In response to a tweet from David Cullen, here’s a summary of SLA research that I think needs to be taken more seriously by the ELT community.
From time to time one sees well known “experts” on ELT refer to SLA research. The standard message is that researchers know nothing about real-world classroom practice, and that most of their findings are either irrelevant or unreliable. A few trinkets from SLA research are trotted out by the ELT experts as evidence of their scholarship, and they include these:
- Using the L1 is OK.
- Teaching lexical sets is not OK.
- Guessing from context is not a reliable way of “accessing meaning”.
- Spaced repetition is a must.
- Getting in the flow really helps learning.
Such accounts of the research are, I think, cynically frivolous, so, within the confines of a blog post, let’s take a slightly more serious look.
The empirical study of how people learn a second language started in the 1960s and I’ve written a series of posts on the history of SLA research which you will find in the Menu on the right of this page. Here, I’ll concentrate on theories which look at the psychological process of learning, and in particular on three areas:
- Interlanguage: how the learner’s mental idea of the target language develops.
- The roles of explicit and implicit knowledge and learning.
- Maturational constraints on adult SLA , including sensitive periods.
Studies of interlanguage development were given a framework by Selinker’s (1972) paper which argues that L2 learners develop their own autonomous mental grammar (IL grammar) with its own internal organising principles. More work on the acquisition of English morphemes, and then studies of developmental stages of negation in English, developmental stages of word order and questions in English, and then Pienemann’s studies of learning German as an L2 where all learners adhered to a five-stage developmental sequence (see Ortega, 2009, for a review) put together an increasingly clear picture of interlanguage development.
By the mid 1980s it was clear that learning an L2 is a process whereby learners slowly develop their own autonomous mental grammar with its own internal organising principles. Thirty years on, after hundreds more studies, it is well established that acquisition of grammatical structures, and also of pronunciation features and many lexical features such as collocation, is typically gradual, incremental and slow. Development of the L2 exhibits plateaus, occasional movement away from, not toward, the L2, and U-shaped or zigzag trajectories rather than smooth, linear contours. No matter what the order or manner in which target-language structures are presented to them by teachers, learners analyze the input and come up with their own interim grammars, the product broadly conforming to developmental sequences observed in naturalistic settings. They master the structures in roughly the same manner and order whether learning in classrooms, on the street, or both. To be perfectly clear: the acquisition sequences displayed in IL development have been shown to be impervious to explicit instruction, and the conclusion is that students don’t learn when and how a teacher decrees that they should, but only when they are developmentally ready to do so.
2. The roles of explicit and implicit knowledge and learning
Two types of knowledge are said to be involved in SLA, and the main difference between them is conscious awareness. Explicit L2 knowledge is knowledge which learners are aware of and which they can retrieve consciously from memory. It’s knowledge about language. In contrast, implicit L2 knowledge is knowledge of how to use language and it’s unconscious – learners don’t know that they know it, and they usually can’t verbalize it. (Note: the terms Declarative and Procedural knowledge are often used. While there are subtle differences, here I take them to mean the same as explicit and implicit knowledge.)
Now here’s the important thing: in terms of cognitive processing, learners need to use attentional resources to retrieve explicit knowledge from memory, which makes using explicit knowledge effortful and slow: the time taken to access explicit knowledge is such that it doesn’t allow for quick and uninterrupted language production. In contrast, learners can access implicit knowledge quickly and unconsciously, allowing it to be used for unplanned language production.
While research into explicit and implicit L2 knowledge started out looking almost exclusively at grammar, we now have lots of evidence to show that these constructs also relate to other areas of language such as vocabulary, pronunciation, and pragmatics. Just for example, while lots of vocab. learning involves learning items by explicit form-meaning mapping (“table” is “mesa” in Spanish), there are very important aspects of vocabulary knowledge, like collocations, and lexical chunks, for example, that involve implicit knowledge.
Three Interface Hypotheses
It’s now generally accepted that these two types of language knowledge are learned in different ways, stored separately and retrieved differently, but disagreement among SLA scholars continues about whether there is any interface between the two. The question is: Can learners turn the explicit knowledge they get in classrooms into implicit knowledge? Or can implicit knowledge be gained only implicitly, while explicit knowledge remains explicit? Those who hold the “No Interface” position argue that there’s a complete inability for explicit knowledge to become implicit. Others take the “Weak Interface” position which assumes that there is a relationship between the two types of knowledge and that they work together during L2 production. Still others take the “Strong Interface” position, based on the assumption that explicit knowledge can and does become implicit.
