Whose Story Is It Anyway? How Improv Informs Agency and Authorship of Emergent Narrative

I.M.T. Swartjes

Abstract

Within the field of interactive digital storytelling, one goal pursued is to build highly interactive fictional worlds in which a user can have the first-person experience of being a character in a story. The story events should unfold based in part on the user’s actions. This introduces a story representation challenge because branching narrative representations used in the games industry are insufficient. Story content must be expressed using novel procedural and generative (AI-based) representations. Emergent narrative is one approach proposed by the field to creating this kind of experience. In this approach, the storyworld is inhabited by a collection of intelligent, autonomous agents, each playing the role of a story character. As in dramatic improvisation, the story is not scripted, but emerges based on the interactions of these characters with each other and with the storyworld they inhabit. This thesis presents conceptual and technical contributions situated within the emergent narrative approach. In order to better understand user agency and authorship of emergent narrative, we took lessons from the theory and practice of dramatic improvisation. User agency in emergent narrative was explored using a human-only, chat-based experiment. Improvisers were given the task to provide participants with an engaging experience as a character in an improvised story. To explore authorship of emergent narrative, we developed The Virtual Storyteller, a story generator based on the emergent narrative approach. Authoring experiments led to the identification of an iterative, improvisational style of authoring in which the emergent properties of the system affect the creative process. Dramatic improvisation also informed character agent modeling. A distinction was proposed between two tightly coupled roles that the agent plays: that of a self-interested character in the storyworld, and that of a drama-interested actor. Modeled after human improvisers, the agents were given the means to delay decisions about the details of the storyworld they inhabit, until they became relevant for the unfolding narrative.
Original languageEnglish
Awarding Institution
  • University of Twente
Supervisors/Advisors
  • Theune, Mariet , Advisor
  • Nijholt, Antinus , Supervisor
Date of Award19 May 2010
Place of PublicationEnschede
Print ISBNs978-90-365-3004-0
DOIs
StatePublished - 19 May 2010

Fingerprint

Improvisation
Authorship
Authoring
Experiment
Industry
Procedural
Modeling
Generative
Interaction
First person
Creative process
Storyteller
Emergent properties
Drama
Fictional worlds
Digital storytelling

Keywords

  • IR-71295
  • Story Generation
  • HMI-IA: Intelligent Agents
  • Intelligent Agents
  • EWI-17893
  • HMI-CI: Computational Intelligence
  • Interactive Storytelling
  • METIS-270817
  • Improvisational theater

Cite this

@misc{adac836998d64d71847741e7ac34d8ff,
title = "Whose Story Is It Anyway? How Improv Informs Agency and Authorship of Emergent Narrative",
abstract = "Within the field of interactive digital storytelling, one goal pursued is to build highly interactive fictional worlds in which a user can have the first-person experience of being a character in a story. The story events should unfold based in part on the user’s actions. This introduces a story representation challenge because branching narrative representations used in the games industry are insufficient. Story content must be expressed using novel procedural and generative (AI-based) representations. Emergent narrative is one approach proposed by the field to creating this kind of experience. In this approach, the storyworld is inhabited by a collection of intelligent, autonomous agents, each playing the role of a story character. As in dramatic improvisation, the story is not scripted, but emerges based on the interactions of these characters with each other and with the storyworld they inhabit. This thesis presents conceptual and technical contributions situated within the emergent narrative approach. In order to better understand user agency and authorship of emergent narrative, we took lessons from the theory and practice of dramatic improvisation. User agency in emergent narrative was explored using a human-only, chat-based experiment. Improvisers were given the task to provide participants with an engaging experience as a character in an improvised story. To explore authorship of emergent narrative, we developed The Virtual Storyteller, a story generator based on the emergent narrative approach. Authoring experiments led to the identification of an iterative, improvisational style of authoring in which the emergent properties of the system affect the creative process. Dramatic improvisation also informed character agent modeling. A distinction was proposed between two tightly coupled roles that the agent plays: that of a self-interested character in the storyworld, and that of a drama-interested actor. Modeled after human improvisers, the agents were given the means to delay decisions about the details of the storyworld they inhabit, until they became relevant for the unfolding narrative.",
keywords = "IR-71295, Story Generation, HMI-IA: Intelligent Agents, Intelligent Agents, EWI-17893, HMI-CI: Computational Intelligence, Interactive Storytelling, METIS-270817, Improvisational theater",
author = "I.M.T. Swartjes",
note = "SIKS Dissertation Series No. 2010-20",
year = "2010",
month = "5",
doi = "10.3990/1.9789036530040",
isbn = "978-90-365-3004-0",
school = "University of Twente",

}

Whose Story Is It Anyway? How Improv Informs Agency and Authorship of Emergent Narrative. / Swartjes, I.M.T.

