The Active Child: Object Play leads to Word Learning

• In Event: Reorienting the study of language learning using naturalistic data

Abstract

Infants actively explore their environment and learn through object play. While playing with objects, infants create rich, perceptual, multimodal feedback in which object properties become simultaneously salient to vision and touch. In turn, responsive caregivers often exploit children’s manual actions by naming or describing the object in the child’s hands (Tamis-LeMonda, Kuchirko & Song, 2014). Optimal naming moments support learning object names in the lab (Pereira et al., 2014); but the association between infants’ object play, caregiver’s naming, and word learning “in the wild” has not been examined. Here we asked: are the objects that infants spontaneously manipulate at home also named by mothers, and are infants likely to learn the words for those objects?

We conducted 2-hour home observations of 18- and 23-month-old infants (N=32) and related infants’ object play to their mothers’ language input and the words in infants’ vocabularies according to the MacArthur-Bates Communicative Development Inventory (MCDI). Further, we asked whether data based on the objects of infants’ play at home predicted the age of acquisition of words by over 5,000 children hosted on Wordbank (Frank et al., 2016). We focused on 271 concrete nouns on the MCDI and coded frame-by-frame in Datavyu: (1) if the child touched the word’s referent during the naturalistic recording; (2) if mother produced the word during the visit; (3) if the infant produced the word during the visit.

Findings revealed that infants heard words for the objects they touched. Specifically, infants heard the words for 68% of the objects that they manipulated during the visit. Figure 1A shows that mothers named on average 21.94 objects out of 32 unique objects that infants touched during the naturalistic recording. Overall, the objects that infants frequently manipulated at home were also objects that mothers frequently named, r = 0.53, p<.001.

Furthermore, words for the objects that infants frequently touched tended to be part of their vocabularies at the group level, r = 0.44, p<.001. Nonetheless, infants manipulated many objects that were not yet part of their vocabularies, and therefore any given infant was equally likely to know and not know the words for the objects they touched. Figure 1B shows infants’ vocabularies included, on average, the words for 15.53 out of 32 unique objects that they touched. Relatedly, the words for frequently touched objects are generally acquired at earlier ages, based on mapping of play data to Wordbank data, r = -0.43, p<.001 (Figure 2). Therefore, although infants may not know the words for all the objects they touch, the words for frequently manipulated objects (such as book, ball, and cup) tended to be acquired earlier by infants observed at home and by infants in the Wordbank repository.

Findings suggest that infant play involves a select set of objects, whose names are frequently produced by mothers and then learned by infants. Most objects of infants’ play are named by mother and/or included in infants’ vocabulary (Figure 1C). Ongoing analyses examine the relation between object play and infants’ in-the-moment object naming.

Authors

• Catalina Suarez-Rivera, New York University
  Presenting Author

• Emily Linn, New York University
  Non-Presenting Author

• Catherine S Tamis-Lemonda, New York University
  Non-Presenting Author