Metaphors do not use associations between concepts,they are used to create them

Does the interpretability and aptness of a metaphor depend on prior existing associative relations between the metaphor topic and vehicle? Lexical decision latencies for pairs of words drawn from apt, comprehensible metaphors were no faster than latencies for randomly paired words. In contrast, lexical decision latencies for associatively related word pairs were faster than latencies for randomly paired words. These data suggest that good metaphors do not use preexisting associations to achieve their effects. Instead, we argue tht people use metaphors to create new relations between concepts. Implications for a theory of metaphor comprehension are discussed.