Modern theories of metaphor seem to be pretty unanimous in taking the "death" of a metaphor literally. By doing so they too easily wipe out sedimented, past meanings and so ignore semantic memory. A further consequence of this stand is that meanings are reduced to a one-dimensional (either metaphoric or literal), static structure. This article, in a procedure that resembles a sort of "archeology of meaning," is critical of such an attitude, for which conventionalization of metaphors means their burial in the graveyard of language. It attempts to provide a ground for the assumption that labeling of metaphors as "dead" is not lethal, for their meanings can be revitalized and recreated, according to our linguistic competence, diachronic perspective, interests, and needs. It contains implicit support for the thesis that semantic mechanisms are flexible, dynamic, and subject to evolutionary growth. Not even the death of a metaphor can be exempt from this overall scheme. (Zdravko Radman)