LINGUIST List 24.2936

Thu Jul 18 2013

Review: Computational Linguistics; Linguistics & Literature: Mani (2012)

Editor for this issue: Joseph Salmons <>

Date: 02-Jun-2013
From: Choonkyu Lee <>
Subject: Computational Modeling of Narrative
E-mail this message to a friend

Discuss this message

Book announced at

AUTHOR: Inderjeet ManiTITLE: Computational Modeling of NarrativeSERIES TITLE: Synthesis Lectures on Human Language TechnologiesPUBLISHER: Morgan & Claypool PublishersYEAR: 2012

REVIEWER: Choonkyu Lee, Utrecht Institute of Linguistics OTS

SUMMARYThis book presents core concepts and insights from humanities narratology,such as time, plot, and the narrator, for automatic narrative understandingand generation in natural language processing and artificial intelligence,including interactive entertainment. It also describes how computationaldevelopments may also contribute to narratological theory and digitalhumanities. The book is mainly a primer on narratology and a review ofcomputational developments for computational linguists and game developers,but may also be of interest to cognitive scientists and narratologistsinterested in applications.

Chapter 1, Narratological Background, provides an overview of the mainnarratological concepts addressed and establishes terminology. Mani defines‘narrative’ with a pre-theoretic notion of ‘story,’ to include parts of blogs,emails, news, and technical and literary works, among others, that involvestorytelling with causal coherence. The relevant modes of presentation (ormedia) also include games, as game developers with an interest in interactivenarrative are among the primary intended audience for the book. The authoralso defines ‘Narrative Structure’ to encompass global dimensions of phenomenarelated to narrative representation, including narrator and audiencevariables. In this way, his view of narrative structure goes beyond discourseformalisms such as Discourse Representation Theory (Kamp, 1984) and RhetoricalStructure Theory (Mann & Thompson, 1988), which focus on relations ofcoreference, time, and/or communicative roles at a more local level. One ofthe most important distinctions that Mani makes -- relevant throughout thebook -- is between ‘fabula’ as the underlying (story) content and ‘discourse’as its expression or form. In a section on narrator characteristics, heintroduces Genette’s (1980) narratological concepts of a homodiegetic narrator(who participates in the story) and a heterodiegetic narrator (who doesn’t),as well as the narrative distance continuum (Genette, 1988), which isrepresented in Mani’s annotation scheme, NarrativeML, with six values, rangingfrom narrated speech to immediate speech. Mani also introduces different casesof narrator perspective: non-focalized (omniscient), internally-focalized, andexternally-focalized (Genette, 1980), with room for ‘other’ possibilities inNarrativeML. He also discusses embedded stories, as in “The Thousand and OneNights,” in which readers often have to revise their beliefs about the factualstatus of characters or events. Subordinated discourse with propositionalattitudes also helps determine the factual status of events. With regard tonarrative time, Mani introduces Genette’s (1980) narrative time relations --subsequent, simultaneous, and prior -- indicated by the dominant tense in thenarrative. He also mentions Chatman’s (1980) distinction between story time(in the fabula) and discourse time -- discourse processing time in Chatman’soriginal formulation, but operationalized as textual length for Mani’sNarrativeML. Mani uses the ratio between the two to represent narrative pace,which indicates pacing techniques such as stretching and speed-up. Seven typesof narrative ordering from Genette’s (1980) account -- chronological(‘Chronicle’) and others -- are also represented in NarrativeML, indicatingthe relationship between discourse order and the actual order of events in thefabula. The issue of audience response, or reader affect, is also mentioned.After a brief discussion of the ontology for a fabula or story world, Manialso introduces accessibility relations, used for representing narrativeembeddings and subordinated discourse with propositional attitudes whileavoiding the representation of multiple worlds (multiple story worlds and theactual world). Changes in accessibility relations can represent the audience’sbelief revision, e.g., in a plot twist. The chapter ends with a description ofNarrativeML, an annotation scheme that incorporates many of the narratologicalconstructs. Such a scheme can be used for human annotation of corpora, and theannotated corpora can then be used for supervised learning of systems forautomatic narrative understanding and generation. The possibility of importingother annotations from the PropBank (Palmer, Gildea & Kingsbury, 2005) andother schemes is also discussed.

