Ontologies for Authoring, or Authoring Ontologies?

Diana Arellano

, Javier Varona

and Volker Helzle

Institute of Animation, Filmakademie Baden-W¨urttemberg, Ludwigsburg, Germany

Mathematics and Computer Science Department, Universitat de les Illes Balears, Palma de Mallorca, Spain

Keywords:

Ontologies, Storytelling, Knowledge Representation, Content Authoring.

Abstract:

In the last years the use of ontologies has broaden to areas that until some time ago were unthinkable, like

storytelling or context/content representation. The main problem with the use of ontologies is that the user

responsible of authoring the story needs to input every single element that is required for the story to make

sense. Depending on the case, this might be a tedious task. However, once it is done, different stories can

be developed by reusing the already deﬁned concepts. The objective of the paper is to provide examples of

applications where the use of ontologies conveyed “authoring effort” with satisfactory results. We also state

our opinion of why is it better to use ontologies for such tasks, explain our own experience with an use case

and propose ideas of what could be enhanced, or taken from other areas, to improve the authoring process.

1 INTRODUCTION

Storytelling, the interactive art of using words and ac-

tions to reveal the elements and images of a story,

while encouraging the listener’s imagination (Na-

tional Storytelling Network, nd) is an area that has

received special attention during the last years. Ac-

cording to (Peinado et al., 2004), the automatic con-

struction of story plots has always been a longed-for

utopian dream in the entertainment industry. Many

factors need to be taken into account to produce en-

gaging narrative, and when it comes to automatic

storytelling, creativity and methodology are two el-

ements that go together hand in hand.

Other reasons that have led researchers to ﬁnd bet-

ter techniques for automatic storytelling, or automatic

content creation, deal with economy, improvement

and independence of the narrative generation.

From the economical point of view, reducing the

times for story creation might translate into reduc-

ing costs. Having a tool that permits fast generation

of stories, without an specialised author, would be a

great asset (e.g. stories in video games for Human-

Computer-Interaction (HCI) applications).

Automatic and intelligent storytelling can also

be beneﬁcial in the resultant quality of the product.

Imagine the plot of a story where the event A causes

the event B, which in turn causes the event C, and so

on. If the author could have a software solution that

allows him/her to change just a portion of the story

and get the subsequent logical changes, then new and

original plots might arise.

If these scenarios were accomplished, that would

lead to an independence of the narrative generation,

where the author and the machine would deal with

the creation of not just one story, but as many as the

logic behind the storytelling generation allows.

Our intention with this paper is twofold. First we

will review and analyse different works that have ap-

proached storytelling through ontologies. Then we

will present our opinion in this matter based on our

experience developing a system for creation of daily

life-based stories. Our position is that although the

use of ontologies results in “authoring effort”, the ob-

tained results can open the insights of story generation

to a broader audience with more creative outcomes.

2 WHAT IS AUTHORING? WHY

ONTOLOGIES?

If we take a look at some of the English dictionaries,

we see that authoring is deﬁned as the process of cre-

ating content in any kind of media: books, websites,

or magazines. This process is performed by a per-

son, the author, who can be aided by using authoring

tools. An authoring tool is usually a piece of software

that facilitates, enhances or guides the author in the

creation of content.

364

Arellano D., Varona J. and Helzle V..

Ontologies for Authoring, or Authoring Ontologies?.

DOI: 10.5220/0004330503640369

In Proceedings of the 5th International Conference on Agents and Artiﬁcial Intelligence (ICAART-2013), pages 364-369

ISBN: 978-989-8565-39-6

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

Conversely, the conceptof ontology belongs to the

ﬁeld of philosophy since Aristotle characterised it by

the study of existence, a compendium of all there is

in the world. Nowadays, it has evolved in great mea-

sure in computer science related ﬁelds as artiﬁcial in-

telligence (L´opez et al., 2008), ubiquitous comput-

ing (Heckmann et al., 2005), body movement of vir-

tual characters (Garc´ıa-Rojas et al., 2006), and con-

text representation (Benta et al., 2007). An ontology

is deﬁned as an explicit speciﬁcation of an abstract,

simpliﬁed view of a world to represent. It speciﬁes

both the concepts related to this view and their inter-

relations (Obrenovic et al., 2005).

In this regard, we will analyse how two different

ﬁelds as ontologies and storytelling have merged as

a response to the problems and needs in the story-

telling world. From our perspective, content and con-

text are part of the storytelling in the sense that “con-

tent” would be what conforms the story, and “context”

what situates the character in a given scenario.

