphasis in original] to non-verbally elicit a self-correction from the learner – e.g., a quizzical look or nodding of the head …” If the signal is given to “elicit a self-correction”, isn’t that Corrective Feedback? Yet, Glisan and Donato place it in the paragraph devoted to examples of using the strategies for a purpose other than Corrective Feedback.
I am not convinced that the authors themselves are clear on the distinction – or that their distinction truly makes a difference for some of the types of feedback.
So, what are the common types of oral corrective feedback?
- Explicit Correction: The teacher gives the correct form or states that the learner’s statement was incorrect.
- Recasts: The teacher responds and reformulates the learner’s statement without calling attention to the fact that it was incorrect.
- Clarification Requests: The teacher indicates that the learner’s statement was not comprehensible and asks for a reformulation.
- Metalingistic Feedback: The teacher asks questions about the statement (specifically about the grammatical error) or provides grammatical metalanguage that points out the nature of the error (i.e. uses technical language to discuss in English the learner’s error).
- Elicitation: The teacher prompts the learner to provide the correct form by repeating the leaner’s statement verbatim up to the point where the error occurred.
- Repetition: The teacher repeats the learner’s utterance but with an inflection that indicates and highlights an error.
Glisan and Donato also place these Corrective Feedback Strategies on a grid with two axes.
The first axis is the distinction between reformulations (which provide the corrected restatement) and prompts (which elicit self-repair from the student).
The second axis is implicit vs explicit correction. In the latter, the teacher states clearly that correction is taking place or otherwise draws direct attention to it. In the former, it is left to the learner to notice the correction.
The authors then place each of the strategies within a quadrant. Since I haven’t yet figured out how to produce a drawing or table, I will simply describe them.
Quadrant A is Implicit Prompts. These are 1) Clarification Requests and 2) Repetition.
Quadrant B is Explicit Prompts. These are 1) Elicitations and 2) Metalinguiastic clues.
Quadrant C is Implicit Reformulations. The sole strategy placed here is Recasts.
Quadrant D is Explicit Reformulations. Here they place Explicit Correction.
My suspicion is that Explicit Correction is the kind of Corrective Feedback that most people mean when they use the term.
In their conclusion to this section of the chapter, Glisan and Donato make some statements that bear further consideration. They write:
… although recasts tend to be the CF type used most by teachers, they have generally been found to be less likely to lead to uptake by learners in comparison to other strategies such as elicitation and clarification requests. The benefit of recasts appears to be i scaling learners’ attention to form without disrupting communication and meaning-making … However the issue is even more complex,. Research has suggested that even though learners might attend to explicit CF more easily, “the effects of implicit CF might be more robust (i.e. longer lasting) than those of explicit CF, which might be more effective in the short term …” (Lyster, Saito, & Sato (2013). “Oral corrective feedback in second language classrooms” in Language Teaching, 46, p. 5; quoted by Glisan and Donato (2017, p. 144).)
Let’s take note of what the authors say (and perhaps don’t say) and the implications. Glisan and Donato report:
When the teacher recasts a learner’s utterance, this is implicit reformulation, and students do not give as great evidence of uptake (i.e. the students do not repair their error at the time of “correction”) as with certain other strategies.
At first glance, we might conclude that the most-used strategy is the least effective.
However, this does not take certain factors into account. First of all, a recast is not necessarily intended to elicit a repair of the utterance on the part of the learner. The recast provides the leaner with a corrected version of the statement but does not prompt the student to repeat it, at least not if we wish to maintain the primary advantage of this strategy – which is that it functions without disrupting communication and meaning-making. (2017, p. 144)
Since we cannot see what is going on within the mind/brain, it is difficult (impossible?) to evaluate the efficacy of a strategy that does not call for production.
This, btw, is one of the difficulties many people have in evaluating the efficacy of comprehension-based strategies and practices. Since acquisition is internal and unconscious even for the student, the traditional markers and indicators used by assessments do not provide sufficient data of the right kind for learning-based tests to make an evaluation.
In this section, I believe that the authors reveal their bias. I may be reading more into the text than is there, but I don’t believe so. Here’s the evidence.
When presenting the basis for the practice, Glisan and Donato are categorical in their assertion that “A body of research has confirmed the effectiveness of corrective feedback …” They are also quite definitive in their assertion that students need to notice grammatical features. Earlier I noted what looks like a (not so) veiled jab at Van Patten’s position on language acquisition.
However, in this instance, where evidence contrasts with their position, the authors present it much more speculatively and attach a qualifier from their position. “Research has suggested … learners might attend … the effects of implicit CF might be more robust … This finding could be the result … and perhaps produce a response.” From their speculation (“This finding could be the result of implicit CF prompting learners to access L2 knowledge and problem-solving as they work to notice the error, correct it, and perhaps produce a response.”), Glisan and Donato produce the following definitive statement: “Therefore [i.e. because of what we speculate], the long-term effects of implicit feedback only apply in cases where learners’ [sic] are motivated to attend to the teacher’s reformulated part of the utterance and process for themselves what they heard and noticed.” (2017, p. 144)
The definitiveness of the conclusion is, to me, unwarranted from the speculation.
Besides, I can postulate another explanation that, I believe, better “preserves the appearances” (cf. Occam’s razor):
Implicit error correction in the form of recasts and reformulations simply provides the learner with additional comprehensible input at a moment when the student is interested in the input. Thus, the long-term effects of implicit “error correction” are attributable to the unconscious process of acquisition rather than the complex and tenuous mechanism of “noticing“.
I will not attempt to ascribe motive or describe a history of the text. Such attempts are overwhelmingly, (nearly) unanimously, erroneous.
Let it suffice to say that I believe Glisan and Donato seriously miss the mark in their evaluation of the effectiveness of both noticing and explicit Corrective Feedback in language acquisition.
Next time we’ll take a look at Glisan and Donato’s Considerations about Providing Corrective Feedback.