We propose a dialogue interpreting model for telephone inquiry dialogues and an application method of predicting the next utterance using the model for a speech-language interface. This method can provide context-sensitive information about the appropriate next utterance on an abstract level. Using this type of information and linguistic and pragmatic knowledge to connect the information to the corresponding surface linguistic expressions, selecting the correct candidate from speech recognition output in a context-sensitive way can be realized. Keywords: spoken dialogue, speech recognition and understanding