Literary Psychology-Modeling Figure-Ground Structures in Narrative Stories: Towards a Narrative Content Generation

Literary Psychology-Modeling Figure-Ground Structures in Narrative Stories: Towards a Narrative Content Generation

Eugene Schneider Kitamura
Copyright: © 2018 |Pages: 36
DOI: 10.4018/978-1-5225-4775-4.ch006
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Today more than ever, narrative content generation has become important. This is due to the advances and accessibility of computational devices. As these devices become more familiar to people and easier to handle, there will be greater expectations for autonomous functioning and desires for a natural communication with the users. To achieve such demands, computational devices need to process and generate higher levels of meanings such as context and abstraction of topics. This chapter gives a background on topics that have been developed so far in content analysis and content generation, but focuses mainly on the figure-ground impression model, for both analysis and generating narrative context. By focusing on the characters and their attributes in the text, not only is this model able to represent the figure-ground impressions qualitatively, but also quantitatively. Such a feature may be useful to execute in computational devices such as artificial intelligence.
Chapter Preview
Top

Background

Some of the early formal efforts of creating stories with computers are TALE-SPIN by Meehan (1976), MINSTREL by Turner (1993), and BRUTUS by Bringsjord and Ferrucci (2000). They were developed with different focus and approaches. TALE-SPIN generates its stories based on a character’s goals and properties. A story unfolds as believable characters with unique traits interact with each other. Since the driving force of the story was the individual characters’ traits and motives, it may realize short term events, but the overall story may not be coherent or dramatic. MINSTREL was designed based on the idea that story generation is a series of incidental problem solving. Solutions are created by combining already existing solutions that are stored in an already existing set or gained from past experiences. This program allows authors to include their intent, giving an overall theme or a moral message to the plot (Swartjes, Vromen, & Bloom, 2007; Kybartas & Bidarra, 2017). BRUTUS is a system that generates complex stories, but limited to plots based on a betrayal situation. The program is designed to follow the logic of betrayal events as it generates stories. Although the stories include figurative language and dialogues, this program is basically reverse- engineered from an already existing story. The creator’s objective was to see how closely the original story can be regenerated (Gervás, 2012; Kybartas & Bidarra, 2017). Although the use of computer programs to generate stories has been around for a while, scholarly efforts of abstracting stories or the use of mathematical methods to model plots go back further. Propp published a book titled Morphology of the Folktale in 1928 (Propp, 1928). He analyzed a large number of Russian folktales and broke them down into smaller story units and organized these subunits according to two categories: functions and characters. There are 31 functions which are story patterns such as “absentation,” “interdiction,” etc. and 8 character types such as “villain,” “hero,” etc. This work is significant in that it abstracted stories into universal patterns that can be applied to construct other stories.

Complete Chapter List

Search this Book:
Reset