Accumulation, transfer, and maturation in bilingual toddlers’ word learning

Bilingual children typically must learn two labels for each concept they encounter, one in each language, effectively doubling the number of words they acquire relative to monolinguals. While bilinguals typically know fewer words in each language than monolinguals, they do not show the delay that would be expected if learning double the words required double the time. Three mechanisms may explain this: the accumulation of language-specific experience with labels (operationalized as language-adjusted word frequency), transfer across and within languages, and maturational changes in learning efficiency. The present studies investigate the interplay of these factors using computational simulations (Study 1) and models of empirical data from toddlers aged 16-30 months acquiring English and/or French (Study 2). Simulations progressively incorporated language exposure and transfer between translation equivalents, establishing that accumulator models require some degree of transfer to produce realistic bilingual vocabulary trajectories. Empirical models first replicated accumulator models with monolingual data, then extended them to bilinguals, confirming that concept frequency, language exposure, and age all predicted word learning, with age emerging as the strongest predictor and modulating the effects of the other two. Moreover, in bilingual children, concept frequency was a stronger predictor of word production than language-adjusted word frequency. Language exposure also independently predicted word production, suggesting that broader language experience supports acquisition beyond the accumulation of experience with individual labels. These findings indicate that accumulation, transfer (both cross-language and within-language), and maturation interact, enabling bilingual children to learn words from two languages without requiring twice the experience.