semantics

This dissertation presents theoretical and technical support for, and implementations of, narrative computational media works with the following characteristics: generative content, semantics-based interaction, reconfigurable narrative structure, and strong cognitive and socio-cultural grounding. A system that can dynamically compose media elements (such as procedural computer graphics, digital video, or text) to result in new media elements can be said to generate content. The GRIOT system, a result of this dissertation, provides an example of this. It has been used to implement computational poetry that generates new narrative poems with varying particular concepts, but fixed themes, upon each execution. This generativity is enabled by the Alloy system, which implements an algorithm that models key aspects of Gilles Fauconnier and Mark Turner's theory of conceptual blending. Alloy is the first implementation of Joseph Goguen's algebraic semiotics approach to blending. (Fauconnier & Turner, 2002; Goguen, 1998) This research also contributes to the theory of algebraic semiotics by developing a blending-based notion of style. Semantics- based interaction means here that (1) media elements are structured according to the meaning of their content, and (2) user interaction can affect content of a computational narrative in a way that produces new meanings that are constrained by the system's author. "Meaning" in this case indicates that the author has provided formal descriptions of domains and concepts pertinent to the media elements and subjective authorial intent. Meaning can also be reconfigured at the level of narrative discourse. The formal structure of a computational narrative can be dynamically restructured, either according to user interaction, or upon execution of the system as in the case of narrative generation. Strong cognitive and socio- cultural grounding here implies that meaning is considered to be contextual, dynamic, and embodied. The formalizations used derive from, and respect, cognitive linguistics theories with such notions of meaning. Furthermore, the notion of narrative here is not biased toward one particular cultural model. Using semantically based media elements as a foundation, a cultural producer can implement a range of culturally specific or experimental narrative structures.

People draw on many diverse sources of real-world knowledge in order to make up stories, including the following: knowledge of the physical world; rules of social behavior and relationships; techniques for solving everyday problems such as transportation, acquisition of objects, and acquisition of information; knowledge about physical needs such as hunger and thirst; knowledge about stories their organization and contents; knowledge about planning behavior and the relationships between kinds of goals; and knowledge about expressing a story in a natural language. This thesis describes a computer program which uses all information to write stories. The areas of knowledge, called problem domains, are defined by a set of representational primitives, a set of problems expressed in terms of those primitives, and a set of procedures for solving those problems. These may vary from one domain to the next. All this specialized knowledge must be integrated in order to accomplish a task such as storytelling. The program, called TALE-SPIN, produces stories in English, interacting with the user, who specifies characters, personality characteristics, and relationships between characters. Operating in a different mode, the program can make those decisions in order to produce Aesop-like fables.

(Source: Author's abstract)