The Contribution Of Phonotactic And Lexical Information In The Segmentation Of Multi-Word Utterances