Simulating Semantics: Are Individual Differences in Motor Imagery Related to Sensorimotor Effects in Language Processing?

In embodied theories of semantic representation, the processes and mechanisms of modal simulations that are engaged during semantic processing have tended to be underspecified. We investigated the possibility that motor imagery may be a mechanism of simulation, using an individual differences approach. In this preregistered study, we assessed motor imagery abilities (n = 161) with implicit and explicit measures and identified two latent factors. We then examined whether those factors account for significant variations in sensorimotor effects observed in three different language tasks: a lexical-decision task, syntactic classification task, and sentence-picture verification task. In the language tasks, when all participants were considered together, we replicated some previously reported sensorimotor effects (e.g., body-object interaction [BOI], effects in semantic processing, wherein words associated with more sensorimotor information were processed more quickly than words associated with less sensorimotor information) and did not replicate others (e.g., BOI effects in the lexical decision task, congruency effects in sentence picture verification task). There were no significant relationships between imagery factor scores and sensorimotor effects. A follow-up analysis using scores from each motor imagery measure revealed a significant interaction between hand movement imagery and BOI effects in the syntactic classification task, with those higher in this imagery ability showing a larger BOI effect. This latter result may suggest that specific types of motor imagery are related to sensorimotor effects in semantic processing; however, further investigation is needed. In general, our findings provide little support for the possibility that motor imagery is an underlying mechanism of sensorimotor simulation during language processing.