A speech recognition system has an operator interface that helps a user navigate new or changing grammars. The speech recognition system includes an application, a vocabulary that holds a set of utterances applicable to the application, and an active grammar that holds a subset of the utterances in the vocabulary. The active grammar includes a help command as well as other options. The speech recognition system has a speech recognition engine to recognize the utterances in the active grammar, including the help command. When the help command is detected, the operator interface verbally enunciates the utterances in the grammar so that the user can hear the available options in a hands-free, eyes-off operating environment. In this way, any time the user does not know or forgets the available options, he/she simply speaks the help command, such as asking "What can I say." The system responds verbally with a list of currently available words and/or phrases in the active grammar.