In Chapter 2, Characters as Intentional Agents, Mani presents the ‘planningperspective’ on narrative, in which the task of narrative understanding is torecognize the characters’ goals and plans, and that of narrative generation isto synthesize a series of plans. Since there have been more developments in AIon intentionality than in humanities narratology, Mani focuses more on the AIside in this chapter while suggesting how these developments may provideinsight for narratology. Interpreting actions and events in terms of goals andplans presumes world knowledge. Preconditions and consequences for actions andstereotypical structures of event sequences have been represented in narrativeunderstanding systems, such as those based on Schank and Abelson’s (1977)scripts, but Mani points out that these systems were too domain-specific andfocused too much on events rather than characters. He also mentions the use ofcase-based reasoning in both understanding and generation, which allowsgeneralization from stored narrative fragments beyond simple retrieval. Inparticular, he points out some general limitations in story planning systems,such as domain specificity, and lack of natural variation in narrative order,distance, etc. For purposes of ‘lightweight’ annotation with NarrativeML, Maniadopts Pavel’s (1982) Move-grammar, which provides a coarse-grained analysisof goal structure in the fabula -- often corresponding to long stretches oftext. Pavel’s analysis decomposes an action into a problem and a solution. Inthe section on interactive narrative, which often involves incremental,non-monotonic plan revision after new events, Mani discusses Mateas andStern’s (2005) FAÇADE to illustrate dynamic re-planning in reaction toaudience feedback, but points out the problem of a large number of branchingpossibilities having to be spelled out. He suggests that a good user model forreader affect or aesthetic preference may help constrain possible paths ininteractive narrative generation, proposing a Boolean model of the reader’sattitude toward the agent of an event -- positive (sympathy), negative(antipathy), or neutral. Mani also mentions the possibility of using sentimentanalysis for a model of narrator attitudes as well, though not represented inhis NarrativeML. He also discusses the problem of balance between authorialand audience control in interactive narrative. The representations ofintentionality discussed in this chapter are closely tied to Chapter 4, Plot.

In Chapter 3, Time, Mani focuses on narrative understanding for story time.After a quick review of rule-based systems making use of tense and aspectualmarking or causal knowledge, he introduces Allen’s (1984) interval calculuswith seven basic temporal relations, which is adopted in NarrativeML and canbe extended with logics for branching time for underspecified relations. Manithen discusses TIMEX2 (Ferro et al., 2005) and TimeML (Pustejovsky et al.,2005) annotation schemes for tagging duration, end time, and relations amongevents and times, in which subscript indices indicate the order of mention,allowing inferences about narrative ordering in relation to the actual orderof events in the story world. Subordinating links, or SLINKs, for relationssuch as remembering require branching time models. Discourse time, measured innumber of words, can be compared to story time in the fabula to measurenarrative pace, or tempo. The author also mentions an interesting possibleextension with estimates of minimum and maximum duration of events torepresent commonsense intuitions (Pan, Mulkar-Mehta & Hobbs, 2011). Issues inhuman annotation with temporal links among times and events, or TLINKs, arediscussed. Successful automatic tagging in these temporal aspects of narrativecan facilitate narratological investigations. In a brief section on automaticnarrative generation, Mani describes two particular systems, Callaway’s (2000)STORYBOOK and Montfort’s (2011) CURVESHIP. Back to narrative understanding,Mani reviews recent success in automatic tagging and resolution of temporalexpressions, as well as recent developments in automatic tagging of events,factuality, coarse-grained duration, and TLINKs. For temporal relationclassifier systems based on local pair-wise decisions, the problem of globalinconsistency may arise. Solutions combining a ranking method with IntegerLinear Programming or Markov Logic Networks are discussed. Habituals andscene-setting descriptions also pose a challenge to narrative timerepresentation.

