ArticlePDF Available

Emergence in Digital Educational Games: A World of Incidents in a Universe of Rules



Using computer games for educational purposes is a compelling idea that is increasingly adopted by researchers, developers, and educators. Still, digital educational games are at an early stage. A crucial factor that must be increasingly addressed by future research is a personalization of learning and gaming experiences in the rich virtual worlds of computer games. In the present paper we introduce an approach to combine frameworks of psycho-pedagogical adaptation, interactive storytelling, and emergent game design in order to provide the individual learners with tailored learning experiences without corrupting the game's storyline and without requiring massive content production.
Emergence in Digital Educational Games: A World of Incidents in a Universe of Rules
Michael D. Kickmeier-Rust, Dietrich Albert
Department of Psychology, University of Graz, Austria
Abstract: Using computer games for educational purposes is a compelling idea that is increasingly
adopted by researchers, developers, and educators. Still, digital educational games are at an early
stage. A crucial factor that must be increasingly addressed by future research is a personalization of
learning and gaming experiences in the rich virtual worlds of computer games. In the present paper we
introduce an approach to combine frameworks of psycho-pedagogical adaptation, interactive
storytelling, and emergent game design in order to provide the individual learners with tailored learning
experiences without corrupting the game’s storyline and without requiring massive content production.
Keywords: Competitive educational games, adaptation, personalization, interactive storytelling,
emergent game design
1. Introduction
Computer games are an outstanding and incredibly successful part of the present entertainment
landscape. The compelling technology of leading edge computer games, at least in our opinion, must
be considered for teaching and learning also. With the increasing time people of all age groups spend
on playing computer games, the idea of utilizing the games’ motivational and educational potential
becomes more and more convincing and fascinating. Still, today’s computer games not only have a
tremendous motivational potential, computer games enable realizing elementary and essential
pedagogical and didactical principles in a very natural way. Computer games, for instance, provide an
emotionally and semantically appealing and meaningful context for learning, rich and immersive
possibilities for visualizing contents, or the possibility for self-directed, active learning. In short,
computer games do have the potential to make knowledge attractive, important, and meaningful.
For several reasons, the vision that digital educational games (DEGs) become a serious part of
educational technology did not come true yet; from today’s perspective, the realization of this vision is
still in its infancy (Oblinger 2006). This is particularly true if educational games for older children and
adolescents are concerned or when considering games related to school curricula. Most existing
DEGs are rather small and often simple games, focusing on a limited set of competencies (e.g., basic
algebra) or addressing specific skills (e.g., job application trainings). They generally do not related to
school curricula or do not attempt to enable learning about school-related subject matter. More
importantly, existing games do not provide sound assessment methods and generally there is an
imbalance between learning and gaming. Finally, while game intelligence is well developed,
educational games do not include adaptation to the learner in terms of knowledge, learning progress,
motivation, or individual preferences. Thus, they cannot compete with their commercial counterparts
and they cannot utilize the full potential of immersive digital games with respect to learning efficacy
and learning experience. In conclusion, a key aspect of the success of an educational game (i.e.,
effective learning and fun) is an intelligent adaptation to the individual learner.
1.1 Around an Inspiring Virtual Learning World in Eighty Days
The psycho-pedagogical personalization in DEGs is in the focus of the European research project
80Days ( Inspired by Jules Verne’s novel “Around the world in eighty days”, the
project aims at developing psycho-pedagogical and technological foundations for intelligent
adaptation. Basically, the project’s endeavours include melding curriculum-related subject matter with
the fun and excitement of an attractive and compelling computer game. In this context, the intrinsic
motivational potential of computer games is the key to learning success in the sense of voluntary and
maybe hidden learning activities.
In the focus of research and development is an intelligent technology that allows an adaptation to
individual learners, their prior knowledge, abilities, preferences, and learning progress, even more, a
technology that allows a so important but so fragile dynamic balance between challenge and ability.
Figure 1: The three act story model and its translation to a sequence of game elements
Figure 2: A formal representation of restrictions in the sequencing of story elements
This task is not trivial; it requires not only the adaptive mechanisms described earlier, it requires a
formal and computable story model. 80Days relies on the classical three-act structure of Aristotle
providing an arc model with ‘exposition’, ‘rising action to climax’ and ‘denouement’ (Figure 1). Thus,
we can combine the story and learning by linking competence structures with story plots (Figure 2).
