Reporting SLA research relevant to teachers: How not to do it. Part 2

In my last post, I commented on the first strand of “recent, interesting research” which Penny Ur addressed at the 2023 IATEFL conference. She’s seen above getting a medal from someone who knows even less about SLA research than she does.

Here, I comment on two more strands, drawing again on Leo Sellivan’s report.    

The Use of the L1

The research paper Ur referred to (de la Fuente & Goldenberg, 2022) is a well-written and well-considered report of a well-designed study, and it confirms the already widely-accepted view that using the L1 is often very helpful. But I doubt many of Ur’s audience needed to be persuaded that judicious use of the L1 is a “good thing”. On the other hand, I bet Ur didn’t point out that the school where the study was carried out is committed to a task-based learning approach (Ur dismisses TBLT as “unproven”), and I bet she didn’t refer to this bit at the end of the paper either: 

An important question, and one that this study did not seek to address, is which types of L1 use (by students, by instructors) and which specific functions had a bigger impact on learning. In other words, what within the +L1 condition (principled approach to L1 use) accounts for the significant gains in learning?  (de la Fuente & Goldenberg, 2022, p. 960).

Surely this is what needs discussion. The types of L1 use that work best depend too much on context for research to uncover any reliable, universal patterns. Consequently, we’re talking here about what Long (2015) calls a “pedagogic procedure”, and, as I argued recently in an exchange of views with Phillip Kerr, research into the efficacy of different pedagogic procedures should be carried out by groups of teachers who share the same context. (Ahem: teachers should be paid for this work, and the results should feed into, and hopefully improve, local teaching practice.)

  Oral corrective feedback

Now this is a very, very important part of what teachers do.

Ur began by citing Lyster and Randa’s (1997) seminal study, “which showed that recast was the most common oral error correction technique used by teachers, but the least effective one”.  That’s not a very accurate summary, but anyway, we’ll let it go and focus on the Big Question: Is it better to provide immediate correction, using relatively unobtrusive procedures such as recasts, or to provide more detailed correction after the task has been completed?

Ur cited studies by Li (2017) and Harmer, P. (2015) which concluded that learners prefer to be corrected at the moment they make the mistake. Ur then cited Fu and Li’s (2022) study, which claims the superiority of immediate corrective feedback. “The authors recommend correcting errors immediately after the first exposure to a new language item in order to rectify errors before they are proceduralized in the learner’s interlanguage”.

according to Sellivan, Ur said no more on the matter.

Discussion

In Long’s TBLT, oral corrective feedback is a vital part of the subservient but important role that explicit teaching plays. It’s a key component of “focus on form”, which is Long’s unfortunate name for his counter to a grammar approach that is based on a “focus on forms”.  Long says that the best way to help students learn an L2 is to involve them in tasks where they “learn by doing”. However, since adults are, in Long’s opinion (an opinion which relies on Nick Ellis’ theory of SLA) “partially disabled” language learners, they benefit from a certain type of explicit instruction and intervention by teachers, and recasts are a key component. The gap – chasm! – between Ur’s and Long’s views on oral correction is hard to exaggerate.

Needless to say, Ur’s perfunctory treatment of “recent, interesting research” into oral corrective feedback gave the audience no insight into its significance, and nothing to chew on. In contrast, Leo Selivan went away, looked up the references, read the articles, thought about the issues, and made some interesting comments in his review. Good for him! But who else in the audience did anything similar? Leo could have emailed Ur and got the suggested reading list in 2 minutes, and the rest of the 2,000+ audience could have done something much better with the time they spent listening to Penny Ur smugly sharing her reading list.

The Fu and Li (2022) paper is worth discussion. It’s an excellent report of a study exploring the differential effects of immediate and delayed corrective feedback on the acquisition of the English past tense by 145 seventh-grade EFL students at a private school in Eastern China. The study is in, IMHO, extremely well-designed, well-reported and well-discussed. It includes discussion of two conflicting SLA theories: Long’s Interaction Hypothesis (Long, 2015), and Skill Acquisition Theory (Lyster, 2004). Long’s Interaction Hypothesis claims that feedback in the form of prompts such as recasts are ideal “because they withhold the correct form and encourage self-repairs according to which the optimal time to address linguistic problems is through critical feedback (CF) during negotiated interaction” (Fu & Lu, 2022, p.4).  On the other hand, Skill Acquisition Theory (Lyster, 2004) emphasises the importance of feedback in “activating and proceduralizing the declarative knowledge which has been presented in the first stage of the teaching”.

Obviously, the two theories contradict each other, but there’s some convergence when it comes to the timing of corrective oral feedback. The Interaction Hypothesis insists that immediate feedback is superior, while Skill Acquisition Theory suggests that feedback can be provided during tasks performed immediately after explicit instruction, or postponed until a later time after learners complete some communicative practice. Given the authors’ and the school’s preference for TBLT, there was no explicit instruction phase prior to performance of the communicative task, and the study found that there was a marked advantage for unobtrusive, immediate critical feedback, including recasts. But the question of whether Long’s interaction hypothesis is a better explanation of L2 learning that skills acquisition theory remains. At least one of the two theories is wrong, and teachers would surely benefit from information about their relative merits and support from research findings. Furthermore, there are still scholars, including Rod Ellis and Peter Skehan & Pauline Foster, who argue for the importance of delayed, post-task feedback.