In Chapter 4, Plot, Mani provides background on important concepts, includingabstract event summaries and Aristotelian mythos, based on a view of plot as acompact structural unit with emphasis on event sequences. Other narratologicalconcepts regarding plot, such as a turning point, a narrative arc with stages,the heroic quest, and Propp’s (1968) narrative functions, are discussed, alongwith applications to interactive narrative. Overly fine-grained distinctionsare often impractical for reliable annotation, but the use of more globalstructural representations based on story grammars (Rumelhart, 1977) andMacrostructures (van Dijk, 1980) in automatic story generation is discussednext. In this section, Mani repeats the distinction between story time at thelevel of fabula and discourse time as text length, pointing out thatgeneration systems often fail to decouple the two and claiming that Pavel’sMove-grammar adopted into Mani’s NarrativeML may improve upon that aspect.Mental states and intentionality in plot are then emphasized, withdescriptions of previous accounts involving affect states (+, -, and a neutralM; Lehnert, 1981), which capture motivations and intentional actualizations.Despite the coarse-grained differentiation of emotional states, classifyingthe (emotional) polarity of verbs based on their arguments may improverepresentation of causality (Goyal, Riloff & Daumé III, 2010). Recurringpatterns of transitions are considered plot units, and applications includingElson’s (2012) Story Intention Graphs for DramaBank are discussed.Acknowledging the inherent difficulty of inferring intentional states that arenot directly expressed, Mani mentions event summarization approaches, such asChambers’ (2011) Narrative Event Chain, combined with salience filtering orevent-based causal reasoning algorithms as a potentially more feasiblealternative. After a quick overview comparing the different plot modelsdiscussed in the chapter, Mani suggests some applications for narratology,such as intelligent searching for stories with similar plots.

In the final chapter, Summary and Future Directions, the author summarizeseach previous chapter, and illustrates the representation of the major aspectsof a narrative in NarrativeML with an example (pp. 96-99), mentioning theproblem that long literary genres with substructural units such as scenes orepisodes would pose a challenge to annotation efforts. In his concludingremarks, Mani speculates that developments in narratological theories ofcharacter psychology may continue the tradition of narratological insights forcomputational applications, and also hints at the opposite direction ofinspiration, namely, automatic narrative computing systems inspiringdevelopments in digital humanities with enhanced search, analysis,translation, clustering, and recommendation, among others.

EVALUATIONThis book offers an easy-to-read introduction to the core issues in narrativerepresentation, both traditional narratological insights and more recentcomputational developments. Mani defines the terminology and explains hischoice of terms carefully, and suggests promising interdisciplinarycontributions between humanities narratology and computational narratologythroughout, achieving his main goals. He concludes each chapter with a clearillustration of how the aspects of narrative discussed in the chapter arerepresented in NarrativeML.

The book is primarily intended for computer scientists working on narrativeprocessing and generation and for narrative theorists interested inapplications, and its direct relevance to other fields of cognitive science,including formal semantics, seems more limited than the back cover suggests.For example, Mani himself points out that he is more concerned with an entirestory at a more global level, compared to the more local focus of DiscourseRepresentation Theory or Rhetorical Structure Theory. In addition, althoughsome interesting experimental findings (e.g., Gerrig & Bernardo, 1994) andhuman annotation studies (e.g., Pan, Mulkar-Mehta & Hobbs, 2011) arementioned, discussion of relevant findings in cognitive psychology (e.g.,Graesser, Singer & Trabasso, 1994, on causality and intentionality; Zwaan &Radvansky, 1998, on the situation model of discourse) is otherwise lacking.Comparing NarrativeML representation to Zwaan and Radvansky’s (1998) situationmodel reveals that the core dimensions of narrative align well between the twomodels, but that in NarrativeML spatial relations are not represented to alevel of granularity that the situation model would predict or to the samelevel of granularity as temporal relations in NarrativeML. IncorporatingMani’s recent work (Mani & Pustejovsky, 2012) on spatial representation may beuseful for narrative representation as well. Another possible addition toNarrativeML is to represent character prominence, which has been found to beimportant for narrative production (e.g., Sanford, Moar & Garrod, 1988).

Narrative computing is an exciting field, very much burgeoning, and it will beinteresting to see how Mani’s proposals, e.g., the use of Pavel’s (1982)Move-grammar for representing character goals, stand the test of time.

REFERENCESAllen, J. (1984). Towards a general theory of action and time. ArtificialIntelligence, 23(2), 123-154.

