3-036 - 30 Years of Shape Bias Research: Lessons for Word Learning, Atypical Development, Categorization and Beyond

Session Type: Paper Symposium

Integrative Statement

Young children are biased to generalize novel nouns for novel solid objects by shape similarity. The accumulation of over 30 years of research on this “shape bias” presents a clear theoretical picture linking its development to statistical regularities in the early vocabulary and basic processes of perception, statistical learning and categorization. The bias has also been used as a test case for computational models of object recognition and one-shot learning, and proposed to be a target for language intervention. Likewise, the three talks in this symposium demonstrate implications of the bias beyond word learning via recent research capitalizing on the latest experimental methodologies and computational tools. Talk 1 will use a novel looking-while-listening task and a computational model to examine the development of the bias from moment-by-moment visual exploration to accumulation of category statistics to higher-order generalizations. Talk 2 will illustrate how understanding of the relation between learning of individual words, the accumulating statistics, and future learning can inform theories of atypical trajectories using both empirical noun generalization studies and a large-scale database analysis with typical and atypical language learners. Talk 3 will present work using a deep neural network to examine the relative roles of neural structure and input on the development of the bias, and showing how the same processing mechanisms that support the bias have applications to classification decisions for medical cardiac images. The discussant, an expert on the bias, will integrate these findings with prior work to illustrate the implications for cognition more generally.

Sub Unit

• Language, Communication

Chair

• Larissa K Samuelson, University of East Anglia

Discussant

• Letitia Naigles, University of Connecticut

Individual Presentations

• Nested Timescales of Looking and Learning: Visual Exploration and Simulations of the Shape Bias - Presenting Author: Larissa K Samuelson, University of East Anglia; Non-Presenting Author: Megan Lorenz, The University of Iowa; Non-Presenting Author: Joshua G Dickinson, University of East Anglia; Non-Presenting Author: Rachel Tremlin, University of East Anglia; Non-Presenting Author: Jessica Bates, University of East Anglia; Non-Presenting Author: John Spencer, University of East Anglia

• Individual Differences and the Shape Bias—Lessons From Late Talkers - Presenting Author: Lynn K Perry, University of Miami; Non-Presenting Author: Sarah Kucker, Oklahoma State University; Non-Presenting Author: Jessica S Horst, University of Sussex; Non-Presenting Author: Larissa K Samuelson, University of East Anglia

• A Deep Learning Account of Shape and Color Biases in Categorization - Presenting Author: Bradley C Love, University College London; Non-Presenting Author: Olivia Guest, University College London