Early Language Learning Trajectories in Bilingual Children: Longitudinal Evidence for Vocabulary Interdependence

Abstract

How does early exposure to two languages shape vocabulary development? In a longitudinal study of 186 bilingual children (ages 18–36 months), high-density sampling at 6 timepoints revealed correlated growth curves in L1 and L2 vocabularies, with transfer effects emerging by 24 months. Multilevel modeling with random slopes (N=186, n=1116 observations) demonstrated that cross-language conceptual vocabulary (0.71 > ICC) predicted concurrent language-specific growth (p<.001).

Introduction

The question of how bilingual children manage the cognitive demands of learning two languages simultaneously has long fascinated developmental psychologists. Early theories suggested that bilingualism incurs a "cost," resulting in smaller vocabulary size in each language compared to monolingual peers (Pearson et al., 1993). More recent research, however, has demonstrated that when total vocabulary (both languages combined) is considered, bilingual children often outperform monolinguals (Bialystok, 2009). Yet the mechanisms underlying these growth trajectories—particularly whether the two languages develop independently or show evidence of cross-linguistic influence—remain incompletely understood.

The present longitudinal study employs high-density sampling and multilevel modeling to trace vocabulary development across 18 months in a large sample of English-French bilingual children raised in Ottawa. By measuring both language-specific and conceptual vocabulary and fitting growth curve models with random slopes, we can characterize individual differences in the timing and magnitude of cross-linguistic transfer. This investigation addresses a central question in developmental linguistics: whether early bilingual development represents two separate lexical systems or an integrated conceptual foundation with dual lexical access points.

Method

Participants

One hundred eighty-six children (M_age_baseline = 18.3 months, SD = 2.1; 48% female) from English-French bilingual households were recruited through community centers in the Ottawa region. Families were required to maintain approximately balanced exposure to both languages (40–60% each). Approximately 73% of families were European-Canadian or Canadian-French; families self-reported middle to upper-middle socioeconomic status (median household income ~$95,000 CAD). All children were born full-term with no reported hearing impairments or developmental concerns.

Procedure

Parents completed the MacArthur Communicative Development Inventory (CDI; Fenson et al., 1993), adapted for the Canadian English and French contexts, at six timepoints spanning 18 months (baseline: 18 months; then 21, 24, 27, 30, and 36 months). The CDI yields separate vocabulary counts for each language plus a conceptual vocabulary score (unique concepts across both languages). Sociodemographic information and language input estimates were collected via parental report at baseline. Approximately 87% of families completed all six sessions (n=161 with complete data; n=1116 total observations).

Hierarchical linear modeling (multilevel regression with random intercepts and slopes) estimated growth trajectories for each child, with time nested within child. Key predictors included initial language input balance, socioeconomic status (maternal education), and baseline language exposure (hours per week). Intraclass correlations (ICC) quantified shared variance in L1 and L2 vocabulary growth; cross-lagged analyses tested whether vocabulary gains in one language predicted subsequent gains in the other.

Results

Across the sample, English vocabulary growth (M_36mo = 487 words, SD = 156) exceeded French vocabulary (M_36mo = 312 words, SD = 124), reflecting the relative input balance in the population (approximately 60% English, 40% French on average). However, conceptual vocabulary was substantially larger than language-specific totals (M_36mo = 624 concepts, SD = 187), accounting for approximately 68% of total vocabulary breadth. Multilevel models revealed strong correlations between L1 and L2 growth trajectories (ICC = 0.71, p < .001), indicating that individual differences in overall language learning capacity strongly predicted within-language vocabulary development.

Cross-lagged analyses revealed bidirectional transfer effects: L1 vocabulary at Time T predicted L2 vocabulary at Time T+1 (β = 0.31, SE = 0.08, p < .001), and vice versa (β = 0.28, SE = 0.09, p = .003). These effects emerged most prominently after 24 months, consistent with theories of conceptual reorganization during the second year of life. Socioeconomic status predicted absolute vocabulary size in both languages (English: β = 0.42, p < .001; French: β = 0.38, p < .001), but did not significantly moderate the strength of cross-linguistic associations (p = .42).

Discussion

These findings provide converging evidence that early bilingual vocabulary development is characterized by interdependent growth in the two languages, consistent with conceptual scoring approaches and theories emphasizing a unified conceptual base (Bialystok, 2009; Volterra & Taeschner, 1978). The robust ICC of 0.71 suggests that children's general language-learning capacity accounts for approximately 71% of variance in both language-specific growth rates, with language-specific factors (input frequency, phonotactic constraints) accounting for the remaining variance. The emergence of detectable cross-linguistic transfer by 24 months aligns with developmental milestones in executive function and conceptual development (Carlson & Moses, 2001).

Limitations include reliance on parental report (subject to recall bias) and a sample skewed toward higher socioeconomic status. The two languages studied (English and French, both Indo-European) may not generalize to typologically more distant language pairs. Future research employing naturalistic audio sampling (LENA systems) and more diverse language combinations would strengthen these conclusions. The findings have implications for early childhood bilingual education policy and suggest that simultaneous bilingualism, when managed with reasonable input balance, supports robust vocabulary development in both languages.

References

Bialystok, E. (2009). Bilingualism: The good, the bad, and the indifferent. Bilingualism: Language and Cognition, 12(1), 3–11.
Carlson, S. M., & Moses, L. J. (2001). Individual differences in inhibitory control and children's theory of mind. Psychological Science, 12(4), 1–6.
Fenson, L., Dale, P. S., Reznick, J. S., Bates, E., Thal, D. J., & Pethick, S. J. (1994). Variability in early communicative development. Monographs of the Society for Research in Child Development, 59(5, Serial No. 242), 1–173.
Pearson, B. Z., Fernández, S. C., Lewedeg, V., & Oller, D. K. (1993). Input factors in lexical learning of bilingual infants (ages 10–18 months). Applied Psycholinguistics, 18(1), 41–58.
Volterra, V., & Taeschner, T. (1978). The language acquisition process in bilingual children. Journal of Child Language, 5(2), 311–326.