This, in turn, generates game paths, possible and meaningful paths through the game accounting for
story model, learning objectives, and pedagogical interventions (see Kickmeier-Rust, Göbel, & Albert
2008 for details).
The outlined approach, unfortunately, has an important drawback: the cost factor. A comprehensive
adaptation throughout an entire game requires massive content (i.e., game elements) production.
However, cost-effectiveness is a crucial factor for a DEG’s success on the market. We address this
problem by extending the approach of adaptive, educational storytelling with ideas of emergent game
2. Emergence in (Educational) Game Design
In regular games, a sequence of scripted events occurs throughout the game. According to Smith
(2002), however, this bears the downside that the game system has a limited awareness of what is
happening and, more importantly, the game is lifelessly determined by what the designers think is
exciting and fun. Emergent behavior, on the other hand, occurs when more or less simple rules
interact to give rise to behavior that was not specifically intended by the developer of a system.
Emergence refers to the process of deriving new but coherent patterns or behaviors in complex
systems. Emergent phenomena occur due to a non-trivial interaction of system components with each
other and with the user. As Johnnson (2001) pointed out, the collective of such kind of interactions
forms novel, complex, and unexpected results. Emergent game design offers a ‚platformand ‘tools’
for gaming, however, without any further blueprint; this is comparable to improvisational theatre or
giving a kid a box of toy cars. The context is fixed but what happens occurs interactively and
One perspective is that emergent gameplay appears due to excellent and comprehensive simulations.
Rich virtual worlds enable the player to interact with a large degree of freedom and, more importantly,
to interact with game entities that respond in a realistic way. Examples might be SimCity, The Sims, or
the interaction with the people in Grand Theft Auto. The key to emergent gameplay and emergent
narrative is a meaningful and “intelligent” interaction with the game and within the game. The
advantage is that each player receives a very unique and personalized gaming experience, which is
potentially enriching the possibilities for educational adaptation/personalization. On the other hand, to
create such intelligent and complete game world may require a significant amount of resources,
perhaps much more than scripted games need.
There exist several techniques from complex systems, machine learning, and artificial life that
potentially enable emergent behavior in games. According to Sweetser (2006a) some examples are
flocking (simulating group behavior such as a flock of birds), cellular automata (discrete time models
simulating complex systems), neural networks (machine learning techniques inspired by the human
brain), or evolutionary algorithms (optimization techniques using concepts from natural selection and
evolution to evolve solutions to problems). Some of those principles have already been transferred to
real games; for example, Half-Life used flocking to give its monsters more lifelike responses. Another
example might be Blade Runner; but also in this example a pre-defined storyline is only “enriched” or
“altered” by accidental aspects, making the game different at each time. Important work in this area
comes from Sweetser (2006b) who developed and evaluated a technically sound framework for
realizing emergent game design. Several authors claim that emergence is the direction game
development is heading, which includes more flexible, realistic, and interactive worlds Sweetser
2.1 Gameplay versus Narrative
Gameplay and narrative are two fundamental dimensions along each game can be described. The
one determines the what and how, the other determines the why. Although both dimensions occur on
a continuum, specific games are either predominantly gameplay-based (e.g., role playing games,
action adventures, or campaign games) or predominantly narrative-based (e.g., simulation games,
management games, or strategy games). Those dimensions also aroused some debate on which a
game should focus more: The ludologists say that games should be played and not perceived like
interactive movies. The narratologists, instead say, games should follow a red story thread. Both, the
gameplay dimension as well as the narrative dimension can be described on a continuum between
open/emergent and predefined/scripted.
With respect to emergent approaches on the gameplay side, intelligent characters play a crucial role.
The “intelligence” of game characters is a essential factor. Those characters are supposed to behave
flexible, challenging, unpredictable, or cunning (Sweetser, Johnson, Sweetser, & Wiles 2003). An
intelligent agent can be considered autonomous if it relies on its own precepts and not on the
predefined ‘will’ or ‘knowledge’ of the game designer (Russel & Norvig 2003). Being autonomous, in
turn, requires situational awareness. An example for such approach in an existing computer game is
the agents in Half Life. Those characters “look” and “listen” to what is happening in their neighboring
areas (Leonard 2003). Still, the realization is rather simple; pre-defined check scripts are processed. In
psychological terms, existing models perform a top-down approach driven by the
designers/developers intelligence. The next generation of artificial in-game intelligence will rather
purse a bottom-up approach, meaningful responses on changes in the agent’s neighborhood.
