Caregiver Behavior That Facilitates Vocal Turn-taking With Prelinguistic Infants

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Language, Communication

Abstract

Early turn-taking interactions, especially the number of conversational turns, predict later language skills (e.g., Romeo et al., 2018). However, we know little about the time course of infant vocalizations within a turn-taking bout. How do different caregiver responses influence the likelihood of infant vocalizing again, and thus extending the turn-taking interactions? Here, we 1) categorized parental responses to infant vocalizations as vocal or non-vocal, as well as sensitive or redirective, 2) calculated the babbling elicitation rate of caregivers’ responses, 3) plotted the response latency (RL) distribution of infant responses to their parents, 4) examine the length of the full turn-taking interaction, and 5) explore the influence of vocal turn-taking on infant vocal quality.

Fifty-one mothers and their 9-month-old infants participated in a 15-min unstructured play session. The dyads were video-recorded and asked to play as at home in a 3.7-m x 5.5-m playroom containing infant toys. We coded the modality and sensitivity of caregiver contingent responses to infant vocalizations (Table 1). Infant vocalizations that occurred within 2 sec before or after the end of the caregiver response were counted as elicited. Infant RL was calculated by subtracting caregiver response offset from infant vocalization onset. Elicitation rate was calculated as the number of elicited infant vocalizations divided by the total number of caregiver responses.

We found that caregiver vocal responses had a higher elicitation rate than non-vocal responses, and sensitive responses than redirective responses. We tested the modality and RL effects as well as sensitivity and RL effects with mixed-effects linear regression models in R (Bates, Maechler & Dai, 2009) of elicitation rate. Type III sum of squares analyses on both models revealed main effects of modality (vocal or non-vocal), sensitivity (sensitive or redirective), RL (9 time bins) and interactions between RL and modality as well as RL and type (all p < 0.0001). Infant RL to caregiver vocal (M = 0.36, SE = 0.05) or sensitive behaviors (M = 0.14, SE = 0.09) peaked at 0-0.5s , while that to non-vocal (M = 0.12, SE = 0.22) or redirective (M = 0.22, SE = 0.19) behaviors was distributed uniformly (Figure 1).

We found that caregivers who respond verbally to their infants and who respond sensitively are more likely to get infants to continue a vocal turn-taking interaction. Our next (preregistered) steps are to understand what extends these turn-taking interactions. First, we will identify and examine the full sequence of turns in each turn-taking bout to determine the average number of turns per bout. Second, we will compare the duration and average number of turns per turn-taking bout as a function of caregivers’ initial vocal versus non-vocal responses. Third, we will assess the role of vocal turn taking in prelinguistic vocal learning, since infant vocal quality tends to become more mature as a result of parental responding (e.g., Goldstein & Schwade, 2008). We will examine whether infants increase consonant production after different caregiver responses. The data are already collected, and these registered analyses will be completed by the meeting.

Authors

• Hanwen Vivian Vivian Zhang, Cornell University
  Presenting Author

• Steven L Elmlinger, Cornell University
  Non-Presenting Author

• Rachel R Albert, Lebanon Valley College
  Non-Presenting Author

• Michael H. Goldstein, Cornell University
  Non-Presenting Author