Phrase-by-phrase self-pacing tasks are not sensitive to structural priming: Methodological and theoretical implications

Over the past fifteen years, researchers have begun using structural priming in language comprehension—rather than language production—to examine language processing and representation (Branigan & Pickering, 2017). At present, there are only limited methods for measuring comprehension structural priming. Here, we tested whether self-paced listening (SPL, Experiments 1a-1a) and self-paced reading (SPR, Experiment 2a-2b) tasks could be an additional way to capture this phenomenon. We tested priming of temporary modifier-goal ambiguities of the form The librarian set the book on the shelf on the cushion this morning, for which comprehension priming has been demonstrated using eye-tracking (Traxler, 2008). In Experiment 1a, we found that SPL is sensitive to priming, but only under certain RT exclusion criteria. In a pre-registered replication (Experiment 1b), we used the same exclusion criteria but found no priming effect. Bayes factors support the null hypothesis in each experiment. In two variants of a self-paced reading (SPR) task (Experiment 2a-2b) using the same modifier-goal construction, we find evidence of priming only when the sentence is presented word-by-word, not when presented phrase-by-phrase. We discuss methodological recommendations for future researchers who may wish to use self-pacing tasks to measure structural priming and the theoretical implications of the absence of the priming effect in phrase-by-phrase presentation.