2.2 The Educational Ways
However, aforementioned approaches were developed in the context of entertainment games.
Educational computer games cannot simply overtake such ideas since a crucial difference between
the two kinds of games is that educational objectives require the learner to pass through certain
learning situations (in whatever way they are realized). This means that pedagogical implications limit
the degree of freedom and randomness in emergent approaches to game design. It is necessary that
a learner is exposed to certain learning situations in a certain sequence.
These limitations contribute to an interactive dilemma (Peinado, Gómez-Martín, & Gómez-Martín
2006) the designers do not want to (and also must not) lose all control and system-only generated
story plots are likely not very convincing. Thus, a subtle balance is required between a global idea of
the story and emergent aspects; research proposed a dual layer model that separates a narrative
layer and an agent/simulation layer (Peinado, Gómez-Martín, & Gómez-Martín 2006). The story
generation is based on the interaction with the beholder, a story-ontology, and vectors of story
elements and relationships.
In this work we want to present a model for involving emergent game design ideas in educational
contexts. This model considers the learning domain, the learner, and it relies on character-based and
plot-based foundations.
3. Educational Adaptation - Interactive Storytelling - Emergent Game Design
First, a narrative context model must be generated. This model is based on the characteristics of the
hero’s journey (Campbell 1993) and the classical three-act story model. It determines a general red
thread through the game and it defines the intro act and the closing act. In-between, we have a large
number of possible story/game paths (Figure 1). These are associated with educational objectives and
pedagogical implications using the cognitive competence-based knowledge space theory
(Kickmeier-Rust & Albert 2008), which establishes a structure of story/game elements that are
meaningful in terms of education and in terms of story. The cognitive model reflects the psycho-
pedagogical requirements and thus determines the admissible game parameters. Formally, we can
summarize the psycho-pedagogical aspects as an “inner state”, which constitutes n-tuples, which in
turn determine transition probabilities (Figure 2). However, in terms of game development this model is
the anti-thesis of cost-effectiveness since it requires massive content production.
As a consequence, we introduce an abstraction layer. On an ontological basis (extending Kickmeier-
Rust & Albert 2008) we separate game play features, story features, and educational features. The
game progresses through a sequence of generic modules (cells) which are sequenced adaptively and
filled with game play, story, and education in real time and system driven.
The theoretical background is similar to the principles of cellular automata. Many of today’s
approaches to modeling real-world phenomena aim to come up with accurate and error-free models.
Often such modeling occurs in the context of scientific applications and forecasts. In games this
complexity is not necessary. It’s all about providing appealing and realistic visual effects (e.g., smoke
or fire) – not necessarily accurate but rather credible. Forsyth (2002), for example has described
methods with which natural processes (e.g., fluid flow) can be simplified for games using cellular
The game elements are seen as cells of a multi-dimensional grid (Figure 3). Each cell must be in one
of a finite set of admissible states (e.g., in terms of story or in terms of knowledge) and each cell has a
set of update rules. The state of a cell is a function of the states of the neighboring cells and it is
sensitive to the actions of the learner. This results in an ebbing and flowing of incidents and it allows
an emergent development of game play as well as narrative – of course limited by the global red
thread through the game and the educational objectives. In more practical terms this means, if the
learner performs an action (e.g., closing the electric circuit) the probability distribution over the
competence states is altered. In combination with other indicators (e.g., intervals between actions or
the number of re-trials) this determines the properties of the game elements (the cells). In turn, altering
the properties of a cell changes the properties of the neighboring cells, comparable to the propagation
of waves when a stone hits the water surface. To give an example, if the learner fails to establish an
electric circuit, the next learning unit automatically adjusts itself to teach the learner about electric
circuits. The advantage of this approach is that the game only needs the assets for the described
adjustments (maybe a set of re-combinable sentences an avatar could say), it is not necessary to
develop all possible learning units.
