There are 2 basic controversies concerning how metaphors are processed. First, are metaphoric mappings more akin to literal comparisons or to literal categorizations? And second, is metaphor comprehension indirect or direct? We believe that these controversies are more apparent than real and that a unified theoretical framework can be offered that reconciles these opposing views. The central idea is that all metaphors involve structural alignment of the target and base domains. Whether these alignments are more akin to comparisons or to categorizations, and whether they are computed directly or indirectly, depends on 2 factors: the type of representation invoked by the base term and the linguistic form of the metaphor.