Contingencies in Maternal Speech and Infant Vocalizations during Naturalistic Play from 4 to 12 Months

• In Event: Poster Session 09
  In Poster Session: PS 09 Topic - Parenting & Parent-Child Relationships

Abstract

Infant language learning depends on temporal and sequential regularities between speech and other joint actions and events in shared social environments, which can be optimized by responsive caregivers during joint engagement. For example, different aspects of maternal responsiveness predict child language milestones (Tamis-LeMonda, Bornstein, & Baumwell, 2001), and maternal responsiveness across the first year structures dyadic interactions to scaffold the development of turn-taking behavior (Leonardi, Nomikou, Rohlfing, & Rączaszek-Leonardi, 2016). Such reciprocal, responsive interactions between parents and infants may create opportunities for learning; notably, contingent responses may increase infants' vocalizations, which have been related to positive language outcomes (Warlaumont, Richards, Gilkerson, & Oller 2014).

In the current study, we use sequential analyses to quantify reciprocal contingencies between infant vocalizations (IV), maternal utterances (MU), and maternal hand-object configurations. Our goal is to identify non-random sequential dependencies that could form exploitable regularities in a dyadic system and promote infant vocalization. Speech and vocalizations of 41 mother-infant dyads, recorded in their homes for a longitudinal study, were transcribed during a 10-minute home session that included free play and pointing segments. We examine whether the frequency or latency of infants' vocal responses differ based on maternal verbal content or object-manipulating actions (Chang, de Barbaro, & Deák, 2016). We also compare contingency statistics between maternal speech, maternal and infant manual actions, and IVs, at 4, 6, 9, and 12 months to characterize developmental trajectories of dyadic interactions.

We found proto-turntaking from four months onward: 3-second time window sequential analyses revealed that IVs follow MUs significantly more than expected (from IV base rates) in nearly all dyads. Odds ratios for 36 dyads at month 4 (and for all dyads at older ages) are >1, meaning that an IV is more probable within 3 seconds after a MU than at other times. Yule’s Q correlation coefficients (based on odds ratio statistics) are shown in Figure 1. They demonstrate strong positive associations between IVs and preceding MUs. By comparison, Yule’s Q values for IVs and preceding maternal hand-object configuration changes within 3 seconds show weak negative correlations (mean Q = -0.25), suggesting that maternal speech drives vocalizing more than changes in maternal object handling during free play.

Moreover, because IV rate increases with age, the mean latency of infant vocalizations that occur contingently (5 seconds after MU onset, preceding MU mean duration = 1.2 seconds, so there is 3 seconds after offset for most MUs) increases. Bootstrapping tests for latency means at each month showed that infants’ contingent responses occur significantly later at 8, 9, and 12 months of age than at 4 and 6 months (Figure 2), suggesting that infant responses become less reflexive, and perhaps more flexible, with age.

Ongoing work is extending analyses to find whether some types (see Chang & Deák, 2019) of maternal utterances predict IVs more than others, and investigating how infant manual actions predict IVs. Finally, we will relate these contingencies to standardized language test outcomes (MBCDI, BSID-3, LUI) at 22 months.

Authors

• Betina Karshaleva, University of California - San Diego
  Presenting Author

• Gedeon O Deak, University of California - San Diego
  Non-Presenting Author