Figure 3: Cellular automata – or a stone hitting the water
4. Conclusion
In conclusion, the presented attempt to emergent game design in educational contexts can be seen as
a hybrid model which tries to combine the best of both worlds, the author driven scripting of the global
context (including the educator driven design of learning) as well as the degree of freedom and cost-
effectiveness of emergent approaches to game design. Apart from the educational context, the hybrid
model provides also ideas for designing virtual environments in general.
Emergence is primarily driven by “intelligent” characters and “smart props” (prop is a term for objects
in the game such as tools, weapons, furniture, etc.). The approach of cellular automata enables
changes in the game context (by actions of the player and by micro or macro adaptive assessments)
affecting not only one specific character or prop but, driven by more or less complex rules,
semantically neighboring characters and props.
The research and development introduced in this work is funded by the European Commission under
the seventh framework programme in the ICT research priority, contract number 215918 (80Days,
Albert, D. and Lukas, J. (1999) Knowledge spaces: theories, empirical research, and applications.
Mahwah, NJ: Lawrence Erlbaum Associates.
Bloom, B. (1984) The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as
One-to-One Tutoring. Educational Researcher, 13 (6), 4–16.
Campbell, J. (1993) Hero with a thousand faces. Pymble, Australia: Harper Collins Religious.
De Bra, P. (2008) Adaptive hypermedia, In H. H. Adelsberger, Kinshuk, J. M. Pawlowski, & D.
Sampson, (Eds.), Handbook on Information Technologies for Education and Training (pp. 29-46),
Berlin: Springer.
Forsyth, T. (2002) Cellular Automata for Physical Modelling. In D. Treglia (Ed.), Game Programming
Gems 3 (pp. 200-214). Hingham, MA: Charles River Media, Inc.
Johnnson, S. (2001) Emergence: the Connected Lives of Ants, Brains, Cities and Software. New York,
NY: Scribner.
Kickmeier-Rust, M.D., and Albert, D. (2008) The ELEKTRA ontology model: A learner-centered
approach to resource description. Advances in Web Based Learning – ICWL 2007 (pp. 78-89). Lecture
Notes in Computer Science, 4823/2008. Springer Berlin / Heidelberg.
Kickmeier-Rust, M.D., Albert, D., Hockemeyer, C. and Augustin, T. (2007) Not breaking the narrative:
Individualized Competence Assessment in Educational Games. In Proceedings of the 1
Conference on Games based Learning (ECGBL), October 25-26, 2007, Paisley, Scotland.
Kickmeier-Rust, M.D., Göbel, S., and Albert, D. (2008) 80Days: Melding adaptive educational
technology and adaptive and interactive storytelling in digital educational games. In R. Klamma, N.
Sharda, B. Fernández-Manjòn, H. Kosch, & M. Spaniol (Eds.), Proceedings of the First International
Workshop on Story-Telling and Educational Games (STEG'08) - The power of narration and
imagination in technology enhanced learning, September 18-19, 2008, Maastricht, The Netherlands.
Kickmeier-Rust, M.D., Marte, B., Linek, S., Lalonde, T., and Albert, D. (2008). The effects of
individualized feedback in digital educational games. In T. Conolly & M. Stansfield (Eds.), Proceedings
of the 2nd European Conference on Games Based Learning (ECGBL), October 16-17, 2008,
Barcelona, Spain.
Leonard, T. (2003) Building an AI Sensory System: Examining the Design of Thief: The Dark
Project,Gamasutra, March 7, 2003.
Oblinger, D. 2008 Simulations, games, and learning. ELI White Paper, May 2006. [online]
Peinado, F., Gómez-Martín, P.P., and Gómez-Martín, M.M. (2005) A game architecture for emergent
story-puzzles in a persistent world. International DiGRA Conference, June 16th - 20th, 2005,
Vancouver, Canada.
Pressey, S.L. 1926 A simple apparatus which gives tests and scores - and teaches. School and
Society, 23 (586), 373-376.
Russel, S. and Norvig, P. (2003) Artificial Intelligence: A Modern Approach. NJ: Prentice Hall.
Smith, H. (2002) Systemic Level Design. Presented at the Game Developers Conference, San Jose,
CA, March 21-23.
Sweetser, P. (2006a) Environmental Awareness in Game Agents. In Rabin, S. (Ed.), AI Game
Programming Wisdom 3, Hingham, MA: Charles River Media, Inc.
