Steven L. Elmlinger

My main research focus concerns the role of parent-child interaction in early acquisition of speech and language. I study human infants' and young children's language and communicative development with a particular interest in how perceptual and neural mechanisms give rise to learning in social environments. I investigate this question in three lines of research:

Infants change their own learning environment. What is the function of infants’ immature behavior in early development? My findings suggest that infants’ immature vocalizations elicit simplified linguistic responses from their caregivers. Speech contingent on early vocalizations contains shorter utterances and fewer unique words than non-contingent, but equally child-directed, speech (Elmlinger et al., 2019, 2021). Using large-scale corpus analysis of parent-child turn-taking, I found that this simplification is a stable pattern across 13 languages and five language families (Elmlinger et al., 2025). For example, Tseltal Mayan mothers, whose socialization style is less child-centric than is seen in the US, nevertheless simplify the content of their speech in response to children’s vocalizations (Elmlinger et al., 2023a). These findings suggest that infants’ immature vocal behavior functions to change the immediate environment in ways that make learning easier to do.

Infants learn to use the social effects of their voices. Do vocal learners have to learn that their voices are socially effective? By 5 months of age, human infants have come to expect that their own vocalizations will elicit changes in the behavior of others around them. In a still-face interaction, 5-month-olds, but not 2-month-olds, increased their vocal rate when a previously responsive adult stopped responding. This burst in vocalizations is a proxy for infants' learned social expectation for responsive social partners. In line with that, I found that the extent of caregiver responsiveness during free play interactions with infants predicted the magnitude of infants' vocal burst during still face (Elmlinger et al., 2023b). The power that vocalizations have over the environment guides what is learned next. In a comparative analysis of vocal development in human infants and songbirds, I found evidence that both of these immature learners use social feedback to navigate their transition into more efficient vocal production. In both taxa, social feedback to immature vocal sequences was related to more efficient production as measured by change in syllables per second over the course of development (Elmlinger et al., 2026). These findings suggest parallel social feedback learning mechanisms that operate across two distantly related taxa.

Neural and perceptual foundations of socially guided learning. Children's voice continues to influence their social environment well beyond the prelinguistic stage. I show evidence of this in parents' brains during real-time interaction with their children. Using functional near-infrared spectroscopy (fNIRS), I found that parents have a rapid neural sensitivity to the communicative maturity of their children's speech, differentiating multiword utterances from babbling (Elmlinger et al., under review). In my current work, I study the precursor skills that are required to learn to talk, that is, to adjust communicative output based on what is effective in the immediate social environment. First, infants need to learn auditory-vocal integration: adjustments to vocal production that reflect the auditory environment. I recently found that 3- and 5-month-olds raise their voices when background noise is present, providing the earliest evidence of the Lombard effect in human infants (Elmlinger et al., in prep). Another capacity is vocal controllability: detecting that one's own voice elicits changes in the environment. In an fNIRS paradigm in which infants' vocalizations trigger visual stimuli, pilot data suggest that infants rapidly build expectations about these effects, as evidenced by increased neural activity when their expectations are violated (Elmlinger & Lew-Williams, in prep). The hypothesis is that these capacities provide a foundation for learning from the social environment. My future research will use these approaches in longitudinal studies to investigate neural development, experiences of communicative success, and later language outcomes.