Events in the world do not map transparently into linguistic expressions. Instead, particular linguistic forms guide our interpretations: different forms guide us to construe events differently. The current experiments investigate children's ability to use telicity information to influence their construal of events. The results reveal that children have an impressive understanding of language specific telicity marking from the age of 2;10, but they also have an equally impressive non-linguistic bias in event construal that persists until as old as age 5 years.
Telicity is the lexical aspect distinction between telic and atelic predicates. Telic predicates describe events with natural end-points ("The door closed", "The bird popped a balloon"); atelic predicates describe events that end at any arbitrary point ("The butterfly flew", "The rabbit moved"). In terms of event construal, telic and atelic predicates provide different criteria for individuation, or what counts as the proper ending of the event: events described with telic predicates end only when they reach the ending specified in the predicate (regardless of any intermediate temporal breaks) while any temporally bounded amount of an event described with an atelic predicate counts as an instance of the event.
Children in these experiments were presented with animated movies portraying events which permitted two different end-point construals: e.g. a girl paints a picture of a flower (the goal ending) but she does so with two, temporally separated brush strokes (the temporal-break endings). Children heard the event described with either a telic predicate ("The girl painted a flower") or an atelic predicate ("The girl painted") and were asked to count how many times it happened. The dependent measure was whether the children counted brush strokes or flowers. Forty-eight children (aged 2, 3, and 5 years) each saw 8 movies, half described with a telic and half with an atelic predicate. The predicates were chosen so as to determine children9s knowledge of the role that the direct object plays in signaling telicity. Children were tested with transitive sentences with telic semantics and intransitive sentences with atelic semantics (the canonical pattern in English) as in the above example; and with violations of this canonical pattern such as intransitive sentences with telic semantics (unaccusatives: e.g. "the vase breaks") and transitive sentences with atelic semantics (e.g. "push a ball"). At issue was whether there would be a preferential mapping of telicity semantics onto the canonical pattern of telicity marking in the language.
The results showed that children at all ages understood the telic and atelic predicates and changed their counting criteria for the events appropriately. The 3 and 5-year olds were equally proficient with all types of predicates; the 2-year-olds had trouble only with transitive sentences with atelic meaning. These results suggest that both the telic and atelic construals are important and available to children at a young age. Moreover, children appear to initially link telic meaning to the linguistic presence of direct objects. Despite the overall linguistic success, all age groups were also found to prefer counting the temporal-breaks (the atelic interpretation) in general, relative to adult controls. This finding echoes previous results that children initially rely on spatio-temporal criteria for individuating objects and may reflect a persistent, perceptually based system that children use for analyzing events.
| Workshop Homepage | Workshop Program | Previous Abstract |