Sweetser, P. (2006b) An Emergent Approach to Game Design - Development and Play. Ph.D. Thesis.
University of Queensland, Australia.
Sweetser, P., Johnson, D., Sweetser, J., and Wiles, J. (2003) Creating Engaging Artificial Characters
for Games. Proceedings of the Second International Conference on Entertainment Computing.
Pittsburgh, PA: Carnegie Mellon University, pp. 1-8.
Blade Runner (1997). Electronic Arts,
Grad Theft Auto (1997-2009). Rockstar Games.
Half Life (1998 - 2007). Sierra Entertainment / Electronic Arts,
The Sims (2000 – 2007). Electronic Arts,
Sim City (1989 – 2007). Electronic Arts,
Until recently, research on video games has been concerned largely with negative effects of play. Increasingly, however, a range of positive psychological effects of playing video games are being reported. Being able to identify positive impacts is an important step toward leveraging the huge appeal of game playing to aid psychological well-being. This chapter discusses a range of benefits associated with playing two genres of popular off-the-shelf video games as opposed to video games that have been constructed to specifically target education or teach health-related lessons. With the video game phenomenon set to have an ever-increasing impact across society on a global scale, knowledge of games that have the potential to meet positive psychological needs will be critical to leveraging positive outcomes from video game play. Full Text Preview Background Although the goals of concern-focused and intervention-focused research are dissimilar, both share a descriptive research approach, in that the methods and theories they employ evaluate the extent to which video games exert positive, negative, or no influence on specified outcomes under a given set of circumstances. Less well understood and less widely studied are the mediating variables that underlie positive links. These include immersion/presence, interaction/movement, and motivation to play. Continue Reading
Conference Paper
Building on existing work on artificial tutors with human-like capabilities, we describe the EMOTE project approach to harnessing benefits of an artificial embodied tutor in a shared physical space. Embodied in robotic platforms or through virtual agents, EMOTE aims to capture some of the empathic and human elements characterising a traditional teacher. As such, empathy and engagement, abilities key to influencing student learning, are at the core of the EMOTE approach. We present non-verbal and adaptive dialogue challenges for such embodied tutors as a foundation for researchers investigating the potential for empathic tutors that will be accepted by students and teachers.
Conference Paper
Full-text available
Significant work has been devoted to the design of artificial tutors with human capabilities with the aim of helping increase the efficiency achieved with a human instructor. Yet, these systems still lack the personal, empathic and human elements that characterise a traditional teacher and fail to engage and motivate students in the same way a human teacher does. The EU-funded project EMOTE (EMbOdied-perceptive Tutors for Empathy-based learning) has recently started, and will continue until the end of 2015. The project aims to design, develop and evaluate a new generation of virtual and robotic embodied tutors that have perceptive capabilities to engage in empathic interactions with learners in a shared physical space. In this paper we wish to discuss the approach we are taking in the project as well as how the project may contribute to knowledge relevant for the Games-Based Learning community.
Full-text available
A crucial factor for successful digital educational games, particularly for older children and adolescents, is an appropriate balance; balance between learning and gaming and balance between challenge and ability. These factors are important to maintain fun, immersion, flow experience, and motivation – the motivation to play and therefore to learn. Moreover, it is important to realize a gaming experience that can compete with that of commercial, non-educational games. A special challenge in this context arises from the need for pedagogical support during learning -and therefore during gaming. At many staves of the learning ladder, from a psych-pedagogical perspective, support and feedback is necessary in order to ensure successful, effective, and complacent learning. Considering the importance of not destroying immersion with the game, the assessment of the learning progress and psycho-pedagogical feedback must occur in a non-invasive way. This, however, requires an intelligent system that is capable of assessing individual competencies and learning progress by observing and interpreting the learner's behaviour in the learning situations within the game. In ELEKTRA, a project funded by the European Commission and aiming at developing a sound psycho-pedagogical framework for immersive educational games, we developed a formal cognitive framework for the non-invasive assessment and interventions within complex learning situations, that is, micro adaptivity. Attuned to the assessed competencies or lack of competencies, meaningful feedback, for example hints, suggestions, reminders, critical questions, or praise, can be triggered, without destroying the gaming experience. Two questions arise with respect to feedback. First, does feedback, although designed to be non-invasive, on educational issues impair gaming experience? Second, can feedback in gaming situations facilitate the learning progress or does it increase the learner's cognitive load, which was suggested be several researchers. In the context of the ELEKTRA project, we implemented the theoretical framework of micro adaptivity in the game demonstrator. This demonstrator is a state-of-the art 3D adventure game teaching physics in relation to national school curricula for the age group of 12 o 14 years. For evaluation purposes, log files of the gaming sessions were recorded and, in addition, questionnaires and performance tests were presented. In this work, we present results from an evaluation session. The results indicate that (micro) adaptive interventions (i.e., appropriate and meaningful interventions/feedback for an individual learner, his/her knowledge and learning progress) are superior to neutral (i.e., non-individualized but semantically correct interventions) and inappropriate interventions (i.e., non-individualized, unsuited interventions) in terms of learning and gaming measures. In addition, we analysed the relationships between learning progress and socio-emotional variables. The results indicate that adaptive feedback not only facilitates learning but also attitude and immersion.