The main theoretical support for the No Interface position is Krashen’s Monitor theory, which has few adherents these days. The Strong Interface case gets its theoretical expression from Skill Acquisition Theory, which describes the process of declarative knowledge becoming proceduralised and is most noteably championed by DeKeyser. This general learning theory clashes completely with evidence from L1 acquistion and with interlanguage findings discussed above. The Weak Interface position seems right to me and to most SLA scholars.
Implicit Knowledge is the driver of IL development
Whatever their differences, there is very general consensus among scholars that SLA has a great deal in common with L1 learning, and that implicit learning is the “default” mechanism of SLA. Wong, Gil & Marsden (2014) point out that implicit knowledge is in fact ‘better’ than explicit knowledge; it is automatic, fast – the basic components of fluency – and more lasting because it’s the result of the deeper entrenchment which comes from repeated activation. The more that any speech act draws from implicit knowledge, the less we need to rely on explicit knowledge. Doughty (2003) concludes:
In sum, the findings of a pervasive implicit mode of learning, and the limited role of explicit learning …, point to a default mode for SLA that is fundamentally implicit, and to the need to avoid declarative knowledge when designing L2 pedagogical procedures.
One more twist. Research shows that although there are doubts about its usefulness, explicit knowledge about the language of the kind teachers habitually offer in class is easy to learn. Equally, the implicit language knowledge that is obtained from engagement in real communicative activities is relatively hard to learn – evidence from Canadian immersion courses, for example, shows that when learners are focused exclusively on meaning, there are parts of the target language that they just don’t learn, even after hundreds of hours of practice. This leads Loewen (2015) to say:
The fact that explicit knowledge is relatively easy to learn but difficult to use for spontaneous L2 production, and that, conversely, implicit knowledge is relatively difficult to learn but important for L2 production is, I feel, one of the most important issues in ISLA and L2 pedagogy. It is essential for L2 learners and teachers to be aware of the different types of knowledge and the roles that they play in L2 acquisition and production. The implication is that the teaching of explicit language rules will not, by itself, result in students who are able to communicate easily or well in the language”.
3. Maturational constraints on adult SLA , including sensitive periods
Our limited ability to learn a second language implcitly (in stark contrast to the way we learn our L1 as young children) brings us to the third area of SLA research I want to look at. Long (2007) in an extensive review of the literature, concludes that there are indeed critical periods for SLA, or “senstive periods”, as they’re called. For most people, the senstive period for native-like phonology closes between age 4 to 6; for the lexicon between 6 and 10; and for morphology and syntax by the mid teens. This is a controversial area, and I’ll be happy to answer any questions you might ask, but there’s general consensus that adults are partially “disabled” language learners who can’t learn in the same way children do. Which is where explicit learning comes in – the right kind of explicit teaching can help adult students learn bits of the language that they are unlikely to learn implicitly. Long calls these bits “fragile” features of the L2 – features that are of low perceptual saliency (because they’re infrequent / irregular / semantically empty / communicatively redundant / involving complex form-meaning mappings), and he says these are likely to be late or never learned without explicit learning.
It seems that the best way teachers can help students to learn an L2 is by concentrating on giving them scaffolded opportunities to use the L2 in the performance of relevant communicative tasks. During the performance of these tasks, where enhanced and genuine texts both written and spoken give the rich input required, teachers should give students help with aspects of the language that they’re having problems with by brief switches to what Long calls “focus on form”, from quick explicit explanations to corrective feedback in the form of recasts.
A relatively inefficacious way of organising ELT courses is to use a General English coursebook. Here, the English language is broken down into constituent parts which are then presented and practiced sequentially following an intuitive easy-to-difficult criterion. The teacher’s main concern is with presenting and practicing bits of grammar, pronunciation and lexis by reading and listening to short texts, doing form-focused exercises, talking about bits of the language, and writing summaries on the whiteboard. The mistake is to suppose that whatever students learn from what they read in the book, hear from the teacher, say in response to prompts, read again on the whiteboard, and maybe even read again in their own notes will lead to communicative competence. It won’t.
The mis-use of SLA findings
Numerous prominent members of the ELT establishment use either carefully selected or misinterpreted research findings in SLA to support their views. Here are 3 quick examples.
