Background: Language development is critical for children's life chances. Promoting parent-child interactions is suggested as one mechanism to support language development in the early years. However, limited evidence exists for a causal effect of parent-child interactions on children's language development. Methods: Data from the Language in Little Ones study, an Australian prospective birth cohort study (n = 296), was used to determine the sustained effect of parent-child interactions over time on children's language development at 36 and 48 months, measured using the Clinical Evaluation of Language Fundamentals Preschool-2 (CELF-P2) language assessment. Marginal structural models and inverse probability of treatment weights were used to allow observational data to emulate a randomised controlled trial by accounting for time-varying exposures and confounding. These results were then used to estimate the effect of several hypothetical scenarios where the exposure was fixed for the whole population at different levels (5th, 25th, 50th, 75th and 95th percentile) across the observed distribution of parent-child interactions. Results: Findings supported a causal effect of parent-child interactions from 6 to 36 (or 48) months on children's language development at 36 and 48 months, in a population of children without language impairment. The counterfactual language score at 48 months increased from 97.21 (95% CI 96.86, 97.56) for the scenario fixed at the 5th percentile to 102.15 (95% CI 101.80, 102.50) at the 50th percentile and 111.41 (95% CI 111.06, 111.76) at the 95th percentile. Conclusions: Although the effects of parent-child interactions on later language were small they do offer one mechanism to support early language development. These findings are discussed within the context of existing interventions to highlight the value of investment into sustained, universal prevention efforts for supporting early language.