Callaway, C. (2000). Narrative prose generation. Doctoral dissertation, NorthCarolina State University.

Chatman, S. (1980). Story and discourse: Narrative structure in fiction andfilm. Ithaca: Cornell University Press.

Ferro, L., Gerber, L., Mani, I., Sundheim, B. & Wilson, G. (2005). TIDES 2005standard for the annotation of temporal expressions.

Genette, G. (1980). Narrative discourse (J. Lewin, Trans.). Ithaca: CornellUniversity Press.

Genette, G. (1988). Narrative discourse revisited (J. Lewin, Trans.). Ithaca:Cornell University Press.

Gerrig, R. & Bernardo, D. (1994). Readers as problem-solvers in the experienceof suspense. Poetics, 22, 459-472.

Goyal, A., Riloff, E. & Daumé III, H. (2010). Automatically producing plotunit representations for narrative text. In Proceedings of EMNLP 2010 (pp.77-86). Cambridge, MA.

Graesser, A. C., Singer, M. & Trabasso, T. (1994) Constructing inferencesduring narrative text comprehension. Psychological Review, 101(3), 371-395.

Kamp, H. (1984). A theory of truth and semantic representation. In J.A.G.Groenendijk, T.M.V. Janssen & M.B.J. Stockhof (Eds.), Truth, Interpretation,and Information (pp. 277-322). Dordrecht: Foris.

Lehnert, W. G. (1981). Plot Units: A narrative summarization strategy. In W.G.Lehnert & M.H. Ringle (Eds.), Strategies for Natural Language Processing.Hillsdale, NJ: Lawrence Erlbaum.

Mani, I. & Pustejovsky, J. (2012). Interpreting motion: Groundedrepresentations for spatial language. New York: Oxford University Press.

Mann, W.C. & Thompson, S.A. (1988). Rhetorical structure theory: Towards afunctional theory of text organization. Text, 8(3), 243-281.

Mateas, M. & Stern, A. (2005). Structuring content in the Façade interactivedrama architecture. In Proceedings of the AIIDE 2005 (pp. 93-98). Menlo Park,CA: AAAI Press.

Montfort, N. (2011). Curveship’s automatic narrative variation. In Proceedingsof the 6th International Conference on the Foundations of Digital Games (pp.211-218).

Palmer, M., Gildea, D. & Kingsbury, P. (2005). The Proposition Bank: A corpusannotated with semantic roles. Computational Linguistics, 31(1), 71-105.

Pan, F., Mulkar-Mehta, R. & Hobbs, J.R. (2011). Annotating and learning eventdurations in text. Computational Linguistics, 37(4), 727-752.

Pavel, T. (1982). Plot-structure and style: Remarks on an unstablerelationship. Canadian Review of Comparative Literature, 9(1), 27-45.

Propp, V. (1968). Morphology of the folktale (L. Scott, Trans.). Austin:University of Texas Press.

Pustejovsky, J., Ingria, B., Saurí, R., Castano, J., Littman, J. Gaizauskas,R., Setzer, A., Katz, G. & Mani, I. (2005). The specification language TimeML.In I. Mani & J. Pustejovsky (Eds)., The language of time: A reader (pp.545-558). New York: Oxford University Press.

Sanford, A.J., Moar, K. & Garrod, S.C. (1988). Proper names as controllers ofdiscourse focus. Language and Speech, 31(1), 43-56.

Schank, R.C. & Abelson, R.P. (1977). Scripts, plans, goals, and understanding:An inquiry into human knowledge structures. Hillsdale, NJ: Lawrence Erlbaum.

Zwaan, R.A. & Radvansky, G.A. (1998). Situation models in languagecomprehension and memory. Psychological Bulletin, 123(2), 162-185.

ABOUT THE REVIEWERChoonkyu Lee is a postdoctoral researcher at the Utrecht Institute ofLinguistics OTS. His research interests include time in narrative discourseand commonsense knowledge in semantics/pragmatics, with interdisciplinaryapproaches involving cognitive psychology, theoretical linguistics, andcomputational linguistics.

Page Updated: 18-Jul-2013