What norming reveals about idioms: Making the case for a presuppositional account.

Abstract

While early accounts of idiomatic expressions proposed that they are compositional or else directly retrievable from memory, the multi-determined view posited that idiom comprehension depends on observable characteristics, such as meaningfulness, familiarity, literal plausibility, global decomposability, and final word predictability. This led researchers to periodically undertake norming tasks in which participants rate idioms on these dimensions. The current study extends this tradition while investigating 36 American English idioms, expressed as She/he verbed x noun (e.g., He fanned the flames). Study 1 introduced a new control (Nonsense idioms), which encourages the exploitation of a scale's lower end, while recruiting sub-samples of participants online for each of the five aforementioned dimensions. Our findings, which primarily concern correlations among dimensions, very largely confirm the prior findings. Study 2 introduced a novel norming dimension that we call presupposition strength. This asks participants to provide a likelihood score about background information that is not conventionally associated with each idiom. The 36 idioms were presented through a vignette (e.g., Tom fanned the flames at the meeting) after which we collected scores to a presuppositional probe question (e.g., How likely is it that there was tension before the meeting?). Participants' mean scores for an individual idiom's presupposition strength were compared to two yoked controls, a paraphrase (from dictionary definitions) and a nonsense idiom. Presuppositional strength for idiomatic expressions led to significantly superior scores, pointing to the importance of this feature to these figures. Intriguingly, correlations between presupposition strength and (Study 1's) meaningfulness and familiarity were statistically significant.