Full-text available
The present paper introduces the 80Days project, an inter-disciplinary European research project endeavoring after pushing the state-of-the-art in digital educational games. The main objectives of the project are enabling curriculum-related education with competitive computer games, realizing non-invasive and educationally meaningful support of the learner, and combining adaptive tutoring with interactive digital storytelling. 80Days' solution to those challenges is an ontology-based linkage between so-called knowledge spaces and atomic narrative elements. On this basis, an intelligent adaptation of storyline, story pace, and game play to the learning progress and the preferences of the learner can be achieved.
Full-text available
Most existing educational games cannot compete with their non-educational counterparts in terms of visual and narrative quality, gameplay, or adaptability. Amongst the most advanced approaches is ELEKTRA, a European project targeting on producing a 3D adventure game teaching physics. The project developed a scientifically sound framework for intelligent and adaptive tutoring, enabling the game to adapt learning/gaming activities to individual learning progress and pedagogical strategies. A crucial aspect, and a weak spot of present educational games, is the individualized assessment of knowledge. Existing approaches frequently rely on typical quiz-like methods, failing to adapt to individual learners and, most likely, they break the game's narrative, what in turn weakens the "natural" advantages of educational games by compromising immersion and motivation to play and learn. In ELEKTRA, assessment occurs in integrated and individualized game situations within which learners have to accomplish adapted and tailored physics-related tasks, for example to hit a light sensor with a narrow beam of light, created with different optical devices, in order to open a door. ELEKTRA's methodology allows providing individualized game situations on the basis of the same pool of game assets. For example, a high performer will be provided with fewer but more complex situations than an underachiever. The set of possible actions and action sequences is modeled in terms of problem spaces. Problem solution states are determined and linked with a skill structure established by prerequisite relations between skills. An ontology holds both information, enabling a "learning engine" to reason about the learner's skills and increase or decrease their probabilities, approaching the true skill state. On this basis, the skills and therefore the learning progress can be assessed without compromising the learner's immersion with the game and, furthermore, subsequent learning and assessment situations can be adapted to the learners' needs. 1. What do you want to play/learn today? The majority of current approaches to technology-enhanced learning are based on traditional, unexciting 2D user interfaces. This perspective is compounded by the proliferation of immersive recreational computer games. In addition, traditional interfaces for educational applications have distinct weaknesses from the perspectives of learning psychology and didactics. For example, they are not intrinsically motivational and it is difficult to retain a learner's interest, to provide a meaningful context throughout learning episodes, or to activate prior knowledge as a basis for learning. Moreover, it is not always possible to provide real-world problems for practicing new knowledge and a purposeful application of new knowledge is difficult without a meaningful and engaging context. Immersive digital educational games (DEGs) offer a highly promising approach to make learning more engaging, satisfying, inspiring, and probably more effective. Thus, it is not surprising that currently there is significant hype over game-based learning (cf. Kickmeier-Rust et al. 2006).
Full-text available
Adaptive hypermedia makes it possible to author learning material once and generate a personalized learning experience for every user. The information that is presented, the way in which it is presented and the possible ways for the user to navigate through it can all be adapted. This chapter presents the most common adaptive hypermedia methods and techniques and shows examples of how they can be (and are) used in existing adaptive hypermedia systems and in adaptive online educational material.