Chapter 10. Designing Prompts
| A VUI dialog is a type of conversation. Perhaps this is even more the case for telephony-based VUIs; after all, the telephone itself is an icon of conversation in modern culture. Even though one of the participants in a VUI dialog is a machine, both the user and the designer who engineered the dialog share a few basic expectations that also hold for human-to-human conversations. Both parties expect the use of a common language; each expects the other to be cooperative; and each expects that the other will possess basic cognitive faculties such as intelligence, short-term memory, and an attention span (which are artificial in the case of the machine but are nonetheless expected). Another level of expectation is implicit in the way users approach spoken language systems, and it has to do with the way prompts are worded. Humans approach all linguistic tasks with unconscious expectations and assumptions not only about content but also about form. Because VUI prompts are communicated through spoken language, their form should satisfy the spoken language norms that users naturally expect. And the spoken language system that VUI users are most familiar with is everyday conversation. We recommend that prompts be modeled after conversational language, in the interest of promoting the design of human-centered, user-friendly, optimally comprehensible VUIs. The alternative is a way of communicating that is less familiar and less natural and therefore less comprehensible. This chapter demonstrates how specific elements of human-to-human conversation can be applied to prompting. We review some basic linguistic concepts for creating a spoken language experience that users will find familiar and engaging. You will also see a number of cases where the intelligibility of prompts is greatly enhanced via the use of conversational language. It is sometimes thought that "conversational" necessarily implies informality and chattiness. For our purposes, conversational refers more technically to the linguistic form of everyday spoken interactions. Indeed, there are many types of conversations in real life, from chatty and informal to serious and formal. The high-level recommendations made throughout this chapter apply to the design of all VUIs, including those that require an air of formality. It is also sometimes thought that the goal of conversational interfaces is to lead unwitting VUI users into thinking that they are interacting with a human being, but this is not what we advocate. Our goal is to leverage the linguistic forms that VUI users are best acquainted with. In fact, linguistic research tells us that conversational language incorporates many design features that are specially adapted and ideally suited to the task of communicating via sound. Thus, many of the principles in this chapter are about improving listening comprehension and enhancing the user's level of comfort by exploiting the linguistically familiar. In designing speech applications, we are also designing a language experience for the user. Although speech technically refers to such matters as acoustics, physiology, sound production, and perception, language involves the intimate connection of structure, meaning, and use, all of which are context-sensitive. Linguistic form is shaped by forces that range from perceptual and cognitive, on the one hand, to social and cultural on the other. The form of conversational language is not haphazard but rather is systematic and principled. Dialog designers should refer to this naturally occurring system as a resource for prompting, at least to the extent that the current technology allows. As much as possible, naturally occurring spoken language should serve as a guide for writing successful speech applications. The ideas presented in this chapter should assist you in creating socially and linguistically appropriate characters for your application (see Chapter 6 for more on persona design). It is a common belief that an application's persona depends on hiring the right voice talent. In fact, not even the most experienced, skilled actor with the most pleasantly resonant voice can make inherently nonconversational prompts sound natural for example, "Please say the duration" instead of a more natural wording such as, "How long do you expect it to last?" As in the movies, good actors depend on good scripts. To create the sense of a familiar, engaging persona to represent your company's brand or image, conversational norms should guide the way prompts are written. Reaping the benefits of a well-designed persona depends on your ability to write messages that suggest natural spoken language. In usability studies of a voice-browser application, callers expressed a sense of loyalty to the professional, matter-of-fact, but personal and familiar character representing the system an administrative assistant in her early 40s. Users felt comfortable entrusting to her their most sensitive personal information, such as credit card details. But a well-designed, likeable persona is essentially an illusion, a mosaic built up of consistent, conversational messages that collectively suggest a coherent, sociolinguistically familiar presence. Usability informants reacted favorably not only to the voice itself, but also to the way prompts were worded (discussed in this chapter) and to the way they were spoken (discussed in Chapter 11). The remainder of this chapter covers the following topics:
|