The following subsections explain the difference

between “authoring ontologies” and “using ontolo-

gies for authoring”. The ﬁrst one deals with the cre-

ation of the ontology itself, and to be effectively done

it requires expert knowledge; the second uses ontolo-

gies as the mechanism or tool for authoring the con-

tent of the stories.

2.1 Ontologies in Authoring

The creation of new stories, scenarios and situations,

requires not only a great amount of creativity, but also

a lot of methodical work in order to keep the story

interesting and meaningful.

In analogy to a puzzle, a ﬁnished story could be

seen as the assembled puzzle, while the ontology

would be the compendium of the pieces of the puzzle

that depending on how they are placed ﬁt together, or

not. Nevertheless, thanks to the power of inference of

the ontologies, these pieces can be correctly assem-

bled in different ways giving a variety of logical re-

sults. This would be achieved through “ontology en-

gineering”, which is described by (Chen et al., 1998)

as a research methodology that enables the accumu-

lation of knowledge and supports knowledge reuse

and sharing. The ultimate purpose of ontology en-

gineering is to “provide a basis for building models of

all things in which information science is interested

in the world” ( (Mizoguchi and Ikeda, 1996) cited

in (Chen et al., 1998)). Therefore, ontologies and

story authoring can be conveniently combined given

that the ﬁrst provides the method and the structure,

while the latter provides a quasi unlimited number of

scenarios to be described and inferred.

Among the researchers that haveworkedin the au-

thoring of ontologies, (Dimitrova et al., 2008) tried to

overcome the author’s need of expertise and knowl-

edge on ontologies for creation of content by devel-

oping ROO. ROO (Rabbit to OWL Ontology author-

ing) is a user-friendly tool that guides the authoring

of a conceptual ontology, which is then converted to

a logical ontology in OWL. As a result, none of the

ontologies created by inexperienced authors were us-

able without modiﬁcations. That led Dimitrova et al.

to consider the involvementof domain experts and ad-

ditional support besides the existent error messages.

(Konstantopoulos et al., 2009) proposed ELEON,

a system for ontology authoring that enables authors

with domain expertise, but no technological exper-

tise, to create a new application domain and the cor-

responding ontologies. Their domain of application

was cultural heritage repositories in order to automat-

ically generate adaptable and customised textual de-

scriptions of the cultural objects for a variety of audi-

ences and purposes. As a result, the ontology helped

to automatically infer missing proﬁle parameters, al-

leviating the burden of explicitly providing all neces-

sary details for large numbers of objects.

Other authoring tools include ONTO-

TRACK (Liebig and Noppens, 2005), which

combined a graph-based hierarchical layout with an

instant reasoning feedback in one single view. Liebig

and Noppens chose this approach because, in their

opinion, the most natural layout for an ontology is a

hierarchical layered graph. As a result ONTOTRACK

constituted a ﬁrst step towards an easy-to-use interac-

tive ontology editor, even for non-experienced users

and especially large ontologies.

In the same direction, the SWAT project devel-

oped a number of techniques and tools focused on

users who wished to encode or query knowledge with-

out deep understanding of RDF, OWL, or other spe-

cialist languages. From a system that generates math-

ematical word problems (Williams, 2011), to a tool

that allows users to read and edit axioms in natu-

ral language - SWAT Tools Verbaliser (Third et al.,

2011), their aim was to facilitate the ontology’s au-

thoring process.

2.2 Authoring based on Ontologies

In this section we review the other side of the coin:

works that haveused ontologies for authoring content,

speciﬁcally for creation of stories and narrative.

One of these works is the The Virtual Story-

teller (Swartjes and Theune, 2009). The Storyteller

is a multi-agent story generation framework based on

the emergent narrative approach, which states that

OntologiesforAuthoring,orAuthoringOntologies?

365

stories emerge from character behaviour. The au-

thoring system is implemented in an three-step it-

erative cycle: idea generation, implementation and

simulation. Regarding ontologies, the Virtual Sto-

ryteller makes use of a number of OWL ontologies:

The Storyworld Core upper ontology for the story-

world simulation knowledge, the Fabula upper on-

tology that deﬁnes elements of the fabula (e.g, Ac-

tions), domain-speciﬁc ontologies that determine sub-

classes of the Storyworld Core or Fabula ontolo-

gies, and Presentation-speciﬁc ontologies containing

classes and properties for specifying a lexicon and

common sense rules for natural language genera-

tion (Swartjes, 2010).

