Jiayi Lu
University of Pennslyvania

Probing the satiation of island effects

Certain structural domains in sentences have been found to restrict syntactic operations (e.g., movements), a phenomenon termed the “island effect”. Recent work has revealed that some island-violated sentences, which are traditionally considered ungrammatical, are susceptible to satiation, an effect whereby speakers come to find initially degraded sentences acceptable after repetition. Previous studies have variably attributed the satiation effect to task artifacts, a domain-general re-allocation of cognitive resources, or structural priming/adapatation. Despite a lack of consensus on its exact mechanism, the satiation effect has been extensively used to inform theories of islands. Its existence has been taken as evidence for working memory–based accounts and frequency–based accounts of certain island effects, and it has also been used to argue both for and against natural classes among island effects. In this talk, I argue that previous literature on satiation has largely conflated two crucially different concepts: the satiation of island-violated sentences and the satiation of island effects. This conflation has muddied past proposals regarding the mechanism of satiation, as well as the theoretical claims about the nature of island effects based on satiation findings. In a series of acceptability judgment experiments, I employ an experimental design that distinguishes between these two concepts and examines the satiation of nine different island types. I will present preliminary results from these experiments, discuss the implications for various accounts of satiation, and revisit past theoretical claims about the nature of island effects based on satiation findings.