ABSTRACT. Infants posess several mechanisms to solve domain-specific problems. However, little is know about infants' abilities to reason beyond these limited domains. I will present evidence that when they witness a scene not previously experienced, infants reason about it by applying basic logical principles. I will argue that such inferences are used to build strategies to inspect the scenes and make inferences to enrich knowledge. I will present data about the behavioral correlates of this inferential processes in infants and adults, focusing on the case of disjuctive reasoning. 

ABSTRACT. Infants’ ability to distinguish between forms of phonetic variation that are relevant to meaning is essential for their language development. Little is known about the developmental course of infants’ perception of pitch, particularity in the presence of segmental variability which entails the ability to extract and generalize contrastive patterns. Using single-bisyllabic utterances, Experiment 1 examined native discrimination of the statement (falling)/yes-no question (rising) tune by European Portuguese (EP)-learning infants, and demonstrated that both 5-6 (t(19) = 6.1, p<.001, d=1.474) and 8-9 month-old infants (t(19) = 4.42, p<.001, d=0.816) were able to discriminate the contrast (Fig.1). Experiments 2 and 3 addressed the question whether falling vs. rising contours would also be perceived in segmentally varied nonnative input. Experiment 2 tested EP-learning infants’ perception of the lexical distinction between Mandarin Chinese Tone 1+Tone 4 and Tone 1+Tone 2. As noted for other languages [1], the EP and Mandarin pitch contrast shows similar contour shapes. However, differences in implementation arise from the different nature of tune versus tone, the latter having the syllable as its domain. Infants failed to discriminate the Mandarin contrast, both at 5-6 months (t(19) = .27, p=.79, d=0.06) and 8-9 months (t(19) = .86, p=.39, d=0.19). Experiment 3 tested EP-learning infants’ perception of the falling/rising contrast in Japanese, a pitch accent language with bisyllabic words distinguished by HL and LH contours [2]. Unlike tone, but like intonation, the Japanese contrast has the bisyllabic-word as its domain. Although infants at 5-6 months failed to discriminate (t(23) = .17, p=.87, d=0.04), they were successful at 8-9 months (t(23) = 2.44, p<.05, d=0.45). These findings show that, beyond segmental variation, infants’ perception of pitch is influenced by language experience as early as 5 months of age, and tune, pitch accent and tone are already differentially perceived in the first year of life.

ABSTRACT. During early childhood, shyness is negatively correlated with measures of language development (e.g. Spere et al., 2004), and this may in part be due to shy children’s reduced ability to learn words (Hilton & Westermann, under review). Shy children also attend more to faces than their less-shy peers (Brunet et al., 2009), so the current work examined whether this increased attention to faces may explain the effect of shyness on children’s word learning. 20-26 month old children (N = 32) took part in an eye-tracking study consisting of twelve onscreen trials. On each trial, one novel and two familiar objects were presented, as an onscreen actor looked at one of the objects and labelled it, using a pseudoword if the target object was novel. Each child’s shyness score was then calculated from the shyness scale of the Early Childhood Behaviour Questionnaire (Putnam, Gartstein & Rothbart, 2006) which was completed by their parent. Results showed that overall, children reliably looked towards the object during labelling. As expected, shyness was positively related to looking at the actor’s face, meaning that on average shy children showed sustained looking towards the face throughout a trial, while less-shy children shifted their attention towards the array of objects as the actor began labelling. Furthermore, less-shy children, but not shy children, showed an increase in looking at the target object in comparison to the two competitor objects. These findings suggest that shy children’s increased attention to faces reduces their ability to attend to objects as they are being labelled. Looking at an object during labelling is a key factor in determining whether the word-object mapping will be learned (e.g. Yu & Smith, 2011), so shy children’s increased attention to faces is a potential mechanism by which shyness affects language acquisition.

ABSTRACT. While infants show sensitivity to the continued existence of hidden objects in looking-times studies from 2.5m, they do not intentionally search for hidden objects until 8m. A proposed explanation is that searching for hidden objects requires forming expectations about how one's actions affect hidden objects. This is not required in looking-time studies. This proposal suggests that experiencing how one’s actions affect hidden objects should lead to generalised improvements in searching.