DGiovanni is another open source multi-agent ar-

chitecture for building interactive dramas. It uses

OWL ontologies to support the creation of different

stories and to feed the system with story-related in-

formation (M¨uller, 2011). In DGiovanni the user is

another actor of the story. Nevertheless in the author-

ing process to generate a new story, the user needs to

create the ontology with all the story-related concepts.

DRAMMAR is an ontology of drama presented

by (Damiano et al., 2005), which integrates the ba-

sic aspects of drama with agent-based theories. The

ontology consists of two levels. The directional level

encodes the speciﬁc traits of drama in a Drama-unit

(<Plot, Direction, Actor>). The actional level un-

folds the rational and the emotional perspective in

terms of the facts in the Drama-unit, the goal of the

Drama-unit and the attributes of the characters. These

facts can be actions, characters and character beliefs,

which allow the elicitation of emotions according to

the OCC model (Ortony et al., 1988).

Nakasone et al. (Nakasone and Ishizuka, 2006)

also presented a generic storytelling ontology model,

the Concept Ontology. They deﬁned a set of topics in

which the story, or part of the story, is based. These

topics are linked through a pseudo-temporal relation

that ensures a smooth transition between them. As

this ontology was for storytelling, the deﬁned classes

were related to scenes, acts, relations, characters of

the story and their roles. Its main advantage was that

it provided the elements for creating a story according

to narrative principles.

3 THE GOOD, THE BAD AND

THE UGLY

As it occurs in most of the situations in life, the so-

lution to a problem might have advantages, disadvan-

tages, and working elements with a not so suitable

approach. In the case of storytelling, content genera-

tion and ontologies something similar occurs; that is

why we have entitled this section “The Good, the Bad

and the Ugly”, because as in the movie of the same

name, the good and the ugly end up working together

to achieve a goal, whereas the bad is always around

until someone makes it disappear.

Our own experience during the development of

ontologies has allowed us to get a better perspec-

tive of the advantages and difﬁculties faced by expert

and non-expert domain users and developers. In the

next subsections we will present our opinion about

the good, the bad and the ugly things when working

with ontology-based storytelling. In further sections

we will express our conclusion about the topic and

propose what would be the path to follow in order to

make the “bad” disappear.

3.1 The Good

Storytelling and content authoring have become a

very mature and engaging area of research thanks to

the efforts of different groups and persons who have

contributed along the years to make it the way it is.

This fact itself is a good sign that the ﬁeld is evolving

and fruitful results have been achieved.

One example is the number of upper ontolo-

gies that have been developed for deﬁnition of story

elements, context representation or emotion elic-

itation. Proof of this are the upper ontologies

DOLCE (Gangemi et al., 2002), SUMO (Niles and

Pease, 2001), SOUPA (Chen et al., 2004), or the up-

per ontologies DRAMMAR and the Concept Ontol-

ogy, mentioned Section 2.

An important aspect of ontologies is that they pro-

vide the knowledge base that is needed to give struc-

ture to the stories, while adding the “surprise” factor

by inferring unexpected situations from this previous

knowledge. Therefore, having well-deﬁned ontolo-

gies with enough instances of their classes and rela-

tions provide an almost unlimited universe of stories

that can be elicited in much less time.

It is precisely this power of inference what makes

ontologies so suitable for automatic storytelling. In-

ferring new knowledge from already existent one pro-

vides the computational creativity needed in this ﬁeld.

Moreover, the new techniques for translating ontol-

ogy language into natural language convertontologies

into a powerful resource for content creation and con-

text representation in storytelling.

In addition to the above, the possibility of deﬁn-

ing affective elements in the ontologies gives stories

another dimension, in which the characters and the el-

ements of the world are seen nearer and closer to our

experiences. The elicitation of emotions, the person-

ICAART2013-InternationalConferenceonAgentsandArtificialIntelligence

366

ality of the characters or the mood of the story are im-

portant for storytelling, and the fact that they can be

achieved through ontologies makes them even more

suitable, as it results in more empathetic and engag-

ing stories.

3.2 The Bad

The drawbacks of ontologies are related to the pro-

cess of authoring by non-expert users and the con-

sequences of bad-deﬁned ontologies. In some cases,

users are presented with the task of enlarging ontolo-

gies by adding new concepts and relations that would

be used in new stories. If the user is a domain expert,

then this task does not represent a major challenge.

The problem is when the user is not an expert and

needs to learn and understand the language and tools

to construct the ontology.

