Linguistic Complexity and Theory of Mind Development in Early Childhood: A Longitudinal Study

Introduction

The developmental relationship between language and theory of mind—the capacity to attribute false beliefs and mental states to others—has emerged as a central question in cognitive development (de Villiers & de Villiers, 2000; Astington & Baird, 2005). While correlational evidence robustly documents associations between language ability and ToM reasoning (Milligan et al., 2007), the causal and mechanistic nature of this relationship remains contested. Competing theoretical accounts propose that linguistic development merely enables expression of preexisting ToM concepts, that language provides the representational scaffolding necessary for ToM reasoning, or that both domains develop semi-independently with domain-specific constraints. Longitudinal investigations distinguishing these mechanisms are sparse, particularly with repeated measurement over extended developmental windows.

The current investigation leverages quarterly assessments of language and false-belief reasoning across 18 months of toddlerhood. We specifically examine whether linguistic complexity—beyond vocabulary breadth—predicts subsequent ToM achievement. We hypothesized that receptive vocabulary would predict ToM onset, consistent with lexical access accounts, but that syntactic complexity (particularly use of subordinate clauses and sentential complements) would provide additional predictive power, supporting representational scaffolding models.

Method

Participants

One hundred fifty-six typically developing children (M age at baseline = 32.1 months, SD = 7.3; 47% female) were recruited from community centers and preschools in the Ottawa metropolitan area. Inclusion criteria were: (a) chronological age 18–48 months, (b) English as primary home language, (c) no diagnosed developmental disorder or hearing impairment, and (d) parental consent. Attrition was minimal (9.6%) with 141 children completing the 18-month protocol. Demographic characteristics: 72% of mothers held bachelor's degrees; median household income CAD $85,000–$95,000; 68% of families identified as European descent, 12% South Asian, 11% East Asian, 9% mixed or other.

Procedure

Children participated in quarterly assessment sessions (baseline + 5 additional sessions) spanning 18 months. Each session lasted approximately 45 minutes and included language testing, false-belief reasoning tasks, and standardized cognitive screening. The Peabody Picture Vocabulary Test-III (PPVT-III) assessed receptive vocabulary. Expressive language and syntactic complexity were measured via spontaneous speech samples (20 minutes of parent-child interaction) transcribed verbatim and analyzed using CHILDES (MacWhinney, 2000) to extract mean length of utterance (MLU), type-token ratio, and frequency of subordinate clauses per 100 utterances. False-belief reasoning was assessed using Wimmer and Perner's (1983) change-of-location task and, at later sessions when children reached sufficient linguistic maturity, Sally-Anne and unexpected contents tasks. Cognitive ability was screened using the Mullen Scales of Early Learning to control for non-linguistic developmental factors. All testing was video-recorded with parental consent and scored by trained, blind coders (inter-rater reliability ICC ≥ .88 across measures).

Results

Hierarchical linear modeling (HLM) with time nested within participants examined longitudinal trajectories. At baseline, receptive vocabulary (M = 487 words, SD = 198) and false-belief task performance (29% pass rate) were substantially correlated (r = .51, p < .001). The primary HLM model predicting false-belief success included baseline age (centered at 32 months), PPVT-III score, subordinate clause frequency, and their interactions with time. Results showed significant effects of age at baseline (γ = 0.08, SE = 0.02, t = 4.2, p < .001) and PPVT-III score (γ = 0.12, SE = 0.03, t = 4.1, p < .001). Critically, subordinate clause frequency predicted incremental variance in ToM success beyond vocabulary (γ = 0.06, SE = 0.02, t = 3.2, p = .002). An interaction emerged between time and syntactic complexity (γ = 0.04, SE = 0.02, t = 2.1, p = .038), indicating that syntactic gains predicted steeper improvements in false-belief reasoning across the 18-month period. Cross-lagged analyses revealed that PPVT-III at time t significantly predicted false-belief task success at time t+1 (β = 0.19, 95% CI [0.12, 0.26]), and subordinate clause use at time t predicted ToM improvement from time t to t+1 (β = 0.14, 95% CI [0.07, 0.21]), even controlling for concurrent false-belief performance. Effect sizes (Cohen's d) for vocabulary effects ranged from 0.32–0.58; for syntactic effects, 0.24–0.41.

Discussion

This 18-month longitudinal investigation provides evidence for domain-specific linguistic constraints on theory of mind reasoning. The predictive relationship between receptive vocabulary and subsequent false-belief success suggests that lexical access to mental-state terms (e.g., "think," "want," "know") provides a necessary substrate for ToM reasoning. However, the additional independent contribution of syntactic complexity—particularly subordinate clauses—implies that linguistic structure itself, not merely vocabulary, scaffolds representational reasoning. Sentences with complement clauses (e.g., "She thinks that...") may provide practice with embedded propositional representations essential for entertaining false beliefs.

These findings align with de Villiers and de Villiers' (2000) syntactic bootstrapping hypothesis while acknowledging that language-ToM correlations may partly reflect shared reliance on general cognitive capacity. The residual effect of subordinate clause use independent of vocabulary suggests domain-specific linguistic contributions rather than purely general cognitive mechanisms. Future investigations should examine whether language intervention studies produce subsequent improvements in ToM reasoning, providing experimental evidence for causality. Cross-linguistic studies comparing languages with different propositional structure would further illuminate whether observed relationships reflect universal developmental mechanisms or language-specific scaffolding.

References

• Astington, J. W., & Baird, J. A. (Eds.). (2005). Why language matters for theory of mind. Oxford University Press.
• de Villiers, J. G., & de Villiers, P. A. (2000). Linguistic determinism and the understanding of false beliefs. In P. Mitchell & K. J. Riggs (Eds.), Children's reasoning and the mind (pp. 191–228). Psychology Press.
• MacWhinney, B. (2000). The CHILDES Project: Tools for analyzing talk (3rd ed.). Lawrence Erlbaum Associates.
• Milligan, B., Astington, J. W., & Dack, L. A. (2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78(3), 622–646.
• Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13(1), 103–128.
• Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a "theory of mind"? Cognition, 21(1), 37–46.
• Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences, 28(5), 675–735.
• Gleitman, L. R., & Papafragou, A. (2005). Language and thought. In K. J. Holyoak & R. G. Morrison (Eds.), The Cambridge handbook of thinking and reasoning (pp. 633–661). Cambridge University Press.