Poster Session C, C24 (Saturday 12:10 – 2PM)
Filler-gap *wh*-dependencies (FGD) have been extensively studied, but the superficially very similar “remnant”-“correlate” *wh*-dependencies (RCD) in backward sprouting have not been. Here we compare their sensitivity to distance and to islands using an acceptability experiment. We find both dependencies show sensitivity to distance (though the degradation is significantly greater with FGDs) but that only FGDs show sensitivity to islands. The difference with distance suggests that FGDs make some demands on working memory that RCDs do not, while the difference in island-sensitivity suggests that islands must stem from something specific to FGD, either a grammatical property or some working memory demand.