Meta-technical Utterances

In a tutorial setting, we often hear expressions such as “The method we are about to discuss will help you solve . . . ” or “Let us consider a subject which demands some more practice,” which are issued by a tutor to motivate a student to attend to forthcoming discourse. In this paper we model the meaning of these expressions in terms of their anticipated influence on the status of a listener’s goals, and use these predictions to produce motivational expressions and embed them in computer generated discourse. In particular, we have recognized relations which are instrumental in determining a listener’s motivational requirements in a hierarchical problem-solving domain. These ideas have been incorporated into a system called FIGMENT which generates commentaries on the solution of algebraic equations. In a tutorial setting a student is constantly exposed to Technical Utterances in the form of explanations, definitions, descriptions and problems to be solved. Intelligent Tutoring Systems (Clancey 1979, Genesereth 1978, Sleeman and Brown 1982) and text generation systems (McKeown 1985, Appelt 1982, Mann and Moore 1982) have addressed the problem of determining the type of technical utterance to be presented and the information it should contain. However, in discourse produced by a human tutor we also notice the presence of expressions such as “however,” “ this technique demands some more practice, ” “as I said before” and “next,” which are not part of the subject matter (Fames 1973, Hallyday and Hassan 1976, Longacre 1976, Hoey 1979, Winter 1968). These expressions, denoted Meta-Technical Utterances (MTUs), carry important information which assists the listener in assimilating the knowledge being transferred (Zukemm 1986, Zukerman and Pearl 1986). In this paper we focus on one type of MT& namely Motivational MTUs, which are used to motivate a student to perform prescribed tasks such as attending to forthcoming discourse and solving given problems. We attempt to gain insight into the mechanisms used by peo* Part of this work was performed at the Cognitive Systems Laboratory, Computer Science Department, University of California, Los Angeles, and supported by the National Science Foundation Grants IST 81-19045 and DCR 83-13875. ple to generate these MTUs by building and implementing a generative model of their meaning. This model has been incorporated into an Intelligent Tutoring System called FIGMENT, enabling it to produce a variety of Motivational MTUs. For instance: 1. “We shall now consider a topic, namely quadratic equations, which we have not seen for a while” This motivation is issued to prompt a student to practice a topic which he may be forgetting. 2. “This alternative serves to introduce the very important and interesting method of factoring out common factors” A tutor uses this motivation to awaken interest in a new item of knowledge. 3. “This type of equation has been practiced a lot, but it still demands some more practice” This motivation is generated to encourage a, probably tired, student to continue practicing a subject in which he lacks proficiency. In the following section we present a goal-based taxonomy of Motivational MTUs. Then we examine the mechanism used by FIGMENT to generate them. During the learning process a student is expected to exhibit the goal of mastering the subject matter. In addition, a typical student usually has a host of other goals, such as: achievement goals (passing a test, getting a good job), social goak (earning the respect of his peers, gaining the approval of the teacher), enjoyment goals (remaining interested and amused during lectures, being able to rest), etc (see Schank and Riesbeck 1981). At any point in time, a goal is either non-existent or it can exhibit varying degrees of activity. A goal is nonexistent if the entities involved are not represented in the listener’s memory. For example, a person who has never heard of university cannot have a goal of studying there. A goal may become active due to the occurrence of an external event, e.g., if we find out there is a new movie featuring our favourite actor, our goal of seeing it becomes active; or by gradual build-up over a period of time, e.g., if a student has been studying for quite a while, his enjoyment goal of being able to rest is strengthened. The level of activity of a goal may decrease over a period of time if it was not reactivated and competing goals became active. In addition, a goal may become inactive if the listener believes that it has been Zaakerman 327 From: AAAI-87 Proceedings. Copyright ©1987, AAAI (www.aaai.org). All rights reserved.