Clarifying the Role of Alignability in Similarity Comparisons

Structure-mapping theory has successfully predicted a number of empirical results concerning ordinary literal similarity processing. In particular, it predicts a distinction between alignable differences—those connected to the common structure derived in a comparison—and nonalignable differences, which are not so connected and which are held to be less salient than alignable differences (Markman & Gentner, 1993). Recently, Estes and Hasson (2004) have challenged the claim that alignable differences are more salient than nonalignable differences. In this paper, we address their criticisms and present data supporting an alternative interpretation of their results.