That is the case of the“Virtual Storyteller” (Swart-

jes and Theune, 2009) where the user needs to work

directly on the ontology, and if necessary, to update

the ﬁles with the ontology rules written in Prolog. Al-

though the results obtained are of great quality thanks

to the iterative process it follows, it might result over-

whelming for a user with no experience in the area.

According to (Liebig and Noppens, 2005), users

are faced with comprehension problems such as the

proper use of role hierarchies, the inﬂuence of tran-

sitive roles on reasoning, or the effect of domain and

range restrictions. All this entails a signiﬁcant amount

of system failures caused by implicit modelling con-

ﬂicts or misunderstandingof the inference algorithms.

Furthermore, the consequencesof those faulty axioms

usually remains hidden until the system is used or in-

tensively tested.

Another issue is the unpredictability of the sys-

tem (Peinado and Gerv´as, 2006). It means that if the

automatic storyteller infers new situations that are not

logic or have no coherence with the narrative struc-

ture, then the research effort can claim little merit.

That is why exhaustive tests should be performed to

measure the originality of the resultant stories and im-

prove the representation of the ﬁctional world.

3.3 The Ugly

As “ugly” we have categorised all those procedures

and methodologies that are tiresome for the user to ac-

complish. For instance, the creation of new instances

or individuals for the ontology classes, or even worse,

the deﬁnition of new classes and relations for a story

that cannot be inferred from the actual knowledge can

be time consuming and bothersome. Specially when

newlogic rules need to be formulated in order to come

up with the correct meaning of a story.

Evidence of this is the work of (M¨uller, 2011),

which despite producingveryreliable results, requires

extensive authoring in order to deﬁne new instances

of story and context elements. Nevertheless, once all

the instances are created, the events can be generated

very straightforward.

A second aspect that can be considered as incon-

venient is the lack of universality in certain concepts.

An example is the representation of the plot struc-

ture. There are a number of theories like the ones

of Vladimir Propp, Lakoff or Barthes (Peinado et al.,

2004), that propose different type of grammars and

representations. Therefore, if an ontology does not

cover a speciﬁc plot structure, then the whole pro-

cess of ontology authoring has to be realised, leading

to the problems faced by non-expert users developing

ontologies.

4 GOOD AND UGLY TOGETHER

- USE CASE

In this section we will give an overview of the de-

velopment and evaluation of ontologies for our Con-

text Representation framework. For more details the

reader can refer to (Arellano et al., 2011).

The objective of our work was to describe the

context (inner context, based on the Belief-Desire-

Intention (BDI) theory plus personality traits; and

outer context, which is the world and its entities) of

virtual characters. To that end, we proposed a se-

mantic framework where we authored two ontologies.

The personalityEmotion ontology considered all the

concepts that deﬁne a character from a psychological-

affective point of view —goals, preferences, social

admiration with other agents, and personality (inner

world). The event ontology described the environ-

ment that surrounds the character (outer world) and

the occurring events in terms of four w-questions:

what happens, where it happens, who are the actors,

and when it happens (Figure 1).

Figure 1: Context Representation with Ontologies.

The difference between our approach and the ones

in the Section 2 is that we combined in one system

both the inﬂuence of the outer and inner world of the

character to simulate its context. Furthermore, we can

OntologiesforAuthoring,orAuthoringOntologies?

367

use these relationships to trigger new emotions in the

character.

To prove that our ontologies and the contextual

module we implemented have the potential to feasible

reproduce any given context and elicit corresponding

emotions, we carried on a perceptual experiment. For

that purpose, we chose three scenes of two movies

—L

eon (1994, directed by Luc Besson) and Downfall

(2004, directed by Oliver Hirschbiegel), which were

simulated in our framework. The goal was to repro-

duce the same context and content manifested in those

scenes, and to achieve the same emotional output as

the one experienced by the real actors. The script ex-

cerpts of both movies (with the scenes emotional out-

put) were taken from (Schapp, 2009).

From both movies scenes, we extracted the events,

character’s preferences, descriptions of the locations

and admiration among characters, in order to intro-

duce them in our Context Representation module.

The correspondence between the resultant emotions

and the emotions manifested by the actors in the

scenes was a hint that our framework indeed repro-

duced the movies context and the right emotional out-

put. The emotions for each scene of L

eon were fear,

pity and sadness, respectively; and for the Downfall

were fear, discourage and disappointment.

The same approach for evaluating our framework

was taken using a a scene extracted from the Robert

Aldrich’s ﬁlm What Ever Happened to Baby Jane?

