Suppression from the Perspective of Distributed Models of Conceptual Representation

Abstract

Looking at metaphor comprehension from the perspective of distributed models of conceptual representation, this article describes the process of suppression through which metaphorically irrelevant features of metaphor’s vehicle are suppressed. Distributed models of conceptual representation hold that meaning of every concept is represented by a set of feature nodes in a connectionist network, and the coactivation of these nodes leads to the understanding of that concept. Based on these models, it is suggested that degree of distinctiveness of features play an important role in the suppression of metaphorically irrelevant features during metaphor comprehension. When the metaphor X is a Y is processed, a salient feature of Y creates a metaphorical class to which both topic (X) and vehicle (Y) belong. The rest of features, which are metaphorically irrelevant, are suppressed. Those irrelevant features which have a high degree of correlational strength are suppressed collectively. Finally, the key role of metaphor’s topic in the suppression of metaphorically irrelevant features is discussed. It is suggested that the set of suppressed features is dependent on the topic. If the defining feature of the metaphorical class of the vehicle (Y) matches the topic (X), the rest of features will be suppressed. In this situation, both topic and vehicle of the metaphor are included in a common metaphorical class. If the defining feature of metaphorical class of the vehicle (Y) does not match the topic (X), X is understood in its literal sense, and the sentence X is a Y will not have a logical metaphorical interpretation.