Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Katz, Albert N.


In a metaphor such as that lawyer is a shark, the concept lawyer, which is the metaphor topic, and the concept shark, which is the metaphor vehicle, interact to produce a figurative meaning such that lawyers are predatory. Some theorists argue that sensorimotor properties of the vehicle are the basis of metaphor comprehension (Gibbs & Matlock, 2008; Paivio, 1979; Wilson & Gibbs, 2007). As such, that lawyer is a shark is processed by an embodied simulation where sensorimotor imagery associated with sharks is simulated (e.g., sharks hunting in deep water). However, the long-standing assumption is that metaphors are processed abstractly and sensorimotor representations play no role (e.g., Gentner & Bowdle, 2008; Glucksberg, 2008). This thesis examines the role of sensorimotor simulation in processing metaphor. In Studies 1 – 2, participants rated metaphors on comprehensibility. The metaphors contained vehicles that varied on a semantic richness variable known as body-object interaction (BOI), which characterizes the degree to which a concept is easy to interact with (Siakaluk et al., 2008). A high-BOI metaphor contains a vehicle concept that is easy-to-interact with (e.g., life is a bicycle) whereas a low-BOI metaphor contains a concept that is difficult-to-interact with (e.g., life is a rainbow). Participants rated low-BOI metaphors to be more comprehensible than high-BOI metaphors, a finding that suggests sensorimotor properties are not heavily involved in metaphor processing. In Study 3 participants created novel metaphors by pairing abstract topics with words that varied on BOI to serve as vehicles. In creating metaphors, participants chose more low-BOI words to serve as vehicles than high-BOI words. However, to interpret their created metaphors, participants used language reflective of an embodied simulation for both high and low-BOI metaphors, indicating that nominal metaphors do indeed involve sensorimotor imagery. In Studies 4 – 7, a priming paradigm showed that processing novel metaphors (e.g., highways are snakes) immediately activates sensorimotor properties (e.g., slither) whereas familiar metaphors (e.g., lawyers are sharks) do not activate sensorimotor properties (e.g., bite) but rather, activate abstract associations (e.g., killer). In sum, the experiments in this dissertation are the first to demonstrate novel metaphors are processed by sensorimotor simulations.