Sentence-Head Drill

The teacher (or a partner) supplies the opening of a semi-fixed frame (The thing is..., What I mean is..., It's worth..., As far as I'm concerned...); the learner completes it with real content on a real topic. Retrieval is the frame, personalisation is the filler. Trains production of discourse-organising chunks under the pressure of saying something true.

Procedure

1. Select 8–10 sentence heads. Pick frames that do discourse work: opening a claim, qualifying, exemplifying, contrasting, summarising. Examples: The thing is..., What I mean is..., It's worth noting that..., The problem with X is that..., What surprises me is..., I'd say..., If you ask me..., To be honest....
2. Agree a live topic. Something the class has genuine opinions on (a local news item, a disagreement from the previous lesson, a question about the coursebook topic). The topic is fuel; the heads are retrieval cues.
3. Round one, teacher-led (4 min): teacher reads a head; nominated learner completes it on the topic. One head per learner; speed matters.
4. Round two, pair drill (6 min): pairs alternate. A reads a head; B completes; swap. Keep the topic constant. Each pair cycles through the eight heads twice.
5. Round three, head-free recycling (4 min): partners continue the conversation on the topic without the list. Count which heads actually appear. The count is the diagnostic: frames that appear unbidden have started to belong to the learner.
6. Consolidation (optional): board the heads that did not appear. Next session's drill foregrounds them.

Why It Works

Discourse-organising chunks are high-frequency in speech and under-produced by learners who have met them only in reading. A sentence head sets up a grammatical and pragmatic commitment: once the learner has said The thing is..., the rest of the sentence has to cash that out. Retrieval is forced from the opener alone. Because the content is the learner's own, the frame gets loaded with personal meaning on first production, which is the condition under which it sticks. The drill is a compressed rehearsal of the Observe-Hypothesise-Experiment cycle Lewis (1993) proposed: the noticing has already happened; this is the experiment.

Variations

• Topic roulette. New topic each round, same heads. Frames prove productive across contents.
• Written first. Learners write one completion per head before speaking. Drops the anxiety floor; raises accuracy.
• Heads-in-context. Learners read a short opinion piece; the drill uses heads that the piece itself contained. Noticing primes retrieval directly.
• Extended heads. Instead of three-word openers, supply six-to-ten-word frames with a clear slot: One thing that really strikes me is how..., I've always found it hard to believe that.... Retrieval of a longer runway.
• Recorded drill. Learners record round two on phones. Self-listening catches the heads that came out unnaturally and the ones that flowed.

Tips

• The topic must be one the learners actually have views on. A borrowed-from-textbook prompt ("Is technology good?") produces borrowed language. A real disagreement produces real retrieval.
• Keep the head set stable for two or three lessons before rotating. Stability is what lets retrieval sediment.
• Resist the urge to correct head-by-head in round two. Correction happens in round-three consolidation, after retrieval has been tested.
• Pair heads with reformulation: note a learner completion that works well, reformulate one that did not, and re-use both in the next drill.
• Great paired with L1-Prompt Pattern Reproduction (retrieval from meaning) and Speed Pattern Retrieval (retrieval under time pressure).

Source

Nattinger & DeCarrico (1992) on lexical phrases and sentence stems; Lewis (1993, 1997); Thornbury (2004) Natural Grammar, which compiles keyword-anchored frames; McCarthy & O'Dell's English Collocations in Use volumes for source material on common heads. Theoretical frame: Pattern Retrieval.