From the speciﬁc scene that presented the two sisters

Blanche and Jane, we deﬁned 5 events: (1) Blanche

is hungry, (2) Jane enters Blanche’s bedroom with

a closed tray, (3) Blanche is alone with Jane in the

house, (4) Blanche does not believe that Jane is capa-

ble of putting a rat in the food, (5) Blanche opens the

tray and sees the rat, (6) Jane hears Blanche opening

the tray. These events were introduced in our frame-

work, generating a set of emotions in the main char-

acter of the scene (Blanche), which corresponded to

the emotional states of the real actress. For instance,

in the event (4) the emotion with greatest intensity

was gratiﬁcation (intensity = 0.64), indicating that

Blanche does not think that a not satisfactory event

will occur, but she is not sure. This emotion can be

manifested by a slightly smiling face, which is the fa-

cial expression of Blanche in this exact part of the

movie.

These results demonstrated that it is not only pos-

sible the use of ontologies to create and simulate sto-

ries. Moreover, they can be used to establish the re-

lationships between both the outer and inner world of

the characters, becoming a tool to extract and manip-

ulate the affective states of those characters, and their

inﬂuence in the development of the story.

The ugly aspects of our framework corresponded

to the authoring of new instances to create new sto-

ries. Although we tried to make the interface as intu-

itive and straightforward as possible, it was still time

consuming. However, automatic techniques, like con-

tent extraction or data mining, could be applied to re-

duce the process to the selection and reuse of existing

instances, instead of creating new ones.

5 DISCUSSION AND

CONCLUSIONS

Ontologies constitute a powerful tool to represent

knowledge and thus can be applied to a large number

of ﬁelds. Nevertheless, this power comes with great

responsibilities, which forces an accurate deﬁnition

of concepts, relations and rules in order to achieve

logical results. For its part, storytelling, content au-

thoring and their automation is a very ambitious goal

that is being approached by leaps and bounds thanks

to a general iterative process, carried on by the re-

searchers.

Nevertheless, there are still too many concepts to

be described, too many theories to be implemented,

too many rules and cases to deﬁne in order to achieve

better inference. All of this would permit the genera-

tion of stories where the user has little to do; or on the

contrary, where the author can have a lot of creative

inﬂuence but little effort.

Ontology-based storytelling is an area that still has

much to offer, but it will be possible only by the join

effort of the professionals that work on it. Ontolo-

gies represent knowledge, and the larger the knowl-

edge the better the results that can be obtained. In

the same direction, ontologies which are compati-

ble among each other, with different levels of reﬁne-

ment regarding concepts deﬁnitions, and that contain

a number of instances for different story domains,

would allow the production of different and numer-

ous stories without the need of low-level authoring.

Perhaps the way to go would include the addition

of other technologies or ﬁelds that can help develop-

ers and authors with little, or no experience to lessen

the burden of creating new ontologies from scratch,

or enlarging them with new data. For instance, com-

bining ontologies with web semantic and data mining

can improve the authoring process, allowing the ex-

traction of information and creation of new concepts

in a more automatic way.

In a different note, there are some areas that can

beneﬁt from the use of ontology-based storytelling

like computer graphics (i.e. video games) or HCI,

where simple stories are required to perform differ-

ICAART2013-InternationalConferenceonAgentsandArtificialIntelligence

368

ent experiments. For instance, a generic evaluation of

user perception usually demands having a context or

a story the user can work with. Thus, having a tool

for the authoring of simple plots within a specialised

domain can enrich and facilitate the whole procedure.

In conclusion, a good alliance between ontologies

and storytelling has been already created, and it is our

task as researchers to enrich and improve it to make it

better and applicable to a larger number of ﬁelds.

REFERENCES

Arellano, D., Varona, J., and Perales, F. J. (2011). Why do i

feel like this?: The importance of context representa-

tion for emotion elicitation. International Journal of

Synthetic Emotions (IJSE), 2(2):28–47.

Benta, K.-I., Rar˘au, A., and Cremene, M. (2007). Ontology

based affective context representation. In Proceedings

of EATIS ’07, pages 46:1–46:9.

Chen, H., Perich, F., Finin, T., and Joshi, A. (2004). Soupa:

Standard ontology for ubiquitous and pervasive appli-

cations. In International Conference on Mobile and

Ubiquitous Systems: Networking and Services, pages

258–267.

Chen, W., Hayashi, Y., Jin, L., Ikeda, M., and Mizoguchi,

R. (1998). An ontology-based intelligent authoring

tool. In Proc. of the Sixth International Conference on

Computers in Education, pages 41–49.

