Skip to main content

The latest research on teaching grammar

This is a summary of a chapter by Hossein Nassaji entitled Grammar Acquisition in the recently published Routledge Handbook of Instructed Second Language Acquisition (2017). The chapter in question succinctly and clearly reviews recent research into classroom grammar teaching and offers some general advice on the topic to language teachers.

Can grammar instruction help develop proficiency?

Nassaji points out that this has been a long-standing controversy in the field. Does learning develop primarily through explicit teaching and conscious manipulation of structures, or merely through unconscious processes when people are exposed to meaningful input (known as implicit learning)? N. Ellis (2007) points out that implicit and explicit learning are functions of separate memory systems in the brain. Scans show that explicit learning is supported by neural networks located in the prefrontal cortex, whereas implicit learning involves other areas of the brain, the perceptual and motor cortex.


1. Two types of grammatical knowledge

Explicit learning is leads to explicit knowledge, often called declarative knowledge, i.e. "being able to explain the rules". This in itself is not much use when it comes to speaking and comprehending in real time. Implicit knowledge is taken to occur based on extensive meaning-focused input, acquired without awareness and stored implicitly (so typically you can speak the language without being able to explain the rules).

2. The Interface

What is the relationship between these two types of knowledge? In particular, of crucial importance to teachers, can explicit gained knowledge become implicit?  There have traditionally been three views about this which involve what's called the interface between explicit and implicit knowledge.  The non-interface position (e.g. Krashen, VanPatten) holds that explicitly, consciously learned language cannot become implicit. Instruction makes no difference - all you need is meaningful exposure. The strong interface position (e.g. DeKeyser, Paradis) holds that implicit knowledge is primarily the result of proceduralisation of explicit knowledge, i.e. through explanation and practising the skills you can become proficient. The weak interface position argues that conscious knowledge can help with gaining implicit knowledge, but does so indirectly by helping learners notice language features which they cab add to their implicit knowledge when they are developmentally ready (e.g. Schmidt, R.Ellis).

(My note: teachers might like to suppose that all of these hypotheses have merits and that teaching which takes into account all three may have its merits.)

3. Empirical evidence?

Does teaching grammar make a difference? Long (1983) looked at 12 studies comparing exposure learning with instructional learning and concluded that overall instruction made a positive difference at all levels with both children and adults. R. Ellis (1990, 1994) and Larsen-Freeman and Long (1991) also found that instruction helped with the rate and ultimate level of acquisition. More recent studies have reached the same conclusion, most famously Norris and Ortega (2000) and Spada and Tomita (2010). Overall then the empirical evidence favours instruction, but...

4. Types of instruction

... not all types of instruction are equally effective. Should the teacher explicitly give rules or let students work them out from input? The evidence so far is that givng and practising rules is generally advantageous, e.g. Norris and Ortega (2000). The data showed a greater effect when using discrete point or grammatical judgment tests, but a lesser effect when spontaneous communication was used to assess teaching success. (Some criticise the Norris and Ortega studies for this reason. It also has to be said that it is much easier to do discrete point texts than tests of general conversation. Do discrete point grammar tests test explicit or implicit knowledge anyway?). In any casre thw later Spada and Tomita (2010) studies usex tests of free production and implicit knowledge.

In short the research evidence so far suggests explicit, conscious learning helps develop implicit knowledge, suggesting that the weak or strong interface positions may be correct.

5. Focus on form or focus on forms?

This distinction was put forward by Long (1991). Focus on forms is traditional, point by point grammar instruction which teachers typically use when using a grammar-translation, audio-lingual or oral-situational approach. Each grammar point is isolated and usually taught in quite a decontextualised, not very meaningful manner, e.g. getting students to do substitution drills where they have to change the tense of the verb in a sentence. In contrast, focus on form is more like dealing with grammar explanations and brief practice when they arise incidentally but where the main focus is on meaningful interaction. It could also involve giving corrective feedback, such as recasts or simple correction.

(My note: many teachers may well use a mix of the above two approaches, depending on the grammar point.)

