A Component-Based Architecture for Suspense Modelling

Abstract

Suspense is a key narrative issue in terms of emotional gratification, influencing directly the way in which the audience experiences a story. Disciplines like psychology, neurology or e-learning study the suspense as the basis of useful techniques for the treatment of mental diseases or improving memory skills and the comprehension. In the field of creativity, it’s an essential cross strategy found in almost any book, film and video-game plots, regardless of technology and genre. With the objective of generating engaging stories, some automatic storytelling systems implement a suspense generation module. These systems are mainly based on narrative theories. However, we observe a lack of aspects from behavioral sciences, involving the study of empathy and emotional effect of scene objects in the audience. Generated plots with an adequate treatment of these features may involve benefits in areas as education and psychology. In this paper, we propose a component-based architectural model that firstly aims to identify and extract all these individual factors of the suspense from a scene; in a second step, the system calculates the level of suspense using a weighted corpus; in the last step, it alters those elements to increase or decrease the original suspense level and reassembles them in a new scene. Further, we discuss the model facing the development challenges and its practical implications.