The Classical System
The Aristotelian privileging of metaphor's formal aspect and its attendant assumptions, emphases, and procedures are together amplified and reified by his classical inheritors. In this tradition, metaphor's creative aspect tends to be viewed with suspicion. For instance, in De oratore (55 B.C.E.; On oratory), through a fabular comparison between metaphor and clothing, Cicero (106–43 B.C.E.) articulates a view of language as a form of ethical conduct that must not stoop to don the potentially corruptive finery of metaphor: "For just as clothes were first invented to protect us against cold and afterwards began to be used for the sake of adornment and dignity as well, so the metaphorical employment of words was begun because of poverty, but was brought into common use for the sake of entertainment." Cicero uses his comparison to caution us against "borrowing" fancy metaphors because they suggest "poverty" of thought and expression (p. 121–123). But it is largely through Cicero's precise, nuanced discussion of the proper and improper forms and uses of intentionally shaped language that his motive to civilize the power of metaphor is transmitted, inspiring the anonymous author of the Rhetorica ad Herennium (first century B.C.E., Rhetoric to Herennius) and Quintilian (c. 35–95/6 C.E.), author of Institutio oratoria (first century C.E.; Institutes of oratory).
Quintilian's terminological and conceptual precision expands the formal context within which metaphor may be understood. Because his primary interest is pedagogically useful classification, Quintilian chooses to treat metaphor as a member of the tropes, which involve "the artistic alteration of a word or phrase from its proper meaning to another" (p. 301). Treating metaphor as a member of a class of similar forms allows Quintilian to note metaphor's uniqueness without committing himself to an equal exploration of both its aspects. For example, while he allows that metaphor is the "commonest and by far the most beautiful of tropes," and praises it for "accomplishing the supremely difficult task of providing a name for anything" (p. 303), he restricts an attribution of cognitive agency to the tropes as a class: "The changes involved [in the use of tropes] concern not merely individual words, but also our thoughts and the structure of our sentences" (p. 301–303). It is noteworthy that here, in the midst of explicating a system whose formal emphasis diminishes metaphor's creative aspect, Quintilian states that tropes in themselves can play a serious role in thinking. Though muted, this connection between figured language and cognition anticipates the ground from which springs the Middle Age's most important contribution to ideas about metaphor: the notion that, when interpreted correctly, the ambiguities and excesses of figured language involve human understanding with a superior order of knowledge and being.