Symmetries and Asymmetries of Age Effects in Naturalistic and Instructed L2 Learning

Barcelona Age Factor (BAF) project cohorts of school learners of English

Catalan/Spanish bilingual school students in Barcelona learning English as a foreign language, differing in the school grade at which English instruction began

The BAF project compares learners who began English instruction at different ages after equal numbers of hours of instruction - the comparison that matters for a school decision and that the naturalistic-immigration literature cannot make. Observational cohort comparison exploiting a curriculum change, not a randomised design.

The age effect in INSTRUCTED foreign-language learning is the opposite of the one in naturalistic immersion. At equal hours of instruction, later starters progress faster and outperform earlier starters on most measures; the early starters' advantage, where it appears, is on implicit/oral measures and is small. Munoz's explanation is that the input in a foreign-language classroom is too thin and too intermittent to engage whatever implicit mechanisms early naturalistic acquisition exploits - so a classroom start is a different treatment from an immersion start, not a smaller dose of the same one.

Medium. Starting age is set by curriculum cohort rather than by family choice, but the cohorts differ in calendar time and in the schools' broader programmes.

The core BAF result - no long-run advantage from an earlier classroom start at equal hours - is reproduced in the Swiss (Pfenninger & Singleton) and German (Jaekel et al.) studies, and contradicted at a longer horizon by Porsch et al. (2023). Recorded as mixed.