Time-dependent effects of decomposability, familiarity and literal plausibility on idiom priming: A cross-modal priming investigation

We address a core question about idioms relevant to formulaic language generally: are the figurative meanings of idioms directly retrieved or compositionally built? An understanding of this question has been previously obscured by the fact that idioms vary in ways that can affect processing, and also because experimental tasks, which differ across studies, probe different kinds of comprehension processes. We thus investigate how linguistic differences among idioms in semantic decomposability, familiarity, and literal plausibility modulate figurative meaning activation using cross-modal semantic priming, which is ideal for tracking activation of a particular target meaning over time. Across two experiments, we obtained two key findings. First, a comparison of different prime-target delay conditions suggests that figurative meaning activation steadily accrues as the idiom unfolds to 1000 ms later. Second, different linguistic attributes of idioms modulate figurative activation at different time points: increased literal plausibility interferes with idiom priming prior to the offset of the phrase, increased familiarity facilitates idiom priming at phrase offset, and increased semantic decomposability (surprisingly) interferes with idiom priming 1000 ms following phrase offset. These results contradict strong decompositional models of idiom processing and rather suggest that multiple linguistic factors jointly constrain figurative meaning retrieval in a time-dependent fashion.