Talkative Learning Opportunities are Nested Within Everyday Activities

• In Event: Lexical Development: environment and mechanisms

Abstract

Children encounter words throughout the everyday business of playing, eating, reading, and more. Ample evidence suggests that each activity presents a distinct environment of what and how many words are available (e.g., Hoff-Ginsburg, 1998; Snow et al., 1976; Soderstrom & Wittebolle, 2013; Tamis-LeMonda et al., 2018). The fact that certain activities (e.g., reading) are more talkative than others (e.g., eating) grounds the prediction that a child’s progress along a curve of cumulating word exposures is constrained by the prevalence of each activity in their daily life (see also Montag et al., 2018). Empirically testing this prediction has been muddled by (1) estimating activity prevalence without sampling densely throughout children’s lived experiences (Franchak, 2019) and (2) identifying ‘talkative’ activities based on short, isolated language samples rather than extended samples characterized by interleaved silence and speech (Cristia et al., 2020; Tamis-LeMonda et al., 2017). Here, we directly address these limitations by sampling both everyday activities and everyday language quantities at scale.

43 English-speaking families with young children (M=5.67, SD=0.72 years) from a larger multi-part investigation (Romeo et al., 2018) completed both activity logging and audio recording at home for two pre-assigned days. For each hour of each sampled day, caregivers reported one or more of the activities listed on their paper-and-pencil log (LENATM; see Figure 1). They also recorded up to 16 hours of each day’s at-home audio using a LENA Pro digital language processor that automatically estimated adult word, child vocalization, and conversational turn quantities (Ford et al., 2008). 1083 of 1376 possible hours (43 families x 2 days/family x 16 hours/day) comprise the analyzed dataset of everyday hours for which caregivers reported one or more activity and audio recorded the entire hour.

Figures 1 and 2 show that one of the most prevalent and talkative activities in children’s everyday lives is play. Overall, children spent the bulk of their time playing, eating, watching TV, sleeping, in group events, and riding in a car (Figure 1). Children encountered dense quantities of words during some less frequent activities like reading, getting ready for bed, and getting dressed, in addition to playing and eating (Figure 2). Talkative activities tended to be dense with adult words, child vocalizations, and conversational turns. The striking variation in language use across different hours of the same activity – spanning zero to many thousands of words per hour – suggests that theories about everyday word learning opportunities will benefit from detailed specification of the multi-dimensional structure of activities (people, places, times, objects, and more) giving rise to word use.

Together with other quantifications of language use outside of researcher-constrained activities (e.g., Rowe, 2012; Roy et al., 2015; Tamis-LeMonda et al., 2018), these data are now available to ground computational theories about how children’s everyday sensory histories matter for building knowledge. We will also share ongoing efforts to model covariation among several dimensions of everyday speech and activities at multiple, nested timescales in this large corpus.

Authors

• Andrea Imhof, University of Oregon
  Presenting Author

• Heather T Anderson, University of Oregon
  Non-Presenting Author

• Rachel R. Romeo, University of Maryland - College Park
  Non-Presenting Author

• Meredith L. Rowe, Harvard University
  Non-Presenting Author

• John D.E. Gabrieli, Massachusetts Institute of Technology
  Non-Presenting Author

• Caitlin Fausey, University of Oregon
  Non-Presenting Author