Hammering out meaning

The thesis contains three papers. In the theoretical paper, I evaluate several proposals on the role of motor information in tool concepts that have been fueled by brain imaging and behavioral studies. The strong embodied cognition hypothesis argues that sensory and motor information determines conceptual content fully: for example, motor representations are the basis of our generalization and categorization of tools and action concepts, our understanding intention, and define the meanings of verbs and tool terms. The weak embodied cognition hypothesis argues that motor information may partially determine the meaning of action and tool concepts but is integrated with other types of information within a conceptual structure. Finally, the abstract cognition hypothesis argues that motor information does not determine conceptual content: abstract notions like intentions and functions provide content for action and tool concepts and are the basis for generalization. I argue that the strong embodied cognition model is inconsistent with (1) our current knowledge of the mirror neuron system (when it is argued to be a critical mechanism of embodied concepts), and (2) our knowledge of conceptual development. Motor information alone is insufficient delineate intended goals or functions. I suggest an alternate hypothesis about the role of mirror neurons in comprehending action concepts and explore different models of how motor information is integrated in the conceptual content.

The empirical papers present findings on the role of motor information the tool concepts of 3-year-olds, 5-year-olds' and adults. I whether children and adults extend novel labels to tools which involved the same motor movement or tools which involved the same function in a two alternative forced choice task. I found that 3-year-olds, 5-year-olds and adults categorized tools based on their function rather than their motor manipulation. One concern is that children may match the tools based on visual information about the tools' function rather than an abstract representation. However, both 3-year-olds and five-year-olds also categorized novel tools by function when the results of the target function were not perceptually accessible. These experiments indicate that abstract properties like function are central to conceptual representation of tools while motor information is not critical for inferring extension of novel tool terms.