Catherine Walter’s (2012) article has been quoted by many. She claims that, while grammar teaching has been under attack for years,
evidence trumps argument, and the evidence is now in. Rigorously conducted meta-analyses of a wide range of studies have shown that, within a generally communicative approach, explicit teaching of grammar rules leads to better learning and to unconscious knowledge, and this knowledge lasts over time. Teaching grammar explicitly is more effective than not teaching it, or than teaching it implicitly; that is now clear. … Taking each class as it comes is not an option. A grammar syllabus is needed.
No meta-analysis in the history of SLA research supports these assertions. Not one, ever.
Likewise, Jason Anderson (2016), in his article defending PPP, says
while research studies conducted between the 1970s and the 1990s cast significant doubt on the validity of more explicit, Focus on Forms-type instruction such as PPP, more recent evidence paints a significantly different picture.
His argument is preposterous and I’ve discussed it here.
The meta-analysis which is most frequently cited to defend the kind of explicit grammar teaching done by teachers using coursebooks is the Norris and Ortega (2000) meta-analysis on the effects of L2 instruction, which found that explicit grammar instruction (Focus on FormS) was more effective than Long’s recommended more discrete focus on form (FoF) approach through procedures like recasts. (Penny Ur is still using this article to “prove” that recasts are ineffective.) However, Norris and Ortega themselves acknowledged, while others like Doughty (2003) reiterated, that the majority of the instruments used to measure acquisition were biased towards explicit knowledge. As they explained, if the goal of the discreet FoF is for learners to develop communicative competence, then it is important to test communicative competence to determine the effects of the treatment. Consequently, explicit tests of grammar don’t provide the best measures of implicit and proceduralized L2 knowledge. Furthermore, the post tests done in the studies used in the meta-analysis were not only grammar tests, they were grammar tests done shortly after the instruction, giving no indication of the lasting effects of this instruction.
Newer, post-2015 meta-analyses have used much better criteria for selecting and evaluating studies. The meta-analysis carried out by Kang, et al (2018) concluded:
implicit instruction (g = 1.76) appeared to have a significantly longer lasting impact on learning … than explicit instruction (g = 0.77). This finding, consistent with Goo et al. (2015), was a major reversal of that of Norris and Ortega (2000).
As for Ur’s typically sweeping assertion that there is “no evidence” to support claims of TBLT’s efficaciousness, a meta-analysis by Bryfonski & McKay (2017) reported:
Findings based on a sample of 52 studies revealed an overall positive and strong effect (d = 0.93) for TBLT implementation on a variety of learning outcomes. … Additionally, synthesizing across both quantitative and qualitative data, results also showed positive stakeholder perceptions towards TBLT programs.
Enough, already! I hope this review of some important areas of SLA research and my comments will generate discussion.
Bryfonski, L., & McKay, T. H. (2017). TBLT implementation and evaluation: A meta-analysis. Language Teaching Research
Corder, S. P. (1967) The significance of learners’ errors.International Review of Applied Linguistics 5, 161-9.
Doughty, C. J. (2003). Instructed SLA: Constraints, Compensation, and Enhancement. In C. J. Doughty, & M. H. Long (Eds.), The Handbook of Second Language Acquisition (pp. 256-310). Malden, MA: Blackwell.
Kang, E., Sok, S., & Hong, Z. (2018) Thirty-five years of ISLA research on form-focused instruction: A meta-analysis. Language Teaching Research Journal.
Krashen, S. (1977) The monitor model of adult second language performance. In Burt, M., Dulay, H. and Finocchiaro, M. (eds.) Viewpoints on English as a second language. New York: Regents, 152-61.
Loewen, S. (2015) Introduction to Instructed Second Language Acquisition.N.Y. Routledge.
Long, M. (2007) Problems in SLA. Lawrence Erlbaum.
Ortega, L. (2009) Sequences and Processes in Language Learning. In Long, M. and Doughty, C. Handbook of Language Teaching. Oxford, Wiley
Seliger, H. (1979) On the nature and function of language rules in language teaching. TESOL Quarterly 13, 359-369.
Selinker, L. (1972) Interlanguage. International Review of Applied Linguistics 10, 209-231.
Wilkins, D. A. (1976) Notional syllabuses. Oxford, Oxford University Press.
Whong, M., Gil, K.H. and Marsden, H., 2014. Beyond paradigm: The ‘what’and the ‘how’of classroom research. Second Language Research, 30(4), pp.551-568.