Word-learning as logical inference: Disjunctive syllogism and the word-learning strategy mutual exclusivity across development
The past fifteen years have seen a proliferation of proposed word-learning constraints that are hypothesized to guide a word-learner towards the correct referent of a newly heard word. Because much of this work has relied on categorical measures of success it has not been possible to observe the processes that lead to success. As such, it has been difficult to specify at what cognitive level a particular constraint might be operating and it has been impossible to document the mental computations that underlie and support the use of these constraints. In this thesis I develop and apply a new methodology that uses eye-tracking to uncover the mental computations that support the word-learning strategy of mapping novel labels (i.e. labels that the child has not heard before) to novel objects (i.e. objects that they child does not know a name for). I document the development of this strategy from 14-months of age through adulthood. This work provides evidence for two developmental shifts. First, children develop from a failure to use this strategy to success at mapping novel labels to novel objects by 17months of age. Second, the mental computations that underlie this strategy change from what may be an associative strategy to the deductive logical strategy Disjunctive Syllogism at 2-years and beyond.
0633: Cognitive therapy
0525: Educational psychology