Abstract

Directional asymmetry is a well-documented phenomenon in research on similarity, metaphor, and analogy. In . this paper, we present an account of this phenomenon based on structural alignment . We propose that a major source of asymmetry is coherence imbalance : that is, a difference in the degree of systematicity of the relational structures being compared. These claims are tested in three experiments which examine the relationship between asymmetry, informativity, and conceptual coherence . The results support the hypothesis that coherence imbalance is a key factor in directional comparison processes. Further, by incorporating the insights offered by structural alignment, coherence imbalance advances a more functional account of asymmetry . Comparison is a fundamental process in cognitive theory, and is crucial to such aspects of mental operation as categorization, decision making, problem solving, and transfer of learning . Intuitively, comparisons might appear to be symmetric : X is similar to Y implies that Y is similar to X. Empirically, however, there is often a preference for one direction of comparison over the other . For example, Tversky (1977) found that subjects preferred the comparison North Korea is similar to Red China over the comparison Red China is similar to North Korea, and rated the similarity of North Korea to Red China as higher than that of Red China to North Korea. Asymmetries have been found not only in literal similarity, but in other forms of comparison, including metaphor and analogy. We begin by reviewing two influential models that have been proposed to explain such asymmetries . According to Tversky's (1977) feature contrast model, the similarity of two entities increases as a function of their common (shared) features, and decreases as a function of their distinctive (non-shared) features . Asymmetries in literal similarity are accounted for in terms of the focusing hypothesis, which states that the target of a directional comparison is the focus of attention . 1 Therefore, the distinctive features of the target are weighted more heavily than those of the base . This predicts that there is a general preference for placing the referent possessing the larger or IIn directional comparisons of the form A is like B, the aterm is referred to as the target and the b-term as the base.