Funded Grants

Elucidating developmental change in the structure of semantic knowledge and inductive reasoning

Structured knowledge representations are a fundamental component of adult cognition. Adults can form concepts not grounded in perceptual commonalities (e.g., empathy, truth, innovation); organize categories into multiple levels of abstraction (e.g., Russian Blue → Cat → Feline → Mammal → Animal → Living Thing); and reason by analogy. There is no evidence that newborns engage in such higher-order thinking. Thus, the challenge for theories of cognitive development is to provide an account of this developmental change. Many developmental psychologists (as well as philosophers, starting with Plato) have posited innate knowledge structures to explain this remarkable change. My research aims to provide an alternative theoretical account of the development of higher-order thinking. The core idea is that humans are remarkable not because of innate knowledge structures, but because of their ability to learn and build structured representations from data. I apply this principle to elucidate pronounced developmental changes in higher-order thinking, with a specific focus on inductive reasoning.

Inductive reasoning involves making generalizations from instances. It is a powerful and effective tool for generating new knowledge. Semantic information (e.g., carried by linguistic labels) has been shown to promote inductive inferences. Consider this example: when told a novel fact about alligators (e.g., ‘alligator embryos lack sex chromosomes’) most adults correctly conclude that crocodile embryos also lack sex chromosomes. One is not given any information about crocodiles; however, making an inductive inference on the basis of what is known creates new knowledge. This mode of inference is not guaranteed to generate correct knowledge (one might incorrectly overgeneralize a fact about alligator embryos to all Oviparous animals). Nevertheless, the ability to make such inferences is a hallmark of mature cognition.

It has long been assumed that even young children are adept in using the semantic information provided by labels to make inductive inferences. However, recent research in my lab challenges this long-held view and points to pronounced developmental changes in this ability. Given the widespread agreement about the fundamental role of inductive reasoning in human cognition, it is crucial to understand the origins of these developmental changes. My research explores the possibility that development of the ability to make inductive inferences about familiar categories is a function of developmental changes in the structure of semantic knowledge.

Computational studies suggest that semantic development entails an increasingly rich organization of concepts into a network of clusters based on taxonomic relations. Until recently, these theoretical claims about developmental changes in semantic structure have not been substantiated empirically. My lab has developed new methodology to examine the development of semantic structure. Studies using this methodology support the insights offered by computational models, making it possible to evaluate the hypothesized link between the structure of semantic knowledge and inductive reasoning.

The goals of this research program are two-fold. First, I aim to explain the mechanisms of developmental change in inductive reasoning, a fundamental aspect of higher-order thinking. Second, I aim to elucidate developmental changes in the structure of semantic knowledge.