Non-Native Language Processing Engages Mental Imagery

The theory of simulation semantics (Bergen & Chang 2005) posits that understanding language, in part, requires activation of mental imagery. This allows understanders to mentally recreate the scene or event to facilitate understanding and prepare for situated action (Glenberg & Kashak 2002). The idea that understanding action language relies on neural circuitry involved in action execution is supported by a crossmodal matching method introduced by Bergen et al. (2003), which demonstrated that specific effectors (hand, mouth, foot) are critical to the motor imagery involved in language understanding. Previous studies, however, have focused exclusively on adult native speakers, which leaves open the question of how language-driven imagery develops during language acquisition. The current study investigates whether non-native English speakers engage in mental simulation during language processing. We used an image-verb forcedchoice matching task, where an image and verb depict different actions using either the same effector (e.g. grab and push) or different effectors (e.g. grab and lick). As in previous work with native speakers, response times were significantly longer when the two actions used the same effector. Moreover, subjects showed a correlation between stimulus comprehension accuracy and the size of the simulation effect. This suggests that non-native speakers not only perform mental imagery like native speakers but do so increasingly as their linguistic competency improves.