Universal Similarities in Children and Adults

Luisa Meroni, Andrea Gualmini and Stephen Crain
University of Maryland at College Park

This study examines adults' on-line processing of sentences containing the universal quantifier ‘every’, using a head-mounted eye-tracking system. It has long been known that even school-age children produce non-adult responses to (1) (e.g., Philip, 1995).

(1) Is every child riding an elephant?

For example, young children sometimes reject (1) as an accurate description of a picture in which every child is riding an elephant if there is an 'extra' elephant (i.e., one that no child is riding). However, child subjects produce adult-like responses in circumstances that arguably satisfy the felicity conditions associated with the test sentences; for example, if
there are other possible animals to ride, besides elephants. This raises the question if adults, too, find sentences like (1) easier to process in contexts that satisfy felicity conditions than in contexts that do not. To address this issue, we examined the on-line patterns of fixation duration on the 'extra' objects corresponding to sentences like (1) by adults, both
in felicitous and infelicitous contexts. Subjects were asked to judge the truth of spoken sentences with the universal quantifier, as descriptions of pictures that were simultaneously presented on a computer screen. Subjects'
eye-movements were monitored as the test sentences unfolded in real time, in order to observe the pattern of fixation on various objects in the pictures, including the 'extra' objects in the two conditions. Twenty English-speaking adults participated in the study, divided in two groups. Subjects in Group I were presented with pictures that satisfied the felicity conditions by providing an additional animal that the children could have ridden, or by providing an additional agent that could have performed the action. Subjects in Group II were presented with pictures that did not satisfy the felicity
conditions. The main result is that the fixation duration on the 'extra' object was significantly longer (p < .01) in response to pictures that failed to satisfy the felicity conditions (i.e., Group II) as compared to pictures that met the felicity conditions. (i.e., Group I). In short, the on-line pattern of fixations by adults mirrors the off-line difficulty experience by children.

The findings support the view that the satisfaction of felicity conditions, rather than differences in the grammatical representations of children, provides a better account of the pattern of behavior shown by children in previous research. Moreover, we show how the eye-tracking technique can be used to further scrutinize the findings of previous studies.