Vocabulary learning is slow and limited by working memory. Fewer items, taught deeper, recycled more often. Chunking doesn’t raise capacity, but it makes each item more functional.
Conti’s targets (his synthesis, not study findings)
| Group | New items per lesson |
|---|---|
| Year 3 (age 7–8) | 2–3 |
| Year 4 | 3–4 |
| Year 5 | 4–5 |
| Year 6 (age 10–11) | 5 (occasionally 6) |
| Secondary | 6–10 (listening-heavy lessons at the low end) |
“Chunking does not allow us to teach more words — it allows us to teach language more effectively.”
Mechanism / theory
Working memory (Miller 1956; Cowan 2001; Gathercole & Alloway 2008); chunking (Wray 2002; N. Ellis 1996, 2003); word-knowledge dimensions (Nation 2001; Schmitt 2008); listening load (Field 2008; Vandergrift & Goh 2012); YL sources (Cameron 2001; Pinter 2017).
Procedure
Set item counts by age; teach as chunks (“I like football”, “on the table”); keep listening-heavy lessons light; teach transparent cognates in bigger batches; recycle across modalities.
Limits & critique
- The numbers are his synthesis. Don’t cite them as findings.
- “Extra items don’t increase retention” is unsourced.
- Partly experience-based (18 years with ages 7–10).
- Transfer caveat: written for UK secondary MFL (French/Spanish/German) with ~2–3 lessons a week and minimal out-of-class exposure. Tashkent EFL learners get far more ambient English (YouTube, games), so input-scarcity arguments apply less strongly. (Claude)
Apply in my context (Claude)
- Audit Kid’s Box / Prepare pages: a typical YL coursebook page introduces 8–12 items. By Conti’s targets that’s 2–3 lessons’ worth. So split the page, don’t rush it.
- The Cambridge YL wordlists are manageable at 3–4 items per lesson if every item comes back repeatedly. What makes the load unrealistic is low recycling, not list length.