Left-to-Right Analysis of Spoken Language