Victor Gomes
Department of Psychology
Harvard University

“Does not compute? The development of real-time negation processing”

Processing negation in real-time speech can be difficult even for adults, however past work has shown that they are able to quickly integrate it into their predictions when given proper contextual support. This is impressive in part because if a particular linguistic context is highly constraining (“Pushing people on the playground is…”), encountering negation (“not”) requires 1) discarding or inhibiting a strong prediction (“mean”) and 2) develop a new prediction from an in-principle larger space to arrive at the likeliest completion (“nice”). While we know that children can process negation in real-time (with some caveats) by the time they’re 3, these findings are based on tasks using a visual world which constraints the set of alternatives to two options (e.g., a broken plate and an unbroken plate). As such, these results can speak to 1 (inhibiting salient prediction up to negator) but not 2 (generating new prediction) because the set of alternatives is highly constrained and given. We seek to investigate children’s ability to both inhibit and generate using supportive linguistic contexts to investigate if, like adults, children are able to quickly integrate negation into their predictions in real-time. I will present on two experiments, an in-progress sentence completion study with children and a related EEG experiment still in piloting. I will present pilot and present data from the sentence completion study, which suggests that children 4-6 are very competent at the task overall but are slightly better with affirmative sentences, and adult pilot data from the EEG study, which is difficult to interpret at present. Audience participation is encouraged to help us uncover what’s going on with the EEG pilot.