Figure 6.

Experiment 3 gaze data, collapsed across ISLANDHOOD, from the onset of the head noun (“Miss Piggy” in the example sentence), the middle subject (“Miss Cat”), the low subject (“Mr. Dog”), and the gap/resumptive pronoun. The four descriptive points we outline in the text are numbered in the right-most plot: (1) participants’ gazes remained on Mr. Dog (i.e. the most recently named character) at the onset of the gap/resumptive pronoun; (2) resumptive pronouns resulted in more looks away from Mr. Dog than gaps; (3) gaps led to more looks to the target than the local interpretation; (4) resumptive pronouns led to roughly equal numbers of looks to the target and local interpretations. The latter two points taken together mean that, although resumptive pronouns led to numerically more target looks than gaps, these looks were less accurate. That is, when participants did look away from Mr. Dog, they were more likely to look at Miss Piggy than Miss Cat when hearing a gap than when hearing a resumptive pronoun.