There are two theories concerning conceptual representation: One is the Amodal Symbol Systems and the other is Modal Symbol Systems. The former one holds that the representation of a concept is amodal and consists of abstract symbols. The abstract symbols are independent from the sensorimotor experience. The latter one is based on embodied cognition and holds that conceptual representation is built on sensorimotor simulation. There are interaction between conceptual processing and perceptual processing. The sensorimotor experience has great influence on the representation of concrete and abstract concepts. The representation of concrete concepts has attained many supports from a mass of studies. However, abstract concepts cannot be associated with sensorimotor experience directly and the representation of abstract concepts is still under-explored. One of the critical questions concerning abstract conception is whether the relationship between conceptual processing and perceptual processing is symmetric or asymmetric. In our experiments, we use emotional concepts and emotional faces to address this question.

All the experiments were carried out using E-prime1.0. In experiment 1, emotional faces were presented to the participants and they were told to judge the valence of the faces. Participants also judged the target emotional face that following the emotional words. The results showed that there was interaction between emotional concept processing and emotional face perception on the deep semantic level. The relationship between emotional concept processing and emotional face perception was symmetric. In order to further explore the influence of semantic depth on emotional concept processing and emotional face perception, we changed the valence judgment task to pseudo-words judgment task in Experiment 2 so as to ensure that readers would only engage in shallow semantic processing in this task. We found then that concept processing affected the processing of emotional face, but not vice versa. Their relationship was asymmetric.

According to the two experiment results, we can conclude that the depth of semantic processing affected the relationship between emotional concept processing and emotional face perception. These results, along with others in the literature, indicate that conceptual processing uses sensorimotor representation, and that the depth of semantic processing affects the relationship between emotional concept processing and emotional face perception. In experiment 3, we further explored the depth of perceptual processing on this relationship. We changed face duration in experiment 3 and we found that when emotional face duration was shortened, emotional face perception did not affect emotion concept processing, and vice versa. There was no interaction between emotional concept processing and emotional face perception. Thus, our results suggested that the depth of perceptual processing also had an influence on the relationship between emotional concept processing and emotional face perception.