"They know the rule but they won't use it"
Ruth teaches a group of 22 adults (A2–B1, nine different first languages) at a private language school. Her standard lesson is presentation–practice–production: she explains a structure on the board, learners complete a ten-item gap-fill, then do a role-play in pairs. On Tuesday the target was past continuous. Ruth's own tally: 19 of the 22 scored at least 8/10 on the gap-fill, but in the recorded role-play the form appeared four times in twelve minutes, and two learners used past simple throughout. In the feedback slot three learners correctly corrected each other's written sentences, then made the same error orally two minutes later. Ruth concludes that the class 'knows the rule but won't use it' and plans to add more gap-fill items next lesson.
1. Ruth's diagnosis is that learners 'know the rule but won't use it'. What is misleading about this diagnosis, and what explanation fits the evidence better?
Read the worked answer
The diagnosis treats the gap as a problem of willingness. The evidence points to a problem of access under pressure, which is a different thing and needs a different remedy.
- The gap-fill is completed with time, a visible prompt and one possible answer. It measures declarative knowledge under monitoring. The role-play measures whether the form is available online, while the learner is also managing meaning, a partner and turn-taking. Those are not the same ability, so 8/10 and zero use are not contradictory findings.
- The feedback episode is the strongest evidence: the same learners self-correct in writing and then fail orally within two minutes. That is a processing constraint, not a motivational one. 'Won't' predicts no change under any condition; 'can't yet, under these conditions' predicts change if the conditions change.
- Past simple for past continuous is not avoidance or laziness. It is a functioning interlanguage strategy that communicates the meaning. Learners will only abandon it when there is a reason to be more precise.
- The role-play itself may create no communicative need for the target: pair role-plays with no information gap let learners say whatever is easiest.
Conclusion: the class does not need more controlled practice of a form it can already produce accurately; it needs conditions where the form is the easiest way to do the task.
2. Design the first ten minutes of the next lesson to address the real problem, and say what you would cut.
Read the worked answer
Cut the board presentation and the gap-fill. They have been shown to work and are not the bottleneck.
A defensible opening:
- Return their own language (3 min). Play two or three short extracts from the recorded role-play. Learners write down what they hear. They will hear past simple where past continuous was expected. No correction yet - just noticing their own output.
- Contrast (3 min). Ask: what was happening / what happened? Elicit one pair of sentences from the extracts and put them side by side. Two examples, not a rule table.
- Set a task with a real need for the form (4 min). An alibi or 'when the power cut happened' task, where each partner must establish what the other was doing at three different times. The form is the only efficient way to get the information.
Reasoning: the change is from practice of the form to use of the form to do something, plus noticing drawn from the learners' own production rather than the teacher's explanation. Noticing their own gap creates the reason to restructure; a rule on the board does not, because they already know the rule.
3. What evidence would convince you the change had worked, and what evidence would not?
Read the worked answer
Would convince me:
- The same task type, recorded at two points, showing the target form used unprompted in a meaning-focused exchange, not just accurately when elicited.
- Evidence that learners repair: they start a past simple, hesitate and reformulate to past continuous without being corrected.
- Longer turns and fewer breakdowns in communication, i.e. the form is doing work for them.
Would not convince me:
- A higher gap-fill score. That is the measure that already looked good while the problem persisted.
- Learners producing the form when the teacher is standing next to them or when the prompt makes it obligatory. That is a monitoring effect, not availability.
- One lesson's data. Proceduralisation is gradual and the first attempt often looks worse, because learners take risks and make more errors than when they play safe with past simple.
Judgement point: I would set a success criterion in advance (for example, the target used at least once unprompted by 12 of 22 learners in a repeated task by week three) so that I am not just collecting impressions that confirm the change was a good idea.