That most errors come from L1 interference is not self-evident. Error analysis showed many errors are developmental and shared across L1s. Transfer is strongest in phonology, weakest at discourse level.
Claim
The L1-transfer assumption persists and is used to justify banning translation. But error analysis (Corder, Richards, 1960s–70s) found many errors are developmental: overgeneralisation, false hypotheses. French (1949, Common Errors in English) already argued that common errors have common causes.
Mechanism / theory
- Contrastive analysis vs error analysis vs interlanguage (Selinker).
- Swan & Smith, Learner English: transfer is greatest between similar languages.
- Cross-linguistic influence (CLI) research: Cenoz, Ringbom (comment).
Classroom moves
- Classify errors by source: L1 transfer / developmental / teaching-induced / fossilised. Treat each differently.
- Prioritise global over local errors (intelligibility).
- Watch for teaching-induced errors: e.g. “What he does for a living?” echoing teacher talk; over-drilling present progressive.
- Test: does the error recur across L1 groups? If so, it’s probably developmental.
Limits & critique
- Jaworski (comment): ~99% of his Turkish learners’ errors look like transfer. Thornbury: the same errors appear in Spanish and Arabic learners.
- The post doesn’t separate grammar, lexis and phonology.
Apply in my context (Claude)
- Uzbek (and Russian, for many Tashkent learners) lack articles. Article omission is therefore both transfer-prone and developmentally late for all L1s, so expect persistence and don’t over-correct at Starters–KET.
- Uzbek is SOV with postpositions and no grammatical gender. Word-order errors in questions are more likely developmental (see Teach grammar in processability order (Pienemann)) than transfer.
- Pronunciation (vowel length, /w/–/v/, /θ ð/) is where transfer dominates. Spend corrective effort there.