CNL Lunch

Luisa Meroni

Felicity Conditions and on-line interpretation of sentences
with quantified NPs

Thursday February 8th, 12:30pm, 3416 Marie Mount Hall

This study examines adults' on-line processing of sentences containing the universal quantifier every, using a head mounted eye-tracking system. In particular, we investigate how the satisfaction of (or failure to satisfy) felicity conditions influence adults' interpretation of spoken sentences associated with visually presented scenes. The interpretation of the universal quantifier has received considerable attention in literature on child language, where it has been found that even school-age children produce non-adult responses in certain circumstances (e.g., Philip, 1995). For example, young children (unlike adults) sometimes reject (1) as an accurate description of a picture in which every child is playing with a ball if there is an 'extra' ball in the picture (i.e., one that no child is playing with).

1) Every child is playing with a ball.

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 participants, besides children, or other possible objects to play with, besides balls. 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, despite the fact that adults ultimately reach the 'correct' interpretation. To address this issue, we examined the on-line patterns of fixation duration by adults in response to the 'extra' objects corresponding to sentences like (1), both in contexts that evoked non-adult responses from children, and in contexts that evoked adult-like responses. Subjects were asked to judge the truth or falsity of spoken sentences with the universal quantifier, every, 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. Two groups of adults were tested. One group of subjects, Group I, was presented with pictures that satisfied the felicity conditions for the test sentences, either by providing an additional object that the children could have played with (see Fig.1), or by providing an additional agent that could have performed the action. A second group of subjects, Group II, was presented with pictures that did not satisfy the felicity conditions in either of these ways (see Fig.2). Twenty English-speaking adults participated in the study, ten in each group. 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 mirror the off-line difficulty children experience in responding to sentences with the universal quantifier. By extension, the findings support the view that the satisfaction of felicity conditions, rather than differences in the grammatical representations of children versus adults, provides a better account of the pattern of behavior shown by children in previous research.