Enschede, 2010. 225 p.

Research output: ScientificPhD Thesis - Research UT, graduation UT

TY - THES

T1 - Whose Story Is It Anyway? How Improv Informs Agency and Authorship of Emergent Narrative

AU - Swartjes,I.M.T.

N1 - SIKS Dissertation Series No. 2010-20

PY - 2010/5/19

Y1 - 2010/5/19

N2 - Within the field of interactive digital storytelling, one goal pursued is to build highly interactive fictional worlds in which a user can have the first-person experience of being a character in a story. The story events should unfold based in part on the user’s actions. This introduces a story representation challenge because branching narrative representations used in the games industry are insufficient. Story content must be expressed using novel procedural and generative (AI-based) representations. Emergent narrative is one approach proposed by the field to creating this kind of experience. In this approach, the storyworld is inhabited by a collection of intelligent, autonomous agents, each playing the role of a story character. As in dramatic improvisation, the story is not scripted, but emerges based on the interactions of these characters with each other and with the storyworld they inhabit. This thesis presents conceptual and technical contributions situated within the emergent narrative approach. In order to better understand user agency and authorship of emergent narrative, we took lessons from the theory and practice of dramatic improvisation. User agency in emergent narrative was explored using a human-only, chat-based experiment. Improvisers were given the task to provide participants with an engaging experience as a character in an improvised story. To explore authorship of emergent narrative, we developed The Virtual Storyteller, a story generator based on the emergent narrative approach. Authoring experiments led to the identification of an iterative, improvisational style of authoring in which the emergent properties of the system affect the creative process. Dramatic improvisation also informed character agent modeling. A distinction was proposed between two tightly coupled roles that the agent plays: that of a self-interested character in the storyworld, and that of a drama-interested actor. Modeled after human improvisers, the agents were given the means to delay decisions about the details of the storyworld they inhabit, until they became relevant for the unfolding narrative.

AB - Within the field of interactive digital storytelling, one goal pursued is to build highly interactive fictional worlds in which a user can have the first-person experience of being a character in a story. The story events should unfold based in part on the user’s actions. This introduces a story representation challenge because branching narrative representations used in the games industry are insufficient. Story content must be expressed using novel procedural and generative (AI-based) representations. Emergent narrative is one approach proposed by the field to creating this kind of experience. In this approach, the storyworld is inhabited by a collection of intelligent, autonomous agents, each playing the role of a story character. As in dramatic improvisation, the story is not scripted, but emerges based on the interactions of these characters with each other and with the storyworld they inhabit. This thesis presents conceptual and technical contributions situated within the emergent narrative approach. In order to better understand user agency and authorship of emergent narrative, we took lessons from the theory and practice of dramatic improvisation. User agency in emergent narrative was explored using a human-only, chat-based experiment. Improvisers were given the task to provide participants with an engaging experience as a character in an improvised story. To explore authorship of emergent narrative, we developed The Virtual Storyteller, a story generator based on the emergent narrative approach. Authoring experiments led to the identification of an iterative, improvisational style of authoring in which the emergent properties of the system affect the creative process. Dramatic improvisation also informed character agent modeling. A distinction was proposed between two tightly coupled roles that the agent plays: that of a self-interested character in the storyworld, and that of a drama-interested actor. Modeled after human improvisers, the agents were given the means to delay decisions about the details of the storyworld they inhabit, until they became relevant for the unfolding narrative.

KW - IR-71295

KW - Story Generation

KW - HMI-IA: Intelligent Agents

KW - Intelligent Agents

KW - EWI-17893

KW - HMI-CI: Computational Intelligence

KW - Interactive Storytelling

KW - METIS-270817

KW - Improvisational theater

U2 - 10.3990/1.9789036530040

DO - 10.3990/1.9789036530040

M3 - PhD Thesis - Research UT, graduation UT

SN - 978-90-365-3004-0

ER -