Catherine Walter started off the second day of E-merging Forum 5 with a plenary reviewing current approaches to teaching grammar and how their efficiency is corroborated by experiments. I’m very interested in this topic, particularly in the overview of evidence, so I was really looking forward to this talk.
Outline:
- How to teach grammar? Overview of approaches
- Why this way and not that way? Overview of evidence
- When – before or after the task? Who chooses what to teach – the teacher beforehand or should we react to the learner’s need?
- How much explicit grammar teaching?
- How to teach x2: what is ‘a good rule‘? A good example? A good exercise? In what order should they appear?
- The role of pronunciation
NB The ideas on teaching grammar outlined here are based on research applicable to middle school and above learners – not YLs.
Catherine Walter started the talk with a quiz for the audience. Grammar: True or false?
- If people learn enough vocab, they’ll acquire the grammar of the language
- The best way to teach grammar is to wait unteel the need for a specific grammar point emerges, and then teach it
- The best way to teach grammar is via tasks
- There is evidence that teaching grammar rules works
For a quick overview of the answers and the talk, here the final slide. For more detail, read on.
How to teach? Overview of approaches:
Explicit (through rules) or implicit (through exposure, examples, etc)? How to provide practice?
- Exposure. Necessary? Sufficient?
Just exposure is insufficient, which was clearly demonstrated by Canadian immersion programmes who were trying Krashen’s idea that comprehensible input is enough: after 12 years of all-day-long immersion the students’ spoken and written production was still non-standard. - Explicit teaching. Input-interaction-output model says conscious knowledge is useful – ‘crutches’ that hold you until you can walk w/o the crutches – which sounds Vygotskian!
>>> Explicit teaching helps noticing: Example: if your L1 has only one word for ‘yes’ and L2 has more, you might not notice that. But if someone tells you about this, you’ll notice;
>>> encourages comparison of noticed input with the learner’s output, can convert directly into unconscious knowledge and
>>> can provide negative feedback, i.e. the knowledge of what doesn’t happen in the language (which is more difficult to get from input) - Tasks. When tasks appeared, they were seen as a panacea. But now it’s not clear how well they work so that’s definitely not ‘the obviously best way’. Explicit grammar instruction can be (and should be) part of task-based instruction.
- Skills approach: the behaviourism is re-emerging in a way: we do need to practice to build our skills.
Why? Overview of the evidence
Is there any evidence that explicit grammar teaching works?
Norris & Ortega review (2000) found that explicit teaching is better than implicit
Gass & Selinker (2008) review: after early childhood, acquiring complex forms requires both meaningful input and explicit grammar focus
Spada & Tomita (2010): explicit better than implicit on simple & complex features, effects last
A possible counter-argument: if you teach them grammar rules, do they only learn grammar rules?
Spada & Lightbown (2008): form-focused input leads to conscious and unconscious knowledge over time
When to teach what?
Two options: first teach then give a task, or feed in grammar exactly when the learners needs it.
Problem with the second approach #1: different people in class will have different needs.
Problem #2: you won’t be able to always think on your feed to respond to the needs.
Relax: Spada & Tomita (2010) found that it doesn’t matter, Both isolated and integrated form-focused instruction lead to conscious and unconscious grammar knowledge; no clear advantage for either kind of form-focuses input.
How much explicit grammar teaching?
Catherine recommends checking out Paul Nation’s (2009) work. The four strands of language instruction, which according to Nation need equal time:
- Meaning-focused input
- Meaning-focused output
- Language-focused learning (rules/pron/how writing works/..)
- Fluency development (activities tat help you get faster and more automatic – activities might be quite artificial, like scales in learning to play a musical instrument).
So according to Nation, and Catherine Walter agrees, Language-focused learning (rules/pron/how writing works/..) takes up one fourth of the course.
How to teach x2. The three Es
- Explanations (rules). However: there are different kinds of rules.
g. ‘No cycles, whether ridden or not!’ is imposed by an authority
‘At sea level water boils at 100 Celsius.’ – an expression of an observed regularity.
very often s/s think that language rules are bicycle rules, but they are boiling water rules - Examples
- Exercises
The order of the 3 Es
- Deductive: explanations > examples > exercises
- Inductive: examples > guiding the s/s towards the rule (explanations) > exercises; useful to demonstrate to s/s that our rules are ‘boiling water’ rules + more appealing to the s/s who don’t like authority + means that they cognitively process the rule more deeply, so they might remember it better
What is ‘a good rule’? According to Swan (‘Design criteria for pedagogical rules’) a good rule should
- be true (not like ‘The past tense refers to a DEFINITE time in the past – what about ‘Once upon a time?’)
- clearly what are the limits on the use (a pika is a smal rodent with small ears – this definition doesn’t differentiate the pika from the mouse; ‘The present perfect continuous tense is used for an action which began in the past and is still continuing, or has only just finished – does not demarcate. ‘I’m speaking to you’ fits that ‘rule’)
- be clear and simple without sacrificing the truth
- (preferably) only uses the knowledge that the learners currently have (e.g. for the first teaching of ‘much and many’ – can we not use ‘countable and uncontable’?
- be relevant: should answer the question that the learner is ‘asking’
My sister Marie-France is hairdresser. She works in Lyon.
My sister Olga is hairdresser. She works in Volgograd.
The two learners who produced that have very different L1s and so they will need completely different rules here!
A good example:
- Needs to be prototypical
- Natural sounding & not containing irrelevant difficulties (inspired by corpora, but not necessarily from a corpus – might contain too difficult vocabulary, cultural references etc) – The oxen are stepping on my feed is not a good example =)
How many examples:
Just one-two not enough: Goldilocks principle (not too big, not too small, just right
Exercises:
- Nothing wrong with exercises! Our concentration capacity is limited (which is exemplified e.g. by the fact that mobile phone use has overtaken alcohol as the biggest cause of accidents in the UK), exercises let one focus. Also, they develop fluency + there’s evidence that focusing on one or two language features is better than focusing on a range of language features.
- A good exercise combines quality and quality to ensure deep cognitive processing (a communicative ‘have you ever’ might look like a drill but not be a drill in terms of processing)
Productive and receptive pronunciation
- Grammar teaching should go hand in hand with teaching receptive pronunciation, because some grammar points are very difficult to hear
- Examples of grammar points that are difficult to distinguish: Regular present simple and regular past simple (She walk to school / She walks to school);
questions: Do you work on Sat / Did you work on Sat, etc - How to teach receptive pronunciation?
Decoding activities: A or B? Same or different? Odd one out (walks, walked, walks, walks); Write the word; how many words? Sentence dictation
Back to the quiz
- If people learn enough vocab, they’ll acquire the grammar of the language
Yes, but they’d need to learn a huge amount of lexis and they’ll never learn enough vocabulary. - The best way to teach grammar is to wait until the need for a specific grammar point emerges, and then teach it
Probably true.. but this isn’t possible in most instructed language - The best way to teach grammar is via tasks
No - There is evidence that teaching grammar rules works
Yes!
________________________
All in all, this was a very enjoyable, informative talk and I’ll definitely be digging up some of the articles that Catherine Walter mentioned.
Olya Sergeeva's ELT blog 