Speech acts redux: Beyond request-response interactions

Abstract

We communicate to (1) express how we feel, (2) share observations about the world, (3) commit to future acts, (4) request others to do things, and (5) change the state of the world according to pragmatics. Of these categories, today's conversational interfaces like Siri and Alexa are mainly designed to fulfill our imperatives, i.e., to respond to our requests on command. Yet, could future conversational interfaces go beyond request-response interactions? One way forward is to consider what conversational interactions allow us to do with language. Not only can we send request to CUIs, but we can also share our emotions, attitudes, beliefs, and promises as speech acts---acts we regularly perform with other humans. To open up pragmatics as an under-investigated design space for conversational technologies, I elaborate on what pragmatics and affective pragmatics are and give examples involving conversational agents. As a theoretical contribution, I provide a taxonomy from pragmatics and affective pragmatics to move beyond request-response interactions. The aim is to extend our conversational experiences with technology to cover the full spectrum of everyday speech acts. Our words can change the world; expressions to CUIs can also do so.