Conclusion

Penny Ur’s 2023 IATEFL talk was a waste of time, typical of what you get at conferences where the organisers blindly rely on “Big Names”. On the other hand, Leo Selivan’s blog post contributes much: he reports on the talk and gives us his considered view of the issues that Ur glided over.

Reporting SLA research relevant to teachers: How not to do it

Penny Ur’s session at the IATEFL 2023 conference was on Interesting recent research. It is, in my opinion, a good example of the poor quality of information that those responsible for teacher education provide the profession. Leo Selivan gave a full review of the session, and I rely on his report here.

Selivan says:

After defining what is meant by recent (the last 5-10 years) and what is considered interesting (what she personally finds interesting), in what followed Penny Ur bombarded the audience with highlights of no less than 30 research studies – all in under 25 minutes!

The studies she presented were divided into six strands:

1. Age and language learning
2. L1 in the language classroom
3. Oral corrective feedback
4. Inferencing meaning from context
5. The flipped classroom
6. The use of pictures

I’ll concentrate here on age and language learning. In a second post, I’ll discuss Ur’s hopeless attempts to inform teachers about the other strands.

Age and language learning

The popular belief that children are better foreign language learners has prompted educational policy makers around the world to advocate for an earlier start of foreign language instruction. However, Ur cites Carmen Muñoz’s (2006, therefore not “recent” by Ur’s own definition) Barcelona Age Effect (BAF) project which showed that in the foreign language classrooms of Barcelona’s primary and secondary schools, “age did not afford the same advantages as it does in a naturalistic language learning setting. Contrariwise, older learners (teenagers) may fare better”.

Ur also cites a review of research on the age factor by Muñoz and Singleton (2010, ditto) which argued that other factors contributing to learner success, including motivation, attitudes, affiliation with the L2 and the amount of exposure and quality of input have been largely ignored. Selivan summarises Ur as follows: “The kind of learning child learners are better at – implicit learning – requires vast amounts of exposure. Unfortunately, in a school setting with a couple of classes a week, exposure is woefully insufficient making it difficult for children to leverage implicit learning mechanisms”.

Finally, Ur cites Lightbrown and Spada’s (2020) paper which concludes that “studies in schools settings around the world have failed to find long-term advantages [ultimate attainment] for an early start …..”.

That’s all Ur had to say.

Discussion

What, I wonder, did the audience make of Ur’s treatment of this important issue? The point she hurriedly made was that research findings question the wisdom of educational policy makers who advocate for an earlier start of foreign language instruction. Not surprisingly, given her attempt to get into the Guiness Book of Records for the most research covered in a 25-minute talk, she neglected to attempt any critical evaluation of the cited research. So she neglected to say anything about what the participants in Muñoz’s study were taught.

Muñoz herself says little about the syllabus and teaching methods used in the English courses she reports on in her articles on the BAF project, and you have to read her book to find out that the courses in question were based on synthetic syllabuses which involve treating the target language as an object of study. The teachers involved in the BAF often had a low proficiency in oral communication, they talked for 70%+ of classroom time, frequently in their L1, and the students had few opportunities to use the target language, to engage in communicative tasks, to learn by doing. Thus, students were taught about the target language far more than they were helped to use it and the tests measured what they knew about the language more than how well they could use it for spontaneous communication. Given all this, it’s hardly surprising that the older students learned what they were taught faster and better than the young ones, because, as Carmen herself indicates, young children rely on implicit learning which leads to procedural knowledge and consequent communicative competence, while older students are more receptive to explicit learning which leads to far less useful declarative knowledge.

So “earlier is better” as an educational policy is only mistaken if ELT is based on the sort of ELT you find in Catalan public schools, which is probably a representative model of how it’s taught just about everywhere. Older children are better than younger children at learning about the language, which is not, in my opinion, what – or how – they should be taught.

The still controversial “Critical Period Hypothesis” (which these days is explored in terms of “sensitive periods”) needs explaining to teachers, and its implications for ELT need attention. Teachers need to be involved in discussions about the research which suggests that young children learn additional languages in the same way that they learn their native language or languages: they make sense of the formal aspects of the additional target language for themselves while being exposed to, and interacting with others in that language. Meanwhile, the research suggests that older students learn slightly differently. If we accept (still disputed) research findings of senstive periods, then the putative windows of purely implicit learning of certain features of the additional target language close by the time children reach the age of 14 or so (the window for pronunciation is said to close at around 8 years old!). In which case, they, like adult learners, need more explicit teaching which pays attention to these features.

Ur did precisely nothing to help teachers appreciate the issues involved in research into how age affects SLA, and I doubt that she did anything to make teachers think about its implications. She was given the biggest stage, star billing and, no doubt, all expenses paid. In his review, Selivan waxes lyrical about Ur’s ability to “make research sound sexy”; he applauds her audacity at seeking to cover so much ground; he drools that he was so “fired with enthusiasm” to report on this session that he made it his “only write-up from IATEFL 2023”.  It’s enough to make you weep!