Models of comparative judgment have assumed that relative magnitude is computed from knowledge about absolute magnitude rather than retrieved directly. In Experiment 1, participants verified the relative size of part-whole pairs (e.g., tree-leaf) and unrelated controls (e.g., tree-penny). The symbolic distance effect was much smaller for part-whole pairs than for unrelated controls. In two subsequent experiments, participants determined either which of two objects was closer in size to a third object or which of two pairs had a greater difference in the size of its constituents. In contrast to the paired comparison task in Experiment 1,judgments of part-whole items were more sensitive to the influence of symbolic distance than were unrelated controls. The fact that the part-whole relation attenuates the effects of symbolic distance in a paired comparison task but not in tasks that require an explicit comparison of size differences suggests that the part-whole relation provides a source of information about relative magnitude that does not depend on knowledge about absolute magnitude.Much of our general world knowledge is knowledge about objects and their attributes. For example, the fact that coffee is darker than milk, that monkeys are smarter than dogs, or that ordering dinner requires much less time than calculating one's taxes can serve to distinguish each ofthese objects or events. An important topic in research on semantic memory is how people represent knowledge about the properties and attributes of objects and events and then use this knowledge to make these relational judgments.