Cognitive Science Colloquium

Fall 2018

All meetings take place on Thursdays, 3.30-5.30 pm in Bioscience Research Building 1103, unless otherwise indicated.

September 6 — Barbara Sarnecka (Cognitive Sciences, UC Irvine).

Title: Numbers and Language

Abstract: After 20 years of studying the relation between language and the mental representation of numbers, I’ve come to the conclusion that three things are true. (1) Numbers are unrelated to language; (2) Numbers are related to language in general; and (3) Numbers are related to particular languages. These sound like contradictory statements, but I will explain during my talk how all of them can be true, depending on what one means by ‘language’ and what one means by ‘number.’

September 20 — Sandra Trehub (Psychology, University of Toronto).

Title: The Musical Infant

Abstract: Despite infants’ immature resolution of pitch and timing, they are surprisingly proficient at holistic processing of pitch and timing patterns. Like adults, they recognize the invariance of melodies across pitch levels and the invariance of rhythmic patterns across tempos. They also exhibit long-term memory for melodies, and their memory is especially detailed and enduring for vocal melodies. In some respects, they are more flexible music listeners and learners than their adult counterparts. Although adults’ musical engagement is commonly unimodal and solitary, infants typically experience music multimodally in dyadic contexts (e.g., Western caregivers sing lively play songs in face-to-face contexts; non-Western caregivers sing soothing lullabies to carried infants). Such singing is more effective than infant-directed speech in regulating infant arousal and affect. Moreover, caregivers’ songs have social as well as emotional significance for infants, influencing their behavior toward unfamiliar individuals who sing such songs.

October 4 — Mina Cikara (Psychology, Harvard University).

Title: Leveraging Choice Architecture to Alter Social Preferences
Abstract: Stereotypes and associated emotions drive discriminatory behavior across numerous consequential contexts. These biases against marginalized social groups have important implications for real-world social decisions, including hiring, voting, health, and housing decisions. Psychologists have traditionally studied how people evaluate different ethnic and cultural groups (and their members) in isolation, but in the real world people commonly make judgments and decisions over sets of people. For example, hiring decisions involve the assessment of multiple candidates at once. Across a series of experiments, we harness insights from computational models of decision-making to examine whether choice set construction—or choice architecture—can be used to influence decision-makers' preferences in consequential social decisions. I will review several findings including a combination of field data and lab experiments to examine the effect of alternatives, or decoys, on social evaluations and decisions in hiring and election contexts. 

October 18 — Percival Matthews (Educational Psychology, University of Wisconsin—Madison).

Title: Rational Feeling: Perceptual foundations for understanding numerical magnitudes
Abstract: Some contemporary theories of numerical cognition posit that a generalized magnitude system may serve as a core primitive foundation for building number concepts. To date, however, these theories have largely privileged whole numbers and whole number analogs, relegating rational numbers to the background. In this talk, I will argue that more explicit attention to nonsymbolic ratio perception can account for the deep connections between whole numbers and other classes of number, while accounting for how continuous magnitudes can be mapped to specific numbers. By carving out a pivotal role for nonsymbolic ratio perception, this correction might help provide the basis for a more unified theory of numerical cognition.

October 25 — Darko Odic (Psychology, University of British Columbia).

Title: Breaking the ATOM: Exploring the relationship between number, time, and space perception
Abstract: Prominent theories in cognitive and developmental psychology have posited that our perceptual sense of number, time, and space form a singular, domain-general representation. In this talk, I provide extensive evidence to suggest that these perceptual representations are distinct: children and adults show no correlations between number, area, density, time, etc., when domain-general factors such as working memory, response conflicts, and general developmental improvements are controlled for, saccadic trajectories show unique encoding algorithms for number vs. area, and individual differences in number perception uniquely predict formal mathematical abilities. But, critically, number, space, and time share a common representational format of noisy Gaussian tuning curves on independent scales, allowing for interactions between them in specific contexts: reasoning about perceptual confidence and learning to map number words to length and area. This independent-representational but shared-format view allows us to reconcile some traditional findings in the perceptual magnitudes literature while preserving the view of domain-specificity in the perception of number, time, and space.

November 29 — Emmanuel Dupoux (L’Ecole des Hautes Etudes en Sciences Sociales, Paris).

Title: Can AI help the study of language development (and vice versa)?
Abstract: The recent advances in artificial intelligence (AI) rely on statistical learning techniques applied to large sets of data. In this presentation, we show that AI can also have a scientific impact in (neuro)cognitive sciences, by offering quantitative models of human learning. We examine the case of cognitive and linguistic development in children and show that, here, the relevant algorithms are essentially unsupervised or weakly supervised (i.e. rely mainly on ambiguous and ambiguous sensory data). We illustrate this type of algorithm with the automatic discovery of linguistic units in an unknown language. We discuss the feasibility of this approach and its ability to generate new hypotheses about child learning processes, as well as the functional role of parent / child interactions.


December 6 — Kevin Ochsner (Psychology, Columbia University). — POSTPONED to April 4, 2019.