The relationship between infants' preference for and learning from adult-directed speech

• In Event: Poster Session 08
  In Poster Session: PS 08 Topic - Language, Communication

Abstract

The current study examines the interaction between children’s attention to and preference for certain kinds of input and the robustness of their learning from this input. Parents typically use an exaggerated register of speech, known as infant-directed speech (IDS), in communication with young infants. Young infants prefer listening to IDS relative to adult-directed speech (ADS; Kaplan et al., 1995). While this IDS preference may vary across development (Segal & Newman, 2015; The ManyBabies Consortium, 2020), even older children learn better from IDS relative to ADS (Ma et al., 2011). Bringing these findings together, using a novel gaze-contingent eye-tracking paradigm, the current study examines the extent to which maternal input and children’s preference for IDS relative to ADS influence children’s learning of novel word-object associations from the two registers.
We tested 48 18- to 24-month-old children in a gaze-contingent preference task and a word-object association learning task. We recorded mothers producing IDS and ADS. In the gaze-contingent preference task, children were presented with still images of two adults on the screen. Children could trigger one of the adults to start speaking with their eyes by fixating one of them for longer than a second. One adult always produced IDS, the other always produced ADS (counterbalanced across children), such that systematic differences in children’s triggering of the different speakers provided an implicit measure of their preference for either register. Subsequently, children were presented with different novel objects and their novel labels. The labels for half of the objects were presented in IDS, the labels for the other objects were presented in ADS (again, counterbalanced). At test, children saw two objects and heard the labels for one of these objects, i.e., the target object, while their eye-movements were recorded to examine the pattern of target fixations upon label presentation.
First, we analysed maternal input and found prosodic differences in maternal IDS and ADS. Second, we analysed whether children systematically triggered the video of the adult speaking IDS relative to the adult speaking ADS and found no systematic differences. With regards to the word learning task, we examined whether children showed differences in their recognition of novel word-object associations that had been introduced in IDS or ADS at training. We found that children learned novel word-object associations from both IDS and ADS. A generalised linear mixed effects model found a positive relationship between the proportion of ADS videos triggered by children and their learning of word-object associations presented in ADS. A similar relationship was not found for IDS. We did not find any relationship between maternal input with children’s preference and their word learning. Together, these results shed light on how the input a child perceives and the preferences a child forms can influence a child's word learning behaviour early in life.

Authors

• Sarah Eiteljoerge, University of Goettingen
  Presenting Author

• Vivien Outters, Georg-August-Universität Göttingen
  Non-Presenting Author

• Jana Schildknecht
  Non-Presenting Author

• Nivedita Mani, Georg-August-Universität Göttingen (GAU)
  Non-Presenting Author