A method for the approximation of incremental understanding of explicit utterance meaning using predictive models in finite domains

David DeVault, David Traum