What do we learn from rating metaphors?

What makes some metaphors easier to understand than others? Theoretical accounts of metaphor processing appeal to dimensions like conventionality and aptness to explain variability in metaphor comprehensibility. In a typical experiment, one group of naive participants rates a set of metaphoric sentences along these dimensions, while another is timed reading the same sentences. Then, the ratings are used to predict response times in order to identify the most relevant linguistic dimension for metaphor comprehension. However, surprisingly high correlations between ratings of theoretically orthogonal constructs and the results of an experiment in which a context manipulation affected ratings of metaphor conventionality and aptness suggest that these measures should be treated as dependent, rather than explanatory, variables. We discuss the implications of this perspective for theories of language processing.