There is research to suggest that focus on form (grammar in a meaning-focused context) can more effective, e.g. R.Ellis (2008), Nasaji and Fotos (2004; 2010), but overall there is not yet any clear evidence that one approach is better than the other, e.g. Shintani (2013). Much of the research in this area is somewhat problematic because there are no clear definitions of what constitutes focus on form and focus on forms instruction.

(My note: could Long have come up with a less confusing terminology for teachers? Almost certainly yes!)

6. Input or output-based instruction?

Nassaji then goes on to look at grammar teaching from another perspective. Should the focus of teaching be on input or output? Input-based instruction refers to strategies which involve the processing of input. This is based on the assumption that learners' attention can be drawn to grammatical forms through activities where the aim is to understand input for meaning (not form), e.g VanPatten (2015). This is known as Processing Instruction (PI). Some studies support this approach over traditional grammar teaching (e.g. give the rule and practise, as in PPP). Critics, e.g. DeKeyser, argue that this type of instruction really only improves comprehension. Input-focused teaching can also involve "flooding the input" with the language features you want the class to develop (e.g. in narrow reading and listening tasks). Research studies have, as yet, yielded little positive results for this approach as attractive as it seems.

In contrast output-based instruction draws attention to structures through eliciting and practising output. This would be more in line with traditional PPP grammar teaching. DeKeyser suggests that output tasks will help develop output whereas input tasks will help develop comprehension.

Some studies have suggested that a combination of input flooding and explicit instruction can have positive results, e.g. Hernandez, 2011, though again the effects on long term learning remain unclear.

(My note: this sounds like "you get better at what you practise". Perhaps a balance of input and output tasks is called for, depending on what the aims of the class are and how the students will be assessed.)

7. Effects of different types of instruction on different knowledge types

Next Nassaji considers whether certain types of instruction benefit different knowledge types. Some recent studies have examined the effects of implicit and explicit instruction on explicit and implicit knowledge. Recall that implicit knowledge is what most teachers would be hoping to develop. R. Ellis (2002) analysed 11 studies that had used free production as a measure of implicit knowledge.  He concluded that explicit knowledge helped (evidence for the interface referred to above). Importantly, however, Ellis noted that explicit instruction worked best with simple target features, e.g. verb forms, but less well with more complex forms, e.g. the English passive. Andringa, de Glopper and Hacquebord (2011) also found that explicit instruction helped with acquiring Dutch grammar, but only with certain structures. Explicit teaching helped most when the L1 and L2 structures were similar.

8. What can teachers take from this?

Let me in bullet point form summarise the points Nassaji makes - bear in mind researchers are better at describing research than suggesting classroom activities:

  • Allow opportunities for implicit and explicit learning, but don't assume explicitly learned language will become implicit.
  • Learning a language takes time. Instruction is important, but more so is long term exposure to meaningful input.
  • Provide opportunities for repeated use of grammatical forms in meaningful contexts. 
  • Give opportunities for both controlled and free practice.
  • Give feedback on errors.
  • Provide exercises where students have to work out a rule from examples. Best if these can be done in the target language in pairs or groups.
  • Bear in mind if your class is developmentally "ready" to do certain types of task. In a mixed ability class this is problematic.
  • Use a wide range of instructional techniques to cater for individual differences.
  • Use formative assessment techniques, e.g. listening for students' errors in conversation or practice to guide future grammar teaching.
  • Bear in mind that some grammatical features are much less "salient" in the input to English learners, e.g. adjective agreement or gender. Make these as salient as possible if you think they are important.
  • Explicit instruction may be particularly useful for rare forms which do not appear often in input.
  • Bear in mind the whole range of learner differences: age, ability, attitudes, cultural background and attitudes towards learning.

(My note: bear in mind too the assessment regime and school culture, e/g/ is pair work common? is most teaching teacher-led?)

9. Concluding remarks

Some readers will find the pedagogical advice from Nassaji pretty obvious. They may also be frustrated by the relatively tentative empirical data from research. Just think how hard these things are to research though! Yet another reminder that research in this field is young.


If you wish to follow up any references just use Google with my brief references from the text. Most will not be available online though.

My reference was Chapter 12 of the Kindle version of:

The Routledge Handbook of Instructed Second Language Acquisition.  Loewen and Sato (2017). London: Routledge.



