Abstract: The concept of meme misidentifies units of cultural information as active agents, which is the same “shorthand” that misleads our understanding of genes and obscures the dynamic logic of evolution. But the meme concept does offer hope by contributing something missing from many semiotic theories. In treating memes as replicators, Dawkins fails to distinguish mere patterns from information (and replication from interpretation), which leads to the problem encountered in all realms of information processing: what counts as information is context dependent. Nothing is intrinsically meaningful, to be so it must be interpreted. In the evolution of both genes and words, replication has always been a multilevel affair in a dynamic system, from which what we conceive as “adapted” or “interpreted” units emerge. Memes are replicas not replicators, and I suggest that the iconic function of signs, as identified by Peirce, is the essence of the meme concept. As in sign function, both gene and meme functions are informed by whatever relationship exists between the physical pattern of the token and the system of processes in which each is embedded (so these are semiotic relationships). I argue that two, not-clearly-articulated aspects of the meme concept could rescue semiotic from mere descriptive taxonomy and lead the way for a general theory of semiosis and a unifying methodology for the semiotic sciences to emerge.