People typically provide higher similarity ratings in response to the question "How similar is the typical preppie to you?" than to the question "How similar are you to the typical preppie?" Observed asymmetries in comparisons of the self and person prototypes have been offered as evidence that the self-concept acts as an habitual reference point in social judgment. However, such a task does not test the habitual placement of a concept in the referent position of a comparison. In this study, participants judged the similarity between the self and person prototypes in response to linguistic (forced directional) queries, or to spatial (nonforced) queries~ which the self was positioned above or below the person concept. Participants also rated pairs 4 familiar and unfamiliar countries in a similar manner (Tversky, 1977). Expected asymmetries v observed in forced comparisons such that the self and the familiar country were seen as more similar to other people and less-familiar countries, respectively, when the former concepts serve as the referent of the comparison than when they served as the subject. Asymmetries were not observed in the nonforced conditions and mean similarity in the these conditions was of the same magnitude as the forced condition in which the more familiar stimulus was the referent of the comparison. This result provides the first direct evidence that the self serves as an habitual referent in similarity judgments.