7m infants (n=44) were taught to spin a turntable to bring back into reach a toy that was either behind an opaque screen (hidden), or behind a transparent screen (visible, but out-of-reach). Before and after this training infants performed a different search task in which a toy was hidden in a hiding-well. Infants who learnt to spin the turntable with the opaque screen showed greater improvement on the hiding-well task. We suggest this improvement was the result of infants experiencing the effects of their actions upon hidden objects, thus learning to direct actions at such objects.

This first-person experience of directing actions at hidden objects might impact infants’ third-person understanding of other agents' actions: understanding hidden objects as goals of other agents’ actions might emerge alongside intentional search. 5-13m infants (n=52) watched an agent repeatedly direct an action at one of two visible objects. Both objects were then hidden by two screens, and the agent directed their action at either the old ‘goal-object’ or the second object.

There was no evidence that infants at any age looked longer when the agent changed goal, suggesting that infants did not interpret the agent’s action as directed at a specific goal-object when the objects were hidden. Possible explanations include methodological problems, working-memory limitations, or that the first-person ability to direct actions at hidden objects does not support the third-person understanding.

ABSTRACT. The violation of expectation paradigm is one of the primary tools used to study infants’ understanding of the world around them. Typically, in these tasks, infants are familiarised with one version of an event before being presented with versions which differ along one or more dimensions of interest, such as the events’ conceptual plausibility. Longer looking to impossible versions of events, where objects do not behave according to physical laws, is commonly interpreted as reflecting infants’ ability to apply their existing knowledge and reason about the events presented to them. In the current study, we examine this assumption by exploring the role of learning during the task itself. The contributions of infants’ conceptual knowledge and of online learning to looking behaviour are contrasted by switching the traditional approach on its head and familiarising infants to impossible events. Nine-month-old infants were presented with video clips of toy trains which either did or did not switch colour following a brief occlusion in a tunnel. Both looking times and pupil dilation were used as dependent measures in a factorial design, in which perceptual (novelty-familiarity) and conceptual (possible-impossible) variables were independently and jointly analysed. We show that infants, generally, reacted more to possible than impossible events following familiarisation. Furthermore, analysis of dynamic changes in pupil diameter whilst infants watched test events uniquely highlights the role of online learning in such tasks, and proves a finer-grained measure of information processing than traditional cumulative looking time measures. The discussion relates those findings to the general problem of measuring cognitive abilities in infants, from both a methodological and theoretical point of view.

ABSTRACT. In her seminal work, Woodward (1998) showed that infants reacted to apparent changes in human-agent goals. Her data were used to suggest an early concept of agency in infants. That work, and subsequent studies, are based on methods derived from habituation and rely on relative looking times as a dependent measure in what are know as Violation of Expectation (VoE) tasks. Methodological criticisms can be levied about such approaches in general (e.g., Heyes, 2014), and about Woodward’s (1998) task in particular (Sirois & Jackson, 2007). In the current study, we replicate the original paradigm but add methodological controls as well as use pupillometry to complement looking time measures. We tested 30 9-month-old infants (15 girls; Mage = 282.9 days, SD = 8.10) on video sequences showing a hand, initially resting between two toys, grabbing one of them. After habituation, infants were shown (in random order) a sequence of four test events that varied target position and target toy. Results show that looking times failed to distinguish between the different test events. However, micro-analyses of gaze data highlight how hand motion, path, and location (but not target) influence how infants look relative to habituation trials, explaining Woodward’s (1998) findings whilst eschewing the notion of goal attribution in infants. Similarly, pupil dilation shows an interaction between target and path, suggesting that babies react more to toys changing location as a function of hand trajectory. These findings highlight the role of online learning in VoE tasks that assess emergent social cognitive abilities. More generally, these data provide additional support to the growing concern that methodological issues (and, in particular, task-dependent arbitrary criteria for familiarisation and testing procedures in VoE studies) have manufactured an inaccurate understanding of the infant social and cognitive mind.