To make a selection of possible responses, the user may request that responses be audibly communicated through speaker 30 or headset 32 to assess the tone, quality, emphasis, accent, empathy, and other non-textual information conveyed in the responses.