Getting ahead of the game: top-down prediction during natural language comprehension in adults and children

Over the last thirty years, psycholinguists have converged on a model of 
language comprehension in which perceptual input (e.g., sounds) 
incrementally activates higher-level representations (e.g., possible 
messages), which in turn constrain the interpretation of subsequent 
percepts.  Such a system should allow for prediction: the use of 
top-down information to anticipate the meaning and form of upcoming 
words.  While form based prediction is sometimes present under 
laboratory conditions, there is no consensus on whether prediction is 
common in ordinary comprehension contexts.  In this talk, I describe a 
new EEG paradigm for studying spoken language comprehension in natural 
contexts (The Storytime Paradigm) and present our recent findings 
exploring the nature of prediction in adults and young children.