Olson (1988) argued that metaphor does not so much depend on an ability to perceive relations between things as on a special use of language. In contrast to Olson, it is pointed out that classification and communication, an intellectual territory that can be approached via perception and expressed via language, could be the home base of metaphor. The study of metaphor, as Olson (1988) noted, uncovers interesting properties of mind. Olson stressed the relationship between metaphor and language in his analysis, although he made an important place in his account for intention, goals, beliefs, and meanings. I believe he overemphasized the importance of language; I argue that the existence of metaphoric pictures, for example, makes any analysis that suggests metaphor is exclusively linguistic highly suspect. At the very least, some broader notion than language is needed. But should one contend simply that metaphor can arise in any communication system? Perhaps not. I try to give reasons for thinking of metaphor as a matter of classification, at heart, rather than language or communication. To accomplish this, I present and briefly expand Olson's thesis and then address and amend each point in reverse order. (John Kennedy)