Damiano, R., Lombardo, V., and Pizzo, A. (2005). For-

mal encoding of drama ontology. In Proceedings of

ICVS’05, pages 95–104.

Dimitrova, V., Denaux, R., Hart, G., Dolbear, C., Holt, I.,

and Cohn, A. (2008). Involving domain experts in au-

thoring owl ontologies. In Proceedings of ISWC ’08,

pages 1–16.

Gangemi, A., Guarino, N., Masolo, C., Oltramari, A., and

Schneider, L. (2002). Sweetening ontologies with

dolce. In Proceedings of EKAW ’02, pages 166–181.

Garc´ıa-Rojas, A., Vexo, F., Thalmann, D., Raouzaiou,

A., Karpouzis, K., Kollias, S., Moccozet, L., and

Magnenat-Thalmann, N. (2006). Emotional face ex-

pression proﬁles supported by virtual human ontol-

ogy. Computer Animation and Virtual Worlds, 17(3-

4):259–269.

Heckmann, D., Schwartz, T., Brandherm, B., Schmitz, M.,

and von WilamowitzMoellendorff, M. (2005). Gumo:

The general user model ontology. User Modeling

2005, pages 428–432.

Konstantopoulos, S., Karkaletsis, V., and Bilidas, D. (2009).

An intelligent authoring environment for abstract se-

mantic representations of cultural object descriptions.

In Proceedings of LaTeCH-SHELT&R ’09, pages 10–

17.

Liebig, T. and Noppens, O. (2005). Ontotrack: A seman-

tic approach for ontology authoring. Web Semantics:

Science, Services and Agents on the World Wide Web,

3(2-3):116–131.

L´opez, J. M., Gil, R., Garc´ıa, R., Cearreta, I., and Garay, N.

(2008). Towards an ontology for describing emotions.

In Proceedings of WSKS ’08, pages 96–104.

Mizoguchi, R. and Ikeda, M. (1996). Towards ontology en-

gineering. Technical report, Osaka University, Japan.

M¨uller, V. M. (2011). An open source architecture for

building interactive dramas. In SBC - Proceedings of

SBGames 2011.

Nakasone, A. and Ishizuka, M. (2006). Storytelling ontol-

ogy model using rst. In Proceedings of IAT ’06, pages

163–169.

National Storytelling Network (n

d). What is story-

telling? Retrieved from www.storynet.org/resources/

whatisstorytelling.html.

Niles, I. and Pease, A. (2001). Towards a standard upper

ontology. In Proceedings of FOIS ’01, pages 2–9.

Obrenovic, Z., Garay, N., Lpez, J., Fajardo, I., and Cear-

reta, I. (2005). An ontology for description of emo-

tional cues. In Tao, J., Tan, T., and Picard, R., edi-

tors, Affective Computing and Intelligent Interaction,

volume 3784 of Lecture Notes in Computer Science,

pages 505–512.

Ortony, A., Clore, G., and Collins, A. (1988). The Cognitive

Structure of Emotions. Cambridge University Press.

Peinado, F. and Gerv´as, P. (2006). Evaluation of automatic

generation of basic stories. New Generation Comput-

ing, 24(3):289–302.

Peinado, F., Gerv´as, P., and D´ıaz-Agudo, B. (2004). A de-

scription logic ontology for fairy tale generation. In

Forth Int. Conf. on Language Resources and Evalua-

tion: Workshop on Language Resources for Linguistic

Creativity, pages 56–61.

Schapp, R. (2009). Enhancing video game characters with

emotion, mood and personality. Master’s thesis, Delft

University of Technology.

Swartjes, I. (2010). Whose story is it anyway? How improv

informs agency and authorship of emergent narrative.

PhD thesis, University of Twente, The Netherlands.

Swartjes, I. and Theune, M. (2009). Iterative authoring

using story generation feedback: debugging or co-

creation? In Second Joint International Conference

on Interactive Digital Storytelling, ICIDS 2009.

Third, A., Williams, S., and Power, R. (2011). Owl to en-

glish: a tool for generating organised easily-navigated

hypertexts from ontologies. In 10th International Se-

mantic Web Conference (ISWC2011), pages 23–27.

Williams, S. (2011). Generating mathematical word prob-

lems. In 2011 Association for the Advancement of Ar-

tiﬁcial Intelligence (AAAI) Fall Symposium on Ques-

tion Generation, pages 61–64.

OntologiesforAuthoring,orAuthoringOntologies?

369