I Knew That: Rules vs Retrieval
Ana, 34, scores 12/20 on the Use of English cloze section of a PEIC B1 mock. Eight of her errors involve prepositions or articles: 'on the morning', 'go to home', a missing 'the' before a specified noun, and 'in Monday'. When you review, she states the article rules correctly and immediately identifies the right answers, saying 'I knew that'. Her reading score is B1+ and her writing is clear, though she makes the same preposition errors in free writing. She has three weeks and five hours a week. She asks for a grammar book chapter on articles and prepositions, believing more rules will fix the problem. She is anxious about 'silly mistakes'.
1. Why is a grammar-book chapter on articles and prepositions unlikely to fix Ana's cloze score, despite her being able to state the rules?
Read the worked answer
Because her problem is retrieval in context, not declarative knowledge. She can state the rules and recognises correct answers on review, which shows the knowledge is stored but not available quickly enough during a cloze. Her free-writing errors with the same prepositions confirm that this is a production and automatisation gap, not a gap in explanation.
A grammar chapter adds more declarative knowledge of the kind she already has; it does not create the repeated, context-bound retrieval that turns 'I knew that' into 'I produced that'. The cloze also tests collocational chunks ('in the morning', 'on Monday', 'at home', 'go home'), which are learned as units rather than derived from rules each time.
2. Design a three-week plan (five hours per week) with proportions, and justify the largest component.
Read the worked answer
Three weeks, five hours each: 3 hours - collocation and chunk work around her actual errors (time expressions, place expressions, fixed phrases); 1 hour - timed cloze practice with a 90-second-per-item cap and an error log; 1 hour - targeted rule review only for patterns that remain unclear after the chunk work.
The largest component is chunk work because it directly addresses retrieval under time pressure and transfers to writing. Activities: sort 40 phrases into 'in/on/at', complete gapped sentences from her own writing, and do two-minute 'say the phrase' drills. Keep the grammar book as a reference for the one-hour slot, not the main course. This respects her request without letting it crowd out the practice that changes performance.
3. How should progress be measured, and why not by rule recitation?
Read the worked answer
Measure performance under time, not recitation. Two measures: (1) cloze accuracy on a parallel 20-item set with the same time limit; (2) target-error count per 100 words in a five-minute free-writing sample, scored blind against her baseline. Track both weekly and show her the trend.
Rule recitation is a poor measure because she already passes it while still making the errors; it gives false reassurance. A useful diagnostic is to ask her to correct her own free writing after 24 hours - if she can correct it then but not during writing, the issue is automaticity. The goal by week 3 is fewer target errors in production and a higher cloze score, not more rules quoted.