This article reviews modern experiments involving chronometric measurements of comprehension processes for various forms of nonliteral language. The goal of such research is to explore hypotheses about cognitive processes and memory systems. The sequence of mental operations involved in comprehending figurative language and the sequence of phenomena experienced by the comprehender depend on the nature of the stimuli, instructions, and reaction-time (RT) task. RT experiments have been conducted on indirect requests, idioms, common metaphors, novel metaphors, similes, common proverbs, and novel proverbs. The experiments have involved such methods as grammaticality decision latency, rhyme monitoring, and true/false decision latency. Much of this research has been motivated by theoretical propositions, such as the claim that metaphors should take longer to comprehend than semantically comparable literal sentences. Except for the research on idioms and indirect requests, the experiments have not been strong tests of such theoretical propositions. The reasons for this include certain methodological pitfalls that are likely to arise in metaphor research, and a lack of clarity about appropriate experimental questions. There is, in particular, a common misunderstanding about so-called two-process theories of metaphor comprehension. An analysis of the logic of such theories shows that the two-process notion is not a testable hypothesis. The analysis of methods suggests advice for researchers interested in metaphor comprehension. (Robert Hoffman and Susan Kemper)