You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster C25, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

Sentence prosody cues object category learning at 6 months

Claudia Männel1,2, Maria Teixido3, Laura Bosch3, Angela D. Friederici1, Manuela Friedrich1,4;1Max Planck Institute for Human Cognitive and Brain Sciences Leipzig, 2University of Leipzig, 3University of Barcelona, 4Humboldt-Universität zu Berlin

In language acquisition, infants face the twin problem of segmenting the speech stream to extract linguistically relevant units and mapping these word-forms to their visual referents, the objects and events in the surrounding world. To complicate matters, infants need to grasp that this referential character extends to categories (e.g. different books are still called books). Behavioral research has revealed some evidence of infants’ simultaneous segmentation and mapping under simplified input conditions, if prosody favors this process (Shukla, White, and Aslin, 2011). However, it is unclear whether infants’ abilities hold for more ecological learning situations and whether infants can generalize their mapping abilities to category members. We evaluated 6-month-old infants’ abilities of mapping sentence-embedded labels to object categories by measuring event-related brain potentials (ERP) in a familiarization-test paradigm. During each familiarization phase, two pseudo-objects were followed by sentences containing a stressed pseudo-word as object label. Moreover, we manipulated the prosodic realization of object labels by placing them at different sentence positions. Infants (n = 64) either heard object labels in the beginning of the sentences (e.g. Dein Mukon ist wirklich schön [Your mukon is really pretty]) or at prosodically more salient sentence-end positions (e.g. Wir spielen mit dem Mukon [We are playing with the mukon]). During test phases, labels were either correctly or incorrectly combined with the pseudo-objects, thus testing whether infants had learned the previous label-object association. Crucially, we did not only present pseudo-objects seen during familiarization, but also new category members testing for generalization. Brain responses recorded during familiarization and test indicated that while label positioning in sentences did not affect infants’ word-form segmentation during familiarization, it impacted on infants’ mapping success at test. Only children, who had heard the object labels at the end of sentences, showed a main effect of mapping during test, apparent in more negative ERP responses at 300-700 ms to labels incorrectly versus correctly following the respective objects. In contrast, children who had heard the object labels at the beginning of sentences, did not show comparable test effects. These results imply that under prosodically salient conditions, infants not only learn previously presented label-object combinations, but also successfully apply this association to new category members. Thus, our study demonstrates for the first time that infants at 6 months can map labels to objects under ecological language learning conditions and, most importantly, show indications of object category learning.

Topic Area: Language Development

Back to Poster Schedule