Popular posts from this blog

The latest research on teaching vocabulary

I've been dipping into The Routledge Handbook of Instructed Second Language Acquisition (2017) edited by Loewen and Sato. This blog is a succinct summary of Chapter 16 by Beatriz González-Fernández and Norbert Schmitt on the topic of teaching vocabulary. I hope you find it useful.

1.  Background

The authors begin by outlining the clear importance of vocabulary knowledge in language acquisition, stating that it's a key predictor of overall language proficiency (e.g. Alderson, 2007). Students often say that their lack of vocabulary is the main reason for their difficulty understanding and using the language (e.g. Nation, 2012). Historically vocabulary has been neglected when compared to grammar, notably in the grammar-translation and audio-lingual traditions as well as  communicative language teaching.

(My note: this is also true, to an extent, of the oral-situational approach which I was trained in where most vocabulary is learned incidentally as part of question-answer sequence…

A zero preparation fluency game

I am grateful to Kayleigh Meyrick, a teacher in Sheffield, for this game which she described in the Languages Today magazine (January, 2018). She called it “Swap It/Add It” and it’s dead simple! I’ve added my own little twist as well as a justification for the activity.

You could use this at almost any level, even advanced level where the language could get a good deal more sophisticated.

Put students into small groups or pairs. If in groups you can have them stand in circles to add a sense of occasion. One student utters a sentence, e.g. “J’aime jouer au foot avec mes copains parce que c’est amusant.” (You could provide the starter sentence or let groups make up their own.) The next student (or partner) has to change one element in the sentence, and so on, until you restart with a different sentence. You could give a time limit of, say, 2 minutes. The sentence could easily relate to the topic you are working on. At advanced level a suitable sentence starter might be:

“Selon un article q…

Google Translate beaters

Google Translate is a really useful tool, but some teachers say that they have stopped setting written work to be done at home because students are cheating by using it. On a number of occasions I have seen teachers asking what tasks can be set which make the use of Google Translate hard or impossible. Having given this some thought I have come up with one possible Google Translate-beating task type. It's a two way gapped translation exercise where students have to complete gaps in two parallel texts, one in French, one in English. There are no complete sentences which can be copied and pasted into Google.

This is what one looks like. Remember to hand out both texts at the same time.


_____. My name is David. _ __ 15 years old and I live in Ripon, a _____ ____ in the north of _______, near York. I have two _______ and one brother. My brother __ ______ David and my _______ are called Erika and Claire. We live in a _____ house in the centre of ____. In ___ house _____ …

Preparing for GCSE speaking: building a repertoire

As your Y11 classes start their final year of GCSE, one potential danger of moving from Controlled Assessment to terminal assessment of speaking is to believe that in this new regime there will be little place for the rote learning or memorisation of language. While it is true that the amount of learning by heart is likely to go down and that greater use of unrehearsed (spontaneous) should be encouraged, there are undoubtedly some good techniques to help your pupils perform well on the day.

I clearly recall, when I marked speaking tests for AQA 15-20 years ago, that schools whose candidates performed the best were often those who had prepared their students with ready-made short paragraphs of language. Candidates who didn't sound particularly like "natural linguists" (e.g. displaying poor accents) nevertheless got high marks. As far as an examiner is concerned is doesn't matter if every single candidate says that last weekend they went to the cinema, saw a James Bond…

Worried about the new GCSEs?

Twitter and MFL Facebook groups are replete with posts expressing concerns about the new GCSEs and, in particular, the difficulty of the exam, grades and tiers. I can only comment from a distance since I am no longer in the classroom, but I have been through a number of sea changes in assessment over the years so may have something useful to say.

Firstly, as far as general difficulty of papers is concerned, I think it’s fair to say that the new assessment is harder (not necessarily in terms of grades though). This is particularly evident in the writing tasks and speaking test. Although it will still be possible to work in some memorised material in these parts of the exam, there is no doubt that weaker candidates will have more problems coping with the greater requirement for unrehearsed language. Past experience working with average to very able students tells me some, even those with reasonable attainment, will flounder on the written questions in the heat of the moment. Others will…