Publications of Gliga, T.
Twelve-month-olds disambiguate new words using mutual-exclusivity inferences. Cognition. 2021;213(104691).
Twelve-month-olds disambiguate new words using mutual-exclusivity inferences
Representing objects in terms of their kinds enables inferences based on the long-term knowledge made available through kind concepts. For example, children readily use lexical knowledge linked to familiar kind concepts to disambiguate new words (e.g., “find the toma”): they exclude members of familiar kinds falling under familiar kind labels (e.g., a ball) as potential referents and link new labels to available unfamiliar objects (e.g., a funnel), a phenomenon dubbed as ‘mutual exclusivity’. Younger infants’ failure in mutual exclusivity tasks has been commonly interpreted as a limitation of early word-learning or inferential abilities. Here, we investigated an alternative explanation, according to which infants do not spontaneously represent familiar objects under kind concepts, hence lacking access to the information necessary for rejecting them as referents of novel labels. Building on findings about conceptual development and communication, we hypothesized that nonverbal communication could prompt infants to set up kind-based representations which, in turn, would promote mutual exclusivity inferences. This hypothesis was tested in a looking-while-listening task involving novel word disambiguation. Twelve-month-olds saw pairs of objects, one familiar and one unfamiliar, and heard familiar kind labels or novel words. Across two experiments providing a cross-lab replication in two different languages, infants successfully disambiguated novel words when the familiar object had been pointed at before labeling, but not when it had been highlighted in a non-communicative manner (Experiment 1) or not highlighted at all (Experiment 2). Nonverbal communication induced infants to recruit kind-based representations of familiar objects that they failed to recruit in its absence and that, once activated, supported mutual-exclusivity inferences. Developmental changes in children’s appreciation of communicative contexts may modulate the expression of early inferential and word learning competences.
Verbal labels modulate perceptual object processing in one-year-old children. Journal of Cognitive Neuroscience. 2010;22(12):2781-9.
Verbal labels modulate perceptual object processing in one-year-old children
It has been debated whether acquiring verbal labels helps infants' visual processing and categorization of objects. Using electroencephalography, we investigated whether possessing or learning verbal labels for objects directly enhances one-year-old infants' neural processes underlying the perception of those objects. We found enhanced gamma-band (20 to 60 Hz) oscillatory activity over the visual cortex in response to seeing objects whose names one-year-old infants knew (Experiment 1), or for which they had just been taught a label (Experiment 2). No such effect was observed for objects with which the infants were simply familiar without having a label for them. These results demonstrate that learning verbal labels modulates how the visual system processes the images of the associated objects, and suggest a possible route of top-down influence of semantic knowledge on object perception.
One-year-old infants appreciate the referential nature of deictic gestures and words. Psychological Science. 2009;20(3):347-53.
One-year-old infants appreciate the referential nature of deictic gestures and words
One-year-old infants have a small receptive vocabulary and follow deictic gestures, but it is still debated whether they appreciate the referential nature of these signals. Demonstrating understanding of the complementary roles of symbolic (word) and indexical (pointing) reference provides evidence of referential interpretation of communicative signals. We presented 13-month-old infants with video sequences of an actress indicating the position of a hidden object while naming it. The infants looked longer when the named object was revealed not at the location indicated by the actress's gestures, but on the opposite side of the display. This finding suggests that infants expect that concurrently occurring communicative signals co-refer to the same object. Another group of infants, who were shown video sequences in which the naming and the deictic cues were provided concurrently but by two different people, displayed no evidence of expectation of co-reference. These findings suggest that a single communicative source, and not simply co-occurrence, is required for mapping the two signals onto each other. By 13 months of age, infants appreciate the referential nature of words and deictic gestures alike.
Seeing the face through the eyes: A developmental perspective on face expertise. Progress in Brain Research. 2007;164:323-39.
Seeing the face through the eyes: A developmental perspective on face expertise
Most people are experts in face recognition. We propose that the special status of this particular body part in telling individuals apart is the result of a developmental process that heavily biases human infants and children to attend towards the eyes of others. We review the evidence supporting this proposal, including neuroimaging results and studies in developmental disorders, like autism. We propose that the most likely explanation of infants’ bias towards eyes is the fact that eye gaze serves important communicative functions in humans.