Narrative Planning: Balancing Plot and Character

By Mark O. Riedl, R. Young. 2010

626 citations recorded by the source corpus.

Narrative, and in particular storytelling, is an important part of the human experience. Consequently, computational systems that can reason about narrative can be more effective communicators, entertainers, educators, and trainers. One of the central challenges in computational narrative reasoning is narrative generation, the automated creation of meaningful event sequences. There are many factors - logical and aesthetic - that contribute to the success of a narrative artifact. Central to this success is its understandability. We argue that the following two attributes of narratives are universal: (a) the logical causal progression of plot, and (b) character believability. Character believability is the perception by the audience that the actions performed by characters do not negatively impact the audience's suspension of disbelief. Specifically, characters must be perceived by the audience to be intentional agents. In this article, we explore the use of refinement search as a technique for solving the narrative generation problem - to find a sound and believable sequence of character actions that transforms an initial world state into a world state in which goal propositions hold. We describe a novel refinement search planning algorithm - the Intent-based Partial Order Causal Link (IPOCL) planner - that, in addition to creating causally sound plot progression, reasons about character intentionality by identifying possible character goals that explain their actions and creating plan structures that explain why those characters commit to their goals. We present the results of an empirical evaluation that demonstrates that narrative plans generated by the IPOCL algorithm support audience comprehension of character intentions better than plans generated by conventional partial-order planners.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Survey of the State of the Art in Natural Language Generation: Core tasks, applications and evaluation (2017, background): “…a knowledge base of story schemas, from which one is selected and elaborated using planning techniques to link causes and effects (see also Young, 2008; Riedl & Young, 2010, among others, for recent examples of the use of planning techniques to model the creative process in narrative generation).”
Plan-And-Write: Towards Better Automatic Storytelling (2018, background): “Prior research on story generation mostly focused on automatically composing a sequence of events that can be told as a story by plot planning (Lebowitz 1987; Perez and Sharples 2001; Porteous and Cavazza 2009; Riedl and Young 2010; Li et al. 2013) or case-based reasoning (Turner 1994; Gervas et al.”
Towards Controllable Story Generation (2018, methodology): “…in this ﬁeld relied on symbolic planning (Meehan, 1977; Lebowitz, 1987; Turner, 1993; Bringsjord and Ferrucci, 1999; Perez and Sharples, 2001; Riedl and Young, 2010), case-based reasoning (Gervas et al., 2005), or generalizing knowledge from existing stories to assemble new ones (Swanson…”
Co-Writing Screenplays and Theatre Scripts with Language Models: Evaluation by Industry Professionals (2022, methodology): “[90] Mark O Riedl and Robert Michael Young.”
Using Stories to Teach Human Values to Artificial Agents (2016, background): “…that can answer questions about stories (Schank and Abelson 1977; Mueller 2004), generate stories (see Gerv´as (2009) for an overview and (Riedl and Young 2010; Swanson and Gordon 2012; Li et al. 2013) for recent related work), respond affectively to stories (O’Neill and Riedl 2014), and…”
Automated Storytelling via Causal, Commonsense Plot Ordering (2020, methodology): “Narrative generation systems that use symbolic planning (Lebowitz 1987; Gervás et al. 2005; Porteous and Cavazza 2009; Riedl and Young 2010; Ware and Young 2011) explicitly ensure causal relations between actions via predicate calculus operations over explicitly modeled action preconditions and post”
Controllable Neural Story Generation via Reinforcement Learning (2018, background): “Thus, many previous approaches to story generation have relied on planning (Lebowitz 1987; Porteous and Cavazza 2009; Gervás et al. 2005; Riedl and Young 2010).”
Controllable Neural Story Plot Generation via Reinforcement Learning (2018, methodology): “Many prior approaches to plot generation relied on planning [Lebowitz, 1987; Gervás et al., 2005; Porteous and Cavazza, 2009; Riedl and Young, 2010].”
Toward Automated Story Generation with Markov Chain Monte Carlo Methods and Deep Neural Networks (2021, methodology): “…techniques include symbolic and logical planning (Meehan 1977; Lebowitz 1987; Cavazza, Charles, and Mead 2002; Porteous and Cavazza 2009; P´erez y P´erez and Sharples 2001; Riedl and Young 2010; Martens et al. 2014; Farrell and Ware 2016) and case-based reasoning (Turner 1994; Gerv´as et al. 2005).”
DramaQA: Character-Centered Video Story Understanding with Hierarchical QA (2020, background): “Furthermore, while focalizing on characters within a story is important to form local story coherence (Riedl and Young 2010; Grosz, Weinstein, and Joshi 1995), previous works did not provide any consistent annotations for characters to model this coherence.”
Goal-Directed Story Generation: Augmenting Generative Language Models with Reinforcement Learning (2021, background): “Early work on story generation used planning (Meehan, 1976; Lebowitz, 1987; Cavazza et al., 2003; Porteous and Cavazza, 2009; Riedl and Young, 2010; Ware and Young, 2010) or case-based reasoning (Pérez y Pérez and Sharples, 2001; Peinado and Gervás, 2005; Turner, 2014).”
Strategies for Structuring Story Generation (2019, neutral): “Traditional approaches organized sequences of character actions using hand crafted models (Riedl and Young, 2010; Porteous and Cavazza, 2009).”
Controllable Neural Story Plot Generation via Reward Shaping (2018, methodology): “Early story and plot generation systems relied on symbolic planning [Meehan, 1977; Lebowitz, 1987; Cavazza et al. , 2002; Porteous and Cavazza, 2009; Riedl and Young, 2010; Ware and Young, 2011] or case-based reasoning [P´erez y P ´ erez and Sharples, 2001; Gerv ´ as et al. , 2005 ] .”
Content Planning for Neural Story Generation with Aristotelian Rescoring (2020, methodology): “Riedl and Young (2010) use reﬁnement search as a technique to balance between character and plot for solving the narrative generation problem.”
Question Answering in the Context of Stories Generated by Computers (2016, methodology): “To that end, Riedl and Young (2010) defined a data structure on top of the base POCL representation that reified character intentions: Definition 6 (Frame of Commitment) A frame of commitment is a tuple 〈S′, P, a, ga, sf 〉 where S′ is a subset of steps in some plan P , a is a character, ga is some…”
Event Representations for Automated Story Generation with Deep Neural Nets (2017, methodology): “To date, most story generation systems have used symbolic planning (Meehan 1976; Lebowitz 1987; P ´ erez y P ´ erez and Sharples 2001; Porteous and Cavazza 2009; Riedl and Young 2010) or case-based reasoning (Gerv ´ as et al. 2005).”